commit dad17e54415b5c7283299bdec860d14faff28377 Author: ModelHub XC Date: Thu Aug 13 01:21:13 2026 +0800 初始化项目，由ModelHub XC社区提供模型 Model: shisa-ai/ablation-147-a128.dpo.armorm.rp.tl.1.5e6-shisa-v2-llama-3.1-8b Source: Original Platform diff --git a/.gitattributes b/.gitattributes new file mode 100644 index 0000000..23c1e42 --- /dev/null +++ b/.gitattributes @@ -0,0 +1,52 @@ +*.7z filter=lfs diff=lfs merge=lfs -text +*.arrow filter=lfs diff=lfs merge=lfs -text +*.bin filter=lfs diff=lfs merge=lfs -text +*.bin.* filter=lfs diff=lfs merge=lfs -text +*.bz2 filter=lfs diff=lfs merge=lfs -text +*.ftz filter=lfs diff=lfs merge=lfs -text +*.gz filter=lfs diff=lfs merge=lfs -text +*.h5 filter=lfs diff=lfs merge=lfs -text +*.joblib filter=lfs diff=lfs merge=lfs -text +*.lfs.* filter=lfs diff=lfs merge=lfs -text +*.model filter=lfs diff=lfs merge=lfs -text +*.msgpack filter=lfs diff=lfs merge=lfs -text +*.onnx filter=lfs diff=lfs merge=lfs -text +*.ot filter=lfs diff=lfs merge=lfs -text +*.parquet filter=lfs diff=lfs merge=lfs -text +*.pb filter=lfs diff=lfs merge=lfs -text +*.pt filter=lfs diff=lfs merge=lfs -text +*.pth filter=lfs diff=lfs merge=lfs -text +*.rar filter=lfs diff=lfs merge=lfs -text +saved_model/**/* filter=lfs diff=lfs merge=lfs -text +*.tar.* filter=lfs diff=lfs merge=lfs -text +*.tflite filter=lfs diff=lfs merge=lfs -text +*.tgz filter=lfs diff=lfs merge=lfs -text +*.xz filter=lfs diff=lfs merge=lfs -text +*.zip filter=lfs diff=lfs merge=lfs -text +*.zstandard filter=lfs diff=lfs merge=lfs -text +*.tfevents* filter=lfs diff=lfs merge=lfs -text +*.db* filter=lfs diff=lfs merge=lfs -text +*.ark* filter=lfs diff=lfs merge=lfs -text +**/*ckpt*data* filter=lfs diff=lfs merge=lfs -text +**/*ckpt*.meta filter=lfs diff=lfs merge=lfs -text +**/*ckpt*.index filter=lfs diff=lfs merge=lfs -text +*.safetensors filter=lfs diff=lfs merge=lfs -text +*.ckpt filter=lfs diff=lfs merge=lfs -text +*.gguf* filter=lfs diff=lfs merge=lfs -text +*.ggml filter=lfs diff=lfs merge=lfs -text +*.llamafile* filter=lfs diff=lfs merge=lfs -text +*.pt2 filter=lfs diff=lfs merge=lfs -text +*.mlmodel filter=lfs diff=lfs merge=lfs -text +*.npy filter=lfs diff=lfs merge=lfs -text +*.npz filter=lfs diff=lfs merge=lfs -text +*.pickle filter=lfs diff=lfs merge=lfs -text +*.pkl filter=lfs diff=lfs merge=lfs -text +*.tar filter=lfs diff=lfs merge=lfs -text +*.wasm filter=lfs diff=lfs merge=lfs -text +*.zst filter=lfs diff=lfs merge=lfs -text +*tfevents* filter=lfs diff=lfs merge=lfs -text + +checkpoint-1630/tokenizer.json filter=lfs diff=lfs merge=lfs -text +checkpoint-2445/tokenizer.json filter=lfs diff=lfs merge=lfs -text +checkpoint-815/tokenizer.json filter=lfs diff=lfs merge=lfs -text +tokenizer.json filter=lfs diff=lfs merge=lfs -text \ No newline at end of file diff --git a/README.md b/README.md new file mode 100644 index 0000000..7a0c67c --- /dev/null +++ b/README.md @@ -0,0 +1,66 @@ +--- +library_name: transformers +model_name: outputs/ablation-147-a128.dpo.armorm.rp.tl.1.5e6-shisa-v2-llama-3.1-8b +tags: +- generated_from_trainer +licence: license +--- + +# Model Card for outputs/ablation-147-a128.dpo.armorm.rp.tl.1.5e6-shisa-v2-llama-3.1-8b + +This model is a fine-tuned version of [None](https://huggingface.co/None). +It has been trained using [TRL](https://github.com/huggingface/trl). + +## Quick start + +```python +from transformers import pipeline + +question = "If you had a time machine, but could only go to the past or the future once and never return, which would you choose and why?" +generator = pipeline("text-generation", model="None", device="cuda") +output = generator([{"role": "user", "content": question}], max_new_tokens=128, return_full_text=False)[0] +print(output["generated_text"]) +``` + +## Training procedure + +[Visualize in Weights & Biases](https://wandb.ai/augmxnt/shisa-v2/runs/c1eolbv9) + + +This model was trained with DPO, a method introduced in [Direct Preference Optimization: Your Language Model is Secretly a Reward Model](https://huggingface.co/papers/2305.18290). + +### Framework versions + +- TRL: 0.15.1 +- Transformers: 4.50.0 +- Pytorch: 2.6.0 +- Datasets: 3.4.1 +- Tokenizers: 0.21.1 + +## Citations + +Cite DPO as: + +```bibtex +@inproceedings{rafailov2023direct, + title = {{Direct Preference Optimization: Your Language Model is Secretly a Reward Model}}, + author = {Rafael Rafailov and Archit Sharma and Eric Mitchell and Christopher D. Manning and Stefano Ermon and Chelsea Finn}, + year = 2023, + booktitle = {Advances in Neural Information Processing Systems 36: Annual Conference on Neural Information Processing Systems 2023, NeurIPS 2023, New Orleans, LA, USA, December 10 - 16, 2023}, + url = {http://papers.nips.cc/paper_files/paper/2023/hash/a85b405ed65c6477a4fe8302b5e06ce7-Abstract-Conference.html}, + editor = {Alice Oh and Tristan Naumann and Amir Globerson and Kate Saenko and Moritz Hardt and Sergey Levine}, +} +``` + +Cite TRL as: + +```bibtex +@misc{vonwerra2022trl, + title = {{TRL: Transformer Reinforcement Learning}}, + author = {Leandro von Werra and Younes Belkada and Lewis Tunstall and Edward Beeching and Tristan Thrush and Nathan Lambert and Shengyi Huang and Kashif Rasul and Quentin Gallouédec}, + year = 2020, + journal = {GitHub repository}, + publisher = {GitHub}, + howpublished = {\url{https://github.com/huggingface/trl}} +} +``` \ No newline at end of file diff --git a/checkpoint-1630/config.json b/checkpoint-1630/config.json new file mode 100644 index 0000000..ded5a8b --- /dev/null +++ b/checkpoint-1630/config.json @@ -0,0 +1,35 @@ +{ + "architectures": [ + "LlamaForCausalLM" + ], + "attention_bias": false, + "attention_dropout": 0.0, + "bos_token_id": 128000, + "eos_token_id": 128009, + "head_dim": 128, + "hidden_act": "silu", + "hidden_size": 4096, + "initializer_range": 0.02, + "intermediate_size": 14336, + "max_position_embeddings": 131072, + "mlp_bias": false, + "model_type": "llama", + "num_attention_heads": 32, + "num_hidden_layers": 32, + "num_key_value_heads": 8, + "pretraining_tp": 1, + "rms_norm_eps": 1e-05, + "rope_scaling": { + "factor": 8.0, + "high_freq_factor": 4.0, + "low_freq_factor": 1.0, + "original_max_position_embeddings": 8192, + "rope_type": "llama3" + }, + "rope_theta": 500000.0, + "tie_word_embeddings": false, + "torch_dtype": "bfloat16", + "transformers_version": "4.50.0", + "use_cache": false, + "vocab_size": 128256 +} diff --git a/checkpoint-1630/generation_config.json b/checkpoint-1630/generation_config.json new file mode 100644 index 0000000..0acdb5a --- /dev/null +++ b/checkpoint-1630/generation_config.json @@ -0,0 +1,12 @@ +{ + "bos_token_id": 128000, + "do_sample": true, + "eos_token_id": [ + 128001, + 128008, + 128009 + ], + "temperature": 0.6, + "top_p": 0.9, + "transformers_version": "4.50.0" +} diff --git a/checkpoint-1630/global_step1630/bf16_zero_pp_rank_0_mp_rank_00_optim_states.pt b/checkpoint-1630/global_step1630/bf16_zero_pp_rank_0_mp_rank_00_optim_states.pt new file mode 100644 index 0000000..f6446db --- /dev/null +++ b/checkpoint-1630/global_step1630/bf16_zero_pp_rank_0_mp_rank_00_optim_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:1a906774bb40a61cac8013150730935c7dc920ad25b9d0fbd24152bfec7942e1 +size 1514015111 diff --git a/checkpoint-1630/global_step1630/bf16_zero_pp_rank_10_mp_rank_00_optim_states.pt b/checkpoint-1630/global_step1630/bf16_zero_pp_rank_10_mp_rank_00_optim_states.pt new file mode 100644 index 0000000..40eec41 --- /dev/null +++ b/checkpoint-1630/global_step1630/bf16_zero_pp_rank_10_mp_rank_00_optim_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:b04c1b3611b157c7a4b1f75a7627b72328d2ecd0873d0d8efcf2d1ae494a7954 +size 1514016442 diff --git a/checkpoint-1630/global_step1630/bf16_zero_pp_rank_11_mp_rank_00_optim_states.pt b/checkpoint-1630/global_step1630/bf16_zero_pp_rank_11_mp_rank_00_optim_states.pt new file mode 100644 index 0000000..ba81569 --- /dev/null +++ b/checkpoint-1630/global_step1630/bf16_zero_pp_rank_11_mp_rank_00_optim_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:02f72bac2a7d1ba37f7cbb8ad9003a91387dd442d53901600532fb371edd4c72 +size 1514016442 diff --git a/checkpoint-1630/global_step1630/bf16_zero_pp_rank_12_mp_rank_00_optim_states.pt b/checkpoint-1630/global_step1630/bf16_zero_pp_rank_12_mp_rank_00_optim_states.pt new file mode 100644 index 0000000..1e84c2c --- /dev/null +++ b/checkpoint-1630/global_step1630/bf16_zero_pp_rank_12_mp_rank_00_optim_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:88ffd8f66e0bcddd4d41755f774002fcbedfaa6f8ee0bd37b3e550322d0c915d +size 1514016442 diff --git a/checkpoint-1630/global_step1630/bf16_zero_pp_rank_13_mp_rank_00_optim_states.pt b/checkpoint-1630/global_step1630/bf16_zero_pp_rank_13_mp_rank_00_optim_states.pt new file mode 100644 index 0000000..0983c9b --- /dev/null +++ b/checkpoint-1630/global_step1630/bf16_zero_pp_rank_13_mp_rank_00_optim_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:39181a12556e0e0161f1f66c2800a4c03051c4e8c9262af30c935a0028501348 +size 1514016442 diff --git a/checkpoint-1630/global_step1630/bf16_zero_pp_rank_14_mp_rank_00_optim_states.pt b/checkpoint-1630/global_step1630/bf16_zero_pp_rank_14_mp_rank_00_optim_states.pt new file mode 100644 index 0000000..2d2716e --- /dev/null +++ b/checkpoint-1630/global_step1630/bf16_zero_pp_rank_14_mp_rank_00_optim_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:85ea97fbc75c34021dcbc1768fa2e953c0e3373bc82e3191d0ad17e03b9b70c0 +size 1514016442 diff --git a/checkpoint-1630/global_step1630/bf16_zero_pp_rank_15_mp_rank_00_optim_states.pt b/checkpoint-1630/global_step1630/bf16_zero_pp_rank_15_mp_rank_00_optim_states.pt new file mode 100644 index 0000000..bb4d781 --- /dev/null +++ b/checkpoint-1630/global_step1630/bf16_zero_pp_rank_15_mp_rank_00_optim_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:2f930ca4babee7fc9d5d4ff47dd2abd11431986dace56f73edd2bbb63ba1b2b7 +size 1514016442 diff --git a/checkpoint-1630/global_step1630/bf16_zero_pp_rank_16_mp_rank_00_optim_states.pt b/checkpoint-1630/global_step1630/bf16_zero_pp_rank_16_mp_rank_00_optim_states.pt new file mode 100644 index 0000000..42a6f3a --- /dev/null +++ b/checkpoint-1630/global_step1630/bf16_zero_pp_rank_16_mp_rank_00_optim_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:0daed32adeb1ef428c164dc7b12ef590b55e770a619ea5d782d19e56090e87ed +size 1514016442 diff --git a/checkpoint-1630/global_step1630/bf16_zero_pp_rank_17_mp_rank_00_optim_states.pt b/checkpoint-1630/global_step1630/bf16_zero_pp_rank_17_mp_rank_00_optim_states.pt new file mode 100644 index 0000000..1fd8497 --- /dev/null +++ b/checkpoint-1630/global_step1630/bf16_zero_pp_rank_17_mp_rank_00_optim_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:031dfd6bd93368f0da6c8ac799e62a380d0e5140dc95c0bbcaa980f9b5dd44a7 +size 1514016442 diff --git a/checkpoint-1630/global_step1630/bf16_zero_pp_rank_18_mp_rank_00_optim_states.pt b/checkpoint-1630/global_step1630/bf16_zero_pp_rank_18_mp_rank_00_optim_states.pt new file mode 100644 index 0000000..5aec751 --- /dev/null +++ b/checkpoint-1630/global_step1630/bf16_zero_pp_rank_18_mp_rank_00_optim_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:e6371356e108eddafa5f4ef35ae4fed90afac1a24ea2a4bbdc35157b8813d60a +size 1514016442 diff --git a/checkpoint-1630/global_step1630/bf16_zero_pp_rank_19_mp_rank_00_optim_states.pt b/checkpoint-1630/global_step1630/bf16_zero_pp_rank_19_mp_rank_00_optim_states.pt new file mode 100644 index 0000000..77b5131 --- /dev/null +++ b/checkpoint-1630/global_step1630/bf16_zero_pp_rank_19_mp_rank_00_optim_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:ac9b2f3864b4e219cd029197f6b469ecb054a6191070db6ae7d4f510543212af +size 1514016442 diff --git a/checkpoint-1630/global_step1630/bf16_zero_pp_rank_1_mp_rank_00_optim_states.pt b/checkpoint-1630/global_step1630/bf16_zero_pp_rank_1_mp_rank_00_optim_states.pt new file mode 100644 index 0000000..3244ea7 --- /dev/null +++ b/checkpoint-1630/global_step1630/bf16_zero_pp_rank_1_mp_rank_00_optim_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:31077488d5bc5dd7d97aa26d805ddd4f0d3ae2400435583b26441a6a5a3c5f48 +size 1514015111 diff --git a/checkpoint-1630/global_step1630/bf16_zero_pp_rank_20_mp_rank_00_optim_states.pt b/checkpoint-1630/global_step1630/bf16_zero_pp_rank_20_mp_rank_00_optim_states.pt new file mode 100644 index 0000000..b576fd5 --- /dev/null +++ b/checkpoint-1630/global_step1630/bf16_zero_pp_rank_20_mp_rank_00_optim_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:20b02e96d9779e3fa3efe13eaefae10747d3d10cada4b9dafaec240885028dcb +size 1514016442 diff --git a/checkpoint-1630/global_step1630/bf16_zero_pp_rank_21_mp_rank_00_optim_states.pt b/checkpoint-1630/global_step1630/bf16_zero_pp_rank_21_mp_rank_00_optim_states.pt new file mode 100644 index 0000000..d4d0bd0 --- /dev/null +++ b/checkpoint-1630/global_step1630/bf16_zero_pp_rank_21_mp_rank_00_optim_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:f68f0d284628baaadd708be4d2626d2ead68022c2e9387d620bf0dc1bfaee068 +size 1514016442 diff --git a/checkpoint-1630/global_step1630/bf16_zero_pp_rank_22_mp_rank_00_optim_states.pt b/checkpoint-1630/global_step1630/bf16_zero_pp_rank_22_mp_rank_00_optim_states.pt new file mode 100644 index 0000000..f8bbafa --- /dev/null +++ b/checkpoint-1630/global_step1630/bf16_zero_pp_rank_22_mp_rank_00_optim_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:b3b29fd9ac3f494573dd56b209962345c04bdac9708424be865865b72a8f2d37 +size 1514016442 diff --git a/checkpoint-1630/global_step1630/bf16_zero_pp_rank_23_mp_rank_00_optim_states.pt b/checkpoint-1630/global_step1630/bf16_zero_pp_rank_23_mp_rank_00_optim_states.pt new file mode 100644 index 0000000..9f11e1a --- /dev/null +++ b/checkpoint-1630/global_step1630/bf16_zero_pp_rank_23_mp_rank_00_optim_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:444754a88e194503f57bbf6da6c307fb70d820440181736c637f306679c0f09b +size 1514016442 diff --git a/checkpoint-1630/global_step1630/bf16_zero_pp_rank_24_mp_rank_00_optim_states.pt b/checkpoint-1630/global_step1630/bf16_zero_pp_rank_24_mp_rank_00_optim_states.pt new file mode 100644 index 0000000..bc6fd98 --- /dev/null +++ b/checkpoint-1630/global_step1630/bf16_zero_pp_rank_24_mp_rank_00_optim_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:44be8ed65ad1a68a8062da3fa59fc5195dd6d078e09648caf575cf04fb82b3ec +size 1514016442 diff --git a/checkpoint-1630/global_step1630/bf16_zero_pp_rank_25_mp_rank_00_optim_states.pt b/checkpoint-1630/global_step1630/bf16_zero_pp_rank_25_mp_rank_00_optim_states.pt new file mode 100644 index 0000000..dff19e9 --- /dev/null +++ b/checkpoint-1630/global_step1630/bf16_zero_pp_rank_25_mp_rank_00_optim_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:7b0fa2dc304ebab6af18c6b26dcd144da70dbd53dd60a50541508273161b59b1 +size 1514016442 diff --git a/checkpoint-1630/global_step1630/bf16_zero_pp_rank_26_mp_rank_00_optim_states.pt b/checkpoint-1630/global_step1630/bf16_zero_pp_rank_26_mp_rank_00_optim_states.pt new file mode 100644 index 0000000..232a8ab --- /dev/null +++ b/checkpoint-1630/global_step1630/bf16_zero_pp_rank_26_mp_rank_00_optim_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:35e50e6bcfdfb72464433ebfff15a9af66b520610eef60fb1cc1645b6a807442 +size 1514016442 diff --git a/checkpoint-1630/global_step1630/bf16_zero_pp_rank_27_mp_rank_00_optim_states.pt b/checkpoint-1630/global_step1630/bf16_zero_pp_rank_27_mp_rank_00_optim_states.pt new file mode 100644 index 0000000..58a318c --- /dev/null +++ b/checkpoint-1630/global_step1630/bf16_zero_pp_rank_27_mp_rank_00_optim_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:0b6904a514d7b3caea21c0cfa8c2ed41d46b5248e2686aec1872c3837d5db012 +size 1514016442 diff --git a/checkpoint-1630/global_step1630/bf16_zero_pp_rank_28_mp_rank_00_optim_states.pt b/checkpoint-1630/global_step1630/bf16_zero_pp_rank_28_mp_rank_00_optim_states.pt new file mode 100644 index 0000000..fedaf60 --- /dev/null +++ b/checkpoint-1630/global_step1630/bf16_zero_pp_rank_28_mp_rank_00_optim_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:ffe964bd340c96dcddad5ee65e4b12398069764cb82469754011404de372418e +size 1514016442 diff --git a/checkpoint-1630/global_step1630/bf16_zero_pp_rank_29_mp_rank_00_optim_states.pt b/checkpoint-1630/global_step1630/bf16_zero_pp_rank_29_mp_rank_00_optim_states.pt new file mode 100644 index 0000000..e48c509 --- /dev/null +++ b/checkpoint-1630/global_step1630/bf16_zero_pp_rank_29_mp_rank_00_optim_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:f96bcec1657b782ac69dc043c46ba93e8aa23d76c1011b7b49d67aa7e26866d5 +size 1514016442 diff --git a/checkpoint-1630/global_step1630/bf16_zero_pp_rank_2_mp_rank_00_optim_states.pt b/checkpoint-1630/global_step1630/bf16_zero_pp_rank_2_mp_rank_00_optim_states.pt new file mode 100644 index 0000000..0f94eea --- /dev/null +++ b/checkpoint-1630/global_step1630/bf16_zero_pp_rank_2_mp_rank_00_optim_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:de893ed7003d6e832e025b8fb0cbe612f5215a8425ec27971c86ecb1a5ed61f5 +size 1514015111 diff --git a/checkpoint-1630/global_step1630/bf16_zero_pp_rank_30_mp_rank_00_optim_states.pt b/checkpoint-1630/global_step1630/bf16_zero_pp_rank_30_mp_rank_00_optim_states.pt new file mode 100644 index 0000000..c56e474 --- /dev/null +++ b/checkpoint-1630/global_step1630/bf16_zero_pp_rank_30_mp_rank_00_optim_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:e7241240eb1fda0be101959f789913734298c8433d77f0b2056d39d9c9e1ee6a +size 1514016442 diff --git a/checkpoint-1630/global_step1630/bf16_zero_pp_rank_31_mp_rank_00_optim_states.pt b/checkpoint-1630/global_step1630/bf16_zero_pp_rank_31_mp_rank_00_optim_states.pt new file mode 100644 index 0000000..59f090c --- /dev/null +++ b/checkpoint-1630/global_step1630/bf16_zero_pp_rank_31_mp_rank_00_optim_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:81cc56b48958e31411f251defa95a385c0bf2f4d88e4facb25c4c59088e63665 +size 1514016442 diff --git a/checkpoint-1630/global_step1630/bf16_zero_pp_rank_3_mp_rank_00_optim_states.pt b/checkpoint-1630/global_step1630/bf16_zero_pp_rank_3_mp_rank_00_optim_states.pt new file mode 100644 index 0000000..993f789 --- /dev/null +++ b/checkpoint-1630/global_step1630/bf16_zero_pp_rank_3_mp_rank_00_optim_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:9eb0445fda0924836047486f031ce439979e6e3097c31c9bb8d39e41e540f339 +size 1514015111 diff --git a/checkpoint-1630/global_step1630/bf16_zero_pp_rank_4_mp_rank_00_optim_states.pt b/checkpoint-1630/global_step1630/bf16_zero_pp_rank_4_mp_rank_00_optim_states.pt new file mode 100644 index 0000000..a2c94a6 --- /dev/null +++ b/checkpoint-1630/global_step1630/bf16_zero_pp_rank_4_mp_rank_00_optim_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:01c14c5108fb8d9a7d4b915f813ecb2981155292fd5df735530fb0268660eca1 +size 1514015111 diff --git a/checkpoint-1630/global_step1630/bf16_zero_pp_rank_5_mp_rank_00_optim_states.pt b/checkpoint-1630/global_step1630/bf16_zero_pp_rank_5_mp_rank_00_optim_states.pt new file mode 100644 index 0000000..4a91c8e --- /dev/null +++ b/checkpoint-1630/global_step1630/bf16_zero_pp_rank_5_mp_rank_00_optim_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:815d5b003d103728516d933da40dd9746a41f4b2c03dd933f7fa1872a770c8be +size 1514015111 diff --git a/checkpoint-1630/global_step1630/bf16_zero_pp_rank_6_mp_rank_00_optim_states.pt b/checkpoint-1630/global_step1630/bf16_zero_pp_rank_6_mp_rank_00_optim_states.pt new file mode 100644 index 0000000..ff297d3 --- /dev/null +++ b/checkpoint-1630/global_step1630/bf16_zero_pp_rank_6_mp_rank_00_optim_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:91c4b2ae2cec76367a7ceac3ad15b3e75375040f7385997b3956a875da9b6e98 +size 1514015111 diff --git a/checkpoint-1630/global_step1630/bf16_zero_pp_rank_7_mp_rank_00_optim_states.pt b/checkpoint-1630/global_step1630/bf16_zero_pp_rank_7_mp_rank_00_optim_states.pt new file mode 100644 index 0000000..f2413d4 --- /dev/null +++ b/checkpoint-1630/global_step1630/bf16_zero_pp_rank_7_mp_rank_00_optim_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:ddf0cde0c2ff34a7ff3b9b8c54295edf0d60a4e332c8f5a79321ec470ae2405e +size 1514015111 diff --git a/checkpoint-1630/global_step1630/bf16_zero_pp_rank_8_mp_rank_00_optim_states.pt b/checkpoint-1630/global_step1630/bf16_zero_pp_rank_8_mp_rank_00_optim_states.pt new file mode 100644 index 0000000..ec22649 --- /dev/null +++ b/checkpoint-1630/global_step1630/bf16_zero_pp_rank_8_mp_rank_00_optim_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:056aa9e88b2cc605e066e487d7b322569f6d735f27e7ff7004c6e7b944b84fe3 +size 1514015111 diff --git a/checkpoint-1630/global_step1630/bf16_zero_pp_rank_9_mp_rank_00_optim_states.pt b/checkpoint-1630/global_step1630/bf16_zero_pp_rank_9_mp_rank_00_optim_states.pt new file mode 100644 index 0000000..e5f9a1e --- /dev/null +++ b/checkpoint-1630/global_step1630/bf16_zero_pp_rank_9_mp_rank_00_optim_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:f63c716b15d2548b8a7f709c8c136759fd5d4eea6f672d564961f08a2b709dcf +size 1514015111 diff --git a/checkpoint-1630/global_step1630/zero_pp_rank_0_mp_rank_00_model_states.pt b/checkpoint-1630/global_step1630/zero_pp_rank_0_mp_rank_00_model_states.pt new file mode 100644 index 0000000..db997e4 --- /dev/null +++ b/checkpoint-1630/global_step1630/zero_pp_rank_0_mp_rank_00_model_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:fa532a7c8189bc930aebd05eba451a267458553a6f3af3bb3b1e98d8523d76be +size 150245 diff --git a/checkpoint-1630/global_step1630/zero_pp_rank_10_mp_rank_00_model_states.pt b/checkpoint-1630/global_step1630/zero_pp_rank_10_mp_rank_00_model_states.pt new file mode 100644 index 0000000..aa9c61b --- /dev/null +++ b/checkpoint-1630/global_step1630/zero_pp_rank_10_mp_rank_00_model_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:f515b12cf75954828dd92c701dcb04425ca7da970eaf6a88653baf1d62011269 +size 150540 diff --git a/checkpoint-1630/global_step1630/zero_pp_rank_11_mp_rank_00_model_states.pt b/checkpoint-1630/global_step1630/zero_pp_rank_11_mp_rank_00_model_states.pt new file mode 100644 index 0000000..7d5ca98 --- /dev/null +++ b/checkpoint-1630/global_step1630/zero_pp_rank_11_mp_rank_00_model_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:84c969c26f5eecbc4c2484d80c3ea790de2bb535e0d02d9935d79c357fbb8f86 +size 150540 diff --git a/checkpoint-1630/global_step1630/zero_pp_rank_12_mp_rank_00_model_states.pt b/checkpoint-1630/global_step1630/zero_pp_rank_12_mp_rank_00_model_states.pt new file mode 100644 index 0000000..c8b5fd8 --- /dev/null +++ b/checkpoint-1630/global_step1630/zero_pp_rank_12_mp_rank_00_model_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:bafcd493f85462e19b9224677c7514bb3b94f423861f053cc4c3fa2be59a53f8 +size 150540 diff --git a/checkpoint-1630/global_step1630/zero_pp_rank_13_mp_rank_00_model_states.pt b/checkpoint-1630/global_step1630/zero_pp_rank_13_mp_rank_00_model_states.pt new file mode 100644 index 0000000..e5d4fff --- /dev/null +++ b/checkpoint-1630/global_step1630/zero_pp_rank_13_mp_rank_00_model_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:6c8717e3c8eaee90ce1e24d39bd0d4d19bc5423035c1a8041bb31c3c2f460158 +size 150540 diff --git a/checkpoint-1630/global_step1630/zero_pp_rank_14_mp_rank_00_model_states.pt b/checkpoint-1630/global_step1630/zero_pp_rank_14_mp_rank_00_model_states.pt new file mode 100644 index 0000000..9740439 --- /dev/null +++ b/checkpoint-1630/global_step1630/zero_pp_rank_14_mp_rank_00_model_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:94de6015b071a4de07754007a639c92a62d94046a025359225e936dcb362950c +size 150540 diff --git a/checkpoint-1630/global_step1630/zero_pp_rank_15_mp_rank_00_model_states.pt b/checkpoint-1630/global_step1630/zero_pp_rank_15_mp_rank_00_model_states.pt new file mode 100644 index 0000000..76bfa5b --- /dev/null +++ b/checkpoint-1630/global_step1630/zero_pp_rank_15_mp_rank_00_model_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:bc643726424c61753779b64ec836f54d53bc3e54c14a897e082cbd4cc1ac39cc +size 150540 diff --git a/checkpoint-1630/global_step1630/zero_pp_rank_16_mp_rank_00_model_states.pt b/checkpoint-1630/global_step1630/zero_pp_rank_16_mp_rank_00_model_states.pt new file mode 100644 index 0000000..3fe6b8d --- /dev/null +++ b/checkpoint-1630/global_step1630/zero_pp_rank_16_mp_rank_00_model_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:45d822837deac3b5b69558d5eb8bf16f733c248644913ba5d1387dfcdbb246e2 +size 150540 diff --git a/checkpoint-1630/global_step1630/zero_pp_rank_17_mp_rank_00_model_states.pt b/checkpoint-1630/global_step1630/zero_pp_rank_17_mp_rank_00_model_states.pt new file mode 100644 index 0000000..141ff82 --- /dev/null +++ b/checkpoint-1630/global_step1630/zero_pp_rank_17_mp_rank_00_model_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:d83791ff8b0cc02d65fd9e6c0250fee80c813bfa93bc74440d0d9546b742e731 +size 150540 diff --git a/checkpoint-1630/global_step1630/zero_pp_rank_18_mp_rank_00_model_states.pt b/checkpoint-1630/global_step1630/zero_pp_rank_18_mp_rank_00_model_states.pt new file mode 100644 index 0000000..67fbf8e --- /dev/null +++ b/checkpoint-1630/global_step1630/zero_pp_rank_18_mp_rank_00_model_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:c1ad648ff2f6632d194310cd590612a34dcee5ae59afc2705330de48f48818fa +size 150540 diff --git a/checkpoint-1630/global_step1630/zero_pp_rank_19_mp_rank_00_model_states.pt b/checkpoint-1630/global_step1630/zero_pp_rank_19_mp_rank_00_model_states.pt new file mode 100644 index 0000000..c8ebd46 --- /dev/null +++ b/checkpoint-1630/global_step1630/zero_pp_rank_19_mp_rank_00_model_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:c5afb79ef6850bc18a5d736840c68a776e930087b38321ffc62455c3ca324957 +size 150540 diff --git a/checkpoint-1630/global_step1630/zero_pp_rank_1_mp_rank_00_model_states.pt b/checkpoint-1630/global_step1630/zero_pp_rank_1_mp_rank_00_model_states.pt new file mode 100644 index 0000000..45d83f1 --- /dev/null +++ b/checkpoint-1630/global_step1630/zero_pp_rank_1_mp_rank_00_model_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:31aa88b36939fbb7eb1353cfb81e6285e052452a3be3f5322a5898606ab6f099 +size 150245 diff --git a/checkpoint-1630/global_step1630/zero_pp_rank_20_mp_rank_00_model_states.pt b/checkpoint-1630/global_step1630/zero_pp_rank_20_mp_rank_00_model_states.pt new file mode 100644 index 0000000..a46f0f7 --- /dev/null +++ b/checkpoint-1630/global_step1630/zero_pp_rank_20_mp_rank_00_model_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:f78a8d21611157ba84d8e933650e4314450e598ef9390e2ef140ffbaebcbb413 +size 150540 diff --git a/checkpoint-1630/global_step1630/zero_pp_rank_21_mp_rank_00_model_states.pt b/checkpoint-1630/global_step1630/zero_pp_rank_21_mp_rank_00_model_states.pt new file mode 100644 index 0000000..7784a83 --- /dev/null +++ b/checkpoint-1630/global_step1630/zero_pp_rank_21_mp_rank_00_model_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:dbd5d9d4ba33afd5b8681d99f7b557498a6bc34b1cf469cd6fdfad70ef3ba462 +size 150540 diff --git a/checkpoint-1630/global_step1630/zero_pp_rank_22_mp_rank_00_model_states.pt b/checkpoint-1630/global_step1630/zero_pp_rank_22_mp_rank_00_model_states.pt new file mode 100644 index 0000000..5bf2872 --- /dev/null +++ b/checkpoint-1630/global_step1630/zero_pp_rank_22_mp_rank_00_model_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:ce1032af673079a83985ef1e3221449808e2e402a401e80878f19ebdefc1fb91 +size 150540 diff --git a/checkpoint-1630/global_step1630/zero_pp_rank_23_mp_rank_00_model_states.pt b/checkpoint-1630/global_step1630/zero_pp_rank_23_mp_rank_00_model_states.pt new file mode 100644 index 0000000..458fd3e --- /dev/null +++ b/checkpoint-1630/global_step1630/zero_pp_rank_23_mp_rank_00_model_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:93c819f2b87b13c7d0365314a8793830cbc3537fce028f88a23171c0a8d139d3 +size 150540 diff --git a/checkpoint-1630/global_step1630/zero_pp_rank_24_mp_rank_00_model_states.pt b/checkpoint-1630/global_step1630/zero_pp_rank_24_mp_rank_00_model_states.pt new file mode 100644 index 0000000..fbce9f2 --- /dev/null +++ b/checkpoint-1630/global_step1630/zero_pp_rank_24_mp_rank_00_model_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:c15e8195851401da7054ba176dee771b7dace27c96d02c3563ed4787041a24c0 +size 150540 diff --git a/checkpoint-1630/global_step1630/zero_pp_rank_25_mp_rank_00_model_states.pt b/checkpoint-1630/global_step1630/zero_pp_rank_25_mp_rank_00_model_states.pt new file mode 100644 index 0000000..ecf2083 --- /dev/null +++ b/checkpoint-1630/global_step1630/zero_pp_rank_25_mp_rank_00_model_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:d4812cff1c52c194835de7c60b4db6ed5af2c99a3eebf7e1e36a50487b485500 +size 150540 diff --git a/checkpoint-1630/global_step1630/zero_pp_rank_26_mp_rank_00_model_states.pt b/checkpoint-1630/global_step1630/zero_pp_rank_26_mp_rank_00_model_states.pt new file mode 100644 index 0000000..89980fc --- /dev/null +++ b/checkpoint-1630/global_step1630/zero_pp_rank_26_mp_rank_00_model_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:1bbdcfe71f050b662d5f093662a16b6fbc07cb1581a11a143e442e34d69d6773 +size 150540 diff --git a/checkpoint-1630/global_step1630/zero_pp_rank_27_mp_rank_00_model_states.pt b/checkpoint-1630/global_step1630/zero_pp_rank_27_mp_rank_00_model_states.pt new file mode 100644 index 0000000..4fd1656 --- /dev/null +++ b/checkpoint-1630/global_step1630/zero_pp_rank_27_mp_rank_00_model_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:3fc965279bf3f1fc656157fe0f3a6ae7d7eff93f4ae7191e81aeb950509b5028 +size 150540 diff --git a/checkpoint-1630/global_step1630/zero_pp_rank_28_mp_rank_00_model_states.pt b/checkpoint-1630/global_step1630/zero_pp_rank_28_mp_rank_00_model_states.pt new file mode 100644 index 0000000..6fbaaa7 --- /dev/null +++ b/checkpoint-1630/global_step1630/zero_pp_rank_28_mp_rank_00_model_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:11fccd0874abe40f5a9c85f018dd6e8fe275cf1ce335e18aa18d328dfcdafe6c +size 150540 diff --git a/checkpoint-1630/global_step1630/zero_pp_rank_29_mp_rank_00_model_states.pt b/checkpoint-1630/global_step1630/zero_pp_rank_29_mp_rank_00_model_states.pt new file mode 100644 index 0000000..d616878 --- /dev/null +++ b/checkpoint-1630/global_step1630/zero_pp_rank_29_mp_rank_00_model_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:d078fc60c854c4f513f1bae8c532e41f3ee95fa8a1627dc892bb71f7eed7b07e +size 150540 diff --git a/checkpoint-1630/global_step1630/zero_pp_rank_2_mp_rank_00_model_states.pt b/checkpoint-1630/global_step1630/zero_pp_rank_2_mp_rank_00_model_states.pt new file mode 100644 index 0000000..bdd2801 --- /dev/null +++ b/checkpoint-1630/global_step1630/zero_pp_rank_2_mp_rank_00_model_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:4c7765bdf55f1a5a1c6e3f10b1fab6afd8bfe9882aa502f584ecb39326329cd5 +size 150245 diff --git a/checkpoint-1630/global_step1630/zero_pp_rank_30_mp_rank_00_model_states.pt b/checkpoint-1630/global_step1630/zero_pp_rank_30_mp_rank_00_model_states.pt new file mode 100644 index 0000000..86eb18a --- /dev/null +++ b/checkpoint-1630/global_step1630/zero_pp_rank_30_mp_rank_00_model_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:cd4fc4625ba7ba5d268581dcd1621b8fad50e98dc1cb0a51ee2d310cf38180f9 +size 150540 diff --git a/checkpoint-1630/global_step1630/zero_pp_rank_31_mp_rank_00_model_states.pt b/checkpoint-1630/global_step1630/zero_pp_rank_31_mp_rank_00_model_states.pt new file mode 100644 index 0000000..2550b22 --- /dev/null +++ b/checkpoint-1630/global_step1630/zero_pp_rank_31_mp_rank_00_model_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:7e9dac2bfd65cf28544ef2c368df6588cb38a0e2aea7ca7309e34803e309a869 +size 150540 diff --git a/checkpoint-1630/global_step1630/zero_pp_rank_3_mp_rank_00_model_states.pt b/checkpoint-1630/global_step1630/zero_pp_rank_3_mp_rank_00_model_states.pt new file mode 100644 index 0000000..5ddc7c4 --- /dev/null +++ b/checkpoint-1630/global_step1630/zero_pp_rank_3_mp_rank_00_model_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:cecb7ce13aae08d3535fed9bc97e11d0a139c2209edf77f3efb1dce2a347b76f +size 150245 diff --git a/checkpoint-1630/global_step1630/zero_pp_rank_4_mp_rank_00_model_states.pt b/checkpoint-1630/global_step1630/zero_pp_rank_4_mp_rank_00_model_states.pt new file mode 100644 index 0000000..d4c6a9b --- /dev/null +++ b/checkpoint-1630/global_step1630/zero_pp_rank_4_mp_rank_00_model_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:d4d1df2d1673f71716cd32eac2fdfeeb293b87695c982654cd52881de39b8609 +size 150245 diff --git a/checkpoint-1630/global_step1630/zero_pp_rank_5_mp_rank_00_model_states.pt b/checkpoint-1630/global_step1630/zero_pp_rank_5_mp_rank_00_model_states.pt new file mode 100644 index 0000000..a04e871 --- /dev/null +++ b/checkpoint-1630/global_step1630/zero_pp_rank_5_mp_rank_00_model_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:4097f82f363f57deca37fc2e94f9db9595a0d3fbff7471e6db2e4996247ea34e +size 150245 diff --git a/checkpoint-1630/global_step1630/zero_pp_rank_6_mp_rank_00_model_states.pt b/checkpoint-1630/global_step1630/zero_pp_rank_6_mp_rank_00_model_states.pt new file mode 100644 index 0000000..0a82596 --- /dev/null +++ b/checkpoint-1630/global_step1630/zero_pp_rank_6_mp_rank_00_model_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:8804c594d8aa48eae8a689acf3aa847a2780956548a12706e750c60449febabd +size 150245 diff --git a/checkpoint-1630/global_step1630/zero_pp_rank_7_mp_rank_00_model_states.pt b/checkpoint-1630/global_step1630/zero_pp_rank_7_mp_rank_00_model_states.pt new file mode 100644 index 0000000..98edf43 --- /dev/null +++ b/checkpoint-1630/global_step1630/zero_pp_rank_7_mp_rank_00_model_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:9be01bda51e54d84107f28aad68934ece7bc9243d22f37bca36d70d5a687c306 +size 150245 diff --git a/checkpoint-1630/global_step1630/zero_pp_rank_8_mp_rank_00_model_states.pt b/checkpoint-1630/global_step1630/zero_pp_rank_8_mp_rank_00_model_states.pt new file mode 100644 index 0000000..44dc7c4 --- /dev/null +++ b/checkpoint-1630/global_step1630/zero_pp_rank_8_mp_rank_00_model_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:965dc1bd2b8bc46f0492c744ea8d80fe74acc33c0a82031d484658924a9188c5 +size 150245 diff --git a/checkpoint-1630/global_step1630/zero_pp_rank_9_mp_rank_00_model_states.pt b/checkpoint-1630/global_step1630/zero_pp_rank_9_mp_rank_00_model_states.pt new file mode 100644 index 0000000..aba92fd --- /dev/null +++ b/checkpoint-1630/global_step1630/zero_pp_rank_9_mp_rank_00_model_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:98208579a411498d49a4075e894283e5b951eee51cec29ca8f3c6dd17dac132c +size 150245 diff --git a/checkpoint-1630/latest b/checkpoint-1630/latest new file mode 100644 index 0000000..b876efb --- /dev/null +++ b/checkpoint-1630/latest @@ -0,0 +1 @@ +global_step1630 \ No newline at end of file diff --git a/checkpoint-1630/model-00001-of-00004.safetensors b/checkpoint-1630/model-00001-of-00004.safetensors new file mode 100644 index 0000000..a16ed17 --- /dev/null +++ b/checkpoint-1630/model-00001-of-00004.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:ca043ed69444e849ba3cbbe7152af1d935d23ceb68302324317defe21b8e7ebe +size 4976698672 diff --git a/checkpoint-1630/model-00002-of-00004.safetensors b/checkpoint-1630/model-00002-of-00004.safetensors new file mode 100644 index 0000000..ad97aa1 --- /dev/null +++ b/checkpoint-1630/model-00002-of-00004.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:0ad9ca5256887c8dde9e7cc78d78284598b2766a69cfa3ac3ba0809e291d786b +size 4999802720 diff --git a/checkpoint-1630/model-00003-of-00004.safetensors b/checkpoint-1630/model-00003-of-00004.safetensors new file mode 100644 index 0000000..8913bec --- /dev/null +++ b/checkpoint-1630/model-00003-of-00004.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:db02621d8d374ea2cf2445b8a20cf494e921dfcbe5e6520b711405683b07dc31 +size 4915916176 diff --git a/checkpoint-1630/model-00004-of-00004.safetensors b/checkpoint-1630/model-00004-of-00004.safetensors new file mode 100644 index 0000000..9db61ba --- /dev/null +++ b/checkpoint-1630/model-00004-of-00004.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:c25ff51f1727eb1fe9a66589dadee1a6aa42082eb5f28e37540c8e92b912fac0 +size 1168138808 diff --git a/checkpoint-1630/model.safetensors.index.json b/checkpoint-1630/model.safetensors.index.json new file mode 100644 index 0000000..0fd8120 --- /dev/null +++ b/checkpoint-1630/model.safetensors.index.json @@ -0,0 +1,298 @@ +{ + "metadata": { + "total_size": 16060522496 + }, + "weight_map": { + "lm_head.weight": "model-00004-of-00004.safetensors", + "model.embed_tokens.weight": "model-00001-of-00004.safetensors", + "model.layers.0.input_layernorm.weight": "model-00001-of-00004.safetensors", + "model.layers.0.mlp.down_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.0.mlp.gate_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.0.mlp.up_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.0.post_attention_layernorm.weight": "model-00001-of-00004.safetensors", + "model.layers.0.self_attn.k_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.0.self_attn.o_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.0.self_attn.q_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.0.self_attn.v_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.1.input_layernorm.weight": "model-00001-of-00004.safetensors", + "model.layers.1.mlp.down_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.1.mlp.gate_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.1.mlp.up_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.1.post_attention_layernorm.weight": "model-00001-of-00004.safetensors", + "model.layers.1.self_attn.k_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.1.self_attn.o_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.1.self_attn.q_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.1.self_attn.v_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.10.input_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.10.mlp.down_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.10.mlp.gate_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.10.mlp.up_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.10.post_attention_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.10.self_attn.k_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.10.self_attn.o_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.10.self_attn.q_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.10.self_attn.v_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.11.input_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.11.mlp.down_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.11.mlp.gate_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.11.mlp.up_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.11.post_attention_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.11.self_attn.k_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.11.self_attn.o_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.11.self_attn.q_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.11.self_attn.v_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.12.input_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.12.mlp.down_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.12.mlp.gate_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.12.mlp.up_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.12.post_attention_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.12.self_attn.k_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.12.self_attn.o_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.12.self_attn.q_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.12.self_attn.v_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.13.input_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.13.mlp.down_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.13.mlp.gate_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.13.mlp.up_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.13.post_attention_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.13.self_attn.k_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.13.self_attn.o_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.13.self_attn.q_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.13.self_attn.v_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.14.input_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.14.mlp.down_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.14.mlp.gate_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.14.mlp.up_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.14.post_attention_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.14.self_attn.k_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.14.self_attn.o_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.14.self_attn.q_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.14.self_attn.v_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.15.input_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.15.mlp.down_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.15.mlp.gate_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.15.mlp.up_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.15.post_attention_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.15.self_attn.k_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.15.self_attn.o_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.15.self_attn.q_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.15.self_attn.v_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.16.input_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.16.mlp.down_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.16.mlp.gate_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.16.mlp.up_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.16.post_attention_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.16.self_attn.k_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.16.self_attn.o_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.16.self_attn.q_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.16.self_attn.v_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.17.input_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.17.mlp.down_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.17.mlp.gate_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.17.mlp.up_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.17.post_attention_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.17.self_attn.k_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.17.self_attn.o_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.17.self_attn.q_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.17.self_attn.v_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.18.input_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.18.mlp.down_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.18.mlp.gate_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.18.mlp.up_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.18.post_attention_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.18.self_attn.k_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.18.self_attn.o_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.18.self_attn.q_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.18.self_attn.v_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.19.input_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.19.mlp.down_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.19.mlp.gate_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.19.mlp.up_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.19.post_attention_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.19.self_attn.k_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.19.self_attn.o_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.19.self_attn.q_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.19.self_attn.v_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.2.input_layernorm.weight": "model-00001-of-00004.safetensors", + "model.layers.2.mlp.down_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.2.mlp.gate_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.2.mlp.up_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.2.post_attention_layernorm.weight": "model-00001-of-00004.safetensors", + "model.layers.2.self_attn.k_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.2.self_attn.o_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.2.self_attn.q_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.2.self_attn.v_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.20.input_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.20.mlp.down_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.20.mlp.gate_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.20.mlp.up_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.20.post_attention_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.20.self_attn.k_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.20.self_attn.o_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.20.self_attn.q_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.20.self_attn.v_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.21.input_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.21.mlp.down_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.21.mlp.gate_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.21.mlp.up_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.21.post_attention_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.21.self_attn.k_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.21.self_attn.o_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.21.self_attn.q_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.21.self_attn.v_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.22.input_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.22.mlp.down_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.22.mlp.gate_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.22.mlp.up_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.22.post_attention_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.22.self_attn.k_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.22.self_attn.o_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.22.self_attn.q_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.22.self_attn.v_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.23.input_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.23.mlp.down_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.23.mlp.gate_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.23.mlp.up_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.23.post_attention_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.23.self_attn.k_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.23.self_attn.o_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.23.self_attn.q_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.23.self_attn.v_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.24.input_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.24.mlp.down_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.24.mlp.gate_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.24.mlp.up_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.24.post_attention_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.24.self_attn.k_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.24.self_attn.o_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.24.self_attn.q_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.24.self_attn.v_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.25.input_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.25.mlp.down_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.25.mlp.gate_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.25.mlp.up_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.25.post_attention_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.25.self_attn.k_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.25.self_attn.o_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.25.self_attn.q_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.25.self_attn.v_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.26.input_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.26.mlp.down_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.26.mlp.gate_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.26.mlp.up_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.26.post_attention_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.26.self_attn.k_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.26.self_attn.o_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.26.self_attn.q_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.26.self_attn.v_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.27.input_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.27.mlp.down_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.27.mlp.gate_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.27.mlp.up_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.27.post_attention_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.27.self_attn.k_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.27.self_attn.o_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.27.self_attn.q_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.27.self_attn.v_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.28.input_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.28.mlp.down_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.28.mlp.gate_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.28.mlp.up_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.28.post_attention_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.28.self_attn.k_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.28.self_attn.o_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.28.self_attn.q_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.28.self_attn.v_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.29.input_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.29.mlp.down_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.29.mlp.gate_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.29.mlp.up_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.29.post_attention_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.29.self_attn.k_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.29.self_attn.o_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.29.self_attn.q_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.29.self_attn.v_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.3.input_layernorm.weight": "model-00001-of-00004.safetensors", + "model.layers.3.mlp.down_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.3.mlp.gate_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.3.mlp.up_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.3.post_attention_layernorm.weight": "model-00001-of-00004.safetensors", + "model.layers.3.self_attn.k_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.3.self_attn.o_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.3.self_attn.q_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.3.self_attn.v_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.30.input_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.30.mlp.down_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.30.mlp.gate_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.30.mlp.up_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.30.post_attention_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.30.self_attn.k_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.30.self_attn.o_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.30.self_attn.q_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.30.self_attn.v_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.31.input_layernorm.weight": "model-00004-of-00004.safetensors", + "model.layers.31.mlp.down_proj.weight": "model-00004-of-00004.safetensors", + "model.layers.31.mlp.gate_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.31.mlp.up_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.31.post_attention_layernorm.weight": "model-00004-of-00004.safetensors", + "model.layers.31.self_attn.k_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.31.self_attn.o_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.31.self_attn.q_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.31.self_attn.v_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.4.input_layernorm.weight": "model-00001-of-00004.safetensors", + "model.layers.4.mlp.down_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.4.mlp.gate_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.4.mlp.up_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.4.post_attention_layernorm.weight": "model-00001-of-00004.safetensors", + "model.layers.4.self_attn.k_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.4.self_attn.o_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.4.self_attn.q_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.4.self_attn.v_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.5.input_layernorm.weight": "model-00001-of-00004.safetensors", + "model.layers.5.mlp.down_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.5.mlp.gate_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.5.mlp.up_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.5.post_attention_layernorm.weight": "model-00001-of-00004.safetensors", + "model.layers.5.self_attn.k_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.5.self_attn.o_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.5.self_attn.q_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.5.self_attn.v_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.6.input_layernorm.weight": "model-00001-of-00004.safetensors", + "model.layers.6.mlp.down_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.6.mlp.gate_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.6.mlp.up_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.6.post_attention_layernorm.weight": "model-00001-of-00004.safetensors", + "model.layers.6.self_attn.k_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.6.self_attn.o_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.6.self_attn.q_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.6.self_attn.v_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.7.input_layernorm.weight": "model-00001-of-00004.safetensors", + "model.layers.7.mlp.down_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.7.mlp.gate_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.7.mlp.up_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.7.post_attention_layernorm.weight": "model-00001-of-00004.safetensors", + "model.layers.7.self_attn.k_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.7.self_attn.o_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.7.self_attn.q_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.7.self_attn.v_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.8.input_layernorm.weight": "model-00001-of-00004.safetensors", + "model.layers.8.mlp.down_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.8.mlp.gate_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.8.mlp.up_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.8.post_attention_layernorm.weight": "model-00001-of-00004.safetensors", + "model.layers.8.self_attn.k_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.8.self_attn.o_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.8.self_attn.q_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.8.self_attn.v_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.9.input_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.9.mlp.down_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.9.mlp.gate_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.9.mlp.up_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.9.post_attention_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.9.self_attn.k_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.9.self_attn.o_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.9.self_attn.q_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.9.self_attn.v_proj.weight": "model-00002-of-00004.safetensors", + "model.norm.weight": "model-00004-of-00004.safetensors" + } +} diff --git a/checkpoint-1630/rng_state_0.pth b/checkpoint-1630/rng_state_0.pth new file mode 100644 index 0000000..b346349 --- /dev/null +++ b/checkpoint-1630/rng_state_0.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:ad8a35afd8967cbb748405387e44426e43ad127028e826eddc9b67d2ca873c85 +size 15984 diff --git a/checkpoint-1630/rng_state_1.pth b/checkpoint-1630/rng_state_1.pth new file mode 100644 index 0000000..68f3c69 --- /dev/null +++ b/checkpoint-1630/rng_state_1.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:f338ce80d7c441076bfc8c53b84067a0181f5a14e80c13d5acb8150b659f4d73 +size 15984 diff --git a/checkpoint-1630/rng_state_10.pth b/checkpoint-1630/rng_state_10.pth new file mode 100644 index 0000000..09612af --- /dev/null +++ b/checkpoint-1630/rng_state_10.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:4d4c67c1d0ec6b889370f634ff29f584f829acaeaf69196a69304fdba936f9d7 +size 15997 diff --git a/checkpoint-1630/rng_state_11.pth b/checkpoint-1630/rng_state_11.pth new file mode 100644 index 0000000..06a46cd --- /dev/null +++ b/checkpoint-1630/rng_state_11.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:afd0c4e2ac5f61d5f4df5d9d5783018b7937173b4afdbb92816c2668982cb351 +size 15997 diff --git a/checkpoint-1630/rng_state_12.pth b/checkpoint-1630/rng_state_12.pth new file mode 100644 index 0000000..bf6f992 --- /dev/null +++ b/checkpoint-1630/rng_state_12.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:a9d1c7f561c5512193e8053ecf371e4fef647a8503481f2dad76332ddbe164fd +size 15997 diff --git a/checkpoint-1630/rng_state_13.pth b/checkpoint-1630/rng_state_13.pth new file mode 100644 index 0000000..eac9c7f --- /dev/null +++ b/checkpoint-1630/rng_state_13.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:7604a3cfaf0e3c3f2aa2b3547660615db82205d4158951bb90f69ffc560ed179 +size 15997 diff --git a/checkpoint-1630/rng_state_14.pth b/checkpoint-1630/rng_state_14.pth new file mode 100644 index 0000000..29fb251 --- /dev/null +++ b/checkpoint-1630/rng_state_14.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:4d984186a3bc06445d6c077fcbf64da96d47d3ed704e68f0212045efb4b91dbb +size 15997 diff --git a/checkpoint-1630/rng_state_15.pth b/checkpoint-1630/rng_state_15.pth new file mode 100644 index 0000000..0bf361b --- /dev/null +++ b/checkpoint-1630/rng_state_15.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:1b3945ba3535361fa13a8460a70ebbd44ece2514a1c9dcb76abfe6f54a775c60 +size 15997 diff --git a/checkpoint-1630/rng_state_16.pth b/checkpoint-1630/rng_state_16.pth new file mode 100644 index 0000000..0e7adea --- /dev/null +++ b/checkpoint-1630/rng_state_16.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:853a02543fd3223edd9334ea75c62e6e99ded22ac864e9930fbf5add26045fe7 +size 15997 diff --git a/checkpoint-1630/rng_state_17.pth b/checkpoint-1630/rng_state_17.pth new file mode 100644 index 0000000..18e8350 --- /dev/null +++ b/checkpoint-1630/rng_state_17.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:9c0420ad6cf62c9fe9cda1807f68e8e40235a8917dbab858f800d24e3ba8eeae +size 15997 diff --git a/checkpoint-1630/rng_state_18.pth b/checkpoint-1630/rng_state_18.pth new file mode 100644 index 0000000..9b6c223 --- /dev/null +++ b/checkpoint-1630/rng_state_18.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:e167524b369e20a3a17ee7f6124322da680ccf2ae553b4adc0e7462d2dfba82b +size 15997 diff --git a/checkpoint-1630/rng_state_19.pth b/checkpoint-1630/rng_state_19.pth new file mode 100644 index 0000000..b0d318f --- /dev/null +++ b/checkpoint-1630/rng_state_19.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:a0915a78eba7c1d2768ecd03818e3c01afd7f3168a3318b1212413bd2c62f7f1 +size 15997 diff --git a/checkpoint-1630/rng_state_2.pth b/checkpoint-1630/rng_state_2.pth new file mode 100644 index 0000000..be044f6 --- /dev/null +++ b/checkpoint-1630/rng_state_2.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:c9fbc9fa428939be10b46779f0eb5cd833e0da426b1cbdee77b3a55b6952235b +size 15984 diff --git a/checkpoint-1630/rng_state_20.pth b/checkpoint-1630/rng_state_20.pth new file mode 100644 index 0000000..ece204f --- /dev/null +++ b/checkpoint-1630/rng_state_20.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:8f8f2fdb90584ebb7247c85570082ac7c6338216de82b07f132a2c679923379a +size 15997 diff --git a/checkpoint-1630/rng_state_21.pth b/checkpoint-1630/rng_state_21.pth new file mode 100644 index 0000000..5b62c48 --- /dev/null +++ b/checkpoint-1630/rng_state_21.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:7a3863ac6512e946a826d479520004385d29f3fd0a2b686ccbfdf488332aa442 +size 15997 diff --git a/checkpoint-1630/rng_state_22.pth b/checkpoint-1630/rng_state_22.pth new file mode 100644 index 0000000..b0fa34f --- /dev/null +++ b/checkpoint-1630/rng_state_22.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:1dfe94fb47194a2a298bed1b33ad18c2132747148814886bd9b8910a7691ba0b +size 15997 diff --git a/checkpoint-1630/rng_state_23.pth b/checkpoint-1630/rng_state_23.pth new file mode 100644 index 0000000..c7c7ed4 --- /dev/null +++ b/checkpoint-1630/rng_state_23.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:c683c09c8883037948fabb143cd972207db3cee22d7efcbcc5866a8ec85c94c6 +size 15997 diff --git a/checkpoint-1630/rng_state_24.pth b/checkpoint-1630/rng_state_24.pth new file mode 100644 index 0000000..be36a1a --- /dev/null +++ b/checkpoint-1630/rng_state_24.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:53c201c89debf1ca694786330ea98ca69da6340ac3b70f40aad77213d1771bf8 +size 15997 diff --git a/checkpoint-1630/rng_state_25.pth b/checkpoint-1630/rng_state_25.pth new file mode 100644 index 0000000..b34b54b --- /dev/null +++ b/checkpoint-1630/rng_state_25.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:b7b8393dc10fb6ab2222179b268184fa5330fd417b66ebb65986818f47334a55 +size 15997 diff --git a/checkpoint-1630/rng_state_26.pth b/checkpoint-1630/rng_state_26.pth new file mode 100644 index 0000000..354820a --- /dev/null +++ b/checkpoint-1630/rng_state_26.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:6b29094f6dcc92be7d58da39df551c573728836c5aec67bef90790810c0d3882 +size 15997 diff --git a/checkpoint-1630/rng_state_27.pth b/checkpoint-1630/rng_state_27.pth new file mode 100644 index 0000000..747b806 --- /dev/null +++ b/checkpoint-1630/rng_state_27.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:b64ff98dfb30bbfa08760a0a79d9be7407917ff0e32f22b03d6528bf7e292e44 +size 15997 diff --git a/checkpoint-1630/rng_state_28.pth b/checkpoint-1630/rng_state_28.pth new file mode 100644 index 0000000..b316dea --- /dev/null +++ b/checkpoint-1630/rng_state_28.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:04e19d4d07b899c9158590672b9acc125c3fb4082f5aed08ddaee573b4dbf33f +size 15997 diff --git a/checkpoint-1630/rng_state_29.pth b/checkpoint-1630/rng_state_29.pth new file mode 100644 index 0000000..7b0fc3c --- /dev/null +++ b/checkpoint-1630/rng_state_29.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:47a620c2380d984b51262d8f8d1853631c1bd13c3da4ea8f73e8dc61dd3fb765 +size 15997 diff --git a/checkpoint-1630/rng_state_3.pth b/checkpoint-1630/rng_state_3.pth new file mode 100644 index 0000000..fc82524 --- /dev/null +++ b/checkpoint-1630/rng_state_3.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:ac55dba0b79d5fa4699d239da2f966d52040d576d31234ac8d4632e6956481bc +size 15984 diff --git a/checkpoint-1630/rng_state_30.pth b/checkpoint-1630/rng_state_30.pth new file mode 100644 index 0000000..f88acd7 --- /dev/null +++ b/checkpoint-1630/rng_state_30.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:1516e58b9c7537b0c41d41487bc1464770e53c3dabdb8cf4f5c07bcc51e07ed6 +size 15997 diff --git a/checkpoint-1630/rng_state_31.pth b/checkpoint-1630/rng_state_31.pth new file mode 100644 index 0000000..4e033f4 --- /dev/null +++ b/checkpoint-1630/rng_state_31.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:58a50bef57066132c2e5b6b53fbb78e101505dd4ad585ec52fb3c11ca0663c22 +size 15997 diff --git a/checkpoint-1630/rng_state_4.pth b/checkpoint-1630/rng_state_4.pth new file mode 100644 index 0000000..d30f52a --- /dev/null +++ b/checkpoint-1630/rng_state_4.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:af2d0c015100768ffa23faf3b6c2d54ea89eb045603e30e55cd211e06ff34972 +size 15984 diff --git a/checkpoint-1630/rng_state_5.pth b/checkpoint-1630/rng_state_5.pth new file mode 100644 index 0000000..c8715d2 --- /dev/null +++ b/checkpoint-1630/rng_state_5.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:c60a1b40608e34bc801c8231f97b81c53b5290dfaed1b9cd0ccbeca29574a991 +size 15984 diff --git a/checkpoint-1630/rng_state_6.pth b/checkpoint-1630/rng_state_6.pth new file mode 100644 index 0000000..1ed791b --- /dev/null +++ b/checkpoint-1630/rng_state_6.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:3ad6a142a403eb9aafc4a3a9a856bca648fe31fd22d796867baca31fb13656aa +size 15984 diff --git a/checkpoint-1630/rng_state_7.pth b/checkpoint-1630/rng_state_7.pth new file mode 100644 index 0000000..800c3bb --- /dev/null +++ b/checkpoint-1630/rng_state_7.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:38bc23a138cc800b22881742c0f3f9a71731a9a7111c6058a0077e6274d21773 +size 15984 diff --git a/checkpoint-1630/rng_state_8.pth b/checkpoint-1630/rng_state_8.pth new file mode 100644 index 0000000..d7068b4 --- /dev/null +++ b/checkpoint-1630/rng_state_8.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:f4b75777b43df70dca87f91deed2e613941188792f244d5eeb06da8ab038bd36 +size 15984 diff --git a/checkpoint-1630/rng_state_9.pth b/checkpoint-1630/rng_state_9.pth new file mode 100644 index 0000000..d42657d --- /dev/null +++ b/checkpoint-1630/rng_state_9.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:a8f9eb93d0ac349ba8b2271c417e5fcc4b70447a424f3f861c3338fbd99a48f4 +size 15984 diff --git a/checkpoint-1630/scheduler.pt b/checkpoint-1630/scheduler.pt new file mode 100644 index 0000000..7d23740 --- /dev/null +++ b/checkpoint-1630/scheduler.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:bf2af7e49c8b1068724ec29892256c434bb0eecdd677615f65de2fef3c9b40a9 +size 1064 diff --git a/checkpoint-1630/special_tokens_map.json b/checkpoint-1630/special_tokens_map.json new file mode 100644 index 0000000..278b7f0 --- /dev/null +++ b/checkpoint-1630/special_tokens_map.json @@ -0,0 +1,23 @@ +{ + "bos_token": { + "content": "<|begin_of_text|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false + }, + "eos_token": { + "content": "<|eot_id|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false + }, + "pad_token": { + "content": "<|end_of_text|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false + } +} diff --git a/checkpoint-1630/tokenizer.json b/checkpoint-1630/tokenizer.json new file mode 100644 index 0000000..1c1d8d5 --- /dev/null +++ b/checkpoint-1630/tokenizer.json @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:6b9e4e7fb171f92fd137b777cc2714bf87d11576700a1dcd7a399e7bbe39537b +size 17209920 diff --git a/checkpoint-1630/tokenizer_config.json b/checkpoint-1630/tokenizer_config.json new file mode 100644 index 0000000..8098161 --- /dev/null +++ b/checkpoint-1630/tokenizer_config.json @@ -0,0 +1,2064 @@ +{ + "added_tokens_decoder": { + "128000": { + "content": "<|begin_of_text|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128001": { + "content": "<|end_of_text|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128002": { + "content": "<|reserved_special_token_0|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128003": { + "content": "<|reserved_special_token_1|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128004": { + "content": "<|finetune_right_pad_id|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128005": { + "content": "<|reserved_special_token_2|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128006": { + "content": "<|start_header_id|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128007": { + "content": "<|end_header_id|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128008": { + "content": "<|eom_id|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128009": { + "content": "<|eot_id|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128010": { + "content": "<|python_tag|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128011": { + "content": "<|reserved_special_token_3|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128012": { + "content": "<|reserved_special_token_4|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128013": { + "content": "<|reserved_special_token_5|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128014": { + "content": "<|reserved_special_token_6|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128015": { + "content": "<|reserved_special_token_7|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128016": { + "content": "<|reserved_special_token_8|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128017": { + "content": "<|reserved_special_token_9|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128018": { + "content": "<|reserved_special_token_10|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128019": { + "content": "<|reserved_special_token_11|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128020": { + "content": "<|reserved_special_token_12|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128021": { + "content": "<|reserved_special_token_13|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128022": { + "content": "<|reserved_special_token_14|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128023": { + "content": "<|reserved_special_token_15|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128024": { + "content": "<|reserved_special_token_16|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128025": { + "content": "<|reserved_special_token_17|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128026": { + "content": "<|reserved_special_token_18|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128027": { + "content": "<|reserved_special_token_19|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128028": { + "content": "<|reserved_special_token_20|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128029": { + "content": "<|reserved_special_token_21|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128030": { + "content": "<|reserved_special_token_22|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128031": { + "content": "<|reserved_special_token_23|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128032": { + "content": "<|reserved_special_token_24|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128033": { + "content": "<|reserved_special_token_25|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128034": { + "content": "<|reserved_special_token_26|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128035": { + "content": "<|reserved_special_token_27|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128036": { + "content": "<|reserved_special_token_28|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128037": { + "content": "<|reserved_special_token_29|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128038": { + "content": "<|reserved_special_token_30|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128039": { + "content": "<|reserved_special_token_31|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128040": { + "content": "<|reserved_special_token_32|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128041": { + "content": "<|reserved_special_token_33|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128042": { + "content": "<|reserved_special_token_34|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128043": { + "content": "<|reserved_special_token_35|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128044": { + "content": "<|reserved_special_token_36|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128045": { + "content": "<|reserved_special_token_37|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128046": { + "content": "<|reserved_special_token_38|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128047": { + "content": "<|reserved_special_token_39|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128048": { + "content": "<|reserved_special_token_40|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128049": { + "content": "<|reserved_special_token_41|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128050": { + "content": "<|reserved_special_token_42|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128051": { + "content": "<|reserved_special_token_43|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128052": { + "content": "<|reserved_special_token_44|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128053": { + "content": "<|reserved_special_token_45|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128054": { + "content": "<|reserved_special_token_46|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128055": { + "content": "<|reserved_special_token_47|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128056": { + "content": "<|reserved_special_token_48|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128057": { + "content": "<|reserved_special_token_49|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128058": { + "content": "<|reserved_special_token_50|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128059": { + "content": "<|reserved_special_token_51|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128060": { + "content": "<|reserved_special_token_52|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128061": { + "content": "<|reserved_special_token_53|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128062": { + "content": "<|reserved_special_token_54|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128063": { + "content": "<|reserved_special_token_55|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128064": { + "content": "<|reserved_special_token_56|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128065": { + "content": "<|reserved_special_token_57|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128066": { + "content": "<|reserved_special_token_58|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128067": { + "content": "<|reserved_special_token_59|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128068": { + "content": "<|reserved_special_token_60|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128069": { + "content": "<|reserved_special_token_61|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128070": { + "content": "<|reserved_special_token_62|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128071": { + "content": "<|reserved_special_token_63|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128072": { + "content": "<|reserved_special_token_64|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128073": { + "content": "<|reserved_special_token_65|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128074": { + "content": "<|reserved_special_token_66|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128075": { + "content": "<|reserved_special_token_67|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128076": { + "content": "<|reserved_special_token_68|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128077": { + "content": "<|reserved_special_token_69|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128078": { + "content": "<|reserved_special_token_70|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128079": { + "content": "<|reserved_special_token_71|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128080": { + "content": "<|reserved_special_token_72|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128081": { + "content": "<|reserved_special_token_73|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128082": { + "content": "<|reserved_special_token_74|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128083": { + "content": "<|reserved_special_token_75|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128084": { + "content": "<|reserved_special_token_76|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128085": { + "content": "<|reserved_special_token_77|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128086": { + "content": "<|reserved_special_token_78|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128087": { + "content": "<|reserved_special_token_79|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128088": { + "content": "<|reserved_special_token_80|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128089": { + "content": "<|reserved_special_token_81|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128090": { + "content": "<|reserved_special_token_82|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128091": { + "content": "<|reserved_special_token_83|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128092": { + "content": "<|reserved_special_token_84|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128093": { + "content": "<|reserved_special_token_85|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128094": { + "content": "<|reserved_special_token_86|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128095": { + "content": "<|reserved_special_token_87|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128096": { + "content": "<|reserved_special_token_88|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128097": { + "content": "<|reserved_special_token_89|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128098": { + "content": "<|reserved_special_token_90|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128099": { + "content": "<|reserved_special_token_91|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128100": { + "content": "<|reserved_special_token_92|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128101": { + "content": "<|reserved_special_token_93|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128102": { + "content": "<|reserved_special_token_94|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128103": { + "content": "<|reserved_special_token_95|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128104": { + "content": "<|reserved_special_token_96|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128105": { + "content": "<|reserved_special_token_97|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128106": { + "content": "<|reserved_special_token_98|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128107": { + "content": "<|reserved_special_token_99|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128108": { + "content": "<|reserved_special_token_100|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128109": { + "content": "<|reserved_special_token_101|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128110": { + "content": "<|reserved_special_token_102|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128111": { + "content": "<|reserved_special_token_103|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128112": { + "content": "<|reserved_special_token_104|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128113": { + "content": "<|reserved_special_token_105|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128114": { + "content": "<|reserved_special_token_106|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128115": { + "content": "<|reserved_special_token_107|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128116": { + "content": "<|reserved_special_token_108|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128117": { + "content": "<|reserved_special_token_109|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128118": { + "content": "<|reserved_special_token_110|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128119": { + "content": "<|reserved_special_token_111|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128120": { + "content": "<|reserved_special_token_112|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128121": { + "content": "<|reserved_special_token_113|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128122": { + "content": "<|reserved_special_token_114|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128123": { + "content": "<|reserved_special_token_115|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128124": { + "content": "<|reserved_special_token_116|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128125": { + "content": "<|reserved_special_token_117|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128126": { + "content": "<|reserved_special_token_118|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128127": { + "content": "<|reserved_special_token_119|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128128": { + "content": "<|reserved_special_token_120|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128129": { + "content": "<|reserved_special_token_121|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128130": { + "content": "<|reserved_special_token_122|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128131": { + "content": "<|reserved_special_token_123|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128132": { + "content": "<|reserved_special_token_124|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128133": { + "content": "<|reserved_special_token_125|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128134": { + "content": "<|reserved_special_token_126|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128135": { + "content": "<|reserved_special_token_127|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128136": { + "content": "<|reserved_special_token_128|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128137": { + "content": "<|reserved_special_token_129|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128138": { + "content": "<|reserved_special_token_130|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128139": { + "content": "<|reserved_special_token_131|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128140": { + "content": "<|reserved_special_token_132|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128141": { + "content": "<|reserved_special_token_133|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128142": { + "content": "<|reserved_special_token_134|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128143": { + "content": "<|reserved_special_token_135|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128144": { + "content": "<|reserved_special_token_136|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128145": { + "content": "<|reserved_special_token_137|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128146": { + "content": "<|reserved_special_token_138|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128147": { + "content": "<|reserved_special_token_139|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128148": { + "content": "<|reserved_special_token_140|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128149": { + "content": "<|reserved_special_token_141|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128150": { + "content": "<|reserved_special_token_142|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128151": { + "content": "<|reserved_special_token_143|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128152": { + "content": "<|reserved_special_token_144|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128153": { + "content": "<|reserved_special_token_145|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128154": { + "content": "<|reserved_special_token_146|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128155": { + "content": "<|reserved_special_token_147|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128156": { + "content": "<|reserved_special_token_148|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128157": { + "content": "<|reserved_special_token_149|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128158": { + "content": "<|reserved_special_token_150|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128159": { + "content": "<|reserved_special_token_151|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128160": { + "content": "<|reserved_special_token_152|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128161": { + "content": "<|reserved_special_token_153|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128162": { + "content": "<|reserved_special_token_154|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128163": { + "content": "<|reserved_special_token_155|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128164": { + "content": "<|reserved_special_token_156|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128165": { + "content": "<|reserved_special_token_157|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128166": { + "content": "<|reserved_special_token_158|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128167": { + "content": "<|reserved_special_token_159|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128168": { + "content": "<|reserved_special_token_160|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128169": { + "content": "<|reserved_special_token_161|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128170": { + "content": "<|reserved_special_token_162|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128171": { + "content": "<|reserved_special_token_163|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128172": { + "content": "<|reserved_special_token_164|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128173": { + "content": "<|reserved_special_token_165|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128174": { + "content": "<|reserved_special_token_166|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128175": { + "content": "<|reserved_special_token_167|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128176": { + "content": "<|reserved_special_token_168|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128177": { + "content": "<|reserved_special_token_169|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128178": { + "content": "<|reserved_special_token_170|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128179": { + "content": "<|reserved_special_token_171|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128180": { + "content": "<|reserved_special_token_172|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128181": { + "content": "<|reserved_special_token_173|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128182": { + "content": "<|reserved_special_token_174|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128183": { + "content": "<|reserved_special_token_175|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128184": { + "content": "<|reserved_special_token_176|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128185": { + "content": "<|reserved_special_token_177|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128186": { + "content": "<|reserved_special_token_178|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128187": { + "content": "<|reserved_special_token_179|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128188": { + "content": "<|reserved_special_token_180|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128189": { + "content": "<|reserved_special_token_181|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128190": { + "content": "<|reserved_special_token_182|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128191": { + "content": "<|reserved_special_token_183|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128192": { + "content": "<|reserved_special_token_184|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128193": { + "content": "<|reserved_special_token_185|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128194": { + "content": "<|reserved_special_token_186|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128195": { + "content": "<|reserved_special_token_187|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128196": { + "content": "<|reserved_special_token_188|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128197": { + "content": "<|reserved_special_token_189|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128198": { + "content": "<|reserved_special_token_190|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128199": { + "content": "<|reserved_special_token_191|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128200": { + "content": "<|reserved_special_token_192|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128201": { + "content": "<|reserved_special_token_193|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128202": { + "content": "<|reserved_special_token_194|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128203": { + "content": "<|reserved_special_token_195|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128204": { + "content": "<|reserved_special_token_196|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128205": { + "content": "<|reserved_special_token_197|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128206": { + "content": "<|reserved_special_token_198|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128207": { + "content": "<|reserved_special_token_199|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128208": { + "content": "<|reserved_special_token_200|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128209": { + "content": "<|reserved_special_token_201|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128210": { + "content": "<|reserved_special_token_202|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128211": { + "content": "<|reserved_special_token_203|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128212": { + "content": "<|reserved_special_token_204|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128213": { + "content": "<|reserved_special_token_205|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128214": { + "content": "<|reserved_special_token_206|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128215": { + "content": "<|reserved_special_token_207|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128216": { + "content": "<|reserved_special_token_208|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128217": { + "content": "<|reserved_special_token_209|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128218": { + "content": "<|reserved_special_token_210|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128219": { + "content": "<|reserved_special_token_211|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128220": { + "content": "<|reserved_special_token_212|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128221": { + "content": "<|reserved_special_token_213|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128222": { + "content": "<|reserved_special_token_214|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128223": { + "content": "<|reserved_special_token_215|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128224": { + "content": "<|reserved_special_token_216|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128225": { + "content": "<|reserved_special_token_217|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128226": { + "content": "<|reserved_special_token_218|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128227": { + "content": "<|reserved_special_token_219|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128228": { + "content": "<|reserved_special_token_220|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128229": { + "content": "<|reserved_special_token_221|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128230": { + "content": "<|reserved_special_token_222|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128231": { + "content": "<|reserved_special_token_223|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128232": { + "content": "<|reserved_special_token_224|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128233": { + "content": "<|reserved_special_token_225|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128234": { + "content": "<|reserved_special_token_226|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128235": { + "content": "<|reserved_special_token_227|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128236": { + "content": "<|reserved_special_token_228|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128237": { + "content": "<|reserved_special_token_229|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128238": { + "content": "<|reserved_special_token_230|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128239": { + "content": "<|reserved_special_token_231|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128240": { + "content": "<|reserved_special_token_232|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128241": { + "content": "<|reserved_special_token_233|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128242": { + "content": "<|reserved_special_token_234|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128243": { + "content": "<|reserved_special_token_235|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128244": { + "content": "<|reserved_special_token_236|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128245": { + "content": "<|reserved_special_token_237|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128246": { + "content": "<|reserved_special_token_238|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128247": { + "content": "<|reserved_special_token_239|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128248": { + "content": "<|reserved_special_token_240|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128249": { + "content": "<|reserved_special_token_241|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128250": { + "content": "<|reserved_special_token_242|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128251": { + "content": "<|reserved_special_token_243|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128252": { + "content": "<|reserved_special_token_244|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128253": { + "content": "<|reserved_special_token_245|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128254": { + "content": "<|reserved_special_token_246|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128255": { + "content": "<|reserved_special_token_247|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + } + }, + "bos_token": "<|begin_of_text|>", + "chat_template": "{% if not add_generation_prompt is defined %}{% set add_generation_prompt = false %}{% endif %}{% set loop_messages = messages %}{% for message in loop_messages %}{% set content = '<|start_header_id|>' + message['role'] + '<|end_header_id|>\n\n'+ message['content'] | trim + '<|eot_id|>' %}{% if loop.index0 == 0 %}{% set content = bos_token + content %}{% endif %}{{ content }}{% endfor %}{% if add_generation_prompt %}{{ '<|start_header_id|>assistant<|end_header_id|>\n\n' }}{% endif %}", + "clean_up_tokenization_spaces": true, + "eos_token": "<|eot_id|>", + "extra_special_tokens": {}, + "model_input_names": [ + "input_ids", + "attention_mask" + ], + "model_max_length": 131072, + "pad_token": "<|end_of_text|>", + "tokenizer_class": "PreTrainedTokenizer" +} diff --git a/checkpoint-1630/trainer_state.json b/checkpoint-1630/trainer_state.json new file mode 100644 index 0000000..3cc093a --- /dev/null +++ b/checkpoint-1630/trainer_state.json @@ -0,0 +1,24484 @@ +{ + "best_global_step": null, + "best_metric": null, + "best_model_checkpoint": null, + "epoch": 2.0, + "eval_steps": 500, + "global_step": 1630, + "is_hyper_param_search": false, + "is_local_process_zero": true, + "is_world_process_zero": true, + "log_history": [ + { + "epoch": 0.001226993865030675, + "grad_norm": 151.20719434086217, + "learning_rate": 6.000000000000001e-08, + "logits/chosen": -0.32421875, + "logits/rejected": -0.515625, + "logps/chosen": -474.0, + "logps/rejected": -316.0, + "loss": 0.6914, + "rewards/accuracies": 0.0, + "rewards/chosen": 0.0, + "rewards/margins": 0.0, + "rewards/rejected": 0.0, + "step": 1 + }, + { + "epoch": 0.00245398773006135, + "grad_norm": 162.171350331327, + "learning_rate": 1.2000000000000002e-07, + "logits/chosen": -0.22265625, + "logits/rejected": -0.439453125, + "logps/chosen": -436.0, + "logps/rejected": -292.0, + "loss": 0.6914, + "rewards/accuracies": 0.0, + "rewards/chosen": 0.0, + "rewards/margins": 0.0, + "rewards/rejected": 0.0, + "step": 2 + }, + { + "epoch": 0.0036809815950920245, + "grad_norm": 121.91181472288706, + "learning_rate": 1.8e-07, + "logits/chosen": -0.1298828125, + "logits/rejected": -0.2578125, + "logps/chosen": -368.0, + "logps/rejected": -272.0, + "loss": 0.7012, + "rewards/accuracies": 0.3359375, + "rewards/chosen": -0.0006103515625, + "rewards/margins": -0.00506591796875, + "rewards/rejected": 0.00445556640625, + "step": 3 + }, + { + "epoch": 0.0049079754601227, + "grad_norm": 138.61232520616198, + "learning_rate": 2.4000000000000003e-07, + "logits/chosen": -0.166015625, + "logits/rejected": -0.326171875, + "logps/chosen": -422.0, + "logps/rejected": -242.0, + "loss": 0.7003, + "rewards/accuracies": 0.359375, + "rewards/chosen": 0.0057373046875, + "rewards/margins": -0.00762939453125, + "rewards/rejected": 0.01336669921875, + "step": 4 + }, + { + "epoch": 0.006134969325153374, + "grad_norm": 318.4843338984337, + "learning_rate": 3.0000000000000004e-07, + "logits/chosen": -0.1826171875, + "logits/rejected": -0.32421875, + "logps/chosen": -380.0, + "logps/rejected": -270.0, + "loss": 0.7002, + "rewards/accuracies": 0.296875, + "rewards/chosen": -0.0037994384765625, + "rewards/margins": -0.01123046875, + "rewards/rejected": 0.007415771484375, + "step": 5 + }, + { + "epoch": 0.007361963190184049, + "grad_norm": 356.6573592678301, + "learning_rate": 3.6e-07, + "logits/chosen": -0.255859375, + "logits/rejected": -0.345703125, + "logps/chosen": -412.0, + "logps/rejected": -306.0, + "loss": 0.692, + "rewards/accuracies": 0.359375, + "rewards/chosen": 0.0208740234375, + "rewards/margins": 0.01153564453125, + "rewards/rejected": 0.00927734375, + "step": 6 + }, + { + "epoch": 0.008588957055214725, + "grad_norm": 228.0843622624859, + "learning_rate": 4.2000000000000006e-07, + "logits/chosen": -0.2060546875, + "logits/rejected": -0.38671875, + "logps/chosen": -378.0, + "logps/rejected": -266.0, + "loss": 0.6962, + "rewards/accuracies": 0.3671875, + "rewards/chosen": 0.006683349609375, + "rewards/margins": -0.001617431640625, + "rewards/rejected": 0.00830078125, + "step": 7 + }, + { + "epoch": 0.0098159509202454, + "grad_norm": 139.9677750400214, + "learning_rate": 4.800000000000001e-07, + "logits/chosen": -0.111328125, + "logits/rejected": -0.3203125, + "logps/chosen": -432.0, + "logps/rejected": -292.0, + "loss": 0.6667, + "rewards/accuracies": 0.453125, + "rewards/chosen": 0.054443359375, + "rewards/margins": 0.06201171875, + "rewards/rejected": -0.00762939453125, + "step": 8 + }, + { + "epoch": 0.011042944785276074, + "grad_norm": 114.26865689975604, + "learning_rate": 5.4e-07, + "logits/chosen": -0.1318359375, + "logits/rejected": -0.359375, + "logps/chosen": -402.0, + "logps/rejected": -294.0, + "loss": 0.6492, + "rewards/accuracies": 0.4921875, + "rewards/chosen": 0.0654296875, + "rewards/margins": 0.10986328125, + "rewards/rejected": -0.044677734375, + "step": 9 + }, + { + "epoch": 0.012269938650306749, + "grad_norm": 150.56729793626556, + "learning_rate": 6.000000000000001e-07, + "logits/chosen": -0.1953125, + "logits/rejected": -0.349609375, + "logps/chosen": -402.0, + "logps/rejected": -278.0, + "loss": 0.6146, + "rewards/accuracies": 0.59375, + "rewards/chosen": 0.158203125, + "rewards/margins": 0.1943359375, + "rewards/rejected": -0.0361328125, + "step": 10 + }, + { + "epoch": 0.013496932515337423, + "grad_norm": 85.94823037139018, + "learning_rate": 6.6e-07, + "logits/chosen": -0.375, + "logits/rejected": -0.515625, + "logps/chosen": -420.0, + "logps/rejected": -298.0, + "loss": 0.5535, + "rewards/accuracies": 0.6875, + "rewards/chosen": 0.2451171875, + "rewards/margins": 0.3984375, + "rewards/rejected": -0.1533203125, + "step": 11 + }, + { + "epoch": 0.014723926380368098, + "grad_norm": 60.96373680958018, + "learning_rate": 7.2e-07, + "logits/chosen": -0.1337890625, + "logits/rejected": -0.244140625, + "logps/chosen": -372.0, + "logps/rejected": -262.0, + "loss": 0.548, + "rewards/accuracies": 0.6953125, + "rewards/chosen": 0.255859375, + "rewards/margins": 0.435546875, + "rewards/rejected": -0.1796875, + "step": 12 + }, + { + "epoch": 0.015950920245398775, + "grad_norm": 92.70871898741228, + "learning_rate": 7.8e-07, + "logits/chosen": -0.240234375, + "logits/rejected": -0.408203125, + "logps/chosen": -390.0, + "logps/rejected": -251.0, + "loss": 0.5255, + "rewards/accuracies": 0.703125, + "rewards/chosen": 0.302734375, + "rewards/margins": 0.56640625, + "rewards/rejected": -0.26171875, + "step": 13 + }, + { + "epoch": 0.01717791411042945, + "grad_norm": 48.59299098972598, + "learning_rate": 8.400000000000001e-07, + "logits/chosen": -0.2138671875, + "logits/rejected": -0.3984375, + "logps/chosen": -466.0, + "logps/rejected": -314.0, + "loss": 0.4612, + "rewards/accuracies": 0.7265625, + "rewards/chosen": 0.765625, + "rewards/margins": 1.078125, + "rewards/rejected": -0.31640625, + "step": 14 + }, + { + "epoch": 0.018404907975460124, + "grad_norm": 59.699183777930486, + "learning_rate": 9e-07, + "logits/chosen": -0.10400390625, + "logits/rejected": -0.177734375, + "logps/chosen": -414.0, + "logps/rejected": -292.0, + "loss": 0.5014, + "rewards/accuracies": 0.734375, + "rewards/chosen": 0.6484375, + "rewards/margins": 0.96875, + "rewards/rejected": -0.31640625, + "step": 15 + }, + { + "epoch": 0.0196319018404908, + "grad_norm": 52.10075632980986, + "learning_rate": 9.600000000000001e-07, + "logits/chosen": -0.25, + "logits/rejected": -0.42578125, + "logps/chosen": -400.0, + "logps/rejected": -310.0, + "loss": 0.4424, + "rewards/accuracies": 0.7265625, + "rewards/chosen": 0.84375, + "rewards/margins": 1.390625, + "rewards/rejected": -0.54296875, + "step": 16 + }, + { + "epoch": 0.020858895705521473, + "grad_norm": 42.16097538458702, + "learning_rate": 1.0200000000000002e-06, + "logits/chosen": -0.1572265625, + "logits/rejected": -0.31640625, + "logps/chosen": -426.0, + "logps/rejected": -260.0, + "loss": 0.4308, + "rewards/accuracies": 0.796875, + "rewards/chosen": 0.984375, + "rewards/margins": 1.578125, + "rewards/rejected": -0.58984375, + "step": 17 + }, + { + "epoch": 0.022085889570552148, + "grad_norm": 46.56295761289983, + "learning_rate": 1.08e-06, + "logits/chosen": -0.1435546875, + "logits/rejected": -0.330078125, + "logps/chosen": -398.0, + "logps/rejected": -282.0, + "loss": 0.432, + "rewards/accuracies": 0.828125, + "rewards/chosen": 0.875, + "rewards/margins": 1.6875, + "rewards/rejected": -0.8125, + "step": 18 + }, + { + "epoch": 0.023312883435582823, + "grad_norm": 49.57615257325465, + "learning_rate": 1.14e-06, + "logits/chosen": -0.185546875, + "logits/rejected": -0.3671875, + "logps/chosen": -382.0, + "logps/rejected": -296.0, + "loss": 0.4802, + "rewards/accuracies": 0.7734375, + "rewards/chosen": 0.76171875, + "rewards/margins": 1.8125, + "rewards/rejected": -1.0546875, + "step": 19 + }, + { + "epoch": 0.024539877300613498, + "grad_norm": 42.63506109619643, + "learning_rate": 1.2000000000000002e-06, + "logits/chosen": -0.263671875, + "logits/rejected": -0.439453125, + "logps/chosen": -456.0, + "logps/rejected": -324.0, + "loss": 0.3801, + "rewards/accuracies": 0.8125, + "rewards/chosen": 1.4375, + "rewards/margins": 3.140625, + "rewards/rejected": -1.703125, + "step": 20 + }, + { + "epoch": 0.025766871165644172, + "grad_norm": 34.456600308133275, + "learning_rate": 1.26e-06, + "logits/chosen": -0.1328125, + "logits/rejected": -0.3203125, + "logps/chosen": -404.0, + "logps/rejected": -272.0, + "loss": 0.3497, + "rewards/accuracies": 0.8515625, + "rewards/chosen": 1.1015625, + "rewards/margins": 2.640625, + "rewards/rejected": -1.53125, + "step": 21 + }, + { + "epoch": 0.026993865030674847, + "grad_norm": 44.983878720566445, + "learning_rate": 1.32e-06, + "logits/chosen": -0.1298828125, + "logits/rejected": -0.271484375, + "logps/chosen": -374.0, + "logps/rejected": -306.0, + "loss": 0.4156, + "rewards/accuracies": 0.7890625, + "rewards/chosen": 1.0546875, + "rewards/margins": 2.84375, + "rewards/rejected": -1.7890625, + "step": 22 + }, + { + "epoch": 0.02822085889570552, + "grad_norm": 56.72818428668396, + "learning_rate": 1.3800000000000001e-06, + "logits/chosen": -0.251953125, + "logits/rejected": -0.314453125, + "logps/chosen": -370.0, + "logps/rejected": -318.0, + "loss": 0.4566, + "rewards/accuracies": 0.7890625, + "rewards/chosen": 1.46875, + "rewards/margins": 3.1875, + "rewards/rejected": -1.7265625, + "step": 23 + }, + { + "epoch": 0.029447852760736196, + "grad_norm": 55.662509241103734, + "learning_rate": 1.44e-06, + "logits/chosen": -0.1884765625, + "logits/rejected": -0.375, + "logps/chosen": -376.0, + "logps/rejected": -288.0, + "loss": 0.4763, + "rewards/accuracies": 0.7421875, + "rewards/chosen": 0.64453125, + "rewards/margins": 2.265625, + "rewards/rejected": -1.625, + "step": 24 + }, + { + "epoch": 0.03067484662576687, + "grad_norm": 35.38065334196675, + "learning_rate": 1.5e-06, + "logits/chosen": -0.2578125, + "logits/rejected": -0.4453125, + "logps/chosen": -430.0, + "logps/rejected": -312.0, + "loss": 0.3378, + "rewards/accuracies": 0.8984375, + "rewards/chosen": 1.2890625, + "rewards/margins": 3.515625, + "rewards/rejected": -2.21875, + "step": 25 + }, + { + "epoch": 0.03190184049079755, + "grad_norm": 42.887450752414956, + "learning_rate": 1.4993801652892562e-06, + "logits/chosen": -0.15625, + "logits/rejected": -0.31640625, + "logps/chosen": -380.0, + "logps/rejected": -288.0, + "loss": 0.4058, + "rewards/accuracies": 0.8046875, + "rewards/chosen": 1.640625, + "rewards/margins": 3.59375, + "rewards/rejected": -1.9609375, + "step": 26 + }, + { + "epoch": 0.033128834355828224, + "grad_norm": 37.07120513953294, + "learning_rate": 1.4987603305785125e-06, + "logits/chosen": -0.095703125, + "logits/rejected": -0.2734375, + "logps/chosen": -354.0, + "logps/rejected": -292.0, + "loss": 0.3683, + "rewards/accuracies": 0.765625, + "rewards/chosen": 1.7578125, + "rewards/margins": 4.09375, + "rewards/rejected": -2.34375, + "step": 27 + }, + { + "epoch": 0.0343558282208589, + "grad_norm": 35.87136073029194, + "learning_rate": 1.4981404958677687e-06, + "logits/chosen": -0.036376953125, + "logits/rejected": -0.1484375, + "logps/chosen": -374.0, + "logps/rejected": -274.0, + "loss": 0.3615, + "rewards/accuracies": 0.8125, + "rewards/chosen": 1.6171875, + "rewards/margins": 3.421875, + "rewards/rejected": -1.796875, + "step": 28 + }, + { + "epoch": 0.03558282208588957, + "grad_norm": 38.17662218776474, + "learning_rate": 1.4975206611570248e-06, + "logits/chosen": -0.263671875, + "logits/rejected": -0.4140625, + "logps/chosen": -414.0, + "logps/rejected": -334.0, + "loss": 0.3367, + "rewards/accuracies": 0.8046875, + "rewards/chosen": 1.765625, + "rewards/margins": 4.0625, + "rewards/rejected": -2.296875, + "step": 29 + }, + { + "epoch": 0.03680981595092025, + "grad_norm": 39.994932388621756, + "learning_rate": 1.496900826446281e-06, + "logits/chosen": -0.1337890625, + "logits/rejected": -0.248046875, + "logps/chosen": -332.0, + "logps/rejected": -278.0, + "loss": 0.3648, + "rewards/accuracies": 0.8203125, + "rewards/chosen": 1.203125, + "rewards/margins": 3.375, + "rewards/rejected": -2.171875, + "step": 30 + }, + { + "epoch": 0.03803680981595092, + "grad_norm": 42.39396872212231, + "learning_rate": 1.496280991735537e-06, + "logits/chosen": -0.10986328125, + "logits/rejected": -0.2275390625, + "logps/chosen": -378.0, + "logps/rejected": -286.0, + "loss": 0.364, + "rewards/accuracies": 0.7890625, + "rewards/chosen": 1.9765625, + "rewards/margins": 4.25, + "rewards/rejected": -2.265625, + "step": 31 + }, + { + "epoch": 0.0392638036809816, + "grad_norm": 43.84495796047598, + "learning_rate": 1.4956611570247935e-06, + "logits/chosen": -0.15625, + "logits/rejected": -0.333984375, + "logps/chosen": -388.0, + "logps/rejected": -298.0, + "loss": 0.3861, + "rewards/accuracies": 0.78125, + "rewards/chosen": 1.71875, + "rewards/margins": 4.0, + "rewards/rejected": -2.28125, + "step": 32 + }, + { + "epoch": 0.04049079754601227, + "grad_norm": 38.440307864705964, + "learning_rate": 1.4950413223140496e-06, + "logits/chosen": -0.04443359375, + "logits/rejected": -0.177734375, + "logps/chosen": -328.0, + "logps/rejected": -248.0, + "loss": 0.3456, + "rewards/accuracies": 0.8515625, + "rewards/chosen": 1.078125, + "rewards/margins": 3.5625, + "rewards/rejected": -2.484375, + "step": 33 + }, + { + "epoch": 0.04171779141104295, + "grad_norm": 42.669938783190766, + "learning_rate": 1.4944214876033057e-06, + "logits/chosen": -0.1796875, + "logits/rejected": -0.296875, + "logps/chosen": -384.0, + "logps/rejected": -316.0, + "loss": 0.3991, + "rewards/accuracies": 0.7890625, + "rewards/chosen": 1.5859375, + "rewards/margins": 4.40625, + "rewards/rejected": -2.8125, + "step": 34 + }, + { + "epoch": 0.04294478527607362, + "grad_norm": 39.82238013261856, + "learning_rate": 1.493801652892562e-06, + "logits/chosen": -0.0810546875, + "logits/rejected": -0.2197265625, + "logps/chosen": -356.0, + "logps/rejected": -278.0, + "loss": 0.3621, + "rewards/accuracies": 0.828125, + "rewards/chosen": 1.1796875, + "rewards/margins": 4.125, + "rewards/rejected": -2.9375, + "step": 35 + }, + { + "epoch": 0.044171779141104296, + "grad_norm": 34.79756258065961, + "learning_rate": 1.4931818181818182e-06, + "logits/chosen": -0.1376953125, + "logits/rejected": -0.20703125, + "logps/chosen": -398.0, + "logps/rejected": -310.0, + "loss": 0.3113, + "rewards/accuracies": 0.8515625, + "rewards/chosen": 1.6640625, + "rewards/margins": 4.78125, + "rewards/rejected": -3.125, + "step": 36 + }, + { + "epoch": 0.04539877300613497, + "grad_norm": 40.21812818356396, + "learning_rate": 1.4925619834710746e-06, + "logits/chosen": -0.091796875, + "logits/rejected": -0.1943359375, + "logps/chosen": -384.0, + "logps/rejected": -322.0, + "loss": 0.3512, + "rewards/accuracies": 0.796875, + "rewards/chosen": 2.3125, + "rewards/margins": 4.625, + "rewards/rejected": -2.328125, + "step": 37 + }, + { + "epoch": 0.046625766871165646, + "grad_norm": 38.815201724599156, + "learning_rate": 1.4919421487603307e-06, + "logits/chosen": -0.032470703125, + "logits/rejected": -0.1484375, + "logps/chosen": -390.0, + "logps/rejected": -312.0, + "loss": 0.36, + "rewards/accuracies": 0.828125, + "rewards/chosen": 2.546875, + "rewards/margins": 5.09375, + "rewards/rejected": -2.546875, + "step": 38 + }, + { + "epoch": 0.04785276073619632, + "grad_norm": 36.01668225984001, + "learning_rate": 1.4913223140495867e-06, + "logits/chosen": -0.142578125, + "logits/rejected": -0.2890625, + "logps/chosen": -418.0, + "logps/rejected": -318.0, + "loss": 0.3347, + "rewards/accuracies": 0.8359375, + "rewards/chosen": 2.296875, + "rewards/margins": 5.21875, + "rewards/rejected": -2.921875, + "step": 39 + }, + { + "epoch": 0.049079754601226995, + "grad_norm": 30.447104323671855, + "learning_rate": 1.490702479338843e-06, + "logits/chosen": -0.0712890625, + "logits/rejected": -0.1865234375, + "logps/chosen": -370.0, + "logps/rejected": -296.0, + "loss": 0.2729, + "rewards/accuracies": 0.8203125, + "rewards/chosen": 2.09375, + "rewards/margins": 4.90625, + "rewards/rejected": -2.796875, + "step": 40 + }, + { + "epoch": 0.05030674846625767, + "grad_norm": 36.96390592719149, + "learning_rate": 1.4900826446280991e-06, + "logits/chosen": -0.0751953125, + "logits/rejected": -0.2490234375, + "logps/chosen": -374.0, + "logps/rejected": -286.0, + "loss": 0.3822, + "rewards/accuracies": 0.8046875, + "rewards/chosen": 1.875, + "rewards/margins": 4.5625, + "rewards/rejected": -2.671875, + "step": 41 + }, + { + "epoch": 0.051533742331288344, + "grad_norm": 33.41932525699284, + "learning_rate": 1.4894628099173555e-06, + "logits/chosen": -0.022705078125, + "logits/rejected": -0.1650390625, + "logps/chosen": -360.0, + "logps/rejected": -294.0, + "loss": 0.3014, + "rewards/accuracies": 0.84375, + "rewards/chosen": 1.75, + "rewards/margins": 4.3125, + "rewards/rejected": -2.5625, + "step": 42 + }, + { + "epoch": 0.05276073619631902, + "grad_norm": 38.41247549359973, + "learning_rate": 1.4888429752066116e-06, + "logits/chosen": -0.09033203125, + "logits/rejected": -0.248046875, + "logps/chosen": -412.0, + "logps/rejected": -336.0, + "loss": 0.2954, + "rewards/accuracies": 0.84375, + "rewards/chosen": 2.09375, + "rewards/margins": 5.09375, + "rewards/rejected": -3.0, + "step": 43 + }, + { + "epoch": 0.053987730061349694, + "grad_norm": 43.07116536752478, + "learning_rate": 1.4882231404958678e-06, + "logits/chosen": -0.1484375, + "logits/rejected": -0.2490234375, + "logps/chosen": -384.0, + "logps/rejected": -304.0, + "loss": 0.3938, + "rewards/accuracies": 0.8046875, + "rewards/chosen": 1.578125, + "rewards/margins": 4.53125, + "rewards/rejected": -2.96875, + "step": 44 + }, + { + "epoch": 0.05521472392638037, + "grad_norm": 39.15973652461684, + "learning_rate": 1.4876033057851241e-06, + "logits/chosen": -0.2197265625, + "logits/rejected": -0.294921875, + "logps/chosen": -350.0, + "logps/rejected": -300.0, + "loss": 0.3339, + "rewards/accuracies": 0.796875, + "rewards/chosen": 1.828125, + "rewards/margins": 4.9375, + "rewards/rejected": -3.109375, + "step": 45 + }, + { + "epoch": 0.05644171779141104, + "grad_norm": 37.261369579802675, + "learning_rate": 1.4869834710743803e-06, + "logits/chosen": -0.09716796875, + "logits/rejected": -0.259765625, + "logps/chosen": -416.0, + "logps/rejected": -302.0, + "loss": 0.2861, + "rewards/accuracies": 0.828125, + "rewards/chosen": 2.46875, + "rewards/margins": 5.375, + "rewards/rejected": -2.890625, + "step": 46 + }, + { + "epoch": 0.05766871165644172, + "grad_norm": 54.78531831058221, + "learning_rate": 1.4863636363636364e-06, + "logits/chosen": -0.15625, + "logits/rejected": -0.216796875, + "logps/chosen": -390.0, + "logps/rejected": -372.0, + "loss": 0.5073, + "rewards/accuracies": 0.7109375, + "rewards/chosen": 1.5234375, + "rewards/margins": 4.1875, + "rewards/rejected": -2.6875, + "step": 47 + }, + { + "epoch": 0.05889570552147239, + "grad_norm": 43.69804279742604, + "learning_rate": 1.4857438016528926e-06, + "logits/chosen": -0.1494140625, + "logits/rejected": -0.26171875, + "logps/chosen": -352.0, + "logps/rejected": -288.0, + "loss": 0.4242, + "rewards/accuracies": 0.7734375, + "rewards/chosen": 1.1015625, + "rewards/margins": 3.84375, + "rewards/rejected": -2.734375, + "step": 48 + }, + { + "epoch": 0.06012269938650307, + "grad_norm": 34.89468290879934, + "learning_rate": 1.4851239669421487e-06, + "logits/chosen": -0.146484375, + "logits/rejected": -0.328125, + "logps/chosen": -400.0, + "logps/rejected": -300.0, + "loss": 0.2721, + "rewards/accuracies": 0.859375, + "rewards/chosen": 2.25, + "rewards/margins": 5.46875, + "rewards/rejected": -3.21875, + "step": 49 + }, + { + "epoch": 0.06134969325153374, + "grad_norm": 54.15414768657872, + "learning_rate": 1.484504132231405e-06, + "logits/chosen": -0.14453125, + "logits/rejected": -0.296875, + "logps/chosen": -404.0, + "logps/rejected": -324.0, + "loss": 0.4387, + "rewards/accuracies": 0.765625, + "rewards/chosen": 1.515625, + "rewards/margins": 4.6875, + "rewards/rejected": -3.171875, + "step": 50 + }, + { + "epoch": 0.06257668711656442, + "grad_norm": 34.25072100129564, + "learning_rate": 1.4838842975206612e-06, + "logits/chosen": -0.08251953125, + "logits/rejected": -0.2177734375, + "logps/chosen": -440.0, + "logps/rejected": -364.0, + "loss": 0.2872, + "rewards/accuracies": 0.8203125, + "rewards/chosen": 1.984375, + "rewards/margins": 5.09375, + "rewards/rejected": -3.125, + "step": 51 + }, + { + "epoch": 0.0638036809815951, + "grad_norm": 38.20166040651749, + "learning_rate": 1.4832644628099173e-06, + "logits/chosen": -0.10693359375, + "logits/rejected": -0.265625, + "logps/chosen": -396.0, + "logps/rejected": -338.0, + "loss": 0.3424, + "rewards/accuracies": 0.78125, + "rewards/chosen": 1.65625, + "rewards/margins": 4.84375, + "rewards/rejected": -3.171875, + "step": 52 + }, + { + "epoch": 0.06503067484662577, + "grad_norm": 37.089420431960214, + "learning_rate": 1.4826446280991737e-06, + "logits/chosen": -0.055908203125, + "logits/rejected": -0.189453125, + "logps/chosen": -398.0, + "logps/rejected": -326.0, + "loss": 0.3026, + "rewards/accuracies": 0.8359375, + "rewards/chosen": 1.953125, + "rewards/margins": 4.71875, + "rewards/rejected": -2.75, + "step": 53 + }, + { + "epoch": 0.06625766871165645, + "grad_norm": 35.67709547242796, + "learning_rate": 1.4820247933884298e-06, + "logits/chosen": 0.045654296875, + "logits/rejected": -0.1162109375, + "logps/chosen": -336.0, + "logps/rejected": -266.0, + "loss": 0.2813, + "rewards/accuracies": 0.8671875, + "rewards/chosen": 1.3046875, + "rewards/margins": 4.625, + "rewards/rejected": -3.296875, + "step": 54 + }, + { + "epoch": 0.06748466257668712, + "grad_norm": 40.58362122970572, + "learning_rate": 1.481404958677686e-06, + "logits/chosen": -0.068359375, + "logits/rejected": -0.2060546875, + "logps/chosen": -376.0, + "logps/rejected": -302.0, + "loss": 0.3515, + "rewards/accuracies": 0.828125, + "rewards/chosen": 1.8046875, + "rewards/margins": 5.03125, + "rewards/rejected": -3.234375, + "step": 55 + }, + { + "epoch": 0.0687116564417178, + "grad_norm": 43.576103099929334, + "learning_rate": 1.4807851239669423e-06, + "logits/chosen": -0.034912109375, + "logits/rejected": -0.1787109375, + "logps/chosen": -384.0, + "logps/rejected": -334.0, + "loss": 0.4139, + "rewards/accuracies": 0.78125, + "rewards/chosen": 2.015625, + "rewards/margins": 4.8125, + "rewards/rejected": -2.796875, + "step": 56 + }, + { + "epoch": 0.06993865030674846, + "grad_norm": 51.562196128944976, + "learning_rate": 1.4801652892561983e-06, + "logits/chosen": -0.042724609375, + "logits/rejected": -0.0966796875, + "logps/chosen": -352.0, + "logps/rejected": -326.0, + "loss": 0.4667, + "rewards/accuracies": 0.78125, + "rewards/chosen": 1.6015625, + "rewards/margins": 4.59375, + "rewards/rejected": -2.984375, + "step": 57 + }, + { + "epoch": 0.07116564417177915, + "grad_norm": 49.38086728368548, + "learning_rate": 1.4795454545454546e-06, + "logits/chosen": -0.046630859375, + "logits/rejected": -0.10009765625, + "logps/chosen": -376.0, + "logps/rejected": -334.0, + "loss": 0.3869, + "rewards/accuracies": 0.765625, + "rewards/chosen": 1.640625, + "rewards/margins": 4.875, + "rewards/rejected": -3.234375, + "step": 58 + }, + { + "epoch": 0.07239263803680981, + "grad_norm": 43.27222220894176, + "learning_rate": 1.4789256198347107e-06, + "logits/chosen": 0.0240478515625, + "logits/rejected": -0.051025390625, + "logps/chosen": -378.0, + "logps/rejected": -302.0, + "loss": 0.3983, + "rewards/accuracies": 0.765625, + "rewards/chosen": 1.3125, + "rewards/margins": 4.25, + "rewards/rejected": -2.953125, + "step": 59 + }, + { + "epoch": 0.0736196319018405, + "grad_norm": 39.673404465674906, + "learning_rate": 1.4783057851239669e-06, + "logits/chosen": -0.049072265625, + "logits/rejected": -0.1669921875, + "logps/chosen": -386.0, + "logps/rejected": -294.0, + "loss": 0.3927, + "rewards/accuracies": 0.8046875, + "rewards/chosen": 0.99609375, + "rewards/margins": 4.1875, + "rewards/rejected": -3.1875, + "step": 60 + }, + { + "epoch": 0.07484662576687116, + "grad_norm": 40.98812217259802, + "learning_rate": 1.4776859504132232e-06, + "logits/chosen": -0.064453125, + "logits/rejected": -0.173828125, + "logps/chosen": -418.0, + "logps/rejected": -328.0, + "loss": 0.3496, + "rewards/accuracies": 0.8203125, + "rewards/chosen": 1.890625, + "rewards/margins": 5.65625, + "rewards/rejected": -3.765625, + "step": 61 + }, + { + "epoch": 0.07607361963190185, + "grad_norm": 40.09146702222396, + "learning_rate": 1.4770661157024794e-06, + "logits/chosen": -0.051513671875, + "logits/rejected": -0.2275390625, + "logps/chosen": -426.0, + "logps/rejected": -316.0, + "loss": 0.3416, + "rewards/accuracies": 0.7890625, + "rewards/chosen": 2.546875, + "rewards/margins": 6.1875, + "rewards/rejected": -3.640625, + "step": 62 + }, + { + "epoch": 0.07730061349693251, + "grad_norm": 46.30805404508898, + "learning_rate": 1.4764462809917355e-06, + "logits/chosen": 0.031982421875, + "logits/rejected": -0.11767578125, + "logps/chosen": -418.0, + "logps/rejected": -314.0, + "loss": 0.3323, + "rewards/accuracies": 0.8203125, + "rewards/chosen": 1.515625, + "rewards/margins": 4.84375, + "rewards/rejected": -3.328125, + "step": 63 + }, + { + "epoch": 0.0785276073619632, + "grad_norm": 47.96974437982722, + "learning_rate": 1.4758264462809919e-06, + "logits/chosen": -0.123046875, + "logits/rejected": -0.27734375, + "logps/chosen": -402.0, + "logps/rejected": -336.0, + "loss": 0.4046, + "rewards/accuracies": 0.78125, + "rewards/chosen": 2.046875, + "rewards/margins": 5.46875, + "rewards/rejected": -3.4375, + "step": 64 + }, + { + "epoch": 0.07975460122699386, + "grad_norm": 35.57381547040094, + "learning_rate": 1.475206611570248e-06, + "logits/chosen": -0.18359375, + "logits/rejected": -0.25390625, + "logps/chosen": -390.0, + "logps/rejected": -310.0, + "loss": 0.2899, + "rewards/accuracies": 0.84375, + "rewards/chosen": 2.046875, + "rewards/margins": 5.53125, + "rewards/rejected": -3.484375, + "step": 65 + }, + { + "epoch": 0.08098159509202454, + "grad_norm": 33.414611619044024, + "learning_rate": 1.4745867768595042e-06, + "logits/chosen": -0.162109375, + "logits/rejected": -0.287109375, + "logps/chosen": -314.0, + "logps/rejected": -270.0, + "loss": 0.2704, + "rewards/accuracies": 0.8515625, + "rewards/chosen": 2.328125, + "rewards/margins": 6.125, + "rewards/rejected": -3.78125, + "step": 66 + }, + { + "epoch": 0.08220858895705521, + "grad_norm": 42.668537708095855, + "learning_rate": 1.4739669421487603e-06, + "logits/chosen": -0.068359375, + "logits/rejected": -0.1767578125, + "logps/chosen": -332.0, + "logps/rejected": -302.0, + "loss": 0.3919, + "rewards/accuracies": 0.8046875, + "rewards/chosen": 1.2578125, + "rewards/margins": 4.84375, + "rewards/rejected": -3.578125, + "step": 67 + }, + { + "epoch": 0.0834355828220859, + "grad_norm": 40.70660622767979, + "learning_rate": 1.4733471074380164e-06, + "logits/chosen": -0.1455078125, + "logits/rejected": -0.193359375, + "logps/chosen": -358.0, + "logps/rejected": -338.0, + "loss": 0.3949, + "rewards/accuracies": 0.7890625, + "rewards/chosen": 1.625, + "rewards/margins": 4.96875, + "rewards/rejected": -3.359375, + "step": 68 + }, + { + "epoch": 0.08466257668711656, + "grad_norm": 42.29365611407333, + "learning_rate": 1.4727272727272728e-06, + "logits/chosen": -0.2216796875, + "logits/rejected": -0.36328125, + "logps/chosen": -406.0, + "logps/rejected": -290.0, + "loss": 0.3163, + "rewards/accuracies": 0.8515625, + "rewards/chosen": 2.484375, + "rewards/margins": 6.0, + "rewards/rejected": -3.515625, + "step": 69 + }, + { + "epoch": 0.08588957055214724, + "grad_norm": 45.16433653526153, + "learning_rate": 1.472107438016529e-06, + "logits/chosen": -0.08447265625, + "logits/rejected": -0.267578125, + "logps/chosen": -406.0, + "logps/rejected": -346.0, + "loss": 0.3817, + "rewards/accuracies": 0.7890625, + "rewards/chosen": 2.515625, + "rewards/margins": 5.6875, + "rewards/rejected": -3.15625, + "step": 70 + }, + { + "epoch": 0.08711656441717791, + "grad_norm": 37.819952615726415, + "learning_rate": 1.4714876033057853e-06, + "logits/chosen": -0.0341796875, + "logits/rejected": -0.1005859375, + "logps/chosen": -344.0, + "logps/rejected": -286.0, + "loss": 0.3666, + "rewards/accuracies": 0.828125, + "rewards/chosen": 1.46875, + "rewards/margins": 4.625, + "rewards/rejected": -3.15625, + "step": 71 + }, + { + "epoch": 0.08834355828220859, + "grad_norm": 41.96140458025261, + "learning_rate": 1.4708677685950414e-06, + "logits/chosen": -0.19921875, + "logits/rejected": -0.33203125, + "logps/chosen": -428.0, + "logps/rejected": -354.0, + "loss": 0.3187, + "rewards/accuracies": 0.8359375, + "rewards/chosen": 2.125, + "rewards/margins": 5.1875, + "rewards/rejected": -3.0625, + "step": 72 + }, + { + "epoch": 0.08957055214723926, + "grad_norm": 45.003627256411235, + "learning_rate": 1.4702479338842976e-06, + "logits/chosen": -0.14453125, + "logits/rejected": -0.220703125, + "logps/chosen": -346.0, + "logps/rejected": -308.0, + "loss": 0.4099, + "rewards/accuracies": 0.78125, + "rewards/chosen": 1.6171875, + "rewards/margins": 5.25, + "rewards/rejected": -3.625, + "step": 73 + }, + { + "epoch": 0.09079754601226994, + "grad_norm": 37.795821367857755, + "learning_rate": 1.4696280991735537e-06, + "logits/chosen": -0.1435546875, + "logits/rejected": -0.212890625, + "logps/chosen": -352.0, + "logps/rejected": -338.0, + "loss": 0.2989, + "rewards/accuracies": 0.8515625, + "rewards/chosen": 1.4296875, + "rewards/margins": 5.0625, + "rewards/rejected": -3.625, + "step": 74 + }, + { + "epoch": 0.09202453987730061, + "grad_norm": 38.6325533255864, + "learning_rate": 1.4690082644628099e-06, + "logits/chosen": -0.1083984375, + "logits/rejected": -0.2578125, + "logps/chosen": -376.0, + "logps/rejected": -326.0, + "loss": 0.3138, + "rewards/accuracies": 0.8125, + "rewards/chosen": 2.265625, + "rewards/margins": 5.9375, + "rewards/rejected": -3.671875, + "step": 75 + }, + { + "epoch": 0.09325153374233129, + "grad_norm": 41.500214372239896, + "learning_rate": 1.4683884297520662e-06, + "logits/chosen": -0.11376953125, + "logits/rejected": -0.283203125, + "logps/chosen": -394.0, + "logps/rejected": -296.0, + "loss": 0.3196, + "rewards/accuracies": 0.8203125, + "rewards/chosen": 2.28125, + "rewards/margins": 5.625, + "rewards/rejected": -3.359375, + "step": 76 + }, + { + "epoch": 0.09447852760736196, + "grad_norm": 37.661301568651126, + "learning_rate": 1.4677685950413223e-06, + "logits/chosen": -0.1337890625, + "logits/rejected": -0.267578125, + "logps/chosen": -350.0, + "logps/rejected": -308.0, + "loss": 0.3298, + "rewards/accuracies": 0.8125, + "rewards/chosen": 1.4453125, + "rewards/margins": 5.5, + "rewards/rejected": -4.0625, + "step": 77 + }, + { + "epoch": 0.09570552147239264, + "grad_norm": 37.401749661903814, + "learning_rate": 1.4671487603305785e-06, + "logits/chosen": -0.08837890625, + "logits/rejected": -0.2412109375, + "logps/chosen": -414.0, + "logps/rejected": -324.0, + "loss": 0.3295, + "rewards/accuracies": 0.8125, + "rewards/chosen": 2.015625, + "rewards/margins": 5.90625, + "rewards/rejected": -3.90625, + "step": 78 + }, + { + "epoch": 0.09693251533742331, + "grad_norm": 41.34926796893681, + "learning_rate": 1.4665289256198348e-06, + "logits/chosen": -0.078125, + "logits/rejected": -0.2109375, + "logps/chosen": -354.0, + "logps/rejected": -300.0, + "loss": 0.3855, + "rewards/accuracies": 0.8046875, + "rewards/chosen": 1.25, + "rewards/margins": 5.53125, + "rewards/rejected": -4.28125, + "step": 79 + }, + { + "epoch": 0.09815950920245399, + "grad_norm": 40.52807889159641, + "learning_rate": 1.465909090909091e-06, + "logits/chosen": -0.1396484375, + "logits/rejected": -0.33984375, + "logps/chosen": -404.0, + "logps/rejected": -340.0, + "loss": 0.3387, + "rewards/accuracies": 0.8515625, + "rewards/chosen": 2.484375, + "rewards/margins": 7.21875, + "rewards/rejected": -4.75, + "step": 80 + }, + { + "epoch": 0.09938650306748466, + "grad_norm": 35.605812619698284, + "learning_rate": 1.4652892561983471e-06, + "logits/chosen": -0.208984375, + "logits/rejected": -0.365234375, + "logps/chosen": -436.0, + "logps/rejected": -356.0, + "loss": 0.3191, + "rewards/accuracies": 0.828125, + "rewards/chosen": 1.8828125, + "rewards/margins": 6.65625, + "rewards/rejected": -4.78125, + "step": 81 + }, + { + "epoch": 0.10061349693251534, + "grad_norm": 47.4617133344908, + "learning_rate": 1.4646694214876035e-06, + "logits/chosen": -0.1171875, + "logits/rejected": -0.228515625, + "logps/chosen": -430.0, + "logps/rejected": -348.0, + "loss": 0.3954, + "rewards/accuracies": 0.8203125, + "rewards/chosen": 1.03125, + "rewards/margins": 5.875, + "rewards/rejected": -4.84375, + "step": 82 + }, + { + "epoch": 0.10184049079754601, + "grad_norm": 42.72902928212931, + "learning_rate": 1.4640495867768594e-06, + "logits/chosen": -0.1025390625, + "logits/rejected": -0.255859375, + "logps/chosen": -352.0, + "logps/rejected": -320.0, + "loss": 0.3735, + "rewards/accuracies": 0.8125, + "rewards/chosen": 1.109375, + "rewards/margins": 5.84375, + "rewards/rejected": -4.71875, + "step": 83 + }, + { + "epoch": 0.10306748466257669, + "grad_norm": 59.351571106495314, + "learning_rate": 1.4634297520661158e-06, + "logits/chosen": -0.0869140625, + "logits/rejected": -0.310546875, + "logps/chosen": -448.0, + "logps/rejected": -360.0, + "loss": 0.2887, + "rewards/accuracies": 0.859375, + "rewards/chosen": 1.75, + "rewards/margins": 6.8125, + "rewards/rejected": -5.0625, + "step": 84 + }, + { + "epoch": 0.10429447852760736, + "grad_norm": 40.67504361329859, + "learning_rate": 1.462809917355372e-06, + "logits/chosen": -0.189453125, + "logits/rejected": -0.3359375, + "logps/chosen": -376.0, + "logps/rejected": -338.0, + "loss": 0.2933, + "rewards/accuracies": 0.875, + "rewards/chosen": 1.03125, + "rewards/margins": 6.03125, + "rewards/rejected": -4.96875, + "step": 85 + }, + { + "epoch": 0.10552147239263804, + "grad_norm": 44.45491286199163, + "learning_rate": 1.462190082644628e-06, + "logits/chosen": -0.12109375, + "logits/rejected": -0.224609375, + "logps/chosen": -424.0, + "logps/rejected": -314.0, + "loss": 0.4003, + "rewards/accuracies": 0.78125, + "rewards/chosen": 1.625, + "rewards/margins": 5.875, + "rewards/rejected": -4.25, + "step": 86 + }, + { + "epoch": 0.1067484662576687, + "grad_norm": 37.230987939956755, + "learning_rate": 1.4615702479338844e-06, + "logits/chosen": -0.1728515625, + "logits/rejected": -0.3046875, + "logps/chosen": -354.0, + "logps/rejected": -274.0, + "loss": 0.3433, + "rewards/accuracies": 0.8203125, + "rewards/chosen": 0.6484375, + "rewards/margins": 5.65625, + "rewards/rejected": -5.0, + "step": 87 + }, + { + "epoch": 0.10797546012269939, + "grad_norm": 35.435756690174514, + "learning_rate": 1.4609504132231405e-06, + "logits/chosen": -0.142578125, + "logits/rejected": -0.302734375, + "logps/chosen": -362.0, + "logps/rejected": -286.0, + "loss": 0.2796, + "rewards/accuracies": 0.859375, + "rewards/chosen": 1.2890625, + "rewards/margins": 5.78125, + "rewards/rejected": -4.5, + "step": 88 + }, + { + "epoch": 0.10920245398773006, + "grad_norm": 39.218068339324205, + "learning_rate": 1.4603305785123967e-06, + "logits/chosen": -0.1708984375, + "logits/rejected": -0.2734375, + "logps/chosen": -392.0, + "logps/rejected": -340.0, + "loss": 0.3661, + "rewards/accuracies": 0.8046875, + "rewards/chosen": 1.609375, + "rewards/margins": 5.875, + "rewards/rejected": -4.28125, + "step": 89 + }, + { + "epoch": 0.11042944785276074, + "grad_norm": 35.86351693081174, + "learning_rate": 1.459710743801653e-06, + "logits/chosen": -0.2021484375, + "logits/rejected": -0.375, + "logps/chosen": -378.0, + "logps/rejected": -334.0, + "loss": 0.3305, + "rewards/accuracies": 0.828125, + "rewards/chosen": 1.0625, + "rewards/margins": 5.6875, + "rewards/rejected": -4.625, + "step": 90 + }, + { + "epoch": 0.1116564417177914, + "grad_norm": 39.260039435042906, + "learning_rate": 1.4590909090909092e-06, + "logits/chosen": -0.028076171875, + "logits/rejected": -0.28125, + "logps/chosen": -392.0, + "logps/rejected": -308.0, + "loss": 0.3104, + "rewards/accuracies": 0.8203125, + "rewards/chosen": 1.5234375, + "rewards/margins": 5.9375, + "rewards/rejected": -4.40625, + "step": 91 + }, + { + "epoch": 0.11288343558282209, + "grad_norm": 35.99771043781269, + "learning_rate": 1.4584710743801653e-06, + "logits/chosen": -0.1259765625, + "logits/rejected": -0.236328125, + "logps/chosen": -390.0, + "logps/rejected": -316.0, + "loss": 0.304, + "rewards/accuracies": 0.84375, + "rewards/chosen": 1.484375, + "rewards/margins": 6.3125, + "rewards/rejected": -4.8125, + "step": 92 + }, + { + "epoch": 0.11411042944785275, + "grad_norm": 38.41867053252873, + "learning_rate": 1.4578512396694215e-06, + "logits/chosen": -0.15234375, + "logits/rejected": -0.36328125, + "logps/chosen": -400.0, + "logps/rejected": -310.0, + "loss": 0.3311, + "rewards/accuracies": 0.8515625, + "rewards/chosen": 2.328125, + "rewards/margins": 6.625, + "rewards/rejected": -4.3125, + "step": 93 + }, + { + "epoch": 0.11533742331288344, + "grad_norm": 37.099757310606805, + "learning_rate": 1.4572314049586776e-06, + "logits/chosen": -0.171875, + "logits/rejected": -0.390625, + "logps/chosen": -426.0, + "logps/rejected": -354.0, + "loss": 0.2709, + "rewards/accuracies": 0.84375, + "rewards/chosen": 2.765625, + "rewards/margins": 7.09375, + "rewards/rejected": -4.3125, + "step": 94 + }, + { + "epoch": 0.1165644171779141, + "grad_norm": 36.042450420967036, + "learning_rate": 1.456611570247934e-06, + "logits/chosen": -0.2490234375, + "logits/rejected": -0.392578125, + "logps/chosen": -378.0, + "logps/rejected": -320.0, + "loss": 0.3138, + "rewards/accuracies": 0.859375, + "rewards/chosen": 2.0625, + "rewards/margins": 6.6875, + "rewards/rejected": -4.625, + "step": 95 + }, + { + "epoch": 0.11779141104294479, + "grad_norm": 37.51605196372948, + "learning_rate": 1.45599173553719e-06, + "logits/chosen": -0.2392578125, + "logits/rejected": -0.400390625, + "logps/chosen": -366.0, + "logps/rejected": -320.0, + "loss": 0.2944, + "rewards/accuracies": 0.84375, + "rewards/chosen": 2.625, + "rewards/margins": 7.125, + "rewards/rejected": -4.5, + "step": 96 + }, + { + "epoch": 0.11901840490797547, + "grad_norm": 47.87658843319023, + "learning_rate": 1.4553719008264462e-06, + "logits/chosen": -0.234375, + "logits/rejected": -0.373046875, + "logps/chosen": -422.0, + "logps/rejected": -362.0, + "loss": 0.4367, + "rewards/accuracies": 0.796875, + "rewards/chosen": 2.25, + "rewards/margins": 6.125, + "rewards/rejected": -3.859375, + "step": 97 + }, + { + "epoch": 0.12024539877300613, + "grad_norm": 37.61598969810632, + "learning_rate": 1.4547520661157026e-06, + "logits/chosen": -0.03759765625, + "logits/rejected": -0.236328125, + "logps/chosen": -354.0, + "logps/rejected": -296.0, + "loss": 0.3071, + "rewards/accuracies": 0.84375, + "rewards/chosen": 1.984375, + "rewards/margins": 6.09375, + "rewards/rejected": -4.09375, + "step": 98 + }, + { + "epoch": 0.12147239263803682, + "grad_norm": 35.33546250741774, + "learning_rate": 1.4541322314049587e-06, + "logits/chosen": -0.189453125, + "logits/rejected": -0.287109375, + "logps/chosen": -338.0, + "logps/rejected": -258.0, + "loss": 0.2669, + "rewards/accuracies": 0.8515625, + "rewards/chosen": 1.359375, + "rewards/margins": 5.5, + "rewards/rejected": -4.15625, + "step": 99 + }, + { + "epoch": 0.12269938650306748, + "grad_norm": 32.14334216613253, + "learning_rate": 1.453512396694215e-06, + "logits/chosen": -0.2197265625, + "logits/rejected": -0.404296875, + "logps/chosen": -394.0, + "logps/rejected": -332.0, + "loss": 0.2858, + "rewards/accuracies": 0.8359375, + "rewards/chosen": 1.8359375, + "rewards/margins": 6.46875, + "rewards/rejected": -4.625, + "step": 100 + }, + { + "epoch": 0.12392638036809817, + "grad_norm": 45.60783475518701, + "learning_rate": 1.452892561983471e-06, + "logits/chosen": -0.220703125, + "logits/rejected": -0.400390625, + "logps/chosen": -412.0, + "logps/rejected": -352.0, + "loss": 0.4076, + "rewards/accuracies": 0.796875, + "rewards/chosen": 1.9140625, + "rewards/margins": 6.3125, + "rewards/rejected": -4.375, + "step": 101 + }, + { + "epoch": 0.12515337423312883, + "grad_norm": 39.740396682339004, + "learning_rate": 1.4522727272727272e-06, + "logits/chosen": -0.1865234375, + "logits/rejected": -0.298828125, + "logps/chosen": -392.0, + "logps/rejected": -330.0, + "loss": 0.3534, + "rewards/accuracies": 0.8046875, + "rewards/chosen": 2.078125, + "rewards/margins": 6.0, + "rewards/rejected": -3.9375, + "step": 102 + }, + { + "epoch": 0.1263803680981595, + "grad_norm": 35.16122653572647, + "learning_rate": 1.4516528925619835e-06, + "logits/chosen": -0.17578125, + "logits/rejected": -0.326171875, + "logps/chosen": -350.0, + "logps/rejected": -310.0, + "loss": 0.3153, + "rewards/accuracies": 0.84375, + "rewards/chosen": 1.4765625, + "rewards/margins": 5.75, + "rewards/rejected": -4.28125, + "step": 103 + }, + { + "epoch": 0.1276073619631902, + "grad_norm": 46.63941850097279, + "learning_rate": 1.4510330578512396e-06, + "logits/chosen": -0.091796875, + "logits/rejected": -0.2255859375, + "logps/chosen": -414.0, + "logps/rejected": -348.0, + "loss": 0.381, + "rewards/accuracies": 0.7890625, + "rewards/chosen": 1.6484375, + "rewards/margins": 5.78125, + "rewards/rejected": -4.125, + "step": 104 + }, + { + "epoch": 0.12883435582822086, + "grad_norm": 35.04246201772006, + "learning_rate": 1.450413223140496e-06, + "logits/chosen": -0.181640625, + "logits/rejected": -0.3828125, + "logps/chosen": -374.0, + "logps/rejected": -322.0, + "loss": 0.2864, + "rewards/accuracies": 0.8515625, + "rewards/chosen": 2.03125, + "rewards/margins": 6.8125, + "rewards/rejected": -4.78125, + "step": 105 + }, + { + "epoch": 0.13006134969325153, + "grad_norm": 39.36563074587024, + "learning_rate": 1.4497933884297521e-06, + "logits/chosen": -0.1591796875, + "logits/rejected": -0.3046875, + "logps/chosen": -392.0, + "logps/rejected": -322.0, + "loss": 0.3462, + "rewards/accuracies": 0.7890625, + "rewards/chosen": 1.7890625, + "rewards/margins": 6.03125, + "rewards/rejected": -4.25, + "step": 106 + }, + { + "epoch": 0.1312883435582822, + "grad_norm": 35.07229699022507, + "learning_rate": 1.4491735537190083e-06, + "logits/chosen": -0.1865234375, + "logits/rejected": -0.26953125, + "logps/chosen": -352.0, + "logps/rejected": -312.0, + "loss": 0.2577, + "rewards/accuracies": 0.8671875, + "rewards/chosen": 1.3515625, + "rewards/margins": 5.84375, + "rewards/rejected": -4.5, + "step": 107 + }, + { + "epoch": 0.1325153374233129, + "grad_norm": 28.019513861884207, + "learning_rate": 1.4485537190082646e-06, + "logits/chosen": -0.19921875, + "logits/rejected": -0.408203125, + "logps/chosen": -382.0, + "logps/rejected": -320.0, + "loss": 0.2438, + "rewards/accuracies": 0.8828125, + "rewards/chosen": 1.3046875, + "rewards/margins": 6.5625, + "rewards/rejected": -5.25, + "step": 108 + }, + { + "epoch": 0.13374233128834356, + "grad_norm": 35.40605672013622, + "learning_rate": 1.4479338842975208e-06, + "logits/chosen": -0.18359375, + "logits/rejected": -0.359375, + "logps/chosen": -412.0, + "logps/rejected": -326.0, + "loss": 0.3584, + "rewards/accuracies": 0.8359375, + "rewards/chosen": 2.5625, + "rewards/margins": 7.1875, + "rewards/rejected": -4.625, + "step": 109 + }, + { + "epoch": 0.13496932515337423, + "grad_norm": 38.14439771999889, + "learning_rate": 1.447314049586777e-06, + "logits/chosen": -0.1728515625, + "logits/rejected": -0.310546875, + "logps/chosen": -406.0, + "logps/rejected": -348.0, + "loss": 0.3481, + "rewards/accuracies": 0.8125, + "rewards/chosen": 1.9140625, + "rewards/margins": 6.4375, + "rewards/rejected": -4.53125, + "step": 110 + }, + { + "epoch": 0.1361963190184049, + "grad_norm": 36.779042619503116, + "learning_rate": 1.446694214876033e-06, + "logits/chosen": -0.09716796875, + "logits/rejected": -0.251953125, + "logps/chosen": -334.0, + "logps/rejected": -282.0, + "loss": 0.3363, + "rewards/accuracies": 0.859375, + "rewards/chosen": 1.3671875, + "rewards/margins": 5.4375, + "rewards/rejected": -4.0625, + "step": 111 + }, + { + "epoch": 0.1374233128834356, + "grad_norm": 36.86435500620325, + "learning_rate": 1.4460743801652892e-06, + "logits/chosen": -0.177734375, + "logits/rejected": -0.29296875, + "logps/chosen": -342.0, + "logps/rejected": -292.0, + "loss": 0.343, + "rewards/accuracies": 0.859375, + "rewards/chosen": 1.3046875, + "rewards/margins": 5.75, + "rewards/rejected": -4.4375, + "step": 112 + }, + { + "epoch": 0.13865030674846626, + "grad_norm": 38.139474814179394, + "learning_rate": 1.4454545454545456e-06, + "logits/chosen": -0.0849609375, + "logits/rejected": -0.169921875, + "logps/chosen": -374.0, + "logps/rejected": -312.0, + "loss": 0.3952, + "rewards/accuracies": 0.828125, + "rewards/chosen": 1.859375, + "rewards/margins": 5.75, + "rewards/rejected": -3.90625, + "step": 113 + }, + { + "epoch": 0.13987730061349693, + "grad_norm": 38.81539599706057, + "learning_rate": 1.4448347107438017e-06, + "logits/chosen": -0.1220703125, + "logits/rejected": -0.275390625, + "logps/chosen": -372.0, + "logps/rejected": -316.0, + "loss": 0.3854, + "rewards/accuracies": 0.859375, + "rewards/chosen": 1.5078125, + "rewards/margins": 5.375, + "rewards/rejected": -3.875, + "step": 114 + }, + { + "epoch": 0.1411042944785276, + "grad_norm": 45.65683573404254, + "learning_rate": 1.4442148760330578e-06, + "logits/chosen": -0.1298828125, + "logits/rejected": -0.2490234375, + "logps/chosen": -338.0, + "logps/rejected": -320.0, + "loss": 0.4779, + "rewards/accuracies": 0.75, + "rewards/chosen": 1.046875, + "rewards/margins": 4.71875, + "rewards/rejected": -3.65625, + "step": 115 + }, + { + "epoch": 0.1423312883435583, + "grad_norm": 32.97644733528144, + "learning_rate": 1.4435950413223142e-06, + "logits/chosen": -0.185546875, + "logits/rejected": -0.408203125, + "logps/chosen": -376.0, + "logps/rejected": -324.0, + "loss": 0.2978, + "rewards/accuracies": 0.8359375, + "rewards/chosen": 2.40625, + "rewards/margins": 6.03125, + "rewards/rejected": -3.625, + "step": 116 + }, + { + "epoch": 0.14355828220858896, + "grad_norm": 38.235071091905795, + "learning_rate": 1.4429752066115703e-06, + "logits/chosen": -0.130859375, + "logits/rejected": -0.26953125, + "logps/chosen": -324.0, + "logps/rejected": -270.0, + "loss": 0.4211, + "rewards/accuracies": 0.796875, + "rewards/chosen": 1.3125, + "rewards/margins": 5.53125, + "rewards/rejected": -4.21875, + "step": 117 + }, + { + "epoch": 0.14478527607361963, + "grad_norm": 41.12118912071453, + "learning_rate": 1.4423553719008265e-06, + "logits/chosen": -0.15234375, + "logits/rejected": -0.298828125, + "logps/chosen": -418.0, + "logps/rejected": -360.0, + "loss": 0.3658, + "rewards/accuracies": 0.8046875, + "rewards/chosen": 2.28125, + "rewards/margins": 6.125, + "rewards/rejected": -3.84375, + "step": 118 + }, + { + "epoch": 0.1460122699386503, + "grad_norm": 35.626041219957244, + "learning_rate": 1.4417355371900826e-06, + "logits/chosen": -0.220703125, + "logits/rejected": -0.30859375, + "logps/chosen": -352.0, + "logps/rejected": -324.0, + "loss": 0.3003, + "rewards/accuracies": 0.8125, + "rewards/chosen": 1.6953125, + "rewards/margins": 5.59375, + "rewards/rejected": -3.90625, + "step": 119 + }, + { + "epoch": 0.147239263803681, + "grad_norm": 40.42176001046704, + "learning_rate": 1.4411157024793388e-06, + "logits/chosen": -0.201171875, + "logits/rejected": -0.37890625, + "logps/chosen": -418.0, + "logps/rejected": -354.0, + "loss": 0.3506, + "rewards/accuracies": 0.8359375, + "rewards/chosen": 2.6875, + "rewards/margins": 6.625, + "rewards/rejected": -3.9375, + "step": 120 + }, + { + "epoch": 0.14846625766871166, + "grad_norm": 33.17627892781642, + "learning_rate": 1.4404958677685951e-06, + "logits/chosen": -0.12060546875, + "logits/rejected": -0.27734375, + "logps/chosen": -404.0, + "logps/rejected": -336.0, + "loss": 0.2824, + "rewards/accuracies": 0.8203125, + "rewards/chosen": 2.625, + "rewards/margins": 6.75, + "rewards/rejected": -4.125, + "step": 121 + }, + { + "epoch": 0.14969325153374233, + "grad_norm": 41.361282319998544, + "learning_rate": 1.4398760330578512e-06, + "logits/chosen": -0.130859375, + "logits/rejected": -0.283203125, + "logps/chosen": -378.0, + "logps/rejected": -298.0, + "loss": 0.3454, + "rewards/accuracies": 0.828125, + "rewards/chosen": 1.7421875, + "rewards/margins": 5.75, + "rewards/rejected": -4.03125, + "step": 122 + }, + { + "epoch": 0.150920245398773, + "grad_norm": 39.0679541541933, + "learning_rate": 1.4392561983471074e-06, + "logits/chosen": -0.0191650390625, + "logits/rejected": -0.1884765625, + "logps/chosen": -380.0, + "logps/rejected": -290.0, + "loss": 0.384, + "rewards/accuracies": 0.7890625, + "rewards/chosen": 1.2734375, + "rewards/margins": 5.625, + "rewards/rejected": -4.375, + "step": 123 + }, + { + "epoch": 0.1521472392638037, + "grad_norm": 30.841344702577548, + "learning_rate": 1.4386363636363637e-06, + "logits/chosen": -0.193359375, + "logits/rejected": -0.255859375, + "logps/chosen": -348.0, + "logps/rejected": -294.0, + "loss": 0.2792, + "rewards/accuracies": 0.890625, + "rewards/chosen": 2.203125, + "rewards/margins": 7.0, + "rewards/rejected": -4.8125, + "step": 124 + }, + { + "epoch": 0.15337423312883436, + "grad_norm": 40.57809799655024, + "learning_rate": 1.4380165289256199e-06, + "logits/chosen": -0.259765625, + "logits/rejected": -0.431640625, + "logps/chosen": -434.0, + "logps/rejected": -340.0, + "loss": 0.3257, + "rewards/accuracies": 0.84375, + "rewards/chosen": 1.6015625, + "rewards/margins": 6.625, + "rewards/rejected": -5.0, + "step": 125 + }, + { + "epoch": 0.15460122699386503, + "grad_norm": 49.176910835221456, + "learning_rate": 1.437396694214876e-06, + "logits/chosen": -0.11279296875, + "logits/rejected": -0.30078125, + "logps/chosen": -368.0, + "logps/rejected": -320.0, + "loss": 0.4071, + "rewards/accuracies": 0.8359375, + "rewards/chosen": 0.6328125, + "rewards/margins": 5.59375, + "rewards/rejected": -4.96875, + "step": 126 + }, + { + "epoch": 0.1558282208588957, + "grad_norm": 42.06253723411312, + "learning_rate": 1.4367768595041324e-06, + "logits/chosen": -0.064453125, + "logits/rejected": -0.2431640625, + "logps/chosen": -388.0, + "logps/rejected": -340.0, + "loss": 0.4543, + "rewards/accuracies": 0.7578125, + "rewards/chosen": 1.75, + "rewards/margins": 6.40625, + "rewards/rejected": -4.65625, + "step": 127 + }, + { + "epoch": 0.1570552147239264, + "grad_norm": 49.94695723772631, + "learning_rate": 1.4361570247933883e-06, + "logits/chosen": -0.1044921875, + "logits/rejected": -0.2890625, + "logps/chosen": -374.0, + "logps/rejected": -344.0, + "loss": 0.3851, + "rewards/accuracies": 0.78125, + "rewards/chosen": 1.5625, + "rewards/margins": 6.6875, + "rewards/rejected": -5.09375, + "step": 128 + }, + { + "epoch": 0.15828220858895706, + "grad_norm": 31.69804344527413, + "learning_rate": 1.4355371900826447e-06, + "logits/chosen": -0.00439453125, + "logits/rejected": -0.1279296875, + "logps/chosen": -342.0, + "logps/rejected": -302.0, + "loss": 0.3139, + "rewards/accuracies": 0.8359375, + "rewards/chosen": 0.71484375, + "rewards/margins": 5.5, + "rewards/rejected": -4.78125, + "step": 129 + }, + { + "epoch": 0.15950920245398773, + "grad_norm": 39.02664413781982, + "learning_rate": 1.4349173553719008e-06, + "logits/chosen": -0.11328125, + "logits/rejected": -0.251953125, + "logps/chosen": -388.0, + "logps/rejected": -324.0, + "loss": 0.3918, + "rewards/accuracies": 0.75, + "rewards/chosen": 0.66015625, + "rewards/margins": 6.03125, + "rewards/rejected": -5.375, + "step": 130 + }, + { + "epoch": 0.1607361963190184, + "grad_norm": 33.65136979573152, + "learning_rate": 1.434297520661157e-06, + "logits/chosen": -0.2255859375, + "logits/rejected": -0.361328125, + "logps/chosen": -396.0, + "logps/rejected": -356.0, + "loss": 0.2429, + "rewards/accuracies": 0.84375, + "rewards/chosen": 1.3671875, + "rewards/margins": 6.96875, + "rewards/rejected": -5.59375, + "step": 131 + }, + { + "epoch": 0.1619631901840491, + "grad_norm": 40.70444309722836, + "learning_rate": 1.4336776859504133e-06, + "logits/chosen": -0.1845703125, + "logits/rejected": -0.3671875, + "logps/chosen": -392.0, + "logps/rejected": -336.0, + "loss": 0.3917, + "rewards/accuracies": 0.796875, + "rewards/chosen": 1.1796875, + "rewards/margins": 6.34375, + "rewards/rejected": -5.15625, + "step": 132 + }, + { + "epoch": 0.16319018404907976, + "grad_norm": 38.39829631153826, + "learning_rate": 1.4330578512396694e-06, + "logits/chosen": -0.220703125, + "logits/rejected": -0.375, + "logps/chosen": -388.0, + "logps/rejected": -342.0, + "loss": 0.3065, + "rewards/accuracies": 0.84375, + "rewards/chosen": 0.94921875, + "rewards/margins": 6.40625, + "rewards/rejected": -5.46875, + "step": 133 + }, + { + "epoch": 0.16441717791411042, + "grad_norm": 54.071557291057594, + "learning_rate": 1.4324380165289258e-06, + "logits/chosen": -0.1416015625, + "logits/rejected": -0.34765625, + "logps/chosen": -458.0, + "logps/rejected": -374.0, + "loss": 0.4193, + "rewards/accuracies": 0.796875, + "rewards/chosen": 0.81640625, + "rewards/margins": 5.4375, + "rewards/rejected": -4.625, + "step": 134 + }, + { + "epoch": 0.1656441717791411, + "grad_norm": 42.09400341898127, + "learning_rate": 1.431818181818182e-06, + "logits/chosen": -0.0140380859375, + "logits/rejected": -0.11767578125, + "logps/chosen": -378.0, + "logps/rejected": -330.0, + "loss": 0.4491, + "rewards/accuracies": 0.8046875, + "rewards/chosen": 0.16015625, + "rewards/margins": 5.21875, + "rewards/rejected": -5.0625, + "step": 135 + }, + { + "epoch": 0.1668711656441718, + "grad_norm": 34.790669543589566, + "learning_rate": 1.431198347107438e-06, + "logits/chosen": -0.259765625, + "logits/rejected": -0.3203125, + "logps/chosen": -392.0, + "logps/rejected": -322.0, + "loss": 0.3098, + "rewards/accuracies": 0.828125, + "rewards/chosen": 0.87890625, + "rewards/margins": 6.34375, + "rewards/rejected": -5.46875, + "step": 136 + }, + { + "epoch": 0.16809815950920245, + "grad_norm": 46.051678091055635, + "learning_rate": 1.4305785123966942e-06, + "logits/chosen": -0.020751953125, + "logits/rejected": -0.1669921875, + "logps/chosen": -398.0, + "logps/rejected": -320.0, + "loss": 0.4495, + "rewards/accuracies": 0.7890625, + "rewards/chosen": 0.609375, + "rewards/margins": 5.21875, + "rewards/rejected": -4.625, + "step": 137 + }, + { + "epoch": 0.16932515337423312, + "grad_norm": 39.60892351140069, + "learning_rate": 1.4299586776859504e-06, + "logits/chosen": -0.10546875, + "logits/rejected": -0.2421875, + "logps/chosen": -406.0, + "logps/rejected": -360.0, + "loss": 0.3662, + "rewards/accuracies": 0.796875, + "rewards/chosen": 0.56640625, + "rewards/margins": 5.59375, + "rewards/rejected": -5.0, + "step": 138 + }, + { + "epoch": 0.1705521472392638, + "grad_norm": 44.067268759137896, + "learning_rate": 1.4293388429752067e-06, + "logits/chosen": -0.10546875, + "logits/rejected": -0.2734375, + "logps/chosen": -418.0, + "logps/rejected": -380.0, + "loss": 0.4261, + "rewards/accuracies": 0.78125, + "rewards/chosen": 1.2109375, + "rewards/margins": 6.8125, + "rewards/rejected": -5.59375, + "step": 139 + }, + { + "epoch": 0.17177914110429449, + "grad_norm": 46.60040949853179, + "learning_rate": 1.4287190082644628e-06, + "logits/chosen": -0.07177734375, + "logits/rejected": -0.251953125, + "logps/chosen": -426.0, + "logps/rejected": -376.0, + "loss": 0.4219, + "rewards/accuracies": 0.765625, + "rewards/chosen": 0.9453125, + "rewards/margins": 6.125, + "rewards/rejected": -5.1875, + "step": 140 + }, + { + "epoch": 0.17300613496932515, + "grad_norm": 34.208928830023055, + "learning_rate": 1.428099173553719e-06, + "logits/chosen": -0.05126953125, + "logits/rejected": -0.173828125, + "logps/chosen": -358.0, + "logps/rejected": -338.0, + "loss": 0.3019, + "rewards/accuracies": 0.8515625, + "rewards/chosen": 0.99609375, + "rewards/margins": 6.28125, + "rewards/rejected": -5.28125, + "step": 141 + }, + { + "epoch": 0.17423312883435582, + "grad_norm": 41.96948447879753, + "learning_rate": 1.4274793388429753e-06, + "logits/chosen": -0.0927734375, + "logits/rejected": -0.27734375, + "logps/chosen": -400.0, + "logps/rejected": -354.0, + "loss": 0.4327, + "rewards/accuracies": 0.796875, + "rewards/chosen": 0.98828125, + "rewards/margins": 6.375, + "rewards/rejected": -5.375, + "step": 142 + }, + { + "epoch": 0.1754601226993865, + "grad_norm": 39.736356252739405, + "learning_rate": 1.4268595041322315e-06, + "logits/chosen": -0.12451171875, + "logits/rejected": -0.33984375, + "logps/chosen": -384.0, + "logps/rejected": -328.0, + "loss": 0.3314, + "rewards/accuracies": 0.828125, + "rewards/chosen": 0.64453125, + "rewards/margins": 6.28125, + "rewards/rejected": -5.625, + "step": 143 + }, + { + "epoch": 0.17668711656441718, + "grad_norm": 40.38397393664774, + "learning_rate": 1.4262396694214876e-06, + "logits/chosen": -0.224609375, + "logits/rejected": -0.3984375, + "logps/chosen": -420.0, + "logps/rejected": -356.0, + "loss": 0.3711, + "rewards/accuracies": 0.828125, + "rewards/chosen": 1.5625, + "rewards/margins": 6.75, + "rewards/rejected": -5.1875, + "step": 144 + }, + { + "epoch": 0.17791411042944785, + "grad_norm": 36.78367346174047, + "learning_rate": 1.425619834710744e-06, + "logits/chosen": -0.20703125, + "logits/rejected": -0.349609375, + "logps/chosen": -380.0, + "logps/rejected": -348.0, + "loss": 0.2992, + "rewards/accuracies": 0.859375, + "rewards/chosen": 1.5625, + "rewards/margins": 6.78125, + "rewards/rejected": -5.21875, + "step": 145 + }, + { + "epoch": 0.17914110429447852, + "grad_norm": 30.78038936070514, + "learning_rate": 1.425e-06, + "logits/chosen": -0.2021484375, + "logits/rejected": -0.30859375, + "logps/chosen": -388.0, + "logps/rejected": -334.0, + "loss": 0.2674, + "rewards/accuracies": 0.875, + "rewards/chosen": 0.53125, + "rewards/margins": 6.03125, + "rewards/rejected": -5.5, + "step": 146 + }, + { + "epoch": 0.18036809815950922, + "grad_norm": 33.88526702598482, + "learning_rate": 1.4243801652892563e-06, + "logits/chosen": -0.2001953125, + "logits/rejected": -0.349609375, + "logps/chosen": -388.0, + "logps/rejected": -332.0, + "loss": 0.3151, + "rewards/accuracies": 0.828125, + "rewards/chosen": 1.5625, + "rewards/margins": 6.875, + "rewards/rejected": -5.3125, + "step": 147 + }, + { + "epoch": 0.18159509202453988, + "grad_norm": 49.54475790087749, + "learning_rate": 1.4237603305785124e-06, + "logits/chosen": -0.2197265625, + "logits/rejected": -0.34765625, + "logps/chosen": -416.0, + "logps/rejected": -368.0, + "loss": 0.4146, + "rewards/accuracies": 0.7734375, + "rewards/chosen": 0.89453125, + "rewards/margins": 6.25, + "rewards/rejected": -5.34375, + "step": 148 + }, + { + "epoch": 0.18282208588957055, + "grad_norm": 43.24095049009772, + "learning_rate": 1.4231404958677685e-06, + "logits/chosen": -0.09521484375, + "logits/rejected": -0.244140625, + "logps/chosen": -406.0, + "logps/rejected": -384.0, + "loss": 0.3173, + "rewards/accuracies": 0.859375, + "rewards/chosen": 1.5859375, + "rewards/margins": 6.625, + "rewards/rejected": -5.03125, + "step": 149 + }, + { + "epoch": 0.18404907975460122, + "grad_norm": 37.7299116363035, + "learning_rate": 1.422520661157025e-06, + "logits/chosen": -0.1181640625, + "logits/rejected": -0.291015625, + "logps/chosen": -424.0, + "logps/rejected": -346.0, + "loss": 0.344, + "rewards/accuracies": 0.8359375, + "rewards/chosen": 1.296875, + "rewards/margins": 5.84375, + "rewards/rejected": -4.53125, + "step": 150 + }, + { + "epoch": 0.18527607361963191, + "grad_norm": 36.98087652686202, + "learning_rate": 1.421900826446281e-06, + "logits/chosen": -0.1376953125, + "logits/rejected": -0.263671875, + "logps/chosen": -382.0, + "logps/rejected": -326.0, + "loss": 0.3467, + "rewards/accuracies": 0.8046875, + "rewards/chosen": 1.1796875, + "rewards/margins": 5.8125, + "rewards/rejected": -4.625, + "step": 151 + }, + { + "epoch": 0.18650306748466258, + "grad_norm": 33.96526048608977, + "learning_rate": 1.4212809917355372e-06, + "logits/chosen": -0.2255859375, + "logits/rejected": -0.462890625, + "logps/chosen": -402.0, + "logps/rejected": -316.0, + "loss": 0.322, + "rewards/accuracies": 0.8515625, + "rewards/chosen": 0.83984375, + "rewards/margins": 6.5625, + "rewards/rejected": -5.71875, + "step": 152 + }, + { + "epoch": 0.18773006134969325, + "grad_norm": 34.59925431133825, + "learning_rate": 1.4206611570247935e-06, + "logits/chosen": -0.1875, + "logits/rejected": -0.298828125, + "logps/chosen": -350.0, + "logps/rejected": -320.0, + "loss": 0.3333, + "rewards/accuracies": 0.84375, + "rewards/chosen": 1.390625, + "rewards/margins": 6.28125, + "rewards/rejected": -4.875, + "step": 153 + }, + { + "epoch": 0.18895705521472392, + "grad_norm": 45.111496339674666, + "learning_rate": 1.4200413223140497e-06, + "logits/chosen": -0.060546875, + "logits/rejected": -0.2431640625, + "logps/chosen": -424.0, + "logps/rejected": -340.0, + "loss": 0.3836, + "rewards/accuracies": 0.7734375, + "rewards/chosen": 1.703125, + "rewards/margins": 6.09375, + "rewards/rejected": -4.375, + "step": 154 + }, + { + "epoch": 0.1901840490797546, + "grad_norm": 28.559326947209748, + "learning_rate": 1.4194214876033058e-06, + "logits/chosen": -0.1748046875, + "logits/rejected": -0.41796875, + "logps/chosen": -382.0, + "logps/rejected": -320.0, + "loss": 0.2127, + "rewards/accuracies": 0.90625, + "rewards/chosen": 2.359375, + "rewards/margins": 7.375, + "rewards/rejected": -5.03125, + "step": 155 + }, + { + "epoch": 0.19141104294478528, + "grad_norm": 41.969901464316806, + "learning_rate": 1.418801652892562e-06, + "logits/chosen": -0.06787109375, + "logits/rejected": -0.1748046875, + "logps/chosen": -370.0, + "logps/rejected": -350.0, + "loss": 0.3849, + "rewards/accuracies": 0.8203125, + "rewards/chosen": 1.1171875, + "rewards/margins": 5.59375, + "rewards/rejected": -4.46875, + "step": 156 + }, + { + "epoch": 0.19263803680981595, + "grad_norm": 35.115688822084685, + "learning_rate": 1.418181818181818e-06, + "logits/chosen": -0.05810546875, + "logits/rejected": -0.1796875, + "logps/chosen": -394.0, + "logps/rejected": -310.0, + "loss": 0.3364, + "rewards/accuracies": 0.828125, + "rewards/chosen": 1.2421875, + "rewards/margins": 5.875, + "rewards/rejected": -4.625, + "step": 157 + }, + { + "epoch": 0.19386503067484662, + "grad_norm": 29.216146857412586, + "learning_rate": 1.4175619834710745e-06, + "logits/chosen": -0.2734375, + "logits/rejected": -0.44921875, + "logps/chosen": -388.0, + "logps/rejected": -324.0, + "loss": 0.2283, + "rewards/accuracies": 0.8515625, + "rewards/chosen": 1.390625, + "rewards/margins": 6.59375, + "rewards/rejected": -5.21875, + "step": 158 + }, + { + "epoch": 0.1950920245398773, + "grad_norm": 39.305920966554595, + "learning_rate": 1.4169421487603306e-06, + "logits/chosen": -0.1630859375, + "logits/rejected": -0.28125, + "logps/chosen": -360.0, + "logps/rejected": -328.0, + "loss": 0.3649, + "rewards/accuracies": 0.84375, + "rewards/chosen": 1.359375, + "rewards/margins": 5.71875, + "rewards/rejected": -4.34375, + "step": 159 + }, + { + "epoch": 0.19631901840490798, + "grad_norm": 26.74081618489141, + "learning_rate": 1.416322314049587e-06, + "logits/chosen": -0.047607421875, + "logits/rejected": -0.26953125, + "logps/chosen": -354.0, + "logps/rejected": -310.0, + "loss": 0.2471, + "rewards/accuracies": 0.890625, + "rewards/chosen": 0.94921875, + "rewards/margins": 6.125, + "rewards/rejected": -5.15625, + "step": 160 + }, + { + "epoch": 0.19754601226993865, + "grad_norm": 32.344289601166764, + "learning_rate": 1.415702479338843e-06, + "logits/chosen": -0.2294921875, + "logits/rejected": -0.4296875, + "logps/chosen": -386.0, + "logps/rejected": -302.0, + "loss": 0.2831, + "rewards/accuracies": 0.8515625, + "rewards/chosen": 1.9453125, + "rewards/margins": 6.6875, + "rewards/rejected": -4.71875, + "step": 161 + }, + { + "epoch": 0.19877300613496932, + "grad_norm": 39.80838608230206, + "learning_rate": 1.4150826446280992e-06, + "logits/chosen": -0.1357421875, + "logits/rejected": -0.2060546875, + "logps/chosen": -350.0, + "logps/rejected": -340.0, + "loss": 0.4018, + "rewards/accuracies": 0.8046875, + "rewards/chosen": 1.4453125, + "rewards/margins": 5.8125, + "rewards/rejected": -4.375, + "step": 162 + }, + { + "epoch": 0.2, + "grad_norm": 38.00229232317159, + "learning_rate": 1.4144628099173554e-06, + "logits/chosen": -0.07373046875, + "logits/rejected": -0.220703125, + "logps/chosen": -392.0, + "logps/rejected": -344.0, + "loss": 0.2926, + "rewards/accuracies": 0.8671875, + "rewards/chosen": 1.421875, + "rewards/margins": 6.25, + "rewards/rejected": -4.8125, + "step": 163 + }, + { + "epoch": 0.20122699386503068, + "grad_norm": 31.02665679742822, + "learning_rate": 1.4138429752066115e-06, + "logits/chosen": -0.236328125, + "logits/rejected": -0.466796875, + "logps/chosen": -392.0, + "logps/rejected": -310.0, + "loss": 0.2813, + "rewards/accuracies": 0.859375, + "rewards/chosen": 2.15625, + "rewards/margins": 7.34375, + "rewards/rejected": -5.1875, + "step": 164 + }, + { + "epoch": 0.20245398773006135, + "grad_norm": 43.699969414714815, + "learning_rate": 1.4132231404958679e-06, + "logits/chosen": -0.021728515625, + "logits/rejected": -0.212890625, + "logps/chosen": -444.0, + "logps/rejected": -326.0, + "loss": 0.377, + "rewards/accuracies": 0.828125, + "rewards/chosen": 1.5078125, + "rewards/margins": 6.3125, + "rewards/rejected": -4.78125, + "step": 165 + }, + { + "epoch": 0.20368098159509201, + "grad_norm": 38.30377721360019, + "learning_rate": 1.412603305785124e-06, + "logits/chosen": -0.09375, + "logits/rejected": -0.314453125, + "logps/chosen": -388.0, + "logps/rejected": -336.0, + "loss": 0.3851, + "rewards/accuracies": 0.828125, + "rewards/chosen": 1.9140625, + "rewards/margins": 6.53125, + "rewards/rejected": -4.625, + "step": 166 + }, + { + "epoch": 0.2049079754601227, + "grad_norm": 38.75636778509963, + "learning_rate": 1.4119834710743801e-06, + "logits/chosen": -0.06591796875, + "logits/rejected": -0.2080078125, + "logps/chosen": -398.0, + "logps/rejected": -318.0, + "loss": 0.3749, + "rewards/accuracies": 0.7890625, + "rewards/chosen": 2.21875, + "rewards/margins": 6.875, + "rewards/rejected": -4.65625, + "step": 167 + }, + { + "epoch": 0.20613496932515338, + "grad_norm": 37.779783793801464, + "learning_rate": 1.4113636363636365e-06, + "logits/chosen": -0.049560546875, + "logits/rejected": -0.1494140625, + "logps/chosen": -348.0, + "logps/rejected": -292.0, + "loss": 0.3337, + "rewards/accuracies": 0.84375, + "rewards/chosen": 0.99609375, + "rewards/margins": 6.46875, + "rewards/rejected": -5.46875, + "step": 168 + }, + { + "epoch": 0.20736196319018405, + "grad_norm": 34.571770475148426, + "learning_rate": 1.4107438016528926e-06, + "logits/chosen": -0.1259765625, + "logits/rejected": -0.27734375, + "logps/chosen": -372.0, + "logps/rejected": -366.0, + "loss": 0.3006, + "rewards/accuracies": 0.8671875, + "rewards/chosen": 1.0078125, + "rewards/margins": 6.25, + "rewards/rejected": -5.25, + "step": 169 + }, + { + "epoch": 0.2085889570552147, + "grad_norm": 42.205138483103795, + "learning_rate": 1.4101239669421488e-06, + "logits/chosen": -0.1123046875, + "logits/rejected": -0.349609375, + "logps/chosen": -450.0, + "logps/rejected": -346.0, + "loss": 0.338, + "rewards/accuracies": 0.8125, + "rewards/chosen": 1.875, + "rewards/margins": 7.34375, + "rewards/rejected": -5.46875, + "step": 170 + }, + { + "epoch": 0.2098159509202454, + "grad_norm": 29.828895289514126, + "learning_rate": 1.4095041322314051e-06, + "logits/chosen": -0.10888671875, + "logits/rejected": -0.279296875, + "logps/chosen": -384.0, + "logps/rejected": -322.0, + "loss": 0.3086, + "rewards/accuracies": 0.859375, + "rewards/chosen": 1.5546875, + "rewards/margins": 7.46875, + "rewards/rejected": -5.90625, + "step": 171 + }, + { + "epoch": 0.21104294478527608, + "grad_norm": 30.980573058674597, + "learning_rate": 1.408884297520661e-06, + "logits/chosen": -0.062255859375, + "logits/rejected": -0.2216796875, + "logps/chosen": -438.0, + "logps/rejected": -354.0, + "loss": 0.2599, + "rewards/accuracies": 0.859375, + "rewards/chosen": 1.84375, + "rewards/margins": 7.84375, + "rewards/rejected": -6.0, + "step": 172 + }, + { + "epoch": 0.21226993865030674, + "grad_norm": 25.22446362010358, + "learning_rate": 1.4082644628099174e-06, + "logits/chosen": -0.123046875, + "logits/rejected": -0.353515625, + "logps/chosen": -428.0, + "logps/rejected": -362.0, + "loss": 0.1997, + "rewards/accuracies": 0.890625, + "rewards/chosen": 1.703125, + "rewards/margins": 8.25, + "rewards/rejected": -6.5625, + "step": 173 + }, + { + "epoch": 0.2134969325153374, + "grad_norm": 43.79333405262764, + "learning_rate": 1.4076446280991736e-06, + "logits/chosen": 0.05126953125, + "logits/rejected": -0.158203125, + "logps/chosen": -384.0, + "logps/rejected": -348.0, + "loss": 0.5204, + "rewards/accuracies": 0.765625, + "rewards/chosen": 0.78125, + "rewards/margins": 6.8125, + "rewards/rejected": -6.03125, + "step": 174 + }, + { + "epoch": 0.2147239263803681, + "grad_norm": 45.27680571467687, + "learning_rate": 1.4070247933884297e-06, + "logits/chosen": -0.10546875, + "logits/rejected": -0.220703125, + "logps/chosen": -402.0, + "logps/rejected": -366.0, + "loss": 0.4454, + "rewards/accuracies": 0.8125, + "rewards/chosen": 1.09375, + "rewards/margins": 7.34375, + "rewards/rejected": -6.25, + "step": 175 + }, + { + "epoch": 0.21595092024539878, + "grad_norm": 35.16824036294685, + "learning_rate": 1.406404958677686e-06, + "logits/chosen": -0.1298828125, + "logits/rejected": -0.30859375, + "logps/chosen": -434.0, + "logps/rejected": -378.0, + "loss": 0.3443, + "rewards/accuracies": 0.796875, + "rewards/chosen": 1.3359375, + "rewards/margins": 7.65625, + "rewards/rejected": -6.3125, + "step": 176 + }, + { + "epoch": 0.21717791411042944, + "grad_norm": 27.30910780814845, + "learning_rate": 1.4057851239669422e-06, + "logits/chosen": -0.08251953125, + "logits/rejected": -0.30859375, + "logps/chosen": -416.0, + "logps/rejected": -356.0, + "loss": 0.2873, + "rewards/accuracies": 0.875, + "rewards/chosen": 2.0625, + "rewards/margins": 8.8125, + "rewards/rejected": -6.71875, + "step": 177 + }, + { + "epoch": 0.2184049079754601, + "grad_norm": 37.84132418939681, + "learning_rate": 1.4051652892561983e-06, + "logits/chosen": -0.07275390625, + "logits/rejected": -0.216796875, + "logps/chosen": -330.0, + "logps/rejected": -318.0, + "loss": 0.4137, + "rewards/accuracies": 0.828125, + "rewards/chosen": 0.1806640625, + "rewards/margins": 6.75, + "rewards/rejected": -6.59375, + "step": 178 + }, + { + "epoch": 0.2196319018404908, + "grad_norm": 32.98852515847511, + "learning_rate": 1.4045454545454547e-06, + "logits/chosen": -0.1064453125, + "logits/rejected": -0.31640625, + "logps/chosen": -372.0, + "logps/rejected": -324.0, + "loss": 0.3252, + "rewards/accuracies": 0.859375, + "rewards/chosen": 0.671875, + "rewards/margins": 6.875, + "rewards/rejected": -6.1875, + "step": 179 + }, + { + "epoch": 0.22085889570552147, + "grad_norm": 38.78908549807395, + "learning_rate": 1.4039256198347108e-06, + "logits/chosen": -0.099609375, + "logits/rejected": -0.21484375, + "logps/chosen": -356.0, + "logps/rejected": -314.0, + "loss": 0.3669, + "rewards/accuracies": 0.828125, + "rewards/chosen": 1.234375, + "rewards/margins": 6.46875, + "rewards/rejected": -5.21875, + "step": 180 + }, + { + "epoch": 0.22208588957055214, + "grad_norm": 35.32530991965649, + "learning_rate": 1.403305785123967e-06, + "logits/chosen": -0.09423828125, + "logits/rejected": -0.2412109375, + "logps/chosen": -410.0, + "logps/rejected": -386.0, + "loss": 0.3373, + "rewards/accuracies": 0.859375, + "rewards/chosen": 1.6171875, + "rewards/margins": 6.875, + "rewards/rejected": -5.25, + "step": 181 + }, + { + "epoch": 0.2233128834355828, + "grad_norm": 36.9736807793252, + "learning_rate": 1.4026859504132231e-06, + "logits/chosen": -0.12890625, + "logits/rejected": -0.2470703125, + "logps/chosen": -370.0, + "logps/rejected": -322.0, + "loss": 0.3661, + "rewards/accuracies": 0.8515625, + "rewards/chosen": 1.109375, + "rewards/margins": 5.875, + "rewards/rejected": -4.75, + "step": 182 + }, + { + "epoch": 0.2245398773006135, + "grad_norm": 40.8878376560248, + "learning_rate": 1.4020661157024793e-06, + "logits/chosen": 0.00762939453125, + "logits/rejected": -0.1669921875, + "logps/chosen": -412.0, + "logps/rejected": -378.0, + "loss": 0.3866, + "rewards/accuracies": 0.8203125, + "rewards/chosen": 0.95703125, + "rewards/margins": 4.90625, + "rewards/rejected": -3.953125, + "step": 183 + }, + { + "epoch": 0.22576687116564417, + "grad_norm": 42.156241310731595, + "learning_rate": 1.4014462809917356e-06, + "logits/chosen": -0.0255126953125, + "logits/rejected": -0.1328125, + "logps/chosen": -390.0, + "logps/rejected": -340.0, + "loss": 0.4166, + "rewards/accuracies": 0.7890625, + "rewards/chosen": 0.87109375, + "rewards/margins": 5.1875, + "rewards/rejected": -4.3125, + "step": 184 + }, + { + "epoch": 0.22699386503067484, + "grad_norm": 44.943258114663124, + "learning_rate": 1.4008264462809917e-06, + "logits/chosen": -0.06787109375, + "logits/rejected": -0.1220703125, + "logps/chosen": -382.0, + "logps/rejected": -350.0, + "loss": 0.4011, + "rewards/accuracies": 0.78125, + "rewards/chosen": 1.9765625, + "rewards/margins": 5.8125, + "rewards/rejected": -3.828125, + "step": 185 + }, + { + "epoch": 0.2282208588957055, + "grad_norm": 33.47267300169336, + "learning_rate": 1.4002066115702479e-06, + "logits/chosen": -0.08203125, + "logits/rejected": -0.1982421875, + "logps/chosen": -336.0, + "logps/rejected": -312.0, + "loss": 0.3163, + "rewards/accuracies": 0.859375, + "rewards/chosen": 1.4453125, + "rewards/margins": 5.78125, + "rewards/rejected": -4.3125, + "step": 186 + }, + { + "epoch": 0.2294478527607362, + "grad_norm": 32.03885108074069, + "learning_rate": 1.3995867768595042e-06, + "logits/chosen": -0.07080078125, + "logits/rejected": -0.2314453125, + "logps/chosen": -370.0, + "logps/rejected": -340.0, + "loss": 0.258, + "rewards/accuracies": 0.8671875, + "rewards/chosen": 1.6328125, + "rewards/margins": 6.4375, + "rewards/rejected": -4.78125, + "step": 187 + }, + { + "epoch": 0.23067484662576687, + "grad_norm": 35.158897439200416, + "learning_rate": 1.3989669421487604e-06, + "logits/chosen": 0.0228271484375, + "logits/rejected": -0.0615234375, + "logps/chosen": -330.0, + "logps/rejected": -292.0, + "loss": 0.3609, + "rewards/accuracies": 0.8046875, + "rewards/chosen": 1.40625, + "rewards/margins": 5.875, + "rewards/rejected": -4.46875, + "step": 188 + }, + { + "epoch": 0.23190184049079754, + "grad_norm": 32.15934523587848, + "learning_rate": 1.3983471074380167e-06, + "logits/chosen": -0.080078125, + "logits/rejected": -0.263671875, + "logps/chosen": -396.0, + "logps/rejected": -322.0, + "loss": 0.2271, + "rewards/accuracies": 0.890625, + "rewards/chosen": 2.046875, + "rewards/margins": 6.78125, + "rewards/rejected": -4.71875, + "step": 189 + }, + { + "epoch": 0.2331288343558282, + "grad_norm": 32.11586724143221, + "learning_rate": 1.3977272727272727e-06, + "logits/chosen": 0.041748046875, + "logits/rejected": -0.10693359375, + "logps/chosen": -340.0, + "logps/rejected": -316.0, + "loss": 0.3271, + "rewards/accuracies": 0.8515625, + "rewards/chosen": 1.4140625, + "rewards/margins": 6.0, + "rewards/rejected": -4.59375, + "step": 190 + }, + { + "epoch": 0.2343558282208589, + "grad_norm": 46.26761231273777, + "learning_rate": 1.3971074380165288e-06, + "logits/chosen": -0.0390625, + "logits/rejected": -0.220703125, + "logps/chosen": -400.0, + "logps/rejected": -306.0, + "loss": 0.3906, + "rewards/accuracies": 0.8046875, + "rewards/chosen": 1.828125, + "rewards/margins": 6.90625, + "rewards/rejected": -5.0625, + "step": 191 + }, + { + "epoch": 0.23558282208588957, + "grad_norm": 39.15535099753523, + "learning_rate": 1.3964876033057852e-06, + "logits/chosen": -0.052490234375, + "logits/rejected": -0.2490234375, + "logps/chosen": -388.0, + "logps/rejected": -302.0, + "loss": 0.3379, + "rewards/accuracies": 0.8046875, + "rewards/chosen": 1.1328125, + "rewards/margins": 5.65625, + "rewards/rejected": -4.5, + "step": 192 + }, + { + "epoch": 0.23680981595092024, + "grad_norm": 31.3365855799527, + "learning_rate": 1.3958677685950413e-06, + "logits/chosen": -0.1767578125, + "logits/rejected": -0.423828125, + "logps/chosen": -410.0, + "logps/rejected": -346.0, + "loss": 0.2485, + "rewards/accuracies": 0.8828125, + "rewards/chosen": 1.7578125, + "rewards/margins": 7.375, + "rewards/rejected": -5.625, + "step": 193 + }, + { + "epoch": 0.23803680981595093, + "grad_norm": 38.5490731428379, + "learning_rate": 1.3952479338842977e-06, + "logits/chosen": -0.21875, + "logits/rejected": -0.326171875, + "logps/chosen": -394.0, + "logps/rejected": -336.0, + "loss": 0.296, + "rewards/accuracies": 0.8203125, + "rewards/chosen": 0.89453125, + "rewards/margins": 6.875, + "rewards/rejected": -5.96875, + "step": 194 + }, + { + "epoch": 0.2392638036809816, + "grad_norm": 35.180225705298604, + "learning_rate": 1.3946280991735538e-06, + "logits/chosen": -0.031982421875, + "logits/rejected": -0.2890625, + "logps/chosen": -408.0, + "logps/rejected": -348.0, + "loss": 0.2727, + "rewards/accuracies": 0.8671875, + "rewards/chosen": 2.046875, + "rewards/margins": 8.1875, + "rewards/rejected": -6.125, + "step": 195 + }, + { + "epoch": 0.24049079754601227, + "grad_norm": 37.22166797602668, + "learning_rate": 1.39400826446281e-06, + "logits/chosen": 0.0157470703125, + "logits/rejected": -0.1845703125, + "logps/chosen": -394.0, + "logps/rejected": -308.0, + "loss": 0.3333, + "rewards/accuracies": 0.8359375, + "rewards/chosen": 1.1171875, + "rewards/margins": 6.65625, + "rewards/rejected": -5.53125, + "step": 196 + }, + { + "epoch": 0.24171779141104294, + "grad_norm": 28.562917988218086, + "learning_rate": 1.3933884297520663e-06, + "logits/chosen": -0.115234375, + "logits/rejected": -0.34765625, + "logps/chosen": -380.0, + "logps/rejected": -334.0, + "loss": 0.3009, + "rewards/accuracies": 0.8828125, + "rewards/chosen": 1.3359375, + "rewards/margins": 7.125, + "rewards/rejected": -5.78125, + "step": 197 + }, + { + "epoch": 0.24294478527607363, + "grad_norm": 36.54466818998776, + "learning_rate": 1.3927685950413224e-06, + "logits/chosen": -0.0174560546875, + "logits/rejected": -0.10205078125, + "logps/chosen": -346.0, + "logps/rejected": -276.0, + "loss": 0.4465, + "rewards/accuracies": 0.8046875, + "rewards/chosen": 0.625, + "rewards/margins": 6.0625, + "rewards/rejected": -5.4375, + "step": 198 + }, + { + "epoch": 0.2441717791411043, + "grad_norm": 25.79094354827764, + "learning_rate": 1.3921487603305786e-06, + "logits/chosen": -0.05615234375, + "logits/rejected": -0.26171875, + "logps/chosen": -434.0, + "logps/rejected": -330.0, + "loss": 0.1969, + "rewards/accuracies": 0.8984375, + "rewards/chosen": 1.7265625, + "rewards/margins": 7.84375, + "rewards/rejected": -6.125, + "step": 199 + }, + { + "epoch": 0.24539877300613497, + "grad_norm": 34.674086597448465, + "learning_rate": 1.3915289256198347e-06, + "logits/chosen": -0.037353515625, + "logits/rejected": -0.25390625, + "logps/chosen": -442.0, + "logps/rejected": -392.0, + "loss": 0.3066, + "rewards/accuracies": 0.8515625, + "rewards/chosen": 1.2734375, + "rewards/margins": 7.4375, + "rewards/rejected": -6.15625, + "step": 200 + }, + { + "epoch": 0.24662576687116564, + "grad_norm": 29.903922318940893, + "learning_rate": 1.3909090909090909e-06, + "logits/chosen": -0.03662109375, + "logits/rejected": -0.259765625, + "logps/chosen": -404.0, + "logps/rejected": -330.0, + "loss": 0.2629, + "rewards/accuracies": 0.8828125, + "rewards/chosen": 1.109375, + "rewards/margins": 7.0625, + "rewards/rejected": -5.9375, + "step": 201 + }, + { + "epoch": 0.24785276073619633, + "grad_norm": 33.833661458600005, + "learning_rate": 1.3902892561983472e-06, + "logits/chosen": -0.02001953125, + "logits/rejected": -0.2333984375, + "logps/chosen": -432.0, + "logps/rejected": -340.0, + "loss": 0.2977, + "rewards/accuracies": 0.8203125, + "rewards/chosen": 1.03125, + "rewards/margins": 6.90625, + "rewards/rejected": -5.875, + "step": 202 + }, + { + "epoch": 0.249079754601227, + "grad_norm": 37.642944458655066, + "learning_rate": 1.3896694214876034e-06, + "logits/chosen": 0.0272216796875, + "logits/rejected": -0.1474609375, + "logps/chosen": -392.0, + "logps/rejected": -354.0, + "loss": 0.482, + "rewards/accuracies": 0.796875, + "rewards/chosen": 0.96875, + "rewards/margins": 6.03125, + "rewards/rejected": -5.0625, + "step": 203 + }, + { + "epoch": 0.25030674846625767, + "grad_norm": 35.44282157848665, + "learning_rate": 1.3890495867768595e-06, + "logits/chosen": 0.0390625, + "logits/rejected": -0.1005859375, + "logps/chosen": -356.0, + "logps/rejected": -290.0, + "loss": 0.4211, + "rewards/accuracies": 0.8359375, + "rewards/chosen": 0.6484375, + "rewards/margins": 6.4375, + "rewards/rejected": -5.78125, + "step": 204 + }, + { + "epoch": 0.25153374233128833, + "grad_norm": 38.48812661842802, + "learning_rate": 1.3884297520661158e-06, + "logits/chosen": 0.08984375, + "logits/rejected": -0.04833984375, + "logps/chosen": -390.0, + "logps/rejected": -316.0, + "loss": 0.4435, + "rewards/accuracies": 0.8046875, + "rewards/chosen": 0.8671875, + "rewards/margins": 6.0, + "rewards/rejected": -5.125, + "step": 205 + }, + { + "epoch": 0.252760736196319, + "grad_norm": 26.640868179250116, + "learning_rate": 1.387809917355372e-06, + "logits/chosen": -0.036376953125, + "logits/rejected": -0.3125, + "logps/chosen": -434.0, + "logps/rejected": -350.0, + "loss": 0.2375, + "rewards/accuracies": 0.9140625, + "rewards/chosen": 1.203125, + "rewards/margins": 7.46875, + "rewards/rejected": -6.28125, + "step": 206 + }, + { + "epoch": 0.25398773006134967, + "grad_norm": 37.72822731879787, + "learning_rate": 1.3871900826446281e-06, + "logits/chosen": -0.16015625, + "logits/rejected": -0.353515625, + "logps/chosen": -434.0, + "logps/rejected": -376.0, + "loss": 0.4045, + "rewards/accuracies": 0.828125, + "rewards/chosen": 1.21875, + "rewards/margins": 6.75, + "rewards/rejected": -5.53125, + "step": 207 + }, + { + "epoch": 0.2552147239263804, + "grad_norm": 39.99822195882057, + "learning_rate": 1.3865702479338843e-06, + "logits/chosen": -0.12255859375, + "logits/rejected": -0.2470703125, + "logps/chosen": -396.0, + "logps/rejected": -340.0, + "loss": 0.4402, + "rewards/accuracies": 0.765625, + "rewards/chosen": 0.83984375, + "rewards/margins": 6.53125, + "rewards/rejected": -5.6875, + "step": 208 + }, + { + "epoch": 0.25644171779141106, + "grad_norm": 31.995520167951547, + "learning_rate": 1.3859504132231404e-06, + "logits/chosen": -0.125, + "logits/rejected": -0.2314453125, + "logps/chosen": -398.0, + "logps/rejected": -352.0, + "loss": 0.3203, + "rewards/accuracies": 0.84375, + "rewards/chosen": 0.76171875, + "rewards/margins": 6.34375, + "rewards/rejected": -5.5625, + "step": 209 + }, + { + "epoch": 0.25766871165644173, + "grad_norm": 42.71588448177596, + "learning_rate": 1.3853305785123968e-06, + "logits/chosen": -0.0732421875, + "logits/rejected": -0.1806640625, + "logps/chosen": -400.0, + "logps/rejected": -382.0, + "loss": 0.4413, + "rewards/accuracies": 0.8203125, + "rewards/chosen": 0.314453125, + "rewards/margins": 5.90625, + "rewards/rejected": -5.59375, + "step": 210 + }, + { + "epoch": 0.2588957055214724, + "grad_norm": 35.38452290000429, + "learning_rate": 1.384710743801653e-06, + "logits/chosen": -0.1806640625, + "logits/rejected": -0.375, + "logps/chosen": -438.0, + "logps/rejected": -362.0, + "loss": 0.3649, + "rewards/accuracies": 0.8125, + "rewards/chosen": 0.9375, + "rewards/margins": 6.84375, + "rewards/rejected": -5.90625, + "step": 211 + }, + { + "epoch": 0.26012269938650306, + "grad_norm": 35.00437697538439, + "learning_rate": 1.384090909090909e-06, + "logits/chosen": -0.099609375, + "logits/rejected": -0.251953125, + "logps/chosen": -404.0, + "logps/rejected": -328.0, + "loss": 0.3606, + "rewards/accuracies": 0.8203125, + "rewards/chosen": 1.125, + "rewards/margins": 6.6875, + "rewards/rejected": -5.5625, + "step": 212 + }, + { + "epoch": 0.26134969325153373, + "grad_norm": 40.33706711227177, + "learning_rate": 1.3834710743801654e-06, + "logits/chosen": -0.031982421875, + "logits/rejected": -0.0888671875, + "logps/chosen": -352.0, + "logps/rejected": -296.0, + "loss": 0.3034, + "rewards/accuracies": 0.828125, + "rewards/chosen": 0.4765625, + "rewards/margins": 5.78125, + "rewards/rejected": -5.3125, + "step": 213 + }, + { + "epoch": 0.2625766871165644, + "grad_norm": 36.278922089821435, + "learning_rate": 1.3828512396694215e-06, + "logits/chosen": -0.0152587890625, + "logits/rejected": -0.171875, + "logps/chosen": -376.0, + "logps/rejected": -326.0, + "loss": 0.3538, + "rewards/accuracies": 0.796875, + "rewards/chosen": 1.1328125, + "rewards/margins": 6.375, + "rewards/rejected": -5.25, + "step": 214 + }, + { + "epoch": 0.26380368098159507, + "grad_norm": 39.61373791294869, + "learning_rate": 1.3822314049586777e-06, + "logits/chosen": 0.0224609375, + "logits/rejected": -0.06689453125, + "logps/chosen": -356.0, + "logps/rejected": -316.0, + "loss": 0.3393, + "rewards/accuracies": 0.8203125, + "rewards/chosen": 0.55859375, + "rewards/margins": 5.5625, + "rewards/rejected": -5.0, + "step": 215 + }, + { + "epoch": 0.2650306748466258, + "grad_norm": 36.52111248304287, + "learning_rate": 1.381611570247934e-06, + "logits/chosen": -0.04443359375, + "logits/rejected": -0.1572265625, + "logps/chosen": -334.0, + "logps/rejected": -302.0, + "loss": 0.3612, + "rewards/accuracies": 0.8203125, + "rewards/chosen": 0.42578125, + "rewards/margins": 5.375, + "rewards/rejected": -4.96875, + "step": 216 + }, + { + "epoch": 0.26625766871165646, + "grad_norm": 35.893159477690816, + "learning_rate": 1.38099173553719e-06, + "logits/chosen": 0.05517578125, + "logits/rejected": -0.08154296875, + "logps/chosen": -402.0, + "logps/rejected": -316.0, + "loss": 0.4174, + "rewards/accuracies": 0.7734375, + "rewards/chosen": 1.2890625, + "rewards/margins": 6.0625, + "rewards/rejected": -4.78125, + "step": 217 + }, + { + "epoch": 0.2674846625766871, + "grad_norm": 43.28791050351139, + "learning_rate": 1.3803719008264463e-06, + "logits/chosen": 0.0108642578125, + "logits/rejected": -0.0908203125, + "logps/chosen": -384.0, + "logps/rejected": -344.0, + "loss": 0.514, + "rewards/accuracies": 0.8125, + "rewards/chosen": 0.79296875, + "rewards/margins": 5.625, + "rewards/rejected": -4.8125, + "step": 218 + }, + { + "epoch": 0.2687116564417178, + "grad_norm": 30.96899305966375, + "learning_rate": 1.3797520661157025e-06, + "logits/chosen": -0.091796875, + "logits/rejected": -0.189453125, + "logps/chosen": -354.0, + "logps/rejected": -310.0, + "loss": 0.3339, + "rewards/accuracies": 0.828125, + "rewards/chosen": 0.81640625, + "rewards/margins": 6.09375, + "rewards/rejected": -5.28125, + "step": 219 + }, + { + "epoch": 0.26993865030674846, + "grad_norm": 49.60415252304962, + "learning_rate": 1.3791322314049586e-06, + "logits/chosen": 0.07373046875, + "logits/rejected": 0.0047607421875, + "logps/chosen": -366.0, + "logps/rejected": -372.0, + "loss": 0.4563, + "rewards/accuracies": 0.765625, + "rewards/chosen": 0.416015625, + "rewards/margins": 5.0625, + "rewards/rejected": -4.65625, + "step": 220 + }, + { + "epoch": 0.27116564417177913, + "grad_norm": 35.817382292638634, + "learning_rate": 1.378512396694215e-06, + "logits/chosen": -0.049560546875, + "logits/rejected": -0.25390625, + "logps/chosen": -364.0, + "logps/rejected": -306.0, + "loss": 0.4136, + "rewards/accuracies": 0.796875, + "rewards/chosen": 0.46484375, + "rewards/margins": 5.75, + "rewards/rejected": -5.28125, + "step": 221 + }, + { + "epoch": 0.2723926380368098, + "grad_norm": 33.842292905676736, + "learning_rate": 1.377892561983471e-06, + "logits/chosen": -0.240234375, + "logits/rejected": -0.3671875, + "logps/chosen": -388.0, + "logps/rejected": -318.0, + "loss": 0.3321, + "rewards/accuracies": 0.859375, + "rewards/chosen": 1.4296875, + "rewards/margins": 7.375, + "rewards/rejected": -5.9375, + "step": 222 + }, + { + "epoch": 0.27361963190184047, + "grad_norm": 34.24344157812324, + "learning_rate": 1.3772727272727274e-06, + "logits/chosen": -0.0262451171875, + "logits/rejected": -0.2060546875, + "logps/chosen": -390.0, + "logps/rejected": -338.0, + "loss": 0.3389, + "rewards/accuracies": 0.84375, + "rewards/chosen": 0.58984375, + "rewards/margins": 6.1875, + "rewards/rejected": -5.59375, + "step": 223 + }, + { + "epoch": 0.2748466257668712, + "grad_norm": 36.32910023908209, + "learning_rate": 1.3766528925619836e-06, + "logits/chosen": -0.1591796875, + "logits/rejected": -0.33203125, + "logps/chosen": -452.0, + "logps/rejected": -352.0, + "loss": 0.3172, + "rewards/accuracies": 0.8046875, + "rewards/chosen": 1.28125, + "rewards/margins": 7.1875, + "rewards/rejected": -5.90625, + "step": 224 + }, + { + "epoch": 0.27607361963190186, + "grad_norm": 33.96451254717418, + "learning_rate": 1.3760330578512397e-06, + "logits/chosen": -0.021484375, + "logits/rejected": -0.2119140625, + "logps/chosen": -344.0, + "logps/rejected": -322.0, + "loss": 0.2991, + "rewards/accuracies": 0.84375, + "rewards/chosen": 0.4609375, + "rewards/margins": 6.5, + "rewards/rejected": -6.0625, + "step": 225 + }, + { + "epoch": 0.2773006134969325, + "grad_norm": 37.753965128366424, + "learning_rate": 1.3754132231404959e-06, + "logits/chosen": -0.035888671875, + "logits/rejected": -0.1689453125, + "logps/chosen": -378.0, + "logps/rejected": -328.0, + "loss": 0.4435, + "rewards/accuracies": 0.8046875, + "rewards/chosen": 0.82421875, + "rewards/margins": 6.4375, + "rewards/rejected": -5.625, + "step": 226 + }, + { + "epoch": 0.2785276073619632, + "grad_norm": 37.97162675437385, + "learning_rate": 1.374793388429752e-06, + "logits/chosen": -0.078125, + "logits/rejected": -0.1416015625, + "logps/chosen": -362.0, + "logps/rejected": -332.0, + "loss": 0.3808, + "rewards/accuracies": 0.859375, + "rewards/chosen": 0.625, + "rewards/margins": 6.625, + "rewards/rejected": -6.0, + "step": 227 + }, + { + "epoch": 0.27975460122699386, + "grad_norm": 42.90390749582614, + "learning_rate": 1.3741735537190084e-06, + "logits/chosen": -0.0908203125, + "logits/rejected": -0.2734375, + "logps/chosen": -366.0, + "logps/rejected": -338.0, + "loss": 0.3568, + "rewards/accuracies": 0.8125, + "rewards/chosen": 0.57421875, + "rewards/margins": 6.03125, + "rewards/rejected": -5.46875, + "step": 228 + }, + { + "epoch": 0.2809815950920245, + "grad_norm": 35.198650610347336, + "learning_rate": 1.3735537190082645e-06, + "logits/chosen": 0.0174560546875, + "logits/rejected": -0.1845703125, + "logps/chosen": -422.0, + "logps/rejected": -344.0, + "loss": 0.3812, + "rewards/accuracies": 0.84375, + "rewards/chosen": 1.640625, + "rewards/margins": 6.03125, + "rewards/rejected": -4.375, + "step": 229 + }, + { + "epoch": 0.2822085889570552, + "grad_norm": 31.76904887089288, + "learning_rate": 1.3729338842975206e-06, + "logits/chosen": 0.0027618408203125, + "logits/rejected": -0.1806640625, + "logps/chosen": -376.0, + "logps/rejected": -334.0, + "loss": 0.2964, + "rewards/accuracies": 0.859375, + "rewards/chosen": 0.99609375, + "rewards/margins": 6.375, + "rewards/rejected": -5.375, + "step": 230 + }, + { + "epoch": 0.28343558282208586, + "grad_norm": 38.2475164063325, + "learning_rate": 1.372314049586777e-06, + "logits/chosen": -0.0615234375, + "logits/rejected": -0.177734375, + "logps/chosen": -370.0, + "logps/rejected": -320.0, + "loss": 0.4068, + "rewards/accuracies": 0.8203125, + "rewards/chosen": 1.03125, + "rewards/margins": 5.6875, + "rewards/rejected": -4.65625, + "step": 231 + }, + { + "epoch": 0.2846625766871166, + "grad_norm": 34.785350120825576, + "learning_rate": 1.3716942148760331e-06, + "logits/chosen": -0.1474609375, + "logits/rejected": -0.27734375, + "logps/chosen": -384.0, + "logps/rejected": -342.0, + "loss": 0.3319, + "rewards/accuracies": 0.8203125, + "rewards/chosen": 0.6796875, + "rewards/margins": 5.5625, + "rewards/rejected": -4.875, + "step": 232 + }, + { + "epoch": 0.28588957055214725, + "grad_norm": 33.915962784771814, + "learning_rate": 1.3710743801652893e-06, + "logits/chosen": -0.04052734375, + "logits/rejected": -0.255859375, + "logps/chosen": -372.0, + "logps/rejected": -296.0, + "loss": 0.3377, + "rewards/accuracies": 0.8359375, + "rewards/chosen": 1.4296875, + "rewards/margins": 6.28125, + "rewards/rejected": -4.84375, + "step": 233 + }, + { + "epoch": 0.2871165644171779, + "grad_norm": 36.944834789970336, + "learning_rate": 1.3704545454545456e-06, + "logits/chosen": -0.189453125, + "logits/rejected": -0.302734375, + "logps/chosen": -374.0, + "logps/rejected": -318.0, + "loss": 0.3474, + "rewards/accuracies": 0.8359375, + "rewards/chosen": 1.203125, + "rewards/margins": 5.625, + "rewards/rejected": -4.40625, + "step": 234 + }, + { + "epoch": 0.2883435582822086, + "grad_norm": 36.36585753307406, + "learning_rate": 1.3698347107438016e-06, + "logits/chosen": -0.201171875, + "logits/rejected": -0.34765625, + "logps/chosen": -362.0, + "logps/rejected": -326.0, + "loss": 0.3987, + "rewards/accuracies": 0.8359375, + "rewards/chosen": 1.1171875, + "rewards/margins": 5.90625, + "rewards/rejected": -4.78125, + "step": 235 + }, + { + "epoch": 0.28957055214723926, + "grad_norm": 39.75324625491103, + "learning_rate": 1.369214876033058e-06, + "logits/chosen": 0.058837890625, + "logits/rejected": -0.1123046875, + "logps/chosen": -396.0, + "logps/rejected": -374.0, + "loss": 0.4644, + "rewards/accuracies": 0.7890625, + "rewards/chosen": 1.7421875, + "rewards/margins": 5.625, + "rewards/rejected": -3.890625, + "step": 236 + }, + { + "epoch": 0.2907975460122699, + "grad_norm": 33.93837750645907, + "learning_rate": 1.368595041322314e-06, + "logits/chosen": -0.1025390625, + "logits/rejected": -0.248046875, + "logps/chosen": -368.0, + "logps/rejected": -314.0, + "loss": 0.3597, + "rewards/accuracies": 0.8515625, + "rewards/chosen": 1.5, + "rewards/margins": 6.1875, + "rewards/rejected": -4.6875, + "step": 237 + }, + { + "epoch": 0.2920245398773006, + "grad_norm": 40.82126535762802, + "learning_rate": 1.3679752066115702e-06, + "logits/chosen": -0.11376953125, + "logits/rejected": -0.259765625, + "logps/chosen": -426.0, + "logps/rejected": -356.0, + "loss": 0.5098, + "rewards/accuracies": 0.8046875, + "rewards/chosen": 1.4453125, + "rewards/margins": 5.6875, + "rewards/rejected": -4.25, + "step": 238 + }, + { + "epoch": 0.29325153374233126, + "grad_norm": 37.96435040796285, + "learning_rate": 1.3673553719008266e-06, + "logits/chosen": -0.130859375, + "logits/rejected": -0.298828125, + "logps/chosen": -382.0, + "logps/rejected": -314.0, + "loss": 0.371, + "rewards/accuracies": 0.796875, + "rewards/chosen": 2.078125, + "rewards/margins": 6.46875, + "rewards/rejected": -4.375, + "step": 239 + }, + { + "epoch": 0.294478527607362, + "grad_norm": 34.11140908926358, + "learning_rate": 1.3667355371900827e-06, + "logits/chosen": -0.15234375, + "logits/rejected": -0.28515625, + "logps/chosen": -418.0, + "logps/rejected": -364.0, + "loss": 0.2283, + "rewards/accuracies": 0.890625, + "rewards/chosen": 1.890625, + "rewards/margins": 6.875, + "rewards/rejected": -5.0, + "step": 240 + }, + { + "epoch": 0.29570552147239265, + "grad_norm": 33.461868421672946, + "learning_rate": 1.3661157024793388e-06, + "logits/chosen": -0.076171875, + "logits/rejected": -0.2001953125, + "logps/chosen": -366.0, + "logps/rejected": -336.0, + "loss": 0.2828, + "rewards/accuracies": 0.84375, + "rewards/chosen": 1.8359375, + "rewards/margins": 6.4375, + "rewards/rejected": -4.59375, + "step": 241 + }, + { + "epoch": 0.2969325153374233, + "grad_norm": 32.77289894542164, + "learning_rate": 1.3654958677685952e-06, + "logits/chosen": -0.10009765625, + "logits/rejected": -0.2490234375, + "logps/chosen": -350.0, + "logps/rejected": -324.0, + "loss": 0.2975, + "rewards/accuracies": 0.84375, + "rewards/chosen": 1.2578125, + "rewards/margins": 6.59375, + "rewards/rejected": -5.34375, + "step": 242 + }, + { + "epoch": 0.298159509202454, + "grad_norm": 38.07813065801456, + "learning_rate": 1.3648760330578511e-06, + "logits/chosen": -0.1923828125, + "logits/rejected": -0.328125, + "logps/chosen": -390.0, + "logps/rejected": -364.0, + "loss": 0.4242, + "rewards/accuracies": 0.8203125, + "rewards/chosen": 1.328125, + "rewards/margins": 5.625, + "rewards/rejected": -4.3125, + "step": 243 + }, + { + "epoch": 0.29938650306748466, + "grad_norm": 39.871497711059654, + "learning_rate": 1.3642561983471075e-06, + "logits/chosen": -0.1806640625, + "logits/rejected": -0.39453125, + "logps/chosen": -386.0, + "logps/rejected": -346.0, + "loss": 0.5523, + "rewards/accuracies": 0.7734375, + "rewards/chosen": 1.265625, + "rewards/margins": 5.46875, + "rewards/rejected": -4.1875, + "step": 244 + }, + { + "epoch": 0.3006134969325153, + "grad_norm": 32.221505039188294, + "learning_rate": 1.3636363636363636e-06, + "logits/chosen": -0.15625, + "logits/rejected": -0.2890625, + "logps/chosen": -398.0, + "logps/rejected": -354.0, + "loss": 0.2967, + "rewards/accuracies": 0.8515625, + "rewards/chosen": 1.9140625, + "rewards/margins": 6.46875, + "rewards/rejected": -4.53125, + "step": 245 + }, + { + "epoch": 0.301840490797546, + "grad_norm": 34.1015902082894, + "learning_rate": 1.3630165289256198e-06, + "logits/chosen": -0.166015625, + "logits/rejected": -0.31640625, + "logps/chosen": -374.0, + "logps/rejected": -346.0, + "loss": 0.3765, + "rewards/accuracies": 0.828125, + "rewards/chosen": 2.171875, + "rewards/margins": 6.46875, + "rewards/rejected": -4.3125, + "step": 246 + }, + { + "epoch": 0.3030674846625767, + "grad_norm": 32.83309359401401, + "learning_rate": 1.3623966942148761e-06, + "logits/chosen": -0.1572265625, + "logits/rejected": -0.337890625, + "logps/chosen": -372.0, + "logps/rejected": -308.0, + "loss": 0.3534, + "rewards/accuracies": 0.8125, + "rewards/chosen": 1.1640625, + "rewards/margins": 5.96875, + "rewards/rejected": -4.78125, + "step": 247 + }, + { + "epoch": 0.3042944785276074, + "grad_norm": 33.089465845458236, + "learning_rate": 1.3617768595041322e-06, + "logits/chosen": -0.076171875, + "logits/rejected": -0.3203125, + "logps/chosen": -388.0, + "logps/rejected": -332.0, + "loss": 0.2682, + "rewards/accuracies": 0.8671875, + "rewards/chosen": 1.703125, + "rewards/margins": 6.03125, + "rewards/rejected": -4.3125, + "step": 248 + }, + { + "epoch": 0.30552147239263805, + "grad_norm": 39.03476607944447, + "learning_rate": 1.3611570247933884e-06, + "logits/chosen": -0.20703125, + "logits/rejected": -0.396484375, + "logps/chosen": -474.0, + "logps/rejected": -368.0, + "loss": 0.3241, + "rewards/accuracies": 0.875, + "rewards/chosen": 2.640625, + "rewards/margins": 6.78125, + "rewards/rejected": -4.15625, + "step": 249 + }, + { + "epoch": 0.3067484662576687, + "grad_norm": 38.29235099515543, + "learning_rate": 1.3605371900826447e-06, + "logits/chosen": -0.30078125, + "logits/rejected": -0.349609375, + "logps/chosen": -340.0, + "logps/rejected": -324.0, + "loss": 0.3804, + "rewards/accuracies": 0.84375, + "rewards/chosen": 1.34375, + "rewards/margins": 5.78125, + "rewards/rejected": -4.4375, + "step": 250 + }, + { + "epoch": 0.3079754601226994, + "grad_norm": 37.34357400280288, + "learning_rate": 1.3599173553719009e-06, + "logits/chosen": 0.0289306640625, + "logits/rejected": -0.134765625, + "logps/chosen": -376.0, + "logps/rejected": -308.0, + "loss": 0.4118, + "rewards/accuracies": 0.7890625, + "rewards/chosen": 1.421875, + "rewards/margins": 5.375, + "rewards/rejected": -3.9375, + "step": 251 + }, + { + "epoch": 0.30920245398773005, + "grad_norm": 37.841749969537275, + "learning_rate": 1.359297520661157e-06, + "logits/chosen": -0.1494140625, + "logits/rejected": -0.29296875, + "logps/chosen": -392.0, + "logps/rejected": -310.0, + "loss": 0.4506, + "rewards/accuracies": 0.8359375, + "rewards/chosen": 1.890625, + "rewards/margins": 5.78125, + "rewards/rejected": -3.875, + "step": 252 + }, + { + "epoch": 0.3104294478527607, + "grad_norm": 25.74895782175542, + "learning_rate": 1.3586776859504132e-06, + "logits/chosen": -0.1630859375, + "logits/rejected": -0.392578125, + "logps/chosen": -336.0, + "logps/rejected": -298.0, + "loss": 0.2694, + "rewards/accuracies": 0.8671875, + "rewards/chosen": 1.46875, + "rewards/margins": 6.0625, + "rewards/rejected": -4.59375, + "step": 253 + }, + { + "epoch": 0.3116564417177914, + "grad_norm": 36.56021325822385, + "learning_rate": 1.3580578512396693e-06, + "logits/chosen": -0.10302734375, + "logits/rejected": -0.25, + "logps/chosen": -352.0, + "logps/rejected": -320.0, + "loss": 0.4183, + "rewards/accuracies": 0.8046875, + "rewards/chosen": 1.8671875, + "rewards/margins": 6.1875, + "rewards/rejected": -4.3125, + "step": 254 + }, + { + "epoch": 0.3128834355828221, + "grad_norm": 37.45936986772829, + "learning_rate": 1.3574380165289257e-06, + "logits/chosen": -0.07958984375, + "logits/rejected": -0.228515625, + "logps/chosen": -384.0, + "logps/rejected": -312.0, + "loss": 0.439, + "rewards/accuracies": 0.796875, + "rewards/chosen": 1.8515625, + "rewards/margins": 5.71875, + "rewards/rejected": -3.859375, + "step": 255 + }, + { + "epoch": 0.3141104294478528, + "grad_norm": 33.113238785557265, + "learning_rate": 1.3568181818181818e-06, + "logits/chosen": -0.1640625, + "logits/rejected": -0.328125, + "logps/chosen": -348.0, + "logps/rejected": -316.0, + "loss": 0.3148, + "rewards/accuracies": 0.875, + "rewards/chosen": 2.3125, + "rewards/margins": 7.0, + "rewards/rejected": -4.6875, + "step": 256 + }, + { + "epoch": 0.31533742331288345, + "grad_norm": 34.825667619250744, + "learning_rate": 1.3561983471074382e-06, + "logits/chosen": -0.1494140625, + "logits/rejected": -0.2890625, + "logps/chosen": -426.0, + "logps/rejected": -356.0, + "loss": 0.3455, + "rewards/accuracies": 0.828125, + "rewards/chosen": 2.25, + "rewards/margins": 6.5625, + "rewards/rejected": -4.3125, + "step": 257 + }, + { + "epoch": 0.3165644171779141, + "grad_norm": 35.08263037322888, + "learning_rate": 1.3555785123966943e-06, + "logits/chosen": -0.0908203125, + "logits/rejected": -0.28125, + "logps/chosen": -396.0, + "logps/rejected": -342.0, + "loss": 0.319, + "rewards/accuracies": 0.875, + "rewards/chosen": 1.9140625, + "rewards/margins": 6.65625, + "rewards/rejected": -4.75, + "step": 258 + }, + { + "epoch": 0.3177914110429448, + "grad_norm": 33.547677568046986, + "learning_rate": 1.3549586776859504e-06, + "logits/chosen": -0.1689453125, + "logits/rejected": -0.349609375, + "logps/chosen": -344.0, + "logps/rejected": -356.0, + "loss": 0.3553, + "rewards/accuracies": 0.84375, + "rewards/chosen": 1.296875, + "rewards/margins": 6.5, + "rewards/rejected": -5.1875, + "step": 259 + }, + { + "epoch": 0.31901840490797545, + "grad_norm": 35.38305104419636, + "learning_rate": 1.3543388429752068e-06, + "logits/chosen": -0.1904296875, + "logits/rejected": -0.26953125, + "logps/chosen": -390.0, + "logps/rejected": -326.0, + "loss": 0.3657, + "rewards/accuracies": 0.8359375, + "rewards/chosen": 1.296875, + "rewards/margins": 6.4375, + "rewards/rejected": -5.15625, + "step": 260 + }, + { + "epoch": 0.3202453987730061, + "grad_norm": 37.62925536561278, + "learning_rate": 1.3537190082644627e-06, + "logits/chosen": -0.11181640625, + "logits/rejected": -0.263671875, + "logps/chosen": -380.0, + "logps/rejected": -330.0, + "loss": 0.3941, + "rewards/accuracies": 0.8125, + "rewards/chosen": 1.296875, + "rewards/margins": 6.8125, + "rewards/rejected": -5.53125, + "step": 261 + }, + { + "epoch": 0.3214723926380368, + "grad_norm": 45.47620412654006, + "learning_rate": 1.353099173553719e-06, + "logits/chosen": -0.1591796875, + "logits/rejected": -0.3359375, + "logps/chosen": -386.0, + "logps/rejected": -340.0, + "loss": 0.4019, + "rewards/accuracies": 0.8125, + "rewards/chosen": 1.875, + "rewards/margins": 6.9375, + "rewards/rejected": -5.0625, + "step": 262 + }, + { + "epoch": 0.3226993865030675, + "grad_norm": 43.246204937096174, + "learning_rate": 1.3524793388429752e-06, + "logits/chosen": -0.1494140625, + "logits/rejected": -0.2734375, + "logps/chosen": -370.0, + "logps/rejected": -322.0, + "loss": 0.4521, + "rewards/accuracies": 0.765625, + "rewards/chosen": 0.6484375, + "rewards/margins": 5.65625, + "rewards/rejected": -5.0, + "step": 263 + }, + { + "epoch": 0.3239263803680982, + "grad_norm": 37.571005698421835, + "learning_rate": 1.3518595041322314e-06, + "logits/chosen": -0.09521484375, + "logits/rejected": -0.20703125, + "logps/chosen": -354.0, + "logps/rejected": -308.0, + "loss": 0.3748, + "rewards/accuracies": 0.8125, + "rewards/chosen": 1.0078125, + "rewards/margins": 6.1875, + "rewards/rejected": -5.15625, + "step": 264 + }, + { + "epoch": 0.32515337423312884, + "grad_norm": 45.351424976255885, + "learning_rate": 1.3512396694214877e-06, + "logits/chosen": -0.1064453125, + "logits/rejected": -0.265625, + "logps/chosen": -360.0, + "logps/rejected": -288.0, + "loss": 0.3745, + "rewards/accuracies": 0.8203125, + "rewards/chosen": 0.87109375, + "rewards/margins": 6.75, + "rewards/rejected": -5.875, + "step": 265 + }, + { + "epoch": 0.3263803680981595, + "grad_norm": 44.31098246879602, + "learning_rate": 1.3506198347107439e-06, + "logits/chosen": -0.07861328125, + "logits/rejected": -0.228515625, + "logps/chosen": -326.0, + "logps/rejected": -316.0, + "loss": 0.5312, + "rewards/accuracies": 0.78125, + "rewards/chosen": -0.27734375, + "rewards/margins": 5.25, + "rewards/rejected": -5.53125, + "step": 266 + }, + { + "epoch": 0.3276073619631902, + "grad_norm": 43.337206726154484, + "learning_rate": 1.35e-06, + "logits/chosen": -0.0966796875, + "logits/rejected": -0.31640625, + "logps/chosen": -406.0, + "logps/rejected": -322.0, + "loss": 0.3998, + "rewards/accuracies": 0.8359375, + "rewards/chosen": 1.2421875, + "rewards/margins": 6.875, + "rewards/rejected": -5.625, + "step": 267 + }, + { + "epoch": 0.32883435582822085, + "grad_norm": 28.822383795146568, + "learning_rate": 1.3493801652892563e-06, + "logits/chosen": -0.25390625, + "logits/rejected": -0.365234375, + "logps/chosen": -392.0, + "logps/rejected": -344.0, + "loss": 0.209, + "rewards/accuracies": 0.90625, + "rewards/chosen": 0.75390625, + "rewards/margins": 7.40625, + "rewards/rejected": -6.65625, + "step": 268 + }, + { + "epoch": 0.3300613496932515, + "grad_norm": 30.291051549986957, + "learning_rate": 1.3487603305785125e-06, + "logits/chosen": -0.1240234375, + "logits/rejected": -0.43359375, + "logps/chosen": -412.0, + "logps/rejected": -322.0, + "loss": 0.237, + "rewards/accuracies": 0.890625, + "rewards/chosen": 0.83203125, + "rewards/margins": 7.125, + "rewards/rejected": -6.3125, + "step": 269 + }, + { + "epoch": 0.3312883435582822, + "grad_norm": 38.1594801225088, + "learning_rate": 1.3481404958677686e-06, + "logits/chosen": -0.09912109375, + "logits/rejected": -0.2138671875, + "logps/chosen": -334.0, + "logps/rejected": -308.0, + "loss": 0.4766, + "rewards/accuracies": 0.8046875, + "rewards/chosen": 0.462890625, + "rewards/margins": 5.625, + "rewards/rejected": -5.15625, + "step": 270 + }, + { + "epoch": 0.3325153374233129, + "grad_norm": 28.374821367974555, + "learning_rate": 1.3475206611570248e-06, + "logits/chosen": -0.12451171875, + "logits/rejected": -0.240234375, + "logps/chosen": -390.0, + "logps/rejected": -344.0, + "loss": 0.3004, + "rewards/accuracies": 0.875, + "rewards/chosen": 1.3984375, + "rewards/margins": 7.5, + "rewards/rejected": -6.09375, + "step": 271 + }, + { + "epoch": 0.3337423312883436, + "grad_norm": 34.17489484053231, + "learning_rate": 1.346900826446281e-06, + "logits/chosen": -0.00823974609375, + "logits/rejected": -0.171875, + "logps/chosen": -366.0, + "logps/rejected": -318.0, + "loss": 0.2861, + "rewards/accuracies": 0.8671875, + "rewards/chosen": 0.423828125, + "rewards/margins": 6.75, + "rewards/rejected": -6.34375, + "step": 272 + }, + { + "epoch": 0.33496932515337424, + "grad_norm": 41.21053534408477, + "learning_rate": 1.3462809917355373e-06, + "logits/chosen": 0.06640625, + "logits/rejected": -0.007110595703125, + "logps/chosen": -344.0, + "logps/rejected": -314.0, + "loss": 0.5165, + "rewards/accuracies": 0.7578125, + "rewards/chosen": 0.166015625, + "rewards/margins": 5.75, + "rewards/rejected": -5.5625, + "step": 273 + }, + { + "epoch": 0.3361963190184049, + "grad_norm": 34.571787806623156, + "learning_rate": 1.3456611570247934e-06, + "logits/chosen": 0.0419921875, + "logits/rejected": -0.1025390625, + "logps/chosen": -410.0, + "logps/rejected": -346.0, + "loss": 0.3839, + "rewards/accuracies": 0.8515625, + "rewards/chosen": 0.83203125, + "rewards/margins": 6.40625, + "rewards/rejected": -5.5625, + "step": 274 + }, + { + "epoch": 0.3374233128834356, + "grad_norm": 31.979565052575825, + "learning_rate": 1.3450413223140495e-06, + "logits/chosen": -0.0302734375, + "logits/rejected": -0.119140625, + "logps/chosen": -326.0, + "logps/rejected": -326.0, + "loss": 0.4175, + "rewards/accuracies": 0.8046875, + "rewards/chosen": 0.412109375, + "rewards/margins": 6.09375, + "rewards/rejected": -5.6875, + "step": 275 + }, + { + "epoch": 0.33865030674846625, + "grad_norm": 28.189301031425032, + "learning_rate": 1.344421487603306e-06, + "logits/chosen": 0.0498046875, + "logits/rejected": -0.162109375, + "logps/chosen": -390.0, + "logps/rejected": -330.0, + "loss": 0.3505, + "rewards/accuracies": 0.8125, + "rewards/chosen": 0.2470703125, + "rewards/margins": 6.21875, + "rewards/rejected": -5.96875, + "step": 276 + }, + { + "epoch": 0.3398773006134969, + "grad_norm": 29.004541701793432, + "learning_rate": 1.343801652892562e-06, + "logits/chosen": -0.15234375, + "logits/rejected": -0.337890625, + "logps/chosen": -414.0, + "logps/rejected": -344.0, + "loss": 0.2468, + "rewards/accuracies": 0.8671875, + "rewards/chosen": 1.1640625, + "rewards/margins": 7.53125, + "rewards/rejected": -6.375, + "step": 277 + }, + { + "epoch": 0.3411042944785276, + "grad_norm": 40.163399091060114, + "learning_rate": 1.3431818181818184e-06, + "logits/chosen": -0.03759765625, + "logits/rejected": -0.08447265625, + "logps/chosen": -384.0, + "logps/rejected": -356.0, + "loss": 0.4008, + "rewards/accuracies": 0.8203125, + "rewards/chosen": 0.69921875, + "rewards/margins": 6.4375, + "rewards/rejected": -5.75, + "step": 278 + }, + { + "epoch": 0.3423312883435583, + "grad_norm": 21.8515947233153, + "learning_rate": 1.3425619834710743e-06, + "logits/chosen": -0.10888671875, + "logits/rejected": -0.2890625, + "logps/chosen": -392.0, + "logps/rejected": -326.0, + "loss": 0.1765, + "rewards/accuracies": 0.921875, + "rewards/chosen": 0.65625, + "rewards/margins": 7.46875, + "rewards/rejected": -6.8125, + "step": 279 + }, + { + "epoch": 0.34355828220858897, + "grad_norm": 34.753521492334585, + "learning_rate": 1.3419421487603305e-06, + "logits/chosen": -0.09228515625, + "logits/rejected": -0.234375, + "logps/chosen": -408.0, + "logps/rejected": -372.0, + "loss": 0.4097, + "rewards/accuracies": 0.8125, + "rewards/chosen": 0.83984375, + "rewards/margins": 6.96875, + "rewards/rejected": -6.15625, + "step": 280 + }, + { + "epoch": 0.34478527607361964, + "grad_norm": 35.589195433139196, + "learning_rate": 1.3413223140495868e-06, + "logits/chosen": -0.00921630859375, + "logits/rejected": -0.201171875, + "logps/chosen": -382.0, + "logps/rejected": -330.0, + "loss": 0.3917, + "rewards/accuracies": 0.8046875, + "rewards/chosen": 0.609375, + "rewards/margins": 5.65625, + "rewards/rejected": -5.0625, + "step": 281 + }, + { + "epoch": 0.3460122699386503, + "grad_norm": 29.302596470256102, + "learning_rate": 1.340702479338843e-06, + "logits/chosen": -0.2021484375, + "logits/rejected": -0.341796875, + "logps/chosen": -344.0, + "logps/rejected": -324.0, + "loss": 0.2481, + "rewards/accuracies": 0.875, + "rewards/chosen": 1.0859375, + "rewards/margins": 7.1875, + "rewards/rejected": -6.09375, + "step": 282 + }, + { + "epoch": 0.347239263803681, + "grad_norm": 33.7375126879184, + "learning_rate": 1.3400826446280993e-06, + "logits/chosen": -0.2265625, + "logits/rejected": -0.3125, + "logps/chosen": -354.0, + "logps/rejected": -298.0, + "loss": 0.3519, + "rewards/accuracies": 0.8203125, + "rewards/chosen": 0.59765625, + "rewards/margins": 6.15625, + "rewards/rejected": -5.5625, + "step": 283 + }, + { + "epoch": 0.34846625766871164, + "grad_norm": 35.97047400633379, + "learning_rate": 1.3394628099173555e-06, + "logits/chosen": -0.034423828125, + "logits/rejected": -0.166015625, + "logps/chosen": -370.0, + "logps/rejected": -320.0, + "loss": 0.2954, + "rewards/accuracies": 0.8515625, + "rewards/chosen": 0.859375, + "rewards/margins": 6.6875, + "rewards/rejected": -5.84375, + "step": 284 + }, + { + "epoch": 0.3496932515337423, + "grad_norm": 29.423070534885003, + "learning_rate": 1.3388429752066116e-06, + "logits/chosen": -0.130859375, + "logits/rejected": -0.322265625, + "logps/chosen": -360.0, + "logps/rejected": -328.0, + "loss": 0.2845, + "rewards/accuracies": 0.84375, + "rewards/chosen": 0.80078125, + "rewards/margins": 6.65625, + "rewards/rejected": -5.875, + "step": 285 + }, + { + "epoch": 0.350920245398773, + "grad_norm": 35.376791661357736, + "learning_rate": 1.338223140495868e-06, + "logits/chosen": -0.04736328125, + "logits/rejected": -0.2412109375, + "logps/chosen": -422.0, + "logps/rejected": -356.0, + "loss": 0.3799, + "rewards/accuracies": 0.796875, + "rewards/chosen": 0.8515625, + "rewards/margins": 6.9375, + "rewards/rejected": -6.0625, + "step": 286 + }, + { + "epoch": 0.3521472392638037, + "grad_norm": 38.51049635159653, + "learning_rate": 1.337603305785124e-06, + "logits/chosen": -0.150390625, + "logits/rejected": -0.283203125, + "logps/chosen": -382.0, + "logps/rejected": -330.0, + "loss": 0.3715, + "rewards/accuracies": 0.8125, + "rewards/chosen": 0.375, + "rewards/margins": 6.1875, + "rewards/rejected": -5.8125, + "step": 287 + }, + { + "epoch": 0.35337423312883437, + "grad_norm": 31.395092155644413, + "learning_rate": 1.3369834710743802e-06, + "logits/chosen": 0.057373046875, + "logits/rejected": -0.0888671875, + "logps/chosen": -386.0, + "logps/rejected": -336.0, + "loss": 0.3837, + "rewards/accuracies": 0.8359375, + "rewards/chosen": 0.298828125, + "rewards/margins": 5.9375, + "rewards/rejected": -5.625, + "step": 288 + }, + { + "epoch": 0.35460122699386504, + "grad_norm": 25.699158986974325, + "learning_rate": 1.3363636363636364e-06, + "logits/chosen": -0.11083984375, + "logits/rejected": -0.248046875, + "logps/chosen": -404.0, + "logps/rejected": -330.0, + "loss": 0.1963, + "rewards/accuracies": 0.921875, + "rewards/chosen": 0.94140625, + "rewards/margins": 7.65625, + "rewards/rejected": -6.71875, + "step": 289 + }, + { + "epoch": 0.3558282208588957, + "grad_norm": 26.172712857085628, + "learning_rate": 1.3357438016528925e-06, + "logits/chosen": -0.283203125, + "logits/rejected": -0.41796875, + "logps/chosen": -380.0, + "logps/rejected": -350.0, + "loss": 0.2207, + "rewards/accuracies": 0.9296875, + "rewards/chosen": 0.65625, + "rewards/margins": 7.65625, + "rewards/rejected": -7.0, + "step": 290 + }, + { + "epoch": 0.3570552147239264, + "grad_norm": 35.706855965144605, + "learning_rate": 1.3351239669421489e-06, + "logits/chosen": -0.2275390625, + "logits/rejected": -0.3359375, + "logps/chosen": -394.0, + "logps/rejected": -340.0, + "loss": 0.4375, + "rewards/accuracies": 0.8046875, + "rewards/chosen": 1.5625, + "rewards/margins": 7.09375, + "rewards/rejected": -5.53125, + "step": 291 + }, + { + "epoch": 0.35828220858895704, + "grad_norm": 32.40631116792141, + "learning_rate": 1.334504132231405e-06, + "logits/chosen": -0.169921875, + "logits/rejected": -0.390625, + "logps/chosen": -406.0, + "logps/rejected": -348.0, + "loss": 0.2821, + "rewards/accuracies": 0.8671875, + "rewards/chosen": 0.7734375, + "rewards/margins": 7.15625, + "rewards/rejected": -6.40625, + "step": 292 + }, + { + "epoch": 0.3595092024539877, + "grad_norm": 27.724555696310663, + "learning_rate": 1.3338842975206611e-06, + "logits/chosen": -0.08251953125, + "logits/rejected": -0.275390625, + "logps/chosen": -382.0, + "logps/rejected": -332.0, + "loss": 0.2357, + "rewards/accuracies": 0.8828125, + "rewards/chosen": 0.033203125, + "rewards/margins": 6.78125, + "rewards/rejected": -6.71875, + "step": 293 + }, + { + "epoch": 0.36073619631901843, + "grad_norm": 26.756162333522596, + "learning_rate": 1.3332644628099175e-06, + "logits/chosen": -0.359375, + "logits/rejected": -0.4921875, + "logps/chosen": -472.0, + "logps/rejected": -368.0, + "loss": 0.3094, + "rewards/accuracies": 0.8515625, + "rewards/chosen": 0.66796875, + "rewards/margins": 7.65625, + "rewards/rejected": -7.0, + "step": 294 + }, + { + "epoch": 0.3619631901840491, + "grad_norm": 27.599547481570436, + "learning_rate": 1.3326446280991736e-06, + "logits/chosen": -0.04345703125, + "logits/rejected": -0.1884765625, + "logps/chosen": -356.0, + "logps/rejected": -332.0, + "loss": 0.2495, + "rewards/accuracies": 0.8828125, + "rewards/chosen": 0.2734375, + "rewards/margins": 6.875, + "rewards/rejected": -6.59375, + "step": 295 + }, + { + "epoch": 0.36319018404907977, + "grad_norm": 35.1081345543023, + "learning_rate": 1.3320247933884298e-06, + "logits/chosen": -0.007598876953125, + "logits/rejected": -0.154296875, + "logps/chosen": -420.0, + "logps/rejected": -338.0, + "loss": 0.3852, + "rewards/accuracies": 0.859375, + "rewards/chosen": 1.1640625, + "rewards/margins": 6.6875, + "rewards/rejected": -5.5, + "step": 296 + }, + { + "epoch": 0.36441717791411044, + "grad_norm": 31.495501539039623, + "learning_rate": 1.331404958677686e-06, + "logits/chosen": -0.154296875, + "logits/rejected": -0.30859375, + "logps/chosen": -360.0, + "logps/rejected": -322.0, + "loss": 0.2553, + "rewards/accuracies": 0.859375, + "rewards/chosen": 0.79296875, + "rewards/margins": 7.0625, + "rewards/rejected": -6.25, + "step": 297 + }, + { + "epoch": 0.3656441717791411, + "grad_norm": 37.56929706779854, + "learning_rate": 1.330785123966942e-06, + "logits/chosen": -0.009521484375, + "logits/rejected": -0.09716796875, + "logps/chosen": -384.0, + "logps/rejected": -336.0, + "loss": 0.3993, + "rewards/accuracies": 0.8203125, + "rewards/chosen": 0.498046875, + "rewards/margins": 6.03125, + "rewards/rejected": -5.53125, + "step": 298 + }, + { + "epoch": 0.36687116564417177, + "grad_norm": 36.46646666012725, + "learning_rate": 1.3301652892561984e-06, + "logits/chosen": -0.058837890625, + "logits/rejected": -0.228515625, + "logps/chosen": -400.0, + "logps/rejected": -334.0, + "loss": 0.3895, + "rewards/accuracies": 0.84375, + "rewards/chosen": -0.1650390625, + "rewards/margins": 5.78125, + "rewards/rejected": -5.9375, + "step": 299 + }, + { + "epoch": 0.36809815950920244, + "grad_norm": 37.01196046043553, + "learning_rate": 1.3295454545454546e-06, + "logits/chosen": -0.173828125, + "logits/rejected": -0.3203125, + "logps/chosen": -372.0, + "logps/rejected": -328.0, + "loss": 0.4644, + "rewards/accuracies": 0.8203125, + "rewards/chosen": 0.482421875, + "rewards/margins": 6.65625, + "rewards/rejected": -6.15625, + "step": 300 + }, + { + "epoch": 0.3693251533742331, + "grad_norm": 32.17417313364543, + "learning_rate": 1.3289256198347107e-06, + "logits/chosen": -0.11865234375, + "logits/rejected": -0.298828125, + "logps/chosen": -460.0, + "logps/rejected": -386.0, + "loss": 0.3417, + "rewards/accuracies": 0.828125, + "rewards/chosen": 0.8515625, + "rewards/margins": 7.1875, + "rewards/rejected": -6.34375, + "step": 301 + }, + { + "epoch": 0.37055214723926383, + "grad_norm": 25.798891470824255, + "learning_rate": 1.328305785123967e-06, + "logits/chosen": -0.150390625, + "logits/rejected": -0.34765625, + "logps/chosen": -382.0, + "logps/rejected": -316.0, + "loss": 0.2452, + "rewards/accuracies": 0.8984375, + "rewards/chosen": 0.287109375, + "rewards/margins": 6.6875, + "rewards/rejected": -6.40625, + "step": 302 + }, + { + "epoch": 0.3717791411042945, + "grad_norm": 40.25697610802204, + "learning_rate": 1.3276859504132232e-06, + "logits/chosen": -0.1474609375, + "logits/rejected": -0.2578125, + "logps/chosen": -388.0, + "logps/rejected": -376.0, + "loss": 0.4569, + "rewards/accuracies": 0.78125, + "rewards/chosen": -0.466796875, + "rewards/margins": 6.1875, + "rewards/rejected": -6.65625, + "step": 303 + }, + { + "epoch": 0.37300613496932516, + "grad_norm": 30.538056754282714, + "learning_rate": 1.3270661157024793e-06, + "logits/chosen": -0.0264892578125, + "logits/rejected": -0.287109375, + "logps/chosen": -396.0, + "logps/rejected": -334.0, + "loss": 0.2974, + "rewards/accuracies": 0.875, + "rewards/chosen": -0.140625, + "rewards/margins": 6.875, + "rewards/rejected": -7.0, + "step": 304 + }, + { + "epoch": 0.37423312883435583, + "grad_norm": 27.228440712710828, + "learning_rate": 1.3264462809917357e-06, + "logits/chosen": -0.109375, + "logits/rejected": -0.2490234375, + "logps/chosen": -416.0, + "logps/rejected": -344.0, + "loss": 0.2623, + "rewards/accuracies": 0.90625, + "rewards/chosen": 0.341796875, + "rewards/margins": 7.71875, + "rewards/rejected": -7.375, + "step": 305 + }, + { + "epoch": 0.3754601226993865, + "grad_norm": 34.02132219179046, + "learning_rate": 1.3258264462809916e-06, + "logits/chosen": -0.2177734375, + "logits/rejected": -0.37890625, + "logps/chosen": -430.0, + "logps/rejected": -364.0, + "loss": 0.2805, + "rewards/accuracies": 0.8828125, + "rewards/chosen": -0.33203125, + "rewards/margins": 6.75, + "rewards/rejected": -7.0625, + "step": 306 + }, + { + "epoch": 0.37668711656441717, + "grad_norm": 34.60907155238812, + "learning_rate": 1.325206611570248e-06, + "logits/chosen": -0.06103515625, + "logits/rejected": -0.19140625, + "logps/chosen": -382.0, + "logps/rejected": -344.0, + "loss": 0.4791, + "rewards/accuracies": 0.796875, + "rewards/chosen": -0.18359375, + "rewards/margins": 6.90625, + "rewards/rejected": -7.09375, + "step": 307 + }, + { + "epoch": 0.37791411042944784, + "grad_norm": 34.523248358542865, + "learning_rate": 1.3245867768595041e-06, + "logits/chosen": -0.00148773193359375, + "logits/rejected": -0.162109375, + "logps/chosen": -398.0, + "logps/rejected": -360.0, + "loss": 0.3857, + "rewards/accuracies": 0.828125, + "rewards/chosen": -0.53515625, + "rewards/margins": 6.0, + "rewards/rejected": -6.5625, + "step": 308 + }, + { + "epoch": 0.3791411042944785, + "grad_norm": 29.962650665710807, + "learning_rate": 1.3239669421487603e-06, + "logits/chosen": -0.208984375, + "logits/rejected": -0.31640625, + "logps/chosen": -388.0, + "logps/rejected": -354.0, + "loss": 0.2543, + "rewards/accuracies": 0.875, + "rewards/chosen": 0.609375, + "rewards/margins": 7.5625, + "rewards/rejected": -6.9375, + "step": 309 + }, + { + "epoch": 0.3803680981595092, + "grad_norm": 33.20481159292772, + "learning_rate": 1.3233471074380166e-06, + "logits/chosen": 0.051513671875, + "logits/rejected": -0.037353515625, + "logps/chosen": -322.0, + "logps/rejected": -312.0, + "loss": 0.4174, + "rewards/accuracies": 0.8359375, + "rewards/chosen": -0.875, + "rewards/margins": 5.5, + "rewards/rejected": -6.375, + "step": 310 + }, + { + "epoch": 0.3815950920245399, + "grad_norm": 32.36453885140526, + "learning_rate": 1.3227272727272727e-06, + "logits/chosen": -0.158203125, + "logits/rejected": -0.3359375, + "logps/chosen": -392.0, + "logps/rejected": -348.0, + "loss": 0.3525, + "rewards/accuracies": 0.8203125, + "rewards/chosen": 0.486328125, + "rewards/margins": 7.78125, + "rewards/rejected": -7.28125, + "step": 311 + }, + { + "epoch": 0.38282208588957056, + "grad_norm": 39.08052846369179, + "learning_rate": 1.322107438016529e-06, + "logits/chosen": 0.0162353515625, + "logits/rejected": -0.1142578125, + "logps/chosen": -392.0, + "logps/rejected": -378.0, + "loss": 0.5608, + "rewards/accuracies": 0.7421875, + "rewards/chosen": 0.0712890625, + "rewards/margins": 6.34375, + "rewards/rejected": -6.28125, + "step": 312 + }, + { + "epoch": 0.38404907975460123, + "grad_norm": 35.15009453427041, + "learning_rate": 1.3214876033057852e-06, + "logits/chosen": -0.0235595703125, + "logits/rejected": -0.298828125, + "logps/chosen": -410.0, + "logps/rejected": -372.0, + "loss": 0.383, + "rewards/accuracies": 0.84375, + "rewards/chosen": 1.1640625, + "rewards/margins": 8.0625, + "rewards/rejected": -6.9375, + "step": 313 + }, + { + "epoch": 0.3852760736196319, + "grad_norm": 28.038472021341548, + "learning_rate": 1.3208677685950414e-06, + "logits/chosen": -0.05517578125, + "logits/rejected": -0.2421875, + "logps/chosen": -458.0, + "logps/rejected": -360.0, + "loss": 0.2315, + "rewards/accuracies": 0.8828125, + "rewards/chosen": 2.03125, + "rewards/margins": 9.375, + "rewards/rejected": -7.34375, + "step": 314 + }, + { + "epoch": 0.38650306748466257, + "grad_norm": 31.702508477036716, + "learning_rate": 1.3202479338842975e-06, + "logits/chosen": 0.076171875, + "logits/rejected": -0.11865234375, + "logps/chosen": -350.0, + "logps/rejected": -300.0, + "loss": 0.3293, + "rewards/accuracies": 0.8203125, + "rewards/chosen": 0.2431640625, + "rewards/margins": 6.84375, + "rewards/rejected": -6.59375, + "step": 315 + }, + { + "epoch": 0.38773006134969323, + "grad_norm": 27.443273044054905, + "learning_rate": 1.3196280991735537e-06, + "logits/chosen": 0.05419921875, + "logits/rejected": -0.08544921875, + "logps/chosen": -360.0, + "logps/rejected": -308.0, + "loss": 0.3129, + "rewards/accuracies": 0.875, + "rewards/chosen": 0.75390625, + "rewards/margins": 7.5625, + "rewards/rejected": -6.8125, + "step": 316 + }, + { + "epoch": 0.3889570552147239, + "grad_norm": 28.339536052474916, + "learning_rate": 1.31900826446281e-06, + "logits/chosen": -0.09326171875, + "logits/rejected": -0.3046875, + "logps/chosen": -400.0, + "logps/rejected": -368.0, + "loss": 0.2873, + "rewards/accuracies": 0.859375, + "rewards/chosen": 0.94140625, + "rewards/margins": 8.125, + "rewards/rejected": -7.15625, + "step": 317 + }, + { + "epoch": 0.3901840490797546, + "grad_norm": 38.792002326692455, + "learning_rate": 1.3183884297520662e-06, + "logits/chosen": -0.142578125, + "logits/rejected": -0.296875, + "logps/chosen": -402.0, + "logps/rejected": -366.0, + "loss": 0.4379, + "rewards/accuracies": 0.828125, + "rewards/chosen": 0.91796875, + "rewards/margins": 7.0625, + "rewards/rejected": -6.125, + "step": 318 + }, + { + "epoch": 0.3914110429447853, + "grad_norm": 17.992013049840406, + "learning_rate": 1.3177685950413223e-06, + "logits/chosen": -0.068359375, + "logits/rejected": -0.255859375, + "logps/chosen": -414.0, + "logps/rejected": -320.0, + "loss": 0.1471, + "rewards/accuracies": 0.9296875, + "rewards/chosen": 1.1640625, + "rewards/margins": 8.8125, + "rewards/rejected": -7.65625, + "step": 319 + }, + { + "epoch": 0.39263803680981596, + "grad_norm": 31.193757820598883, + "learning_rate": 1.3171487603305787e-06, + "logits/chosen": -0.056884765625, + "logits/rejected": -0.166015625, + "logps/chosen": -320.0, + "logps/rejected": -286.0, + "loss": 0.325, + "rewards/accuracies": 0.859375, + "rewards/chosen": 0.31640625, + "rewards/margins": 6.84375, + "rewards/rejected": -6.53125, + "step": 320 + }, + { + "epoch": 0.39386503067484663, + "grad_norm": 39.71075166734815, + "learning_rate": 1.3165289256198348e-06, + "logits/chosen": -0.201171875, + "logits/rejected": -0.35546875, + "logps/chosen": -410.0, + "logps/rejected": -398.0, + "loss": 0.3784, + "rewards/accuracies": 0.8515625, + "rewards/chosen": 0.439453125, + "rewards/margins": 7.84375, + "rewards/rejected": -7.375, + "step": 321 + }, + { + "epoch": 0.3950920245398773, + "grad_norm": 26.7824487712344, + "learning_rate": 1.315909090909091e-06, + "logits/chosen": -0.107421875, + "logits/rejected": -0.2578125, + "logps/chosen": -394.0, + "logps/rejected": -330.0, + "loss": 0.2893, + "rewards/accuracies": 0.8984375, + "rewards/chosen": 1.203125, + "rewards/margins": 7.84375, + "rewards/rejected": -6.65625, + "step": 322 + }, + { + "epoch": 0.39631901840490796, + "grad_norm": 36.45438287117724, + "learning_rate": 1.3152892561983473e-06, + "logits/chosen": -0.1318359375, + "logits/rejected": -0.265625, + "logps/chosen": -354.0, + "logps/rejected": -334.0, + "loss": 0.4474, + "rewards/accuracies": 0.8203125, + "rewards/chosen": 0.6875, + "rewards/margins": 6.15625, + "rewards/rejected": -5.46875, + "step": 323 + }, + { + "epoch": 0.39754601226993863, + "grad_norm": 31.15880722522094, + "learning_rate": 1.3146694214876032e-06, + "logits/chosen": -0.193359375, + "logits/rejected": -0.361328125, + "logps/chosen": -356.0, + "logps/rejected": -328.0, + "loss": 0.3486, + "rewards/accuracies": 0.8515625, + "rewards/chosen": 0.62890625, + "rewards/margins": 7.4375, + "rewards/rejected": -6.8125, + "step": 324 + }, + { + "epoch": 0.3987730061349693, + "grad_norm": 36.47850348671778, + "learning_rate": 1.3140495867768596e-06, + "logits/chosen": -0.0272216796875, + "logits/rejected": -0.09326171875, + "logps/chosen": -330.0, + "logps/rejected": -304.0, + "loss": 0.5542, + "rewards/accuracies": 0.8046875, + "rewards/chosen": 0.33203125, + "rewards/margins": 6.15625, + "rewards/rejected": -5.84375, + "step": 325 + }, + { + "epoch": 0.4, + "grad_norm": 26.62936328977598, + "learning_rate": 1.3134297520661157e-06, + "logits/chosen": -0.2294921875, + "logits/rejected": -0.369140625, + "logps/chosen": -398.0, + "logps/rejected": -374.0, + "loss": 0.1993, + "rewards/accuracies": 0.8671875, + "rewards/chosen": 1.5546875, + "rewards/margins": 8.375, + "rewards/rejected": -6.84375, + "step": 326 + }, + { + "epoch": 0.4012269938650307, + "grad_norm": 39.76264342514329, + "learning_rate": 1.3128099173553719e-06, + "logits/chosen": -0.1015625, + "logits/rejected": -0.271484375, + "logps/chosen": -430.0, + "logps/rejected": -352.0, + "loss": 0.4557, + "rewards/accuracies": 0.8046875, + "rewards/chosen": 0.921875, + "rewards/margins": 6.125, + "rewards/rejected": -5.1875, + "step": 327 + }, + { + "epoch": 0.40245398773006136, + "grad_norm": 33.62855666124103, + "learning_rate": 1.3121900826446282e-06, + "logits/chosen": -0.1328125, + "logits/rejected": -0.259765625, + "logps/chosen": -362.0, + "logps/rejected": -288.0, + "loss": 0.363, + "rewards/accuracies": 0.84375, + "rewards/chosen": 0.228515625, + "rewards/margins": 6.15625, + "rewards/rejected": -5.9375, + "step": 328 + }, + { + "epoch": 0.403680981595092, + "grad_norm": 36.1359556726495, + "learning_rate": 1.3115702479338844e-06, + "logits/chosen": -0.08203125, + "logits/rejected": -0.283203125, + "logps/chosen": -466.0, + "logps/rejected": -386.0, + "loss": 0.3754, + "rewards/accuracies": 0.796875, + "rewards/chosen": 1.078125, + "rewards/margins": 8.3125, + "rewards/rejected": -7.25, + "step": 329 + }, + { + "epoch": 0.4049079754601227, + "grad_norm": 27.601952208856858, + "learning_rate": 1.3109504132231405e-06, + "logits/chosen": -0.22265625, + "logits/rejected": -0.423828125, + "logps/chosen": -418.0, + "logps/rejected": -352.0, + "loss": 0.2524, + "rewards/accuracies": 0.8515625, + "rewards/chosen": 1.0078125, + "rewards/margins": 8.25, + "rewards/rejected": -7.25, + "step": 330 + }, + { + "epoch": 0.40613496932515336, + "grad_norm": 26.526762211128467, + "learning_rate": 1.3103305785123968e-06, + "logits/chosen": -0.1767578125, + "logits/rejected": -0.283203125, + "logps/chosen": -408.0, + "logps/rejected": -356.0, + "loss": 0.2755, + "rewards/accuracies": 0.859375, + "rewards/chosen": 1.09375, + "rewards/margins": 7.625, + "rewards/rejected": -6.53125, + "step": 331 + }, + { + "epoch": 0.40736196319018403, + "grad_norm": 24.05844405102659, + "learning_rate": 1.3097107438016528e-06, + "logits/chosen": -0.1826171875, + "logits/rejected": -0.37109375, + "logps/chosen": -430.0, + "logps/rejected": -332.0, + "loss": 0.2196, + "rewards/accuracies": 0.8984375, + "rewards/chosen": 1.109375, + "rewards/margins": 7.96875, + "rewards/rejected": -6.84375, + "step": 332 + }, + { + "epoch": 0.4085889570552147, + "grad_norm": 36.41468851723906, + "learning_rate": 1.3090909090909091e-06, + "logits/chosen": 0.017822265625, + "logits/rejected": -0.1044921875, + "logps/chosen": -372.0, + "logps/rejected": -346.0, + "loss": 0.4229, + "rewards/accuracies": 0.7734375, + "rewards/chosen": 0.48046875, + "rewards/margins": 6.40625, + "rewards/rejected": -5.90625, + "step": 333 + }, + { + "epoch": 0.4098159509202454, + "grad_norm": 30.33603786358444, + "learning_rate": 1.3084710743801653e-06, + "logits/chosen": -0.1181640625, + "logits/rejected": -0.349609375, + "logps/chosen": -382.0, + "logps/rejected": -366.0, + "loss": 0.3109, + "rewards/accuracies": 0.8671875, + "rewards/chosen": 0.458984375, + "rewards/margins": 7.75, + "rewards/rejected": -7.28125, + "step": 334 + }, + { + "epoch": 0.4110429447852761, + "grad_norm": 33.01078113652063, + "learning_rate": 1.3078512396694214e-06, + "logits/chosen": -0.0213623046875, + "logits/rejected": -0.28125, + "logps/chosen": -432.0, + "logps/rejected": -332.0, + "loss": 0.2319, + "rewards/accuracies": 0.859375, + "rewards/chosen": 0.34375, + "rewards/margins": 7.0625, + "rewards/rejected": -6.71875, + "step": 335 + }, + { + "epoch": 0.41226993865030676, + "grad_norm": 31.965240510538873, + "learning_rate": 1.3072314049586778e-06, + "logits/chosen": -0.177734375, + "logits/rejected": -0.326171875, + "logps/chosen": -402.0, + "logps/rejected": -352.0, + "loss": 0.3352, + "rewards/accuracies": 0.859375, + "rewards/chosen": 0.3515625, + "rewards/margins": 7.15625, + "rewards/rejected": -6.8125, + "step": 336 + }, + { + "epoch": 0.4134969325153374, + "grad_norm": 36.98349242868308, + "learning_rate": 1.306611570247934e-06, + "logits/chosen": 0.0264892578125, + "logits/rejected": -0.1318359375, + "logps/chosen": -358.0, + "logps/rejected": -352.0, + "loss": 0.4393, + "rewards/accuracies": 0.796875, + "rewards/chosen": 0.1748046875, + "rewards/margins": 6.90625, + "rewards/rejected": -6.75, + "step": 337 + }, + { + "epoch": 0.4147239263803681, + "grad_norm": 31.08894150655132, + "learning_rate": 1.30599173553719e-06, + "logits/chosen": -0.005157470703125, + "logits/rejected": -0.1484375, + "logps/chosen": -392.0, + "logps/rejected": -354.0, + "loss": 0.3686, + "rewards/accuracies": 0.8515625, + "rewards/chosen": 0.2216796875, + "rewards/margins": 6.96875, + "rewards/rejected": -6.75, + "step": 338 + }, + { + "epoch": 0.41595092024539876, + "grad_norm": 34.40525022311832, + "learning_rate": 1.3053719008264464e-06, + "logits/chosen": 0.0654296875, + "logits/rejected": -0.06103515625, + "logps/chosen": -378.0, + "logps/rejected": -346.0, + "loss": 0.3699, + "rewards/accuracies": 0.796875, + "rewards/chosen": -0.02880859375, + "rewards/margins": 6.53125, + "rewards/rejected": -6.5625, + "step": 339 + }, + { + "epoch": 0.4171779141104294, + "grad_norm": 35.90111385254676, + "learning_rate": 1.3047520661157025e-06, + "logits/chosen": -0.03759765625, + "logits/rejected": -0.267578125, + "logps/chosen": -366.0, + "logps/rejected": -350.0, + "loss": 0.3448, + "rewards/accuracies": 0.828125, + "rewards/chosen": -0.058349609375, + "rewards/margins": 6.125, + "rewards/rejected": -6.1875, + "step": 340 + }, + { + "epoch": 0.41840490797546015, + "grad_norm": 27.984626902771293, + "learning_rate": 1.3041322314049587e-06, + "logits/chosen": -0.09521484375, + "logits/rejected": -0.3125, + "logps/chosen": -398.0, + "logps/rejected": -334.0, + "loss": 0.344, + "rewards/accuracies": 0.8515625, + "rewards/chosen": 0.56640625, + "rewards/margins": 7.3125, + "rewards/rejected": -6.75, + "step": 341 + }, + { + "epoch": 0.4196319018404908, + "grad_norm": 34.66561631382139, + "learning_rate": 1.3035123966942148e-06, + "logits/chosen": -0.021240234375, + "logits/rejected": -0.119140625, + "logps/chosen": -340.0, + "logps/rejected": -318.0, + "loss": 0.3745, + "rewards/accuracies": 0.78125, + "rewards/chosen": 0.049072265625, + "rewards/margins": 6.09375, + "rewards/rejected": -6.03125, + "step": 342 + }, + { + "epoch": 0.4208588957055215, + "grad_norm": 34.35806991693024, + "learning_rate": 1.302892561983471e-06, + "logits/chosen": -0.06396484375, + "logits/rejected": -0.1513671875, + "logps/chosen": -386.0, + "logps/rejected": -356.0, + "loss": 0.4138, + "rewards/accuracies": 0.8046875, + "rewards/chosen": -0.251953125, + "rewards/margins": 5.53125, + "rewards/rejected": -5.8125, + "step": 343 + }, + { + "epoch": 0.42208588957055215, + "grad_norm": 36.57762286698119, + "learning_rate": 1.3022727272727273e-06, + "logits/chosen": -0.0654296875, + "logits/rejected": -0.2041015625, + "logps/chosen": -408.0, + "logps/rejected": -340.0, + "loss": 0.4341, + "rewards/accuracies": 0.8125, + "rewards/chosen": 0.036376953125, + "rewards/margins": 5.875, + "rewards/rejected": -5.84375, + "step": 344 + }, + { + "epoch": 0.4233128834355828, + "grad_norm": 32.698706365667604, + "learning_rate": 1.3016528925619835e-06, + "logits/chosen": -0.04931640625, + "logits/rejected": -0.2265625, + "logps/chosen": -414.0, + "logps/rejected": -340.0, + "loss": 0.3793, + "rewards/accuracies": 0.84375, + "rewards/chosen": 0.2177734375, + "rewards/margins": 6.4375, + "rewards/rejected": -6.21875, + "step": 345 + }, + { + "epoch": 0.4245398773006135, + "grad_norm": 34.95248105852957, + "learning_rate": 1.3010330578512398e-06, + "logits/chosen": -0.036376953125, + "logits/rejected": -0.23046875, + "logps/chosen": -344.0, + "logps/rejected": -306.0, + "loss": 0.4481, + "rewards/accuracies": 0.828125, + "rewards/chosen": -0.19921875, + "rewards/margins": 5.84375, + "rewards/rejected": -6.03125, + "step": 346 + }, + { + "epoch": 0.42576687116564416, + "grad_norm": 27.628512071923325, + "learning_rate": 1.300413223140496e-06, + "logits/chosen": -0.06884765625, + "logits/rejected": -0.1875, + "logps/chosen": -370.0, + "logps/rejected": -356.0, + "loss": 0.317, + "rewards/accuracies": 0.8515625, + "rewards/chosen": 0.20703125, + "rewards/margins": 6.125, + "rewards/rejected": -5.90625, + "step": 347 + }, + { + "epoch": 0.4269938650306748, + "grad_norm": 24.80952384589789, + "learning_rate": 1.299793388429752e-06, + "logits/chosen": -0.12451171875, + "logits/rejected": -0.318359375, + "logps/chosen": -418.0, + "logps/rejected": -328.0, + "loss": 0.2563, + "rewards/accuracies": 0.8671875, + "rewards/chosen": 0.126953125, + "rewards/margins": 7.4375, + "rewards/rejected": -7.3125, + "step": 348 + }, + { + "epoch": 0.42822085889570555, + "grad_norm": 38.377577445863466, + "learning_rate": 1.2991735537190084e-06, + "logits/chosen": -0.0986328125, + "logits/rejected": -0.2412109375, + "logps/chosen": -406.0, + "logps/rejected": -322.0, + "loss": 0.4136, + "rewards/accuracies": 0.8046875, + "rewards/chosen": 0.046875, + "rewards/margins": 6.90625, + "rewards/rejected": -6.84375, + "step": 349 + }, + { + "epoch": 0.4294478527607362, + "grad_norm": 37.75213306544608, + "learning_rate": 1.2985537190082644e-06, + "logits/chosen": 0.06689453125, + "logits/rejected": -0.12890625, + "logps/chosen": -412.0, + "logps/rejected": -318.0, + "loss": 0.3697, + "rewards/accuracies": 0.796875, + "rewards/chosen": 0.045654296875, + "rewards/margins": 6.03125, + "rewards/rejected": -6.0, + "step": 350 + }, + { + "epoch": 0.4306748466257669, + "grad_norm": 35.53586893463626, + "learning_rate": 1.2979338842975207e-06, + "logits/chosen": -0.11279296875, + "logits/rejected": -0.251953125, + "logps/chosen": -342.0, + "logps/rejected": -326.0, + "loss": 0.4086, + "rewards/accuracies": 0.8125, + "rewards/chosen": -0.1337890625, + "rewards/margins": 6.03125, + "rewards/rejected": -6.1875, + "step": 351 + }, + { + "epoch": 0.43190184049079755, + "grad_norm": 30.59296136721169, + "learning_rate": 1.2973140495867769e-06, + "logits/chosen": -0.0230712890625, + "logits/rejected": -0.2177734375, + "logps/chosen": -386.0, + "logps/rejected": -334.0, + "loss": 0.4358, + "rewards/accuracies": 0.828125, + "rewards/chosen": 1.0625, + "rewards/margins": 6.34375, + "rewards/rejected": -5.28125, + "step": 352 + }, + { + "epoch": 0.4331288343558282, + "grad_norm": 24.76896622328251, + "learning_rate": 1.296694214876033e-06, + "logits/chosen": -0.07666015625, + "logits/rejected": -0.244140625, + "logps/chosen": -360.0, + "logps/rejected": -330.0, + "loss": 0.2411, + "rewards/accuracies": 0.859375, + "rewards/chosen": 0.921875, + "rewards/margins": 7.46875, + "rewards/rejected": -6.5625, + "step": 353 + }, + { + "epoch": 0.4343558282208589, + "grad_norm": 31.01854459387497, + "learning_rate": 1.2960743801652894e-06, + "logits/chosen": 0.01043701171875, + "logits/rejected": -0.162109375, + "logps/chosen": -444.0, + "logps/rejected": -378.0, + "loss": 0.3525, + "rewards/accuracies": 0.8359375, + "rewards/chosen": 0.99609375, + "rewards/margins": 6.53125, + "rewards/rejected": -5.53125, + "step": 354 + }, + { + "epoch": 0.43558282208588955, + "grad_norm": 32.456178590454805, + "learning_rate": 1.2954545454545455e-06, + "logits/chosen": -0.1572265625, + "logits/rejected": -0.24609375, + "logps/chosen": -348.0, + "logps/rejected": -356.0, + "loss": 0.4542, + "rewards/accuracies": 0.8671875, + "rewards/chosen": 0.296875, + "rewards/margins": 6.3125, + "rewards/rejected": -6.0, + "step": 355 + }, + { + "epoch": 0.4368098159509202, + "grad_norm": 31.595656958351967, + "learning_rate": 1.2948347107438016e-06, + "logits/chosen": 0.06396484375, + "logits/rejected": -0.0859375, + "logps/chosen": -396.0, + "logps/rejected": -332.0, + "loss": 0.4115, + "rewards/accuracies": 0.8046875, + "rewards/chosen": 0.46484375, + "rewards/margins": 6.0625, + "rewards/rejected": -5.59375, + "step": 356 + }, + { + "epoch": 0.43803680981595094, + "grad_norm": 28.72400215925028, + "learning_rate": 1.294214876033058e-06, + "logits/chosen": -0.23046875, + "logits/rejected": -0.384765625, + "logps/chosen": -402.0, + "logps/rejected": -340.0, + "loss": 0.3662, + "rewards/accuracies": 0.875, + "rewards/chosen": 0.63671875, + "rewards/margins": 7.40625, + "rewards/rejected": -6.78125, + "step": 357 + }, + { + "epoch": 0.4392638036809816, + "grad_norm": 29.5841283121426, + "learning_rate": 1.2935950413223141e-06, + "logits/chosen": -0.1708984375, + "logits/rejected": -0.328125, + "logps/chosen": -446.0, + "logps/rejected": -330.0, + "loss": 0.3109, + "rewards/accuracies": 0.8671875, + "rewards/chosen": 1.2109375, + "rewards/margins": 7.125, + "rewards/rejected": -5.90625, + "step": 358 + }, + { + "epoch": 0.4404907975460123, + "grad_norm": 33.37778751879116, + "learning_rate": 1.2929752066115703e-06, + "logits/chosen": -0.043212890625, + "logits/rejected": -0.154296875, + "logps/chosen": -424.0, + "logps/rejected": -334.0, + "loss": 0.3631, + "rewards/accuracies": 0.8515625, + "rewards/chosen": 0.89453125, + "rewards/margins": 6.71875, + "rewards/rejected": -5.8125, + "step": 359 + }, + { + "epoch": 0.44171779141104295, + "grad_norm": 32.72250588219352, + "learning_rate": 1.2923553719008264e-06, + "logits/chosen": -0.02587890625, + "logits/rejected": -0.1484375, + "logps/chosen": -396.0, + "logps/rejected": -358.0, + "loss": 0.2778, + "rewards/accuracies": 0.8515625, + "rewards/chosen": 0.33984375, + "rewards/margins": 6.9375, + "rewards/rejected": -6.59375, + "step": 360 + }, + { + "epoch": 0.4429447852760736, + "grad_norm": 29.99311979639803, + "learning_rate": 1.2917355371900826e-06, + "logits/chosen": 0.010498046875, + "logits/rejected": -0.177734375, + "logps/chosen": -438.0, + "logps/rejected": -382.0, + "loss": 0.3143, + "rewards/accuracies": 0.859375, + "rewards/chosen": 0.6015625, + "rewards/margins": 7.5, + "rewards/rejected": -6.875, + "step": 361 + }, + { + "epoch": 0.4441717791411043, + "grad_norm": 30.82537911112007, + "learning_rate": 1.291115702479339e-06, + "logits/chosen": -0.11083984375, + "logits/rejected": -0.267578125, + "logps/chosen": -418.0, + "logps/rejected": -344.0, + "loss": 0.3606, + "rewards/accuracies": 0.8203125, + "rewards/chosen": 0.62109375, + "rewards/margins": 7.21875, + "rewards/rejected": -6.59375, + "step": 362 + }, + { + "epoch": 0.44539877300613495, + "grad_norm": 26.541431319509446, + "learning_rate": 1.290495867768595e-06, + "logits/chosen": -0.018798828125, + "logits/rejected": -0.171875, + "logps/chosen": -392.0, + "logps/rejected": -344.0, + "loss": 0.2712, + "rewards/accuracies": 0.8984375, + "rewards/chosen": 0.77734375, + "rewards/margins": 6.8125, + "rewards/rejected": -6.03125, + "step": 363 + }, + { + "epoch": 0.4466257668711656, + "grad_norm": 36.761947246861894, + "learning_rate": 1.2898760330578512e-06, + "logits/chosen": 0.06591796875, + "logits/rejected": -0.109375, + "logps/chosen": -392.0, + "logps/rejected": -314.0, + "loss": 0.3953, + "rewards/accuracies": 0.8671875, + "rewards/chosen": 0.765625, + "rewards/margins": 6.75, + "rewards/rejected": -6.0, + "step": 364 + }, + { + "epoch": 0.44785276073619634, + "grad_norm": 27.87437703363369, + "learning_rate": 1.2892561983471076e-06, + "logits/chosen": -0.0018463134765625, + "logits/rejected": -0.1435546875, + "logps/chosen": -352.0, + "logps/rejected": -314.0, + "loss": 0.3086, + "rewards/accuracies": 0.8828125, + "rewards/chosen": 0.53125, + "rewards/margins": 6.65625, + "rewards/rejected": -6.125, + "step": 365 + }, + { + "epoch": 0.449079754601227, + "grad_norm": 32.787747807253844, + "learning_rate": 1.2886363636363637e-06, + "logits/chosen": 0.0291748046875, + "logits/rejected": -0.06787109375, + "logps/chosen": -366.0, + "logps/rejected": -332.0, + "loss": 0.447, + "rewards/accuracies": 0.8203125, + "rewards/chosen": 0.87890625, + "rewards/margins": 6.0, + "rewards/rejected": -5.125, + "step": 366 + }, + { + "epoch": 0.4503067484662577, + "grad_norm": 26.662012319210124, + "learning_rate": 1.2880165289256198e-06, + "logits/chosen": -0.142578125, + "logits/rejected": -0.326171875, + "logps/chosen": -394.0, + "logps/rejected": -336.0, + "loss": 0.2357, + "rewards/accuracies": 0.8984375, + "rewards/chosen": 0.9453125, + "rewards/margins": 6.65625, + "rewards/rejected": -5.71875, + "step": 367 + }, + { + "epoch": 0.45153374233128835, + "grad_norm": 36.83996177643189, + "learning_rate": 1.287396694214876e-06, + "logits/chosen": 0.06787109375, + "logits/rejected": 0.0234375, + "logps/chosen": -368.0, + "logps/rejected": -328.0, + "loss": 0.3909, + "rewards/accuracies": 0.84375, + "rewards/chosen": 0.6796875, + "rewards/margins": 6.0, + "rewards/rejected": -5.3125, + "step": 368 + }, + { + "epoch": 0.452760736196319, + "grad_norm": 35.726856908761434, + "learning_rate": 1.2867768595041321e-06, + "logits/chosen": 0.05322265625, + "logits/rejected": -0.06591796875, + "logps/chosen": -344.0, + "logps/rejected": -298.0, + "loss": 0.3738, + "rewards/accuracies": 0.8359375, + "rewards/chosen": 0.169921875, + "rewards/margins": 5.71875, + "rewards/rejected": -5.5625, + "step": 369 + }, + { + "epoch": 0.4539877300613497, + "grad_norm": 31.108323501043053, + "learning_rate": 1.2861570247933885e-06, + "logits/chosen": -0.1484375, + "logits/rejected": -0.328125, + "logps/chosen": -402.0, + "logps/rejected": -348.0, + "loss": 0.369, + "rewards/accuracies": 0.84375, + "rewards/chosen": 1.75, + "rewards/margins": 6.65625, + "rewards/rejected": -4.90625, + "step": 370 + }, + { + "epoch": 0.45521472392638035, + "grad_norm": 31.06340590217051, + "learning_rate": 1.2855371900826446e-06, + "logits/chosen": -0.130859375, + "logits/rejected": -0.298828125, + "logps/chosen": -460.0, + "logps/rejected": -364.0, + "loss": 0.3181, + "rewards/accuracies": 0.8984375, + "rewards/chosen": 1.625, + "rewards/margins": 6.0, + "rewards/rejected": -4.375, + "step": 371 + }, + { + "epoch": 0.456441717791411, + "grad_norm": 28.189166139003635, + "learning_rate": 1.2849173553719008e-06, + "logits/chosen": -0.0301513671875, + "logits/rejected": -0.1875, + "logps/chosen": -410.0, + "logps/rejected": -324.0, + "loss": 0.2419, + "rewards/accuracies": 0.890625, + "rewards/chosen": 1.5078125, + "rewards/margins": 6.4375, + "rewards/rejected": -4.9375, + "step": 372 + }, + { + "epoch": 0.45766871165644174, + "grad_norm": 33.21287873989905, + "learning_rate": 1.2842975206611571e-06, + "logits/chosen": -0.053466796875, + "logits/rejected": -0.1591796875, + "logps/chosen": -362.0, + "logps/rejected": -322.0, + "loss": 0.4071, + "rewards/accuracies": 0.8359375, + "rewards/chosen": 0.58203125, + "rewards/margins": 5.71875, + "rewards/rejected": -5.125, + "step": 373 + }, + { + "epoch": 0.4588957055214724, + "grad_norm": 26.151862774833457, + "learning_rate": 1.2836776859504132e-06, + "logits/chosen": -0.0908203125, + "logits/rejected": -0.2197265625, + "logps/chosen": -382.0, + "logps/rejected": -316.0, + "loss": 0.2534, + "rewards/accuracies": 0.8984375, + "rewards/chosen": 2.0, + "rewards/margins": 6.78125, + "rewards/rejected": -4.78125, + "step": 374 + }, + { + "epoch": 0.4601226993865031, + "grad_norm": 32.35488575008133, + "learning_rate": 1.2830578512396696e-06, + "logits/chosen": -0.14453125, + "logits/rejected": -0.2177734375, + "logps/chosen": -392.0, + "logps/rejected": -350.0, + "loss": 0.2771, + "rewards/accuracies": 0.859375, + "rewards/chosen": 1.7109375, + "rewards/margins": 5.8125, + "rewards/rejected": -4.09375, + "step": 375 + }, + { + "epoch": 0.46134969325153374, + "grad_norm": 34.001922340864844, + "learning_rate": 1.2824380165289257e-06, + "logits/chosen": -0.00946044921875, + "logits/rejected": -0.15625, + "logps/chosen": -376.0, + "logps/rejected": -344.0, + "loss": 0.3408, + "rewards/accuracies": 0.8515625, + "rewards/chosen": 1.6953125, + "rewards/margins": 5.75, + "rewards/rejected": -4.03125, + "step": 376 + }, + { + "epoch": 0.4625766871165644, + "grad_norm": 35.73955039729585, + "learning_rate": 1.2818181818181817e-06, + "logits/chosen": -0.0927734375, + "logits/rejected": -0.150390625, + "logps/chosen": -352.0, + "logps/rejected": -318.0, + "loss": 0.4247, + "rewards/accuracies": 0.7734375, + "rewards/chosen": 1.5625, + "rewards/margins": 6.65625, + "rewards/rejected": -5.0625, + "step": 377 + }, + { + "epoch": 0.4638036809815951, + "grad_norm": 33.82458737973786, + "learning_rate": 1.281198347107438e-06, + "logits/chosen": -0.1396484375, + "logits/rejected": -0.376953125, + "logps/chosen": -376.0, + "logps/rejected": -332.0, + "loss": 0.3987, + "rewards/accuracies": 0.796875, + "rewards/chosen": 1.625, + "rewards/margins": 6.625, + "rewards/rejected": -5.0, + "step": 378 + }, + { + "epoch": 0.46503067484662575, + "grad_norm": 31.86429872362659, + "learning_rate": 1.2805785123966942e-06, + "logits/chosen": -0.1953125, + "logits/rejected": -0.359375, + "logps/chosen": -408.0, + "logps/rejected": -344.0, + "loss": 0.3028, + "rewards/accuracies": 0.8671875, + "rewards/chosen": 1.71875, + "rewards/margins": 7.1875, + "rewards/rejected": -5.46875, + "step": 379 + }, + { + "epoch": 0.4662576687116564, + "grad_norm": 31.446849732786838, + "learning_rate": 1.2799586776859505e-06, + "logits/chosen": -0.09326171875, + "logits/rejected": -0.2314453125, + "logps/chosen": -382.0, + "logps/rejected": -326.0, + "loss": 0.3068, + "rewards/accuracies": 0.859375, + "rewards/chosen": 1.625, + "rewards/margins": 6.78125, + "rewards/rejected": -5.15625, + "step": 380 + }, + { + "epoch": 0.46748466257668714, + "grad_norm": 31.679856008810177, + "learning_rate": 1.2793388429752067e-06, + "logits/chosen": -0.189453125, + "logits/rejected": -0.3671875, + "logps/chosen": -372.0, + "logps/rejected": -328.0, + "loss": 0.374, + "rewards/accuracies": 0.8515625, + "rewards/chosen": 1.046875, + "rewards/margins": 7.0, + "rewards/rejected": -5.9375, + "step": 381 + }, + { + "epoch": 0.4687116564417178, + "grad_norm": 31.436122523167693, + "learning_rate": 1.2787190082644628e-06, + "logits/chosen": -0.185546875, + "logits/rejected": -0.279296875, + "logps/chosen": -380.0, + "logps/rejected": -340.0, + "loss": 0.3158, + "rewards/accuracies": 0.8515625, + "rewards/chosen": 1.078125, + "rewards/margins": 6.875, + "rewards/rejected": -5.8125, + "step": 382 + }, + { + "epoch": 0.4699386503067485, + "grad_norm": 35.065632595795535, + "learning_rate": 1.2780991735537192e-06, + "logits/chosen": -0.09912109375, + "logits/rejected": -0.2578125, + "logps/chosen": -402.0, + "logps/rejected": -378.0, + "loss": 0.4456, + "rewards/accuracies": 0.8359375, + "rewards/chosen": 1.1015625, + "rewards/margins": 6.78125, + "rewards/rejected": -5.6875, + "step": 383 + }, + { + "epoch": 0.47116564417177914, + "grad_norm": 33.79101177267741, + "learning_rate": 1.2774793388429753e-06, + "logits/chosen": -0.2001953125, + "logits/rejected": -0.435546875, + "logps/chosen": -398.0, + "logps/rejected": -342.0, + "loss": 0.2961, + "rewards/accuracies": 0.890625, + "rewards/chosen": 0.640625, + "rewards/margins": 7.6875, + "rewards/rejected": -7.03125, + "step": 384 + }, + { + "epoch": 0.4723926380368098, + "grad_norm": 36.518171637968166, + "learning_rate": 1.2768595041322314e-06, + "logits/chosen": -0.10302734375, + "logits/rejected": -0.283203125, + "logps/chosen": -394.0, + "logps/rejected": -330.0, + "loss": 0.3771, + "rewards/accuracies": 0.8046875, + "rewards/chosen": 0.283203125, + "rewards/margins": 6.34375, + "rewards/rejected": -6.0625, + "step": 385 + }, + { + "epoch": 0.4736196319018405, + "grad_norm": 34.95682432591222, + "learning_rate": 1.2762396694214876e-06, + "logits/chosen": -0.043701171875, + "logits/rejected": -0.2041015625, + "logps/chosen": -342.0, + "logps/rejected": -310.0, + "loss": 0.369, + "rewards/accuracies": 0.875, + "rewards/chosen": -0.3203125, + "rewards/margins": 6.09375, + "rewards/rejected": -6.40625, + "step": 386 + }, + { + "epoch": 0.47484662576687114, + "grad_norm": 64.7195376952873, + "learning_rate": 1.2756198347107437e-06, + "logits/chosen": -0.068359375, + "logits/rejected": -0.271484375, + "logps/chosen": -422.0, + "logps/rejected": -350.0, + "loss": 0.4035, + "rewards/accuracies": 0.8359375, + "rewards/chosen": 0.6171875, + "rewards/margins": 7.34375, + "rewards/rejected": -6.75, + "step": 387 + }, + { + "epoch": 0.47607361963190187, + "grad_norm": 42.357455920572896, + "learning_rate": 1.275e-06, + "logits/chosen": -0.08349609375, + "logits/rejected": -0.1826171875, + "logps/chosen": -406.0, + "logps/rejected": -326.0, + "loss": 0.5163, + "rewards/accuracies": 0.7890625, + "rewards/chosen": -0.04736328125, + "rewards/margins": 5.53125, + "rewards/rejected": -5.59375, + "step": 388 + }, + { + "epoch": 0.47730061349693254, + "grad_norm": 38.953026072404185, + "learning_rate": 1.2743801652892562e-06, + "logits/chosen": -0.1435546875, + "logits/rejected": -0.2236328125, + "logps/chosen": -412.0, + "logps/rejected": -340.0, + "loss": 0.4455, + "rewards/accuracies": 0.8203125, + "rewards/chosen": 0.099609375, + "rewards/margins": 6.3125, + "rewards/rejected": -6.1875, + "step": 389 + }, + { + "epoch": 0.4785276073619632, + "grad_norm": 38.5573431700193, + "learning_rate": 1.2737603305785124e-06, + "logits/chosen": -0.2216796875, + "logits/rejected": -0.34375, + "logps/chosen": -372.0, + "logps/rejected": -344.0, + "loss": 0.492, + "rewards/accuracies": 0.828125, + "rewards/chosen": -0.62890625, + "rewards/margins": 6.0, + "rewards/rejected": -6.65625, + "step": 390 + }, + { + "epoch": 0.47975460122699387, + "grad_norm": 37.845045364551986, + "learning_rate": 1.2731404958677687e-06, + "logits/chosen": -0.1201171875, + "logits/rejected": -0.23828125, + "logps/chosen": -390.0, + "logps/rejected": -344.0, + "loss": 0.4807, + "rewards/accuracies": 0.8203125, + "rewards/chosen": -0.181640625, + "rewards/margins": 6.4375, + "rewards/rejected": -6.625, + "step": 391 + }, + { + "epoch": 0.48098159509202454, + "grad_norm": 29.781954844950388, + "learning_rate": 1.2725206611570249e-06, + "logits/chosen": -0.060791015625, + "logits/rejected": -0.208984375, + "logps/chosen": -362.0, + "logps/rejected": -330.0, + "loss": 0.3103, + "rewards/accuracies": 0.875, + "rewards/chosen": -0.06982421875, + "rewards/margins": 6.59375, + "rewards/rejected": -6.65625, + "step": 392 + }, + { + "epoch": 0.4822085889570552, + "grad_norm": 33.5926238632541, + "learning_rate": 1.271900826446281e-06, + "logits/chosen": -0.0286865234375, + "logits/rejected": -0.166015625, + "logps/chosen": -416.0, + "logps/rejected": -374.0, + "loss": 0.3323, + "rewards/accuracies": 0.8359375, + "rewards/chosen": -0.263671875, + "rewards/margins": 6.4375, + "rewards/rejected": -6.6875, + "step": 393 + }, + { + "epoch": 0.4834355828220859, + "grad_norm": 32.087849602750985, + "learning_rate": 1.2712809917355373e-06, + "logits/chosen": -0.09521484375, + "logits/rejected": -0.298828125, + "logps/chosen": -398.0, + "logps/rejected": -342.0, + "loss": 0.262, + "rewards/accuracies": 0.84375, + "rewards/chosen": -0.166015625, + "rewards/margins": 6.1875, + "rewards/rejected": -6.375, + "step": 394 + }, + { + "epoch": 0.48466257668711654, + "grad_norm": 34.842476262528464, + "learning_rate": 1.2706611570247933e-06, + "logits/chosen": -0.1025390625, + "logits/rejected": -0.216796875, + "logps/chosen": -448.0, + "logps/rejected": -356.0, + "loss": 0.3764, + "rewards/accuracies": 0.8046875, + "rewards/chosen": -0.09375, + "rewards/margins": 6.125, + "rewards/rejected": -6.21875, + "step": 395 + }, + { + "epoch": 0.48588957055214727, + "grad_norm": 22.42844686295793, + "learning_rate": 1.2700413223140496e-06, + "logits/chosen": -0.1806640625, + "logits/rejected": -0.388671875, + "logps/chosen": -348.0, + "logps/rejected": -332.0, + "loss": 0.2277, + "rewards/accuracies": 0.90625, + "rewards/chosen": -0.54296875, + "rewards/margins": 6.3125, + "rewards/rejected": -6.875, + "step": 396 + }, + { + "epoch": 0.48711656441717793, + "grad_norm": 33.97555781703141, + "learning_rate": 1.2694214876033058e-06, + "logits/chosen": -0.1513671875, + "logits/rejected": -0.314453125, + "logps/chosen": -388.0, + "logps/rejected": -350.0, + "loss": 0.3848, + "rewards/accuracies": 0.828125, + "rewards/chosen": -0.625, + "rewards/margins": 6.78125, + "rewards/rejected": -7.40625, + "step": 397 + }, + { + "epoch": 0.4883435582822086, + "grad_norm": 39.08030474072834, + "learning_rate": 1.268801652892562e-06, + "logits/chosen": -0.1328125, + "logits/rejected": -0.27734375, + "logps/chosen": -428.0, + "logps/rejected": -346.0, + "loss": 0.4427, + "rewards/accuracies": 0.796875, + "rewards/chosen": -0.08740234375, + "rewards/margins": 6.375, + "rewards/rejected": -6.46875, + "step": 398 + }, + { + "epoch": 0.48957055214723927, + "grad_norm": 33.83840388728663, + "learning_rate": 1.2681818181818183e-06, + "logits/chosen": -0.1611328125, + "logits/rejected": -0.291015625, + "logps/chosen": -394.0, + "logps/rejected": -346.0, + "loss": 0.2963, + "rewards/accuracies": 0.859375, + "rewards/chosen": -0.0035858154296875, + "rewards/margins": 6.53125, + "rewards/rejected": -6.53125, + "step": 399 + }, + { + "epoch": 0.49079754601226994, + "grad_norm": 41.49480537946823, + "learning_rate": 1.2675619834710744e-06, + "logits/chosen": -0.09130859375, + "logits/rejected": -0.1650390625, + "logps/chosen": -384.0, + "logps/rejected": -368.0, + "loss": 0.5062, + "rewards/accuracies": 0.796875, + "rewards/chosen": -0.5390625, + "rewards/margins": 5.59375, + "rewards/rejected": -6.15625, + "step": 400 + }, + { + "epoch": 0.4920245398773006, + "grad_norm": 27.800803034272004, + "learning_rate": 1.2669421487603308e-06, + "logits/chosen": -0.23828125, + "logits/rejected": -0.373046875, + "logps/chosen": -456.0, + "logps/rejected": -390.0, + "loss": 0.2938, + "rewards/accuracies": 0.8671875, + "rewards/chosen": 0.40625, + "rewards/margins": 7.59375, + "rewards/rejected": -7.1875, + "step": 401 + }, + { + "epoch": 0.49325153374233127, + "grad_norm": 32.09882908595564, + "learning_rate": 1.266322314049587e-06, + "logits/chosen": -0.0791015625, + "logits/rejected": -0.1796875, + "logps/chosen": -394.0, + "logps/rejected": -350.0, + "loss": 0.349, + "rewards/accuracies": 0.8359375, + "rewards/chosen": -0.04443359375, + "rewards/margins": 6.5, + "rewards/rejected": -6.53125, + "step": 402 + }, + { + "epoch": 0.49447852760736194, + "grad_norm": 29.930227321008882, + "learning_rate": 1.265702479338843e-06, + "logits/chosen": -0.07958984375, + "logits/rejected": -0.15625, + "logps/chosen": -396.0, + "logps/rejected": -374.0, + "loss": 0.341, + "rewards/accuracies": 0.8671875, + "rewards/chosen": -0.033447265625, + "rewards/margins": 6.78125, + "rewards/rejected": -6.8125, + "step": 403 + }, + { + "epoch": 0.49570552147239266, + "grad_norm": 25.987833840695732, + "learning_rate": 1.2650826446280992e-06, + "logits/chosen": -0.021728515625, + "logits/rejected": -0.150390625, + "logps/chosen": -372.0, + "logps/rejected": -348.0, + "loss": 0.2606, + "rewards/accuracies": 0.8984375, + "rewards/chosen": 0.1669921875, + "rewards/margins": 7.0625, + "rewards/rejected": -6.90625, + "step": 404 + }, + { + "epoch": 0.49693251533742333, + "grad_norm": 38.14679307536266, + "learning_rate": 1.2644628099173553e-06, + "logits/chosen": -0.1064453125, + "logits/rejected": -0.328125, + "logps/chosen": -370.0, + "logps/rejected": -322.0, + "loss": 0.4214, + "rewards/accuracies": 0.8125, + "rewards/chosen": 0.6484375, + "rewards/margins": 6.65625, + "rewards/rejected": -6.03125, + "step": 405 + }, + { + "epoch": 0.498159509202454, + "grad_norm": 37.82975021877776, + "learning_rate": 1.2638429752066117e-06, + "logits/chosen": 0.0166015625, + "logits/rejected": -0.0625, + "logps/chosen": -378.0, + "logps/rejected": -340.0, + "loss": 0.5427, + "rewards/accuracies": 0.8046875, + "rewards/chosen": 0.3515625, + "rewards/margins": 6.09375, + "rewards/rejected": -5.75, + "step": 406 + }, + { + "epoch": 0.49938650306748467, + "grad_norm": 38.332851373691135, + "learning_rate": 1.2632231404958678e-06, + "logits/chosen": 0.0115966796875, + "logits/rejected": -0.134765625, + "logps/chosen": -408.0, + "logps/rejected": -344.0, + "loss": 0.4382, + "rewards/accuracies": 0.8046875, + "rewards/chosen": 1.34375, + "rewards/margins": 7.40625, + "rewards/rejected": -6.0625, + "step": 407 + }, + { + "epoch": 0.5006134969325153, + "grad_norm": 41.6208959235071, + "learning_rate": 1.262603305785124e-06, + "logits/chosen": -0.083984375, + "logits/rejected": -0.1474609375, + "logps/chosen": -426.0, + "logps/rejected": -360.0, + "loss": 0.5081, + "rewards/accuracies": 0.7734375, + "rewards/chosen": -0.091796875, + "rewards/margins": 5.1875, + "rewards/rejected": -5.28125, + "step": 408 + }, + { + "epoch": 0.501840490797546, + "grad_norm": 42.88666645295333, + "learning_rate": 1.2619834710743803e-06, + "logits/chosen": -0.00860595703125, + "logits/rejected": -0.205078125, + "logps/chosen": -426.0, + "logps/rejected": -342.0, + "loss": 0.4119, + "rewards/accuracies": 0.8203125, + "rewards/chosen": 0.56640625, + "rewards/margins": 6.71875, + "rewards/rejected": -6.15625, + "step": 409 + }, + { + "epoch": 0.5030674846625767, + "grad_norm": 36.428858153344706, + "learning_rate": 1.2613636363636365e-06, + "logits/chosen": -0.00095367431640625, + "logits/rejected": -0.1943359375, + "logps/chosen": -388.0, + "logps/rejected": -360.0, + "loss": 0.4038, + "rewards/accuracies": 0.796875, + "rewards/chosen": 0.3828125, + "rewards/margins": 6.125, + "rewards/rejected": -5.75, + "step": 410 + }, + { + "epoch": 0.5042944785276073, + "grad_norm": 33.21094894864319, + "learning_rate": 1.2607438016528926e-06, + "logits/chosen": 0.0140380859375, + "logits/rejected": -0.12890625, + "logps/chosen": -342.0, + "logps/rejected": -320.0, + "loss": 0.3058, + "rewards/accuracies": 0.8359375, + "rewards/chosen": 0.1298828125, + "rewards/margins": 6.46875, + "rewards/rejected": -6.34375, + "step": 411 + }, + { + "epoch": 0.505521472392638, + "grad_norm": 42.15288062503197, + "learning_rate": 1.260123966942149e-06, + "logits/chosen": 0.007080078125, + "logits/rejected": -0.1259765625, + "logps/chosen": -428.0, + "logps/rejected": -368.0, + "loss": 0.4444, + "rewards/accuracies": 0.796875, + "rewards/chosen": 0.30078125, + "rewards/margins": 5.96875, + "rewards/rejected": -5.6875, + "step": 412 + }, + { + "epoch": 0.5067484662576687, + "grad_norm": 37.795734491735246, + "learning_rate": 1.2595041322314049e-06, + "logits/chosen": -0.042236328125, + "logits/rejected": -0.26171875, + "logps/chosen": -408.0, + "logps/rejected": -332.0, + "loss": 0.3878, + "rewards/accuracies": 0.8125, + "rewards/chosen": 0.9140625, + "rewards/margins": 6.96875, + "rewards/rejected": -6.0625, + "step": 413 + }, + { + "epoch": 0.5079754601226993, + "grad_norm": 37.27154344231617, + "learning_rate": 1.2588842975206612e-06, + "logits/chosen": -0.0859375, + "logits/rejected": -0.2265625, + "logps/chosen": -406.0, + "logps/rejected": -354.0, + "loss": 0.3559, + "rewards/accuracies": 0.84375, + "rewards/chosen": 0.85546875, + "rewards/margins": 6.71875, + "rewards/rejected": -5.84375, + "step": 414 + }, + { + "epoch": 0.50920245398773, + "grad_norm": 30.678885373774655, + "learning_rate": 1.2582644628099174e-06, + "logits/chosen": -0.041748046875, + "logits/rejected": -0.212890625, + "logps/chosen": -350.0, + "logps/rejected": -332.0, + "loss": 0.3048, + "rewards/accuracies": 0.828125, + "rewards/chosen": 0.52734375, + "rewards/margins": 6.75, + "rewards/rejected": -6.21875, + "step": 415 + }, + { + "epoch": 0.5104294478527608, + "grad_norm": 26.882010464050307, + "learning_rate": 1.2576446280991735e-06, + "logits/chosen": -0.004730224609375, + "logits/rejected": -0.2197265625, + "logps/chosen": -360.0, + "logps/rejected": -324.0, + "loss": 0.1899, + "rewards/accuracies": 0.890625, + "rewards/chosen": 0.458984375, + "rewards/margins": 6.75, + "rewards/rejected": -6.28125, + "step": 416 + }, + { + "epoch": 0.5116564417177915, + "grad_norm": 30.354406082685717, + "learning_rate": 1.2570247933884299e-06, + "logits/chosen": -0.2197265625, + "logits/rejected": -0.384765625, + "logps/chosen": -438.0, + "logps/rejected": -382.0, + "loss": 0.3453, + "rewards/accuracies": 0.84375, + "rewards/chosen": 1.4765625, + "rewards/margins": 8.125, + "rewards/rejected": -6.65625, + "step": 417 + }, + { + "epoch": 0.5128834355828221, + "grad_norm": 33.53763392215037, + "learning_rate": 1.256404958677686e-06, + "logits/chosen": -0.126953125, + "logits/rejected": -0.2099609375, + "logps/chosen": -368.0, + "logps/rejected": -306.0, + "loss": 0.3744, + "rewards/accuracies": 0.859375, + "rewards/chosen": 0.7890625, + "rewards/margins": 7.1875, + "rewards/rejected": -6.375, + "step": 418 + }, + { + "epoch": 0.5141104294478528, + "grad_norm": 28.10440726715766, + "learning_rate": 1.2557851239669421e-06, + "logits/chosen": -0.0546875, + "logits/rejected": -0.21875, + "logps/chosen": -380.0, + "logps/rejected": -342.0, + "loss": 0.3206, + "rewards/accuracies": 0.8671875, + "rewards/chosen": 0.99609375, + "rewards/margins": 7.78125, + "rewards/rejected": -6.78125, + "step": 419 + }, + { + "epoch": 0.5153374233128835, + "grad_norm": 30.627621824355007, + "learning_rate": 1.2551652892561985e-06, + "logits/chosen": 0.0262451171875, + "logits/rejected": -0.134765625, + "logps/chosen": -346.0, + "logps/rejected": -312.0, + "loss": 0.3514, + "rewards/accuracies": 0.828125, + "rewards/chosen": 0.796875, + "rewards/margins": 6.4375, + "rewards/rejected": -5.65625, + "step": 420 + }, + { + "epoch": 0.5165644171779141, + "grad_norm": 31.844076287237254, + "learning_rate": 1.2545454545454544e-06, + "logits/chosen": -0.01544189453125, + "logits/rejected": -0.2158203125, + "logps/chosen": -364.0, + "logps/rejected": -320.0, + "loss": 0.4123, + "rewards/accuracies": 0.8125, + "rewards/chosen": 1.1875, + "rewards/margins": 6.8125, + "rewards/rejected": -5.625, + "step": 421 + }, + { + "epoch": 0.5177914110429448, + "grad_norm": 33.522822368907185, + "learning_rate": 1.2539256198347108e-06, + "logits/chosen": 0.0595703125, + "logits/rejected": -0.0986328125, + "logps/chosen": -360.0, + "logps/rejected": -302.0, + "loss": 0.3724, + "rewards/accuracies": 0.84375, + "rewards/chosen": 0.52734375, + "rewards/margins": 6.6875, + "rewards/rejected": -6.15625, + "step": 422 + }, + { + "epoch": 0.5190184049079755, + "grad_norm": 37.40841671544669, + "learning_rate": 1.253305785123967e-06, + "logits/chosen": -0.0595703125, + "logits/rejected": -0.20703125, + "logps/chosen": -398.0, + "logps/rejected": -362.0, + "loss": 0.4749, + "rewards/accuracies": 0.8125, + "rewards/chosen": 0.578125, + "rewards/margins": 6.0, + "rewards/rejected": -5.4375, + "step": 423 + }, + { + "epoch": 0.5202453987730061, + "grad_norm": 34.546118881756456, + "learning_rate": 1.252685950413223e-06, + "logits/chosen": 0.0235595703125, + "logits/rejected": -0.166015625, + "logps/chosen": -378.0, + "logps/rejected": -342.0, + "loss": 0.4503, + "rewards/accuracies": 0.8359375, + "rewards/chosen": 1.2109375, + "rewards/margins": 7.09375, + "rewards/rejected": -5.875, + "step": 424 + }, + { + "epoch": 0.5214723926380368, + "grad_norm": 37.84213132068162, + "learning_rate": 1.2520661157024794e-06, + "logits/chosen": -0.0123291015625, + "logits/rejected": -0.11572265625, + "logps/chosen": -396.0, + "logps/rejected": -348.0, + "loss": 0.5144, + "rewards/accuracies": 0.796875, + "rewards/chosen": 1.3203125, + "rewards/margins": 6.84375, + "rewards/rejected": -5.53125, + "step": 425 + }, + { + "epoch": 0.5226993865030675, + "grad_norm": 36.174280199739684, + "learning_rate": 1.2514462809917356e-06, + "logits/chosen": -0.06396484375, + "logits/rejected": -0.166015625, + "logps/chosen": -370.0, + "logps/rejected": -348.0, + "loss": 0.4214, + "rewards/accuracies": 0.875, + "rewards/chosen": 1.015625, + "rewards/margins": 7.0625, + "rewards/rejected": -6.0625, + "step": 426 + }, + { + "epoch": 0.5239263803680981, + "grad_norm": 45.37106746020955, + "learning_rate": 1.2508264462809917e-06, + "logits/chosen": -0.01483154296875, + "logits/rejected": -0.1279296875, + "logps/chosen": -378.0, + "logps/rejected": -334.0, + "loss": 0.4318, + "rewards/accuracies": 0.8203125, + "rewards/chosen": 1.1171875, + "rewards/margins": 6.75, + "rewards/rejected": -5.625, + "step": 427 + }, + { + "epoch": 0.5251533742331288, + "grad_norm": 33.589023647082186, + "learning_rate": 1.250206611570248e-06, + "logits/chosen": -0.0419921875, + "logits/rejected": -0.1533203125, + "logps/chosen": -364.0, + "logps/rejected": -324.0, + "loss": 0.4488, + "rewards/accuracies": 0.8046875, + "rewards/chosen": 0.462890625, + "rewards/margins": 6.46875, + "rewards/rejected": -6.0, + "step": 428 + }, + { + "epoch": 0.5263803680981595, + "grad_norm": 30.11424815064351, + "learning_rate": 1.2495867768595042e-06, + "logits/chosen": -0.044677734375, + "logits/rejected": -0.1494140625, + "logps/chosen": -360.0, + "logps/rejected": -376.0, + "loss": 0.2639, + "rewards/accuracies": 0.8671875, + "rewards/chosen": 0.373046875, + "rewards/margins": 7.96875, + "rewards/rejected": -7.59375, + "step": 429 + }, + { + "epoch": 0.5276073619631901, + "grad_norm": 23.36575464672381, + "learning_rate": 1.2489669421487603e-06, + "logits/chosen": -0.11767578125, + "logits/rejected": -0.28125, + "logps/chosen": -394.0, + "logps/rejected": -350.0, + "loss": 0.2208, + "rewards/accuracies": 0.875, + "rewards/chosen": 0.921875, + "rewards/margins": 7.5625, + "rewards/rejected": -6.625, + "step": 430 + }, + { + "epoch": 0.5288343558282208, + "grad_norm": 39.63938191210347, + "learning_rate": 1.2483471074380165e-06, + "logits/chosen": 0.01806640625, + "logits/rejected": -0.150390625, + "logps/chosen": -386.0, + "logps/rejected": -346.0, + "loss": 0.4684, + "rewards/accuracies": 0.8203125, + "rewards/chosen": 0.12255859375, + "rewards/margins": 6.4375, + "rewards/rejected": -6.3125, + "step": 431 + }, + { + "epoch": 0.5300613496932516, + "grad_norm": 35.90106246867024, + "learning_rate": 1.2477272727272726e-06, + "logits/chosen": 0.06689453125, + "logits/rejected": 0.035888671875, + "logps/chosen": -382.0, + "logps/rejected": -354.0, + "loss": 0.4356, + "rewards/accuracies": 0.84375, + "rewards/chosen": 0.78515625, + "rewards/margins": 6.71875, + "rewards/rejected": -5.9375, + "step": 432 + }, + { + "epoch": 0.5312883435582823, + "grad_norm": 43.12860297142818, + "learning_rate": 1.247107438016529e-06, + "logits/chosen": 0.08740234375, + "logits/rejected": -0.1328125, + "logps/chosen": -388.0, + "logps/rejected": -320.0, + "loss": 0.4803, + "rewards/accuracies": 0.8046875, + "rewards/chosen": -0.12353515625, + "rewards/margins": 6.4375, + "rewards/rejected": -6.5625, + "step": 433 + }, + { + "epoch": 0.5325153374233129, + "grad_norm": 37.083326074229944, + "learning_rate": 1.2464876033057851e-06, + "logits/chosen": -0.0869140625, + "logits/rejected": -0.1611328125, + "logps/chosen": -390.0, + "logps/rejected": -356.0, + "loss": 0.3206, + "rewards/accuracies": 0.859375, + "rewards/chosen": 0.07666015625, + "rewards/margins": 6.375, + "rewards/rejected": -6.28125, + "step": 434 + }, + { + "epoch": 0.5337423312883436, + "grad_norm": 37.88064967580844, + "learning_rate": 1.2458677685950415e-06, + "logits/chosen": -0.01458740234375, + "logits/rejected": -0.208984375, + "logps/chosen": -394.0, + "logps/rejected": -362.0, + "loss": 0.3939, + "rewards/accuracies": 0.8359375, + "rewards/chosen": 0.51953125, + "rewards/margins": 7.15625, + "rewards/rejected": -6.65625, + "step": 435 + }, + { + "epoch": 0.5349693251533743, + "grad_norm": 30.102357096787667, + "learning_rate": 1.2452479338842976e-06, + "logits/chosen": -0.0250244140625, + "logits/rejected": -0.2255859375, + "logps/chosen": -422.0, + "logps/rejected": -358.0, + "loss": 0.2562, + "rewards/accuracies": 0.8828125, + "rewards/chosen": 1.1171875, + "rewards/margins": 7.75, + "rewards/rejected": -6.625, + "step": 436 + }, + { + "epoch": 0.5361963190184049, + "grad_norm": 25.03840429478704, + "learning_rate": 1.2446280991735538e-06, + "logits/chosen": -0.080078125, + "logits/rejected": -0.25, + "logps/chosen": -388.0, + "logps/rejected": -350.0, + "loss": 0.241, + "rewards/accuracies": 0.8671875, + "rewards/chosen": 0.41796875, + "rewards/margins": 7.28125, + "rewards/rejected": -6.875, + "step": 437 + }, + { + "epoch": 0.5374233128834356, + "grad_norm": 36.15225304248156, + "learning_rate": 1.24400826446281e-06, + "logits/chosen": 0.02001953125, + "logits/rejected": -0.06982421875, + "logps/chosen": -366.0, + "logps/rejected": -334.0, + "loss": 0.5276, + "rewards/accuracies": 0.8125, + "rewards/chosen": -0.0250244140625, + "rewards/margins": 5.65625, + "rewards/rejected": -5.6875, + "step": 438 + }, + { + "epoch": 0.5386503067484663, + "grad_norm": 34.58997013752512, + "learning_rate": 1.243388429752066e-06, + "logits/chosen": 0.0625, + "logits/rejected": -0.126953125, + "logps/chosen": -372.0, + "logps/rejected": -348.0, + "loss": 0.3671, + "rewards/accuracies": 0.84375, + "rewards/chosen": -0.0703125, + "rewards/margins": 6.65625, + "rewards/rejected": -6.71875, + "step": 439 + }, + { + "epoch": 0.5398773006134969, + "grad_norm": 43.86154514789445, + "learning_rate": 1.2427685950413224e-06, + "logits/chosen": 0.0218505859375, + "logits/rejected": -0.0751953125, + "logps/chosen": -392.0, + "logps/rejected": -344.0, + "loss": 0.4825, + "rewards/accuracies": 0.8046875, + "rewards/chosen": -0.515625, + "rewards/margins": 5.4375, + "rewards/rejected": -5.96875, + "step": 440 + }, + { + "epoch": 0.5411042944785276, + "grad_norm": 32.02976017157544, + "learning_rate": 1.2421487603305785e-06, + "logits/chosen": 0.09619140625, + "logits/rejected": -0.0108642578125, + "logps/chosen": -340.0, + "logps/rejected": -314.0, + "loss": 0.4408, + "rewards/accuracies": 0.828125, + "rewards/chosen": -0.0869140625, + "rewards/margins": 6.15625, + "rewards/rejected": -6.21875, + "step": 441 + }, + { + "epoch": 0.5423312883435583, + "grad_norm": 31.122024927447573, + "learning_rate": 1.2415289256198347e-06, + "logits/chosen": -0.064453125, + "logits/rejected": -0.1982421875, + "logps/chosen": -390.0, + "logps/rejected": -352.0, + "loss": 0.3842, + "rewards/accuracies": 0.859375, + "rewards/chosen": 0.51171875, + "rewards/margins": 6.5625, + "rewards/rejected": -6.0625, + "step": 442 + }, + { + "epoch": 0.5435582822085889, + "grad_norm": 35.35206800809496, + "learning_rate": 1.240909090909091e-06, + "logits/chosen": -0.024169921875, + "logits/rejected": -0.1435546875, + "logps/chosen": -372.0, + "logps/rejected": -332.0, + "loss": 0.3707, + "rewards/accuracies": 0.828125, + "rewards/chosen": 0.69140625, + "rewards/margins": 6.8125, + "rewards/rejected": -6.125, + "step": 443 + }, + { + "epoch": 0.5447852760736196, + "grad_norm": 35.692369145811284, + "learning_rate": 1.2402892561983472e-06, + "logits/chosen": -0.06201171875, + "logits/rejected": -0.298828125, + "logps/chosen": -454.0, + "logps/rejected": -382.0, + "loss": 0.3655, + "rewards/accuracies": 0.859375, + "rewards/chosen": 1.2890625, + "rewards/margins": 8.0, + "rewards/rejected": -6.71875, + "step": 444 + }, + { + "epoch": 0.5460122699386503, + "grad_norm": 32.62912831625137, + "learning_rate": 1.2396694214876033e-06, + "logits/chosen": -0.016357421875, + "logits/rejected": -0.208984375, + "logps/chosen": -358.0, + "logps/rejected": -314.0, + "loss": 0.4199, + "rewards/accuracies": 0.796875, + "rewards/chosen": 0.66796875, + "rewards/margins": 6.21875, + "rewards/rejected": -5.5625, + "step": 445 + }, + { + "epoch": 0.5472392638036809, + "grad_norm": 25.78605234898617, + "learning_rate": 1.2390495867768597e-06, + "logits/chosen": -0.040283203125, + "logits/rejected": -0.203125, + "logps/chosen": -438.0, + "logps/rejected": -370.0, + "loss": 0.2432, + "rewards/accuracies": 0.890625, + "rewards/chosen": 0.81640625, + "rewards/margins": 6.46875, + "rewards/rejected": -5.65625, + "step": 446 + }, + { + "epoch": 0.5484662576687117, + "grad_norm": 28.526549681854004, + "learning_rate": 1.2384297520661158e-06, + "logits/chosen": -0.0341796875, + "logits/rejected": -0.23828125, + "logps/chosen": -414.0, + "logps/rejected": -360.0, + "loss": 0.3087, + "rewards/accuracies": 0.8671875, + "rewards/chosen": 1.359375, + "rewards/margins": 8.1875, + "rewards/rejected": -6.84375, + "step": 447 + }, + { + "epoch": 0.5496932515337424, + "grad_norm": 33.26832004918433, + "learning_rate": 1.237809917355372e-06, + "logits/chosen": 0.042724609375, + "logits/rejected": -0.10009765625, + "logps/chosen": -346.0, + "logps/rejected": -314.0, + "loss": 0.3362, + "rewards/accuracies": 0.84375, + "rewards/chosen": 0.1796875, + "rewards/margins": 6.03125, + "rewards/rejected": -5.84375, + "step": 448 + }, + { + "epoch": 0.550920245398773, + "grad_norm": 29.824788916298026, + "learning_rate": 1.237190082644628e-06, + "logits/chosen": -0.0341796875, + "logits/rejected": -0.1650390625, + "logps/chosen": -400.0, + "logps/rejected": -346.0, + "loss": 0.3203, + "rewards/accuracies": 0.8515625, + "rewards/chosen": 1.234375, + "rewards/margins": 7.34375, + "rewards/rejected": -6.09375, + "step": 449 + }, + { + "epoch": 0.5521472392638037, + "grad_norm": 29.646416826102993, + "learning_rate": 1.2365702479338842e-06, + "logits/chosen": -0.1103515625, + "logits/rejected": -0.1494140625, + "logps/chosen": -350.0, + "logps/rejected": -302.0, + "loss": 0.2642, + "rewards/accuracies": 0.8359375, + "rewards/chosen": 0.66015625, + "rewards/margins": 6.78125, + "rewards/rejected": -6.125, + "step": 450 + }, + { + "epoch": 0.5533742331288344, + "grad_norm": 31.583737211583365, + "learning_rate": 1.2359504132231406e-06, + "logits/chosen": 0.03173828125, + "logits/rejected": -0.1396484375, + "logps/chosen": -348.0, + "logps/rejected": -294.0, + "loss": 0.3883, + "rewards/accuracies": 0.84375, + "rewards/chosen": 0.54296875, + "rewards/margins": 6.40625, + "rewards/rejected": -5.875, + "step": 451 + }, + { + "epoch": 0.554601226993865, + "grad_norm": 27.43380959875228, + "learning_rate": 1.2353305785123967e-06, + "logits/chosen": -0.0439453125, + "logits/rejected": -0.28125, + "logps/chosen": -460.0, + "logps/rejected": -342.0, + "loss": 0.2556, + "rewards/accuracies": 0.859375, + "rewards/chosen": 1.9453125, + "rewards/margins": 7.53125, + "rewards/rejected": -5.59375, + "step": 452 + }, + { + "epoch": 0.5558282208588957, + "grad_norm": 24.613642812642805, + "learning_rate": 1.2347107438016529e-06, + "logits/chosen": -0.130859375, + "logits/rejected": -0.33984375, + "logps/chosen": -440.0, + "logps/rejected": -378.0, + "loss": 0.2702, + "rewards/accuracies": 0.90625, + "rewards/chosen": 2.125, + "rewards/margins": 8.1875, + "rewards/rejected": -6.0625, + "step": 453 + }, + { + "epoch": 0.5570552147239264, + "grad_norm": 31.985648649098213, + "learning_rate": 1.2340909090909092e-06, + "logits/chosen": 0.046875, + "logits/rejected": -0.0206298828125, + "logps/chosen": -346.0, + "logps/rejected": -320.0, + "loss": 0.3305, + "rewards/accuracies": 0.8359375, + "rewards/chosen": 0.0859375, + "rewards/margins": 6.0625, + "rewards/rejected": -6.0, + "step": 454 + }, + { + "epoch": 0.558282208588957, + "grad_norm": 26.277923249806037, + "learning_rate": 1.2334710743801654e-06, + "logits/chosen": -0.0771484375, + "logits/rejected": -0.201171875, + "logps/chosen": -372.0, + "logps/rejected": -362.0, + "loss": 0.2299, + "rewards/accuracies": 0.8828125, + "rewards/chosen": 1.5, + "rewards/margins": 7.59375, + "rewards/rejected": -6.09375, + "step": 455 + }, + { + "epoch": 0.5595092024539877, + "grad_norm": 35.777155080443045, + "learning_rate": 1.2328512396694215e-06, + "logits/chosen": 0.0419921875, + "logits/rejected": -0.0576171875, + "logps/chosen": -318.0, + "logps/rejected": -314.0, + "loss": 0.4054, + "rewards/accuracies": 0.8125, + "rewards/chosen": 0.412109375, + "rewards/margins": 6.8125, + "rewards/rejected": -6.375, + "step": 456 + }, + { + "epoch": 0.5607361963190184, + "grad_norm": 33.41411882209408, + "learning_rate": 1.2322314049586776e-06, + "logits/chosen": 0.0126953125, + "logits/rejected": -0.1142578125, + "logps/chosen": -360.0, + "logps/rejected": -330.0, + "loss": 0.3362, + "rewards/accuracies": 0.8359375, + "rewards/chosen": 0.63671875, + "rewards/margins": 6.21875, + "rewards/rejected": -5.59375, + "step": 457 + }, + { + "epoch": 0.561963190184049, + "grad_norm": 27.510163645115238, + "learning_rate": 1.2316115702479338e-06, + "logits/chosen": -0.0264892578125, + "logits/rejected": -0.1845703125, + "logps/chosen": -376.0, + "logps/rejected": -302.0, + "loss": 0.2733, + "rewards/accuracies": 0.8828125, + "rewards/chosen": 1.109375, + "rewards/margins": 7.40625, + "rewards/rejected": -6.3125, + "step": 458 + }, + { + "epoch": 0.5631901840490797, + "grad_norm": 32.20829218573613, + "learning_rate": 1.2309917355371901e-06, + "logits/chosen": -0.0172119140625, + "logits/rejected": -0.1796875, + "logps/chosen": -396.0, + "logps/rejected": -356.0, + "loss": 0.3698, + "rewards/accuracies": 0.8515625, + "rewards/chosen": 1.171875, + "rewards/margins": 7.375, + "rewards/rejected": -6.1875, + "step": 459 + }, + { + "epoch": 0.5644171779141104, + "grad_norm": 28.463767397560563, + "learning_rate": 1.2303719008264463e-06, + "logits/chosen": -0.0159912109375, + "logits/rejected": -0.11572265625, + "logps/chosen": -354.0, + "logps/rejected": -342.0, + "loss": 0.2835, + "rewards/accuracies": 0.859375, + "rewards/chosen": 0.146484375, + "rewards/margins": 7.46875, + "rewards/rejected": -7.3125, + "step": 460 + }, + { + "epoch": 0.5656441717791411, + "grad_norm": 30.885477551125014, + "learning_rate": 1.2297520661157024e-06, + "logits/chosen": -0.2158203125, + "logits/rejected": -0.330078125, + "logps/chosen": -380.0, + "logps/rejected": -352.0, + "loss": 0.2845, + "rewards/accuracies": 0.8671875, + "rewards/chosen": 0.1884765625, + "rewards/margins": 6.5625, + "rewards/rejected": -6.375, + "step": 461 + }, + { + "epoch": 0.5668711656441717, + "grad_norm": 44.21195903099332, + "learning_rate": 1.2291322314049588e-06, + "logits/chosen": 0.047607421875, + "logits/rejected": -0.1552734375, + "logps/chosen": -370.0, + "logps/rejected": -310.0, + "loss": 0.4151, + "rewards/accuracies": 0.8203125, + "rewards/chosen": -0.1337890625, + "rewards/margins": 5.875, + "rewards/rejected": -6.03125, + "step": 462 + }, + { + "epoch": 0.5680981595092025, + "grad_norm": 33.208183287178784, + "learning_rate": 1.228512396694215e-06, + "logits/chosen": -0.0849609375, + "logits/rejected": -0.18359375, + "logps/chosen": -370.0, + "logps/rejected": -372.0, + "loss": 0.3252, + "rewards/accuracies": 0.8203125, + "rewards/chosen": 0.4140625, + "rewards/margins": 7.09375, + "rewards/rejected": -6.6875, + "step": 463 + }, + { + "epoch": 0.5693251533742332, + "grad_norm": 37.13310236043479, + "learning_rate": 1.2278925619834713e-06, + "logits/chosen": 0.05224609375, + "logits/rejected": -0.119140625, + "logps/chosen": -398.0, + "logps/rejected": -334.0, + "loss": 0.3492, + "rewards/accuracies": 0.859375, + "rewards/chosen": 0.3359375, + "rewards/margins": 6.90625, + "rewards/rejected": -6.59375, + "step": 464 + }, + { + "epoch": 0.5705521472392638, + "grad_norm": 21.80723319666503, + "learning_rate": 1.2272727272727274e-06, + "logits/chosen": -0.06396484375, + "logits/rejected": -0.302734375, + "logps/chosen": -404.0, + "logps/rejected": -362.0, + "loss": 0.1834, + "rewards/accuracies": 0.9140625, + "rewards/chosen": 0.6171875, + "rewards/margins": 8.25, + "rewards/rejected": -7.625, + "step": 465 + }, + { + "epoch": 0.5717791411042945, + "grad_norm": 32.116293843093004, + "learning_rate": 1.2266528925619833e-06, + "logits/chosen": -0.016845703125, + "logits/rejected": -0.1416015625, + "logps/chosen": -424.0, + "logps/rejected": -356.0, + "loss": 0.3118, + "rewards/accuracies": 0.8203125, + "rewards/chosen": 0.0458984375, + "rewards/margins": 7.5, + "rewards/rejected": -7.46875, + "step": 466 + }, + { + "epoch": 0.5730061349693252, + "grad_norm": 37.63110345633159, + "learning_rate": 1.2260330578512397e-06, + "logits/chosen": -0.0279541015625, + "logits/rejected": -0.2373046875, + "logps/chosen": -388.0, + "logps/rejected": -364.0, + "loss": 0.3871, + "rewards/accuracies": 0.828125, + "rewards/chosen": 0.8046875, + "rewards/margins": 7.21875, + "rewards/rejected": -6.40625, + "step": 467 + }, + { + "epoch": 0.5742331288343558, + "grad_norm": 33.08460200253717, + "learning_rate": 1.2254132231404958e-06, + "logits/chosen": -0.0419921875, + "logits/rejected": -0.1806640625, + "logps/chosen": -372.0, + "logps/rejected": -358.0, + "loss": 0.3099, + "rewards/accuracies": 0.8671875, + "rewards/chosen": 0.5234375, + "rewards/margins": 7.625, + "rewards/rejected": -7.09375, + "step": 468 + }, + { + "epoch": 0.5754601226993865, + "grad_norm": 22.648898471514507, + "learning_rate": 1.2247933884297522e-06, + "logits/chosen": 0.0230712890625, + "logits/rejected": -0.15234375, + "logps/chosen": -390.0, + "logps/rejected": -320.0, + "loss": 0.2207, + "rewards/accuracies": 0.90625, + "rewards/chosen": 0.796875, + "rewards/margins": 7.84375, + "rewards/rejected": -7.0625, + "step": 469 + }, + { + "epoch": 0.5766871165644172, + "grad_norm": 33.108021332496286, + "learning_rate": 1.2241735537190083e-06, + "logits/chosen": -0.010498046875, + "logits/rejected": -0.07373046875, + "logps/chosen": -356.0, + "logps/rejected": -338.0, + "loss": 0.367, + "rewards/accuracies": 0.828125, + "rewards/chosen": 0.69921875, + "rewards/margins": 7.5, + "rewards/rejected": -6.8125, + "step": 470 + }, + { + "epoch": 0.5779141104294478, + "grad_norm": 30.281427499200404, + "learning_rate": 1.2235537190082645e-06, + "logits/chosen": 0.1171875, + "logits/rejected": -0.0458984375, + "logps/chosen": -408.0, + "logps/rejected": -350.0, + "loss": 0.3469, + "rewards/accuracies": 0.8125, + "rewards/chosen": 0.328125, + "rewards/margins": 6.90625, + "rewards/rejected": -6.5625, + "step": 471 + }, + { + "epoch": 0.5791411042944785, + "grad_norm": 30.766178072522326, + "learning_rate": 1.2229338842975208e-06, + "logits/chosen": 0.0654296875, + "logits/rejected": -0.15234375, + "logps/chosen": -414.0, + "logps/rejected": -336.0, + "loss": 0.341, + "rewards/accuracies": 0.84375, + "rewards/chosen": 0.400390625, + "rewards/margins": 7.125, + "rewards/rejected": -6.71875, + "step": 472 + }, + { + "epoch": 0.5803680981595092, + "grad_norm": 35.82872232913366, + "learning_rate": 1.222314049586777e-06, + "logits/chosen": -0.10986328125, + "logits/rejected": -0.2138671875, + "logps/chosen": -406.0, + "logps/rejected": -394.0, + "loss": 0.3929, + "rewards/accuracies": 0.796875, + "rewards/chosen": 0.267578125, + "rewards/margins": 6.78125, + "rewards/rejected": -6.5, + "step": 473 + }, + { + "epoch": 0.5815950920245399, + "grad_norm": 38.235476934077774, + "learning_rate": 1.221694214876033e-06, + "logits/chosen": -0.06591796875, + "logits/rejected": -0.271484375, + "logps/chosen": -440.0, + "logps/rejected": -362.0, + "loss": 0.3842, + "rewards/accuracies": 0.8359375, + "rewards/chosen": 0.0673828125, + "rewards/margins": 6.875, + "rewards/rejected": -6.8125, + "step": 474 + }, + { + "epoch": 0.5828220858895705, + "grad_norm": 28.310340512331347, + "learning_rate": 1.2210743801652892e-06, + "logits/chosen": -0.06689453125, + "logits/rejected": -0.2734375, + "logps/chosen": -434.0, + "logps/rejected": -362.0, + "loss": 0.2596, + "rewards/accuracies": 0.8828125, + "rewards/chosen": 1.140625, + "rewards/margins": 8.4375, + "rewards/rejected": -7.3125, + "step": 475 + }, + { + "epoch": 0.5840490797546012, + "grad_norm": 31.21556050331037, + "learning_rate": 1.2204545454545454e-06, + "logits/chosen": 0.00927734375, + "logits/rejected": -0.1806640625, + "logps/chosen": -420.0, + "logps/rejected": -360.0, + "loss": 0.3346, + "rewards/accuracies": 0.8359375, + "rewards/chosen": -1.078125, + "rewards/margins": 6.53125, + "rewards/rejected": -7.625, + "step": 476 + }, + { + "epoch": 0.5852760736196319, + "grad_norm": 31.773096981894017, + "learning_rate": 1.2198347107438017e-06, + "logits/chosen": -0.11474609375, + "logits/rejected": -0.255859375, + "logps/chosen": -432.0, + "logps/rejected": -372.0, + "loss": 0.3232, + "rewards/accuracies": 0.890625, + "rewards/chosen": 0.1484375, + "rewards/margins": 7.8125, + "rewards/rejected": -7.65625, + "step": 477 + }, + { + "epoch": 0.5865030674846625, + "grad_norm": 30.507836762025494, + "learning_rate": 1.2192148760330579e-06, + "logits/chosen": -0.04150390625, + "logits/rejected": -0.2080078125, + "logps/chosen": -430.0, + "logps/rejected": -368.0, + "loss": 0.3843, + "rewards/accuracies": 0.8671875, + "rewards/chosen": -0.024658203125, + "rewards/margins": 7.625, + "rewards/rejected": -7.65625, + "step": 478 + }, + { + "epoch": 0.5877300613496933, + "grad_norm": 32.261887607989564, + "learning_rate": 1.218595041322314e-06, + "logits/chosen": -0.03759765625, + "logits/rejected": -0.298828125, + "logps/chosen": -370.0, + "logps/rejected": -352.0, + "loss": 0.322, + "rewards/accuracies": 0.84375, + "rewards/chosen": -1.1796875, + "rewards/margins": 6.625, + "rewards/rejected": -7.78125, + "step": 479 + }, + { + "epoch": 0.588957055214724, + "grad_norm": 28.188949742159306, + "learning_rate": 1.2179752066115704e-06, + "logits/chosen": -0.0791015625, + "logits/rejected": -0.208984375, + "logps/chosen": -372.0, + "logps/rejected": -348.0, + "loss": 0.2558, + "rewards/accuracies": 0.8828125, + "rewards/chosen": -0.8125, + "rewards/margins": 7.21875, + "rewards/rejected": -8.0, + "step": 480 + }, + { + "epoch": 0.5901840490797546, + "grad_norm": 37.17733184183342, + "learning_rate": 1.2173553719008265e-06, + "logits/chosen": -0.07568359375, + "logits/rejected": -0.1826171875, + "logps/chosen": -438.0, + "logps/rejected": -398.0, + "loss": 0.4973, + "rewards/accuracies": 0.796875, + "rewards/chosen": -0.578125, + "rewards/margins": 6.71875, + "rewards/rejected": -7.3125, + "step": 481 + }, + { + "epoch": 0.5914110429447853, + "grad_norm": 34.97683920834337, + "learning_rate": 1.2167355371900826e-06, + "logits/chosen": -0.07373046875, + "logits/rejected": -0.244140625, + "logps/chosen": -424.0, + "logps/rejected": -352.0, + "loss": 0.3731, + "rewards/accuracies": 0.8125, + "rewards/chosen": -0.255859375, + "rewards/margins": 6.875, + "rewards/rejected": -7.125, + "step": 482 + }, + { + "epoch": 0.592638036809816, + "grad_norm": 32.698615963886375, + "learning_rate": 1.216115702479339e-06, + "logits/chosen": 0.07763671875, + "logits/rejected": -0.0556640625, + "logps/chosen": -358.0, + "logps/rejected": -330.0, + "loss": 0.3304, + "rewards/accuracies": 0.8046875, + "rewards/chosen": -0.55859375, + "rewards/margins": 6.78125, + "rewards/rejected": -7.34375, + "step": 483 + }, + { + "epoch": 0.5938650306748466, + "grad_norm": 31.73235234166197, + "learning_rate": 1.215495867768595e-06, + "logits/chosen": 0.00390625, + "logits/rejected": -0.142578125, + "logps/chosen": -380.0, + "logps/rejected": -340.0, + "loss": 0.3937, + "rewards/accuracies": 0.828125, + "rewards/chosen": 0.1611328125, + "rewards/margins": 6.46875, + "rewards/rejected": -6.3125, + "step": 484 + }, + { + "epoch": 0.5950920245398773, + "grad_norm": 28.55956265398308, + "learning_rate": 1.2148760330578513e-06, + "logits/chosen": 0.023681640625, + "logits/rejected": -0.12109375, + "logps/chosen": -370.0, + "logps/rejected": -324.0, + "loss": 0.3173, + "rewards/accuracies": 0.859375, + "rewards/chosen": -0.36328125, + "rewards/margins": 6.9375, + "rewards/rejected": -7.3125, + "step": 485 + }, + { + "epoch": 0.596319018404908, + "grad_norm": 33.81986931883939, + "learning_rate": 1.2142561983471074e-06, + "logits/chosen": -0.0050048828125, + "logits/rejected": -0.14453125, + "logps/chosen": -414.0, + "logps/rejected": -352.0, + "loss": 0.3778, + "rewards/accuracies": 0.875, + "rewards/chosen": 0.455078125, + "rewards/margins": 7.0, + "rewards/rejected": -6.53125, + "step": 486 + }, + { + "epoch": 0.5975460122699386, + "grad_norm": 30.028596310368773, + "learning_rate": 1.2136363636363636e-06, + "logits/chosen": -0.08251953125, + "logits/rejected": -0.27734375, + "logps/chosen": -410.0, + "logps/rejected": -366.0, + "loss": 0.2781, + "rewards/accuracies": 0.875, + "rewards/chosen": 0.33984375, + "rewards/margins": 7.375, + "rewards/rejected": -7.03125, + "step": 487 + }, + { + "epoch": 0.5987730061349693, + "grad_norm": 35.71061050092003, + "learning_rate": 1.21301652892562e-06, + "logits/chosen": -0.0693359375, + "logits/rejected": -0.138671875, + "logps/chosen": -424.0, + "logps/rejected": -370.0, + "loss": 0.3544, + "rewards/accuracies": 0.84375, + "rewards/chosen": 0.5078125, + "rewards/margins": 6.96875, + "rewards/rejected": -6.46875, + "step": 488 + }, + { + "epoch": 0.6, + "grad_norm": 26.138379124419743, + "learning_rate": 1.212396694214876e-06, + "logits/chosen": 0.01324462890625, + "logits/rejected": -0.224609375, + "logps/chosen": -384.0, + "logps/rejected": -340.0, + "loss": 0.3008, + "rewards/accuracies": 0.859375, + "rewards/chosen": 0.92578125, + "rewards/margins": 7.875, + "rewards/rejected": -6.9375, + "step": 489 + }, + { + "epoch": 0.6012269938650306, + "grad_norm": 38.68817736099095, + "learning_rate": 1.2117768595041322e-06, + "logits/chosen": 0.055419921875, + "logits/rejected": -0.1689453125, + "logps/chosen": -432.0, + "logps/rejected": -364.0, + "loss": 0.404, + "rewards/accuracies": 0.8125, + "rewards/chosen": 0.263671875, + "rewards/margins": 6.84375, + "rewards/rejected": -6.5625, + "step": 490 + }, + { + "epoch": 0.6024539877300613, + "grad_norm": 30.239394074669093, + "learning_rate": 1.2111570247933886e-06, + "logits/chosen": 0.07666015625, + "logits/rejected": -0.134765625, + "logps/chosen": -372.0, + "logps/rejected": -328.0, + "loss": 0.3068, + "rewards/accuracies": 0.8828125, + "rewards/chosen": 0.443359375, + "rewards/margins": 6.96875, + "rewards/rejected": -6.53125, + "step": 491 + }, + { + "epoch": 0.603680981595092, + "grad_norm": 33.80787533439265, + "learning_rate": 1.2105371900826447e-06, + "logits/chosen": -0.013427734375, + "logits/rejected": -0.1533203125, + "logps/chosen": -398.0, + "logps/rejected": -390.0, + "loss": 0.3711, + "rewards/accuracies": 0.8515625, + "rewards/chosen": 0.5625, + "rewards/margins": 6.6875, + "rewards/rejected": -6.125, + "step": 492 + }, + { + "epoch": 0.6049079754601226, + "grad_norm": 41.137878058973136, + "learning_rate": 1.2099173553719008e-06, + "logits/chosen": 0.14453125, + "logits/rejected": 0.043701171875, + "logps/chosen": -392.0, + "logps/rejected": -340.0, + "loss": 0.5049, + "rewards/accuracies": 0.796875, + "rewards/chosen": 0.19921875, + "rewards/margins": 5.65625, + "rewards/rejected": -5.4375, + "step": 493 + }, + { + "epoch": 0.6061349693251534, + "grad_norm": 28.550910744803765, + "learning_rate": 1.209297520661157e-06, + "logits/chosen": 0.044921875, + "logits/rejected": -0.1435546875, + "logps/chosen": -428.0, + "logps/rejected": -340.0, + "loss": 0.3041, + "rewards/accuracies": 0.875, + "rewards/chosen": 1.1796875, + "rewards/margins": 7.21875, + "rewards/rejected": -6.0625, + "step": 494 + }, + { + "epoch": 0.6073619631901841, + "grad_norm": 32.18887165372733, + "learning_rate": 1.2086776859504131e-06, + "logits/chosen": 0.031494140625, + "logits/rejected": -0.1064453125, + "logps/chosen": -402.0, + "logps/rejected": -350.0, + "loss": 0.3067, + "rewards/accuracies": 0.859375, + "rewards/chosen": 1.0859375, + "rewards/margins": 7.25, + "rewards/rejected": -6.15625, + "step": 495 + }, + { + "epoch": 0.6085889570552148, + "grad_norm": 34.7968597037722, + "learning_rate": 1.2080578512396695e-06, + "logits/chosen": -0.0025634765625, + "logits/rejected": -0.21484375, + "logps/chosen": -382.0, + "logps/rejected": -322.0, + "loss": 0.3605, + "rewards/accuracies": 0.859375, + "rewards/chosen": 0.1708984375, + "rewards/margins": 6.25, + "rewards/rejected": -6.09375, + "step": 496 + }, + { + "epoch": 0.6098159509202454, + "grad_norm": 36.914820937247434, + "learning_rate": 1.2074380165289256e-06, + "logits/chosen": 0.0361328125, + "logits/rejected": -0.125, + "logps/chosen": -404.0, + "logps/rejected": -366.0, + "loss": 0.4718, + "rewards/accuracies": 0.78125, + "rewards/chosen": 0.1337890625, + "rewards/margins": 6.5, + "rewards/rejected": -6.375, + "step": 497 + }, + { + "epoch": 0.6110429447852761, + "grad_norm": 27.898760612078714, + "learning_rate": 1.206818181818182e-06, + "logits/chosen": -0.10302734375, + "logits/rejected": -0.265625, + "logps/chosen": -400.0, + "logps/rejected": -366.0, + "loss": 0.2974, + "rewards/accuracies": 0.8359375, + "rewards/chosen": 0.9765625, + "rewards/margins": 7.46875, + "rewards/rejected": -6.5, + "step": 498 + }, + { + "epoch": 0.6122699386503068, + "grad_norm": 33.514362873417944, + "learning_rate": 1.2061983471074381e-06, + "logits/chosen": 0.041015625, + "logits/rejected": -0.177734375, + "logps/chosen": -426.0, + "logps/rejected": -350.0, + "loss": 0.3147, + "rewards/accuracies": 0.8671875, + "rewards/chosen": 0.984375, + "rewards/margins": 6.96875, + "rewards/rejected": -6.0, + "step": 499 + }, + { + "epoch": 0.6134969325153374, + "grad_norm": 30.177881260758163, + "learning_rate": 1.2055785123966943e-06, + "logits/chosen": 0.0224609375, + "logits/rejected": -0.09375, + "logps/chosen": -418.0, + "logps/rejected": -338.0, + "loss": 0.3575, + "rewards/accuracies": 0.796875, + "rewards/chosen": 1.0390625, + "rewards/margins": 7.3125, + "rewards/rejected": -6.28125, + "step": 500 + }, + { + "epoch": 0.6147239263803681, + "grad_norm": 34.255808614960706, + "learning_rate": 1.2049586776859504e-06, + "logits/chosen": -0.0390625, + "logits/rejected": -0.171875, + "logps/chosen": -410.0, + "logps/rejected": -352.0, + "loss": 0.4018, + "rewards/accuracies": 0.8359375, + "rewards/chosen": 1.234375, + "rewards/margins": 7.59375, + "rewards/rejected": -6.34375, + "step": 501 + }, + { + "epoch": 0.6159509202453988, + "grad_norm": 28.29113723996875, + "learning_rate": 1.2043388429752065e-06, + "logits/chosen": -0.01165771484375, + "logits/rejected": -0.28125, + "logps/chosen": -386.0, + "logps/rejected": -340.0, + "loss": 0.3467, + "rewards/accuracies": 0.8828125, + "rewards/chosen": 0.6796875, + "rewards/margins": 7.71875, + "rewards/rejected": -7.03125, + "step": 502 + }, + { + "epoch": 0.6171779141104294, + "grad_norm": 29.752413229458206, + "learning_rate": 1.2037190082644629e-06, + "logits/chosen": 0.1904296875, + "logits/rejected": -0.0013275146484375, + "logps/chosen": -334.0, + "logps/rejected": -324.0, + "loss": 0.2996, + "rewards/accuracies": 0.8046875, + "rewards/chosen": 0.28515625, + "rewards/margins": 6.6875, + "rewards/rejected": -6.40625, + "step": 503 + }, + { + "epoch": 0.6184049079754601, + "grad_norm": 24.506669229979327, + "learning_rate": 1.203099173553719e-06, + "logits/chosen": -0.06201171875, + "logits/rejected": -0.234375, + "logps/chosen": -388.0, + "logps/rejected": -340.0, + "loss": 0.2718, + "rewards/accuracies": 0.8984375, + "rewards/chosen": -0.1025390625, + "rewards/margins": 6.21875, + "rewards/rejected": -6.3125, + "step": 504 + }, + { + "epoch": 0.6196319018404908, + "grad_norm": 38.20086418972538, + "learning_rate": 1.2024793388429752e-06, + "logits/chosen": -0.07666015625, + "logits/rejected": -0.298828125, + "logps/chosen": -434.0, + "logps/rejected": -406.0, + "loss": 0.4423, + "rewards/accuracies": 0.828125, + "rewards/chosen": -0.609375, + "rewards/margins": 6.78125, + "rewards/rejected": -7.375, + "step": 505 + }, + { + "epoch": 0.6208588957055214, + "grad_norm": 67.15816433272404, + "learning_rate": 1.2018595041322315e-06, + "logits/chosen": 0.06982421875, + "logits/rejected": -0.2373046875, + "logps/chosen": -428.0, + "logps/rejected": -340.0, + "loss": 0.3369, + "rewards/accuracies": 0.859375, + "rewards/chosen": -0.0081787109375, + "rewards/margins": 7.03125, + "rewards/rejected": -7.03125, + "step": 506 + }, + { + "epoch": 0.6220858895705521, + "grad_norm": 25.2127178538785, + "learning_rate": 1.2012396694214877e-06, + "logits/chosen": -0.04296875, + "logits/rejected": -0.150390625, + "logps/chosen": -394.0, + "logps/rejected": -318.0, + "loss": 0.1989, + "rewards/accuracies": 0.8984375, + "rewards/chosen": 0.11474609375, + "rewards/margins": 7.34375, + "rewards/rejected": -7.21875, + "step": 507 + }, + { + "epoch": 0.6233128834355828, + "grad_norm": 31.70942946416519, + "learning_rate": 1.2006198347107438e-06, + "logits/chosen": 0.14453125, + "logits/rejected": -0.033935546875, + "logps/chosen": -396.0, + "logps/rejected": -320.0, + "loss": 0.3214, + "rewards/accuracies": 0.84375, + "rewards/chosen": -0.1494140625, + "rewards/margins": 6.40625, + "rewards/rejected": -6.5625, + "step": 508 + }, + { + "epoch": 0.6245398773006134, + "grad_norm": 39.83330773441419, + "learning_rate": 1.2000000000000002e-06, + "logits/chosen": -0.0225830078125, + "logits/rejected": -0.2578125, + "logps/chosen": -474.0, + "logps/rejected": -396.0, + "loss": 0.569, + "rewards/accuracies": 0.8125, + "rewards/chosen": 1.0390625, + "rewards/margins": 8.25, + "rewards/rejected": -7.21875, + "step": 509 + }, + { + "epoch": 0.6257668711656442, + "grad_norm": 29.863730506340346, + "learning_rate": 1.199380165289256e-06, + "logits/chosen": -0.11181640625, + "logits/rejected": -0.291015625, + "logps/chosen": -368.0, + "logps/rejected": -314.0, + "loss": 0.2926, + "rewards/accuracies": 0.875, + "rewards/chosen": -0.375, + "rewards/margins": 6.3125, + "rewards/rejected": -6.6875, + "step": 510 + }, + { + "epoch": 0.6269938650306749, + "grad_norm": 35.188221718693796, + "learning_rate": 1.1987603305785124e-06, + "logits/chosen": -0.1328125, + "logits/rejected": -0.259765625, + "logps/chosen": -402.0, + "logps/rejected": -366.0, + "loss": 0.4071, + "rewards/accuracies": 0.8359375, + "rewards/chosen": 0.0859375, + "rewards/margins": 7.3125, + "rewards/rejected": -7.21875, + "step": 511 + }, + { + "epoch": 0.6282208588957056, + "grad_norm": 35.23659404136466, + "learning_rate": 1.1981404958677686e-06, + "logits/chosen": 0.0791015625, + "logits/rejected": -0.0654296875, + "logps/chosen": -388.0, + "logps/rejected": -398.0, + "loss": 0.4354, + "rewards/accuracies": 0.8203125, + "rewards/chosen": -0.388671875, + "rewards/margins": 5.90625, + "rewards/rejected": -6.28125, + "step": 512 + }, + { + "epoch": 0.6294478527607362, + "grad_norm": 33.94396537230905, + "learning_rate": 1.1975206611570247e-06, + "logits/chosen": 0.07958984375, + "logits/rejected": -0.0517578125, + "logps/chosen": -384.0, + "logps/rejected": -326.0, + "loss": 0.4151, + "rewards/accuracies": 0.796875, + "rewards/chosen": 0.1328125, + "rewards/margins": 6.96875, + "rewards/rejected": -6.8125, + "step": 513 + }, + { + "epoch": 0.6306748466257669, + "grad_norm": 27.33280439239978, + "learning_rate": 1.196900826446281e-06, + "logits/chosen": 0.09765625, + "logits/rejected": -0.0537109375, + "logps/chosen": -340.0, + "logps/rejected": -302.0, + "loss": 0.2677, + "rewards/accuracies": 0.8828125, + "rewards/chosen": -1.015625, + "rewards/margins": 6.5, + "rewards/rejected": -7.53125, + "step": 514 + }, + { + "epoch": 0.6319018404907976, + "grad_norm": 32.16724360242484, + "learning_rate": 1.1962809917355372e-06, + "logits/chosen": 0.07275390625, + "logits/rejected": -0.07763671875, + "logps/chosen": -358.0, + "logps/rejected": -362.0, + "loss": 0.3029, + "rewards/accuracies": 0.8515625, + "rewards/chosen": -1.0703125, + "rewards/margins": 6.5, + "rewards/rejected": -7.5625, + "step": 515 + }, + { + "epoch": 0.6331288343558282, + "grad_norm": 26.23429258203521, + "learning_rate": 1.1956611570247934e-06, + "logits/chosen": 0.0279541015625, + "logits/rejected": -0.07861328125, + "logps/chosen": -378.0, + "logps/rejected": -346.0, + "loss": 0.2424, + "rewards/accuracies": 0.8828125, + "rewards/chosen": -0.8828125, + "rewards/margins": 6.65625, + "rewards/rejected": -7.53125, + "step": 516 + }, + { + "epoch": 0.6343558282208589, + "grad_norm": 37.00771401511282, + "learning_rate": 1.1950413223140497e-06, + "logits/chosen": 0.0703125, + "logits/rejected": -0.08642578125, + "logps/chosen": -400.0, + "logps/rejected": -352.0, + "loss": 0.4735, + "rewards/accuracies": 0.78125, + "rewards/chosen": -0.48828125, + "rewards/margins": 6.28125, + "rewards/rejected": -6.78125, + "step": 517 + }, + { + "epoch": 0.6355828220858896, + "grad_norm": 28.05157133110184, + "learning_rate": 1.1944214876033059e-06, + "logits/chosen": -0.04541015625, + "logits/rejected": -0.150390625, + "logps/chosen": -424.0, + "logps/rejected": -360.0, + "loss": 0.2475, + "rewards/accuracies": 0.8671875, + "rewards/chosen": 0.396484375, + "rewards/margins": 7.625, + "rewards/rejected": -7.21875, + "step": 518 + }, + { + "epoch": 0.6368098159509202, + "grad_norm": 38.286039724245256, + "learning_rate": 1.193801652892562e-06, + "logits/chosen": -0.026123046875, + "logits/rejected": -0.1630859375, + "logps/chosen": -396.0, + "logps/rejected": -354.0, + "loss": 0.3562, + "rewards/accuracies": 0.828125, + "rewards/chosen": -0.60546875, + "rewards/margins": 6.9375, + "rewards/rejected": -7.53125, + "step": 519 + }, + { + "epoch": 0.6380368098159509, + "grad_norm": 34.57537616564543, + "learning_rate": 1.1931818181818181e-06, + "logits/chosen": -0.0517578125, + "logits/rejected": -0.279296875, + "logps/chosen": -380.0, + "logps/rejected": -350.0, + "loss": 0.37, + "rewards/accuracies": 0.84375, + "rewards/chosen": -0.271484375, + "rewards/margins": 7.71875, + "rewards/rejected": -8.0, + "step": 520 + }, + { + "epoch": 0.6392638036809816, + "grad_norm": 29.132838683315903, + "learning_rate": 1.1925619834710743e-06, + "logits/chosen": 0.158203125, + "logits/rejected": 0.01312255859375, + "logps/chosen": -428.0, + "logps/rejected": -378.0, + "loss": 0.2975, + "rewards/accuracies": 0.859375, + "rewards/chosen": 0.01263427734375, + "rewards/margins": 7.8125, + "rewards/rejected": -7.8125, + "step": 521 + }, + { + "epoch": 0.6404907975460122, + "grad_norm": 30.38859952385869, + "learning_rate": 1.1919421487603306e-06, + "logits/chosen": -0.0546875, + "logits/rejected": -0.2060546875, + "logps/chosen": -438.0, + "logps/rejected": -374.0, + "loss": 0.3466, + "rewards/accuracies": 0.8359375, + "rewards/chosen": -0.439453125, + "rewards/margins": 8.1875, + "rewards/rejected": -8.625, + "step": 522 + }, + { + "epoch": 0.6417177914110429, + "grad_norm": 36.440630501558566, + "learning_rate": 1.1913223140495868e-06, + "logits/chosen": 0.0238037109375, + "logits/rejected": -0.12890625, + "logps/chosen": -380.0, + "logps/rejected": -346.0, + "loss": 0.4467, + "rewards/accuracies": 0.8125, + "rewards/chosen": -1.09375, + "rewards/margins": 6.3125, + "rewards/rejected": -7.40625, + "step": 523 + }, + { + "epoch": 0.6429447852760736, + "grad_norm": 39.231543678462224, + "learning_rate": 1.1907024793388431e-06, + "logits/chosen": 0.05615234375, + "logits/rejected": -0.1474609375, + "logps/chosen": -440.0, + "logps/rejected": -380.0, + "loss": 0.4136, + "rewards/accuracies": 0.859375, + "rewards/chosen": -1.1953125, + "rewards/margins": 6.40625, + "rewards/rejected": -7.59375, + "step": 524 + }, + { + "epoch": 0.6441717791411042, + "grad_norm": 47.6608831234207, + "learning_rate": 1.1900826446280993e-06, + "logits/chosen": 0.07666015625, + "logits/rejected": 0.02587890625, + "logps/chosen": -406.0, + "logps/rejected": -382.0, + "loss": 0.5676, + "rewards/accuracies": 0.7421875, + "rewards/chosen": -1.3125, + "rewards/margins": 5.53125, + "rewards/rejected": -6.84375, + "step": 525 + }, + { + "epoch": 0.645398773006135, + "grad_norm": 31.7891594700196, + "learning_rate": 1.1894628099173554e-06, + "logits/chosen": 0.0458984375, + "logits/rejected": -0.10107421875, + "logps/chosen": -402.0, + "logps/rejected": -356.0, + "loss": 0.3316, + "rewards/accuracies": 0.8203125, + "rewards/chosen": -0.6171875, + "rewards/margins": 7.21875, + "rewards/rejected": -7.84375, + "step": 526 + }, + { + "epoch": 0.6466257668711657, + "grad_norm": 32.21728261594982, + "learning_rate": 1.1888429752066118e-06, + "logits/chosen": 0.1201171875, + "logits/rejected": -0.103515625, + "logps/chosen": -438.0, + "logps/rejected": -348.0, + "loss": 0.3529, + "rewards/accuracies": 0.8359375, + "rewards/chosen": -0.7265625, + "rewards/margins": 7.03125, + "rewards/rejected": -7.75, + "step": 527 + }, + { + "epoch": 0.6478527607361964, + "grad_norm": 26.60182420517136, + "learning_rate": 1.1882231404958677e-06, + "logits/chosen": -0.0380859375, + "logits/rejected": -0.220703125, + "logps/chosen": -368.0, + "logps/rejected": -344.0, + "loss": 0.2928, + "rewards/accuracies": 0.859375, + "rewards/chosen": -0.671875, + "rewards/margins": 7.375, + "rewards/rejected": -8.0625, + "step": 528 + }, + { + "epoch": 0.649079754601227, + "grad_norm": 32.109755070015275, + "learning_rate": 1.187603305785124e-06, + "logits/chosen": 0.030029296875, + "logits/rejected": -0.14453125, + "logps/chosen": -352.0, + "logps/rejected": -352.0, + "loss": 0.418, + "rewards/accuracies": 0.8203125, + "rewards/chosen": -0.3984375, + "rewards/margins": 6.9375, + "rewards/rejected": -7.34375, + "step": 529 + }, + { + "epoch": 0.6503067484662577, + "grad_norm": 30.441331665820965, + "learning_rate": 1.1869834710743802e-06, + "logits/chosen": 0.07373046875, + "logits/rejected": -0.0517578125, + "logps/chosen": -370.0, + "logps/rejected": -366.0, + "loss": 0.2962, + "rewards/accuracies": 0.8515625, + "rewards/chosen": -0.1103515625, + "rewards/margins": 7.5, + "rewards/rejected": -7.59375, + "step": 530 + }, + { + "epoch": 0.6515337423312884, + "grad_norm": 25.97427159310545, + "learning_rate": 1.1863636363636363e-06, + "logits/chosen": -0.0005950927734375, + "logits/rejected": -0.1767578125, + "logps/chosen": -394.0, + "logps/rejected": -338.0, + "loss": 0.3013, + "rewards/accuracies": 0.8671875, + "rewards/chosen": -0.2001953125, + "rewards/margins": 7.0625, + "rewards/rejected": -7.25, + "step": 531 + }, + { + "epoch": 0.652760736196319, + "grad_norm": 29.45586726470446, + "learning_rate": 1.1857438016528927e-06, + "logits/chosen": 0.12060546875, + "logits/rejected": -0.01611328125, + "logps/chosen": -394.0, + "logps/rejected": -342.0, + "loss": 0.3256, + "rewards/accuracies": 0.8671875, + "rewards/chosen": -0.1650390625, + "rewards/margins": 6.5625, + "rewards/rejected": -6.71875, + "step": 532 + }, + { + "epoch": 0.6539877300613497, + "grad_norm": 31.510396061786174, + "learning_rate": 1.1851239669421488e-06, + "logits/chosen": -0.0634765625, + "logits/rejected": -0.265625, + "logps/chosen": -414.0, + "logps/rejected": -360.0, + "loss": 0.322, + "rewards/accuracies": 0.8359375, + "rewards/chosen": 0.71484375, + "rewards/margins": 7.78125, + "rewards/rejected": -7.09375, + "step": 533 + }, + { + "epoch": 0.6552147239263804, + "grad_norm": 29.028681159770606, + "learning_rate": 1.184504132231405e-06, + "logits/chosen": 0.0908203125, + "logits/rejected": -0.02685546875, + "logps/chosen": -374.0, + "logps/rejected": -338.0, + "loss": 0.3243, + "rewards/accuracies": 0.8203125, + "rewards/chosen": 0.0341796875, + "rewards/margins": 6.40625, + "rewards/rejected": -6.375, + "step": 534 + }, + { + "epoch": 0.656441717791411, + "grad_norm": 33.385986566673, + "learning_rate": 1.1838842975206613e-06, + "logits/chosen": -0.022705078125, + "logits/rejected": -0.125, + "logps/chosen": -378.0, + "logps/rejected": -348.0, + "loss": 0.3502, + "rewards/accuracies": 0.84375, + "rewards/chosen": -0.7421875, + "rewards/margins": 6.40625, + "rewards/rejected": -7.125, + "step": 535 + }, + { + "epoch": 0.6576687116564417, + "grad_norm": 27.982276879832543, + "learning_rate": 1.1832644628099175e-06, + "logits/chosen": 0.0201416015625, + "logits/rejected": -0.07568359375, + "logps/chosen": -394.0, + "logps/rejected": -330.0, + "loss": 0.2519, + "rewards/accuracies": 0.890625, + "rewards/chosen": 0.6015625, + "rewards/margins": 7.40625, + "rewards/rejected": -6.78125, + "step": 536 + }, + { + "epoch": 0.6588957055214724, + "grad_norm": 40.68442319965527, + "learning_rate": 1.1826446280991736e-06, + "logits/chosen": 0.06201171875, + "logits/rejected": -0.0927734375, + "logps/chosen": -386.0, + "logps/rejected": -318.0, + "loss": 0.5685, + "rewards/accuracies": 0.7734375, + "rewards/chosen": -0.50390625, + "rewards/margins": 5.28125, + "rewards/rejected": -5.78125, + "step": 537 + }, + { + "epoch": 0.660122699386503, + "grad_norm": 35.32326255751808, + "learning_rate": 1.1820247933884297e-06, + "logits/chosen": 0.068359375, + "logits/rejected": -0.09912109375, + "logps/chosen": -402.0, + "logps/rejected": -360.0, + "loss": 0.4164, + "rewards/accuracies": 0.84375, + "rewards/chosen": 0.78515625, + "rewards/margins": 6.75, + "rewards/rejected": -5.96875, + "step": 538 + }, + { + "epoch": 0.6613496932515337, + "grad_norm": 32.525023600229474, + "learning_rate": 1.1814049586776859e-06, + "logits/chosen": 0.07666015625, + "logits/rejected": -0.04541015625, + "logps/chosen": -400.0, + "logps/rejected": -346.0, + "loss": 0.4519, + "rewards/accuracies": 0.8125, + "rewards/chosen": 0.80859375, + "rewards/margins": 5.84375, + "rewards/rejected": -5.03125, + "step": 539 + }, + { + "epoch": 0.6625766871165644, + "grad_norm": 35.156155067546834, + "learning_rate": 1.1807851239669422e-06, + "logits/chosen": 0.0986328125, + "logits/rejected": -0.06787109375, + "logps/chosen": -396.0, + "logps/rejected": -332.0, + "loss": 0.432, + "rewards/accuracies": 0.8515625, + "rewards/chosen": 0.80859375, + "rewards/margins": 6.71875, + "rewards/rejected": -5.9375, + "step": 540 + }, + { + "epoch": 0.6638036809815951, + "grad_norm": 45.099896848619814, + "learning_rate": 1.1801652892561984e-06, + "logits/chosen": 0.08447265625, + "logits/rejected": -0.0064697265625, + "logps/chosen": -390.0, + "logps/rejected": -354.0, + "loss": 0.3808, + "rewards/accuracies": 0.828125, + "rewards/chosen": 0.279296875, + "rewards/margins": 5.84375, + "rewards/rejected": -5.5625, + "step": 541 + }, + { + "epoch": 0.6650306748466258, + "grad_norm": 30.118145811004524, + "learning_rate": 1.1795454545454545e-06, + "logits/chosen": 0.031494140625, + "logits/rejected": -0.053466796875, + "logps/chosen": -356.0, + "logps/rejected": -320.0, + "loss": 0.3659, + "rewards/accuracies": 0.859375, + "rewards/chosen": 0.87109375, + "rewards/margins": 6.03125, + "rewards/rejected": -5.15625, + "step": 542 + }, + { + "epoch": 0.6662576687116565, + "grad_norm": 32.83770736751292, + "learning_rate": 1.1789256198347109e-06, + "logits/chosen": -0.02294921875, + "logits/rejected": -0.201171875, + "logps/chosen": -450.0, + "logps/rejected": -354.0, + "loss": 0.3396, + "rewards/accuracies": 0.8359375, + "rewards/chosen": 1.5078125, + "rewards/margins": 6.875, + "rewards/rejected": -5.375, + "step": 543 + }, + { + "epoch": 0.6674846625766871, + "grad_norm": 35.490916209773005, + "learning_rate": 1.178305785123967e-06, + "logits/chosen": 0.007598876953125, + "logits/rejected": -0.1318359375, + "logps/chosen": -378.0, + "logps/rejected": -318.0, + "loss": 0.434, + "rewards/accuracies": 0.8359375, + "rewards/chosen": 0.376953125, + "rewards/margins": 6.15625, + "rewards/rejected": -5.78125, + "step": 544 + }, + { + "epoch": 0.6687116564417178, + "grad_norm": 33.60719342971152, + "learning_rate": 1.1776859504132231e-06, + "logits/chosen": -0.054443359375, + "logits/rejected": -0.21875, + "logps/chosen": -458.0, + "logps/rejected": -368.0, + "loss": 0.3855, + "rewards/accuracies": 0.8125, + "rewards/chosen": 0.94140625, + "rewards/margins": 6.1875, + "rewards/rejected": -5.25, + "step": 545 + }, + { + "epoch": 0.6699386503067485, + "grad_norm": 33.972456048977826, + "learning_rate": 1.1770661157024793e-06, + "logits/chosen": 0.0693359375, + "logits/rejected": -0.028564453125, + "logps/chosen": -374.0, + "logps/rejected": -338.0, + "loss": 0.3815, + "rewards/accuracies": 0.828125, + "rewards/chosen": 0.6953125, + "rewards/margins": 6.40625, + "rewards/rejected": -5.71875, + "step": 546 + }, + { + "epoch": 0.6711656441717792, + "grad_norm": 28.303815223441166, + "learning_rate": 1.1764462809917354e-06, + "logits/chosen": 0.0133056640625, + "logits/rejected": -0.166015625, + "logps/chosen": -386.0, + "logps/rejected": -312.0, + "loss": 0.2992, + "rewards/accuracies": 0.84375, + "rewards/chosen": 0.88671875, + "rewards/margins": 6.4375, + "rewards/rejected": -5.53125, + "step": 547 + }, + { + "epoch": 0.6723926380368098, + "grad_norm": 38.31133772907115, + "learning_rate": 1.1758264462809918e-06, + "logits/chosen": -0.06494140625, + "logits/rejected": -0.265625, + "logps/chosen": -402.0, + "logps/rejected": -358.0, + "loss": 0.3532, + "rewards/accuracies": 0.8515625, + "rewards/chosen": 1.015625, + "rewards/margins": 6.125, + "rewards/rejected": -5.125, + "step": 548 + }, + { + "epoch": 0.6736196319018405, + "grad_norm": 33.292222723297044, + "learning_rate": 1.175206611570248e-06, + "logits/chosen": 0.047607421875, + "logits/rejected": -0.1572265625, + "logps/chosen": -414.0, + "logps/rejected": -344.0, + "loss": 0.3292, + "rewards/accuracies": 0.8359375, + "rewards/chosen": 0.5625, + "rewards/margins": 5.96875, + "rewards/rejected": -5.40625, + "step": 549 + }, + { + "epoch": 0.6748466257668712, + "grad_norm": 30.807870745155437, + "learning_rate": 1.174586776859504e-06, + "logits/chosen": -0.024658203125, + "logits/rejected": -0.08642578125, + "logps/chosen": -402.0, + "logps/rejected": -330.0, + "loss": 0.3926, + "rewards/accuracies": 0.8125, + "rewards/chosen": 0.12890625, + "rewards/margins": 5.71875, + "rewards/rejected": -5.5625, + "step": 550 + }, + { + "epoch": 0.6760736196319018, + "grad_norm": 28.384977078060157, + "learning_rate": 1.1739669421487604e-06, + "logits/chosen": 0.06787109375, + "logits/rejected": -0.04541015625, + "logps/chosen": -410.0, + "logps/rejected": -346.0, + "loss": 0.3704, + "rewards/accuracies": 0.859375, + "rewards/chosen": 1.2109375, + "rewards/margins": 6.75, + "rewards/rejected": -5.53125, + "step": 551 + }, + { + "epoch": 0.6773006134969325, + "grad_norm": 29.63861597355486, + "learning_rate": 1.1733471074380166e-06, + "logits/chosen": -0.022705078125, + "logits/rejected": -0.158203125, + "logps/chosen": -416.0, + "logps/rejected": -360.0, + "loss": 0.3331, + "rewards/accuracies": 0.828125, + "rewards/chosen": 0.71875, + "rewards/margins": 6.25, + "rewards/rejected": -5.53125, + "step": 552 + }, + { + "epoch": 0.6785276073619632, + "grad_norm": 32.698792932077886, + "learning_rate": 1.172727272727273e-06, + "logits/chosen": 0.01806640625, + "logits/rejected": -0.1259765625, + "logps/chosen": -366.0, + "logps/rejected": -342.0, + "loss": 0.4129, + "rewards/accuracies": 0.8359375, + "rewards/chosen": -0.1474609375, + "rewards/margins": 5.625, + "rewards/rejected": -5.78125, + "step": 553 + }, + { + "epoch": 0.6797546012269938, + "grad_norm": 32.72020008420759, + "learning_rate": 1.172107438016529e-06, + "logits/chosen": 0.0791015625, + "logits/rejected": -0.1171875, + "logps/chosen": -434.0, + "logps/rejected": -334.0, + "loss": 0.3464, + "rewards/accuracies": 0.84375, + "rewards/chosen": 0.83984375, + "rewards/margins": 6.4375, + "rewards/rejected": -5.59375, + "step": 554 + }, + { + "epoch": 0.6809815950920245, + "grad_norm": 30.608371906092913, + "learning_rate": 1.171487603305785e-06, + "logits/chosen": 0.2021484375, + "logits/rejected": 0.09130859375, + "logps/chosen": -364.0, + "logps/rejected": -318.0, + "loss": 0.3666, + "rewards/accuracies": 0.8046875, + "rewards/chosen": 0.3515625, + "rewards/margins": 6.34375, + "rewards/rejected": -6.0, + "step": 555 + }, + { + "epoch": 0.6822085889570552, + "grad_norm": 26.18493705591653, + "learning_rate": 1.1708677685950413e-06, + "logits/chosen": 0.053466796875, + "logits/rejected": -0.162109375, + "logps/chosen": -424.0, + "logps/rejected": -360.0, + "loss": 0.2487, + "rewards/accuracies": 0.890625, + "rewards/chosen": 0.96875, + "rewards/margins": 7.5625, + "rewards/rejected": -6.59375, + "step": 556 + }, + { + "epoch": 0.6834355828220859, + "grad_norm": 27.830299010779843, + "learning_rate": 1.1702479338842975e-06, + "logits/chosen": 0.05029296875, + "logits/rejected": -0.16796875, + "logps/chosen": -398.0, + "logps/rejected": -346.0, + "loss": 0.3343, + "rewards/accuracies": 0.828125, + "rewards/chosen": 0.53515625, + "rewards/margins": 6.6875, + "rewards/rejected": -6.15625, + "step": 557 + }, + { + "epoch": 0.6846625766871166, + "grad_norm": 25.854408165941038, + "learning_rate": 1.1696280991735538e-06, + "logits/chosen": 0.087890625, + "logits/rejected": -0.078125, + "logps/chosen": -378.0, + "logps/rejected": -340.0, + "loss": 0.2782, + "rewards/accuracies": 0.8671875, + "rewards/chosen": -0.3125, + "rewards/margins": 6.9375, + "rewards/rejected": -7.25, + "step": 558 + }, + { + "epoch": 0.6858895705521473, + "grad_norm": 33.3339739767785, + "learning_rate": 1.16900826446281e-06, + "logits/chosen": -0.072265625, + "logits/rejected": -0.1591796875, + "logps/chosen": -388.0, + "logps/rejected": -354.0, + "loss": 0.3848, + "rewards/accuracies": 0.828125, + "rewards/chosen": -0.255859375, + "rewards/margins": 6.5, + "rewards/rejected": -6.78125, + "step": 559 + }, + { + "epoch": 0.6871165644171779, + "grad_norm": 21.74474553758103, + "learning_rate": 1.1683884297520661e-06, + "logits/chosen": -0.1669921875, + "logits/rejected": -0.302734375, + "logps/chosen": -394.0, + "logps/rejected": -344.0, + "loss": 0.2247, + "rewards/accuracies": 0.9140625, + "rewards/chosen": 0.29296875, + "rewards/margins": 7.96875, + "rewards/rejected": -7.6875, + "step": 560 + }, + { + "epoch": 0.6883435582822086, + "grad_norm": 24.604934270788146, + "learning_rate": 1.1677685950413225e-06, + "logits/chosen": 0.00164031982421875, + "logits/rejected": -0.177734375, + "logps/chosen": -360.0, + "logps/rejected": -314.0, + "loss": 0.2132, + "rewards/accuracies": 0.890625, + "rewards/chosen": 0.171875, + "rewards/margins": 7.25, + "rewards/rejected": -7.09375, + "step": 561 + }, + { + "epoch": 0.6895705521472393, + "grad_norm": 25.33619721263671, + "learning_rate": 1.1671487603305786e-06, + "logits/chosen": 0.142578125, + "logits/rejected": -0.064453125, + "logps/chosen": -382.0, + "logps/rejected": -336.0, + "loss": 0.2353, + "rewards/accuracies": 0.875, + "rewards/chosen": -0.55078125, + "rewards/margins": 6.75, + "rewards/rejected": -7.3125, + "step": 562 + }, + { + "epoch": 0.69079754601227, + "grad_norm": 28.2123146430669, + "learning_rate": 1.1665289256198348e-06, + "logits/chosen": -0.055908203125, + "logits/rejected": -0.2041015625, + "logps/chosen": -398.0, + "logps/rejected": -386.0, + "loss": 0.3157, + "rewards/accuracies": 0.8828125, + "rewards/chosen": 0.71484375, + "rewards/margins": 8.0, + "rewards/rejected": -7.3125, + "step": 563 + }, + { + "epoch": 0.6920245398773006, + "grad_norm": 24.610651917154996, + "learning_rate": 1.1659090909090909e-06, + "logits/chosen": -0.040283203125, + "logits/rejected": -0.1552734375, + "logps/chosen": -396.0, + "logps/rejected": -350.0, + "loss": 0.3066, + "rewards/accuracies": 0.859375, + "rewards/chosen": -0.21875, + "rewards/margins": 6.625, + "rewards/rejected": -6.84375, + "step": 564 + }, + { + "epoch": 0.6932515337423313, + "grad_norm": 30.241252760906917, + "learning_rate": 1.165289256198347e-06, + "logits/chosen": 0.0081787109375, + "logits/rejected": -0.1796875, + "logps/chosen": -404.0, + "logps/rejected": -354.0, + "loss": 0.4933, + "rewards/accuracies": 0.828125, + "rewards/chosen": -0.396484375, + "rewards/margins": 6.03125, + "rewards/rejected": -6.40625, + "step": 565 + }, + { + "epoch": 0.694478527607362, + "grad_norm": 30.063355062368128, + "learning_rate": 1.1646694214876034e-06, + "logits/chosen": 0.004852294921875, + "logits/rejected": -0.08544921875, + "logps/chosen": -420.0, + "logps/rejected": -348.0, + "loss": 0.2711, + "rewards/accuracies": 0.84375, + "rewards/chosen": 0.765625, + "rewards/margins": 7.3125, + "rewards/rejected": -6.53125, + "step": 566 + }, + { + "epoch": 0.6957055214723926, + "grad_norm": 23.27581640569879, + "learning_rate": 1.1640495867768595e-06, + "logits/chosen": 0.11865234375, + "logits/rejected": -0.0096435546875, + "logps/chosen": -322.0, + "logps/rejected": -288.0, + "loss": 0.2144, + "rewards/accuracies": 0.890625, + "rewards/chosen": -0.30859375, + "rewards/margins": 6.03125, + "rewards/rejected": -6.34375, + "step": 567 + }, + { + "epoch": 0.6969325153374233, + "grad_norm": 30.55160170642105, + "learning_rate": 1.1634297520661157e-06, + "logits/chosen": -0.0390625, + "logits/rejected": -0.0869140625, + "logps/chosen": -388.0, + "logps/rejected": -334.0, + "loss": 0.3414, + "rewards/accuracies": 0.8203125, + "rewards/chosen": 0.357421875, + "rewards/margins": 6.71875, + "rewards/rejected": -6.375, + "step": 568 + }, + { + "epoch": 0.698159509202454, + "grad_norm": 29.293298679377934, + "learning_rate": 1.162809917355372e-06, + "logits/chosen": 0.04736328125, + "logits/rejected": -0.0986328125, + "logps/chosen": -388.0, + "logps/rejected": -316.0, + "loss": 0.2723, + "rewards/accuracies": 0.84375, + "rewards/chosen": -0.0361328125, + "rewards/margins": 6.28125, + "rewards/rejected": -6.3125, + "step": 569 + }, + { + "epoch": 0.6993865030674846, + "grad_norm": 29.217886094805987, + "learning_rate": 1.1621900826446282e-06, + "logits/chosen": -0.1298828125, + "logits/rejected": -0.2490234375, + "logps/chosen": -406.0, + "logps/rejected": -348.0, + "loss": 0.2078, + "rewards/accuracies": 0.8671875, + "rewards/chosen": 0.9375, + "rewards/margins": 7.40625, + "rewards/rejected": -6.46875, + "step": 570 + }, + { + "epoch": 0.7006134969325153, + "grad_norm": 29.496729411168463, + "learning_rate": 1.1615702479338843e-06, + "logits/chosen": 0.0179443359375, + "logits/rejected": -0.10205078125, + "logps/chosen": -332.0, + "logps/rejected": -314.0, + "loss": 0.3188, + "rewards/accuracies": 0.8359375, + "rewards/chosen": 0.0294189453125, + "rewards/margins": 5.875, + "rewards/rejected": -5.84375, + "step": 571 + }, + { + "epoch": 0.701840490797546, + "grad_norm": 28.740254642158334, + "learning_rate": 1.1609504132231407e-06, + "logits/chosen": 0.03173828125, + "logits/rejected": -0.11669921875, + "logps/chosen": -388.0, + "logps/rejected": -378.0, + "loss": 0.2732, + "rewards/accuracies": 0.890625, + "rewards/chosen": 1.1484375, + "rewards/margins": 7.40625, + "rewards/rejected": -6.25, + "step": 572 + }, + { + "epoch": 0.7030674846625767, + "grad_norm": 35.19506105279461, + "learning_rate": 1.1603305785123966e-06, + "logits/chosen": 0.078125, + "logits/rejected": -0.050048828125, + "logps/chosen": -394.0, + "logps/rejected": -346.0, + "loss": 0.5289, + "rewards/accuracies": 0.8125, + "rewards/chosen": -0.150390625, + "rewards/margins": 6.46875, + "rewards/rejected": -6.59375, + "step": 573 + }, + { + "epoch": 0.7042944785276074, + "grad_norm": 30.319920980400575, + "learning_rate": 1.159710743801653e-06, + "logits/chosen": -0.06201171875, + "logits/rejected": -0.1865234375, + "logps/chosen": -328.0, + "logps/rejected": -320.0, + "loss": 0.34, + "rewards/accuracies": 0.8515625, + "rewards/chosen": 0.28125, + "rewards/margins": 6.78125, + "rewards/rejected": -6.5, + "step": 574 + }, + { + "epoch": 0.7055214723926381, + "grad_norm": 29.615184087034994, + "learning_rate": 1.159090909090909e-06, + "logits/chosen": -0.119140625, + "logits/rejected": -0.2734375, + "logps/chosen": -412.0, + "logps/rejected": -388.0, + "loss": 0.2626, + "rewards/accuracies": 0.8828125, + "rewards/chosen": 1.03125, + "rewards/margins": 8.375, + "rewards/rejected": -7.34375, + "step": 575 + }, + { + "epoch": 0.7067484662576687, + "grad_norm": 22.847983936498952, + "learning_rate": 1.1584710743801652e-06, + "logits/chosen": 0.03271484375, + "logits/rejected": -0.1787109375, + "logps/chosen": -386.0, + "logps/rejected": -344.0, + "loss": 0.1933, + "rewards/accuracies": 0.90625, + "rewards/chosen": 1.0625, + "rewards/margins": 8.5, + "rewards/rejected": -7.46875, + "step": 576 + }, + { + "epoch": 0.7079754601226994, + "grad_norm": 33.54348915562207, + "learning_rate": 1.1578512396694216e-06, + "logits/chosen": 0.08935546875, + "logits/rejected": -0.0296630859375, + "logps/chosen": -368.0, + "logps/rejected": -334.0, + "loss": 0.4071, + "rewards/accuracies": 0.875, + "rewards/chosen": 0.828125, + "rewards/margins": 8.1875, + "rewards/rejected": -7.34375, + "step": 577 + }, + { + "epoch": 0.7092024539877301, + "grad_norm": 31.17463611360663, + "learning_rate": 1.1572314049586777e-06, + "logits/chosen": -0.06591796875, + "logits/rejected": -0.208984375, + "logps/chosen": -420.0, + "logps/rejected": -354.0, + "loss": 0.3672, + "rewards/accuracies": 0.8515625, + "rewards/chosen": 0.921875, + "rewards/margins": 8.6875, + "rewards/rejected": -7.75, + "step": 578 + }, + { + "epoch": 0.7104294478527607, + "grad_norm": 37.58239380773895, + "learning_rate": 1.1566115702479339e-06, + "logits/chosen": -0.025390625, + "logits/rejected": -0.1181640625, + "logps/chosen": -440.0, + "logps/rejected": -352.0, + "loss": 0.4655, + "rewards/accuracies": 0.765625, + "rewards/chosen": -0.6953125, + "rewards/margins": 6.28125, + "rewards/rejected": -7.0, + "step": 579 + }, + { + "epoch": 0.7116564417177914, + "grad_norm": 34.56148267033386, + "learning_rate": 1.1559917355371902e-06, + "logits/chosen": 0.0673828125, + "logits/rejected": -0.10791015625, + "logps/chosen": -390.0, + "logps/rejected": -374.0, + "loss": 0.2813, + "rewards/accuracies": 0.859375, + "rewards/chosen": -0.21484375, + "rewards/margins": 8.0625, + "rewards/rejected": -8.3125, + "step": 580 + }, + { + "epoch": 0.7128834355828221, + "grad_norm": 26.418081399921064, + "learning_rate": 1.1553719008264464e-06, + "logits/chosen": -0.0498046875, + "logits/rejected": -0.244140625, + "logps/chosen": -380.0, + "logps/rejected": -338.0, + "loss": 0.2267, + "rewards/accuracies": 0.8515625, + "rewards/chosen": 0.271484375, + "rewards/margins": 8.3125, + "rewards/rejected": -8.0625, + "step": 581 + }, + { + "epoch": 0.7141104294478527, + "grad_norm": 30.496748663730113, + "learning_rate": 1.1547520661157025e-06, + "logits/chosen": 0.197265625, + "logits/rejected": -0.02587890625, + "logps/chosen": -402.0, + "logps/rejected": -350.0, + "loss": 0.3169, + "rewards/accuracies": 0.859375, + "rewards/chosen": -0.01220703125, + "rewards/margins": 6.96875, + "rewards/rejected": -7.0, + "step": 582 + }, + { + "epoch": 0.7153374233128834, + "grad_norm": 31.96559757214927, + "learning_rate": 1.1541322314049586e-06, + "logits/chosen": -0.02880859375, + "logits/rejected": -0.1748046875, + "logps/chosen": -392.0, + "logps/rejected": -364.0, + "loss": 0.3362, + "rewards/accuracies": 0.8359375, + "rewards/chosen": 0.3671875, + "rewards/margins": 7.5, + "rewards/rejected": -7.125, + "step": 583 + }, + { + "epoch": 0.7165644171779141, + "grad_norm": 36.407577582473756, + "learning_rate": 1.1535123966942148e-06, + "logits/chosen": 0.08154296875, + "logits/rejected": 0.0218505859375, + "logps/chosen": -380.0, + "logps/rejected": -392.0, + "loss": 0.431, + "rewards/accuracies": 0.8125, + "rewards/chosen": -0.58203125, + "rewards/margins": 5.8125, + "rewards/rejected": -6.375, + "step": 584 + }, + { + "epoch": 0.7177914110429447, + "grad_norm": 41.799396220975545, + "learning_rate": 1.1528925619834711e-06, + "logits/chosen": 0.1796875, + "logits/rejected": 0.0634765625, + "logps/chosen": -410.0, + "logps/rejected": -344.0, + "loss": 0.5833, + "rewards/accuracies": 0.7578125, + "rewards/chosen": 0.2255859375, + "rewards/margins": 6.4375, + "rewards/rejected": -6.21875, + "step": 585 + }, + { + "epoch": 0.7190184049079754, + "grad_norm": 32.85377702391839, + "learning_rate": 1.1522727272727273e-06, + "logits/chosen": 0.035888671875, + "logits/rejected": -0.07373046875, + "logps/chosen": -356.0, + "logps/rejected": -324.0, + "loss": 0.3732, + "rewards/accuracies": 0.8515625, + "rewards/chosen": 0.00927734375, + "rewards/margins": 6.9375, + "rewards/rejected": -6.90625, + "step": 586 + }, + { + "epoch": 0.7202453987730061, + "grad_norm": 28.585914695021586, + "learning_rate": 1.1516528925619836e-06, + "logits/chosen": 0.01116943359375, + "logits/rejected": -0.09912109375, + "logps/chosen": -342.0, + "logps/rejected": -332.0, + "loss": 0.3594, + "rewards/accuracies": 0.859375, + "rewards/chosen": 0.08935546875, + "rewards/margins": 7.0625, + "rewards/rejected": -6.96875, + "step": 587 + }, + { + "epoch": 0.7214723926380369, + "grad_norm": 24.10336561473994, + "learning_rate": 1.1510330578512398e-06, + "logits/chosen": -0.0262451171875, + "logits/rejected": -0.1826171875, + "logps/chosen": -388.0, + "logps/rejected": -342.0, + "loss": 0.2892, + "rewards/accuracies": 0.8671875, + "rewards/chosen": 1.2109375, + "rewards/margins": 8.0625, + "rewards/rejected": -6.875, + "step": 588 + }, + { + "epoch": 0.7226993865030675, + "grad_norm": 28.27312313115512, + "learning_rate": 1.150413223140496e-06, + "logits/chosen": -0.0986328125, + "logits/rejected": -0.2041015625, + "logps/chosen": -372.0, + "logps/rejected": -354.0, + "loss": 0.3598, + "rewards/accuracies": 0.8359375, + "rewards/chosen": -0.37890625, + "rewards/margins": 6.28125, + "rewards/rejected": -6.65625, + "step": 589 + }, + { + "epoch": 0.7239263803680982, + "grad_norm": 30.343626428171632, + "learning_rate": 1.149793388429752e-06, + "logits/chosen": 0.00927734375, + "logits/rejected": -0.103515625, + "logps/chosen": -374.0, + "logps/rejected": -330.0, + "loss": 0.3324, + "rewards/accuracies": 0.8359375, + "rewards/chosen": 0.0986328125, + "rewards/margins": 6.875, + "rewards/rejected": -6.78125, + "step": 590 + }, + { + "epoch": 0.7251533742331289, + "grad_norm": 36.86660943740023, + "learning_rate": 1.1491735537190082e-06, + "logits/chosen": -0.169921875, + "logits/rejected": -0.373046875, + "logps/chosen": -414.0, + "logps/rejected": -354.0, + "loss": 0.28, + "rewards/accuracies": 0.8828125, + "rewards/chosen": 0.71484375, + "rewards/margins": 8.0, + "rewards/rejected": -7.3125, + "step": 591 + }, + { + "epoch": 0.7263803680981595, + "grad_norm": 27.76314433981488, + "learning_rate": 1.1485537190082645e-06, + "logits/chosen": -0.02490234375, + "logits/rejected": -0.2138671875, + "logps/chosen": -378.0, + "logps/rejected": -328.0, + "loss": 0.2629, + "rewards/accuracies": 0.8828125, + "rewards/chosen": 0.64453125, + "rewards/margins": 8.4375, + "rewards/rejected": -7.78125, + "step": 592 + }, + { + "epoch": 0.7276073619631902, + "grad_norm": 36.82046387528082, + "learning_rate": 1.1479338842975207e-06, + "logits/chosen": 0.022705078125, + "logits/rejected": -0.2333984375, + "logps/chosen": -454.0, + "logps/rejected": -392.0, + "loss": 0.4117, + "rewards/accuracies": 0.859375, + "rewards/chosen": 0.76953125, + "rewards/margins": 7.84375, + "rewards/rejected": -7.09375, + "step": 593 + }, + { + "epoch": 0.7288343558282209, + "grad_norm": 30.96499084941041, + "learning_rate": 1.1473140495867768e-06, + "logits/chosen": -0.036865234375, + "logits/rejected": -0.197265625, + "logps/chosen": -366.0, + "logps/rejected": -334.0, + "loss": 0.3336, + "rewards/accuracies": 0.84375, + "rewards/chosen": -0.294921875, + "rewards/margins": 7.375, + "rewards/rejected": -7.65625, + "step": 594 + }, + { + "epoch": 0.7300613496932515, + "grad_norm": 33.881983550723454, + "learning_rate": 1.1466942148760332e-06, + "logits/chosen": -0.059814453125, + "logits/rejected": -0.26953125, + "logps/chosen": -454.0, + "logps/rejected": -372.0, + "loss": 0.2673, + "rewards/accuracies": 0.890625, + "rewards/chosen": 0.5234375, + "rewards/margins": 8.25, + "rewards/rejected": -7.71875, + "step": 595 + }, + { + "epoch": 0.7312883435582822, + "grad_norm": 34.4461650357885, + "learning_rate": 1.1460743801652893e-06, + "logits/chosen": -0.0286865234375, + "logits/rejected": -0.18359375, + "logps/chosen": -422.0, + "logps/rejected": -368.0, + "loss": 0.3909, + "rewards/accuracies": 0.8125, + "rewards/chosen": -0.004608154296875, + "rewards/margins": 7.25, + "rewards/rejected": -7.25, + "step": 596 + }, + { + "epoch": 0.7325153374233129, + "grad_norm": 32.96451484117692, + "learning_rate": 1.1454545454545455e-06, + "logits/chosen": 0.0830078125, + "logits/rejected": -0.09228515625, + "logps/chosen": -388.0, + "logps/rejected": -306.0, + "loss": 0.4178, + "rewards/accuracies": 0.8203125, + "rewards/chosen": 0.271484375, + "rewards/margins": 6.8125, + "rewards/rejected": -6.53125, + "step": 597 + }, + { + "epoch": 0.7337423312883435, + "grad_norm": 25.225306353625232, + "learning_rate": 1.1448347107438018e-06, + "logits/chosen": -0.0211181640625, + "logits/rejected": -0.216796875, + "logps/chosen": -446.0, + "logps/rejected": -338.0, + "loss": 0.2285, + "rewards/accuracies": 0.875, + "rewards/chosen": 1.296875, + "rewards/margins": 8.375, + "rewards/rejected": -7.09375, + "step": 598 + }, + { + "epoch": 0.7349693251533742, + "grad_norm": 30.51091893610521, + "learning_rate": 1.1442148760330577e-06, + "logits/chosen": -0.169921875, + "logits/rejected": -0.3125, + "logps/chosen": -456.0, + "logps/rejected": -386.0, + "loss": 0.3423, + "rewards/accuracies": 0.8359375, + "rewards/chosen": 0.82421875, + "rewards/margins": 8.375, + "rewards/rejected": -7.53125, + "step": 599 + }, + { + "epoch": 0.7361963190184049, + "grad_norm": 31.65374536927, + "learning_rate": 1.143595041322314e-06, + "logits/chosen": -0.1796875, + "logits/rejected": -0.25390625, + "logps/chosen": -364.0, + "logps/rejected": -334.0, + "loss": 0.437, + "rewards/accuracies": 0.84375, + "rewards/chosen": 0.3515625, + "rewards/margins": 7.5625, + "rewards/rejected": -7.1875, + "step": 600 + }, + { + "epoch": 0.7374233128834355, + "grad_norm": 29.257508369320092, + "learning_rate": 1.1429752066115702e-06, + "logits/chosen": -0.03759765625, + "logits/rejected": -0.224609375, + "logps/chosen": -348.0, + "logps/rejected": -324.0, + "loss": 0.3337, + "rewards/accuracies": 0.84375, + "rewards/chosen": -0.51953125, + "rewards/margins": 7.21875, + "rewards/rejected": -7.71875, + "step": 601 + }, + { + "epoch": 0.7386503067484662, + "grad_norm": 23.637703488224467, + "learning_rate": 1.1423553719008264e-06, + "logits/chosen": 0.1103515625, + "logits/rejected": -0.0517578125, + "logps/chosen": -406.0, + "logps/rejected": -350.0, + "loss": 0.2192, + "rewards/accuracies": 0.9296875, + "rewards/chosen": 0.64453125, + "rewards/margins": 8.4375, + "rewards/rejected": -7.78125, + "step": 602 + }, + { + "epoch": 0.7398773006134969, + "grad_norm": 24.798455930230393, + "learning_rate": 1.1417355371900827e-06, + "logits/chosen": -0.0615234375, + "logits/rejected": -0.1943359375, + "logps/chosen": -382.0, + "logps/rejected": -336.0, + "loss": 0.3279, + "rewards/accuracies": 0.8671875, + "rewards/chosen": 0.82421875, + "rewards/margins": 8.3125, + "rewards/rejected": -7.5, + "step": 603 + }, + { + "epoch": 0.7411042944785277, + "grad_norm": 28.661443673376716, + "learning_rate": 1.1411157024793389e-06, + "logits/chosen": -0.11376953125, + "logits/rejected": -0.275390625, + "logps/chosen": -420.0, + "logps/rejected": -374.0, + "loss": 0.2818, + "rewards/accuracies": 0.8828125, + "rewards/chosen": 0.267578125, + "rewards/margins": 8.125, + "rewards/rejected": -7.8125, + "step": 604 + }, + { + "epoch": 0.7423312883435583, + "grad_norm": 28.682789953669307, + "learning_rate": 1.140495867768595e-06, + "logits/chosen": -0.057373046875, + "logits/rejected": -0.24609375, + "logps/chosen": -380.0, + "logps/rejected": -348.0, + "loss": 0.2877, + "rewards/accuracies": 0.8828125, + "rewards/chosen": 0.349609375, + "rewards/margins": 8.5, + "rewards/rejected": -8.125, + "step": 605 + }, + { + "epoch": 0.743558282208589, + "grad_norm": 31.866662459492442, + "learning_rate": 1.1398760330578514e-06, + "logits/chosen": -0.138671875, + "logits/rejected": -0.263671875, + "logps/chosen": -390.0, + "logps/rejected": -348.0, + "loss": 0.3751, + "rewards/accuracies": 0.828125, + "rewards/chosen": 0.443359375, + "rewards/margins": 7.65625, + "rewards/rejected": -7.21875, + "step": 606 + }, + { + "epoch": 0.7447852760736197, + "grad_norm": 29.660350580126263, + "learning_rate": 1.1392561983471075e-06, + "logits/chosen": -0.0289306640625, + "logits/rejected": -0.25, + "logps/chosen": -404.0, + "logps/rejected": -386.0, + "loss": 0.3196, + "rewards/accuracies": 0.890625, + "rewards/chosen": -0.265625, + "rewards/margins": 8.0625, + "rewards/rejected": -8.3125, + "step": 607 + }, + { + "epoch": 0.7460122699386503, + "grad_norm": 31.21931912524627, + "learning_rate": 1.1386363636363636e-06, + "logits/chosen": 0.10498046875, + "logits/rejected": -0.1416015625, + "logps/chosen": -404.0, + "logps/rejected": -360.0, + "loss": 0.3793, + "rewards/accuracies": 0.8671875, + "rewards/chosen": -0.6953125, + "rewards/margins": 6.8125, + "rewards/rejected": -7.5, + "step": 608 + }, + { + "epoch": 0.747239263803681, + "grad_norm": 35.278030303126314, + "learning_rate": 1.1380165289256198e-06, + "logits/chosen": -0.02978515625, + "logits/rejected": -0.1953125, + "logps/chosen": -412.0, + "logps/rejected": -342.0, + "loss": 0.4025, + "rewards/accuracies": 0.8515625, + "rewards/chosen": -0.255859375, + "rewards/margins": 7.9375, + "rewards/rejected": -8.1875, + "step": 609 + }, + { + "epoch": 0.7484662576687117, + "grad_norm": 29.748859331814412, + "learning_rate": 1.137396694214876e-06, + "logits/chosen": -0.10302734375, + "logits/rejected": -0.3203125, + "logps/chosen": -452.0, + "logps/rejected": -396.0, + "loss": 0.2827, + "rewards/accuracies": 0.859375, + "rewards/chosen": 0.2216796875, + "rewards/margins": 8.625, + "rewards/rejected": -8.375, + "step": 610 + }, + { + "epoch": 0.7496932515337423, + "grad_norm": 42.36265371910942, + "learning_rate": 1.1367768595041323e-06, + "logits/chosen": -0.050537109375, + "logits/rejected": -0.2216796875, + "logps/chosen": -442.0, + "logps/rejected": -384.0, + "loss": 0.493, + "rewards/accuracies": 0.828125, + "rewards/chosen": -1.234375, + "rewards/margins": 6.84375, + "rewards/rejected": -8.0625, + "step": 611 + }, + { + "epoch": 0.750920245398773, + "grad_norm": 36.03030744089288, + "learning_rate": 1.1361570247933884e-06, + "logits/chosen": -0.042236328125, + "logits/rejected": -0.1796875, + "logps/chosen": -402.0, + "logps/rejected": -376.0, + "loss": 0.4916, + "rewards/accuracies": 0.8203125, + "rewards/chosen": -0.462890625, + "rewards/margins": 6.78125, + "rewards/rejected": -7.25, + "step": 612 + }, + { + "epoch": 0.7521472392638037, + "grad_norm": 35.23308877108843, + "learning_rate": 1.1355371900826446e-06, + "logits/chosen": -0.04052734375, + "logits/rejected": -0.08447265625, + "logps/chosen": -388.0, + "logps/rejected": -348.0, + "loss": 0.4598, + "rewards/accuracies": 0.8359375, + "rewards/chosen": -1.0078125, + "rewards/margins": 6.84375, + "rewards/rejected": -7.875, + "step": 613 + }, + { + "epoch": 0.7533742331288343, + "grad_norm": 35.827377930375896, + "learning_rate": 1.134917355371901e-06, + "logits/chosen": -0.095703125, + "logits/rejected": -0.2216796875, + "logps/chosen": -430.0, + "logps/rejected": -364.0, + "loss": 0.3803, + "rewards/accuracies": 0.828125, + "rewards/chosen": -0.333984375, + "rewards/margins": 7.375, + "rewards/rejected": -7.71875, + "step": 614 + }, + { + "epoch": 0.754601226993865, + "grad_norm": 31.70255072714647, + "learning_rate": 1.134297520661157e-06, + "logits/chosen": -0.07373046875, + "logits/rejected": -0.166015625, + "logps/chosen": -376.0, + "logps/rejected": -364.0, + "loss": 0.3435, + "rewards/accuracies": 0.84375, + "rewards/chosen": 0.037841796875, + "rewards/margins": 7.84375, + "rewards/rejected": -7.8125, + "step": 615 + }, + { + "epoch": 0.7558282208588957, + "grad_norm": 30.125086244839487, + "learning_rate": 1.1336776859504134e-06, + "logits/chosen": -0.004241943359375, + "logits/rejected": -0.15625, + "logps/chosen": -332.0, + "logps/rejected": -310.0, + "loss": 0.3151, + "rewards/accuracies": 0.8359375, + "rewards/chosen": -0.6328125, + "rewards/margins": 6.53125, + "rewards/rejected": -7.15625, + "step": 616 + }, + { + "epoch": 0.7570552147239263, + "grad_norm": 32.7474160937551, + "learning_rate": 1.1330578512396693e-06, + "logits/chosen": 0.030517578125, + "logits/rejected": -0.0390625, + "logps/chosen": -378.0, + "logps/rejected": -336.0, + "loss": 0.3705, + "rewards/accuracies": 0.828125, + "rewards/chosen": -0.419921875, + "rewards/margins": 6.75, + "rewards/rejected": -7.15625, + "step": 617 + }, + { + "epoch": 0.758282208588957, + "grad_norm": 45.056199177526004, + "learning_rate": 1.1324380165289255e-06, + "logits/chosen": 0.109375, + "logits/rejected": -0.056396484375, + "logps/chosen": -426.0, + "logps/rejected": -332.0, + "loss": 0.5324, + "rewards/accuracies": 0.78125, + "rewards/chosen": -0.427734375, + "rewards/margins": 6.1875, + "rewards/rejected": -6.625, + "step": 618 + }, + { + "epoch": 0.7595092024539877, + "grad_norm": 29.95096958444537, + "learning_rate": 1.1318181818181818e-06, + "logits/chosen": 0.08056640625, + "logits/rejected": -0.12060546875, + "logps/chosen": -432.0, + "logps/rejected": -350.0, + "loss": 0.2823, + "rewards/accuracies": 0.8828125, + "rewards/chosen": 0.36328125, + "rewards/margins": 7.71875, + "rewards/rejected": -7.34375, + "step": 619 + }, + { + "epoch": 0.7607361963190185, + "grad_norm": 37.035466575479155, + "learning_rate": 1.131198347107438e-06, + "logits/chosen": 0.08349609375, + "logits/rejected": -0.11181640625, + "logps/chosen": -410.0, + "logps/rejected": -346.0, + "loss": 0.3202, + "rewards/accuracies": 0.828125, + "rewards/chosen": -0.03271484375, + "rewards/margins": 6.65625, + "rewards/rejected": -6.71875, + "step": 620 + }, + { + "epoch": 0.7619631901840491, + "grad_norm": 30.546031931386054, + "learning_rate": 1.1305785123966943e-06, + "logits/chosen": -0.052490234375, + "logits/rejected": -0.2060546875, + "logps/chosen": -482.0, + "logps/rejected": -380.0, + "loss": 0.3488, + "rewards/accuracies": 0.8828125, + "rewards/chosen": 0.7109375, + "rewards/margins": 8.125, + "rewards/rejected": -7.40625, + "step": 621 + }, + { + "epoch": 0.7631901840490798, + "grad_norm": 30.102626290429743, + "learning_rate": 1.1299586776859505e-06, + "logits/chosen": 0.07666015625, + "logits/rejected": -0.0791015625, + "logps/chosen": -382.0, + "logps/rejected": -354.0, + "loss": 0.3683, + "rewards/accuracies": 0.8671875, + "rewards/chosen": 0.2060546875, + "rewards/margins": 7.375, + "rewards/rejected": -7.15625, + "step": 622 + }, + { + "epoch": 0.7644171779141105, + "grad_norm": 31.272882227536257, + "learning_rate": 1.1293388429752066e-06, + "logits/chosen": 0.041259765625, + "logits/rejected": -0.1318359375, + "logps/chosen": -456.0, + "logps/rejected": -390.0, + "loss": 0.338, + "rewards/accuracies": 0.8125, + "rewards/chosen": 1.0078125, + "rewards/margins": 7.78125, + "rewards/rejected": -6.78125, + "step": 623 + }, + { + "epoch": 0.7656441717791411, + "grad_norm": 29.25638333606671, + "learning_rate": 1.128719008264463e-06, + "logits/chosen": -0.036865234375, + "logits/rejected": -0.208984375, + "logps/chosen": -382.0, + "logps/rejected": -334.0, + "loss": 0.302, + "rewards/accuracies": 0.890625, + "rewards/chosen": 0.8671875, + "rewards/margins": 8.5, + "rewards/rejected": -7.59375, + "step": 624 + }, + { + "epoch": 0.7668711656441718, + "grad_norm": 27.76458750020888, + "learning_rate": 1.1280991735537191e-06, + "logits/chosen": 0.01214599609375, + "logits/rejected": -0.123046875, + "logps/chosen": -408.0, + "logps/rejected": -334.0, + "loss": 0.2518, + "rewards/accuracies": 0.859375, + "rewards/chosen": 0.69140625, + "rewards/margins": 7.15625, + "rewards/rejected": -6.46875, + "step": 625 + }, + { + "epoch": 0.7680981595092025, + "grad_norm": 34.03667353357896, + "learning_rate": 1.1274793388429753e-06, + "logits/chosen": -0.119140625, + "logits/rejected": -0.2451171875, + "logps/chosen": -416.0, + "logps/rejected": -358.0, + "loss": 0.4042, + "rewards/accuracies": 0.859375, + "rewards/chosen": 1.140625, + "rewards/margins": 7.03125, + "rewards/rejected": -5.875, + "step": 626 + }, + { + "epoch": 0.7693251533742331, + "grad_norm": 30.346756983376576, + "learning_rate": 1.1268595041322314e-06, + "logits/chosen": 0.095703125, + "logits/rejected": 0.0223388671875, + "logps/chosen": -314.0, + "logps/rejected": -332.0, + "loss": 0.3854, + "rewards/accuracies": 0.8046875, + "rewards/chosen": -0.65625, + "rewards/margins": 5.28125, + "rewards/rejected": -5.9375, + "step": 627 + }, + { + "epoch": 0.7705521472392638, + "grad_norm": 31.196230115400443, + "learning_rate": 1.1262396694214875e-06, + "logits/chosen": 0.0732421875, + "logits/rejected": -0.06884765625, + "logps/chosen": -380.0, + "logps/rejected": -304.0, + "loss": 0.2959, + "rewards/accuracies": 0.8515625, + "rewards/chosen": 0.70703125, + "rewards/margins": 6.40625, + "rewards/rejected": -5.71875, + "step": 628 + }, + { + "epoch": 0.7717791411042945, + "grad_norm": 27.109929558619076, + "learning_rate": 1.1256198347107439e-06, + "logits/chosen": 0.0262451171875, + "logits/rejected": -0.09033203125, + "logps/chosen": -352.0, + "logps/rejected": -292.0, + "loss": 0.2072, + "rewards/accuracies": 0.90625, + "rewards/chosen": 0.353515625, + "rewards/margins": 6.5, + "rewards/rejected": -6.15625, + "step": 629 + }, + { + "epoch": 0.7730061349693251, + "grad_norm": 27.79699333558513, + "learning_rate": 1.125e-06, + "logits/chosen": 0.0947265625, + "logits/rejected": -0.1318359375, + "logps/chosen": -374.0, + "logps/rejected": -372.0, + "loss": 0.3309, + "rewards/accuracies": 0.8671875, + "rewards/chosen": 1.21875, + "rewards/margins": 7.09375, + "rewards/rejected": -5.875, + "step": 630 + }, + { + "epoch": 0.7742331288343558, + "grad_norm": 32.55842929304895, + "learning_rate": 1.1243801652892562e-06, + "logits/chosen": -0.11474609375, + "logits/rejected": -0.2001953125, + "logps/chosen": -394.0, + "logps/rejected": -326.0, + "loss": 0.4115, + "rewards/accuracies": 0.8046875, + "rewards/chosen": 1.03125, + "rewards/margins": 6.71875, + "rewards/rejected": -5.6875, + "step": 631 + }, + { + "epoch": 0.7754601226993865, + "grad_norm": 30.882292248466843, + "learning_rate": 1.1237603305785125e-06, + "logits/chosen": -0.0111083984375, + "logits/rejected": -0.134765625, + "logps/chosen": -406.0, + "logps/rejected": -344.0, + "loss": 0.3282, + "rewards/accuracies": 0.828125, + "rewards/chosen": 1.125, + "rewards/margins": 6.21875, + "rewards/rejected": -5.09375, + "step": 632 + }, + { + "epoch": 0.7766871165644171, + "grad_norm": 32.30019011289937, + "learning_rate": 1.1231404958677687e-06, + "logits/chosen": 0.072265625, + "logits/rejected": -0.03857421875, + "logps/chosen": -342.0, + "logps/rejected": -338.0, + "loss": 0.3515, + "rewards/accuracies": 0.8046875, + "rewards/chosen": 0.39453125, + "rewards/margins": 5.71875, + "rewards/rejected": -5.34375, + "step": 633 + }, + { + "epoch": 0.7779141104294478, + "grad_norm": 26.84260857948927, + "learning_rate": 1.1225206611570248e-06, + "logits/chosen": -0.052490234375, + "logits/rejected": -0.248046875, + "logps/chosen": -416.0, + "logps/rejected": -328.0, + "loss": 0.2366, + "rewards/accuracies": 0.9140625, + "rewards/chosen": 2.546875, + "rewards/margins": 8.3125, + "rewards/rejected": -5.78125, + "step": 634 + }, + { + "epoch": 0.7791411042944786, + "grad_norm": 23.744218806065337, + "learning_rate": 1.121900826446281e-06, + "logits/chosen": -0.0380859375, + "logits/rejected": -0.09912109375, + "logps/chosen": -362.0, + "logps/rejected": -302.0, + "loss": 0.2194, + "rewards/accuracies": 0.90625, + "rewards/chosen": 0.6484375, + "rewards/margins": 6.1875, + "rewards/rejected": -5.5625, + "step": 635 + }, + { + "epoch": 0.7803680981595092, + "grad_norm": 34.66381046216169, + "learning_rate": 1.121280991735537e-06, + "logits/chosen": -0.06201171875, + "logits/rejected": -0.166015625, + "logps/chosen": -414.0, + "logps/rejected": -368.0, + "loss": 0.4205, + "rewards/accuracies": 0.84375, + "rewards/chosen": 1.46875, + "rewards/margins": 6.65625, + "rewards/rejected": -5.1875, + "step": 636 + }, + { + "epoch": 0.7815950920245399, + "grad_norm": 41.24072249858344, + "learning_rate": 1.1206611570247934e-06, + "logits/chosen": 0.11962890625, + "logits/rejected": -0.0615234375, + "logps/chosen": -382.0, + "logps/rejected": -306.0, + "loss": 0.4805, + "rewards/accuracies": 0.8125, + "rewards/chosen": 0.439453125, + "rewards/margins": 6.125, + "rewards/rejected": -5.6875, + "step": 637 + }, + { + "epoch": 0.7828220858895706, + "grad_norm": 31.604533876936294, + "learning_rate": 1.1200413223140496e-06, + "logits/chosen": -0.044677734375, + "logits/rejected": -0.2451171875, + "logps/chosen": -406.0, + "logps/rejected": -340.0, + "loss": 0.2858, + "rewards/accuracies": 0.859375, + "rewards/chosen": 1.3046875, + "rewards/margins": 6.9375, + "rewards/rejected": -5.625, + "step": 638 + }, + { + "epoch": 0.7840490797546013, + "grad_norm": 34.862589865848754, + "learning_rate": 1.1194214876033057e-06, + "logits/chosen": 0.0869140625, + "logits/rejected": -0.080078125, + "logps/chosen": -408.0, + "logps/rejected": -342.0, + "loss": 0.4829, + "rewards/accuracies": 0.828125, + "rewards/chosen": 1.046875, + "rewards/margins": 6.34375, + "rewards/rejected": -5.28125, + "step": 639 + }, + { + "epoch": 0.7852760736196319, + "grad_norm": 29.526720756728196, + "learning_rate": 1.118801652892562e-06, + "logits/chosen": -0.08837890625, + "logits/rejected": -0.2119140625, + "logps/chosen": -446.0, + "logps/rejected": -394.0, + "loss": 0.2982, + "rewards/accuracies": 0.890625, + "rewards/chosen": 1.6328125, + "rewards/margins": 7.5, + "rewards/rejected": -5.875, + "step": 640 + }, + { + "epoch": 0.7865030674846626, + "grad_norm": 38.16790195191018, + "learning_rate": 1.1181818181818182e-06, + "logits/chosen": -0.1259765625, + "logits/rejected": -0.251953125, + "logps/chosen": -398.0, + "logps/rejected": -360.0, + "loss": 0.5882, + "rewards/accuracies": 0.8046875, + "rewards/chosen": 0.90625, + "rewards/margins": 6.15625, + "rewards/rejected": -5.25, + "step": 641 + }, + { + "epoch": 0.7877300613496933, + "grad_norm": 25.77138922359556, + "learning_rate": 1.1175619834710746e-06, + "logits/chosen": -0.01422119140625, + "logits/rejected": -0.2333984375, + "logps/chosen": -352.0, + "logps/rejected": -304.0, + "loss": 0.2926, + "rewards/accuracies": 0.875, + "rewards/chosen": 0.9140625, + "rewards/margins": 6.65625, + "rewards/rejected": -5.75, + "step": 642 + }, + { + "epoch": 0.7889570552147239, + "grad_norm": 35.220692662985385, + "learning_rate": 1.1169421487603307e-06, + "logits/chosen": -0.1748046875, + "logits/rejected": -0.333984375, + "logps/chosen": -406.0, + "logps/rejected": -386.0, + "loss": 0.327, + "rewards/accuracies": 0.8671875, + "rewards/chosen": 0.92578125, + "rewards/margins": 6.84375, + "rewards/rejected": -5.9375, + "step": 643 + }, + { + "epoch": 0.7901840490797546, + "grad_norm": 30.676171080381682, + "learning_rate": 1.1163223140495866e-06, + "logits/chosen": -0.07666015625, + "logits/rejected": -0.21875, + "logps/chosen": -392.0, + "logps/rejected": -338.0, + "loss": 0.3619, + "rewards/accuracies": 0.8515625, + "rewards/chosen": 0.75390625, + "rewards/margins": 7.25, + "rewards/rejected": -6.5, + "step": 644 + }, + { + "epoch": 0.7914110429447853, + "grad_norm": 29.15888390878358, + "learning_rate": 1.115702479338843e-06, + "logits/chosen": -0.03955078125, + "logits/rejected": -0.234375, + "logps/chosen": -336.0, + "logps/rejected": -346.0, + "loss": 0.2816, + "rewards/accuracies": 0.859375, + "rewards/chosen": 0.98046875, + "rewards/margins": 7.53125, + "rewards/rejected": -6.5625, + "step": 645 + }, + { + "epoch": 0.7926380368098159, + "grad_norm": 28.593454046054227, + "learning_rate": 1.1150826446280991e-06, + "logits/chosen": 0.047119140625, + "logits/rejected": -0.134765625, + "logps/chosen": -386.0, + "logps/rejected": -340.0, + "loss": 0.2765, + "rewards/accuracies": 0.8671875, + "rewards/chosen": 0.68359375, + "rewards/margins": 7.28125, + "rewards/rejected": -6.59375, + "step": 646 + }, + { + "epoch": 0.7938650306748466, + "grad_norm": 32.656529890149606, + "learning_rate": 1.1144628099173555e-06, + "logits/chosen": -0.033935546875, + "logits/rejected": -0.1708984375, + "logps/chosen": -386.0, + "logps/rejected": -334.0, + "loss": 0.3573, + "rewards/accuracies": 0.84375, + "rewards/chosen": 0.9296875, + "rewards/margins": 7.78125, + "rewards/rejected": -6.875, + "step": 647 + }, + { + "epoch": 0.7950920245398773, + "grad_norm": 36.18303049965862, + "learning_rate": 1.1138429752066116e-06, + "logits/chosen": -0.0908203125, + "logits/rejected": -0.08837890625, + "logps/chosen": -334.0, + "logps/rejected": -336.0, + "loss": 0.438, + "rewards/accuracies": 0.8125, + "rewards/chosen": -0.0986328125, + "rewards/margins": 6.375, + "rewards/rejected": -6.46875, + "step": 648 + }, + { + "epoch": 0.7963190184049079, + "grad_norm": 33.937125436202074, + "learning_rate": 1.1132231404958678e-06, + "logits/chosen": 0.061279296875, + "logits/rejected": -0.027587890625, + "logps/chosen": -372.0, + "logps/rejected": -342.0, + "loss": 0.5003, + "rewards/accuracies": 0.8125, + "rewards/chosen": 0.0732421875, + "rewards/margins": 6.0, + "rewards/rejected": -5.9375, + "step": 649 + }, + { + "epoch": 0.7975460122699386, + "grad_norm": 36.01649609906715, + "learning_rate": 1.1126033057851241e-06, + "logits/chosen": 0.0264892578125, + "logits/rejected": -0.1298828125, + "logps/chosen": -432.0, + "logps/rejected": -364.0, + "loss": 0.4749, + "rewards/accuracies": 0.8203125, + "rewards/chosen": 0.71875, + "rewards/margins": 6.5, + "rewards/rejected": -5.78125, + "step": 650 + }, + { + "epoch": 0.7987730061349694, + "grad_norm": 37.059100709512336, + "learning_rate": 1.1119834710743803e-06, + "logits/chosen": -0.12109375, + "logits/rejected": -0.1845703125, + "logps/chosen": -376.0, + "logps/rejected": -362.0, + "loss": 0.4208, + "rewards/accuracies": 0.8359375, + "rewards/chosen": 0.9609375, + "rewards/margins": 8.0625, + "rewards/rejected": -7.125, + "step": 651 + }, + { + "epoch": 0.8, + "grad_norm": 40.43433181753938, + "learning_rate": 1.1113636363636364e-06, + "logits/chosen": -0.031494140625, + "logits/rejected": -0.1455078125, + "logps/chosen": -376.0, + "logps/rejected": -360.0, + "loss": 0.4169, + "rewards/accuracies": 0.8515625, + "rewards/chosen": -0.5078125, + "rewards/margins": 6.625, + "rewards/rejected": -7.125, + "step": 652 + }, + { + "epoch": 0.8012269938650307, + "grad_norm": 32.30744337627645, + "learning_rate": 1.1107438016528925e-06, + "logits/chosen": 0.032958984375, + "logits/rejected": -0.142578125, + "logps/chosen": -358.0, + "logps/rejected": -346.0, + "loss": 0.2863, + "rewards/accuracies": 0.8359375, + "rewards/chosen": 0.0498046875, + "rewards/margins": 7.625, + "rewards/rejected": -7.59375, + "step": 653 + }, + { + "epoch": 0.8024539877300614, + "grad_norm": 23.943395122229603, + "learning_rate": 1.1101239669421487e-06, + "logits/chosen": 0.050537109375, + "logits/rejected": -0.15234375, + "logps/chosen": -400.0, + "logps/rejected": -336.0, + "loss": 0.226, + "rewards/accuracies": 0.890625, + "rewards/chosen": 0.333984375, + "rewards/margins": 7.96875, + "rewards/rejected": -7.625, + "step": 654 + }, + { + "epoch": 0.803680981595092, + "grad_norm": 31.423267487900187, + "learning_rate": 1.109504132231405e-06, + "logits/chosen": -0.007659912109375, + "logits/rejected": -0.1787109375, + "logps/chosen": -396.0, + "logps/rejected": -366.0, + "loss": 0.3332, + "rewards/accuracies": 0.8359375, + "rewards/chosen": -0.1943359375, + "rewards/margins": 7.125, + "rewards/rejected": -7.3125, + "step": 655 + }, + { + "epoch": 0.8049079754601227, + "grad_norm": 35.483670047647635, + "learning_rate": 1.1088842975206612e-06, + "logits/chosen": 0.055419921875, + "logits/rejected": -0.08740234375, + "logps/chosen": -380.0, + "logps/rejected": -344.0, + "loss": 0.4163, + "rewards/accuracies": 0.828125, + "rewards/chosen": -0.419921875, + "rewards/margins": 7.09375, + "rewards/rejected": -7.5, + "step": 656 + }, + { + "epoch": 0.8061349693251534, + "grad_norm": 27.21341640433181, + "learning_rate": 1.1082644628099173e-06, + "logits/chosen": -0.10107421875, + "logits/rejected": -0.2431640625, + "logps/chosen": -372.0, + "logps/rejected": -354.0, + "loss": 0.3384, + "rewards/accuracies": 0.8359375, + "rewards/chosen": -0.388671875, + "rewards/margins": 7.1875, + "rewards/rejected": -7.5625, + "step": 657 + }, + { + "epoch": 0.807361963190184, + "grad_norm": 31.34176371809714, + "learning_rate": 1.1076446280991737e-06, + "logits/chosen": -0.08349609375, + "logits/rejected": -0.212890625, + "logps/chosen": -402.0, + "logps/rejected": -366.0, + "loss": 0.3063, + "rewards/accuracies": 0.875, + "rewards/chosen": -0.412109375, + "rewards/margins": 7.21875, + "rewards/rejected": -7.625, + "step": 658 + }, + { + "epoch": 0.8085889570552147, + "grad_norm": 32.62830203002507, + "learning_rate": 1.1070247933884298e-06, + "logits/chosen": -0.0291748046875, + "logits/rejected": -0.16796875, + "logps/chosen": -374.0, + "logps/rejected": -348.0, + "loss": 0.3798, + "rewards/accuracies": 0.84375, + "rewards/chosen": -1.1171875, + "rewards/margins": 6.59375, + "rewards/rejected": -7.71875, + "step": 659 + }, + { + "epoch": 0.8098159509202454, + "grad_norm": 29.335771689820604, + "learning_rate": 1.106404958677686e-06, + "logits/chosen": -0.0380859375, + "logits/rejected": -0.220703125, + "logps/chosen": -412.0, + "logps/rejected": -374.0, + "loss": 0.3215, + "rewards/accuracies": 0.8359375, + "rewards/chosen": 0.053466796875, + "rewards/margins": 7.8125, + "rewards/rejected": -7.75, + "step": 660 + }, + { + "epoch": 0.811042944785276, + "grad_norm": 39.19555957381655, + "learning_rate": 1.1057851239669423e-06, + "logits/chosen": 0.054443359375, + "logits/rejected": -0.052734375, + "logps/chosen": -374.0, + "logps/rejected": -318.0, + "loss": 0.5135, + "rewards/accuracies": 0.8046875, + "rewards/chosen": -1.09375, + "rewards/margins": 6.46875, + "rewards/rejected": -7.5625, + "step": 661 + }, + { + "epoch": 0.8122699386503067, + "grad_norm": 25.59951308246572, + "learning_rate": 1.1051652892561982e-06, + "logits/chosen": 0.05224609375, + "logits/rejected": -0.06884765625, + "logps/chosen": -372.0, + "logps/rejected": -342.0, + "loss": 0.218, + "rewards/accuracies": 0.890625, + "rewards/chosen": 0.3671875, + "rewards/margins": 8.1875, + "rewards/rejected": -7.8125, + "step": 662 + }, + { + "epoch": 0.8134969325153374, + "grad_norm": 29.103568092770104, + "learning_rate": 1.1045454545454546e-06, + "logits/chosen": 0.1142578125, + "logits/rejected": -0.080078125, + "logps/chosen": -362.0, + "logps/rejected": -340.0, + "loss": 0.3666, + "rewards/accuracies": 0.8125, + "rewards/chosen": 0.54296875, + "rewards/margins": 7.375, + "rewards/rejected": -6.8125, + "step": 663 + }, + { + "epoch": 0.8147239263803681, + "grad_norm": 24.88151445188299, + "learning_rate": 1.1039256198347107e-06, + "logits/chosen": -0.2255859375, + "logits/rejected": -0.35546875, + "logps/chosen": -388.0, + "logps/rejected": -376.0, + "loss": 0.213, + "rewards/accuracies": 0.9140625, + "rewards/chosen": 0.7578125, + "rewards/margins": 9.0625, + "rewards/rejected": -8.3125, + "step": 664 + }, + { + "epoch": 0.8159509202453987, + "grad_norm": 34.48345990804285, + "learning_rate": 1.1033057851239669e-06, + "logits/chosen": 0.1259765625, + "logits/rejected": -0.0419921875, + "logps/chosen": -342.0, + "logps/rejected": -308.0, + "loss": 0.4057, + "rewards/accuracies": 0.8125, + "rewards/chosen": -0.2470703125, + "rewards/margins": 7.125, + "rewards/rejected": -7.375, + "step": 665 + }, + { + "epoch": 0.8171779141104294, + "grad_norm": 35.79692152845337, + "learning_rate": 1.1026859504132232e-06, + "logits/chosen": 0.0791015625, + "logits/rejected": -0.0927734375, + "logps/chosen": -422.0, + "logps/rejected": -362.0, + "loss": 0.4703, + "rewards/accuracies": 0.8125, + "rewards/chosen": 0.68359375, + "rewards/margins": 7.28125, + "rewards/rejected": -6.59375, + "step": 666 + }, + { + "epoch": 0.8184049079754602, + "grad_norm": 29.967347606106397, + "learning_rate": 1.1020661157024794e-06, + "logits/chosen": -0.010986328125, + "logits/rejected": -0.2109375, + "logps/chosen": -376.0, + "logps/rejected": -344.0, + "loss": 0.2918, + "rewards/accuracies": 0.890625, + "rewards/chosen": 0.7421875, + "rewards/margins": 7.53125, + "rewards/rejected": -6.78125, + "step": 667 + }, + { + "epoch": 0.8196319018404908, + "grad_norm": 30.862796893024488, + "learning_rate": 1.1014462809917355e-06, + "logits/chosen": 0.0123291015625, + "logits/rejected": -0.119140625, + "logps/chosen": -392.0, + "logps/rejected": -360.0, + "loss": 0.3721, + "rewards/accuracies": 0.8359375, + "rewards/chosen": 0.65625, + "rewards/margins": 7.09375, + "rewards/rejected": -6.4375, + "step": 668 + }, + { + "epoch": 0.8208588957055215, + "grad_norm": 35.17919160040694, + "learning_rate": 1.1008264462809919e-06, + "logits/chosen": -0.052490234375, + "logits/rejected": -0.2060546875, + "logps/chosen": -340.0, + "logps/rejected": -338.0, + "loss": 0.3167, + "rewards/accuracies": 0.875, + "rewards/chosen": 0.6953125, + "rewards/margins": 7.09375, + "rewards/rejected": -6.40625, + "step": 669 + }, + { + "epoch": 0.8220858895705522, + "grad_norm": 25.88170844832394, + "learning_rate": 1.100206611570248e-06, + "logits/chosen": 0.17578125, + "logits/rejected": 0.0250244140625, + "logps/chosen": -324.0, + "logps/rejected": -302.0, + "loss": 0.2514, + "rewards/accuracies": 0.8984375, + "rewards/chosen": 1.0390625, + "rewards/margins": 6.8125, + "rewards/rejected": -5.78125, + "step": 670 + }, + { + "epoch": 0.8233128834355828, + "grad_norm": 30.762328731871737, + "learning_rate": 1.0995867768595042e-06, + "logits/chosen": 0.0390625, + "logits/rejected": -0.12109375, + "logps/chosen": -372.0, + "logps/rejected": -332.0, + "loss": 0.2995, + "rewards/accuracies": 0.890625, + "rewards/chosen": 1.7421875, + "rewards/margins": 7.5, + "rewards/rejected": -5.78125, + "step": 671 + }, + { + "epoch": 0.8245398773006135, + "grad_norm": 30.48600462690235, + "learning_rate": 1.0989669421487603e-06, + "logits/chosen": 0.054931640625, + "logits/rejected": -0.2197265625, + "logps/chosen": -446.0, + "logps/rejected": -332.0, + "loss": 0.3199, + "rewards/accuracies": 0.8515625, + "rewards/chosen": 2.015625, + "rewards/margins": 7.71875, + "rewards/rejected": -5.6875, + "step": 672 + }, + { + "epoch": 0.8257668711656442, + "grad_norm": 32.03977574623459, + "learning_rate": 1.0983471074380164e-06, + "logits/chosen": 0.1416015625, + "logits/rejected": -0.0986328125, + "logps/chosen": -390.0, + "logps/rejected": -328.0, + "loss": 0.3096, + "rewards/accuracies": 0.859375, + "rewards/chosen": 1.109375, + "rewards/margins": 6.21875, + "rewards/rejected": -5.125, + "step": 673 + }, + { + "epoch": 0.8269938650306748, + "grad_norm": 32.93059543816188, + "learning_rate": 1.0977272727272728e-06, + "logits/chosen": 0.162109375, + "logits/rejected": 0.00286865234375, + "logps/chosen": -358.0, + "logps/rejected": -330.0, + "loss": 0.3985, + "rewards/accuracies": 0.8359375, + "rewards/chosen": 1.0546875, + "rewards/margins": 6.59375, + "rewards/rejected": -5.53125, + "step": 674 + }, + { + "epoch": 0.8282208588957055, + "grad_norm": 34.971531039669564, + "learning_rate": 1.097107438016529e-06, + "logits/chosen": 0.08203125, + "logits/rejected": -0.130859375, + "logps/chosen": -426.0, + "logps/rejected": -368.0, + "loss": 0.4407, + "rewards/accuracies": 0.84375, + "rewards/chosen": 1.3828125, + "rewards/margins": 7.40625, + "rewards/rejected": -6.0, + "step": 675 + }, + { + "epoch": 0.8294478527607362, + "grad_norm": 34.2224078373622, + "learning_rate": 1.0964876033057853e-06, + "logits/chosen": -0.037353515625, + "logits/rejected": -0.14453125, + "logps/chosen": -394.0, + "logps/rejected": -352.0, + "loss": 0.3813, + "rewards/accuracies": 0.859375, + "rewards/chosen": 0.4375, + "rewards/margins": 6.8125, + "rewards/rejected": -6.375, + "step": 676 + }, + { + "epoch": 0.8306748466257668, + "grad_norm": 22.049121537748125, + "learning_rate": 1.0958677685950414e-06, + "logits/chosen": 0.09423828125, + "logits/rejected": -0.08447265625, + "logps/chosen": -338.0, + "logps/rejected": -316.0, + "loss": 0.2974, + "rewards/accuracies": 0.875, + "rewards/chosen": 0.2099609375, + "rewards/margins": 7.1875, + "rewards/rejected": -6.96875, + "step": 677 + }, + { + "epoch": 0.8319018404907975, + "grad_norm": 31.042587162348006, + "learning_rate": 1.0952479338842976e-06, + "logits/chosen": 0.07421875, + "logits/rejected": -0.1376953125, + "logps/chosen": -422.0, + "logps/rejected": -362.0, + "loss": 0.4105, + "rewards/accuracies": 0.84375, + "rewards/chosen": 0.439453125, + "rewards/margins": 6.96875, + "rewards/rejected": -6.53125, + "step": 678 + }, + { + "epoch": 0.8331288343558282, + "grad_norm": 32.125565947605324, + "learning_rate": 1.0946280991735537e-06, + "logits/chosen": 0.1689453125, + "logits/rejected": 0.06201171875, + "logps/chosen": -344.0, + "logps/rejected": -328.0, + "loss": 0.3816, + "rewards/accuracies": 0.859375, + "rewards/chosen": 0.44921875, + "rewards/margins": 6.53125, + "rewards/rejected": -6.0625, + "step": 679 + }, + { + "epoch": 0.8343558282208589, + "grad_norm": 30.54691503767555, + "learning_rate": 1.0940082644628098e-06, + "logits/chosen": 0.025390625, + "logits/rejected": -0.150390625, + "logps/chosen": -454.0, + "logps/rejected": -374.0, + "loss": 0.3146, + "rewards/accuracies": 0.84375, + "rewards/chosen": 1.25, + "rewards/margins": 8.125, + "rewards/rejected": -6.875, + "step": 680 + }, + { + "epoch": 0.8355828220858895, + "grad_norm": 25.35697096492706, + "learning_rate": 1.0933884297520662e-06, + "logits/chosen": -0.0262451171875, + "logits/rejected": -0.1796875, + "logps/chosen": -348.0, + "logps/rejected": -338.0, + "loss": 0.2471, + "rewards/accuracies": 0.890625, + "rewards/chosen": 0.72265625, + "rewards/margins": 8.4375, + "rewards/rejected": -7.71875, + "step": 681 + }, + { + "epoch": 0.8368098159509203, + "grad_norm": 29.372252652247152, + "learning_rate": 1.0927685950413223e-06, + "logits/chosen": -0.000934600830078125, + "logits/rejected": -0.115234375, + "logps/chosen": -352.0, + "logps/rejected": -334.0, + "loss": 0.3998, + "rewards/accuracies": 0.84375, + "rewards/chosen": 0.5546875, + "rewards/margins": 7.4375, + "rewards/rejected": -6.90625, + "step": 682 + }, + { + "epoch": 0.838036809815951, + "grad_norm": 28.48014927993986, + "learning_rate": 1.0921487603305785e-06, + "logits/chosen": -0.0264892578125, + "logits/rejected": -0.1328125, + "logps/chosen": -374.0, + "logps/rejected": -330.0, + "loss": 0.3192, + "rewards/accuracies": 0.8359375, + "rewards/chosen": 0.67578125, + "rewards/margins": 8.0625, + "rewards/rejected": -7.375, + "step": 683 + }, + { + "epoch": 0.8392638036809816, + "grad_norm": 30.9956365019602, + "learning_rate": 1.0915289256198348e-06, + "logits/chosen": 0.2001953125, + "logits/rejected": 0.08544921875, + "logps/chosen": -322.0, + "logps/rejected": -352.0, + "loss": 0.3553, + "rewards/accuracies": 0.8515625, + "rewards/chosen": -0.6953125, + "rewards/margins": 6.3125, + "rewards/rejected": -7.03125, + "step": 684 + }, + { + "epoch": 0.8404907975460123, + "grad_norm": 31.024167516184594, + "learning_rate": 1.090909090909091e-06, + "logits/chosen": 0.040283203125, + "logits/rejected": -0.12255859375, + "logps/chosen": -404.0, + "logps/rejected": -380.0, + "loss": 0.3423, + "rewards/accuracies": 0.875, + "rewards/chosen": 0.400390625, + "rewards/margins": 8.625, + "rewards/rejected": -8.25, + "step": 685 + }, + { + "epoch": 0.841717791411043, + "grad_norm": 25.880225574187453, + "learning_rate": 1.0902892561983471e-06, + "logits/chosen": 0.031982421875, + "logits/rejected": -0.095703125, + "logps/chosen": -400.0, + "logps/rejected": -348.0, + "loss": 0.2926, + "rewards/accuracies": 0.8515625, + "rewards/chosen": 0.27734375, + "rewards/margins": 7.9375, + "rewards/rejected": -7.65625, + "step": 686 + }, + { + "epoch": 0.8429447852760736, + "grad_norm": 28.80339614616542, + "learning_rate": 1.0896694214876035e-06, + "logits/chosen": 0.03759765625, + "logits/rejected": -0.1044921875, + "logps/chosen": -380.0, + "logps/rejected": -364.0, + "loss": 0.2973, + "rewards/accuracies": 0.8359375, + "rewards/chosen": 0.484375, + "rewards/margins": 7.90625, + "rewards/rejected": -7.40625, + "step": 687 + }, + { + "epoch": 0.8441717791411043, + "grad_norm": 30.11912577456189, + "learning_rate": 1.0890495867768594e-06, + "logits/chosen": -0.000240325927734375, + "logits/rejected": -0.158203125, + "logps/chosen": -384.0, + "logps/rejected": -328.0, + "loss": 0.327, + "rewards/accuracies": 0.8203125, + "rewards/chosen": -0.5625, + "rewards/margins": 7.5, + "rewards/rejected": -8.0625, + "step": 688 + }, + { + "epoch": 0.845398773006135, + "grad_norm": 30.49529821286284, + "learning_rate": 1.0884297520661158e-06, + "logits/chosen": 0.027099609375, + "logits/rejected": -0.058837890625, + "logps/chosen": -382.0, + "logps/rejected": -348.0, + "loss": 0.344, + "rewards/accuracies": 0.8359375, + "rewards/chosen": 0.271484375, + "rewards/margins": 7.65625, + "rewards/rejected": -7.40625, + "step": 689 + }, + { + "epoch": 0.8466257668711656, + "grad_norm": 41.86773593768432, + "learning_rate": 1.0878099173553719e-06, + "logits/chosen": 0.171875, + "logits/rejected": -0.0174560546875, + "logps/chosen": -414.0, + "logps/rejected": -370.0, + "loss": 0.5772, + "rewards/accuracies": 0.7890625, + "rewards/chosen": 0.408203125, + "rewards/margins": 7.625, + "rewards/rejected": -7.21875, + "step": 690 + }, + { + "epoch": 0.8478527607361963, + "grad_norm": 27.177968327720407, + "learning_rate": 1.087190082644628e-06, + "logits/chosen": -0.162109375, + "logits/rejected": -0.2734375, + "logps/chosen": -366.0, + "logps/rejected": -368.0, + "loss": 0.3098, + "rewards/accuracies": 0.859375, + "rewards/chosen": 0.29296875, + "rewards/margins": 8.1875, + "rewards/rejected": -7.90625, + "step": 691 + }, + { + "epoch": 0.849079754601227, + "grad_norm": 38.42433716881971, + "learning_rate": 1.0865702479338844e-06, + "logits/chosen": 0.0179443359375, + "logits/rejected": -0.142578125, + "logps/chosen": -416.0, + "logps/rejected": -358.0, + "loss": 0.3292, + "rewards/accuracies": 0.859375, + "rewards/chosen": 0.8203125, + "rewards/margins": 8.375, + "rewards/rejected": -7.53125, + "step": 692 + }, + { + "epoch": 0.8503067484662576, + "grad_norm": 31.708672912333903, + "learning_rate": 1.0859504132231405e-06, + "logits/chosen": -0.00286865234375, + "logits/rejected": -0.1904296875, + "logps/chosen": -452.0, + "logps/rejected": -386.0, + "loss": 0.3787, + "rewards/accuracies": 0.84375, + "rewards/chosen": 1.0625, + "rewards/margins": 8.875, + "rewards/rejected": -7.8125, + "step": 693 + }, + { + "epoch": 0.8515337423312883, + "grad_norm": 39.65580173728978, + "learning_rate": 1.0853305785123967e-06, + "logits/chosen": 0.024658203125, + "logits/rejected": -0.162109375, + "logps/chosen": -384.0, + "logps/rejected": -352.0, + "loss": 0.5333, + "rewards/accuracies": 0.8203125, + "rewards/chosen": -0.671875, + "rewards/margins": 6.09375, + "rewards/rejected": -6.75, + "step": 694 + }, + { + "epoch": 0.852760736196319, + "grad_norm": 38.14890644660088, + "learning_rate": 1.084710743801653e-06, + "logits/chosen": 0.06103515625, + "logits/rejected": -0.0179443359375, + "logps/chosen": -338.0, + "logps/rejected": -322.0, + "loss": 0.5352, + "rewards/accuracies": 0.75, + "rewards/chosen": -0.4609375, + "rewards/margins": 5.375, + "rewards/rejected": -5.8125, + "step": 695 + }, + { + "epoch": 0.8539877300613496, + "grad_norm": 32.915799758521196, + "learning_rate": 1.0840909090909092e-06, + "logits/chosen": 0.0947265625, + "logits/rejected": -0.08154296875, + "logps/chosen": -390.0, + "logps/rejected": -336.0, + "loss": 0.3241, + "rewards/accuracies": 0.84375, + "rewards/chosen": 0.365234375, + "rewards/margins": 7.125, + "rewards/rejected": -6.78125, + "step": 696 + }, + { + "epoch": 0.8552147239263803, + "grad_norm": 35.9913417584187, + "learning_rate": 1.0834710743801653e-06, + "logits/chosen": -0.0771484375, + "logits/rejected": -0.10009765625, + "logps/chosen": -398.0, + "logps/rejected": -384.0, + "loss": 0.4041, + "rewards/accuracies": 0.84375, + "rewards/chosen": 0.9375, + "rewards/margins": 8.0, + "rewards/rejected": -7.0625, + "step": 697 + }, + { + "epoch": 0.8564417177914111, + "grad_norm": 31.064185654999054, + "learning_rate": 1.0828512396694214e-06, + "logits/chosen": 0.08984375, + "logits/rejected": -0.027587890625, + "logps/chosen": -404.0, + "logps/rejected": -366.0, + "loss": 0.3837, + "rewards/accuracies": 0.875, + "rewards/chosen": 0.392578125, + "rewards/margins": 7.5, + "rewards/rejected": -7.09375, + "step": 698 + }, + { + "epoch": 0.8576687116564418, + "grad_norm": 26.06263909227069, + "learning_rate": 1.0822314049586776e-06, + "logits/chosen": 0.0260009765625, + "logits/rejected": -0.053955078125, + "logps/chosen": -346.0, + "logps/rejected": -340.0, + "loss": 0.2587, + "rewards/accuracies": 0.8671875, + "rewards/chosen": -0.228515625, + "rewards/margins": 6.46875, + "rewards/rejected": -6.6875, + "step": 699 + }, + { + "epoch": 0.8588957055214724, + "grad_norm": 33.7346004258113, + "learning_rate": 1.081611570247934e-06, + "logits/chosen": 0.1484375, + "logits/rejected": 0.06005859375, + "logps/chosen": -380.0, + "logps/rejected": -342.0, + "loss": 0.398, + "rewards/accuracies": 0.8359375, + "rewards/chosen": -0.0072021484375, + "rewards/margins": 6.28125, + "rewards/rejected": -6.28125, + "step": 700 + }, + { + "epoch": 0.8601226993865031, + "grad_norm": 36.508608167851875, + "learning_rate": 1.08099173553719e-06, + "logits/chosen": 0.1630859375, + "logits/rejected": 0.05810546875, + "logps/chosen": -390.0, + "logps/rejected": -342.0, + "loss": 0.4454, + "rewards/accuracies": 0.78125, + "rewards/chosen": -0.267578125, + "rewards/margins": 6.5625, + "rewards/rejected": -6.8125, + "step": 701 + }, + { + "epoch": 0.8613496932515338, + "grad_norm": 29.752428717784724, + "learning_rate": 1.0803719008264462e-06, + "logits/chosen": -0.005767822265625, + "logits/rejected": -0.30078125, + "logps/chosen": -406.0, + "logps/rejected": -348.0, + "loss": 0.2347, + "rewards/accuracies": 0.859375, + "rewards/chosen": 0.91015625, + "rewards/margins": 8.0, + "rewards/rejected": -7.09375, + "step": 702 + }, + { + "epoch": 0.8625766871165644, + "grad_norm": 28.46276315635888, + "learning_rate": 1.0797520661157026e-06, + "logits/chosen": 0.1513671875, + "logits/rejected": 0.0250244140625, + "logps/chosen": -424.0, + "logps/rejected": -358.0, + "loss": 0.3098, + "rewards/accuracies": 0.859375, + "rewards/chosen": 0.1396484375, + "rewards/margins": 6.53125, + "rewards/rejected": -6.375, + "step": 703 + }, + { + "epoch": 0.8638036809815951, + "grad_norm": 28.331719307949253, + "learning_rate": 1.0791322314049587e-06, + "logits/chosen": 0.06689453125, + "logits/rejected": -0.09033203125, + "logps/chosen": -332.0, + "logps/rejected": -316.0, + "loss": 0.3639, + "rewards/accuracies": 0.859375, + "rewards/chosen": 0.07861328125, + "rewards/margins": 6.96875, + "rewards/rejected": -6.875, + "step": 704 + }, + { + "epoch": 0.8650306748466258, + "grad_norm": 33.73975373967079, + "learning_rate": 1.078512396694215e-06, + "logits/chosen": -0.0615234375, + "logits/rejected": -0.177734375, + "logps/chosen": -424.0, + "logps/rejected": -376.0, + "loss": 0.3238, + "rewards/accuracies": 0.8515625, + "rewards/chosen": 0.75, + "rewards/margins": 8.0, + "rewards/rejected": -7.25, + "step": 705 + }, + { + "epoch": 0.8662576687116564, + "grad_norm": 32.06391249106975, + "learning_rate": 1.077892561983471e-06, + "logits/chosen": 0.01544189453125, + "logits/rejected": -0.1796875, + "logps/chosen": -410.0, + "logps/rejected": -316.0, + "loss": 0.3915, + "rewards/accuracies": 0.8515625, + "rewards/chosen": 0.1748046875, + "rewards/margins": 6.84375, + "rewards/rejected": -6.6875, + "step": 706 + }, + { + "epoch": 0.8674846625766871, + "grad_norm": 32.50030520028076, + "learning_rate": 1.0772727272727271e-06, + "logits/chosen": -0.00185394287109375, + "logits/rejected": -0.1650390625, + "logps/chosen": -394.0, + "logps/rejected": -326.0, + "loss": 0.3089, + "rewards/accuracies": 0.8515625, + "rewards/chosen": -0.0947265625, + "rewards/margins": 7.28125, + "rewards/rejected": -7.375, + "step": 707 + }, + { + "epoch": 0.8687116564417178, + "grad_norm": 27.35860386441147, + "learning_rate": 1.0766528925619835e-06, + "logits/chosen": 0.0245361328125, + "logits/rejected": -0.193359375, + "logps/chosen": -380.0, + "logps/rejected": -328.0, + "loss": 0.2856, + "rewards/accuracies": 0.8359375, + "rewards/chosen": 0.65234375, + "rewards/margins": 7.28125, + "rewards/rejected": -6.625, + "step": 708 + }, + { + "epoch": 0.8699386503067484, + "grad_norm": 22.967542887627133, + "learning_rate": 1.0760330578512396e-06, + "logits/chosen": 0.11376953125, + "logits/rejected": -0.031982421875, + "logps/chosen": -444.0, + "logps/rejected": -360.0, + "loss": 0.2311, + "rewards/accuracies": 0.875, + "rewards/chosen": 0.984375, + "rewards/margins": 8.4375, + "rewards/rejected": -7.46875, + "step": 709 + }, + { + "epoch": 0.8711656441717791, + "grad_norm": 30.012370581713046, + "learning_rate": 1.075413223140496e-06, + "logits/chosen": 0.1845703125, + "logits/rejected": 0.06103515625, + "logps/chosen": -372.0, + "logps/rejected": -330.0, + "loss": 0.3445, + "rewards/accuracies": 0.8125, + "rewards/chosen": 0.1064453125, + "rewards/margins": 6.46875, + "rewards/rejected": -6.375, + "step": 710 + }, + { + "epoch": 0.8723926380368098, + "grad_norm": 26.59602404250697, + "learning_rate": 1.0747933884297521e-06, + "logits/chosen": 0.15625, + "logits/rejected": -0.039306640625, + "logps/chosen": -390.0, + "logps/rejected": -332.0, + "loss": 0.3562, + "rewards/accuracies": 0.8515625, + "rewards/chosen": 0.5703125, + "rewards/margins": 7.0, + "rewards/rejected": -6.4375, + "step": 711 + }, + { + "epoch": 0.8736196319018404, + "grad_norm": 34.082753455253695, + "learning_rate": 1.0741735537190083e-06, + "logits/chosen": 0.1484375, + "logits/rejected": 0.00482177734375, + "logps/chosen": -410.0, + "logps/rejected": -356.0, + "loss": 0.4269, + "rewards/accuracies": 0.8125, + "rewards/chosen": 1.328125, + "rewards/margins": 7.03125, + "rewards/rejected": -5.71875, + "step": 712 + }, + { + "epoch": 0.8748466257668711, + "grad_norm": 24.8779582799026, + "learning_rate": 1.0735537190082646e-06, + "logits/chosen": 0.05810546875, + "logits/rejected": -0.1484375, + "logps/chosen": -386.0, + "logps/rejected": -336.0, + "loss": 0.2908, + "rewards/accuracies": 0.875, + "rewards/chosen": 0.625, + "rewards/margins": 7.28125, + "rewards/rejected": -6.65625, + "step": 713 + }, + { + "epoch": 0.8760736196319019, + "grad_norm": 27.823114674287325, + "learning_rate": 1.0729338842975208e-06, + "logits/chosen": 0.181640625, + "logits/rejected": -0.038818359375, + "logps/chosen": -370.0, + "logps/rejected": -328.0, + "loss": 0.3057, + "rewards/accuracies": 0.890625, + "rewards/chosen": 0.87109375, + "rewards/margins": 6.84375, + "rewards/rejected": -6.0, + "step": 714 + }, + { + "epoch": 0.8773006134969326, + "grad_norm": 29.027125190845624, + "learning_rate": 1.072314049586777e-06, + "logits/chosen": -0.06494140625, + "logits/rejected": -0.1318359375, + "logps/chosen": -384.0, + "logps/rejected": -328.0, + "loss": 0.3391, + "rewards/accuracies": 0.8203125, + "rewards/chosen": 1.015625, + "rewards/margins": 7.75, + "rewards/rejected": -6.71875, + "step": 715 + }, + { + "epoch": 0.8785276073619632, + "grad_norm": 26.494530473617345, + "learning_rate": 1.071694214876033e-06, + "logits/chosen": -0.018798828125, + "logits/rejected": -0.1865234375, + "logps/chosen": -384.0, + "logps/rejected": -324.0, + "loss": 0.3073, + "rewards/accuracies": 0.8515625, + "rewards/chosen": 1.1640625, + "rewards/margins": 6.84375, + "rewards/rejected": -5.65625, + "step": 716 + }, + { + "epoch": 0.8797546012269939, + "grad_norm": 29.02718718589355, + "learning_rate": 1.0710743801652892e-06, + "logits/chosen": 0.07470703125, + "logits/rejected": -0.08056640625, + "logps/chosen": -344.0, + "logps/rejected": -284.0, + "loss": 0.3344, + "rewards/accuracies": 0.84375, + "rewards/chosen": 0.6875, + "rewards/margins": 6.90625, + "rewards/rejected": -6.21875, + "step": 717 + }, + { + "epoch": 0.8809815950920246, + "grad_norm": 23.193149611746794, + "learning_rate": 1.0704545454545455e-06, + "logits/chosen": -0.034912109375, + "logits/rejected": -0.1640625, + "logps/chosen": -402.0, + "logps/rejected": -356.0, + "loss": 0.2308, + "rewards/accuracies": 0.8828125, + "rewards/chosen": 1.5859375, + "rewards/margins": 7.84375, + "rewards/rejected": -6.25, + "step": 718 + }, + { + "epoch": 0.8822085889570552, + "grad_norm": 36.13316553444024, + "learning_rate": 1.0698347107438017e-06, + "logits/chosen": -0.07763671875, + "logits/rejected": -0.19921875, + "logps/chosen": -410.0, + "logps/rejected": -368.0, + "loss": 0.3568, + "rewards/accuracies": 0.8203125, + "rewards/chosen": 1.8046875, + "rewards/margins": 8.0, + "rewards/rejected": -6.1875, + "step": 719 + }, + { + "epoch": 0.8834355828220859, + "grad_norm": 25.470910809931194, + "learning_rate": 1.0692148760330578e-06, + "logits/chosen": -0.07421875, + "logits/rejected": -0.279296875, + "logps/chosen": -412.0, + "logps/rejected": -350.0, + "loss": 0.2706, + "rewards/accuracies": 0.8828125, + "rewards/chosen": 1.9140625, + "rewards/margins": 7.9375, + "rewards/rejected": -6.03125, + "step": 720 + }, + { + "epoch": 0.8846625766871166, + "grad_norm": 24.45362594207483, + "learning_rate": 1.0685950413223142e-06, + "logits/chosen": -0.0810546875, + "logits/rejected": -0.337890625, + "logps/chosen": -414.0, + "logps/rejected": -364.0, + "loss": 0.2805, + "rewards/accuracies": 0.921875, + "rewards/chosen": 1.5625, + "rewards/margins": 8.8125, + "rewards/rejected": -7.25, + "step": 721 + }, + { + "epoch": 0.8858895705521472, + "grad_norm": 26.653798150160988, + "learning_rate": 1.0679752066115703e-06, + "logits/chosen": -0.049072265625, + "logits/rejected": -0.1376953125, + "logps/chosen": -380.0, + "logps/rejected": -358.0, + "loss": 0.2495, + "rewards/accuracies": 0.890625, + "rewards/chosen": 1.609375, + "rewards/margins": 7.875, + "rewards/rejected": -6.28125, + "step": 722 + }, + { + "epoch": 0.8871165644171779, + "grad_norm": 29.259615066962372, + "learning_rate": 1.0673553719008265e-06, + "logits/chosen": -0.043212890625, + "logits/rejected": -0.14453125, + "logps/chosen": -376.0, + "logps/rejected": -334.0, + "loss": 0.3287, + "rewards/accuracies": 0.859375, + "rewards/chosen": 0.94921875, + "rewards/margins": 7.4375, + "rewards/rejected": -6.5, + "step": 723 + }, + { + "epoch": 0.8883435582822086, + "grad_norm": 30.45569089414315, + "learning_rate": 1.0667355371900826e-06, + "logits/chosen": -0.0308837890625, + "logits/rejected": -0.251953125, + "logps/chosen": -382.0, + "logps/rejected": -320.0, + "loss": 0.3219, + "rewards/accuracies": 0.859375, + "rewards/chosen": 1.3359375, + "rewards/margins": 7.8125, + "rewards/rejected": -6.5, + "step": 724 + }, + { + "epoch": 0.8895705521472392, + "grad_norm": 30.8228410953136, + "learning_rate": 1.0661157024793387e-06, + "logits/chosen": 0.12255859375, + "logits/rejected": -0.05419921875, + "logps/chosen": -404.0, + "logps/rejected": -340.0, + "loss": 0.3833, + "rewards/accuracies": 0.8359375, + "rewards/chosen": 0.984375, + "rewards/margins": 7.375, + "rewards/rejected": -6.375, + "step": 725 + }, + { + "epoch": 0.8907975460122699, + "grad_norm": 32.44931458712784, + "learning_rate": 1.065495867768595e-06, + "logits/chosen": -0.030029296875, + "logits/rejected": -0.154296875, + "logps/chosen": -340.0, + "logps/rejected": -302.0, + "loss": 0.2852, + "rewards/accuracies": 0.8515625, + "rewards/chosen": 0.5703125, + "rewards/margins": 7.53125, + "rewards/rejected": -6.96875, + "step": 726 + }, + { + "epoch": 0.8920245398773006, + "grad_norm": 30.00382649809932, + "learning_rate": 1.0648760330578512e-06, + "logits/chosen": -0.0013427734375, + "logits/rejected": -0.1474609375, + "logps/chosen": -374.0, + "logps/rejected": -338.0, + "loss": 0.2567, + "rewards/accuracies": 0.890625, + "rewards/chosen": 0.84765625, + "rewards/margins": 7.5625, + "rewards/rejected": -6.71875, + "step": 727 + }, + { + "epoch": 0.8932515337423312, + "grad_norm": 32.084882438300085, + "learning_rate": 1.0642561983471074e-06, + "logits/chosen": -0.0244140625, + "logits/rejected": -0.1611328125, + "logps/chosen": -406.0, + "logps/rejected": -384.0, + "loss": 0.3182, + "rewards/accuracies": 0.8515625, + "rewards/chosen": 1.234375, + "rewards/margins": 8.875, + "rewards/rejected": -7.625, + "step": 728 + }, + { + "epoch": 0.894478527607362, + "grad_norm": 24.427532348622456, + "learning_rate": 1.0636363636363637e-06, + "logits/chosen": -0.040771484375, + "logits/rejected": -0.1708984375, + "logps/chosen": -396.0, + "logps/rejected": -372.0, + "loss": 0.2099, + "rewards/accuracies": 0.8984375, + "rewards/chosen": 0.9765625, + "rewards/margins": 8.875, + "rewards/rejected": -7.90625, + "step": 729 + }, + { + "epoch": 0.8957055214723927, + "grad_norm": 24.535854490724105, + "learning_rate": 1.0630165289256199e-06, + "logits/chosen": -0.103515625, + "logits/rejected": -0.2060546875, + "logps/chosen": -404.0, + "logps/rejected": -380.0, + "loss": 0.2127, + "rewards/accuracies": 0.9140625, + "rewards/chosen": 1.515625, + "rewards/margins": 8.8125, + "rewards/rejected": -7.3125, + "step": 730 + }, + { + "epoch": 0.8969325153374234, + "grad_norm": 34.14852950381922, + "learning_rate": 1.062396694214876e-06, + "logits/chosen": 0.1171875, + "logits/rejected": 0.002349853515625, + "logps/chosen": -438.0, + "logps/rejected": -366.0, + "loss": 0.3465, + "rewards/accuracies": 0.859375, + "rewards/chosen": 0.62890625, + "rewards/margins": 8.1875, + "rewards/rejected": -7.5625, + "step": 731 + }, + { + "epoch": 0.898159509202454, + "grad_norm": 34.434393814707754, + "learning_rate": 1.0617768595041324e-06, + "logits/chosen": -0.058837890625, + "logits/rejected": -0.2099609375, + "logps/chosen": -380.0, + "logps/rejected": -372.0, + "loss": 0.4237, + "rewards/accuracies": 0.8046875, + "rewards/chosen": 0.6875, + "rewards/margins": 7.71875, + "rewards/rejected": -7.03125, + "step": 732 + }, + { + "epoch": 0.8993865030674847, + "grad_norm": 28.68332287996042, + "learning_rate": 1.0611570247933883e-06, + "logits/chosen": 0.0673828125, + "logits/rejected": -0.083984375, + "logps/chosen": -366.0, + "logps/rejected": -362.0, + "loss": 0.2815, + "rewards/accuracies": 0.859375, + "rewards/chosen": 0.1015625, + "rewards/margins": 7.65625, + "rewards/rejected": -7.5625, + "step": 733 + }, + { + "epoch": 0.9006134969325154, + "grad_norm": 32.670441266137765, + "learning_rate": 1.0605371900826447e-06, + "logits/chosen": 0.06884765625, + "logits/rejected": -0.1044921875, + "logps/chosen": -384.0, + "logps/rejected": -354.0, + "loss": 0.3319, + "rewards/accuracies": 0.859375, + "rewards/chosen": 0.408203125, + "rewards/margins": 8.3125, + "rewards/rejected": -7.90625, + "step": 734 + }, + { + "epoch": 0.901840490797546, + "grad_norm": 28.766622871975752, + "learning_rate": 1.0599173553719008e-06, + "logits/chosen": -0.037841796875, + "logits/rejected": -0.259765625, + "logps/chosen": -434.0, + "logps/rejected": -390.0, + "loss": 0.2447, + "rewards/accuracies": 0.875, + "rewards/chosen": 0.2490234375, + "rewards/margins": 8.875, + "rewards/rejected": -8.625, + "step": 735 + }, + { + "epoch": 0.9030674846625767, + "grad_norm": 31.10013526298491, + "learning_rate": 1.059297520661157e-06, + "logits/chosen": 0.1533203125, + "logits/rejected": 0.029541015625, + "logps/chosen": -362.0, + "logps/rejected": -340.0, + "loss": 0.4211, + "rewards/accuracies": 0.8359375, + "rewards/chosen": -0.53515625, + "rewards/margins": 7.03125, + "rewards/rejected": -7.5625, + "step": 736 + }, + { + "epoch": 0.9042944785276074, + "grad_norm": 38.03816635846729, + "learning_rate": 1.0586776859504133e-06, + "logits/chosen": 0.04541015625, + "logits/rejected": -0.051513671875, + "logps/chosen": -360.0, + "logps/rejected": -348.0, + "loss": 0.4641, + "rewards/accuracies": 0.8046875, + "rewards/chosen": -0.478515625, + "rewards/margins": 7.09375, + "rewards/rejected": -7.5625, + "step": 737 + }, + { + "epoch": 0.905521472392638, + "grad_norm": 32.716309617319716, + "learning_rate": 1.0580578512396694e-06, + "logits/chosen": -0.0380859375, + "logits/rejected": -0.1474609375, + "logps/chosen": -450.0, + "logps/rejected": -364.0, + "loss": 0.3437, + "rewards/accuracies": 0.8359375, + "rewards/chosen": 0.60546875, + "rewards/margins": 8.1875, + "rewards/rejected": -7.59375, + "step": 738 + }, + { + "epoch": 0.9067484662576687, + "grad_norm": 29.627927045089628, + "learning_rate": 1.0574380165289258e-06, + "logits/chosen": 0.060302734375, + "logits/rejected": -0.0654296875, + "logps/chosen": -346.0, + "logps/rejected": -340.0, + "loss": 0.2827, + "rewards/accuracies": 0.84375, + "rewards/chosen": -0.06640625, + "rewards/margins": 7.1875, + "rewards/rejected": -7.25, + "step": 739 + }, + { + "epoch": 0.9079754601226994, + "grad_norm": 29.046321525927297, + "learning_rate": 1.056818181818182e-06, + "logits/chosen": -0.0308837890625, + "logits/rejected": -0.18359375, + "logps/chosen": -344.0, + "logps/rejected": -326.0, + "loss": 0.3577, + "rewards/accuracies": 0.8515625, + "rewards/chosen": 0.02392578125, + "rewards/margins": 7.5625, + "rewards/rejected": -7.53125, + "step": 740 + }, + { + "epoch": 0.90920245398773, + "grad_norm": 29.826427625052336, + "learning_rate": 1.056198347107438e-06, + "logits/chosen": -0.00982666015625, + "logits/rejected": -0.162109375, + "logps/chosen": -356.0, + "logps/rejected": -328.0, + "loss": 0.3514, + "rewards/accuracies": 0.84375, + "rewards/chosen": 0.8359375, + "rewards/margins": 8.75, + "rewards/rejected": -7.875, + "step": 741 + }, + { + "epoch": 0.9104294478527607, + "grad_norm": 28.054610141633944, + "learning_rate": 1.0555785123966942e-06, + "logits/chosen": -0.033935546875, + "logits/rejected": -0.1982421875, + "logps/chosen": -368.0, + "logps/rejected": -372.0, + "loss": 0.2813, + "rewards/accuracies": 0.8828125, + "rewards/chosen": 0.6640625, + "rewards/margins": 8.5625, + "rewards/rejected": -7.90625, + "step": 742 + }, + { + "epoch": 0.9116564417177914, + "grad_norm": 35.487753263657986, + "learning_rate": 1.0549586776859503e-06, + "logits/chosen": 0.01055908203125, + "logits/rejected": -0.040771484375, + "logps/chosen": -424.0, + "logps/rejected": -392.0, + "loss": 0.3821, + "rewards/accuracies": 0.8203125, + "rewards/chosen": 0.2265625, + "rewards/margins": 7.21875, + "rewards/rejected": -7.0, + "step": 743 + }, + { + "epoch": 0.912883435582822, + "grad_norm": 28.496366117513464, + "learning_rate": 1.0543388429752067e-06, + "logits/chosen": 0.035400390625, + "logits/rejected": -0.14453125, + "logps/chosen": -380.0, + "logps/rejected": -340.0, + "loss": 0.2775, + "rewards/accuracies": 0.859375, + "rewards/chosen": 0.9609375, + "rewards/margins": 8.6875, + "rewards/rejected": -7.71875, + "step": 744 + }, + { + "epoch": 0.9141104294478528, + "grad_norm": 30.005979043165016, + "learning_rate": 1.0537190082644628e-06, + "logits/chosen": 0.08642578125, + "logits/rejected": -0.0003509521484375, + "logps/chosen": -346.0, + "logps/rejected": -318.0, + "loss": 0.3381, + "rewards/accuracies": 0.8515625, + "rewards/chosen": 0.1044921875, + "rewards/margins": 6.84375, + "rewards/rejected": -6.75, + "step": 745 + }, + { + "epoch": 0.9153374233128835, + "grad_norm": 36.365949373279655, + "learning_rate": 1.053099173553719e-06, + "logits/chosen": 0.09033203125, + "logits/rejected": -0.07958984375, + "logps/chosen": -408.0, + "logps/rejected": -356.0, + "loss": 0.4283, + "rewards/accuracies": 0.84375, + "rewards/chosen": 0.4375, + "rewards/margins": 7.4375, + "rewards/rejected": -7.0, + "step": 746 + }, + { + "epoch": 0.9165644171779141, + "grad_norm": 30.46892734256714, + "learning_rate": 1.0524793388429753e-06, + "logits/chosen": -0.1123046875, + "logits/rejected": -0.21875, + "logps/chosen": -342.0, + "logps/rejected": -344.0, + "loss": 0.2905, + "rewards/accuracies": 0.8515625, + "rewards/chosen": -0.0234375, + "rewards/margins": 6.75, + "rewards/rejected": -6.75, + "step": 747 + }, + { + "epoch": 0.9177914110429448, + "grad_norm": 37.226894144134455, + "learning_rate": 1.0518595041322315e-06, + "logits/chosen": 0.125, + "logits/rejected": 0.01458740234375, + "logps/chosen": -380.0, + "logps/rejected": -354.0, + "loss": 0.4992, + "rewards/accuracies": 0.78125, + "rewards/chosen": -0.50390625, + "rewards/margins": 6.28125, + "rewards/rejected": -6.78125, + "step": 748 + }, + { + "epoch": 0.9190184049079755, + "grad_norm": 34.24194812875213, + "learning_rate": 1.0512396694214876e-06, + "logits/chosen": -0.04541015625, + "logits/rejected": -0.2431640625, + "logps/chosen": -424.0, + "logps/rejected": -358.0, + "loss": 0.4049, + "rewards/accuracies": 0.8125, + "rewards/chosen": 0.2470703125, + "rewards/margins": 7.59375, + "rewards/rejected": -7.34375, + "step": 749 + }, + { + "epoch": 0.9202453987730062, + "grad_norm": 35.848596312284045, + "learning_rate": 1.050619834710744e-06, + "logits/chosen": 0.01190185546875, + "logits/rejected": -0.119140625, + "logps/chosen": -412.0, + "logps/rejected": -348.0, + "loss": 0.4472, + "rewards/accuracies": 0.828125, + "rewards/chosen": 0.09326171875, + "rewards/margins": 7.375, + "rewards/rejected": -7.28125, + "step": 750 + }, + { + "epoch": 0.9214723926380368, + "grad_norm": 36.49684940562663, + "learning_rate": 1.05e-06, + "logits/chosen": -0.0064697265625, + "logits/rejected": -0.169921875, + "logps/chosen": -428.0, + "logps/rejected": -366.0, + "loss": 0.3626, + "rewards/accuracies": 0.84375, + "rewards/chosen": 0.5234375, + "rewards/margins": 8.3125, + "rewards/rejected": -7.8125, + "step": 751 + }, + { + "epoch": 0.9226993865030675, + "grad_norm": 31.9902592360704, + "learning_rate": 1.0493801652892563e-06, + "logits/chosen": -0.028564453125, + "logits/rejected": -0.275390625, + "logps/chosen": -424.0, + "logps/rejected": -368.0, + "loss": 0.3427, + "rewards/accuracies": 0.8515625, + "rewards/chosen": 0.640625, + "rewards/margins": 8.375, + "rewards/rejected": -7.71875, + "step": 752 + }, + { + "epoch": 0.9239263803680982, + "grad_norm": 30.210419690529548, + "learning_rate": 1.0487603305785124e-06, + "logits/chosen": 0.142578125, + "logits/rejected": 0.01361083984375, + "logps/chosen": -408.0, + "logps/rejected": -362.0, + "loss": 0.3032, + "rewards/accuracies": 0.859375, + "rewards/chosen": 0.126953125, + "rewards/margins": 7.375, + "rewards/rejected": -7.25, + "step": 753 + }, + { + "epoch": 0.9251533742331288, + "grad_norm": 33.70924536763799, + "learning_rate": 1.0481404958677685e-06, + "logits/chosen": 0.06640625, + "logits/rejected": -0.1044921875, + "logps/chosen": -394.0, + "logps/rejected": -348.0, + "loss": 0.4405, + "rewards/accuracies": 0.796875, + "rewards/chosen": -0.1416015625, + "rewards/margins": 7.09375, + "rewards/rejected": -7.21875, + "step": 754 + }, + { + "epoch": 0.9263803680981595, + "grad_norm": 28.923057351049156, + "learning_rate": 1.0475206611570249e-06, + "logits/chosen": 0.01611328125, + "logits/rejected": -0.1064453125, + "logps/chosen": -388.0, + "logps/rejected": -366.0, + "loss": 0.313, + "rewards/accuracies": 0.875, + "rewards/chosen": -0.291015625, + "rewards/margins": 7.21875, + "rewards/rejected": -7.5, + "step": 755 + }, + { + "epoch": 0.9276073619631902, + "grad_norm": 26.956563960777117, + "learning_rate": 1.046900826446281e-06, + "logits/chosen": 0.115234375, + "logits/rejected": 0.02978515625, + "logps/chosen": -342.0, + "logps/rejected": -326.0, + "loss": 0.2891, + "rewards/accuracies": 0.8828125, + "rewards/chosen": -0.640625, + "rewards/margins": 6.8125, + "rewards/rejected": -7.46875, + "step": 756 + }, + { + "epoch": 0.9288343558282208, + "grad_norm": 25.54044780724446, + "learning_rate": 1.0462809917355372e-06, + "logits/chosen": 0.15625, + "logits/rejected": -0.0595703125, + "logps/chosen": -404.0, + "logps/rejected": -344.0, + "loss": 0.27, + "rewards/accuracies": 0.875, + "rewards/chosen": 0.7578125, + "rewards/margins": 8.0625, + "rewards/rejected": -7.3125, + "step": 757 + }, + { + "epoch": 0.9300613496932515, + "grad_norm": 35.658646153584634, + "learning_rate": 1.0456611570247935e-06, + "logits/chosen": 0.08740234375, + "logits/rejected": -0.080078125, + "logps/chosen": -376.0, + "logps/rejected": -320.0, + "loss": 0.3867, + "rewards/accuracies": 0.8125, + "rewards/chosen": -0.55859375, + "rewards/margins": 6.71875, + "rewards/rejected": -7.28125, + "step": 758 + }, + { + "epoch": 0.9312883435582822, + "grad_norm": 29.003566311314795, + "learning_rate": 1.0450413223140497e-06, + "logits/chosen": 0.1123046875, + "logits/rejected": -0.11572265625, + "logps/chosen": -396.0, + "logps/rejected": -340.0, + "loss": 0.2888, + "rewards/accuracies": 0.84375, + "rewards/chosen": 0.16015625, + "rewards/margins": 6.9375, + "rewards/rejected": -6.78125, + "step": 759 + }, + { + "epoch": 0.9325153374233128, + "grad_norm": 25.36330419002366, + "learning_rate": 1.0444214876033058e-06, + "logits/chosen": -0.0986328125, + "logits/rejected": -0.35546875, + "logps/chosen": -442.0, + "logps/rejected": -366.0, + "loss": 0.2887, + "rewards/accuracies": 0.875, + "rewards/chosen": 1.1796875, + "rewards/margins": 8.625, + "rewards/rejected": -7.4375, + "step": 760 + }, + { + "epoch": 0.9337423312883436, + "grad_norm": 31.72216949313869, + "learning_rate": 1.043801652892562e-06, + "logits/chosen": 0.1484375, + "logits/rejected": 0.0277099609375, + "logps/chosen": -382.0, + "logps/rejected": -354.0, + "loss": 0.3707, + "rewards/accuracies": 0.8671875, + "rewards/chosen": 0.3984375, + "rewards/margins": 7.21875, + "rewards/rejected": -6.8125, + "step": 761 + }, + { + "epoch": 0.9349693251533743, + "grad_norm": 31.01314079949462, + "learning_rate": 1.043181818181818e-06, + "logits/chosen": 0.1591796875, + "logits/rejected": -0.020751953125, + "logps/chosen": -428.0, + "logps/rejected": -364.0, + "loss": 0.3149, + "rewards/accuracies": 0.8984375, + "rewards/chosen": 0.2373046875, + "rewards/margins": 7.53125, + "rewards/rejected": -7.28125, + "step": 762 + }, + { + "epoch": 0.9361963190184049, + "grad_norm": 28.119069224585317, + "learning_rate": 1.0425619834710744e-06, + "logits/chosen": 0.0986328125, + "logits/rejected": 0.0164794921875, + "logps/chosen": -400.0, + "logps/rejected": -364.0, + "loss": 0.3398, + "rewards/accuracies": 0.84375, + "rewards/chosen": 0.4375, + "rewards/margins": 7.5625, + "rewards/rejected": -7.125, + "step": 763 + }, + { + "epoch": 0.9374233128834356, + "grad_norm": 26.42805050753053, + "learning_rate": 1.0419421487603306e-06, + "logits/chosen": -0.1328125, + "logits/rejected": -0.287109375, + "logps/chosen": -330.0, + "logps/rejected": -316.0, + "loss": 0.2955, + "rewards/accuracies": 0.90625, + "rewards/chosen": -0.494140625, + "rewards/margins": 7.0, + "rewards/rejected": -7.46875, + "step": 764 + }, + { + "epoch": 0.9386503067484663, + "grad_norm": 35.659703021872964, + "learning_rate": 1.041322314049587e-06, + "logits/chosen": 0.1298828125, + "logits/rejected": 0.005828857421875, + "logps/chosen": -404.0, + "logps/rejected": -320.0, + "loss": 0.3833, + "rewards/accuracies": 0.796875, + "rewards/chosen": -0.2041015625, + "rewards/margins": 6.125, + "rewards/rejected": -6.34375, + "step": 765 + }, + { + "epoch": 0.939877300613497, + "grad_norm": 27.867245312963597, + "learning_rate": 1.040702479338843e-06, + "logits/chosen": 0.04345703125, + "logits/rejected": -0.1494140625, + "logps/chosen": -446.0, + "logps/rejected": -340.0, + "loss": 0.3188, + "rewards/accuracies": 0.8671875, + "rewards/chosen": 0.515625, + "rewards/margins": 7.5, + "rewards/rejected": -7.0, + "step": 766 + }, + { + "epoch": 0.9411042944785276, + "grad_norm": 33.0423901915957, + "learning_rate": 1.0400826446280992e-06, + "logits/chosen": 0.0390625, + "logits/rejected": -0.10498046875, + "logps/chosen": -386.0, + "logps/rejected": -338.0, + "loss": 0.355, + "rewards/accuracies": 0.796875, + "rewards/chosen": 0.1689453125, + "rewards/margins": 7.5625, + "rewards/rejected": -7.375, + "step": 767 + }, + { + "epoch": 0.9423312883435583, + "grad_norm": 31.83881144027911, + "learning_rate": 1.0394628099173554e-06, + "logits/chosen": 0.003570556640625, + "logits/rejected": -0.1611328125, + "logps/chosen": -410.0, + "logps/rejected": -384.0, + "loss": 0.3826, + "rewards/accuracies": 0.875, + "rewards/chosen": 0.6484375, + "rewards/margins": 7.59375, + "rewards/rejected": -6.96875, + "step": 768 + }, + { + "epoch": 0.943558282208589, + "grad_norm": 24.722102647279602, + "learning_rate": 1.0388429752066115e-06, + "logits/chosen": 0.00079345703125, + "logits/rejected": -0.138671875, + "logps/chosen": -378.0, + "logps/rejected": -344.0, + "loss": 0.2777, + "rewards/accuracies": 0.859375, + "rewards/chosen": -0.068359375, + "rewards/margins": 7.53125, + "rewards/rejected": -7.59375, + "step": 769 + }, + { + "epoch": 0.9447852760736196, + "grad_norm": 33.3608331180557, + "learning_rate": 1.0382231404958679e-06, + "logits/chosen": -0.021728515625, + "logits/rejected": -0.265625, + "logps/chosen": -416.0, + "logps/rejected": -372.0, + "loss": 0.321, + "rewards/accuracies": 0.8515625, + "rewards/chosen": 0.92578125, + "rewards/margins": 7.5, + "rewards/rejected": -6.59375, + "step": 770 + }, + { + "epoch": 0.9460122699386503, + "grad_norm": 36.53871156772584, + "learning_rate": 1.037603305785124e-06, + "logits/chosen": -0.00537109375, + "logits/rejected": -0.01507568359375, + "logps/chosen": -430.0, + "logps/rejected": -360.0, + "loss": 0.5217, + "rewards/accuracies": 0.78125, + "rewards/chosen": -0.6484375, + "rewards/margins": 6.21875, + "rewards/rejected": -6.875, + "step": 771 + }, + { + "epoch": 0.947239263803681, + "grad_norm": 30.200058873190507, + "learning_rate": 1.0369834710743801e-06, + "logits/chosen": 0.1328125, + "logits/rejected": -0.0869140625, + "logps/chosen": -424.0, + "logps/rejected": -362.0, + "loss": 0.3536, + "rewards/accuracies": 0.84375, + "rewards/chosen": 0.38671875, + "rewards/margins": 7.28125, + "rewards/rejected": -6.90625, + "step": 772 + }, + { + "epoch": 0.9484662576687116, + "grad_norm": 31.95788929467944, + "learning_rate": 1.0363636363636365e-06, + "logits/chosen": 0.1484375, + "logits/rejected": -0.037841796875, + "logps/chosen": -374.0, + "logps/rejected": -324.0, + "loss": 0.3406, + "rewards/accuracies": 0.828125, + "rewards/chosen": -0.65234375, + "rewards/margins": 6.5, + "rewards/rejected": -7.15625, + "step": 773 + }, + { + "epoch": 0.9496932515337423, + "grad_norm": 34.11224513726983, + "learning_rate": 1.0357438016528926e-06, + "logits/chosen": 0.1396484375, + "logits/rejected": -0.0037078857421875, + "logps/chosen": -362.0, + "logps/rejected": -326.0, + "loss": 0.4295, + "rewards/accuracies": 0.7890625, + "rewards/chosen": -1.359375, + "rewards/margins": 5.1875, + "rewards/rejected": -6.5625, + "step": 774 + }, + { + "epoch": 0.950920245398773, + "grad_norm": 31.83399815065193, + "learning_rate": 1.0351239669421488e-06, + "logits/chosen": 0.2197265625, + "logits/rejected": 0.04296875, + "logps/chosen": -418.0, + "logps/rejected": -362.0, + "loss": 0.3519, + "rewards/accuracies": 0.84375, + "rewards/chosen": -0.32421875, + "rewards/margins": 6.21875, + "rewards/rejected": -6.53125, + "step": 775 + }, + { + "epoch": 0.9521472392638037, + "grad_norm": 29.278356141500073, + "learning_rate": 1.0345041322314051e-06, + "logits/chosen": -0.01409912109375, + "logits/rejected": -0.1708984375, + "logps/chosen": -414.0, + "logps/rejected": -366.0, + "loss": 0.3978, + "rewards/accuracies": 0.84375, + "rewards/chosen": 0.0888671875, + "rewards/margins": 6.8125, + "rewards/rejected": -6.75, + "step": 776 + }, + { + "epoch": 0.9533742331288344, + "grad_norm": 33.3868265617977, + "learning_rate": 1.033884297520661e-06, + "logits/chosen": 0.0267333984375, + "logits/rejected": -0.1015625, + "logps/chosen": -408.0, + "logps/rejected": -346.0, + "loss": 0.3944, + "rewards/accuracies": 0.8046875, + "rewards/chosen": -0.58203125, + "rewards/margins": 6.71875, + "rewards/rejected": -7.3125, + "step": 777 + }, + { + "epoch": 0.9546012269938651, + "grad_norm": 30.60810394473529, + "learning_rate": 1.0332644628099174e-06, + "logits/chosen": 0.0703125, + "logits/rejected": -0.130859375, + "logps/chosen": -400.0, + "logps/rejected": -328.0, + "loss": 0.2893, + "rewards/accuracies": 0.8671875, + "rewards/chosen": -0.74609375, + "rewards/margins": 6.0, + "rewards/rejected": -6.75, + "step": 778 + }, + { + "epoch": 0.9558282208588957, + "grad_norm": 31.819603506171298, + "learning_rate": 1.0326446280991735e-06, + "logits/chosen": 0.00994873046875, + "logits/rejected": -0.1533203125, + "logps/chosen": -442.0, + "logps/rejected": -362.0, + "loss": 0.331, + "rewards/accuracies": 0.8671875, + "rewards/chosen": 0.16796875, + "rewards/margins": 7.59375, + "rewards/rejected": -7.40625, + "step": 779 + }, + { + "epoch": 0.9570552147239264, + "grad_norm": 33.50608635895235, + "learning_rate": 1.0320247933884297e-06, + "logits/chosen": 0.13671875, + "logits/rejected": -0.037353515625, + "logps/chosen": -386.0, + "logps/rejected": -360.0, + "loss": 0.4421, + "rewards/accuracies": 0.7890625, + "rewards/chosen": -0.91015625, + "rewards/margins": 5.96875, + "rewards/rejected": -6.875, + "step": 780 + }, + { + "epoch": 0.9582822085889571, + "grad_norm": 27.175413708110298, + "learning_rate": 1.031404958677686e-06, + "logits/chosen": -0.0087890625, + "logits/rejected": -0.2392578125, + "logps/chosen": -426.0, + "logps/rejected": -356.0, + "loss": 0.2289, + "rewards/accuracies": 0.8828125, + "rewards/chosen": 0.9140625, + "rewards/margins": 8.75, + "rewards/rejected": -7.8125, + "step": 781 + }, + { + "epoch": 0.9595092024539877, + "grad_norm": 27.961517528382746, + "learning_rate": 1.0307851239669422e-06, + "logits/chosen": 0.0341796875, + "logits/rejected": -0.055419921875, + "logps/chosen": -370.0, + "logps/rejected": -354.0, + "loss": 0.2727, + "rewards/accuracies": 0.890625, + "rewards/chosen": 0.0152587890625, + "rewards/margins": 7.21875, + "rewards/rejected": -7.1875, + "step": 782 + }, + { + "epoch": 0.9607361963190184, + "grad_norm": 36.53707624363368, + "learning_rate": 1.0301652892561983e-06, + "logits/chosen": 0.0115966796875, + "logits/rejected": -0.12353515625, + "logps/chosen": -420.0, + "logps/rejected": -370.0, + "loss": 0.3333, + "rewards/accuracies": 0.859375, + "rewards/chosen": 0.408203125, + "rewards/margins": 7.9375, + "rewards/rejected": -7.5, + "step": 783 + }, + { + "epoch": 0.9619631901840491, + "grad_norm": 33.01812506650061, + "learning_rate": 1.0295454545454547e-06, + "logits/chosen": 0.09716796875, + "logits/rejected": 0.056640625, + "logps/chosen": -386.0, + "logps/rejected": -354.0, + "loss": 0.3887, + "rewards/accuracies": 0.828125, + "rewards/chosen": -0.69140625, + "rewards/margins": 6.4375, + "rewards/rejected": -7.125, + "step": 784 + }, + { + "epoch": 0.9631901840490797, + "grad_norm": 25.64548904648115, + "learning_rate": 1.0289256198347108e-06, + "logits/chosen": 0.126953125, + "logits/rejected": -0.0400390625, + "logps/chosen": -384.0, + "logps/rejected": -324.0, + "loss": 0.2131, + "rewards/accuracies": 0.875, + "rewards/chosen": 0.0047607421875, + "rewards/margins": 7.40625, + "rewards/rejected": -7.375, + "step": 785 + }, + { + "epoch": 0.9644171779141104, + "grad_norm": 24.085836304747, + "learning_rate": 1.028305785123967e-06, + "logits/chosen": 0.11376953125, + "logits/rejected": -0.09033203125, + "logps/chosen": -372.0, + "logps/rejected": -334.0, + "loss": 0.2197, + "rewards/accuracies": 0.8984375, + "rewards/chosen": -0.427734375, + "rewards/margins": 7.375, + "rewards/rejected": -7.8125, + "step": 786 + }, + { + "epoch": 0.9656441717791411, + "grad_norm": 32.97798807698007, + "learning_rate": 1.027685950413223e-06, + "logits/chosen": 0.0269775390625, + "logits/rejected": -0.08740234375, + "logps/chosen": -412.0, + "logps/rejected": -348.0, + "loss": 0.3407, + "rewards/accuracies": 0.859375, + "rewards/chosen": -0.271484375, + "rewards/margins": 7.65625, + "rewards/rejected": -7.9375, + "step": 787 + }, + { + "epoch": 0.9668711656441717, + "grad_norm": 24.19044914351049, + "learning_rate": 1.0270661157024792e-06, + "logits/chosen": -0.00860595703125, + "logits/rejected": -0.1787109375, + "logps/chosen": -390.0, + "logps/rejected": -366.0, + "loss": 0.2123, + "rewards/accuracies": 0.9140625, + "rewards/chosen": 0.486328125, + "rewards/margins": 8.1875, + "rewards/rejected": -7.6875, + "step": 788 + }, + { + "epoch": 0.9680981595092024, + "grad_norm": 30.061433994827276, + "learning_rate": 1.0264462809917356e-06, + "logits/chosen": 0.12060546875, + "logits/rejected": -0.06396484375, + "logps/chosen": -384.0, + "logps/rejected": -332.0, + "loss": 0.383, + "rewards/accuracies": 0.859375, + "rewards/chosen": -0.52734375, + "rewards/margins": 6.71875, + "rewards/rejected": -7.25, + "step": 789 + }, + { + "epoch": 0.9693251533742331, + "grad_norm": 27.089223909733843, + "learning_rate": 1.0258264462809917e-06, + "logits/chosen": -0.028564453125, + "logits/rejected": -0.2373046875, + "logps/chosen": -388.0, + "logps/rejected": -330.0, + "loss": 0.1804, + "rewards/accuracies": 0.8984375, + "rewards/chosen": -0.1533203125, + "rewards/margins": 7.3125, + "rewards/rejected": -7.46875, + "step": 790 + }, + { + "epoch": 0.9705521472392638, + "grad_norm": 34.35477577844613, + "learning_rate": 1.0252066115702479e-06, + "logits/chosen": -0.0303955078125, + "logits/rejected": -0.23046875, + "logps/chosen": -398.0, + "logps/rejected": -346.0, + "loss": 0.3642, + "rewards/accuracies": 0.8125, + "rewards/chosen": 0.349609375, + "rewards/margins": 8.25, + "rewards/rejected": -7.90625, + "step": 791 + }, + { + "epoch": 0.9717791411042945, + "grad_norm": 33.05641026053793, + "learning_rate": 1.0245867768595042e-06, + "logits/chosen": 0.0006256103515625, + "logits/rejected": -0.1279296875, + "logps/chosen": -438.0, + "logps/rejected": -374.0, + "loss": 0.406, + "rewards/accuracies": 0.8125, + "rewards/chosen": 0.380859375, + "rewards/margins": 6.8125, + "rewards/rejected": -6.4375, + "step": 792 + }, + { + "epoch": 0.9730061349693252, + "grad_norm": 32.69405943160363, + "learning_rate": 1.0239669421487604e-06, + "logits/chosen": 0.0284423828125, + "logits/rejected": -0.09765625, + "logps/chosen": -392.0, + "logps/rejected": -362.0, + "loss": 0.4018, + "rewards/accuracies": 0.8671875, + "rewards/chosen": -0.10546875, + "rewards/margins": 6.75, + "rewards/rejected": -6.84375, + "step": 793 + }, + { + "epoch": 0.9742331288343559, + "grad_norm": 33.74613547302303, + "learning_rate": 1.0233471074380167e-06, + "logits/chosen": 0.0830078125, + "logits/rejected": -0.043701171875, + "logps/chosen": -358.0, + "logps/rejected": -336.0, + "loss": 0.4877, + "rewards/accuracies": 0.8046875, + "rewards/chosen": -0.6015625, + "rewards/margins": 6.59375, + "rewards/rejected": -7.1875, + "step": 794 + }, + { + "epoch": 0.9754601226993865, + "grad_norm": 22.619679414305867, + "learning_rate": 1.0227272727272727e-06, + "logits/chosen": -0.1513671875, + "logits/rejected": -0.296875, + "logps/chosen": -392.0, + "logps/rejected": -372.0, + "loss": 0.281, + "rewards/accuracies": 0.875, + "rewards/chosen": 0.80859375, + "rewards/margins": 9.25, + "rewards/rejected": -8.4375, + "step": 795 + }, + { + "epoch": 0.9766871165644172, + "grad_norm": 35.85933901249144, + "learning_rate": 1.0221074380165288e-06, + "logits/chosen": -0.0260009765625, + "logits/rejected": -0.2353515625, + "logps/chosen": -410.0, + "logps/rejected": -338.0, + "loss": 0.3447, + "rewards/accuracies": 0.8828125, + "rewards/chosen": -0.78515625, + "rewards/margins": 7.25, + "rewards/rejected": -8.0625, + "step": 796 + }, + { + "epoch": 0.9779141104294479, + "grad_norm": 24.484140263550707, + "learning_rate": 1.0214876033057852e-06, + "logits/chosen": 0.059814453125, + "logits/rejected": -0.11962890625, + "logps/chosen": -404.0, + "logps/rejected": -324.0, + "loss": 0.243, + "rewards/accuracies": 0.8671875, + "rewards/chosen": -0.0966796875, + "rewards/margins": 7.15625, + "rewards/rejected": -7.25, + "step": 797 + }, + { + "epoch": 0.9791411042944785, + "grad_norm": 35.333603529709684, + "learning_rate": 1.0208677685950413e-06, + "logits/chosen": 0.00506591796875, + "logits/rejected": -0.173828125, + "logps/chosen": -430.0, + "logps/rejected": -378.0, + "loss": 0.3546, + "rewards/accuracies": 0.8203125, + "rewards/chosen": 0.1455078125, + "rewards/margins": 7.78125, + "rewards/rejected": -7.625, + "step": 798 + }, + { + "epoch": 0.9803680981595092, + "grad_norm": 32.17706869798208, + "learning_rate": 1.0202479338842976e-06, + "logits/chosen": -0.05029296875, + "logits/rejected": -0.201171875, + "logps/chosen": -400.0, + "logps/rejected": -360.0, + "loss": 0.3323, + "rewards/accuracies": 0.8203125, + "rewards/chosen": 0.4296875, + "rewards/margins": 7.71875, + "rewards/rejected": -7.28125, + "step": 799 + }, + { + "epoch": 0.9815950920245399, + "grad_norm": 32.44337864899576, + "learning_rate": 1.0196280991735538e-06, + "logits/chosen": -0.0224609375, + "logits/rejected": -0.208984375, + "logps/chosen": -386.0, + "logps/rejected": -344.0, + "loss": 0.3141, + "rewards/accuracies": 0.8828125, + "rewards/chosen": -0.0654296875, + "rewards/margins": 7.28125, + "rewards/rejected": -7.375, + "step": 800 + }, + { + "epoch": 0.9828220858895705, + "grad_norm": 37.380385495976945, + "learning_rate": 1.01900826446281e-06, + "logits/chosen": -0.024658203125, + "logits/rejected": -0.17578125, + "logps/chosen": -436.0, + "logps/rejected": -382.0, + "loss": 0.4546, + "rewards/accuracies": 0.828125, + "rewards/chosen": 0.0001010894775390625, + "rewards/margins": 6.6875, + "rewards/rejected": -6.6875, + "step": 801 + }, + { + "epoch": 0.9840490797546012, + "grad_norm": 35.755684948235334, + "learning_rate": 1.0183884297520663e-06, + "logits/chosen": -0.07421875, + "logits/rejected": -0.21484375, + "logps/chosen": -386.0, + "logps/rejected": -348.0, + "loss": 0.472, + "rewards/accuracies": 0.84375, + "rewards/chosen": -0.1552734375, + "rewards/margins": 7.0, + "rewards/rejected": -7.15625, + "step": 802 + }, + { + "epoch": 0.9852760736196319, + "grad_norm": 29.08898198089266, + "learning_rate": 1.0177685950413224e-06, + "logits/chosen": -0.1513671875, + "logits/rejected": -0.302734375, + "logps/chosen": -398.0, + "logps/rejected": -348.0, + "loss": 0.3346, + "rewards/accuracies": 0.84375, + "rewards/chosen": 0.54296875, + "rewards/margins": 7.75, + "rewards/rejected": -7.21875, + "step": 803 + }, + { + "epoch": 0.9865030674846625, + "grad_norm": 30.37947277261302, + "learning_rate": 1.0171487603305786e-06, + "logits/chosen": 0.04150390625, + "logits/rejected": -0.06689453125, + "logps/chosen": -358.0, + "logps/rejected": -346.0, + "loss": 0.3638, + "rewards/accuracies": 0.859375, + "rewards/chosen": -0.392578125, + "rewards/margins": 6.9375, + "rewards/rejected": -7.34375, + "step": 804 + }, + { + "epoch": 0.9877300613496932, + "grad_norm": 34.04488703971709, + "learning_rate": 1.0165289256198347e-06, + "logits/chosen": -0.078125, + "logits/rejected": -0.2119140625, + "logps/chosen": -434.0, + "logps/rejected": -394.0, + "loss": 0.4569, + "rewards/accuracies": 0.828125, + "rewards/chosen": 0.4296875, + "rewards/margins": 7.78125, + "rewards/rejected": -7.375, + "step": 805 + }, + { + "epoch": 0.9889570552147239, + "grad_norm": 35.48502129676359, + "learning_rate": 1.0159090909090908e-06, + "logits/chosen": 0.16015625, + "logits/rejected": 0.03564453125, + "logps/chosen": -412.0, + "logps/rejected": -384.0, + "loss": 0.4253, + "rewards/accuracies": 0.8046875, + "rewards/chosen": 0.31640625, + "rewards/margins": 7.15625, + "rewards/rejected": -6.84375, + "step": 806 + }, + { + "epoch": 0.9901840490797545, + "grad_norm": 33.61353577378301, + "learning_rate": 1.0152892561983472e-06, + "logits/chosen": 0.061767578125, + "logits/rejected": -0.0908203125, + "logps/chosen": -400.0, + "logps/rejected": -350.0, + "loss": 0.3144, + "rewards/accuracies": 0.8828125, + "rewards/chosen": 0.484375, + "rewards/margins": 8.25, + "rewards/rejected": -7.75, + "step": 807 + }, + { + "epoch": 0.9914110429447853, + "grad_norm": 31.091169744856, + "learning_rate": 1.0146694214876033e-06, + "logits/chosen": -0.083984375, + "logits/rejected": -0.0966796875, + "logps/chosen": -364.0, + "logps/rejected": -352.0, + "loss": 0.3745, + "rewards/accuracies": 0.84375, + "rewards/chosen": -0.640625, + "rewards/margins": 7.5625, + "rewards/rejected": -8.1875, + "step": 808 + }, + { + "epoch": 0.992638036809816, + "grad_norm": 34.39338760048118, + "learning_rate": 1.0140495867768595e-06, + "logits/chosen": -0.00872802734375, + "logits/rejected": -0.2314453125, + "logps/chosen": -438.0, + "logps/rejected": -360.0, + "loss": 0.319, + "rewards/accuracies": 0.859375, + "rewards/chosen": 0.8515625, + "rewards/margins": 8.5625, + "rewards/rejected": -7.71875, + "step": 809 + }, + { + "epoch": 0.9938650306748467, + "grad_norm": 26.950235054794184, + "learning_rate": 1.0134297520661158e-06, + "logits/chosen": -0.11572265625, + "logits/rejected": -0.28515625, + "logps/chosen": -446.0, + "logps/rejected": -384.0, + "loss": 0.2346, + "rewards/accuracies": 0.8828125, + "rewards/chosen": 0.44140625, + "rewards/margins": 7.9375, + "rewards/rejected": -7.5, + "step": 810 + }, + { + "epoch": 0.9950920245398773, + "grad_norm": 26.578554179921486, + "learning_rate": 1.012809917355372e-06, + "logits/chosen": -0.08203125, + "logits/rejected": -0.21484375, + "logps/chosen": -394.0, + "logps/rejected": -380.0, + "loss": 0.3194, + "rewards/accuracies": 0.84375, + "rewards/chosen": 0.44921875, + "rewards/margins": 7.46875, + "rewards/rejected": -7.03125, + "step": 811 + }, + { + "epoch": 0.996319018404908, + "grad_norm": 34.32887556980461, + "learning_rate": 1.0121900826446281e-06, + "logits/chosen": -0.05908203125, + "logits/rejected": -0.125, + "logps/chosen": -364.0, + "logps/rejected": -340.0, + "loss": 0.3746, + "rewards/accuracies": 0.875, + "rewards/chosen": -0.09619140625, + "rewards/margins": 7.625, + "rewards/rejected": -7.71875, + "step": 812 + }, + { + "epoch": 0.9975460122699387, + "grad_norm": 25.82868308838759, + "learning_rate": 1.0115702479338843e-06, + "logits/chosen": 0.01611328125, + "logits/rejected": -0.1123046875, + "logps/chosen": -342.0, + "logps/rejected": -336.0, + "loss": 0.2067, + "rewards/accuracies": 0.8984375, + "rewards/chosen": 0.46484375, + "rewards/margins": 7.84375, + "rewards/rejected": -7.375, + "step": 813 + }, + { + "epoch": 0.9987730061349693, + "grad_norm": 35.210336167923685, + "learning_rate": 1.0109504132231404e-06, + "logits/chosen": 0.054443359375, + "logits/rejected": -0.0206298828125, + "logps/chosen": -388.0, + "logps/rejected": -370.0, + "loss": 0.3789, + "rewards/accuracies": 0.78125, + "rewards/chosen": -0.80859375, + "rewards/margins": 6.125, + "rewards/rejected": -6.90625, + "step": 814 + }, + { + "epoch": 1.0, + "grad_norm": 31.93988486773496, + "learning_rate": 1.0103305785123968e-06, + "logits/chosen": 0.1279296875, + "logits/rejected": -0.061279296875, + "logps/chosen": -388.0, + "logps/rejected": -364.0, + "loss": 0.2533, + "rewards/accuracies": 0.8991597294807434, + "rewards/chosen": -0.095703125, + "rewards/margins": 7.53125, + "rewards/rejected": -7.625, + "step": 815 + }, + { + "epoch": 1.0012269938650307, + "grad_norm": 5.842576996186414, + "learning_rate": 1.009710743801653e-06, + "logits/chosen": 0.0537109375, + "logits/rejected": -0.08154296875, + "logps/chosen": -376.0, + "logps/rejected": -332.0, + "loss": 0.034, + "rewards/accuracies": 1.0, + "rewards/chosen": 1.2890625, + "rewards/margins": 9.5, + "rewards/rejected": -8.1875, + "step": 816 + }, + { + "epoch": 1.0024539877300613, + "grad_norm": 8.73132250150435, + "learning_rate": 1.009090909090909e-06, + "logits/chosen": 0.03857421875, + "logits/rejected": -0.107421875, + "logps/chosen": -362.0, + "logps/rejected": -334.0, + "loss": 0.0486, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.7109375, + "rewards/margins": 8.8125, + "rewards/rejected": -8.125, + "step": 817 + }, + { + "epoch": 1.003680981595092, + "grad_norm": 4.472623669255691, + "learning_rate": 1.0084710743801654e-06, + "logits/chosen": -0.09326171875, + "logits/rejected": -0.1728515625, + "logps/chosen": -364.0, + "logps/rejected": -362.0, + "loss": 0.0318, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 1.125, + "rewards/margins": 9.1875, + "rewards/rejected": -8.0625, + "step": 818 + }, + { + "epoch": 1.0049079754601227, + "grad_norm": 10.555989313773281, + "learning_rate": 1.0078512396694215e-06, + "logits/chosen": 0.0576171875, + "logits/rejected": -0.12109375, + "logps/chosen": -390.0, + "logps/rejected": -366.0, + "loss": 0.043, + "rewards/accuracies": 0.9765625, + "rewards/chosen": 1.4453125, + "rewards/margins": 9.875, + "rewards/rejected": -8.4375, + "step": 819 + }, + { + "epoch": 1.0061349693251533, + "grad_norm": 7.844244829439074, + "learning_rate": 1.0072314049586777e-06, + "logits/chosen": 0.09375, + "logits/rejected": -0.02880859375, + "logps/chosen": -396.0, + "logps/rejected": -356.0, + "loss": 0.0346, + "rewards/accuracies": 1.0, + "rewards/chosen": 1.6796875, + "rewards/margins": 9.25, + "rewards/rejected": -7.5625, + "step": 820 + }, + { + "epoch": 1.007361963190184, + "grad_norm": 13.093642235831343, + "learning_rate": 1.006611570247934e-06, + "logits/chosen": -0.0054931640625, + "logits/rejected": -0.126953125, + "logps/chosen": -350.0, + "logps/rejected": -342.0, + "loss": 0.0919, + "rewards/accuracies": 0.984375, + "rewards/chosen": 1.578125, + "rewards/margins": 9.5, + "rewards/rejected": -7.90625, + "step": 821 + }, + { + "epoch": 1.0085889570552147, + "grad_norm": 11.258353605754225, + "learning_rate": 1.00599173553719e-06, + "logits/chosen": 0.00653076171875, + "logits/rejected": -0.126953125, + "logps/chosen": -404.0, + "logps/rejected": -370.0, + "loss": 0.0581, + "rewards/accuracies": 0.9609375, + "rewards/chosen": 1.6484375, + "rewards/margins": 9.0, + "rewards/rejected": -7.375, + "step": 822 + }, + { + "epoch": 1.0098159509202453, + "grad_norm": 7.2063058089923135, + "learning_rate": 1.0053719008264463e-06, + "logits/chosen": 0.03466796875, + "logits/rejected": -0.1474609375, + "logps/chosen": -364.0, + "logps/rejected": -350.0, + "loss": 0.0354, + "rewards/accuracies": 0.984375, + "rewards/chosen": 1.84375, + "rewards/margins": 9.4375, + "rewards/rejected": -7.5625, + "step": 823 + }, + { + "epoch": 1.011042944785276, + "grad_norm": 10.387457112253443, + "learning_rate": 1.0047520661157024e-06, + "logits/chosen": -0.04833984375, + "logits/rejected": -0.1904296875, + "logps/chosen": -378.0, + "logps/rejected": -362.0, + "loss": 0.0277, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 1.921875, + "rewards/margins": 8.875, + "rewards/rejected": -7.0, + "step": 824 + }, + { + "epoch": 1.0122699386503067, + "grad_norm": 6.9895041600763665, + "learning_rate": 1.0041322314049586e-06, + "logits/chosen": 0.140625, + "logits/rejected": -0.130859375, + "logps/chosen": -396.0, + "logps/rejected": -356.0, + "loss": 0.0298, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 1.5546875, + "rewards/margins": 9.625, + "rewards/rejected": -8.0625, + "step": 825 + }, + { + "epoch": 1.0134969325153373, + "grad_norm": 12.841731144722432, + "learning_rate": 1.003512396694215e-06, + "logits/chosen": -0.064453125, + "logits/rejected": -0.197265625, + "logps/chosen": -372.0, + "logps/rejected": -354.0, + "loss": 0.0693, + "rewards/accuracies": 0.9609375, + "rewards/chosen": 1.1640625, + "rewards/margins": 8.6875, + "rewards/rejected": -7.5, + "step": 826 + }, + { + "epoch": 1.014723926380368, + "grad_norm": 6.4682689307001455, + "learning_rate": 1.002892561983471e-06, + "logits/chosen": 0.01708984375, + "logits/rejected": -0.091796875, + "logps/chosen": -344.0, + "logps/rejected": -338.0, + "loss": 0.0373, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.65625, + "rewards/margins": 8.4375, + "rewards/rejected": -7.75, + "step": 827 + }, + { + "epoch": 1.0159509202453987, + "grad_norm": 4.914698955857516, + "learning_rate": 1.0022727272727274e-06, + "logits/chosen": 0.15625, + "logits/rejected": -0.036865234375, + "logps/chosen": -382.0, + "logps/rejected": -330.0, + "loss": 0.0321, + "rewards/accuracies": 1.0, + "rewards/chosen": 1.15625, + "rewards/margins": 8.9375, + "rewards/rejected": -7.75, + "step": 828 + }, + { + "epoch": 1.0171779141104293, + "grad_norm": 13.706631134085173, + "learning_rate": 1.0016528925619836e-06, + "logits/chosen": -0.1396484375, + "logits/rejected": -0.1748046875, + "logps/chosen": -376.0, + "logps/rejected": -386.0, + "loss": 0.0804, + "rewards/accuracies": 0.96875, + "rewards/chosen": 1.5078125, + "rewards/margins": 9.875, + "rewards/rejected": -8.375, + "step": 829 + }, + { + "epoch": 1.01840490797546, + "grad_norm": 4.7042664986817755, + "learning_rate": 1.0010330578512397e-06, + "logits/chosen": -0.059814453125, + "logits/rejected": -0.25390625, + "logps/chosen": -410.0, + "logps/rejected": -384.0, + "loss": 0.0274, + "rewards/accuracies": 1.0, + "rewards/chosen": 2.03125, + "rewards/margins": 11.125, + "rewards/rejected": -9.125, + "step": 830 + }, + { + "epoch": 1.019631901840491, + "grad_norm": 10.294544241161669, + "learning_rate": 1.0004132231404959e-06, + "logits/chosen": -0.0341796875, + "logits/rejected": -0.0810546875, + "logps/chosen": -404.0, + "logps/rejected": -404.0, + "loss": 0.0445, + "rewards/accuracies": 0.984375, + "rewards/chosen": 1.7109375, + "rewards/margins": 11.1875, + "rewards/rejected": -9.4375, + "step": 831 + }, + { + "epoch": 1.0208588957055216, + "grad_norm": 11.737112567032785, + "learning_rate": 9.99793388429752e-07, + "logits/chosen": -0.037353515625, + "logits/rejected": -0.2412109375, + "logps/chosen": -406.0, + "logps/rejected": -388.0, + "loss": 0.0521, + "rewards/accuracies": 0.9765625, + "rewards/chosen": 1.15625, + "rewards/margins": 10.0, + "rewards/rejected": -8.875, + "step": 832 + }, + { + "epoch": 1.0220858895705522, + "grad_norm": 22.2024732224313, + "learning_rate": 9.991735537190084e-07, + "logits/chosen": -0.111328125, + "logits/rejected": -0.26953125, + "logps/chosen": -450.0, + "logps/rejected": -410.0, + "loss": 0.1041, + "rewards/accuracies": 0.9765625, + "rewards/chosen": 1.09375, + "rewards/margins": 10.5, + "rewards/rejected": -9.4375, + "step": 833 + }, + { + "epoch": 1.023312883435583, + "grad_norm": 10.90677573357752, + "learning_rate": 9.985537190082645e-07, + "logits/chosen": 0.024658203125, + "logits/rejected": 0.00970458984375, + "logps/chosen": -356.0, + "logps/rejected": -366.0, + "loss": 0.0368, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.1435546875, + "rewards/margins": 9.0, + "rewards/rejected": -8.875, + "step": 834 + }, + { + "epoch": 1.0245398773006136, + "grad_norm": 12.900773925818388, + "learning_rate": 9.979338842975206e-07, + "logits/chosen": 0.0712890625, + "logits/rejected": -0.14453125, + "logps/chosen": -414.0, + "logps/rejected": -404.0, + "loss": 0.0634, + "rewards/accuracies": 0.9765625, + "rewards/chosen": 0.859375, + "rewards/margins": 10.1875, + "rewards/rejected": -9.3125, + "step": 835 + }, + { + "epoch": 1.0257668711656442, + "grad_norm": 16.38159113775479, + "learning_rate": 9.97314049586777e-07, + "logits/chosen": -0.0419921875, + "logits/rejected": -0.123046875, + "logps/chosen": -376.0, + "logps/rejected": -378.0, + "loss": 0.0545, + "rewards/accuracies": 0.9765625, + "rewards/chosen": 0.703125, + "rewards/margins": 10.4375, + "rewards/rejected": -9.75, + "step": 836 + }, + { + "epoch": 1.026993865030675, + "grad_norm": 5.759934115077334, + "learning_rate": 9.966942148760331e-07, + "logits/chosen": 0.12353515625, + "logits/rejected": -0.13671875, + "logps/chosen": -444.0, + "logps/rejected": -408.0, + "loss": 0.0291, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 1.15625, + "rewards/margins": 10.5625, + "rewards/rejected": -9.4375, + "step": 837 + }, + { + "epoch": 1.0282208588957056, + "grad_norm": 10.475656005242545, + "learning_rate": 9.960743801652893e-07, + "logits/chosen": -0.043212890625, + "logits/rejected": -0.2265625, + "logps/chosen": -394.0, + "logps/rejected": -386.0, + "loss": 0.0452, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.30078125, + "rewards/margins": 10.4375, + "rewards/rejected": -10.125, + "step": 838 + }, + { + "epoch": 1.0294478527607362, + "grad_norm": 16.273519813105544, + "learning_rate": 9.954545454545456e-07, + "logits/chosen": 0.0142822265625, + "logits/rejected": -0.1328125, + "logps/chosen": -398.0, + "logps/rejected": -388.0, + "loss": 0.052, + "rewards/accuracies": 0.96875, + "rewards/chosen": 0.640625, + "rewards/margins": 10.25, + "rewards/rejected": -9.625, + "step": 839 + }, + { + "epoch": 1.030674846625767, + "grad_norm": 15.221579329063106, + "learning_rate": 9.948347107438016e-07, + "logits/chosen": -0.0167236328125, + "logits/rejected": -0.2470703125, + "logps/chosen": -446.0, + "logps/rejected": -422.0, + "loss": 0.0552, + "rewards/accuracies": 0.96875, + "rewards/chosen": 0.91015625, + "rewards/margins": 11.875, + "rewards/rejected": -11.0, + "step": 840 + }, + { + "epoch": 1.0319018404907976, + "grad_norm": 8.194954853669916, + "learning_rate": 9.94214876033058e-07, + "logits/chosen": 0.054443359375, + "logits/rejected": -0.09716796875, + "logps/chosen": -388.0, + "logps/rejected": -396.0, + "loss": 0.042, + "rewards/accuracies": 0.984375, + "rewards/chosen": -0.0072021484375, + "rewards/margins": 10.125, + "rewards/rejected": -10.125, + "step": 841 + }, + { + "epoch": 1.0331288343558283, + "grad_norm": 6.243363966929884, + "learning_rate": 9.93595041322314e-07, + "logits/chosen": -0.03955078125, + "logits/rejected": -0.16015625, + "logps/chosen": -456.0, + "logps/rejected": -364.0, + "loss": 0.0377, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.39453125, + "rewards/margins": 10.0625, + "rewards/rejected": -9.625, + "step": 842 + }, + { + "epoch": 1.034355828220859, + "grad_norm": 13.025481965324344, + "learning_rate": 9.929752066115702e-07, + "logits/chosen": -0.03955078125, + "logits/rejected": -0.1533203125, + "logps/chosen": -460.0, + "logps/rejected": -416.0, + "loss": 0.0486, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.5390625, + "rewards/margins": 10.6875, + "rewards/rejected": -10.125, + "step": 843 + }, + { + "epoch": 1.0355828220858896, + "grad_norm": 9.843030269655861, + "learning_rate": 9.923553719008265e-07, + "logits/chosen": 0.0703125, + "logits/rejected": -0.15625, + "logps/chosen": -434.0, + "logps/rejected": -380.0, + "loss": 0.0284, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 1.1953125, + "rewards/margins": 11.4375, + "rewards/rejected": -10.25, + "step": 844 + }, + { + "epoch": 1.0368098159509203, + "grad_norm": 6.820305865616003, + "learning_rate": 9.917355371900827e-07, + "logits/chosen": -0.00494384765625, + "logits/rejected": -0.11767578125, + "logps/chosen": -374.0, + "logps/rejected": -352.0, + "loss": 0.0243, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.41015625, + "rewards/margins": 9.1875, + "rewards/rejected": -8.75, + "step": 845 + }, + { + "epoch": 1.038036809815951, + "grad_norm": 10.389123582255607, + "learning_rate": 9.911157024793388e-07, + "logits/chosen": 0.1025390625, + "logits/rejected": -0.11181640625, + "logps/chosen": -422.0, + "logps/rejected": -356.0, + "loss": 0.0361, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 1.109375, + "rewards/margins": 11.4375, + "rewards/rejected": -10.3125, + "step": 846 + }, + { + "epoch": 1.0392638036809816, + "grad_norm": 6.3892793464793165, + "learning_rate": 9.904958677685952e-07, + "logits/chosen": 0.07666015625, + "logits/rejected": -0.0830078125, + "logps/chosen": -374.0, + "logps/rejected": -366.0, + "loss": 0.0319, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.63671875, + "rewards/margins": 10.0625, + "rewards/rejected": -9.375, + "step": 847 + }, + { + "epoch": 1.0404907975460123, + "grad_norm": 6.1623878145786435, + "learning_rate": 9.898760330578511e-07, + "logits/chosen": -0.03955078125, + "logits/rejected": -0.1357421875, + "logps/chosen": -354.0, + "logps/rejected": -362.0, + "loss": 0.0272, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.205078125, + "rewards/margins": 10.625, + "rewards/rejected": -10.4375, + "step": 848 + }, + { + "epoch": 1.041717791411043, + "grad_norm": 9.98252857798321, + "learning_rate": 9.892561983471075e-07, + "logits/chosen": 0.01361083984375, + "logits/rejected": -0.05859375, + "logps/chosen": -384.0, + "logps/rejected": -418.0, + "loss": 0.044, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.06396484375, + "rewards/margins": 10.1875, + "rewards/rejected": -10.125, + "step": 849 + }, + { + "epoch": 1.0429447852760736, + "grad_norm": 6.152234316016775, + "learning_rate": 9.886363636363636e-07, + "logits/chosen": -0.0224609375, + "logits/rejected": -0.26171875, + "logps/chosen": -374.0, + "logps/rejected": -368.0, + "loss": 0.0233, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.7734375, + "rewards/margins": 10.875, + "rewards/rejected": -10.125, + "step": 850 + }, + { + "epoch": 1.0441717791411043, + "grad_norm": 5.157350169674604, + "learning_rate": 9.880165289256197e-07, + "logits/chosen": -0.07666015625, + "logits/rejected": -0.275390625, + "logps/chosen": -382.0, + "logps/rejected": -394.0, + "loss": 0.016, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.6875, + "rewards/margins": 11.0, + "rewards/rejected": -10.3125, + "step": 851 + }, + { + "epoch": 1.045398773006135, + "grad_norm": 4.950192991710521, + "learning_rate": 9.87396694214876e-07, + "logits/chosen": -0.138671875, + "logits/rejected": -0.287109375, + "logps/chosen": -432.0, + "logps/rejected": -418.0, + "loss": 0.0193, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.90234375, + "rewards/margins": 10.875, + "rewards/rejected": -10.0, + "step": 852 + }, + { + "epoch": 1.0466257668711656, + "grad_norm": 6.8291284978183375, + "learning_rate": 9.867768595041322e-07, + "logits/chosen": -0.05615234375, + "logits/rejected": -0.15625, + "logps/chosen": -416.0, + "logps/rejected": -372.0, + "loss": 0.0293, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.71875, + "rewards/margins": 10.1875, + "rewards/rejected": -9.4375, + "step": 853 + }, + { + "epoch": 1.0478527607361963, + "grad_norm": 21.523576031133718, + "learning_rate": 9.861570247933884e-07, + "logits/chosen": 0.037109375, + "logits/rejected": -0.09814453125, + "logps/chosen": -396.0, + "logps/rejected": -388.0, + "loss": 0.128, + "rewards/accuracies": 0.9609375, + "rewards/chosen": 0.2265625, + "rewards/margins": 9.25, + "rewards/rejected": -9.0625, + "step": 854 + }, + { + "epoch": 1.049079754601227, + "grad_norm": 7.615435271592111, + "learning_rate": 9.855371900826447e-07, + "logits/chosen": 0.0947265625, + "logits/rejected": -0.045654296875, + "logps/chosen": -398.0, + "logps/rejected": -412.0, + "loss": 0.0403, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 1.3125, + "rewards/margins": 10.8125, + "rewards/rejected": -9.5, + "step": 855 + }, + { + "epoch": 1.0503067484662576, + "grad_norm": 13.390214785568759, + "learning_rate": 9.849173553719009e-07, + "logits/chosen": -0.020751953125, + "logits/rejected": -0.2021484375, + "logps/chosen": -408.0, + "logps/rejected": -364.0, + "loss": 0.0522, + "rewards/accuracies": 0.96875, + "rewards/chosen": 0.89453125, + "rewards/margins": 10.3125, + "rewards/rejected": -9.375, + "step": 856 + }, + { + "epoch": 1.0515337423312883, + "grad_norm": 7.721662927396619, + "learning_rate": 9.84297520661157e-07, + "logits/chosen": -0.1396484375, + "logits/rejected": -0.31640625, + "logps/chosen": -404.0, + "logps/rejected": -382.0, + "loss": 0.0336, + "rewards/accuracies": 0.9765625, + "rewards/chosen": 0.98828125, + "rewards/margins": 10.4375, + "rewards/rejected": -9.5, + "step": 857 + }, + { + "epoch": 1.052760736196319, + "grad_norm": 5.024633893121141, + "learning_rate": 9.836776859504132e-07, + "logits/chosen": -0.008056640625, + "logits/rejected": -0.06884765625, + "logps/chosen": -322.0, + "logps/rejected": -328.0, + "loss": 0.0242, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.2109375, + "rewards/margins": 10.375, + "rewards/rejected": -10.125, + "step": 858 + }, + { + "epoch": 1.0539877300613496, + "grad_norm": 11.331768445922425, + "learning_rate": 9.830578512396693e-07, + "logits/chosen": -0.05029296875, + "logits/rejected": -0.267578125, + "logps/chosen": -400.0, + "logps/rejected": -352.0, + "loss": 0.0499, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.263671875, + "rewards/margins": 10.1875, + "rewards/rejected": -9.9375, + "step": 859 + }, + { + "epoch": 1.0552147239263803, + "grad_norm": 11.938945819408568, + "learning_rate": 9.824380165289257e-07, + "logits/chosen": 0.042724609375, + "logits/rejected": -0.16796875, + "logps/chosen": -416.0, + "logps/rejected": -384.0, + "loss": 0.0571, + "rewards/accuracies": 0.984375, + "rewards/chosen": -0.267578125, + "rewards/margins": 9.125, + "rewards/rejected": -9.375, + "step": 860 + }, + { + "epoch": 1.056441717791411, + "grad_norm": 10.622500605691952, + "learning_rate": 9.818181818181818e-07, + "logits/chosen": -0.00323486328125, + "logits/rejected": -0.1953125, + "logps/chosen": -404.0, + "logps/rejected": -358.0, + "loss": 0.0358, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.65625, + "rewards/margins": 10.25, + "rewards/rejected": -9.5625, + "step": 861 + }, + { + "epoch": 1.0576687116564418, + "grad_norm": 13.987502566864515, + "learning_rate": 9.811983471074381e-07, + "logits/chosen": -0.123046875, + "logits/rejected": -0.2021484375, + "logps/chosen": -386.0, + "logps/rejected": -378.0, + "loss": 0.053, + "rewards/accuracies": 0.9765625, + "rewards/chosen": -0.5078125, + "rewards/margins": 10.0625, + "rewards/rejected": -10.5625, + "step": 862 + }, + { + "epoch": 1.0588957055214725, + "grad_norm": 5.222070055196097, + "learning_rate": 9.805785123966943e-07, + "logits/chosen": -0.12890625, + "logits/rejected": -0.345703125, + "logps/chosen": -428.0, + "logps/rejected": -404.0, + "loss": 0.0159, + "rewards/accuracies": 1.0, + "rewards/chosen": 1.640625, + "rewards/margins": 11.875, + "rewards/rejected": -10.25, + "step": 863 + }, + { + "epoch": 1.0601226993865032, + "grad_norm": 4.4447902099136645, + "learning_rate": 9.799586776859504e-07, + "logits/chosen": -0.051513671875, + "logits/rejected": -0.248046875, + "logps/chosen": -382.0, + "logps/rejected": -390.0, + "loss": 0.022, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.193359375, + "rewards/margins": 11.0, + "rewards/rejected": -10.75, + "step": 864 + }, + { + "epoch": 1.0613496932515338, + "grad_norm": 6.7260760377615, + "learning_rate": 9.793388429752068e-07, + "logits/chosen": -0.2177734375, + "logits/rejected": -0.30859375, + "logps/chosen": -432.0, + "logps/rejected": -370.0, + "loss": 0.0262, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.1982421875, + "rewards/margins": 10.6875, + "rewards/rejected": -10.5, + "step": 865 + }, + { + "epoch": 1.0625766871165645, + "grad_norm": 7.412871849821934, + "learning_rate": 9.787190082644627e-07, + "logits/chosen": -0.166015625, + "logits/rejected": -0.3515625, + "logps/chosen": -356.0, + "logps/rejected": -364.0, + "loss": 0.0325, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.455078125, + "rewards/margins": 10.8125, + "rewards/rejected": -10.375, + "step": 866 + }, + { + "epoch": 1.0638036809815952, + "grad_norm": 12.621207505635246, + "learning_rate": 9.78099173553719e-07, + "logits/chosen": -0.1025390625, + "logits/rejected": -0.251953125, + "logps/chosen": -434.0, + "logps/rejected": -378.0, + "loss": 0.0623, + "rewards/accuracies": 0.9765625, + "rewards/chosen": 0.1123046875, + "rewards/margins": 9.9375, + "rewards/rejected": -9.8125, + "step": 867 + }, + { + "epoch": 1.0650306748466258, + "grad_norm": 6.401402652027187, + "learning_rate": 9.774793388429752e-07, + "logits/chosen": 0.1396484375, + "logits/rejected": -0.06396484375, + "logps/chosen": -358.0, + "logps/rejected": -326.0, + "loss": 0.0388, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.042236328125, + "rewards/margins": 9.1875, + "rewards/rejected": -9.125, + "step": 868 + }, + { + "epoch": 1.0662576687116565, + "grad_norm": 8.548172524646416, + "learning_rate": 9.768595041322313e-07, + "logits/chosen": -0.10205078125, + "logits/rejected": -0.2158203125, + "logps/chosen": -410.0, + "logps/rejected": -394.0, + "loss": 0.0415, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.5, + "rewards/margins": 10.8125, + "rewards/rejected": -10.3125, + "step": 869 + }, + { + "epoch": 1.0674846625766872, + "grad_norm": 10.984226790709622, + "learning_rate": 9.762396694214877e-07, + "logits/chosen": 0.06396484375, + "logits/rejected": -0.11474609375, + "logps/chosen": -390.0, + "logps/rejected": -374.0, + "loss": 0.0438, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.6015625, + "rewards/margins": 10.625, + "rewards/rejected": -10.0, + "step": 870 + }, + { + "epoch": 1.0687116564417178, + "grad_norm": 17.83709695663358, + "learning_rate": 9.756198347107438e-07, + "logits/chosen": 0.0732421875, + "logits/rejected": -0.09130859375, + "logps/chosen": -410.0, + "logps/rejected": -364.0, + "loss": 0.0866, + "rewards/accuracies": 0.96875, + "rewards/chosen": 0.1845703125, + "rewards/margins": 9.375, + "rewards/rejected": -9.25, + "step": 871 + }, + { + "epoch": 1.0699386503067485, + "grad_norm": 6.167620411597598, + "learning_rate": 9.75e-07, + "logits/chosen": 0.03466796875, + "logits/rejected": -0.0859375, + "logps/chosen": -418.0, + "logps/rejected": -338.0, + "loss": 0.0272, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.53125, + "rewards/margins": 10.1875, + "rewards/rejected": -9.625, + "step": 872 + }, + { + "epoch": 1.0711656441717792, + "grad_norm": 7.7202347168828265, + "learning_rate": 9.743801652892563e-07, + "logits/chosen": -0.056640625, + "logits/rejected": -0.21484375, + "logps/chosen": -380.0, + "logps/rejected": -394.0, + "loss": 0.0402, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.6484375, + "rewards/margins": 11.4375, + "rewards/rejected": -10.8125, + "step": 873 + }, + { + "epoch": 1.0723926380368098, + "grad_norm": 20.769443880639034, + "learning_rate": 9.737603305785125e-07, + "logits/chosen": -0.08837890625, + "logits/rejected": -0.2158203125, + "logps/chosen": -436.0, + "logps/rejected": -412.0, + "loss": 0.0993, + "rewards/accuracies": 0.9609375, + "rewards/chosen": 1.1328125, + "rewards/margins": 11.3125, + "rewards/rejected": -10.1875, + "step": 874 + }, + { + "epoch": 1.0736196319018405, + "grad_norm": 4.947185378698119, + "learning_rate": 9.731404958677686e-07, + "logits/chosen": 0.03564453125, + "logits/rejected": -0.11572265625, + "logps/chosen": -354.0, + "logps/rejected": -388.0, + "loss": 0.0241, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.474609375, + "rewards/margins": 10.375, + "rewards/rejected": -9.9375, + "step": 875 + }, + { + "epoch": 1.0748466257668712, + "grad_norm": 5.995360821265753, + "learning_rate": 9.725206611570248e-07, + "logits/chosen": -0.09814453125, + "logits/rejected": -0.279296875, + "logps/chosen": -420.0, + "logps/rejected": -418.0, + "loss": 0.0269, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.443359375, + "rewards/margins": 11.375, + "rewards/rejected": -10.9375, + "step": 876 + }, + { + "epoch": 1.0760736196319018, + "grad_norm": 2.795180251678899, + "learning_rate": 9.71900826446281e-07, + "logits/chosen": 0.0986328125, + "logits/rejected": -0.050048828125, + "logps/chosen": -408.0, + "logps/rejected": -368.0, + "loss": 0.0117, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.0380859375, + "rewards/margins": 9.9375, + "rewards/rejected": -9.9375, + "step": 877 + }, + { + "epoch": 1.0773006134969325, + "grad_norm": 11.795724961832926, + "learning_rate": 9.712809917355373e-07, + "logits/chosen": -0.08056640625, + "logits/rejected": -0.287109375, + "logps/chosen": -384.0, + "logps/rejected": -380.0, + "loss": 0.0589, + "rewards/accuracies": 0.9609375, + "rewards/chosen": -0.21484375, + "rewards/margins": 10.0, + "rewards/rejected": -10.25, + "step": 878 + }, + { + "epoch": 1.0785276073619632, + "grad_norm": 4.020220366948393, + "learning_rate": 9.706611570247934e-07, + "logits/chosen": 0.040283203125, + "logits/rejected": -0.2216796875, + "logps/chosen": -414.0, + "logps/rejected": -362.0, + "loss": 0.0148, + "rewards/accuracies": 1.0, + "rewards/chosen": 1.09375, + "rewards/margins": 11.0625, + "rewards/rejected": -10.0, + "step": 879 + }, + { + "epoch": 1.0797546012269938, + "grad_norm": 7.945458923967855, + "learning_rate": 9.700413223140495e-07, + "logits/chosen": 0.037109375, + "logits/rejected": -0.205078125, + "logps/chosen": -354.0, + "logps/rejected": -354.0, + "loss": 0.0369, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.421875, + "rewards/margins": 10.625, + "rewards/rejected": -11.0625, + "step": 880 + }, + { + "epoch": 1.0809815950920245, + "grad_norm": 7.655504936651129, + "learning_rate": 9.694214876033059e-07, + "logits/chosen": 0.08447265625, + "logits/rejected": -0.055908203125, + "logps/chosen": -350.0, + "logps/rejected": -320.0, + "loss": 0.0397, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.267578125, + "rewards/margins": 8.8125, + "rewards/rejected": -8.5625, + "step": 881 + }, + { + "epoch": 1.0822085889570552, + "grad_norm": 17.14199925290297, + "learning_rate": 9.68801652892562e-07, + "logits/chosen": -0.06787109375, + "logits/rejected": -0.205078125, + "logps/chosen": -402.0, + "logps/rejected": -398.0, + "loss": 0.092, + "rewards/accuracies": 0.9765625, + "rewards/chosen": 1.296875, + "rewards/margins": 10.625, + "rewards/rejected": -9.375, + "step": 882 + }, + { + "epoch": 1.0834355828220859, + "grad_norm": 6.074505313460856, + "learning_rate": 9.681818181818184e-07, + "logits/chosen": -0.11181640625, + "logits/rejected": -0.232421875, + "logps/chosen": -388.0, + "logps/rejected": -364.0, + "loss": 0.0179, + "rewards/accuracies": 1.0, + "rewards/chosen": 1.4375, + "rewards/margins": 9.9375, + "rewards/rejected": -8.5, + "step": 883 + }, + { + "epoch": 1.0846625766871165, + "grad_norm": 5.184873219214592, + "learning_rate": 9.675619834710743e-07, + "logits/chosen": -0.1015625, + "logits/rejected": -0.263671875, + "logps/chosen": -428.0, + "logps/rejected": -410.0, + "loss": 0.015, + "rewards/accuracies": 1.0, + "rewards/chosen": 1.8515625, + "rewards/margins": 11.625, + "rewards/rejected": -9.75, + "step": 884 + }, + { + "epoch": 1.0858895705521472, + "grad_norm": 5.141932376248221, + "learning_rate": 9.669421487603305e-07, + "logits/chosen": -0.0250244140625, + "logits/rejected": -0.16015625, + "logps/chosen": -392.0, + "logps/rejected": -354.0, + "loss": 0.0325, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 1.5, + "rewards/margins": 10.25, + "rewards/rejected": -8.75, + "step": 885 + }, + { + "epoch": 1.0871165644171779, + "grad_norm": 10.51661040179474, + "learning_rate": 9.663223140495868e-07, + "logits/chosen": 0.07666015625, + "logits/rejected": -0.126953125, + "logps/chosen": -384.0, + "logps/rejected": -384.0, + "loss": 0.0416, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.75390625, + "rewards/margins": 9.3125, + "rewards/rejected": -8.5625, + "step": 886 + }, + { + "epoch": 1.0883435582822085, + "grad_norm": 10.214979385844872, + "learning_rate": 9.65702479338843e-07, + "logits/chosen": -0.11376953125, + "logits/rejected": -0.294921875, + "logps/chosen": -402.0, + "logps/rejected": -390.0, + "loss": 0.0423, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.80859375, + "rewards/margins": 10.9375, + "rewards/rejected": -10.125, + "step": 887 + }, + { + "epoch": 1.0895705521472392, + "grad_norm": 12.886011610183632, + "learning_rate": 9.650826446280993e-07, + "logits/chosen": -0.08544921875, + "logits/rejected": -0.357421875, + "logps/chosen": -400.0, + "logps/rejected": -380.0, + "loss": 0.0493, + "rewards/accuracies": 0.9765625, + "rewards/chosen": 1.0859375, + "rewards/margins": 10.5, + "rewards/rejected": -9.4375, + "step": 888 + }, + { + "epoch": 1.0907975460122699, + "grad_norm": 21.84348705309875, + "learning_rate": 9.644628099173554e-07, + "logits/chosen": -0.044677734375, + "logits/rejected": -0.1611328125, + "logps/chosen": -454.0, + "logps/rejected": -420.0, + "loss": 0.0864, + "rewards/accuracies": 0.9609375, + "rewards/chosen": 1.3828125, + "rewards/margins": 10.875, + "rewards/rejected": -9.5, + "step": 889 + }, + { + "epoch": 1.0920245398773005, + "grad_norm": 7.008527769291815, + "learning_rate": 9.638429752066116e-07, + "logits/chosen": -0.0947265625, + "logits/rejected": -0.2158203125, + "logps/chosen": -366.0, + "logps/rejected": -358.0, + "loss": 0.0303, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 1.140625, + "rewards/margins": 10.3125, + "rewards/rejected": -9.125, + "step": 890 + }, + { + "epoch": 1.0932515337423312, + "grad_norm": 9.38763327259464, + "learning_rate": 9.63223140495868e-07, + "logits/chosen": -0.1416015625, + "logits/rejected": -0.28125, + "logps/chosen": -368.0, + "logps/rejected": -386.0, + "loss": 0.0416, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.435546875, + "rewards/margins": 10.5625, + "rewards/rejected": -10.125, + "step": 891 + }, + { + "epoch": 1.0944785276073619, + "grad_norm": 7.190745784431904, + "learning_rate": 9.62603305785124e-07, + "logits/chosen": -0.04248046875, + "logits/rejected": -0.26953125, + "logps/chosen": -358.0, + "logps/rejected": -364.0, + "loss": 0.0387, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.2099609375, + "rewards/margins": 10.5625, + "rewards/rejected": -10.375, + "step": 892 + }, + { + "epoch": 1.0957055214723925, + "grad_norm": 10.096733031161396, + "learning_rate": 9.619834710743802e-07, + "logits/chosen": -0.0869140625, + "logits/rejected": -0.09326171875, + "logps/chosen": -320.0, + "logps/rejected": -332.0, + "loss": 0.0549, + "rewards/accuracies": 0.953125, + "rewards/chosen": 0.54296875, + "rewards/margins": 9.8125, + "rewards/rejected": -9.3125, + "step": 893 + }, + { + "epoch": 1.0969325153374232, + "grad_norm": 15.853564017698188, + "learning_rate": 9.613636363636364e-07, + "logits/chosen": -0.06689453125, + "logits/rejected": -0.28125, + "logps/chosen": -442.0, + "logps/rejected": -416.0, + "loss": 0.0575, + "rewards/accuracies": 0.984375, + "rewards/chosen": 1.1640625, + "rewards/margins": 11.5, + "rewards/rejected": -10.3125, + "step": 894 + }, + { + "epoch": 1.098159509202454, + "grad_norm": 10.254886625086469, + "learning_rate": 9.607438016528925e-07, + "logits/chosen": -0.045166015625, + "logits/rejected": -0.20703125, + "logps/chosen": -388.0, + "logps/rejected": -376.0, + "loss": 0.0352, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 1.515625, + "rewards/margins": 11.0, + "rewards/rejected": -9.4375, + "step": 895 + }, + { + "epoch": 1.0993865030674848, + "grad_norm": 7.1661826290113755, + "learning_rate": 9.601239669421489e-07, + "logits/chosen": -0.09033203125, + "logits/rejected": -0.248046875, + "logps/chosen": -370.0, + "logps/rejected": -374.0, + "loss": 0.0272, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 1.0546875, + "rewards/margins": 10.5, + "rewards/rejected": -9.4375, + "step": 896 + }, + { + "epoch": 1.1006134969325154, + "grad_norm": 13.118106354556259, + "learning_rate": 9.59504132231405e-07, + "logits/chosen": -0.12060546875, + "logits/rejected": -0.357421875, + "logps/chosen": -442.0, + "logps/rejected": -392.0, + "loss": 0.0388, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.90234375, + "rewards/margins": 11.25, + "rewards/rejected": -10.375, + "step": 897 + }, + { + "epoch": 1.101840490797546, + "grad_norm": 7.692840365219393, + "learning_rate": 9.588842975206611e-07, + "logits/chosen": -0.12353515625, + "logits/rejected": -0.33203125, + "logps/chosen": -420.0, + "logps/rejected": -388.0, + "loss": 0.0326, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 1.0859375, + "rewards/margins": 11.4375, + "rewards/rejected": -10.375, + "step": 898 + }, + { + "epoch": 1.1030674846625768, + "grad_norm": 4.182192778148996, + "learning_rate": 9.582644628099175e-07, + "logits/chosen": -0.053466796875, + "logits/rejected": -0.1806640625, + "logps/chosen": -340.0, + "logps/rejected": -348.0, + "loss": 0.0244, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.06298828125, + "rewards/margins": 10.25, + "rewards/rejected": -10.3125, + "step": 899 + }, + { + "epoch": 1.1042944785276074, + "grad_norm": 17.13756577176759, + "learning_rate": 9.576446280991736e-07, + "logits/chosen": 0.0125732421875, + "logits/rejected": -0.1748046875, + "logps/chosen": -404.0, + "logps/rejected": -398.0, + "loss": 0.0925, + "rewards/accuracies": 0.953125, + "rewards/chosen": 0.447265625, + "rewards/margins": 10.1875, + "rewards/rejected": -9.75, + "step": 900 + }, + { + "epoch": 1.105521472392638, + "grad_norm": 6.771958107284894, + "learning_rate": 9.570247933884298e-07, + "logits/chosen": -0.1455078125, + "logits/rejected": -0.36328125, + "logps/chosen": -490.0, + "logps/rejected": -428.0, + "loss": 0.0284, + "rewards/accuracies": 0.984375, + "rewards/chosen": 1.890625, + "rewards/margins": 13.125, + "rewards/rejected": -11.25, + "step": 901 + }, + { + "epoch": 1.1067484662576688, + "grad_norm": 7.061170750215255, + "learning_rate": 9.56404958677686e-07, + "logits/chosen": -0.18359375, + "logits/rejected": -0.341796875, + "logps/chosen": -376.0, + "logps/rejected": -384.0, + "loss": 0.0349, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.5390625, + "rewards/margins": 11.0625, + "rewards/rejected": -10.5625, + "step": 902 + }, + { + "epoch": 1.1079754601226994, + "grad_norm": 8.245692452787445, + "learning_rate": 9.55785123966942e-07, + "logits/chosen": -0.08642578125, + "logits/rejected": -0.30078125, + "logps/chosen": -364.0, + "logps/rejected": -370.0, + "loss": 0.0341, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.75390625, + "rewards/margins": 11.1875, + "rewards/rejected": -10.4375, + "step": 903 + }, + { + "epoch": 1.10920245398773, + "grad_norm": 7.285502821165709, + "learning_rate": 9.551652892561984e-07, + "logits/chosen": -0.09228515625, + "logits/rejected": -0.22265625, + "logps/chosen": -378.0, + "logps/rejected": -416.0, + "loss": 0.0233, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 1.109375, + "rewards/margins": 11.8125, + "rewards/rejected": -10.6875, + "step": 904 + }, + { + "epoch": 1.1104294478527608, + "grad_norm": 11.766334881942175, + "learning_rate": 9.545454545454546e-07, + "logits/chosen": -0.26171875, + "logits/rejected": -0.458984375, + "logps/chosen": -392.0, + "logps/rejected": -392.0, + "loss": 0.0414, + "rewards/accuracies": 0.984375, + "rewards/chosen": 1.453125, + "rewards/margins": 12.0625, + "rewards/rejected": -10.5625, + "step": 905 + }, + { + "epoch": 1.1116564417177914, + "grad_norm": 11.885836553849193, + "learning_rate": 9.539256198347107e-07, + "logits/chosen": -0.1806640625, + "logits/rejected": -0.26171875, + "logps/chosen": -342.0, + "logps/rejected": -382.0, + "loss": 0.0611, + "rewards/accuracies": 0.984375, + "rewards/chosen": -0.337890625, + "rewards/margins": 9.5, + "rewards/rejected": -9.875, + "step": 906 + }, + { + "epoch": 1.112883435582822, + "grad_norm": 8.496822948479098, + "learning_rate": 9.533057851239669e-07, + "logits/chosen": -0.11376953125, + "logits/rejected": -0.388671875, + "logps/chosen": -386.0, + "logps/rejected": -366.0, + "loss": 0.0392, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.4921875, + "rewards/margins": 11.6875, + "rewards/rejected": -11.1875, + "step": 907 + }, + { + "epoch": 1.1141104294478528, + "grad_norm": 4.416563202380208, + "learning_rate": 9.526859504132232e-07, + "logits/chosen": -0.16796875, + "logits/rejected": -0.23828125, + "logps/chosen": -392.0, + "logps/rejected": -380.0, + "loss": 0.0153, + "rewards/accuracies": 1.0, + "rewards/chosen": 1.4453125, + "rewards/margins": 11.5, + "rewards/rejected": -10.0, + "step": 908 + }, + { + "epoch": 1.1153374233128834, + "grad_norm": 12.914705090980133, + "learning_rate": 9.520661157024794e-07, + "logits/chosen": -0.014892578125, + "logits/rejected": -0.1123046875, + "logps/chosen": -418.0, + "logps/rejected": -524.0, + "loss": 0.0577, + "rewards/accuracies": 0.9765625, + "rewards/chosen": 0.7265625, + "rewards/margins": 19.375, + "rewards/rejected": -18.625, + "step": 909 + }, + { + "epoch": 1.116564417177914, + "grad_norm": 11.193369005763312, + "learning_rate": 9.514462809917357e-07, + "logits/chosen": -0.1591796875, + "logits/rejected": -0.3359375, + "logps/chosen": -392.0, + "logps/rejected": -400.0, + "loss": 0.0381, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.9296875, + "rewards/margins": 11.625, + "rewards/rejected": -10.6875, + "step": 910 + }, + { + "epoch": 1.1177914110429448, + "grad_norm": 10.08348423519109, + "learning_rate": 9.508264462809917e-07, + "logits/chosen": -0.109375, + "logits/rejected": -0.3515625, + "logps/chosen": -472.0, + "logps/rejected": -428.0, + "loss": 0.0619, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 1.6953125, + "rewards/margins": 13.375, + "rewards/rejected": -11.6875, + "step": 911 + }, + { + "epoch": 1.1190184049079754, + "grad_norm": 14.853944201654002, + "learning_rate": 9.502066115702479e-07, + "logits/chosen": 0.01153564453125, + "logits/rejected": -0.0849609375, + "logps/chosen": -362.0, + "logps/rejected": -384.0, + "loss": 0.0499, + "rewards/accuracies": 0.9765625, + "rewards/chosen": 0.07080078125, + "rewards/margins": 10.25, + "rewards/rejected": -10.1875, + "step": 912 + }, + { + "epoch": 1.120245398773006, + "grad_norm": 7.41210069844449, + "learning_rate": 9.495867768595041e-07, + "logits/chosen": -0.1845703125, + "logits/rejected": -0.28125, + "logps/chosen": -328.0, + "logps/rejected": -376.0, + "loss": 0.0393, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.56640625, + "rewards/margins": 11.625, + "rewards/rejected": -11.0, + "step": 913 + }, + { + "epoch": 1.1214723926380368, + "grad_norm": 17.23126040997008, + "learning_rate": 9.489669421487604e-07, + "logits/chosen": 0.0013275146484375, + "logits/rejected": -0.1748046875, + "logps/chosen": -412.0, + "logps/rejected": -376.0, + "loss": 0.0818, + "rewards/accuracies": 0.9375, + "rewards/chosen": -0.037109375, + "rewards/margins": 10.125, + "rewards/rejected": -10.125, + "step": 914 + }, + { + "epoch": 1.1226993865030674, + "grad_norm": 7.728386443202459, + "learning_rate": 9.483471074380166e-07, + "logits/chosen": -0.034912109375, + "logits/rejected": -0.291015625, + "logps/chosen": -408.0, + "logps/rejected": -402.0, + "loss": 0.0253, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.81640625, + "rewards/margins": 11.625, + "rewards/rejected": -10.8125, + "step": 915 + }, + { + "epoch": 1.123926380368098, + "grad_norm": 8.320575830299521, + "learning_rate": 9.477272727272727e-07, + "logits/chosen": -0.2001953125, + "logits/rejected": -0.34765625, + "logps/chosen": -394.0, + "logps/rejected": -376.0, + "loss": 0.0332, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.51171875, + "rewards/margins": 11.375, + "rewards/rejected": -10.875, + "step": 916 + }, + { + "epoch": 1.1251533742331288, + "grad_norm": 10.220447898199359, + "learning_rate": 9.47107438016529e-07, + "logits/chosen": -0.1650390625, + "logits/rejected": -0.392578125, + "logps/chosen": -428.0, + "logps/rejected": -430.0, + "loss": 0.0592, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.291015625, + "rewards/margins": 10.9375, + "rewards/rejected": -10.625, + "step": 917 + }, + { + "epoch": 1.1263803680981594, + "grad_norm": 16.327456489125716, + "learning_rate": 9.464876033057852e-07, + "logits/chosen": -0.040771484375, + "logits/rejected": -0.1806640625, + "logps/chosen": -388.0, + "logps/rejected": -384.0, + "loss": 0.0625, + "rewards/accuracies": 0.9609375, + "rewards/chosen": 0.287109375, + "rewards/margins": 10.625, + "rewards/rejected": -10.375, + "step": 918 + }, + { + "epoch": 1.1276073619631901, + "grad_norm": 9.650345793046794, + "learning_rate": 9.458677685950414e-07, + "logits/chosen": -0.2412109375, + "logits/rejected": -0.294921875, + "logps/chosen": -384.0, + "logps/rejected": -408.0, + "loss": 0.0513, + "rewards/accuracies": 0.984375, + "rewards/chosen": -0.44140625, + "rewards/margins": 10.625, + "rewards/rejected": -11.0625, + "step": 919 + }, + { + "epoch": 1.1288343558282208, + "grad_norm": 5.645777430480824, + "learning_rate": 9.452479338842975e-07, + "logits/chosen": -0.07373046875, + "logits/rejected": -0.287109375, + "logps/chosen": -410.0, + "logps/rejected": -396.0, + "loss": 0.0183, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.0281982421875, + "rewards/margins": 11.0, + "rewards/rejected": -11.0, + "step": 920 + }, + { + "epoch": 1.1300613496932514, + "grad_norm": 6.67262482291788, + "learning_rate": 9.446280991735537e-07, + "logits/chosen": -0.08935546875, + "logits/rejected": -0.2421875, + "logps/chosen": -398.0, + "logps/rejected": -396.0, + "loss": 0.0236, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 1.078125, + "rewards/margins": 11.6875, + "rewards/rejected": -10.5625, + "step": 921 + }, + { + "epoch": 1.1312883435582821, + "grad_norm": 6.984229781332186, + "learning_rate": 9.440082644628099e-07, + "logits/chosen": -0.036865234375, + "logits/rejected": -0.15234375, + "logps/chosen": -392.0, + "logps/rejected": -402.0, + "loss": 0.0334, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.98046875, + "rewards/margins": 10.9375, + "rewards/rejected": -9.9375, + "step": 922 + }, + { + "epoch": 1.132515337423313, + "grad_norm": 12.662651710075986, + "learning_rate": 9.433884297520662e-07, + "logits/chosen": 0.005828857421875, + "logits/rejected": -0.08349609375, + "logps/chosen": -364.0, + "logps/rejected": -378.0, + "loss": 0.0303, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.9609375, + "rewards/margins": 10.5625, + "rewards/rejected": -9.625, + "step": 923 + }, + { + "epoch": 1.1337423312883437, + "grad_norm": 9.002647218488805, + "learning_rate": 9.427685950413223e-07, + "logits/chosen": -0.142578125, + "logits/rejected": -0.2265625, + "logps/chosen": -384.0, + "logps/rejected": -368.0, + "loss": 0.0351, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.91015625, + "rewards/margins": 11.25, + "rewards/rejected": -10.375, + "step": 924 + }, + { + "epoch": 1.1349693251533743, + "grad_norm": 12.813399120951164, + "learning_rate": 9.421487603305785e-07, + "logits/chosen": -0.203125, + "logits/rejected": -0.404296875, + "logps/chosen": -378.0, + "logps/rejected": -364.0, + "loss": 0.0329, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.0306396484375, + "rewards/margins": 10.875, + "rewards/rejected": -10.8125, + "step": 925 + }, + { + "epoch": 1.136196319018405, + "grad_norm": 13.681237077440016, + "learning_rate": 9.415289256198348e-07, + "logits/chosen": -0.2138671875, + "logits/rejected": -0.384765625, + "logps/chosen": -426.0, + "logps/rejected": -410.0, + "loss": 0.0828, + "rewards/accuracies": 0.9765625, + "rewards/chosen": 0.9921875, + "rewards/margins": 11.1875, + "rewards/rejected": -10.1875, + "step": 926 + }, + { + "epoch": 1.1374233128834357, + "grad_norm": 14.007812307598066, + "learning_rate": 9.40909090909091e-07, + "logits/chosen": -0.091796875, + "logits/rejected": -0.2431640625, + "logps/chosen": -426.0, + "logps/rejected": -406.0, + "loss": 0.0454, + "rewards/accuracies": 0.9765625, + "rewards/chosen": 0.921875, + "rewards/margins": 11.0625, + "rewards/rejected": -10.1875, + "step": 927 + }, + { + "epoch": 1.1386503067484663, + "grad_norm": 5.804375472545515, + "learning_rate": 9.402892561983472e-07, + "logits/chosen": -0.06982421875, + "logits/rejected": -0.2734375, + "logps/chosen": -416.0, + "logps/rejected": -398.0, + "loss": 0.0282, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.67578125, + "rewards/margins": 10.8125, + "rewards/rejected": -10.125, + "step": 928 + }, + { + "epoch": 1.139877300613497, + "grad_norm": 7.0706705773655205, + "learning_rate": 9.396694214876032e-07, + "logits/chosen": -0.038330078125, + "logits/rejected": -0.212890625, + "logps/chosen": -430.0, + "logps/rejected": -348.0, + "loss": 0.0216, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.90625, + "rewards/margins": 10.75, + "rewards/rejected": -9.8125, + "step": 929 + }, + { + "epoch": 1.1411042944785277, + "grad_norm": 20.03419906761046, + "learning_rate": 9.390495867768595e-07, + "logits/chosen": -0.0673828125, + "logits/rejected": -0.302734375, + "logps/chosen": -408.0, + "logps/rejected": -392.0, + "loss": 0.0678, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.546875, + "rewards/margins": 11.5, + "rewards/rejected": -10.9375, + "step": 930 + }, + { + "epoch": 1.1423312883435583, + "grad_norm": 8.79614969540993, + "learning_rate": 9.384297520661157e-07, + "logits/chosen": -0.0712890625, + "logits/rejected": -0.154296875, + "logps/chosen": -412.0, + "logps/rejected": -376.0, + "loss": 0.0392, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.041748046875, + "rewards/margins": 9.5, + "rewards/rejected": -9.5625, + "step": 931 + }, + { + "epoch": 1.143558282208589, + "grad_norm": 5.533709744700053, + "learning_rate": 9.37809917355372e-07, + "logits/chosen": -0.1923828125, + "logits/rejected": -0.3671875, + "logps/chosen": -380.0, + "logps/rejected": -380.0, + "loss": 0.0302, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.045166015625, + "rewards/margins": 11.125, + "rewards/rejected": -11.0625, + "step": 932 + }, + { + "epoch": 1.1447852760736197, + "grad_norm": 17.392392128691636, + "learning_rate": 9.371900826446281e-07, + "logits/chosen": -0.04052734375, + "logits/rejected": -0.1318359375, + "logps/chosen": -372.0, + "logps/rejected": -400.0, + "loss": 0.0582, + "rewards/accuracies": 0.9765625, + "rewards/chosen": 0.4296875, + "rewards/margins": 10.3125, + "rewards/rejected": -9.875, + "step": 933 + }, + { + "epoch": 1.1460122699386504, + "grad_norm": 11.220639802991899, + "learning_rate": 9.365702479338843e-07, + "logits/chosen": -0.076171875, + "logits/rejected": -0.23828125, + "logps/chosen": -386.0, + "logps/rejected": -366.0, + "loss": 0.0466, + "rewards/accuracies": 0.9765625, + "rewards/chosen": 0.94140625, + "rewards/margins": 10.25, + "rewards/rejected": -9.3125, + "step": 934 + }, + { + "epoch": 1.147239263803681, + "grad_norm": 3.127170097265917, + "learning_rate": 9.359504132231406e-07, + "logits/chosen": -0.125, + "logits/rejected": -0.26171875, + "logps/chosen": -416.0, + "logps/rejected": -408.0, + "loss": 0.0156, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.703125, + "rewards/margins": 10.6875, + "rewards/rejected": -10.0, + "step": 935 + }, + { + "epoch": 1.1484662576687117, + "grad_norm": 8.797510711759905, + "learning_rate": 9.353305785123968e-07, + "logits/chosen": 0.04833984375, + "logits/rejected": -0.1953125, + "logps/chosen": -366.0, + "logps/rejected": -382.0, + "loss": 0.0399, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.0927734375, + "rewards/margins": 10.5625, + "rewards/rejected": -10.6875, + "step": 936 + }, + { + "epoch": 1.1496932515337424, + "grad_norm": 8.508479799139819, + "learning_rate": 9.347107438016529e-07, + "logits/chosen": -0.032958984375, + "logits/rejected": -0.267578125, + "logps/chosen": -434.0, + "logps/rejected": -406.0, + "loss": 0.0222, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.255859375, + "rewards/margins": 11.0625, + "rewards/rejected": -10.8125, + "step": 937 + }, + { + "epoch": 1.150920245398773, + "grad_norm": 11.734844982011332, + "learning_rate": 9.34090909090909e-07, + "logits/chosen": -0.01031494140625, + "logits/rejected": -0.058837890625, + "logps/chosen": -374.0, + "logps/rejected": -378.0, + "loss": 0.0576, + "rewards/accuracies": 0.9765625, + "rewards/chosen": 0.294921875, + "rewards/margins": 10.6875, + "rewards/rejected": -10.375, + "step": 938 + }, + { + "epoch": 1.1521472392638037, + "grad_norm": 6.686498950384566, + "learning_rate": 9.334710743801653e-07, + "logits/chosen": -0.1796875, + "logits/rejected": -0.318359375, + "logps/chosen": -450.0, + "logps/rejected": -420.0, + "loss": 0.022, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 1.0703125, + "rewards/margins": 12.0625, + "rewards/rejected": -11.0, + "step": 939 + }, + { + "epoch": 1.1533742331288344, + "grad_norm": 9.333043593563687, + "learning_rate": 9.328512396694215e-07, + "logits/chosen": -0.169921875, + "logits/rejected": -0.361328125, + "logps/chosen": -420.0, + "logps/rejected": -388.0, + "loss": 0.0298, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 1.0234375, + "rewards/margins": 12.125, + "rewards/rejected": -11.125, + "step": 940 + }, + { + "epoch": 1.154601226993865, + "grad_norm": 13.328010603903623, + "learning_rate": 9.322314049586778e-07, + "logits/chosen": -0.0150146484375, + "logits/rejected": -0.220703125, + "logps/chosen": -418.0, + "logps/rejected": -390.0, + "loss": 0.058, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 1.25, + "rewards/margins": 10.9375, + "rewards/rejected": -9.6875, + "step": 941 + }, + { + "epoch": 1.1558282208588957, + "grad_norm": 10.748405478390469, + "learning_rate": 9.316115702479339e-07, + "logits/chosen": -0.2431640625, + "logits/rejected": -0.380859375, + "logps/chosen": -418.0, + "logps/rejected": -406.0, + "loss": 0.0373, + "rewards/accuracies": 0.984375, + "rewards/chosen": 1.2109375, + "rewards/margins": 12.25, + "rewards/rejected": -11.0625, + "step": 942 + }, + { + "epoch": 1.1570552147239264, + "grad_norm": 6.66797822267693, + "learning_rate": 9.309917355371901e-07, + "logits/chosen": 0.0140380859375, + "logits/rejected": -0.1474609375, + "logps/chosen": -354.0, + "logps/rejected": -350.0, + "loss": 0.021, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.359375, + "rewards/margins": 10.25, + "rewards/rejected": -9.875, + "step": 943 + }, + { + "epoch": 1.158282208588957, + "grad_norm": 6.910993717404971, + "learning_rate": 9.303719008264464e-07, + "logits/chosen": -0.0791015625, + "logits/rejected": -0.2490234375, + "logps/chosen": -404.0, + "logps/rejected": -412.0, + "loss": 0.0213, + "rewards/accuracies": 1.0, + "rewards/chosen": 1.1171875, + "rewards/margins": 11.5625, + "rewards/rejected": -10.4375, + "step": 944 + }, + { + "epoch": 1.1595092024539877, + "grad_norm": 8.94698116889764, + "learning_rate": 9.297520661157025e-07, + "logits/chosen": -0.1640625, + "logits/rejected": -0.33984375, + "logps/chosen": -416.0, + "logps/rejected": -380.0, + "loss": 0.0401, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 1.390625, + "rewards/margins": 10.75, + "rewards/rejected": -9.375, + "step": 945 + }, + { + "epoch": 1.1607361963190184, + "grad_norm": 5.757411937782045, + "learning_rate": 9.291322314049587e-07, + "logits/chosen": -0.0016937255859375, + "logits/rejected": -0.07177734375, + "logps/chosen": -380.0, + "logps/rejected": -400.0, + "loss": 0.0224, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.291015625, + "rewards/margins": 9.8125, + "rewards/rejected": -9.5625, + "step": 946 + }, + { + "epoch": 1.161963190184049, + "grad_norm": 10.946703026968246, + "learning_rate": 9.285123966942148e-07, + "logits/chosen": -0.2138671875, + "logits/rejected": -0.3984375, + "logps/chosen": -346.0, + "logps/rejected": -368.0, + "loss": 0.0589, + "rewards/accuracies": 0.9765625, + "rewards/chosen": 0.7734375, + "rewards/margins": 10.125, + "rewards/rejected": -9.375, + "step": 947 + }, + { + "epoch": 1.1631901840490797, + "grad_norm": 13.673425770565624, + "learning_rate": 9.278925619834711e-07, + "logits/chosen": -0.1962890625, + "logits/rejected": -0.427734375, + "logps/chosen": -428.0, + "logps/rejected": -380.0, + "loss": 0.0544, + "rewards/accuracies": 0.984375, + "rewards/chosen": 1.5, + "rewards/margins": 10.6875, + "rewards/rejected": -9.1875, + "step": 948 + }, + { + "epoch": 1.1644171779141104, + "grad_norm": 10.764489540205084, + "learning_rate": 9.272727272727273e-07, + "logits/chosen": 0.0032958984375, + "logits/rejected": -0.265625, + "logps/chosen": -414.0, + "logps/rejected": -410.0, + "loss": 0.0627, + "rewards/accuracies": 0.96875, + "rewards/chosen": 1.46875, + "rewards/margins": 11.625, + "rewards/rejected": -10.125, + "step": 949 + }, + { + "epoch": 1.165644171779141, + "grad_norm": 6.163510999316528, + "learning_rate": 9.266528925619834e-07, + "logits/chosen": -0.134765625, + "logits/rejected": -0.23828125, + "logps/chosen": -338.0, + "logps/rejected": -336.0, + "loss": 0.0233, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.265625, + "rewards/margins": 10.5, + "rewards/rejected": -10.25, + "step": 950 + }, + { + "epoch": 1.1668711656441717, + "grad_norm": 7.49094764434787, + "learning_rate": 9.260330578512397e-07, + "logits/chosen": -0.021240234375, + "logits/rejected": -0.146484375, + "logps/chosen": -350.0, + "logps/rejected": -354.0, + "loss": 0.0353, + "rewards/accuracies": 0.9765625, + "rewards/chosen": 0.259765625, + "rewards/margins": 9.9375, + "rewards/rejected": -9.6875, + "step": 951 + }, + { + "epoch": 1.1680981595092024, + "grad_norm": 11.812912807851047, + "learning_rate": 9.254132231404959e-07, + "logits/chosen": -0.2197265625, + "logits/rejected": -0.494140625, + "logps/chosen": -424.0, + "logps/rejected": -386.0, + "loss": 0.0565, + "rewards/accuracies": 0.984375, + "rewards/chosen": 1.09375, + "rewards/margins": 11.1875, + "rewards/rejected": -10.0625, + "step": 952 + }, + { + "epoch": 1.169325153374233, + "grad_norm": 9.899530945112486, + "learning_rate": 9.247933884297522e-07, + "logits/chosen": -0.07080078125, + "logits/rejected": -0.271484375, + "logps/chosen": -388.0, + "logps/rejected": -360.0, + "loss": 0.038, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.96484375, + "rewards/margins": 10.5625, + "rewards/rejected": -9.5625, + "step": 953 + }, + { + "epoch": 1.1705521472392637, + "grad_norm": 14.874148700852642, + "learning_rate": 9.241735537190083e-07, + "logits/chosen": -0.0771484375, + "logits/rejected": -0.283203125, + "logps/chosen": -322.0, + "logps/rejected": -320.0, + "loss": 0.0591, + "rewards/accuracies": 0.96875, + "rewards/chosen": -0.439453125, + "rewards/margins": 9.8125, + "rewards/rejected": -10.25, + "step": 954 + }, + { + "epoch": 1.1717791411042944, + "grad_norm": 14.69776784288544, + "learning_rate": 9.235537190082644e-07, + "logits/chosen": -0.005157470703125, + "logits/rejected": -0.0771484375, + "logps/chosen": -380.0, + "logps/rejected": -378.0, + "loss": 0.0538, + "rewards/accuracies": 0.984375, + "rewards/chosen": 1.015625, + "rewards/margins": 10.8125, + "rewards/rejected": -9.8125, + "step": 955 + }, + { + "epoch": 1.173006134969325, + "grad_norm": 8.526200653171099, + "learning_rate": 9.229338842975206e-07, + "logits/chosen": -0.10009765625, + "logits/rejected": -0.2041015625, + "logps/chosen": -316.0, + "logps/rejected": -344.0, + "loss": 0.0306, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.30859375, + "rewards/margins": 9.75, + "rewards/rejected": -9.4375, + "step": 956 + }, + { + "epoch": 1.1742331288343557, + "grad_norm": 7.973457405245253, + "learning_rate": 9.223140495867769e-07, + "logits/chosen": -0.134765625, + "logits/rejected": -0.357421875, + "logps/chosen": -390.0, + "logps/rejected": -404.0, + "loss": 0.0368, + "rewards/accuracies": 0.984375, + "rewards/chosen": 1.0859375, + "rewards/margins": 10.875, + "rewards/rejected": -9.8125, + "step": 957 + }, + { + "epoch": 1.1754601226993864, + "grad_norm": 14.106346468292243, + "learning_rate": 9.216942148760331e-07, + "logits/chosen": -0.055908203125, + "logits/rejected": -0.15625, + "logps/chosen": -364.0, + "logps/rejected": -404.0, + "loss": 0.0479, + "rewards/accuracies": 0.9765625, + "rewards/chosen": 0.404296875, + "rewards/margins": 11.0, + "rewards/rejected": -10.625, + "step": 958 + }, + { + "epoch": 1.1766871165644173, + "grad_norm": 11.510149735049389, + "learning_rate": 9.210743801652892e-07, + "logits/chosen": -0.095703125, + "logits/rejected": -0.267578125, + "logps/chosen": -396.0, + "logps/rejected": -388.0, + "loss": 0.0431, + "rewards/accuracies": 0.9765625, + "rewards/chosen": 0.63671875, + "rewards/margins": 10.8125, + "rewards/rejected": -10.1875, + "step": 959 + }, + { + "epoch": 1.177914110429448, + "grad_norm": 3.8552096147370842, + "learning_rate": 9.204545454545455e-07, + "logits/chosen": -0.1728515625, + "logits/rejected": -0.330078125, + "logps/chosen": -346.0, + "logps/rejected": -350.0, + "loss": 0.0149, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.546875, + "rewards/margins": 10.25, + "rewards/rejected": -10.75, + "step": 960 + }, + { + "epoch": 1.1791411042944786, + "grad_norm": 13.533375661298507, + "learning_rate": 9.198347107438017e-07, + "logits/chosen": -0.12451171875, + "logits/rejected": -0.2353515625, + "logps/chosen": -404.0, + "logps/rejected": -380.0, + "loss": 0.0407, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.85546875, + "rewards/margins": 11.5, + "rewards/rejected": -10.625, + "step": 961 + }, + { + "epoch": 1.1803680981595093, + "grad_norm": 9.63092706255587, + "learning_rate": 9.192148760330579e-07, + "logits/chosen": 0.00921630859375, + "logits/rejected": -0.16015625, + "logps/chosen": -440.0, + "logps/rejected": -430.0, + "loss": 0.0318, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.609375, + "rewards/margins": 11.0, + "rewards/rejected": -10.375, + "step": 962 + }, + { + "epoch": 1.18159509202454, + "grad_norm": 13.10088160617771, + "learning_rate": 9.185950413223141e-07, + "logits/chosen": -0.11474609375, + "logits/rejected": -0.306640625, + "logps/chosen": -378.0, + "logps/rejected": -352.0, + "loss": 0.0397, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.61328125, + "rewards/margins": 10.875, + "rewards/rejected": -10.25, + "step": 963 + }, + { + "epoch": 1.1828220858895706, + "grad_norm": 7.358279878141561, + "learning_rate": 9.179752066115702e-07, + "logits/chosen": -0.049560546875, + "logits/rejected": -0.2314453125, + "logps/chosen": -380.0, + "logps/rejected": -368.0, + "loss": 0.0233, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.69921875, + "rewards/margins": 11.375, + "rewards/rejected": -10.6875, + "step": 964 + }, + { + "epoch": 1.1840490797546013, + "grad_norm": 10.963522347339602, + "learning_rate": 9.173553719008264e-07, + "logits/chosen": -0.12158203125, + "logits/rejected": -0.255859375, + "logps/chosen": -366.0, + "logps/rejected": -386.0, + "loss": 0.0329, + "rewards/accuracies": 0.984375, + "rewards/chosen": -0.287109375, + "rewards/margins": 10.75, + "rewards/rejected": -11.0625, + "step": 965 + }, + { + "epoch": 1.185276073619632, + "grad_norm": 11.155993123724087, + "learning_rate": 9.167355371900827e-07, + "logits/chosen": -0.1875, + "logits/rejected": -0.322265625, + "logps/chosen": -396.0, + "logps/rejected": -360.0, + "loss": 0.0412, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.00421142578125, + "rewards/margins": 10.6875, + "rewards/rejected": -10.625, + "step": 966 + }, + { + "epoch": 1.1865030674846626, + "grad_norm": 7.534460905603958, + "learning_rate": 9.161157024793388e-07, + "logits/chosen": -0.11279296875, + "logits/rejected": -0.28515625, + "logps/chosen": -412.0, + "logps/rejected": -378.0, + "loss": 0.027, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.95703125, + "rewards/margins": 11.4375, + "rewards/rejected": -10.5, + "step": 967 + }, + { + "epoch": 1.1877300613496933, + "grad_norm": 17.6032434079415, + "learning_rate": 9.15495867768595e-07, + "logits/chosen": -0.041015625, + "logits/rejected": -0.224609375, + "logps/chosen": -424.0, + "logps/rejected": -396.0, + "loss": 0.0462, + "rewards/accuracies": 0.9765625, + "rewards/chosen": 0.875, + "rewards/margins": 12.25, + "rewards/rejected": -11.375, + "step": 968 + }, + { + "epoch": 1.188957055214724, + "grad_norm": 15.694096678580106, + "learning_rate": 9.148760330578513e-07, + "logits/chosen": -0.1455078125, + "logits/rejected": -0.3046875, + "logps/chosen": -388.0, + "logps/rejected": -422.0, + "loss": 0.0629, + "rewards/accuracies": 0.96875, + "rewards/chosen": -1.390625, + "rewards/margins": 10.0, + "rewards/rejected": -11.375, + "step": 969 + }, + { + "epoch": 1.1901840490797546, + "grad_norm": 11.921476909261088, + "learning_rate": 9.142561983471075e-07, + "logits/chosen": -0.10595703125, + "logits/rejected": -0.2451171875, + "logps/chosen": -372.0, + "logps/rejected": -502.0, + "loss": 0.0453, + "rewards/accuracies": 0.984375, + "rewards/chosen": -0.01531982421875, + "rewards/margins": 23.125, + "rewards/rejected": -23.125, + "step": 970 + }, + { + "epoch": 1.1914110429447853, + "grad_norm": 13.62660815818298, + "learning_rate": 9.136363636363637e-07, + "logits/chosen": -0.0091552734375, + "logits/rejected": -0.25, + "logps/chosen": -358.0, + "logps/rejected": -356.0, + "loss": 0.0572, + "rewards/accuracies": 0.984375, + "rewards/chosen": -0.09375, + "rewards/margins": 10.5625, + "rewards/rejected": -10.625, + "step": 971 + }, + { + "epoch": 1.192638036809816, + "grad_norm": 7.237399332254699, + "learning_rate": 9.130165289256199e-07, + "logits/chosen": -0.057373046875, + "logits/rejected": -0.1982421875, + "logps/chosen": -342.0, + "logps/rejected": -312.0, + "loss": 0.0381, + "rewards/accuracies": 0.984375, + "rewards/chosen": -0.3515625, + "rewards/margins": 9.5625, + "rewards/rejected": -9.875, + "step": 972 + }, + { + "epoch": 1.1938650306748466, + "grad_norm": 9.089933594933969, + "learning_rate": 9.12396694214876e-07, + "logits/chosen": -0.1533203125, + "logits/rejected": -0.279296875, + "logps/chosen": -412.0, + "logps/rejected": -432.0, + "loss": 0.029, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.7109375, + "rewards/margins": 11.625, + "rewards/rejected": -10.875, + "step": 973 + }, + { + "epoch": 1.1950920245398773, + "grad_norm": 3.7486390717851217, + "learning_rate": 9.117768595041322e-07, + "logits/chosen": -0.2275390625, + "logits/rejected": -0.41796875, + "logps/chosen": -398.0, + "logps/rejected": -406.0, + "loss": 0.0146, + "rewards/accuracies": 1.0, + "rewards/chosen": 1.09375, + "rewards/margins": 11.9375, + "rewards/rejected": -10.8125, + "step": 974 + }, + { + "epoch": 1.196319018404908, + "grad_norm": 17.692363179531846, + "learning_rate": 9.111570247933885e-07, + "logits/chosen": 0.045654296875, + "logits/rejected": -0.07861328125, + "logps/chosen": -360.0, + "logps/rejected": -332.0, + "loss": 0.0651, + "rewards/accuracies": 0.9765625, + "rewards/chosen": 0.2265625, + "rewards/margins": 9.9375, + "rewards/rejected": -9.6875, + "step": 975 + }, + { + "epoch": 1.1975460122699386, + "grad_norm": 13.042080391360054, + "learning_rate": 9.105371900826446e-07, + "logits/chosen": -0.173828125, + "logits/rejected": -0.392578125, + "logps/chosen": -422.0, + "logps/rejected": -390.0, + "loss": 0.0488, + "rewards/accuracies": 0.984375, + "rewards/chosen": 1.0078125, + "rewards/margins": 10.4375, + "rewards/rejected": -9.4375, + "step": 976 + }, + { + "epoch": 1.1987730061349693, + "grad_norm": 17.283449726274785, + "learning_rate": 9.099173553719009e-07, + "logits/chosen": -0.06787109375, + "logits/rejected": -0.267578125, + "logps/chosen": -388.0, + "logps/rejected": -374.0, + "loss": 0.083, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.91796875, + "rewards/margins": 11.5, + "rewards/rejected": -10.5625, + "step": 977 + }, + { + "epoch": 1.2, + "grad_norm": 10.327665992150298, + "learning_rate": 9.092975206611571e-07, + "logits/chosen": -0.11865234375, + "logits/rejected": -0.1416015625, + "logps/chosen": -364.0, + "logps/rejected": -344.0, + "loss": 0.0616, + "rewards/accuracies": 0.96875, + "rewards/chosen": -0.046630859375, + "rewards/margins": 10.125, + "rewards/rejected": -10.125, + "step": 978 + }, + { + "epoch": 1.2012269938650306, + "grad_norm": 6.639005543404221, + "learning_rate": 9.086776859504132e-07, + "logits/chosen": -0.00439453125, + "logits/rejected": -0.126953125, + "logps/chosen": -410.0, + "logps/rejected": -384.0, + "loss": 0.0294, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.76953125, + "rewards/margins": 11.25, + "rewards/rejected": -10.4375, + "step": 979 + }, + { + "epoch": 1.2024539877300613, + "grad_norm": 3.7048596375075307, + "learning_rate": 9.080578512396695e-07, + "logits/chosen": -0.173828125, + "logits/rejected": -0.35546875, + "logps/chosen": -362.0, + "logps/rejected": -374.0, + "loss": 0.0142, + "rewards/accuracies": 1.0, + "rewards/chosen": 1.6484375, + "rewards/margins": 11.625, + "rewards/rejected": -10.0, + "step": 980 + }, + { + "epoch": 1.203680981595092, + "grad_norm": 9.842884355659873, + "learning_rate": 9.074380165289257e-07, + "logits/chosen": -0.0908203125, + "logits/rejected": -0.1796875, + "logps/chosen": -386.0, + "logps/rejected": -386.0, + "loss": 0.0228, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.8671875, + "rewards/margins": 10.9375, + "rewards/rejected": -10.0625, + "step": 981 + }, + { + "epoch": 1.2049079754601226, + "grad_norm": 8.509249883161877, + "learning_rate": 9.068181818181818e-07, + "logits/chosen": -0.09130859375, + "logits/rejected": -0.267578125, + "logps/chosen": -408.0, + "logps/rejected": -408.0, + "loss": 0.0327, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.734375, + "rewards/margins": 10.75, + "rewards/rejected": -10.0, + "step": 982 + }, + { + "epoch": 1.2061349693251533, + "grad_norm": 17.32331392424542, + "learning_rate": 9.06198347107438e-07, + "logits/chosen": -0.09814453125, + "logits/rejected": -0.28125, + "logps/chosen": -382.0, + "logps/rejected": -346.0, + "loss": 0.0823, + "rewards/accuracies": 0.953125, + "rewards/chosen": 0.671875, + "rewards/margins": 9.625, + "rewards/rejected": -9.0, + "step": 983 + }, + { + "epoch": 1.207361963190184, + "grad_norm": 6.308837245908633, + "learning_rate": 9.055785123966942e-07, + "logits/chosen": -0.1328125, + "logits/rejected": -0.310546875, + "logps/chosen": -362.0, + "logps/rejected": -346.0, + "loss": 0.0285, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.447265625, + "rewards/margins": 10.3125, + "rewards/rejected": -9.875, + "step": 984 + }, + { + "epoch": 1.2085889570552146, + "grad_norm": 11.159178783645137, + "learning_rate": 9.049586776859504e-07, + "logits/chosen": -0.04443359375, + "logits/rejected": -0.212890625, + "logps/chosen": -350.0, + "logps/rejected": -352.0, + "loss": 0.0427, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.76171875, + "rewards/margins": 10.3125, + "rewards/rejected": -9.5625, + "step": 985 + }, + { + "epoch": 1.2098159509202455, + "grad_norm": 9.051373556481083, + "learning_rate": 9.043388429752067e-07, + "logits/chosen": 0.0625, + "logits/rejected": -0.08935546875, + "logps/chosen": -390.0, + "logps/rejected": -356.0, + "loss": 0.0347, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 1.0546875, + "rewards/margins": 10.0, + "rewards/rejected": -8.9375, + "step": 986 + }, + { + "epoch": 1.2110429447852762, + "grad_norm": 6.198302604249075, + "learning_rate": 9.037190082644629e-07, + "logits/chosen": -0.1435546875, + "logits/rejected": -0.390625, + "logps/chosen": -458.0, + "logps/rejected": -406.0, + "loss": 0.0282, + "rewards/accuracies": 0.984375, + "rewards/chosen": 1.28125, + "rewards/margins": 11.1875, + "rewards/rejected": -9.875, + "step": 987 + }, + { + "epoch": 1.2122699386503069, + "grad_norm": 4.102469498417344, + "learning_rate": 9.03099173553719e-07, + "logits/chosen": -0.047607421875, + "logits/rejected": -0.26171875, + "logps/chosen": -372.0, + "logps/rejected": -348.0, + "loss": 0.0143, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.240234375, + "rewards/margins": 10.5625, + "rewards/rejected": -10.3125, + "step": 988 + }, + { + "epoch": 1.2134969325153375, + "grad_norm": 8.46030603702519, + "learning_rate": 9.024793388429753e-07, + "logits/chosen": -0.0023345947265625, + "logits/rejected": -0.16796875, + "logps/chosen": -354.0, + "logps/rejected": -342.0, + "loss": 0.0404, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.21484375, + "rewards/margins": 9.4375, + "rewards/rejected": -9.6875, + "step": 989 + }, + { + "epoch": 1.2147239263803682, + "grad_norm": 10.249236383125274, + "learning_rate": 9.018595041322315e-07, + "logits/chosen": -0.1494140625, + "logits/rejected": -0.326171875, + "logps/chosen": -402.0, + "logps/rejected": -396.0, + "loss": 0.0387, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.58984375, + "rewards/margins": 11.5, + "rewards/rejected": -10.875, + "step": 990 + }, + { + "epoch": 1.2159509202453989, + "grad_norm": 20.10350249363253, + "learning_rate": 9.012396694214876e-07, + "logits/chosen": -0.09130859375, + "logits/rejected": -0.2197265625, + "logps/chosen": -416.0, + "logps/rejected": -408.0, + "loss": 0.0671, + "rewards/accuracies": 0.96875, + "rewards/chosen": -0.263671875, + "rewards/margins": 11.0, + "rewards/rejected": -11.25, + "step": 991 + }, + { + "epoch": 1.2171779141104295, + "grad_norm": 15.353242393368603, + "learning_rate": 9.006198347107438e-07, + "logits/chosen": -0.11572265625, + "logits/rejected": -0.302734375, + "logps/chosen": -388.0, + "logps/rejected": -392.0, + "loss": 0.067, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.11962890625, + "rewards/margins": 10.875, + "rewards/rejected": -10.75, + "step": 992 + }, + { + "epoch": 1.2184049079754602, + "grad_norm": 14.594366015899684, + "learning_rate": 9e-07, + "logits/chosen": -0.1689453125, + "logits/rejected": -0.314453125, + "logps/chosen": -376.0, + "logps/rejected": -376.0, + "loss": 0.0584, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.7265625, + "rewards/margins": 10.8125, + "rewards/rejected": -10.0625, + "step": 993 + }, + { + "epoch": 1.2196319018404909, + "grad_norm": 17.819835095986992, + "learning_rate": 8.993801652892562e-07, + "logits/chosen": -0.169921875, + "logits/rejected": -0.341796875, + "logps/chosen": -428.0, + "logps/rejected": -408.0, + "loss": 0.0555, + "rewards/accuracies": 0.9765625, + "rewards/chosen": 0.251953125, + "rewards/margins": 11.0625, + "rewards/rejected": -10.8125, + "step": 994 + }, + { + "epoch": 1.2208588957055215, + "grad_norm": 15.26610473864711, + "learning_rate": 8.987603305785125e-07, + "logits/chosen": -0.099609375, + "logits/rejected": -0.2412109375, + "logps/chosen": -362.0, + "logps/rejected": -398.0, + "loss": 0.046, + "rewards/accuracies": 0.9765625, + "rewards/chosen": -0.52734375, + "rewards/margins": 10.3125, + "rewards/rejected": -10.8125, + "step": 995 + }, + { + "epoch": 1.2220858895705522, + "grad_norm": 19.263245504901857, + "learning_rate": 8.981404958677686e-07, + "logits/chosen": -0.023193359375, + "logits/rejected": -0.287109375, + "logps/chosen": -388.0, + "logps/rejected": -372.0, + "loss": 0.0498, + "rewards/accuracies": 0.984375, + "rewards/chosen": -0.2021484375, + "rewards/margins": 10.125, + "rewards/rejected": -10.375, + "step": 996 + }, + { + "epoch": 1.2233128834355829, + "grad_norm": 12.681921452742241, + "learning_rate": 8.975206611570248e-07, + "logits/chosen": -0.18359375, + "logits/rejected": -0.37109375, + "logps/chosen": -420.0, + "logps/rejected": -376.0, + "loss": 0.0371, + "rewards/accuracies": 0.984375, + "rewards/chosen": 1.171875, + "rewards/margins": 10.9375, + "rewards/rejected": -9.8125, + "step": 997 + }, + { + "epoch": 1.2245398773006135, + "grad_norm": 6.006248666573442, + "learning_rate": 8.969008264462811e-07, + "logits/chosen": -0.017578125, + "logits/rejected": -0.21484375, + "logps/chosen": -446.0, + "logps/rejected": -414.0, + "loss": 0.0184, + "rewards/accuracies": 1.0, + "rewards/chosen": 1.171875, + "rewards/margins": 11.25, + "rewards/rejected": -10.0625, + "step": 998 + }, + { + "epoch": 1.2257668711656442, + "grad_norm": 11.143450935670302, + "learning_rate": 8.962809917355373e-07, + "logits/chosen": -0.07177734375, + "logits/rejected": -0.240234375, + "logps/chosen": -410.0, + "logps/rejected": -386.0, + "loss": 0.0398, + "rewards/accuracies": 0.984375, + "rewards/chosen": 1.21875, + "rewards/margins": 11.25, + "rewards/rejected": -10.0625, + "step": 999 + }, + { + "epoch": 1.2269938650306749, + "grad_norm": 8.965376937967298, + "learning_rate": 8.956611570247934e-07, + "logits/chosen": -0.01385498046875, + "logits/rejected": -0.1015625, + "logps/chosen": -386.0, + "logps/rejected": -386.0, + "loss": 0.0469, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 1.0234375, + "rewards/margins": 10.8125, + "rewards/rejected": -9.75, + "step": 1000 + }, + { + "epoch": 1.2282208588957055, + "grad_norm": 11.244494297541378, + "learning_rate": 8.950413223140495e-07, + "logits/chosen": -0.185546875, + "logits/rejected": -0.3515625, + "logps/chosen": -410.0, + "logps/rejected": -378.0, + "loss": 0.0447, + "rewards/accuracies": 0.984375, + "rewards/chosen": 1.4453125, + "rewards/margins": 11.75, + "rewards/rejected": -10.3125, + "step": 1001 + }, + { + "epoch": 1.2294478527607362, + "grad_norm": 16.003709589565055, + "learning_rate": 8.944214876033058e-07, + "logits/chosen": -0.059326171875, + "logits/rejected": -0.09375, + "logps/chosen": -378.0, + "logps/rejected": -410.0, + "loss": 0.0566, + "rewards/accuracies": 0.9765625, + "rewards/chosen": 0.875, + "rewards/margins": 10.8125, + "rewards/rejected": -9.9375, + "step": 1002 + }, + { + "epoch": 1.2306748466257669, + "grad_norm": 12.870892763097368, + "learning_rate": 8.93801652892562e-07, + "logits/chosen": -0.1552734375, + "logits/rejected": -0.283203125, + "logps/chosen": -362.0, + "logps/rejected": -376.0, + "loss": 0.0662, + "rewards/accuracies": 0.9765625, + "rewards/chosen": 0.48828125, + "rewards/margins": 10.625, + "rewards/rejected": -10.125, + "step": 1003 + }, + { + "epoch": 1.2319018404907975, + "grad_norm": 5.2378730711968435, + "learning_rate": 8.931818181818183e-07, + "logits/chosen": -0.03076171875, + "logits/rejected": -0.2451171875, + "logps/chosen": -360.0, + "logps/rejected": -328.0, + "loss": 0.0257, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.55859375, + "rewards/margins": 9.875, + "rewards/rejected": -9.3125, + "step": 1004 + }, + { + "epoch": 1.2331288343558282, + "grad_norm": 6.60847796992945, + "learning_rate": 8.925619834710744e-07, + "logits/chosen": -0.0322265625, + "logits/rejected": -0.2177734375, + "logps/chosen": -396.0, + "logps/rejected": -360.0, + "loss": 0.031, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.443359375, + "rewards/margins": 10.5625, + "rewards/rejected": -10.125, + "step": 1005 + }, + { + "epoch": 1.2343558282208589, + "grad_norm": 7.331811572998349, + "learning_rate": 8.919421487603306e-07, + "logits/chosen": -0.0927734375, + "logits/rejected": -0.26171875, + "logps/chosen": -368.0, + "logps/rejected": -360.0, + "loss": 0.03, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.57421875, + "rewards/margins": 10.9375, + "rewards/rejected": -10.375, + "step": 1006 + }, + { + "epoch": 1.2355828220858895, + "grad_norm": 12.728485046758394, + "learning_rate": 8.913223140495869e-07, + "logits/chosen": -0.0595703125, + "logits/rejected": -0.2177734375, + "logps/chosen": -404.0, + "logps/rejected": -404.0, + "loss": 0.0543, + "rewards/accuracies": 0.96875, + "rewards/chosen": 0.8671875, + "rewards/margins": 11.125, + "rewards/rejected": -10.25, + "step": 1007 + }, + { + "epoch": 1.2368098159509202, + "grad_norm": 4.4503501042063816, + "learning_rate": 8.90702479338843e-07, + "logits/chosen": -0.0751953125, + "logits/rejected": -0.1904296875, + "logps/chosen": -358.0, + "logps/rejected": -356.0, + "loss": 0.0251, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.1806640625, + "rewards/margins": 10.125, + "rewards/rejected": -9.9375, + "step": 1008 + }, + { + "epoch": 1.2380368098159509, + "grad_norm": 12.812283743763317, + "learning_rate": 8.900826446280992e-07, + "logits/chosen": 0.0169677734375, + "logits/rejected": -0.185546875, + "logps/chosen": -426.0, + "logps/rejected": -380.0, + "loss": 0.0431, + "rewards/accuracies": 0.984375, + "rewards/chosen": 1.40625, + "rewards/margins": 11.1875, + "rewards/rejected": -9.8125, + "step": 1009 + }, + { + "epoch": 1.2392638036809815, + "grad_norm": 6.899754042678334, + "learning_rate": 8.894628099173553e-07, + "logits/chosen": -0.0181884765625, + "logits/rejected": -0.10205078125, + "logps/chosen": -354.0, + "logps/rejected": -374.0, + "loss": 0.03, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.10595703125, + "rewards/margins": 10.0, + "rewards/rejected": -10.0625, + "step": 1010 + }, + { + "epoch": 1.2404907975460122, + "grad_norm": 6.300076629714054, + "learning_rate": 8.888429752066116e-07, + "logits/chosen": -0.0289306640625, + "logits/rejected": -0.21875, + "logps/chosen": -412.0, + "logps/rejected": -390.0, + "loss": 0.0272, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.4375, + "rewards/margins": 10.6875, + "rewards/rejected": -10.25, + "step": 1011 + }, + { + "epoch": 1.2417177914110429, + "grad_norm": 15.956703887603753, + "learning_rate": 8.882231404958678e-07, + "logits/chosen": -0.10107421875, + "logits/rejected": -0.25390625, + "logps/chosen": -384.0, + "logps/rejected": -372.0, + "loss": 0.0733, + "rewards/accuracies": 0.9609375, + "rewards/chosen": 0.462890625, + "rewards/margins": 10.25, + "rewards/rejected": -9.75, + "step": 1012 + }, + { + "epoch": 1.2429447852760735, + "grad_norm": 7.738683663773001, + "learning_rate": 8.87603305785124e-07, + "logits/chosen": 0.01318359375, + "logits/rejected": -0.049560546875, + "logps/chosen": -398.0, + "logps/rejected": -380.0, + "loss": 0.0336, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.11328125, + "rewards/margins": 9.0625, + "rewards/rejected": -9.1875, + "step": 1013 + }, + { + "epoch": 1.2441717791411042, + "grad_norm": 8.908202927372313, + "learning_rate": 8.869834710743802e-07, + "logits/chosen": -0.09765625, + "logits/rejected": -0.2275390625, + "logps/chosen": -386.0, + "logps/rejected": -406.0, + "loss": 0.0403, + "rewards/accuracies": 0.9765625, + "rewards/chosen": -0.00567626953125, + "rewards/margins": 10.0, + "rewards/rejected": -10.0, + "step": 1014 + }, + { + "epoch": 1.2453987730061349, + "grad_norm": 8.437111565369221, + "learning_rate": 8.863636363636364e-07, + "logits/chosen": -0.06689453125, + "logits/rejected": -0.1826171875, + "logps/chosen": -370.0, + "logps/rejected": -388.0, + "loss": 0.036, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 1.21875, + "rewards/margins": 11.6875, + "rewards/rejected": -10.4375, + "step": 1015 + }, + { + "epoch": 1.2466257668711656, + "grad_norm": 19.806315774782945, + "learning_rate": 8.857438016528927e-07, + "logits/chosen": -0.0859375, + "logits/rejected": -0.1669921875, + "logps/chosen": -364.0, + "logps/rejected": -348.0, + "loss": 0.0999, + "rewards/accuracies": 0.9609375, + "rewards/chosen": -0.06787109375, + "rewards/margins": 9.4375, + "rewards/rejected": -9.5, + "step": 1016 + }, + { + "epoch": 1.2478527607361962, + "grad_norm": 1.9621085105816412, + "learning_rate": 8.851239669421488e-07, + "logits/chosen": -0.08740234375, + "logits/rejected": -0.31640625, + "logps/chosen": -404.0, + "logps/rejected": -362.0, + "loss": 0.0089, + "rewards/accuracies": 1.0, + "rewards/chosen": 1.234375, + "rewards/margins": 11.8125, + "rewards/rejected": -10.5625, + "step": 1017 + }, + { + "epoch": 1.2490797546012269, + "grad_norm": 7.315995365487366, + "learning_rate": 8.845041322314049e-07, + "logits/chosen": -0.11865234375, + "logits/rejected": -0.1962890625, + "logps/chosen": -408.0, + "logps/rejected": -416.0, + "loss": 0.0283, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.11572265625, + "rewards/margins": 11.125, + "rewards/rejected": -11.25, + "step": 1018 + }, + { + "epoch": 1.2503067484662576, + "grad_norm": 6.452661280214959, + "learning_rate": 8.838842975206611e-07, + "logits/chosen": -0.15234375, + "logits/rejected": -0.302734375, + "logps/chosen": -376.0, + "logps/rejected": -378.0, + "loss": 0.0318, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.296875, + "rewards/margins": 11.1875, + "rewards/rejected": -10.875, + "step": 1019 + }, + { + "epoch": 1.2515337423312882, + "grad_norm": 4.8866762625917, + "learning_rate": 8.832644628099174e-07, + "logits/chosen": -0.09033203125, + "logits/rejected": -0.193359375, + "logps/chosen": -414.0, + "logps/rejected": -402.0, + "loss": 0.0215, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.33203125, + "rewards/margins": 11.75, + "rewards/rejected": -11.4375, + "step": 1020 + }, + { + "epoch": 1.252760736196319, + "grad_norm": 7.044566658635158, + "learning_rate": 8.826446280991736e-07, + "logits/chosen": -0.05908203125, + "logits/rejected": -0.13671875, + "logps/chosen": -370.0, + "logps/rejected": -378.0, + "loss": 0.0313, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.171875, + "rewards/margins": 11.4375, + "rewards/rejected": -11.25, + "step": 1021 + }, + { + "epoch": 1.2539877300613496, + "grad_norm": 2.074684049190498, + "learning_rate": 8.820247933884298e-07, + "logits/chosen": -0.1708984375, + "logits/rejected": -0.30859375, + "logps/chosen": -420.0, + "logps/rejected": -396.0, + "loss": 0.0134, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.400390625, + "rewards/margins": 11.3125, + "rewards/rejected": -10.9375, + "step": 1022 + }, + { + "epoch": 1.2552147239263804, + "grad_norm": 11.38544236701507, + "learning_rate": 8.81404958677686e-07, + "logits/chosen": -0.142578125, + "logits/rejected": -0.314453125, + "logps/chosen": -402.0, + "logps/rejected": -398.0, + "loss": 0.0566, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.9453125, + "rewards/margins": 11.8125, + "rewards/rejected": -10.875, + "step": 1023 + }, + { + "epoch": 1.2564417177914111, + "grad_norm": 7.64965729920436, + "learning_rate": 8.807851239669422e-07, + "logits/chosen": -0.028564453125, + "logits/rejected": -0.275390625, + "logps/chosen": -410.0, + "logps/rejected": -420.0, + "loss": 0.0254, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.408203125, + "rewards/margins": 11.9375, + "rewards/rejected": -11.5, + "step": 1024 + }, + { + "epoch": 1.2576687116564418, + "grad_norm": 23.522395476453028, + "learning_rate": 8.801652892561984e-07, + "logits/chosen": -0.03466796875, + "logits/rejected": -0.1962890625, + "logps/chosen": -414.0, + "logps/rejected": -418.0, + "loss": 0.1276, + "rewards/accuracies": 0.96875, + "rewards/chosen": 1.0078125, + "rewards/margins": 11.3125, + "rewards/rejected": -10.3125, + "step": 1025 + }, + { + "epoch": 1.2588957055214725, + "grad_norm": 4.975922218453701, + "learning_rate": 8.795454545454545e-07, + "logits/chosen": -0.1318359375, + "logits/rejected": -0.30078125, + "logps/chosen": -370.0, + "logps/rejected": -376.0, + "loss": 0.0313, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.30859375, + "rewards/margins": 10.9375, + "rewards/rejected": -10.625, + "step": 1026 + }, + { + "epoch": 1.2601226993865031, + "grad_norm": 4.952502473285682, + "learning_rate": 8.789256198347107e-07, + "logits/chosen": -0.17578125, + "logits/rejected": -0.41015625, + "logps/chosen": -448.0, + "logps/rejected": -412.0, + "loss": 0.0181, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.6015625, + "rewards/margins": 11.375, + "rewards/rejected": -10.75, + "step": 1027 + }, + { + "epoch": 1.2613496932515338, + "grad_norm": 7.5494432139799015, + "learning_rate": 8.783057851239669e-07, + "logits/chosen": -0.158203125, + "logits/rejected": -0.3125, + "logps/chosen": -396.0, + "logps/rejected": -396.0, + "loss": 0.0222, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 1.2421875, + "rewards/margins": 12.3125, + "rewards/rejected": -11.0625, + "step": 1028 + }, + { + "epoch": 1.2625766871165645, + "grad_norm": 8.079926823970226, + "learning_rate": 8.776859504132232e-07, + "logits/chosen": -0.1083984375, + "logits/rejected": -0.298828125, + "logps/chosen": -390.0, + "logps/rejected": -384.0, + "loss": 0.0315, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.455078125, + "rewards/margins": 10.5, + "rewards/rejected": -10.0, + "step": 1029 + }, + { + "epoch": 1.2638036809815951, + "grad_norm": 6.946808552773218, + "learning_rate": 8.770661157024793e-07, + "logits/chosen": -0.3046875, + "logits/rejected": -0.38671875, + "logps/chosen": -414.0, + "logps/rejected": -410.0, + "loss": 0.0214, + "rewards/accuracies": 1.0, + "rewards/chosen": 1.1796875, + "rewards/margins": 12.125, + "rewards/rejected": -10.9375, + "step": 1030 + }, + { + "epoch": 1.2650306748466258, + "grad_norm": 13.376232318087313, + "learning_rate": 8.764462809917356e-07, + "logits/chosen": -0.0703125, + "logits/rejected": -0.30859375, + "logps/chosen": -382.0, + "logps/rejected": -360.0, + "loss": 0.0746, + "rewards/accuracies": 0.9765625, + "rewards/chosen": 0.55859375, + "rewards/margins": 10.375, + "rewards/rejected": -9.875, + "step": 1031 + }, + { + "epoch": 1.2662576687116565, + "grad_norm": 6.898643809079758, + "learning_rate": 8.758264462809918e-07, + "logits/chosen": 0.042724609375, + "logits/rejected": -0.1669921875, + "logps/chosen": -376.0, + "logps/rejected": -358.0, + "loss": 0.0263, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.76171875, + "rewards/margins": 10.5625, + "rewards/rejected": -9.8125, + "step": 1032 + }, + { + "epoch": 1.2674846625766871, + "grad_norm": 14.180165703858819, + "learning_rate": 8.75206611570248e-07, + "logits/chosen": -0.123046875, + "logits/rejected": -0.33984375, + "logps/chosen": -438.0, + "logps/rejected": -392.0, + "loss": 0.0718, + "rewards/accuracies": 0.9609375, + "rewards/chosen": -0.004547119140625, + "rewards/margins": 10.0, + "rewards/rejected": -10.0625, + "step": 1033 + }, + { + "epoch": 1.2687116564417178, + "grad_norm": 5.086342064289492, + "learning_rate": 8.745867768595042e-07, + "logits/chosen": -0.1572265625, + "logits/rejected": -0.38671875, + "logps/chosen": -464.0, + "logps/rejected": -420.0, + "loss": 0.0192, + "rewards/accuracies": 1.0, + "rewards/chosen": 1.25, + "rewards/margins": 12.1875, + "rewards/rejected": -10.9375, + "step": 1034 + }, + { + "epoch": 1.2699386503067485, + "grad_norm": 14.981826471029674, + "learning_rate": 8.739669421487602e-07, + "logits/chosen": -0.0712890625, + "logits/rejected": -0.279296875, + "logps/chosen": -408.0, + "logps/rejected": -388.0, + "loss": 0.0436, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.8046875, + "rewards/margins": 11.375, + "rewards/rejected": -10.5625, + "step": 1035 + }, + { + "epoch": 1.2711656441717791, + "grad_norm": 11.555096518453992, + "learning_rate": 8.733471074380165e-07, + "logits/chosen": -0.0693359375, + "logits/rejected": -0.20703125, + "logps/chosen": -380.0, + "logps/rejected": -368.0, + "loss": 0.0478, + "rewards/accuracies": 0.96875, + "rewards/chosen": 1.1640625, + "rewards/margins": 10.875, + "rewards/rejected": -9.6875, + "step": 1036 + }, + { + "epoch": 1.2723926380368098, + "grad_norm": 4.796864490403106, + "learning_rate": 8.727272727272727e-07, + "logits/chosen": -0.27734375, + "logits/rejected": -0.41015625, + "logps/chosen": -416.0, + "logps/rejected": -432.0, + "loss": 0.0167, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.984375, + "rewards/margins": 11.5625, + "rewards/rejected": -10.5625, + "step": 1037 + }, + { + "epoch": 1.2736196319018405, + "grad_norm": 7.4606854944626555, + "learning_rate": 8.72107438016529e-07, + "logits/chosen": -0.1103515625, + "logits/rejected": -0.2734375, + "logps/chosen": -382.0, + "logps/rejected": -392.0, + "loss": 0.0268, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.60546875, + "rewards/margins": 11.6875, + "rewards/rejected": -11.0625, + "step": 1038 + }, + { + "epoch": 1.2748466257668711, + "grad_norm": 11.522250663977559, + "learning_rate": 8.714876033057851e-07, + "logits/chosen": -0.146484375, + "logits/rejected": -0.2392578125, + "logps/chosen": -382.0, + "logps/rejected": -414.0, + "loss": 0.0475, + "rewards/accuracies": 0.9765625, + "rewards/chosen": 0.0751953125, + "rewards/margins": 10.375, + "rewards/rejected": -10.3125, + "step": 1039 + }, + { + "epoch": 1.2760736196319018, + "grad_norm": 13.412392481033551, + "learning_rate": 8.708677685950414e-07, + "logits/chosen": -0.18359375, + "logits/rejected": -0.470703125, + "logps/chosen": -426.0, + "logps/rejected": -366.0, + "loss": 0.0344, + "rewards/accuracies": 0.9765625, + "rewards/chosen": 0.7734375, + "rewards/margins": 11.875, + "rewards/rejected": -11.125, + "step": 1040 + }, + { + "epoch": 1.2773006134969325, + "grad_norm": 4.076327393298983, + "learning_rate": 8.702479338842976e-07, + "logits/chosen": -0.1357421875, + "logits/rejected": -0.2734375, + "logps/chosen": -382.0, + "logps/rejected": -378.0, + "loss": 0.013, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.279296875, + "rewards/margins": 11.3125, + "rewards/rejected": -11.0625, + "step": 1041 + }, + { + "epoch": 1.2785276073619631, + "grad_norm": 12.198994819526218, + "learning_rate": 8.696280991735537e-07, + "logits/chosen": -0.3046875, + "logits/rejected": -0.5, + "logps/chosen": -452.0, + "logps/rejected": -434.0, + "loss": 0.0344, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.68359375, + "rewards/margins": 12.6875, + "rewards/rejected": -12.0, + "step": 1042 + }, + { + "epoch": 1.2797546012269938, + "grad_norm": 15.876129225419682, + "learning_rate": 8.6900826446281e-07, + "logits/chosen": -0.10400390625, + "logits/rejected": -0.287109375, + "logps/chosen": -380.0, + "logps/rejected": -386.0, + "loss": 0.0536, + "rewards/accuracies": 0.9765625, + "rewards/chosen": -0.63671875, + "rewards/margins": 10.125, + "rewards/rejected": -10.8125, + "step": 1043 + }, + { + "epoch": 1.2809815950920245, + "grad_norm": 5.177920548454582, + "learning_rate": 8.68388429752066e-07, + "logits/chosen": -0.2041015625, + "logits/rejected": -0.421875, + "logps/chosen": -432.0, + "logps/rejected": -402.0, + "loss": 0.0149, + "rewards/accuracies": 1.0, + "rewards/chosen": 1.1796875, + "rewards/margins": 12.25, + "rewards/rejected": -11.0625, + "step": 1044 + }, + { + "epoch": 1.2822085889570551, + "grad_norm": 8.9147491979472, + "learning_rate": 8.677685950413223e-07, + "logits/chosen": -0.251953125, + "logits/rejected": -0.52734375, + "logps/chosen": -430.0, + "logps/rejected": -374.0, + "loss": 0.0292, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.734375, + "rewards/margins": 11.75, + "rewards/rejected": -11.0, + "step": 1045 + }, + { + "epoch": 1.2834355828220858, + "grad_norm": 6.882145843965057, + "learning_rate": 8.671487603305785e-07, + "logits/chosen": -0.017578125, + "logits/rejected": -0.1689453125, + "logps/chosen": -396.0, + "logps/rejected": -404.0, + "loss": 0.0288, + "rewards/accuracies": 0.984375, + "rewards/chosen": -0.0260009765625, + "rewards/margins": 10.625, + "rewards/rejected": -10.6875, + "step": 1046 + }, + { + "epoch": 1.2846625766871167, + "grad_norm": 12.443133314198835, + "learning_rate": 8.665289256198347e-07, + "logits/chosen": -0.1826171875, + "logits/rejected": -0.330078125, + "logps/chosen": -404.0, + "logps/rejected": -404.0, + "loss": 0.0533, + "rewards/accuracies": 0.9765625, + "rewards/chosen": 0.68359375, + "rewards/margins": 12.25, + "rewards/rejected": -11.5625, + "step": 1047 + }, + { + "epoch": 1.2858895705521474, + "grad_norm": 4.187308341376978, + "learning_rate": 8.659090909090909e-07, + "logits/chosen": -0.0703125, + "logits/rejected": -0.265625, + "logps/chosen": -342.0, + "logps/rejected": -348.0, + "loss": 0.019, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.1494140625, + "rewards/margins": 10.9375, + "rewards/rejected": -10.8125, + "step": 1048 + }, + { + "epoch": 1.287116564417178, + "grad_norm": 12.339404257771985, + "learning_rate": 8.652892561983472e-07, + "logits/chosen": -0.10693359375, + "logits/rejected": -0.326171875, + "logps/chosen": -376.0, + "logps/rejected": -374.0, + "loss": 0.0446, + "rewards/accuracies": 0.9765625, + "rewards/chosen": -0.419921875, + "rewards/margins": 10.4375, + "rewards/rejected": -10.8125, + "step": 1049 + }, + { + "epoch": 1.2883435582822087, + "grad_norm": 7.946831033807572, + "learning_rate": 8.646694214876034e-07, + "logits/chosen": -0.2451171875, + "logits/rejected": -0.384765625, + "logps/chosen": -398.0, + "logps/rejected": -406.0, + "loss": 0.0372, + "rewards/accuracies": 0.984375, + "rewards/chosen": -0.25390625, + "rewards/margins": 11.0625, + "rewards/rejected": -11.3125, + "step": 1050 + }, + { + "epoch": 1.2895705521472394, + "grad_norm": 12.451016602757491, + "learning_rate": 8.640495867768595e-07, + "logits/chosen": -0.279296875, + "logits/rejected": -0.466796875, + "logps/chosen": -420.0, + "logps/rejected": -424.0, + "loss": 0.037, + "rewards/accuracies": 0.9765625, + "rewards/chosen": 0.71484375, + "rewards/margins": 12.6875, + "rewards/rejected": -12.0, + "step": 1051 + }, + { + "epoch": 1.29079754601227, + "grad_norm": 14.98941632692573, + "learning_rate": 8.634297520661158e-07, + "logits/chosen": -0.1328125, + "logits/rejected": -0.3359375, + "logps/chosen": -432.0, + "logps/rejected": -414.0, + "loss": 0.0593, + "rewards/accuracies": 0.9609375, + "rewards/chosen": -0.12353515625, + "rewards/margins": 10.6875, + "rewards/rejected": -10.8125, + "step": 1052 + }, + { + "epoch": 1.2920245398773007, + "grad_norm": 17.525115785642342, + "learning_rate": 8.628099173553718e-07, + "logits/chosen": -0.06640625, + "logits/rejected": -0.287109375, + "logps/chosen": -366.0, + "logps/rejected": -366.0, + "loss": 0.069, + "rewards/accuracies": 0.96875, + "rewards/chosen": 0.26171875, + "rewards/margins": 11.3125, + "rewards/rejected": -11.0625, + "step": 1053 + }, + { + "epoch": 1.2932515337423314, + "grad_norm": 11.652127715235839, + "learning_rate": 8.621900826446281e-07, + "logits/chosen": -0.11767578125, + "logits/rejected": -0.3359375, + "logps/chosen": -402.0, + "logps/rejected": -394.0, + "loss": 0.0503, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.263671875, + "rewards/margins": 11.9375, + "rewards/rejected": -11.6875, + "step": 1054 + }, + { + "epoch": 1.294478527607362, + "grad_norm": 5.171446811615976, + "learning_rate": 8.615702479338843e-07, + "logits/chosen": -0.11962890625, + "logits/rejected": -0.33203125, + "logps/chosen": -386.0, + "logps/rejected": -386.0, + "loss": 0.0178, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.0439453125, + "rewards/margins": 11.6875, + "rewards/rejected": -11.75, + "step": 1055 + }, + { + "epoch": 1.2957055214723927, + "grad_norm": 3.3720116007928245, + "learning_rate": 8.609504132231405e-07, + "logits/chosen": -0.15625, + "logits/rejected": -0.34375, + "logps/chosen": -368.0, + "logps/rejected": -346.0, + "loss": 0.0109, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.0181884765625, + "rewards/margins": 10.875, + "rewards/rejected": -10.875, + "step": 1056 + }, + { + "epoch": 1.2969325153374234, + "grad_norm": 12.84642536946459, + "learning_rate": 8.603305785123967e-07, + "logits/chosen": -0.1298828125, + "logits/rejected": -0.2734375, + "logps/chosen": -404.0, + "logps/rejected": -390.0, + "loss": 0.0402, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.365234375, + "rewards/margins": 11.25, + "rewards/rejected": -10.9375, + "step": 1057 + }, + { + "epoch": 1.298159509202454, + "grad_norm": 9.137246685227527, + "learning_rate": 8.59710743801653e-07, + "logits/chosen": -0.130859375, + "logits/rejected": -0.365234375, + "logps/chosen": -390.0, + "logps/rejected": -388.0, + "loss": 0.0341, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.84765625, + "rewards/margins": 12.375, + "rewards/rejected": -11.5, + "step": 1058 + }, + { + "epoch": 1.2993865030674847, + "grad_norm": 8.50866186346429, + "learning_rate": 8.590909090909092e-07, + "logits/chosen": -0.1845703125, + "logits/rejected": -0.38671875, + "logps/chosen": -416.0, + "logps/rejected": -384.0, + "loss": 0.0325, + "rewards/accuracies": 0.984375, + "rewards/chosen": -0.341796875, + "rewards/margins": 11.3125, + "rewards/rejected": -11.6875, + "step": 1059 + }, + { + "epoch": 1.3006134969325154, + "grad_norm": 8.4621363068171, + "learning_rate": 8.584710743801653e-07, + "logits/chosen": -0.15234375, + "logits/rejected": -0.2421875, + "logps/chosen": -338.0, + "logps/rejected": -350.0, + "loss": 0.0316, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.392578125, + "rewards/margins": 11.3125, + "rewards/rejected": -10.875, + "step": 1060 + }, + { + "epoch": 1.301840490797546, + "grad_norm": 12.798702675578042, + "learning_rate": 8.578512396694216e-07, + "logits/chosen": -0.059326171875, + "logits/rejected": -0.1630859375, + "logps/chosen": -342.0, + "logps/rejected": -372.0, + "loss": 0.0443, + "rewards/accuracies": 0.984375, + "rewards/chosen": -0.052734375, + "rewards/margins": 10.6875, + "rewards/rejected": -10.6875, + "step": 1061 + }, + { + "epoch": 1.3030674846625767, + "grad_norm": 10.362126702943744, + "learning_rate": 8.572314049586776e-07, + "logits/chosen": -0.1015625, + "logits/rejected": -0.2373046875, + "logps/chosen": -422.0, + "logps/rejected": -416.0, + "loss": 0.0494, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.369140625, + "rewards/margins": 11.0625, + "rewards/rejected": -10.6875, + "step": 1062 + }, + { + "epoch": 1.3042944785276074, + "grad_norm": 13.766378342401254, + "learning_rate": 8.566115702479339e-07, + "logits/chosen": -0.1630859375, + "logits/rejected": -0.296875, + "logps/chosen": -370.0, + "logps/rejected": -366.0, + "loss": 0.0643, + "rewards/accuracies": 0.9765625, + "rewards/chosen": -0.228515625, + "rewards/margins": 10.75, + "rewards/rejected": -10.9375, + "step": 1063 + }, + { + "epoch": 1.305521472392638, + "grad_norm": 8.950574541603398, + "learning_rate": 8.559917355371901e-07, + "logits/chosen": -0.1640625, + "logits/rejected": -0.3046875, + "logps/chosen": -386.0, + "logps/rejected": -392.0, + "loss": 0.034, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.057861328125, + "rewards/margins": 11.5, + "rewards/rejected": -11.4375, + "step": 1064 + }, + { + "epoch": 1.3067484662576687, + "grad_norm": 8.164498598328498, + "learning_rate": 8.553719008264463e-07, + "logits/chosen": -0.11279296875, + "logits/rejected": -0.4296875, + "logps/chosen": -448.0, + "logps/rejected": -408.0, + "loss": 0.0283, + "rewards/accuracies": 0.984375, + "rewards/chosen": 1.2734375, + "rewards/margins": 13.3125, + "rewards/rejected": -12.0, + "step": 1065 + }, + { + "epoch": 1.3079754601226994, + "grad_norm": 9.939864149496472, + "learning_rate": 8.547520661157025e-07, + "logits/chosen": -0.1435546875, + "logits/rejected": -0.435546875, + "logps/chosen": -464.0, + "logps/rejected": -430.0, + "loss": 0.0316, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.328125, + "rewards/margins": 11.5, + "rewards/rejected": -11.1875, + "step": 1066 + }, + { + "epoch": 1.30920245398773, + "grad_norm": 5.026276230946228, + "learning_rate": 8.541322314049588e-07, + "logits/chosen": -0.1640625, + "logits/rejected": -0.27734375, + "logps/chosen": -376.0, + "logps/rejected": -386.0, + "loss": 0.0294, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.359375, + "rewards/margins": 9.75, + "rewards/rejected": -10.125, + "step": 1067 + }, + { + "epoch": 1.3104294478527607, + "grad_norm": 11.031737625122464, + "learning_rate": 8.535123966942149e-07, + "logits/chosen": -0.1767578125, + "logits/rejected": -0.32421875, + "logps/chosen": -388.0, + "logps/rejected": -366.0, + "loss": 0.0537, + "rewards/accuracies": 0.9765625, + "rewards/chosen": 0.283203125, + "rewards/margins": 11.3125, + "rewards/rejected": -11.0625, + "step": 1068 + }, + { + "epoch": 1.3116564417177914, + "grad_norm": 8.905100714168142, + "learning_rate": 8.528925619834711e-07, + "logits/chosen": -0.07666015625, + "logits/rejected": -0.3046875, + "logps/chosen": -406.0, + "logps/rejected": -402.0, + "loss": 0.0429, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.48828125, + "rewards/margins": 10.6875, + "rewards/rejected": -10.1875, + "step": 1069 + }, + { + "epoch": 1.312883435582822, + "grad_norm": 11.752525584916722, + "learning_rate": 8.522727272727274e-07, + "logits/chosen": -0.19140625, + "logits/rejected": -0.259765625, + "logps/chosen": -386.0, + "logps/rejected": -406.0, + "loss": 0.0348, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.96484375, + "rewards/margins": 12.125, + "rewards/rejected": -11.1875, + "step": 1070 + }, + { + "epoch": 1.3141104294478527, + "grad_norm": 12.851082658918303, + "learning_rate": 8.516528925619834e-07, + "logits/chosen": -0.2255859375, + "logits/rejected": -0.408203125, + "logps/chosen": -392.0, + "logps/rejected": -388.0, + "loss": 0.0656, + "rewards/accuracies": 0.9765625, + "rewards/chosen": 0.4375, + "rewards/margins": 11.25, + "rewards/rejected": -10.8125, + "step": 1071 + }, + { + "epoch": 1.3153374233128834, + "grad_norm": 9.860743385521058, + "learning_rate": 8.510330578512397e-07, + "logits/chosen": -0.002685546875, + "logits/rejected": -0.197265625, + "logps/chosen": -408.0, + "logps/rejected": -398.0, + "loss": 0.0258, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.66015625, + "rewards/margins": 11.125, + "rewards/rejected": -10.4375, + "step": 1072 + }, + { + "epoch": 1.316564417177914, + "grad_norm": 8.94951495923183, + "learning_rate": 8.504132231404958e-07, + "logits/chosen": -0.1982421875, + "logits/rejected": -0.30859375, + "logps/chosen": -330.0, + "logps/rejected": -372.0, + "loss": 0.0282, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.0400390625, + "rewards/margins": 10.25, + "rewards/rejected": -10.1875, + "step": 1073 + }, + { + "epoch": 1.3177914110429447, + "grad_norm": 5.703540754764619, + "learning_rate": 8.497933884297521e-07, + "logits/chosen": -0.2099609375, + "logits/rejected": -0.3203125, + "logps/chosen": -390.0, + "logps/rejected": -388.0, + "loss": 0.0261, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.291015625, + "rewards/margins": 11.375, + "rewards/rejected": -11.125, + "step": 1074 + }, + { + "epoch": 1.3190184049079754, + "grad_norm": 9.876644774460352, + "learning_rate": 8.491735537190083e-07, + "logits/chosen": -0.060791015625, + "logits/rejected": -0.224609375, + "logps/chosen": -398.0, + "logps/rejected": -396.0, + "loss": 0.0274, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.3046875, + "rewards/margins": 11.0625, + "rewards/rejected": -10.75, + "step": 1075 + }, + { + "epoch": 1.320245398773006, + "grad_norm": 4.967869121366148, + "learning_rate": 8.485537190082646e-07, + "logits/chosen": -0.037353515625, + "logits/rejected": -0.123046875, + "logps/chosen": -400.0, + "logps/rejected": -398.0, + "loss": 0.0212, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.28125, + "rewards/margins": 11.4375, + "rewards/rejected": -11.1875, + "step": 1076 + }, + { + "epoch": 1.3214723926380367, + "grad_norm": 9.322961880086396, + "learning_rate": 8.479338842975207e-07, + "logits/chosen": -0.2099609375, + "logits/rejected": -0.3359375, + "logps/chosen": -302.0, + "logps/rejected": -312.0, + "loss": 0.0452, + "rewards/accuracies": 0.984375, + "rewards/chosen": -0.3984375, + "rewards/margins": 9.875, + "rewards/rejected": -10.25, + "step": 1077 + }, + { + "epoch": 1.3226993865030674, + "grad_norm": 5.738661119438272, + "learning_rate": 8.473140495867769e-07, + "logits/chosen": -0.049072265625, + "logits/rejected": -0.158203125, + "logps/chosen": -386.0, + "logps/rejected": -396.0, + "loss": 0.0334, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.5078125, + "rewards/margins": 9.3125, + "rewards/rejected": -9.8125, + "step": 1078 + }, + { + "epoch": 1.323926380368098, + "grad_norm": 6.333892508394478, + "learning_rate": 8.466942148760332e-07, + "logits/chosen": -0.1201171875, + "logits/rejected": -0.259765625, + "logps/chosen": -384.0, + "logps/rejected": -388.0, + "loss": 0.0328, + "rewards/accuracies": 0.9765625, + "rewards/chosen": 0.16796875, + "rewards/margins": 11.6875, + "rewards/rejected": -11.5, + "step": 1079 + }, + { + "epoch": 1.3251533742331287, + "grad_norm": 13.790501614554328, + "learning_rate": 8.460743801652892e-07, + "logits/chosen": -0.0654296875, + "logits/rejected": -0.263671875, + "logps/chosen": -366.0, + "logps/rejected": -358.0, + "loss": 0.047, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.0322265625, + "rewards/margins": 10.0625, + "rewards/rejected": -10.0625, + "step": 1080 + }, + { + "epoch": 1.3263803680981594, + "grad_norm": 7.653539387358885, + "learning_rate": 8.454545454545455e-07, + "logits/chosen": -0.1005859375, + "logits/rejected": -0.259765625, + "logps/chosen": -392.0, + "logps/rejected": -396.0, + "loss": 0.0364, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.1640625, + "rewards/margins": 10.9375, + "rewards/rejected": -10.75, + "step": 1081 + }, + { + "epoch": 1.32760736196319, + "grad_norm": 10.410957750314225, + "learning_rate": 8.448347107438016e-07, + "logits/chosen": -0.2021484375, + "logits/rejected": -0.435546875, + "logps/chosen": -412.0, + "logps/rejected": -366.0, + "loss": 0.0374, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.2119140625, + "rewards/margins": 10.375, + "rewards/rejected": -10.125, + "step": 1082 + }, + { + "epoch": 1.3288343558282207, + "grad_norm": 5.66314951524206, + "learning_rate": 8.442148760330579e-07, + "logits/chosen": -0.1025390625, + "logits/rejected": -0.283203125, + "logps/chosen": -372.0, + "logps/rejected": -328.0, + "loss": 0.0214, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.142578125, + "rewards/margins": 9.875, + "rewards/rejected": -10.0, + "step": 1083 + }, + { + "epoch": 1.3300613496932514, + "grad_norm": 11.42793523481754, + "learning_rate": 8.435950413223141e-07, + "logits/chosen": -0.15625, + "logits/rejected": -0.16015625, + "logps/chosen": -354.0, + "logps/rejected": -362.0, + "loss": 0.0449, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.1640625, + "rewards/margins": 10.625, + "rewards/rejected": -10.5, + "step": 1084 + }, + { + "epoch": 1.331288343558282, + "grad_norm": 6.950168336723294, + "learning_rate": 8.429752066115703e-07, + "logits/chosen": -0.033203125, + "logits/rejected": -0.181640625, + "logps/chosen": -386.0, + "logps/rejected": -376.0, + "loss": 0.0221, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.125, + "rewards/margins": 10.125, + "rewards/rejected": -10.25, + "step": 1085 + }, + { + "epoch": 1.332515337423313, + "grad_norm": 7.903285776855893, + "learning_rate": 8.423553719008265e-07, + "logits/chosen": -0.10107421875, + "logits/rejected": -0.228515625, + "logps/chosen": -316.0, + "logps/rejected": -344.0, + "loss": 0.0384, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.365234375, + "rewards/margins": 10.75, + "rewards/rejected": -10.375, + "step": 1086 + }, + { + "epoch": 1.3337423312883436, + "grad_norm": 14.294247050939283, + "learning_rate": 8.417355371900827e-07, + "logits/chosen": -0.2099609375, + "logits/rejected": -0.33984375, + "logps/chosen": -322.0, + "logps/rejected": -376.0, + "loss": 0.0381, + "rewards/accuracies": 0.984375, + "rewards/chosen": -0.251953125, + "rewards/margins": 10.375, + "rewards/rejected": -10.625, + "step": 1087 + }, + { + "epoch": 1.3349693251533743, + "grad_norm": 12.270443455514881, + "learning_rate": 8.41115702479339e-07, + "logits/chosen": -0.1376953125, + "logits/rejected": -0.365234375, + "logps/chosen": -446.0, + "logps/rejected": -378.0, + "loss": 0.0352, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.65234375, + "rewards/margins": 11.25, + "rewards/rejected": -10.625, + "step": 1088 + }, + { + "epoch": 1.336196319018405, + "grad_norm": 9.38631227726379, + "learning_rate": 8.40495867768595e-07, + "logits/chosen": -0.1884765625, + "logits/rejected": -0.34375, + "logps/chosen": -390.0, + "logps/rejected": -360.0, + "loss": 0.039, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.6015625, + "rewards/margins": 9.875, + "rewards/rejected": -10.5, + "step": 1089 + }, + { + "epoch": 1.3374233128834356, + "grad_norm": 13.010546372728042, + "learning_rate": 8.398760330578512e-07, + "logits/chosen": -0.1357421875, + "logits/rejected": -0.33984375, + "logps/chosen": -374.0, + "logps/rejected": -352.0, + "loss": 0.0449, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.1728515625, + "rewards/margins": 10.125, + "rewards/rejected": -10.25, + "step": 1090 + }, + { + "epoch": 1.3386503067484663, + "grad_norm": 17.2682088486596, + "learning_rate": 8.392561983471074e-07, + "logits/chosen": 0.013427734375, + "logits/rejected": -0.1552734375, + "logps/chosen": -424.0, + "logps/rejected": -368.0, + "loss": 0.0874, + "rewards/accuracies": 0.9765625, + "rewards/chosen": 0.1650390625, + "rewards/margins": 10.125, + "rewards/rejected": -9.9375, + "step": 1091 + }, + { + "epoch": 1.339877300613497, + "grad_norm": 6.844436331855588, + "learning_rate": 8.386363636363637e-07, + "logits/chosen": -0.1455078125, + "logits/rejected": -0.31640625, + "logps/chosen": -414.0, + "logps/rejected": -398.0, + "loss": 0.0256, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.484375, + "rewards/margins": 10.75, + "rewards/rejected": -10.25, + "step": 1092 + }, + { + "epoch": 1.3411042944785276, + "grad_norm": 5.767510463883874, + "learning_rate": 8.380165289256199e-07, + "logits/chosen": -0.2041015625, + "logits/rejected": -0.34375, + "logps/chosen": -378.0, + "logps/rejected": -372.0, + "loss": 0.0207, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.328125, + "rewards/margins": 10.25, + "rewards/rejected": -9.9375, + "step": 1093 + }, + { + "epoch": 1.3423312883435583, + "grad_norm": 8.158470602909508, + "learning_rate": 8.37396694214876e-07, + "logits/chosen": -0.08251953125, + "logits/rejected": -0.30859375, + "logps/chosen": -426.0, + "logps/rejected": -420.0, + "loss": 0.0291, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.546875, + "rewards/margins": 10.875, + "rewards/rejected": -10.3125, + "step": 1094 + }, + { + "epoch": 1.343558282208589, + "grad_norm": 5.508962768344179, + "learning_rate": 8.367768595041323e-07, + "logits/chosen": -0.138671875, + "logits/rejected": -0.349609375, + "logps/chosen": -446.0, + "logps/rejected": -430.0, + "loss": 0.0202, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.890625, + "rewards/margins": 12.4375, + "rewards/rejected": -11.5, + "step": 1095 + }, + { + "epoch": 1.3447852760736196, + "grad_norm": 18.396109473702488, + "learning_rate": 8.361570247933885e-07, + "logits/chosen": -0.146484375, + "logits/rejected": -0.279296875, + "logps/chosen": -382.0, + "logps/rejected": -374.0, + "loss": 0.0744, + "rewards/accuracies": 0.9609375, + "rewards/chosen": -0.091796875, + "rewards/margins": 9.8125, + "rewards/rejected": -9.9375, + "step": 1096 + }, + { + "epoch": 1.3460122699386503, + "grad_norm": 4.9008028557309355, + "learning_rate": 8.355371900826447e-07, + "logits/chosen": -0.1015625, + "logits/rejected": -0.279296875, + "logps/chosen": -432.0, + "logps/rejected": -380.0, + "loss": 0.0189, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.66015625, + "rewards/margins": 10.6875, + "rewards/rejected": -10.0625, + "step": 1097 + }, + { + "epoch": 1.347239263803681, + "grad_norm": 17.180940504647815, + "learning_rate": 8.349173553719008e-07, + "logits/chosen": -0.0299072265625, + "logits/rejected": -0.1826171875, + "logps/chosen": -364.0, + "logps/rejected": -384.0, + "loss": 0.0604, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.1845703125, + "rewards/margins": 10.8125, + "rewards/rejected": -10.625, + "step": 1098 + }, + { + "epoch": 1.3484662576687116, + "grad_norm": 15.090380221292753, + "learning_rate": 8.34297520661157e-07, + "logits/chosen": -0.09619140625, + "logits/rejected": -0.2080078125, + "logps/chosen": -354.0, + "logps/rejected": -344.0, + "loss": 0.0831, + "rewards/accuracies": 0.984375, + "rewards/chosen": -0.130859375, + "rewards/margins": 9.8125, + "rewards/rejected": -9.9375, + "step": 1099 + }, + { + "epoch": 1.3496932515337423, + "grad_norm": 8.957606377667311, + "learning_rate": 8.336776859504132e-07, + "logits/chosen": -0.1572265625, + "logits/rejected": -0.37890625, + "logps/chosen": -406.0, + "logps/rejected": -374.0, + "loss": 0.0292, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.92578125, + "rewards/margins": 10.875, + "rewards/rejected": -9.9375, + "step": 1100 + }, + { + "epoch": 1.350920245398773, + "grad_norm": 12.12291867951929, + "learning_rate": 8.330578512396695e-07, + "logits/chosen": 0.0400390625, + "logits/rejected": -0.1171875, + "logps/chosen": -370.0, + "logps/rejected": -348.0, + "loss": 0.0654, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.4921875, + "rewards/margins": 10.75, + "rewards/rejected": -10.25, + "step": 1101 + }, + { + "epoch": 1.3521472392638036, + "grad_norm": 18.332708975287847, + "learning_rate": 8.324380165289256e-07, + "logits/chosen": -0.251953125, + "logits/rejected": -0.431640625, + "logps/chosen": -370.0, + "logps/rejected": -352.0, + "loss": 0.0716, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.671875, + "rewards/margins": 11.75, + "rewards/rejected": -11.0625, + "step": 1102 + }, + { + "epoch": 1.3533742331288343, + "grad_norm": 10.7430697279738, + "learning_rate": 8.318181818181819e-07, + "logits/chosen": -0.119140625, + "logits/rejected": -0.275390625, + "logps/chosen": -416.0, + "logps/rejected": -400.0, + "loss": 0.0442, + "rewards/accuracies": 0.9765625, + "rewards/chosen": 0.3984375, + "rewards/margins": 10.8125, + "rewards/rejected": -10.4375, + "step": 1103 + }, + { + "epoch": 1.354601226993865, + "grad_norm": 12.033995818293517, + "learning_rate": 8.311983471074381e-07, + "logits/chosen": -0.078125, + "logits/rejected": -0.2265625, + "logps/chosen": -364.0, + "logps/rejected": -374.0, + "loss": 0.0536, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.4296875, + "rewards/margins": 10.375, + "rewards/rejected": -9.9375, + "step": 1104 + }, + { + "epoch": 1.3558282208588956, + "grad_norm": 4.495393110896528, + "learning_rate": 8.305785123966943e-07, + "logits/chosen": -0.2177734375, + "logits/rejected": -0.361328125, + "logps/chosen": -408.0, + "logps/rejected": -414.0, + "loss": 0.0158, + "rewards/accuracies": 1.0, + "rewards/chosen": 1.09375, + "rewards/margins": 11.3125, + "rewards/rejected": -10.25, + "step": 1105 + }, + { + "epoch": 1.3570552147239263, + "grad_norm": 10.27676045940915, + "learning_rate": 8.299586776859504e-07, + "logits/chosen": -0.150390625, + "logits/rejected": -0.263671875, + "logps/chosen": -360.0, + "logps/rejected": -374.0, + "loss": 0.0323, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.177734375, + "rewards/margins": 10.6875, + "rewards/rejected": -10.5, + "step": 1106 + }, + { + "epoch": 1.358282208588957, + "grad_norm": 15.415657480989857, + "learning_rate": 8.293388429752065e-07, + "logits/chosen": 0.03759765625, + "logits/rejected": -0.07861328125, + "logps/chosen": -402.0, + "logps/rejected": -374.0, + "loss": 0.0477, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.76953125, + "rewards/margins": 10.3125, + "rewards/rejected": -9.5625, + "step": 1107 + }, + { + "epoch": 1.3595092024539877, + "grad_norm": 5.235763636929729, + "learning_rate": 8.287190082644628e-07, + "logits/chosen": -0.1552734375, + "logits/rejected": -0.3359375, + "logps/chosen": -440.0, + "logps/rejected": -402.0, + "loss": 0.0197, + "rewards/accuracies": 1.0, + "rewards/chosen": 1.1484375, + "rewards/margins": 12.1875, + "rewards/rejected": -11.0, + "step": 1108 + }, + { + "epoch": 1.3607361963190185, + "grad_norm": 10.622377911792848, + "learning_rate": 8.28099173553719e-07, + "logits/chosen": -0.0361328125, + "logits/rejected": -0.1787109375, + "logps/chosen": -400.0, + "logps/rejected": -382.0, + "loss": 0.0409, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.6328125, + "rewards/margins": 10.5, + "rewards/rejected": -9.875, + "step": 1109 + }, + { + "epoch": 1.3619631901840492, + "grad_norm": 15.219013654594981, + "learning_rate": 8.274793388429753e-07, + "logits/chosen": -0.11572265625, + "logits/rejected": -0.330078125, + "logps/chosen": -398.0, + "logps/rejected": -392.0, + "loss": 0.0485, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.482421875, + "rewards/margins": 11.4375, + "rewards/rejected": -10.9375, + "step": 1110 + }, + { + "epoch": 1.3631901840490799, + "grad_norm": 5.052963784815555, + "learning_rate": 8.268595041322314e-07, + "logits/chosen": -0.150390625, + "logits/rejected": -0.333984375, + "logps/chosen": -432.0, + "logps/rejected": -380.0, + "loss": 0.0166, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.6640625, + "rewards/margins": 11.3125, + "rewards/rejected": -10.625, + "step": 1111 + }, + { + "epoch": 1.3644171779141105, + "grad_norm": 13.661073432102611, + "learning_rate": 8.262396694214877e-07, + "logits/chosen": -0.138671875, + "logits/rejected": -0.23046875, + "logps/chosen": -418.0, + "logps/rejected": -406.0, + "loss": 0.0562, + "rewards/accuracies": 0.96875, + "rewards/chosen": 0.298828125, + "rewards/margins": 11.0625, + "rewards/rejected": -10.75, + "step": 1112 + }, + { + "epoch": 1.3656441717791412, + "grad_norm": 11.164562514831367, + "learning_rate": 8.256198347107439e-07, + "logits/chosen": -0.10888671875, + "logits/rejected": -0.16015625, + "logps/chosen": -356.0, + "logps/rejected": -360.0, + "loss": 0.0353, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.263671875, + "rewards/margins": 10.1875, + "rewards/rejected": -9.9375, + "step": 1113 + }, + { + "epoch": 1.3668711656441719, + "grad_norm": 12.65975505597321, + "learning_rate": 8.25e-07, + "logits/chosen": -0.25390625, + "logits/rejected": -0.3671875, + "logps/chosen": -344.0, + "logps/rejected": -358.0, + "loss": 0.0419, + "rewards/accuracies": 0.9765625, + "rewards/chosen": -0.45703125, + "rewards/margins": 10.0625, + "rewards/rejected": -10.5, + "step": 1114 + }, + { + "epoch": 1.3680981595092025, + "grad_norm": 9.32147436119456, + "learning_rate": 8.243801652892562e-07, + "logits/chosen": -0.1396484375, + "logits/rejected": -0.25, + "logps/chosen": -376.0, + "logps/rejected": -368.0, + "loss": 0.0374, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.388671875, + "rewards/margins": 11.1875, + "rewards/rejected": -10.8125, + "step": 1115 + }, + { + "epoch": 1.3693251533742332, + "grad_norm": 7.035037640219457, + "learning_rate": 8.237603305785123e-07, + "logits/chosen": -0.06982421875, + "logits/rejected": -0.2490234375, + "logps/chosen": -430.0, + "logps/rejected": -402.0, + "loss": 0.0362, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.326171875, + "rewards/margins": 12.0, + "rewards/rejected": -11.6875, + "step": 1116 + }, + { + "epoch": 1.3705521472392639, + "grad_norm": 7.072591615437579, + "learning_rate": 8.231404958677686e-07, + "logits/chosen": -0.251953125, + "logits/rejected": -0.47265625, + "logps/chosen": -434.0, + "logps/rejected": -400.0, + "loss": 0.028, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.671875, + "rewards/margins": 11.875, + "rewards/rejected": -11.1875, + "step": 1117 + }, + { + "epoch": 1.3717791411042946, + "grad_norm": 15.059088219830123, + "learning_rate": 8.225206611570248e-07, + "logits/chosen": -0.056884765625, + "logits/rejected": -0.2294921875, + "logps/chosen": -380.0, + "logps/rejected": -382.0, + "loss": 0.0791, + "rewards/accuracies": 0.9609375, + "rewards/chosen": -0.40234375, + "rewards/margins": 10.5, + "rewards/rejected": -10.875, + "step": 1118 + }, + { + "epoch": 1.3730061349693252, + "grad_norm": 7.182852167794888, + "learning_rate": 8.21900826446281e-07, + "logits/chosen": -0.150390625, + "logits/rejected": -0.328125, + "logps/chosen": -424.0, + "logps/rejected": -420.0, + "loss": 0.0365, + "rewards/accuracies": 0.984375, + "rewards/chosen": -0.271484375, + "rewards/margins": 10.6875, + "rewards/rejected": -11.0, + "step": 1119 + }, + { + "epoch": 1.3742331288343559, + "grad_norm": 6.6804769663365855, + "learning_rate": 8.212809917355372e-07, + "logits/chosen": -0.1103515625, + "logits/rejected": -0.400390625, + "logps/chosen": -430.0, + "logps/rejected": -398.0, + "loss": 0.0239, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.58984375, + "rewards/margins": 12.0, + "rewards/rejected": -11.4375, + "step": 1120 + }, + { + "epoch": 1.3754601226993866, + "grad_norm": 16.092490568103823, + "learning_rate": 8.206611570247935e-07, + "logits/chosen": -0.216796875, + "logits/rejected": -0.375, + "logps/chosen": -444.0, + "logps/rejected": -406.0, + "loss": 0.0841, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.1826171875, + "rewards/margins": 11.3125, + "rewards/rejected": -11.125, + "step": 1121 + }, + { + "epoch": 1.3766871165644172, + "grad_norm": 45.62488030196555, + "learning_rate": 8.200413223140497e-07, + "logits/chosen": -0.2314453125, + "logits/rejected": -0.361328125, + "logps/chosen": -372.0, + "logps/rejected": -374.0, + "loss": 0.1206, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -1.2421875, + "rewards/margins": 10.1875, + "rewards/rejected": -11.375, + "step": 1122 + }, + { + "epoch": 1.377914110429448, + "grad_norm": 7.987049229922856, + "learning_rate": 8.194214876033058e-07, + "logits/chosen": -0.23046875, + "logits/rejected": -0.380859375, + "logps/chosen": -446.0, + "logps/rejected": -384.0, + "loss": 0.0291, + "rewards/accuracies": 0.984375, + "rewards/chosen": -0.59765625, + "rewards/margins": 11.0, + "rewards/rejected": -11.625, + "step": 1123 + }, + { + "epoch": 1.3791411042944786, + "grad_norm": 10.84925435010219, + "learning_rate": 8.188016528925619e-07, + "logits/chosen": -0.1494140625, + "logits/rejected": -0.33203125, + "logps/chosen": -398.0, + "logps/rejected": -372.0, + "loss": 0.045, + "rewards/accuracies": 0.9765625, + "rewards/chosen": -0.53515625, + "rewards/margins": 11.4375, + "rewards/rejected": -12.0, + "step": 1124 + }, + { + "epoch": 1.3803680981595092, + "grad_norm": 12.438670127459194, + "learning_rate": 8.181818181818181e-07, + "logits/chosen": -0.2197265625, + "logits/rejected": -0.3671875, + "logps/chosen": -388.0, + "logps/rejected": -382.0, + "loss": 0.0514, + "rewards/accuracies": 0.96875, + "rewards/chosen": -0.70703125, + "rewards/margins": 10.6875, + "rewards/rejected": -11.4375, + "step": 1125 + }, + { + "epoch": 1.38159509202454, + "grad_norm": 7.702209911740994, + "learning_rate": 8.175619834710744e-07, + "logits/chosen": -0.21875, + "logits/rejected": -0.345703125, + "logps/chosen": -378.0, + "logps/rejected": -372.0, + "loss": 0.0244, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.5, + "rewards/margins": 11.4375, + "rewards/rejected": -11.9375, + "step": 1126 + }, + { + "epoch": 1.3828220858895706, + "grad_norm": 9.98545953456932, + "learning_rate": 8.169421487603306e-07, + "logits/chosen": -0.177734375, + "logits/rejected": -0.392578125, + "logps/chosen": -452.0, + "logps/rejected": -438.0, + "loss": 0.0506, + "rewards/accuracies": 0.9765625, + "rewards/chosen": 0.515625, + "rewards/margins": 12.0625, + "rewards/rejected": -11.5625, + "step": 1127 + }, + { + "epoch": 1.3840490797546012, + "grad_norm": 14.966530393829544, + "learning_rate": 8.163223140495868e-07, + "logits/chosen": -0.1396484375, + "logits/rejected": -0.294921875, + "logps/chosen": -398.0, + "logps/rejected": -416.0, + "loss": 0.0573, + "rewards/accuracies": 0.96875, + "rewards/chosen": -0.83203125, + "rewards/margins": 10.0625, + "rewards/rejected": -10.875, + "step": 1128 + }, + { + "epoch": 1.385276073619632, + "grad_norm": 6.960083817877731, + "learning_rate": 8.15702479338843e-07, + "logits/chosen": -0.0244140625, + "logits/rejected": -0.166015625, + "logps/chosen": -400.0, + "logps/rejected": -372.0, + "loss": 0.022, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.29296875, + "rewards/margins": 10.625, + "rewards/rejected": -10.9375, + "step": 1129 + }, + { + "epoch": 1.3865030674846626, + "grad_norm": 15.732715609813184, + "learning_rate": 8.150826446280993e-07, + "logits/chosen": 0.0286865234375, + "logits/rejected": -0.1923828125, + "logps/chosen": -438.0, + "logps/rejected": -408.0, + "loss": 0.0498, + "rewards/accuracies": 0.984375, + "rewards/chosen": -0.361328125, + "rewards/margins": 10.5625, + "rewards/rejected": -10.9375, + "step": 1130 + }, + { + "epoch": 1.3877300613496932, + "grad_norm": 5.221583127777315, + "learning_rate": 8.144628099173554e-07, + "logits/chosen": -0.232421875, + "logits/rejected": -0.5234375, + "logps/chosen": -434.0, + "logps/rejected": -400.0, + "loss": 0.0197, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.04541015625, + "rewards/margins": 10.9375, + "rewards/rejected": -10.9375, + "step": 1131 + }, + { + "epoch": 1.388957055214724, + "grad_norm": 12.813444795393385, + "learning_rate": 8.138429752066116e-07, + "logits/chosen": -0.134765625, + "logits/rejected": -0.36328125, + "logps/chosen": -494.0, + "logps/rejected": -408.0, + "loss": 0.0313, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.890625, + "rewards/margins": 11.3125, + "rewards/rejected": -10.375, + "step": 1132 + }, + { + "epoch": 1.3901840490797546, + "grad_norm": 15.660321100678845, + "learning_rate": 8.132231404958677e-07, + "logits/chosen": -0.2099609375, + "logits/rejected": -0.34765625, + "logps/chosen": -410.0, + "logps/rejected": -410.0, + "loss": 0.0504, + "rewards/accuracies": 0.9765625, + "rewards/chosen": 0.72265625, + "rewards/margins": 11.875, + "rewards/rejected": -11.125, + "step": 1133 + }, + { + "epoch": 1.3914110429447852, + "grad_norm": 7.9073668160942745, + "learning_rate": 8.126033057851239e-07, + "logits/chosen": -0.1962890625, + "logits/rejected": -0.404296875, + "logps/chosen": -324.0, + "logps/rejected": -340.0, + "loss": 0.0393, + "rewards/accuracies": 0.984375, + "rewards/chosen": -0.91015625, + "rewards/margins": 10.0625, + "rewards/rejected": -10.9375, + "step": 1134 + }, + { + "epoch": 1.392638036809816, + "grad_norm": 7.004247584369281, + "learning_rate": 8.119834710743802e-07, + "logits/chosen": -0.146484375, + "logits/rejected": -0.3359375, + "logps/chosen": -410.0, + "logps/rejected": -374.0, + "loss": 0.0317, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.69140625, + "rewards/margins": 11.5625, + "rewards/rejected": -10.875, + "step": 1135 + }, + { + "epoch": 1.3938650306748466, + "grad_norm": 13.652544230864027, + "learning_rate": 8.113636363636363e-07, + "logits/chosen": -0.29296875, + "logits/rejected": -0.4375, + "logps/chosen": -466.0, + "logps/rejected": -408.0, + "loss": 0.0464, + "rewards/accuracies": 0.96875, + "rewards/chosen": 1.1796875, + "rewards/margins": 12.125, + "rewards/rejected": -10.9375, + "step": 1136 + }, + { + "epoch": 1.3950920245398772, + "grad_norm": 13.746874338763654, + "learning_rate": 8.107438016528926e-07, + "logits/chosen": -0.2314453125, + "logits/rejected": -0.48046875, + "logps/chosen": -466.0, + "logps/rejected": -406.0, + "loss": 0.0681, + "rewards/accuracies": 0.9765625, + "rewards/chosen": 0.345703125, + "rewards/margins": 10.9375, + "rewards/rejected": -10.5625, + "step": 1137 + }, + { + "epoch": 1.396319018404908, + "grad_norm": 13.137416168579717, + "learning_rate": 8.101239669421488e-07, + "logits/chosen": -0.05126953125, + "logits/rejected": -0.25, + "logps/chosen": -386.0, + "logps/rejected": -356.0, + "loss": 0.0423, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.051513671875, + "rewards/margins": 9.875, + "rewards/rejected": -9.8125, + "step": 1138 + }, + { + "epoch": 1.3975460122699386, + "grad_norm": 14.313209901423209, + "learning_rate": 8.095041322314051e-07, + "logits/chosen": -0.2001953125, + "logits/rejected": -0.404296875, + "logps/chosen": -360.0, + "logps/rejected": -368.0, + "loss": 0.0629, + "rewards/accuracies": 0.96875, + "rewards/chosen": 0.2392578125, + "rewards/margins": 11.1875, + "rewards/rejected": -10.9375, + "step": 1139 + }, + { + "epoch": 1.3987730061349692, + "grad_norm": 12.684020520423767, + "learning_rate": 8.088842975206612e-07, + "logits/chosen": -0.10498046875, + "logits/rejected": -0.232421875, + "logps/chosen": -352.0, + "logps/rejected": -360.0, + "loss": 0.0614, + "rewards/accuracies": 0.984375, + "rewards/chosen": -0.2294921875, + "rewards/margins": 10.625, + "rewards/rejected": -10.8125, + "step": 1140 + }, + { + "epoch": 1.4, + "grad_norm": 8.366317687764122, + "learning_rate": 8.082644628099174e-07, + "logits/chosen": -0.1689453125, + "logits/rejected": -0.396484375, + "logps/chosen": -354.0, + "logps/rejected": -356.0, + "loss": 0.0322, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.154296875, + "rewards/margins": 10.4375, + "rewards/rejected": -10.25, + "step": 1141 + }, + { + "epoch": 1.4012269938650306, + "grad_norm": 9.78962851865542, + "learning_rate": 8.076446280991735e-07, + "logits/chosen": -0.057861328125, + "logits/rejected": -0.2001953125, + "logps/chosen": -346.0, + "logps/rejected": -382.0, + "loss": 0.0316, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.34765625, + "rewards/margins": 11.0, + "rewards/rejected": -10.625, + "step": 1142 + }, + { + "epoch": 1.4024539877300612, + "grad_norm": 15.82156683988277, + "learning_rate": 8.070247933884297e-07, + "logits/chosen": -0.1328125, + "logits/rejected": -0.248046875, + "logps/chosen": -396.0, + "logps/rejected": -416.0, + "loss": 0.0618, + "rewards/accuracies": 0.9765625, + "rewards/chosen": -0.51953125, + "rewards/margins": 10.75, + "rewards/rejected": -11.25, + "step": 1143 + }, + { + "epoch": 1.403680981595092, + "grad_norm": 17.350320312048925, + "learning_rate": 8.06404958677686e-07, + "logits/chosen": -0.255859375, + "logits/rejected": -0.38671875, + "logps/chosen": -396.0, + "logps/rejected": -388.0, + "loss": 0.0738, + "rewards/accuracies": 0.984375, + "rewards/chosen": -0.169921875, + "rewards/margins": 10.9375, + "rewards/rejected": -11.125, + "step": 1144 + }, + { + "epoch": 1.4049079754601226, + "grad_norm": 13.43693939499879, + "learning_rate": 8.057851239669421e-07, + "logits/chosen": -0.1572265625, + "logits/rejected": -0.333984375, + "logps/chosen": -456.0, + "logps/rejected": -434.0, + "loss": 0.0376, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.86328125, + "rewards/margins": 11.8125, + "rewards/rejected": -10.9375, + "step": 1145 + }, + { + "epoch": 1.4061349693251532, + "grad_norm": 10.84486528809094, + "learning_rate": 8.051652892561984e-07, + "logits/chosen": -0.046875, + "logits/rejected": -0.16796875, + "logps/chosen": -376.0, + "logps/rejected": -400.0, + "loss": 0.0344, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.05322265625, + "rewards/margins": 11.625, + "rewards/rejected": -11.6875, + "step": 1146 + }, + { + "epoch": 1.407361963190184, + "grad_norm": 7.584140894053648, + "learning_rate": 8.045454545454546e-07, + "logits/chosen": -0.1572265625, + "logits/rejected": -0.361328125, + "logps/chosen": -394.0, + "logps/rejected": -386.0, + "loss": 0.0283, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.5, + "rewards/margins": 12.125, + "rewards/rejected": -11.625, + "step": 1147 + }, + { + "epoch": 1.4085889570552146, + "grad_norm": 9.922196775746215, + "learning_rate": 8.039256198347108e-07, + "logits/chosen": -0.0849609375, + "logits/rejected": -0.2353515625, + "logps/chosen": -370.0, + "logps/rejected": -378.0, + "loss": 0.0426, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.052978515625, + "rewards/margins": 11.3125, + "rewards/rejected": -11.25, + "step": 1148 + }, + { + "epoch": 1.4098159509202455, + "grad_norm": 6.148272524003038, + "learning_rate": 8.03305785123967e-07, + "logits/chosen": -0.140625, + "logits/rejected": -0.30859375, + "logps/chosen": -384.0, + "logps/rejected": -398.0, + "loss": 0.0231, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.35546875, + "rewards/margins": 11.8125, + "rewards/rejected": -11.5, + "step": 1149 + }, + { + "epoch": 1.4110429447852761, + "grad_norm": 6.406336886570724, + "learning_rate": 8.026859504132232e-07, + "logits/chosen": -0.06689453125, + "logits/rejected": -0.240234375, + "logps/chosen": -390.0, + "logps/rejected": -386.0, + "loss": 0.0284, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.0020904541015625, + "rewards/margins": 11.125, + "rewards/rejected": -11.125, + "step": 1150 + }, + { + "epoch": 1.4122699386503068, + "grad_norm": 5.500976110615787, + "learning_rate": 8.020661157024793e-07, + "logits/chosen": -0.1259765625, + "logits/rejected": -0.330078125, + "logps/chosen": -388.0, + "logps/rejected": -406.0, + "loss": 0.022, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.138671875, + "rewards/margins": 11.625, + "rewards/rejected": -11.75, + "step": 1151 + }, + { + "epoch": 1.4134969325153375, + "grad_norm": 15.46899721034435, + "learning_rate": 8.014462809917355e-07, + "logits/chosen": -0.2041015625, + "logits/rejected": -0.318359375, + "logps/chosen": -420.0, + "logps/rejected": -436.0, + "loss": 0.0474, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.5859375, + "rewards/margins": 12.25, + "rewards/rejected": -11.6875, + "step": 1152 + }, + { + "epoch": 1.4147239263803681, + "grad_norm": 29.448301297633176, + "learning_rate": 8.008264462809917e-07, + "logits/chosen": -0.12109375, + "logits/rejected": -0.26171875, + "logps/chosen": -406.0, + "logps/rejected": -406.0, + "loss": 0.0778, + "rewards/accuracies": 0.9765625, + "rewards/chosen": 0.349609375, + "rewards/margins": 11.375, + "rewards/rejected": -11.0, + "step": 1153 + }, + { + "epoch": 1.4159509202453988, + "grad_norm": 8.55335933719368, + "learning_rate": 8.002066115702479e-07, + "logits/chosen": -0.2041015625, + "logits/rejected": -0.373046875, + "logps/chosen": -446.0, + "logps/rejected": -424.0, + "loss": 0.0282, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.62109375, + "rewards/margins": 12.6875, + "rewards/rejected": -12.0625, + "step": 1154 + }, + { + "epoch": 1.4171779141104295, + "grad_norm": 11.91106417252758, + "learning_rate": 7.995867768595042e-07, + "logits/chosen": -0.1962890625, + "logits/rejected": -0.310546875, + "logps/chosen": -378.0, + "logps/rejected": -386.0, + "loss": 0.0437, + "rewards/accuracies": 0.984375, + "rewards/chosen": -0.146484375, + "rewards/margins": 10.8125, + "rewards/rejected": -10.9375, + "step": 1155 + }, + { + "epoch": 1.4184049079754601, + "grad_norm": 14.345078375136291, + "learning_rate": 7.989669421487604e-07, + "logits/chosen": -0.1103515625, + "logits/rejected": -0.369140625, + "logps/chosen": -440.0, + "logps/rejected": -396.0, + "loss": 0.0455, + "rewards/accuracies": 0.9765625, + "rewards/chosen": 0.7421875, + "rewards/margins": 12.3125, + "rewards/rejected": -11.625, + "step": 1156 + }, + { + "epoch": 1.4196319018404908, + "grad_norm": 13.702503887945511, + "learning_rate": 7.983471074380166e-07, + "logits/chosen": -0.150390625, + "logits/rejected": -0.28125, + "logps/chosen": -450.0, + "logps/rejected": -424.0, + "loss": 0.0482, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.93359375, + "rewards/margins": 11.9375, + "rewards/rejected": -11.0, + "step": 1157 + }, + { + "epoch": 1.4208588957055215, + "grad_norm": 8.654555323915172, + "learning_rate": 7.977272727272728e-07, + "logits/chosen": -0.1650390625, + "logits/rejected": -0.306640625, + "logps/chosen": -346.0, + "logps/rejected": -354.0, + "loss": 0.0362, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.498046875, + "rewards/margins": 11.375, + "rewards/rejected": -10.875, + "step": 1158 + }, + { + "epoch": 1.4220858895705522, + "grad_norm": 14.010487033573805, + "learning_rate": 7.97107438016529e-07, + "logits/chosen": -0.11376953125, + "logits/rejected": -0.1875, + "logps/chosen": -354.0, + "logps/rejected": -388.0, + "loss": 0.0558, + "rewards/accuracies": 0.9765625, + "rewards/chosen": -0.10986328125, + "rewards/margins": 10.6875, + "rewards/rejected": -10.8125, + "step": 1159 + }, + { + "epoch": 1.4233128834355828, + "grad_norm": 16.133153822756856, + "learning_rate": 7.964876033057851e-07, + "logits/chosen": -0.2373046875, + "logits/rejected": -0.3984375, + "logps/chosen": -414.0, + "logps/rejected": -398.0, + "loss": 0.0599, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.23828125, + "rewards/margins": 11.125, + "rewards/rejected": -10.875, + "step": 1160 + }, + { + "epoch": 1.4245398773006135, + "grad_norm": 9.461066515172439, + "learning_rate": 7.958677685950413e-07, + "logits/chosen": -0.0791015625, + "logits/rejected": -0.1962890625, + "logps/chosen": -378.0, + "logps/rejected": -414.0, + "loss": 0.07, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.07421875, + "rewards/margins": 10.6875, + "rewards/rejected": -10.625, + "step": 1161 + }, + { + "epoch": 1.4257668711656442, + "grad_norm": 13.085163631657917, + "learning_rate": 7.952479338842975e-07, + "logits/chosen": -0.1923828125, + "logits/rejected": -0.314453125, + "logps/chosen": -394.0, + "logps/rejected": -382.0, + "loss": 0.0558, + "rewards/accuracies": 0.9765625, + "rewards/chosen": 0.059814453125, + "rewards/margins": 10.8125, + "rewards/rejected": -10.75, + "step": 1162 + }, + { + "epoch": 1.4269938650306748, + "grad_norm": 9.56097265388881, + "learning_rate": 7.946280991735537e-07, + "logits/chosen": -0.279296875, + "logits/rejected": -0.52734375, + "logps/chosen": -402.0, + "logps/rejected": -378.0, + "loss": 0.0236, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.341796875, + "rewards/margins": 11.5625, + "rewards/rejected": -11.25, + "step": 1163 + }, + { + "epoch": 1.4282208588957055, + "grad_norm": 16.722388625925955, + "learning_rate": 7.9400826446281e-07, + "logits/chosen": -0.091796875, + "logits/rejected": -0.126953125, + "logps/chosen": -336.0, + "logps/rejected": -368.0, + "loss": 0.0682, + "rewards/accuracies": 0.96875, + "rewards/chosen": -0.017822265625, + "rewards/margins": 10.4375, + "rewards/rejected": -10.5, + "step": 1164 + }, + { + "epoch": 1.4294478527607362, + "grad_norm": 7.348322611328401, + "learning_rate": 7.933884297520661e-07, + "logits/chosen": -0.0986328125, + "logits/rejected": -0.2373046875, + "logps/chosen": -388.0, + "logps/rejected": -370.0, + "loss": 0.0227, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.08154296875, + "rewards/margins": 10.75, + "rewards/rejected": -10.625, + "step": 1165 + }, + { + "epoch": 1.4306748466257668, + "grad_norm": 5.34988661629868, + "learning_rate": 7.927685950413224e-07, + "logits/chosen": -0.138671875, + "logits/rejected": -0.369140625, + "logps/chosen": -410.0, + "logps/rejected": -386.0, + "loss": 0.0213, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.2353515625, + "rewards/margins": 10.0625, + "rewards/rejected": -10.3125, + "step": 1166 + }, + { + "epoch": 1.4319018404907975, + "grad_norm": 2.351503468437324, + "learning_rate": 7.921487603305786e-07, + "logits/chosen": -0.1630859375, + "logits/rejected": -0.3515625, + "logps/chosen": -404.0, + "logps/rejected": -410.0, + "loss": 0.0096, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.2734375, + "rewards/margins": 11.6875, + "rewards/rejected": -11.4375, + "step": 1167 + }, + { + "epoch": 1.4331288343558282, + "grad_norm": 10.275425558023933, + "learning_rate": 7.915289256198348e-07, + "logits/chosen": -0.150390625, + "logits/rejected": -0.2734375, + "logps/chosen": -372.0, + "logps/rejected": -374.0, + "loss": 0.0434, + "rewards/accuracies": 0.9765625, + "rewards/chosen": -0.357421875, + "rewards/margins": 11.1875, + "rewards/rejected": -11.5625, + "step": 1168 + }, + { + "epoch": 1.4343558282208588, + "grad_norm": 9.69113419509306, + "learning_rate": 7.909090909090909e-07, + "logits/chosen": -0.07568359375, + "logits/rejected": -0.236328125, + "logps/chosen": -376.0, + "logps/rejected": -398.0, + "loss": 0.0378, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.390625, + "rewards/margins": 10.25, + "rewards/rejected": -9.875, + "step": 1169 + }, + { + "epoch": 1.4355828220858895, + "grad_norm": 11.100318005337881, + "learning_rate": 7.90289256198347e-07, + "logits/chosen": -0.171875, + "logits/rejected": -0.421875, + "logps/chosen": -396.0, + "logps/rejected": -364.0, + "loss": 0.0327, + "rewards/accuracies": 0.984375, + "rewards/chosen": -0.095703125, + "rewards/margins": 10.6875, + "rewards/rejected": -10.75, + "step": 1170 + }, + { + "epoch": 1.4368098159509202, + "grad_norm": 13.42384166771764, + "learning_rate": 7.896694214876033e-07, + "logits/chosen": -0.126953125, + "logits/rejected": -0.28515625, + "logps/chosen": -380.0, + "logps/rejected": -372.0, + "loss": 0.0555, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.400390625, + "rewards/margins": 10.6875, + "rewards/rejected": -10.3125, + "step": 1171 + }, + { + "epoch": 1.438036809815951, + "grad_norm": 16.33374781104785, + "learning_rate": 7.890495867768595e-07, + "logits/chosen": -0.055419921875, + "logits/rejected": -0.173828125, + "logps/chosen": -436.0, + "logps/rejected": -400.0, + "loss": 0.0706, + "rewards/accuracies": 0.96875, + "rewards/chosen": -0.2421875, + "rewards/margins": 9.25, + "rewards/rejected": -9.4375, + "step": 1172 + }, + { + "epoch": 1.4392638036809817, + "grad_norm": 17.67963845713936, + "learning_rate": 7.884297520661158e-07, + "logits/chosen": -0.125, + "logits/rejected": -0.275390625, + "logps/chosen": -428.0, + "logps/rejected": -410.0, + "loss": 0.0608, + "rewards/accuracies": 0.9765625, + "rewards/chosen": 0.275390625, + "rewards/margins": 10.875, + "rewards/rejected": -10.625, + "step": 1173 + }, + { + "epoch": 1.4404907975460124, + "grad_norm": 5.1384201359953, + "learning_rate": 7.878099173553719e-07, + "logits/chosen": -0.138671875, + "logits/rejected": -0.166015625, + "logps/chosen": -362.0, + "logps/rejected": -366.0, + "loss": 0.0129, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.017578125, + "rewards/margins": 10.625, + "rewards/rejected": -10.5625, + "step": 1174 + }, + { + "epoch": 1.441717791411043, + "grad_norm": 5.148940884458012, + "learning_rate": 7.871900826446282e-07, + "logits/chosen": -0.1083984375, + "logits/rejected": -0.173828125, + "logps/chosen": -346.0, + "logps/rejected": -364.0, + "loss": 0.0273, + "rewards/accuracies": 0.984375, + "rewards/chosen": -0.35546875, + "rewards/margins": 10.0, + "rewards/rejected": -10.375, + "step": 1175 + }, + { + "epoch": 1.4429447852760737, + "grad_norm": 11.603523639095071, + "learning_rate": 7.865702479338844e-07, + "logits/chosen": -0.1923828125, + "logits/rejected": -0.310546875, + "logps/chosen": -392.0, + "logps/rejected": -402.0, + "loss": 0.0389, + "rewards/accuracies": 0.984375, + "rewards/chosen": -0.306640625, + "rewards/margins": 9.875, + "rewards/rejected": -10.1875, + "step": 1176 + }, + { + "epoch": 1.4441717791411044, + "grad_norm": 10.040926511598048, + "learning_rate": 7.859504132231406e-07, + "logits/chosen": -0.0517578125, + "logits/rejected": -0.244140625, + "logps/chosen": -384.0, + "logps/rejected": -354.0, + "loss": 0.0383, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.04296875, + "rewards/margins": 9.5, + "rewards/rejected": -9.5, + "step": 1177 + }, + { + "epoch": 1.445398773006135, + "grad_norm": 11.028023088957683, + "learning_rate": 7.853305785123967e-07, + "logits/chosen": -0.1484375, + "logits/rejected": -0.35546875, + "logps/chosen": -396.0, + "logps/rejected": -364.0, + "loss": 0.0425, + "rewards/accuracies": 0.984375, + "rewards/chosen": 1.0859375, + "rewards/margins": 10.9375, + "rewards/rejected": -9.875, + "step": 1178 + }, + { + "epoch": 1.4466257668711657, + "grad_norm": 8.150158411239879, + "learning_rate": 7.847107438016528e-07, + "logits/chosen": -0.2412109375, + "logits/rejected": -0.443359375, + "logps/chosen": -438.0, + "logps/rejected": -394.0, + "loss": 0.0374, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.515625, + "rewards/margins": 10.75, + "rewards/rejected": -10.1875, + "step": 1179 + }, + { + "epoch": 1.4478527607361964, + "grad_norm": 6.894255501689617, + "learning_rate": 7.840909090909091e-07, + "logits/chosen": -0.1279296875, + "logits/rejected": -0.1748046875, + "logps/chosen": -344.0, + "logps/rejected": -386.0, + "loss": 0.0309, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.05029296875, + "rewards/margins": 10.75, + "rewards/rejected": -10.6875, + "step": 1180 + }, + { + "epoch": 1.449079754601227, + "grad_norm": 8.019249451553959, + "learning_rate": 7.834710743801653e-07, + "logits/chosen": -0.162109375, + "logits/rejected": -0.2890625, + "logps/chosen": -434.0, + "logps/rejected": -430.0, + "loss": 0.0239, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.87890625, + "rewards/margins": 10.8125, + "rewards/rejected": -9.9375, + "step": 1181 + }, + { + "epoch": 1.4503067484662577, + "grad_norm": 8.340607467826109, + "learning_rate": 7.828512396694216e-07, + "logits/chosen": -0.1083984375, + "logits/rejected": -0.2490234375, + "logps/chosen": -426.0, + "logps/rejected": -404.0, + "loss": 0.0307, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.427734375, + "rewards/margins": 10.1875, + "rewards/rejected": -10.625, + "step": 1182 + }, + { + "epoch": 1.4515337423312884, + "grad_norm": 11.72673950966423, + "learning_rate": 7.822314049586777e-07, + "logits/chosen": -0.142578125, + "logits/rejected": -0.19921875, + "logps/chosen": -326.0, + "logps/rejected": -372.0, + "loss": 0.0563, + "rewards/accuracies": 0.984375, + "rewards/chosen": -0.06787109375, + "rewards/margins": 9.875, + "rewards/rejected": -9.9375, + "step": 1183 + }, + { + "epoch": 1.452760736196319, + "grad_norm": 9.058514886167021, + "learning_rate": 7.81611570247934e-07, + "logits/chosen": -0.15234375, + "logits/rejected": -0.34765625, + "logps/chosen": -366.0, + "logps/rejected": -352.0, + "loss": 0.0407, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.330078125, + "rewards/margins": 10.9375, + "rewards/rejected": -10.625, + "step": 1184 + }, + { + "epoch": 1.4539877300613497, + "grad_norm": 11.422467086533596, + "learning_rate": 7.809917355371902e-07, + "logits/chosen": -0.041259765625, + "logits/rejected": -0.19140625, + "logps/chosen": -420.0, + "logps/rejected": -402.0, + "loss": 0.0448, + "rewards/accuracies": 0.9765625, + "rewards/chosen": 0.5625, + "rewards/margins": 10.8125, + "rewards/rejected": -10.25, + "step": 1185 + }, + { + "epoch": 1.4552147239263804, + "grad_norm": 9.735079419768947, + "learning_rate": 7.803719008264463e-07, + "logits/chosen": -0.11328125, + "logits/rejected": -0.263671875, + "logps/chosen": -378.0, + "logps/rejected": -390.0, + "loss": 0.0379, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.11328125, + "rewards/margins": 10.375, + "rewards/rejected": -10.25, + "step": 1186 + }, + { + "epoch": 1.456441717791411, + "grad_norm": 6.162772380389009, + "learning_rate": 7.797520661157025e-07, + "logits/chosen": -0.1396484375, + "logits/rejected": -0.3359375, + "logps/chosen": -456.0, + "logps/rejected": -410.0, + "loss": 0.0253, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.6875, + "rewards/margins": 10.375, + "rewards/rejected": -9.6875, + "step": 1187 + }, + { + "epoch": 1.4576687116564417, + "grad_norm": 15.719129543958939, + "learning_rate": 7.791322314049586e-07, + "logits/chosen": -0.0654296875, + "logits/rejected": -0.1787109375, + "logps/chosen": -432.0, + "logps/rejected": -434.0, + "loss": 0.0721, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.08447265625, + "rewards/margins": 10.25, + "rewards/rejected": -10.125, + "step": 1188 + }, + { + "epoch": 1.4588957055214724, + "grad_norm": 12.022319651435222, + "learning_rate": 7.785123966942149e-07, + "logits/chosen": -0.1689453125, + "logits/rejected": -0.349609375, + "logps/chosen": -428.0, + "logps/rejected": -400.0, + "loss": 0.082, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.380859375, + "rewards/margins": 10.875, + "rewards/rejected": -10.5, + "step": 1189 + }, + { + "epoch": 1.460122699386503, + "grad_norm": 10.515256662579745, + "learning_rate": 7.778925619834711e-07, + "logits/chosen": -0.173828125, + "logits/rejected": -0.39453125, + "logps/chosen": -436.0, + "logps/rejected": -410.0, + "loss": 0.0408, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.80078125, + "rewards/margins": 12.125, + "rewards/rejected": -11.3125, + "step": 1190 + }, + { + "epoch": 1.4613496932515337, + "grad_norm": 14.208293904428151, + "learning_rate": 7.772727272727273e-07, + "logits/chosen": 0.002960205078125, + "logits/rejected": -0.13671875, + "logps/chosen": -370.0, + "logps/rejected": -390.0, + "loss": 0.0511, + "rewards/accuracies": 0.96875, + "rewards/chosen": 0.140625, + "rewards/margins": 11.375, + "rewards/rejected": -11.25, + "step": 1191 + }, + { + "epoch": 1.4625766871165644, + "grad_norm": 4.726495134099289, + "learning_rate": 7.766528925619835e-07, + "logits/chosen": -0.1162109375, + "logits/rejected": -0.29296875, + "logps/chosen": -400.0, + "logps/rejected": -406.0, + "loss": 0.0213, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.0208740234375, + "rewards/margins": 11.4375, + "rewards/rejected": -11.375, + "step": 1192 + }, + { + "epoch": 1.463803680981595, + "grad_norm": 7.11137377163575, + "learning_rate": 7.760330578512398e-07, + "logits/chosen": -0.181640625, + "logits/rejected": -0.373046875, + "logps/chosen": -458.0, + "logps/rejected": -420.0, + "loss": 0.0249, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.44921875, + "rewards/margins": 11.4375, + "rewards/rejected": -11.0, + "step": 1193 + }, + { + "epoch": 1.4650306748466257, + "grad_norm": 7.389252182902934, + "learning_rate": 7.75413223140496e-07, + "logits/chosen": -0.15625, + "logits/rejected": -0.275390625, + "logps/chosen": -420.0, + "logps/rejected": -404.0, + "loss": 0.0133, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.326171875, + "rewards/margins": 10.9375, + "rewards/rejected": -10.625, + "step": 1194 + }, + { + "epoch": 1.4662576687116564, + "grad_norm": 10.994717707834955, + "learning_rate": 7.74793388429752e-07, + "logits/chosen": -0.12451171875, + "logits/rejected": -0.2451171875, + "logps/chosen": -402.0, + "logps/rejected": -410.0, + "loss": 0.0418, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.08251953125, + "rewards/margins": 10.6875, + "rewards/rejected": -10.8125, + "step": 1195 + }, + { + "epoch": 1.467484662576687, + "grad_norm": 8.858341862725984, + "learning_rate": 7.741735537190082e-07, + "logits/chosen": -0.04638671875, + "logits/rejected": -0.1708984375, + "logps/chosen": -390.0, + "logps/rejected": -392.0, + "loss": 0.0296, + "rewards/accuracies": 0.984375, + "rewards/chosen": -0.1884765625, + "rewards/margins": 10.125, + "rewards/rejected": -10.3125, + "step": 1196 + }, + { + "epoch": 1.4687116564417177, + "grad_norm": 15.001467453574548, + "learning_rate": 7.735537190082644e-07, + "logits/chosen": -0.0546875, + "logits/rejected": -0.2333984375, + "logps/chosen": -422.0, + "logps/rejected": -382.0, + "loss": 0.0572, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.05810546875, + "rewards/margins": 10.875, + "rewards/rejected": -10.8125, + "step": 1197 + }, + { + "epoch": 1.4699386503067484, + "grad_norm": 11.467996330718492, + "learning_rate": 7.729338842975207e-07, + "logits/chosen": 0.01531982421875, + "logits/rejected": -0.126953125, + "logps/chosen": -384.0, + "logps/rejected": -356.0, + "loss": 0.0766, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.07177734375, + "rewards/margins": 10.8125, + "rewards/rejected": -10.6875, + "step": 1198 + }, + { + "epoch": 1.471165644171779, + "grad_norm": 8.297507128344717, + "learning_rate": 7.723140495867769e-07, + "logits/chosen": -0.2041015625, + "logits/rejected": -0.3671875, + "logps/chosen": -394.0, + "logps/rejected": -376.0, + "loss": 0.0247, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.376953125, + "rewards/margins": 11.625, + "rewards/rejected": -11.1875, + "step": 1199 + }, + { + "epoch": 1.4723926380368098, + "grad_norm": 6.708462240020914, + "learning_rate": 7.716942148760331e-07, + "logits/chosen": -0.13671875, + "logits/rejected": -0.2890625, + "logps/chosen": -330.0, + "logps/rejected": -350.0, + "loss": 0.0204, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.1572265625, + "rewards/margins": 10.125, + "rewards/rejected": -10.25, + "step": 1200 + }, + { + "epoch": 1.4736196319018404, + "grad_norm": 10.300151959699189, + "learning_rate": 7.710743801652893e-07, + "logits/chosen": -0.041015625, + "logits/rejected": -0.18359375, + "logps/chosen": -378.0, + "logps/rejected": -410.0, + "loss": 0.0294, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.1494140625, + "rewards/margins": 10.8125, + "rewards/rejected": -10.6875, + "step": 1201 + }, + { + "epoch": 1.474846625766871, + "grad_norm": 10.130166213309426, + "learning_rate": 7.704545454545456e-07, + "logits/chosen": -0.1337890625, + "logits/rejected": -0.2734375, + "logps/chosen": -428.0, + "logps/rejected": -422.0, + "loss": 0.0315, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.81640625, + "rewards/margins": 11.5625, + "rewards/rejected": -10.75, + "step": 1202 + }, + { + "epoch": 1.4760736196319018, + "grad_norm": 9.219567443331156, + "learning_rate": 7.698347107438017e-07, + "logits/chosen": -0.2314453125, + "logits/rejected": -0.408203125, + "logps/chosen": -410.0, + "logps/rejected": -372.0, + "loss": 0.0433, + "rewards/accuracies": 0.9765625, + "rewards/chosen": 0.2060546875, + "rewards/margins": 11.125, + "rewards/rejected": -10.9375, + "step": 1203 + }, + { + "epoch": 1.4773006134969324, + "grad_norm": 9.51752284318554, + "learning_rate": 7.692148760330578e-07, + "logits/chosen": -0.08203125, + "logits/rejected": -0.30859375, + "logps/chosen": -374.0, + "logps/rejected": -392.0, + "loss": 0.041, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.546875, + "rewards/margins": 11.125, + "rewards/rejected": -10.625, + "step": 1204 + }, + { + "epoch": 1.478527607361963, + "grad_norm": 23.546428660464922, + "learning_rate": 7.68595041322314e-07, + "logits/chosen": 0.02734375, + "logits/rejected": -0.2021484375, + "logps/chosen": -364.0, + "logps/rejected": -378.0, + "loss": 0.1268, + "rewards/accuracies": 0.9453125, + "rewards/chosen": 0.2470703125, + "rewards/margins": 10.4375, + "rewards/rejected": -10.1875, + "step": 1205 + }, + { + "epoch": 1.4797546012269938, + "grad_norm": 6.082107723283603, + "learning_rate": 7.679752066115702e-07, + "logits/chosen": -0.02685546875, + "logits/rejected": -0.3046875, + "logps/chosen": -350.0, + "logps/rejected": -344.0, + "loss": 0.0275, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.0791015625, + "rewards/margins": 10.25, + "rewards/rejected": -10.1875, + "step": 1206 + }, + { + "epoch": 1.4809815950920244, + "grad_norm": 10.622559047787306, + "learning_rate": 7.673553719008265e-07, + "logits/chosen": -0.09375, + "logits/rejected": -0.27734375, + "logps/chosen": -406.0, + "logps/rejected": -348.0, + "loss": 0.0353, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.05859375, + "rewards/margins": 10.9375, + "rewards/rejected": -10.875, + "step": 1207 + }, + { + "epoch": 1.482208588957055, + "grad_norm": 13.683817866460686, + "learning_rate": 7.667355371900826e-07, + "logits/chosen": -0.1123046875, + "logits/rejected": -0.26953125, + "logps/chosen": -376.0, + "logps/rejected": -382.0, + "loss": 0.0619, + "rewards/accuracies": 0.953125, + "rewards/chosen": -0.11474609375, + "rewards/margins": 10.25, + "rewards/rejected": -10.3125, + "step": 1208 + }, + { + "epoch": 1.4834355828220858, + "grad_norm": 8.88934565774754, + "learning_rate": 7.661157024793389e-07, + "logits/chosen": -0.08154296875, + "logits/rejected": -0.26171875, + "logps/chosen": -362.0, + "logps/rejected": -364.0, + "loss": 0.042, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.0166015625, + "rewards/margins": 11.3125, + "rewards/rejected": -11.375, + "step": 1209 + }, + { + "epoch": 1.4846625766871164, + "grad_norm": 13.517803461526613, + "learning_rate": 7.654958677685951e-07, + "logits/chosen": -0.036865234375, + "logits/rejected": -0.19140625, + "logps/chosen": -366.0, + "logps/rejected": -364.0, + "loss": 0.048, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.57421875, + "rewards/margins": 11.4375, + "rewards/rejected": -10.875, + "step": 1210 + }, + { + "epoch": 1.4858895705521473, + "grad_norm": 3.0410283674490257, + "learning_rate": 7.648760330578514e-07, + "logits/chosen": -0.06640625, + "logits/rejected": -0.283203125, + "logps/chosen": -358.0, + "logps/rejected": -382.0, + "loss": 0.0121, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.2021484375, + "rewards/margins": 11.0, + "rewards/rejected": -10.8125, + "step": 1211 + }, + { + "epoch": 1.487116564417178, + "grad_norm": 5.560187356994683, + "learning_rate": 7.642561983471075e-07, + "logits/chosen": -0.1767578125, + "logits/rejected": -0.28125, + "logps/chosen": -356.0, + "logps/rejected": -386.0, + "loss": 0.0259, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.12158203125, + "rewards/margins": 11.25, + "rewards/rejected": -11.375, + "step": 1212 + }, + { + "epoch": 1.4883435582822087, + "grad_norm": 7.567720346296838, + "learning_rate": 7.636363636363635e-07, + "logits/chosen": -0.1669921875, + "logits/rejected": -0.42578125, + "logps/chosen": -432.0, + "logps/rejected": -396.0, + "loss": 0.0322, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.7890625, + "rewards/margins": 11.5625, + "rewards/rejected": -10.75, + "step": 1213 + }, + { + "epoch": 1.4895705521472393, + "grad_norm": 10.044279259721861, + "learning_rate": 7.630165289256198e-07, + "logits/chosen": -0.1259765625, + "logits/rejected": -0.2421875, + "logps/chosen": -348.0, + "logps/rejected": -380.0, + "loss": 0.0415, + "rewards/accuracies": 0.984375, + "rewards/chosen": -0.369140625, + "rewards/margins": 11.0625, + "rewards/rejected": -11.4375, + "step": 1214 + }, + { + "epoch": 1.49079754601227, + "grad_norm": 11.888900857643913, + "learning_rate": 7.62396694214876e-07, + "logits/chosen": -0.181640625, + "logits/rejected": -0.390625, + "logps/chosen": -382.0, + "logps/rejected": -392.0, + "loss": 0.0302, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.31640625, + "rewards/margins": 12.25, + "rewards/rejected": -11.9375, + "step": 1215 + }, + { + "epoch": 1.4920245398773007, + "grad_norm": 11.234520606273014, + "learning_rate": 7.617768595041323e-07, + "logits/chosen": -0.140625, + "logits/rejected": -0.1845703125, + "logps/chosen": -378.0, + "logps/rejected": -370.0, + "loss": 0.0389, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.396484375, + "rewards/margins": 10.625, + "rewards/rejected": -11.0, + "step": 1216 + }, + { + "epoch": 1.4932515337423313, + "grad_norm": 5.600087268075799, + "learning_rate": 7.611570247933884e-07, + "logits/chosen": -0.1484375, + "logits/rejected": -0.26953125, + "logps/chosen": -404.0, + "logps/rejected": -406.0, + "loss": 0.0147, + "rewards/accuracies": 1.0, + "rewards/chosen": 1.078125, + "rewards/margins": 12.0625, + "rewards/rejected": -11.0, + "step": 1217 + }, + { + "epoch": 1.494478527607362, + "grad_norm": 13.512440969064604, + "learning_rate": 7.605371900826447e-07, + "logits/chosen": -0.0634765625, + "logits/rejected": -0.2333984375, + "logps/chosen": -374.0, + "logps/rejected": -386.0, + "loss": 0.0486, + "rewards/accuracies": 0.9765625, + "rewards/chosen": -0.76171875, + "rewards/margins": 10.0625, + "rewards/rejected": -10.8125, + "step": 1218 + }, + { + "epoch": 1.4957055214723927, + "grad_norm": 11.083011010632404, + "learning_rate": 7.599173553719009e-07, + "logits/chosen": -0.1513671875, + "logits/rejected": -0.302734375, + "logps/chosen": -394.0, + "logps/rejected": -408.0, + "loss": 0.0355, + "rewards/accuracies": 0.984375, + "rewards/chosen": -0.1767578125, + "rewards/margins": 10.8125, + "rewards/rejected": -10.9375, + "step": 1219 + }, + { + "epoch": 1.4969325153374233, + "grad_norm": 12.719289884310554, + "learning_rate": 7.59297520661157e-07, + "logits/chosen": -0.181640625, + "logits/rejected": -0.271484375, + "logps/chosen": -404.0, + "logps/rejected": -420.0, + "loss": 0.0511, + "rewards/accuracies": 0.96875, + "rewards/chosen": -0.53515625, + "rewards/margins": 11.0, + "rewards/rejected": -11.5, + "step": 1220 + }, + { + "epoch": 1.498159509202454, + "grad_norm": 11.10401374516513, + "learning_rate": 7.586776859504133e-07, + "logits/chosen": -0.2109375, + "logits/rejected": -0.3671875, + "logps/chosen": -412.0, + "logps/rejected": -382.0, + "loss": 0.038, + "rewards/accuracies": 0.96875, + "rewards/chosen": 0.087890625, + "rewards/margins": 11.5625, + "rewards/rejected": -11.4375, + "step": 1221 + }, + { + "epoch": 1.4993865030674847, + "grad_norm": 10.550256846021844, + "learning_rate": 7.580578512396693e-07, + "logits/chosen": -0.08935546875, + "logits/rejected": -0.263671875, + "logps/chosen": -342.0, + "logps/rejected": -370.0, + "loss": 0.0459, + "rewards/accuracies": 0.984375, + "rewards/chosen": -1.015625, + "rewards/margins": 10.0625, + "rewards/rejected": -11.0625, + "step": 1222 + }, + { + "epoch": 1.5006134969325153, + "grad_norm": 4.907032447202447, + "learning_rate": 7.574380165289256e-07, + "logits/chosen": -0.11328125, + "logits/rejected": -0.2421875, + "logps/chosen": -428.0, + "logps/rejected": -406.0, + "loss": 0.0214, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.4609375, + "rewards/margins": 11.8125, + "rewards/rejected": -11.3125, + "step": 1223 + }, + { + "epoch": 1.501840490797546, + "grad_norm": 7.212415747597125, + "learning_rate": 7.568181818181818e-07, + "logits/chosen": -0.1044921875, + "logits/rejected": -0.2041015625, + "logps/chosen": -360.0, + "logps/rejected": -374.0, + "loss": 0.0386, + "rewards/accuracies": 0.9765625, + "rewards/chosen": -0.96875, + "rewards/margins": 9.8125, + "rewards/rejected": -10.8125, + "step": 1224 + }, + { + "epoch": 1.5030674846625767, + "grad_norm": 5.207118696961701, + "learning_rate": 7.56198347107438e-07, + "logits/chosen": -0.09033203125, + "logits/rejected": -0.220703125, + "logps/chosen": -362.0, + "logps/rejected": -378.0, + "loss": 0.0192, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.35546875, + "rewards/margins": 10.5625, + "rewards/rejected": -10.9375, + "step": 1225 + }, + { + "epoch": 1.5042944785276073, + "grad_norm": 12.469106555425835, + "learning_rate": 7.555785123966942e-07, + "logits/chosen": -0.1669921875, + "logits/rejected": -0.34765625, + "logps/chosen": -394.0, + "logps/rejected": -384.0, + "loss": 0.0364, + "rewards/accuracies": 0.9765625, + "rewards/chosen": -0.58203125, + "rewards/margins": 10.75, + "rewards/rejected": -11.3125, + "step": 1226 + }, + { + "epoch": 1.505521472392638, + "grad_norm": 6.703095964441331, + "learning_rate": 7.549586776859505e-07, + "logits/chosen": -0.31640625, + "logits/rejected": -0.396484375, + "logps/chosen": -398.0, + "logps/rejected": -422.0, + "loss": 0.0161, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.228515625, + "rewards/margins": 11.625, + "rewards/rejected": -11.375, + "step": 1227 + }, + { + "epoch": 1.5067484662576687, + "grad_norm": 20.182875127429853, + "learning_rate": 7.543388429752067e-07, + "logits/chosen": -0.14453125, + "logits/rejected": -0.251953125, + "logps/chosen": -416.0, + "logps/rejected": -426.0, + "loss": 0.1114, + "rewards/accuracies": 0.9609375, + "rewards/chosen": -0.458984375, + "rewards/margins": 10.4375, + "rewards/rejected": -10.875, + "step": 1228 + }, + { + "epoch": 1.5079754601226993, + "grad_norm": 7.168144452671762, + "learning_rate": 7.537190082644629e-07, + "logits/chosen": -0.2314453125, + "logits/rejected": -0.34765625, + "logps/chosen": -382.0, + "logps/rejected": -394.0, + "loss": 0.0249, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.330078125, + "rewards/margins": 11.0, + "rewards/rejected": -11.3125, + "step": 1229 + }, + { + "epoch": 1.50920245398773, + "grad_norm": 8.739084266889561, + "learning_rate": 7.530991735537191e-07, + "logits/chosen": -0.1220703125, + "logits/rejected": -0.29296875, + "logps/chosen": -370.0, + "logps/rejected": -370.0, + "loss": 0.041, + "rewards/accuracies": 0.984375, + "rewards/chosen": -0.0230712890625, + "rewards/margins": 10.5, + "rewards/rejected": -10.5625, + "step": 1230 + }, + { + "epoch": 1.510429447852761, + "grad_norm": 14.93632729231571, + "learning_rate": 7.524793388429751e-07, + "logits/chosen": -0.0654296875, + "logits/rejected": -0.23046875, + "logps/chosen": -356.0, + "logps/rejected": -356.0, + "loss": 0.0501, + "rewards/accuracies": 0.9609375, + "rewards/chosen": -0.1796875, + "rewards/margins": 10.375, + "rewards/rejected": -10.5625, + "step": 1231 + }, + { + "epoch": 1.5116564417177916, + "grad_norm": 6.602158622702926, + "learning_rate": 7.518595041322314e-07, + "logits/chosen": -0.296875, + "logits/rejected": -0.486328125, + "logps/chosen": -400.0, + "logps/rejected": -396.0, + "loss": 0.0195, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.482421875, + "rewards/margins": 12.1875, + "rewards/rejected": -11.75, + "step": 1232 + }, + { + "epoch": 1.5128834355828222, + "grad_norm": 6.2838103080905015, + "learning_rate": 7.512396694214876e-07, + "logits/chosen": -0.193359375, + "logits/rejected": -0.37890625, + "logps/chosen": -384.0, + "logps/rejected": -374.0, + "loss": 0.0189, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.302734375, + "rewards/margins": 10.875, + "rewards/rejected": -11.1875, + "step": 1233 + }, + { + "epoch": 1.514110429447853, + "grad_norm": 18.0136951443616, + "learning_rate": 7.506198347107438e-07, + "logits/chosen": -0.1318359375, + "logits/rejected": -0.328125, + "logps/chosen": -426.0, + "logps/rejected": -394.0, + "loss": 0.0583, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.65234375, + "rewards/margins": 11.375, + "rewards/rejected": -10.6875, + "step": 1234 + }, + { + "epoch": 1.5153374233128836, + "grad_norm": 7.846641320419013, + "learning_rate": 7.5e-07, + "logits/chosen": -0.0546875, + "logits/rejected": -0.19140625, + "logps/chosen": -392.0, + "logps/rejected": -382.0, + "loss": 0.0282, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.9609375, + "rewards/margins": 10.375, + "rewards/rejected": -11.375, + "step": 1235 + }, + { + "epoch": 1.5165644171779142, + "grad_norm": 14.299732747614243, + "learning_rate": 7.493801652892563e-07, + "logits/chosen": -0.138671875, + "logits/rejected": -0.310546875, + "logps/chosen": -392.0, + "logps/rejected": -388.0, + "loss": 0.0557, + "rewards/accuracies": 0.96875, + "rewards/chosen": 0.0291748046875, + "rewards/margins": 10.8125, + "rewards/rejected": -10.8125, + "step": 1236 + }, + { + "epoch": 1.517791411042945, + "grad_norm": 15.326417785804917, + "learning_rate": 7.487603305785124e-07, + "logits/chosen": -0.24609375, + "logits/rejected": -0.40625, + "logps/chosen": -434.0, + "logps/rejected": -402.0, + "loss": 0.0642, + "rewards/accuracies": 0.9609375, + "rewards/chosen": -0.23046875, + "rewards/margins": 12.125, + "rewards/rejected": -12.3125, + "step": 1237 + }, + { + "epoch": 1.5190184049079756, + "grad_norm": 10.442301727950197, + "learning_rate": 7.481404958677685e-07, + "logits/chosen": -0.07373046875, + "logits/rejected": -0.2216796875, + "logps/chosen": -464.0, + "logps/rejected": -396.0, + "loss": 0.0259, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.7734375, + "rewards/margins": 11.625, + "rewards/rejected": -10.875, + "step": 1238 + }, + { + "epoch": 1.5202453987730062, + "grad_norm": 10.812028014205586, + "learning_rate": 7.475206611570248e-07, + "logits/chosen": -0.193359375, + "logits/rejected": -0.40625, + "logps/chosen": -404.0, + "logps/rejected": -382.0, + "loss": 0.0373, + "rewards/accuracies": 0.984375, + "rewards/chosen": -0.10302734375, + "rewards/margins": 11.0625, + "rewards/rejected": -11.1875, + "step": 1239 + }, + { + "epoch": 1.521472392638037, + "grad_norm": 7.399322537588399, + "learning_rate": 7.46900826446281e-07, + "logits/chosen": -0.0830078125, + "logits/rejected": -0.28125, + "logps/chosen": -410.0, + "logps/rejected": -402.0, + "loss": 0.0326, + "rewards/accuracies": 0.9765625, + "rewards/chosen": -0.353515625, + "rewards/margins": 10.9375, + "rewards/rejected": -11.25, + "step": 1240 + }, + { + "epoch": 1.5226993865030676, + "grad_norm": 11.737470905903368, + "learning_rate": 7.462809917355373e-07, + "logits/chosen": -0.033447265625, + "logits/rejected": -0.1806640625, + "logps/chosen": -360.0, + "logps/rejected": -368.0, + "loss": 0.0451, + "rewards/accuracies": 0.984375, + "rewards/chosen": -0.46484375, + "rewards/margins": 10.0625, + "rewards/rejected": -10.5625, + "step": 1241 + }, + { + "epoch": 1.5239263803680982, + "grad_norm": 11.68475946384956, + "learning_rate": 7.456611570247933e-07, + "logits/chosen": -0.263671875, + "logits/rejected": -0.42578125, + "logps/chosen": -388.0, + "logps/rejected": -386.0, + "loss": 0.0484, + "rewards/accuracies": 0.9765625, + "rewards/chosen": -0.5703125, + "rewards/margins": 11.25, + "rewards/rejected": -11.8125, + "step": 1242 + }, + { + "epoch": 1.525153374233129, + "grad_norm": 7.544728556493131, + "learning_rate": 7.450413223140496e-07, + "logits/chosen": -0.059814453125, + "logits/rejected": -0.2158203125, + "logps/chosen": -334.0, + "logps/rejected": -354.0, + "loss": 0.028, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.73828125, + "rewards/margins": 10.125, + "rewards/rejected": -10.875, + "step": 1243 + }, + { + "epoch": 1.5263803680981596, + "grad_norm": 8.662327485339514, + "learning_rate": 7.444214876033058e-07, + "logits/chosen": -0.06787109375, + "logits/rejected": -0.16796875, + "logps/chosen": -342.0, + "logps/rejected": -376.0, + "loss": 0.0232, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.59375, + "rewards/margins": 10.0, + "rewards/rejected": -10.5625, + "step": 1244 + }, + { + "epoch": 1.5276073619631902, + "grad_norm": 14.324574848335546, + "learning_rate": 7.438016528925621e-07, + "logits/chosen": -0.10498046875, + "logits/rejected": -0.302734375, + "logps/chosen": -422.0, + "logps/rejected": -418.0, + "loss": 0.0653, + "rewards/accuracies": 0.96875, + "rewards/chosen": -0.08154296875, + "rewards/margins": 11.25, + "rewards/rejected": -11.3125, + "step": 1245 + }, + { + "epoch": 1.528834355828221, + "grad_norm": 8.939887633566299, + "learning_rate": 7.431818181818182e-07, + "logits/chosen": -0.23046875, + "logits/rejected": -0.396484375, + "logps/chosen": -352.0, + "logps/rejected": -370.0, + "loss": 0.0281, + "rewards/accuracies": 0.984375, + "rewards/chosen": -0.05908203125, + "rewards/margins": 11.125, + "rewards/rejected": -11.1875, + "step": 1246 + }, + { + "epoch": 1.5300613496932516, + "grad_norm": 2.737977434539333, + "learning_rate": 7.425619834710743e-07, + "logits/chosen": -0.189453125, + "logits/rejected": -0.302734375, + "logps/chosen": -434.0, + "logps/rejected": -404.0, + "loss": 0.0103, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.318359375, + "rewards/margins": 12.0625, + "rewards/rejected": -11.75, + "step": 1247 + }, + { + "epoch": 1.5312883435582823, + "grad_norm": 7.955215828013632, + "learning_rate": 7.419421487603306e-07, + "logits/chosen": -0.1494140625, + "logits/rejected": -0.3359375, + "logps/chosen": -376.0, + "logps/rejected": -386.0, + "loss": 0.0229, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.3828125, + "rewards/margins": 11.375, + "rewards/rejected": -11.0, + "step": 1248 + }, + { + "epoch": 1.532515337423313, + "grad_norm": 13.148164884666349, + "learning_rate": 7.413223140495868e-07, + "logits/chosen": 0.036376953125, + "logits/rejected": -0.0673828125, + "logps/chosen": -350.0, + "logps/rejected": -378.0, + "loss": 0.0618, + "rewards/accuracies": 0.984375, + "rewards/chosen": -0.54296875, + "rewards/margins": 10.875, + "rewards/rejected": -11.4375, + "step": 1249 + }, + { + "epoch": 1.5337423312883436, + "grad_norm": 6.7014580907570505, + "learning_rate": 7.40702479338843e-07, + "logits/chosen": -0.2333984375, + "logits/rejected": -0.384765625, + "logps/chosen": -446.0, + "logps/rejected": -472.0, + "loss": 0.028, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.39453125, + "rewards/margins": 13.125, + "rewards/rejected": -12.75, + "step": 1250 + }, + { + "epoch": 1.5349693251533743, + "grad_norm": 9.188465721829859, + "learning_rate": 7.400826446280991e-07, + "logits/chosen": -0.291015625, + "logits/rejected": -0.39453125, + "logps/chosen": -376.0, + "logps/rejected": -414.0, + "loss": 0.0271, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.68359375, + "rewards/margins": 12.625, + "rewards/rejected": -11.9375, + "step": 1251 + }, + { + "epoch": 1.536196319018405, + "grad_norm": 8.458678355370827, + "learning_rate": 7.394628099173554e-07, + "logits/chosen": -0.181640625, + "logits/rejected": -0.34765625, + "logps/chosen": -390.0, + "logps/rejected": -372.0, + "loss": 0.0304, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.158203125, + "rewards/margins": 12.375, + "rewards/rejected": -12.1875, + "step": 1252 + }, + { + "epoch": 1.5374233128834356, + "grad_norm": 8.012255387049295, + "learning_rate": 7.388429752066116e-07, + "logits/chosen": -0.16796875, + "logits/rejected": -0.326171875, + "logps/chosen": -420.0, + "logps/rejected": -410.0, + "loss": 0.0317, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.126953125, + "rewards/margins": 11.3125, + "rewards/rejected": -11.1875, + "step": 1253 + }, + { + "epoch": 1.5386503067484663, + "grad_norm": 9.41643727038603, + "learning_rate": 7.382231404958678e-07, + "logits/chosen": -0.0654296875, + "logits/rejected": -0.19921875, + "logps/chosen": -378.0, + "logps/rejected": -360.0, + "loss": 0.0366, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.267578125, + "rewards/margins": 11.0, + "rewards/rejected": -10.75, + "step": 1254 + }, + { + "epoch": 1.539877300613497, + "grad_norm": 7.434007112296638, + "learning_rate": 7.37603305785124e-07, + "logits/chosen": -0.203125, + "logits/rejected": -0.427734375, + "logps/chosen": -410.0, + "logps/rejected": -432.0, + "loss": 0.0221, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.62890625, + "rewards/margins": 12.6875, + "rewards/rejected": -12.0625, + "step": 1255 + }, + { + "epoch": 1.5411042944785276, + "grad_norm": 8.974145176497167, + "learning_rate": 7.369834710743802e-07, + "logits/chosen": -0.1142578125, + "logits/rejected": -0.328125, + "logps/chosen": -410.0, + "logps/rejected": -398.0, + "loss": 0.0302, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.06494140625, + "rewards/margins": 10.6875, + "rewards/rejected": -10.625, + "step": 1256 + }, + { + "epoch": 1.5423312883435583, + "grad_norm": 13.782445851616703, + "learning_rate": 7.363636363636364e-07, + "logits/chosen": -0.0732421875, + "logits/rejected": -0.26171875, + "logps/chosen": -390.0, + "logps/rejected": -386.0, + "loss": 0.0591, + "rewards/accuracies": 0.9765625, + "rewards/chosen": 0.0693359375, + "rewards/margins": 11.0, + "rewards/rejected": -10.9375, + "step": 1257 + }, + { + "epoch": 1.543558282208589, + "grad_norm": 8.478001588142458, + "learning_rate": 7.357438016528926e-07, + "logits/chosen": -0.10791015625, + "logits/rejected": -0.3125, + "logps/chosen": -404.0, + "logps/rejected": -368.0, + "loss": 0.0315, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.82421875, + "rewards/margins": 11.375, + "rewards/rejected": -10.5625, + "step": 1258 + }, + { + "epoch": 1.5447852760736196, + "grad_norm": 12.343587239938957, + "learning_rate": 7.351239669421488e-07, + "logits/chosen": -0.11181640625, + "logits/rejected": -0.34375, + "logps/chosen": -404.0, + "logps/rejected": -380.0, + "loss": 0.0382, + "rewards/accuracies": 0.96875, + "rewards/chosen": 0.6484375, + "rewards/margins": 11.25, + "rewards/rejected": -10.625, + "step": 1259 + }, + { + "epoch": 1.5460122699386503, + "grad_norm": 10.786647686676162, + "learning_rate": 7.345041322314049e-07, + "logits/chosen": -0.169921875, + "logits/rejected": -0.34765625, + "logps/chosen": -450.0, + "logps/rejected": -396.0, + "loss": 0.0273, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.9453125, + "rewards/margins": 11.6875, + "rewards/rejected": -10.75, + "step": 1260 + }, + { + "epoch": 1.547239263803681, + "grad_norm": 11.69554028398202, + "learning_rate": 7.338842975206612e-07, + "logits/chosen": -0.2236328125, + "logits/rejected": -0.345703125, + "logps/chosen": -406.0, + "logps/rejected": -388.0, + "loss": 0.0333, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.490234375, + "rewards/margins": 10.375, + "rewards/rejected": -9.875, + "step": 1261 + }, + { + "epoch": 1.5484662576687116, + "grad_norm": 7.864292149216722, + "learning_rate": 7.332644628099174e-07, + "logits/chosen": -0.146484375, + "logits/rejected": -0.333984375, + "logps/chosen": -386.0, + "logps/rejected": -362.0, + "loss": 0.038, + "rewards/accuracies": 0.984375, + "rewards/chosen": -0.13671875, + "rewards/margins": 10.9375, + "rewards/rejected": -11.0625, + "step": 1262 + }, + { + "epoch": 1.5496932515337423, + "grad_norm": 13.897917738806647, + "learning_rate": 7.326446280991736e-07, + "logits/chosen": -0.0830078125, + "logits/rejected": -0.298828125, + "logps/chosen": -394.0, + "logps/rejected": -380.0, + "loss": 0.0455, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.69921875, + "rewards/margins": 11.625, + "rewards/rejected": -10.9375, + "step": 1263 + }, + { + "epoch": 1.550920245398773, + "grad_norm": 11.512996209757597, + "learning_rate": 7.320247933884297e-07, + "logits/chosen": -0.0218505859375, + "logits/rejected": -0.1875, + "logps/chosen": -390.0, + "logps/rejected": -398.0, + "loss": 0.0581, + "rewards/accuracies": 0.984375, + "rewards/chosen": -0.1591796875, + "rewards/margins": 10.1875, + "rewards/rejected": -10.3125, + "step": 1264 + }, + { + "epoch": 1.5521472392638036, + "grad_norm": 13.705005476191044, + "learning_rate": 7.31404958677686e-07, + "logits/chosen": -0.1689453125, + "logits/rejected": -0.341796875, + "logps/chosen": -380.0, + "logps/rejected": -390.0, + "loss": 0.0413, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.3359375, + "rewards/margins": 11.125, + "rewards/rejected": -10.8125, + "step": 1265 + }, + { + "epoch": 1.5533742331288343, + "grad_norm": 10.295352755658561, + "learning_rate": 7.307851239669422e-07, + "logits/chosen": -0.2021484375, + "logits/rejected": -0.296875, + "logps/chosen": -352.0, + "logps/rejected": -372.0, + "loss": 0.0297, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.09130859375, + "rewards/margins": 10.5625, + "rewards/rejected": -10.5, + "step": 1266 + }, + { + "epoch": 1.554601226993865, + "grad_norm": 18.45895537867682, + "learning_rate": 7.301652892561983e-07, + "logits/chosen": -0.130859375, + "logits/rejected": -0.271484375, + "logps/chosen": -368.0, + "logps/rejected": -366.0, + "loss": 0.1103, + "rewards/accuracies": 0.96875, + "rewards/chosen": 0.29296875, + "rewards/margins": 10.8125, + "rewards/rejected": -10.5, + "step": 1267 + }, + { + "epoch": 1.5558282208588956, + "grad_norm": 14.03503002507992, + "learning_rate": 7.295454545454546e-07, + "logits/chosen": -0.14453125, + "logits/rejected": -0.255859375, + "logps/chosen": -498.0, + "logps/rejected": -422.0, + "loss": 0.0495, + "rewards/accuracies": 0.9765625, + "rewards/chosen": 0.9609375, + "rewards/margins": 12.125, + "rewards/rejected": -11.1875, + "step": 1268 + }, + { + "epoch": 1.5570552147239263, + "grad_norm": 3.7230278336867006, + "learning_rate": 7.289256198347107e-07, + "logits/chosen": -0.1806640625, + "logits/rejected": -0.314453125, + "logps/chosen": -378.0, + "logps/rejected": -362.0, + "loss": 0.0144, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.8515625, + "rewards/margins": 11.5625, + "rewards/rejected": -10.75, + "step": 1269 + }, + { + "epoch": 1.558282208588957, + "grad_norm": 8.890488407699403, + "learning_rate": 7.28305785123967e-07, + "logits/chosen": -0.1689453125, + "logits/rejected": -0.22265625, + "logps/chosen": -406.0, + "logps/rejected": -406.0, + "loss": 0.026, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.396484375, + "rewards/margins": 10.75, + "rewards/rejected": -10.375, + "step": 1270 + }, + { + "epoch": 1.5595092024539876, + "grad_norm": 8.516198696845315, + "learning_rate": 7.276859504132231e-07, + "logits/chosen": -0.0830078125, + "logits/rejected": -0.1650390625, + "logps/chosen": -356.0, + "logps/rejected": -366.0, + "loss": 0.0302, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.56640625, + "rewards/margins": 10.8125, + "rewards/rejected": -10.25, + "step": 1271 + }, + { + "epoch": 1.5607361963190183, + "grad_norm": 3.950798580884482, + "learning_rate": 7.270661157024794e-07, + "logits/chosen": -0.2578125, + "logits/rejected": -0.412109375, + "logps/chosen": -426.0, + "logps/rejected": -392.0, + "loss": 0.0131, + "rewards/accuracies": 1.0, + "rewards/chosen": 1.46875, + "rewards/margins": 12.25, + "rewards/rejected": -10.75, + "step": 1272 + }, + { + "epoch": 1.561963190184049, + "grad_norm": 12.063277382793855, + "learning_rate": 7.264462809917355e-07, + "logits/chosen": -0.2080078125, + "logits/rejected": -0.3046875, + "logps/chosen": -426.0, + "logps/rejected": -412.0, + "loss": 0.0509, + "rewards/accuracies": 0.9765625, + "rewards/chosen": 0.365234375, + "rewards/margins": 10.6875, + "rewards/rejected": -10.3125, + "step": 1273 + }, + { + "epoch": 1.5631901840490796, + "grad_norm": 5.052885121682078, + "learning_rate": 7.258264462809918e-07, + "logits/chosen": -0.31640625, + "logits/rejected": -0.5078125, + "logps/chosen": -400.0, + "logps/rejected": -388.0, + "loss": 0.0173, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.263671875, + "rewards/margins": 11.0, + "rewards/rejected": -10.75, + "step": 1274 + }, + { + "epoch": 1.5644171779141103, + "grad_norm": 15.33141936648023, + "learning_rate": 7.25206611570248e-07, + "logits/chosen": -0.10791015625, + "logits/rejected": -0.2578125, + "logps/chosen": -388.0, + "logps/rejected": -394.0, + "loss": 0.0744, + "rewards/accuracies": 0.96875, + "rewards/chosen": -0.111328125, + "rewards/margins": 10.6875, + "rewards/rejected": -10.8125, + "step": 1275 + }, + { + "epoch": 1.565644171779141, + "grad_norm": 7.988279468708803, + "learning_rate": 7.245867768595041e-07, + "logits/chosen": -0.025146484375, + "logits/rejected": -0.2138671875, + "logps/chosen": -418.0, + "logps/rejected": -374.0, + "loss": 0.0247, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.48046875, + "rewards/margins": 11.0, + "rewards/rejected": -10.5, + "step": 1276 + }, + { + "epoch": 1.5668711656441716, + "grad_norm": 12.554173231892838, + "learning_rate": 7.239669421487604e-07, + "logits/chosen": -0.185546875, + "logits/rejected": -0.314453125, + "logps/chosen": -358.0, + "logps/rejected": -354.0, + "loss": 0.043, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.034423828125, + "rewards/margins": 10.0, + "rewards/rejected": -10.0625, + "step": 1277 + }, + { + "epoch": 1.5680981595092025, + "grad_norm": 15.815759980577592, + "learning_rate": 7.233471074380165e-07, + "logits/chosen": -0.15625, + "logits/rejected": -0.365234375, + "logps/chosen": -404.0, + "logps/rejected": -396.0, + "loss": 0.028, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.77734375, + "rewards/margins": 12.125, + "rewards/rejected": -11.3125, + "step": 1278 + }, + { + "epoch": 1.5693251533742332, + "grad_norm": 13.88863047530027, + "learning_rate": 7.227272727272728e-07, + "logits/chosen": -0.169921875, + "logits/rejected": -0.310546875, + "logps/chosen": -356.0, + "logps/rejected": -394.0, + "loss": 0.0486, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.1826171875, + "rewards/margins": 10.6875, + "rewards/rejected": -10.5, + "step": 1279 + }, + { + "epoch": 1.5705521472392638, + "grad_norm": 11.958915852962361, + "learning_rate": 7.221074380165289e-07, + "logits/chosen": -0.2294921875, + "logits/rejected": -0.3515625, + "logps/chosen": -434.0, + "logps/rejected": -432.0, + "loss": 0.0382, + "rewards/accuracies": 0.9765625, + "rewards/chosen": 0.470703125, + "rewards/margins": 12.1875, + "rewards/rejected": -11.75, + "step": 1280 + }, + { + "epoch": 1.5717791411042945, + "grad_norm": 13.174861138807707, + "learning_rate": 7.214876033057852e-07, + "logits/chosen": -0.2236328125, + "logits/rejected": -0.384765625, + "logps/chosen": -438.0, + "logps/rejected": -440.0, + "loss": 0.0556, + "rewards/accuracies": 0.96875, + "rewards/chosen": 1.125, + "rewards/margins": 12.6875, + "rewards/rejected": -11.5625, + "step": 1281 + }, + { + "epoch": 1.5730061349693252, + "grad_norm": 9.377186147771738, + "learning_rate": 7.208677685950413e-07, + "logits/chosen": 0.048828125, + "logits/rejected": -0.185546875, + "logps/chosen": -394.0, + "logps/rejected": -374.0, + "loss": 0.0309, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.032470703125, + "rewards/margins": 11.125, + "rewards/rejected": -11.125, + "step": 1282 + }, + { + "epoch": 1.5742331288343558, + "grad_norm": 9.78074406651563, + "learning_rate": 7.202479338842976e-07, + "logits/chosen": -0.0028228759765625, + "logits/rejected": -0.11865234375, + "logps/chosen": -354.0, + "logps/rejected": -380.0, + "loss": 0.0371, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.88671875, + "rewards/margins": 10.8125, + "rewards/rejected": -11.6875, + "step": 1283 + }, + { + "epoch": 1.5754601226993865, + "grad_norm": 15.650011079839588, + "learning_rate": 7.196280991735537e-07, + "logits/chosen": -0.072265625, + "logits/rejected": -0.2490234375, + "logps/chosen": -404.0, + "logps/rejected": -420.0, + "loss": 0.0681, + "rewards/accuracies": 0.9765625, + "rewards/chosen": -0.33203125, + "rewards/margins": 12.1875, + "rewards/rejected": -12.5, + "step": 1284 + }, + { + "epoch": 1.5766871165644172, + "grad_norm": 15.27608982764389, + "learning_rate": 7.190082644628099e-07, + "logits/chosen": -0.130859375, + "logits/rejected": -0.373046875, + "logps/chosen": -422.0, + "logps/rejected": -420.0, + "loss": 0.0468, + "rewards/accuracies": 0.9765625, + "rewards/chosen": -0.109375, + "rewards/margins": 12.1875, + "rewards/rejected": -12.3125, + "step": 1285 + }, + { + "epoch": 1.5779141104294478, + "grad_norm": 15.001506307085636, + "learning_rate": 7.183884297520662e-07, + "logits/chosen": -0.0771484375, + "logits/rejected": -0.271484375, + "logps/chosen": -372.0, + "logps/rejected": -356.0, + "loss": 0.0383, + "rewards/accuracies": 0.984375, + "rewards/chosen": -0.921875, + "rewards/margins": 10.75, + "rewards/rejected": -11.6875, + "step": 1286 + }, + { + "epoch": 1.5791411042944785, + "grad_norm": 5.225854839258308, + "learning_rate": 7.177685950413223e-07, + "logits/chosen": -0.1748046875, + "logits/rejected": -0.251953125, + "logps/chosen": -330.0, + "logps/rejected": -380.0, + "loss": 0.0262, + "rewards/accuracies": 0.984375, + "rewards/chosen": -1.375, + "rewards/margins": 11.0, + "rewards/rejected": -12.375, + "step": 1287 + }, + { + "epoch": 1.5803680981595092, + "grad_norm": 11.803831089679342, + "learning_rate": 7.171487603305785e-07, + "logits/chosen": -0.2255859375, + "logits/rejected": -0.400390625, + "logps/chosen": -394.0, + "logps/rejected": -412.0, + "loss": 0.0419, + "rewards/accuracies": 0.96875, + "rewards/chosen": -0.2236328125, + "rewards/margins": 12.125, + "rewards/rejected": -12.3125, + "step": 1288 + }, + { + "epoch": 1.5815950920245399, + "grad_norm": 9.347363939888405, + "learning_rate": 7.165289256198347e-07, + "logits/chosen": -0.087890625, + "logits/rejected": -0.2353515625, + "logps/chosen": -334.0, + "logps/rejected": -342.0, + "loss": 0.0396, + "rewards/accuracies": 0.984375, + "rewards/chosen": -1.0390625, + "rewards/margins": 10.75, + "rewards/rejected": -11.8125, + "step": 1289 + }, + { + "epoch": 1.5828220858895705, + "grad_norm": 4.3767999812751714, + "learning_rate": 7.15909090909091e-07, + "logits/chosen": -0.2294921875, + "logits/rejected": -0.40234375, + "logps/chosen": -396.0, + "logps/rejected": -416.0, + "loss": 0.0096, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.8984375, + "rewards/margins": 12.375, + "rewards/rejected": -13.25, + "step": 1290 + }, + { + "epoch": 1.5840490797546012, + "grad_norm": 4.228705291026931, + "learning_rate": 7.152892561983471e-07, + "logits/chosen": -0.099609375, + "logits/rejected": -0.21484375, + "logps/chosen": -372.0, + "logps/rejected": -376.0, + "loss": 0.0152, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.875, + "rewards/margins": 11.3125, + "rewards/rejected": -12.1875, + "step": 1291 + }, + { + "epoch": 1.5852760736196319, + "grad_norm": 7.584547777512873, + "learning_rate": 7.146694214876034e-07, + "logits/chosen": -0.25, + "logits/rejected": -0.35546875, + "logps/chosen": -390.0, + "logps/rejected": -428.0, + "loss": 0.0277, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.67578125, + "rewards/margins": 12.1875, + "rewards/rejected": -12.875, + "step": 1292 + }, + { + "epoch": 1.5865030674846625, + "grad_norm": 10.655223291637148, + "learning_rate": 7.140495867768595e-07, + "logits/chosen": -0.0137939453125, + "logits/rejected": -0.28515625, + "logps/chosen": -416.0, + "logps/rejected": -384.0, + "loss": 0.0202, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.25390625, + "rewards/margins": 11.375, + "rewards/rejected": -11.625, + "step": 1293 + }, + { + "epoch": 1.5877300613496934, + "grad_norm": 4.2898047053372155, + "learning_rate": 7.134297520661157e-07, + "logits/chosen": -0.11279296875, + "logits/rejected": -0.30078125, + "logps/chosen": -412.0, + "logps/rejected": -428.0, + "loss": 0.0154, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.51953125, + "rewards/margins": 12.25, + "rewards/rejected": -12.8125, + "step": 1294 + }, + { + "epoch": 1.588957055214724, + "grad_norm": 8.499456654824181, + "learning_rate": 7.12809917355372e-07, + "logits/chosen": -0.2470703125, + "logits/rejected": -0.431640625, + "logps/chosen": -370.0, + "logps/rejected": -378.0, + "loss": 0.0387, + "rewards/accuracies": 0.9765625, + "rewards/chosen": -0.59375, + "rewards/margins": 11.0, + "rewards/rejected": -11.5625, + "step": 1295 + }, + { + "epoch": 1.5901840490797547, + "grad_norm": 16.676691340584743, + "learning_rate": 7.121900826446281e-07, + "logits/chosen": -0.1220703125, + "logits/rejected": -0.2099609375, + "logps/chosen": -394.0, + "logps/rejected": -402.0, + "loss": 0.0806, + "rewards/accuracies": 0.96875, + "rewards/chosen": -1.234375, + "rewards/margins": 10.75, + "rewards/rejected": -12.0, + "step": 1296 + }, + { + "epoch": 1.5914110429447854, + "grad_norm": 9.185721305153498, + "learning_rate": 7.115702479338843e-07, + "logits/chosen": -0.042236328125, + "logits/rejected": -0.2080078125, + "logps/chosen": -414.0, + "logps/rejected": -384.0, + "loss": 0.0365, + "rewards/accuracies": 0.984375, + "rewards/chosen": -0.41015625, + "rewards/margins": 9.875, + "rewards/rejected": -10.25, + "step": 1297 + }, + { + "epoch": 1.592638036809816, + "grad_norm": 9.516915353409866, + "learning_rate": 7.109504132231405e-07, + "logits/chosen": -0.07568359375, + "logits/rejected": -0.197265625, + "logps/chosen": -410.0, + "logps/rejected": -414.0, + "loss": 0.0373, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.279296875, + "rewards/margins": 11.75, + "rewards/rejected": -11.4375, + "step": 1298 + }, + { + "epoch": 1.5938650306748468, + "grad_norm": 12.688329159049632, + "learning_rate": 7.103305785123968e-07, + "logits/chosen": -0.197265625, + "logits/rejected": -0.28515625, + "logps/chosen": -394.0, + "logps/rejected": -402.0, + "loss": 0.0515, + "rewards/accuracies": 0.984375, + "rewards/chosen": -0.294921875, + "rewards/margins": 10.9375, + "rewards/rejected": -11.1875, + "step": 1299 + }, + { + "epoch": 1.5950920245398774, + "grad_norm": 9.480010594113034, + "learning_rate": 7.097107438016529e-07, + "logits/chosen": -0.08447265625, + "logits/rejected": -0.283203125, + "logps/chosen": -406.0, + "logps/rejected": -410.0, + "loss": 0.0238, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.345703125, + "rewards/margins": 11.8125, + "rewards/rejected": -11.4375, + "step": 1300 + }, + { + "epoch": 1.596319018404908, + "grad_norm": 6.743619593496769, + "learning_rate": 7.09090909090909e-07, + "logits/chosen": -0.07958984375, + "logits/rejected": -0.27734375, + "logps/chosen": -384.0, + "logps/rejected": -394.0, + "loss": 0.0206, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.045166015625, + "rewards/margins": 11.1875, + "rewards/rejected": -11.25, + "step": 1301 + }, + { + "epoch": 1.5975460122699388, + "grad_norm": 10.079704999041756, + "learning_rate": 7.084710743801653e-07, + "logits/chosen": -0.283203125, + "logits/rejected": -0.396484375, + "logps/chosen": -404.0, + "logps/rejected": -402.0, + "loss": 0.0333, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.49609375, + "rewards/margins": 11.8125, + "rewards/rejected": -11.3125, + "step": 1302 + }, + { + "epoch": 1.5987730061349694, + "grad_norm": 4.128642305901009, + "learning_rate": 7.078512396694215e-07, + "logits/chosen": -0.15625, + "logits/rejected": -0.369140625, + "logps/chosen": -440.0, + "logps/rejected": -408.0, + "loss": 0.0206, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.6328125, + "rewards/margins": 12.5625, + "rewards/rejected": -11.9375, + "step": 1303 + }, + { + "epoch": 1.6, + "grad_norm": 6.006214161041272, + "learning_rate": 7.072314049586777e-07, + "logits/chosen": -0.2255859375, + "logits/rejected": -0.388671875, + "logps/chosen": -380.0, + "logps/rejected": -408.0, + "loss": 0.0223, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.7421875, + "rewards/margins": 12.3125, + "rewards/rejected": -11.5625, + "step": 1304 + }, + { + "epoch": 1.6012269938650308, + "grad_norm": 15.374011330047086, + "learning_rate": 7.066115702479339e-07, + "logits/chosen": -0.12451171875, + "logits/rejected": -0.259765625, + "logps/chosen": -346.0, + "logps/rejected": -366.0, + "loss": 0.068, + "rewards/accuracies": 0.9765625, + "rewards/chosen": 0.423828125, + "rewards/margins": 10.875, + "rewards/rejected": -10.4375, + "step": 1305 + }, + { + "epoch": 1.6024539877300614, + "grad_norm": 10.751324968165987, + "learning_rate": 7.059917355371901e-07, + "logits/chosen": -0.1953125, + "logits/rejected": -0.294921875, + "logps/chosen": -406.0, + "logps/rejected": -384.0, + "loss": 0.0244, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.44921875, + "rewards/margins": 11.3125, + "rewards/rejected": -10.875, + "step": 1306 + }, + { + "epoch": 1.603680981595092, + "grad_norm": 7.9285821638566905, + "learning_rate": 7.053719008264463e-07, + "logits/chosen": -0.1953125, + "logits/rejected": -0.38671875, + "logps/chosen": -440.0, + "logps/rejected": -402.0, + "loss": 0.0207, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 1.0078125, + "rewards/margins": 11.25, + "rewards/rejected": -10.25, + "step": 1307 + }, + { + "epoch": 1.6049079754601228, + "grad_norm": 4.924145204402143, + "learning_rate": 7.047520661157026e-07, + "logits/chosen": -0.25, + "logits/rejected": -0.392578125, + "logps/chosen": -380.0, + "logps/rejected": -372.0, + "loss": 0.0136, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.875, + "rewards/margins": 11.5625, + "rewards/rejected": -10.6875, + "step": 1308 + }, + { + "epoch": 1.6061349693251534, + "grad_norm": 11.95129952622703, + "learning_rate": 7.041322314049587e-07, + "logits/chosen": -0.185546875, + "logits/rejected": -0.3046875, + "logps/chosen": -392.0, + "logps/rejected": -394.0, + "loss": 0.0371, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.263671875, + "rewards/margins": 11.25, + "rewards/rejected": -11.0, + "step": 1309 + }, + { + "epoch": 1.607361963190184, + "grad_norm": 6.935480593251807, + "learning_rate": 7.035123966942149e-07, + "logits/chosen": -0.09619140625, + "logits/rejected": -0.2578125, + "logps/chosen": -334.0, + "logps/rejected": -356.0, + "loss": 0.0257, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.7109375, + "rewards/margins": 11.125, + "rewards/rejected": -10.4375, + "step": 1310 + }, + { + "epoch": 1.6085889570552148, + "grad_norm": 7.070447805786784, + "learning_rate": 7.028925619834711e-07, + "logits/chosen": -0.046142578125, + "logits/rejected": -0.24609375, + "logps/chosen": -418.0, + "logps/rejected": -380.0, + "loss": 0.0364, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.0751953125, + "rewards/margins": 9.9375, + "rewards/rejected": -9.8125, + "step": 1311 + }, + { + "epoch": 1.6098159509202454, + "grad_norm": 6.5989010041947065, + "learning_rate": 7.022727272727273e-07, + "logits/chosen": -0.2216796875, + "logits/rejected": -0.453125, + "logps/chosen": -404.0, + "logps/rejected": -392.0, + "loss": 0.0213, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.75390625, + "rewards/margins": 12.3125, + "rewards/rejected": -11.5625, + "step": 1312 + }, + { + "epoch": 1.611042944785276, + "grad_norm": 5.053991836327145, + "learning_rate": 7.016528925619835e-07, + "logits/chosen": -0.1376953125, + "logits/rejected": -0.28125, + "logps/chosen": -366.0, + "logps/rejected": -374.0, + "loss": 0.0231, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.2294921875, + "rewards/margins": 10.75, + "rewards/rejected": -10.5, + "step": 1313 + }, + { + "epoch": 1.6122699386503068, + "grad_norm": 7.196334947724689, + "learning_rate": 7.010330578512396e-07, + "logits/chosen": -0.310546875, + "logits/rejected": -0.390625, + "logps/chosen": -392.0, + "logps/rejected": -408.0, + "loss": 0.0192, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.99609375, + "rewards/margins": 12.4375, + "rewards/rejected": -11.4375, + "step": 1314 + }, + { + "epoch": 1.6134969325153374, + "grad_norm": 7.456124142101998, + "learning_rate": 7.004132231404959e-07, + "logits/chosen": -0.158203125, + "logits/rejected": -0.384765625, + "logps/chosen": -452.0, + "logps/rejected": -422.0, + "loss": 0.0295, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.396484375, + "rewards/margins": 11.5625, + "rewards/rejected": -11.1875, + "step": 1315 + }, + { + "epoch": 1.614723926380368, + "grad_norm": 14.771588037378546, + "learning_rate": 6.997933884297521e-07, + "logits/chosen": -0.022705078125, + "logits/rejected": -0.2138671875, + "logps/chosen": -378.0, + "logps/rejected": -366.0, + "loss": 0.0521, + "rewards/accuracies": 0.984375, + "rewards/chosen": -0.09912109375, + "rewards/margins": 10.1875, + "rewards/rejected": -10.3125, + "step": 1316 + }, + { + "epoch": 1.6159509202453988, + "grad_norm": 6.5021730031025555, + "learning_rate": 6.991735537190084e-07, + "logits/chosen": -0.19140625, + "logits/rejected": -0.296875, + "logps/chosen": -388.0, + "logps/rejected": -414.0, + "loss": 0.024, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.146484375, + "rewards/margins": 11.1875, + "rewards/rejected": -11.0625, + "step": 1317 + }, + { + "epoch": 1.6171779141104294, + "grad_norm": 7.036484087537271, + "learning_rate": 6.985537190082644e-07, + "logits/chosen": -0.126953125, + "logits/rejected": -0.333984375, + "logps/chosen": -388.0, + "logps/rejected": -382.0, + "loss": 0.0165, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.037109375, + "rewards/margins": 11.375, + "rewards/rejected": -11.3125, + "step": 1318 + }, + { + "epoch": 1.61840490797546, + "grad_norm": 12.457147867029073, + "learning_rate": 6.979338842975207e-07, + "logits/chosen": -0.1552734375, + "logits/rejected": -0.30859375, + "logps/chosen": -390.0, + "logps/rejected": -424.0, + "loss": 0.0243, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.474609375, + "rewards/margins": 12.0, + "rewards/rejected": -11.5625, + "step": 1319 + }, + { + "epoch": 1.6196319018404908, + "grad_norm": 18.65684903039485, + "learning_rate": 6.973140495867769e-07, + "logits/chosen": -0.123046875, + "logits/rejected": -0.314453125, + "logps/chosen": -438.0, + "logps/rejected": -406.0, + "loss": 0.0705, + "rewards/accuracies": 0.9765625, + "rewards/chosen": 0.4375, + "rewards/margins": 11.625, + "rewards/rejected": -11.1875, + "step": 1320 + }, + { + "epoch": 1.6208588957055214, + "grad_norm": 15.725190139889248, + "learning_rate": 6.966942148760331e-07, + "logits/chosen": -0.16796875, + "logits/rejected": -0.296875, + "logps/chosen": -384.0, + "logps/rejected": -376.0, + "loss": 0.0439, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.17578125, + "rewards/margins": 11.375, + "rewards/rejected": -11.25, + "step": 1321 + }, + { + "epoch": 1.622085889570552, + "grad_norm": 13.624128902680484, + "learning_rate": 6.960743801652893e-07, + "logits/chosen": -0.189453125, + "logits/rejected": -0.345703125, + "logps/chosen": -366.0, + "logps/rejected": -396.0, + "loss": 0.0327, + "rewards/accuracies": 0.984375, + "rewards/chosen": -0.1865234375, + "rewards/margins": 10.9375, + "rewards/rejected": -11.125, + "step": 1322 + }, + { + "epoch": 1.6233128834355828, + "grad_norm": 2.854089790732775, + "learning_rate": 6.954545454545454e-07, + "logits/chosen": -0.0157470703125, + "logits/rejected": -0.208984375, + "logps/chosen": -336.0, + "logps/rejected": -376.0, + "loss": 0.0118, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.03955078125, + "rewards/margins": 11.375, + "rewards/rejected": -11.375, + "step": 1323 + }, + { + "epoch": 1.6245398773006134, + "grad_norm": 13.205081682962938, + "learning_rate": 6.948347107438017e-07, + "logits/chosen": -0.1826171875, + "logits/rejected": -0.404296875, + "logps/chosen": -416.0, + "logps/rejected": -394.0, + "loss": 0.0504, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.265625, + "rewards/margins": 11.875, + "rewards/rejected": -11.625, + "step": 1324 + }, + { + "epoch": 1.6257668711656441, + "grad_norm": 10.512892902569012, + "learning_rate": 6.942148760330579e-07, + "logits/chosen": -0.07275390625, + "logits/rejected": -0.33203125, + "logps/chosen": -388.0, + "logps/rejected": -390.0, + "loss": 0.0409, + "rewards/accuracies": 0.984375, + "rewards/chosen": -0.427734375, + "rewards/margins": 11.3125, + "rewards/rejected": -11.75, + "step": 1325 + }, + { + "epoch": 1.6269938650306748, + "grad_norm": 16.75758409246438, + "learning_rate": 6.935950413223141e-07, + "logits/chosen": -0.19921875, + "logits/rejected": -0.33203125, + "logps/chosen": -412.0, + "logps/rejected": -418.0, + "loss": 0.0433, + "rewards/accuracies": 0.9765625, + "rewards/chosen": -0.076171875, + "rewards/margins": 12.25, + "rewards/rejected": -12.3125, + "step": 1326 + }, + { + "epoch": 1.6282208588957054, + "grad_norm": 17.757254444197297, + "learning_rate": 6.929752066115702e-07, + "logits/chosen": -0.1201171875, + "logits/rejected": -0.2265625, + "logps/chosen": -372.0, + "logps/rejected": -406.0, + "loss": 0.095, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.80859375, + "rewards/margins": 10.875, + "rewards/rejected": -11.6875, + "step": 1327 + }, + { + "epoch": 1.6294478527607361, + "grad_norm": 13.567940362966437, + "learning_rate": 6.923553719008265e-07, + "logits/chosen": -0.1953125, + "logits/rejected": -0.40625, + "logps/chosen": -414.0, + "logps/rejected": -398.0, + "loss": 0.0402, + "rewards/accuracies": 0.9765625, + "rewards/chosen": -0.330078125, + "rewards/margins": 11.75, + "rewards/rejected": -12.0625, + "step": 1328 + }, + { + "epoch": 1.6306748466257668, + "grad_norm": 15.662951369347224, + "learning_rate": 6.917355371900827e-07, + "logits/chosen": -0.2353515625, + "logits/rejected": -0.41015625, + "logps/chosen": -422.0, + "logps/rejected": -414.0, + "loss": 0.0593, + "rewards/accuracies": 0.96875, + "rewards/chosen": -0.6640625, + "rewards/margins": 10.75, + "rewards/rejected": -11.4375, + "step": 1329 + }, + { + "epoch": 1.6319018404907975, + "grad_norm": 10.57654744394464, + "learning_rate": 6.911157024793388e-07, + "logits/chosen": -0.12158203125, + "logits/rejected": -0.33203125, + "logps/chosen": -404.0, + "logps/rejected": -404.0, + "loss": 0.032, + "rewards/accuracies": 0.9765625, + "rewards/chosen": -0.37109375, + "rewards/margins": 11.5625, + "rewards/rejected": -11.875, + "step": 1330 + }, + { + "epoch": 1.6331288343558281, + "grad_norm": 7.812525368210755, + "learning_rate": 6.90495867768595e-07, + "logits/chosen": -0.169921875, + "logits/rejected": -0.2392578125, + "logps/chosen": -366.0, + "logps/rejected": -388.0, + "loss": 0.0254, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -1.0546875, + "rewards/margins": 11.8125, + "rewards/rejected": -12.875, + "step": 1331 + }, + { + "epoch": 1.6343558282208588, + "grad_norm": 12.529370972774968, + "learning_rate": 6.898760330578512e-07, + "logits/chosen": -0.263671875, + "logits/rejected": -0.416015625, + "logps/chosen": -386.0, + "logps/rejected": -390.0, + "loss": 0.0267, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.328125, + "rewards/margins": 11.375, + "rewards/rejected": -11.75, + "step": 1332 + }, + { + "epoch": 1.6355828220858895, + "grad_norm": 15.494182278895307, + "learning_rate": 6.892561983471075e-07, + "logits/chosen": -0.27734375, + "logits/rejected": -0.416015625, + "logps/chosen": -420.0, + "logps/rejected": -440.0, + "loss": 0.0527, + "rewards/accuracies": 0.96875, + "rewards/chosen": -0.6953125, + "rewards/margins": 11.1875, + "rewards/rejected": -11.875, + "step": 1333 + }, + { + "epoch": 1.6368098159509201, + "grad_norm": 8.698893043286152, + "learning_rate": 6.886363636363637e-07, + "logits/chosen": -0.10009765625, + "logits/rejected": -0.244140625, + "logps/chosen": -382.0, + "logps/rejected": -400.0, + "loss": 0.0306, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.5234375, + "rewards/margins": 12.1875, + "rewards/rejected": -11.6875, + "step": 1334 + }, + { + "epoch": 1.6380368098159508, + "grad_norm": 14.987956201874853, + "learning_rate": 6.880165289256199e-07, + "logits/chosen": -0.2451171875, + "logits/rejected": -0.44921875, + "logps/chosen": -432.0, + "logps/rejected": -386.0, + "loss": 0.0336, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 1.109375, + "rewards/margins": 11.625, + "rewards/rejected": -10.5625, + "step": 1335 + }, + { + "epoch": 1.6392638036809815, + "grad_norm": 15.60191378628464, + "learning_rate": 6.87396694214876e-07, + "logits/chosen": -0.361328125, + "logits/rejected": -0.53125, + "logps/chosen": -436.0, + "logps/rejected": -436.0, + "loss": 0.0541, + "rewards/accuracies": 0.9765625, + "rewards/chosen": -0.53515625, + "rewards/margins": 12.0625, + "rewards/rejected": -12.5625, + "step": 1336 + }, + { + "epoch": 1.6404907975460121, + "grad_norm": 6.108403854935857, + "learning_rate": 6.867768595041323e-07, + "logits/chosen": -0.2001953125, + "logits/rejected": -0.376953125, + "logps/chosen": -414.0, + "logps/rejected": -384.0, + "loss": 0.0178, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.1640625, + "rewards/margins": 11.8125, + "rewards/rejected": -11.6875, + "step": 1337 + }, + { + "epoch": 1.6417177914110428, + "grad_norm": 6.235103209750328, + "learning_rate": 6.861570247933885e-07, + "logits/chosen": -0.2734375, + "logits/rejected": -0.400390625, + "logps/chosen": -432.0, + "logps/rejected": -410.0, + "loss": 0.0253, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.435546875, + "rewards/margins": 11.625, + "rewards/rejected": -11.1875, + "step": 1338 + }, + { + "epoch": 1.6429447852760735, + "grad_norm": 13.809775053491462, + "learning_rate": 6.855371900826446e-07, + "logits/chosen": -0.13671875, + "logits/rejected": -0.1923828125, + "logps/chosen": -390.0, + "logps/rejected": -346.0, + "loss": 0.0495, + "rewards/accuracies": 0.984375, + "rewards/chosen": -0.033935546875, + "rewards/margins": 10.375, + "rewards/rejected": -10.375, + "step": 1339 + }, + { + "epoch": 1.6441717791411041, + "grad_norm": 14.649539845278797, + "learning_rate": 6.849173553719008e-07, + "logits/chosen": -0.060791015625, + "logits/rejected": -0.220703125, + "logps/chosen": -412.0, + "logps/rejected": -394.0, + "loss": 0.0626, + "rewards/accuracies": 0.9765625, + "rewards/chosen": 0.271484375, + "rewards/margins": 10.4375, + "rewards/rejected": -10.1875, + "step": 1340 + }, + { + "epoch": 1.645398773006135, + "grad_norm": 6.096103726931023, + "learning_rate": 6.84297520661157e-07, + "logits/chosen": -0.224609375, + "logits/rejected": -0.453125, + "logps/chosen": -438.0, + "logps/rejected": -398.0, + "loss": 0.0186, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.68359375, + "rewards/margins": 12.1875, + "rewards/rejected": -11.5, + "step": 1341 + }, + { + "epoch": 1.6466257668711657, + "grad_norm": 4.043549433090365, + "learning_rate": 6.836776859504133e-07, + "logits/chosen": -0.251953125, + "logits/rejected": -0.470703125, + "logps/chosen": -410.0, + "logps/rejected": -378.0, + "loss": 0.0267, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.60546875, + "rewards/margins": 11.0625, + "rewards/rejected": -10.4375, + "step": 1342 + }, + { + "epoch": 1.6478527607361964, + "grad_norm": 22.56474705026688, + "learning_rate": 6.830578512396694e-07, + "logits/chosen": -0.1708984375, + "logits/rejected": -0.40625, + "logps/chosen": -408.0, + "logps/rejected": -404.0, + "loss": 0.0661, + "rewards/accuracies": 0.9765625, + "rewards/chosen": -0.1650390625, + "rewards/margins": 10.75, + "rewards/rejected": -10.9375, + "step": 1343 + }, + { + "epoch": 1.649079754601227, + "grad_norm": 7.990667626479867, + "learning_rate": 6.824380165289256e-07, + "logits/chosen": -0.158203125, + "logits/rejected": -0.416015625, + "logps/chosen": -434.0, + "logps/rejected": -388.0, + "loss": 0.0183, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 1.203125, + "rewards/margins": 11.9375, + "rewards/rejected": -10.75, + "step": 1344 + }, + { + "epoch": 1.6503067484662577, + "grad_norm": 9.134005678853116, + "learning_rate": 6.818181818181818e-07, + "logits/chosen": -0.1943359375, + "logits/rejected": -0.34375, + "logps/chosen": -422.0, + "logps/rejected": -376.0, + "loss": 0.0366, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.361328125, + "rewards/margins": 10.0, + "rewards/rejected": -9.625, + "step": 1345 + }, + { + "epoch": 1.6515337423312884, + "grad_norm": 19.441190331880875, + "learning_rate": 6.811983471074381e-07, + "logits/chosen": -0.1259765625, + "logits/rejected": -0.32421875, + "logps/chosen": -438.0, + "logps/rejected": -396.0, + "loss": 0.0749, + "rewards/accuracies": 0.96875, + "rewards/chosen": 0.8125, + "rewards/margins": 11.125, + "rewards/rejected": -10.3125, + "step": 1346 + }, + { + "epoch": 1.652760736196319, + "grad_norm": 14.390784632473302, + "learning_rate": 6.805785123966942e-07, + "logits/chosen": -0.1416015625, + "logits/rejected": -0.3046875, + "logps/chosen": -420.0, + "logps/rejected": -420.0, + "loss": 0.0671, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.55078125, + "rewards/margins": 11.0, + "rewards/rejected": -10.4375, + "step": 1347 + }, + { + "epoch": 1.6539877300613497, + "grad_norm": 8.537444327620346, + "learning_rate": 6.799586776859504e-07, + "logits/chosen": -0.1455078125, + "logits/rejected": -0.32421875, + "logps/chosen": -340.0, + "logps/rejected": -382.0, + "loss": 0.0348, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.048095703125, + "rewards/margins": 10.6875, + "rewards/rejected": -10.6875, + "step": 1348 + }, + { + "epoch": 1.6552147239263804, + "grad_norm": 7.326685705649235, + "learning_rate": 6.793388429752066e-07, + "logits/chosen": -0.123046875, + "logits/rejected": -0.337890625, + "logps/chosen": -368.0, + "logps/rejected": -380.0, + "loss": 0.0303, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.80078125, + "rewards/margins": 11.1875, + "rewards/rejected": -10.375, + "step": 1349 + }, + { + "epoch": 1.656441717791411, + "grad_norm": 11.253855573956272, + "learning_rate": 6.787190082644628e-07, + "logits/chosen": -0.1669921875, + "logits/rejected": -0.3515625, + "logps/chosen": -422.0, + "logps/rejected": -432.0, + "loss": 0.0313, + "rewards/accuracies": 0.984375, + "rewards/chosen": 1.375, + "rewards/margins": 12.5, + "rewards/rejected": -11.125, + "step": 1350 + }, + { + "epoch": 1.6576687116564417, + "grad_norm": 20.220907692170936, + "learning_rate": 6.780991735537191e-07, + "logits/chosen": -0.11962890625, + "logits/rejected": -0.26171875, + "logps/chosen": -396.0, + "logps/rejected": -392.0, + "loss": 0.1149, + "rewards/accuracies": 0.953125, + "rewards/chosen": 0.1650390625, + "rewards/margins": 10.875, + "rewards/rejected": -10.75, + "step": 1351 + }, + { + "epoch": 1.6588957055214724, + "grad_norm": 8.156843128459297, + "learning_rate": 6.774793388429752e-07, + "logits/chosen": -0.2119140625, + "logits/rejected": -0.369140625, + "logps/chosen": -388.0, + "logps/rejected": -398.0, + "loss": 0.0315, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.2314453125, + "rewards/margins": 11.4375, + "rewards/rejected": -11.25, + "step": 1352 + }, + { + "epoch": 1.660122699386503, + "grad_norm": 6.558928813231135, + "learning_rate": 6.768595041322314e-07, + "logits/chosen": -0.10986328125, + "logits/rejected": -0.146484375, + "logps/chosen": -376.0, + "logps/rejected": -412.0, + "loss": 0.0286, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.0908203125, + "rewards/margins": 10.75, + "rewards/rejected": -10.8125, + "step": 1353 + }, + { + "epoch": 1.6613496932515337, + "grad_norm": 4.479021963688224, + "learning_rate": 6.762396694214876e-07, + "logits/chosen": -0.1826171875, + "logits/rejected": -0.333984375, + "logps/chosen": -368.0, + "logps/rejected": -394.0, + "loss": 0.025, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.41796875, + "rewards/margins": 11.6875, + "rewards/rejected": -11.25, + "step": 1354 + }, + { + "epoch": 1.6625766871165644, + "grad_norm": 12.69028856164417, + "learning_rate": 6.756198347107439e-07, + "logits/chosen": -0.20703125, + "logits/rejected": -0.345703125, + "logps/chosen": -424.0, + "logps/rejected": -422.0, + "loss": 0.0465, + "rewards/accuracies": 0.9765625, + "rewards/chosen": -0.20703125, + "rewards/margins": 10.875, + "rewards/rejected": -11.0625, + "step": 1355 + }, + { + "epoch": 1.6638036809815953, + "grad_norm": 6.416852390478722, + "learning_rate": 6.75e-07, + "logits/chosen": -0.16015625, + "logits/rejected": -0.322265625, + "logps/chosen": -408.0, + "logps/rejected": -412.0, + "loss": 0.0263, + "rewards/accuracies": 0.984375, + "rewards/chosen": -0.396484375, + "rewards/margins": 11.875, + "rewards/rejected": -12.25, + "step": 1356 + }, + { + "epoch": 1.665030674846626, + "grad_norm": 15.843916290648238, + "learning_rate": 6.743801652892562e-07, + "logits/chosen": -0.1259765625, + "logits/rejected": -0.353515625, + "logps/chosen": -436.0, + "logps/rejected": -410.0, + "loss": 0.0297, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.30859375, + "rewards/margins": 12.0, + "rewards/rejected": -11.6875, + "step": 1357 + }, + { + "epoch": 1.6662576687116566, + "grad_norm": 15.902047502639894, + "learning_rate": 6.737603305785124e-07, + "logits/chosen": -0.2080078125, + "logits/rejected": -0.453125, + "logps/chosen": -400.0, + "logps/rejected": -388.0, + "loss": 0.06, + "rewards/accuracies": 0.9765625, + "rewards/chosen": -0.11328125, + "rewards/margins": 10.8125, + "rewards/rejected": -10.9375, + "step": 1358 + }, + { + "epoch": 1.6674846625766873, + "grad_norm": 3.3586369592233085, + "learning_rate": 6.731404958677686e-07, + "logits/chosen": -0.1201171875, + "logits/rejected": -0.318359375, + "logps/chosen": -410.0, + "logps/rejected": -412.0, + "loss": 0.0118, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.6171875, + "rewards/margins": 12.3125, + "rewards/rejected": -11.6875, + "step": 1359 + }, + { + "epoch": 1.668711656441718, + "grad_norm": 12.937160200064577, + "learning_rate": 6.725206611570248e-07, + "logits/chosen": -0.388671875, + "logits/rejected": -0.6171875, + "logps/chosen": -392.0, + "logps/rejected": -380.0, + "loss": 0.0478, + "rewards/accuracies": 0.96875, + "rewards/chosen": 0.66015625, + "rewards/margins": 11.9375, + "rewards/rejected": -11.3125, + "step": 1360 + }, + { + "epoch": 1.6699386503067486, + "grad_norm": 11.349224605689916, + "learning_rate": 6.71900826446281e-07, + "logits/chosen": -0.0869140625, + "logits/rejected": -0.251953125, + "logps/chosen": -416.0, + "logps/rejected": -422.0, + "loss": 0.0391, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.201171875, + "rewards/margins": 11.0625, + "rewards/rejected": -10.875, + "step": 1361 + }, + { + "epoch": 1.6711656441717793, + "grad_norm": 16.146708757099134, + "learning_rate": 6.712809917355372e-07, + "logits/chosen": -0.119140625, + "logits/rejected": -0.2578125, + "logps/chosen": -402.0, + "logps/rejected": -376.0, + "loss": 0.051, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.474609375, + "rewards/margins": 11.5625, + "rewards/rejected": -11.0625, + "step": 1362 + }, + { + "epoch": 1.67239263803681, + "grad_norm": 3.943632661486204, + "learning_rate": 6.706611570247934e-07, + "logits/chosen": -0.224609375, + "logits/rejected": -0.41796875, + "logps/chosen": -398.0, + "logps/rejected": -366.0, + "loss": 0.0129, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.48046875, + "rewards/margins": 11.5625, + "rewards/rejected": -11.125, + "step": 1363 + }, + { + "epoch": 1.6736196319018406, + "grad_norm": 4.1389200081135655, + "learning_rate": 6.700413223140497e-07, + "logits/chosen": -0.09033203125, + "logits/rejected": -0.279296875, + "logps/chosen": -374.0, + "logps/rejected": -356.0, + "loss": 0.0111, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.453125, + "rewards/margins": 11.5625, + "rewards/rejected": -11.125, + "step": 1364 + }, + { + "epoch": 1.6748466257668713, + "grad_norm": 10.368495733373145, + "learning_rate": 6.694214876033058e-07, + "logits/chosen": -0.0458984375, + "logits/rejected": -0.2275390625, + "logps/chosen": -408.0, + "logps/rejected": -370.0, + "loss": 0.033, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.6640625, + "rewards/margins": 11.5625, + "rewards/rejected": -10.875, + "step": 1365 + }, + { + "epoch": 1.676073619631902, + "grad_norm": 10.90227263887412, + "learning_rate": 6.68801652892562e-07, + "logits/chosen": -0.12353515625, + "logits/rejected": -0.34765625, + "logps/chosen": -398.0, + "logps/rejected": -368.0, + "loss": 0.034, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.7109375, + "rewards/margins": 10.875, + "rewards/rejected": -10.125, + "step": 1366 + }, + { + "epoch": 1.6773006134969326, + "grad_norm": 14.33752123603866, + "learning_rate": 6.681818181818182e-07, + "logits/chosen": -0.1025390625, + "logits/rejected": -0.25390625, + "logps/chosen": -394.0, + "logps/rejected": -400.0, + "loss": 0.0614, + "rewards/accuracies": 0.9609375, + "rewards/chosen": 0.62109375, + "rewards/margins": 11.75, + "rewards/rejected": -11.125, + "step": 1367 + }, + { + "epoch": 1.6785276073619633, + "grad_norm": 10.4754734935581, + "learning_rate": 6.675619834710744e-07, + "logits/chosen": -0.3203125, + "logits/rejected": -0.462890625, + "logps/chosen": -400.0, + "logps/rejected": -390.0, + "loss": 0.0374, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.365234375, + "rewards/margins": 11.5, + "rewards/rejected": -11.125, + "step": 1368 + }, + { + "epoch": 1.679754601226994, + "grad_norm": 13.976696266239935, + "learning_rate": 6.669421487603306e-07, + "logits/chosen": -0.2451171875, + "logits/rejected": -0.4453125, + "logps/chosen": -404.0, + "logps/rejected": -394.0, + "loss": 0.0565, + "rewards/accuracies": 0.9765625, + "rewards/chosen": 0.10986328125, + "rewards/margins": 10.8125, + "rewards/rejected": -10.6875, + "step": 1369 + }, + { + "epoch": 1.6809815950920246, + "grad_norm": 10.93083816897367, + "learning_rate": 6.663223140495868e-07, + "logits/chosen": -0.1865234375, + "logits/rejected": -0.296875, + "logps/chosen": -382.0, + "logps/rejected": -402.0, + "loss": 0.0301, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.66796875, + "rewards/margins": 11.5, + "rewards/rejected": -10.875, + "step": 1370 + }, + { + "epoch": 1.6822085889570553, + "grad_norm": 23.02977282251546, + "learning_rate": 6.65702479338843e-07, + "logits/chosen": -0.1201171875, + "logits/rejected": -0.388671875, + "logps/chosen": -468.0, + "logps/rejected": -428.0, + "loss": 0.0674, + "rewards/accuracies": 0.9765625, + "rewards/chosen": 0.79296875, + "rewards/margins": 11.8125, + "rewards/rejected": -11.0, + "step": 1371 + }, + { + "epoch": 1.683435582822086, + "grad_norm": 6.529767256435788, + "learning_rate": 6.650826446280992e-07, + "logits/chosen": -0.1787109375, + "logits/rejected": -0.404296875, + "logps/chosen": -392.0, + "logps/rejected": -384.0, + "loss": 0.0215, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.6328125, + "rewards/margins": 11.6875, + "rewards/rejected": -11.0625, + "step": 1372 + }, + { + "epoch": 1.6846625766871166, + "grad_norm": 16.218767103827304, + "learning_rate": 6.644628099173554e-07, + "logits/chosen": -0.1318359375, + "logits/rejected": -0.216796875, + "logps/chosen": -376.0, + "logps/rejected": -410.0, + "loss": 0.0695, + "rewards/accuracies": 0.96875, + "rewards/chosen": 0.41796875, + "rewards/margins": 10.9375, + "rewards/rejected": -10.5, + "step": 1373 + }, + { + "epoch": 1.6858895705521473, + "grad_norm": 12.743040998410756, + "learning_rate": 6.638429752066116e-07, + "logits/chosen": -0.056396484375, + "logits/rejected": -0.29296875, + "logps/chosen": -394.0, + "logps/rejected": -368.0, + "loss": 0.0465, + "rewards/accuracies": 0.9765625, + "rewards/chosen": 0.1943359375, + "rewards/margins": 10.625, + "rewards/rejected": -10.4375, + "step": 1374 + }, + { + "epoch": 1.687116564417178, + "grad_norm": 11.631482703134425, + "learning_rate": 6.632231404958678e-07, + "logits/chosen": -0.0947265625, + "logits/rejected": -0.1962890625, + "logps/chosen": -328.0, + "logps/rejected": -358.0, + "loss": 0.0442, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.162109375, + "rewards/margins": 10.4375, + "rewards/rejected": -10.25, + "step": 1375 + }, + { + "epoch": 1.6883435582822086, + "grad_norm": 8.378575051965539, + "learning_rate": 6.62603305785124e-07, + "logits/chosen": -0.25, + "logits/rejected": -0.349609375, + "logps/chosen": -358.0, + "logps/rejected": -402.0, + "loss": 0.0241, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.51953125, + "rewards/margins": 11.875, + "rewards/rejected": -11.3125, + "step": 1376 + }, + { + "epoch": 1.6895705521472393, + "grad_norm": 6.071796287867807, + "learning_rate": 6.619834710743801e-07, + "logits/chosen": -0.2138671875, + "logits/rejected": -0.3984375, + "logps/chosen": -360.0, + "logps/rejected": -380.0, + "loss": 0.0204, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.75390625, + "rewards/margins": 12.3125, + "rewards/rejected": -11.5, + "step": 1377 + }, + { + "epoch": 1.69079754601227, + "grad_norm": 6.709699176048537, + "learning_rate": 6.613636363636364e-07, + "logits/chosen": -0.1767578125, + "logits/rejected": -0.35546875, + "logps/chosen": -366.0, + "logps/rejected": -362.0, + "loss": 0.0212, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.259765625, + "rewards/margins": 10.25, + "rewards/rejected": -10.5, + "step": 1378 + }, + { + "epoch": 1.6920245398773006, + "grad_norm": 9.699757128977033, + "learning_rate": 6.607438016528926e-07, + "logits/chosen": -0.1943359375, + "logits/rejected": -0.33984375, + "logps/chosen": -410.0, + "logps/rejected": -378.0, + "loss": 0.0463, + "rewards/accuracies": 0.9765625, + "rewards/chosen": 0.62890625, + "rewards/margins": 10.5, + "rewards/rejected": -9.875, + "step": 1379 + }, + { + "epoch": 1.6932515337423313, + "grad_norm": 13.747966139148284, + "learning_rate": 6.601239669421488e-07, + "logits/chosen": -0.14453125, + "logits/rejected": -0.37109375, + "logps/chosen": -422.0, + "logps/rejected": -354.0, + "loss": 0.0229, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.68359375, + "rewards/margins": 10.8125, + "rewards/rejected": -10.125, + "step": 1380 + }, + { + "epoch": 1.694478527607362, + "grad_norm": 12.861297927766412, + "learning_rate": 6.59504132231405e-07, + "logits/chosen": -0.0712890625, + "logits/rejected": -0.0966796875, + "logps/chosen": -362.0, + "logps/rejected": -364.0, + "loss": 0.0549, + "rewards/accuracies": 0.9765625, + "rewards/chosen": 0.26171875, + "rewards/margins": 9.9375, + "rewards/rejected": -9.6875, + "step": 1381 + }, + { + "epoch": 1.6957055214723926, + "grad_norm": 17.309330036067298, + "learning_rate": 6.588842975206612e-07, + "logits/chosen": -0.1376953125, + "logits/rejected": -0.25, + "logps/chosen": -348.0, + "logps/rejected": -376.0, + "loss": 0.0564, + "rewards/accuracies": 0.9765625, + "rewards/chosen": -0.046630859375, + "rewards/margins": 10.1875, + "rewards/rejected": -10.25, + "step": 1382 + }, + { + "epoch": 1.6969325153374233, + "grad_norm": 9.03911314369842, + "learning_rate": 6.582644628099174e-07, + "logits/chosen": -0.15234375, + "logits/rejected": -0.33203125, + "logps/chosen": -386.0, + "logps/rejected": -392.0, + "loss": 0.0248, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.95703125, + "rewards/margins": 11.375, + "rewards/rejected": -10.375, + "step": 1383 + }, + { + "epoch": 1.698159509202454, + "grad_norm": 16.803262144198918, + "learning_rate": 6.576446280991736e-07, + "logits/chosen": -0.01507568359375, + "logits/rejected": -0.130859375, + "logps/chosen": -370.0, + "logps/rejected": -352.0, + "loss": 0.0558, + "rewards/accuracies": 0.9765625, + "rewards/chosen": -0.04931640625, + "rewards/margins": 9.875, + "rewards/rejected": -9.875, + "step": 1384 + }, + { + "epoch": 1.6993865030674846, + "grad_norm": 9.539239180399765, + "learning_rate": 6.570247933884298e-07, + "logits/chosen": -0.146484375, + "logits/rejected": -0.265625, + "logps/chosen": -328.0, + "logps/rejected": -342.0, + "loss": 0.0495, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.462890625, + "rewards/margins": 10.375, + "rewards/rejected": -9.875, + "step": 1385 + }, + { + "epoch": 1.7006134969325153, + "grad_norm": 8.811291557763298, + "learning_rate": 6.564049586776859e-07, + "logits/chosen": -0.294921875, + "logits/rejected": -0.50390625, + "logps/chosen": -416.0, + "logps/rejected": -410.0, + "loss": 0.0357, + "rewards/accuracies": 0.984375, + "rewards/chosen": 1.1484375, + "rewards/margins": 12.0625, + "rewards/rejected": -10.9375, + "step": 1386 + }, + { + "epoch": 1.701840490797546, + "grad_norm": 9.305768357984167, + "learning_rate": 6.557851239669422e-07, + "logits/chosen": -0.2294921875, + "logits/rejected": -0.404296875, + "logps/chosen": -428.0, + "logps/rejected": -440.0, + "loss": 0.029, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.68359375, + "rewards/margins": 11.375, + "rewards/rejected": -10.6875, + "step": 1387 + }, + { + "epoch": 1.7030674846625766, + "grad_norm": 7.800969393038809, + "learning_rate": 6.551652892561984e-07, + "logits/chosen": -0.01025390625, + "logits/rejected": -0.1474609375, + "logps/chosen": -372.0, + "logps/rejected": -350.0, + "loss": 0.0219, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.27734375, + "rewards/margins": 9.875, + "rewards/rejected": -9.625, + "step": 1388 + }, + { + "epoch": 1.7042944785276073, + "grad_norm": 7.529656036678668, + "learning_rate": 6.545454545454546e-07, + "logits/chosen": -0.29296875, + "logits/rejected": -0.3984375, + "logps/chosen": -408.0, + "logps/rejected": -412.0, + "loss": 0.0288, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.83203125, + "rewards/margins": 11.75, + "rewards/rejected": -10.9375, + "step": 1389 + }, + { + "epoch": 1.705521472392638, + "grad_norm": 10.948563401414935, + "learning_rate": 6.539256198347107e-07, + "logits/chosen": -0.158203125, + "logits/rejected": -0.29296875, + "logps/chosen": -374.0, + "logps/rejected": -386.0, + "loss": 0.032, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.345703125, + "rewards/margins": 11.375, + "rewards/rejected": -11.0, + "step": 1390 + }, + { + "epoch": 1.7067484662576686, + "grad_norm": 8.365223158634247, + "learning_rate": 6.53305785123967e-07, + "logits/chosen": -0.158203125, + "logits/rejected": -0.40625, + "logps/chosen": -392.0, + "logps/rejected": -386.0, + "loss": 0.031, + "rewards/accuracies": 0.9765625, + "rewards/chosen": 1.125, + "rewards/margins": 11.125, + "rewards/rejected": -10.0, + "step": 1391 + }, + { + "epoch": 1.7079754601226993, + "grad_norm": 7.910192843345908, + "learning_rate": 6.526859504132232e-07, + "logits/chosen": 0.01153564453125, + "logits/rejected": -0.162109375, + "logps/chosen": -360.0, + "logps/rejected": -346.0, + "loss": 0.037, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.1953125, + "rewards/margins": 9.5, + "rewards/rejected": -9.6875, + "step": 1392 + }, + { + "epoch": 1.70920245398773, + "grad_norm": 10.155639345314109, + "learning_rate": 6.520661157024793e-07, + "logits/chosen": -0.19921875, + "logits/rejected": -0.388671875, + "logps/chosen": -372.0, + "logps/rejected": -344.0, + "loss": 0.0319, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.7734375, + "rewards/margins": 10.5, + "rewards/rejected": -9.75, + "step": 1393 + }, + { + "epoch": 1.7104294478527606, + "grad_norm": 11.604341441937358, + "learning_rate": 6.514462809917355e-07, + "logits/chosen": -0.248046875, + "logits/rejected": -0.41015625, + "logps/chosen": -406.0, + "logps/rejected": -382.0, + "loss": 0.0466, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.2890625, + "rewards/margins": 11.125, + "rewards/rejected": -10.8125, + "step": 1394 + }, + { + "epoch": 1.7116564417177913, + "grad_norm": 6.312301480394843, + "learning_rate": 6.508264462809917e-07, + "logits/chosen": -0.15234375, + "logits/rejected": -0.349609375, + "logps/chosen": -408.0, + "logps/rejected": -382.0, + "loss": 0.0145, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.66796875, + "rewards/margins": 11.0, + "rewards/rejected": -10.3125, + "step": 1395 + }, + { + "epoch": 1.712883435582822, + "grad_norm": 2.1889604215084524, + "learning_rate": 6.50206611570248e-07, + "logits/chosen": 0.01226806640625, + "logits/rejected": -0.2021484375, + "logps/chosen": -382.0, + "logps/rejected": -372.0, + "loss": 0.0102, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.60546875, + "rewards/margins": 12.0, + "rewards/rejected": -11.4375, + "step": 1396 + }, + { + "epoch": 1.7141104294478526, + "grad_norm": 8.251001626937791, + "learning_rate": 6.495867768595042e-07, + "logits/chosen": -0.1611328125, + "logits/rejected": -0.306640625, + "logps/chosen": -388.0, + "logps/rejected": -388.0, + "loss": 0.0266, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 1.09375, + "rewards/margins": 12.5625, + "rewards/rejected": -11.4375, + "step": 1397 + }, + { + "epoch": 1.7153374233128833, + "grad_norm": 12.58694243429673, + "learning_rate": 6.489669421487604e-07, + "logits/chosen": -0.1044921875, + "logits/rejected": -0.21484375, + "logps/chosen": -432.0, + "logps/rejected": -426.0, + "loss": 0.0351, + "rewards/accuracies": 0.9765625, + "rewards/chosen": 0.365234375, + "rewards/margins": 11.625, + "rewards/rejected": -11.25, + "step": 1398 + }, + { + "epoch": 1.716564417177914, + "grad_norm": 7.217759877360516, + "learning_rate": 6.483471074380165e-07, + "logits/chosen": -0.07275390625, + "logits/rejected": -0.181640625, + "logps/chosen": -362.0, + "logps/rejected": -402.0, + "loss": 0.0306, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.018798828125, + "rewards/margins": 10.5, + "rewards/rejected": -10.5625, + "step": 1399 + }, + { + "epoch": 1.7177914110429446, + "grad_norm": 21.648162026893676, + "learning_rate": 6.477272727272728e-07, + "logits/chosen": -0.072265625, + "logits/rejected": -0.271484375, + "logps/chosen": -344.0, + "logps/rejected": -352.0, + "loss": 0.0312, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.5, + "rewards/margins": 10.3125, + "rewards/rejected": -10.8125, + "step": 1400 + }, + { + "epoch": 1.7190184049079753, + "grad_norm": 6.36710851995934, + "learning_rate": 6.47107438016529e-07, + "logits/chosen": -0.134765625, + "logits/rejected": -0.36328125, + "logps/chosen": -446.0, + "logps/rejected": -442.0, + "loss": 0.0253, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.6796875, + "rewards/margins": 12.5, + "rewards/rejected": -11.8125, + "step": 1401 + }, + { + "epoch": 1.720245398773006, + "grad_norm": 6.937087144639035, + "learning_rate": 6.464876033057851e-07, + "logits/chosen": -0.1611328125, + "logits/rejected": -0.37890625, + "logps/chosen": -406.0, + "logps/rejected": -372.0, + "loss": 0.0245, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.4453125, + "rewards/margins": 11.875, + "rewards/rejected": -11.4375, + "step": 1402 + }, + { + "epoch": 1.7214723926380369, + "grad_norm": 10.928603486238107, + "learning_rate": 6.458677685950413e-07, + "logits/chosen": -0.0498046875, + "logits/rejected": -0.19921875, + "logps/chosen": -406.0, + "logps/rejected": -386.0, + "loss": 0.036, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.10888671875, + "rewards/margins": 10.9375, + "rewards/rejected": -10.8125, + "step": 1403 + }, + { + "epoch": 1.7226993865030675, + "grad_norm": 15.583870620895178, + "learning_rate": 6.452479338842975e-07, + "logits/chosen": -0.07568359375, + "logits/rejected": -0.228515625, + "logps/chosen": -336.0, + "logps/rejected": -330.0, + "loss": 0.0482, + "rewards/accuracies": 0.9765625, + "rewards/chosen": 0.7734375, + "rewards/margins": 11.75, + "rewards/rejected": -11.0, + "step": 1404 + }, + { + "epoch": 1.7239263803680982, + "grad_norm": 11.143077881808916, + "learning_rate": 6.446280991735538e-07, + "logits/chosen": -0.173828125, + "logits/rejected": -0.34765625, + "logps/chosen": -428.0, + "logps/rejected": -412.0, + "loss": 0.0398, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.240234375, + "rewards/margins": 11.75, + "rewards/rejected": -11.5, + "step": 1405 + }, + { + "epoch": 1.7251533742331289, + "grad_norm": 4.224299099232638, + "learning_rate": 6.440082644628099e-07, + "logits/chosen": -0.302734375, + "logits/rejected": -0.421875, + "logps/chosen": -328.0, + "logps/rejected": -374.0, + "loss": 0.0104, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.39453125, + "rewards/margins": 11.75, + "rewards/rejected": -12.125, + "step": 1406 + }, + { + "epoch": 1.7263803680981595, + "grad_norm": 13.330656715870173, + "learning_rate": 6.433884297520661e-07, + "logits/chosen": -0.111328125, + "logits/rejected": -0.390625, + "logps/chosen": -442.0, + "logps/rejected": -426.0, + "loss": 0.0462, + "rewards/accuracies": 0.9765625, + "rewards/chosen": 0.3671875, + "rewards/margins": 11.75, + "rewards/rejected": -11.375, + "step": 1407 + }, + { + "epoch": 1.7276073619631902, + "grad_norm": 7.332633738244965, + "learning_rate": 6.427685950413223e-07, + "logits/chosen": -0.1328125, + "logits/rejected": -0.30078125, + "logps/chosen": -470.0, + "logps/rejected": -440.0, + "loss": 0.0309, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.765625, + "rewards/margins": 13.0, + "rewards/rejected": -12.1875, + "step": 1408 + }, + { + "epoch": 1.7288343558282209, + "grad_norm": 12.35088317930436, + "learning_rate": 6.421487603305786e-07, + "logits/chosen": -0.0927734375, + "logits/rejected": -0.255859375, + "logps/chosen": -380.0, + "logps/rejected": -376.0, + "loss": 0.0446, + "rewards/accuracies": 0.9765625, + "rewards/chosen": 0.0240478515625, + "rewards/margins": 11.75, + "rewards/rejected": -11.75, + "step": 1409 + }, + { + "epoch": 1.7300613496932515, + "grad_norm": 12.863248116987783, + "learning_rate": 6.415289256198348e-07, + "logits/chosen": -0.1787109375, + "logits/rejected": -0.353515625, + "logps/chosen": -400.0, + "logps/rejected": -388.0, + "loss": 0.0733, + "rewards/accuracies": 0.9609375, + "rewards/chosen": -0.5859375, + "rewards/margins": 10.9375, + "rewards/rejected": -11.5, + "step": 1410 + }, + { + "epoch": 1.7312883435582822, + "grad_norm": 5.395301448670732, + "learning_rate": 6.409090909090908e-07, + "logits/chosen": -0.130859375, + "logits/rejected": -0.279296875, + "logps/chosen": -360.0, + "logps/rejected": -382.0, + "loss": 0.0197, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.451171875, + "rewards/margins": 11.4375, + "rewards/rejected": -11.875, + "step": 1411 + }, + { + "epoch": 1.7325153374233129, + "grad_norm": 11.512023543876147, + "learning_rate": 6.402892561983471e-07, + "logits/chosen": -0.0908203125, + "logits/rejected": -0.330078125, + "logps/chosen": -430.0, + "logps/rejected": -398.0, + "loss": 0.063, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.1904296875, + "rewards/margins": 11.8125, + "rewards/rejected": -11.625, + "step": 1412 + }, + { + "epoch": 1.7337423312883435, + "grad_norm": 15.349502769641875, + "learning_rate": 6.396694214876033e-07, + "logits/chosen": -0.2333984375, + "logits/rejected": -0.4453125, + "logps/chosen": -400.0, + "logps/rejected": -392.0, + "loss": 0.0712, + "rewards/accuracies": 0.96875, + "rewards/chosen": -0.734375, + "rewards/margins": 11.875, + "rewards/rejected": -12.625, + "step": 1413 + }, + { + "epoch": 1.7349693251533742, + "grad_norm": 10.27908988260348, + "learning_rate": 6.390495867768596e-07, + "logits/chosen": -0.10595703125, + "logits/rejected": -0.287109375, + "logps/chosen": -436.0, + "logps/rejected": -390.0, + "loss": 0.0453, + "rewards/accuracies": 0.9765625, + "rewards/chosen": 0.158203125, + "rewards/margins": 11.4375, + "rewards/rejected": -11.3125, + "step": 1414 + }, + { + "epoch": 1.7361963190184049, + "grad_norm": 9.292914335874663, + "learning_rate": 6.384297520661157e-07, + "logits/chosen": -0.169921875, + "logits/rejected": -0.349609375, + "logps/chosen": -350.0, + "logps/rejected": -380.0, + "loss": 0.041, + "rewards/accuracies": 0.9765625, + "rewards/chosen": -0.15625, + "rewards/margins": 11.5, + "rewards/rejected": -11.6875, + "step": 1415 + }, + { + "epoch": 1.7374233128834355, + "grad_norm": 7.447145147683031, + "learning_rate": 6.378099173553719e-07, + "logits/chosen": -0.09521484375, + "logits/rejected": -0.361328125, + "logps/chosen": -420.0, + "logps/rejected": -406.0, + "loss": 0.027, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.609375, + "rewards/margins": 12.625, + "rewards/rejected": -12.0625, + "step": 1416 + }, + { + "epoch": 1.7386503067484662, + "grad_norm": 8.792777557704559, + "learning_rate": 6.371900826446281e-07, + "logits/chosen": -0.2353515625, + "logits/rejected": -0.37109375, + "logps/chosen": -374.0, + "logps/rejected": -386.0, + "loss": 0.0291, + "rewards/accuracies": 0.984375, + "rewards/chosen": -0.162109375, + "rewards/margins": 11.875, + "rewards/rejected": -12.0, + "step": 1417 + }, + { + "epoch": 1.7398773006134969, + "grad_norm": 13.38073206624744, + "learning_rate": 6.365702479338844e-07, + "logits/chosen": -0.12060546875, + "logits/rejected": -0.3671875, + "logps/chosen": -396.0, + "logps/rejected": -408.0, + "loss": 0.0355, + "rewards/accuracies": 0.984375, + "rewards/chosen": -0.02587890625, + "rewards/margins": 12.4375, + "rewards/rejected": -12.4375, + "step": 1418 + }, + { + "epoch": 1.7411042944785278, + "grad_norm": 10.607683414699668, + "learning_rate": 6.359504132231405e-07, + "logits/chosen": -0.0908203125, + "logits/rejected": -0.21875, + "logps/chosen": -346.0, + "logps/rejected": -342.0, + "loss": 0.0453, + "rewards/accuracies": 0.9765625, + "rewards/chosen": -0.63671875, + "rewards/margins": 10.5625, + "rewards/rejected": -11.1875, + "step": 1419 + }, + { + "epoch": 1.7423312883435584, + "grad_norm": 10.995863145556806, + "learning_rate": 6.353305785123966e-07, + "logits/chosen": -0.3203125, + "logits/rejected": -0.466796875, + "logps/chosen": -430.0, + "logps/rejected": -422.0, + "loss": 0.0369, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.00186920166015625, + "rewards/margins": 12.375, + "rewards/rejected": -12.375, + "step": 1420 + }, + { + "epoch": 1.743558282208589, + "grad_norm": 4.993783324754888, + "learning_rate": 6.347107438016529e-07, + "logits/chosen": -0.2119140625, + "logits/rejected": -0.412109375, + "logps/chosen": -378.0, + "logps/rejected": -392.0, + "loss": 0.0189, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.1796875, + "rewards/margins": 12.5625, + "rewards/rejected": -12.375, + "step": 1421 + }, + { + "epoch": 1.7447852760736198, + "grad_norm": 9.402562866512993, + "learning_rate": 6.340909090909091e-07, + "logits/chosen": -0.08935546875, + "logits/rejected": -0.251953125, + "logps/chosen": -398.0, + "logps/rejected": -392.0, + "loss": 0.034, + "rewards/accuracies": 0.984375, + "rewards/chosen": -0.1494140625, + "rewards/margins": 11.125, + "rewards/rejected": -11.3125, + "step": 1422 + }, + { + "epoch": 1.7460122699386504, + "grad_norm": 7.9284649955491595, + "learning_rate": 6.334710743801654e-07, + "logits/chosen": -0.16015625, + "logits/rejected": -0.33984375, + "logps/chosen": -382.0, + "logps/rejected": -422.0, + "loss": 0.0293, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.91796875, + "rewards/margins": 12.3125, + "rewards/rejected": -13.25, + "step": 1423 + }, + { + "epoch": 1.747239263803681, + "grad_norm": 6.1006118198416805, + "learning_rate": 6.328512396694215e-07, + "logits/chosen": -0.255859375, + "logits/rejected": -0.49609375, + "logps/chosen": -424.0, + "logps/rejected": -410.0, + "loss": 0.0223, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.23046875, + "rewards/margins": 12.125, + "rewards/rejected": -11.875, + "step": 1424 + }, + { + "epoch": 1.7484662576687118, + "grad_norm": 12.036819296468309, + "learning_rate": 6.322314049586777e-07, + "logits/chosen": -0.12255859375, + "logits/rejected": -0.2373046875, + "logps/chosen": -386.0, + "logps/rejected": -392.0, + "loss": 0.0329, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.1484375, + "rewards/margins": 11.875, + "rewards/rejected": -11.75, + "step": 1425 + }, + { + "epoch": 1.7496932515337424, + "grad_norm": 18.4014829822398, + "learning_rate": 6.316115702479339e-07, + "logits/chosen": -0.1259765625, + "logits/rejected": -0.30078125, + "logps/chosen": -374.0, + "logps/rejected": -388.0, + "loss": 0.0692, + "rewards/accuracies": 0.9765625, + "rewards/chosen": 0.271484375, + "rewards/margins": 11.4375, + "rewards/rejected": -11.1875, + "step": 1426 + }, + { + "epoch": 1.7509202453987731, + "grad_norm": 12.513859109903631, + "learning_rate": 6.309917355371902e-07, + "logits/chosen": -0.1298828125, + "logits/rejected": -0.265625, + "logps/chosen": -380.0, + "logps/rejected": -386.0, + "loss": 0.0557, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.427734375, + "rewards/margins": 11.3125, + "rewards/rejected": -10.875, + "step": 1427 + }, + { + "epoch": 1.7521472392638038, + "grad_norm": 11.939507817491306, + "learning_rate": 6.303719008264463e-07, + "logits/chosen": -0.1845703125, + "logits/rejected": -0.298828125, + "logps/chosen": -420.0, + "logps/rejected": -410.0, + "loss": 0.0502, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.8671875, + "rewards/margins": 12.375, + "rewards/rejected": -11.5, + "step": 1428 + }, + { + "epoch": 1.7533742331288344, + "grad_norm": 6.914751025127418, + "learning_rate": 6.297520661157024e-07, + "logits/chosen": -0.134765625, + "logits/rejected": -0.296875, + "logps/chosen": -354.0, + "logps/rejected": -360.0, + "loss": 0.0278, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.37109375, + "rewards/margins": 11.5, + "rewards/rejected": -11.125, + "step": 1429 + }, + { + "epoch": 1.7546012269938651, + "grad_norm": 5.616187235790868, + "learning_rate": 6.291322314049587e-07, + "logits/chosen": -0.212890625, + "logits/rejected": -0.298828125, + "logps/chosen": -388.0, + "logps/rejected": -416.0, + "loss": 0.0209, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.22265625, + "rewards/margins": 11.625, + "rewards/rejected": -11.4375, + "step": 1430 + }, + { + "epoch": 1.7558282208588958, + "grad_norm": 9.370787243161711, + "learning_rate": 6.285123966942149e-07, + "logits/chosen": -0.0196533203125, + "logits/rejected": -0.181640625, + "logps/chosen": -396.0, + "logps/rejected": -380.0, + "loss": 0.0259, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.0771484375, + "rewards/margins": 10.4375, + "rewards/rejected": -10.3125, + "step": 1431 + }, + { + "epoch": 1.7570552147239265, + "grad_norm": 4.016040968990093, + "learning_rate": 6.278925619834711e-07, + "logits/chosen": -0.181640625, + "logits/rejected": -0.22265625, + "logps/chosen": -336.0, + "logps/rejected": -380.0, + "loss": 0.012, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.41796875, + "rewards/margins": 11.875, + "rewards/rejected": -11.5, + "step": 1432 + }, + { + "epoch": 1.7582822085889571, + "grad_norm": 12.729230743822482, + "learning_rate": 6.272727272727272e-07, + "logits/chosen": 0.0184326171875, + "logits/rejected": -0.0966796875, + "logps/chosen": -360.0, + "logps/rejected": -354.0, + "loss": 0.0589, + "rewards/accuracies": 0.984375, + "rewards/chosen": -0.134765625, + "rewards/margins": 10.0, + "rewards/rejected": -10.1875, + "step": 1433 + }, + { + "epoch": 1.7595092024539878, + "grad_norm": 18.378377289476965, + "learning_rate": 6.266528925619835e-07, + "logits/chosen": -0.0145263671875, + "logits/rejected": -0.1103515625, + "logps/chosen": -344.0, + "logps/rejected": -410.0, + "loss": 0.0787, + "rewards/accuracies": 0.96875, + "rewards/chosen": 0.2216796875, + "rewards/margins": 10.3125, + "rewards/rejected": -10.125, + "step": 1434 + }, + { + "epoch": 1.7607361963190185, + "grad_norm": 6.317649451474902, + "learning_rate": 6.260330578512397e-07, + "logits/chosen": -0.05810546875, + "logits/rejected": -0.169921875, + "logps/chosen": -352.0, + "logps/rejected": -368.0, + "loss": 0.0262, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.31640625, + "rewards/margins": 10.625, + "rewards/rejected": -10.3125, + "step": 1435 + }, + { + "epoch": 1.7619631901840491, + "grad_norm": 4.69256168204721, + "learning_rate": 6.254132231404959e-07, + "logits/chosen": -0.09619140625, + "logits/rejected": -0.306640625, + "logps/chosen": -388.0, + "logps/rejected": -404.0, + "loss": 0.0142, + "rewards/accuracies": 1.0, + "rewards/chosen": 1.46875, + "rewards/margins": 12.4375, + "rewards/rejected": -10.9375, + "step": 1436 + }, + { + "epoch": 1.7631901840490798, + "grad_norm": 29.257144870088368, + "learning_rate": 6.247933884297521e-07, + "logits/chosen": -0.248046875, + "logits/rejected": -0.453125, + "logps/chosen": -396.0, + "logps/rejected": -416.0, + "loss": 0.0505, + "rewards/accuracies": 0.96875, + "rewards/chosen": 0.53125, + "rewards/margins": 11.9375, + "rewards/rejected": -11.375, + "step": 1437 + }, + { + "epoch": 1.7644171779141105, + "grad_norm": 3.93681972201715, + "learning_rate": 6.241735537190082e-07, + "logits/chosen": -0.12109375, + "logits/rejected": -0.27734375, + "logps/chosen": -382.0, + "logps/rejected": -380.0, + "loss": 0.0156, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.2138671875, + "rewards/margins": 10.625, + "rewards/rejected": -10.375, + "step": 1438 + }, + { + "epoch": 1.7656441717791411, + "grad_norm": 7.385811382795677, + "learning_rate": 6.235537190082645e-07, + "logits/chosen": -0.1728515625, + "logits/rejected": -0.41015625, + "logps/chosen": -368.0, + "logps/rejected": -372.0, + "loss": 0.0188, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.13671875, + "rewards/margins": 10.5625, + "rewards/rejected": -10.6875, + "step": 1439 + }, + { + "epoch": 1.7668711656441718, + "grad_norm": 5.11550436529797, + "learning_rate": 6.229338842975207e-07, + "logits/chosen": -0.138671875, + "logits/rejected": -0.34375, + "logps/chosen": -470.0, + "logps/rejected": -432.0, + "loss": 0.0149, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.81640625, + "rewards/margins": 12.5, + "rewards/rejected": -11.6875, + "step": 1440 + }, + { + "epoch": 1.7680981595092025, + "grad_norm": 8.599376446703577, + "learning_rate": 6.223140495867769e-07, + "logits/chosen": -0.28515625, + "logits/rejected": -0.359375, + "logps/chosen": -360.0, + "logps/rejected": -398.0, + "loss": 0.0371, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.2001953125, + "rewards/margins": 11.8125, + "rewards/rejected": -11.625, + "step": 1441 + }, + { + "epoch": 1.7693251533742331, + "grad_norm": 20.234174712489157, + "learning_rate": 6.21694214876033e-07, + "logits/chosen": -0.21875, + "logits/rejected": -0.43359375, + "logps/chosen": -434.0, + "logps/rejected": -404.0, + "loss": 0.0897, + "rewards/accuracies": 0.9609375, + "rewards/chosen": 0.057373046875, + "rewards/margins": 11.4375, + "rewards/rejected": -11.375, + "step": 1442 + }, + { + "epoch": 1.7705521472392638, + "grad_norm": 8.089818483065772, + "learning_rate": 6.210743801652893e-07, + "logits/chosen": -0.06640625, + "logits/rejected": -0.21484375, + "logps/chosen": -422.0, + "logps/rejected": -416.0, + "loss": 0.0331, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.46875, + "rewards/margins": 12.0625, + "rewards/rejected": -11.5625, + "step": 1443 + }, + { + "epoch": 1.7717791411042945, + "grad_norm": 7.490857640483026, + "learning_rate": 6.204545454545455e-07, + "logits/chosen": -0.1435546875, + "logits/rejected": -0.3046875, + "logps/chosen": -398.0, + "logps/rejected": -398.0, + "loss": 0.023, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.69921875, + "rewards/margins": 11.9375, + "rewards/rejected": -11.25, + "step": 1444 + }, + { + "epoch": 1.7730061349693251, + "grad_norm": 4.507209000487265, + "learning_rate": 6.198347107438017e-07, + "logits/chosen": -0.09375, + "logits/rejected": -0.267578125, + "logps/chosen": -330.0, + "logps/rejected": -350.0, + "loss": 0.0205, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.20703125, + "rewards/margins": 10.75, + "rewards/rejected": -11.0, + "step": 1445 + }, + { + "epoch": 1.7742331288343558, + "grad_norm": 9.506356568202962, + "learning_rate": 6.192148760330579e-07, + "logits/chosen": -0.0830078125, + "logits/rejected": -0.2412109375, + "logps/chosen": -418.0, + "logps/rejected": -416.0, + "loss": 0.0362, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.055908203125, + "rewards/margins": 12.25, + "rewards/rejected": -12.1875, + "step": 1446 + }, + { + "epoch": 1.7754601226993865, + "grad_norm": 10.56612980003067, + "learning_rate": 6.18595041322314e-07, + "logits/chosen": -0.24609375, + "logits/rejected": -0.40625, + "logps/chosen": -386.0, + "logps/rejected": -392.0, + "loss": 0.0424, + "rewards/accuracies": 0.96875, + "rewards/chosen": -0.2314453125, + "rewards/margins": 11.8125, + "rewards/rejected": -12.0625, + "step": 1447 + }, + { + "epoch": 1.7766871165644171, + "grad_norm": 17.112227162070322, + "learning_rate": 6.179752066115703e-07, + "logits/chosen": -0.0888671875, + "logits/rejected": -0.208984375, + "logps/chosen": -346.0, + "logps/rejected": -376.0, + "loss": 0.0438, + "rewards/accuracies": 0.984375, + "rewards/chosen": -0.75390625, + "rewards/margins": 10.8125, + "rewards/rejected": -11.5625, + "step": 1448 + }, + { + "epoch": 1.7779141104294478, + "grad_norm": 7.673926079275721, + "learning_rate": 6.173553719008264e-07, + "logits/chosen": -0.003936767578125, + "logits/rejected": -0.201171875, + "logps/chosen": -430.0, + "logps/rejected": -420.0, + "loss": 0.0294, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.091796875, + "rewards/margins": 11.25, + "rewards/rejected": -11.3125, + "step": 1449 + }, + { + "epoch": 1.7791411042944785, + "grad_norm": 10.143124845251792, + "learning_rate": 6.167355371900827e-07, + "logits/chosen": -0.166015625, + "logits/rejected": -0.333984375, + "logps/chosen": -428.0, + "logps/rejected": -398.0, + "loss": 0.0361, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.099609375, + "rewards/margins": 11.8125, + "rewards/rejected": -11.6875, + "step": 1450 + }, + { + "epoch": 1.7803680981595091, + "grad_norm": 6.237913509142783, + "learning_rate": 6.161157024793388e-07, + "logits/chosen": -0.0751953125, + "logits/rejected": -0.23046875, + "logps/chosen": -422.0, + "logps/rejected": -410.0, + "loss": 0.0221, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.3515625, + "rewards/margins": 12.1875, + "rewards/rejected": -11.8125, + "step": 1451 + }, + { + "epoch": 1.7815950920245398, + "grad_norm": 13.734009163796884, + "learning_rate": 6.154958677685951e-07, + "logits/chosen": -0.29296875, + "logits/rejected": -0.55078125, + "logps/chosen": -458.0, + "logps/rejected": -422.0, + "loss": 0.0286, + "rewards/accuracies": 0.984375, + "rewards/chosen": 1.2265625, + "rewards/margins": 13.5625, + "rewards/rejected": -12.375, + "step": 1452 + }, + { + "epoch": 1.7828220858895705, + "grad_norm": 5.142550278798512, + "learning_rate": 6.148760330578512e-07, + "logits/chosen": -0.166015625, + "logits/rejected": -0.3671875, + "logps/chosen": -412.0, + "logps/rejected": -388.0, + "loss": 0.0161, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.265625, + "rewards/margins": 12.375, + "rewards/rejected": -12.125, + "step": 1453 + }, + { + "epoch": 1.7840490797546011, + "grad_norm": 11.252004249668094, + "learning_rate": 6.142561983471075e-07, + "logits/chosen": -0.053955078125, + "logits/rejected": -0.25390625, + "logps/chosen": -402.0, + "logps/rejected": -396.0, + "loss": 0.0341, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.04833984375, + "rewards/margins": 12.0, + "rewards/rejected": -12.0, + "step": 1454 + }, + { + "epoch": 1.7852760736196318, + "grad_norm": 5.828112394054352, + "learning_rate": 6.136363636363637e-07, + "logits/chosen": -0.1416015625, + "logits/rejected": -0.267578125, + "logps/chosen": -398.0, + "logps/rejected": -418.0, + "loss": 0.024, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.1318359375, + "rewards/margins": 11.5625, + "rewards/rejected": -11.4375, + "step": 1455 + }, + { + "epoch": 1.7865030674846625, + "grad_norm": 8.622121139819185, + "learning_rate": 6.130165289256198e-07, + "logits/chosen": -0.177734375, + "logits/rejected": -0.29296875, + "logps/chosen": -370.0, + "logps/rejected": -366.0, + "loss": 0.0437, + "rewards/accuracies": 0.984375, + "rewards/chosen": -0.232421875, + "rewards/margins": 10.3125, + "rewards/rejected": -10.5625, + "step": 1456 + }, + { + "epoch": 1.7877300613496931, + "grad_norm": 10.346638960597081, + "learning_rate": 6.123966942148761e-07, + "logits/chosen": -0.19140625, + "logits/rejected": -0.41796875, + "logps/chosen": -424.0, + "logps/rejected": -418.0, + "loss": 0.0402, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.09521484375, + "rewards/margins": 11.8125, + "rewards/rejected": -11.75, + "step": 1457 + }, + { + "epoch": 1.7889570552147238, + "grad_norm": 9.510862849605985, + "learning_rate": 6.117768595041322e-07, + "logits/chosen": -0.04736328125, + "logits/rejected": -0.130859375, + "logps/chosen": -360.0, + "logps/rejected": -396.0, + "loss": 0.0352, + "rewards/accuracies": 0.984375, + "rewards/chosen": -0.19921875, + "rewards/margins": 10.3125, + "rewards/rejected": -10.5, + "step": 1458 + }, + { + "epoch": 1.7901840490797545, + "grad_norm": 2985.9302248963727, + "learning_rate": 6.111570247933885e-07, + "logits/chosen": -0.1572265625, + "logits/rejected": -0.376953125, + "logps/chosen": -404.0, + "logps/rejected": -386.0, + "loss": 1.516, + "rewards/accuracies": 0.984375, + "rewards/chosen": -1.1484375, + "rewards/margins": 10.6875, + "rewards/rejected": -11.8125, + "step": 1459 + }, + { + "epoch": 1.7914110429447851, + "grad_norm": 6.456485664548984, + "learning_rate": 6.105371900826446e-07, + "logits/chosen": -0.12060546875, + "logits/rejected": -0.265625, + "logps/chosen": -422.0, + "logps/rejected": -406.0, + "loss": 0.0201, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.5234375, + "rewards/margins": 11.625, + "rewards/rejected": -11.125, + "step": 1460 + }, + { + "epoch": 1.7926380368098158, + "grad_norm": 12.803301057131963, + "learning_rate": 6.099173553719009e-07, + "logits/chosen": -0.11328125, + "logits/rejected": -0.32421875, + "logps/chosen": -430.0, + "logps/rejected": -416.0, + "loss": 0.0432, + "rewards/accuracies": 0.9765625, + "rewards/chosen": 0.52734375, + "rewards/margins": 13.625, + "rewards/rejected": -13.125, + "step": 1461 + }, + { + "epoch": 1.7938650306748465, + "grad_norm": 8.137637416090907, + "learning_rate": 6.09297520661157e-07, + "logits/chosen": -0.13671875, + "logits/rejected": -0.2021484375, + "logps/chosen": -326.0, + "logps/rejected": -372.0, + "loss": 0.0317, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.201171875, + "rewards/margins": 11.6875, + "rewards/rejected": -11.5, + "step": 1462 + }, + { + "epoch": 1.7950920245398772, + "grad_norm": 9.5142759213274, + "learning_rate": 6.086776859504133e-07, + "logits/chosen": -0.058837890625, + "logits/rejected": -0.29296875, + "logps/chosen": -406.0, + "logps/rejected": -392.0, + "loss": 0.0229, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.33203125, + "rewards/margins": 11.375, + "rewards/rejected": -11.0625, + "step": 1463 + }, + { + "epoch": 1.7963190184049078, + "grad_norm": 16.72592677806428, + "learning_rate": 6.080578512396695e-07, + "logits/chosen": -0.1884765625, + "logits/rejected": -0.32421875, + "logps/chosen": -346.0, + "logps/rejected": -364.0, + "loss": 0.0569, + "rewards/accuracies": 0.9765625, + "rewards/chosen": 0.1328125, + "rewards/margins": 11.625, + "rewards/rejected": -11.5, + "step": 1464 + }, + { + "epoch": 1.7975460122699385, + "grad_norm": 8.34969796778253, + "learning_rate": 6.074380165289256e-07, + "logits/chosen": -0.2294921875, + "logits/rejected": -0.3515625, + "logps/chosen": -396.0, + "logps/rejected": -414.0, + "loss": 0.0236, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.52734375, + "rewards/margins": 12.5625, + "rewards/rejected": -12.0, + "step": 1465 + }, + { + "epoch": 1.7987730061349694, + "grad_norm": 6.129551015730449, + "learning_rate": 6.068181818181818e-07, + "logits/chosen": -0.1455078125, + "logits/rejected": -0.359375, + "logps/chosen": -364.0, + "logps/rejected": -378.0, + "loss": 0.0164, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.10205078125, + "rewards/margins": 11.9375, + "rewards/rejected": -12.0, + "step": 1466 + }, + { + "epoch": 1.8, + "grad_norm": 6.8967202729590245, + "learning_rate": 6.06198347107438e-07, + "logits/chosen": -0.2060546875, + "logits/rejected": -0.3046875, + "logps/chosen": -426.0, + "logps/rejected": -420.0, + "loss": 0.0195, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.76171875, + "rewards/margins": 11.6875, + "rewards/rejected": -10.9375, + "step": 1467 + }, + { + "epoch": 1.8012269938650307, + "grad_norm": 8.361065202748346, + "learning_rate": 6.055785123966943e-07, + "logits/chosen": -0.10107421875, + "logits/rejected": -0.29296875, + "logps/chosen": -412.0, + "logps/rejected": -402.0, + "loss": 0.0182, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 1.03125, + "rewards/margins": 12.6875, + "rewards/rejected": -11.625, + "step": 1468 + }, + { + "epoch": 1.8024539877300614, + "grad_norm": 12.455616695494426, + "learning_rate": 6.049586776859504e-07, + "logits/chosen": -0.08203125, + "logits/rejected": -0.224609375, + "logps/chosen": -402.0, + "logps/rejected": -390.0, + "loss": 0.0396, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.06494140625, + "rewards/margins": 11.75, + "rewards/rejected": -11.6875, + "step": 1469 + }, + { + "epoch": 1.803680981595092, + "grad_norm": 9.292975088281931, + "learning_rate": 6.043388429752066e-07, + "logits/chosen": -0.05126953125, + "logits/rejected": -0.2734375, + "logps/chosen": -404.0, + "logps/rejected": -384.0, + "loss": 0.0445, + "rewards/accuracies": 0.984375, + "rewards/chosen": -0.419921875, + "rewards/margins": 10.9375, + "rewards/rejected": -11.375, + "step": 1470 + }, + { + "epoch": 1.8049079754601227, + "grad_norm": 9.047461001821528, + "learning_rate": 6.037190082644628e-07, + "logits/chosen": -0.07177734375, + "logits/rejected": -0.1748046875, + "logps/chosen": -372.0, + "logps/rejected": -396.0, + "loss": 0.0295, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.478515625, + "rewards/margins": 11.125, + "rewards/rejected": -10.6875, + "step": 1471 + }, + { + "epoch": 1.8061349693251534, + "grad_norm": 13.42466742331635, + "learning_rate": 6.030991735537191e-07, + "logits/chosen": -0.1083984375, + "logits/rejected": -0.2470703125, + "logps/chosen": -392.0, + "logps/rejected": -398.0, + "loss": 0.04, + "rewards/accuracies": 0.9765625, + "rewards/chosen": 0.34765625, + "rewards/margins": 12.25, + "rewards/rejected": -11.875, + "step": 1472 + }, + { + "epoch": 1.807361963190184, + "grad_norm": 6.913536686973756, + "learning_rate": 6.024793388429752e-07, + "logits/chosen": -0.14453125, + "logits/rejected": -0.40234375, + "logps/chosen": -384.0, + "logps/rejected": -390.0, + "loss": 0.0335, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.287109375, + "rewards/margins": 12.5625, + "rewards/rejected": -12.25, + "step": 1473 + }, + { + "epoch": 1.8085889570552147, + "grad_norm": 17.128950617994402, + "learning_rate": 6.018595041322314e-07, + "logits/chosen": -0.017822265625, + "logits/rejected": -0.234375, + "logps/chosen": -426.0, + "logps/rejected": -418.0, + "loss": 0.0425, + "rewards/accuracies": 0.9765625, + "rewards/chosen": 0.318359375, + "rewards/margins": 12.25, + "rewards/rejected": -11.9375, + "step": 1474 + }, + { + "epoch": 1.8098159509202454, + "grad_norm": 13.951944722139784, + "learning_rate": 6.012396694214876e-07, + "logits/chosen": -0.154296875, + "logits/rejected": -0.349609375, + "logps/chosen": -402.0, + "logps/rejected": -398.0, + "loss": 0.0357, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.1357421875, + "rewards/margins": 12.0625, + "rewards/rejected": -11.9375, + "step": 1475 + }, + { + "epoch": 1.811042944785276, + "grad_norm": 12.939142286970506, + "learning_rate": 6.006198347107438e-07, + "logits/chosen": -0.181640625, + "logits/rejected": -0.3359375, + "logps/chosen": -368.0, + "logps/rejected": -392.0, + "loss": 0.0343, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.032470703125, + "rewards/margins": 12.5, + "rewards/rejected": -12.4375, + "step": 1476 + }, + { + "epoch": 1.8122699386503067, + "grad_norm": 15.184402213006727, + "learning_rate": 6.000000000000001e-07, + "logits/chosen": -0.057861328125, + "logits/rejected": -0.185546875, + "logps/chosen": -390.0, + "logps/rejected": -404.0, + "loss": 0.0564, + "rewards/accuracies": 0.9765625, + "rewards/chosen": -0.54296875, + "rewards/margins": 11.125, + "rewards/rejected": -11.6875, + "step": 1477 + }, + { + "epoch": 1.8134969325153374, + "grad_norm": 7.953555230394229, + "learning_rate": 5.993801652892562e-07, + "logits/chosen": -0.09619140625, + "logits/rejected": -0.267578125, + "logps/chosen": -402.0, + "logps/rejected": -424.0, + "loss": 0.027, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.8828125, + "rewards/margins": 12.375, + "rewards/rejected": -11.5, + "step": 1478 + }, + { + "epoch": 1.814723926380368, + "grad_norm": 11.656764281029574, + "learning_rate": 5.987603305785124e-07, + "logits/chosen": -0.1884765625, + "logits/rejected": -0.359375, + "logps/chosen": -384.0, + "logps/rejected": -396.0, + "loss": 0.0344, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.25, + "rewards/margins": 12.1875, + "rewards/rejected": -11.9375, + "step": 1479 + }, + { + "epoch": 1.8159509202453987, + "grad_norm": 10.973049014636235, + "learning_rate": 5.981404958677686e-07, + "logits/chosen": -0.267578125, + "logits/rejected": -0.408203125, + "logps/chosen": -396.0, + "logps/rejected": -424.0, + "loss": 0.0437, + "rewards/accuracies": 0.9765625, + "rewards/chosen": 0.72265625, + "rewards/margins": 11.375, + "rewards/rejected": -10.6875, + "step": 1480 + }, + { + "epoch": 1.8171779141104294, + "grad_norm": 16.33917392069797, + "learning_rate": 5.975206611570249e-07, + "logits/chosen": -0.2197265625, + "logits/rejected": -0.333984375, + "logps/chosen": -434.0, + "logps/rejected": -434.0, + "loss": 0.0484, + "rewards/accuracies": 0.9765625, + "rewards/chosen": -0.1962890625, + "rewards/margins": 11.9375, + "rewards/rejected": -12.125, + "step": 1481 + }, + { + "epoch": 1.8184049079754603, + "grad_norm": 13.067184042335276, + "learning_rate": 5.96900826446281e-07, + "logits/chosen": -0.046142578125, + "logits/rejected": -0.203125, + "logps/chosen": -374.0, + "logps/rejected": -376.0, + "loss": 0.0686, + "rewards/accuracies": 0.9765625, + "rewards/chosen": 0.08984375, + "rewards/margins": 11.375, + "rewards/rejected": -11.25, + "step": 1482 + }, + { + "epoch": 1.819631901840491, + "grad_norm": 14.380040955958671, + "learning_rate": 5.962809917355371e-07, + "logits/chosen": 0.0086669921875, + "logits/rejected": -0.146484375, + "logps/chosen": -356.0, + "logps/rejected": -342.0, + "loss": 0.0636, + "rewards/accuracies": 0.96875, + "rewards/chosen": -0.322265625, + "rewards/margins": 11.0, + "rewards/rejected": -11.3125, + "step": 1483 + }, + { + "epoch": 1.8208588957055216, + "grad_norm": 13.600405309118111, + "learning_rate": 5.956611570247934e-07, + "logits/chosen": -0.2158203125, + "logits/rejected": -0.373046875, + "logps/chosen": -354.0, + "logps/rejected": -374.0, + "loss": 0.0529, + "rewards/accuracies": 0.984375, + "rewards/chosen": -0.5078125, + "rewards/margins": 10.9375, + "rewards/rejected": -11.4375, + "step": 1484 + }, + { + "epoch": 1.8220858895705523, + "grad_norm": 19.36891506988035, + "learning_rate": 5.950413223140496e-07, + "logits/chosen": -0.2021484375, + "logits/rejected": -0.23046875, + "logps/chosen": -350.0, + "logps/rejected": -424.0, + "loss": 0.0361, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.5859375, + "rewards/margins": 11.0625, + "rewards/rejected": -11.6875, + "step": 1485 + }, + { + "epoch": 1.823312883435583, + "grad_norm": 13.246751492795608, + "learning_rate": 5.944214876033059e-07, + "logits/chosen": -0.2890625, + "logits/rejected": -0.447265625, + "logps/chosen": -382.0, + "logps/rejected": -402.0, + "loss": 0.0601, + "rewards/accuracies": 0.9765625, + "rewards/chosen": -0.38671875, + "rewards/margins": 11.75, + "rewards/rejected": -12.125, + "step": 1486 + }, + { + "epoch": 1.8245398773006136, + "grad_norm": 3.3945752544087964, + "learning_rate": 5.93801652892562e-07, + "logits/chosen": -0.34375, + "logits/rejected": -0.5078125, + "logps/chosen": -416.0, + "logps/rejected": -388.0, + "loss": 0.013, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.07958984375, + "rewards/margins": 12.5, + "rewards/rejected": -12.4375, + "step": 1487 + }, + { + "epoch": 1.8257668711656443, + "grad_norm": 8.043522667448203, + "learning_rate": 5.931818181818182e-07, + "logits/chosen": -0.201171875, + "logits/rejected": -0.462890625, + "logps/chosen": -416.0, + "logps/rejected": -378.0, + "loss": 0.0227, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.1240234375, + "rewards/margins": 12.1875, + "rewards/rejected": -12.0625, + "step": 1488 + }, + { + "epoch": 1.826993865030675, + "grad_norm": 10.658513612446942, + "learning_rate": 5.925619834710744e-07, + "logits/chosen": -0.333984375, + "logits/rejected": -0.54296875, + "logps/chosen": -410.0, + "logps/rejected": -408.0, + "loss": 0.0417, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.162109375, + "rewards/margins": 12.75, + "rewards/rejected": -12.5625, + "step": 1489 + }, + { + "epoch": 1.8282208588957056, + "grad_norm": 5.907167171601785, + "learning_rate": 5.919421487603307e-07, + "logits/chosen": -0.193359375, + "logits/rejected": -0.416015625, + "logps/chosen": -398.0, + "logps/rejected": -404.0, + "loss": 0.0174, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.71875, + "rewards/margins": 12.875, + "rewards/rejected": -12.125, + "step": 1490 + }, + { + "epoch": 1.8294478527607363, + "grad_norm": 7.708311095918759, + "learning_rate": 5.913223140495868e-07, + "logits/chosen": -0.2109375, + "logits/rejected": -0.466796875, + "logps/chosen": -484.0, + "logps/rejected": -444.0, + "loss": 0.0265, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.9140625, + "rewards/margins": 13.0625, + "rewards/rejected": -12.125, + "step": 1491 + }, + { + "epoch": 1.830674846625767, + "grad_norm": 11.650261479643984, + "learning_rate": 5.907024793388429e-07, + "logits/chosen": -0.349609375, + "logits/rejected": -0.498046875, + "logps/chosen": -390.0, + "logps/rejected": -396.0, + "loss": 0.0365, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.03369140625, + "rewards/margins": 11.75, + "rewards/rejected": -11.8125, + "step": 1492 + }, + { + "epoch": 1.8319018404907976, + "grad_norm": 6.3612166084027235, + "learning_rate": 5.900826446280992e-07, + "logits/chosen": -0.3125, + "logits/rejected": -0.42578125, + "logps/chosen": -330.0, + "logps/rejected": -366.0, + "loss": 0.0247, + "rewards/accuracies": 0.984375, + "rewards/chosen": -0.97265625, + "rewards/margins": 11.375, + "rewards/rejected": -12.375, + "step": 1493 + }, + { + "epoch": 1.8331288343558283, + "grad_norm": 7.617904477738232, + "learning_rate": 5.894628099173554e-07, + "logits/chosen": -0.251953125, + "logits/rejected": -0.38671875, + "logps/chosen": -442.0, + "logps/rejected": -438.0, + "loss": 0.0286, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.1708984375, + "rewards/margins": 12.125, + "rewards/rejected": -12.3125, + "step": 1494 + }, + { + "epoch": 1.834355828220859, + "grad_norm": 14.805530582972839, + "learning_rate": 5.888429752066116e-07, + "logits/chosen": -0.2138671875, + "logits/rejected": -0.369140625, + "logps/chosen": -366.0, + "logps/rejected": -378.0, + "loss": 0.0427, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.051025390625, + "rewards/margins": 11.0625, + "rewards/rejected": -11.0, + "step": 1495 + }, + { + "epoch": 1.8355828220858896, + "grad_norm": 8.603548186700504, + "learning_rate": 5.882231404958677e-07, + "logits/chosen": -0.1533203125, + "logits/rejected": -0.2890625, + "logps/chosen": -374.0, + "logps/rejected": -394.0, + "loss": 0.0246, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.57421875, + "rewards/margins": 11.875, + "rewards/rejected": -11.3125, + "step": 1496 + }, + { + "epoch": 1.8368098159509203, + "grad_norm": 5.119681631341246, + "learning_rate": 5.87603305785124e-07, + "logits/chosen": -0.14453125, + "logits/rejected": -0.375, + "logps/chosen": -414.0, + "logps/rejected": -380.0, + "loss": 0.0172, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.466796875, + "rewards/margins": 11.5625, + "rewards/rejected": -11.0625, + "step": 1497 + }, + { + "epoch": 1.838036809815951, + "grad_norm": 13.256747203318977, + "learning_rate": 5.869834710743802e-07, + "logits/chosen": -0.31640625, + "logits/rejected": -0.49609375, + "logps/chosen": -424.0, + "logps/rejected": -436.0, + "loss": 0.0546, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.328125, + "rewards/margins": 11.625, + "rewards/rejected": -11.3125, + "step": 1498 + }, + { + "epoch": 1.8392638036809816, + "grad_norm": 12.35912732127943, + "learning_rate": 5.863636363636365e-07, + "logits/chosen": -0.1181640625, + "logits/rejected": -0.298828125, + "logps/chosen": -388.0, + "logps/rejected": -396.0, + "loss": 0.0348, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.58984375, + "rewards/margins": 11.9375, + "rewards/rejected": -11.375, + "step": 1499 + }, + { + "epoch": 1.8404907975460123, + "grad_norm": 7.9066533942342945, + "learning_rate": 5.857438016528925e-07, + "logits/chosen": -0.1474609375, + "logits/rejected": -0.361328125, + "logps/chosen": -370.0, + "logps/rejected": -348.0, + "loss": 0.034, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 1.015625, + "rewards/margins": 11.75, + "rewards/rejected": -10.75, + "step": 1500 + }, + { + "epoch": 1.841717791411043, + "grad_norm": 7.790234578512824, + "learning_rate": 5.851239669421487e-07, + "logits/chosen": -0.14453125, + "logits/rejected": -0.27734375, + "logps/chosen": -374.0, + "logps/rejected": -398.0, + "loss": 0.0245, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.61328125, + "rewards/margins": 11.8125, + "rewards/rejected": -11.25, + "step": 1501 + }, + { + "epoch": 1.8429447852760736, + "grad_norm": 8.476340764027082, + "learning_rate": 5.84504132231405e-07, + "logits/chosen": -0.087890625, + "logits/rejected": -0.1953125, + "logps/chosen": -366.0, + "logps/rejected": -394.0, + "loss": 0.0284, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.6953125, + "rewards/margins": 10.875, + "rewards/rejected": -11.5625, + "step": 1502 + }, + { + "epoch": 1.8441717791411043, + "grad_norm": 8.931095134479317, + "learning_rate": 5.838842975206612e-07, + "logits/chosen": -0.201171875, + "logits/rejected": -0.412109375, + "logps/chosen": -430.0, + "logps/rejected": -456.0, + "loss": 0.0258, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.4296875, + "rewards/margins": 12.125, + "rewards/rejected": -11.6875, + "step": 1503 + }, + { + "epoch": 1.845398773006135, + "grad_norm": 9.61981018831853, + "learning_rate": 5.832644628099174e-07, + "logits/chosen": -0.259765625, + "logits/rejected": -0.408203125, + "logps/chosen": -388.0, + "logps/rejected": -408.0, + "loss": 0.0334, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.58203125, + "rewards/margins": 11.5625, + "rewards/rejected": -10.9375, + "step": 1504 + }, + { + "epoch": 1.8466257668711656, + "grad_norm": 8.898568846872344, + "learning_rate": 5.826446280991735e-07, + "logits/chosen": -0.23828125, + "logits/rejected": -0.404296875, + "logps/chosen": -416.0, + "logps/rejected": -420.0, + "loss": 0.0376, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.2099609375, + "rewards/margins": 11.625, + "rewards/rejected": -11.4375, + "step": 1505 + }, + { + "epoch": 1.8478527607361963, + "grad_norm": 8.234391016597, + "learning_rate": 5.820247933884298e-07, + "logits/chosen": -0.1787109375, + "logits/rejected": -0.353515625, + "logps/chosen": -358.0, + "logps/rejected": -398.0, + "loss": 0.0295, + "rewards/accuracies": 0.984375, + "rewards/chosen": -0.48046875, + "rewards/margins": 11.125, + "rewards/rejected": -11.625, + "step": 1506 + }, + { + "epoch": 1.849079754601227, + "grad_norm": 12.615811463348589, + "learning_rate": 5.81404958677686e-07, + "logits/chosen": -0.15625, + "logits/rejected": -0.365234375, + "logps/chosen": -400.0, + "logps/rejected": -390.0, + "loss": 0.0434, + "rewards/accuracies": 0.9765625, + "rewards/chosen": 0.609375, + "rewards/margins": 11.1875, + "rewards/rejected": -10.625, + "step": 1507 + }, + { + "epoch": 1.8503067484662576, + "grad_norm": 11.413630239199696, + "learning_rate": 5.807851239669422e-07, + "logits/chosen": -0.059326171875, + "logits/rejected": -0.2353515625, + "logps/chosen": -392.0, + "logps/rejected": -404.0, + "loss": 0.0325, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.55859375, + "rewards/margins": 10.875, + "rewards/rejected": -11.4375, + "step": 1508 + }, + { + "epoch": 1.8515337423312883, + "grad_norm": 7.238823334535515, + "learning_rate": 5.801652892561983e-07, + "logits/chosen": -0.1611328125, + "logits/rejected": -0.361328125, + "logps/chosen": -444.0, + "logps/rejected": -422.0, + "loss": 0.0186, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.453125, + "rewards/margins": 11.3125, + "rewards/rejected": -10.875, + "step": 1509 + }, + { + "epoch": 1.852760736196319, + "grad_norm": 4.424946915077485, + "learning_rate": 5.795454545454545e-07, + "logits/chosen": -0.126953125, + "logits/rejected": -0.375, + "logps/chosen": -432.0, + "logps/rejected": -398.0, + "loss": 0.0131, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.53515625, + "rewards/margins": 12.1875, + "rewards/rejected": -11.625, + "step": 1510 + }, + { + "epoch": 1.8539877300613496, + "grad_norm": 7.053221824008459, + "learning_rate": 5.789256198347108e-07, + "logits/chosen": -0.203125, + "logits/rejected": -0.267578125, + "logps/chosen": -420.0, + "logps/rejected": -456.0, + "loss": 0.0169, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.322265625, + "rewards/margins": 12.5, + "rewards/rejected": -12.1875, + "step": 1511 + }, + { + "epoch": 1.8552147239263803, + "grad_norm": 3.5547744931254903, + "learning_rate": 5.783057851239669e-07, + "logits/chosen": -0.1494140625, + "logits/rejected": -0.373046875, + "logps/chosen": -358.0, + "logps/rejected": -364.0, + "loss": 0.0135, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.9921875, + "rewards/margins": 13.0, + "rewards/rejected": -12.0, + "step": 1512 + }, + { + "epoch": 1.856441717791411, + "grad_norm": 15.688605244420566, + "learning_rate": 5.776859504132232e-07, + "logits/chosen": -0.150390625, + "logits/rejected": -0.376953125, + "logps/chosen": -404.0, + "logps/rejected": -378.0, + "loss": 0.0586, + "rewards/accuracies": 0.9765625, + "rewards/chosen": 0.12451171875, + "rewards/margins": 11.75, + "rewards/rejected": -11.625, + "step": 1513 + }, + { + "epoch": 1.8576687116564417, + "grad_norm": 5.4110839912069, + "learning_rate": 5.770661157024793e-07, + "logits/chosen": -0.1953125, + "logits/rejected": -0.37109375, + "logps/chosen": -412.0, + "logps/rejected": -426.0, + "loss": 0.0181, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.369140625, + "rewards/margins": 13.125, + "rewards/rejected": -12.75, + "step": 1514 + }, + { + "epoch": 1.8588957055214723, + "grad_norm": 10.478869181509733, + "learning_rate": 5.764462809917356e-07, + "logits/chosen": -0.2265625, + "logits/rejected": -0.4453125, + "logps/chosen": -408.0, + "logps/rejected": -390.0, + "loss": 0.0311, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.94140625, + "rewards/margins": 13.0625, + "rewards/rejected": -12.125, + "step": 1515 + }, + { + "epoch": 1.860122699386503, + "grad_norm": 9.184505008497789, + "learning_rate": 5.758264462809918e-07, + "logits/chosen": -0.1806640625, + "logits/rejected": -0.328125, + "logps/chosen": -420.0, + "logps/rejected": -416.0, + "loss": 0.0329, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.8046875, + "rewards/margins": 11.125, + "rewards/rejected": -11.9375, + "step": 1516 + }, + { + "epoch": 1.8613496932515337, + "grad_norm": 4.539913274268795, + "learning_rate": 5.75206611570248e-07, + "logits/chosen": -0.224609375, + "logits/rejected": -0.42578125, + "logps/chosen": -418.0, + "logps/rejected": -398.0, + "loss": 0.0172, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.69921875, + "rewards/margins": 13.0, + "rewards/rejected": -12.3125, + "step": 1517 + }, + { + "epoch": 1.8625766871165643, + "grad_norm": 14.818256983584963, + "learning_rate": 5.745867768595041e-07, + "logits/chosen": -0.2451171875, + "logits/rejected": -0.443359375, + "logps/chosen": -384.0, + "logps/rejected": -372.0, + "loss": 0.0612, + "rewards/accuracies": 0.96875, + "rewards/chosen": -0.3125, + "rewards/margins": 12.3125, + "rewards/rejected": -12.625, + "step": 1518 + }, + { + "epoch": 1.863803680981595, + "grad_norm": 21.719699854423048, + "learning_rate": 5.739669421487603e-07, + "logits/chosen": -0.2314453125, + "logits/rejected": -0.349609375, + "logps/chosen": -374.0, + "logps/rejected": -412.0, + "loss": 0.0598, + "rewards/accuracies": 0.9765625, + "rewards/chosen": -1.1953125, + "rewards/margins": 10.875, + "rewards/rejected": -12.125, + "step": 1519 + }, + { + "epoch": 1.8650306748466257, + "grad_norm": 4.643580726862002, + "learning_rate": 5.733471074380166e-07, + "logits/chosen": -0.2236328125, + "logits/rejected": -0.30078125, + "logps/chosen": -396.0, + "logps/rejected": -404.0, + "loss": 0.0153, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.41796875, + "rewards/margins": 12.0625, + "rewards/rejected": -12.5, + "step": 1520 + }, + { + "epoch": 1.8662576687116563, + "grad_norm": 9.963670178751276, + "learning_rate": 5.727272727272727e-07, + "logits/chosen": -0.1376953125, + "logits/rejected": -0.3359375, + "logps/chosen": -374.0, + "logps/rejected": -372.0, + "loss": 0.0385, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.28125, + "rewards/margins": 11.5625, + "rewards/rejected": -11.8125, + "step": 1521 + }, + { + "epoch": 1.867484662576687, + "grad_norm": 13.279186623430544, + "learning_rate": 5.721074380165289e-07, + "logits/chosen": -0.1640625, + "logits/rejected": -0.35546875, + "logps/chosen": -370.0, + "logps/rejected": -374.0, + "loss": 0.0478, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.0673828125, + "rewards/margins": 11.5625, + "rewards/rejected": -11.5, + "step": 1522 + }, + { + "epoch": 1.8687116564417177, + "grad_norm": 6.578971949059929, + "learning_rate": 5.714876033057851e-07, + "logits/chosen": -0.2412109375, + "logits/rejected": -0.373046875, + "logps/chosen": -376.0, + "logps/rejected": -376.0, + "loss": 0.0198, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.1591796875, + "rewards/margins": 12.0625, + "rewards/rejected": -11.875, + "step": 1523 + }, + { + "epoch": 1.8699386503067483, + "grad_norm": 13.581236746412541, + "learning_rate": 5.708677685950414e-07, + "logits/chosen": -0.083984375, + "logits/rejected": -0.236328125, + "logps/chosen": -348.0, + "logps/rejected": -400.0, + "loss": 0.04, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.75, + "rewards/margins": 11.4375, + "rewards/rejected": -12.1875, + "step": 1524 + }, + { + "epoch": 1.871165644171779, + "grad_norm": 8.911113085416726, + "learning_rate": 5.702479338842975e-07, + "logits/chosen": -0.2099609375, + "logits/rejected": -0.36328125, + "logps/chosen": -352.0, + "logps/rejected": -390.0, + "loss": 0.0247, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.470703125, + "rewards/margins": 12.375, + "rewards/rejected": -11.9375, + "step": 1525 + }, + { + "epoch": 1.8723926380368097, + "grad_norm": 12.467591280037412, + "learning_rate": 5.696280991735538e-07, + "logits/chosen": -0.07470703125, + "logits/rejected": -0.26953125, + "logps/chosen": -410.0, + "logps/rejected": -390.0, + "loss": 0.0599, + "rewards/accuracies": 0.9765625, + "rewards/chosen": 0.671875, + "rewards/margins": 11.6875, + "rewards/rejected": -11.0625, + "step": 1526 + }, + { + "epoch": 1.8736196319018403, + "grad_norm": 7.9523490622772, + "learning_rate": 5.690082644628099e-07, + "logits/chosen": -0.11279296875, + "logits/rejected": -0.3203125, + "logps/chosen": -412.0, + "logps/rejected": -366.0, + "loss": 0.0304, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.48046875, + "rewards/margins": 11.75, + "rewards/rejected": -11.25, + "step": 1527 + }, + { + "epoch": 1.874846625766871, + "grad_norm": 16.90843634388907, + "learning_rate": 5.683884297520661e-07, + "logits/chosen": -0.07470703125, + "logits/rejected": -0.29296875, + "logps/chosen": -398.0, + "logps/rejected": -362.0, + "loss": 0.0687, + "rewards/accuracies": 0.9609375, + "rewards/chosen": 0.443359375, + "rewards/margins": 11.9375, + "rewards/rejected": -11.5, + "step": 1528 + }, + { + "epoch": 1.876073619631902, + "grad_norm": 5.8636813967032415, + "learning_rate": 5.677685950413223e-07, + "logits/chosen": -0.2021484375, + "logits/rejected": -0.447265625, + "logps/chosen": -422.0, + "logps/rejected": -402.0, + "loss": 0.0195, + "rewards/accuracies": 1.0, + "rewards/chosen": 1.359375, + "rewards/margins": 13.8125, + "rewards/rejected": -12.4375, + "step": 1529 + }, + { + "epoch": 1.8773006134969326, + "grad_norm": 10.180871320115374, + "learning_rate": 5.671487603305785e-07, + "logits/chosen": -0.189453125, + "logits/rejected": -0.326171875, + "logps/chosen": -388.0, + "logps/rejected": -368.0, + "loss": 0.0404, + "rewards/accuracies": 0.9765625, + "rewards/chosen": -0.7265625, + "rewards/margins": 11.8125, + "rewards/rejected": -12.5625, + "step": 1530 + }, + { + "epoch": 1.8785276073619632, + "grad_norm": 9.962751714095937, + "learning_rate": 5.665289256198347e-07, + "logits/chosen": -0.042236328125, + "logits/rejected": -0.1953125, + "logps/chosen": -378.0, + "logps/rejected": -374.0, + "loss": 0.0435, + "rewards/accuracies": 0.9765625, + "rewards/chosen": -0.3984375, + "rewards/margins": 11.5, + "rewards/rejected": -11.875, + "step": 1531 + }, + { + "epoch": 1.879754601226994, + "grad_norm": 13.344235509044795, + "learning_rate": 5.659090909090909e-07, + "logits/chosen": -0.328125, + "logits/rejected": -0.4921875, + "logps/chosen": -436.0, + "logps/rejected": -450.0, + "loss": 0.0678, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.6953125, + "rewards/margins": 13.625, + "rewards/rejected": -12.9375, + "step": 1532 + }, + { + "epoch": 1.8809815950920246, + "grad_norm": 12.313294846355545, + "learning_rate": 5.652892561983472e-07, + "logits/chosen": -0.12060546875, + "logits/rejected": -0.26171875, + "logps/chosen": -382.0, + "logps/rejected": -402.0, + "loss": 0.0423, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.057861328125, + "rewards/margins": 11.0625, + "rewards/rejected": -11.0, + "step": 1533 + }, + { + "epoch": 1.8822085889570552, + "grad_norm": 7.89566619418515, + "learning_rate": 5.646694214876033e-07, + "logits/chosen": -0.205078125, + "logits/rejected": -0.337890625, + "logps/chosen": -392.0, + "logps/rejected": -414.0, + "loss": 0.025, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.59375, + "rewards/margins": 12.0, + "rewards/rejected": -12.5625, + "step": 1534 + }, + { + "epoch": 1.883435582822086, + "grad_norm": 11.645948617484482, + "learning_rate": 5.640495867768596e-07, + "logits/chosen": -0.046875, + "logits/rejected": -0.3046875, + "logps/chosen": -374.0, + "logps/rejected": -360.0, + "loss": 0.0563, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.404296875, + "rewards/margins": 11.9375, + "rewards/rejected": -11.5625, + "step": 1535 + }, + { + "epoch": 1.8846625766871166, + "grad_norm": 12.956631655705353, + "learning_rate": 5.634297520661157e-07, + "logits/chosen": -0.056640625, + "logits/rejected": -0.2412109375, + "logps/chosen": -458.0, + "logps/rejected": -440.0, + "loss": 0.0515, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.34375, + "rewards/margins": 11.875, + "rewards/rejected": -11.5625, + "step": 1536 + }, + { + "epoch": 1.8858895705521472, + "grad_norm": 4.634384436264029, + "learning_rate": 5.628099173553719e-07, + "logits/chosen": -0.1123046875, + "logits/rejected": -0.34375, + "logps/chosen": -380.0, + "logps/rejected": -390.0, + "loss": 0.0124, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.037353515625, + "rewards/margins": 12.1875, + "rewards/rejected": -12.125, + "step": 1537 + }, + { + "epoch": 1.887116564417178, + "grad_norm": 5.6435058037855965, + "learning_rate": 5.621900826446281e-07, + "logits/chosen": -0.10498046875, + "logits/rejected": -0.251953125, + "logps/chosen": -366.0, + "logps/rejected": -394.0, + "loss": 0.0103, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.51953125, + "rewards/margins": 12.375, + "rewards/rejected": -11.875, + "step": 1538 + }, + { + "epoch": 1.8883435582822086, + "grad_norm": 8.930748067631301, + "learning_rate": 5.615702479338843e-07, + "logits/chosen": -0.01275634765625, + "logits/rejected": -0.240234375, + "logps/chosen": -380.0, + "logps/rejected": -390.0, + "loss": 0.0294, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.201171875, + "rewards/margins": 11.5625, + "rewards/rejected": -11.375, + "step": 1539 + }, + { + "epoch": 1.8895705521472392, + "grad_norm": 10.323806155698314, + "learning_rate": 5.609504132231405e-07, + "logits/chosen": -0.302734375, + "logits/rejected": -0.4375, + "logps/chosen": -438.0, + "logps/rejected": -468.0, + "loss": 0.0427, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.486328125, + "rewards/margins": 12.625, + "rewards/rejected": -12.125, + "step": 1540 + }, + { + "epoch": 1.89079754601227, + "grad_norm": 11.351287245365558, + "learning_rate": 5.603305785123967e-07, + "logits/chosen": -0.2421875, + "logits/rejected": -0.396484375, + "logps/chosen": -384.0, + "logps/rejected": -406.0, + "loss": 0.0483, + "rewards/accuracies": 0.9765625, + "rewards/chosen": 0.11669921875, + "rewards/margins": 12.0, + "rewards/rejected": -11.875, + "step": 1541 + }, + { + "epoch": 1.8920245398773006, + "grad_norm": 3.2489476684935736, + "learning_rate": 5.597107438016529e-07, + "logits/chosen": -0.2041015625, + "logits/rejected": -0.375, + "logps/chosen": -404.0, + "logps/rejected": -406.0, + "loss": 0.0088, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.80078125, + "rewards/margins": 13.1875, + "rewards/rejected": -12.375, + "step": 1542 + }, + { + "epoch": 1.8932515337423312, + "grad_norm": 8.504627067419271, + "learning_rate": 5.590909090909091e-07, + "logits/chosen": -0.1884765625, + "logits/rejected": -0.357421875, + "logps/chosen": -378.0, + "logps/rejected": -396.0, + "loss": 0.0291, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.158203125, + "rewards/margins": 12.5, + "rewards/rejected": -12.3125, + "step": 1543 + }, + { + "epoch": 1.8944785276073621, + "grad_norm": 8.19609184280983, + "learning_rate": 5.584710743801654e-07, + "logits/chosen": -0.2451171875, + "logits/rejected": -0.408203125, + "logps/chosen": -422.0, + "logps/rejected": -450.0, + "loss": 0.0282, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.314453125, + "rewards/margins": 12.625, + "rewards/rejected": -12.3125, + "step": 1544 + }, + { + "epoch": 1.8957055214723928, + "grad_norm": 8.691100794961908, + "learning_rate": 5.578512396694215e-07, + "logits/chosen": -0.12451171875, + "logits/rejected": -0.326171875, + "logps/chosen": -392.0, + "logps/rejected": -358.0, + "loss": 0.0275, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.294921875, + "rewards/margins": 10.75, + "rewards/rejected": -11.0625, + "step": 1545 + }, + { + "epoch": 1.8969325153374235, + "grad_norm": 8.614624685931496, + "learning_rate": 5.572314049586777e-07, + "logits/chosen": -0.1640625, + "logits/rejected": -0.3359375, + "logps/chosen": -376.0, + "logps/rejected": -372.0, + "loss": 0.024, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.1435546875, + "rewards/margins": 11.5, + "rewards/rejected": -11.375, + "step": 1546 + }, + { + "epoch": 1.8981595092024541, + "grad_norm": 15.386153310990915, + "learning_rate": 5.566115702479339e-07, + "logits/chosen": -0.1494140625, + "logits/rejected": -0.484375, + "logps/chosen": -490.0, + "logps/rejected": -418.0, + "loss": 0.0277, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.361328125, + "rewards/margins": 13.1875, + "rewards/rejected": -12.8125, + "step": 1547 + }, + { + "epoch": 1.8993865030674848, + "grad_norm": 21.182540544956645, + "learning_rate": 5.559917355371901e-07, + "logits/chosen": -0.1572265625, + "logits/rejected": -0.2001953125, + "logps/chosen": -400.0, + "logps/rejected": -430.0, + "loss": 0.0662, + "rewards/accuracies": 0.984375, + "rewards/chosen": -0.11376953125, + "rewards/margins": 11.9375, + "rewards/rejected": -12.0625, + "step": 1548 + }, + { + "epoch": 1.9006134969325155, + "grad_norm": 9.056705219686792, + "learning_rate": 5.553719008264463e-07, + "logits/chosen": -0.046875, + "logits/rejected": -0.26953125, + "logps/chosen": -388.0, + "logps/rejected": -374.0, + "loss": 0.032, + "rewards/accuracies": 0.9765625, + "rewards/chosen": -0.0185546875, + "rewards/margins": 11.4375, + "rewards/rejected": -11.4375, + "step": 1549 + }, + { + "epoch": 1.9018404907975461, + "grad_norm": 12.174899827400916, + "learning_rate": 5.547520661157025e-07, + "logits/chosen": -0.1923828125, + "logits/rejected": -0.2890625, + "logps/chosen": -394.0, + "logps/rejected": -396.0, + "loss": 0.0361, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.0478515625, + "rewards/margins": 12.6875, + "rewards/rejected": -12.6875, + "step": 1550 + }, + { + "epoch": 1.9030674846625768, + "grad_norm": 11.776464119336328, + "learning_rate": 5.541322314049587e-07, + "logits/chosen": -0.2177734375, + "logits/rejected": -0.42578125, + "logps/chosen": -434.0, + "logps/rejected": -416.0, + "loss": 0.0399, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.466796875, + "rewards/margins": 11.375, + "rewards/rejected": -11.8125, + "step": 1551 + }, + { + "epoch": 1.9042944785276075, + "grad_norm": 15.301018527355753, + "learning_rate": 5.535123966942149e-07, + "logits/chosen": -0.1416015625, + "logits/rejected": -0.26953125, + "logps/chosen": -358.0, + "logps/rejected": -362.0, + "loss": 0.071, + "rewards/accuracies": 0.96875, + "rewards/chosen": -0.19140625, + "rewards/margins": 10.375, + "rewards/rejected": -10.5625, + "step": 1552 + }, + { + "epoch": 1.9055214723926381, + "grad_norm": 13.713477834316057, + "learning_rate": 5.528925619834712e-07, + "logits/chosen": -0.078125, + "logits/rejected": -0.2041015625, + "logps/chosen": -392.0, + "logps/rejected": -356.0, + "loss": 0.0288, + "rewards/accuracies": 0.984375, + "rewards/chosen": -0.015625, + "rewards/margins": 11.5, + "rewards/rejected": -11.5, + "step": 1553 + }, + { + "epoch": 1.9067484662576688, + "grad_norm": 13.970935293353676, + "learning_rate": 5.522727272727273e-07, + "logits/chosen": -0.29296875, + "logits/rejected": -0.55078125, + "logps/chosen": -426.0, + "logps/rejected": -406.0, + "loss": 0.0417, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.05810546875, + "rewards/margins": 12.1875, + "rewards/rejected": -12.125, + "step": 1554 + }, + { + "epoch": 1.9079754601226995, + "grad_norm": 17.08022297387505, + "learning_rate": 5.516528925619834e-07, + "logits/chosen": -0.306640625, + "logits/rejected": -0.318359375, + "logps/chosen": -388.0, + "logps/rejected": -414.0, + "loss": 0.0317, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.18359375, + "rewards/margins": 11.875, + "rewards/rejected": -12.0625, + "step": 1555 + }, + { + "epoch": 1.9092024539877301, + "grad_norm": 5.512042362005046, + "learning_rate": 5.510330578512397e-07, + "logits/chosen": -0.099609375, + "logits/rejected": -0.3359375, + "logps/chosen": -410.0, + "logps/rejected": -392.0, + "loss": 0.0214, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.23828125, + "rewards/margins": 11.5625, + "rewards/rejected": -11.375, + "step": 1556 + }, + { + "epoch": 1.9104294478527608, + "grad_norm": 9.862650604155213, + "learning_rate": 5.504132231404959e-07, + "logits/chosen": -0.01397705078125, + "logits/rejected": -0.263671875, + "logps/chosen": -404.0, + "logps/rejected": -414.0, + "loss": 0.0364, + "rewards/accuracies": 0.9765625, + "rewards/chosen": 0.041748046875, + "rewards/margins": 11.25, + "rewards/rejected": -11.1875, + "step": 1557 + }, + { + "epoch": 1.9116564417177915, + "grad_norm": 9.10335392738119, + "learning_rate": 5.497933884297521e-07, + "logits/chosen": -0.173828125, + "logits/rejected": -0.337890625, + "logps/chosen": -394.0, + "logps/rejected": -432.0, + "loss": 0.0229, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.296875, + "rewards/margins": 12.375, + "rewards/rejected": -12.0625, + "step": 1558 + }, + { + "epoch": 1.9128834355828221, + "grad_norm": 10.053624701482729, + "learning_rate": 5.491735537190082e-07, + "logits/chosen": -0.032958984375, + "logits/rejected": -0.1162109375, + "logps/chosen": -396.0, + "logps/rejected": -354.0, + "loss": 0.0385, + "rewards/accuracies": 0.984375, + "rewards/chosen": -0.1142578125, + "rewards/margins": 10.9375, + "rewards/rejected": -11.0625, + "step": 1559 + }, + { + "epoch": 1.9141104294478528, + "grad_norm": 6.811810662950314, + "learning_rate": 5.485537190082645e-07, + "logits/chosen": -0.09228515625, + "logits/rejected": -0.287109375, + "logps/chosen": -376.0, + "logps/rejected": -370.0, + "loss": 0.0255, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.91796875, + "rewards/margins": 11.1875, + "rewards/rejected": -12.125, + "step": 1560 + }, + { + "epoch": 1.9153374233128835, + "grad_norm": 10.499380910825229, + "learning_rate": 5.479338842975207e-07, + "logits/chosen": -0.1923828125, + "logits/rejected": -0.470703125, + "logps/chosen": -456.0, + "logps/rejected": -410.0, + "loss": 0.0363, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.734375, + "rewards/margins": 12.9375, + "rewards/rejected": -12.1875, + "step": 1561 + }, + { + "epoch": 1.9165644171779141, + "grad_norm": 12.232131041485115, + "learning_rate": 5.473140495867769e-07, + "logits/chosen": -0.21875, + "logits/rejected": -0.4296875, + "logps/chosen": -438.0, + "logps/rejected": -414.0, + "loss": 0.03, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 1.34375, + "rewards/margins": 12.625, + "rewards/rejected": -11.25, + "step": 1562 + }, + { + "epoch": 1.9177914110429448, + "grad_norm": 14.943326703058906, + "learning_rate": 5.466942148760331e-07, + "logits/chosen": -0.173828125, + "logits/rejected": -0.408203125, + "logps/chosen": -404.0, + "logps/rejected": -404.0, + "loss": 0.0469, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.7421875, + "rewards/margins": 12.5, + "rewards/rejected": -11.75, + "step": 1563 + }, + { + "epoch": 1.9190184049079755, + "grad_norm": 2.974048087039378, + "learning_rate": 5.460743801652892e-07, + "logits/chosen": -0.1650390625, + "logits/rejected": -0.361328125, + "logps/chosen": -388.0, + "logps/rejected": -402.0, + "loss": 0.0101, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.10302734375, + "rewards/margins": 11.1875, + "rewards/rejected": -11.25, + "step": 1564 + }, + { + "epoch": 1.9202453987730062, + "grad_norm": 6.834929386098197, + "learning_rate": 5.454545454545455e-07, + "logits/chosen": -0.298828125, + "logits/rejected": -0.490234375, + "logps/chosen": -322.0, + "logps/rejected": -342.0, + "loss": 0.0246, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.474609375, + "rewards/margins": 12.0625, + "rewards/rejected": -11.5625, + "step": 1565 + }, + { + "epoch": 1.9214723926380368, + "grad_norm": 19.709537367801847, + "learning_rate": 5.448347107438017e-07, + "logits/chosen": -0.1279296875, + "logits/rejected": -0.267578125, + "logps/chosen": -406.0, + "logps/rejected": -422.0, + "loss": 0.0777, + "rewards/accuracies": 0.9609375, + "rewards/chosen": 0.890625, + "rewards/margins": 12.3125, + "rewards/rejected": -11.4375, + "step": 1566 + }, + { + "epoch": 1.9226993865030675, + "grad_norm": 10.759644336799223, + "learning_rate": 5.442148760330579e-07, + "logits/chosen": -0.15234375, + "logits/rejected": -0.302734375, + "logps/chosen": -376.0, + "logps/rejected": -388.0, + "loss": 0.0364, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.220703125, + "rewards/margins": 12.0, + "rewards/rejected": -11.75, + "step": 1567 + }, + { + "epoch": 1.9239263803680982, + "grad_norm": 20.138956412255148, + "learning_rate": 5.43595041322314e-07, + "logits/chosen": -0.10107421875, + "logits/rejected": -0.328125, + "logps/chosen": -392.0, + "logps/rejected": -404.0, + "loss": 0.0545, + "rewards/accuracies": 0.96875, + "rewards/chosen": 1.2890625, + "rewards/margins": 12.75, + "rewards/rejected": -11.4375, + "step": 1568 + }, + { + "epoch": 1.9251533742331288, + "grad_norm": 4.621297903485169, + "learning_rate": 5.429752066115703e-07, + "logits/chosen": 0.022705078125, + "logits/rejected": -0.1220703125, + "logps/chosen": -390.0, + "logps/rejected": -370.0, + "loss": 0.0098, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.396484375, + "rewards/margins": 12.1875, + "rewards/rejected": -11.8125, + "step": 1569 + }, + { + "epoch": 1.9263803680981595, + "grad_norm": 12.208310607847823, + "learning_rate": 5.423553719008265e-07, + "logits/chosen": -0.109375, + "logits/rejected": -0.33984375, + "logps/chosen": -384.0, + "logps/rejected": -388.0, + "loss": 0.0424, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.12158203125, + "rewards/margins": 11.1875, + "rewards/rejected": -11.3125, + "step": 1570 + }, + { + "epoch": 1.9276073619631902, + "grad_norm": 13.248924763237904, + "learning_rate": 5.417355371900827e-07, + "logits/chosen": -0.1416015625, + "logits/rejected": -0.294921875, + "logps/chosen": -426.0, + "logps/rejected": -430.0, + "loss": 0.0522, + "rewards/accuracies": 0.9609375, + "rewards/chosen": 0.76171875, + "rewards/margins": 12.3125, + "rewards/rejected": -11.5, + "step": 1571 + }, + { + "epoch": 1.9288343558282208, + "grad_norm": 7.855413156125594, + "learning_rate": 5.411157024793388e-07, + "logits/chosen": -0.185546875, + "logits/rejected": -0.392578125, + "logps/chosen": -414.0, + "logps/rejected": -406.0, + "loss": 0.0196, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.1494140625, + "rewards/margins": 12.6875, + "rewards/rejected": -12.5625, + "step": 1572 + }, + { + "epoch": 1.9300613496932515, + "grad_norm": 11.846684123578562, + "learning_rate": 5.40495867768595e-07, + "logits/chosen": -0.17578125, + "logits/rejected": -0.3046875, + "logps/chosen": -402.0, + "logps/rejected": -396.0, + "loss": 0.0502, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.279296875, + "rewards/margins": 11.375, + "rewards/rejected": -11.625, + "step": 1573 + }, + { + "epoch": 1.9312883435582822, + "grad_norm": 15.381676115792011, + "learning_rate": 5.398760330578513e-07, + "logits/chosen": -0.02880859375, + "logits/rejected": -0.162109375, + "logps/chosen": -366.0, + "logps/rejected": -404.0, + "loss": 0.0444, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.078125, + "rewards/margins": 11.3125, + "rewards/rejected": -11.1875, + "step": 1574 + }, + { + "epoch": 1.9325153374233128, + "grad_norm": 16.45905252471069, + "learning_rate": 5.392561983471075e-07, + "logits/chosen": -0.1669921875, + "logits/rejected": -0.30078125, + "logps/chosen": -398.0, + "logps/rejected": -384.0, + "loss": 0.0753, + "rewards/accuracies": 0.9765625, + "rewards/chosen": 0.57421875, + "rewards/margins": 12.5, + "rewards/rejected": -11.9375, + "step": 1575 + }, + { + "epoch": 1.9337423312883435, + "grad_norm": 5.91372888725709, + "learning_rate": 5.386363636363636e-07, + "logits/chosen": -0.1630859375, + "logits/rejected": -0.37890625, + "logps/chosen": -378.0, + "logps/rejected": -384.0, + "loss": 0.0169, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.69921875, + "rewards/margins": 11.5625, + "rewards/rejected": -12.3125, + "step": 1576 + }, + { + "epoch": 1.9349693251533742, + "grad_norm": 13.068737835207726, + "learning_rate": 5.380165289256198e-07, + "logits/chosen": -0.10693359375, + "logits/rejected": -0.294921875, + "logps/chosen": -388.0, + "logps/rejected": -346.0, + "loss": 0.0301, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.82421875, + "rewards/margins": 12.4375, + "rewards/rejected": -11.625, + "step": 1577 + }, + { + "epoch": 1.9361963190184048, + "grad_norm": 12.574938700274403, + "learning_rate": 5.373966942148761e-07, + "logits/chosen": -0.12255859375, + "logits/rejected": -0.296875, + "logps/chosen": -382.0, + "logps/rejected": -420.0, + "loss": 0.0401, + "rewards/accuracies": 0.984375, + "rewards/chosen": -0.0927734375, + "rewards/margins": 12.125, + "rewards/rejected": -12.1875, + "step": 1578 + }, + { + "epoch": 1.9374233128834355, + "grad_norm": 6.1456398779932035, + "learning_rate": 5.367768595041323e-07, + "logits/chosen": -0.1494140625, + "logits/rejected": -0.255859375, + "logps/chosen": -384.0, + "logps/rejected": -386.0, + "loss": 0.0222, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.19140625, + "rewards/margins": 12.1875, + "rewards/rejected": -12.0, + "step": 1579 + }, + { + "epoch": 1.9386503067484662, + "grad_norm": 7.144532600637711, + "learning_rate": 5.361570247933885e-07, + "logits/chosen": -0.15625, + "logits/rejected": -0.376953125, + "logps/chosen": -444.0, + "logps/rejected": -416.0, + "loss": 0.0192, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.78515625, + "rewards/margins": 12.4375, + "rewards/rejected": -11.6875, + "step": 1580 + }, + { + "epoch": 1.9398773006134968, + "grad_norm": 5.065784309723247, + "learning_rate": 5.355371900826446e-07, + "logits/chosen": -0.09765625, + "logits/rejected": -0.2470703125, + "logps/chosen": -416.0, + "logps/rejected": -408.0, + "loss": 0.0165, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.6015625, + "rewards/margins": 12.375, + "rewards/rejected": -11.75, + "step": 1581 + }, + { + "epoch": 1.9411042944785275, + "grad_norm": 8.432710268891027, + "learning_rate": 5.349173553719008e-07, + "logits/chosen": -0.0242919921875, + "logits/rejected": -0.142578125, + "logps/chosen": -332.0, + "logps/rejected": -352.0, + "loss": 0.0494, + "rewards/accuracies": 0.9765625, + "rewards/chosen": -0.31640625, + "rewards/margins": 10.6875, + "rewards/rejected": -11.0, + "step": 1582 + }, + { + "epoch": 1.9423312883435582, + "grad_norm": 14.414561246083233, + "learning_rate": 5.342975206611571e-07, + "logits/chosen": -0.12255859375, + "logits/rejected": -0.30078125, + "logps/chosen": -418.0, + "logps/rejected": -400.0, + "loss": 0.0637, + "rewards/accuracies": 0.9765625, + "rewards/chosen": 0.7734375, + "rewards/margins": 12.1875, + "rewards/rejected": -11.4375, + "step": 1583 + }, + { + "epoch": 1.9435582822085888, + "grad_norm": 13.517917688736201, + "learning_rate": 5.336776859504132e-07, + "logits/chosen": -0.08740234375, + "logits/rejected": -0.271484375, + "logps/chosen": -436.0, + "logps/rejected": -394.0, + "loss": 0.064, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.37890625, + "rewards/margins": 11.0625, + "rewards/rejected": -10.6875, + "step": 1584 + }, + { + "epoch": 1.9447852760736195, + "grad_norm": 18.795243167760674, + "learning_rate": 5.330578512396694e-07, + "logits/chosen": -0.181640625, + "logits/rejected": -0.326171875, + "logps/chosen": -418.0, + "logps/rejected": -434.0, + "loss": 0.073, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.82421875, + "rewards/margins": 13.1875, + "rewards/rejected": -12.375, + "step": 1585 + }, + { + "epoch": 1.9460122699386502, + "grad_norm": 9.671395753436496, + "learning_rate": 5.324380165289256e-07, + "logits/chosen": -0.203125, + "logits/rejected": -0.37890625, + "logps/chosen": -404.0, + "logps/rejected": -386.0, + "loss": 0.0373, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.169921875, + "rewards/margins": 11.125, + "rewards/rejected": -10.9375, + "step": 1586 + }, + { + "epoch": 1.9472392638036808, + "grad_norm": 10.924001936653777, + "learning_rate": 5.318181818181819e-07, + "logits/chosen": -0.1669921875, + "logits/rejected": -0.34765625, + "logps/chosen": -358.0, + "logps/rejected": -366.0, + "loss": 0.0462, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.13671875, + "rewards/margins": 11.1875, + "rewards/rejected": -11.3125, + "step": 1587 + }, + { + "epoch": 1.9484662576687115, + "grad_norm": 10.287013784424305, + "learning_rate": 5.31198347107438e-07, + "logits/chosen": -0.0869140625, + "logits/rejected": -0.251953125, + "logps/chosen": -374.0, + "logps/rejected": -378.0, + "loss": 0.0323, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.35546875, + "rewards/margins": 11.375, + "rewards/rejected": -11.0, + "step": 1588 + }, + { + "epoch": 1.9496932515337422, + "grad_norm": 11.345438684638955, + "learning_rate": 5.305785123966941e-07, + "logits/chosen": -0.1748046875, + "logits/rejected": -0.33984375, + "logps/chosen": -360.0, + "logps/rejected": -388.0, + "loss": 0.028, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.26953125, + "rewards/margins": 12.5, + "rewards/rejected": -12.1875, + "step": 1589 + }, + { + "epoch": 1.9509202453987728, + "grad_norm": 12.380473368422328, + "learning_rate": 5.299586776859504e-07, + "logits/chosen": -0.091796875, + "logits/rejected": -0.232421875, + "logps/chosen": -398.0, + "logps/rejected": -420.0, + "loss": 0.0306, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.4765625, + "rewards/margins": 12.375, + "rewards/rejected": -11.875, + "step": 1590 + }, + { + "epoch": 1.9521472392638037, + "grad_norm": 6.3846910410496855, + "learning_rate": 5.293388429752066e-07, + "logits/chosen": -0.1796875, + "logits/rejected": -0.328125, + "logps/chosen": -364.0, + "logps/rejected": -368.0, + "loss": 0.0163, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.02490234375, + "rewards/margins": 11.875, + "rewards/rejected": -11.875, + "step": 1591 + }, + { + "epoch": 1.9533742331288344, + "grad_norm": 12.161899209714043, + "learning_rate": 5.287190082644629e-07, + "logits/chosen": -0.07666015625, + "logits/rejected": -0.2158203125, + "logps/chosen": -378.0, + "logps/rejected": -402.0, + "loss": 0.0384, + "rewards/accuracies": 0.9765625, + "rewards/chosen": -0.6015625, + "rewards/margins": 11.5, + "rewards/rejected": -12.0625, + "step": 1592 + }, + { + "epoch": 1.954601226993865, + "grad_norm": 16.096985565614556, + "learning_rate": 5.28099173553719e-07, + "logits/chosen": -0.1083984375, + "logits/rejected": -0.263671875, + "logps/chosen": -406.0, + "logps/rejected": -396.0, + "loss": 0.0467, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.478515625, + "rewards/margins": 11.75, + "rewards/rejected": -12.25, + "step": 1593 + }, + { + "epoch": 1.9558282208588957, + "grad_norm": 8.516823030452967, + "learning_rate": 5.274793388429752e-07, + "logits/chosen": -0.2001953125, + "logits/rejected": -0.34765625, + "logps/chosen": -332.0, + "logps/rejected": -344.0, + "loss": 0.0568, + "rewards/accuracies": 0.984375, + "rewards/chosen": -0.1103515625, + "rewards/margins": 11.3125, + "rewards/rejected": -11.4375, + "step": 1594 + }, + { + "epoch": 1.9570552147239264, + "grad_norm": 20.676589588590403, + "learning_rate": 5.268595041322314e-07, + "logits/chosen": -0.146484375, + "logits/rejected": -0.21875, + "logps/chosen": -382.0, + "logps/rejected": -404.0, + "loss": 0.0608, + "rewards/accuracies": 0.984375, + "rewards/chosen": -0.953125, + "rewards/margins": 11.8125, + "rewards/rejected": -12.8125, + "step": 1595 + }, + { + "epoch": 1.958282208588957, + "grad_norm": 11.402306615191995, + "learning_rate": 5.262396694214877e-07, + "logits/chosen": -0.2119140625, + "logits/rejected": -0.25390625, + "logps/chosen": -362.0, + "logps/rejected": -370.0, + "loss": 0.0339, + "rewards/accuracies": 0.9765625, + "rewards/chosen": -0.5234375, + "rewards/margins": 11.8125, + "rewards/rejected": -12.375, + "step": 1596 + }, + { + "epoch": 1.9595092024539877, + "grad_norm": 20.832383714357192, + "learning_rate": 5.256198347107438e-07, + "logits/chosen": -0.12353515625, + "logits/rejected": -0.330078125, + "logps/chosen": -402.0, + "logps/rejected": -394.0, + "loss": 0.0374, + "rewards/accuracies": 0.96875, + "rewards/chosen": -0.25390625, + "rewards/margins": 12.5, + "rewards/rejected": -12.75, + "step": 1597 + }, + { + "epoch": 1.9607361963190184, + "grad_norm": 17.67799399509982, + "learning_rate": 5.25e-07, + "logits/chosen": -0.2216796875, + "logits/rejected": -0.30078125, + "logps/chosen": -396.0, + "logps/rejected": -432.0, + "loss": 0.0746, + "rewards/accuracies": 0.9765625, + "rewards/chosen": -1.0078125, + "rewards/margins": 11.625, + "rewards/rejected": -12.6875, + "step": 1598 + }, + { + "epoch": 1.961963190184049, + "grad_norm": 2.4424485737982744, + "learning_rate": 5.243801652892562e-07, + "logits/chosen": -0.123046875, + "logits/rejected": -0.36328125, + "logps/chosen": -402.0, + "logps/rejected": -388.0, + "loss": 0.011, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.27734375, + "rewards/margins": 12.0, + "rewards/rejected": -12.25, + "step": 1599 + }, + { + "epoch": 1.9631901840490797, + "grad_norm": 13.907963965939771, + "learning_rate": 5.237603305785124e-07, + "logits/chosen": -0.06884765625, + "logits/rejected": -0.2392578125, + "logps/chosen": -416.0, + "logps/rejected": -410.0, + "loss": 0.0418, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.5078125, + "rewards/margins": 11.625, + "rewards/rejected": -12.125, + "step": 1600 + }, + { + "epoch": 1.9644171779141104, + "grad_norm": 13.814689419274158, + "learning_rate": 5.231404958677686e-07, + "logits/chosen": -0.033447265625, + "logits/rejected": -0.1435546875, + "logps/chosen": -346.0, + "logps/rejected": -372.0, + "loss": 0.0472, + "rewards/accuracies": 0.9765625, + "rewards/chosen": -0.98046875, + "rewards/margins": 10.6875, + "rewards/rejected": -11.6875, + "step": 1601 + }, + { + "epoch": 1.965644171779141, + "grad_norm": 11.545621891190732, + "learning_rate": 5.225206611570248e-07, + "logits/chosen": -0.11962890625, + "logits/rejected": -0.326171875, + "logps/chosen": -442.0, + "logps/rejected": -380.0, + "loss": 0.0297, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.2578125, + "rewards/margins": 12.1875, + "rewards/rejected": -11.9375, + "step": 1602 + }, + { + "epoch": 1.9668711656441717, + "grad_norm": 8.879321275799807, + "learning_rate": 5.21900826446281e-07, + "logits/chosen": -0.146484375, + "logits/rejected": -0.341796875, + "logps/chosen": -472.0, + "logps/rejected": -436.0, + "loss": 0.0337, + "rewards/accuracies": 0.984375, + "rewards/chosen": -0.373046875, + "rewards/margins": 13.0, + "rewards/rejected": -13.3125, + "step": 1603 + }, + { + "epoch": 1.9680981595092024, + "grad_norm": 10.570209103673362, + "learning_rate": 5.212809917355372e-07, + "logits/chosen": -0.080078125, + "logits/rejected": -0.2734375, + "logps/chosen": -394.0, + "logps/rejected": -392.0, + "loss": 0.0386, + "rewards/accuracies": 0.984375, + "rewards/chosen": -0.07861328125, + "rewards/margins": 11.1875, + "rewards/rejected": -11.3125, + "step": 1604 + }, + { + "epoch": 1.969325153374233, + "grad_norm": 11.659167158794315, + "learning_rate": 5.206611570247935e-07, + "logits/chosen": -0.08837890625, + "logits/rejected": -0.232421875, + "logps/chosen": -434.0, + "logps/rejected": -436.0, + "loss": 0.0439, + "rewards/accuracies": 0.984375, + "rewards/chosen": -0.875, + "rewards/margins": 11.625, + "rewards/rejected": -12.5, + "step": 1605 + }, + { + "epoch": 1.9705521472392638, + "grad_norm": 4.504606116838098, + "learning_rate": 5.200413223140496e-07, + "logits/chosen": -0.2138671875, + "logits/rejected": -0.427734375, + "logps/chosen": -418.0, + "logps/rejected": -434.0, + "loss": 0.0155, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.2470703125, + "rewards/margins": 12.5625, + "rewards/rejected": -12.3125, + "step": 1606 + }, + { + "epoch": 1.9717791411042946, + "grad_norm": 3.9125260909977104, + "learning_rate": 5.194214876033058e-07, + "logits/chosen": -0.16015625, + "logits/rejected": -0.396484375, + "logps/chosen": -384.0, + "logps/rejected": -370.0, + "loss": 0.0139, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.2578125, + "rewards/margins": 12.75, + "rewards/rejected": -12.4375, + "step": 1607 + }, + { + "epoch": 1.9730061349693253, + "grad_norm": 12.198720753980819, + "learning_rate": 5.18801652892562e-07, + "logits/chosen": -0.166015625, + "logits/rejected": -0.369140625, + "logps/chosen": -392.0, + "logps/rejected": -414.0, + "loss": 0.0264, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.2236328125, + "rewards/margins": 12.75, + "rewards/rejected": -12.5625, + "step": 1608 + }, + { + "epoch": 1.974233128834356, + "grad_norm": 13.187714303157518, + "learning_rate": 5.181818181818182e-07, + "logits/chosen": -0.1806640625, + "logits/rejected": -0.3125, + "logps/chosen": -378.0, + "logps/rejected": -398.0, + "loss": 0.0497, + "rewards/accuracies": 0.984375, + "rewards/chosen": -0.57421875, + "rewards/margins": 12.0625, + "rewards/rejected": -12.625, + "step": 1609 + }, + { + "epoch": 1.9754601226993866, + "grad_norm": 13.077595266689482, + "learning_rate": 5.175619834710744e-07, + "logits/chosen": -0.22265625, + "logits/rejected": -0.3125, + "logps/chosen": -366.0, + "logps/rejected": -362.0, + "loss": 0.0473, + "rewards/accuracies": 0.984375, + "rewards/chosen": -0.06982421875, + "rewards/margins": 12.0, + "rewards/rejected": -12.0625, + "step": 1610 + }, + { + "epoch": 1.9766871165644173, + "grad_norm": 11.034197753449089, + "learning_rate": 5.169421487603305e-07, + "logits/chosen": -0.0849609375, + "logits/rejected": -0.216796875, + "logps/chosen": -328.0, + "logps/rejected": -342.0, + "loss": 0.0548, + "rewards/accuracies": 0.9765625, + "rewards/chosen": -0.78125, + "rewards/margins": 10.5, + "rewards/rejected": -11.25, + "step": 1611 + }, + { + "epoch": 1.977914110429448, + "grad_norm": 19.25493001940352, + "learning_rate": 5.163223140495868e-07, + "logits/chosen": -0.07275390625, + "logits/rejected": -0.275390625, + "logps/chosen": -436.0, + "logps/rejected": -462.0, + "loss": 0.0788, + "rewards/accuracies": 0.96875, + "rewards/chosen": 0.478515625, + "rewards/margins": 12.75, + "rewards/rejected": -12.25, + "step": 1612 + }, + { + "epoch": 1.9791411042944786, + "grad_norm": 2.50058852484782, + "learning_rate": 5.15702479338843e-07, + "logits/chosen": -0.041259765625, + "logits/rejected": -0.2119140625, + "logps/chosen": -442.0, + "logps/rejected": -420.0, + "loss": 0.0116, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.1943359375, + "rewards/margins": 13.5, + "rewards/rejected": -13.3125, + "step": 1613 + }, + { + "epoch": 1.9803680981595093, + "grad_norm": 9.641767117339855, + "learning_rate": 5.150826446280992e-07, + "logits/chosen": -0.259765625, + "logits/rejected": -0.470703125, + "logps/chosen": -436.0, + "logps/rejected": -432.0, + "loss": 0.0409, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.1953125, + "rewards/margins": 13.0, + "rewards/rejected": -12.8125, + "step": 1614 + }, + { + "epoch": 1.98159509202454, + "grad_norm": 13.7846054101889, + "learning_rate": 5.144628099173554e-07, + "logits/chosen": -0.283203125, + "logits/rejected": -0.37890625, + "logps/chosen": -376.0, + "logps/rejected": -414.0, + "loss": 0.0534, + "rewards/accuracies": 0.96875, + "rewards/chosen": 0.126953125, + "rewards/margins": 12.6875, + "rewards/rejected": -12.5625, + "step": 1615 + }, + { + "epoch": 1.9828220858895707, + "grad_norm": 10.800122873462064, + "learning_rate": 5.138429752066116e-07, + "logits/chosen": -0.0927734375, + "logits/rejected": -0.2470703125, + "logps/chosen": -352.0, + "logps/rejected": -348.0, + "loss": 0.0361, + "rewards/accuracies": 0.984375, + "rewards/chosen": -0.28515625, + "rewards/margins": 11.3125, + "rewards/rejected": -11.625, + "step": 1616 + }, + { + "epoch": 1.9840490797546013, + "grad_norm": 10.77210403419169, + "learning_rate": 5.132231404958678e-07, + "logits/chosen": -0.033935546875, + "logits/rejected": -0.2216796875, + "logps/chosen": -360.0, + "logps/rejected": -372.0, + "loss": 0.0398, + "rewards/accuracies": 0.9765625, + "rewards/chosen": -0.0654296875, + "rewards/margins": 10.875, + "rewards/rejected": -10.9375, + "step": 1617 + }, + { + "epoch": 1.985276073619632, + "grad_norm": 9.529792792248653, + "learning_rate": 5.126033057851239e-07, + "logits/chosen": -0.2080078125, + "logits/rejected": -0.3203125, + "logps/chosen": -370.0, + "logps/rejected": -378.0, + "loss": 0.037, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.2373046875, + "rewards/margins": 11.875, + "rewards/rejected": -11.6875, + "step": 1618 + }, + { + "epoch": 1.9865030674846627, + "grad_norm": 8.741040335110851, + "learning_rate": 5.119834710743802e-07, + "logits/chosen": -0.2314453125, + "logits/rejected": -0.384765625, + "logps/chosen": -416.0, + "logps/rejected": -388.0, + "loss": 0.0321, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.0927734375, + "rewards/margins": 11.75, + "rewards/rejected": -11.875, + "step": 1619 + }, + { + "epoch": 1.9877300613496933, + "grad_norm": 7.785518167433976, + "learning_rate": 5.113636363636363e-07, + "logits/chosen": -0.134765625, + "logits/rejected": -0.3203125, + "logps/chosen": -372.0, + "logps/rejected": -372.0, + "loss": 0.0185, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.33984375, + "rewards/margins": 11.75, + "rewards/rejected": -11.375, + "step": 1620 + }, + { + "epoch": 1.988957055214724, + "grad_norm": 5.464211487320141, + "learning_rate": 5.107438016528926e-07, + "logits/chosen": -0.10888671875, + "logits/rejected": -0.2353515625, + "logps/chosen": -408.0, + "logps/rejected": -422.0, + "loss": 0.0187, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.39453125, + "rewards/margins": 12.4375, + "rewards/rejected": -12.0625, + "step": 1621 + }, + { + "epoch": 1.9901840490797547, + "grad_norm": 13.071939990292023, + "learning_rate": 5.101239669421488e-07, + "logits/chosen": -0.3828125, + "logits/rejected": -0.5078125, + "logps/chosen": -408.0, + "logps/rejected": -442.0, + "loss": 0.0443, + "rewards/accuracies": 0.9765625, + "rewards/chosen": -0.10546875, + "rewards/margins": 11.9375, + "rewards/rejected": -12.0625, + "step": 1622 + }, + { + "epoch": 1.9914110429447853, + "grad_norm": 3.434637787021447, + "learning_rate": 5.09504132231405e-07, + "logits/chosen": -0.0947265625, + "logits/rejected": -0.30078125, + "logps/chosen": -402.0, + "logps/rejected": -402.0, + "loss": 0.0095, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.61328125, + "rewards/margins": 12.5, + "rewards/rejected": -11.875, + "step": 1623 + }, + { + "epoch": 1.992638036809816, + "grad_norm": 3.315305852103551, + "learning_rate": 5.088842975206612e-07, + "logits/chosen": -0.212890625, + "logits/rejected": -0.474609375, + "logps/chosen": -376.0, + "logps/rejected": -394.0, + "loss": 0.0142, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.75390625, + "rewards/margins": 12.75, + "rewards/rejected": -12.0, + "step": 1624 + }, + { + "epoch": 1.9938650306748467, + "grad_norm": 22.7278950349499, + "learning_rate": 5.082644628099174e-07, + "logits/chosen": -0.043212890625, + "logits/rejected": -0.2578125, + "logps/chosen": -440.0, + "logps/rejected": -408.0, + "loss": 0.0652, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.7734375, + "rewards/margins": 12.375, + "rewards/rejected": -11.625, + "step": 1625 + }, + { + "epoch": 1.9950920245398773, + "grad_norm": 7.501231527411867, + "learning_rate": 5.076446280991736e-07, + "logits/chosen": -0.17578125, + "logits/rejected": -0.28125, + "logps/chosen": -432.0, + "logps/rejected": -426.0, + "loss": 0.0195, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.251953125, + "rewards/margins": 12.0, + "rewards/rejected": -11.75, + "step": 1626 + }, + { + "epoch": 1.996319018404908, + "grad_norm": 8.75791968734159, + "learning_rate": 5.070247933884297e-07, + "logits/chosen": -0.24609375, + "logits/rejected": -0.345703125, + "logps/chosen": -372.0, + "logps/rejected": -368.0, + "loss": 0.0302, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.4765625, + "rewards/margins": 12.1875, + "rewards/rejected": -11.6875, + "step": 1627 + }, + { + "epoch": 1.9975460122699387, + "grad_norm": 12.320799734184225, + "learning_rate": 5.06404958677686e-07, + "logits/chosen": -0.1689453125, + "logits/rejected": -0.296875, + "logps/chosen": -368.0, + "logps/rejected": -406.0, + "loss": 0.0386, + "rewards/accuracies": 0.9765625, + "rewards/chosen": 0.65625, + "rewards/margins": 12.6875, + "rewards/rejected": -12.0, + "step": 1628 + }, + { + "epoch": 1.9987730061349693, + "grad_norm": 16.231968818300622, + "learning_rate": 5.057851239669421e-07, + "logits/chosen": -0.05126953125, + "logits/rejected": -0.216796875, + "logps/chosen": -398.0, + "logps/rejected": -408.0, + "loss": 0.0634, + "rewards/accuracies": 0.9765625, + "rewards/chosen": -0.11181640625, + "rewards/margins": 11.5, + "rewards/rejected": -11.625, + "step": 1629 + }, + { + "epoch": 2.0, + "grad_norm": 7.251135077046663, + "learning_rate": 5.051652892561984e-07, + "logits/chosen": -0.1787109375, + "logits/rejected": -0.341796875, + "logps/chosen": -428.0, + "logps/rejected": -424.0, + "loss": 0.0224, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.1337890625, + "rewards/margins": 12.75, + "rewards/rejected": -12.625, + "step": 1630 + } + ], + "logging_steps": 1, + "max_steps": 2445, + "num_input_tokens_seen": 0, + "num_train_epochs": 3, + "save_steps": 500, + "stateful_callbacks": { + "TrainerControl": { + "args": { + "should_epoch_stop": false, + "should_evaluate": false, + "should_log": false, + "should_save": true, + "should_training_stop": false + }, + "attributes": {} + } + }, + "total_flos": 0.0, + "train_batch_size": 4, + "trial_name": null, + "trial_params": null +} diff --git a/checkpoint-1630/training_args.bin b/checkpoint-1630/training_args.bin new file mode 100644 index 0000000..a88cbd3 --- /dev/null +++ b/checkpoint-1630/training_args.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:e0cffaa6f04a26499dd6171106fabe250a7e89bd923590f4fdeb897bc28870a9 +size 9528 diff --git a/checkpoint-1630/zero_to_fp32.py b/checkpoint-1630/zero_to_fp32.py new file mode 100644 index 0000000..24cc342 --- /dev/null +++ b/checkpoint-1630/zero_to_fp32.py @@ -0,0 +1,604 @@ +#!/usr/bin/env python + +# Copyright (c) Microsoft Corporation. +# SPDX-License-Identifier: Apache-2.0 + +# DeepSpeed Team + +# This script extracts fp32 consolidated weights from a zero 1, 2 and 3 DeepSpeed checkpoints. It gets +# copied into the top level checkpoint dir, so the user can easily do the conversion at any point in +# the future. Once extracted, the weights don't require DeepSpeed and can be used in any +# application. +# +# example: python zero_to_fp32.py . pytorch_model.bin + +import argparse +import torch +import glob +import math +import os +import re +from collections import OrderedDict +from dataclasses import dataclass + +# while this script doesn't use deepspeed to recover data, since the checkpoints are pickled with +# DeepSpeed data structures it has to be available in the current python environment. +from deepspeed.utils import logger +from deepspeed.checkpoint.constants import (DS_VERSION, OPTIMIZER_STATE_DICT, SINGLE_PARTITION_OF_FP32_GROUPS, + FP32_FLAT_GROUPS, ZERO_STAGE, PARTITION_COUNT, PARAM_SHAPES, BUFFER_NAMES, + FROZEN_PARAM_SHAPES, FROZEN_PARAM_FRAGMENTS) + + +@dataclass +class zero_model_state: + buffers: dict() + param_shapes: dict() + shared_params: list + ds_version: int + frozen_param_shapes: dict() + frozen_param_fragments: dict() + + +debug = 0 + +# load to cpu +device = torch.device('cpu') + + +def atoi(text): + return int(text) if text.isdigit() else text + + +def natural_keys(text): + ''' + alist.sort(key=natural_keys) sorts in human order + http://nedbatchelder.com/blog/200712/human_sorting.html + (See Toothy's implementation in the comments) + ''' + return [atoi(c) for c in re.split(r'(\d+)', text)] + + +def get_model_state_file(checkpoint_dir, zero_stage): + if not os.path.isdir(checkpoint_dir): + raise FileNotFoundError(f"Directory '{checkpoint_dir}' doesn't exist") + + # there should be only one file + if zero_stage <= 2: + file = os.path.join(checkpoint_dir, "mp_rank_00_model_states.pt") + elif zero_stage == 3: + file = os.path.join(checkpoint_dir, "zero_pp_rank_0_mp_rank_00_model_states.pt") + + if not os.path.exists(file): + raise FileNotFoundError(f"can't find model states file at '{file}'") + + return file + + +def get_checkpoint_files(checkpoint_dir, glob_pattern): + # XXX: need to test that this simple glob rule works for multi-node setup too + ckpt_files = sorted(glob.glob(os.path.join(checkpoint_dir, glob_pattern)), key=natural_keys) + + if len(ckpt_files) == 0: + raise FileNotFoundError(f"can't find {glob_pattern} files in directory '{checkpoint_dir}'") + + return ckpt_files + + +def get_optim_files(checkpoint_dir): + return get_checkpoint_files(checkpoint_dir, "*_optim_states.pt") + + +def get_model_state_files(checkpoint_dir): + return get_checkpoint_files(checkpoint_dir, "*_model_states.pt") + + +def parse_model_states(files): + zero_model_states = [] + for file in files: + state_dict = torch.load(file, map_location=device) + + if BUFFER_NAMES not in state_dict: + raise ValueError(f"{file} is not a model state checkpoint") + buffer_names = state_dict[BUFFER_NAMES] + if debug: + print("Found buffers:", buffer_names) + + # recover just the buffers while restoring them to fp32 if they were saved in fp16 + buffers = {k: v.float() for k, v in state_dict["module"].items() if k in buffer_names} + param_shapes = state_dict[PARAM_SHAPES] + + # collect parameters that are included in param_shapes + param_names = [] + for s in param_shapes: + for name in s.keys(): + param_names.append(name) + + # update with frozen parameters + frozen_param_shapes = state_dict.get(FROZEN_PARAM_SHAPES, None) + if frozen_param_shapes is not None: + if debug: + print(f"Found frozen_param_shapes: {frozen_param_shapes}") + param_names += list(frozen_param_shapes.keys()) + + # handle shared params + shared_params = [[k, v] for k, v in state_dict["shared_params"].items()] + + ds_version = state_dict.get(DS_VERSION, None) + + frozen_param_fragments = state_dict.get(FROZEN_PARAM_FRAGMENTS, None) + + z_model_state = zero_model_state(buffers=buffers, + param_shapes=param_shapes, + shared_params=shared_params, + ds_version=ds_version, + frozen_param_shapes=frozen_param_shapes, + frozen_param_fragments=frozen_param_fragments) + zero_model_states.append(z_model_state) + + return zero_model_states + + +def parse_optim_states(files, ds_checkpoint_dir): + + total_files = len(files) + state_dicts = [] + for f in files: + state_dict = torch.load(f, map_location=device) + # immediately discard the potentially huge 2 optimizer states as we only care for fp32 master weights + # and also handle the case where it was already removed by another helper script + state_dict["optimizer_state_dict"].pop("optimizer_state_dict", None) + state_dicts.append(state_dict) + + if not ZERO_STAGE in state_dicts[0][OPTIMIZER_STATE_DICT]: + raise ValueError(f"{files[0]} is not a zero checkpoint") + zero_stage = state_dicts[0][OPTIMIZER_STATE_DICT][ZERO_STAGE] + world_size = state_dicts[0][OPTIMIZER_STATE_DICT][PARTITION_COUNT] + + # For ZeRO-2 each param group can have different partition_count as data parallelism for expert + # parameters can be different from data parallelism for non-expert parameters. So we can just + # use the max of the partition_count to get the dp world_size. + + if type(world_size) is list: + world_size = max(world_size) + + if world_size != total_files: + raise ValueError( + f"Expected {world_size} of '*_optim_states.pt' under '{ds_checkpoint_dir}' but found {total_files} files. " + "Possibly due to an overwrite of an old checkpoint, or a checkpoint didn't get saved by one or more processes." + ) + + # the groups are named differently in each stage + if zero_stage <= 2: + fp32_groups_key = SINGLE_PARTITION_OF_FP32_GROUPS + elif zero_stage == 3: + fp32_groups_key = FP32_FLAT_GROUPS + else: + raise ValueError(f"unknown zero stage {zero_stage}") + + if zero_stage <= 2: + fp32_flat_groups = [state_dicts[i][OPTIMIZER_STATE_DICT][fp32_groups_key] for i in range(len(state_dicts))] + elif zero_stage == 3: + # if there is more than one param group, there will be multiple flattened tensors - one + # flattened tensor per group - for simplicity merge them into a single tensor + # + # XXX: could make the script more memory efficient for when there are multiple groups - it + # will require matching the sub-lists of param_shapes for each param group flattened tensor + + fp32_flat_groups = [ + torch.cat(state_dicts[i][OPTIMIZER_STATE_DICT][fp32_groups_key], 0) for i in range(len(state_dicts)) + ] + + return zero_stage, world_size, fp32_flat_groups + + +def _get_fp32_state_dict_from_zero_checkpoint(ds_checkpoint_dir, exclude_frozen_parameters): + """ + Returns fp32 state_dict reconstructed from ds checkpoint + + Args: + - ``ds_checkpoint_dir``: path to the deepspeed checkpoint folder (where the optimizer files are) + + """ + print(f"Processing zero checkpoint '{ds_checkpoint_dir}'") + + optim_files = get_optim_files(ds_checkpoint_dir) + zero_stage, world_size, fp32_flat_groups = parse_optim_states(optim_files, ds_checkpoint_dir) + print(f"Detected checkpoint of type zero stage {zero_stage}, world_size: {world_size}") + + model_files = get_model_state_files(ds_checkpoint_dir) + + zero_model_states = parse_model_states(model_files) + print(f'Parsing checkpoint created by deepspeed=={zero_model_states[0].ds_version}') + + if zero_stage <= 2: + return _get_fp32_state_dict_from_zero2_checkpoint(world_size, fp32_flat_groups, zero_model_states, + exclude_frozen_parameters) + elif zero_stage == 3: + return _get_fp32_state_dict_from_zero3_checkpoint(world_size, fp32_flat_groups, zero_model_states, + exclude_frozen_parameters) + + +def _zero2_merge_frozen_params(state_dict, zero_model_states): + if zero_model_states[0].frozen_param_shapes is None or len(zero_model_states[0].frozen_param_shapes) == 0: + return + + frozen_param_shapes = zero_model_states[0].frozen_param_shapes + frozen_param_fragments = zero_model_states[0].frozen_param_fragments + + if debug: + num_elem = sum(s.numel() for s in frozen_param_shapes.values()) + print(f'rank 0: {FROZEN_PARAM_SHAPES}.numel = {num_elem}') + + wanted_params = len(frozen_param_shapes) + wanted_numel = sum(s.numel() for s in frozen_param_shapes.values()) + avail_numel = sum([p.numel() for p in frozen_param_fragments.values()]) + print(f'Frozen params: Have {avail_numel} numels to process.') + print(f'Frozen params: Need {wanted_numel} numels in {wanted_params} params') + + total_params = 0 + total_numel = 0 + for name, shape in frozen_param_shapes.items(): + total_params += 1 + unpartitioned_numel = shape.numel() + total_numel += unpartitioned_numel + + state_dict[name] = frozen_param_fragments[name] + + if debug: + print(f"{name} full shape: {shape} unpartitioned numel {unpartitioned_numel} ") + + print(f"Reconstructed Frozen fp32 state dict with {total_params} params {total_numel} elements") + + +def _has_callable(obj, fn): + attr = getattr(obj, fn, None) + return callable(attr) + + +def _zero2_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states): + param_shapes = zero_model_states[0].param_shapes + + # Reconstruction protocol: + # + # XXX: document this + + if debug: + for i in range(world_size): + for j in range(len(fp32_flat_groups[0])): + print(f"{FP32_FLAT_GROUPS}[{i}][{j}].shape={fp32_flat_groups[i][j].shape}") + + # XXX: memory usage doubles here (zero2) + num_param_groups = len(fp32_flat_groups[0]) + merged_single_partition_of_fp32_groups = [] + for i in range(num_param_groups): + merged_partitions = [sd[i] for sd in fp32_flat_groups] + full_single_fp32_vector = torch.cat(merged_partitions, 0) + merged_single_partition_of_fp32_groups.append(full_single_fp32_vector) + avail_numel = sum( + [full_single_fp32_vector.numel() for full_single_fp32_vector in merged_single_partition_of_fp32_groups]) + + if debug: + wanted_params = sum([len(shapes) for shapes in param_shapes]) + wanted_numel = sum([sum(shape.numel() for shape in shapes.values()) for shapes in param_shapes]) + # not asserting if there is a mismatch due to possible padding + print(f"Have {avail_numel} numels to process.") + print(f"Need {wanted_numel} numels in {wanted_params} params.") + + # params + # XXX: for huge models that can't fit into the host's RAM we will have to recode this to support + # out-of-core computing solution + total_numel = 0 + total_params = 0 + for shapes, full_single_fp32_vector in zip(param_shapes, merged_single_partition_of_fp32_groups): + offset = 0 + avail_numel = full_single_fp32_vector.numel() + for name, shape in shapes.items(): + + unpartitioned_numel = shape.numel() if _has_callable(shape, 'numel') else math.prod(shape) + total_numel += unpartitioned_numel + total_params += 1 + + if debug: + print(f"{name} full shape: {shape} unpartitioned numel {unpartitioned_numel} ") + state_dict[name] = full_single_fp32_vector.narrow(0, offset, unpartitioned_numel).view(shape) + offset += unpartitioned_numel + + # Z2 started to align to 2*world_size to improve nccl performance. Therefore both offset and + # avail_numel can differ by anywhere between 0..2*world_size. Due to two unrelated complex + # paddings performed in the code it's almost impossible to predict the exact numbers w/o the + # live optimizer object, so we are checking that the numbers are within the right range + align_to = 2 * world_size + + def zero2_align(x): + return align_to * math.ceil(x / align_to) + + if debug: + print(f"original offset={offset}, avail_numel={avail_numel}") + + offset = zero2_align(offset) + avail_numel = zero2_align(avail_numel) + + if debug: + print(f"aligned offset={offset}, avail_numel={avail_numel}") + + # Sanity check + if offset != avail_numel: + raise ValueError(f"consumed {offset} numels out of {avail_numel} - something is wrong") + + print(f"Reconstructed fp32 state dict with {total_params} params {total_numel} elements") + + +def _get_fp32_state_dict_from_zero2_checkpoint(world_size, fp32_flat_groups, zero_model_states, + exclude_frozen_parameters): + state_dict = OrderedDict() + + # buffers + buffers = zero_model_states[0].buffers + state_dict.update(buffers) + if debug: + print(f"added {len(buffers)} buffers") + + if not exclude_frozen_parameters: + _zero2_merge_frozen_params(state_dict, zero_model_states) + + _zero2_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states) + + # recover shared parameters + for pair in zero_model_states[0].shared_params: + if pair[1] in state_dict: + state_dict[pair[0]] = state_dict[pair[1]] + + return state_dict + + +def zero3_partitioned_param_info(unpartitioned_numel, world_size): + remainder = unpartitioned_numel % world_size + padding_numel = (world_size - remainder) if remainder else 0 + partitioned_numel = math.ceil(unpartitioned_numel / world_size) + return partitioned_numel, padding_numel + + +def _zero3_merge_frozen_params(state_dict, world_size, zero_model_states): + if zero_model_states[0].frozen_param_shapes is None or len(zero_model_states[0].frozen_param_shapes) == 0: + return + + if debug: + for i in range(world_size): + num_elem = sum(s.numel() for s in zero_model_states[i].frozen_param_fragments.values()) + print(f'rank {i}: {FROZEN_PARAM_SHAPES}.numel = {num_elem}') + + frozen_param_shapes = zero_model_states[0].frozen_param_shapes + wanted_params = len(frozen_param_shapes) + wanted_numel = sum(s.numel() for s in frozen_param_shapes.values()) + avail_numel = sum([p.numel() for p in zero_model_states[0].frozen_param_fragments.values()]) * world_size + print(f'Frozen params: Have {avail_numel} numels to process.') + print(f'Frozen params: Need {wanted_numel} numels in {wanted_params} params') + + total_params = 0 + total_numel = 0 + for name, shape in zero_model_states[0].frozen_param_shapes.items(): + total_params += 1 + unpartitioned_numel = shape.numel() + total_numel += unpartitioned_numel + + param_frags = tuple(model_state.frozen_param_fragments[name] for model_state in zero_model_states) + state_dict[name] = torch.cat(param_frags, 0).narrow(0, 0, unpartitioned_numel).view(shape) + + partitioned_numel, partitioned_padding_numel = zero3_partitioned_param_info(unpartitioned_numel, world_size) + + if debug: + print( + f"Frozen params: {total_params} {name} full shape: {shape} partition0 numel={partitioned_numel} partitioned_padding_numel={partitioned_padding_numel}" + ) + + print(f"Reconstructed Frozen fp32 state dict with {total_params} params {total_numel} elements") + + +def _zero3_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states): + param_shapes = zero_model_states[0].param_shapes + avail_numel = fp32_flat_groups[0].numel() * world_size + # Reconstruction protocol: For zero3 we need to zip the partitions together at boundary of each + # param, re-consolidating each param, while dealing with padding if any + + # merge list of dicts, preserving order + param_shapes = {k: v for d in param_shapes for k, v in d.items()} + + if debug: + for i in range(world_size): + print(f"{FP32_FLAT_GROUPS}[{i}].shape={fp32_flat_groups[i].shape}") + + wanted_params = len(param_shapes) + wanted_numel = sum(shape.numel() for shape in param_shapes.values()) + # not asserting if there is a mismatch due to possible padding + avail_numel = fp32_flat_groups[0].numel() * world_size + print(f"Trainable params: Have {avail_numel} numels to process.") + print(f"Trainable params: Need {wanted_numel} numels in {wanted_params} params.") + + # params + # XXX: for huge models that can't fit into the host's RAM we will have to recode this to support + # out-of-core computing solution + offset = 0 + total_numel = 0 + total_params = 0 + for name, shape in param_shapes.items(): + + unpartitioned_numel = shape.numel() + total_numel += unpartitioned_numel + total_params += 1 + + partitioned_numel, partitioned_padding_numel = zero3_partitioned_param_info(unpartitioned_numel, world_size) + + if debug: + print( + f"Trainable params: {total_params} {name} full shape: {shape} partition0 numel={partitioned_numel} partitioned_padding_numel={partitioned_padding_numel}" + ) + + # XXX: memory usage doubles here + state_dict[name] = torch.cat( + tuple(fp32_flat_groups[i].narrow(0, offset, partitioned_numel) for i in range(world_size)), + 0).narrow(0, 0, unpartitioned_numel).view(shape) + offset += partitioned_numel + + offset *= world_size + + # Sanity check + if offset != avail_numel: + raise ValueError(f"consumed {offset} numels out of {avail_numel} - something is wrong") + + print(f"Reconstructed Trainable fp32 state dict with {total_params} params {total_numel} elements") + + +def _get_fp32_state_dict_from_zero3_checkpoint(world_size, fp32_flat_groups, zero_model_states, + exclude_frozen_parameters): + state_dict = OrderedDict() + + # buffers + buffers = zero_model_states[0].buffers + state_dict.update(buffers) + if debug: + print(f"added {len(buffers)} buffers") + + if not exclude_frozen_parameters: + _zero3_merge_frozen_params(state_dict, world_size, zero_model_states) + + _zero3_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states) + + # recover shared parameters + for pair in zero_model_states[0].shared_params: + if pair[1] in state_dict: + state_dict[pair[0]] = state_dict[pair[1]] + + return state_dict + + +def get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag=None, exclude_frozen_parameters=False): + """ + Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated state_dict that can be loaded with + ``load_state_dict()`` and used for training without DeepSpeed or shared with others, for example + via a model hub. + + Args: + - ``checkpoint_dir``: path to the desired checkpoint folder + - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in 'latest' file. e.g., ``global_step14`` + - ``exclude_frozen_parameters``: exclude frozen parameters + + Returns: + - pytorch ``state_dict`` + + Note: this approach may not work if your application doesn't have sufficient free CPU memory and + you may need to use the offline approach using the ``zero_to_fp32.py`` script that is saved with + the checkpoint. + + A typical usage might be :: + + from deepspeed.utils.zero_to_fp32 import get_fp32_state_dict_from_zero_checkpoint + # do the training and checkpoint saving + state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir) # already on cpu + model = model.cpu() # move to cpu + model.load_state_dict(state_dict) + # submit to model hub or save the model to share with others + + In this example the ``model`` will no longer be usable in the deepspeed context of the same + application. i.e. you will need to re-initialize the deepspeed engine, since + ``model.load_state_dict(state_dict)`` will remove all the deepspeed magic from it. + + If you want it all done for you, use ``load_state_dict_from_zero_checkpoint`` instead. + + """ + if tag is None: + latest_path = os.path.join(checkpoint_dir, 'latest') + if os.path.isfile(latest_path): + with open(latest_path, 'r') as fd: + tag = fd.read().strip() + else: + raise ValueError(f"Unable to find 'latest' file at {latest_path}") + + ds_checkpoint_dir = os.path.join(checkpoint_dir, tag) + + if not os.path.isdir(ds_checkpoint_dir): + raise FileNotFoundError(f"Directory '{ds_checkpoint_dir}' doesn't exist") + + return _get_fp32_state_dict_from_zero_checkpoint(ds_checkpoint_dir, exclude_frozen_parameters) + + +def convert_zero_checkpoint_to_fp32_state_dict(checkpoint_dir, output_file, tag=None, exclude_frozen_parameters=False): + """ + Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated ``state_dict`` file that can be + loaded with ``torch.load(file)`` + ``load_state_dict()`` and used for training without DeepSpeed. + + Args: + - ``checkpoint_dir``: path to the desired checkpoint folder. (one that contains the tag-folder, like ``global_step14``) + - ``output_file``: path to the pytorch fp32 state_dict output file (e.g. path/pytorch_model.bin) + - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in the file named ``latest`` in the checkpoint folder, e.g., ``global_step14`` + - ``exclude_frozen_parameters``: exclude frozen parameters + """ + + state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag, exclude_frozen_parameters) + print(f"Saving fp32 state dict to {output_file}") + torch.save(state_dict, output_file) + + +def load_state_dict_from_zero_checkpoint(model, checkpoint_dir, tag=None): + """ + 1. Put the provided model to cpu + 2. Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated ``state_dict`` + 3. Load it into the provided model + + Args: + - ``model``: the model object to update + - ``checkpoint_dir``: path to the desired checkpoint folder. (one that contains the tag-folder, like ``global_step14``) + - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in the file named ``latest`` in the checkpoint folder, e.g., ``global_step14`` + + Returns: + - ``model`: modified model + + Make sure you have plenty of CPU memory available before you call this function. If you don't + have enough use the ``zero_to_fp32.py`` utility to do the conversion. You will find it + conveniently placed for you in the checkpoint folder. + + A typical usage might be :: + + from deepspeed.utils.zero_to_fp32 import load_state_dict_from_zero_checkpoint + model = load_state_dict_from_zero_checkpoint(trainer.model, checkpoint_dir) + # submit to model hub or save the model to share with others + + Note, that once this was run, the ``model`` will no longer be usable in the deepspeed context + of the same application. i.e. you will need to re-initialize the deepspeed engine, since + ``model.load_state_dict(state_dict)`` will remove all the deepspeed magic from it. + + """ + logger.info(f"Extracting fp32 weights") + state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag) + + logger.info(f"Overwriting model with fp32 weights") + model = model.cpu() + model.load_state_dict(state_dict, strict=False) + + return model + + +if __name__ == "__main__": + + parser = argparse.ArgumentParser() + parser.add_argument("checkpoint_dir", + type=str, + help="path to the desired checkpoint folder, e.g., path/checkpoint-12") + parser.add_argument( + "output_file", + type=str, + help="path to the pytorch fp32 state_dict output file (e.g. path/checkpoint-12/pytorch_model.bin)") + parser.add_argument("-t", + "--tag", + type=str, + default=None, + help="checkpoint tag used as a unique identifier for checkpoint. e.g., global_step1") + parser.add_argument("--exclude_frozen_parameters", action='store_true', help="exclude frozen parameters") + parser.add_argument("-d", "--debug", action='store_true', help="enable debug") + args = parser.parse_args() + + debug = args.debug + + convert_zero_checkpoint_to_fp32_state_dict(args.checkpoint_dir, + args.output_file, + tag=args.tag, + exclude_frozen_parameters=args.exclude_frozen_parameters) diff --git a/checkpoint-2445/config.json b/checkpoint-2445/config.json new file mode 100644 index 0000000..ded5a8b --- /dev/null +++ b/checkpoint-2445/config.json @@ -0,0 +1,35 @@ +{ + "architectures": [ + "LlamaForCausalLM" + ], + "attention_bias": false, + "attention_dropout": 0.0, + "bos_token_id": 128000, + "eos_token_id": 128009, + "head_dim": 128, + "hidden_act": "silu", + "hidden_size": 4096, + "initializer_range": 0.02, + "intermediate_size": 14336, + "max_position_embeddings": 131072, + "mlp_bias": false, + "model_type": "llama", + "num_attention_heads": 32, + "num_hidden_layers": 32, + "num_key_value_heads": 8, + "pretraining_tp": 1, + "rms_norm_eps": 1e-05, + "rope_scaling": { + "factor": 8.0, + "high_freq_factor": 4.0, + "low_freq_factor": 1.0, + "original_max_position_embeddings": 8192, + "rope_type": "llama3" + }, + "rope_theta": 500000.0, + "tie_word_embeddings": false, + "torch_dtype": "bfloat16", + "transformers_version": "4.50.0", + "use_cache": false, + "vocab_size": 128256 +} diff --git a/checkpoint-2445/generation_config.json b/checkpoint-2445/generation_config.json new file mode 100644 index 0000000..0acdb5a --- /dev/null +++ b/checkpoint-2445/generation_config.json @@ -0,0 +1,12 @@ +{ + "bos_token_id": 128000, + "do_sample": true, + "eos_token_id": [ + 128001, + 128008, + 128009 + ], + "temperature": 0.6, + "top_p": 0.9, + "transformers_version": "4.50.0" +} diff --git a/checkpoint-2445/global_step2445/bf16_zero_pp_rank_0_mp_rank_00_optim_states.pt b/checkpoint-2445/global_step2445/bf16_zero_pp_rank_0_mp_rank_00_optim_states.pt new file mode 100644 index 0000000..0e41978 --- /dev/null +++ b/checkpoint-2445/global_step2445/bf16_zero_pp_rank_0_mp_rank_00_optim_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:6792578d88474bc5587deb51212d52bb0e309b6477af43f5be5521799453956a +size 1514015111 diff --git a/checkpoint-2445/global_step2445/bf16_zero_pp_rank_10_mp_rank_00_optim_states.pt b/checkpoint-2445/global_step2445/bf16_zero_pp_rank_10_mp_rank_00_optim_states.pt new file mode 100644 index 0000000..547bac6 --- /dev/null +++ b/checkpoint-2445/global_step2445/bf16_zero_pp_rank_10_mp_rank_00_optim_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:a850d6ee5f6d88d98d3a88a6dc4749a05cd23daaeebf13b3e6b41ce2ede15e10 +size 1514016442 diff --git a/checkpoint-2445/global_step2445/bf16_zero_pp_rank_11_mp_rank_00_optim_states.pt b/checkpoint-2445/global_step2445/bf16_zero_pp_rank_11_mp_rank_00_optim_states.pt new file mode 100644 index 0000000..8943db6 --- /dev/null +++ b/checkpoint-2445/global_step2445/bf16_zero_pp_rank_11_mp_rank_00_optim_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:5c036c88b20a676f617b5729910d2ca9ef8caceda3640a13d4da8ad26f9b7f53 +size 1514016442 diff --git a/checkpoint-2445/global_step2445/bf16_zero_pp_rank_12_mp_rank_00_optim_states.pt b/checkpoint-2445/global_step2445/bf16_zero_pp_rank_12_mp_rank_00_optim_states.pt new file mode 100644 index 0000000..2b0e19e --- /dev/null +++ b/checkpoint-2445/global_step2445/bf16_zero_pp_rank_12_mp_rank_00_optim_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:1377b40f426e870963d954b77ee06bc8224571d518482191f3ea10cb9495ff6c +size 1514016442 diff --git a/checkpoint-2445/global_step2445/bf16_zero_pp_rank_13_mp_rank_00_optim_states.pt b/checkpoint-2445/global_step2445/bf16_zero_pp_rank_13_mp_rank_00_optim_states.pt new file mode 100644 index 0000000..ba1d1ae --- /dev/null +++ b/checkpoint-2445/global_step2445/bf16_zero_pp_rank_13_mp_rank_00_optim_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:e1f8b920921997cdbaf694828200423c6addfb2e6ca488a9a866582615a12cfe +size 1514016442 diff --git a/checkpoint-2445/global_step2445/bf16_zero_pp_rank_14_mp_rank_00_optim_states.pt b/checkpoint-2445/global_step2445/bf16_zero_pp_rank_14_mp_rank_00_optim_states.pt new file mode 100644 index 0000000..dec4782 --- /dev/null +++ b/checkpoint-2445/global_step2445/bf16_zero_pp_rank_14_mp_rank_00_optim_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:587635760355854bdb923cb34e09fd725c5bdbd46e1963479c5e4a25edd6b5db +size 1514016442 diff --git a/checkpoint-2445/global_step2445/bf16_zero_pp_rank_15_mp_rank_00_optim_states.pt b/checkpoint-2445/global_step2445/bf16_zero_pp_rank_15_mp_rank_00_optim_states.pt new file mode 100644 index 0000000..4a1590a --- /dev/null +++ b/checkpoint-2445/global_step2445/bf16_zero_pp_rank_15_mp_rank_00_optim_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:1934e4b5501be77d9a1052ea77f500c6f265905bcd0dbaa365a28230d603e76b +size 1514016442 diff --git a/checkpoint-2445/global_step2445/bf16_zero_pp_rank_16_mp_rank_00_optim_states.pt b/checkpoint-2445/global_step2445/bf16_zero_pp_rank_16_mp_rank_00_optim_states.pt new file mode 100644 index 0000000..0a5319c --- /dev/null +++ b/checkpoint-2445/global_step2445/bf16_zero_pp_rank_16_mp_rank_00_optim_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:c23a039cc74dec6f52bd8d959de9215e092d9aaab93893f6908ccaaf67b4d67d +size 1514016442 diff --git a/checkpoint-2445/global_step2445/bf16_zero_pp_rank_17_mp_rank_00_optim_states.pt b/checkpoint-2445/global_step2445/bf16_zero_pp_rank_17_mp_rank_00_optim_states.pt new file mode 100644 index 0000000..3d399c0 --- /dev/null +++ b/checkpoint-2445/global_step2445/bf16_zero_pp_rank_17_mp_rank_00_optim_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:c652ab82f8a3014d2ff01d40229a0ebbdd510b75eeb86576780fdd8cce8f74eb +size 1514016442 diff --git a/checkpoint-2445/global_step2445/bf16_zero_pp_rank_18_mp_rank_00_optim_states.pt b/checkpoint-2445/global_step2445/bf16_zero_pp_rank_18_mp_rank_00_optim_states.pt new file mode 100644 index 0000000..30a3b64 --- /dev/null +++ b/checkpoint-2445/global_step2445/bf16_zero_pp_rank_18_mp_rank_00_optim_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:a0a07d85bde4c1eb15fc1ecea554b1d15b897159f83b7a214758c4acae8fd2ab +size 1514016442 diff --git a/checkpoint-2445/global_step2445/bf16_zero_pp_rank_19_mp_rank_00_optim_states.pt b/checkpoint-2445/global_step2445/bf16_zero_pp_rank_19_mp_rank_00_optim_states.pt new file mode 100644 index 0000000..b0dbfeb --- /dev/null +++ b/checkpoint-2445/global_step2445/bf16_zero_pp_rank_19_mp_rank_00_optim_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:b83642279921662c4c29ca739bf3473adb4e8f7bf5a459109391880c8e14b6ce +size 1514016442 diff --git a/checkpoint-2445/global_step2445/bf16_zero_pp_rank_1_mp_rank_00_optim_states.pt b/checkpoint-2445/global_step2445/bf16_zero_pp_rank_1_mp_rank_00_optim_states.pt new file mode 100644 index 0000000..b7ebb43 --- /dev/null +++ b/checkpoint-2445/global_step2445/bf16_zero_pp_rank_1_mp_rank_00_optim_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:3aeda3cf43e8afc35929eecdecb43e9fd1aef95ad516283a11da8e242558a2d7 +size 1514015111 diff --git a/checkpoint-2445/global_step2445/bf16_zero_pp_rank_20_mp_rank_00_optim_states.pt b/checkpoint-2445/global_step2445/bf16_zero_pp_rank_20_mp_rank_00_optim_states.pt new file mode 100644 index 0000000..471bf7f --- /dev/null +++ b/checkpoint-2445/global_step2445/bf16_zero_pp_rank_20_mp_rank_00_optim_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:29750d10d0fc857c0e5efcc80c7d016a7313e8a44d08ee6adfc4389f2d853d8e +size 1514016442 diff --git a/checkpoint-2445/global_step2445/bf16_zero_pp_rank_21_mp_rank_00_optim_states.pt b/checkpoint-2445/global_step2445/bf16_zero_pp_rank_21_mp_rank_00_optim_states.pt new file mode 100644 index 0000000..4ae4d53 --- /dev/null +++ b/checkpoint-2445/global_step2445/bf16_zero_pp_rank_21_mp_rank_00_optim_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:8213ccbb496d780cda0372f3a49c6d59c8f0a8912d97e4f9a92d8dba321a8e16 +size 1514016442 diff --git a/checkpoint-2445/global_step2445/bf16_zero_pp_rank_22_mp_rank_00_optim_states.pt b/checkpoint-2445/global_step2445/bf16_zero_pp_rank_22_mp_rank_00_optim_states.pt new file mode 100644 index 0000000..efbd028 --- /dev/null +++ b/checkpoint-2445/global_step2445/bf16_zero_pp_rank_22_mp_rank_00_optim_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:aaccf699686ee3230b4e7b42fae6b3ad8b39748854758cab216566bb273aa8ff +size 1514016442 diff --git a/checkpoint-2445/global_step2445/bf16_zero_pp_rank_23_mp_rank_00_optim_states.pt b/checkpoint-2445/global_step2445/bf16_zero_pp_rank_23_mp_rank_00_optim_states.pt new file mode 100644 index 0000000..fd04b3c --- /dev/null +++ b/checkpoint-2445/global_step2445/bf16_zero_pp_rank_23_mp_rank_00_optim_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:0ae699dbcb0b9d36cd64086fbdc815e2746385712bc864c2ee7800e872e37254 +size 1514016442 diff --git a/checkpoint-2445/global_step2445/bf16_zero_pp_rank_24_mp_rank_00_optim_states.pt b/checkpoint-2445/global_step2445/bf16_zero_pp_rank_24_mp_rank_00_optim_states.pt new file mode 100644 index 0000000..d028053 --- /dev/null +++ b/checkpoint-2445/global_step2445/bf16_zero_pp_rank_24_mp_rank_00_optim_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:1373e15a93f569e3a9641c463448decdcaa149f1d87bd721bd3da9931e8be1c8 +size 1514016442 diff --git a/checkpoint-2445/global_step2445/bf16_zero_pp_rank_25_mp_rank_00_optim_states.pt b/checkpoint-2445/global_step2445/bf16_zero_pp_rank_25_mp_rank_00_optim_states.pt new file mode 100644 index 0000000..3c39f1a --- /dev/null +++ b/checkpoint-2445/global_step2445/bf16_zero_pp_rank_25_mp_rank_00_optim_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:7675f015a617965fb2ef0348c52aacb116cc9b2f06a6f4623ff322bcfae63ca9 +size 1514016442 diff --git a/checkpoint-2445/global_step2445/bf16_zero_pp_rank_26_mp_rank_00_optim_states.pt b/checkpoint-2445/global_step2445/bf16_zero_pp_rank_26_mp_rank_00_optim_states.pt new file mode 100644 index 0000000..267e5e8 --- /dev/null +++ b/checkpoint-2445/global_step2445/bf16_zero_pp_rank_26_mp_rank_00_optim_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:3d5447f5daaceab60b9d7b28f0db07a47ec305659279db005414bc419f4a94b1 +size 1514016442 diff --git a/checkpoint-2445/global_step2445/bf16_zero_pp_rank_27_mp_rank_00_optim_states.pt b/checkpoint-2445/global_step2445/bf16_zero_pp_rank_27_mp_rank_00_optim_states.pt new file mode 100644 index 0000000..0599465 --- /dev/null +++ b/checkpoint-2445/global_step2445/bf16_zero_pp_rank_27_mp_rank_00_optim_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:c3f632be01fd59c9943f923aa47ccdeb12511d86166a8ab56332623b9af453e4 +size 1514016442 diff --git a/checkpoint-2445/global_step2445/bf16_zero_pp_rank_28_mp_rank_00_optim_states.pt b/checkpoint-2445/global_step2445/bf16_zero_pp_rank_28_mp_rank_00_optim_states.pt new file mode 100644 index 0000000..81b40ff --- /dev/null +++ b/checkpoint-2445/global_step2445/bf16_zero_pp_rank_28_mp_rank_00_optim_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:3237122aef14287f7deeba8f0c3f000c16f0ad8f2db0aeb5c12b576774041daa +size 1514016442 diff --git a/checkpoint-2445/global_step2445/bf16_zero_pp_rank_29_mp_rank_00_optim_states.pt b/checkpoint-2445/global_step2445/bf16_zero_pp_rank_29_mp_rank_00_optim_states.pt new file mode 100644 index 0000000..6b9a06d --- /dev/null +++ b/checkpoint-2445/global_step2445/bf16_zero_pp_rank_29_mp_rank_00_optim_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:75b0b4beba51ee3ff7b9700f1ad3de6f3d931d2e1d9bcafdab1830222f1c2073 +size 1514016442 diff --git a/checkpoint-2445/global_step2445/bf16_zero_pp_rank_2_mp_rank_00_optim_states.pt b/checkpoint-2445/global_step2445/bf16_zero_pp_rank_2_mp_rank_00_optim_states.pt new file mode 100644 index 0000000..23c2937 --- /dev/null +++ b/checkpoint-2445/global_step2445/bf16_zero_pp_rank_2_mp_rank_00_optim_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:e697b57de47b0ec3f8754ea3297c108a6281f70fc35e2cfed245d86818a161c2 +size 1514015111 diff --git a/checkpoint-2445/global_step2445/bf16_zero_pp_rank_30_mp_rank_00_optim_states.pt b/checkpoint-2445/global_step2445/bf16_zero_pp_rank_30_mp_rank_00_optim_states.pt new file mode 100644 index 0000000..45d92dc --- /dev/null +++ b/checkpoint-2445/global_step2445/bf16_zero_pp_rank_30_mp_rank_00_optim_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:30c1e5064f651a74558e93b936a5e160db2aeabf5182857fb48d001b3490b5ba +size 1514016442 diff --git a/checkpoint-2445/global_step2445/bf16_zero_pp_rank_31_mp_rank_00_optim_states.pt b/checkpoint-2445/global_step2445/bf16_zero_pp_rank_31_mp_rank_00_optim_states.pt new file mode 100644 index 0000000..f0b0662 --- /dev/null +++ b/checkpoint-2445/global_step2445/bf16_zero_pp_rank_31_mp_rank_00_optim_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:3497160edba4271ba82153afb553e83d70b7155c01e49fea1669604f85199297 +size 1514016442 diff --git a/checkpoint-2445/global_step2445/bf16_zero_pp_rank_3_mp_rank_00_optim_states.pt b/checkpoint-2445/global_step2445/bf16_zero_pp_rank_3_mp_rank_00_optim_states.pt new file mode 100644 index 0000000..8361a9f --- /dev/null +++ b/checkpoint-2445/global_step2445/bf16_zero_pp_rank_3_mp_rank_00_optim_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:0c403979393817a89d42012d3836d84b320bdac23d27ce5b6d1e3d98c591eb3b +size 1514015111 diff --git a/checkpoint-2445/global_step2445/bf16_zero_pp_rank_4_mp_rank_00_optim_states.pt b/checkpoint-2445/global_step2445/bf16_zero_pp_rank_4_mp_rank_00_optim_states.pt new file mode 100644 index 0000000..7df72dc --- /dev/null +++ b/checkpoint-2445/global_step2445/bf16_zero_pp_rank_4_mp_rank_00_optim_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:3db5082a4e546bb475dd2ec28adf8f1a0d3be7ffbd489442f30163d3d1dca7e6 +size 1514015111 diff --git a/checkpoint-2445/global_step2445/bf16_zero_pp_rank_5_mp_rank_00_optim_states.pt b/checkpoint-2445/global_step2445/bf16_zero_pp_rank_5_mp_rank_00_optim_states.pt new file mode 100644 index 0000000..e0aa057 --- /dev/null +++ b/checkpoint-2445/global_step2445/bf16_zero_pp_rank_5_mp_rank_00_optim_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:9b4dfa44f039eb318d40a166189085af40109da9e887d7eb788d06ec8a782491 +size 1514015111 diff --git a/checkpoint-2445/global_step2445/bf16_zero_pp_rank_6_mp_rank_00_optim_states.pt b/checkpoint-2445/global_step2445/bf16_zero_pp_rank_6_mp_rank_00_optim_states.pt new file mode 100644 index 0000000..d6d64c0 --- /dev/null +++ b/checkpoint-2445/global_step2445/bf16_zero_pp_rank_6_mp_rank_00_optim_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:e7c3b2a76ca342ba6ed6fbd2b2a4fe0ed9ad8423d7c160f5bb82c9305790c874 +size 1514015111 diff --git a/checkpoint-2445/global_step2445/bf16_zero_pp_rank_7_mp_rank_00_optim_states.pt b/checkpoint-2445/global_step2445/bf16_zero_pp_rank_7_mp_rank_00_optim_states.pt new file mode 100644 index 0000000..b25612b --- /dev/null +++ b/checkpoint-2445/global_step2445/bf16_zero_pp_rank_7_mp_rank_00_optim_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:54e439425a054dbadc2b0b783f6095e595be3897e879f9c7149d0d91057e90d4 +size 1514015111 diff --git a/checkpoint-2445/global_step2445/bf16_zero_pp_rank_8_mp_rank_00_optim_states.pt b/checkpoint-2445/global_step2445/bf16_zero_pp_rank_8_mp_rank_00_optim_states.pt new file mode 100644 index 0000000..60e1fff --- /dev/null +++ b/checkpoint-2445/global_step2445/bf16_zero_pp_rank_8_mp_rank_00_optim_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:754195b694c1b4a8933f828b4bb86d8fbdb40d3ae0b8149799845df685908bf2 +size 1514015111 diff --git a/checkpoint-2445/global_step2445/bf16_zero_pp_rank_9_mp_rank_00_optim_states.pt b/checkpoint-2445/global_step2445/bf16_zero_pp_rank_9_mp_rank_00_optim_states.pt new file mode 100644 index 0000000..2c4e9e0 --- /dev/null +++ b/checkpoint-2445/global_step2445/bf16_zero_pp_rank_9_mp_rank_00_optim_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:ddfbe28d0e6cf7204c1a9c00cf9d7e0bbf3163e0c0de825774e47fa02e636922 +size 1514015111 diff --git a/checkpoint-2445/global_step2445/zero_pp_rank_0_mp_rank_00_model_states.pt b/checkpoint-2445/global_step2445/zero_pp_rank_0_mp_rank_00_model_states.pt new file mode 100644 index 0000000..b74ff49 --- /dev/null +++ b/checkpoint-2445/global_step2445/zero_pp_rank_0_mp_rank_00_model_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:ffe2ec38edbe0b5d6e6150e6e9c4874bad566b9a275825fae36b115dfb705507 +size 150245 diff --git a/checkpoint-2445/global_step2445/zero_pp_rank_10_mp_rank_00_model_states.pt b/checkpoint-2445/global_step2445/zero_pp_rank_10_mp_rank_00_model_states.pt new file mode 100644 index 0000000..9f111a7 --- /dev/null +++ b/checkpoint-2445/global_step2445/zero_pp_rank_10_mp_rank_00_model_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:c4e658b292c123f55b3d55c45e964bd863e856250b17842c831ffef0c85f16a1 +size 150540 diff --git a/checkpoint-2445/global_step2445/zero_pp_rank_11_mp_rank_00_model_states.pt b/checkpoint-2445/global_step2445/zero_pp_rank_11_mp_rank_00_model_states.pt new file mode 100644 index 0000000..06fe07b --- /dev/null +++ b/checkpoint-2445/global_step2445/zero_pp_rank_11_mp_rank_00_model_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:9fecdf063c4da0cf60d7f6b35ac9ea6a055abe701f722bc2c3d2bb7b044c2ea1 +size 150540 diff --git a/checkpoint-2445/global_step2445/zero_pp_rank_12_mp_rank_00_model_states.pt b/checkpoint-2445/global_step2445/zero_pp_rank_12_mp_rank_00_model_states.pt new file mode 100644 index 0000000..a7b1e94 --- /dev/null +++ b/checkpoint-2445/global_step2445/zero_pp_rank_12_mp_rank_00_model_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:bfc82216501ff83ea47d43f9f20b16bc0d56eb698d65763fa53af58261283b40 +size 150540 diff --git a/checkpoint-2445/global_step2445/zero_pp_rank_13_mp_rank_00_model_states.pt b/checkpoint-2445/global_step2445/zero_pp_rank_13_mp_rank_00_model_states.pt new file mode 100644 index 0000000..76a6636 --- /dev/null +++ b/checkpoint-2445/global_step2445/zero_pp_rank_13_mp_rank_00_model_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:d5ca14f3ce2aa6e8432025d9bcfae023f4028772903aedf27251a73ee89d9d9c +size 150540 diff --git a/checkpoint-2445/global_step2445/zero_pp_rank_14_mp_rank_00_model_states.pt b/checkpoint-2445/global_step2445/zero_pp_rank_14_mp_rank_00_model_states.pt new file mode 100644 index 0000000..1a1c1d9 --- /dev/null +++ b/checkpoint-2445/global_step2445/zero_pp_rank_14_mp_rank_00_model_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:b499c7332ae7a32106b346ae639177132aedc19afabb2738ca0e64bc3d7de2c1 +size 150540 diff --git a/checkpoint-2445/global_step2445/zero_pp_rank_15_mp_rank_00_model_states.pt b/checkpoint-2445/global_step2445/zero_pp_rank_15_mp_rank_00_model_states.pt new file mode 100644 index 0000000..cc19f64 --- /dev/null +++ b/checkpoint-2445/global_step2445/zero_pp_rank_15_mp_rank_00_model_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:329e41b68c60e83e871a716f6082a4be3c906e00da7ae73e051bd4337426c778 +size 150540 diff --git a/checkpoint-2445/global_step2445/zero_pp_rank_16_mp_rank_00_model_states.pt b/checkpoint-2445/global_step2445/zero_pp_rank_16_mp_rank_00_model_states.pt new file mode 100644 index 0000000..413b1e6 --- /dev/null +++ b/checkpoint-2445/global_step2445/zero_pp_rank_16_mp_rank_00_model_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:2100aeaab7b7a3f79e3a5f65f4b0921ccb135f81fa1e051446156c3a3bfd73bb +size 150540 diff --git a/checkpoint-2445/global_step2445/zero_pp_rank_17_mp_rank_00_model_states.pt b/checkpoint-2445/global_step2445/zero_pp_rank_17_mp_rank_00_model_states.pt new file mode 100644 index 0000000..0f2822c --- /dev/null +++ b/checkpoint-2445/global_step2445/zero_pp_rank_17_mp_rank_00_model_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:e2fbc8d92199cf5ea2edfdebf931db8adfe2a5f8765c18e9d8179778be59c539 +size 150540 diff --git a/checkpoint-2445/global_step2445/zero_pp_rank_18_mp_rank_00_model_states.pt b/checkpoint-2445/global_step2445/zero_pp_rank_18_mp_rank_00_model_states.pt new file mode 100644 index 0000000..7601315 --- /dev/null +++ b/checkpoint-2445/global_step2445/zero_pp_rank_18_mp_rank_00_model_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:3db66e4fe950e5c4b5246e87074224980c95347d1ba107cbfc86679f6d97445d +size 150540 diff --git a/checkpoint-2445/global_step2445/zero_pp_rank_19_mp_rank_00_model_states.pt b/checkpoint-2445/global_step2445/zero_pp_rank_19_mp_rank_00_model_states.pt new file mode 100644 index 0000000..85f488c --- /dev/null +++ b/checkpoint-2445/global_step2445/zero_pp_rank_19_mp_rank_00_model_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:8fb63cd572f258454684e6ca311b00501e7d86139f94138968c00257036fa155 +size 150540 diff --git a/checkpoint-2445/global_step2445/zero_pp_rank_1_mp_rank_00_model_states.pt b/checkpoint-2445/global_step2445/zero_pp_rank_1_mp_rank_00_model_states.pt new file mode 100644 index 0000000..4ae6595 --- /dev/null +++ b/checkpoint-2445/global_step2445/zero_pp_rank_1_mp_rank_00_model_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:bae542fcea121575e03aef13180259386db45d8338cfb8e170c1b1e936f07f56 +size 150245 diff --git a/checkpoint-2445/global_step2445/zero_pp_rank_20_mp_rank_00_model_states.pt b/checkpoint-2445/global_step2445/zero_pp_rank_20_mp_rank_00_model_states.pt new file mode 100644 index 0000000..7af3882 --- /dev/null +++ b/checkpoint-2445/global_step2445/zero_pp_rank_20_mp_rank_00_model_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:5868fd1a75aa1dd0d95ee8c620b62c2278b143828cbebe4eccd1407be9677f7b +size 150540 diff --git a/checkpoint-2445/global_step2445/zero_pp_rank_21_mp_rank_00_model_states.pt b/checkpoint-2445/global_step2445/zero_pp_rank_21_mp_rank_00_model_states.pt new file mode 100644 index 0000000..40ace16 --- /dev/null +++ b/checkpoint-2445/global_step2445/zero_pp_rank_21_mp_rank_00_model_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:9ddc82f8cf5a751cf0e3262640ed49766df68af472a6d4814e51dac88862d27d +size 150540 diff --git a/checkpoint-2445/global_step2445/zero_pp_rank_22_mp_rank_00_model_states.pt b/checkpoint-2445/global_step2445/zero_pp_rank_22_mp_rank_00_model_states.pt new file mode 100644 index 0000000..a4ae388 --- /dev/null +++ b/checkpoint-2445/global_step2445/zero_pp_rank_22_mp_rank_00_model_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:12f2337e40e00274ed590c1fe24b40149ea7e94cffbef40a7cbf47645389cb3f +size 150540 diff --git a/checkpoint-2445/global_step2445/zero_pp_rank_23_mp_rank_00_model_states.pt b/checkpoint-2445/global_step2445/zero_pp_rank_23_mp_rank_00_model_states.pt new file mode 100644 index 0000000..26d41ca --- /dev/null +++ b/checkpoint-2445/global_step2445/zero_pp_rank_23_mp_rank_00_model_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:3d4886a3dfdec0f0415d38692a739a753ab7e29dc88fa108e8b6fe42625ef192 +size 150540 diff --git a/checkpoint-2445/global_step2445/zero_pp_rank_24_mp_rank_00_model_states.pt b/checkpoint-2445/global_step2445/zero_pp_rank_24_mp_rank_00_model_states.pt new file mode 100644 index 0000000..1e1c670 --- /dev/null +++ b/checkpoint-2445/global_step2445/zero_pp_rank_24_mp_rank_00_model_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:a5fb4ecdb0b270ff0eced437c781d31b0cb014809d32917e168940877bfb03f4 +size 150540 diff --git a/checkpoint-2445/global_step2445/zero_pp_rank_25_mp_rank_00_model_states.pt b/checkpoint-2445/global_step2445/zero_pp_rank_25_mp_rank_00_model_states.pt new file mode 100644 index 0000000..f2774b4 --- /dev/null +++ b/checkpoint-2445/global_step2445/zero_pp_rank_25_mp_rank_00_model_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:0901e901fe596b3138d2ccc32204539f794ef6f279c6b42be27a40c895c787d3 +size 150540 diff --git a/checkpoint-2445/global_step2445/zero_pp_rank_26_mp_rank_00_model_states.pt b/checkpoint-2445/global_step2445/zero_pp_rank_26_mp_rank_00_model_states.pt new file mode 100644 index 0000000..7e7363b --- /dev/null +++ b/checkpoint-2445/global_step2445/zero_pp_rank_26_mp_rank_00_model_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:6f207ac4a816c43bf844c2c8c80644cbc46e16a4111a38c58827903cf3d4bcdd +size 150540 diff --git a/checkpoint-2445/global_step2445/zero_pp_rank_27_mp_rank_00_model_states.pt b/checkpoint-2445/global_step2445/zero_pp_rank_27_mp_rank_00_model_states.pt new file mode 100644 index 0000000..cbdd845 --- /dev/null +++ b/checkpoint-2445/global_step2445/zero_pp_rank_27_mp_rank_00_model_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:bd23381e14373662f70a9dce1c0eeb15924cb8077c967ddc0e6d1e1a572ce43c +size 150540 diff --git a/checkpoint-2445/global_step2445/zero_pp_rank_28_mp_rank_00_model_states.pt b/checkpoint-2445/global_step2445/zero_pp_rank_28_mp_rank_00_model_states.pt new file mode 100644 index 0000000..b4a85c8 --- /dev/null +++ b/checkpoint-2445/global_step2445/zero_pp_rank_28_mp_rank_00_model_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:059bf1ba37ee92ccca47187fded153a9af22ee7044a0cee33b17a8c25c31b2c1 +size 150540 diff --git a/checkpoint-2445/global_step2445/zero_pp_rank_29_mp_rank_00_model_states.pt b/checkpoint-2445/global_step2445/zero_pp_rank_29_mp_rank_00_model_states.pt new file mode 100644 index 0000000..387d79d --- /dev/null +++ b/checkpoint-2445/global_step2445/zero_pp_rank_29_mp_rank_00_model_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:f3f05ba702f76a781cc0cca9f4d3fab13ea3f597fc2ea91a5201fba24d0419b0 +size 150540 diff --git a/checkpoint-2445/global_step2445/zero_pp_rank_2_mp_rank_00_model_states.pt b/checkpoint-2445/global_step2445/zero_pp_rank_2_mp_rank_00_model_states.pt new file mode 100644 index 0000000..d0b24ac --- /dev/null +++ b/checkpoint-2445/global_step2445/zero_pp_rank_2_mp_rank_00_model_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:0bf6f1b5fd7f395e28561eb3210bc8f3fa0d6cb589439ed820880d6b6a48bd6b +size 150245 diff --git a/checkpoint-2445/global_step2445/zero_pp_rank_30_mp_rank_00_model_states.pt b/checkpoint-2445/global_step2445/zero_pp_rank_30_mp_rank_00_model_states.pt new file mode 100644 index 0000000..0f81b1e --- /dev/null +++ b/checkpoint-2445/global_step2445/zero_pp_rank_30_mp_rank_00_model_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:60601369222e025ec7ace86e4faa98f72f42033dbddedc18dcc9e4a05d45b2c3 +size 150540 diff --git a/checkpoint-2445/global_step2445/zero_pp_rank_31_mp_rank_00_model_states.pt b/checkpoint-2445/global_step2445/zero_pp_rank_31_mp_rank_00_model_states.pt new file mode 100644 index 0000000..94614c4 --- /dev/null +++ b/checkpoint-2445/global_step2445/zero_pp_rank_31_mp_rank_00_model_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:6cf71cecae2998dfe48359806530fde46071708a330bf5da58ac545c673467de +size 150540 diff --git a/checkpoint-2445/global_step2445/zero_pp_rank_3_mp_rank_00_model_states.pt b/checkpoint-2445/global_step2445/zero_pp_rank_3_mp_rank_00_model_states.pt new file mode 100644 index 0000000..ad35869 --- /dev/null +++ b/checkpoint-2445/global_step2445/zero_pp_rank_3_mp_rank_00_model_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:e70e74735403d741bd55d1ece22450b62be54ef37652769657aa404eed47d2ee +size 150245 diff --git a/checkpoint-2445/global_step2445/zero_pp_rank_4_mp_rank_00_model_states.pt b/checkpoint-2445/global_step2445/zero_pp_rank_4_mp_rank_00_model_states.pt new file mode 100644 index 0000000..64d925c --- /dev/null +++ b/checkpoint-2445/global_step2445/zero_pp_rank_4_mp_rank_00_model_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:4991f0ead2f78d5715727716ecf7bfe732ecfd7ab68f34f25f15fcc68c5a6267 +size 150245 diff --git a/checkpoint-2445/global_step2445/zero_pp_rank_5_mp_rank_00_model_states.pt b/checkpoint-2445/global_step2445/zero_pp_rank_5_mp_rank_00_model_states.pt new file mode 100644 index 0000000..cbcf333 --- /dev/null +++ b/checkpoint-2445/global_step2445/zero_pp_rank_5_mp_rank_00_model_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:c5497e5c0a1f39abf9e032be83d4142aaca609dcce04ee031ffb3d11fd1e3c1a +size 150245 diff --git a/checkpoint-2445/global_step2445/zero_pp_rank_6_mp_rank_00_model_states.pt b/checkpoint-2445/global_step2445/zero_pp_rank_6_mp_rank_00_model_states.pt new file mode 100644 index 0000000..fdd724e --- /dev/null +++ b/checkpoint-2445/global_step2445/zero_pp_rank_6_mp_rank_00_model_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:5b97d0b7bc65b7c10b9af30baaecc9e58d79b2cb0861ae27187b4b0b3ba2674b +size 150245 diff --git a/checkpoint-2445/global_step2445/zero_pp_rank_7_mp_rank_00_model_states.pt b/checkpoint-2445/global_step2445/zero_pp_rank_7_mp_rank_00_model_states.pt new file mode 100644 index 0000000..2b06dbe --- /dev/null +++ b/checkpoint-2445/global_step2445/zero_pp_rank_7_mp_rank_00_model_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:1e2de5673e2d582e3de81f7b93016335f067b70ac5717d7651cb85afa1756486 +size 150245 diff --git a/checkpoint-2445/global_step2445/zero_pp_rank_8_mp_rank_00_model_states.pt b/checkpoint-2445/global_step2445/zero_pp_rank_8_mp_rank_00_model_states.pt new file mode 100644 index 0000000..83d25ae --- /dev/null +++ b/checkpoint-2445/global_step2445/zero_pp_rank_8_mp_rank_00_model_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:d89c67316075467476c1eb25aa00637bc97f5c0b241d464e6ad18d5544dc314b +size 150245 diff --git a/checkpoint-2445/global_step2445/zero_pp_rank_9_mp_rank_00_model_states.pt b/checkpoint-2445/global_step2445/zero_pp_rank_9_mp_rank_00_model_states.pt new file mode 100644 index 0000000..ee1482a --- /dev/null +++ b/checkpoint-2445/global_step2445/zero_pp_rank_9_mp_rank_00_model_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:909fce6df9e30ff0e3b1913a1313d8414cff3352514e4d46eb424a280c594a27 +size 150245 diff --git a/checkpoint-2445/latest b/checkpoint-2445/latest new file mode 100644 index 0000000..5247c35 --- /dev/null +++ b/checkpoint-2445/latest @@ -0,0 +1 @@ +global_step2445 \ No newline at end of file diff --git a/checkpoint-2445/model-00001-of-00004.safetensors b/checkpoint-2445/model-00001-of-00004.safetensors new file mode 100644 index 0000000..e954b8d --- /dev/null +++ b/checkpoint-2445/model-00001-of-00004.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:0bd99094aa0d9de1c56cec4a5cfe8c2278dd5fa844cd5eb17a98f4eb663b9a9d +size 4976698672 diff --git a/checkpoint-2445/model-00002-of-00004.safetensors b/checkpoint-2445/model-00002-of-00004.safetensors new file mode 100644 index 0000000..479cac7 --- /dev/null +++ b/checkpoint-2445/model-00002-of-00004.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:26f9f2f17af25c67509c30d8c79c3e9bca63682789638a940b9ad429db428593 +size 4999802720 diff --git a/checkpoint-2445/model-00003-of-00004.safetensors b/checkpoint-2445/model-00003-of-00004.safetensors new file mode 100644 index 0000000..725f564 --- /dev/null +++ b/checkpoint-2445/model-00003-of-00004.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:6207337975b4b7d327d6135e93cde49007c00a5c630eadb2a8ce685f83b431a2 +size 4915916176 diff --git a/checkpoint-2445/model-00004-of-00004.safetensors b/checkpoint-2445/model-00004-of-00004.safetensors new file mode 100644 index 0000000..b4c25ca --- /dev/null +++ b/checkpoint-2445/model-00004-of-00004.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:c2e5860979ea435ace5595020a69b0f0a26d819b061ba8455b510448b4226ba1 +size 1168138808 diff --git a/checkpoint-2445/model.safetensors.index.json b/checkpoint-2445/model.safetensors.index.json new file mode 100644 index 0000000..0fd8120 --- /dev/null +++ b/checkpoint-2445/model.safetensors.index.json @@ -0,0 +1,298 @@ +{ + "metadata": { + "total_size": 16060522496 + }, + "weight_map": { + "lm_head.weight": "model-00004-of-00004.safetensors", + "model.embed_tokens.weight": "model-00001-of-00004.safetensors", + "model.layers.0.input_layernorm.weight": "model-00001-of-00004.safetensors", + "model.layers.0.mlp.down_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.0.mlp.gate_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.0.mlp.up_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.0.post_attention_layernorm.weight": "model-00001-of-00004.safetensors", + "model.layers.0.self_attn.k_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.0.self_attn.o_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.0.self_attn.q_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.0.self_attn.v_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.1.input_layernorm.weight": "model-00001-of-00004.safetensors", + "model.layers.1.mlp.down_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.1.mlp.gate_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.1.mlp.up_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.1.post_attention_layernorm.weight": "model-00001-of-00004.safetensors", + "model.layers.1.self_attn.k_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.1.self_attn.o_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.1.self_attn.q_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.1.self_attn.v_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.10.input_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.10.mlp.down_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.10.mlp.gate_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.10.mlp.up_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.10.post_attention_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.10.self_attn.k_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.10.self_attn.o_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.10.self_attn.q_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.10.self_attn.v_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.11.input_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.11.mlp.down_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.11.mlp.gate_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.11.mlp.up_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.11.post_attention_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.11.self_attn.k_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.11.self_attn.o_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.11.self_attn.q_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.11.self_attn.v_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.12.input_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.12.mlp.down_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.12.mlp.gate_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.12.mlp.up_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.12.post_attention_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.12.self_attn.k_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.12.self_attn.o_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.12.self_attn.q_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.12.self_attn.v_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.13.input_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.13.mlp.down_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.13.mlp.gate_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.13.mlp.up_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.13.post_attention_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.13.self_attn.k_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.13.self_attn.o_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.13.self_attn.q_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.13.self_attn.v_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.14.input_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.14.mlp.down_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.14.mlp.gate_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.14.mlp.up_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.14.post_attention_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.14.self_attn.k_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.14.self_attn.o_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.14.self_attn.q_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.14.self_attn.v_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.15.input_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.15.mlp.down_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.15.mlp.gate_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.15.mlp.up_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.15.post_attention_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.15.self_attn.k_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.15.self_attn.o_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.15.self_attn.q_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.15.self_attn.v_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.16.input_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.16.mlp.down_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.16.mlp.gate_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.16.mlp.up_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.16.post_attention_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.16.self_attn.k_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.16.self_attn.o_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.16.self_attn.q_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.16.self_attn.v_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.17.input_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.17.mlp.down_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.17.mlp.gate_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.17.mlp.up_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.17.post_attention_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.17.self_attn.k_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.17.self_attn.o_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.17.self_attn.q_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.17.self_attn.v_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.18.input_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.18.mlp.down_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.18.mlp.gate_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.18.mlp.up_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.18.post_attention_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.18.self_attn.k_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.18.self_attn.o_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.18.self_attn.q_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.18.self_attn.v_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.19.input_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.19.mlp.down_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.19.mlp.gate_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.19.mlp.up_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.19.post_attention_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.19.self_attn.k_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.19.self_attn.o_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.19.self_attn.q_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.19.self_attn.v_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.2.input_layernorm.weight": "model-00001-of-00004.safetensors", + "model.layers.2.mlp.down_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.2.mlp.gate_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.2.mlp.up_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.2.post_attention_layernorm.weight": "model-00001-of-00004.safetensors", + "model.layers.2.self_attn.k_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.2.self_attn.o_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.2.self_attn.q_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.2.self_attn.v_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.20.input_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.20.mlp.down_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.20.mlp.gate_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.20.mlp.up_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.20.post_attention_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.20.self_attn.k_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.20.self_attn.o_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.20.self_attn.q_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.20.self_attn.v_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.21.input_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.21.mlp.down_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.21.mlp.gate_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.21.mlp.up_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.21.post_attention_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.21.self_attn.k_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.21.self_attn.o_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.21.self_attn.q_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.21.self_attn.v_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.22.input_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.22.mlp.down_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.22.mlp.gate_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.22.mlp.up_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.22.post_attention_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.22.self_attn.k_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.22.self_attn.o_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.22.self_attn.q_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.22.self_attn.v_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.23.input_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.23.mlp.down_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.23.mlp.gate_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.23.mlp.up_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.23.post_attention_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.23.self_attn.k_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.23.self_attn.o_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.23.self_attn.q_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.23.self_attn.v_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.24.input_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.24.mlp.down_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.24.mlp.gate_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.24.mlp.up_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.24.post_attention_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.24.self_attn.k_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.24.self_attn.o_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.24.self_attn.q_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.24.self_attn.v_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.25.input_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.25.mlp.down_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.25.mlp.gate_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.25.mlp.up_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.25.post_attention_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.25.self_attn.k_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.25.self_attn.o_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.25.self_attn.q_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.25.self_attn.v_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.26.input_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.26.mlp.down_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.26.mlp.gate_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.26.mlp.up_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.26.post_attention_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.26.self_attn.k_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.26.self_attn.o_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.26.self_attn.q_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.26.self_attn.v_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.27.input_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.27.mlp.down_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.27.mlp.gate_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.27.mlp.up_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.27.post_attention_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.27.self_attn.k_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.27.self_attn.o_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.27.self_attn.q_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.27.self_attn.v_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.28.input_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.28.mlp.down_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.28.mlp.gate_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.28.mlp.up_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.28.post_attention_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.28.self_attn.k_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.28.self_attn.o_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.28.self_attn.q_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.28.self_attn.v_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.29.input_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.29.mlp.down_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.29.mlp.gate_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.29.mlp.up_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.29.post_attention_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.29.self_attn.k_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.29.self_attn.o_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.29.self_attn.q_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.29.self_attn.v_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.3.input_layernorm.weight": "model-00001-of-00004.safetensors", + "model.layers.3.mlp.down_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.3.mlp.gate_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.3.mlp.up_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.3.post_attention_layernorm.weight": "model-00001-of-00004.safetensors", + "model.layers.3.self_attn.k_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.3.self_attn.o_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.3.self_attn.q_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.3.self_attn.v_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.30.input_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.30.mlp.down_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.30.mlp.gate_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.30.mlp.up_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.30.post_attention_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.30.self_attn.k_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.30.self_attn.o_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.30.self_attn.q_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.30.self_attn.v_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.31.input_layernorm.weight": "model-00004-of-00004.safetensors", + "model.layers.31.mlp.down_proj.weight": "model-00004-of-00004.safetensors", + "model.layers.31.mlp.gate_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.31.mlp.up_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.31.post_attention_layernorm.weight": "model-00004-of-00004.safetensors", + "model.layers.31.self_attn.k_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.31.self_attn.o_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.31.self_attn.q_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.31.self_attn.v_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.4.input_layernorm.weight": "model-00001-of-00004.safetensors", + "model.layers.4.mlp.down_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.4.mlp.gate_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.4.mlp.up_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.4.post_attention_layernorm.weight": "model-00001-of-00004.safetensors", + "model.layers.4.self_attn.k_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.4.self_attn.o_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.4.self_attn.q_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.4.self_attn.v_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.5.input_layernorm.weight": "model-00001-of-00004.safetensors", + "model.layers.5.mlp.down_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.5.mlp.gate_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.5.mlp.up_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.5.post_attention_layernorm.weight": "model-00001-of-00004.safetensors", + "model.layers.5.self_attn.k_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.5.self_attn.o_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.5.self_attn.q_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.5.self_attn.v_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.6.input_layernorm.weight": "model-00001-of-00004.safetensors", + "model.layers.6.mlp.down_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.6.mlp.gate_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.6.mlp.up_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.6.post_attention_layernorm.weight": "model-00001-of-00004.safetensors", + "model.layers.6.self_attn.k_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.6.self_attn.o_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.6.self_attn.q_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.6.self_attn.v_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.7.input_layernorm.weight": "model-00001-of-00004.safetensors", + "model.layers.7.mlp.down_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.7.mlp.gate_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.7.mlp.up_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.7.post_attention_layernorm.weight": "model-00001-of-00004.safetensors", + "model.layers.7.self_attn.k_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.7.self_attn.o_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.7.self_attn.q_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.7.self_attn.v_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.8.input_layernorm.weight": "model-00001-of-00004.safetensors", + "model.layers.8.mlp.down_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.8.mlp.gate_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.8.mlp.up_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.8.post_attention_layernorm.weight": "model-00001-of-00004.safetensors", + "model.layers.8.self_attn.k_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.8.self_attn.o_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.8.self_attn.q_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.8.self_attn.v_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.9.input_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.9.mlp.down_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.9.mlp.gate_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.9.mlp.up_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.9.post_attention_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.9.self_attn.k_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.9.self_attn.o_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.9.self_attn.q_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.9.self_attn.v_proj.weight": "model-00002-of-00004.safetensors", + "model.norm.weight": "model-00004-of-00004.safetensors" + } +} diff --git a/checkpoint-2445/rng_state_0.pth b/checkpoint-2445/rng_state_0.pth new file mode 100644 index 0000000..3525225 --- /dev/null +++ b/checkpoint-2445/rng_state_0.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:08282b46825aa78d10fe10e3fea89555c5b5a691b261a3ddfd58fcb58370edff +size 15984 diff --git a/checkpoint-2445/rng_state_1.pth b/checkpoint-2445/rng_state_1.pth new file mode 100644 index 0000000..eeb7d8d --- /dev/null +++ b/checkpoint-2445/rng_state_1.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:dbab71d98a3a9a92df82a6bba463947327c3a1bcf35cd9f4f46114641fc42dd9 +size 15984 diff --git a/checkpoint-2445/rng_state_10.pth b/checkpoint-2445/rng_state_10.pth new file mode 100644 index 0000000..338c1db --- /dev/null +++ b/checkpoint-2445/rng_state_10.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:54e5c2d65c14df39137f6a3bd8314f534dad93d72efe8e40e23a61187ba74ce8 +size 15997 diff --git a/checkpoint-2445/rng_state_11.pth b/checkpoint-2445/rng_state_11.pth new file mode 100644 index 0000000..4bbb9bf --- /dev/null +++ b/checkpoint-2445/rng_state_11.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:f578146bcaaed333cf4637aa3b454e2f6238e691e45b4160001b63b4f8b5ce4e +size 15997 diff --git a/checkpoint-2445/rng_state_12.pth b/checkpoint-2445/rng_state_12.pth new file mode 100644 index 0000000..82f488f --- /dev/null +++ b/checkpoint-2445/rng_state_12.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:8463662162a837871bb50af43a407175ca95f3f364bc13c7d065e8b6929bad11 +size 15997 diff --git a/checkpoint-2445/rng_state_13.pth b/checkpoint-2445/rng_state_13.pth new file mode 100644 index 0000000..1b4dd63 --- /dev/null +++ b/checkpoint-2445/rng_state_13.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:fe9bafd75244b907f13aa5b102963f4d018d7b80b94ddb0262313774362d9305 +size 15997 diff --git a/checkpoint-2445/rng_state_14.pth b/checkpoint-2445/rng_state_14.pth new file mode 100644 index 0000000..b7f3c97 --- /dev/null +++ b/checkpoint-2445/rng_state_14.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:c5e353c1ca54c464e3464d695507da7da6c134e4ad9d6acab6f72e52d5bf13c7 +size 15997 diff --git a/checkpoint-2445/rng_state_15.pth b/checkpoint-2445/rng_state_15.pth new file mode 100644 index 0000000..8e4dee4 --- /dev/null +++ b/checkpoint-2445/rng_state_15.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:24f1ef20cfc904ac8f811ceffab91f00cccadca0e878c8ef8a7335e83ef9513f +size 15997 diff --git a/checkpoint-2445/rng_state_16.pth b/checkpoint-2445/rng_state_16.pth new file mode 100644 index 0000000..4033ee0 --- /dev/null +++ b/checkpoint-2445/rng_state_16.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:cecf995eb1b13a85a673384ed67c75acc989e9d373ac7da2d90cbc7af2c1b4a7 +size 15997 diff --git a/checkpoint-2445/rng_state_17.pth b/checkpoint-2445/rng_state_17.pth new file mode 100644 index 0000000..0a50e02 --- /dev/null +++ b/checkpoint-2445/rng_state_17.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:375986c2a348f319aeed3322fcf23be7787f801c585046ff217d7855463fbc0d +size 15997 diff --git a/checkpoint-2445/rng_state_18.pth b/checkpoint-2445/rng_state_18.pth new file mode 100644 index 0000000..075e295 --- /dev/null +++ b/checkpoint-2445/rng_state_18.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:a50dfdcb902d78e55850cabc410552a4e651090792c2e086926cbda383b4f35f +size 15997 diff --git a/checkpoint-2445/rng_state_19.pth b/checkpoint-2445/rng_state_19.pth new file mode 100644 index 0000000..ac69e75 --- /dev/null +++ b/checkpoint-2445/rng_state_19.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:79850d41cfed8d83ef96222b2b956a0ec273b7fa78630d3cb0e86f2ab29bb934 +size 15997 diff --git a/checkpoint-2445/rng_state_2.pth b/checkpoint-2445/rng_state_2.pth new file mode 100644 index 0000000..e144a44 --- /dev/null +++ b/checkpoint-2445/rng_state_2.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:caac82d57d878d30219a4f9ec289a97ff90c53afc160b968f251b3fd3454b8d8 +size 15984 diff --git a/checkpoint-2445/rng_state_20.pth b/checkpoint-2445/rng_state_20.pth new file mode 100644 index 0000000..f4588d3 --- /dev/null +++ b/checkpoint-2445/rng_state_20.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:60201c3050f5fddd8ba67c9cc8ea589b8296872acef041a46972a75489ba48f2 +size 15997 diff --git a/checkpoint-2445/rng_state_21.pth b/checkpoint-2445/rng_state_21.pth new file mode 100644 index 0000000..94a9f24 --- /dev/null +++ b/checkpoint-2445/rng_state_21.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:9b45ccf4c1d4fa82ab73d5b1084c8895bb7a843bb65c39bb0b2bb2a6192e6df3 +size 15997 diff --git a/checkpoint-2445/rng_state_22.pth b/checkpoint-2445/rng_state_22.pth new file mode 100644 index 0000000..7d803e4 --- /dev/null +++ b/checkpoint-2445/rng_state_22.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:54521e12e4ad123a5cdba9d3c092e63ee37b099ff802bbe417dfb47861244ab9 +size 15997 diff --git a/checkpoint-2445/rng_state_23.pth b/checkpoint-2445/rng_state_23.pth new file mode 100644 index 0000000..10c1a68 --- /dev/null +++ b/checkpoint-2445/rng_state_23.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:f4917fcd537e822a9012f654e3131d5dbd358e1916dec368f464d89c1223a2e5 +size 15997 diff --git a/checkpoint-2445/rng_state_24.pth b/checkpoint-2445/rng_state_24.pth new file mode 100644 index 0000000..67c352c --- /dev/null +++ b/checkpoint-2445/rng_state_24.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:e6b078914d8fc5b3568727c1679a97889d10f35460aaa0147c1ab343e0a8576f +size 15997 diff --git a/checkpoint-2445/rng_state_25.pth b/checkpoint-2445/rng_state_25.pth new file mode 100644 index 0000000..4e42fe3 --- /dev/null +++ b/checkpoint-2445/rng_state_25.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:58235624c008372a4489d1dc4d4a5d582e83b5954529d532a8ac9e27cc1117eb +size 15997 diff --git a/checkpoint-2445/rng_state_26.pth b/checkpoint-2445/rng_state_26.pth new file mode 100644 index 0000000..58fc36f --- /dev/null +++ b/checkpoint-2445/rng_state_26.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:6d0a9e394ce0c35df215df576763794e20589536fd7d2e8aae0d2bfaac14eb95 +size 15997 diff --git a/checkpoint-2445/rng_state_27.pth b/checkpoint-2445/rng_state_27.pth new file mode 100644 index 0000000..0e9b2ee --- /dev/null +++ b/checkpoint-2445/rng_state_27.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:5a40fc515e550756e1b3497c9e62fa179918e76adb87ab3233db714b992268c8 +size 15997 diff --git a/checkpoint-2445/rng_state_28.pth b/checkpoint-2445/rng_state_28.pth new file mode 100644 index 0000000..befeed4 --- /dev/null +++ b/checkpoint-2445/rng_state_28.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:6dadf1f4ae507c5e5ab940441273a3552a7e59fcc516e7dfddd1956d1c5c5a3d +size 15997 diff --git a/checkpoint-2445/rng_state_29.pth b/checkpoint-2445/rng_state_29.pth new file mode 100644 index 0000000..dfca012 --- /dev/null +++ b/checkpoint-2445/rng_state_29.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:73dfc967a343045291988f3bcbbdee93e0877bb869db48fd675e9946c2c43507 +size 15997 diff --git a/checkpoint-2445/rng_state_3.pth b/checkpoint-2445/rng_state_3.pth new file mode 100644 index 0000000..a10f352 --- /dev/null +++ b/checkpoint-2445/rng_state_3.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:19762d2d370222b01817da11bbaa6665d542293373186d66f754e7246bb861ed +size 15984 diff --git a/checkpoint-2445/rng_state_30.pth b/checkpoint-2445/rng_state_30.pth new file mode 100644 index 0000000..a6f6699 --- /dev/null +++ b/checkpoint-2445/rng_state_30.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:6890d56ed291634fd39a3901c77e10c249a355684a11b3e34af07a400ff9189c +size 15997 diff --git a/checkpoint-2445/rng_state_31.pth b/checkpoint-2445/rng_state_31.pth new file mode 100644 index 0000000..5c916cb --- /dev/null +++ b/checkpoint-2445/rng_state_31.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:25d52ac86e2e15b3ec76429e008562b498d6431769f7f3ad24c182d032f7701b +size 15997 diff --git a/checkpoint-2445/rng_state_4.pth b/checkpoint-2445/rng_state_4.pth new file mode 100644 index 0000000..e6ef215 --- /dev/null +++ b/checkpoint-2445/rng_state_4.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:00c7508b346a7d3c5c23392845f1d013331114ade778794b76e919cb3ed5d33e +size 15984 diff --git a/checkpoint-2445/rng_state_5.pth b/checkpoint-2445/rng_state_5.pth new file mode 100644 index 0000000..126662e --- /dev/null +++ b/checkpoint-2445/rng_state_5.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:b89de7d14dd20a191f56b74c816ef8b7fe5c171e31efbeadbf321c4539ed68c3 +size 15984 diff --git a/checkpoint-2445/rng_state_6.pth b/checkpoint-2445/rng_state_6.pth new file mode 100644 index 0000000..d4e6b27 --- /dev/null +++ b/checkpoint-2445/rng_state_6.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:1c71152053553e6e22d670fbc4fd7550bf8a046b54cad7b71869787986a6a42c +size 15984 diff --git a/checkpoint-2445/rng_state_7.pth b/checkpoint-2445/rng_state_7.pth new file mode 100644 index 0000000..14e214a --- /dev/null +++ b/checkpoint-2445/rng_state_7.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:4b67db12a26a26ffe03d9afc84a43857eb2e5b2fec2dd189653b415f74208190 +size 15984 diff --git a/checkpoint-2445/rng_state_8.pth b/checkpoint-2445/rng_state_8.pth new file mode 100644 index 0000000..22cf6c8 --- /dev/null +++ b/checkpoint-2445/rng_state_8.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:50d06b29f11da5c2e4a388e6620296a543a68f4d46bb2bb978188a8e374f8925 +size 15984 diff --git a/checkpoint-2445/rng_state_9.pth b/checkpoint-2445/rng_state_9.pth new file mode 100644 index 0000000..a4fd554 --- /dev/null +++ b/checkpoint-2445/rng_state_9.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:bde660db671106d0d18882a4db0faf4fe72837df15f5537708435e1d684a549e +size 15984 diff --git a/checkpoint-2445/scheduler.pt b/checkpoint-2445/scheduler.pt new file mode 100644 index 0000000..f6fbac9 --- /dev/null +++ b/checkpoint-2445/scheduler.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:4acc52b35683a6241808314fd968e9578d4c370cd73326a54dd5d23bbee4bc52 +size 1064 diff --git a/checkpoint-2445/special_tokens_map.json b/checkpoint-2445/special_tokens_map.json new file mode 100644 index 0000000..278b7f0 --- /dev/null +++ b/checkpoint-2445/special_tokens_map.json @@ -0,0 +1,23 @@ +{ + "bos_token": { + "content": "<|begin_of_text|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false + }, + "eos_token": { + "content": "<|eot_id|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false + }, + "pad_token": { + "content": "<|end_of_text|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false + } +} diff --git a/checkpoint-2445/tokenizer.json b/checkpoint-2445/tokenizer.json new file mode 100644 index 0000000..1c1d8d5 --- /dev/null +++ b/checkpoint-2445/tokenizer.json @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:6b9e4e7fb171f92fd137b777cc2714bf87d11576700a1dcd7a399e7bbe39537b +size 17209920 diff --git a/checkpoint-2445/tokenizer_config.json b/checkpoint-2445/tokenizer_config.json new file mode 100644 index 0000000..8098161 --- /dev/null +++ b/checkpoint-2445/tokenizer_config.json @@ -0,0 +1,2064 @@ +{ + "added_tokens_decoder": { + "128000": { + "content": "<|begin_of_text|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128001": { + "content": "<|end_of_text|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128002": { + "content": "<|reserved_special_token_0|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128003": { + "content": "<|reserved_special_token_1|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128004": { + "content": "<|finetune_right_pad_id|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128005": { + "content": "<|reserved_special_token_2|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128006": { + "content": "<|start_header_id|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128007": { + "content": "<|end_header_id|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128008": { + "content": "<|eom_id|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128009": { + "content": "<|eot_id|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128010": { + "content": "<|python_tag|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128011": { + "content": "<|reserved_special_token_3|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128012": { + "content": "<|reserved_special_token_4|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128013": { + "content": "<|reserved_special_token_5|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128014": { + "content": "<|reserved_special_token_6|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128015": { + "content": "<|reserved_special_token_7|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128016": { + "content": "<|reserved_special_token_8|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128017": { + "content": "<|reserved_special_token_9|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128018": { + "content": "<|reserved_special_token_10|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128019": { + "content": "<|reserved_special_token_11|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128020": { + "content": "<|reserved_special_token_12|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128021": { + "content": "<|reserved_special_token_13|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128022": { + "content": "<|reserved_special_token_14|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128023": { + "content": "<|reserved_special_token_15|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128024": { + "content": "<|reserved_special_token_16|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128025": { + "content": "<|reserved_special_token_17|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128026": { + "content": "<|reserved_special_token_18|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128027": { + "content": "<|reserved_special_token_19|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128028": { + "content": "<|reserved_special_token_20|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128029": { + "content": "<|reserved_special_token_21|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128030": { + "content": "<|reserved_special_token_22|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128031": { + "content": "<|reserved_special_token_23|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128032": { + "content": "<|reserved_special_token_24|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128033": { + "content": "<|reserved_special_token_25|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128034": { + "content": "<|reserved_special_token_26|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128035": { + "content": "<|reserved_special_token_27|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128036": { + "content": "<|reserved_special_token_28|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128037": { + "content": "<|reserved_special_token_29|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128038": { + "content": "<|reserved_special_token_30|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128039": { + "content": "<|reserved_special_token_31|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128040": { + "content": "<|reserved_special_token_32|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128041": { + "content": "<|reserved_special_token_33|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128042": { + "content": "<|reserved_special_token_34|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128043": { + "content": "<|reserved_special_token_35|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128044": { + "content": "<|reserved_special_token_36|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128045": { + "content": "<|reserved_special_token_37|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128046": { + "content": "<|reserved_special_token_38|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128047": { + "content": "<|reserved_special_token_39|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128048": { + "content": "<|reserved_special_token_40|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128049": { + "content": "<|reserved_special_token_41|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128050": { + "content": "<|reserved_special_token_42|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128051": { + "content": "<|reserved_special_token_43|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128052": { + "content": "<|reserved_special_token_44|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128053": { + "content": "<|reserved_special_token_45|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128054": { + "content": "<|reserved_special_token_46|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128055": { + "content": "<|reserved_special_token_47|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128056": { + "content": "<|reserved_special_token_48|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128057": { + "content": "<|reserved_special_token_49|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128058": { + "content": "<|reserved_special_token_50|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128059": { + "content": "<|reserved_special_token_51|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128060": { + "content": "<|reserved_special_token_52|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128061": { + "content": "<|reserved_special_token_53|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128062": { + "content": "<|reserved_special_token_54|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128063": { + "content": "<|reserved_special_token_55|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128064": { + "content": "<|reserved_special_token_56|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128065": { + "content": "<|reserved_special_token_57|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128066": { + "content": "<|reserved_special_token_58|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128067": { + "content": "<|reserved_special_token_59|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128068": { + "content": "<|reserved_special_token_60|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128069": { + "content": "<|reserved_special_token_61|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128070": { + "content": "<|reserved_special_token_62|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128071": { + "content": "<|reserved_special_token_63|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128072": { + "content": "<|reserved_special_token_64|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128073": { + "content": "<|reserved_special_token_65|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128074": { + "content": "<|reserved_special_token_66|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128075": { + "content": "<|reserved_special_token_67|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128076": { + "content": "<|reserved_special_token_68|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128077": { + "content": "<|reserved_special_token_69|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128078": { + "content": "<|reserved_special_token_70|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128079": { + "content": "<|reserved_special_token_71|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128080": { + "content": "<|reserved_special_token_72|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128081": { + "content": "<|reserved_special_token_73|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128082": { + "content": "<|reserved_special_token_74|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128083": { + "content": "<|reserved_special_token_75|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128084": { + "content": "<|reserved_special_token_76|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128085": { + "content": "<|reserved_special_token_77|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128086": { + "content": "<|reserved_special_token_78|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128087": { + "content": "<|reserved_special_token_79|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128088": { + "content": "<|reserved_special_token_80|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128089": { + "content": "<|reserved_special_token_81|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128090": { + "content": "<|reserved_special_token_82|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128091": { + "content": "<|reserved_special_token_83|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128092": { + "content": "<|reserved_special_token_84|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128093": { + "content": "<|reserved_special_token_85|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128094": { + "content": "<|reserved_special_token_86|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128095": { + "content": "<|reserved_special_token_87|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128096": { + "content": "<|reserved_special_token_88|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128097": { + "content": "<|reserved_special_token_89|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128098": { + "content": "<|reserved_special_token_90|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128099": { + "content": "<|reserved_special_token_91|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128100": { + "content": "<|reserved_special_token_92|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128101": { + "content": "<|reserved_special_token_93|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128102": { + "content": "<|reserved_special_token_94|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128103": { + "content": "<|reserved_special_token_95|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128104": { + "content": "<|reserved_special_token_96|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128105": { + "content": "<|reserved_special_token_97|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128106": { + "content": "<|reserved_special_token_98|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128107": { + "content": "<|reserved_special_token_99|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128108": { + "content": "<|reserved_special_token_100|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128109": { + "content": "<|reserved_special_token_101|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128110": { + "content": "<|reserved_special_token_102|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128111": { + "content": "<|reserved_special_token_103|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128112": { + "content": "<|reserved_special_token_104|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128113": { + "content": "<|reserved_special_token_105|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128114": { + "content": "<|reserved_special_token_106|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128115": { + "content": "<|reserved_special_token_107|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128116": { + "content": "<|reserved_special_token_108|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128117": { + "content": "<|reserved_special_token_109|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128118": { + "content": "<|reserved_special_token_110|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128119": { + "content": "<|reserved_special_token_111|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128120": { + "content": "<|reserved_special_token_112|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128121": { + "content": "<|reserved_special_token_113|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128122": { + "content": "<|reserved_special_token_114|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128123": { + "content": "<|reserved_special_token_115|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128124": { + "content": "<|reserved_special_token_116|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128125": { + "content": "<|reserved_special_token_117|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128126": { + "content": "<|reserved_special_token_118|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128127": { + "content": "<|reserved_special_token_119|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128128": { + "content": "<|reserved_special_token_120|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128129": { + "content": "<|reserved_special_token_121|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128130": { + "content": "<|reserved_special_token_122|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128131": { + "content": "<|reserved_special_token_123|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128132": { + "content": "<|reserved_special_token_124|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128133": { + "content": "<|reserved_special_token_125|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128134": { + "content": "<|reserved_special_token_126|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128135": { + "content": "<|reserved_special_token_127|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128136": { + "content": "<|reserved_special_token_128|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128137": { + "content": "<|reserved_special_token_129|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128138": { + "content": "<|reserved_special_token_130|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128139": { + "content": "<|reserved_special_token_131|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128140": { + "content": "<|reserved_special_token_132|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128141": { + "content": "<|reserved_special_token_133|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128142": { + "content": "<|reserved_special_token_134|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128143": { + "content": "<|reserved_special_token_135|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128144": { + "content": "<|reserved_special_token_136|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128145": { + "content": "<|reserved_special_token_137|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128146": { + "content": "<|reserved_special_token_138|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128147": { + "content": "<|reserved_special_token_139|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128148": { + "content": "<|reserved_special_token_140|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128149": { + "content": "<|reserved_special_token_141|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128150": { + "content": "<|reserved_special_token_142|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128151": { + "content": "<|reserved_special_token_143|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128152": { + "content": "<|reserved_special_token_144|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128153": { + "content": "<|reserved_special_token_145|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128154": { + "content": "<|reserved_special_token_146|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128155": { + "content": "<|reserved_special_token_147|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128156": { + "content": "<|reserved_special_token_148|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128157": { + "content": "<|reserved_special_token_149|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128158": { + "content": "<|reserved_special_token_150|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128159": { + "content": "<|reserved_special_token_151|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128160": { + "content": "<|reserved_special_token_152|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128161": { + "content": "<|reserved_special_token_153|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128162": { + "content": "<|reserved_special_token_154|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128163": { + "content": "<|reserved_special_token_155|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128164": { + "content": "<|reserved_special_token_156|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128165": { + "content": "<|reserved_special_token_157|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128166": { + "content": "<|reserved_special_token_158|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128167": { + "content": "<|reserved_special_token_159|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128168": { + "content": "<|reserved_special_token_160|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128169": { + "content": "<|reserved_special_token_161|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128170": { + "content": "<|reserved_special_token_162|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128171": { + "content": "<|reserved_special_token_163|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128172": { + "content": "<|reserved_special_token_164|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128173": { + "content": "<|reserved_special_token_165|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128174": { + "content": "<|reserved_special_token_166|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128175": { + "content": "<|reserved_special_token_167|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128176": { + "content": "<|reserved_special_token_168|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128177": { + "content": "<|reserved_special_token_169|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128178": { + "content": "<|reserved_special_token_170|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128179": { + "content": "<|reserved_special_token_171|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128180": { + "content": "<|reserved_special_token_172|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128181": { + "content": "<|reserved_special_token_173|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128182": { + "content": "<|reserved_special_token_174|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128183": { + "content": "<|reserved_special_token_175|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128184": { + "content": "<|reserved_special_token_176|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128185": { + "content": "<|reserved_special_token_177|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128186": { + "content": "<|reserved_special_token_178|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128187": { + "content": "<|reserved_special_token_179|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128188": { + "content": "<|reserved_special_token_180|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128189": { + "content": "<|reserved_special_token_181|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128190": { + "content": "<|reserved_special_token_182|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128191": { + "content": "<|reserved_special_token_183|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128192": { + "content": "<|reserved_special_token_184|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128193": { + "content": "<|reserved_special_token_185|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128194": { + "content": "<|reserved_special_token_186|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128195": { + "content": "<|reserved_special_token_187|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128196": { + "content": "<|reserved_special_token_188|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128197": { + "content": "<|reserved_special_token_189|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128198": { + "content": "<|reserved_special_token_190|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128199": { + "content": "<|reserved_special_token_191|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128200": { + "content": "<|reserved_special_token_192|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128201": { + "content": "<|reserved_special_token_193|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128202": { + "content": "<|reserved_special_token_194|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128203": { + "content": "<|reserved_special_token_195|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128204": { + "content": "<|reserved_special_token_196|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128205": { + "content": "<|reserved_special_token_197|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128206": { + "content": "<|reserved_special_token_198|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128207": { + "content": "<|reserved_special_token_199|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128208": { + "content": "<|reserved_special_token_200|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128209": { + "content": "<|reserved_special_token_201|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128210": { + "content": "<|reserved_special_token_202|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128211": { + "content": "<|reserved_special_token_203|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128212": { + "content": "<|reserved_special_token_204|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128213": { + "content": "<|reserved_special_token_205|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128214": { + "content": "<|reserved_special_token_206|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128215": { + "content": "<|reserved_special_token_207|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128216": { + "content": "<|reserved_special_token_208|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128217": { + "content": "<|reserved_special_token_209|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128218": { + "content": "<|reserved_special_token_210|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128219": { + "content": "<|reserved_special_token_211|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128220": { + "content": "<|reserved_special_token_212|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128221": { + "content": "<|reserved_special_token_213|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128222": { + "content": "<|reserved_special_token_214|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128223": { + "content": "<|reserved_special_token_215|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128224": { + "content": "<|reserved_special_token_216|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128225": { + "content": "<|reserved_special_token_217|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128226": { + "content": "<|reserved_special_token_218|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128227": { + "content": "<|reserved_special_token_219|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128228": { + "content": "<|reserved_special_token_220|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128229": { + "content": "<|reserved_special_token_221|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128230": { + "content": "<|reserved_special_token_222|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128231": { + "content": "<|reserved_special_token_223|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128232": { + "content": "<|reserved_special_token_224|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128233": { + "content": "<|reserved_special_token_225|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128234": { + "content": "<|reserved_special_token_226|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128235": { + "content": "<|reserved_special_token_227|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128236": { + "content": "<|reserved_special_token_228|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128237": { + "content": "<|reserved_special_token_229|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128238": { + "content": "<|reserved_special_token_230|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128239": { + "content": "<|reserved_special_token_231|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128240": { + "content": "<|reserved_special_token_232|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128241": { + "content": "<|reserved_special_token_233|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128242": { + "content": "<|reserved_special_token_234|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128243": { + "content": "<|reserved_special_token_235|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128244": { + "content": "<|reserved_special_token_236|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128245": { + "content": "<|reserved_special_token_237|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128246": { + "content": "<|reserved_special_token_238|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128247": { + "content": "<|reserved_special_token_239|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128248": { + "content": "<|reserved_special_token_240|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128249": { + "content": "<|reserved_special_token_241|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128250": { + "content": "<|reserved_special_token_242|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128251": { + "content": "<|reserved_special_token_243|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128252": { + "content": "<|reserved_special_token_244|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128253": { + "content": "<|reserved_special_token_245|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128254": { + "content": "<|reserved_special_token_246|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128255": { + "content": "<|reserved_special_token_247|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + } + }, + "bos_token": "<|begin_of_text|>", + "chat_template": "{% if not add_generation_prompt is defined %}{% set add_generation_prompt = false %}{% endif %}{% set loop_messages = messages %}{% for message in loop_messages %}{% set content = '<|start_header_id|>' + message['role'] + '<|end_header_id|>\n\n'+ message['content'] | trim + '<|eot_id|>' %}{% if loop.index0 == 0 %}{% set content = bos_token + content %}{% endif %}{{ content }}{% endfor %}{% if add_generation_prompt %}{{ '<|start_header_id|>assistant<|end_header_id|>\n\n' }}{% endif %}", + "clean_up_tokenization_spaces": true, + "eos_token": "<|eot_id|>", + "extra_special_tokens": {}, + "model_input_names": [ + "input_ids", + "attention_mask" + ], + "model_max_length": 131072, + "pad_token": "<|end_of_text|>", + "tokenizer_class": "PreTrainedTokenizer" +} diff --git a/checkpoint-2445/trainer_state.json b/checkpoint-2445/trainer_state.json new file mode 100644 index 0000000..3866020 --- /dev/null +++ b/checkpoint-2445/trainer_state.json @@ -0,0 +1,36709 @@ +{ + "best_global_step": null, + "best_metric": null, + "best_model_checkpoint": null, + "epoch": 3.0, + "eval_steps": 500, + "global_step": 2445, + "is_hyper_param_search": false, + "is_local_process_zero": true, + "is_world_process_zero": true, + "log_history": [ + { + "epoch": 0.001226993865030675, + "grad_norm": 151.20719434086217, + "learning_rate": 6.000000000000001e-08, + "logits/chosen": -0.32421875, + "logits/rejected": -0.515625, + "logps/chosen": -474.0, + "logps/rejected": -316.0, + "loss": 0.6914, + "rewards/accuracies": 0.0, + "rewards/chosen": 0.0, + "rewards/margins": 0.0, + "rewards/rejected": 0.0, + "step": 1 + }, + { + "epoch": 0.00245398773006135, + "grad_norm": 162.171350331327, + "learning_rate": 1.2000000000000002e-07, + "logits/chosen": -0.22265625, + "logits/rejected": -0.439453125, + "logps/chosen": -436.0, + "logps/rejected": -292.0, + "loss": 0.6914, + "rewards/accuracies": 0.0, + "rewards/chosen": 0.0, + "rewards/margins": 0.0, + "rewards/rejected": 0.0, + "step": 2 + }, + { + "epoch": 0.0036809815950920245, + "grad_norm": 121.91181472288706, + "learning_rate": 1.8e-07, + "logits/chosen": -0.1298828125, + "logits/rejected": -0.2578125, + "logps/chosen": -368.0, + "logps/rejected": -272.0, + "loss": 0.7012, + "rewards/accuracies": 0.3359375, + "rewards/chosen": -0.0006103515625, + "rewards/margins": -0.00506591796875, + "rewards/rejected": 0.00445556640625, + "step": 3 + }, + { + "epoch": 0.0049079754601227, + "grad_norm": 138.61232520616198, + "learning_rate": 2.4000000000000003e-07, + "logits/chosen": -0.166015625, + "logits/rejected": -0.326171875, + "logps/chosen": -422.0, + "logps/rejected": -242.0, + "loss": 0.7003, + "rewards/accuracies": 0.359375, + "rewards/chosen": 0.0057373046875, + "rewards/margins": -0.00762939453125, + "rewards/rejected": 0.01336669921875, + "step": 4 + }, + { + "epoch": 0.006134969325153374, + "grad_norm": 318.4843338984337, + "learning_rate": 3.0000000000000004e-07, + "logits/chosen": -0.1826171875, + "logits/rejected": -0.32421875, + "logps/chosen": -380.0, + "logps/rejected": -270.0, + "loss": 0.7002, + "rewards/accuracies": 0.296875, + "rewards/chosen": -0.0037994384765625, + "rewards/margins": -0.01123046875, + "rewards/rejected": 0.007415771484375, + "step": 5 + }, + { + "epoch": 0.007361963190184049, + "grad_norm": 356.6573592678301, + "learning_rate": 3.6e-07, + "logits/chosen": -0.255859375, + "logits/rejected": -0.345703125, + "logps/chosen": -412.0, + "logps/rejected": -306.0, + "loss": 0.692, + "rewards/accuracies": 0.359375, + "rewards/chosen": 0.0208740234375, + "rewards/margins": 0.01153564453125, + "rewards/rejected": 0.00927734375, + "step": 6 + }, + { + "epoch": 0.008588957055214725, + "grad_norm": 228.0843622624859, + "learning_rate": 4.2000000000000006e-07, + "logits/chosen": -0.2060546875, + "logits/rejected": -0.38671875, + "logps/chosen": -378.0, + "logps/rejected": -266.0, + "loss": 0.6962, + "rewards/accuracies": 0.3671875, + "rewards/chosen": 0.006683349609375, + "rewards/margins": -0.001617431640625, + "rewards/rejected": 0.00830078125, + "step": 7 + }, + { + "epoch": 0.0098159509202454, + "grad_norm": 139.9677750400214, + "learning_rate": 4.800000000000001e-07, + "logits/chosen": -0.111328125, + "logits/rejected": -0.3203125, + "logps/chosen": -432.0, + "logps/rejected": -292.0, + "loss": 0.6667, + "rewards/accuracies": 0.453125, + "rewards/chosen": 0.054443359375, + "rewards/margins": 0.06201171875, + "rewards/rejected": -0.00762939453125, + "step": 8 + }, + { + "epoch": 0.011042944785276074, + "grad_norm": 114.26865689975604, + "learning_rate": 5.4e-07, + "logits/chosen": -0.1318359375, + "logits/rejected": -0.359375, + "logps/chosen": -402.0, + "logps/rejected": -294.0, + "loss": 0.6492, + "rewards/accuracies": 0.4921875, + "rewards/chosen": 0.0654296875, + "rewards/margins": 0.10986328125, + "rewards/rejected": -0.044677734375, + "step": 9 + }, + { + "epoch": 0.012269938650306749, + "grad_norm": 150.56729793626556, + "learning_rate": 6.000000000000001e-07, + "logits/chosen": -0.1953125, + "logits/rejected": -0.349609375, + "logps/chosen": -402.0, + "logps/rejected": -278.0, + "loss": 0.6146, + "rewards/accuracies": 0.59375, + "rewards/chosen": 0.158203125, + "rewards/margins": 0.1943359375, + "rewards/rejected": -0.0361328125, + "step": 10 + }, + { + "epoch": 0.013496932515337423, + "grad_norm": 85.94823037139018, + "learning_rate": 6.6e-07, + "logits/chosen": -0.375, + "logits/rejected": -0.515625, + "logps/chosen": -420.0, + "logps/rejected": -298.0, + "loss": 0.5535, + "rewards/accuracies": 0.6875, + "rewards/chosen": 0.2451171875, + "rewards/margins": 0.3984375, + "rewards/rejected": -0.1533203125, + "step": 11 + }, + { + "epoch": 0.014723926380368098, + "grad_norm": 60.96373680958018, + "learning_rate": 7.2e-07, + "logits/chosen": -0.1337890625, + "logits/rejected": -0.244140625, + "logps/chosen": -372.0, + "logps/rejected": -262.0, + "loss": 0.548, + "rewards/accuracies": 0.6953125, + "rewards/chosen": 0.255859375, + "rewards/margins": 0.435546875, + "rewards/rejected": -0.1796875, + "step": 12 + }, + { + "epoch": 0.015950920245398775, + "grad_norm": 92.70871898741228, + "learning_rate": 7.8e-07, + "logits/chosen": -0.240234375, + "logits/rejected": -0.408203125, + "logps/chosen": -390.0, + "logps/rejected": -251.0, + "loss": 0.5255, + "rewards/accuracies": 0.703125, + "rewards/chosen": 0.302734375, + "rewards/margins": 0.56640625, + "rewards/rejected": -0.26171875, + "step": 13 + }, + { + "epoch": 0.01717791411042945, + "grad_norm": 48.59299098972598, + "learning_rate": 8.400000000000001e-07, + "logits/chosen": -0.2138671875, + "logits/rejected": -0.3984375, + "logps/chosen": -466.0, + "logps/rejected": -314.0, + "loss": 0.4612, + "rewards/accuracies": 0.7265625, + "rewards/chosen": 0.765625, + "rewards/margins": 1.078125, + "rewards/rejected": -0.31640625, + "step": 14 + }, + { + "epoch": 0.018404907975460124, + "grad_norm": 59.699183777930486, + "learning_rate": 9e-07, + "logits/chosen": -0.10400390625, + "logits/rejected": -0.177734375, + "logps/chosen": -414.0, + "logps/rejected": -292.0, + "loss": 0.5014, + "rewards/accuracies": 0.734375, + "rewards/chosen": 0.6484375, + "rewards/margins": 0.96875, + "rewards/rejected": -0.31640625, + "step": 15 + }, + { + "epoch": 0.0196319018404908, + "grad_norm": 52.10075632980986, + "learning_rate": 9.600000000000001e-07, + "logits/chosen": -0.25, + "logits/rejected": -0.42578125, + "logps/chosen": -400.0, + "logps/rejected": -310.0, + "loss": 0.4424, + "rewards/accuracies": 0.7265625, + "rewards/chosen": 0.84375, + "rewards/margins": 1.390625, + "rewards/rejected": -0.54296875, + "step": 16 + }, + { + "epoch": 0.020858895705521473, + "grad_norm": 42.16097538458702, + "learning_rate": 1.0200000000000002e-06, + "logits/chosen": -0.1572265625, + "logits/rejected": -0.31640625, + "logps/chosen": -426.0, + "logps/rejected": -260.0, + "loss": 0.4308, + "rewards/accuracies": 0.796875, + "rewards/chosen": 0.984375, + "rewards/margins": 1.578125, + "rewards/rejected": -0.58984375, + "step": 17 + }, + { + "epoch": 0.022085889570552148, + "grad_norm": 46.56295761289983, + "learning_rate": 1.08e-06, + "logits/chosen": -0.1435546875, + "logits/rejected": -0.330078125, + "logps/chosen": -398.0, + "logps/rejected": -282.0, + "loss": 0.432, + "rewards/accuracies": 0.828125, + "rewards/chosen": 0.875, + "rewards/margins": 1.6875, + "rewards/rejected": -0.8125, + "step": 18 + }, + { + "epoch": 0.023312883435582823, + "grad_norm": 49.57615257325465, + "learning_rate": 1.14e-06, + "logits/chosen": -0.185546875, + "logits/rejected": -0.3671875, + "logps/chosen": -382.0, + "logps/rejected": -296.0, + "loss": 0.4802, + "rewards/accuracies": 0.7734375, + "rewards/chosen": 0.76171875, + "rewards/margins": 1.8125, + "rewards/rejected": -1.0546875, + "step": 19 + }, + { + "epoch": 0.024539877300613498, + "grad_norm": 42.63506109619643, + "learning_rate": 1.2000000000000002e-06, + "logits/chosen": -0.263671875, + "logits/rejected": -0.439453125, + "logps/chosen": -456.0, + "logps/rejected": -324.0, + "loss": 0.3801, + "rewards/accuracies": 0.8125, + "rewards/chosen": 1.4375, + "rewards/margins": 3.140625, + "rewards/rejected": -1.703125, + "step": 20 + }, + { + "epoch": 0.025766871165644172, + "grad_norm": 34.456600308133275, + "learning_rate": 1.26e-06, + "logits/chosen": -0.1328125, + "logits/rejected": -0.3203125, + "logps/chosen": -404.0, + "logps/rejected": -272.0, + "loss": 0.3497, + "rewards/accuracies": 0.8515625, + "rewards/chosen": 1.1015625, + "rewards/margins": 2.640625, + "rewards/rejected": -1.53125, + "step": 21 + }, + { + "epoch": 0.026993865030674847, + "grad_norm": 44.983878720566445, + "learning_rate": 1.32e-06, + "logits/chosen": -0.1298828125, + "logits/rejected": -0.271484375, + "logps/chosen": -374.0, + "logps/rejected": -306.0, + "loss": 0.4156, + "rewards/accuracies": 0.7890625, + "rewards/chosen": 1.0546875, + "rewards/margins": 2.84375, + "rewards/rejected": -1.7890625, + "step": 22 + }, + { + "epoch": 0.02822085889570552, + "grad_norm": 56.72818428668396, + "learning_rate": 1.3800000000000001e-06, + "logits/chosen": -0.251953125, + "logits/rejected": -0.314453125, + "logps/chosen": -370.0, + "logps/rejected": -318.0, + "loss": 0.4566, + "rewards/accuracies": 0.7890625, + "rewards/chosen": 1.46875, + "rewards/margins": 3.1875, + "rewards/rejected": -1.7265625, + "step": 23 + }, + { + "epoch": 0.029447852760736196, + "grad_norm": 55.662509241103734, + "learning_rate": 1.44e-06, + "logits/chosen": -0.1884765625, + "logits/rejected": -0.375, + "logps/chosen": -376.0, + "logps/rejected": -288.0, + "loss": 0.4763, + "rewards/accuracies": 0.7421875, + "rewards/chosen": 0.64453125, + "rewards/margins": 2.265625, + "rewards/rejected": -1.625, + "step": 24 + }, + { + "epoch": 0.03067484662576687, + "grad_norm": 35.38065334196675, + "learning_rate": 1.5e-06, + "logits/chosen": -0.2578125, + "logits/rejected": -0.4453125, + "logps/chosen": -430.0, + "logps/rejected": -312.0, + "loss": 0.3378, + "rewards/accuracies": 0.8984375, + "rewards/chosen": 1.2890625, + "rewards/margins": 3.515625, + "rewards/rejected": -2.21875, + "step": 25 + }, + { + "epoch": 0.03190184049079755, + "grad_norm": 42.887450752414956, + "learning_rate": 1.4993801652892562e-06, + "logits/chosen": -0.15625, + "logits/rejected": -0.31640625, + "logps/chosen": -380.0, + "logps/rejected": -288.0, + "loss": 0.4058, + "rewards/accuracies": 0.8046875, + "rewards/chosen": 1.640625, + "rewards/margins": 3.59375, + "rewards/rejected": -1.9609375, + "step": 26 + }, + { + "epoch": 0.033128834355828224, + "grad_norm": 37.07120513953294, + "learning_rate": 1.4987603305785125e-06, + "logits/chosen": -0.095703125, + "logits/rejected": -0.2734375, + "logps/chosen": -354.0, + "logps/rejected": -292.0, + "loss": 0.3683, + "rewards/accuracies": 0.765625, + "rewards/chosen": 1.7578125, + "rewards/margins": 4.09375, + "rewards/rejected": -2.34375, + "step": 27 + }, + { + "epoch": 0.0343558282208589, + "grad_norm": 35.87136073029194, + "learning_rate": 1.4981404958677687e-06, + "logits/chosen": -0.036376953125, + "logits/rejected": -0.1484375, + "logps/chosen": -374.0, + "logps/rejected": -274.0, + "loss": 0.3615, + "rewards/accuracies": 0.8125, + "rewards/chosen": 1.6171875, + "rewards/margins": 3.421875, + "rewards/rejected": -1.796875, + "step": 28 + }, + { + "epoch": 0.03558282208588957, + "grad_norm": 38.17662218776474, + "learning_rate": 1.4975206611570248e-06, + "logits/chosen": -0.263671875, + "logits/rejected": -0.4140625, + "logps/chosen": -414.0, + "logps/rejected": -334.0, + "loss": 0.3367, + "rewards/accuracies": 0.8046875, + "rewards/chosen": 1.765625, + "rewards/margins": 4.0625, + "rewards/rejected": -2.296875, + "step": 29 + }, + { + "epoch": 0.03680981595092025, + "grad_norm": 39.994932388621756, + "learning_rate": 1.496900826446281e-06, + "logits/chosen": -0.1337890625, + "logits/rejected": -0.248046875, + "logps/chosen": -332.0, + "logps/rejected": -278.0, + "loss": 0.3648, + "rewards/accuracies": 0.8203125, + "rewards/chosen": 1.203125, + "rewards/margins": 3.375, + "rewards/rejected": -2.171875, + "step": 30 + }, + { + "epoch": 0.03803680981595092, + "grad_norm": 42.39396872212231, + "learning_rate": 1.496280991735537e-06, + "logits/chosen": -0.10986328125, + "logits/rejected": -0.2275390625, + "logps/chosen": -378.0, + "logps/rejected": -286.0, + "loss": 0.364, + "rewards/accuracies": 0.7890625, + "rewards/chosen": 1.9765625, + "rewards/margins": 4.25, + "rewards/rejected": -2.265625, + "step": 31 + }, + { + "epoch": 0.0392638036809816, + "grad_norm": 43.84495796047598, + "learning_rate": 1.4956611570247935e-06, + "logits/chosen": -0.15625, + "logits/rejected": -0.333984375, + "logps/chosen": -388.0, + "logps/rejected": -298.0, + "loss": 0.3861, + "rewards/accuracies": 0.78125, + "rewards/chosen": 1.71875, + "rewards/margins": 4.0, + "rewards/rejected": -2.28125, + "step": 32 + }, + { + "epoch": 0.04049079754601227, + "grad_norm": 38.440307864705964, + "learning_rate": 1.4950413223140496e-06, + "logits/chosen": -0.04443359375, + "logits/rejected": -0.177734375, + "logps/chosen": -328.0, + "logps/rejected": -248.0, + "loss": 0.3456, + "rewards/accuracies": 0.8515625, + "rewards/chosen": 1.078125, + "rewards/margins": 3.5625, + "rewards/rejected": -2.484375, + "step": 33 + }, + { + "epoch": 0.04171779141104295, + "grad_norm": 42.669938783190766, + "learning_rate": 1.4944214876033057e-06, + "logits/chosen": -0.1796875, + "logits/rejected": -0.296875, + "logps/chosen": -384.0, + "logps/rejected": -316.0, + "loss": 0.3991, + "rewards/accuracies": 0.7890625, + "rewards/chosen": 1.5859375, + "rewards/margins": 4.40625, + "rewards/rejected": -2.8125, + "step": 34 + }, + { + "epoch": 0.04294478527607362, + "grad_norm": 39.82238013261856, + "learning_rate": 1.493801652892562e-06, + "logits/chosen": -0.0810546875, + "logits/rejected": -0.2197265625, + "logps/chosen": -356.0, + "logps/rejected": -278.0, + "loss": 0.3621, + "rewards/accuracies": 0.828125, + "rewards/chosen": 1.1796875, + "rewards/margins": 4.125, + "rewards/rejected": -2.9375, + "step": 35 + }, + { + "epoch": 0.044171779141104296, + "grad_norm": 34.79756258065961, + "learning_rate": 1.4931818181818182e-06, + "logits/chosen": -0.1376953125, + "logits/rejected": -0.20703125, + "logps/chosen": -398.0, + "logps/rejected": -310.0, + "loss": 0.3113, + "rewards/accuracies": 0.8515625, + "rewards/chosen": 1.6640625, + "rewards/margins": 4.78125, + "rewards/rejected": -3.125, + "step": 36 + }, + { + "epoch": 0.04539877300613497, + "grad_norm": 40.21812818356396, + "learning_rate": 1.4925619834710746e-06, + "logits/chosen": -0.091796875, + "logits/rejected": -0.1943359375, + "logps/chosen": -384.0, + "logps/rejected": -322.0, + "loss": 0.3512, + "rewards/accuracies": 0.796875, + "rewards/chosen": 2.3125, + "rewards/margins": 4.625, + "rewards/rejected": -2.328125, + "step": 37 + }, + { + "epoch": 0.046625766871165646, + "grad_norm": 38.815201724599156, + "learning_rate": 1.4919421487603307e-06, + "logits/chosen": -0.032470703125, + "logits/rejected": -0.1484375, + "logps/chosen": -390.0, + "logps/rejected": -312.0, + "loss": 0.36, + "rewards/accuracies": 0.828125, + "rewards/chosen": 2.546875, + "rewards/margins": 5.09375, + "rewards/rejected": -2.546875, + "step": 38 + }, + { + "epoch": 0.04785276073619632, + "grad_norm": 36.01668225984001, + "learning_rate": 1.4913223140495867e-06, + "logits/chosen": -0.142578125, + "logits/rejected": -0.2890625, + "logps/chosen": -418.0, + "logps/rejected": -318.0, + "loss": 0.3347, + "rewards/accuracies": 0.8359375, + "rewards/chosen": 2.296875, + "rewards/margins": 5.21875, + "rewards/rejected": -2.921875, + "step": 39 + }, + { + "epoch": 0.049079754601226995, + "grad_norm": 30.447104323671855, + "learning_rate": 1.490702479338843e-06, + "logits/chosen": -0.0712890625, + "logits/rejected": -0.1865234375, + "logps/chosen": -370.0, + "logps/rejected": -296.0, + "loss": 0.2729, + "rewards/accuracies": 0.8203125, + "rewards/chosen": 2.09375, + "rewards/margins": 4.90625, + "rewards/rejected": -2.796875, + "step": 40 + }, + { + "epoch": 0.05030674846625767, + "grad_norm": 36.96390592719149, + "learning_rate": 1.4900826446280991e-06, + "logits/chosen": -0.0751953125, + "logits/rejected": -0.2490234375, + "logps/chosen": -374.0, + "logps/rejected": -286.0, + "loss": 0.3822, + "rewards/accuracies": 0.8046875, + "rewards/chosen": 1.875, + "rewards/margins": 4.5625, + "rewards/rejected": -2.671875, + "step": 41 + }, + { + "epoch": 0.051533742331288344, + "grad_norm": 33.41932525699284, + "learning_rate": 1.4894628099173555e-06, + "logits/chosen": -0.022705078125, + "logits/rejected": -0.1650390625, + "logps/chosen": -360.0, + "logps/rejected": -294.0, + "loss": 0.3014, + "rewards/accuracies": 0.84375, + "rewards/chosen": 1.75, + "rewards/margins": 4.3125, + "rewards/rejected": -2.5625, + "step": 42 + }, + { + "epoch": 0.05276073619631902, + "grad_norm": 38.41247549359973, + "learning_rate": 1.4888429752066116e-06, + "logits/chosen": -0.09033203125, + "logits/rejected": -0.248046875, + "logps/chosen": -412.0, + "logps/rejected": -336.0, + "loss": 0.2954, + "rewards/accuracies": 0.84375, + "rewards/chosen": 2.09375, + "rewards/margins": 5.09375, + "rewards/rejected": -3.0, + "step": 43 + }, + { + "epoch": 0.053987730061349694, + "grad_norm": 43.07116536752478, + "learning_rate": 1.4882231404958678e-06, + "logits/chosen": -0.1484375, + "logits/rejected": -0.2490234375, + "logps/chosen": -384.0, + "logps/rejected": -304.0, + "loss": 0.3938, + "rewards/accuracies": 0.8046875, + "rewards/chosen": 1.578125, + "rewards/margins": 4.53125, + "rewards/rejected": -2.96875, + "step": 44 + }, + { + "epoch": 0.05521472392638037, + "grad_norm": 39.15973652461684, + "learning_rate": 1.4876033057851241e-06, + "logits/chosen": -0.2197265625, + "logits/rejected": -0.294921875, + "logps/chosen": -350.0, + "logps/rejected": -300.0, + "loss": 0.3339, + "rewards/accuracies": 0.796875, + "rewards/chosen": 1.828125, + "rewards/margins": 4.9375, + "rewards/rejected": -3.109375, + "step": 45 + }, + { + "epoch": 0.05644171779141104, + "grad_norm": 37.261369579802675, + "learning_rate": 1.4869834710743803e-06, + "logits/chosen": -0.09716796875, + "logits/rejected": -0.259765625, + "logps/chosen": -416.0, + "logps/rejected": -302.0, + "loss": 0.2861, + "rewards/accuracies": 0.828125, + "rewards/chosen": 2.46875, + "rewards/margins": 5.375, + "rewards/rejected": -2.890625, + "step": 46 + }, + { + "epoch": 0.05766871165644172, + "grad_norm": 54.78531831058221, + "learning_rate": 1.4863636363636364e-06, + "logits/chosen": -0.15625, + "logits/rejected": -0.216796875, + "logps/chosen": -390.0, + "logps/rejected": -372.0, + "loss": 0.5073, + "rewards/accuracies": 0.7109375, + "rewards/chosen": 1.5234375, + "rewards/margins": 4.1875, + "rewards/rejected": -2.6875, + "step": 47 + }, + { + "epoch": 0.05889570552147239, + "grad_norm": 43.69804279742604, + "learning_rate": 1.4857438016528926e-06, + "logits/chosen": -0.1494140625, + "logits/rejected": -0.26171875, + "logps/chosen": -352.0, + "logps/rejected": -288.0, + "loss": 0.4242, + "rewards/accuracies": 0.7734375, + "rewards/chosen": 1.1015625, + "rewards/margins": 3.84375, + "rewards/rejected": -2.734375, + "step": 48 + }, + { + "epoch": 0.06012269938650307, + "grad_norm": 34.89468290879934, + "learning_rate": 1.4851239669421487e-06, + "logits/chosen": -0.146484375, + "logits/rejected": -0.328125, + "logps/chosen": -400.0, + "logps/rejected": -300.0, + "loss": 0.2721, + "rewards/accuracies": 0.859375, + "rewards/chosen": 2.25, + "rewards/margins": 5.46875, + "rewards/rejected": -3.21875, + "step": 49 + }, + { + "epoch": 0.06134969325153374, + "grad_norm": 54.15414768657872, + "learning_rate": 1.484504132231405e-06, + "logits/chosen": -0.14453125, + "logits/rejected": -0.296875, + "logps/chosen": -404.0, + "logps/rejected": -324.0, + "loss": 0.4387, + "rewards/accuracies": 0.765625, + "rewards/chosen": 1.515625, + "rewards/margins": 4.6875, + "rewards/rejected": -3.171875, + "step": 50 + }, + { + "epoch": 0.06257668711656442, + "grad_norm": 34.25072100129564, + "learning_rate": 1.4838842975206612e-06, + "logits/chosen": -0.08251953125, + "logits/rejected": -0.2177734375, + "logps/chosen": -440.0, + "logps/rejected": -364.0, + "loss": 0.2872, + "rewards/accuracies": 0.8203125, + "rewards/chosen": 1.984375, + "rewards/margins": 5.09375, + "rewards/rejected": -3.125, + "step": 51 + }, + { + "epoch": 0.0638036809815951, + "grad_norm": 38.20166040651749, + "learning_rate": 1.4832644628099173e-06, + "logits/chosen": -0.10693359375, + "logits/rejected": -0.265625, + "logps/chosen": -396.0, + "logps/rejected": -338.0, + "loss": 0.3424, + "rewards/accuracies": 0.78125, + "rewards/chosen": 1.65625, + "rewards/margins": 4.84375, + "rewards/rejected": -3.171875, + "step": 52 + }, + { + "epoch": 0.06503067484662577, + "grad_norm": 37.089420431960214, + "learning_rate": 1.4826446280991737e-06, + "logits/chosen": -0.055908203125, + "logits/rejected": -0.189453125, + "logps/chosen": -398.0, + "logps/rejected": -326.0, + "loss": 0.3026, + "rewards/accuracies": 0.8359375, + "rewards/chosen": 1.953125, + "rewards/margins": 4.71875, + "rewards/rejected": -2.75, + "step": 53 + }, + { + "epoch": 0.06625766871165645, + "grad_norm": 35.67709547242796, + "learning_rate": 1.4820247933884298e-06, + "logits/chosen": 0.045654296875, + "logits/rejected": -0.1162109375, + "logps/chosen": -336.0, + "logps/rejected": -266.0, + "loss": 0.2813, + "rewards/accuracies": 0.8671875, + "rewards/chosen": 1.3046875, + "rewards/margins": 4.625, + "rewards/rejected": -3.296875, + "step": 54 + }, + { + "epoch": 0.06748466257668712, + "grad_norm": 40.58362122970572, + "learning_rate": 1.481404958677686e-06, + "logits/chosen": -0.068359375, + "logits/rejected": -0.2060546875, + "logps/chosen": -376.0, + "logps/rejected": -302.0, + "loss": 0.3515, + "rewards/accuracies": 0.828125, + "rewards/chosen": 1.8046875, + "rewards/margins": 5.03125, + "rewards/rejected": -3.234375, + "step": 55 + }, + { + "epoch": 0.0687116564417178, + "grad_norm": 43.576103099929334, + "learning_rate": 1.4807851239669423e-06, + "logits/chosen": -0.034912109375, + "logits/rejected": -0.1787109375, + "logps/chosen": -384.0, + "logps/rejected": -334.0, + "loss": 0.4139, + "rewards/accuracies": 0.78125, + "rewards/chosen": 2.015625, + "rewards/margins": 4.8125, + "rewards/rejected": -2.796875, + "step": 56 + }, + { + "epoch": 0.06993865030674846, + "grad_norm": 51.562196128944976, + "learning_rate": 1.4801652892561983e-06, + "logits/chosen": -0.042724609375, + "logits/rejected": -0.0966796875, + "logps/chosen": -352.0, + "logps/rejected": -326.0, + "loss": 0.4667, + "rewards/accuracies": 0.78125, + "rewards/chosen": 1.6015625, + "rewards/margins": 4.59375, + "rewards/rejected": -2.984375, + "step": 57 + }, + { + "epoch": 0.07116564417177915, + "grad_norm": 49.38086728368548, + "learning_rate": 1.4795454545454546e-06, + "logits/chosen": -0.046630859375, + "logits/rejected": -0.10009765625, + "logps/chosen": -376.0, + "logps/rejected": -334.0, + "loss": 0.3869, + "rewards/accuracies": 0.765625, + "rewards/chosen": 1.640625, + "rewards/margins": 4.875, + "rewards/rejected": -3.234375, + "step": 58 + }, + { + "epoch": 0.07239263803680981, + "grad_norm": 43.27222220894176, + "learning_rate": 1.4789256198347107e-06, + "logits/chosen": 0.0240478515625, + "logits/rejected": -0.051025390625, + "logps/chosen": -378.0, + "logps/rejected": -302.0, + "loss": 0.3983, + "rewards/accuracies": 0.765625, + "rewards/chosen": 1.3125, + "rewards/margins": 4.25, + "rewards/rejected": -2.953125, + "step": 59 + }, + { + "epoch": 0.0736196319018405, + "grad_norm": 39.673404465674906, + "learning_rate": 1.4783057851239669e-06, + "logits/chosen": -0.049072265625, + "logits/rejected": -0.1669921875, + "logps/chosen": -386.0, + "logps/rejected": -294.0, + "loss": 0.3927, + "rewards/accuracies": 0.8046875, + "rewards/chosen": 0.99609375, + "rewards/margins": 4.1875, + "rewards/rejected": -3.1875, + "step": 60 + }, + { + "epoch": 0.07484662576687116, + "grad_norm": 40.98812217259802, + "learning_rate": 1.4776859504132232e-06, + "logits/chosen": -0.064453125, + "logits/rejected": -0.173828125, + "logps/chosen": -418.0, + "logps/rejected": -328.0, + "loss": 0.3496, + "rewards/accuracies": 0.8203125, + "rewards/chosen": 1.890625, + "rewards/margins": 5.65625, + "rewards/rejected": -3.765625, + "step": 61 + }, + { + "epoch": 0.07607361963190185, + "grad_norm": 40.09146702222396, + "learning_rate": 1.4770661157024794e-06, + "logits/chosen": -0.051513671875, + "logits/rejected": -0.2275390625, + "logps/chosen": -426.0, + "logps/rejected": -316.0, + "loss": 0.3416, + "rewards/accuracies": 0.7890625, + "rewards/chosen": 2.546875, + "rewards/margins": 6.1875, + "rewards/rejected": -3.640625, + "step": 62 + }, + { + "epoch": 0.07730061349693251, + "grad_norm": 46.30805404508898, + "learning_rate": 1.4764462809917355e-06, + "logits/chosen": 0.031982421875, + "logits/rejected": -0.11767578125, + "logps/chosen": -418.0, + "logps/rejected": -314.0, + "loss": 0.3323, + "rewards/accuracies": 0.8203125, + "rewards/chosen": 1.515625, + "rewards/margins": 4.84375, + "rewards/rejected": -3.328125, + "step": 63 + }, + { + "epoch": 0.0785276073619632, + "grad_norm": 47.96974437982722, + "learning_rate": 1.4758264462809919e-06, + "logits/chosen": -0.123046875, + "logits/rejected": -0.27734375, + "logps/chosen": -402.0, + "logps/rejected": -336.0, + "loss": 0.4046, + "rewards/accuracies": 0.78125, + "rewards/chosen": 2.046875, + "rewards/margins": 5.46875, + "rewards/rejected": -3.4375, + "step": 64 + }, + { + "epoch": 0.07975460122699386, + "grad_norm": 35.57381547040094, + "learning_rate": 1.475206611570248e-06, + "logits/chosen": -0.18359375, + "logits/rejected": -0.25390625, + "logps/chosen": -390.0, + "logps/rejected": -310.0, + "loss": 0.2899, + "rewards/accuracies": 0.84375, + "rewards/chosen": 2.046875, + "rewards/margins": 5.53125, + "rewards/rejected": -3.484375, + "step": 65 + }, + { + "epoch": 0.08098159509202454, + "grad_norm": 33.414611619044024, + "learning_rate": 1.4745867768595042e-06, + "logits/chosen": -0.162109375, + "logits/rejected": -0.287109375, + "logps/chosen": -314.0, + "logps/rejected": -270.0, + "loss": 0.2704, + "rewards/accuracies": 0.8515625, + "rewards/chosen": 2.328125, + "rewards/margins": 6.125, + "rewards/rejected": -3.78125, + "step": 66 + }, + { + "epoch": 0.08220858895705521, + "grad_norm": 42.668537708095855, + "learning_rate": 1.4739669421487603e-06, + "logits/chosen": -0.068359375, + "logits/rejected": -0.1767578125, + "logps/chosen": -332.0, + "logps/rejected": -302.0, + "loss": 0.3919, + "rewards/accuracies": 0.8046875, + "rewards/chosen": 1.2578125, + "rewards/margins": 4.84375, + "rewards/rejected": -3.578125, + "step": 67 + }, + { + "epoch": 0.0834355828220859, + "grad_norm": 40.70660622767979, + "learning_rate": 1.4733471074380164e-06, + "logits/chosen": -0.1455078125, + "logits/rejected": -0.193359375, + "logps/chosen": -358.0, + "logps/rejected": -338.0, + "loss": 0.3949, + "rewards/accuracies": 0.7890625, + "rewards/chosen": 1.625, + "rewards/margins": 4.96875, + "rewards/rejected": -3.359375, + "step": 68 + }, + { + "epoch": 0.08466257668711656, + "grad_norm": 42.29365611407333, + "learning_rate": 1.4727272727272728e-06, + "logits/chosen": -0.2216796875, + "logits/rejected": -0.36328125, + "logps/chosen": -406.0, + "logps/rejected": -290.0, + "loss": 0.3163, + "rewards/accuracies": 0.8515625, + "rewards/chosen": 2.484375, + "rewards/margins": 6.0, + "rewards/rejected": -3.515625, + "step": 69 + }, + { + "epoch": 0.08588957055214724, + "grad_norm": 45.16433653526153, + "learning_rate": 1.472107438016529e-06, + "logits/chosen": -0.08447265625, + "logits/rejected": -0.267578125, + "logps/chosen": -406.0, + "logps/rejected": -346.0, + "loss": 0.3817, + "rewards/accuracies": 0.7890625, + "rewards/chosen": 2.515625, + "rewards/margins": 5.6875, + "rewards/rejected": -3.15625, + "step": 70 + }, + { + "epoch": 0.08711656441717791, + "grad_norm": 37.819952615726415, + "learning_rate": 1.4714876033057853e-06, + "logits/chosen": -0.0341796875, + "logits/rejected": -0.1005859375, + "logps/chosen": -344.0, + "logps/rejected": -286.0, + "loss": 0.3666, + "rewards/accuracies": 0.828125, + "rewards/chosen": 1.46875, + "rewards/margins": 4.625, + "rewards/rejected": -3.15625, + "step": 71 + }, + { + "epoch": 0.08834355828220859, + "grad_norm": 41.96140458025261, + "learning_rate": 1.4708677685950414e-06, + "logits/chosen": -0.19921875, + "logits/rejected": -0.33203125, + "logps/chosen": -428.0, + "logps/rejected": -354.0, + "loss": 0.3187, + "rewards/accuracies": 0.8359375, + "rewards/chosen": 2.125, + "rewards/margins": 5.1875, + "rewards/rejected": -3.0625, + "step": 72 + }, + { + "epoch": 0.08957055214723926, + "grad_norm": 45.003627256411235, + "learning_rate": 1.4702479338842976e-06, + "logits/chosen": -0.14453125, + "logits/rejected": -0.220703125, + "logps/chosen": -346.0, + "logps/rejected": -308.0, + "loss": 0.4099, + "rewards/accuracies": 0.78125, + "rewards/chosen": 1.6171875, + "rewards/margins": 5.25, + "rewards/rejected": -3.625, + "step": 73 + }, + { + "epoch": 0.09079754601226994, + "grad_norm": 37.795821367857755, + "learning_rate": 1.4696280991735537e-06, + "logits/chosen": -0.1435546875, + "logits/rejected": -0.212890625, + "logps/chosen": -352.0, + "logps/rejected": -338.0, + "loss": 0.2989, + "rewards/accuracies": 0.8515625, + "rewards/chosen": 1.4296875, + "rewards/margins": 5.0625, + "rewards/rejected": -3.625, + "step": 74 + }, + { + "epoch": 0.09202453987730061, + "grad_norm": 38.6325533255864, + "learning_rate": 1.4690082644628099e-06, + "logits/chosen": -0.1083984375, + "logits/rejected": -0.2578125, + "logps/chosen": -376.0, + "logps/rejected": -326.0, + "loss": 0.3138, + "rewards/accuracies": 0.8125, + "rewards/chosen": 2.265625, + "rewards/margins": 5.9375, + "rewards/rejected": -3.671875, + "step": 75 + }, + { + "epoch": 0.09325153374233129, + "grad_norm": 41.500214372239896, + "learning_rate": 1.4683884297520662e-06, + "logits/chosen": -0.11376953125, + "logits/rejected": -0.283203125, + "logps/chosen": -394.0, + "logps/rejected": -296.0, + "loss": 0.3196, + "rewards/accuracies": 0.8203125, + "rewards/chosen": 2.28125, + "rewards/margins": 5.625, + "rewards/rejected": -3.359375, + "step": 76 + }, + { + "epoch": 0.09447852760736196, + "grad_norm": 37.661301568651126, + "learning_rate": 1.4677685950413223e-06, + "logits/chosen": -0.1337890625, + "logits/rejected": -0.267578125, + "logps/chosen": -350.0, + "logps/rejected": -308.0, + "loss": 0.3298, + "rewards/accuracies": 0.8125, + "rewards/chosen": 1.4453125, + "rewards/margins": 5.5, + "rewards/rejected": -4.0625, + "step": 77 + }, + { + "epoch": 0.09570552147239264, + "grad_norm": 37.401749661903814, + "learning_rate": 1.4671487603305785e-06, + "logits/chosen": -0.08837890625, + "logits/rejected": -0.2412109375, + "logps/chosen": -414.0, + "logps/rejected": -324.0, + "loss": 0.3295, + "rewards/accuracies": 0.8125, + "rewards/chosen": 2.015625, + "rewards/margins": 5.90625, + "rewards/rejected": -3.90625, + "step": 78 + }, + { + "epoch": 0.09693251533742331, + "grad_norm": 41.34926796893681, + "learning_rate": 1.4665289256198348e-06, + "logits/chosen": -0.078125, + "logits/rejected": -0.2109375, + "logps/chosen": -354.0, + "logps/rejected": -300.0, + "loss": 0.3855, + "rewards/accuracies": 0.8046875, + "rewards/chosen": 1.25, + "rewards/margins": 5.53125, + "rewards/rejected": -4.28125, + "step": 79 + }, + { + "epoch": 0.09815950920245399, + "grad_norm": 40.52807889159641, + "learning_rate": 1.465909090909091e-06, + "logits/chosen": -0.1396484375, + "logits/rejected": -0.33984375, + "logps/chosen": -404.0, + "logps/rejected": -340.0, + "loss": 0.3387, + "rewards/accuracies": 0.8515625, + "rewards/chosen": 2.484375, + "rewards/margins": 7.21875, + "rewards/rejected": -4.75, + "step": 80 + }, + { + "epoch": 0.09938650306748466, + "grad_norm": 35.605812619698284, + "learning_rate": 1.4652892561983471e-06, + "logits/chosen": -0.208984375, + "logits/rejected": -0.365234375, + "logps/chosen": -436.0, + "logps/rejected": -356.0, + "loss": 0.3191, + "rewards/accuracies": 0.828125, + "rewards/chosen": 1.8828125, + "rewards/margins": 6.65625, + "rewards/rejected": -4.78125, + "step": 81 + }, + { + "epoch": 0.10061349693251534, + "grad_norm": 47.4617133344908, + "learning_rate": 1.4646694214876035e-06, + "logits/chosen": -0.1171875, + "logits/rejected": -0.228515625, + "logps/chosen": -430.0, + "logps/rejected": -348.0, + "loss": 0.3954, + "rewards/accuracies": 0.8203125, + "rewards/chosen": 1.03125, + "rewards/margins": 5.875, + "rewards/rejected": -4.84375, + "step": 82 + }, + { + "epoch": 0.10184049079754601, + "grad_norm": 42.72902928212931, + "learning_rate": 1.4640495867768594e-06, + "logits/chosen": -0.1025390625, + "logits/rejected": -0.255859375, + "logps/chosen": -352.0, + "logps/rejected": -320.0, + "loss": 0.3735, + "rewards/accuracies": 0.8125, + "rewards/chosen": 1.109375, + "rewards/margins": 5.84375, + "rewards/rejected": -4.71875, + "step": 83 + }, + { + "epoch": 0.10306748466257669, + "grad_norm": 59.351571106495314, + "learning_rate": 1.4634297520661158e-06, + "logits/chosen": -0.0869140625, + "logits/rejected": -0.310546875, + "logps/chosen": -448.0, + "logps/rejected": -360.0, + "loss": 0.2887, + "rewards/accuracies": 0.859375, + "rewards/chosen": 1.75, + "rewards/margins": 6.8125, + "rewards/rejected": -5.0625, + "step": 84 + }, + { + "epoch": 0.10429447852760736, + "grad_norm": 40.67504361329859, + "learning_rate": 1.462809917355372e-06, + "logits/chosen": -0.189453125, + "logits/rejected": -0.3359375, + "logps/chosen": -376.0, + "logps/rejected": -338.0, + "loss": 0.2933, + "rewards/accuracies": 0.875, + "rewards/chosen": 1.03125, + "rewards/margins": 6.03125, + "rewards/rejected": -4.96875, + "step": 85 + }, + { + "epoch": 0.10552147239263804, + "grad_norm": 44.45491286199163, + "learning_rate": 1.462190082644628e-06, + "logits/chosen": -0.12109375, + "logits/rejected": -0.224609375, + "logps/chosen": -424.0, + "logps/rejected": -314.0, + "loss": 0.4003, + "rewards/accuracies": 0.78125, + "rewards/chosen": 1.625, + "rewards/margins": 5.875, + "rewards/rejected": -4.25, + "step": 86 + }, + { + "epoch": 0.1067484662576687, + "grad_norm": 37.230987939956755, + "learning_rate": 1.4615702479338844e-06, + "logits/chosen": -0.1728515625, + "logits/rejected": -0.3046875, + "logps/chosen": -354.0, + "logps/rejected": -274.0, + "loss": 0.3433, + "rewards/accuracies": 0.8203125, + "rewards/chosen": 0.6484375, + "rewards/margins": 5.65625, + "rewards/rejected": -5.0, + "step": 87 + }, + { + "epoch": 0.10797546012269939, + "grad_norm": 35.435756690174514, + "learning_rate": 1.4609504132231405e-06, + "logits/chosen": -0.142578125, + "logits/rejected": -0.302734375, + "logps/chosen": -362.0, + "logps/rejected": -286.0, + "loss": 0.2796, + "rewards/accuracies": 0.859375, + "rewards/chosen": 1.2890625, + "rewards/margins": 5.78125, + "rewards/rejected": -4.5, + "step": 88 + }, + { + "epoch": 0.10920245398773006, + "grad_norm": 39.218068339324205, + "learning_rate": 1.4603305785123967e-06, + "logits/chosen": -0.1708984375, + "logits/rejected": -0.2734375, + "logps/chosen": -392.0, + "logps/rejected": -340.0, + "loss": 0.3661, + "rewards/accuracies": 0.8046875, + "rewards/chosen": 1.609375, + "rewards/margins": 5.875, + "rewards/rejected": -4.28125, + "step": 89 + }, + { + "epoch": 0.11042944785276074, + "grad_norm": 35.86351693081174, + "learning_rate": 1.459710743801653e-06, + "logits/chosen": -0.2021484375, + "logits/rejected": -0.375, + "logps/chosen": -378.0, + "logps/rejected": -334.0, + "loss": 0.3305, + "rewards/accuracies": 0.828125, + "rewards/chosen": 1.0625, + "rewards/margins": 5.6875, + "rewards/rejected": -4.625, + "step": 90 + }, + { + "epoch": 0.1116564417177914, + "grad_norm": 39.260039435042906, + "learning_rate": 1.4590909090909092e-06, + "logits/chosen": -0.028076171875, + "logits/rejected": -0.28125, + "logps/chosen": -392.0, + "logps/rejected": -308.0, + "loss": 0.3104, + "rewards/accuracies": 0.8203125, + "rewards/chosen": 1.5234375, + "rewards/margins": 5.9375, + "rewards/rejected": -4.40625, + "step": 91 + }, + { + "epoch": 0.11288343558282209, + "grad_norm": 35.99771043781269, + "learning_rate": 1.4584710743801653e-06, + "logits/chosen": -0.1259765625, + "logits/rejected": -0.236328125, + "logps/chosen": -390.0, + "logps/rejected": -316.0, + "loss": 0.304, + "rewards/accuracies": 0.84375, + "rewards/chosen": 1.484375, + "rewards/margins": 6.3125, + "rewards/rejected": -4.8125, + "step": 92 + }, + { + "epoch": 0.11411042944785275, + "grad_norm": 38.41867053252873, + "learning_rate": 1.4578512396694215e-06, + "logits/chosen": -0.15234375, + "logits/rejected": -0.36328125, + "logps/chosen": -400.0, + "logps/rejected": -310.0, + "loss": 0.3311, + "rewards/accuracies": 0.8515625, + "rewards/chosen": 2.328125, + "rewards/margins": 6.625, + "rewards/rejected": -4.3125, + "step": 93 + }, + { + "epoch": 0.11533742331288344, + "grad_norm": 37.099757310606805, + "learning_rate": 1.4572314049586776e-06, + "logits/chosen": -0.171875, + "logits/rejected": -0.390625, + "logps/chosen": -426.0, + "logps/rejected": -354.0, + "loss": 0.2709, + "rewards/accuracies": 0.84375, + "rewards/chosen": 2.765625, + "rewards/margins": 7.09375, + "rewards/rejected": -4.3125, + "step": 94 + }, + { + "epoch": 0.1165644171779141, + "grad_norm": 36.042450420967036, + "learning_rate": 1.456611570247934e-06, + "logits/chosen": -0.2490234375, + "logits/rejected": -0.392578125, + "logps/chosen": -378.0, + "logps/rejected": -320.0, + "loss": 0.3138, + "rewards/accuracies": 0.859375, + "rewards/chosen": 2.0625, + "rewards/margins": 6.6875, + "rewards/rejected": -4.625, + "step": 95 + }, + { + "epoch": 0.11779141104294479, + "grad_norm": 37.51605196372948, + "learning_rate": 1.45599173553719e-06, + "logits/chosen": -0.2392578125, + "logits/rejected": -0.400390625, + "logps/chosen": -366.0, + "logps/rejected": -320.0, + "loss": 0.2944, + "rewards/accuracies": 0.84375, + "rewards/chosen": 2.625, + "rewards/margins": 7.125, + "rewards/rejected": -4.5, + "step": 96 + }, + { + "epoch": 0.11901840490797547, + "grad_norm": 47.87658843319023, + "learning_rate": 1.4553719008264462e-06, + "logits/chosen": -0.234375, + "logits/rejected": -0.373046875, + "logps/chosen": -422.0, + "logps/rejected": -362.0, + "loss": 0.4367, + "rewards/accuracies": 0.796875, + "rewards/chosen": 2.25, + "rewards/margins": 6.125, + "rewards/rejected": -3.859375, + "step": 97 + }, + { + "epoch": 0.12024539877300613, + "grad_norm": 37.61598969810632, + "learning_rate": 1.4547520661157026e-06, + "logits/chosen": -0.03759765625, + "logits/rejected": -0.236328125, + "logps/chosen": -354.0, + "logps/rejected": -296.0, + "loss": 0.3071, + "rewards/accuracies": 0.84375, + "rewards/chosen": 1.984375, + "rewards/margins": 6.09375, + "rewards/rejected": -4.09375, + "step": 98 + }, + { + "epoch": 0.12147239263803682, + "grad_norm": 35.33546250741774, + "learning_rate": 1.4541322314049587e-06, + "logits/chosen": -0.189453125, + "logits/rejected": -0.287109375, + "logps/chosen": -338.0, + "logps/rejected": -258.0, + "loss": 0.2669, + "rewards/accuracies": 0.8515625, + "rewards/chosen": 1.359375, + "rewards/margins": 5.5, + "rewards/rejected": -4.15625, + "step": 99 + }, + { + "epoch": 0.12269938650306748, + "grad_norm": 32.14334216613253, + "learning_rate": 1.453512396694215e-06, + "logits/chosen": -0.2197265625, + "logits/rejected": -0.404296875, + "logps/chosen": -394.0, + "logps/rejected": -332.0, + "loss": 0.2858, + "rewards/accuracies": 0.8359375, + "rewards/chosen": 1.8359375, + "rewards/margins": 6.46875, + "rewards/rejected": -4.625, + "step": 100 + }, + { + "epoch": 0.12392638036809817, + "grad_norm": 45.60783475518701, + "learning_rate": 1.452892561983471e-06, + "logits/chosen": -0.220703125, + "logits/rejected": -0.400390625, + "logps/chosen": -412.0, + "logps/rejected": -352.0, + "loss": 0.4076, + "rewards/accuracies": 0.796875, + "rewards/chosen": 1.9140625, + "rewards/margins": 6.3125, + "rewards/rejected": -4.375, + "step": 101 + }, + { + "epoch": 0.12515337423312883, + "grad_norm": 39.740396682339004, + "learning_rate": 1.4522727272727272e-06, + "logits/chosen": -0.1865234375, + "logits/rejected": -0.298828125, + "logps/chosen": -392.0, + "logps/rejected": -330.0, + "loss": 0.3534, + "rewards/accuracies": 0.8046875, + "rewards/chosen": 2.078125, + "rewards/margins": 6.0, + "rewards/rejected": -3.9375, + "step": 102 + }, + { + "epoch": 0.1263803680981595, + "grad_norm": 35.16122653572647, + "learning_rate": 1.4516528925619835e-06, + "logits/chosen": -0.17578125, + "logits/rejected": -0.326171875, + "logps/chosen": -350.0, + "logps/rejected": -310.0, + "loss": 0.3153, + "rewards/accuracies": 0.84375, + "rewards/chosen": 1.4765625, + "rewards/margins": 5.75, + "rewards/rejected": -4.28125, + "step": 103 + }, + { + "epoch": 0.1276073619631902, + "grad_norm": 46.63941850097279, + "learning_rate": 1.4510330578512396e-06, + "logits/chosen": -0.091796875, + "logits/rejected": -0.2255859375, + "logps/chosen": -414.0, + "logps/rejected": -348.0, + "loss": 0.381, + "rewards/accuracies": 0.7890625, + "rewards/chosen": 1.6484375, + "rewards/margins": 5.78125, + "rewards/rejected": -4.125, + "step": 104 + }, + { + "epoch": 0.12883435582822086, + "grad_norm": 35.04246201772006, + "learning_rate": 1.450413223140496e-06, + "logits/chosen": -0.181640625, + "logits/rejected": -0.3828125, + "logps/chosen": -374.0, + "logps/rejected": -322.0, + "loss": 0.2864, + "rewards/accuracies": 0.8515625, + "rewards/chosen": 2.03125, + "rewards/margins": 6.8125, + "rewards/rejected": -4.78125, + "step": 105 + }, + { + "epoch": 0.13006134969325153, + "grad_norm": 39.36563074587024, + "learning_rate": 1.4497933884297521e-06, + "logits/chosen": -0.1591796875, + "logits/rejected": -0.3046875, + "logps/chosen": -392.0, + "logps/rejected": -322.0, + "loss": 0.3462, + "rewards/accuracies": 0.7890625, + "rewards/chosen": 1.7890625, + "rewards/margins": 6.03125, + "rewards/rejected": -4.25, + "step": 106 + }, + { + "epoch": 0.1312883435582822, + "grad_norm": 35.07229699022507, + "learning_rate": 1.4491735537190083e-06, + "logits/chosen": -0.1865234375, + "logits/rejected": -0.26953125, + "logps/chosen": -352.0, + "logps/rejected": -312.0, + "loss": 0.2577, + "rewards/accuracies": 0.8671875, + "rewards/chosen": 1.3515625, + "rewards/margins": 5.84375, + "rewards/rejected": -4.5, + "step": 107 + }, + { + "epoch": 0.1325153374233129, + "grad_norm": 28.019513861884207, + "learning_rate": 1.4485537190082646e-06, + "logits/chosen": -0.19921875, + "logits/rejected": -0.408203125, + "logps/chosen": -382.0, + "logps/rejected": -320.0, + "loss": 0.2438, + "rewards/accuracies": 0.8828125, + "rewards/chosen": 1.3046875, + "rewards/margins": 6.5625, + "rewards/rejected": -5.25, + "step": 108 + }, + { + "epoch": 0.13374233128834356, + "grad_norm": 35.40605672013622, + "learning_rate": 1.4479338842975208e-06, + "logits/chosen": -0.18359375, + "logits/rejected": -0.359375, + "logps/chosen": -412.0, + "logps/rejected": -326.0, + "loss": 0.3584, + "rewards/accuracies": 0.8359375, + "rewards/chosen": 2.5625, + "rewards/margins": 7.1875, + "rewards/rejected": -4.625, + "step": 109 + }, + { + "epoch": 0.13496932515337423, + "grad_norm": 38.14439771999889, + "learning_rate": 1.447314049586777e-06, + "logits/chosen": -0.1728515625, + "logits/rejected": -0.310546875, + "logps/chosen": -406.0, + "logps/rejected": -348.0, + "loss": 0.3481, + "rewards/accuracies": 0.8125, + "rewards/chosen": 1.9140625, + "rewards/margins": 6.4375, + "rewards/rejected": -4.53125, + "step": 110 + }, + { + "epoch": 0.1361963190184049, + "grad_norm": 36.779042619503116, + "learning_rate": 1.446694214876033e-06, + "logits/chosen": -0.09716796875, + "logits/rejected": -0.251953125, + "logps/chosen": -334.0, + "logps/rejected": -282.0, + "loss": 0.3363, + "rewards/accuracies": 0.859375, + "rewards/chosen": 1.3671875, + "rewards/margins": 5.4375, + "rewards/rejected": -4.0625, + "step": 111 + }, + { + "epoch": 0.1374233128834356, + "grad_norm": 36.86435500620325, + "learning_rate": 1.4460743801652892e-06, + "logits/chosen": -0.177734375, + "logits/rejected": -0.29296875, + "logps/chosen": -342.0, + "logps/rejected": -292.0, + "loss": 0.343, + "rewards/accuracies": 0.859375, + "rewards/chosen": 1.3046875, + "rewards/margins": 5.75, + "rewards/rejected": -4.4375, + "step": 112 + }, + { + "epoch": 0.13865030674846626, + "grad_norm": 38.139474814179394, + "learning_rate": 1.4454545454545456e-06, + "logits/chosen": -0.0849609375, + "logits/rejected": -0.169921875, + "logps/chosen": -374.0, + "logps/rejected": -312.0, + "loss": 0.3952, + "rewards/accuracies": 0.828125, + "rewards/chosen": 1.859375, + "rewards/margins": 5.75, + "rewards/rejected": -3.90625, + "step": 113 + }, + { + "epoch": 0.13987730061349693, + "grad_norm": 38.81539599706057, + "learning_rate": 1.4448347107438017e-06, + "logits/chosen": -0.1220703125, + "logits/rejected": -0.275390625, + "logps/chosen": -372.0, + "logps/rejected": -316.0, + "loss": 0.3854, + "rewards/accuracies": 0.859375, + "rewards/chosen": 1.5078125, + "rewards/margins": 5.375, + "rewards/rejected": -3.875, + "step": 114 + }, + { + "epoch": 0.1411042944785276, + "grad_norm": 45.65683573404254, + "learning_rate": 1.4442148760330578e-06, + "logits/chosen": -0.1298828125, + "logits/rejected": -0.2490234375, + "logps/chosen": -338.0, + "logps/rejected": -320.0, + "loss": 0.4779, + "rewards/accuracies": 0.75, + "rewards/chosen": 1.046875, + "rewards/margins": 4.71875, + "rewards/rejected": -3.65625, + "step": 115 + }, + { + "epoch": 0.1423312883435583, + "grad_norm": 32.97644733528144, + "learning_rate": 1.4435950413223142e-06, + "logits/chosen": -0.185546875, + "logits/rejected": -0.408203125, + "logps/chosen": -376.0, + "logps/rejected": -324.0, + "loss": 0.2978, + "rewards/accuracies": 0.8359375, + "rewards/chosen": 2.40625, + "rewards/margins": 6.03125, + "rewards/rejected": -3.625, + "step": 116 + }, + { + "epoch": 0.14355828220858896, + "grad_norm": 38.235071091905795, + "learning_rate": 1.4429752066115703e-06, + "logits/chosen": -0.130859375, + "logits/rejected": -0.26953125, + "logps/chosen": -324.0, + "logps/rejected": -270.0, + "loss": 0.4211, + "rewards/accuracies": 0.796875, + "rewards/chosen": 1.3125, + "rewards/margins": 5.53125, + "rewards/rejected": -4.21875, + "step": 117 + }, + { + "epoch": 0.14478527607361963, + "grad_norm": 41.12118912071453, + "learning_rate": 1.4423553719008265e-06, + "logits/chosen": -0.15234375, + "logits/rejected": -0.298828125, + "logps/chosen": -418.0, + "logps/rejected": -360.0, + "loss": 0.3658, + "rewards/accuracies": 0.8046875, + "rewards/chosen": 2.28125, + "rewards/margins": 6.125, + "rewards/rejected": -3.84375, + "step": 118 + }, + { + "epoch": 0.1460122699386503, + "grad_norm": 35.626041219957244, + "learning_rate": 1.4417355371900826e-06, + "logits/chosen": -0.220703125, + "logits/rejected": -0.30859375, + "logps/chosen": -352.0, + "logps/rejected": -324.0, + "loss": 0.3003, + "rewards/accuracies": 0.8125, + "rewards/chosen": 1.6953125, + "rewards/margins": 5.59375, + "rewards/rejected": -3.90625, + "step": 119 + }, + { + "epoch": 0.147239263803681, + "grad_norm": 40.42176001046704, + "learning_rate": 1.4411157024793388e-06, + "logits/chosen": -0.201171875, + "logits/rejected": -0.37890625, + "logps/chosen": -418.0, + "logps/rejected": -354.0, + "loss": 0.3506, + "rewards/accuracies": 0.8359375, + "rewards/chosen": 2.6875, + "rewards/margins": 6.625, + "rewards/rejected": -3.9375, + "step": 120 + }, + { + "epoch": 0.14846625766871166, + "grad_norm": 33.17627892781642, + "learning_rate": 1.4404958677685951e-06, + "logits/chosen": -0.12060546875, + "logits/rejected": -0.27734375, + "logps/chosen": -404.0, + "logps/rejected": -336.0, + "loss": 0.2824, + "rewards/accuracies": 0.8203125, + "rewards/chosen": 2.625, + "rewards/margins": 6.75, + "rewards/rejected": -4.125, + "step": 121 + }, + { + "epoch": 0.14969325153374233, + "grad_norm": 41.361282319998544, + "learning_rate": 1.4398760330578512e-06, + "logits/chosen": -0.130859375, + "logits/rejected": -0.283203125, + "logps/chosen": -378.0, + "logps/rejected": -298.0, + "loss": 0.3454, + "rewards/accuracies": 0.828125, + "rewards/chosen": 1.7421875, + "rewards/margins": 5.75, + "rewards/rejected": -4.03125, + "step": 122 + }, + { + "epoch": 0.150920245398773, + "grad_norm": 39.0679541541933, + "learning_rate": 1.4392561983471074e-06, + "logits/chosen": -0.0191650390625, + "logits/rejected": -0.1884765625, + "logps/chosen": -380.0, + "logps/rejected": -290.0, + "loss": 0.384, + "rewards/accuracies": 0.7890625, + "rewards/chosen": 1.2734375, + "rewards/margins": 5.625, + "rewards/rejected": -4.375, + "step": 123 + }, + { + "epoch": 0.1521472392638037, + "grad_norm": 30.841344702577548, + "learning_rate": 1.4386363636363637e-06, + "logits/chosen": -0.193359375, + "logits/rejected": -0.255859375, + "logps/chosen": -348.0, + "logps/rejected": -294.0, + "loss": 0.2792, + "rewards/accuracies": 0.890625, + "rewards/chosen": 2.203125, + "rewards/margins": 7.0, + "rewards/rejected": -4.8125, + "step": 124 + }, + { + "epoch": 0.15337423312883436, + "grad_norm": 40.57809799655024, + "learning_rate": 1.4380165289256199e-06, + "logits/chosen": -0.259765625, + "logits/rejected": -0.431640625, + "logps/chosen": -434.0, + "logps/rejected": -340.0, + "loss": 0.3257, + "rewards/accuracies": 0.84375, + "rewards/chosen": 1.6015625, + "rewards/margins": 6.625, + "rewards/rejected": -5.0, + "step": 125 + }, + { + "epoch": 0.15460122699386503, + "grad_norm": 49.176910835221456, + "learning_rate": 1.437396694214876e-06, + "logits/chosen": -0.11279296875, + "logits/rejected": -0.30078125, + "logps/chosen": -368.0, + "logps/rejected": -320.0, + "loss": 0.4071, + "rewards/accuracies": 0.8359375, + "rewards/chosen": 0.6328125, + "rewards/margins": 5.59375, + "rewards/rejected": -4.96875, + "step": 126 + }, + { + "epoch": 0.1558282208588957, + "grad_norm": 42.06253723411312, + "learning_rate": 1.4367768595041324e-06, + "logits/chosen": -0.064453125, + "logits/rejected": -0.2431640625, + "logps/chosen": -388.0, + "logps/rejected": -340.0, + "loss": 0.4543, + "rewards/accuracies": 0.7578125, + "rewards/chosen": 1.75, + "rewards/margins": 6.40625, + "rewards/rejected": -4.65625, + "step": 127 + }, + { + "epoch": 0.1570552147239264, + "grad_norm": 49.94695723772631, + "learning_rate": 1.4361570247933883e-06, + "logits/chosen": -0.1044921875, + "logits/rejected": -0.2890625, + "logps/chosen": -374.0, + "logps/rejected": -344.0, + "loss": 0.3851, + "rewards/accuracies": 0.78125, + "rewards/chosen": 1.5625, + "rewards/margins": 6.6875, + "rewards/rejected": -5.09375, + "step": 128 + }, + { + "epoch": 0.15828220858895706, + "grad_norm": 31.69804344527413, + "learning_rate": 1.4355371900826447e-06, + "logits/chosen": -0.00439453125, + "logits/rejected": -0.1279296875, + "logps/chosen": -342.0, + "logps/rejected": -302.0, + "loss": 0.3139, + "rewards/accuracies": 0.8359375, + "rewards/chosen": 0.71484375, + "rewards/margins": 5.5, + "rewards/rejected": -4.78125, + "step": 129 + }, + { + "epoch": 0.15950920245398773, + "grad_norm": 39.02664413781982, + "learning_rate": 1.4349173553719008e-06, + "logits/chosen": -0.11328125, + "logits/rejected": -0.251953125, + "logps/chosen": -388.0, + "logps/rejected": -324.0, + "loss": 0.3918, + "rewards/accuracies": 0.75, + "rewards/chosen": 0.66015625, + "rewards/margins": 6.03125, + "rewards/rejected": -5.375, + "step": 130 + }, + { + "epoch": 0.1607361963190184, + "grad_norm": 33.65136979573152, + "learning_rate": 1.434297520661157e-06, + "logits/chosen": -0.2255859375, + "logits/rejected": -0.361328125, + "logps/chosen": -396.0, + "logps/rejected": -356.0, + "loss": 0.2429, + "rewards/accuracies": 0.84375, + "rewards/chosen": 1.3671875, + "rewards/margins": 6.96875, + "rewards/rejected": -5.59375, + "step": 131 + }, + { + "epoch": 0.1619631901840491, + "grad_norm": 40.70444309722836, + "learning_rate": 1.4336776859504133e-06, + "logits/chosen": -0.1845703125, + "logits/rejected": -0.3671875, + "logps/chosen": -392.0, + "logps/rejected": -336.0, + "loss": 0.3917, + "rewards/accuracies": 0.796875, + "rewards/chosen": 1.1796875, + "rewards/margins": 6.34375, + "rewards/rejected": -5.15625, + "step": 132 + }, + { + "epoch": 0.16319018404907976, + "grad_norm": 38.39829631153826, + "learning_rate": 1.4330578512396694e-06, + "logits/chosen": -0.220703125, + "logits/rejected": -0.375, + "logps/chosen": -388.0, + "logps/rejected": -342.0, + "loss": 0.3065, + "rewards/accuracies": 0.84375, + "rewards/chosen": 0.94921875, + "rewards/margins": 6.40625, + "rewards/rejected": -5.46875, + "step": 133 + }, + { + "epoch": 0.16441717791411042, + "grad_norm": 54.071557291057594, + "learning_rate": 1.4324380165289258e-06, + "logits/chosen": -0.1416015625, + "logits/rejected": -0.34765625, + "logps/chosen": -458.0, + "logps/rejected": -374.0, + "loss": 0.4193, + "rewards/accuracies": 0.796875, + "rewards/chosen": 0.81640625, + "rewards/margins": 5.4375, + "rewards/rejected": -4.625, + "step": 134 + }, + { + "epoch": 0.1656441717791411, + "grad_norm": 42.09400341898127, + "learning_rate": 1.431818181818182e-06, + "logits/chosen": -0.0140380859375, + "logits/rejected": -0.11767578125, + "logps/chosen": -378.0, + "logps/rejected": -330.0, + "loss": 0.4491, + "rewards/accuracies": 0.8046875, + "rewards/chosen": 0.16015625, + "rewards/margins": 5.21875, + "rewards/rejected": -5.0625, + "step": 135 + }, + { + "epoch": 0.1668711656441718, + "grad_norm": 34.790669543589566, + "learning_rate": 1.431198347107438e-06, + "logits/chosen": -0.259765625, + "logits/rejected": -0.3203125, + "logps/chosen": -392.0, + "logps/rejected": -322.0, + "loss": 0.3098, + "rewards/accuracies": 0.828125, + "rewards/chosen": 0.87890625, + "rewards/margins": 6.34375, + "rewards/rejected": -5.46875, + "step": 136 + }, + { + "epoch": 0.16809815950920245, + "grad_norm": 46.051678091055635, + "learning_rate": 1.4305785123966942e-06, + "logits/chosen": -0.020751953125, + "logits/rejected": -0.1669921875, + "logps/chosen": -398.0, + "logps/rejected": -320.0, + "loss": 0.4495, + "rewards/accuracies": 0.7890625, + "rewards/chosen": 0.609375, + "rewards/margins": 5.21875, + "rewards/rejected": -4.625, + "step": 137 + }, + { + "epoch": 0.16932515337423312, + "grad_norm": 39.60892351140069, + "learning_rate": 1.4299586776859504e-06, + "logits/chosen": -0.10546875, + "logits/rejected": -0.2421875, + "logps/chosen": -406.0, + "logps/rejected": -360.0, + "loss": 0.3662, + "rewards/accuracies": 0.796875, + "rewards/chosen": 0.56640625, + "rewards/margins": 5.59375, + "rewards/rejected": -5.0, + "step": 138 + }, + { + "epoch": 0.1705521472392638, + "grad_norm": 44.067268759137896, + "learning_rate": 1.4293388429752067e-06, + "logits/chosen": -0.10546875, + "logits/rejected": -0.2734375, + "logps/chosen": -418.0, + "logps/rejected": -380.0, + "loss": 0.4261, + "rewards/accuracies": 0.78125, + "rewards/chosen": 1.2109375, + "rewards/margins": 6.8125, + "rewards/rejected": -5.59375, + "step": 139 + }, + { + "epoch": 0.17177914110429449, + "grad_norm": 46.60040949853179, + "learning_rate": 1.4287190082644628e-06, + "logits/chosen": -0.07177734375, + "logits/rejected": -0.251953125, + "logps/chosen": -426.0, + "logps/rejected": -376.0, + "loss": 0.4219, + "rewards/accuracies": 0.765625, + "rewards/chosen": 0.9453125, + "rewards/margins": 6.125, + "rewards/rejected": -5.1875, + "step": 140 + }, + { + "epoch": 0.17300613496932515, + "grad_norm": 34.208928830023055, + "learning_rate": 1.428099173553719e-06, + "logits/chosen": -0.05126953125, + "logits/rejected": -0.173828125, + "logps/chosen": -358.0, + "logps/rejected": -338.0, + "loss": 0.3019, + "rewards/accuracies": 0.8515625, + "rewards/chosen": 0.99609375, + "rewards/margins": 6.28125, + "rewards/rejected": -5.28125, + "step": 141 + }, + { + "epoch": 0.17423312883435582, + "grad_norm": 41.96948447879753, + "learning_rate": 1.4274793388429753e-06, + "logits/chosen": -0.0927734375, + "logits/rejected": -0.27734375, + "logps/chosen": -400.0, + "logps/rejected": -354.0, + "loss": 0.4327, + "rewards/accuracies": 0.796875, + "rewards/chosen": 0.98828125, + "rewards/margins": 6.375, + "rewards/rejected": -5.375, + "step": 142 + }, + { + "epoch": 0.1754601226993865, + "grad_norm": 39.736356252739405, + "learning_rate": 1.4268595041322315e-06, + "logits/chosen": -0.12451171875, + "logits/rejected": -0.33984375, + "logps/chosen": -384.0, + "logps/rejected": -328.0, + "loss": 0.3314, + "rewards/accuracies": 0.828125, + "rewards/chosen": 0.64453125, + "rewards/margins": 6.28125, + "rewards/rejected": -5.625, + "step": 143 + }, + { + "epoch": 0.17668711656441718, + "grad_norm": 40.38397393664774, + "learning_rate": 1.4262396694214876e-06, + "logits/chosen": -0.224609375, + "logits/rejected": -0.3984375, + "logps/chosen": -420.0, + "logps/rejected": -356.0, + "loss": 0.3711, + "rewards/accuracies": 0.828125, + "rewards/chosen": 1.5625, + "rewards/margins": 6.75, + "rewards/rejected": -5.1875, + "step": 144 + }, + { + "epoch": 0.17791411042944785, + "grad_norm": 36.78367346174047, + "learning_rate": 1.425619834710744e-06, + "logits/chosen": -0.20703125, + "logits/rejected": -0.349609375, + "logps/chosen": -380.0, + "logps/rejected": -348.0, + "loss": 0.2992, + "rewards/accuracies": 0.859375, + "rewards/chosen": 1.5625, + "rewards/margins": 6.78125, + "rewards/rejected": -5.21875, + "step": 145 + }, + { + "epoch": 0.17914110429447852, + "grad_norm": 30.78038936070514, + "learning_rate": 1.425e-06, + "logits/chosen": -0.2021484375, + "logits/rejected": -0.30859375, + "logps/chosen": -388.0, + "logps/rejected": -334.0, + "loss": 0.2674, + "rewards/accuracies": 0.875, + "rewards/chosen": 0.53125, + "rewards/margins": 6.03125, + "rewards/rejected": -5.5, + "step": 146 + }, + { + "epoch": 0.18036809815950922, + "grad_norm": 33.88526702598482, + "learning_rate": 1.4243801652892563e-06, + "logits/chosen": -0.2001953125, + "logits/rejected": -0.349609375, + "logps/chosen": -388.0, + "logps/rejected": -332.0, + "loss": 0.3151, + "rewards/accuracies": 0.828125, + "rewards/chosen": 1.5625, + "rewards/margins": 6.875, + "rewards/rejected": -5.3125, + "step": 147 + }, + { + "epoch": 0.18159509202453988, + "grad_norm": 49.54475790087749, + "learning_rate": 1.4237603305785124e-06, + "logits/chosen": -0.2197265625, + "logits/rejected": -0.34765625, + "logps/chosen": -416.0, + "logps/rejected": -368.0, + "loss": 0.4146, + "rewards/accuracies": 0.7734375, + "rewards/chosen": 0.89453125, + "rewards/margins": 6.25, + "rewards/rejected": -5.34375, + "step": 148 + }, + { + "epoch": 0.18282208588957055, + "grad_norm": 43.24095049009772, + "learning_rate": 1.4231404958677685e-06, + "logits/chosen": -0.09521484375, + "logits/rejected": -0.244140625, + "logps/chosen": -406.0, + "logps/rejected": -384.0, + "loss": 0.3173, + "rewards/accuracies": 0.859375, + "rewards/chosen": 1.5859375, + "rewards/margins": 6.625, + "rewards/rejected": -5.03125, + "step": 149 + }, + { + "epoch": 0.18404907975460122, + "grad_norm": 37.7299116363035, + "learning_rate": 1.422520661157025e-06, + "logits/chosen": -0.1181640625, + "logits/rejected": -0.291015625, + "logps/chosen": -424.0, + "logps/rejected": -346.0, + "loss": 0.344, + "rewards/accuracies": 0.8359375, + "rewards/chosen": 1.296875, + "rewards/margins": 5.84375, + "rewards/rejected": -4.53125, + "step": 150 + }, + { + "epoch": 0.18527607361963191, + "grad_norm": 36.98087652686202, + "learning_rate": 1.421900826446281e-06, + "logits/chosen": -0.1376953125, + "logits/rejected": -0.263671875, + "logps/chosen": -382.0, + "logps/rejected": -326.0, + "loss": 0.3467, + "rewards/accuracies": 0.8046875, + "rewards/chosen": 1.1796875, + "rewards/margins": 5.8125, + "rewards/rejected": -4.625, + "step": 151 + }, + { + "epoch": 0.18650306748466258, + "grad_norm": 33.96526048608977, + "learning_rate": 1.4212809917355372e-06, + "logits/chosen": -0.2255859375, + "logits/rejected": -0.462890625, + "logps/chosen": -402.0, + "logps/rejected": -316.0, + "loss": 0.322, + "rewards/accuracies": 0.8515625, + "rewards/chosen": 0.83984375, + "rewards/margins": 6.5625, + "rewards/rejected": -5.71875, + "step": 152 + }, + { + "epoch": 0.18773006134969325, + "grad_norm": 34.59925431133825, + "learning_rate": 1.4206611570247935e-06, + "logits/chosen": -0.1875, + "logits/rejected": -0.298828125, + "logps/chosen": -350.0, + "logps/rejected": -320.0, + "loss": 0.3333, + "rewards/accuracies": 0.84375, + "rewards/chosen": 1.390625, + "rewards/margins": 6.28125, + "rewards/rejected": -4.875, + "step": 153 + }, + { + "epoch": 0.18895705521472392, + "grad_norm": 45.111496339674666, + "learning_rate": 1.4200413223140497e-06, + "logits/chosen": -0.060546875, + "logits/rejected": -0.2431640625, + "logps/chosen": -424.0, + "logps/rejected": -340.0, + "loss": 0.3836, + "rewards/accuracies": 0.7734375, + "rewards/chosen": 1.703125, + "rewards/margins": 6.09375, + "rewards/rejected": -4.375, + "step": 154 + }, + { + "epoch": 0.1901840490797546, + "grad_norm": 28.559326947209748, + "learning_rate": 1.4194214876033058e-06, + "logits/chosen": -0.1748046875, + "logits/rejected": -0.41796875, + "logps/chosen": -382.0, + "logps/rejected": -320.0, + "loss": 0.2127, + "rewards/accuracies": 0.90625, + "rewards/chosen": 2.359375, + "rewards/margins": 7.375, + "rewards/rejected": -5.03125, + "step": 155 + }, + { + "epoch": 0.19141104294478528, + "grad_norm": 41.969901464316806, + "learning_rate": 1.418801652892562e-06, + "logits/chosen": -0.06787109375, + "logits/rejected": -0.1748046875, + "logps/chosen": -370.0, + "logps/rejected": -350.0, + "loss": 0.3849, + "rewards/accuracies": 0.8203125, + "rewards/chosen": 1.1171875, + "rewards/margins": 5.59375, + "rewards/rejected": -4.46875, + "step": 156 + }, + { + "epoch": 0.19263803680981595, + "grad_norm": 35.115688822084685, + "learning_rate": 1.418181818181818e-06, + "logits/chosen": -0.05810546875, + "logits/rejected": -0.1796875, + "logps/chosen": -394.0, + "logps/rejected": -310.0, + "loss": 0.3364, + "rewards/accuracies": 0.828125, + "rewards/chosen": 1.2421875, + "rewards/margins": 5.875, + "rewards/rejected": -4.625, + "step": 157 + }, + { + "epoch": 0.19386503067484662, + "grad_norm": 29.216146857412586, + "learning_rate": 1.4175619834710745e-06, + "logits/chosen": -0.2734375, + "logits/rejected": -0.44921875, + "logps/chosen": -388.0, + "logps/rejected": -324.0, + "loss": 0.2283, + "rewards/accuracies": 0.8515625, + "rewards/chosen": 1.390625, + "rewards/margins": 6.59375, + "rewards/rejected": -5.21875, + "step": 158 + }, + { + "epoch": 0.1950920245398773, + "grad_norm": 39.305920966554595, + "learning_rate": 1.4169421487603306e-06, + "logits/chosen": -0.1630859375, + "logits/rejected": -0.28125, + "logps/chosen": -360.0, + "logps/rejected": -328.0, + "loss": 0.3649, + "rewards/accuracies": 0.84375, + "rewards/chosen": 1.359375, + "rewards/margins": 5.71875, + "rewards/rejected": -4.34375, + "step": 159 + }, + { + "epoch": 0.19631901840490798, + "grad_norm": 26.74081618489141, + "learning_rate": 1.416322314049587e-06, + "logits/chosen": -0.047607421875, + "logits/rejected": -0.26953125, + "logps/chosen": -354.0, + "logps/rejected": -310.0, + "loss": 0.2471, + "rewards/accuracies": 0.890625, + "rewards/chosen": 0.94921875, + "rewards/margins": 6.125, + "rewards/rejected": -5.15625, + "step": 160 + }, + { + "epoch": 0.19754601226993865, + "grad_norm": 32.344289601166764, + "learning_rate": 1.415702479338843e-06, + "logits/chosen": -0.2294921875, + "logits/rejected": -0.4296875, + "logps/chosen": -386.0, + "logps/rejected": -302.0, + "loss": 0.2831, + "rewards/accuracies": 0.8515625, + "rewards/chosen": 1.9453125, + "rewards/margins": 6.6875, + "rewards/rejected": -4.71875, + "step": 161 + }, + { + "epoch": 0.19877300613496932, + "grad_norm": 39.80838608230206, + "learning_rate": 1.4150826446280992e-06, + "logits/chosen": -0.1357421875, + "logits/rejected": -0.2060546875, + "logps/chosen": -350.0, + "logps/rejected": -340.0, + "loss": 0.4018, + "rewards/accuracies": 0.8046875, + "rewards/chosen": 1.4453125, + "rewards/margins": 5.8125, + "rewards/rejected": -4.375, + "step": 162 + }, + { + "epoch": 0.2, + "grad_norm": 38.00229232317159, + "learning_rate": 1.4144628099173554e-06, + "logits/chosen": -0.07373046875, + "logits/rejected": -0.220703125, + "logps/chosen": -392.0, + "logps/rejected": -344.0, + "loss": 0.2926, + "rewards/accuracies": 0.8671875, + "rewards/chosen": 1.421875, + "rewards/margins": 6.25, + "rewards/rejected": -4.8125, + "step": 163 + }, + { + "epoch": 0.20122699386503068, + "grad_norm": 31.02665679742822, + "learning_rate": 1.4138429752066115e-06, + "logits/chosen": -0.236328125, + "logits/rejected": -0.466796875, + "logps/chosen": -392.0, + "logps/rejected": -310.0, + "loss": 0.2813, + "rewards/accuracies": 0.859375, + "rewards/chosen": 2.15625, + "rewards/margins": 7.34375, + "rewards/rejected": -5.1875, + "step": 164 + }, + { + "epoch": 0.20245398773006135, + "grad_norm": 43.699969414714815, + "learning_rate": 1.4132231404958679e-06, + "logits/chosen": -0.021728515625, + "logits/rejected": -0.212890625, + "logps/chosen": -444.0, + "logps/rejected": -326.0, + "loss": 0.377, + "rewards/accuracies": 0.828125, + "rewards/chosen": 1.5078125, + "rewards/margins": 6.3125, + "rewards/rejected": -4.78125, + "step": 165 + }, + { + "epoch": 0.20368098159509201, + "grad_norm": 38.30377721360019, + "learning_rate": 1.412603305785124e-06, + "logits/chosen": -0.09375, + "logits/rejected": -0.314453125, + "logps/chosen": -388.0, + "logps/rejected": -336.0, + "loss": 0.3851, + "rewards/accuracies": 0.828125, + "rewards/chosen": 1.9140625, + "rewards/margins": 6.53125, + "rewards/rejected": -4.625, + "step": 166 + }, + { + "epoch": 0.2049079754601227, + "grad_norm": 38.75636778509963, + "learning_rate": 1.4119834710743801e-06, + "logits/chosen": -0.06591796875, + "logits/rejected": -0.2080078125, + "logps/chosen": -398.0, + "logps/rejected": -318.0, + "loss": 0.3749, + "rewards/accuracies": 0.7890625, + "rewards/chosen": 2.21875, + "rewards/margins": 6.875, + "rewards/rejected": -4.65625, + "step": 167 + }, + { + "epoch": 0.20613496932515338, + "grad_norm": 37.779783793801464, + "learning_rate": 1.4113636363636365e-06, + "logits/chosen": -0.049560546875, + "logits/rejected": -0.1494140625, + "logps/chosen": -348.0, + "logps/rejected": -292.0, + "loss": 0.3337, + "rewards/accuracies": 0.84375, + "rewards/chosen": 0.99609375, + "rewards/margins": 6.46875, + "rewards/rejected": -5.46875, + "step": 168 + }, + { + "epoch": 0.20736196319018405, + "grad_norm": 34.571770475148426, + "learning_rate": 1.4107438016528926e-06, + "logits/chosen": -0.1259765625, + "logits/rejected": -0.27734375, + "logps/chosen": -372.0, + "logps/rejected": -366.0, + "loss": 0.3006, + "rewards/accuracies": 0.8671875, + "rewards/chosen": 1.0078125, + "rewards/margins": 6.25, + "rewards/rejected": -5.25, + "step": 169 + }, + { + "epoch": 0.2085889570552147, + "grad_norm": 42.205138483103795, + "learning_rate": 1.4101239669421488e-06, + "logits/chosen": -0.1123046875, + "logits/rejected": -0.349609375, + "logps/chosen": -450.0, + "logps/rejected": -346.0, + "loss": 0.338, + "rewards/accuracies": 0.8125, + "rewards/chosen": 1.875, + "rewards/margins": 7.34375, + "rewards/rejected": -5.46875, + "step": 170 + }, + { + "epoch": 0.2098159509202454, + "grad_norm": 29.828895289514126, + "learning_rate": 1.4095041322314051e-06, + "logits/chosen": -0.10888671875, + "logits/rejected": -0.279296875, + "logps/chosen": -384.0, + "logps/rejected": -322.0, + "loss": 0.3086, + "rewards/accuracies": 0.859375, + "rewards/chosen": 1.5546875, + "rewards/margins": 7.46875, + "rewards/rejected": -5.90625, + "step": 171 + }, + { + "epoch": 0.21104294478527608, + "grad_norm": 30.980573058674597, + "learning_rate": 1.408884297520661e-06, + "logits/chosen": -0.062255859375, + "logits/rejected": -0.2216796875, + "logps/chosen": -438.0, + "logps/rejected": -354.0, + "loss": 0.2599, + "rewards/accuracies": 0.859375, + "rewards/chosen": 1.84375, + "rewards/margins": 7.84375, + "rewards/rejected": -6.0, + "step": 172 + }, + { + "epoch": 0.21226993865030674, + "grad_norm": 25.22446362010358, + "learning_rate": 1.4082644628099174e-06, + "logits/chosen": -0.123046875, + "logits/rejected": -0.353515625, + "logps/chosen": -428.0, + "logps/rejected": -362.0, + "loss": 0.1997, + "rewards/accuracies": 0.890625, + "rewards/chosen": 1.703125, + "rewards/margins": 8.25, + "rewards/rejected": -6.5625, + "step": 173 + }, + { + "epoch": 0.2134969325153374, + "grad_norm": 43.79333405262764, + "learning_rate": 1.4076446280991736e-06, + "logits/chosen": 0.05126953125, + "logits/rejected": -0.158203125, + "logps/chosen": -384.0, + "logps/rejected": -348.0, + "loss": 0.5204, + "rewards/accuracies": 0.765625, + "rewards/chosen": 0.78125, + "rewards/margins": 6.8125, + "rewards/rejected": -6.03125, + "step": 174 + }, + { + "epoch": 0.2147239263803681, + "grad_norm": 45.27680571467687, + "learning_rate": 1.4070247933884297e-06, + "logits/chosen": -0.10546875, + "logits/rejected": -0.220703125, + "logps/chosen": -402.0, + "logps/rejected": -366.0, + "loss": 0.4454, + "rewards/accuracies": 0.8125, + "rewards/chosen": 1.09375, + "rewards/margins": 7.34375, + "rewards/rejected": -6.25, + "step": 175 + }, + { + "epoch": 0.21595092024539878, + "grad_norm": 35.16824036294685, + "learning_rate": 1.406404958677686e-06, + "logits/chosen": -0.1298828125, + "logits/rejected": -0.30859375, + "logps/chosen": -434.0, + "logps/rejected": -378.0, + "loss": 0.3443, + "rewards/accuracies": 0.796875, + "rewards/chosen": 1.3359375, + "rewards/margins": 7.65625, + "rewards/rejected": -6.3125, + "step": 176 + }, + { + "epoch": 0.21717791411042944, + "grad_norm": 27.30910780814845, + "learning_rate": 1.4057851239669422e-06, + "logits/chosen": -0.08251953125, + "logits/rejected": -0.30859375, + "logps/chosen": -416.0, + "logps/rejected": -356.0, + "loss": 0.2873, + "rewards/accuracies": 0.875, + "rewards/chosen": 2.0625, + "rewards/margins": 8.8125, + "rewards/rejected": -6.71875, + "step": 177 + }, + { + "epoch": 0.2184049079754601, + "grad_norm": 37.84132418939681, + "learning_rate": 1.4051652892561983e-06, + "logits/chosen": -0.07275390625, + "logits/rejected": -0.216796875, + "logps/chosen": -330.0, + "logps/rejected": -318.0, + "loss": 0.4137, + "rewards/accuracies": 0.828125, + "rewards/chosen": 0.1806640625, + "rewards/margins": 6.75, + "rewards/rejected": -6.59375, + "step": 178 + }, + { + "epoch": 0.2196319018404908, + "grad_norm": 32.98852515847511, + "learning_rate": 1.4045454545454547e-06, + "logits/chosen": -0.1064453125, + "logits/rejected": -0.31640625, + "logps/chosen": -372.0, + "logps/rejected": -324.0, + "loss": 0.3252, + "rewards/accuracies": 0.859375, + "rewards/chosen": 0.671875, + "rewards/margins": 6.875, + "rewards/rejected": -6.1875, + "step": 179 + }, + { + "epoch": 0.22085889570552147, + "grad_norm": 38.78908549807395, + "learning_rate": 1.4039256198347108e-06, + "logits/chosen": -0.099609375, + "logits/rejected": -0.21484375, + "logps/chosen": -356.0, + "logps/rejected": -314.0, + "loss": 0.3669, + "rewards/accuracies": 0.828125, + "rewards/chosen": 1.234375, + "rewards/margins": 6.46875, + "rewards/rejected": -5.21875, + "step": 180 + }, + { + "epoch": 0.22208588957055214, + "grad_norm": 35.32530991965649, + "learning_rate": 1.403305785123967e-06, + "logits/chosen": -0.09423828125, + "logits/rejected": -0.2412109375, + "logps/chosen": -410.0, + "logps/rejected": -386.0, + "loss": 0.3373, + "rewards/accuracies": 0.859375, + "rewards/chosen": 1.6171875, + "rewards/margins": 6.875, + "rewards/rejected": -5.25, + "step": 181 + }, + { + "epoch": 0.2233128834355828, + "grad_norm": 36.9736807793252, + "learning_rate": 1.4026859504132231e-06, + "logits/chosen": -0.12890625, + "logits/rejected": -0.2470703125, + "logps/chosen": -370.0, + "logps/rejected": -322.0, + "loss": 0.3661, + "rewards/accuracies": 0.8515625, + "rewards/chosen": 1.109375, + "rewards/margins": 5.875, + "rewards/rejected": -4.75, + "step": 182 + }, + { + "epoch": 0.2245398773006135, + "grad_norm": 40.8878376560248, + "learning_rate": 1.4020661157024793e-06, + "logits/chosen": 0.00762939453125, + "logits/rejected": -0.1669921875, + "logps/chosen": -412.0, + "logps/rejected": -378.0, + "loss": 0.3866, + "rewards/accuracies": 0.8203125, + "rewards/chosen": 0.95703125, + "rewards/margins": 4.90625, + "rewards/rejected": -3.953125, + "step": 183 + }, + { + "epoch": 0.22576687116564417, + "grad_norm": 42.156241310731595, + "learning_rate": 1.4014462809917356e-06, + "logits/chosen": -0.0255126953125, + "logits/rejected": -0.1328125, + "logps/chosen": -390.0, + "logps/rejected": -340.0, + "loss": 0.4166, + "rewards/accuracies": 0.7890625, + "rewards/chosen": 0.87109375, + "rewards/margins": 5.1875, + "rewards/rejected": -4.3125, + "step": 184 + }, + { + "epoch": 0.22699386503067484, + "grad_norm": 44.943258114663124, + "learning_rate": 1.4008264462809917e-06, + "logits/chosen": -0.06787109375, + "logits/rejected": -0.1220703125, + "logps/chosen": -382.0, + "logps/rejected": -350.0, + "loss": 0.4011, + "rewards/accuracies": 0.78125, + "rewards/chosen": 1.9765625, + "rewards/margins": 5.8125, + "rewards/rejected": -3.828125, + "step": 185 + }, + { + "epoch": 0.2282208588957055, + "grad_norm": 33.47267300169336, + "learning_rate": 1.4002066115702479e-06, + "logits/chosen": -0.08203125, + "logits/rejected": -0.1982421875, + "logps/chosen": -336.0, + "logps/rejected": -312.0, + "loss": 0.3163, + "rewards/accuracies": 0.859375, + "rewards/chosen": 1.4453125, + "rewards/margins": 5.78125, + "rewards/rejected": -4.3125, + "step": 186 + }, + { + "epoch": 0.2294478527607362, + "grad_norm": 32.03885108074069, + "learning_rate": 1.3995867768595042e-06, + "logits/chosen": -0.07080078125, + "logits/rejected": -0.2314453125, + "logps/chosen": -370.0, + "logps/rejected": -340.0, + "loss": 0.258, + "rewards/accuracies": 0.8671875, + "rewards/chosen": 1.6328125, + "rewards/margins": 6.4375, + "rewards/rejected": -4.78125, + "step": 187 + }, + { + "epoch": 0.23067484662576687, + "grad_norm": 35.158897439200416, + "learning_rate": 1.3989669421487604e-06, + "logits/chosen": 0.0228271484375, + "logits/rejected": -0.0615234375, + "logps/chosen": -330.0, + "logps/rejected": -292.0, + "loss": 0.3609, + "rewards/accuracies": 0.8046875, + "rewards/chosen": 1.40625, + "rewards/margins": 5.875, + "rewards/rejected": -4.46875, + "step": 188 + }, + { + "epoch": 0.23190184049079754, + "grad_norm": 32.15934523587848, + "learning_rate": 1.3983471074380167e-06, + "logits/chosen": -0.080078125, + "logits/rejected": -0.263671875, + "logps/chosen": -396.0, + "logps/rejected": -322.0, + "loss": 0.2271, + "rewards/accuracies": 0.890625, + "rewards/chosen": 2.046875, + "rewards/margins": 6.78125, + "rewards/rejected": -4.71875, + "step": 189 + }, + { + "epoch": 0.2331288343558282, + "grad_norm": 32.11586724143221, + "learning_rate": 1.3977272727272727e-06, + "logits/chosen": 0.041748046875, + "logits/rejected": -0.10693359375, + "logps/chosen": -340.0, + "logps/rejected": -316.0, + "loss": 0.3271, + "rewards/accuracies": 0.8515625, + "rewards/chosen": 1.4140625, + "rewards/margins": 6.0, + "rewards/rejected": -4.59375, + "step": 190 + }, + { + "epoch": 0.2343558282208589, + "grad_norm": 46.26761231273777, + "learning_rate": 1.3971074380165288e-06, + "logits/chosen": -0.0390625, + "logits/rejected": -0.220703125, + "logps/chosen": -400.0, + "logps/rejected": -306.0, + "loss": 0.3906, + "rewards/accuracies": 0.8046875, + "rewards/chosen": 1.828125, + "rewards/margins": 6.90625, + "rewards/rejected": -5.0625, + "step": 191 + }, + { + "epoch": 0.23558282208588957, + "grad_norm": 39.15535099753523, + "learning_rate": 1.3964876033057852e-06, + "logits/chosen": -0.052490234375, + "logits/rejected": -0.2490234375, + "logps/chosen": -388.0, + "logps/rejected": -302.0, + "loss": 0.3379, + "rewards/accuracies": 0.8046875, + "rewards/chosen": 1.1328125, + "rewards/margins": 5.65625, + "rewards/rejected": -4.5, + "step": 192 + }, + { + "epoch": 0.23680981595092024, + "grad_norm": 31.3365855799527, + "learning_rate": 1.3958677685950413e-06, + "logits/chosen": -0.1767578125, + "logits/rejected": -0.423828125, + "logps/chosen": -410.0, + "logps/rejected": -346.0, + "loss": 0.2485, + "rewards/accuracies": 0.8828125, + "rewards/chosen": 1.7578125, + "rewards/margins": 7.375, + "rewards/rejected": -5.625, + "step": 193 + }, + { + "epoch": 0.23803680981595093, + "grad_norm": 38.5490731428379, + "learning_rate": 1.3952479338842977e-06, + "logits/chosen": -0.21875, + "logits/rejected": -0.326171875, + "logps/chosen": -394.0, + "logps/rejected": -336.0, + "loss": 0.296, + "rewards/accuracies": 0.8203125, + "rewards/chosen": 0.89453125, + "rewards/margins": 6.875, + "rewards/rejected": -5.96875, + "step": 194 + }, + { + "epoch": 0.2392638036809816, + "grad_norm": 35.180225705298604, + "learning_rate": 1.3946280991735538e-06, + "logits/chosen": -0.031982421875, + "logits/rejected": -0.2890625, + "logps/chosen": -408.0, + "logps/rejected": -348.0, + "loss": 0.2727, + "rewards/accuracies": 0.8671875, + "rewards/chosen": 2.046875, + "rewards/margins": 8.1875, + "rewards/rejected": -6.125, + "step": 195 + }, + { + "epoch": 0.24049079754601227, + "grad_norm": 37.22166797602668, + "learning_rate": 1.39400826446281e-06, + "logits/chosen": 0.0157470703125, + "logits/rejected": -0.1845703125, + "logps/chosen": -394.0, + "logps/rejected": -308.0, + "loss": 0.3333, + "rewards/accuracies": 0.8359375, + "rewards/chosen": 1.1171875, + "rewards/margins": 6.65625, + "rewards/rejected": -5.53125, + "step": 196 + }, + { + "epoch": 0.24171779141104294, + "grad_norm": 28.562917988218086, + "learning_rate": 1.3933884297520663e-06, + "logits/chosen": -0.115234375, + "logits/rejected": -0.34765625, + "logps/chosen": -380.0, + "logps/rejected": -334.0, + "loss": 0.3009, + "rewards/accuracies": 0.8828125, + "rewards/chosen": 1.3359375, + "rewards/margins": 7.125, + "rewards/rejected": -5.78125, + "step": 197 + }, + { + "epoch": 0.24294478527607363, + "grad_norm": 36.54466818998776, + "learning_rate": 1.3927685950413224e-06, + "logits/chosen": -0.0174560546875, + "logits/rejected": -0.10205078125, + "logps/chosen": -346.0, + "logps/rejected": -276.0, + "loss": 0.4465, + "rewards/accuracies": 0.8046875, + "rewards/chosen": 0.625, + "rewards/margins": 6.0625, + "rewards/rejected": -5.4375, + "step": 198 + }, + { + "epoch": 0.2441717791411043, + "grad_norm": 25.79094354827764, + "learning_rate": 1.3921487603305786e-06, + "logits/chosen": -0.05615234375, + "logits/rejected": -0.26171875, + "logps/chosen": -434.0, + "logps/rejected": -330.0, + "loss": 0.1969, + "rewards/accuracies": 0.8984375, + "rewards/chosen": 1.7265625, + "rewards/margins": 7.84375, + "rewards/rejected": -6.125, + "step": 199 + }, + { + "epoch": 0.24539877300613497, + "grad_norm": 34.674086597448465, + "learning_rate": 1.3915289256198347e-06, + "logits/chosen": -0.037353515625, + "logits/rejected": -0.25390625, + "logps/chosen": -442.0, + "logps/rejected": -392.0, + "loss": 0.3066, + "rewards/accuracies": 0.8515625, + "rewards/chosen": 1.2734375, + "rewards/margins": 7.4375, + "rewards/rejected": -6.15625, + "step": 200 + }, + { + "epoch": 0.24662576687116564, + "grad_norm": 29.903922318940893, + "learning_rate": 1.3909090909090909e-06, + "logits/chosen": -0.03662109375, + "logits/rejected": -0.259765625, + "logps/chosen": -404.0, + "logps/rejected": -330.0, + "loss": 0.2629, + "rewards/accuracies": 0.8828125, + "rewards/chosen": 1.109375, + "rewards/margins": 7.0625, + "rewards/rejected": -5.9375, + "step": 201 + }, + { + "epoch": 0.24785276073619633, + "grad_norm": 33.833661458600005, + "learning_rate": 1.3902892561983472e-06, + "logits/chosen": -0.02001953125, + "logits/rejected": -0.2333984375, + "logps/chosen": -432.0, + "logps/rejected": -340.0, + "loss": 0.2977, + "rewards/accuracies": 0.8203125, + "rewards/chosen": 1.03125, + "rewards/margins": 6.90625, + "rewards/rejected": -5.875, + "step": 202 + }, + { + "epoch": 0.249079754601227, + "grad_norm": 37.642944458655066, + "learning_rate": 1.3896694214876034e-06, + "logits/chosen": 0.0272216796875, + "logits/rejected": -0.1474609375, + "logps/chosen": -392.0, + "logps/rejected": -354.0, + "loss": 0.482, + "rewards/accuracies": 0.796875, + "rewards/chosen": 0.96875, + "rewards/margins": 6.03125, + "rewards/rejected": -5.0625, + "step": 203 + }, + { + "epoch": 0.25030674846625767, + "grad_norm": 35.44282157848665, + "learning_rate": 1.3890495867768595e-06, + "logits/chosen": 0.0390625, + "logits/rejected": -0.1005859375, + "logps/chosen": -356.0, + "logps/rejected": -290.0, + "loss": 0.4211, + "rewards/accuracies": 0.8359375, + "rewards/chosen": 0.6484375, + "rewards/margins": 6.4375, + "rewards/rejected": -5.78125, + "step": 204 + }, + { + "epoch": 0.25153374233128833, + "grad_norm": 38.48812661842802, + "learning_rate": 1.3884297520661158e-06, + "logits/chosen": 0.08984375, + "logits/rejected": -0.04833984375, + "logps/chosen": -390.0, + "logps/rejected": -316.0, + "loss": 0.4435, + "rewards/accuracies": 0.8046875, + "rewards/chosen": 0.8671875, + "rewards/margins": 6.0, + "rewards/rejected": -5.125, + "step": 205 + }, + { + "epoch": 0.252760736196319, + "grad_norm": 26.640868179250116, + "learning_rate": 1.387809917355372e-06, + "logits/chosen": -0.036376953125, + "logits/rejected": -0.3125, + "logps/chosen": -434.0, + "logps/rejected": -350.0, + "loss": 0.2375, + "rewards/accuracies": 0.9140625, + "rewards/chosen": 1.203125, + "rewards/margins": 7.46875, + "rewards/rejected": -6.28125, + "step": 206 + }, + { + "epoch": 0.25398773006134967, + "grad_norm": 37.72822731879787, + "learning_rate": 1.3871900826446281e-06, + "logits/chosen": -0.16015625, + "logits/rejected": -0.353515625, + "logps/chosen": -434.0, + "logps/rejected": -376.0, + "loss": 0.4045, + "rewards/accuracies": 0.828125, + "rewards/chosen": 1.21875, + "rewards/margins": 6.75, + "rewards/rejected": -5.53125, + "step": 207 + }, + { + "epoch": 0.2552147239263804, + "grad_norm": 39.99822195882057, + "learning_rate": 1.3865702479338843e-06, + "logits/chosen": -0.12255859375, + "logits/rejected": -0.2470703125, + "logps/chosen": -396.0, + "logps/rejected": -340.0, + "loss": 0.4402, + "rewards/accuracies": 0.765625, + "rewards/chosen": 0.83984375, + "rewards/margins": 6.53125, + "rewards/rejected": -5.6875, + "step": 208 + }, + { + "epoch": 0.25644171779141106, + "grad_norm": 31.995520167951547, + "learning_rate": 1.3859504132231404e-06, + "logits/chosen": -0.125, + "logits/rejected": -0.2314453125, + "logps/chosen": -398.0, + "logps/rejected": -352.0, + "loss": 0.3203, + "rewards/accuracies": 0.84375, + "rewards/chosen": 0.76171875, + "rewards/margins": 6.34375, + "rewards/rejected": -5.5625, + "step": 209 + }, + { + "epoch": 0.25766871165644173, + "grad_norm": 42.71588448177596, + "learning_rate": 1.3853305785123968e-06, + "logits/chosen": -0.0732421875, + "logits/rejected": -0.1806640625, + "logps/chosen": -400.0, + "logps/rejected": -382.0, + "loss": 0.4413, + "rewards/accuracies": 0.8203125, + "rewards/chosen": 0.314453125, + "rewards/margins": 5.90625, + "rewards/rejected": -5.59375, + "step": 210 + }, + { + "epoch": 0.2588957055214724, + "grad_norm": 35.38452290000429, + "learning_rate": 1.384710743801653e-06, + "logits/chosen": -0.1806640625, + "logits/rejected": -0.375, + "logps/chosen": -438.0, + "logps/rejected": -362.0, + "loss": 0.3649, + "rewards/accuracies": 0.8125, + "rewards/chosen": 0.9375, + "rewards/margins": 6.84375, + "rewards/rejected": -5.90625, + "step": 211 + }, + { + "epoch": 0.26012269938650306, + "grad_norm": 35.00437697538439, + "learning_rate": 1.384090909090909e-06, + "logits/chosen": -0.099609375, + "logits/rejected": -0.251953125, + "logps/chosen": -404.0, + "logps/rejected": -328.0, + "loss": 0.3606, + "rewards/accuracies": 0.8203125, + "rewards/chosen": 1.125, + "rewards/margins": 6.6875, + "rewards/rejected": -5.5625, + "step": 212 + }, + { + "epoch": 0.26134969325153373, + "grad_norm": 40.33706711227177, + "learning_rate": 1.3834710743801654e-06, + "logits/chosen": -0.031982421875, + "logits/rejected": -0.0888671875, + "logps/chosen": -352.0, + "logps/rejected": -296.0, + "loss": 0.3034, + "rewards/accuracies": 0.828125, + "rewards/chosen": 0.4765625, + "rewards/margins": 5.78125, + "rewards/rejected": -5.3125, + "step": 213 + }, + { + "epoch": 0.2625766871165644, + "grad_norm": 36.278922089821435, + "learning_rate": 1.3828512396694215e-06, + "logits/chosen": -0.0152587890625, + "logits/rejected": -0.171875, + "logps/chosen": -376.0, + "logps/rejected": -326.0, + "loss": 0.3538, + "rewards/accuracies": 0.796875, + "rewards/chosen": 1.1328125, + "rewards/margins": 6.375, + "rewards/rejected": -5.25, + "step": 214 + }, + { + "epoch": 0.26380368098159507, + "grad_norm": 39.61373791294869, + "learning_rate": 1.3822314049586777e-06, + "logits/chosen": 0.0224609375, + "logits/rejected": -0.06689453125, + "logps/chosen": -356.0, + "logps/rejected": -316.0, + "loss": 0.3393, + "rewards/accuracies": 0.8203125, + "rewards/chosen": 0.55859375, + "rewards/margins": 5.5625, + "rewards/rejected": -5.0, + "step": 215 + }, + { + "epoch": 0.2650306748466258, + "grad_norm": 36.52111248304287, + "learning_rate": 1.381611570247934e-06, + "logits/chosen": -0.04443359375, + "logits/rejected": -0.1572265625, + "logps/chosen": -334.0, + "logps/rejected": -302.0, + "loss": 0.3612, + "rewards/accuracies": 0.8203125, + "rewards/chosen": 0.42578125, + "rewards/margins": 5.375, + "rewards/rejected": -4.96875, + "step": 216 + }, + { + "epoch": 0.26625766871165646, + "grad_norm": 35.893159477690816, + "learning_rate": 1.38099173553719e-06, + "logits/chosen": 0.05517578125, + "logits/rejected": -0.08154296875, + "logps/chosen": -402.0, + "logps/rejected": -316.0, + "loss": 0.4174, + "rewards/accuracies": 0.7734375, + "rewards/chosen": 1.2890625, + "rewards/margins": 6.0625, + "rewards/rejected": -4.78125, + "step": 217 + }, + { + "epoch": 0.2674846625766871, + "grad_norm": 43.28791050351139, + "learning_rate": 1.3803719008264463e-06, + "logits/chosen": 0.0108642578125, + "logits/rejected": -0.0908203125, + "logps/chosen": -384.0, + "logps/rejected": -344.0, + "loss": 0.514, + "rewards/accuracies": 0.8125, + "rewards/chosen": 0.79296875, + "rewards/margins": 5.625, + "rewards/rejected": -4.8125, + "step": 218 + }, + { + "epoch": 0.2687116564417178, + "grad_norm": 30.96899305966375, + "learning_rate": 1.3797520661157025e-06, + "logits/chosen": -0.091796875, + "logits/rejected": -0.189453125, + "logps/chosen": -354.0, + "logps/rejected": -310.0, + "loss": 0.3339, + "rewards/accuracies": 0.828125, + "rewards/chosen": 0.81640625, + "rewards/margins": 6.09375, + "rewards/rejected": -5.28125, + "step": 219 + }, + { + "epoch": 0.26993865030674846, + "grad_norm": 49.60415252304962, + "learning_rate": 1.3791322314049586e-06, + "logits/chosen": 0.07373046875, + "logits/rejected": 0.0047607421875, + "logps/chosen": -366.0, + "logps/rejected": -372.0, + "loss": 0.4563, + "rewards/accuracies": 0.765625, + "rewards/chosen": 0.416015625, + "rewards/margins": 5.0625, + "rewards/rejected": -4.65625, + "step": 220 + }, + { + "epoch": 0.27116564417177913, + "grad_norm": 35.817382292638634, + "learning_rate": 1.378512396694215e-06, + "logits/chosen": -0.049560546875, + "logits/rejected": -0.25390625, + "logps/chosen": -364.0, + "logps/rejected": -306.0, + "loss": 0.4136, + "rewards/accuracies": 0.796875, + "rewards/chosen": 0.46484375, + "rewards/margins": 5.75, + "rewards/rejected": -5.28125, + "step": 221 + }, + { + "epoch": 0.2723926380368098, + "grad_norm": 33.842292905676736, + "learning_rate": 1.377892561983471e-06, + "logits/chosen": -0.240234375, + "logits/rejected": -0.3671875, + "logps/chosen": -388.0, + "logps/rejected": -318.0, + "loss": 0.3321, + "rewards/accuracies": 0.859375, + "rewards/chosen": 1.4296875, + "rewards/margins": 7.375, + "rewards/rejected": -5.9375, + "step": 222 + }, + { + "epoch": 0.27361963190184047, + "grad_norm": 34.24344157812324, + "learning_rate": 1.3772727272727274e-06, + "logits/chosen": -0.0262451171875, + "logits/rejected": -0.2060546875, + "logps/chosen": -390.0, + "logps/rejected": -338.0, + "loss": 0.3389, + "rewards/accuracies": 0.84375, + "rewards/chosen": 0.58984375, + "rewards/margins": 6.1875, + "rewards/rejected": -5.59375, + "step": 223 + }, + { + "epoch": 0.2748466257668712, + "grad_norm": 36.32910023908209, + "learning_rate": 1.3766528925619836e-06, + "logits/chosen": -0.1591796875, + "logits/rejected": -0.33203125, + "logps/chosen": -452.0, + "logps/rejected": -352.0, + "loss": 0.3172, + "rewards/accuracies": 0.8046875, + "rewards/chosen": 1.28125, + "rewards/margins": 7.1875, + "rewards/rejected": -5.90625, + "step": 224 + }, + { + "epoch": 0.27607361963190186, + "grad_norm": 33.96451254717418, + "learning_rate": 1.3760330578512397e-06, + "logits/chosen": -0.021484375, + "logits/rejected": -0.2119140625, + "logps/chosen": -344.0, + "logps/rejected": -322.0, + "loss": 0.2991, + "rewards/accuracies": 0.84375, + "rewards/chosen": 0.4609375, + "rewards/margins": 6.5, + "rewards/rejected": -6.0625, + "step": 225 + }, + { + "epoch": 0.2773006134969325, + "grad_norm": 37.753965128366424, + "learning_rate": 1.3754132231404959e-06, + "logits/chosen": -0.035888671875, + "logits/rejected": -0.1689453125, + "logps/chosen": -378.0, + "logps/rejected": -328.0, + "loss": 0.4435, + "rewards/accuracies": 0.8046875, + "rewards/chosen": 0.82421875, + "rewards/margins": 6.4375, + "rewards/rejected": -5.625, + "step": 226 + }, + { + "epoch": 0.2785276073619632, + "grad_norm": 37.97162675437385, + "learning_rate": 1.374793388429752e-06, + "logits/chosen": -0.078125, + "logits/rejected": -0.1416015625, + "logps/chosen": -362.0, + "logps/rejected": -332.0, + "loss": 0.3808, + "rewards/accuracies": 0.859375, + "rewards/chosen": 0.625, + "rewards/margins": 6.625, + "rewards/rejected": -6.0, + "step": 227 + }, + { + "epoch": 0.27975460122699386, + "grad_norm": 42.90390749582614, + "learning_rate": 1.3741735537190084e-06, + "logits/chosen": -0.0908203125, + "logits/rejected": -0.2734375, + "logps/chosen": -366.0, + "logps/rejected": -338.0, + "loss": 0.3568, + "rewards/accuracies": 0.8125, + "rewards/chosen": 0.57421875, + "rewards/margins": 6.03125, + "rewards/rejected": -5.46875, + "step": 228 + }, + { + "epoch": 0.2809815950920245, + "grad_norm": 35.198650610347336, + "learning_rate": 1.3735537190082645e-06, + "logits/chosen": 0.0174560546875, + "logits/rejected": -0.1845703125, + "logps/chosen": -422.0, + "logps/rejected": -344.0, + "loss": 0.3812, + "rewards/accuracies": 0.84375, + "rewards/chosen": 1.640625, + "rewards/margins": 6.03125, + "rewards/rejected": -4.375, + "step": 229 + }, + { + "epoch": 0.2822085889570552, + "grad_norm": 31.76904887089288, + "learning_rate": 1.3729338842975206e-06, + "logits/chosen": 0.0027618408203125, + "logits/rejected": -0.1806640625, + "logps/chosen": -376.0, + "logps/rejected": -334.0, + "loss": 0.2964, + "rewards/accuracies": 0.859375, + "rewards/chosen": 0.99609375, + "rewards/margins": 6.375, + "rewards/rejected": -5.375, + "step": 230 + }, + { + "epoch": 0.28343558282208586, + "grad_norm": 38.2475164063325, + "learning_rate": 1.372314049586777e-06, + "logits/chosen": -0.0615234375, + "logits/rejected": -0.177734375, + "logps/chosen": -370.0, + "logps/rejected": -320.0, + "loss": 0.4068, + "rewards/accuracies": 0.8203125, + "rewards/chosen": 1.03125, + "rewards/margins": 5.6875, + "rewards/rejected": -4.65625, + "step": 231 + }, + { + "epoch": 0.2846625766871166, + "grad_norm": 34.785350120825576, + "learning_rate": 1.3716942148760331e-06, + "logits/chosen": -0.1474609375, + "logits/rejected": -0.27734375, + "logps/chosen": -384.0, + "logps/rejected": -342.0, + "loss": 0.3319, + "rewards/accuracies": 0.8203125, + "rewards/chosen": 0.6796875, + "rewards/margins": 5.5625, + "rewards/rejected": -4.875, + "step": 232 + }, + { + "epoch": 0.28588957055214725, + "grad_norm": 33.915962784771814, + "learning_rate": 1.3710743801652893e-06, + "logits/chosen": -0.04052734375, + "logits/rejected": -0.255859375, + "logps/chosen": -372.0, + "logps/rejected": -296.0, + "loss": 0.3377, + "rewards/accuracies": 0.8359375, + "rewards/chosen": 1.4296875, + "rewards/margins": 6.28125, + "rewards/rejected": -4.84375, + "step": 233 + }, + { + "epoch": 0.2871165644171779, + "grad_norm": 36.944834789970336, + "learning_rate": 1.3704545454545456e-06, + "logits/chosen": -0.189453125, + "logits/rejected": -0.302734375, + "logps/chosen": -374.0, + "logps/rejected": -318.0, + "loss": 0.3474, + "rewards/accuracies": 0.8359375, + "rewards/chosen": 1.203125, + "rewards/margins": 5.625, + "rewards/rejected": -4.40625, + "step": 234 + }, + { + "epoch": 0.2883435582822086, + "grad_norm": 36.36585753307406, + "learning_rate": 1.3698347107438016e-06, + "logits/chosen": -0.201171875, + "logits/rejected": -0.34765625, + "logps/chosen": -362.0, + "logps/rejected": -326.0, + "loss": 0.3987, + "rewards/accuracies": 0.8359375, + "rewards/chosen": 1.1171875, + "rewards/margins": 5.90625, + "rewards/rejected": -4.78125, + "step": 235 + }, + { + "epoch": 0.28957055214723926, + "grad_norm": 39.75324625491103, + "learning_rate": 1.369214876033058e-06, + "logits/chosen": 0.058837890625, + "logits/rejected": -0.1123046875, + "logps/chosen": -396.0, + "logps/rejected": -374.0, + "loss": 0.4644, + "rewards/accuracies": 0.7890625, + "rewards/chosen": 1.7421875, + "rewards/margins": 5.625, + "rewards/rejected": -3.890625, + "step": 236 + }, + { + "epoch": 0.2907975460122699, + "grad_norm": 33.93837750645907, + "learning_rate": 1.368595041322314e-06, + "logits/chosen": -0.1025390625, + "logits/rejected": -0.248046875, + "logps/chosen": -368.0, + "logps/rejected": -314.0, + "loss": 0.3597, + "rewards/accuracies": 0.8515625, + "rewards/chosen": 1.5, + "rewards/margins": 6.1875, + "rewards/rejected": -4.6875, + "step": 237 + }, + { + "epoch": 0.2920245398773006, + "grad_norm": 40.82126535762802, + "learning_rate": 1.3679752066115702e-06, + "logits/chosen": -0.11376953125, + "logits/rejected": -0.259765625, + "logps/chosen": -426.0, + "logps/rejected": -356.0, + "loss": 0.5098, + "rewards/accuracies": 0.8046875, + "rewards/chosen": 1.4453125, + "rewards/margins": 5.6875, + "rewards/rejected": -4.25, + "step": 238 + }, + { + "epoch": 0.29325153374233126, + "grad_norm": 37.96435040796285, + "learning_rate": 1.3673553719008266e-06, + "logits/chosen": -0.130859375, + "logits/rejected": -0.298828125, + "logps/chosen": -382.0, + "logps/rejected": -314.0, + "loss": 0.371, + "rewards/accuracies": 0.796875, + "rewards/chosen": 2.078125, + "rewards/margins": 6.46875, + "rewards/rejected": -4.375, + "step": 239 + }, + { + "epoch": 0.294478527607362, + "grad_norm": 34.11140908926358, + "learning_rate": 1.3667355371900827e-06, + "logits/chosen": -0.15234375, + "logits/rejected": -0.28515625, + "logps/chosen": -418.0, + "logps/rejected": -364.0, + "loss": 0.2283, + "rewards/accuracies": 0.890625, + "rewards/chosen": 1.890625, + "rewards/margins": 6.875, + "rewards/rejected": -5.0, + "step": 240 + }, + { + "epoch": 0.29570552147239265, + "grad_norm": 33.461868421672946, + "learning_rate": 1.3661157024793388e-06, + "logits/chosen": -0.076171875, + "logits/rejected": -0.2001953125, + "logps/chosen": -366.0, + "logps/rejected": -336.0, + "loss": 0.2828, + "rewards/accuracies": 0.84375, + "rewards/chosen": 1.8359375, + "rewards/margins": 6.4375, + "rewards/rejected": -4.59375, + "step": 241 + }, + { + "epoch": 0.2969325153374233, + "grad_norm": 32.77289894542164, + "learning_rate": 1.3654958677685952e-06, + "logits/chosen": -0.10009765625, + "logits/rejected": -0.2490234375, + "logps/chosen": -350.0, + "logps/rejected": -324.0, + "loss": 0.2975, + "rewards/accuracies": 0.84375, + "rewards/chosen": 1.2578125, + "rewards/margins": 6.59375, + "rewards/rejected": -5.34375, + "step": 242 + }, + { + "epoch": 0.298159509202454, + "grad_norm": 38.07813065801456, + "learning_rate": 1.3648760330578511e-06, + "logits/chosen": -0.1923828125, + "logits/rejected": -0.328125, + "logps/chosen": -390.0, + "logps/rejected": -364.0, + "loss": 0.4242, + "rewards/accuracies": 0.8203125, + "rewards/chosen": 1.328125, + "rewards/margins": 5.625, + "rewards/rejected": -4.3125, + "step": 243 + }, + { + "epoch": 0.29938650306748466, + "grad_norm": 39.871497711059654, + "learning_rate": 1.3642561983471075e-06, + "logits/chosen": -0.1806640625, + "logits/rejected": -0.39453125, + "logps/chosen": -386.0, + "logps/rejected": -346.0, + "loss": 0.5523, + "rewards/accuracies": 0.7734375, + "rewards/chosen": 1.265625, + "rewards/margins": 5.46875, + "rewards/rejected": -4.1875, + "step": 244 + }, + { + "epoch": 0.3006134969325153, + "grad_norm": 32.221505039188294, + "learning_rate": 1.3636363636363636e-06, + "logits/chosen": -0.15625, + "logits/rejected": -0.2890625, + "logps/chosen": -398.0, + "logps/rejected": -354.0, + "loss": 0.2967, + "rewards/accuracies": 0.8515625, + "rewards/chosen": 1.9140625, + "rewards/margins": 6.46875, + "rewards/rejected": -4.53125, + "step": 245 + }, + { + "epoch": 0.301840490797546, + "grad_norm": 34.1015902082894, + "learning_rate": 1.3630165289256198e-06, + "logits/chosen": -0.166015625, + "logits/rejected": -0.31640625, + "logps/chosen": -374.0, + "logps/rejected": -346.0, + "loss": 0.3765, + "rewards/accuracies": 0.828125, + "rewards/chosen": 2.171875, + "rewards/margins": 6.46875, + "rewards/rejected": -4.3125, + "step": 246 + }, + { + "epoch": 0.3030674846625767, + "grad_norm": 32.83309359401401, + "learning_rate": 1.3623966942148761e-06, + "logits/chosen": -0.1572265625, + "logits/rejected": -0.337890625, + "logps/chosen": -372.0, + "logps/rejected": -308.0, + "loss": 0.3534, + "rewards/accuracies": 0.8125, + "rewards/chosen": 1.1640625, + "rewards/margins": 5.96875, + "rewards/rejected": -4.78125, + "step": 247 + }, + { + "epoch": 0.3042944785276074, + "grad_norm": 33.089465845458236, + "learning_rate": 1.3617768595041322e-06, + "logits/chosen": -0.076171875, + "logits/rejected": -0.3203125, + "logps/chosen": -388.0, + "logps/rejected": -332.0, + "loss": 0.2682, + "rewards/accuracies": 0.8671875, + "rewards/chosen": 1.703125, + "rewards/margins": 6.03125, + "rewards/rejected": -4.3125, + "step": 248 + }, + { + "epoch": 0.30552147239263805, + "grad_norm": 39.03476607944447, + "learning_rate": 1.3611570247933884e-06, + "logits/chosen": -0.20703125, + "logits/rejected": -0.396484375, + "logps/chosen": -474.0, + "logps/rejected": -368.0, + "loss": 0.3241, + "rewards/accuracies": 0.875, + "rewards/chosen": 2.640625, + "rewards/margins": 6.78125, + "rewards/rejected": -4.15625, + "step": 249 + }, + { + "epoch": 0.3067484662576687, + "grad_norm": 38.29235099515543, + "learning_rate": 1.3605371900826447e-06, + "logits/chosen": -0.30078125, + "logits/rejected": -0.349609375, + "logps/chosen": -340.0, + "logps/rejected": -324.0, + "loss": 0.3804, + "rewards/accuracies": 0.84375, + "rewards/chosen": 1.34375, + "rewards/margins": 5.78125, + "rewards/rejected": -4.4375, + "step": 250 + }, + { + "epoch": 0.3079754601226994, + "grad_norm": 37.34357400280288, + "learning_rate": 1.3599173553719009e-06, + "logits/chosen": 0.0289306640625, + "logits/rejected": -0.134765625, + "logps/chosen": -376.0, + "logps/rejected": -308.0, + "loss": 0.4118, + "rewards/accuracies": 0.7890625, + "rewards/chosen": 1.421875, + "rewards/margins": 5.375, + "rewards/rejected": -3.9375, + "step": 251 + }, + { + "epoch": 0.30920245398773005, + "grad_norm": 37.841749969537275, + "learning_rate": 1.359297520661157e-06, + "logits/chosen": -0.1494140625, + "logits/rejected": -0.29296875, + "logps/chosen": -392.0, + "logps/rejected": -310.0, + "loss": 0.4506, + "rewards/accuracies": 0.8359375, + "rewards/chosen": 1.890625, + "rewards/margins": 5.78125, + "rewards/rejected": -3.875, + "step": 252 + }, + { + "epoch": 0.3104294478527607, + "grad_norm": 25.74895782175542, + "learning_rate": 1.3586776859504132e-06, + "logits/chosen": -0.1630859375, + "logits/rejected": -0.392578125, + "logps/chosen": -336.0, + "logps/rejected": -298.0, + "loss": 0.2694, + "rewards/accuracies": 0.8671875, + "rewards/chosen": 1.46875, + "rewards/margins": 6.0625, + "rewards/rejected": -4.59375, + "step": 253 + }, + { + "epoch": 0.3116564417177914, + "grad_norm": 36.56021325822385, + "learning_rate": 1.3580578512396693e-06, + "logits/chosen": -0.10302734375, + "logits/rejected": -0.25, + "logps/chosen": -352.0, + "logps/rejected": -320.0, + "loss": 0.4183, + "rewards/accuracies": 0.8046875, + "rewards/chosen": 1.8671875, + "rewards/margins": 6.1875, + "rewards/rejected": -4.3125, + "step": 254 + }, + { + "epoch": 0.3128834355828221, + "grad_norm": 37.45936986772829, + "learning_rate": 1.3574380165289257e-06, + "logits/chosen": -0.07958984375, + "logits/rejected": -0.228515625, + "logps/chosen": -384.0, + "logps/rejected": -312.0, + "loss": 0.439, + "rewards/accuracies": 0.796875, + "rewards/chosen": 1.8515625, + "rewards/margins": 5.71875, + "rewards/rejected": -3.859375, + "step": 255 + }, + { + "epoch": 0.3141104294478528, + "grad_norm": 33.113238785557265, + "learning_rate": 1.3568181818181818e-06, + "logits/chosen": -0.1640625, + "logits/rejected": -0.328125, + "logps/chosen": -348.0, + "logps/rejected": -316.0, + "loss": 0.3148, + "rewards/accuracies": 0.875, + "rewards/chosen": 2.3125, + "rewards/margins": 7.0, + "rewards/rejected": -4.6875, + "step": 256 + }, + { + "epoch": 0.31533742331288345, + "grad_norm": 34.825667619250744, + "learning_rate": 1.3561983471074382e-06, + "logits/chosen": -0.1494140625, + "logits/rejected": -0.2890625, + "logps/chosen": -426.0, + "logps/rejected": -356.0, + "loss": 0.3455, + "rewards/accuracies": 0.828125, + "rewards/chosen": 2.25, + "rewards/margins": 6.5625, + "rewards/rejected": -4.3125, + "step": 257 + }, + { + "epoch": 0.3165644171779141, + "grad_norm": 35.08263037322888, + "learning_rate": 1.3555785123966943e-06, + "logits/chosen": -0.0908203125, + "logits/rejected": -0.28125, + "logps/chosen": -396.0, + "logps/rejected": -342.0, + "loss": 0.319, + "rewards/accuracies": 0.875, + "rewards/chosen": 1.9140625, + "rewards/margins": 6.65625, + "rewards/rejected": -4.75, + "step": 258 + }, + { + "epoch": 0.3177914110429448, + "grad_norm": 33.547677568046986, + "learning_rate": 1.3549586776859504e-06, + "logits/chosen": -0.1689453125, + "logits/rejected": -0.349609375, + "logps/chosen": -344.0, + "logps/rejected": -356.0, + "loss": 0.3553, + "rewards/accuracies": 0.84375, + "rewards/chosen": 1.296875, + "rewards/margins": 6.5, + "rewards/rejected": -5.1875, + "step": 259 + }, + { + "epoch": 0.31901840490797545, + "grad_norm": 35.38305104419636, + "learning_rate": 1.3543388429752068e-06, + "logits/chosen": -0.1904296875, + "logits/rejected": -0.26953125, + "logps/chosen": -390.0, + "logps/rejected": -326.0, + "loss": 0.3657, + "rewards/accuracies": 0.8359375, + "rewards/chosen": 1.296875, + "rewards/margins": 6.4375, + "rewards/rejected": -5.15625, + "step": 260 + }, + { + "epoch": 0.3202453987730061, + "grad_norm": 37.62925536561278, + "learning_rate": 1.3537190082644627e-06, + "logits/chosen": -0.11181640625, + "logits/rejected": -0.263671875, + "logps/chosen": -380.0, + "logps/rejected": -330.0, + "loss": 0.3941, + "rewards/accuracies": 0.8125, + "rewards/chosen": 1.296875, + "rewards/margins": 6.8125, + "rewards/rejected": -5.53125, + "step": 261 + }, + { + "epoch": 0.3214723926380368, + "grad_norm": 45.47620412654006, + "learning_rate": 1.353099173553719e-06, + "logits/chosen": -0.1591796875, + "logits/rejected": -0.3359375, + "logps/chosen": -386.0, + "logps/rejected": -340.0, + "loss": 0.4019, + "rewards/accuracies": 0.8125, + "rewards/chosen": 1.875, + "rewards/margins": 6.9375, + "rewards/rejected": -5.0625, + "step": 262 + }, + { + "epoch": 0.3226993865030675, + "grad_norm": 43.246204937096174, + "learning_rate": 1.3524793388429752e-06, + "logits/chosen": -0.1494140625, + "logits/rejected": -0.2734375, + "logps/chosen": -370.0, + "logps/rejected": -322.0, + "loss": 0.4521, + "rewards/accuracies": 0.765625, + "rewards/chosen": 0.6484375, + "rewards/margins": 5.65625, + "rewards/rejected": -5.0, + "step": 263 + }, + { + "epoch": 0.3239263803680982, + "grad_norm": 37.571005698421835, + "learning_rate": 1.3518595041322314e-06, + "logits/chosen": -0.09521484375, + "logits/rejected": -0.20703125, + "logps/chosen": -354.0, + "logps/rejected": -308.0, + "loss": 0.3748, + "rewards/accuracies": 0.8125, + "rewards/chosen": 1.0078125, + "rewards/margins": 6.1875, + "rewards/rejected": -5.15625, + "step": 264 + }, + { + "epoch": 0.32515337423312884, + "grad_norm": 45.351424976255885, + "learning_rate": 1.3512396694214877e-06, + "logits/chosen": -0.1064453125, + "logits/rejected": -0.265625, + "logps/chosen": -360.0, + "logps/rejected": -288.0, + "loss": 0.3745, + "rewards/accuracies": 0.8203125, + "rewards/chosen": 0.87109375, + "rewards/margins": 6.75, + "rewards/rejected": -5.875, + "step": 265 + }, + { + "epoch": 0.3263803680981595, + "grad_norm": 44.31098246879602, + "learning_rate": 1.3506198347107439e-06, + "logits/chosen": -0.07861328125, + "logits/rejected": -0.228515625, + "logps/chosen": -326.0, + "logps/rejected": -316.0, + "loss": 0.5312, + "rewards/accuracies": 0.78125, + "rewards/chosen": -0.27734375, + "rewards/margins": 5.25, + "rewards/rejected": -5.53125, + "step": 266 + }, + { + "epoch": 0.3276073619631902, + "grad_norm": 43.337206726154484, + "learning_rate": 1.35e-06, + "logits/chosen": -0.0966796875, + "logits/rejected": -0.31640625, + "logps/chosen": -406.0, + "logps/rejected": -322.0, + "loss": 0.3998, + "rewards/accuracies": 0.8359375, + "rewards/chosen": 1.2421875, + "rewards/margins": 6.875, + "rewards/rejected": -5.625, + "step": 267 + }, + { + "epoch": 0.32883435582822085, + "grad_norm": 28.822383795146568, + "learning_rate": 1.3493801652892563e-06, + "logits/chosen": -0.25390625, + "logits/rejected": -0.365234375, + "logps/chosen": -392.0, + "logps/rejected": -344.0, + "loss": 0.209, + "rewards/accuracies": 0.90625, + "rewards/chosen": 0.75390625, + "rewards/margins": 7.40625, + "rewards/rejected": -6.65625, + "step": 268 + }, + { + "epoch": 0.3300613496932515, + "grad_norm": 30.291051549986957, + "learning_rate": 1.3487603305785125e-06, + "logits/chosen": -0.1240234375, + "logits/rejected": -0.43359375, + "logps/chosen": -412.0, + "logps/rejected": -322.0, + "loss": 0.237, + "rewards/accuracies": 0.890625, + "rewards/chosen": 0.83203125, + "rewards/margins": 7.125, + "rewards/rejected": -6.3125, + "step": 269 + }, + { + "epoch": 0.3312883435582822, + "grad_norm": 38.1594801225088, + "learning_rate": 1.3481404958677686e-06, + "logits/chosen": -0.09912109375, + "logits/rejected": -0.2138671875, + "logps/chosen": -334.0, + "logps/rejected": -308.0, + "loss": 0.4766, + "rewards/accuracies": 0.8046875, + "rewards/chosen": 0.462890625, + "rewards/margins": 5.625, + "rewards/rejected": -5.15625, + "step": 270 + }, + { + "epoch": 0.3325153374233129, + "grad_norm": 28.374821367974555, + "learning_rate": 1.3475206611570248e-06, + "logits/chosen": -0.12451171875, + "logits/rejected": -0.240234375, + "logps/chosen": -390.0, + "logps/rejected": -344.0, + "loss": 0.3004, + "rewards/accuracies": 0.875, + "rewards/chosen": 1.3984375, + "rewards/margins": 7.5, + "rewards/rejected": -6.09375, + "step": 271 + }, + { + "epoch": 0.3337423312883436, + "grad_norm": 34.17489484053231, + "learning_rate": 1.346900826446281e-06, + "logits/chosen": -0.00823974609375, + "logits/rejected": -0.171875, + "logps/chosen": -366.0, + "logps/rejected": -318.0, + "loss": 0.2861, + "rewards/accuracies": 0.8671875, + "rewards/chosen": 0.423828125, + "rewards/margins": 6.75, + "rewards/rejected": -6.34375, + "step": 272 + }, + { + "epoch": 0.33496932515337424, + "grad_norm": 41.21053534408477, + "learning_rate": 1.3462809917355373e-06, + "logits/chosen": 0.06640625, + "logits/rejected": -0.007110595703125, + "logps/chosen": -344.0, + "logps/rejected": -314.0, + "loss": 0.5165, + "rewards/accuracies": 0.7578125, + "rewards/chosen": 0.166015625, + "rewards/margins": 5.75, + "rewards/rejected": -5.5625, + "step": 273 + }, + { + "epoch": 0.3361963190184049, + "grad_norm": 34.571787806623156, + "learning_rate": 1.3456611570247934e-06, + "logits/chosen": 0.0419921875, + "logits/rejected": -0.1025390625, + "logps/chosen": -410.0, + "logps/rejected": -346.0, + "loss": 0.3839, + "rewards/accuracies": 0.8515625, + "rewards/chosen": 0.83203125, + "rewards/margins": 6.40625, + "rewards/rejected": -5.5625, + "step": 274 + }, + { + "epoch": 0.3374233128834356, + "grad_norm": 31.979565052575825, + "learning_rate": 1.3450413223140495e-06, + "logits/chosen": -0.0302734375, + "logits/rejected": -0.119140625, + "logps/chosen": -326.0, + "logps/rejected": -326.0, + "loss": 0.4175, + "rewards/accuracies": 0.8046875, + "rewards/chosen": 0.412109375, + "rewards/margins": 6.09375, + "rewards/rejected": -5.6875, + "step": 275 + }, + { + "epoch": 0.33865030674846625, + "grad_norm": 28.189301031425032, + "learning_rate": 1.344421487603306e-06, + "logits/chosen": 0.0498046875, + "logits/rejected": -0.162109375, + "logps/chosen": -390.0, + "logps/rejected": -330.0, + "loss": 0.3505, + "rewards/accuracies": 0.8125, + "rewards/chosen": 0.2470703125, + "rewards/margins": 6.21875, + "rewards/rejected": -5.96875, + "step": 276 + }, + { + "epoch": 0.3398773006134969, + "grad_norm": 29.004541701793432, + "learning_rate": 1.343801652892562e-06, + "logits/chosen": -0.15234375, + "logits/rejected": -0.337890625, + "logps/chosen": -414.0, + "logps/rejected": -344.0, + "loss": 0.2468, + "rewards/accuracies": 0.8671875, + "rewards/chosen": 1.1640625, + "rewards/margins": 7.53125, + "rewards/rejected": -6.375, + "step": 277 + }, + { + "epoch": 0.3411042944785276, + "grad_norm": 40.163399091060114, + "learning_rate": 1.3431818181818184e-06, + "logits/chosen": -0.03759765625, + "logits/rejected": -0.08447265625, + "logps/chosen": -384.0, + "logps/rejected": -356.0, + "loss": 0.4008, + "rewards/accuracies": 0.8203125, + "rewards/chosen": 0.69921875, + "rewards/margins": 6.4375, + "rewards/rejected": -5.75, + "step": 278 + }, + { + "epoch": 0.3423312883435583, + "grad_norm": 21.8515947233153, + "learning_rate": 1.3425619834710743e-06, + "logits/chosen": -0.10888671875, + "logits/rejected": -0.2890625, + "logps/chosen": -392.0, + "logps/rejected": -326.0, + "loss": 0.1765, + "rewards/accuracies": 0.921875, + "rewards/chosen": 0.65625, + "rewards/margins": 7.46875, + "rewards/rejected": -6.8125, + "step": 279 + }, + { + "epoch": 0.34355828220858897, + "grad_norm": 34.753521492334585, + "learning_rate": 1.3419421487603305e-06, + "logits/chosen": -0.09228515625, + "logits/rejected": -0.234375, + "logps/chosen": -408.0, + "logps/rejected": -372.0, + "loss": 0.4097, + "rewards/accuracies": 0.8125, + "rewards/chosen": 0.83984375, + "rewards/margins": 6.96875, + "rewards/rejected": -6.15625, + "step": 280 + }, + { + "epoch": 0.34478527607361964, + "grad_norm": 35.589195433139196, + "learning_rate": 1.3413223140495868e-06, + "logits/chosen": -0.00921630859375, + "logits/rejected": -0.201171875, + "logps/chosen": -382.0, + "logps/rejected": -330.0, + "loss": 0.3917, + "rewards/accuracies": 0.8046875, + "rewards/chosen": 0.609375, + "rewards/margins": 5.65625, + "rewards/rejected": -5.0625, + "step": 281 + }, + { + "epoch": 0.3460122699386503, + "grad_norm": 29.302596470256102, + "learning_rate": 1.340702479338843e-06, + "logits/chosen": -0.2021484375, + "logits/rejected": -0.341796875, + "logps/chosen": -344.0, + "logps/rejected": -324.0, + "loss": 0.2481, + "rewards/accuracies": 0.875, + "rewards/chosen": 1.0859375, + "rewards/margins": 7.1875, + "rewards/rejected": -6.09375, + "step": 282 + }, + { + "epoch": 0.347239263803681, + "grad_norm": 33.7375126879184, + "learning_rate": 1.3400826446280993e-06, + "logits/chosen": -0.2265625, + "logits/rejected": -0.3125, + "logps/chosen": -354.0, + "logps/rejected": -298.0, + "loss": 0.3519, + "rewards/accuracies": 0.8203125, + "rewards/chosen": 0.59765625, + "rewards/margins": 6.15625, + "rewards/rejected": -5.5625, + "step": 283 + }, + { + "epoch": 0.34846625766871164, + "grad_norm": 35.97047400633379, + "learning_rate": 1.3394628099173555e-06, + "logits/chosen": -0.034423828125, + "logits/rejected": -0.166015625, + "logps/chosen": -370.0, + "logps/rejected": -320.0, + "loss": 0.2954, + "rewards/accuracies": 0.8515625, + "rewards/chosen": 0.859375, + "rewards/margins": 6.6875, + "rewards/rejected": -5.84375, + "step": 284 + }, + { + "epoch": 0.3496932515337423, + "grad_norm": 29.423070534885003, + "learning_rate": 1.3388429752066116e-06, + "logits/chosen": -0.130859375, + "logits/rejected": -0.322265625, + "logps/chosen": -360.0, + "logps/rejected": -328.0, + "loss": 0.2845, + "rewards/accuracies": 0.84375, + "rewards/chosen": 0.80078125, + "rewards/margins": 6.65625, + "rewards/rejected": -5.875, + "step": 285 + }, + { + "epoch": 0.350920245398773, + "grad_norm": 35.376791661357736, + "learning_rate": 1.338223140495868e-06, + "logits/chosen": -0.04736328125, + "logits/rejected": -0.2412109375, + "logps/chosen": -422.0, + "logps/rejected": -356.0, + "loss": 0.3799, + "rewards/accuracies": 0.796875, + "rewards/chosen": 0.8515625, + "rewards/margins": 6.9375, + "rewards/rejected": -6.0625, + "step": 286 + }, + { + "epoch": 0.3521472392638037, + "grad_norm": 38.51049635159653, + "learning_rate": 1.337603305785124e-06, + "logits/chosen": -0.150390625, + "logits/rejected": -0.283203125, + "logps/chosen": -382.0, + "logps/rejected": -330.0, + "loss": 0.3715, + "rewards/accuracies": 0.8125, + "rewards/chosen": 0.375, + "rewards/margins": 6.1875, + "rewards/rejected": -5.8125, + "step": 287 + }, + { + "epoch": 0.35337423312883437, + "grad_norm": 31.395092155644413, + "learning_rate": 1.3369834710743802e-06, + "logits/chosen": 0.057373046875, + "logits/rejected": -0.0888671875, + "logps/chosen": -386.0, + "logps/rejected": -336.0, + "loss": 0.3837, + "rewards/accuracies": 0.8359375, + "rewards/chosen": 0.298828125, + "rewards/margins": 5.9375, + "rewards/rejected": -5.625, + "step": 288 + }, + { + "epoch": 0.35460122699386504, + "grad_norm": 25.699158986974325, + "learning_rate": 1.3363636363636364e-06, + "logits/chosen": -0.11083984375, + "logits/rejected": -0.248046875, + "logps/chosen": -404.0, + "logps/rejected": -330.0, + "loss": 0.1963, + "rewards/accuracies": 0.921875, + "rewards/chosen": 0.94140625, + "rewards/margins": 7.65625, + "rewards/rejected": -6.71875, + "step": 289 + }, + { + "epoch": 0.3558282208588957, + "grad_norm": 26.172712857085628, + "learning_rate": 1.3357438016528925e-06, + "logits/chosen": -0.283203125, + "logits/rejected": -0.41796875, + "logps/chosen": -380.0, + "logps/rejected": -350.0, + "loss": 0.2207, + "rewards/accuracies": 0.9296875, + "rewards/chosen": 0.65625, + "rewards/margins": 7.65625, + "rewards/rejected": -7.0, + "step": 290 + }, + { + "epoch": 0.3570552147239264, + "grad_norm": 35.706855965144605, + "learning_rate": 1.3351239669421489e-06, + "logits/chosen": -0.2275390625, + "logits/rejected": -0.3359375, + "logps/chosen": -394.0, + "logps/rejected": -340.0, + "loss": 0.4375, + "rewards/accuracies": 0.8046875, + "rewards/chosen": 1.5625, + "rewards/margins": 7.09375, + "rewards/rejected": -5.53125, + "step": 291 + }, + { + "epoch": 0.35828220858895704, + "grad_norm": 32.40631116792141, + "learning_rate": 1.334504132231405e-06, + "logits/chosen": -0.169921875, + "logits/rejected": -0.390625, + "logps/chosen": -406.0, + "logps/rejected": -348.0, + "loss": 0.2821, + "rewards/accuracies": 0.8671875, + "rewards/chosen": 0.7734375, + "rewards/margins": 7.15625, + "rewards/rejected": -6.40625, + "step": 292 + }, + { + "epoch": 0.3595092024539877, + "grad_norm": 27.724555696310663, + "learning_rate": 1.3338842975206611e-06, + "logits/chosen": -0.08251953125, + "logits/rejected": -0.275390625, + "logps/chosen": -382.0, + "logps/rejected": -332.0, + "loss": 0.2357, + "rewards/accuracies": 0.8828125, + "rewards/chosen": 0.033203125, + "rewards/margins": 6.78125, + "rewards/rejected": -6.71875, + "step": 293 + }, + { + "epoch": 0.36073619631901843, + "grad_norm": 26.756162333522596, + "learning_rate": 1.3332644628099175e-06, + "logits/chosen": -0.359375, + "logits/rejected": -0.4921875, + "logps/chosen": -472.0, + "logps/rejected": -368.0, + "loss": 0.3094, + "rewards/accuracies": 0.8515625, + "rewards/chosen": 0.66796875, + "rewards/margins": 7.65625, + "rewards/rejected": -7.0, + "step": 294 + }, + { + "epoch": 0.3619631901840491, + "grad_norm": 27.599547481570436, + "learning_rate": 1.3326446280991736e-06, + "logits/chosen": -0.04345703125, + "logits/rejected": -0.1884765625, + "logps/chosen": -356.0, + "logps/rejected": -332.0, + "loss": 0.2495, + "rewards/accuracies": 0.8828125, + "rewards/chosen": 0.2734375, + "rewards/margins": 6.875, + "rewards/rejected": -6.59375, + "step": 295 + }, + { + "epoch": 0.36319018404907977, + "grad_norm": 35.1081345543023, + "learning_rate": 1.3320247933884298e-06, + "logits/chosen": -0.007598876953125, + "logits/rejected": -0.154296875, + "logps/chosen": -420.0, + "logps/rejected": -338.0, + "loss": 0.3852, + "rewards/accuracies": 0.859375, + "rewards/chosen": 1.1640625, + "rewards/margins": 6.6875, + "rewards/rejected": -5.5, + "step": 296 + }, + { + "epoch": 0.36441717791411044, + "grad_norm": 31.495501539039623, + "learning_rate": 1.331404958677686e-06, + "logits/chosen": -0.154296875, + "logits/rejected": -0.30859375, + "logps/chosen": -360.0, + "logps/rejected": -322.0, + "loss": 0.2553, + "rewards/accuracies": 0.859375, + "rewards/chosen": 0.79296875, + "rewards/margins": 7.0625, + "rewards/rejected": -6.25, + "step": 297 + }, + { + "epoch": 0.3656441717791411, + "grad_norm": 37.56929706779854, + "learning_rate": 1.330785123966942e-06, + "logits/chosen": -0.009521484375, + "logits/rejected": -0.09716796875, + "logps/chosen": -384.0, + "logps/rejected": -336.0, + "loss": 0.3993, + "rewards/accuracies": 0.8203125, + "rewards/chosen": 0.498046875, + "rewards/margins": 6.03125, + "rewards/rejected": -5.53125, + "step": 298 + }, + { + "epoch": 0.36687116564417177, + "grad_norm": 36.46646666012725, + "learning_rate": 1.3301652892561984e-06, + "logits/chosen": -0.058837890625, + "logits/rejected": -0.228515625, + "logps/chosen": -400.0, + "logps/rejected": -334.0, + "loss": 0.3895, + "rewards/accuracies": 0.84375, + "rewards/chosen": -0.1650390625, + "rewards/margins": 5.78125, + "rewards/rejected": -5.9375, + "step": 299 + }, + { + "epoch": 0.36809815950920244, + "grad_norm": 37.01196046043553, + "learning_rate": 1.3295454545454546e-06, + "logits/chosen": -0.173828125, + "logits/rejected": -0.3203125, + "logps/chosen": -372.0, + "logps/rejected": -328.0, + "loss": 0.4644, + "rewards/accuracies": 0.8203125, + "rewards/chosen": 0.482421875, + "rewards/margins": 6.65625, + "rewards/rejected": -6.15625, + "step": 300 + }, + { + "epoch": 0.3693251533742331, + "grad_norm": 32.17417313364543, + "learning_rate": 1.3289256198347107e-06, + "logits/chosen": -0.11865234375, + "logits/rejected": -0.298828125, + "logps/chosen": -460.0, + "logps/rejected": -386.0, + "loss": 0.3417, + "rewards/accuracies": 0.828125, + "rewards/chosen": 0.8515625, + "rewards/margins": 7.1875, + "rewards/rejected": -6.34375, + "step": 301 + }, + { + "epoch": 0.37055214723926383, + "grad_norm": 25.798891470824255, + "learning_rate": 1.328305785123967e-06, + "logits/chosen": -0.150390625, + "logits/rejected": -0.34765625, + "logps/chosen": -382.0, + "logps/rejected": -316.0, + "loss": 0.2452, + "rewards/accuracies": 0.8984375, + "rewards/chosen": 0.287109375, + "rewards/margins": 6.6875, + "rewards/rejected": -6.40625, + "step": 302 + }, + { + "epoch": 0.3717791411042945, + "grad_norm": 40.25697610802204, + "learning_rate": 1.3276859504132232e-06, + "logits/chosen": -0.1474609375, + "logits/rejected": -0.2578125, + "logps/chosen": -388.0, + "logps/rejected": -376.0, + "loss": 0.4569, + "rewards/accuracies": 0.78125, + "rewards/chosen": -0.466796875, + "rewards/margins": 6.1875, + "rewards/rejected": -6.65625, + "step": 303 + }, + { + "epoch": 0.37300613496932516, + "grad_norm": 30.538056754282714, + "learning_rate": 1.3270661157024793e-06, + "logits/chosen": -0.0264892578125, + "logits/rejected": -0.287109375, + "logps/chosen": -396.0, + "logps/rejected": -334.0, + "loss": 0.2974, + "rewards/accuracies": 0.875, + "rewards/chosen": -0.140625, + "rewards/margins": 6.875, + "rewards/rejected": -7.0, + "step": 304 + }, + { + "epoch": 0.37423312883435583, + "grad_norm": 27.228440712710828, + "learning_rate": 1.3264462809917357e-06, + "logits/chosen": -0.109375, + "logits/rejected": -0.2490234375, + "logps/chosen": -416.0, + "logps/rejected": -344.0, + "loss": 0.2623, + "rewards/accuracies": 0.90625, + "rewards/chosen": 0.341796875, + "rewards/margins": 7.71875, + "rewards/rejected": -7.375, + "step": 305 + }, + { + "epoch": 0.3754601226993865, + "grad_norm": 34.02132219179046, + "learning_rate": 1.3258264462809916e-06, + "logits/chosen": -0.2177734375, + "logits/rejected": -0.37890625, + "logps/chosen": -430.0, + "logps/rejected": -364.0, + "loss": 0.2805, + "rewards/accuracies": 0.8828125, + "rewards/chosen": -0.33203125, + "rewards/margins": 6.75, + "rewards/rejected": -7.0625, + "step": 306 + }, + { + "epoch": 0.37668711656441717, + "grad_norm": 34.60907155238812, + "learning_rate": 1.325206611570248e-06, + "logits/chosen": -0.06103515625, + "logits/rejected": -0.19140625, + "logps/chosen": -382.0, + "logps/rejected": -344.0, + "loss": 0.4791, + "rewards/accuracies": 0.796875, + "rewards/chosen": -0.18359375, + "rewards/margins": 6.90625, + "rewards/rejected": -7.09375, + "step": 307 + }, + { + "epoch": 0.37791411042944784, + "grad_norm": 34.523248358542865, + "learning_rate": 1.3245867768595041e-06, + "logits/chosen": -0.00148773193359375, + "logits/rejected": -0.162109375, + "logps/chosen": -398.0, + "logps/rejected": -360.0, + "loss": 0.3857, + "rewards/accuracies": 0.828125, + "rewards/chosen": -0.53515625, + "rewards/margins": 6.0, + "rewards/rejected": -6.5625, + "step": 308 + }, + { + "epoch": 0.3791411042944785, + "grad_norm": 29.962650665710807, + "learning_rate": 1.3239669421487603e-06, + "logits/chosen": -0.208984375, + "logits/rejected": -0.31640625, + "logps/chosen": -388.0, + "logps/rejected": -354.0, + "loss": 0.2543, + "rewards/accuracies": 0.875, + "rewards/chosen": 0.609375, + "rewards/margins": 7.5625, + "rewards/rejected": -6.9375, + "step": 309 + }, + { + "epoch": 0.3803680981595092, + "grad_norm": 33.20481159292772, + "learning_rate": 1.3233471074380166e-06, + "logits/chosen": 0.051513671875, + "logits/rejected": -0.037353515625, + "logps/chosen": -322.0, + "logps/rejected": -312.0, + "loss": 0.4174, + "rewards/accuracies": 0.8359375, + "rewards/chosen": -0.875, + "rewards/margins": 5.5, + "rewards/rejected": -6.375, + "step": 310 + }, + { + "epoch": 0.3815950920245399, + "grad_norm": 32.36453885140526, + "learning_rate": 1.3227272727272727e-06, + "logits/chosen": -0.158203125, + "logits/rejected": -0.3359375, + "logps/chosen": -392.0, + "logps/rejected": -348.0, + "loss": 0.3525, + "rewards/accuracies": 0.8203125, + "rewards/chosen": 0.486328125, + "rewards/margins": 7.78125, + "rewards/rejected": -7.28125, + "step": 311 + }, + { + "epoch": 0.38282208588957056, + "grad_norm": 39.08052846369179, + "learning_rate": 1.322107438016529e-06, + "logits/chosen": 0.0162353515625, + "logits/rejected": -0.1142578125, + "logps/chosen": -392.0, + "logps/rejected": -378.0, + "loss": 0.5608, + "rewards/accuracies": 0.7421875, + "rewards/chosen": 0.0712890625, + "rewards/margins": 6.34375, + "rewards/rejected": -6.28125, + "step": 312 + }, + { + "epoch": 0.38404907975460123, + "grad_norm": 35.15009453427041, + "learning_rate": 1.3214876033057852e-06, + "logits/chosen": -0.0235595703125, + "logits/rejected": -0.298828125, + "logps/chosen": -410.0, + "logps/rejected": -372.0, + "loss": 0.383, + "rewards/accuracies": 0.84375, + "rewards/chosen": 1.1640625, + "rewards/margins": 8.0625, + "rewards/rejected": -6.9375, + "step": 313 + }, + { + "epoch": 0.3852760736196319, + "grad_norm": 28.038472021341548, + "learning_rate": 1.3208677685950414e-06, + "logits/chosen": -0.05517578125, + "logits/rejected": -0.2421875, + "logps/chosen": -458.0, + "logps/rejected": -360.0, + "loss": 0.2315, + "rewards/accuracies": 0.8828125, + "rewards/chosen": 2.03125, + "rewards/margins": 9.375, + "rewards/rejected": -7.34375, + "step": 314 + }, + { + "epoch": 0.38650306748466257, + "grad_norm": 31.702508477036716, + "learning_rate": 1.3202479338842975e-06, + "logits/chosen": 0.076171875, + "logits/rejected": -0.11865234375, + "logps/chosen": -350.0, + "logps/rejected": -300.0, + "loss": 0.3293, + "rewards/accuracies": 0.8203125, + "rewards/chosen": 0.2431640625, + "rewards/margins": 6.84375, + "rewards/rejected": -6.59375, + "step": 315 + }, + { + "epoch": 0.38773006134969323, + "grad_norm": 27.443273044054905, + "learning_rate": 1.3196280991735537e-06, + "logits/chosen": 0.05419921875, + "logits/rejected": -0.08544921875, + "logps/chosen": -360.0, + "logps/rejected": -308.0, + "loss": 0.3129, + "rewards/accuracies": 0.875, + "rewards/chosen": 0.75390625, + "rewards/margins": 7.5625, + "rewards/rejected": -6.8125, + "step": 316 + }, + { + "epoch": 0.3889570552147239, + "grad_norm": 28.339536052474916, + "learning_rate": 1.31900826446281e-06, + "logits/chosen": -0.09326171875, + "logits/rejected": -0.3046875, + "logps/chosen": -400.0, + "logps/rejected": -368.0, + "loss": 0.2873, + "rewards/accuracies": 0.859375, + "rewards/chosen": 0.94140625, + "rewards/margins": 8.125, + "rewards/rejected": -7.15625, + "step": 317 + }, + { + "epoch": 0.3901840490797546, + "grad_norm": 38.792002326692455, + "learning_rate": 1.3183884297520662e-06, + "logits/chosen": -0.142578125, + "logits/rejected": -0.296875, + "logps/chosen": -402.0, + "logps/rejected": -366.0, + "loss": 0.4379, + "rewards/accuracies": 0.828125, + "rewards/chosen": 0.91796875, + "rewards/margins": 7.0625, + "rewards/rejected": -6.125, + "step": 318 + }, + { + "epoch": 0.3914110429447853, + "grad_norm": 17.992013049840406, + "learning_rate": 1.3177685950413223e-06, + "logits/chosen": -0.068359375, + "logits/rejected": -0.255859375, + "logps/chosen": -414.0, + "logps/rejected": -320.0, + "loss": 0.1471, + "rewards/accuracies": 0.9296875, + "rewards/chosen": 1.1640625, + "rewards/margins": 8.8125, + "rewards/rejected": -7.65625, + "step": 319 + }, + { + "epoch": 0.39263803680981596, + "grad_norm": 31.193757820598883, + "learning_rate": 1.3171487603305787e-06, + "logits/chosen": -0.056884765625, + "logits/rejected": -0.166015625, + "logps/chosen": -320.0, + "logps/rejected": -286.0, + "loss": 0.325, + "rewards/accuracies": 0.859375, + "rewards/chosen": 0.31640625, + "rewards/margins": 6.84375, + "rewards/rejected": -6.53125, + "step": 320 + }, + { + "epoch": 0.39386503067484663, + "grad_norm": 39.71075166734815, + "learning_rate": 1.3165289256198348e-06, + "logits/chosen": -0.201171875, + "logits/rejected": -0.35546875, + "logps/chosen": -410.0, + "logps/rejected": -398.0, + "loss": 0.3784, + "rewards/accuracies": 0.8515625, + "rewards/chosen": 0.439453125, + "rewards/margins": 7.84375, + "rewards/rejected": -7.375, + "step": 321 + }, + { + "epoch": 0.3950920245398773, + "grad_norm": 26.7824487712344, + "learning_rate": 1.315909090909091e-06, + "logits/chosen": -0.107421875, + "logits/rejected": -0.2578125, + "logps/chosen": -394.0, + "logps/rejected": -330.0, + "loss": 0.2893, + "rewards/accuracies": 0.8984375, + "rewards/chosen": 1.203125, + "rewards/margins": 7.84375, + "rewards/rejected": -6.65625, + "step": 322 + }, + { + "epoch": 0.39631901840490796, + "grad_norm": 36.45438287117724, + "learning_rate": 1.3152892561983473e-06, + "logits/chosen": -0.1318359375, + "logits/rejected": -0.265625, + "logps/chosen": -354.0, + "logps/rejected": -334.0, + "loss": 0.4474, + "rewards/accuracies": 0.8203125, + "rewards/chosen": 0.6875, + "rewards/margins": 6.15625, + "rewards/rejected": -5.46875, + "step": 323 + }, + { + "epoch": 0.39754601226993863, + "grad_norm": 31.15880722522094, + "learning_rate": 1.3146694214876032e-06, + "logits/chosen": -0.193359375, + "logits/rejected": -0.361328125, + "logps/chosen": -356.0, + "logps/rejected": -328.0, + "loss": 0.3486, + "rewards/accuracies": 0.8515625, + "rewards/chosen": 0.62890625, + "rewards/margins": 7.4375, + "rewards/rejected": -6.8125, + "step": 324 + }, + { + "epoch": 0.3987730061349693, + "grad_norm": 36.47850348671778, + "learning_rate": 1.3140495867768596e-06, + "logits/chosen": -0.0272216796875, + "logits/rejected": -0.09326171875, + "logps/chosen": -330.0, + "logps/rejected": -304.0, + "loss": 0.5542, + "rewards/accuracies": 0.8046875, + "rewards/chosen": 0.33203125, + "rewards/margins": 6.15625, + "rewards/rejected": -5.84375, + "step": 325 + }, + { + "epoch": 0.4, + "grad_norm": 26.62936328977598, + "learning_rate": 1.3134297520661157e-06, + "logits/chosen": -0.2294921875, + "logits/rejected": -0.369140625, + "logps/chosen": -398.0, + "logps/rejected": -374.0, + "loss": 0.1993, + "rewards/accuracies": 0.8671875, + "rewards/chosen": 1.5546875, + "rewards/margins": 8.375, + "rewards/rejected": -6.84375, + "step": 326 + }, + { + "epoch": 0.4012269938650307, + "grad_norm": 39.76264342514329, + "learning_rate": 1.3128099173553719e-06, + "logits/chosen": -0.1015625, + "logits/rejected": -0.271484375, + "logps/chosen": -430.0, + "logps/rejected": -352.0, + "loss": 0.4557, + "rewards/accuracies": 0.8046875, + "rewards/chosen": 0.921875, + "rewards/margins": 6.125, + "rewards/rejected": -5.1875, + "step": 327 + }, + { + "epoch": 0.40245398773006136, + "grad_norm": 33.62855666124103, + "learning_rate": 1.3121900826446282e-06, + "logits/chosen": -0.1328125, + "logits/rejected": -0.259765625, + "logps/chosen": -362.0, + "logps/rejected": -288.0, + "loss": 0.363, + "rewards/accuracies": 0.84375, + "rewards/chosen": 0.228515625, + "rewards/margins": 6.15625, + "rewards/rejected": -5.9375, + "step": 328 + }, + { + "epoch": 0.403680981595092, + "grad_norm": 36.1359556726495, + "learning_rate": 1.3115702479338844e-06, + "logits/chosen": -0.08203125, + "logits/rejected": -0.283203125, + "logps/chosen": -466.0, + "logps/rejected": -386.0, + "loss": 0.3754, + "rewards/accuracies": 0.796875, + "rewards/chosen": 1.078125, + "rewards/margins": 8.3125, + "rewards/rejected": -7.25, + "step": 329 + }, + { + "epoch": 0.4049079754601227, + "grad_norm": 27.601952208856858, + "learning_rate": 1.3109504132231405e-06, + "logits/chosen": -0.22265625, + "logits/rejected": -0.423828125, + "logps/chosen": -418.0, + "logps/rejected": -352.0, + "loss": 0.2524, + "rewards/accuracies": 0.8515625, + "rewards/chosen": 1.0078125, + "rewards/margins": 8.25, + "rewards/rejected": -7.25, + "step": 330 + }, + { + "epoch": 0.40613496932515336, + "grad_norm": 26.526762211128467, + "learning_rate": 1.3103305785123968e-06, + "logits/chosen": -0.1767578125, + "logits/rejected": -0.283203125, + "logps/chosen": -408.0, + "logps/rejected": -356.0, + "loss": 0.2755, + "rewards/accuracies": 0.859375, + "rewards/chosen": 1.09375, + "rewards/margins": 7.625, + "rewards/rejected": -6.53125, + "step": 331 + }, + { + "epoch": 0.40736196319018403, + "grad_norm": 24.05844405102659, + "learning_rate": 1.3097107438016528e-06, + "logits/chosen": -0.1826171875, + "logits/rejected": -0.37109375, + "logps/chosen": -430.0, + "logps/rejected": -332.0, + "loss": 0.2196, + "rewards/accuracies": 0.8984375, + "rewards/chosen": 1.109375, + "rewards/margins": 7.96875, + "rewards/rejected": -6.84375, + "step": 332 + }, + { + "epoch": 0.4085889570552147, + "grad_norm": 36.41468851723906, + "learning_rate": 1.3090909090909091e-06, + "logits/chosen": 0.017822265625, + "logits/rejected": -0.1044921875, + "logps/chosen": -372.0, + "logps/rejected": -346.0, + "loss": 0.4229, + "rewards/accuracies": 0.7734375, + "rewards/chosen": 0.48046875, + "rewards/margins": 6.40625, + "rewards/rejected": -5.90625, + "step": 333 + }, + { + "epoch": 0.4098159509202454, + "grad_norm": 30.33603786358444, + "learning_rate": 1.3084710743801653e-06, + "logits/chosen": -0.1181640625, + "logits/rejected": -0.349609375, + "logps/chosen": -382.0, + "logps/rejected": -366.0, + "loss": 0.3109, + "rewards/accuracies": 0.8671875, + "rewards/chosen": 0.458984375, + "rewards/margins": 7.75, + "rewards/rejected": -7.28125, + "step": 334 + }, + { + "epoch": 0.4110429447852761, + "grad_norm": 33.01078113652063, + "learning_rate": 1.3078512396694214e-06, + "logits/chosen": -0.0213623046875, + "logits/rejected": -0.28125, + "logps/chosen": -432.0, + "logps/rejected": -332.0, + "loss": 0.2319, + "rewards/accuracies": 0.859375, + "rewards/chosen": 0.34375, + "rewards/margins": 7.0625, + "rewards/rejected": -6.71875, + "step": 335 + }, + { + "epoch": 0.41226993865030676, + "grad_norm": 31.965240510538873, + "learning_rate": 1.3072314049586778e-06, + "logits/chosen": -0.177734375, + "logits/rejected": -0.326171875, + "logps/chosen": -402.0, + "logps/rejected": -352.0, + "loss": 0.3352, + "rewards/accuracies": 0.859375, + "rewards/chosen": 0.3515625, + "rewards/margins": 7.15625, + "rewards/rejected": -6.8125, + "step": 336 + }, + { + "epoch": 0.4134969325153374, + "grad_norm": 36.98349242868308, + "learning_rate": 1.306611570247934e-06, + "logits/chosen": 0.0264892578125, + "logits/rejected": -0.1318359375, + "logps/chosen": -358.0, + "logps/rejected": -352.0, + "loss": 0.4393, + "rewards/accuracies": 0.796875, + "rewards/chosen": 0.1748046875, + "rewards/margins": 6.90625, + "rewards/rejected": -6.75, + "step": 337 + }, + { + "epoch": 0.4147239263803681, + "grad_norm": 31.08894150655132, + "learning_rate": 1.30599173553719e-06, + "logits/chosen": -0.005157470703125, + "logits/rejected": -0.1484375, + "logps/chosen": -392.0, + "logps/rejected": -354.0, + "loss": 0.3686, + "rewards/accuracies": 0.8515625, + "rewards/chosen": 0.2216796875, + "rewards/margins": 6.96875, + "rewards/rejected": -6.75, + "step": 338 + }, + { + "epoch": 0.41595092024539876, + "grad_norm": 34.40525022311832, + "learning_rate": 1.3053719008264464e-06, + "logits/chosen": 0.0654296875, + "logits/rejected": -0.06103515625, + "logps/chosen": -378.0, + "logps/rejected": -346.0, + "loss": 0.3699, + "rewards/accuracies": 0.796875, + "rewards/chosen": -0.02880859375, + "rewards/margins": 6.53125, + "rewards/rejected": -6.5625, + "step": 339 + }, + { + "epoch": 0.4171779141104294, + "grad_norm": 35.90111385254676, + "learning_rate": 1.3047520661157025e-06, + "logits/chosen": -0.03759765625, + "logits/rejected": -0.267578125, + "logps/chosen": -366.0, + "logps/rejected": -350.0, + "loss": 0.3448, + "rewards/accuracies": 0.828125, + "rewards/chosen": -0.058349609375, + "rewards/margins": 6.125, + "rewards/rejected": -6.1875, + "step": 340 + }, + { + "epoch": 0.41840490797546015, + "grad_norm": 27.984626902771293, + "learning_rate": 1.3041322314049587e-06, + "logits/chosen": -0.09521484375, + "logits/rejected": -0.3125, + "logps/chosen": -398.0, + "logps/rejected": -334.0, + "loss": 0.344, + "rewards/accuracies": 0.8515625, + "rewards/chosen": 0.56640625, + "rewards/margins": 7.3125, + "rewards/rejected": -6.75, + "step": 341 + }, + { + "epoch": 0.4196319018404908, + "grad_norm": 34.66561631382139, + "learning_rate": 1.3035123966942148e-06, + "logits/chosen": -0.021240234375, + "logits/rejected": -0.119140625, + "logps/chosen": -340.0, + "logps/rejected": -318.0, + "loss": 0.3745, + "rewards/accuracies": 0.78125, + "rewards/chosen": 0.049072265625, + "rewards/margins": 6.09375, + "rewards/rejected": -6.03125, + "step": 342 + }, + { + "epoch": 0.4208588957055215, + "grad_norm": 34.35806991693024, + "learning_rate": 1.302892561983471e-06, + "logits/chosen": -0.06396484375, + "logits/rejected": -0.1513671875, + "logps/chosen": -386.0, + "logps/rejected": -356.0, + "loss": 0.4138, + "rewards/accuracies": 0.8046875, + "rewards/chosen": -0.251953125, + "rewards/margins": 5.53125, + "rewards/rejected": -5.8125, + "step": 343 + }, + { + "epoch": 0.42208588957055215, + "grad_norm": 36.57762286698119, + "learning_rate": 1.3022727272727273e-06, + "logits/chosen": -0.0654296875, + "logits/rejected": -0.2041015625, + "logps/chosen": -408.0, + "logps/rejected": -340.0, + "loss": 0.4341, + "rewards/accuracies": 0.8125, + "rewards/chosen": 0.036376953125, + "rewards/margins": 5.875, + "rewards/rejected": -5.84375, + "step": 344 + }, + { + "epoch": 0.4233128834355828, + "grad_norm": 32.698706365667604, + "learning_rate": 1.3016528925619835e-06, + "logits/chosen": -0.04931640625, + "logits/rejected": -0.2265625, + "logps/chosen": -414.0, + "logps/rejected": -340.0, + "loss": 0.3793, + "rewards/accuracies": 0.84375, + "rewards/chosen": 0.2177734375, + "rewards/margins": 6.4375, + "rewards/rejected": -6.21875, + "step": 345 + }, + { + "epoch": 0.4245398773006135, + "grad_norm": 34.95248105852957, + "learning_rate": 1.3010330578512398e-06, + "logits/chosen": -0.036376953125, + "logits/rejected": -0.23046875, + "logps/chosen": -344.0, + "logps/rejected": -306.0, + "loss": 0.4481, + "rewards/accuracies": 0.828125, + "rewards/chosen": -0.19921875, + "rewards/margins": 5.84375, + "rewards/rejected": -6.03125, + "step": 346 + }, + { + "epoch": 0.42576687116564416, + "grad_norm": 27.628512071923325, + "learning_rate": 1.300413223140496e-06, + "logits/chosen": -0.06884765625, + "logits/rejected": -0.1875, + "logps/chosen": -370.0, + "logps/rejected": -356.0, + "loss": 0.317, + "rewards/accuracies": 0.8515625, + "rewards/chosen": 0.20703125, + "rewards/margins": 6.125, + "rewards/rejected": -5.90625, + "step": 347 + }, + { + "epoch": 0.4269938650306748, + "grad_norm": 24.80952384589789, + "learning_rate": 1.299793388429752e-06, + "logits/chosen": -0.12451171875, + "logits/rejected": -0.318359375, + "logps/chosen": -418.0, + "logps/rejected": -328.0, + "loss": 0.2563, + "rewards/accuracies": 0.8671875, + "rewards/chosen": 0.126953125, + "rewards/margins": 7.4375, + "rewards/rejected": -7.3125, + "step": 348 + }, + { + "epoch": 0.42822085889570555, + "grad_norm": 38.377577445863466, + "learning_rate": 1.2991735537190084e-06, + "logits/chosen": -0.0986328125, + "logits/rejected": -0.2412109375, + "logps/chosen": -406.0, + "logps/rejected": -322.0, + "loss": 0.4136, + "rewards/accuracies": 0.8046875, + "rewards/chosen": 0.046875, + "rewards/margins": 6.90625, + "rewards/rejected": -6.84375, + "step": 349 + }, + { + "epoch": 0.4294478527607362, + "grad_norm": 37.75213306544608, + "learning_rate": 1.2985537190082644e-06, + "logits/chosen": 0.06689453125, + "logits/rejected": -0.12890625, + "logps/chosen": -412.0, + "logps/rejected": -318.0, + "loss": 0.3697, + "rewards/accuracies": 0.796875, + "rewards/chosen": 0.045654296875, + "rewards/margins": 6.03125, + "rewards/rejected": -6.0, + "step": 350 + }, + { + "epoch": 0.4306748466257669, + "grad_norm": 35.53586893463626, + "learning_rate": 1.2979338842975207e-06, + "logits/chosen": -0.11279296875, + "logits/rejected": -0.251953125, + "logps/chosen": -342.0, + "logps/rejected": -326.0, + "loss": 0.4086, + "rewards/accuracies": 0.8125, + "rewards/chosen": -0.1337890625, + "rewards/margins": 6.03125, + "rewards/rejected": -6.1875, + "step": 351 + }, + { + "epoch": 0.43190184049079755, + "grad_norm": 30.59296136721169, + "learning_rate": 1.2973140495867769e-06, + "logits/chosen": -0.0230712890625, + "logits/rejected": -0.2177734375, + "logps/chosen": -386.0, + "logps/rejected": -334.0, + "loss": 0.4358, + "rewards/accuracies": 0.828125, + "rewards/chosen": 1.0625, + "rewards/margins": 6.34375, + "rewards/rejected": -5.28125, + "step": 352 + }, + { + "epoch": 0.4331288343558282, + "grad_norm": 24.76896622328251, + "learning_rate": 1.296694214876033e-06, + "logits/chosen": -0.07666015625, + "logits/rejected": -0.244140625, + "logps/chosen": -360.0, + "logps/rejected": -330.0, + "loss": 0.2411, + "rewards/accuracies": 0.859375, + "rewards/chosen": 0.921875, + "rewards/margins": 7.46875, + "rewards/rejected": -6.5625, + "step": 353 + }, + { + "epoch": 0.4343558282208589, + "grad_norm": 31.01854459387497, + "learning_rate": 1.2960743801652894e-06, + "logits/chosen": 0.01043701171875, + "logits/rejected": -0.162109375, + "logps/chosen": -444.0, + "logps/rejected": -378.0, + "loss": 0.3525, + "rewards/accuracies": 0.8359375, + "rewards/chosen": 0.99609375, + "rewards/margins": 6.53125, + "rewards/rejected": -5.53125, + "step": 354 + }, + { + "epoch": 0.43558282208588955, + "grad_norm": 32.456178590454805, + "learning_rate": 1.2954545454545455e-06, + "logits/chosen": -0.1572265625, + "logits/rejected": -0.24609375, + "logps/chosen": -348.0, + "logps/rejected": -356.0, + "loss": 0.4542, + "rewards/accuracies": 0.8671875, + "rewards/chosen": 0.296875, + "rewards/margins": 6.3125, + "rewards/rejected": -6.0, + "step": 355 + }, + { + "epoch": 0.4368098159509202, + "grad_norm": 31.595656958351967, + "learning_rate": 1.2948347107438016e-06, + "logits/chosen": 0.06396484375, + "logits/rejected": -0.0859375, + "logps/chosen": -396.0, + "logps/rejected": -332.0, + "loss": 0.4115, + "rewards/accuracies": 0.8046875, + "rewards/chosen": 0.46484375, + "rewards/margins": 6.0625, + "rewards/rejected": -5.59375, + "step": 356 + }, + { + "epoch": 0.43803680981595094, + "grad_norm": 28.72400215925028, + "learning_rate": 1.294214876033058e-06, + "logits/chosen": -0.23046875, + "logits/rejected": -0.384765625, + "logps/chosen": -402.0, + "logps/rejected": -340.0, + "loss": 0.3662, + "rewards/accuracies": 0.875, + "rewards/chosen": 0.63671875, + "rewards/margins": 7.40625, + "rewards/rejected": -6.78125, + "step": 357 + }, + { + "epoch": 0.4392638036809816, + "grad_norm": 29.5841283121426, + "learning_rate": 1.2935950413223141e-06, + "logits/chosen": -0.1708984375, + "logits/rejected": -0.328125, + "logps/chosen": -446.0, + "logps/rejected": -330.0, + "loss": 0.3109, + "rewards/accuracies": 0.8671875, + "rewards/chosen": 1.2109375, + "rewards/margins": 7.125, + "rewards/rejected": -5.90625, + "step": 358 + }, + { + "epoch": 0.4404907975460123, + "grad_norm": 33.37778751879116, + "learning_rate": 1.2929752066115703e-06, + "logits/chosen": -0.043212890625, + "logits/rejected": -0.154296875, + "logps/chosen": -424.0, + "logps/rejected": -334.0, + "loss": 0.3631, + "rewards/accuracies": 0.8515625, + "rewards/chosen": 0.89453125, + "rewards/margins": 6.71875, + "rewards/rejected": -5.8125, + "step": 359 + }, + { + "epoch": 0.44171779141104295, + "grad_norm": 32.72250588219352, + "learning_rate": 1.2923553719008264e-06, + "logits/chosen": -0.02587890625, + "logits/rejected": -0.1484375, + "logps/chosen": -396.0, + "logps/rejected": -358.0, + "loss": 0.2778, + "rewards/accuracies": 0.8515625, + "rewards/chosen": 0.33984375, + "rewards/margins": 6.9375, + "rewards/rejected": -6.59375, + "step": 360 + }, + { + "epoch": 0.4429447852760736, + "grad_norm": 29.99311979639803, + "learning_rate": 1.2917355371900826e-06, + "logits/chosen": 0.010498046875, + "logits/rejected": -0.177734375, + "logps/chosen": -438.0, + "logps/rejected": -382.0, + "loss": 0.3143, + "rewards/accuracies": 0.859375, + "rewards/chosen": 0.6015625, + "rewards/margins": 7.5, + "rewards/rejected": -6.875, + "step": 361 + }, + { + "epoch": 0.4441717791411043, + "grad_norm": 30.82537911112007, + "learning_rate": 1.291115702479339e-06, + "logits/chosen": -0.11083984375, + "logits/rejected": -0.267578125, + "logps/chosen": -418.0, + "logps/rejected": -344.0, + "loss": 0.3606, + "rewards/accuracies": 0.8203125, + "rewards/chosen": 0.62109375, + "rewards/margins": 7.21875, + "rewards/rejected": -6.59375, + "step": 362 + }, + { + "epoch": 0.44539877300613495, + "grad_norm": 26.541431319509446, + "learning_rate": 1.290495867768595e-06, + "logits/chosen": -0.018798828125, + "logits/rejected": -0.171875, + "logps/chosen": -392.0, + "logps/rejected": -344.0, + "loss": 0.2712, + "rewards/accuracies": 0.8984375, + "rewards/chosen": 0.77734375, + "rewards/margins": 6.8125, + "rewards/rejected": -6.03125, + "step": 363 + }, + { + "epoch": 0.4466257668711656, + "grad_norm": 36.761947246861894, + "learning_rate": 1.2898760330578512e-06, + "logits/chosen": 0.06591796875, + "logits/rejected": -0.109375, + "logps/chosen": -392.0, + "logps/rejected": -314.0, + "loss": 0.3953, + "rewards/accuracies": 0.8671875, + "rewards/chosen": 0.765625, + "rewards/margins": 6.75, + "rewards/rejected": -6.0, + "step": 364 + }, + { + "epoch": 0.44785276073619634, + "grad_norm": 27.87437703363369, + "learning_rate": 1.2892561983471076e-06, + "logits/chosen": -0.0018463134765625, + "logits/rejected": -0.1435546875, + "logps/chosen": -352.0, + "logps/rejected": -314.0, + "loss": 0.3086, + "rewards/accuracies": 0.8828125, + "rewards/chosen": 0.53125, + "rewards/margins": 6.65625, + "rewards/rejected": -6.125, + "step": 365 + }, + { + "epoch": 0.449079754601227, + "grad_norm": 32.787747807253844, + "learning_rate": 1.2886363636363637e-06, + "logits/chosen": 0.0291748046875, + "logits/rejected": -0.06787109375, + "logps/chosen": -366.0, + "logps/rejected": -332.0, + "loss": 0.447, + "rewards/accuracies": 0.8203125, + "rewards/chosen": 0.87890625, + "rewards/margins": 6.0, + "rewards/rejected": -5.125, + "step": 366 + }, + { + "epoch": 0.4503067484662577, + "grad_norm": 26.662012319210124, + "learning_rate": 1.2880165289256198e-06, + "logits/chosen": -0.142578125, + "logits/rejected": -0.326171875, + "logps/chosen": -394.0, + "logps/rejected": -336.0, + "loss": 0.2357, + "rewards/accuracies": 0.8984375, + "rewards/chosen": 0.9453125, + "rewards/margins": 6.65625, + "rewards/rejected": -5.71875, + "step": 367 + }, + { + "epoch": 0.45153374233128835, + "grad_norm": 36.83996177643189, + "learning_rate": 1.287396694214876e-06, + "logits/chosen": 0.06787109375, + "logits/rejected": 0.0234375, + "logps/chosen": -368.0, + "logps/rejected": -328.0, + "loss": 0.3909, + "rewards/accuracies": 0.84375, + "rewards/chosen": 0.6796875, + "rewards/margins": 6.0, + "rewards/rejected": -5.3125, + "step": 368 + }, + { + "epoch": 0.452760736196319, + "grad_norm": 35.726856908761434, + "learning_rate": 1.2867768595041321e-06, + "logits/chosen": 0.05322265625, + "logits/rejected": -0.06591796875, + "logps/chosen": -344.0, + "logps/rejected": -298.0, + "loss": 0.3738, + "rewards/accuracies": 0.8359375, + "rewards/chosen": 0.169921875, + "rewards/margins": 5.71875, + "rewards/rejected": -5.5625, + "step": 369 + }, + { + "epoch": 0.4539877300613497, + "grad_norm": 31.108323501043053, + "learning_rate": 1.2861570247933885e-06, + "logits/chosen": -0.1484375, + "logits/rejected": -0.328125, + "logps/chosen": -402.0, + "logps/rejected": -348.0, + "loss": 0.369, + "rewards/accuracies": 0.84375, + "rewards/chosen": 1.75, + "rewards/margins": 6.65625, + "rewards/rejected": -4.90625, + "step": 370 + }, + { + "epoch": 0.45521472392638035, + "grad_norm": 31.06340590217051, + "learning_rate": 1.2855371900826446e-06, + "logits/chosen": -0.130859375, + "logits/rejected": -0.298828125, + "logps/chosen": -460.0, + "logps/rejected": -364.0, + "loss": 0.3181, + "rewards/accuracies": 0.8984375, + "rewards/chosen": 1.625, + "rewards/margins": 6.0, + "rewards/rejected": -4.375, + "step": 371 + }, + { + "epoch": 0.456441717791411, + "grad_norm": 28.189166139003635, + "learning_rate": 1.2849173553719008e-06, + "logits/chosen": -0.0301513671875, + "logits/rejected": -0.1875, + "logps/chosen": -410.0, + "logps/rejected": -324.0, + "loss": 0.2419, + "rewards/accuracies": 0.890625, + "rewards/chosen": 1.5078125, + "rewards/margins": 6.4375, + "rewards/rejected": -4.9375, + "step": 372 + }, + { + "epoch": 0.45766871165644174, + "grad_norm": 33.21287873989905, + "learning_rate": 1.2842975206611571e-06, + "logits/chosen": -0.053466796875, + "logits/rejected": -0.1591796875, + "logps/chosen": -362.0, + "logps/rejected": -322.0, + "loss": 0.4071, + "rewards/accuracies": 0.8359375, + "rewards/chosen": 0.58203125, + "rewards/margins": 5.71875, + "rewards/rejected": -5.125, + "step": 373 + }, + { + "epoch": 0.4588957055214724, + "grad_norm": 26.151862774833457, + "learning_rate": 1.2836776859504132e-06, + "logits/chosen": -0.0908203125, + "logits/rejected": -0.2197265625, + "logps/chosen": -382.0, + "logps/rejected": -316.0, + "loss": 0.2534, + "rewards/accuracies": 0.8984375, + "rewards/chosen": 2.0, + "rewards/margins": 6.78125, + "rewards/rejected": -4.78125, + "step": 374 + }, + { + "epoch": 0.4601226993865031, + "grad_norm": 32.35488575008133, + "learning_rate": 1.2830578512396696e-06, + "logits/chosen": -0.14453125, + "logits/rejected": -0.2177734375, + "logps/chosen": -392.0, + "logps/rejected": -350.0, + "loss": 0.2771, + "rewards/accuracies": 0.859375, + "rewards/chosen": 1.7109375, + "rewards/margins": 5.8125, + "rewards/rejected": -4.09375, + "step": 375 + }, + { + "epoch": 0.46134969325153374, + "grad_norm": 34.001922340864844, + "learning_rate": 1.2824380165289257e-06, + "logits/chosen": -0.00946044921875, + "logits/rejected": -0.15625, + "logps/chosen": -376.0, + "logps/rejected": -344.0, + "loss": 0.3408, + "rewards/accuracies": 0.8515625, + "rewards/chosen": 1.6953125, + "rewards/margins": 5.75, + "rewards/rejected": -4.03125, + "step": 376 + }, + { + "epoch": 0.4625766871165644, + "grad_norm": 35.73955039729585, + "learning_rate": 1.2818181818181817e-06, + "logits/chosen": -0.0927734375, + "logits/rejected": -0.150390625, + "logps/chosen": -352.0, + "logps/rejected": -318.0, + "loss": 0.4247, + "rewards/accuracies": 0.7734375, + "rewards/chosen": 1.5625, + "rewards/margins": 6.65625, + "rewards/rejected": -5.0625, + "step": 377 + }, + { + "epoch": 0.4638036809815951, + "grad_norm": 33.82458737973786, + "learning_rate": 1.281198347107438e-06, + "logits/chosen": -0.1396484375, + "logits/rejected": -0.376953125, + "logps/chosen": -376.0, + "logps/rejected": -332.0, + "loss": 0.3987, + "rewards/accuracies": 0.796875, + "rewards/chosen": 1.625, + "rewards/margins": 6.625, + "rewards/rejected": -5.0, + "step": 378 + }, + { + "epoch": 0.46503067484662575, + "grad_norm": 31.86429872362659, + "learning_rate": 1.2805785123966942e-06, + "logits/chosen": -0.1953125, + "logits/rejected": -0.359375, + "logps/chosen": -408.0, + "logps/rejected": -344.0, + "loss": 0.3028, + "rewards/accuracies": 0.8671875, + "rewards/chosen": 1.71875, + "rewards/margins": 7.1875, + "rewards/rejected": -5.46875, + "step": 379 + }, + { + "epoch": 0.4662576687116564, + "grad_norm": 31.446849732786838, + "learning_rate": 1.2799586776859505e-06, + "logits/chosen": -0.09326171875, + "logits/rejected": -0.2314453125, + "logps/chosen": -382.0, + "logps/rejected": -326.0, + "loss": 0.3068, + "rewards/accuracies": 0.859375, + "rewards/chosen": 1.625, + "rewards/margins": 6.78125, + "rewards/rejected": -5.15625, + "step": 380 + }, + { + "epoch": 0.46748466257668714, + "grad_norm": 31.679856008810177, + "learning_rate": 1.2793388429752067e-06, + "logits/chosen": -0.189453125, + "logits/rejected": -0.3671875, + "logps/chosen": -372.0, + "logps/rejected": -328.0, + "loss": 0.374, + "rewards/accuracies": 0.8515625, + "rewards/chosen": 1.046875, + "rewards/margins": 7.0, + "rewards/rejected": -5.9375, + "step": 381 + }, + { + "epoch": 0.4687116564417178, + "grad_norm": 31.436122523167693, + "learning_rate": 1.2787190082644628e-06, + "logits/chosen": -0.185546875, + "logits/rejected": -0.279296875, + "logps/chosen": -380.0, + "logps/rejected": -340.0, + "loss": 0.3158, + "rewards/accuracies": 0.8515625, + "rewards/chosen": 1.078125, + "rewards/margins": 6.875, + "rewards/rejected": -5.8125, + "step": 382 + }, + { + "epoch": 0.4699386503067485, + "grad_norm": 35.065632595795535, + "learning_rate": 1.2780991735537192e-06, + "logits/chosen": -0.09912109375, + "logits/rejected": -0.2578125, + "logps/chosen": -402.0, + "logps/rejected": -378.0, + "loss": 0.4456, + "rewards/accuracies": 0.8359375, + "rewards/chosen": 1.1015625, + "rewards/margins": 6.78125, + "rewards/rejected": -5.6875, + "step": 383 + }, + { + "epoch": 0.47116564417177914, + "grad_norm": 33.79101177267741, + "learning_rate": 1.2774793388429753e-06, + "logits/chosen": -0.2001953125, + "logits/rejected": -0.435546875, + "logps/chosen": -398.0, + "logps/rejected": -342.0, + "loss": 0.2961, + "rewards/accuracies": 0.890625, + "rewards/chosen": 0.640625, + "rewards/margins": 7.6875, + "rewards/rejected": -7.03125, + "step": 384 + }, + { + "epoch": 0.4723926380368098, + "grad_norm": 36.518171637968166, + "learning_rate": 1.2768595041322314e-06, + "logits/chosen": -0.10302734375, + "logits/rejected": -0.283203125, + "logps/chosen": -394.0, + "logps/rejected": -330.0, + "loss": 0.3771, + "rewards/accuracies": 0.8046875, + "rewards/chosen": 0.283203125, + "rewards/margins": 6.34375, + "rewards/rejected": -6.0625, + "step": 385 + }, + { + "epoch": 0.4736196319018405, + "grad_norm": 34.95682432591222, + "learning_rate": 1.2762396694214876e-06, + "logits/chosen": -0.043701171875, + "logits/rejected": -0.2041015625, + "logps/chosen": -342.0, + "logps/rejected": -310.0, + "loss": 0.369, + "rewards/accuracies": 0.875, + "rewards/chosen": -0.3203125, + "rewards/margins": 6.09375, + "rewards/rejected": -6.40625, + "step": 386 + }, + { + "epoch": 0.47484662576687114, + "grad_norm": 64.7195376952873, + "learning_rate": 1.2756198347107437e-06, + "logits/chosen": -0.068359375, + "logits/rejected": -0.271484375, + "logps/chosen": -422.0, + "logps/rejected": -350.0, + "loss": 0.4035, + "rewards/accuracies": 0.8359375, + "rewards/chosen": 0.6171875, + "rewards/margins": 7.34375, + "rewards/rejected": -6.75, + "step": 387 + }, + { + "epoch": 0.47607361963190187, + "grad_norm": 42.357455920572896, + "learning_rate": 1.275e-06, + "logits/chosen": -0.08349609375, + "logits/rejected": -0.1826171875, + "logps/chosen": -406.0, + "logps/rejected": -326.0, + "loss": 0.5163, + "rewards/accuracies": 0.7890625, + "rewards/chosen": -0.04736328125, + "rewards/margins": 5.53125, + "rewards/rejected": -5.59375, + "step": 388 + }, + { + "epoch": 0.47730061349693254, + "grad_norm": 38.953026072404185, + "learning_rate": 1.2743801652892562e-06, + "logits/chosen": -0.1435546875, + "logits/rejected": -0.2236328125, + "logps/chosen": -412.0, + "logps/rejected": -340.0, + "loss": 0.4455, + "rewards/accuracies": 0.8203125, + "rewards/chosen": 0.099609375, + "rewards/margins": 6.3125, + "rewards/rejected": -6.1875, + "step": 389 + }, + { + "epoch": 0.4785276073619632, + "grad_norm": 38.5573431700193, + "learning_rate": 1.2737603305785124e-06, + "logits/chosen": -0.2216796875, + "logits/rejected": -0.34375, + "logps/chosen": -372.0, + "logps/rejected": -344.0, + "loss": 0.492, + "rewards/accuracies": 0.828125, + "rewards/chosen": -0.62890625, + "rewards/margins": 6.0, + "rewards/rejected": -6.65625, + "step": 390 + }, + { + "epoch": 0.47975460122699387, + "grad_norm": 37.845045364551986, + "learning_rate": 1.2731404958677687e-06, + "logits/chosen": -0.1201171875, + "logits/rejected": -0.23828125, + "logps/chosen": -390.0, + "logps/rejected": -344.0, + "loss": 0.4807, + "rewards/accuracies": 0.8203125, + "rewards/chosen": -0.181640625, + "rewards/margins": 6.4375, + "rewards/rejected": -6.625, + "step": 391 + }, + { + "epoch": 0.48098159509202454, + "grad_norm": 29.781954844950388, + "learning_rate": 1.2725206611570249e-06, + "logits/chosen": -0.060791015625, + "logits/rejected": -0.208984375, + "logps/chosen": -362.0, + "logps/rejected": -330.0, + "loss": 0.3103, + "rewards/accuracies": 0.875, + "rewards/chosen": -0.06982421875, + "rewards/margins": 6.59375, + "rewards/rejected": -6.65625, + "step": 392 + }, + { + "epoch": 0.4822085889570552, + "grad_norm": 33.5926238632541, + "learning_rate": 1.271900826446281e-06, + "logits/chosen": -0.0286865234375, + "logits/rejected": -0.166015625, + "logps/chosen": -416.0, + "logps/rejected": -374.0, + "loss": 0.3323, + "rewards/accuracies": 0.8359375, + "rewards/chosen": -0.263671875, + "rewards/margins": 6.4375, + "rewards/rejected": -6.6875, + "step": 393 + }, + { + "epoch": 0.4834355828220859, + "grad_norm": 32.087849602750985, + "learning_rate": 1.2712809917355373e-06, + "logits/chosen": -0.09521484375, + "logits/rejected": -0.298828125, + "logps/chosen": -398.0, + "logps/rejected": -342.0, + "loss": 0.262, + "rewards/accuracies": 0.84375, + "rewards/chosen": -0.166015625, + "rewards/margins": 6.1875, + "rewards/rejected": -6.375, + "step": 394 + }, + { + "epoch": 0.48466257668711654, + "grad_norm": 34.842476262528464, + "learning_rate": 1.2706611570247933e-06, + "logits/chosen": -0.1025390625, + "logits/rejected": -0.216796875, + "logps/chosen": -448.0, + "logps/rejected": -356.0, + "loss": 0.3764, + "rewards/accuracies": 0.8046875, + "rewards/chosen": -0.09375, + "rewards/margins": 6.125, + "rewards/rejected": -6.21875, + "step": 395 + }, + { + "epoch": 0.48588957055214727, + "grad_norm": 22.42844686295793, + "learning_rate": 1.2700413223140496e-06, + "logits/chosen": -0.1806640625, + "logits/rejected": -0.388671875, + "logps/chosen": -348.0, + "logps/rejected": -332.0, + "loss": 0.2277, + "rewards/accuracies": 0.90625, + "rewards/chosen": -0.54296875, + "rewards/margins": 6.3125, + "rewards/rejected": -6.875, + "step": 396 + }, + { + "epoch": 0.48711656441717793, + "grad_norm": 33.97555781703141, + "learning_rate": 1.2694214876033058e-06, + "logits/chosen": -0.1513671875, + "logits/rejected": -0.314453125, + "logps/chosen": -388.0, + "logps/rejected": -350.0, + "loss": 0.3848, + "rewards/accuracies": 0.828125, + "rewards/chosen": -0.625, + "rewards/margins": 6.78125, + "rewards/rejected": -7.40625, + "step": 397 + }, + { + "epoch": 0.4883435582822086, + "grad_norm": 39.08030474072834, + "learning_rate": 1.268801652892562e-06, + "logits/chosen": -0.1328125, + "logits/rejected": -0.27734375, + "logps/chosen": -428.0, + "logps/rejected": -346.0, + "loss": 0.4427, + "rewards/accuracies": 0.796875, + "rewards/chosen": -0.08740234375, + "rewards/margins": 6.375, + "rewards/rejected": -6.46875, + "step": 398 + }, + { + "epoch": 0.48957055214723927, + "grad_norm": 33.83840388728663, + "learning_rate": 1.2681818181818183e-06, + "logits/chosen": -0.1611328125, + "logits/rejected": -0.291015625, + "logps/chosen": -394.0, + "logps/rejected": -346.0, + "loss": 0.2963, + "rewards/accuracies": 0.859375, + "rewards/chosen": -0.0035858154296875, + "rewards/margins": 6.53125, + "rewards/rejected": -6.53125, + "step": 399 + }, + { + "epoch": 0.49079754601226994, + "grad_norm": 41.49480537946823, + "learning_rate": 1.2675619834710744e-06, + "logits/chosen": -0.09130859375, + "logits/rejected": -0.1650390625, + "logps/chosen": -384.0, + "logps/rejected": -368.0, + "loss": 0.5062, + "rewards/accuracies": 0.796875, + "rewards/chosen": -0.5390625, + "rewards/margins": 5.59375, + "rewards/rejected": -6.15625, + "step": 400 + }, + { + "epoch": 0.4920245398773006, + "grad_norm": 27.800803034272004, + "learning_rate": 1.2669421487603308e-06, + "logits/chosen": -0.23828125, + "logits/rejected": -0.373046875, + "logps/chosen": -456.0, + "logps/rejected": -390.0, + "loss": 0.2938, + "rewards/accuracies": 0.8671875, + "rewards/chosen": 0.40625, + "rewards/margins": 7.59375, + "rewards/rejected": -7.1875, + "step": 401 + }, + { + "epoch": 0.49325153374233127, + "grad_norm": 32.09882908595564, + "learning_rate": 1.266322314049587e-06, + "logits/chosen": -0.0791015625, + "logits/rejected": -0.1796875, + "logps/chosen": -394.0, + "logps/rejected": -350.0, + "loss": 0.349, + "rewards/accuracies": 0.8359375, + "rewards/chosen": -0.04443359375, + "rewards/margins": 6.5, + "rewards/rejected": -6.53125, + "step": 402 + }, + { + "epoch": 0.49447852760736194, + "grad_norm": 29.930227321008882, + "learning_rate": 1.265702479338843e-06, + "logits/chosen": -0.07958984375, + "logits/rejected": -0.15625, + "logps/chosen": -396.0, + "logps/rejected": -374.0, + "loss": 0.341, + "rewards/accuracies": 0.8671875, + "rewards/chosen": -0.033447265625, + "rewards/margins": 6.78125, + "rewards/rejected": -6.8125, + "step": 403 + }, + { + "epoch": 0.49570552147239266, + "grad_norm": 25.987833840695732, + "learning_rate": 1.2650826446280992e-06, + "logits/chosen": -0.021728515625, + "logits/rejected": -0.150390625, + "logps/chosen": -372.0, + "logps/rejected": -348.0, + "loss": 0.2606, + "rewards/accuracies": 0.8984375, + "rewards/chosen": 0.1669921875, + "rewards/margins": 7.0625, + "rewards/rejected": -6.90625, + "step": 404 + }, + { + "epoch": 0.49693251533742333, + "grad_norm": 38.14679307536266, + "learning_rate": 1.2644628099173553e-06, + "logits/chosen": -0.1064453125, + "logits/rejected": -0.328125, + "logps/chosen": -370.0, + "logps/rejected": -322.0, + "loss": 0.4214, + "rewards/accuracies": 0.8125, + "rewards/chosen": 0.6484375, + "rewards/margins": 6.65625, + "rewards/rejected": -6.03125, + "step": 405 + }, + { + "epoch": 0.498159509202454, + "grad_norm": 37.82975021877776, + "learning_rate": 1.2638429752066117e-06, + "logits/chosen": 0.0166015625, + "logits/rejected": -0.0625, + "logps/chosen": -378.0, + "logps/rejected": -340.0, + "loss": 0.5427, + "rewards/accuracies": 0.8046875, + "rewards/chosen": 0.3515625, + "rewards/margins": 6.09375, + "rewards/rejected": -5.75, + "step": 406 + }, + { + "epoch": 0.49938650306748467, + "grad_norm": 38.332851373691135, + "learning_rate": 1.2632231404958678e-06, + "logits/chosen": 0.0115966796875, + "logits/rejected": -0.134765625, + "logps/chosen": -408.0, + "logps/rejected": -344.0, + "loss": 0.4382, + "rewards/accuracies": 0.8046875, + "rewards/chosen": 1.34375, + "rewards/margins": 7.40625, + "rewards/rejected": -6.0625, + "step": 407 + }, + { + "epoch": 0.5006134969325153, + "grad_norm": 41.6208959235071, + "learning_rate": 1.262603305785124e-06, + "logits/chosen": -0.083984375, + "logits/rejected": -0.1474609375, + "logps/chosen": -426.0, + "logps/rejected": -360.0, + "loss": 0.5081, + "rewards/accuracies": 0.7734375, + "rewards/chosen": -0.091796875, + "rewards/margins": 5.1875, + "rewards/rejected": -5.28125, + "step": 408 + }, + { + "epoch": 0.501840490797546, + "grad_norm": 42.88666645295333, + "learning_rate": 1.2619834710743803e-06, + "logits/chosen": -0.00860595703125, + "logits/rejected": -0.205078125, + "logps/chosen": -426.0, + "logps/rejected": -342.0, + "loss": 0.4119, + "rewards/accuracies": 0.8203125, + "rewards/chosen": 0.56640625, + "rewards/margins": 6.71875, + "rewards/rejected": -6.15625, + "step": 409 + }, + { + "epoch": 0.5030674846625767, + "grad_norm": 36.428858153344706, + "learning_rate": 1.2613636363636365e-06, + "logits/chosen": -0.00095367431640625, + "logits/rejected": -0.1943359375, + "logps/chosen": -388.0, + "logps/rejected": -360.0, + "loss": 0.4038, + "rewards/accuracies": 0.796875, + "rewards/chosen": 0.3828125, + "rewards/margins": 6.125, + "rewards/rejected": -5.75, + "step": 410 + }, + { + "epoch": 0.5042944785276073, + "grad_norm": 33.21094894864319, + "learning_rate": 1.2607438016528926e-06, + "logits/chosen": 0.0140380859375, + "logits/rejected": -0.12890625, + "logps/chosen": -342.0, + "logps/rejected": -320.0, + "loss": 0.3058, + "rewards/accuracies": 0.8359375, + "rewards/chosen": 0.1298828125, + "rewards/margins": 6.46875, + "rewards/rejected": -6.34375, + "step": 411 + }, + { + "epoch": 0.505521472392638, + "grad_norm": 42.15288062503197, + "learning_rate": 1.260123966942149e-06, + "logits/chosen": 0.007080078125, + "logits/rejected": -0.1259765625, + "logps/chosen": -428.0, + "logps/rejected": -368.0, + "loss": 0.4444, + "rewards/accuracies": 0.796875, + "rewards/chosen": 0.30078125, + "rewards/margins": 5.96875, + "rewards/rejected": -5.6875, + "step": 412 + }, + { + "epoch": 0.5067484662576687, + "grad_norm": 37.795734491735246, + "learning_rate": 1.2595041322314049e-06, + "logits/chosen": -0.042236328125, + "logits/rejected": -0.26171875, + "logps/chosen": -408.0, + "logps/rejected": -332.0, + "loss": 0.3878, + "rewards/accuracies": 0.8125, + "rewards/chosen": 0.9140625, + "rewards/margins": 6.96875, + "rewards/rejected": -6.0625, + "step": 413 + }, + { + "epoch": 0.5079754601226993, + "grad_norm": 37.27154344231617, + "learning_rate": 1.2588842975206612e-06, + "logits/chosen": -0.0859375, + "logits/rejected": -0.2265625, + "logps/chosen": -406.0, + "logps/rejected": -354.0, + "loss": 0.3559, + "rewards/accuracies": 0.84375, + "rewards/chosen": 0.85546875, + "rewards/margins": 6.71875, + "rewards/rejected": -5.84375, + "step": 414 + }, + { + "epoch": 0.50920245398773, + "grad_norm": 30.678885373774655, + "learning_rate": 1.2582644628099174e-06, + "logits/chosen": -0.041748046875, + "logits/rejected": -0.212890625, + "logps/chosen": -350.0, + "logps/rejected": -332.0, + "loss": 0.3048, + "rewards/accuracies": 0.828125, + "rewards/chosen": 0.52734375, + "rewards/margins": 6.75, + "rewards/rejected": -6.21875, + "step": 415 + }, + { + "epoch": 0.5104294478527608, + "grad_norm": 26.882010464050307, + "learning_rate": 1.2576446280991735e-06, + "logits/chosen": -0.004730224609375, + "logits/rejected": -0.2197265625, + "logps/chosen": -360.0, + "logps/rejected": -324.0, + "loss": 0.1899, + "rewards/accuracies": 0.890625, + "rewards/chosen": 0.458984375, + "rewards/margins": 6.75, + "rewards/rejected": -6.28125, + "step": 416 + }, + { + "epoch": 0.5116564417177915, + "grad_norm": 30.354406082685717, + "learning_rate": 1.2570247933884299e-06, + "logits/chosen": -0.2197265625, + "logits/rejected": -0.384765625, + "logps/chosen": -438.0, + "logps/rejected": -382.0, + "loss": 0.3453, + "rewards/accuracies": 0.84375, + "rewards/chosen": 1.4765625, + "rewards/margins": 8.125, + "rewards/rejected": -6.65625, + "step": 417 + }, + { + "epoch": 0.5128834355828221, + "grad_norm": 33.53763392215037, + "learning_rate": 1.256404958677686e-06, + "logits/chosen": -0.126953125, + "logits/rejected": -0.2099609375, + "logps/chosen": -368.0, + "logps/rejected": -306.0, + "loss": 0.3744, + "rewards/accuracies": 0.859375, + "rewards/chosen": 0.7890625, + "rewards/margins": 7.1875, + "rewards/rejected": -6.375, + "step": 418 + }, + { + "epoch": 0.5141104294478528, + "grad_norm": 28.10440726715766, + "learning_rate": 1.2557851239669421e-06, + "logits/chosen": -0.0546875, + "logits/rejected": -0.21875, + "logps/chosen": -380.0, + "logps/rejected": -342.0, + "loss": 0.3206, + "rewards/accuracies": 0.8671875, + "rewards/chosen": 0.99609375, + "rewards/margins": 7.78125, + "rewards/rejected": -6.78125, + "step": 419 + }, + { + "epoch": 0.5153374233128835, + "grad_norm": 30.627621824355007, + "learning_rate": 1.2551652892561985e-06, + "logits/chosen": 0.0262451171875, + "logits/rejected": -0.134765625, + "logps/chosen": -346.0, + "logps/rejected": -312.0, + "loss": 0.3514, + "rewards/accuracies": 0.828125, + "rewards/chosen": 0.796875, + "rewards/margins": 6.4375, + "rewards/rejected": -5.65625, + "step": 420 + }, + { + "epoch": 0.5165644171779141, + "grad_norm": 31.844076287237254, + "learning_rate": 1.2545454545454544e-06, + "logits/chosen": -0.01544189453125, + "logits/rejected": -0.2158203125, + "logps/chosen": -364.0, + "logps/rejected": -320.0, + "loss": 0.4123, + "rewards/accuracies": 0.8125, + "rewards/chosen": 1.1875, + "rewards/margins": 6.8125, + "rewards/rejected": -5.625, + "step": 421 + }, + { + "epoch": 0.5177914110429448, + "grad_norm": 33.522822368907185, + "learning_rate": 1.2539256198347108e-06, + "logits/chosen": 0.0595703125, + "logits/rejected": -0.0986328125, + "logps/chosen": -360.0, + "logps/rejected": -302.0, + "loss": 0.3724, + "rewards/accuracies": 0.84375, + "rewards/chosen": 0.52734375, + "rewards/margins": 6.6875, + "rewards/rejected": -6.15625, + "step": 422 + }, + { + "epoch": 0.5190184049079755, + "grad_norm": 37.40841671544669, + "learning_rate": 1.253305785123967e-06, + "logits/chosen": -0.0595703125, + "logits/rejected": -0.20703125, + "logps/chosen": -398.0, + "logps/rejected": -362.0, + "loss": 0.4749, + "rewards/accuracies": 0.8125, + "rewards/chosen": 0.578125, + "rewards/margins": 6.0, + "rewards/rejected": -5.4375, + "step": 423 + }, + { + "epoch": 0.5202453987730061, + "grad_norm": 34.546118881756456, + "learning_rate": 1.252685950413223e-06, + "logits/chosen": 0.0235595703125, + "logits/rejected": -0.166015625, + "logps/chosen": -378.0, + "logps/rejected": -342.0, + "loss": 0.4503, + "rewards/accuracies": 0.8359375, + "rewards/chosen": 1.2109375, + "rewards/margins": 7.09375, + "rewards/rejected": -5.875, + "step": 424 + }, + { + "epoch": 0.5214723926380368, + "grad_norm": 37.84213132068162, + "learning_rate": 1.2520661157024794e-06, + "logits/chosen": -0.0123291015625, + "logits/rejected": -0.11572265625, + "logps/chosen": -396.0, + "logps/rejected": -348.0, + "loss": 0.5144, + "rewards/accuracies": 0.796875, + "rewards/chosen": 1.3203125, + "rewards/margins": 6.84375, + "rewards/rejected": -5.53125, + "step": 425 + }, + { + "epoch": 0.5226993865030675, + "grad_norm": 36.174280199739684, + "learning_rate": 1.2514462809917356e-06, + "logits/chosen": -0.06396484375, + "logits/rejected": -0.166015625, + "logps/chosen": -370.0, + "logps/rejected": -348.0, + "loss": 0.4214, + "rewards/accuracies": 0.875, + "rewards/chosen": 1.015625, + "rewards/margins": 7.0625, + "rewards/rejected": -6.0625, + "step": 426 + }, + { + "epoch": 0.5239263803680981, + "grad_norm": 45.37106746020955, + "learning_rate": 1.2508264462809917e-06, + "logits/chosen": -0.01483154296875, + "logits/rejected": -0.1279296875, + "logps/chosen": -378.0, + "logps/rejected": -334.0, + "loss": 0.4318, + "rewards/accuracies": 0.8203125, + "rewards/chosen": 1.1171875, + "rewards/margins": 6.75, + "rewards/rejected": -5.625, + "step": 427 + }, + { + "epoch": 0.5251533742331288, + "grad_norm": 33.589023647082186, + "learning_rate": 1.250206611570248e-06, + "logits/chosen": -0.0419921875, + "logits/rejected": -0.1533203125, + "logps/chosen": -364.0, + "logps/rejected": -324.0, + "loss": 0.4488, + "rewards/accuracies": 0.8046875, + "rewards/chosen": 0.462890625, + "rewards/margins": 6.46875, + "rewards/rejected": -6.0, + "step": 428 + }, + { + "epoch": 0.5263803680981595, + "grad_norm": 30.11424815064351, + "learning_rate": 1.2495867768595042e-06, + "logits/chosen": -0.044677734375, + "logits/rejected": -0.1494140625, + "logps/chosen": -360.0, + "logps/rejected": -376.0, + "loss": 0.2639, + "rewards/accuracies": 0.8671875, + "rewards/chosen": 0.373046875, + "rewards/margins": 7.96875, + "rewards/rejected": -7.59375, + "step": 429 + }, + { + "epoch": 0.5276073619631901, + "grad_norm": 23.36575464672381, + "learning_rate": 1.2489669421487603e-06, + "logits/chosen": -0.11767578125, + "logits/rejected": -0.28125, + "logps/chosen": -394.0, + "logps/rejected": -350.0, + "loss": 0.2208, + "rewards/accuracies": 0.875, + "rewards/chosen": 0.921875, + "rewards/margins": 7.5625, + "rewards/rejected": -6.625, + "step": 430 + }, + { + "epoch": 0.5288343558282208, + "grad_norm": 39.63938191210347, + "learning_rate": 1.2483471074380165e-06, + "logits/chosen": 0.01806640625, + "logits/rejected": -0.150390625, + "logps/chosen": -386.0, + "logps/rejected": -346.0, + "loss": 0.4684, + "rewards/accuracies": 0.8203125, + "rewards/chosen": 0.12255859375, + "rewards/margins": 6.4375, + "rewards/rejected": -6.3125, + "step": 431 + }, + { + "epoch": 0.5300613496932516, + "grad_norm": 35.90106246867024, + "learning_rate": 1.2477272727272726e-06, + "logits/chosen": 0.06689453125, + "logits/rejected": 0.035888671875, + "logps/chosen": -382.0, + "logps/rejected": -354.0, + "loss": 0.4356, + "rewards/accuracies": 0.84375, + "rewards/chosen": 0.78515625, + "rewards/margins": 6.71875, + "rewards/rejected": -5.9375, + "step": 432 + }, + { + "epoch": 0.5312883435582823, + "grad_norm": 43.12860297142818, + "learning_rate": 1.247107438016529e-06, + "logits/chosen": 0.08740234375, + "logits/rejected": -0.1328125, + "logps/chosen": -388.0, + "logps/rejected": -320.0, + "loss": 0.4803, + "rewards/accuracies": 0.8046875, + "rewards/chosen": -0.12353515625, + "rewards/margins": 6.4375, + "rewards/rejected": -6.5625, + "step": 433 + }, + { + "epoch": 0.5325153374233129, + "grad_norm": 37.083326074229944, + "learning_rate": 1.2464876033057851e-06, + "logits/chosen": -0.0869140625, + "logits/rejected": -0.1611328125, + "logps/chosen": -390.0, + "logps/rejected": -356.0, + "loss": 0.3206, + "rewards/accuracies": 0.859375, + "rewards/chosen": 0.07666015625, + "rewards/margins": 6.375, + "rewards/rejected": -6.28125, + "step": 434 + }, + { + "epoch": 0.5337423312883436, + "grad_norm": 37.88064967580844, + "learning_rate": 1.2458677685950415e-06, + "logits/chosen": -0.01458740234375, + "logits/rejected": -0.208984375, + "logps/chosen": -394.0, + "logps/rejected": -362.0, + "loss": 0.3939, + "rewards/accuracies": 0.8359375, + "rewards/chosen": 0.51953125, + "rewards/margins": 7.15625, + "rewards/rejected": -6.65625, + "step": 435 + }, + { + "epoch": 0.5349693251533743, + "grad_norm": 30.102357096787667, + "learning_rate": 1.2452479338842976e-06, + "logits/chosen": -0.0250244140625, + "logits/rejected": -0.2255859375, + "logps/chosen": -422.0, + "logps/rejected": -358.0, + "loss": 0.2562, + "rewards/accuracies": 0.8828125, + "rewards/chosen": 1.1171875, + "rewards/margins": 7.75, + "rewards/rejected": -6.625, + "step": 436 + }, + { + "epoch": 0.5361963190184049, + "grad_norm": 25.03840429478704, + "learning_rate": 1.2446280991735538e-06, + "logits/chosen": -0.080078125, + "logits/rejected": -0.25, + "logps/chosen": -388.0, + "logps/rejected": -350.0, + "loss": 0.241, + "rewards/accuracies": 0.8671875, + "rewards/chosen": 0.41796875, + "rewards/margins": 7.28125, + "rewards/rejected": -6.875, + "step": 437 + }, + { + "epoch": 0.5374233128834356, + "grad_norm": 36.15225304248156, + "learning_rate": 1.24400826446281e-06, + "logits/chosen": 0.02001953125, + "logits/rejected": -0.06982421875, + "logps/chosen": -366.0, + "logps/rejected": -334.0, + "loss": 0.5276, + "rewards/accuracies": 0.8125, + "rewards/chosen": -0.0250244140625, + "rewards/margins": 5.65625, + "rewards/rejected": -5.6875, + "step": 438 + }, + { + "epoch": 0.5386503067484663, + "grad_norm": 34.58997013752512, + "learning_rate": 1.243388429752066e-06, + "logits/chosen": 0.0625, + "logits/rejected": -0.126953125, + "logps/chosen": -372.0, + "logps/rejected": -348.0, + "loss": 0.3671, + "rewards/accuracies": 0.84375, + "rewards/chosen": -0.0703125, + "rewards/margins": 6.65625, + "rewards/rejected": -6.71875, + "step": 439 + }, + { + "epoch": 0.5398773006134969, + "grad_norm": 43.86154514789445, + "learning_rate": 1.2427685950413224e-06, + "logits/chosen": 0.0218505859375, + "logits/rejected": -0.0751953125, + "logps/chosen": -392.0, + "logps/rejected": -344.0, + "loss": 0.4825, + "rewards/accuracies": 0.8046875, + "rewards/chosen": -0.515625, + "rewards/margins": 5.4375, + "rewards/rejected": -5.96875, + "step": 440 + }, + { + "epoch": 0.5411042944785276, + "grad_norm": 32.02976017157544, + "learning_rate": 1.2421487603305785e-06, + "logits/chosen": 0.09619140625, + "logits/rejected": -0.0108642578125, + "logps/chosen": -340.0, + "logps/rejected": -314.0, + "loss": 0.4408, + "rewards/accuracies": 0.828125, + "rewards/chosen": -0.0869140625, + "rewards/margins": 6.15625, + "rewards/rejected": -6.21875, + "step": 441 + }, + { + "epoch": 0.5423312883435583, + "grad_norm": 31.122024927447573, + "learning_rate": 1.2415289256198347e-06, + "logits/chosen": -0.064453125, + "logits/rejected": -0.1982421875, + "logps/chosen": -390.0, + "logps/rejected": -352.0, + "loss": 0.3842, + "rewards/accuracies": 0.859375, + "rewards/chosen": 0.51171875, + "rewards/margins": 6.5625, + "rewards/rejected": -6.0625, + "step": 442 + }, + { + "epoch": 0.5435582822085889, + "grad_norm": 35.35206800809496, + "learning_rate": 1.240909090909091e-06, + "logits/chosen": -0.024169921875, + "logits/rejected": -0.1435546875, + "logps/chosen": -372.0, + "logps/rejected": -332.0, + "loss": 0.3707, + "rewards/accuracies": 0.828125, + "rewards/chosen": 0.69140625, + "rewards/margins": 6.8125, + "rewards/rejected": -6.125, + "step": 443 + }, + { + "epoch": 0.5447852760736196, + "grad_norm": 35.692369145811284, + "learning_rate": 1.2402892561983472e-06, + "logits/chosen": -0.06201171875, + "logits/rejected": -0.298828125, + "logps/chosen": -454.0, + "logps/rejected": -382.0, + "loss": 0.3655, + "rewards/accuracies": 0.859375, + "rewards/chosen": 1.2890625, + "rewards/margins": 8.0, + "rewards/rejected": -6.71875, + "step": 444 + }, + { + "epoch": 0.5460122699386503, + "grad_norm": 32.62912831625137, + "learning_rate": 1.2396694214876033e-06, + "logits/chosen": -0.016357421875, + "logits/rejected": -0.208984375, + "logps/chosen": -358.0, + "logps/rejected": -314.0, + "loss": 0.4199, + "rewards/accuracies": 0.796875, + "rewards/chosen": 0.66796875, + "rewards/margins": 6.21875, + "rewards/rejected": -5.5625, + "step": 445 + }, + { + "epoch": 0.5472392638036809, + "grad_norm": 25.78605234898617, + "learning_rate": 1.2390495867768597e-06, + "logits/chosen": -0.040283203125, + "logits/rejected": -0.203125, + "logps/chosen": -438.0, + "logps/rejected": -370.0, + "loss": 0.2432, + "rewards/accuracies": 0.890625, + "rewards/chosen": 0.81640625, + "rewards/margins": 6.46875, + "rewards/rejected": -5.65625, + "step": 446 + }, + { + "epoch": 0.5484662576687117, + "grad_norm": 28.526549681854004, + "learning_rate": 1.2384297520661158e-06, + "logits/chosen": -0.0341796875, + "logits/rejected": -0.23828125, + "logps/chosen": -414.0, + "logps/rejected": -360.0, + "loss": 0.3087, + "rewards/accuracies": 0.8671875, + "rewards/chosen": 1.359375, + "rewards/margins": 8.1875, + "rewards/rejected": -6.84375, + "step": 447 + }, + { + "epoch": 0.5496932515337424, + "grad_norm": 33.26832004918433, + "learning_rate": 1.237809917355372e-06, + "logits/chosen": 0.042724609375, + "logits/rejected": -0.10009765625, + "logps/chosen": -346.0, + "logps/rejected": -314.0, + "loss": 0.3362, + "rewards/accuracies": 0.84375, + "rewards/chosen": 0.1796875, + "rewards/margins": 6.03125, + "rewards/rejected": -5.84375, + "step": 448 + }, + { + "epoch": 0.550920245398773, + "grad_norm": 29.824788916298026, + "learning_rate": 1.237190082644628e-06, + "logits/chosen": -0.0341796875, + "logits/rejected": -0.1650390625, + "logps/chosen": -400.0, + "logps/rejected": -346.0, + "loss": 0.3203, + "rewards/accuracies": 0.8515625, + "rewards/chosen": 1.234375, + "rewards/margins": 7.34375, + "rewards/rejected": -6.09375, + "step": 449 + }, + { + "epoch": 0.5521472392638037, + "grad_norm": 29.646416826102993, + "learning_rate": 1.2365702479338842e-06, + "logits/chosen": -0.1103515625, + "logits/rejected": -0.1494140625, + "logps/chosen": -350.0, + "logps/rejected": -302.0, + "loss": 0.2642, + "rewards/accuracies": 0.8359375, + "rewards/chosen": 0.66015625, + "rewards/margins": 6.78125, + "rewards/rejected": -6.125, + "step": 450 + }, + { + "epoch": 0.5533742331288344, + "grad_norm": 31.583737211583365, + "learning_rate": 1.2359504132231406e-06, + "logits/chosen": 0.03173828125, + "logits/rejected": -0.1396484375, + "logps/chosen": -348.0, + "logps/rejected": -294.0, + "loss": 0.3883, + "rewards/accuracies": 0.84375, + "rewards/chosen": 0.54296875, + "rewards/margins": 6.40625, + "rewards/rejected": -5.875, + "step": 451 + }, + { + "epoch": 0.554601226993865, + "grad_norm": 27.43380959875228, + "learning_rate": 1.2353305785123967e-06, + "logits/chosen": -0.0439453125, + "logits/rejected": -0.28125, + "logps/chosen": -460.0, + "logps/rejected": -342.0, + "loss": 0.2556, + "rewards/accuracies": 0.859375, + "rewards/chosen": 1.9453125, + "rewards/margins": 7.53125, + "rewards/rejected": -5.59375, + "step": 452 + }, + { + "epoch": 0.5558282208588957, + "grad_norm": 24.613642812642805, + "learning_rate": 1.2347107438016529e-06, + "logits/chosen": -0.130859375, + "logits/rejected": -0.33984375, + "logps/chosen": -440.0, + "logps/rejected": -378.0, + "loss": 0.2702, + "rewards/accuracies": 0.90625, + "rewards/chosen": 2.125, + "rewards/margins": 8.1875, + "rewards/rejected": -6.0625, + "step": 453 + }, + { + "epoch": 0.5570552147239264, + "grad_norm": 31.985648649098213, + "learning_rate": 1.2340909090909092e-06, + "logits/chosen": 0.046875, + "logits/rejected": -0.0206298828125, + "logps/chosen": -346.0, + "logps/rejected": -320.0, + "loss": 0.3305, + "rewards/accuracies": 0.8359375, + "rewards/chosen": 0.0859375, + "rewards/margins": 6.0625, + "rewards/rejected": -6.0, + "step": 454 + }, + { + "epoch": 0.558282208588957, + "grad_norm": 26.277923249806037, + "learning_rate": 1.2334710743801654e-06, + "logits/chosen": -0.0771484375, + "logits/rejected": -0.201171875, + "logps/chosen": -372.0, + "logps/rejected": -362.0, + "loss": 0.2299, + "rewards/accuracies": 0.8828125, + "rewards/chosen": 1.5, + "rewards/margins": 7.59375, + "rewards/rejected": -6.09375, + "step": 455 + }, + { + "epoch": 0.5595092024539877, + "grad_norm": 35.777155080443045, + "learning_rate": 1.2328512396694215e-06, + "logits/chosen": 0.0419921875, + "logits/rejected": -0.0576171875, + "logps/chosen": -318.0, + "logps/rejected": -314.0, + "loss": 0.4054, + "rewards/accuracies": 0.8125, + "rewards/chosen": 0.412109375, + "rewards/margins": 6.8125, + "rewards/rejected": -6.375, + "step": 456 + }, + { + "epoch": 0.5607361963190184, + "grad_norm": 33.41411882209408, + "learning_rate": 1.2322314049586776e-06, + "logits/chosen": 0.0126953125, + "logits/rejected": -0.1142578125, + "logps/chosen": -360.0, + "logps/rejected": -330.0, + "loss": 0.3362, + "rewards/accuracies": 0.8359375, + "rewards/chosen": 0.63671875, + "rewards/margins": 6.21875, + "rewards/rejected": -5.59375, + "step": 457 + }, + { + "epoch": 0.561963190184049, + "grad_norm": 27.510163645115238, + "learning_rate": 1.2316115702479338e-06, + "logits/chosen": -0.0264892578125, + "logits/rejected": -0.1845703125, + "logps/chosen": -376.0, + "logps/rejected": -302.0, + "loss": 0.2733, + "rewards/accuracies": 0.8828125, + "rewards/chosen": 1.109375, + "rewards/margins": 7.40625, + "rewards/rejected": -6.3125, + "step": 458 + }, + { + "epoch": 0.5631901840490797, + "grad_norm": 32.20829218573613, + "learning_rate": 1.2309917355371901e-06, + "logits/chosen": -0.0172119140625, + "logits/rejected": -0.1796875, + "logps/chosen": -396.0, + "logps/rejected": -356.0, + "loss": 0.3698, + "rewards/accuracies": 0.8515625, + "rewards/chosen": 1.171875, + "rewards/margins": 7.375, + "rewards/rejected": -6.1875, + "step": 459 + }, + { + "epoch": 0.5644171779141104, + "grad_norm": 28.463767397560563, + "learning_rate": 1.2303719008264463e-06, + "logits/chosen": -0.0159912109375, + "logits/rejected": -0.11572265625, + "logps/chosen": -354.0, + "logps/rejected": -342.0, + "loss": 0.2835, + "rewards/accuracies": 0.859375, + "rewards/chosen": 0.146484375, + "rewards/margins": 7.46875, + "rewards/rejected": -7.3125, + "step": 460 + }, + { + "epoch": 0.5656441717791411, + "grad_norm": 30.885477551125014, + "learning_rate": 1.2297520661157024e-06, + "logits/chosen": -0.2158203125, + "logits/rejected": -0.330078125, + "logps/chosen": -380.0, + "logps/rejected": -352.0, + "loss": 0.2845, + "rewards/accuracies": 0.8671875, + "rewards/chosen": 0.1884765625, + "rewards/margins": 6.5625, + "rewards/rejected": -6.375, + "step": 461 + }, + { + "epoch": 0.5668711656441717, + "grad_norm": 44.21195903099332, + "learning_rate": 1.2291322314049588e-06, + "logits/chosen": 0.047607421875, + "logits/rejected": -0.1552734375, + "logps/chosen": -370.0, + "logps/rejected": -310.0, + "loss": 0.4151, + "rewards/accuracies": 0.8203125, + "rewards/chosen": -0.1337890625, + "rewards/margins": 5.875, + "rewards/rejected": -6.03125, + "step": 462 + }, + { + "epoch": 0.5680981595092025, + "grad_norm": 33.208183287178784, + "learning_rate": 1.228512396694215e-06, + "logits/chosen": -0.0849609375, + "logits/rejected": -0.18359375, + "logps/chosen": -370.0, + "logps/rejected": -372.0, + "loss": 0.3252, + "rewards/accuracies": 0.8203125, + "rewards/chosen": 0.4140625, + "rewards/margins": 7.09375, + "rewards/rejected": -6.6875, + "step": 463 + }, + { + "epoch": 0.5693251533742332, + "grad_norm": 37.13310236043479, + "learning_rate": 1.2278925619834713e-06, + "logits/chosen": 0.05224609375, + "logits/rejected": -0.119140625, + "logps/chosen": -398.0, + "logps/rejected": -334.0, + "loss": 0.3492, + "rewards/accuracies": 0.859375, + "rewards/chosen": 0.3359375, + "rewards/margins": 6.90625, + "rewards/rejected": -6.59375, + "step": 464 + }, + { + "epoch": 0.5705521472392638, + "grad_norm": 21.80723319666503, + "learning_rate": 1.2272727272727274e-06, + "logits/chosen": -0.06396484375, + "logits/rejected": -0.302734375, + "logps/chosen": -404.0, + "logps/rejected": -362.0, + "loss": 0.1834, + "rewards/accuracies": 0.9140625, + "rewards/chosen": 0.6171875, + "rewards/margins": 8.25, + "rewards/rejected": -7.625, + "step": 465 + }, + { + "epoch": 0.5717791411042945, + "grad_norm": 32.116293843093004, + "learning_rate": 1.2266528925619833e-06, + "logits/chosen": -0.016845703125, + "logits/rejected": -0.1416015625, + "logps/chosen": -424.0, + "logps/rejected": -356.0, + "loss": 0.3118, + "rewards/accuracies": 0.8203125, + "rewards/chosen": 0.0458984375, + "rewards/margins": 7.5, + "rewards/rejected": -7.46875, + "step": 466 + }, + { + "epoch": 0.5730061349693252, + "grad_norm": 37.63110345633159, + "learning_rate": 1.2260330578512397e-06, + "logits/chosen": -0.0279541015625, + "logits/rejected": -0.2373046875, + "logps/chosen": -388.0, + "logps/rejected": -364.0, + "loss": 0.3871, + "rewards/accuracies": 0.828125, + "rewards/chosen": 0.8046875, + "rewards/margins": 7.21875, + "rewards/rejected": -6.40625, + "step": 467 + }, + { + "epoch": 0.5742331288343558, + "grad_norm": 33.08460200253717, + "learning_rate": 1.2254132231404958e-06, + "logits/chosen": -0.0419921875, + "logits/rejected": -0.1806640625, + "logps/chosen": -372.0, + "logps/rejected": -358.0, + "loss": 0.3099, + "rewards/accuracies": 0.8671875, + "rewards/chosen": 0.5234375, + "rewards/margins": 7.625, + "rewards/rejected": -7.09375, + "step": 468 + }, + { + "epoch": 0.5754601226993865, + "grad_norm": 22.648898471514507, + "learning_rate": 1.2247933884297522e-06, + "logits/chosen": 0.0230712890625, + "logits/rejected": -0.15234375, + "logps/chosen": -390.0, + "logps/rejected": -320.0, + "loss": 0.2207, + "rewards/accuracies": 0.90625, + "rewards/chosen": 0.796875, + "rewards/margins": 7.84375, + "rewards/rejected": -7.0625, + "step": 469 + }, + { + "epoch": 0.5766871165644172, + "grad_norm": 33.108021332496286, + "learning_rate": 1.2241735537190083e-06, + "logits/chosen": -0.010498046875, + "logits/rejected": -0.07373046875, + "logps/chosen": -356.0, + "logps/rejected": -338.0, + "loss": 0.367, + "rewards/accuracies": 0.828125, + "rewards/chosen": 0.69921875, + "rewards/margins": 7.5, + "rewards/rejected": -6.8125, + "step": 470 + }, + { + "epoch": 0.5779141104294478, + "grad_norm": 30.281427499200404, + "learning_rate": 1.2235537190082645e-06, + "logits/chosen": 0.1171875, + "logits/rejected": -0.0458984375, + "logps/chosen": -408.0, + "logps/rejected": -350.0, + "loss": 0.3469, + "rewards/accuracies": 0.8125, + "rewards/chosen": 0.328125, + "rewards/margins": 6.90625, + "rewards/rejected": -6.5625, + "step": 471 + }, + { + "epoch": 0.5791411042944785, + "grad_norm": 30.766178072522326, + "learning_rate": 1.2229338842975208e-06, + "logits/chosen": 0.0654296875, + "logits/rejected": -0.15234375, + "logps/chosen": -414.0, + "logps/rejected": -336.0, + "loss": 0.341, + "rewards/accuracies": 0.84375, + "rewards/chosen": 0.400390625, + "rewards/margins": 7.125, + "rewards/rejected": -6.71875, + "step": 472 + }, + { + "epoch": 0.5803680981595092, + "grad_norm": 35.82872232913366, + "learning_rate": 1.222314049586777e-06, + "logits/chosen": -0.10986328125, + "logits/rejected": -0.2138671875, + "logps/chosen": -406.0, + "logps/rejected": -394.0, + "loss": 0.3929, + "rewards/accuracies": 0.796875, + "rewards/chosen": 0.267578125, + "rewards/margins": 6.78125, + "rewards/rejected": -6.5, + "step": 473 + }, + { + "epoch": 0.5815950920245399, + "grad_norm": 38.235476934077774, + "learning_rate": 1.221694214876033e-06, + "logits/chosen": -0.06591796875, + "logits/rejected": -0.271484375, + "logps/chosen": -440.0, + "logps/rejected": -362.0, + "loss": 0.3842, + "rewards/accuracies": 0.8359375, + "rewards/chosen": 0.0673828125, + "rewards/margins": 6.875, + "rewards/rejected": -6.8125, + "step": 474 + }, + { + "epoch": 0.5828220858895705, + "grad_norm": 28.310340512331347, + "learning_rate": 1.2210743801652892e-06, + "logits/chosen": -0.06689453125, + "logits/rejected": -0.2734375, + "logps/chosen": -434.0, + "logps/rejected": -362.0, + "loss": 0.2596, + "rewards/accuracies": 0.8828125, + "rewards/chosen": 1.140625, + "rewards/margins": 8.4375, + "rewards/rejected": -7.3125, + "step": 475 + }, + { + "epoch": 0.5840490797546012, + "grad_norm": 31.21556050331037, + "learning_rate": 1.2204545454545454e-06, + "logits/chosen": 0.00927734375, + "logits/rejected": -0.1806640625, + "logps/chosen": -420.0, + "logps/rejected": -360.0, + "loss": 0.3346, + "rewards/accuracies": 0.8359375, + "rewards/chosen": -1.078125, + "rewards/margins": 6.53125, + "rewards/rejected": -7.625, + "step": 476 + }, + { + "epoch": 0.5852760736196319, + "grad_norm": 31.773096981894017, + "learning_rate": 1.2198347107438017e-06, + "logits/chosen": -0.11474609375, + "logits/rejected": -0.255859375, + "logps/chosen": -432.0, + "logps/rejected": -372.0, + "loss": 0.3232, + "rewards/accuracies": 0.890625, + "rewards/chosen": 0.1484375, + "rewards/margins": 7.8125, + "rewards/rejected": -7.65625, + "step": 477 + }, + { + "epoch": 0.5865030674846625, + "grad_norm": 30.507836762025494, + "learning_rate": 1.2192148760330579e-06, + "logits/chosen": -0.04150390625, + "logits/rejected": -0.2080078125, + "logps/chosen": -430.0, + "logps/rejected": -368.0, + "loss": 0.3843, + "rewards/accuracies": 0.8671875, + "rewards/chosen": -0.024658203125, + "rewards/margins": 7.625, + "rewards/rejected": -7.65625, + "step": 478 + }, + { + "epoch": 0.5877300613496933, + "grad_norm": 32.261887607989564, + "learning_rate": 1.218595041322314e-06, + "logits/chosen": -0.03759765625, + "logits/rejected": -0.298828125, + "logps/chosen": -370.0, + "logps/rejected": -352.0, + "loss": 0.322, + "rewards/accuracies": 0.84375, + "rewards/chosen": -1.1796875, + "rewards/margins": 6.625, + "rewards/rejected": -7.78125, + "step": 479 + }, + { + "epoch": 0.588957055214724, + "grad_norm": 28.188949742159306, + "learning_rate": 1.2179752066115704e-06, + "logits/chosen": -0.0791015625, + "logits/rejected": -0.208984375, + "logps/chosen": -372.0, + "logps/rejected": -348.0, + "loss": 0.2558, + "rewards/accuracies": 0.8828125, + "rewards/chosen": -0.8125, + "rewards/margins": 7.21875, + "rewards/rejected": -8.0, + "step": 480 + }, + { + "epoch": 0.5901840490797546, + "grad_norm": 37.17733184183342, + "learning_rate": 1.2173553719008265e-06, + "logits/chosen": -0.07568359375, + "logits/rejected": -0.1826171875, + "logps/chosen": -438.0, + "logps/rejected": -398.0, + "loss": 0.4973, + "rewards/accuracies": 0.796875, + "rewards/chosen": -0.578125, + "rewards/margins": 6.71875, + "rewards/rejected": -7.3125, + "step": 481 + }, + { + "epoch": 0.5914110429447853, + "grad_norm": 34.97683920834337, + "learning_rate": 1.2167355371900826e-06, + "logits/chosen": -0.07373046875, + "logits/rejected": -0.244140625, + "logps/chosen": -424.0, + "logps/rejected": -352.0, + "loss": 0.3731, + "rewards/accuracies": 0.8125, + "rewards/chosen": -0.255859375, + "rewards/margins": 6.875, + "rewards/rejected": -7.125, + "step": 482 + }, + { + "epoch": 0.592638036809816, + "grad_norm": 32.698615963886375, + "learning_rate": 1.216115702479339e-06, + "logits/chosen": 0.07763671875, + "logits/rejected": -0.0556640625, + "logps/chosen": -358.0, + "logps/rejected": -330.0, + "loss": 0.3304, + "rewards/accuracies": 0.8046875, + "rewards/chosen": -0.55859375, + "rewards/margins": 6.78125, + "rewards/rejected": -7.34375, + "step": 483 + }, + { + "epoch": 0.5938650306748466, + "grad_norm": 31.73235234166197, + "learning_rate": 1.215495867768595e-06, + "logits/chosen": 0.00390625, + "logits/rejected": -0.142578125, + "logps/chosen": -380.0, + "logps/rejected": -340.0, + "loss": 0.3937, + "rewards/accuracies": 0.828125, + "rewards/chosen": 0.1611328125, + "rewards/margins": 6.46875, + "rewards/rejected": -6.3125, + "step": 484 + }, + { + "epoch": 0.5950920245398773, + "grad_norm": 28.55956265398308, + "learning_rate": 1.2148760330578513e-06, + "logits/chosen": 0.023681640625, + "logits/rejected": -0.12109375, + "logps/chosen": -370.0, + "logps/rejected": -324.0, + "loss": 0.3173, + "rewards/accuracies": 0.859375, + "rewards/chosen": -0.36328125, + "rewards/margins": 6.9375, + "rewards/rejected": -7.3125, + "step": 485 + }, + { + "epoch": 0.596319018404908, + "grad_norm": 33.81986931883939, + "learning_rate": 1.2142561983471074e-06, + "logits/chosen": -0.0050048828125, + "logits/rejected": -0.14453125, + "logps/chosen": -414.0, + "logps/rejected": -352.0, + "loss": 0.3778, + "rewards/accuracies": 0.875, + "rewards/chosen": 0.455078125, + "rewards/margins": 7.0, + "rewards/rejected": -6.53125, + "step": 486 + }, + { + "epoch": 0.5975460122699386, + "grad_norm": 30.028596310368773, + "learning_rate": 1.2136363636363636e-06, + "logits/chosen": -0.08251953125, + "logits/rejected": -0.27734375, + "logps/chosen": -410.0, + "logps/rejected": -366.0, + "loss": 0.2781, + "rewards/accuracies": 0.875, + "rewards/chosen": 0.33984375, + "rewards/margins": 7.375, + "rewards/rejected": -7.03125, + "step": 487 + }, + { + "epoch": 0.5987730061349693, + "grad_norm": 35.71061050092003, + "learning_rate": 1.21301652892562e-06, + "logits/chosen": -0.0693359375, + "logits/rejected": -0.138671875, + "logps/chosen": -424.0, + "logps/rejected": -370.0, + "loss": 0.3544, + "rewards/accuracies": 0.84375, + "rewards/chosen": 0.5078125, + "rewards/margins": 6.96875, + "rewards/rejected": -6.46875, + "step": 488 + }, + { + "epoch": 0.6, + "grad_norm": 26.138379124419743, + "learning_rate": 1.212396694214876e-06, + "logits/chosen": 0.01324462890625, + "logits/rejected": -0.224609375, + "logps/chosen": -384.0, + "logps/rejected": -340.0, + "loss": 0.3008, + "rewards/accuracies": 0.859375, + "rewards/chosen": 0.92578125, + "rewards/margins": 7.875, + "rewards/rejected": -6.9375, + "step": 489 + }, + { + "epoch": 0.6012269938650306, + "grad_norm": 38.68817736099095, + "learning_rate": 1.2117768595041322e-06, + "logits/chosen": 0.055419921875, + "logits/rejected": -0.1689453125, + "logps/chosen": -432.0, + "logps/rejected": -364.0, + "loss": 0.404, + "rewards/accuracies": 0.8125, + "rewards/chosen": 0.263671875, + "rewards/margins": 6.84375, + "rewards/rejected": -6.5625, + "step": 490 + }, + { + "epoch": 0.6024539877300613, + "grad_norm": 30.239394074669093, + "learning_rate": 1.2111570247933886e-06, + "logits/chosen": 0.07666015625, + "logits/rejected": -0.134765625, + "logps/chosen": -372.0, + "logps/rejected": -328.0, + "loss": 0.3068, + "rewards/accuracies": 0.8828125, + "rewards/chosen": 0.443359375, + "rewards/margins": 6.96875, + "rewards/rejected": -6.53125, + "step": 491 + }, + { + "epoch": 0.603680981595092, + "grad_norm": 33.80787533439265, + "learning_rate": 1.2105371900826447e-06, + "logits/chosen": -0.013427734375, + "logits/rejected": -0.1533203125, + "logps/chosen": -398.0, + "logps/rejected": -390.0, + "loss": 0.3711, + "rewards/accuracies": 0.8515625, + "rewards/chosen": 0.5625, + "rewards/margins": 6.6875, + "rewards/rejected": -6.125, + "step": 492 + }, + { + "epoch": 0.6049079754601226, + "grad_norm": 41.137878058973136, + "learning_rate": 1.2099173553719008e-06, + "logits/chosen": 0.14453125, + "logits/rejected": 0.043701171875, + "logps/chosen": -392.0, + "logps/rejected": -340.0, + "loss": 0.5049, + "rewards/accuracies": 0.796875, + "rewards/chosen": 0.19921875, + "rewards/margins": 5.65625, + "rewards/rejected": -5.4375, + "step": 493 + }, + { + "epoch": 0.6061349693251534, + "grad_norm": 28.550910744803765, + "learning_rate": 1.209297520661157e-06, + "logits/chosen": 0.044921875, + "logits/rejected": -0.1435546875, + "logps/chosen": -428.0, + "logps/rejected": -340.0, + "loss": 0.3041, + "rewards/accuracies": 0.875, + "rewards/chosen": 1.1796875, + "rewards/margins": 7.21875, + "rewards/rejected": -6.0625, + "step": 494 + }, + { + "epoch": 0.6073619631901841, + "grad_norm": 32.18887165372733, + "learning_rate": 1.2086776859504131e-06, + "logits/chosen": 0.031494140625, + "logits/rejected": -0.1064453125, + "logps/chosen": -402.0, + "logps/rejected": -350.0, + "loss": 0.3067, + "rewards/accuracies": 0.859375, + "rewards/chosen": 1.0859375, + "rewards/margins": 7.25, + "rewards/rejected": -6.15625, + "step": 495 + }, + { + "epoch": 0.6085889570552148, + "grad_norm": 34.7968597037722, + "learning_rate": 1.2080578512396695e-06, + "logits/chosen": -0.0025634765625, + "logits/rejected": -0.21484375, + "logps/chosen": -382.0, + "logps/rejected": -322.0, + "loss": 0.3605, + "rewards/accuracies": 0.859375, + "rewards/chosen": 0.1708984375, + "rewards/margins": 6.25, + "rewards/rejected": -6.09375, + "step": 496 + }, + { + "epoch": 0.6098159509202454, + "grad_norm": 36.914820937247434, + "learning_rate": 1.2074380165289256e-06, + "logits/chosen": 0.0361328125, + "logits/rejected": -0.125, + "logps/chosen": -404.0, + "logps/rejected": -366.0, + "loss": 0.4718, + "rewards/accuracies": 0.78125, + "rewards/chosen": 0.1337890625, + "rewards/margins": 6.5, + "rewards/rejected": -6.375, + "step": 497 + }, + { + "epoch": 0.6110429447852761, + "grad_norm": 27.898760612078714, + "learning_rate": 1.206818181818182e-06, + "logits/chosen": -0.10302734375, + "logits/rejected": -0.265625, + "logps/chosen": -400.0, + "logps/rejected": -366.0, + "loss": 0.2974, + "rewards/accuracies": 0.8359375, + "rewards/chosen": 0.9765625, + "rewards/margins": 7.46875, + "rewards/rejected": -6.5, + "step": 498 + }, + { + "epoch": 0.6122699386503068, + "grad_norm": 33.514362873417944, + "learning_rate": 1.2061983471074381e-06, + "logits/chosen": 0.041015625, + "logits/rejected": -0.177734375, + "logps/chosen": -426.0, + "logps/rejected": -350.0, + "loss": 0.3147, + "rewards/accuracies": 0.8671875, + "rewards/chosen": 0.984375, + "rewards/margins": 6.96875, + "rewards/rejected": -6.0, + "step": 499 + }, + { + "epoch": 0.6134969325153374, + "grad_norm": 30.177881260758163, + "learning_rate": 1.2055785123966943e-06, + "logits/chosen": 0.0224609375, + "logits/rejected": -0.09375, + "logps/chosen": -418.0, + "logps/rejected": -338.0, + "loss": 0.3575, + "rewards/accuracies": 0.796875, + "rewards/chosen": 1.0390625, + "rewards/margins": 7.3125, + "rewards/rejected": -6.28125, + "step": 500 + }, + { + "epoch": 0.6147239263803681, + "grad_norm": 34.255808614960706, + "learning_rate": 1.2049586776859504e-06, + "logits/chosen": -0.0390625, + "logits/rejected": -0.171875, + "logps/chosen": -410.0, + "logps/rejected": -352.0, + "loss": 0.4018, + "rewards/accuracies": 0.8359375, + "rewards/chosen": 1.234375, + "rewards/margins": 7.59375, + "rewards/rejected": -6.34375, + "step": 501 + }, + { + "epoch": 0.6159509202453988, + "grad_norm": 28.29113723996875, + "learning_rate": 1.2043388429752065e-06, + "logits/chosen": -0.01165771484375, + "logits/rejected": -0.28125, + "logps/chosen": -386.0, + "logps/rejected": -340.0, + "loss": 0.3467, + "rewards/accuracies": 0.8828125, + "rewards/chosen": 0.6796875, + "rewards/margins": 7.71875, + "rewards/rejected": -7.03125, + "step": 502 + }, + { + "epoch": 0.6171779141104294, + "grad_norm": 29.752413229458206, + "learning_rate": 1.2037190082644629e-06, + "logits/chosen": 0.1904296875, + "logits/rejected": -0.0013275146484375, + "logps/chosen": -334.0, + "logps/rejected": -324.0, + "loss": 0.2996, + "rewards/accuracies": 0.8046875, + "rewards/chosen": 0.28515625, + "rewards/margins": 6.6875, + "rewards/rejected": -6.40625, + "step": 503 + }, + { + "epoch": 0.6184049079754601, + "grad_norm": 24.506669229979327, + "learning_rate": 1.203099173553719e-06, + "logits/chosen": -0.06201171875, + "logits/rejected": -0.234375, + "logps/chosen": -388.0, + "logps/rejected": -340.0, + "loss": 0.2718, + "rewards/accuracies": 0.8984375, + "rewards/chosen": -0.1025390625, + "rewards/margins": 6.21875, + "rewards/rejected": -6.3125, + "step": 504 + }, + { + "epoch": 0.6196319018404908, + "grad_norm": 38.20086418972538, + "learning_rate": 1.2024793388429752e-06, + "logits/chosen": -0.07666015625, + "logits/rejected": -0.298828125, + "logps/chosen": -434.0, + "logps/rejected": -406.0, + "loss": 0.4423, + "rewards/accuracies": 0.828125, + "rewards/chosen": -0.609375, + "rewards/margins": 6.78125, + "rewards/rejected": -7.375, + "step": 505 + }, + { + "epoch": 0.6208588957055214, + "grad_norm": 67.15816433272404, + "learning_rate": 1.2018595041322315e-06, + "logits/chosen": 0.06982421875, + "logits/rejected": -0.2373046875, + "logps/chosen": -428.0, + "logps/rejected": -340.0, + "loss": 0.3369, + "rewards/accuracies": 0.859375, + "rewards/chosen": -0.0081787109375, + "rewards/margins": 7.03125, + "rewards/rejected": -7.03125, + "step": 506 + }, + { + "epoch": 0.6220858895705521, + "grad_norm": 25.2127178538785, + "learning_rate": 1.2012396694214877e-06, + "logits/chosen": -0.04296875, + "logits/rejected": -0.150390625, + "logps/chosen": -394.0, + "logps/rejected": -318.0, + "loss": 0.1989, + "rewards/accuracies": 0.8984375, + "rewards/chosen": 0.11474609375, + "rewards/margins": 7.34375, + "rewards/rejected": -7.21875, + "step": 507 + }, + { + "epoch": 0.6233128834355828, + "grad_norm": 31.70942946416519, + "learning_rate": 1.2006198347107438e-06, + "logits/chosen": 0.14453125, + "logits/rejected": -0.033935546875, + "logps/chosen": -396.0, + "logps/rejected": -320.0, + "loss": 0.3214, + "rewards/accuracies": 0.84375, + "rewards/chosen": -0.1494140625, + "rewards/margins": 6.40625, + "rewards/rejected": -6.5625, + "step": 508 + }, + { + "epoch": 0.6245398773006134, + "grad_norm": 39.83330773441419, + "learning_rate": 1.2000000000000002e-06, + "logits/chosen": -0.0225830078125, + "logits/rejected": -0.2578125, + "logps/chosen": -474.0, + "logps/rejected": -396.0, + "loss": 0.569, + "rewards/accuracies": 0.8125, + "rewards/chosen": 1.0390625, + "rewards/margins": 8.25, + "rewards/rejected": -7.21875, + "step": 509 + }, + { + "epoch": 0.6257668711656442, + "grad_norm": 29.863730506340346, + "learning_rate": 1.199380165289256e-06, + "logits/chosen": -0.11181640625, + "logits/rejected": -0.291015625, + "logps/chosen": -368.0, + "logps/rejected": -314.0, + "loss": 0.2926, + "rewards/accuracies": 0.875, + "rewards/chosen": -0.375, + "rewards/margins": 6.3125, + "rewards/rejected": -6.6875, + "step": 510 + }, + { + "epoch": 0.6269938650306749, + "grad_norm": 35.188221718693796, + "learning_rate": 1.1987603305785124e-06, + "logits/chosen": -0.1328125, + "logits/rejected": -0.259765625, + "logps/chosen": -402.0, + "logps/rejected": -366.0, + "loss": 0.4071, + "rewards/accuracies": 0.8359375, + "rewards/chosen": 0.0859375, + "rewards/margins": 7.3125, + "rewards/rejected": -7.21875, + "step": 511 + }, + { + "epoch": 0.6282208588957056, + "grad_norm": 35.23659404136466, + "learning_rate": 1.1981404958677686e-06, + "logits/chosen": 0.0791015625, + "logits/rejected": -0.0654296875, + "logps/chosen": -388.0, + "logps/rejected": -398.0, + "loss": 0.4354, + "rewards/accuracies": 0.8203125, + "rewards/chosen": -0.388671875, + "rewards/margins": 5.90625, + "rewards/rejected": -6.28125, + "step": 512 + }, + { + "epoch": 0.6294478527607362, + "grad_norm": 33.94396537230905, + "learning_rate": 1.1975206611570247e-06, + "logits/chosen": 0.07958984375, + "logits/rejected": -0.0517578125, + "logps/chosen": -384.0, + "logps/rejected": -326.0, + "loss": 0.4151, + "rewards/accuracies": 0.796875, + "rewards/chosen": 0.1328125, + "rewards/margins": 6.96875, + "rewards/rejected": -6.8125, + "step": 513 + }, + { + "epoch": 0.6306748466257669, + "grad_norm": 27.33280439239978, + "learning_rate": 1.196900826446281e-06, + "logits/chosen": 0.09765625, + "logits/rejected": -0.0537109375, + "logps/chosen": -340.0, + "logps/rejected": -302.0, + "loss": 0.2677, + "rewards/accuracies": 0.8828125, + "rewards/chosen": -1.015625, + "rewards/margins": 6.5, + "rewards/rejected": -7.53125, + "step": 514 + }, + { + "epoch": 0.6319018404907976, + "grad_norm": 32.16724360242484, + "learning_rate": 1.1962809917355372e-06, + "logits/chosen": 0.07275390625, + "logits/rejected": -0.07763671875, + "logps/chosen": -358.0, + "logps/rejected": -362.0, + "loss": 0.3029, + "rewards/accuracies": 0.8515625, + "rewards/chosen": -1.0703125, + "rewards/margins": 6.5, + "rewards/rejected": -7.5625, + "step": 515 + }, + { + "epoch": 0.6331288343558282, + "grad_norm": 26.23429258203521, + "learning_rate": 1.1956611570247934e-06, + "logits/chosen": 0.0279541015625, + "logits/rejected": -0.07861328125, + "logps/chosen": -378.0, + "logps/rejected": -346.0, + "loss": 0.2424, + "rewards/accuracies": 0.8828125, + "rewards/chosen": -0.8828125, + "rewards/margins": 6.65625, + "rewards/rejected": -7.53125, + "step": 516 + }, + { + "epoch": 0.6343558282208589, + "grad_norm": 37.00771401511282, + "learning_rate": 1.1950413223140497e-06, + "logits/chosen": 0.0703125, + "logits/rejected": -0.08642578125, + "logps/chosen": -400.0, + "logps/rejected": -352.0, + "loss": 0.4735, + "rewards/accuracies": 0.78125, + "rewards/chosen": -0.48828125, + "rewards/margins": 6.28125, + "rewards/rejected": -6.78125, + "step": 517 + }, + { + "epoch": 0.6355828220858896, + "grad_norm": 28.05157133110184, + "learning_rate": 1.1944214876033059e-06, + "logits/chosen": -0.04541015625, + "logits/rejected": -0.150390625, + "logps/chosen": -424.0, + "logps/rejected": -360.0, + "loss": 0.2475, + "rewards/accuracies": 0.8671875, + "rewards/chosen": 0.396484375, + "rewards/margins": 7.625, + "rewards/rejected": -7.21875, + "step": 518 + }, + { + "epoch": 0.6368098159509202, + "grad_norm": 38.286039724245256, + "learning_rate": 1.193801652892562e-06, + "logits/chosen": -0.026123046875, + "logits/rejected": -0.1630859375, + "logps/chosen": -396.0, + "logps/rejected": -354.0, + "loss": 0.3562, + "rewards/accuracies": 0.828125, + "rewards/chosen": -0.60546875, + "rewards/margins": 6.9375, + "rewards/rejected": -7.53125, + "step": 519 + }, + { + "epoch": 0.6380368098159509, + "grad_norm": 34.57537616564543, + "learning_rate": 1.1931818181818181e-06, + "logits/chosen": -0.0517578125, + "logits/rejected": -0.279296875, + "logps/chosen": -380.0, + "logps/rejected": -350.0, + "loss": 0.37, + "rewards/accuracies": 0.84375, + "rewards/chosen": -0.271484375, + "rewards/margins": 7.71875, + "rewards/rejected": -8.0, + "step": 520 + }, + { + "epoch": 0.6392638036809816, + "grad_norm": 29.132838683315903, + "learning_rate": 1.1925619834710743e-06, + "logits/chosen": 0.158203125, + "logits/rejected": 0.01312255859375, + "logps/chosen": -428.0, + "logps/rejected": -378.0, + "loss": 0.2975, + "rewards/accuracies": 0.859375, + "rewards/chosen": 0.01263427734375, + "rewards/margins": 7.8125, + "rewards/rejected": -7.8125, + "step": 521 + }, + { + "epoch": 0.6404907975460122, + "grad_norm": 30.38859952385869, + "learning_rate": 1.1919421487603306e-06, + "logits/chosen": -0.0546875, + "logits/rejected": -0.2060546875, + "logps/chosen": -438.0, + "logps/rejected": -374.0, + "loss": 0.3466, + "rewards/accuracies": 0.8359375, + "rewards/chosen": -0.439453125, + "rewards/margins": 8.1875, + "rewards/rejected": -8.625, + "step": 522 + }, + { + "epoch": 0.6417177914110429, + "grad_norm": 36.440630501558566, + "learning_rate": 1.1913223140495868e-06, + "logits/chosen": 0.0238037109375, + "logits/rejected": -0.12890625, + "logps/chosen": -380.0, + "logps/rejected": -346.0, + "loss": 0.4467, + "rewards/accuracies": 0.8125, + "rewards/chosen": -1.09375, + "rewards/margins": 6.3125, + "rewards/rejected": -7.40625, + "step": 523 + }, + { + "epoch": 0.6429447852760736, + "grad_norm": 39.231543678462224, + "learning_rate": 1.1907024793388431e-06, + "logits/chosen": 0.05615234375, + "logits/rejected": -0.1474609375, + "logps/chosen": -440.0, + "logps/rejected": -380.0, + "loss": 0.4136, + "rewards/accuracies": 0.859375, + "rewards/chosen": -1.1953125, + "rewards/margins": 6.40625, + "rewards/rejected": -7.59375, + "step": 524 + }, + { + "epoch": 0.6441717791411042, + "grad_norm": 47.6608831234207, + "learning_rate": 1.1900826446280993e-06, + "logits/chosen": 0.07666015625, + "logits/rejected": 0.02587890625, + "logps/chosen": -406.0, + "logps/rejected": -382.0, + "loss": 0.5676, + "rewards/accuracies": 0.7421875, + "rewards/chosen": -1.3125, + "rewards/margins": 5.53125, + "rewards/rejected": -6.84375, + "step": 525 + }, + { + "epoch": 0.645398773006135, + "grad_norm": 31.7891594700196, + "learning_rate": 1.1894628099173554e-06, + "logits/chosen": 0.0458984375, + "logits/rejected": -0.10107421875, + "logps/chosen": -402.0, + "logps/rejected": -356.0, + "loss": 0.3316, + "rewards/accuracies": 0.8203125, + "rewards/chosen": -0.6171875, + "rewards/margins": 7.21875, + "rewards/rejected": -7.84375, + "step": 526 + }, + { + "epoch": 0.6466257668711657, + "grad_norm": 32.21728261594982, + "learning_rate": 1.1888429752066118e-06, + "logits/chosen": 0.1201171875, + "logits/rejected": -0.103515625, + "logps/chosen": -438.0, + "logps/rejected": -348.0, + "loss": 0.3529, + "rewards/accuracies": 0.8359375, + "rewards/chosen": -0.7265625, + "rewards/margins": 7.03125, + "rewards/rejected": -7.75, + "step": 527 + }, + { + "epoch": 0.6478527607361964, + "grad_norm": 26.60182420517136, + "learning_rate": 1.1882231404958677e-06, + "logits/chosen": -0.0380859375, + "logits/rejected": -0.220703125, + "logps/chosen": -368.0, + "logps/rejected": -344.0, + "loss": 0.2928, + "rewards/accuracies": 0.859375, + "rewards/chosen": -0.671875, + "rewards/margins": 7.375, + "rewards/rejected": -8.0625, + "step": 528 + }, + { + "epoch": 0.649079754601227, + "grad_norm": 32.109755070015275, + "learning_rate": 1.187603305785124e-06, + "logits/chosen": 0.030029296875, + "logits/rejected": -0.14453125, + "logps/chosen": -352.0, + "logps/rejected": -352.0, + "loss": 0.418, + "rewards/accuracies": 0.8203125, + "rewards/chosen": -0.3984375, + "rewards/margins": 6.9375, + "rewards/rejected": -7.34375, + "step": 529 + }, + { + "epoch": 0.6503067484662577, + "grad_norm": 30.441331665820965, + "learning_rate": 1.1869834710743802e-06, + "logits/chosen": 0.07373046875, + "logits/rejected": -0.0517578125, + "logps/chosen": -370.0, + "logps/rejected": -366.0, + "loss": 0.2962, + "rewards/accuracies": 0.8515625, + "rewards/chosen": -0.1103515625, + "rewards/margins": 7.5, + "rewards/rejected": -7.59375, + "step": 530 + }, + { + "epoch": 0.6515337423312884, + "grad_norm": 25.97427159310545, + "learning_rate": 1.1863636363636363e-06, + "logits/chosen": -0.0005950927734375, + "logits/rejected": -0.1767578125, + "logps/chosen": -394.0, + "logps/rejected": -338.0, + "loss": 0.3013, + "rewards/accuracies": 0.8671875, + "rewards/chosen": -0.2001953125, + "rewards/margins": 7.0625, + "rewards/rejected": -7.25, + "step": 531 + }, + { + "epoch": 0.652760736196319, + "grad_norm": 29.45586726470446, + "learning_rate": 1.1857438016528927e-06, + "logits/chosen": 0.12060546875, + "logits/rejected": -0.01611328125, + "logps/chosen": -394.0, + "logps/rejected": -342.0, + "loss": 0.3256, + "rewards/accuracies": 0.8671875, + "rewards/chosen": -0.1650390625, + "rewards/margins": 6.5625, + "rewards/rejected": -6.71875, + "step": 532 + }, + { + "epoch": 0.6539877300613497, + "grad_norm": 31.510396061786174, + "learning_rate": 1.1851239669421488e-06, + "logits/chosen": -0.0634765625, + "logits/rejected": -0.265625, + "logps/chosen": -414.0, + "logps/rejected": -360.0, + "loss": 0.322, + "rewards/accuracies": 0.8359375, + "rewards/chosen": 0.71484375, + "rewards/margins": 7.78125, + "rewards/rejected": -7.09375, + "step": 533 + }, + { + "epoch": 0.6552147239263804, + "grad_norm": 29.028681159770606, + "learning_rate": 1.184504132231405e-06, + "logits/chosen": 0.0908203125, + "logits/rejected": -0.02685546875, + "logps/chosen": -374.0, + "logps/rejected": -338.0, + "loss": 0.3243, + "rewards/accuracies": 0.8203125, + "rewards/chosen": 0.0341796875, + "rewards/margins": 6.40625, + "rewards/rejected": -6.375, + "step": 534 + }, + { + "epoch": 0.656441717791411, + "grad_norm": 33.385986566673, + "learning_rate": 1.1838842975206613e-06, + "logits/chosen": -0.022705078125, + "logits/rejected": -0.125, + "logps/chosen": -378.0, + "logps/rejected": -348.0, + "loss": 0.3502, + "rewards/accuracies": 0.84375, + "rewards/chosen": -0.7421875, + "rewards/margins": 6.40625, + "rewards/rejected": -7.125, + "step": 535 + }, + { + "epoch": 0.6576687116564417, + "grad_norm": 27.982276879832543, + "learning_rate": 1.1832644628099175e-06, + "logits/chosen": 0.0201416015625, + "logits/rejected": -0.07568359375, + "logps/chosen": -394.0, + "logps/rejected": -330.0, + "loss": 0.2519, + "rewards/accuracies": 0.890625, + "rewards/chosen": 0.6015625, + "rewards/margins": 7.40625, + "rewards/rejected": -6.78125, + "step": 536 + }, + { + "epoch": 0.6588957055214724, + "grad_norm": 40.68442319965527, + "learning_rate": 1.1826446280991736e-06, + "logits/chosen": 0.06201171875, + "logits/rejected": -0.0927734375, + "logps/chosen": -386.0, + "logps/rejected": -318.0, + "loss": 0.5685, + "rewards/accuracies": 0.7734375, + "rewards/chosen": -0.50390625, + "rewards/margins": 5.28125, + "rewards/rejected": -5.78125, + "step": 537 + }, + { + "epoch": 0.660122699386503, + "grad_norm": 35.32326255751808, + "learning_rate": 1.1820247933884297e-06, + "logits/chosen": 0.068359375, + "logits/rejected": -0.09912109375, + "logps/chosen": -402.0, + "logps/rejected": -360.0, + "loss": 0.4164, + "rewards/accuracies": 0.84375, + "rewards/chosen": 0.78515625, + "rewards/margins": 6.75, + "rewards/rejected": -5.96875, + "step": 538 + }, + { + "epoch": 0.6613496932515337, + "grad_norm": 32.525023600229474, + "learning_rate": 1.1814049586776859e-06, + "logits/chosen": 0.07666015625, + "logits/rejected": -0.04541015625, + "logps/chosen": -400.0, + "logps/rejected": -346.0, + "loss": 0.4519, + "rewards/accuracies": 0.8125, + "rewards/chosen": 0.80859375, + "rewards/margins": 5.84375, + "rewards/rejected": -5.03125, + "step": 539 + }, + { + "epoch": 0.6625766871165644, + "grad_norm": 35.156155067546834, + "learning_rate": 1.1807851239669422e-06, + "logits/chosen": 0.0986328125, + "logits/rejected": -0.06787109375, + "logps/chosen": -396.0, + "logps/rejected": -332.0, + "loss": 0.432, + "rewards/accuracies": 0.8515625, + "rewards/chosen": 0.80859375, + "rewards/margins": 6.71875, + "rewards/rejected": -5.9375, + "step": 540 + }, + { + "epoch": 0.6638036809815951, + "grad_norm": 45.099896848619814, + "learning_rate": 1.1801652892561984e-06, + "logits/chosen": 0.08447265625, + "logits/rejected": -0.0064697265625, + "logps/chosen": -390.0, + "logps/rejected": -354.0, + "loss": 0.3808, + "rewards/accuracies": 0.828125, + "rewards/chosen": 0.279296875, + "rewards/margins": 5.84375, + "rewards/rejected": -5.5625, + "step": 541 + }, + { + "epoch": 0.6650306748466258, + "grad_norm": 30.118145811004524, + "learning_rate": 1.1795454545454545e-06, + "logits/chosen": 0.031494140625, + "logits/rejected": -0.053466796875, + "logps/chosen": -356.0, + "logps/rejected": -320.0, + "loss": 0.3659, + "rewards/accuracies": 0.859375, + "rewards/chosen": 0.87109375, + "rewards/margins": 6.03125, + "rewards/rejected": -5.15625, + "step": 542 + }, + { + "epoch": 0.6662576687116565, + "grad_norm": 32.83770736751292, + "learning_rate": 1.1789256198347109e-06, + "logits/chosen": -0.02294921875, + "logits/rejected": -0.201171875, + "logps/chosen": -450.0, + "logps/rejected": -354.0, + "loss": 0.3396, + "rewards/accuracies": 0.8359375, + "rewards/chosen": 1.5078125, + "rewards/margins": 6.875, + "rewards/rejected": -5.375, + "step": 543 + }, + { + "epoch": 0.6674846625766871, + "grad_norm": 35.490916209773005, + "learning_rate": 1.178305785123967e-06, + "logits/chosen": 0.007598876953125, + "logits/rejected": -0.1318359375, + "logps/chosen": -378.0, + "logps/rejected": -318.0, + "loss": 0.434, + "rewards/accuracies": 0.8359375, + "rewards/chosen": 0.376953125, + "rewards/margins": 6.15625, + "rewards/rejected": -5.78125, + "step": 544 + }, + { + "epoch": 0.6687116564417178, + "grad_norm": 33.60719342971152, + "learning_rate": 1.1776859504132231e-06, + "logits/chosen": -0.054443359375, + "logits/rejected": -0.21875, + "logps/chosen": -458.0, + "logps/rejected": -368.0, + "loss": 0.3855, + "rewards/accuracies": 0.8125, + "rewards/chosen": 0.94140625, + "rewards/margins": 6.1875, + "rewards/rejected": -5.25, + "step": 545 + }, + { + "epoch": 0.6699386503067485, + "grad_norm": 33.972456048977826, + "learning_rate": 1.1770661157024793e-06, + "logits/chosen": 0.0693359375, + "logits/rejected": -0.028564453125, + "logps/chosen": -374.0, + "logps/rejected": -338.0, + "loss": 0.3815, + "rewards/accuracies": 0.828125, + "rewards/chosen": 0.6953125, + "rewards/margins": 6.40625, + "rewards/rejected": -5.71875, + "step": 546 + }, + { + "epoch": 0.6711656441717792, + "grad_norm": 28.303815223441166, + "learning_rate": 1.1764462809917354e-06, + "logits/chosen": 0.0133056640625, + "logits/rejected": -0.166015625, + "logps/chosen": -386.0, + "logps/rejected": -312.0, + "loss": 0.2992, + "rewards/accuracies": 0.84375, + "rewards/chosen": 0.88671875, + "rewards/margins": 6.4375, + "rewards/rejected": -5.53125, + "step": 547 + }, + { + "epoch": 0.6723926380368098, + "grad_norm": 38.31133772907115, + "learning_rate": 1.1758264462809918e-06, + "logits/chosen": -0.06494140625, + "logits/rejected": -0.265625, + "logps/chosen": -402.0, + "logps/rejected": -358.0, + "loss": 0.3532, + "rewards/accuracies": 0.8515625, + "rewards/chosen": 1.015625, + "rewards/margins": 6.125, + "rewards/rejected": -5.125, + "step": 548 + }, + { + "epoch": 0.6736196319018405, + "grad_norm": 33.292222723297044, + "learning_rate": 1.175206611570248e-06, + "logits/chosen": 0.047607421875, + "logits/rejected": -0.1572265625, + "logps/chosen": -414.0, + "logps/rejected": -344.0, + "loss": 0.3292, + "rewards/accuracies": 0.8359375, + "rewards/chosen": 0.5625, + "rewards/margins": 5.96875, + "rewards/rejected": -5.40625, + "step": 549 + }, + { + "epoch": 0.6748466257668712, + "grad_norm": 30.807870745155437, + "learning_rate": 1.174586776859504e-06, + "logits/chosen": -0.024658203125, + "logits/rejected": -0.08642578125, + "logps/chosen": -402.0, + "logps/rejected": -330.0, + "loss": 0.3926, + "rewards/accuracies": 0.8125, + "rewards/chosen": 0.12890625, + "rewards/margins": 5.71875, + "rewards/rejected": -5.5625, + "step": 550 + }, + { + "epoch": 0.6760736196319018, + "grad_norm": 28.384977078060157, + "learning_rate": 1.1739669421487604e-06, + "logits/chosen": 0.06787109375, + "logits/rejected": -0.04541015625, + "logps/chosen": -410.0, + "logps/rejected": -346.0, + "loss": 0.3704, + "rewards/accuracies": 0.859375, + "rewards/chosen": 1.2109375, + "rewards/margins": 6.75, + "rewards/rejected": -5.53125, + "step": 551 + }, + { + "epoch": 0.6773006134969325, + "grad_norm": 29.63861597355486, + "learning_rate": 1.1733471074380166e-06, + "logits/chosen": -0.022705078125, + "logits/rejected": -0.158203125, + "logps/chosen": -416.0, + "logps/rejected": -360.0, + "loss": 0.3331, + "rewards/accuracies": 0.828125, + "rewards/chosen": 0.71875, + "rewards/margins": 6.25, + "rewards/rejected": -5.53125, + "step": 552 + }, + { + "epoch": 0.6785276073619632, + "grad_norm": 32.698792932077886, + "learning_rate": 1.172727272727273e-06, + "logits/chosen": 0.01806640625, + "logits/rejected": -0.1259765625, + "logps/chosen": -366.0, + "logps/rejected": -342.0, + "loss": 0.4129, + "rewards/accuracies": 0.8359375, + "rewards/chosen": -0.1474609375, + "rewards/margins": 5.625, + "rewards/rejected": -5.78125, + "step": 553 + }, + { + "epoch": 0.6797546012269938, + "grad_norm": 32.72020008420759, + "learning_rate": 1.172107438016529e-06, + "logits/chosen": 0.0791015625, + "logits/rejected": -0.1171875, + "logps/chosen": -434.0, + "logps/rejected": -334.0, + "loss": 0.3464, + "rewards/accuracies": 0.84375, + "rewards/chosen": 0.83984375, + "rewards/margins": 6.4375, + "rewards/rejected": -5.59375, + "step": 554 + }, + { + "epoch": 0.6809815950920245, + "grad_norm": 30.608371906092913, + "learning_rate": 1.171487603305785e-06, + "logits/chosen": 0.2021484375, + "logits/rejected": 0.09130859375, + "logps/chosen": -364.0, + "logps/rejected": -318.0, + "loss": 0.3666, + "rewards/accuracies": 0.8046875, + "rewards/chosen": 0.3515625, + "rewards/margins": 6.34375, + "rewards/rejected": -6.0, + "step": 555 + }, + { + "epoch": 0.6822085889570552, + "grad_norm": 26.18493705591653, + "learning_rate": 1.1708677685950413e-06, + "logits/chosen": 0.053466796875, + "logits/rejected": -0.162109375, + "logps/chosen": -424.0, + "logps/rejected": -360.0, + "loss": 0.2487, + "rewards/accuracies": 0.890625, + "rewards/chosen": 0.96875, + "rewards/margins": 7.5625, + "rewards/rejected": -6.59375, + "step": 556 + }, + { + "epoch": 0.6834355828220859, + "grad_norm": 27.830299010779843, + "learning_rate": 1.1702479338842975e-06, + "logits/chosen": 0.05029296875, + "logits/rejected": -0.16796875, + "logps/chosen": -398.0, + "logps/rejected": -346.0, + "loss": 0.3343, + "rewards/accuracies": 0.828125, + "rewards/chosen": 0.53515625, + "rewards/margins": 6.6875, + "rewards/rejected": -6.15625, + "step": 557 + }, + { + "epoch": 0.6846625766871166, + "grad_norm": 25.854408165941038, + "learning_rate": 1.1696280991735538e-06, + "logits/chosen": 0.087890625, + "logits/rejected": -0.078125, + "logps/chosen": -378.0, + "logps/rejected": -340.0, + "loss": 0.2782, + "rewards/accuracies": 0.8671875, + "rewards/chosen": -0.3125, + "rewards/margins": 6.9375, + "rewards/rejected": -7.25, + "step": 558 + }, + { + "epoch": 0.6858895705521473, + "grad_norm": 33.3339739767785, + "learning_rate": 1.16900826446281e-06, + "logits/chosen": -0.072265625, + "logits/rejected": -0.1591796875, + "logps/chosen": -388.0, + "logps/rejected": -354.0, + "loss": 0.3848, + "rewards/accuracies": 0.828125, + "rewards/chosen": -0.255859375, + "rewards/margins": 6.5, + "rewards/rejected": -6.78125, + "step": 559 + }, + { + "epoch": 0.6871165644171779, + "grad_norm": 21.74474553758103, + "learning_rate": 1.1683884297520661e-06, + "logits/chosen": -0.1669921875, + "logits/rejected": -0.302734375, + "logps/chosen": -394.0, + "logps/rejected": -344.0, + "loss": 0.2247, + "rewards/accuracies": 0.9140625, + "rewards/chosen": 0.29296875, + "rewards/margins": 7.96875, + "rewards/rejected": -7.6875, + "step": 560 + }, + { + "epoch": 0.6883435582822086, + "grad_norm": 24.604934270788146, + "learning_rate": 1.1677685950413225e-06, + "logits/chosen": 0.00164031982421875, + "logits/rejected": -0.177734375, + "logps/chosen": -360.0, + "logps/rejected": -314.0, + "loss": 0.2132, + "rewards/accuracies": 0.890625, + "rewards/chosen": 0.171875, + "rewards/margins": 7.25, + "rewards/rejected": -7.09375, + "step": 561 + }, + { + "epoch": 0.6895705521472393, + "grad_norm": 25.33619721263671, + "learning_rate": 1.1671487603305786e-06, + "logits/chosen": 0.142578125, + "logits/rejected": -0.064453125, + "logps/chosen": -382.0, + "logps/rejected": -336.0, + "loss": 0.2353, + "rewards/accuracies": 0.875, + "rewards/chosen": -0.55078125, + "rewards/margins": 6.75, + "rewards/rejected": -7.3125, + "step": 562 + }, + { + "epoch": 0.69079754601227, + "grad_norm": 28.2123146430669, + "learning_rate": 1.1665289256198348e-06, + "logits/chosen": -0.055908203125, + "logits/rejected": -0.2041015625, + "logps/chosen": -398.0, + "logps/rejected": -386.0, + "loss": 0.3157, + "rewards/accuracies": 0.8828125, + "rewards/chosen": 0.71484375, + "rewards/margins": 8.0, + "rewards/rejected": -7.3125, + "step": 563 + }, + { + "epoch": 0.6920245398773006, + "grad_norm": 24.610651917154996, + "learning_rate": 1.1659090909090909e-06, + "logits/chosen": -0.040283203125, + "logits/rejected": -0.1552734375, + "logps/chosen": -396.0, + "logps/rejected": -350.0, + "loss": 0.3066, + "rewards/accuracies": 0.859375, + "rewards/chosen": -0.21875, + "rewards/margins": 6.625, + "rewards/rejected": -6.84375, + "step": 564 + }, + { + "epoch": 0.6932515337423313, + "grad_norm": 30.241252760906917, + "learning_rate": 1.165289256198347e-06, + "logits/chosen": 0.0081787109375, + "logits/rejected": -0.1796875, + "logps/chosen": -404.0, + "logps/rejected": -354.0, + "loss": 0.4933, + "rewards/accuracies": 0.828125, + "rewards/chosen": -0.396484375, + "rewards/margins": 6.03125, + "rewards/rejected": -6.40625, + "step": 565 + }, + { + "epoch": 0.694478527607362, + "grad_norm": 30.063355062368128, + "learning_rate": 1.1646694214876034e-06, + "logits/chosen": 0.004852294921875, + "logits/rejected": -0.08544921875, + "logps/chosen": -420.0, + "logps/rejected": -348.0, + "loss": 0.2711, + "rewards/accuracies": 0.84375, + "rewards/chosen": 0.765625, + "rewards/margins": 7.3125, + "rewards/rejected": -6.53125, + "step": 566 + }, + { + "epoch": 0.6957055214723926, + "grad_norm": 23.27581640569879, + "learning_rate": 1.1640495867768595e-06, + "logits/chosen": 0.11865234375, + "logits/rejected": -0.0096435546875, + "logps/chosen": -322.0, + "logps/rejected": -288.0, + "loss": 0.2144, + "rewards/accuracies": 0.890625, + "rewards/chosen": -0.30859375, + "rewards/margins": 6.03125, + "rewards/rejected": -6.34375, + "step": 567 + }, + { + "epoch": 0.6969325153374233, + "grad_norm": 30.55160170642105, + "learning_rate": 1.1634297520661157e-06, + "logits/chosen": -0.0390625, + "logits/rejected": -0.0869140625, + "logps/chosen": -388.0, + "logps/rejected": -334.0, + "loss": 0.3414, + "rewards/accuracies": 0.8203125, + "rewards/chosen": 0.357421875, + "rewards/margins": 6.71875, + "rewards/rejected": -6.375, + "step": 568 + }, + { + "epoch": 0.698159509202454, + "grad_norm": 29.293298679377934, + "learning_rate": 1.162809917355372e-06, + "logits/chosen": 0.04736328125, + "logits/rejected": -0.0986328125, + "logps/chosen": -388.0, + "logps/rejected": -316.0, + "loss": 0.2723, + "rewards/accuracies": 0.84375, + "rewards/chosen": -0.0361328125, + "rewards/margins": 6.28125, + "rewards/rejected": -6.3125, + "step": 569 + }, + { + "epoch": 0.6993865030674846, + "grad_norm": 29.217886094805987, + "learning_rate": 1.1621900826446282e-06, + "logits/chosen": -0.1298828125, + "logits/rejected": -0.2490234375, + "logps/chosen": -406.0, + "logps/rejected": -348.0, + "loss": 0.2078, + "rewards/accuracies": 0.8671875, + "rewards/chosen": 0.9375, + "rewards/margins": 7.40625, + "rewards/rejected": -6.46875, + "step": 570 + }, + { + "epoch": 0.7006134969325153, + "grad_norm": 29.496729411168463, + "learning_rate": 1.1615702479338843e-06, + "logits/chosen": 0.0179443359375, + "logits/rejected": -0.10205078125, + "logps/chosen": -332.0, + "logps/rejected": -314.0, + "loss": 0.3188, + "rewards/accuracies": 0.8359375, + "rewards/chosen": 0.0294189453125, + "rewards/margins": 5.875, + "rewards/rejected": -5.84375, + "step": 571 + }, + { + "epoch": 0.701840490797546, + "grad_norm": 28.740254642158334, + "learning_rate": 1.1609504132231407e-06, + "logits/chosen": 0.03173828125, + "logits/rejected": -0.11669921875, + "logps/chosen": -388.0, + "logps/rejected": -378.0, + "loss": 0.2732, + "rewards/accuracies": 0.890625, + "rewards/chosen": 1.1484375, + "rewards/margins": 7.40625, + "rewards/rejected": -6.25, + "step": 572 + }, + { + "epoch": 0.7030674846625767, + "grad_norm": 35.19506105279461, + "learning_rate": 1.1603305785123966e-06, + "logits/chosen": 0.078125, + "logits/rejected": -0.050048828125, + "logps/chosen": -394.0, + "logps/rejected": -346.0, + "loss": 0.5289, + "rewards/accuracies": 0.8125, + "rewards/chosen": -0.150390625, + "rewards/margins": 6.46875, + "rewards/rejected": -6.59375, + "step": 573 + }, + { + "epoch": 0.7042944785276074, + "grad_norm": 30.319920980400575, + "learning_rate": 1.159710743801653e-06, + "logits/chosen": -0.06201171875, + "logits/rejected": -0.1865234375, + "logps/chosen": -328.0, + "logps/rejected": -320.0, + "loss": 0.34, + "rewards/accuracies": 0.8515625, + "rewards/chosen": 0.28125, + "rewards/margins": 6.78125, + "rewards/rejected": -6.5, + "step": 574 + }, + { + "epoch": 0.7055214723926381, + "grad_norm": 29.615184087034994, + "learning_rate": 1.159090909090909e-06, + "logits/chosen": -0.119140625, + "logits/rejected": -0.2734375, + "logps/chosen": -412.0, + "logps/rejected": -388.0, + "loss": 0.2626, + "rewards/accuracies": 0.8828125, + "rewards/chosen": 1.03125, + "rewards/margins": 8.375, + "rewards/rejected": -7.34375, + "step": 575 + }, + { + "epoch": 0.7067484662576687, + "grad_norm": 22.847983936498952, + "learning_rate": 1.1584710743801652e-06, + "logits/chosen": 0.03271484375, + "logits/rejected": -0.1787109375, + "logps/chosen": -386.0, + "logps/rejected": -344.0, + "loss": 0.1933, + "rewards/accuracies": 0.90625, + "rewards/chosen": 1.0625, + "rewards/margins": 8.5, + "rewards/rejected": -7.46875, + "step": 576 + }, + { + "epoch": 0.7079754601226994, + "grad_norm": 33.54348915562207, + "learning_rate": 1.1578512396694216e-06, + "logits/chosen": 0.08935546875, + "logits/rejected": -0.0296630859375, + "logps/chosen": -368.0, + "logps/rejected": -334.0, + "loss": 0.4071, + "rewards/accuracies": 0.875, + "rewards/chosen": 0.828125, + "rewards/margins": 8.1875, + "rewards/rejected": -7.34375, + "step": 577 + }, + { + "epoch": 0.7092024539877301, + "grad_norm": 31.17463611360663, + "learning_rate": 1.1572314049586777e-06, + "logits/chosen": -0.06591796875, + "logits/rejected": -0.208984375, + "logps/chosen": -420.0, + "logps/rejected": -354.0, + "loss": 0.3672, + "rewards/accuracies": 0.8515625, + "rewards/chosen": 0.921875, + "rewards/margins": 8.6875, + "rewards/rejected": -7.75, + "step": 578 + }, + { + "epoch": 0.7104294478527607, + "grad_norm": 37.58239380773895, + "learning_rate": 1.1566115702479339e-06, + "logits/chosen": -0.025390625, + "logits/rejected": -0.1181640625, + "logps/chosen": -440.0, + "logps/rejected": -352.0, + "loss": 0.4655, + "rewards/accuracies": 0.765625, + "rewards/chosen": -0.6953125, + "rewards/margins": 6.28125, + "rewards/rejected": -7.0, + "step": 579 + }, + { + "epoch": 0.7116564417177914, + "grad_norm": 34.56148267033386, + "learning_rate": 1.1559917355371902e-06, + "logits/chosen": 0.0673828125, + "logits/rejected": -0.10791015625, + "logps/chosen": -390.0, + "logps/rejected": -374.0, + "loss": 0.2813, + "rewards/accuracies": 0.859375, + "rewards/chosen": -0.21484375, + "rewards/margins": 8.0625, + "rewards/rejected": -8.3125, + "step": 580 + }, + { + "epoch": 0.7128834355828221, + "grad_norm": 26.418081399921064, + "learning_rate": 1.1553719008264464e-06, + "logits/chosen": -0.0498046875, + "logits/rejected": -0.244140625, + "logps/chosen": -380.0, + "logps/rejected": -338.0, + "loss": 0.2267, + "rewards/accuracies": 0.8515625, + "rewards/chosen": 0.271484375, + "rewards/margins": 8.3125, + "rewards/rejected": -8.0625, + "step": 581 + }, + { + "epoch": 0.7141104294478527, + "grad_norm": 30.496748663730113, + "learning_rate": 1.1547520661157025e-06, + "logits/chosen": 0.197265625, + "logits/rejected": -0.02587890625, + "logps/chosen": -402.0, + "logps/rejected": -350.0, + "loss": 0.3169, + "rewards/accuracies": 0.859375, + "rewards/chosen": -0.01220703125, + "rewards/margins": 6.96875, + "rewards/rejected": -7.0, + "step": 582 + }, + { + "epoch": 0.7153374233128834, + "grad_norm": 31.96559757214927, + "learning_rate": 1.1541322314049586e-06, + "logits/chosen": -0.02880859375, + "logits/rejected": -0.1748046875, + "logps/chosen": -392.0, + "logps/rejected": -364.0, + "loss": 0.3362, + "rewards/accuracies": 0.8359375, + "rewards/chosen": 0.3671875, + "rewards/margins": 7.5, + "rewards/rejected": -7.125, + "step": 583 + }, + { + "epoch": 0.7165644171779141, + "grad_norm": 36.407577582473756, + "learning_rate": 1.1535123966942148e-06, + "logits/chosen": 0.08154296875, + "logits/rejected": 0.0218505859375, + "logps/chosen": -380.0, + "logps/rejected": -392.0, + "loss": 0.431, + "rewards/accuracies": 0.8125, + "rewards/chosen": -0.58203125, + "rewards/margins": 5.8125, + "rewards/rejected": -6.375, + "step": 584 + }, + { + "epoch": 0.7177914110429447, + "grad_norm": 41.799396220975545, + "learning_rate": 1.1528925619834711e-06, + "logits/chosen": 0.1796875, + "logits/rejected": 0.0634765625, + "logps/chosen": -410.0, + "logps/rejected": -344.0, + "loss": 0.5833, + "rewards/accuracies": 0.7578125, + "rewards/chosen": 0.2255859375, + "rewards/margins": 6.4375, + "rewards/rejected": -6.21875, + "step": 585 + }, + { + "epoch": 0.7190184049079754, + "grad_norm": 32.85377702391839, + "learning_rate": 1.1522727272727273e-06, + "logits/chosen": 0.035888671875, + "logits/rejected": -0.07373046875, + "logps/chosen": -356.0, + "logps/rejected": -324.0, + "loss": 0.3732, + "rewards/accuracies": 0.8515625, + "rewards/chosen": 0.00927734375, + "rewards/margins": 6.9375, + "rewards/rejected": -6.90625, + "step": 586 + }, + { + "epoch": 0.7202453987730061, + "grad_norm": 28.585914695021586, + "learning_rate": 1.1516528925619836e-06, + "logits/chosen": 0.01116943359375, + "logits/rejected": -0.09912109375, + "logps/chosen": -342.0, + "logps/rejected": -332.0, + "loss": 0.3594, + "rewards/accuracies": 0.859375, + "rewards/chosen": 0.08935546875, + "rewards/margins": 7.0625, + "rewards/rejected": -6.96875, + "step": 587 + }, + { + "epoch": 0.7214723926380369, + "grad_norm": 24.10336561473994, + "learning_rate": 1.1510330578512398e-06, + "logits/chosen": -0.0262451171875, + "logits/rejected": -0.1826171875, + "logps/chosen": -388.0, + "logps/rejected": -342.0, + "loss": 0.2892, + "rewards/accuracies": 0.8671875, + "rewards/chosen": 1.2109375, + "rewards/margins": 8.0625, + "rewards/rejected": -6.875, + "step": 588 + }, + { + "epoch": 0.7226993865030675, + "grad_norm": 28.27312313115512, + "learning_rate": 1.150413223140496e-06, + "logits/chosen": -0.0986328125, + "logits/rejected": -0.2041015625, + "logps/chosen": -372.0, + "logps/rejected": -354.0, + "loss": 0.3598, + "rewards/accuracies": 0.8359375, + "rewards/chosen": -0.37890625, + "rewards/margins": 6.28125, + "rewards/rejected": -6.65625, + "step": 589 + }, + { + "epoch": 0.7239263803680982, + "grad_norm": 30.343626428171632, + "learning_rate": 1.149793388429752e-06, + "logits/chosen": 0.00927734375, + "logits/rejected": -0.103515625, + "logps/chosen": -374.0, + "logps/rejected": -330.0, + "loss": 0.3324, + "rewards/accuracies": 0.8359375, + "rewards/chosen": 0.0986328125, + "rewards/margins": 6.875, + "rewards/rejected": -6.78125, + "step": 590 + }, + { + "epoch": 0.7251533742331289, + "grad_norm": 36.86660943740023, + "learning_rate": 1.1491735537190082e-06, + "logits/chosen": -0.169921875, + "logits/rejected": -0.373046875, + "logps/chosen": -414.0, + "logps/rejected": -354.0, + "loss": 0.28, + "rewards/accuracies": 0.8828125, + "rewards/chosen": 0.71484375, + "rewards/margins": 8.0, + "rewards/rejected": -7.3125, + "step": 591 + }, + { + "epoch": 0.7263803680981595, + "grad_norm": 27.76314433981488, + "learning_rate": 1.1485537190082645e-06, + "logits/chosen": -0.02490234375, + "logits/rejected": -0.2138671875, + "logps/chosen": -378.0, + "logps/rejected": -328.0, + "loss": 0.2629, + "rewards/accuracies": 0.8828125, + "rewards/chosen": 0.64453125, + "rewards/margins": 8.4375, + "rewards/rejected": -7.78125, + "step": 592 + }, + { + "epoch": 0.7276073619631902, + "grad_norm": 36.82046387528082, + "learning_rate": 1.1479338842975207e-06, + "logits/chosen": 0.022705078125, + "logits/rejected": -0.2333984375, + "logps/chosen": -454.0, + "logps/rejected": -392.0, + "loss": 0.4117, + "rewards/accuracies": 0.859375, + "rewards/chosen": 0.76953125, + "rewards/margins": 7.84375, + "rewards/rejected": -7.09375, + "step": 593 + }, + { + "epoch": 0.7288343558282209, + "grad_norm": 30.96499084941041, + "learning_rate": 1.1473140495867768e-06, + "logits/chosen": -0.036865234375, + "logits/rejected": -0.197265625, + "logps/chosen": -366.0, + "logps/rejected": -334.0, + "loss": 0.3336, + "rewards/accuracies": 0.84375, + "rewards/chosen": -0.294921875, + "rewards/margins": 7.375, + "rewards/rejected": -7.65625, + "step": 594 + }, + { + "epoch": 0.7300613496932515, + "grad_norm": 33.881983550723454, + "learning_rate": 1.1466942148760332e-06, + "logits/chosen": -0.059814453125, + "logits/rejected": -0.26953125, + "logps/chosen": -454.0, + "logps/rejected": -372.0, + "loss": 0.2673, + "rewards/accuracies": 0.890625, + "rewards/chosen": 0.5234375, + "rewards/margins": 8.25, + "rewards/rejected": -7.71875, + "step": 595 + }, + { + "epoch": 0.7312883435582822, + "grad_norm": 34.4461650357885, + "learning_rate": 1.1460743801652893e-06, + "logits/chosen": -0.0286865234375, + "logits/rejected": -0.18359375, + "logps/chosen": -422.0, + "logps/rejected": -368.0, + "loss": 0.3909, + "rewards/accuracies": 0.8125, + "rewards/chosen": -0.004608154296875, + "rewards/margins": 7.25, + "rewards/rejected": -7.25, + "step": 596 + }, + { + "epoch": 0.7325153374233129, + "grad_norm": 32.96451484117692, + "learning_rate": 1.1454545454545455e-06, + "logits/chosen": 0.0830078125, + "logits/rejected": -0.09228515625, + "logps/chosen": -388.0, + "logps/rejected": -306.0, + "loss": 0.4178, + "rewards/accuracies": 0.8203125, + "rewards/chosen": 0.271484375, + "rewards/margins": 6.8125, + "rewards/rejected": -6.53125, + "step": 597 + }, + { + "epoch": 0.7337423312883435, + "grad_norm": 25.225306353625232, + "learning_rate": 1.1448347107438018e-06, + "logits/chosen": -0.0211181640625, + "logits/rejected": -0.216796875, + "logps/chosen": -446.0, + "logps/rejected": -338.0, + "loss": 0.2285, + "rewards/accuracies": 0.875, + "rewards/chosen": 1.296875, + "rewards/margins": 8.375, + "rewards/rejected": -7.09375, + "step": 598 + }, + { + "epoch": 0.7349693251533742, + "grad_norm": 30.51091893610521, + "learning_rate": 1.1442148760330577e-06, + "logits/chosen": -0.169921875, + "logits/rejected": -0.3125, + "logps/chosen": -456.0, + "logps/rejected": -386.0, + "loss": 0.3423, + "rewards/accuracies": 0.8359375, + "rewards/chosen": 0.82421875, + "rewards/margins": 8.375, + "rewards/rejected": -7.53125, + "step": 599 + }, + { + "epoch": 0.7361963190184049, + "grad_norm": 31.65374536927, + "learning_rate": 1.143595041322314e-06, + "logits/chosen": -0.1796875, + "logits/rejected": -0.25390625, + "logps/chosen": -364.0, + "logps/rejected": -334.0, + "loss": 0.437, + "rewards/accuracies": 0.84375, + "rewards/chosen": 0.3515625, + "rewards/margins": 7.5625, + "rewards/rejected": -7.1875, + "step": 600 + }, + { + "epoch": 0.7374233128834355, + "grad_norm": 29.257508369320092, + "learning_rate": 1.1429752066115702e-06, + "logits/chosen": -0.03759765625, + "logits/rejected": -0.224609375, + "logps/chosen": -348.0, + "logps/rejected": -324.0, + "loss": 0.3337, + "rewards/accuracies": 0.84375, + "rewards/chosen": -0.51953125, + "rewards/margins": 7.21875, + "rewards/rejected": -7.71875, + "step": 601 + }, + { + "epoch": 0.7386503067484662, + "grad_norm": 23.637703488224467, + "learning_rate": 1.1423553719008264e-06, + "logits/chosen": 0.1103515625, + "logits/rejected": -0.0517578125, + "logps/chosen": -406.0, + "logps/rejected": -350.0, + "loss": 0.2192, + "rewards/accuracies": 0.9296875, + "rewards/chosen": 0.64453125, + "rewards/margins": 8.4375, + "rewards/rejected": -7.78125, + "step": 602 + }, + { + "epoch": 0.7398773006134969, + "grad_norm": 24.798455930230393, + "learning_rate": 1.1417355371900827e-06, + "logits/chosen": -0.0615234375, + "logits/rejected": -0.1943359375, + "logps/chosen": -382.0, + "logps/rejected": -336.0, + "loss": 0.3279, + "rewards/accuracies": 0.8671875, + "rewards/chosen": 0.82421875, + "rewards/margins": 8.3125, + "rewards/rejected": -7.5, + "step": 603 + }, + { + "epoch": 0.7411042944785277, + "grad_norm": 28.661443673376716, + "learning_rate": 1.1411157024793389e-06, + "logits/chosen": -0.11376953125, + "logits/rejected": -0.275390625, + "logps/chosen": -420.0, + "logps/rejected": -374.0, + "loss": 0.2818, + "rewards/accuracies": 0.8828125, + "rewards/chosen": 0.267578125, + "rewards/margins": 8.125, + "rewards/rejected": -7.8125, + "step": 604 + }, + { + "epoch": 0.7423312883435583, + "grad_norm": 28.682789953669307, + "learning_rate": 1.140495867768595e-06, + "logits/chosen": -0.057373046875, + "logits/rejected": -0.24609375, + "logps/chosen": -380.0, + "logps/rejected": -348.0, + "loss": 0.2877, + "rewards/accuracies": 0.8828125, + "rewards/chosen": 0.349609375, + "rewards/margins": 8.5, + "rewards/rejected": -8.125, + "step": 605 + }, + { + "epoch": 0.743558282208589, + "grad_norm": 31.866662459492442, + "learning_rate": 1.1398760330578514e-06, + "logits/chosen": -0.138671875, + "logits/rejected": -0.263671875, + "logps/chosen": -390.0, + "logps/rejected": -348.0, + "loss": 0.3751, + "rewards/accuracies": 0.828125, + "rewards/chosen": 0.443359375, + "rewards/margins": 7.65625, + "rewards/rejected": -7.21875, + "step": 606 + }, + { + "epoch": 0.7447852760736197, + "grad_norm": 29.660350580126263, + "learning_rate": 1.1392561983471075e-06, + "logits/chosen": -0.0289306640625, + "logits/rejected": -0.25, + "logps/chosen": -404.0, + "logps/rejected": -386.0, + "loss": 0.3196, + "rewards/accuracies": 0.890625, + "rewards/chosen": -0.265625, + "rewards/margins": 8.0625, + "rewards/rejected": -8.3125, + "step": 607 + }, + { + "epoch": 0.7460122699386503, + "grad_norm": 31.21931912524627, + "learning_rate": 1.1386363636363636e-06, + "logits/chosen": 0.10498046875, + "logits/rejected": -0.1416015625, + "logps/chosen": -404.0, + "logps/rejected": -360.0, + "loss": 0.3793, + "rewards/accuracies": 0.8671875, + "rewards/chosen": -0.6953125, + "rewards/margins": 6.8125, + "rewards/rejected": -7.5, + "step": 608 + }, + { + "epoch": 0.747239263803681, + "grad_norm": 35.278030303126314, + "learning_rate": 1.1380165289256198e-06, + "logits/chosen": -0.02978515625, + "logits/rejected": -0.1953125, + "logps/chosen": -412.0, + "logps/rejected": -342.0, + "loss": 0.4025, + "rewards/accuracies": 0.8515625, + "rewards/chosen": -0.255859375, + "rewards/margins": 7.9375, + "rewards/rejected": -8.1875, + "step": 609 + }, + { + "epoch": 0.7484662576687117, + "grad_norm": 29.748859331814412, + "learning_rate": 1.137396694214876e-06, + "logits/chosen": -0.10302734375, + "logits/rejected": -0.3203125, + "logps/chosen": -452.0, + "logps/rejected": -396.0, + "loss": 0.2827, + "rewards/accuracies": 0.859375, + "rewards/chosen": 0.2216796875, + "rewards/margins": 8.625, + "rewards/rejected": -8.375, + "step": 610 + }, + { + "epoch": 0.7496932515337423, + "grad_norm": 42.36265371910942, + "learning_rate": 1.1367768595041323e-06, + "logits/chosen": -0.050537109375, + "logits/rejected": -0.2216796875, + "logps/chosen": -442.0, + "logps/rejected": -384.0, + "loss": 0.493, + "rewards/accuracies": 0.828125, + "rewards/chosen": -1.234375, + "rewards/margins": 6.84375, + "rewards/rejected": -8.0625, + "step": 611 + }, + { + "epoch": 0.750920245398773, + "grad_norm": 36.03030744089288, + "learning_rate": 1.1361570247933884e-06, + "logits/chosen": -0.042236328125, + "logits/rejected": -0.1796875, + "logps/chosen": -402.0, + "logps/rejected": -376.0, + "loss": 0.4916, + "rewards/accuracies": 0.8203125, + "rewards/chosen": -0.462890625, + "rewards/margins": 6.78125, + "rewards/rejected": -7.25, + "step": 612 + }, + { + "epoch": 0.7521472392638037, + "grad_norm": 35.23308877108843, + "learning_rate": 1.1355371900826446e-06, + "logits/chosen": -0.04052734375, + "logits/rejected": -0.08447265625, + "logps/chosen": -388.0, + "logps/rejected": -348.0, + "loss": 0.4598, + "rewards/accuracies": 0.8359375, + "rewards/chosen": -1.0078125, + "rewards/margins": 6.84375, + "rewards/rejected": -7.875, + "step": 613 + }, + { + "epoch": 0.7533742331288343, + "grad_norm": 35.827377930375896, + "learning_rate": 1.134917355371901e-06, + "logits/chosen": -0.095703125, + "logits/rejected": -0.2216796875, + "logps/chosen": -430.0, + "logps/rejected": -364.0, + "loss": 0.3803, + "rewards/accuracies": 0.828125, + "rewards/chosen": -0.333984375, + "rewards/margins": 7.375, + "rewards/rejected": -7.71875, + "step": 614 + }, + { + "epoch": 0.754601226993865, + "grad_norm": 31.70255072714647, + "learning_rate": 1.134297520661157e-06, + "logits/chosen": -0.07373046875, + "logits/rejected": -0.166015625, + "logps/chosen": -376.0, + "logps/rejected": -364.0, + "loss": 0.3435, + "rewards/accuracies": 0.84375, + "rewards/chosen": 0.037841796875, + "rewards/margins": 7.84375, + "rewards/rejected": -7.8125, + "step": 615 + }, + { + "epoch": 0.7558282208588957, + "grad_norm": 30.125086244839487, + "learning_rate": 1.1336776859504134e-06, + "logits/chosen": -0.004241943359375, + "logits/rejected": -0.15625, + "logps/chosen": -332.0, + "logps/rejected": -310.0, + "loss": 0.3151, + "rewards/accuracies": 0.8359375, + "rewards/chosen": -0.6328125, + "rewards/margins": 6.53125, + "rewards/rejected": -7.15625, + "step": 616 + }, + { + "epoch": 0.7570552147239263, + "grad_norm": 32.7474160937551, + "learning_rate": 1.1330578512396693e-06, + "logits/chosen": 0.030517578125, + "logits/rejected": -0.0390625, + "logps/chosen": -378.0, + "logps/rejected": -336.0, + "loss": 0.3705, + "rewards/accuracies": 0.828125, + "rewards/chosen": -0.419921875, + "rewards/margins": 6.75, + "rewards/rejected": -7.15625, + "step": 617 + }, + { + "epoch": 0.758282208588957, + "grad_norm": 45.056199177526004, + "learning_rate": 1.1324380165289255e-06, + "logits/chosen": 0.109375, + "logits/rejected": -0.056396484375, + "logps/chosen": -426.0, + "logps/rejected": -332.0, + "loss": 0.5324, + "rewards/accuracies": 0.78125, + "rewards/chosen": -0.427734375, + "rewards/margins": 6.1875, + "rewards/rejected": -6.625, + "step": 618 + }, + { + "epoch": 0.7595092024539877, + "grad_norm": 29.95096958444537, + "learning_rate": 1.1318181818181818e-06, + "logits/chosen": 0.08056640625, + "logits/rejected": -0.12060546875, + "logps/chosen": -432.0, + "logps/rejected": -350.0, + "loss": 0.2823, + "rewards/accuracies": 0.8828125, + "rewards/chosen": 0.36328125, + "rewards/margins": 7.71875, + "rewards/rejected": -7.34375, + "step": 619 + }, + { + "epoch": 0.7607361963190185, + "grad_norm": 37.035466575479155, + "learning_rate": 1.131198347107438e-06, + "logits/chosen": 0.08349609375, + "logits/rejected": -0.11181640625, + "logps/chosen": -410.0, + "logps/rejected": -346.0, + "loss": 0.3202, + "rewards/accuracies": 0.828125, + "rewards/chosen": -0.03271484375, + "rewards/margins": 6.65625, + "rewards/rejected": -6.71875, + "step": 620 + }, + { + "epoch": 0.7619631901840491, + "grad_norm": 30.546031931386054, + "learning_rate": 1.1305785123966943e-06, + "logits/chosen": -0.052490234375, + "logits/rejected": -0.2060546875, + "logps/chosen": -482.0, + "logps/rejected": -380.0, + "loss": 0.3488, + "rewards/accuracies": 0.8828125, + "rewards/chosen": 0.7109375, + "rewards/margins": 8.125, + "rewards/rejected": -7.40625, + "step": 621 + }, + { + "epoch": 0.7631901840490798, + "grad_norm": 30.102626290429743, + "learning_rate": 1.1299586776859505e-06, + "logits/chosen": 0.07666015625, + "logits/rejected": -0.0791015625, + "logps/chosen": -382.0, + "logps/rejected": -354.0, + "loss": 0.3683, + "rewards/accuracies": 0.8671875, + "rewards/chosen": 0.2060546875, + "rewards/margins": 7.375, + "rewards/rejected": -7.15625, + "step": 622 + }, + { + "epoch": 0.7644171779141105, + "grad_norm": 31.272882227536257, + "learning_rate": 1.1293388429752066e-06, + "logits/chosen": 0.041259765625, + "logits/rejected": -0.1318359375, + "logps/chosen": -456.0, + "logps/rejected": -390.0, + "loss": 0.338, + "rewards/accuracies": 0.8125, + "rewards/chosen": 1.0078125, + "rewards/margins": 7.78125, + "rewards/rejected": -6.78125, + "step": 623 + }, + { + "epoch": 0.7656441717791411, + "grad_norm": 29.25638333606671, + "learning_rate": 1.128719008264463e-06, + "logits/chosen": -0.036865234375, + "logits/rejected": -0.208984375, + "logps/chosen": -382.0, + "logps/rejected": -334.0, + "loss": 0.302, + "rewards/accuracies": 0.890625, + "rewards/chosen": 0.8671875, + "rewards/margins": 8.5, + "rewards/rejected": -7.59375, + "step": 624 + }, + { + "epoch": 0.7668711656441718, + "grad_norm": 27.76458750020888, + "learning_rate": 1.1280991735537191e-06, + "logits/chosen": 0.01214599609375, + "logits/rejected": -0.123046875, + "logps/chosen": -408.0, + "logps/rejected": -334.0, + "loss": 0.2518, + "rewards/accuracies": 0.859375, + "rewards/chosen": 0.69140625, + "rewards/margins": 7.15625, + "rewards/rejected": -6.46875, + "step": 625 + }, + { + "epoch": 0.7680981595092025, + "grad_norm": 34.03667353357896, + "learning_rate": 1.1274793388429753e-06, + "logits/chosen": -0.119140625, + "logits/rejected": -0.2451171875, + "logps/chosen": -416.0, + "logps/rejected": -358.0, + "loss": 0.4042, + "rewards/accuracies": 0.859375, + "rewards/chosen": 1.140625, + "rewards/margins": 7.03125, + "rewards/rejected": -5.875, + "step": 626 + }, + { + "epoch": 0.7693251533742331, + "grad_norm": 30.346756983376576, + "learning_rate": 1.1268595041322314e-06, + "logits/chosen": 0.095703125, + "logits/rejected": 0.0223388671875, + "logps/chosen": -314.0, + "logps/rejected": -332.0, + "loss": 0.3854, + "rewards/accuracies": 0.8046875, + "rewards/chosen": -0.65625, + "rewards/margins": 5.28125, + "rewards/rejected": -5.9375, + "step": 627 + }, + { + "epoch": 0.7705521472392638, + "grad_norm": 31.196230115400443, + "learning_rate": 1.1262396694214875e-06, + "logits/chosen": 0.0732421875, + "logits/rejected": -0.06884765625, + "logps/chosen": -380.0, + "logps/rejected": -304.0, + "loss": 0.2959, + "rewards/accuracies": 0.8515625, + "rewards/chosen": 0.70703125, + "rewards/margins": 6.40625, + "rewards/rejected": -5.71875, + "step": 628 + }, + { + "epoch": 0.7717791411042945, + "grad_norm": 27.109929558619076, + "learning_rate": 1.1256198347107439e-06, + "logits/chosen": 0.0262451171875, + "logits/rejected": -0.09033203125, + "logps/chosen": -352.0, + "logps/rejected": -292.0, + "loss": 0.2072, + "rewards/accuracies": 0.90625, + "rewards/chosen": 0.353515625, + "rewards/margins": 6.5, + "rewards/rejected": -6.15625, + "step": 629 + }, + { + "epoch": 0.7730061349693251, + "grad_norm": 27.79699333558513, + "learning_rate": 1.125e-06, + "logits/chosen": 0.0947265625, + "logits/rejected": -0.1318359375, + "logps/chosen": -374.0, + "logps/rejected": -372.0, + "loss": 0.3309, + "rewards/accuracies": 0.8671875, + "rewards/chosen": 1.21875, + "rewards/margins": 7.09375, + "rewards/rejected": -5.875, + "step": 630 + }, + { + "epoch": 0.7742331288343558, + "grad_norm": 32.55842929304895, + "learning_rate": 1.1243801652892562e-06, + "logits/chosen": -0.11474609375, + "logits/rejected": -0.2001953125, + "logps/chosen": -394.0, + "logps/rejected": -326.0, + "loss": 0.4115, + "rewards/accuracies": 0.8046875, + "rewards/chosen": 1.03125, + "rewards/margins": 6.71875, + "rewards/rejected": -5.6875, + "step": 631 + }, + { + "epoch": 0.7754601226993865, + "grad_norm": 30.882292248466843, + "learning_rate": 1.1237603305785125e-06, + "logits/chosen": -0.0111083984375, + "logits/rejected": -0.134765625, + "logps/chosen": -406.0, + "logps/rejected": -344.0, + "loss": 0.3282, + "rewards/accuracies": 0.828125, + "rewards/chosen": 1.125, + "rewards/margins": 6.21875, + "rewards/rejected": -5.09375, + "step": 632 + }, + { + "epoch": 0.7766871165644171, + "grad_norm": 32.30019011289937, + "learning_rate": 1.1231404958677687e-06, + "logits/chosen": 0.072265625, + "logits/rejected": -0.03857421875, + "logps/chosen": -342.0, + "logps/rejected": -338.0, + "loss": 0.3515, + "rewards/accuracies": 0.8046875, + "rewards/chosen": 0.39453125, + "rewards/margins": 5.71875, + "rewards/rejected": -5.34375, + "step": 633 + }, + { + "epoch": 0.7779141104294478, + "grad_norm": 26.84260857948927, + "learning_rate": 1.1225206611570248e-06, + "logits/chosen": -0.052490234375, + "logits/rejected": -0.248046875, + "logps/chosen": -416.0, + "logps/rejected": -328.0, + "loss": 0.2366, + "rewards/accuracies": 0.9140625, + "rewards/chosen": 2.546875, + "rewards/margins": 8.3125, + "rewards/rejected": -5.78125, + "step": 634 + }, + { + "epoch": 0.7791411042944786, + "grad_norm": 23.744218806065337, + "learning_rate": 1.121900826446281e-06, + "logits/chosen": -0.0380859375, + "logits/rejected": -0.09912109375, + "logps/chosen": -362.0, + "logps/rejected": -302.0, + "loss": 0.2194, + "rewards/accuracies": 0.90625, + "rewards/chosen": 0.6484375, + "rewards/margins": 6.1875, + "rewards/rejected": -5.5625, + "step": 635 + }, + { + "epoch": 0.7803680981595092, + "grad_norm": 34.66381046216169, + "learning_rate": 1.121280991735537e-06, + "logits/chosen": -0.06201171875, + "logits/rejected": -0.166015625, + "logps/chosen": -414.0, + "logps/rejected": -368.0, + "loss": 0.4205, + "rewards/accuracies": 0.84375, + "rewards/chosen": 1.46875, + "rewards/margins": 6.65625, + "rewards/rejected": -5.1875, + "step": 636 + }, + { + "epoch": 0.7815950920245399, + "grad_norm": 41.24072249858344, + "learning_rate": 1.1206611570247934e-06, + "logits/chosen": 0.11962890625, + "logits/rejected": -0.0615234375, + "logps/chosen": -382.0, + "logps/rejected": -306.0, + "loss": 0.4805, + "rewards/accuracies": 0.8125, + "rewards/chosen": 0.439453125, + "rewards/margins": 6.125, + "rewards/rejected": -5.6875, + "step": 637 + }, + { + "epoch": 0.7828220858895706, + "grad_norm": 31.604533876936294, + "learning_rate": 1.1200413223140496e-06, + "logits/chosen": -0.044677734375, + "logits/rejected": -0.2451171875, + "logps/chosen": -406.0, + "logps/rejected": -340.0, + "loss": 0.2858, + "rewards/accuracies": 0.859375, + "rewards/chosen": 1.3046875, + "rewards/margins": 6.9375, + "rewards/rejected": -5.625, + "step": 638 + }, + { + "epoch": 0.7840490797546013, + "grad_norm": 34.862589865848754, + "learning_rate": 1.1194214876033057e-06, + "logits/chosen": 0.0869140625, + "logits/rejected": -0.080078125, + "logps/chosen": -408.0, + "logps/rejected": -342.0, + "loss": 0.4829, + "rewards/accuracies": 0.828125, + "rewards/chosen": 1.046875, + "rewards/margins": 6.34375, + "rewards/rejected": -5.28125, + "step": 639 + }, + { + "epoch": 0.7852760736196319, + "grad_norm": 29.526720756728196, + "learning_rate": 1.118801652892562e-06, + "logits/chosen": -0.08837890625, + "logits/rejected": -0.2119140625, + "logps/chosen": -446.0, + "logps/rejected": -394.0, + "loss": 0.2982, + "rewards/accuracies": 0.890625, + "rewards/chosen": 1.6328125, + "rewards/margins": 7.5, + "rewards/rejected": -5.875, + "step": 640 + }, + { + "epoch": 0.7865030674846626, + "grad_norm": 38.16790195191018, + "learning_rate": 1.1181818181818182e-06, + "logits/chosen": -0.1259765625, + "logits/rejected": -0.251953125, + "logps/chosen": -398.0, + "logps/rejected": -360.0, + "loss": 0.5882, + "rewards/accuracies": 0.8046875, + "rewards/chosen": 0.90625, + "rewards/margins": 6.15625, + "rewards/rejected": -5.25, + "step": 641 + }, + { + "epoch": 0.7877300613496933, + "grad_norm": 25.77138922359556, + "learning_rate": 1.1175619834710746e-06, + "logits/chosen": -0.01422119140625, + "logits/rejected": -0.2333984375, + "logps/chosen": -352.0, + "logps/rejected": -304.0, + "loss": 0.2926, + "rewards/accuracies": 0.875, + "rewards/chosen": 0.9140625, + "rewards/margins": 6.65625, + "rewards/rejected": -5.75, + "step": 642 + }, + { + "epoch": 0.7889570552147239, + "grad_norm": 35.220692662985385, + "learning_rate": 1.1169421487603307e-06, + "logits/chosen": -0.1748046875, + "logits/rejected": -0.333984375, + "logps/chosen": -406.0, + "logps/rejected": -386.0, + "loss": 0.327, + "rewards/accuracies": 0.8671875, + "rewards/chosen": 0.92578125, + "rewards/margins": 6.84375, + "rewards/rejected": -5.9375, + "step": 643 + }, + { + "epoch": 0.7901840490797546, + "grad_norm": 30.676171080381682, + "learning_rate": 1.1163223140495866e-06, + "logits/chosen": -0.07666015625, + "logits/rejected": -0.21875, + "logps/chosen": -392.0, + "logps/rejected": -338.0, + "loss": 0.3619, + "rewards/accuracies": 0.8515625, + "rewards/chosen": 0.75390625, + "rewards/margins": 7.25, + "rewards/rejected": -6.5, + "step": 644 + }, + { + "epoch": 0.7914110429447853, + "grad_norm": 29.15888390878358, + "learning_rate": 1.115702479338843e-06, + "logits/chosen": -0.03955078125, + "logits/rejected": -0.234375, + "logps/chosen": -336.0, + "logps/rejected": -346.0, + "loss": 0.2816, + "rewards/accuracies": 0.859375, + "rewards/chosen": 0.98046875, + "rewards/margins": 7.53125, + "rewards/rejected": -6.5625, + "step": 645 + }, + { + "epoch": 0.7926380368098159, + "grad_norm": 28.593454046054227, + "learning_rate": 1.1150826446280991e-06, + "logits/chosen": 0.047119140625, + "logits/rejected": -0.134765625, + "logps/chosen": -386.0, + "logps/rejected": -340.0, + "loss": 0.2765, + "rewards/accuracies": 0.8671875, + "rewards/chosen": 0.68359375, + "rewards/margins": 7.28125, + "rewards/rejected": -6.59375, + "step": 646 + }, + { + "epoch": 0.7938650306748466, + "grad_norm": 32.656529890149606, + "learning_rate": 1.1144628099173555e-06, + "logits/chosen": -0.033935546875, + "logits/rejected": -0.1708984375, + "logps/chosen": -386.0, + "logps/rejected": -334.0, + "loss": 0.3573, + "rewards/accuracies": 0.84375, + "rewards/chosen": 0.9296875, + "rewards/margins": 7.78125, + "rewards/rejected": -6.875, + "step": 647 + }, + { + "epoch": 0.7950920245398773, + "grad_norm": 36.18303049965862, + "learning_rate": 1.1138429752066116e-06, + "logits/chosen": -0.0908203125, + "logits/rejected": -0.08837890625, + "logps/chosen": -334.0, + "logps/rejected": -336.0, + "loss": 0.438, + "rewards/accuracies": 0.8125, + "rewards/chosen": -0.0986328125, + "rewards/margins": 6.375, + "rewards/rejected": -6.46875, + "step": 648 + }, + { + "epoch": 0.7963190184049079, + "grad_norm": 33.937125436202074, + "learning_rate": 1.1132231404958678e-06, + "logits/chosen": 0.061279296875, + "logits/rejected": -0.027587890625, + "logps/chosen": -372.0, + "logps/rejected": -342.0, + "loss": 0.5003, + "rewards/accuracies": 0.8125, + "rewards/chosen": 0.0732421875, + "rewards/margins": 6.0, + "rewards/rejected": -5.9375, + "step": 649 + }, + { + "epoch": 0.7975460122699386, + "grad_norm": 36.01649609906715, + "learning_rate": 1.1126033057851241e-06, + "logits/chosen": 0.0264892578125, + "logits/rejected": -0.1298828125, + "logps/chosen": -432.0, + "logps/rejected": -364.0, + "loss": 0.4749, + "rewards/accuracies": 0.8203125, + "rewards/chosen": 0.71875, + "rewards/margins": 6.5, + "rewards/rejected": -5.78125, + "step": 650 + }, + { + "epoch": 0.7987730061349694, + "grad_norm": 37.059100709512336, + "learning_rate": 1.1119834710743803e-06, + "logits/chosen": -0.12109375, + "logits/rejected": -0.1845703125, + "logps/chosen": -376.0, + "logps/rejected": -362.0, + "loss": 0.4208, + "rewards/accuracies": 0.8359375, + "rewards/chosen": 0.9609375, + "rewards/margins": 8.0625, + "rewards/rejected": -7.125, + "step": 651 + }, + { + "epoch": 0.8, + "grad_norm": 40.43433181753938, + "learning_rate": 1.1113636363636364e-06, + "logits/chosen": -0.031494140625, + "logits/rejected": -0.1455078125, + "logps/chosen": -376.0, + "logps/rejected": -360.0, + "loss": 0.4169, + "rewards/accuracies": 0.8515625, + "rewards/chosen": -0.5078125, + "rewards/margins": 6.625, + "rewards/rejected": -7.125, + "step": 652 + }, + { + "epoch": 0.8012269938650307, + "grad_norm": 32.30744337627645, + "learning_rate": 1.1107438016528925e-06, + "logits/chosen": 0.032958984375, + "logits/rejected": -0.142578125, + "logps/chosen": -358.0, + "logps/rejected": -346.0, + "loss": 0.2863, + "rewards/accuracies": 0.8359375, + "rewards/chosen": 0.0498046875, + "rewards/margins": 7.625, + "rewards/rejected": -7.59375, + "step": 653 + }, + { + "epoch": 0.8024539877300614, + "grad_norm": 23.943395122229603, + "learning_rate": 1.1101239669421487e-06, + "logits/chosen": 0.050537109375, + "logits/rejected": -0.15234375, + "logps/chosen": -400.0, + "logps/rejected": -336.0, + "loss": 0.226, + "rewards/accuracies": 0.890625, + "rewards/chosen": 0.333984375, + "rewards/margins": 7.96875, + "rewards/rejected": -7.625, + "step": 654 + }, + { + "epoch": 0.803680981595092, + "grad_norm": 31.423267487900187, + "learning_rate": 1.109504132231405e-06, + "logits/chosen": -0.007659912109375, + "logits/rejected": -0.1787109375, + "logps/chosen": -396.0, + "logps/rejected": -366.0, + "loss": 0.3332, + "rewards/accuracies": 0.8359375, + "rewards/chosen": -0.1943359375, + "rewards/margins": 7.125, + "rewards/rejected": -7.3125, + "step": 655 + }, + { + "epoch": 0.8049079754601227, + "grad_norm": 35.483670047647635, + "learning_rate": 1.1088842975206612e-06, + "logits/chosen": 0.055419921875, + "logits/rejected": -0.08740234375, + "logps/chosen": -380.0, + "logps/rejected": -344.0, + "loss": 0.4163, + "rewards/accuracies": 0.828125, + "rewards/chosen": -0.419921875, + "rewards/margins": 7.09375, + "rewards/rejected": -7.5, + "step": 656 + }, + { + "epoch": 0.8061349693251534, + "grad_norm": 27.21341640433181, + "learning_rate": 1.1082644628099173e-06, + "logits/chosen": -0.10107421875, + "logits/rejected": -0.2431640625, + "logps/chosen": -372.0, + "logps/rejected": -354.0, + "loss": 0.3384, + "rewards/accuracies": 0.8359375, + "rewards/chosen": -0.388671875, + "rewards/margins": 7.1875, + "rewards/rejected": -7.5625, + "step": 657 + }, + { + "epoch": 0.807361963190184, + "grad_norm": 31.34176371809714, + "learning_rate": 1.1076446280991737e-06, + "logits/chosen": -0.08349609375, + "logits/rejected": -0.212890625, + "logps/chosen": -402.0, + "logps/rejected": -366.0, + "loss": 0.3063, + "rewards/accuracies": 0.875, + "rewards/chosen": -0.412109375, + "rewards/margins": 7.21875, + "rewards/rejected": -7.625, + "step": 658 + }, + { + "epoch": 0.8085889570552147, + "grad_norm": 32.62830203002507, + "learning_rate": 1.1070247933884298e-06, + "logits/chosen": -0.0291748046875, + "logits/rejected": -0.16796875, + "logps/chosen": -374.0, + "logps/rejected": -348.0, + "loss": 0.3798, + "rewards/accuracies": 0.84375, + "rewards/chosen": -1.1171875, + "rewards/margins": 6.59375, + "rewards/rejected": -7.71875, + "step": 659 + }, + { + "epoch": 0.8098159509202454, + "grad_norm": 29.335771689820604, + "learning_rate": 1.106404958677686e-06, + "logits/chosen": -0.0380859375, + "logits/rejected": -0.220703125, + "logps/chosen": -412.0, + "logps/rejected": -374.0, + "loss": 0.3215, + "rewards/accuracies": 0.8359375, + "rewards/chosen": 0.053466796875, + "rewards/margins": 7.8125, + "rewards/rejected": -7.75, + "step": 660 + }, + { + "epoch": 0.811042944785276, + "grad_norm": 39.19555957381655, + "learning_rate": 1.1057851239669423e-06, + "logits/chosen": 0.054443359375, + "logits/rejected": -0.052734375, + "logps/chosen": -374.0, + "logps/rejected": -318.0, + "loss": 0.5135, + "rewards/accuracies": 0.8046875, + "rewards/chosen": -1.09375, + "rewards/margins": 6.46875, + "rewards/rejected": -7.5625, + "step": 661 + }, + { + "epoch": 0.8122699386503067, + "grad_norm": 25.59951308246572, + "learning_rate": 1.1051652892561982e-06, + "logits/chosen": 0.05224609375, + "logits/rejected": -0.06884765625, + "logps/chosen": -372.0, + "logps/rejected": -342.0, + "loss": 0.218, + "rewards/accuracies": 0.890625, + "rewards/chosen": 0.3671875, + "rewards/margins": 8.1875, + "rewards/rejected": -7.8125, + "step": 662 + }, + { + "epoch": 0.8134969325153374, + "grad_norm": 29.103568092770104, + "learning_rate": 1.1045454545454546e-06, + "logits/chosen": 0.1142578125, + "logits/rejected": -0.080078125, + "logps/chosen": -362.0, + "logps/rejected": -340.0, + "loss": 0.3666, + "rewards/accuracies": 0.8125, + "rewards/chosen": 0.54296875, + "rewards/margins": 7.375, + "rewards/rejected": -6.8125, + "step": 663 + }, + { + "epoch": 0.8147239263803681, + "grad_norm": 24.88151445188299, + "learning_rate": 1.1039256198347107e-06, + "logits/chosen": -0.2255859375, + "logits/rejected": -0.35546875, + "logps/chosen": -388.0, + "logps/rejected": -376.0, + "loss": 0.213, + "rewards/accuracies": 0.9140625, + "rewards/chosen": 0.7578125, + "rewards/margins": 9.0625, + "rewards/rejected": -8.3125, + "step": 664 + }, + { + "epoch": 0.8159509202453987, + "grad_norm": 34.48345990804285, + "learning_rate": 1.1033057851239669e-06, + "logits/chosen": 0.1259765625, + "logits/rejected": -0.0419921875, + "logps/chosen": -342.0, + "logps/rejected": -308.0, + "loss": 0.4057, + "rewards/accuracies": 0.8125, + "rewards/chosen": -0.2470703125, + "rewards/margins": 7.125, + "rewards/rejected": -7.375, + "step": 665 + }, + { + "epoch": 0.8171779141104294, + "grad_norm": 35.79692152845337, + "learning_rate": 1.1026859504132232e-06, + "logits/chosen": 0.0791015625, + "logits/rejected": -0.0927734375, + "logps/chosen": -422.0, + "logps/rejected": -362.0, + "loss": 0.4703, + "rewards/accuracies": 0.8125, + "rewards/chosen": 0.68359375, + "rewards/margins": 7.28125, + "rewards/rejected": -6.59375, + "step": 666 + }, + { + "epoch": 0.8184049079754602, + "grad_norm": 29.967347606106397, + "learning_rate": 1.1020661157024794e-06, + "logits/chosen": -0.010986328125, + "logits/rejected": -0.2109375, + "logps/chosen": -376.0, + "logps/rejected": -344.0, + "loss": 0.2918, + "rewards/accuracies": 0.890625, + "rewards/chosen": 0.7421875, + "rewards/margins": 7.53125, + "rewards/rejected": -6.78125, + "step": 667 + }, + { + "epoch": 0.8196319018404908, + "grad_norm": 30.862796893024488, + "learning_rate": 1.1014462809917355e-06, + "logits/chosen": 0.0123291015625, + "logits/rejected": -0.119140625, + "logps/chosen": -392.0, + "logps/rejected": -360.0, + "loss": 0.3721, + "rewards/accuracies": 0.8359375, + "rewards/chosen": 0.65625, + "rewards/margins": 7.09375, + "rewards/rejected": -6.4375, + "step": 668 + }, + { + "epoch": 0.8208588957055215, + "grad_norm": 35.17919160040694, + "learning_rate": 1.1008264462809919e-06, + "logits/chosen": -0.052490234375, + "logits/rejected": -0.2060546875, + "logps/chosen": -340.0, + "logps/rejected": -338.0, + "loss": 0.3167, + "rewards/accuracies": 0.875, + "rewards/chosen": 0.6953125, + "rewards/margins": 7.09375, + "rewards/rejected": -6.40625, + "step": 669 + }, + { + "epoch": 0.8220858895705522, + "grad_norm": 25.88170844832394, + "learning_rate": 1.100206611570248e-06, + "logits/chosen": 0.17578125, + "logits/rejected": 0.0250244140625, + "logps/chosen": -324.0, + "logps/rejected": -302.0, + "loss": 0.2514, + "rewards/accuracies": 0.8984375, + "rewards/chosen": 1.0390625, + "rewards/margins": 6.8125, + "rewards/rejected": -5.78125, + "step": 670 + }, + { + "epoch": 0.8233128834355828, + "grad_norm": 30.762328731871737, + "learning_rate": 1.0995867768595042e-06, + "logits/chosen": 0.0390625, + "logits/rejected": -0.12109375, + "logps/chosen": -372.0, + "logps/rejected": -332.0, + "loss": 0.2995, + "rewards/accuracies": 0.890625, + "rewards/chosen": 1.7421875, + "rewards/margins": 7.5, + "rewards/rejected": -5.78125, + "step": 671 + }, + { + "epoch": 0.8245398773006135, + "grad_norm": 30.48600462690235, + "learning_rate": 1.0989669421487603e-06, + "logits/chosen": 0.054931640625, + "logits/rejected": -0.2197265625, + "logps/chosen": -446.0, + "logps/rejected": -332.0, + "loss": 0.3199, + "rewards/accuracies": 0.8515625, + "rewards/chosen": 2.015625, + "rewards/margins": 7.71875, + "rewards/rejected": -5.6875, + "step": 672 + }, + { + "epoch": 0.8257668711656442, + "grad_norm": 32.03977574623459, + "learning_rate": 1.0983471074380164e-06, + "logits/chosen": 0.1416015625, + "logits/rejected": -0.0986328125, + "logps/chosen": -390.0, + "logps/rejected": -328.0, + "loss": 0.3096, + "rewards/accuracies": 0.859375, + "rewards/chosen": 1.109375, + "rewards/margins": 6.21875, + "rewards/rejected": -5.125, + "step": 673 + }, + { + "epoch": 0.8269938650306748, + "grad_norm": 32.93059543816188, + "learning_rate": 1.0977272727272728e-06, + "logits/chosen": 0.162109375, + "logits/rejected": 0.00286865234375, + "logps/chosen": -358.0, + "logps/rejected": -330.0, + "loss": 0.3985, + "rewards/accuracies": 0.8359375, + "rewards/chosen": 1.0546875, + "rewards/margins": 6.59375, + "rewards/rejected": -5.53125, + "step": 674 + }, + { + "epoch": 0.8282208588957055, + "grad_norm": 34.971531039669564, + "learning_rate": 1.097107438016529e-06, + "logits/chosen": 0.08203125, + "logits/rejected": -0.130859375, + "logps/chosen": -426.0, + "logps/rejected": -368.0, + "loss": 0.4407, + "rewards/accuracies": 0.84375, + "rewards/chosen": 1.3828125, + "rewards/margins": 7.40625, + "rewards/rejected": -6.0, + "step": 675 + }, + { + "epoch": 0.8294478527607362, + "grad_norm": 34.2224078373622, + "learning_rate": 1.0964876033057853e-06, + "logits/chosen": -0.037353515625, + "logits/rejected": -0.14453125, + "logps/chosen": -394.0, + "logps/rejected": -352.0, + "loss": 0.3813, + "rewards/accuracies": 0.859375, + "rewards/chosen": 0.4375, + "rewards/margins": 6.8125, + "rewards/rejected": -6.375, + "step": 676 + }, + { + "epoch": 0.8306748466257668, + "grad_norm": 22.049121537748125, + "learning_rate": 1.0958677685950414e-06, + "logits/chosen": 0.09423828125, + "logits/rejected": -0.08447265625, + "logps/chosen": -338.0, + "logps/rejected": -316.0, + "loss": 0.2974, + "rewards/accuracies": 0.875, + "rewards/chosen": 0.2099609375, + "rewards/margins": 7.1875, + "rewards/rejected": -6.96875, + "step": 677 + }, + { + "epoch": 0.8319018404907975, + "grad_norm": 31.042587162348006, + "learning_rate": 1.0952479338842976e-06, + "logits/chosen": 0.07421875, + "logits/rejected": -0.1376953125, + "logps/chosen": -422.0, + "logps/rejected": -362.0, + "loss": 0.4105, + "rewards/accuracies": 0.84375, + "rewards/chosen": 0.439453125, + "rewards/margins": 6.96875, + "rewards/rejected": -6.53125, + "step": 678 + }, + { + "epoch": 0.8331288343558282, + "grad_norm": 32.125565947605324, + "learning_rate": 1.0946280991735537e-06, + "logits/chosen": 0.1689453125, + "logits/rejected": 0.06201171875, + "logps/chosen": -344.0, + "logps/rejected": -328.0, + "loss": 0.3816, + "rewards/accuracies": 0.859375, + "rewards/chosen": 0.44921875, + "rewards/margins": 6.53125, + "rewards/rejected": -6.0625, + "step": 679 + }, + { + "epoch": 0.8343558282208589, + "grad_norm": 30.54691503767555, + "learning_rate": 1.0940082644628098e-06, + "logits/chosen": 0.025390625, + "logits/rejected": -0.150390625, + "logps/chosen": -454.0, + "logps/rejected": -374.0, + "loss": 0.3146, + "rewards/accuracies": 0.84375, + "rewards/chosen": 1.25, + "rewards/margins": 8.125, + "rewards/rejected": -6.875, + "step": 680 + }, + { + "epoch": 0.8355828220858895, + "grad_norm": 25.35697096492706, + "learning_rate": 1.0933884297520662e-06, + "logits/chosen": -0.0262451171875, + "logits/rejected": -0.1796875, + "logps/chosen": -348.0, + "logps/rejected": -338.0, + "loss": 0.2471, + "rewards/accuracies": 0.890625, + "rewards/chosen": 0.72265625, + "rewards/margins": 8.4375, + "rewards/rejected": -7.71875, + "step": 681 + }, + { + "epoch": 0.8368098159509203, + "grad_norm": 29.372252652247152, + "learning_rate": 1.0927685950413223e-06, + "logits/chosen": -0.000934600830078125, + "logits/rejected": -0.115234375, + "logps/chosen": -352.0, + "logps/rejected": -334.0, + "loss": 0.3998, + "rewards/accuracies": 0.84375, + "rewards/chosen": 0.5546875, + "rewards/margins": 7.4375, + "rewards/rejected": -6.90625, + "step": 682 + }, + { + "epoch": 0.838036809815951, + "grad_norm": 28.48014927993986, + "learning_rate": 1.0921487603305785e-06, + "logits/chosen": -0.0264892578125, + "logits/rejected": -0.1328125, + "logps/chosen": -374.0, + "logps/rejected": -330.0, + "loss": 0.3192, + "rewards/accuracies": 0.8359375, + "rewards/chosen": 0.67578125, + "rewards/margins": 8.0625, + "rewards/rejected": -7.375, + "step": 683 + }, + { + "epoch": 0.8392638036809816, + "grad_norm": 30.9956365019602, + "learning_rate": 1.0915289256198348e-06, + "logits/chosen": 0.2001953125, + "logits/rejected": 0.08544921875, + "logps/chosen": -322.0, + "logps/rejected": -352.0, + "loss": 0.3553, + "rewards/accuracies": 0.8515625, + "rewards/chosen": -0.6953125, + "rewards/margins": 6.3125, + "rewards/rejected": -7.03125, + "step": 684 + }, + { + "epoch": 0.8404907975460123, + "grad_norm": 31.024167516184594, + "learning_rate": 1.090909090909091e-06, + "logits/chosen": 0.040283203125, + "logits/rejected": -0.12255859375, + "logps/chosen": -404.0, + "logps/rejected": -380.0, + "loss": 0.3423, + "rewards/accuracies": 0.875, + "rewards/chosen": 0.400390625, + "rewards/margins": 8.625, + "rewards/rejected": -8.25, + "step": 685 + }, + { + "epoch": 0.841717791411043, + "grad_norm": 25.880225574187453, + "learning_rate": 1.0902892561983471e-06, + "logits/chosen": 0.031982421875, + "logits/rejected": -0.095703125, + "logps/chosen": -400.0, + "logps/rejected": -348.0, + "loss": 0.2926, + "rewards/accuracies": 0.8515625, + "rewards/chosen": 0.27734375, + "rewards/margins": 7.9375, + "rewards/rejected": -7.65625, + "step": 686 + }, + { + "epoch": 0.8429447852760736, + "grad_norm": 28.80339614616542, + "learning_rate": 1.0896694214876035e-06, + "logits/chosen": 0.03759765625, + "logits/rejected": -0.1044921875, + "logps/chosen": -380.0, + "logps/rejected": -364.0, + "loss": 0.2973, + "rewards/accuracies": 0.8359375, + "rewards/chosen": 0.484375, + "rewards/margins": 7.90625, + "rewards/rejected": -7.40625, + "step": 687 + }, + { + "epoch": 0.8441717791411043, + "grad_norm": 30.11912577456189, + "learning_rate": 1.0890495867768594e-06, + "logits/chosen": -0.000240325927734375, + "logits/rejected": -0.158203125, + "logps/chosen": -384.0, + "logps/rejected": -328.0, + "loss": 0.327, + "rewards/accuracies": 0.8203125, + "rewards/chosen": -0.5625, + "rewards/margins": 7.5, + "rewards/rejected": -8.0625, + "step": 688 + }, + { + "epoch": 0.845398773006135, + "grad_norm": 30.49529821286284, + "learning_rate": 1.0884297520661158e-06, + "logits/chosen": 0.027099609375, + "logits/rejected": -0.058837890625, + "logps/chosen": -382.0, + "logps/rejected": -348.0, + "loss": 0.344, + "rewards/accuracies": 0.8359375, + "rewards/chosen": 0.271484375, + "rewards/margins": 7.65625, + "rewards/rejected": -7.40625, + "step": 689 + }, + { + "epoch": 0.8466257668711656, + "grad_norm": 41.86773593768432, + "learning_rate": 1.0878099173553719e-06, + "logits/chosen": 0.171875, + "logits/rejected": -0.0174560546875, + "logps/chosen": -414.0, + "logps/rejected": -370.0, + "loss": 0.5772, + "rewards/accuracies": 0.7890625, + "rewards/chosen": 0.408203125, + "rewards/margins": 7.625, + "rewards/rejected": -7.21875, + "step": 690 + }, + { + "epoch": 0.8478527607361963, + "grad_norm": 27.177968327720407, + "learning_rate": 1.087190082644628e-06, + "logits/chosen": -0.162109375, + "logits/rejected": -0.2734375, + "logps/chosen": -366.0, + "logps/rejected": -368.0, + "loss": 0.3098, + "rewards/accuracies": 0.859375, + "rewards/chosen": 0.29296875, + "rewards/margins": 8.1875, + "rewards/rejected": -7.90625, + "step": 691 + }, + { + "epoch": 0.849079754601227, + "grad_norm": 38.42433716881971, + "learning_rate": 1.0865702479338844e-06, + "logits/chosen": 0.0179443359375, + "logits/rejected": -0.142578125, + "logps/chosen": -416.0, + "logps/rejected": -358.0, + "loss": 0.3292, + "rewards/accuracies": 0.859375, + "rewards/chosen": 0.8203125, + "rewards/margins": 8.375, + "rewards/rejected": -7.53125, + "step": 692 + }, + { + "epoch": 0.8503067484662576, + "grad_norm": 31.708672912333903, + "learning_rate": 1.0859504132231405e-06, + "logits/chosen": -0.00286865234375, + "logits/rejected": -0.1904296875, + "logps/chosen": -452.0, + "logps/rejected": -386.0, + "loss": 0.3787, + "rewards/accuracies": 0.84375, + "rewards/chosen": 1.0625, + "rewards/margins": 8.875, + "rewards/rejected": -7.8125, + "step": 693 + }, + { + "epoch": 0.8515337423312883, + "grad_norm": 39.65580173728978, + "learning_rate": 1.0853305785123967e-06, + "logits/chosen": 0.024658203125, + "logits/rejected": -0.162109375, + "logps/chosen": -384.0, + "logps/rejected": -352.0, + "loss": 0.5333, + "rewards/accuracies": 0.8203125, + "rewards/chosen": -0.671875, + "rewards/margins": 6.09375, + "rewards/rejected": -6.75, + "step": 694 + }, + { + "epoch": 0.852760736196319, + "grad_norm": 38.14890644660088, + "learning_rate": 1.084710743801653e-06, + "logits/chosen": 0.06103515625, + "logits/rejected": -0.0179443359375, + "logps/chosen": -338.0, + "logps/rejected": -322.0, + "loss": 0.5352, + "rewards/accuracies": 0.75, + "rewards/chosen": -0.4609375, + "rewards/margins": 5.375, + "rewards/rejected": -5.8125, + "step": 695 + }, + { + "epoch": 0.8539877300613496, + "grad_norm": 32.915799758521196, + "learning_rate": 1.0840909090909092e-06, + "logits/chosen": 0.0947265625, + "logits/rejected": -0.08154296875, + "logps/chosen": -390.0, + "logps/rejected": -336.0, + "loss": 0.3241, + "rewards/accuracies": 0.84375, + "rewards/chosen": 0.365234375, + "rewards/margins": 7.125, + "rewards/rejected": -6.78125, + "step": 696 + }, + { + "epoch": 0.8552147239263803, + "grad_norm": 35.9913417584187, + "learning_rate": 1.0834710743801653e-06, + "logits/chosen": -0.0771484375, + "logits/rejected": -0.10009765625, + "logps/chosen": -398.0, + "logps/rejected": -384.0, + "loss": 0.4041, + "rewards/accuracies": 0.84375, + "rewards/chosen": 0.9375, + "rewards/margins": 8.0, + "rewards/rejected": -7.0625, + "step": 697 + }, + { + "epoch": 0.8564417177914111, + "grad_norm": 31.064185654999054, + "learning_rate": 1.0828512396694214e-06, + "logits/chosen": 0.08984375, + "logits/rejected": -0.027587890625, + "logps/chosen": -404.0, + "logps/rejected": -366.0, + "loss": 0.3837, + "rewards/accuracies": 0.875, + "rewards/chosen": 0.392578125, + "rewards/margins": 7.5, + "rewards/rejected": -7.09375, + "step": 698 + }, + { + "epoch": 0.8576687116564418, + "grad_norm": 26.06263909227069, + "learning_rate": 1.0822314049586776e-06, + "logits/chosen": 0.0260009765625, + "logits/rejected": -0.053955078125, + "logps/chosen": -346.0, + "logps/rejected": -340.0, + "loss": 0.2587, + "rewards/accuracies": 0.8671875, + "rewards/chosen": -0.228515625, + "rewards/margins": 6.46875, + "rewards/rejected": -6.6875, + "step": 699 + }, + { + "epoch": 0.8588957055214724, + "grad_norm": 33.7346004258113, + "learning_rate": 1.081611570247934e-06, + "logits/chosen": 0.1484375, + "logits/rejected": 0.06005859375, + "logps/chosen": -380.0, + "logps/rejected": -342.0, + "loss": 0.398, + "rewards/accuracies": 0.8359375, + "rewards/chosen": -0.0072021484375, + "rewards/margins": 6.28125, + "rewards/rejected": -6.28125, + "step": 700 + }, + { + "epoch": 0.8601226993865031, + "grad_norm": 36.508608167851875, + "learning_rate": 1.08099173553719e-06, + "logits/chosen": 0.1630859375, + "logits/rejected": 0.05810546875, + "logps/chosen": -390.0, + "logps/rejected": -342.0, + "loss": 0.4454, + "rewards/accuracies": 0.78125, + "rewards/chosen": -0.267578125, + "rewards/margins": 6.5625, + "rewards/rejected": -6.8125, + "step": 701 + }, + { + "epoch": 0.8613496932515338, + "grad_norm": 29.752428717784724, + "learning_rate": 1.0803719008264462e-06, + "logits/chosen": -0.005767822265625, + "logits/rejected": -0.30078125, + "logps/chosen": -406.0, + "logps/rejected": -348.0, + "loss": 0.2347, + "rewards/accuracies": 0.859375, + "rewards/chosen": 0.91015625, + "rewards/margins": 8.0, + "rewards/rejected": -7.09375, + "step": 702 + }, + { + "epoch": 0.8625766871165644, + "grad_norm": 28.46276315635888, + "learning_rate": 1.0797520661157026e-06, + "logits/chosen": 0.1513671875, + "logits/rejected": 0.0250244140625, + "logps/chosen": -424.0, + "logps/rejected": -358.0, + "loss": 0.3098, + "rewards/accuracies": 0.859375, + "rewards/chosen": 0.1396484375, + "rewards/margins": 6.53125, + "rewards/rejected": -6.375, + "step": 703 + }, + { + "epoch": 0.8638036809815951, + "grad_norm": 28.331719307949253, + "learning_rate": 1.0791322314049587e-06, + "logits/chosen": 0.06689453125, + "logits/rejected": -0.09033203125, + "logps/chosen": -332.0, + "logps/rejected": -316.0, + "loss": 0.3639, + "rewards/accuracies": 0.859375, + "rewards/chosen": 0.07861328125, + "rewards/margins": 6.96875, + "rewards/rejected": -6.875, + "step": 704 + }, + { + "epoch": 0.8650306748466258, + "grad_norm": 33.73975373967079, + "learning_rate": 1.078512396694215e-06, + "logits/chosen": -0.0615234375, + "logits/rejected": -0.177734375, + "logps/chosen": -424.0, + "logps/rejected": -376.0, + "loss": 0.3238, + "rewards/accuracies": 0.8515625, + "rewards/chosen": 0.75, + "rewards/margins": 8.0, + "rewards/rejected": -7.25, + "step": 705 + }, + { + "epoch": 0.8662576687116564, + "grad_norm": 32.06391249106975, + "learning_rate": 1.077892561983471e-06, + "logits/chosen": 0.01544189453125, + "logits/rejected": -0.1796875, + "logps/chosen": -410.0, + "logps/rejected": -316.0, + "loss": 0.3915, + "rewards/accuracies": 0.8515625, + "rewards/chosen": 0.1748046875, + "rewards/margins": 6.84375, + "rewards/rejected": -6.6875, + "step": 706 + }, + { + "epoch": 0.8674846625766871, + "grad_norm": 32.50030520028076, + "learning_rate": 1.0772727272727271e-06, + "logits/chosen": -0.00185394287109375, + "logits/rejected": -0.1650390625, + "logps/chosen": -394.0, + "logps/rejected": -326.0, + "loss": 0.3089, + "rewards/accuracies": 0.8515625, + "rewards/chosen": -0.0947265625, + "rewards/margins": 7.28125, + "rewards/rejected": -7.375, + "step": 707 + }, + { + "epoch": 0.8687116564417178, + "grad_norm": 27.35860386441147, + "learning_rate": 1.0766528925619835e-06, + "logits/chosen": 0.0245361328125, + "logits/rejected": -0.193359375, + "logps/chosen": -380.0, + "logps/rejected": -328.0, + "loss": 0.2856, + "rewards/accuracies": 0.8359375, + "rewards/chosen": 0.65234375, + "rewards/margins": 7.28125, + "rewards/rejected": -6.625, + "step": 708 + }, + { + "epoch": 0.8699386503067484, + "grad_norm": 22.967542887627133, + "learning_rate": 1.0760330578512396e-06, + "logits/chosen": 0.11376953125, + "logits/rejected": -0.031982421875, + "logps/chosen": -444.0, + "logps/rejected": -360.0, + "loss": 0.2311, + "rewards/accuracies": 0.875, + "rewards/chosen": 0.984375, + "rewards/margins": 8.4375, + "rewards/rejected": -7.46875, + "step": 709 + }, + { + "epoch": 0.8711656441717791, + "grad_norm": 30.012370581713046, + "learning_rate": 1.075413223140496e-06, + "logits/chosen": 0.1845703125, + "logits/rejected": 0.06103515625, + "logps/chosen": -372.0, + "logps/rejected": -330.0, + "loss": 0.3445, + "rewards/accuracies": 0.8125, + "rewards/chosen": 0.1064453125, + "rewards/margins": 6.46875, + "rewards/rejected": -6.375, + "step": 710 + }, + { + "epoch": 0.8723926380368098, + "grad_norm": 26.59602404250697, + "learning_rate": 1.0747933884297521e-06, + "logits/chosen": 0.15625, + "logits/rejected": -0.039306640625, + "logps/chosen": -390.0, + "logps/rejected": -332.0, + "loss": 0.3562, + "rewards/accuracies": 0.8515625, + "rewards/chosen": 0.5703125, + "rewards/margins": 7.0, + "rewards/rejected": -6.4375, + "step": 711 + }, + { + "epoch": 0.8736196319018404, + "grad_norm": 34.082753455253695, + "learning_rate": 1.0741735537190083e-06, + "logits/chosen": 0.1484375, + "logits/rejected": 0.00482177734375, + "logps/chosen": -410.0, + "logps/rejected": -356.0, + "loss": 0.4269, + "rewards/accuracies": 0.8125, + "rewards/chosen": 1.328125, + "rewards/margins": 7.03125, + "rewards/rejected": -5.71875, + "step": 712 + }, + { + "epoch": 0.8748466257668711, + "grad_norm": 24.8779582799026, + "learning_rate": 1.0735537190082646e-06, + "logits/chosen": 0.05810546875, + "logits/rejected": -0.1484375, + "logps/chosen": -386.0, + "logps/rejected": -336.0, + "loss": 0.2908, + "rewards/accuracies": 0.875, + "rewards/chosen": 0.625, + "rewards/margins": 7.28125, + "rewards/rejected": -6.65625, + "step": 713 + }, + { + "epoch": 0.8760736196319019, + "grad_norm": 27.823114674287325, + "learning_rate": 1.0729338842975208e-06, + "logits/chosen": 0.181640625, + "logits/rejected": -0.038818359375, + "logps/chosen": -370.0, + "logps/rejected": -328.0, + "loss": 0.3057, + "rewards/accuracies": 0.890625, + "rewards/chosen": 0.87109375, + "rewards/margins": 6.84375, + "rewards/rejected": -6.0, + "step": 714 + }, + { + "epoch": 0.8773006134969326, + "grad_norm": 29.027125190845624, + "learning_rate": 1.072314049586777e-06, + "logits/chosen": -0.06494140625, + "logits/rejected": -0.1318359375, + "logps/chosen": -384.0, + "logps/rejected": -328.0, + "loss": 0.3391, + "rewards/accuracies": 0.8203125, + "rewards/chosen": 1.015625, + "rewards/margins": 7.75, + "rewards/rejected": -6.71875, + "step": 715 + }, + { + "epoch": 0.8785276073619632, + "grad_norm": 26.494530473617345, + "learning_rate": 1.071694214876033e-06, + "logits/chosen": -0.018798828125, + "logits/rejected": -0.1865234375, + "logps/chosen": -384.0, + "logps/rejected": -324.0, + "loss": 0.3073, + "rewards/accuracies": 0.8515625, + "rewards/chosen": 1.1640625, + "rewards/margins": 6.84375, + "rewards/rejected": -5.65625, + "step": 716 + }, + { + "epoch": 0.8797546012269939, + "grad_norm": 29.02718718589355, + "learning_rate": 1.0710743801652892e-06, + "logits/chosen": 0.07470703125, + "logits/rejected": -0.08056640625, + "logps/chosen": -344.0, + "logps/rejected": -284.0, + "loss": 0.3344, + "rewards/accuracies": 0.84375, + "rewards/chosen": 0.6875, + "rewards/margins": 6.90625, + "rewards/rejected": -6.21875, + "step": 717 + }, + { + "epoch": 0.8809815950920246, + "grad_norm": 23.193149611746794, + "learning_rate": 1.0704545454545455e-06, + "logits/chosen": -0.034912109375, + "logits/rejected": -0.1640625, + "logps/chosen": -402.0, + "logps/rejected": -356.0, + "loss": 0.2308, + "rewards/accuracies": 0.8828125, + "rewards/chosen": 1.5859375, + "rewards/margins": 7.84375, + "rewards/rejected": -6.25, + "step": 718 + }, + { + "epoch": 0.8822085889570552, + "grad_norm": 36.13316553444024, + "learning_rate": 1.0698347107438017e-06, + "logits/chosen": -0.07763671875, + "logits/rejected": -0.19921875, + "logps/chosen": -410.0, + "logps/rejected": -368.0, + "loss": 0.3568, + "rewards/accuracies": 0.8203125, + "rewards/chosen": 1.8046875, + "rewards/margins": 8.0, + "rewards/rejected": -6.1875, + "step": 719 + }, + { + "epoch": 0.8834355828220859, + "grad_norm": 25.470910809931194, + "learning_rate": 1.0692148760330578e-06, + "logits/chosen": -0.07421875, + "logits/rejected": -0.279296875, + "logps/chosen": -412.0, + "logps/rejected": -350.0, + "loss": 0.2706, + "rewards/accuracies": 0.8828125, + "rewards/chosen": 1.9140625, + "rewards/margins": 7.9375, + "rewards/rejected": -6.03125, + "step": 720 + }, + { + "epoch": 0.8846625766871166, + "grad_norm": 24.45362594207483, + "learning_rate": 1.0685950413223142e-06, + "logits/chosen": -0.0810546875, + "logits/rejected": -0.337890625, + "logps/chosen": -414.0, + "logps/rejected": -364.0, + "loss": 0.2805, + "rewards/accuracies": 0.921875, + "rewards/chosen": 1.5625, + "rewards/margins": 8.8125, + "rewards/rejected": -7.25, + "step": 721 + }, + { + "epoch": 0.8858895705521472, + "grad_norm": 26.653798150160988, + "learning_rate": 1.0679752066115703e-06, + "logits/chosen": -0.049072265625, + "logits/rejected": -0.1376953125, + "logps/chosen": -380.0, + "logps/rejected": -358.0, + "loss": 0.2495, + "rewards/accuracies": 0.890625, + "rewards/chosen": 1.609375, + "rewards/margins": 7.875, + "rewards/rejected": -6.28125, + "step": 722 + }, + { + "epoch": 0.8871165644171779, + "grad_norm": 29.259615066962372, + "learning_rate": 1.0673553719008265e-06, + "logits/chosen": -0.043212890625, + "logits/rejected": -0.14453125, + "logps/chosen": -376.0, + "logps/rejected": -334.0, + "loss": 0.3287, + "rewards/accuracies": 0.859375, + "rewards/chosen": 0.94921875, + "rewards/margins": 7.4375, + "rewards/rejected": -6.5, + "step": 723 + }, + { + "epoch": 0.8883435582822086, + "grad_norm": 30.45569089414315, + "learning_rate": 1.0667355371900826e-06, + "logits/chosen": -0.0308837890625, + "logits/rejected": -0.251953125, + "logps/chosen": -382.0, + "logps/rejected": -320.0, + "loss": 0.3219, + "rewards/accuracies": 0.859375, + "rewards/chosen": 1.3359375, + "rewards/margins": 7.8125, + "rewards/rejected": -6.5, + "step": 724 + }, + { + "epoch": 0.8895705521472392, + "grad_norm": 30.8228410953136, + "learning_rate": 1.0661157024793387e-06, + "logits/chosen": 0.12255859375, + "logits/rejected": -0.05419921875, + "logps/chosen": -404.0, + "logps/rejected": -340.0, + "loss": 0.3833, + "rewards/accuracies": 0.8359375, + "rewards/chosen": 0.984375, + "rewards/margins": 7.375, + "rewards/rejected": -6.375, + "step": 725 + }, + { + "epoch": 0.8907975460122699, + "grad_norm": 32.44931458712784, + "learning_rate": 1.065495867768595e-06, + "logits/chosen": -0.030029296875, + "logits/rejected": -0.154296875, + "logps/chosen": -340.0, + "logps/rejected": -302.0, + "loss": 0.2852, + "rewards/accuracies": 0.8515625, + "rewards/chosen": 0.5703125, + "rewards/margins": 7.53125, + "rewards/rejected": -6.96875, + "step": 726 + }, + { + "epoch": 0.8920245398773006, + "grad_norm": 30.00382649809932, + "learning_rate": 1.0648760330578512e-06, + "logits/chosen": -0.0013427734375, + "logits/rejected": -0.1474609375, + "logps/chosen": -374.0, + "logps/rejected": -338.0, + "loss": 0.2567, + "rewards/accuracies": 0.890625, + "rewards/chosen": 0.84765625, + "rewards/margins": 7.5625, + "rewards/rejected": -6.71875, + "step": 727 + }, + { + "epoch": 0.8932515337423312, + "grad_norm": 32.084882438300085, + "learning_rate": 1.0642561983471074e-06, + "logits/chosen": -0.0244140625, + "logits/rejected": -0.1611328125, + "logps/chosen": -406.0, + "logps/rejected": -384.0, + "loss": 0.3182, + "rewards/accuracies": 0.8515625, + "rewards/chosen": 1.234375, + "rewards/margins": 8.875, + "rewards/rejected": -7.625, + "step": 728 + }, + { + "epoch": 0.894478527607362, + "grad_norm": 24.427532348622456, + "learning_rate": 1.0636363636363637e-06, + "logits/chosen": -0.040771484375, + "logits/rejected": -0.1708984375, + "logps/chosen": -396.0, + "logps/rejected": -372.0, + "loss": 0.2099, + "rewards/accuracies": 0.8984375, + "rewards/chosen": 0.9765625, + "rewards/margins": 8.875, + "rewards/rejected": -7.90625, + "step": 729 + }, + { + "epoch": 0.8957055214723927, + "grad_norm": 24.535854490724105, + "learning_rate": 1.0630165289256199e-06, + "logits/chosen": -0.103515625, + "logits/rejected": -0.2060546875, + "logps/chosen": -404.0, + "logps/rejected": -380.0, + "loss": 0.2127, + "rewards/accuracies": 0.9140625, + "rewards/chosen": 1.515625, + "rewards/margins": 8.8125, + "rewards/rejected": -7.3125, + "step": 730 + }, + { + "epoch": 0.8969325153374234, + "grad_norm": 34.14852950381922, + "learning_rate": 1.062396694214876e-06, + "logits/chosen": 0.1171875, + "logits/rejected": 0.002349853515625, + "logps/chosen": -438.0, + "logps/rejected": -366.0, + "loss": 0.3465, + "rewards/accuracies": 0.859375, + "rewards/chosen": 0.62890625, + "rewards/margins": 8.1875, + "rewards/rejected": -7.5625, + "step": 731 + }, + { + "epoch": 0.898159509202454, + "grad_norm": 34.434393814707754, + "learning_rate": 1.0617768595041324e-06, + "logits/chosen": -0.058837890625, + "logits/rejected": -0.2099609375, + "logps/chosen": -380.0, + "logps/rejected": -372.0, + "loss": 0.4237, + "rewards/accuracies": 0.8046875, + "rewards/chosen": 0.6875, + "rewards/margins": 7.71875, + "rewards/rejected": -7.03125, + "step": 732 + }, + { + "epoch": 0.8993865030674847, + "grad_norm": 28.68332287996042, + "learning_rate": 1.0611570247933883e-06, + "logits/chosen": 0.0673828125, + "logits/rejected": -0.083984375, + "logps/chosen": -366.0, + "logps/rejected": -362.0, + "loss": 0.2815, + "rewards/accuracies": 0.859375, + "rewards/chosen": 0.1015625, + "rewards/margins": 7.65625, + "rewards/rejected": -7.5625, + "step": 733 + }, + { + "epoch": 0.9006134969325154, + "grad_norm": 32.670441266137765, + "learning_rate": 1.0605371900826447e-06, + "logits/chosen": 0.06884765625, + "logits/rejected": -0.1044921875, + "logps/chosen": -384.0, + "logps/rejected": -354.0, + "loss": 0.3319, + "rewards/accuracies": 0.859375, + "rewards/chosen": 0.408203125, + "rewards/margins": 8.3125, + "rewards/rejected": -7.90625, + "step": 734 + }, + { + "epoch": 0.901840490797546, + "grad_norm": 28.766622871975752, + "learning_rate": 1.0599173553719008e-06, + "logits/chosen": -0.037841796875, + "logits/rejected": -0.259765625, + "logps/chosen": -434.0, + "logps/rejected": -390.0, + "loss": 0.2447, + "rewards/accuracies": 0.875, + "rewards/chosen": 0.2490234375, + "rewards/margins": 8.875, + "rewards/rejected": -8.625, + "step": 735 + }, + { + "epoch": 0.9030674846625767, + "grad_norm": 31.10013526298491, + "learning_rate": 1.059297520661157e-06, + "logits/chosen": 0.1533203125, + "logits/rejected": 0.029541015625, + "logps/chosen": -362.0, + "logps/rejected": -340.0, + "loss": 0.4211, + "rewards/accuracies": 0.8359375, + "rewards/chosen": -0.53515625, + "rewards/margins": 7.03125, + "rewards/rejected": -7.5625, + "step": 736 + }, + { + "epoch": 0.9042944785276074, + "grad_norm": 38.03816635846729, + "learning_rate": 1.0586776859504133e-06, + "logits/chosen": 0.04541015625, + "logits/rejected": -0.051513671875, + "logps/chosen": -360.0, + "logps/rejected": -348.0, + "loss": 0.4641, + "rewards/accuracies": 0.8046875, + "rewards/chosen": -0.478515625, + "rewards/margins": 7.09375, + "rewards/rejected": -7.5625, + "step": 737 + }, + { + "epoch": 0.905521472392638, + "grad_norm": 32.716309617319716, + "learning_rate": 1.0580578512396694e-06, + "logits/chosen": -0.0380859375, + "logits/rejected": -0.1474609375, + "logps/chosen": -450.0, + "logps/rejected": -364.0, + "loss": 0.3437, + "rewards/accuracies": 0.8359375, + "rewards/chosen": 0.60546875, + "rewards/margins": 8.1875, + "rewards/rejected": -7.59375, + "step": 738 + }, + { + "epoch": 0.9067484662576687, + "grad_norm": 29.627927045089628, + "learning_rate": 1.0574380165289258e-06, + "logits/chosen": 0.060302734375, + "logits/rejected": -0.0654296875, + "logps/chosen": -346.0, + "logps/rejected": -340.0, + "loss": 0.2827, + "rewards/accuracies": 0.84375, + "rewards/chosen": -0.06640625, + "rewards/margins": 7.1875, + "rewards/rejected": -7.25, + "step": 739 + }, + { + "epoch": 0.9079754601226994, + "grad_norm": 29.046321525927297, + "learning_rate": 1.056818181818182e-06, + "logits/chosen": -0.0308837890625, + "logits/rejected": -0.18359375, + "logps/chosen": -344.0, + "logps/rejected": -326.0, + "loss": 0.3577, + "rewards/accuracies": 0.8515625, + "rewards/chosen": 0.02392578125, + "rewards/margins": 7.5625, + "rewards/rejected": -7.53125, + "step": 740 + }, + { + "epoch": 0.90920245398773, + "grad_norm": 29.826427625052336, + "learning_rate": 1.056198347107438e-06, + "logits/chosen": -0.00982666015625, + "logits/rejected": -0.162109375, + "logps/chosen": -356.0, + "logps/rejected": -328.0, + "loss": 0.3514, + "rewards/accuracies": 0.84375, + "rewards/chosen": 0.8359375, + "rewards/margins": 8.75, + "rewards/rejected": -7.875, + "step": 741 + }, + { + "epoch": 0.9104294478527607, + "grad_norm": 28.054610141633944, + "learning_rate": 1.0555785123966942e-06, + "logits/chosen": -0.033935546875, + "logits/rejected": -0.1982421875, + "logps/chosen": -368.0, + "logps/rejected": -372.0, + "loss": 0.2813, + "rewards/accuracies": 0.8828125, + "rewards/chosen": 0.6640625, + "rewards/margins": 8.5625, + "rewards/rejected": -7.90625, + "step": 742 + }, + { + "epoch": 0.9116564417177914, + "grad_norm": 35.487753263657986, + "learning_rate": 1.0549586776859503e-06, + "logits/chosen": 0.01055908203125, + "logits/rejected": -0.040771484375, + "logps/chosen": -424.0, + "logps/rejected": -392.0, + "loss": 0.3821, + "rewards/accuracies": 0.8203125, + "rewards/chosen": 0.2265625, + "rewards/margins": 7.21875, + "rewards/rejected": -7.0, + "step": 743 + }, + { + "epoch": 0.912883435582822, + "grad_norm": 28.496366117513464, + "learning_rate": 1.0543388429752067e-06, + "logits/chosen": 0.035400390625, + "logits/rejected": -0.14453125, + "logps/chosen": -380.0, + "logps/rejected": -340.0, + "loss": 0.2775, + "rewards/accuracies": 0.859375, + "rewards/chosen": 0.9609375, + "rewards/margins": 8.6875, + "rewards/rejected": -7.71875, + "step": 744 + }, + { + "epoch": 0.9141104294478528, + "grad_norm": 30.005979043165016, + "learning_rate": 1.0537190082644628e-06, + "logits/chosen": 0.08642578125, + "logits/rejected": -0.0003509521484375, + "logps/chosen": -346.0, + "logps/rejected": -318.0, + "loss": 0.3381, + "rewards/accuracies": 0.8515625, + "rewards/chosen": 0.1044921875, + "rewards/margins": 6.84375, + "rewards/rejected": -6.75, + "step": 745 + }, + { + "epoch": 0.9153374233128835, + "grad_norm": 36.365949373279655, + "learning_rate": 1.053099173553719e-06, + "logits/chosen": 0.09033203125, + "logits/rejected": -0.07958984375, + "logps/chosen": -408.0, + "logps/rejected": -356.0, + "loss": 0.4283, + "rewards/accuracies": 0.84375, + "rewards/chosen": 0.4375, + "rewards/margins": 7.4375, + "rewards/rejected": -7.0, + "step": 746 + }, + { + "epoch": 0.9165644171779141, + "grad_norm": 30.46892734256714, + "learning_rate": 1.0524793388429753e-06, + "logits/chosen": -0.1123046875, + "logits/rejected": -0.21875, + "logps/chosen": -342.0, + "logps/rejected": -344.0, + "loss": 0.2905, + "rewards/accuracies": 0.8515625, + "rewards/chosen": -0.0234375, + "rewards/margins": 6.75, + "rewards/rejected": -6.75, + "step": 747 + }, + { + "epoch": 0.9177914110429448, + "grad_norm": 37.226894144134455, + "learning_rate": 1.0518595041322315e-06, + "logits/chosen": 0.125, + "logits/rejected": 0.01458740234375, + "logps/chosen": -380.0, + "logps/rejected": -354.0, + "loss": 0.4992, + "rewards/accuracies": 0.78125, + "rewards/chosen": -0.50390625, + "rewards/margins": 6.28125, + "rewards/rejected": -6.78125, + "step": 748 + }, + { + "epoch": 0.9190184049079755, + "grad_norm": 34.24194812875213, + "learning_rate": 1.0512396694214876e-06, + "logits/chosen": -0.04541015625, + "logits/rejected": -0.2431640625, + "logps/chosen": -424.0, + "logps/rejected": -358.0, + "loss": 0.4049, + "rewards/accuracies": 0.8125, + "rewards/chosen": 0.2470703125, + "rewards/margins": 7.59375, + "rewards/rejected": -7.34375, + "step": 749 + }, + { + "epoch": 0.9202453987730062, + "grad_norm": 35.848596312284045, + "learning_rate": 1.050619834710744e-06, + "logits/chosen": 0.01190185546875, + "logits/rejected": -0.119140625, + "logps/chosen": -412.0, + "logps/rejected": -348.0, + "loss": 0.4472, + "rewards/accuracies": 0.828125, + "rewards/chosen": 0.09326171875, + "rewards/margins": 7.375, + "rewards/rejected": -7.28125, + "step": 750 + }, + { + "epoch": 0.9214723926380368, + "grad_norm": 36.49684940562663, + "learning_rate": 1.05e-06, + "logits/chosen": -0.0064697265625, + "logits/rejected": -0.169921875, + "logps/chosen": -428.0, + "logps/rejected": -366.0, + "loss": 0.3626, + "rewards/accuracies": 0.84375, + "rewards/chosen": 0.5234375, + "rewards/margins": 8.3125, + "rewards/rejected": -7.8125, + "step": 751 + }, + { + "epoch": 0.9226993865030675, + "grad_norm": 31.9902592360704, + "learning_rate": 1.0493801652892563e-06, + "logits/chosen": -0.028564453125, + "logits/rejected": -0.275390625, + "logps/chosen": -424.0, + "logps/rejected": -368.0, + "loss": 0.3427, + "rewards/accuracies": 0.8515625, + "rewards/chosen": 0.640625, + "rewards/margins": 8.375, + "rewards/rejected": -7.71875, + "step": 752 + }, + { + "epoch": 0.9239263803680982, + "grad_norm": 30.210419690529548, + "learning_rate": 1.0487603305785124e-06, + "logits/chosen": 0.142578125, + "logits/rejected": 0.01361083984375, + "logps/chosen": -408.0, + "logps/rejected": -362.0, + "loss": 0.3032, + "rewards/accuracies": 0.859375, + "rewards/chosen": 0.126953125, + "rewards/margins": 7.375, + "rewards/rejected": -7.25, + "step": 753 + }, + { + "epoch": 0.9251533742331288, + "grad_norm": 33.70924536763799, + "learning_rate": 1.0481404958677685e-06, + "logits/chosen": 0.06640625, + "logits/rejected": -0.1044921875, + "logps/chosen": -394.0, + "logps/rejected": -348.0, + "loss": 0.4405, + "rewards/accuracies": 0.796875, + "rewards/chosen": -0.1416015625, + "rewards/margins": 7.09375, + "rewards/rejected": -7.21875, + "step": 754 + }, + { + "epoch": 0.9263803680981595, + "grad_norm": 28.923057351049156, + "learning_rate": 1.0475206611570249e-06, + "logits/chosen": 0.01611328125, + "logits/rejected": -0.1064453125, + "logps/chosen": -388.0, + "logps/rejected": -366.0, + "loss": 0.313, + "rewards/accuracies": 0.875, + "rewards/chosen": -0.291015625, + "rewards/margins": 7.21875, + "rewards/rejected": -7.5, + "step": 755 + }, + { + "epoch": 0.9276073619631902, + "grad_norm": 26.956563960777117, + "learning_rate": 1.046900826446281e-06, + "logits/chosen": 0.115234375, + "logits/rejected": 0.02978515625, + "logps/chosen": -342.0, + "logps/rejected": -326.0, + "loss": 0.2891, + "rewards/accuracies": 0.8828125, + "rewards/chosen": -0.640625, + "rewards/margins": 6.8125, + "rewards/rejected": -7.46875, + "step": 756 + }, + { + "epoch": 0.9288343558282208, + "grad_norm": 25.54044780724446, + "learning_rate": 1.0462809917355372e-06, + "logits/chosen": 0.15625, + "logits/rejected": -0.0595703125, + "logps/chosen": -404.0, + "logps/rejected": -344.0, + "loss": 0.27, + "rewards/accuracies": 0.875, + "rewards/chosen": 0.7578125, + "rewards/margins": 8.0625, + "rewards/rejected": -7.3125, + "step": 757 + }, + { + "epoch": 0.9300613496932515, + "grad_norm": 35.658646153584634, + "learning_rate": 1.0456611570247935e-06, + "logits/chosen": 0.08740234375, + "logits/rejected": -0.080078125, + "logps/chosen": -376.0, + "logps/rejected": -320.0, + "loss": 0.3867, + "rewards/accuracies": 0.8125, + "rewards/chosen": -0.55859375, + "rewards/margins": 6.71875, + "rewards/rejected": -7.28125, + "step": 758 + }, + { + "epoch": 0.9312883435582822, + "grad_norm": 29.003566311314795, + "learning_rate": 1.0450413223140497e-06, + "logits/chosen": 0.1123046875, + "logits/rejected": -0.11572265625, + "logps/chosen": -396.0, + "logps/rejected": -340.0, + "loss": 0.2888, + "rewards/accuracies": 0.84375, + "rewards/chosen": 0.16015625, + "rewards/margins": 6.9375, + "rewards/rejected": -6.78125, + "step": 759 + }, + { + "epoch": 0.9325153374233128, + "grad_norm": 25.36330419002366, + "learning_rate": 1.0444214876033058e-06, + "logits/chosen": -0.0986328125, + "logits/rejected": -0.35546875, + "logps/chosen": -442.0, + "logps/rejected": -366.0, + "loss": 0.2887, + "rewards/accuracies": 0.875, + "rewards/chosen": 1.1796875, + "rewards/margins": 8.625, + "rewards/rejected": -7.4375, + "step": 760 + }, + { + "epoch": 0.9337423312883436, + "grad_norm": 31.72216949313869, + "learning_rate": 1.043801652892562e-06, + "logits/chosen": 0.1484375, + "logits/rejected": 0.0277099609375, + "logps/chosen": -382.0, + "logps/rejected": -354.0, + "loss": 0.3707, + "rewards/accuracies": 0.8671875, + "rewards/chosen": 0.3984375, + "rewards/margins": 7.21875, + "rewards/rejected": -6.8125, + "step": 761 + }, + { + "epoch": 0.9349693251533743, + "grad_norm": 31.01314079949462, + "learning_rate": 1.043181818181818e-06, + "logits/chosen": 0.1591796875, + "logits/rejected": -0.020751953125, + "logps/chosen": -428.0, + "logps/rejected": -364.0, + "loss": 0.3149, + "rewards/accuracies": 0.8984375, + "rewards/chosen": 0.2373046875, + "rewards/margins": 7.53125, + "rewards/rejected": -7.28125, + "step": 762 + }, + { + "epoch": 0.9361963190184049, + "grad_norm": 28.119069224585317, + "learning_rate": 1.0425619834710744e-06, + "logits/chosen": 0.0986328125, + "logits/rejected": 0.0164794921875, + "logps/chosen": -400.0, + "logps/rejected": -364.0, + "loss": 0.3398, + "rewards/accuracies": 0.84375, + "rewards/chosen": 0.4375, + "rewards/margins": 7.5625, + "rewards/rejected": -7.125, + "step": 763 + }, + { + "epoch": 0.9374233128834356, + "grad_norm": 26.42805050753053, + "learning_rate": 1.0419421487603306e-06, + "logits/chosen": -0.1328125, + "logits/rejected": -0.287109375, + "logps/chosen": -330.0, + "logps/rejected": -316.0, + "loss": 0.2955, + "rewards/accuracies": 0.90625, + "rewards/chosen": -0.494140625, + "rewards/margins": 7.0, + "rewards/rejected": -7.46875, + "step": 764 + }, + { + "epoch": 0.9386503067484663, + "grad_norm": 35.659703021872964, + "learning_rate": 1.041322314049587e-06, + "logits/chosen": 0.1298828125, + "logits/rejected": 0.005828857421875, + "logps/chosen": -404.0, + "logps/rejected": -320.0, + "loss": 0.3833, + "rewards/accuracies": 0.796875, + "rewards/chosen": -0.2041015625, + "rewards/margins": 6.125, + "rewards/rejected": -6.34375, + "step": 765 + }, + { + "epoch": 0.939877300613497, + "grad_norm": 27.867245312963597, + "learning_rate": 1.040702479338843e-06, + "logits/chosen": 0.04345703125, + "logits/rejected": -0.1494140625, + "logps/chosen": -446.0, + "logps/rejected": -340.0, + "loss": 0.3188, + "rewards/accuracies": 0.8671875, + "rewards/chosen": 0.515625, + "rewards/margins": 7.5, + "rewards/rejected": -7.0, + "step": 766 + }, + { + "epoch": 0.9411042944785276, + "grad_norm": 33.0423901915957, + "learning_rate": 1.0400826446280992e-06, + "logits/chosen": 0.0390625, + "logits/rejected": -0.10498046875, + "logps/chosen": -386.0, + "logps/rejected": -338.0, + "loss": 0.355, + "rewards/accuracies": 0.796875, + "rewards/chosen": 0.1689453125, + "rewards/margins": 7.5625, + "rewards/rejected": -7.375, + "step": 767 + }, + { + "epoch": 0.9423312883435583, + "grad_norm": 31.83881144027911, + "learning_rate": 1.0394628099173554e-06, + "logits/chosen": 0.003570556640625, + "logits/rejected": -0.1611328125, + "logps/chosen": -410.0, + "logps/rejected": -384.0, + "loss": 0.3826, + "rewards/accuracies": 0.875, + "rewards/chosen": 0.6484375, + "rewards/margins": 7.59375, + "rewards/rejected": -6.96875, + "step": 768 + }, + { + "epoch": 0.943558282208589, + "grad_norm": 24.722102647279602, + "learning_rate": 1.0388429752066115e-06, + "logits/chosen": 0.00079345703125, + "logits/rejected": -0.138671875, + "logps/chosen": -378.0, + "logps/rejected": -344.0, + "loss": 0.2777, + "rewards/accuracies": 0.859375, + "rewards/chosen": -0.068359375, + "rewards/margins": 7.53125, + "rewards/rejected": -7.59375, + "step": 769 + }, + { + "epoch": 0.9447852760736196, + "grad_norm": 33.3608331180557, + "learning_rate": 1.0382231404958679e-06, + "logits/chosen": -0.021728515625, + "logits/rejected": -0.265625, + "logps/chosen": -416.0, + "logps/rejected": -372.0, + "loss": 0.321, + "rewards/accuracies": 0.8515625, + "rewards/chosen": 0.92578125, + "rewards/margins": 7.5, + "rewards/rejected": -6.59375, + "step": 770 + }, + { + "epoch": 0.9460122699386503, + "grad_norm": 36.53871156772584, + "learning_rate": 1.037603305785124e-06, + "logits/chosen": -0.00537109375, + "logits/rejected": -0.01507568359375, + "logps/chosen": -430.0, + "logps/rejected": -360.0, + "loss": 0.5217, + "rewards/accuracies": 0.78125, + "rewards/chosen": -0.6484375, + "rewards/margins": 6.21875, + "rewards/rejected": -6.875, + "step": 771 + }, + { + "epoch": 0.947239263803681, + "grad_norm": 30.200058873190507, + "learning_rate": 1.0369834710743801e-06, + "logits/chosen": 0.1328125, + "logits/rejected": -0.0869140625, + "logps/chosen": -424.0, + "logps/rejected": -362.0, + "loss": 0.3536, + "rewards/accuracies": 0.84375, + "rewards/chosen": 0.38671875, + "rewards/margins": 7.28125, + "rewards/rejected": -6.90625, + "step": 772 + }, + { + "epoch": 0.9484662576687116, + "grad_norm": 31.95788929467944, + "learning_rate": 1.0363636363636365e-06, + "logits/chosen": 0.1484375, + "logits/rejected": -0.037841796875, + "logps/chosen": -374.0, + "logps/rejected": -324.0, + "loss": 0.3406, + "rewards/accuracies": 0.828125, + "rewards/chosen": -0.65234375, + "rewards/margins": 6.5, + "rewards/rejected": -7.15625, + "step": 773 + }, + { + "epoch": 0.9496932515337423, + "grad_norm": 34.11224513726983, + "learning_rate": 1.0357438016528926e-06, + "logits/chosen": 0.1396484375, + "logits/rejected": -0.0037078857421875, + "logps/chosen": -362.0, + "logps/rejected": -326.0, + "loss": 0.4295, + "rewards/accuracies": 0.7890625, + "rewards/chosen": -1.359375, + "rewards/margins": 5.1875, + "rewards/rejected": -6.5625, + "step": 774 + }, + { + "epoch": 0.950920245398773, + "grad_norm": 31.83399815065193, + "learning_rate": 1.0351239669421488e-06, + "logits/chosen": 0.2197265625, + "logits/rejected": 0.04296875, + "logps/chosen": -418.0, + "logps/rejected": -362.0, + "loss": 0.3519, + "rewards/accuracies": 0.84375, + "rewards/chosen": -0.32421875, + "rewards/margins": 6.21875, + "rewards/rejected": -6.53125, + "step": 775 + }, + { + "epoch": 0.9521472392638037, + "grad_norm": 29.278356141500073, + "learning_rate": 1.0345041322314051e-06, + "logits/chosen": -0.01409912109375, + "logits/rejected": -0.1708984375, + "logps/chosen": -414.0, + "logps/rejected": -366.0, + "loss": 0.3978, + "rewards/accuracies": 0.84375, + "rewards/chosen": 0.0888671875, + "rewards/margins": 6.8125, + "rewards/rejected": -6.75, + "step": 776 + }, + { + "epoch": 0.9533742331288344, + "grad_norm": 33.3868265617977, + "learning_rate": 1.033884297520661e-06, + "logits/chosen": 0.0267333984375, + "logits/rejected": -0.1015625, + "logps/chosen": -408.0, + "logps/rejected": -346.0, + "loss": 0.3944, + "rewards/accuracies": 0.8046875, + "rewards/chosen": -0.58203125, + "rewards/margins": 6.71875, + "rewards/rejected": -7.3125, + "step": 777 + }, + { + "epoch": 0.9546012269938651, + "grad_norm": 30.60810394473529, + "learning_rate": 1.0332644628099174e-06, + "logits/chosen": 0.0703125, + "logits/rejected": -0.130859375, + "logps/chosen": -400.0, + "logps/rejected": -328.0, + "loss": 0.2893, + "rewards/accuracies": 0.8671875, + "rewards/chosen": -0.74609375, + "rewards/margins": 6.0, + "rewards/rejected": -6.75, + "step": 778 + }, + { + "epoch": 0.9558282208588957, + "grad_norm": 31.819603506171298, + "learning_rate": 1.0326446280991735e-06, + "logits/chosen": 0.00994873046875, + "logits/rejected": -0.1533203125, + "logps/chosen": -442.0, + "logps/rejected": -362.0, + "loss": 0.331, + "rewards/accuracies": 0.8671875, + "rewards/chosen": 0.16796875, + "rewards/margins": 7.59375, + "rewards/rejected": -7.40625, + "step": 779 + }, + { + "epoch": 0.9570552147239264, + "grad_norm": 33.50608635895235, + "learning_rate": 1.0320247933884297e-06, + "logits/chosen": 0.13671875, + "logits/rejected": -0.037353515625, + "logps/chosen": -386.0, + "logps/rejected": -360.0, + "loss": 0.4421, + "rewards/accuracies": 0.7890625, + "rewards/chosen": -0.91015625, + "rewards/margins": 5.96875, + "rewards/rejected": -6.875, + "step": 780 + }, + { + "epoch": 0.9582822085889571, + "grad_norm": 27.175413708110298, + "learning_rate": 1.031404958677686e-06, + "logits/chosen": -0.0087890625, + "logits/rejected": -0.2392578125, + "logps/chosen": -426.0, + "logps/rejected": -356.0, + "loss": 0.2289, + "rewards/accuracies": 0.8828125, + "rewards/chosen": 0.9140625, + "rewards/margins": 8.75, + "rewards/rejected": -7.8125, + "step": 781 + }, + { + "epoch": 0.9595092024539877, + "grad_norm": 27.961517528382746, + "learning_rate": 1.0307851239669422e-06, + "logits/chosen": 0.0341796875, + "logits/rejected": -0.055419921875, + "logps/chosen": -370.0, + "logps/rejected": -354.0, + "loss": 0.2727, + "rewards/accuracies": 0.890625, + "rewards/chosen": 0.0152587890625, + "rewards/margins": 7.21875, + "rewards/rejected": -7.1875, + "step": 782 + }, + { + "epoch": 0.9607361963190184, + "grad_norm": 36.53707624363368, + "learning_rate": 1.0301652892561983e-06, + "logits/chosen": 0.0115966796875, + "logits/rejected": -0.12353515625, + "logps/chosen": -420.0, + "logps/rejected": -370.0, + "loss": 0.3333, + "rewards/accuracies": 0.859375, + "rewards/chosen": 0.408203125, + "rewards/margins": 7.9375, + "rewards/rejected": -7.5, + "step": 783 + }, + { + "epoch": 0.9619631901840491, + "grad_norm": 33.01812506650061, + "learning_rate": 1.0295454545454547e-06, + "logits/chosen": 0.09716796875, + "logits/rejected": 0.056640625, + "logps/chosen": -386.0, + "logps/rejected": -354.0, + "loss": 0.3887, + "rewards/accuracies": 0.828125, + "rewards/chosen": -0.69140625, + "rewards/margins": 6.4375, + "rewards/rejected": -7.125, + "step": 784 + }, + { + "epoch": 0.9631901840490797, + "grad_norm": 25.64548904648115, + "learning_rate": 1.0289256198347108e-06, + "logits/chosen": 0.126953125, + "logits/rejected": -0.0400390625, + "logps/chosen": -384.0, + "logps/rejected": -324.0, + "loss": 0.2131, + "rewards/accuracies": 0.875, + "rewards/chosen": 0.0047607421875, + "rewards/margins": 7.40625, + "rewards/rejected": -7.375, + "step": 785 + }, + { + "epoch": 0.9644171779141104, + "grad_norm": 24.085836304747, + "learning_rate": 1.028305785123967e-06, + "logits/chosen": 0.11376953125, + "logits/rejected": -0.09033203125, + "logps/chosen": -372.0, + "logps/rejected": -334.0, + "loss": 0.2197, + "rewards/accuracies": 0.8984375, + "rewards/chosen": -0.427734375, + "rewards/margins": 7.375, + "rewards/rejected": -7.8125, + "step": 786 + }, + { + "epoch": 0.9656441717791411, + "grad_norm": 32.97798807698007, + "learning_rate": 1.027685950413223e-06, + "logits/chosen": 0.0269775390625, + "logits/rejected": -0.08740234375, + "logps/chosen": -412.0, + "logps/rejected": -348.0, + "loss": 0.3407, + "rewards/accuracies": 0.859375, + "rewards/chosen": -0.271484375, + "rewards/margins": 7.65625, + "rewards/rejected": -7.9375, + "step": 787 + }, + { + "epoch": 0.9668711656441717, + "grad_norm": 24.19044914351049, + "learning_rate": 1.0270661157024792e-06, + "logits/chosen": -0.00860595703125, + "logits/rejected": -0.1787109375, + "logps/chosen": -390.0, + "logps/rejected": -366.0, + "loss": 0.2123, + "rewards/accuracies": 0.9140625, + "rewards/chosen": 0.486328125, + "rewards/margins": 8.1875, + "rewards/rejected": -7.6875, + "step": 788 + }, + { + "epoch": 0.9680981595092024, + "grad_norm": 30.061433994827276, + "learning_rate": 1.0264462809917356e-06, + "logits/chosen": 0.12060546875, + "logits/rejected": -0.06396484375, + "logps/chosen": -384.0, + "logps/rejected": -332.0, + "loss": 0.383, + "rewards/accuracies": 0.859375, + "rewards/chosen": -0.52734375, + "rewards/margins": 6.71875, + "rewards/rejected": -7.25, + "step": 789 + }, + { + "epoch": 0.9693251533742331, + "grad_norm": 27.089223909733843, + "learning_rate": 1.0258264462809917e-06, + "logits/chosen": -0.028564453125, + "logits/rejected": -0.2373046875, + "logps/chosen": -388.0, + "logps/rejected": -330.0, + "loss": 0.1804, + "rewards/accuracies": 0.8984375, + "rewards/chosen": -0.1533203125, + "rewards/margins": 7.3125, + "rewards/rejected": -7.46875, + "step": 790 + }, + { + "epoch": 0.9705521472392638, + "grad_norm": 34.35477577844613, + "learning_rate": 1.0252066115702479e-06, + "logits/chosen": -0.0303955078125, + "logits/rejected": -0.23046875, + "logps/chosen": -398.0, + "logps/rejected": -346.0, + "loss": 0.3642, + "rewards/accuracies": 0.8125, + "rewards/chosen": 0.349609375, + "rewards/margins": 8.25, + "rewards/rejected": -7.90625, + "step": 791 + }, + { + "epoch": 0.9717791411042945, + "grad_norm": 33.05641026053793, + "learning_rate": 1.0245867768595042e-06, + "logits/chosen": 0.0006256103515625, + "logits/rejected": -0.1279296875, + "logps/chosen": -438.0, + "logps/rejected": -374.0, + "loss": 0.406, + "rewards/accuracies": 0.8125, + "rewards/chosen": 0.380859375, + "rewards/margins": 6.8125, + "rewards/rejected": -6.4375, + "step": 792 + }, + { + "epoch": 0.9730061349693252, + "grad_norm": 32.69405943160363, + "learning_rate": 1.0239669421487604e-06, + "logits/chosen": 0.0284423828125, + "logits/rejected": -0.09765625, + "logps/chosen": -392.0, + "logps/rejected": -362.0, + "loss": 0.4018, + "rewards/accuracies": 0.8671875, + "rewards/chosen": -0.10546875, + "rewards/margins": 6.75, + "rewards/rejected": -6.84375, + "step": 793 + }, + { + "epoch": 0.9742331288343559, + "grad_norm": 33.74613547302303, + "learning_rate": 1.0233471074380167e-06, + "logits/chosen": 0.0830078125, + "logits/rejected": -0.043701171875, + "logps/chosen": -358.0, + "logps/rejected": -336.0, + "loss": 0.4877, + "rewards/accuracies": 0.8046875, + "rewards/chosen": -0.6015625, + "rewards/margins": 6.59375, + "rewards/rejected": -7.1875, + "step": 794 + }, + { + "epoch": 0.9754601226993865, + "grad_norm": 22.619679414305867, + "learning_rate": 1.0227272727272727e-06, + "logits/chosen": -0.1513671875, + "logits/rejected": -0.296875, + "logps/chosen": -392.0, + "logps/rejected": -372.0, + "loss": 0.281, + "rewards/accuracies": 0.875, + "rewards/chosen": 0.80859375, + "rewards/margins": 9.25, + "rewards/rejected": -8.4375, + "step": 795 + }, + { + "epoch": 0.9766871165644172, + "grad_norm": 35.85933901249144, + "learning_rate": 1.0221074380165288e-06, + "logits/chosen": -0.0260009765625, + "logits/rejected": -0.2353515625, + "logps/chosen": -410.0, + "logps/rejected": -338.0, + "loss": 0.3447, + "rewards/accuracies": 0.8828125, + "rewards/chosen": -0.78515625, + "rewards/margins": 7.25, + "rewards/rejected": -8.0625, + "step": 796 + }, + { + "epoch": 0.9779141104294479, + "grad_norm": 24.484140263550707, + "learning_rate": 1.0214876033057852e-06, + "logits/chosen": 0.059814453125, + "logits/rejected": -0.11962890625, + "logps/chosen": -404.0, + "logps/rejected": -324.0, + "loss": 0.243, + "rewards/accuracies": 0.8671875, + "rewards/chosen": -0.0966796875, + "rewards/margins": 7.15625, + "rewards/rejected": -7.25, + "step": 797 + }, + { + "epoch": 0.9791411042944785, + "grad_norm": 35.333603529709684, + "learning_rate": 1.0208677685950413e-06, + "logits/chosen": 0.00506591796875, + "logits/rejected": -0.173828125, + "logps/chosen": -430.0, + "logps/rejected": -378.0, + "loss": 0.3546, + "rewards/accuracies": 0.8203125, + "rewards/chosen": 0.1455078125, + "rewards/margins": 7.78125, + "rewards/rejected": -7.625, + "step": 798 + }, + { + "epoch": 0.9803680981595092, + "grad_norm": 32.17706869798208, + "learning_rate": 1.0202479338842976e-06, + "logits/chosen": -0.05029296875, + "logits/rejected": -0.201171875, + "logps/chosen": -400.0, + "logps/rejected": -360.0, + "loss": 0.3323, + "rewards/accuracies": 0.8203125, + "rewards/chosen": 0.4296875, + "rewards/margins": 7.71875, + "rewards/rejected": -7.28125, + "step": 799 + }, + { + "epoch": 0.9815950920245399, + "grad_norm": 32.44337864899576, + "learning_rate": 1.0196280991735538e-06, + "logits/chosen": -0.0224609375, + "logits/rejected": -0.208984375, + "logps/chosen": -386.0, + "logps/rejected": -344.0, + "loss": 0.3141, + "rewards/accuracies": 0.8828125, + "rewards/chosen": -0.0654296875, + "rewards/margins": 7.28125, + "rewards/rejected": -7.375, + "step": 800 + }, + { + "epoch": 0.9828220858895705, + "grad_norm": 37.380385495976945, + "learning_rate": 1.01900826446281e-06, + "logits/chosen": -0.024658203125, + "logits/rejected": -0.17578125, + "logps/chosen": -436.0, + "logps/rejected": -382.0, + "loss": 0.4546, + "rewards/accuracies": 0.828125, + "rewards/chosen": 0.0001010894775390625, + "rewards/margins": 6.6875, + "rewards/rejected": -6.6875, + "step": 801 + }, + { + "epoch": 0.9840490797546012, + "grad_norm": 35.755684948235334, + "learning_rate": 1.0183884297520663e-06, + "logits/chosen": -0.07421875, + "logits/rejected": -0.21484375, + "logps/chosen": -386.0, + "logps/rejected": -348.0, + "loss": 0.472, + "rewards/accuracies": 0.84375, + "rewards/chosen": -0.1552734375, + "rewards/margins": 7.0, + "rewards/rejected": -7.15625, + "step": 802 + }, + { + "epoch": 0.9852760736196319, + "grad_norm": 29.08898198089266, + "learning_rate": 1.0177685950413224e-06, + "logits/chosen": -0.1513671875, + "logits/rejected": -0.302734375, + "logps/chosen": -398.0, + "logps/rejected": -348.0, + "loss": 0.3346, + "rewards/accuracies": 0.84375, + "rewards/chosen": 0.54296875, + "rewards/margins": 7.75, + "rewards/rejected": -7.21875, + "step": 803 + }, + { + "epoch": 0.9865030674846625, + "grad_norm": 30.37947277261302, + "learning_rate": 1.0171487603305786e-06, + "logits/chosen": 0.04150390625, + "logits/rejected": -0.06689453125, + "logps/chosen": -358.0, + "logps/rejected": -346.0, + "loss": 0.3638, + "rewards/accuracies": 0.859375, + "rewards/chosen": -0.392578125, + "rewards/margins": 6.9375, + "rewards/rejected": -7.34375, + "step": 804 + }, + { + "epoch": 0.9877300613496932, + "grad_norm": 34.04488703971709, + "learning_rate": 1.0165289256198347e-06, + "logits/chosen": -0.078125, + "logits/rejected": -0.2119140625, + "logps/chosen": -434.0, + "logps/rejected": -394.0, + "loss": 0.4569, + "rewards/accuracies": 0.828125, + "rewards/chosen": 0.4296875, + "rewards/margins": 7.78125, + "rewards/rejected": -7.375, + "step": 805 + }, + { + "epoch": 0.9889570552147239, + "grad_norm": 35.48502129676359, + "learning_rate": 1.0159090909090908e-06, + "logits/chosen": 0.16015625, + "logits/rejected": 0.03564453125, + "logps/chosen": -412.0, + "logps/rejected": -384.0, + "loss": 0.4253, + "rewards/accuracies": 0.8046875, + "rewards/chosen": 0.31640625, + "rewards/margins": 7.15625, + "rewards/rejected": -6.84375, + "step": 806 + }, + { + "epoch": 0.9901840490797545, + "grad_norm": 33.61353577378301, + "learning_rate": 1.0152892561983472e-06, + "logits/chosen": 0.061767578125, + "logits/rejected": -0.0908203125, + "logps/chosen": -400.0, + "logps/rejected": -350.0, + "loss": 0.3144, + "rewards/accuracies": 0.8828125, + "rewards/chosen": 0.484375, + "rewards/margins": 8.25, + "rewards/rejected": -7.75, + "step": 807 + }, + { + "epoch": 0.9914110429447853, + "grad_norm": 31.091169744856, + "learning_rate": 1.0146694214876033e-06, + "logits/chosen": -0.083984375, + "logits/rejected": -0.0966796875, + "logps/chosen": -364.0, + "logps/rejected": -352.0, + "loss": 0.3745, + "rewards/accuracies": 0.84375, + "rewards/chosen": -0.640625, + "rewards/margins": 7.5625, + "rewards/rejected": -8.1875, + "step": 808 + }, + { + "epoch": 0.992638036809816, + "grad_norm": 34.39338760048118, + "learning_rate": 1.0140495867768595e-06, + "logits/chosen": -0.00872802734375, + "logits/rejected": -0.2314453125, + "logps/chosen": -438.0, + "logps/rejected": -360.0, + "loss": 0.319, + "rewards/accuracies": 0.859375, + "rewards/chosen": 0.8515625, + "rewards/margins": 8.5625, + "rewards/rejected": -7.71875, + "step": 809 + }, + { + "epoch": 0.9938650306748467, + "grad_norm": 26.950235054794184, + "learning_rate": 1.0134297520661158e-06, + "logits/chosen": -0.11572265625, + "logits/rejected": -0.28515625, + "logps/chosen": -446.0, + "logps/rejected": -384.0, + "loss": 0.2346, + "rewards/accuracies": 0.8828125, + "rewards/chosen": 0.44140625, + "rewards/margins": 7.9375, + "rewards/rejected": -7.5, + "step": 810 + }, + { + "epoch": 0.9950920245398773, + "grad_norm": 26.578554179921486, + "learning_rate": 1.012809917355372e-06, + "logits/chosen": -0.08203125, + "logits/rejected": -0.21484375, + "logps/chosen": -394.0, + "logps/rejected": -380.0, + "loss": 0.3194, + "rewards/accuracies": 0.84375, + "rewards/chosen": 0.44921875, + "rewards/margins": 7.46875, + "rewards/rejected": -7.03125, + "step": 811 + }, + { + "epoch": 0.996319018404908, + "grad_norm": 34.32887556980461, + "learning_rate": 1.0121900826446281e-06, + "logits/chosen": -0.05908203125, + "logits/rejected": -0.125, + "logps/chosen": -364.0, + "logps/rejected": -340.0, + "loss": 0.3746, + "rewards/accuracies": 0.875, + "rewards/chosen": -0.09619140625, + "rewards/margins": 7.625, + "rewards/rejected": -7.71875, + "step": 812 + }, + { + "epoch": 0.9975460122699387, + "grad_norm": 25.82868308838759, + "learning_rate": 1.0115702479338843e-06, + "logits/chosen": 0.01611328125, + "logits/rejected": -0.1123046875, + "logps/chosen": -342.0, + "logps/rejected": -336.0, + "loss": 0.2067, + "rewards/accuracies": 0.8984375, + "rewards/chosen": 0.46484375, + "rewards/margins": 7.84375, + "rewards/rejected": -7.375, + "step": 813 + }, + { + "epoch": 0.9987730061349693, + "grad_norm": 35.210336167923685, + "learning_rate": 1.0109504132231404e-06, + "logits/chosen": 0.054443359375, + "logits/rejected": -0.0206298828125, + "logps/chosen": -388.0, + "logps/rejected": -370.0, + "loss": 0.3789, + "rewards/accuracies": 0.78125, + "rewards/chosen": -0.80859375, + "rewards/margins": 6.125, + "rewards/rejected": -6.90625, + "step": 814 + }, + { + "epoch": 1.0, + "grad_norm": 31.93988486773496, + "learning_rate": 1.0103305785123968e-06, + "logits/chosen": 0.1279296875, + "logits/rejected": -0.061279296875, + "logps/chosen": -388.0, + "logps/rejected": -364.0, + "loss": 0.2533, + "rewards/accuracies": 0.8991597294807434, + "rewards/chosen": -0.095703125, + "rewards/margins": 7.53125, + "rewards/rejected": -7.625, + "step": 815 + }, + { + "epoch": 1.0012269938650307, + "grad_norm": 5.842576996186414, + "learning_rate": 1.009710743801653e-06, + "logits/chosen": 0.0537109375, + "logits/rejected": -0.08154296875, + "logps/chosen": -376.0, + "logps/rejected": -332.0, + "loss": 0.034, + "rewards/accuracies": 1.0, + "rewards/chosen": 1.2890625, + "rewards/margins": 9.5, + "rewards/rejected": -8.1875, + "step": 816 + }, + { + "epoch": 1.0024539877300613, + "grad_norm": 8.73132250150435, + "learning_rate": 1.009090909090909e-06, + "logits/chosen": 0.03857421875, + "logits/rejected": -0.107421875, + "logps/chosen": -362.0, + "logps/rejected": -334.0, + "loss": 0.0486, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.7109375, + "rewards/margins": 8.8125, + "rewards/rejected": -8.125, + "step": 817 + }, + { + "epoch": 1.003680981595092, + "grad_norm": 4.472623669255691, + "learning_rate": 1.0084710743801654e-06, + "logits/chosen": -0.09326171875, + "logits/rejected": -0.1728515625, + "logps/chosen": -364.0, + "logps/rejected": -362.0, + "loss": 0.0318, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 1.125, + "rewards/margins": 9.1875, + "rewards/rejected": -8.0625, + "step": 818 + }, + { + "epoch": 1.0049079754601227, + "grad_norm": 10.555989313773281, + "learning_rate": 1.0078512396694215e-06, + "logits/chosen": 0.0576171875, + "logits/rejected": -0.12109375, + "logps/chosen": -390.0, + "logps/rejected": -366.0, + "loss": 0.043, + "rewards/accuracies": 0.9765625, + "rewards/chosen": 1.4453125, + "rewards/margins": 9.875, + "rewards/rejected": -8.4375, + "step": 819 + }, + { + "epoch": 1.0061349693251533, + "grad_norm": 7.844244829439074, + "learning_rate": 1.0072314049586777e-06, + "logits/chosen": 0.09375, + "logits/rejected": -0.02880859375, + "logps/chosen": -396.0, + "logps/rejected": -356.0, + "loss": 0.0346, + "rewards/accuracies": 1.0, + "rewards/chosen": 1.6796875, + "rewards/margins": 9.25, + "rewards/rejected": -7.5625, + "step": 820 + }, + { + "epoch": 1.007361963190184, + "grad_norm": 13.093642235831343, + "learning_rate": 1.006611570247934e-06, + "logits/chosen": -0.0054931640625, + "logits/rejected": -0.126953125, + "logps/chosen": -350.0, + "logps/rejected": -342.0, + "loss": 0.0919, + "rewards/accuracies": 0.984375, + "rewards/chosen": 1.578125, + "rewards/margins": 9.5, + "rewards/rejected": -7.90625, + "step": 821 + }, + { + "epoch": 1.0085889570552147, + "grad_norm": 11.258353605754225, + "learning_rate": 1.00599173553719e-06, + "logits/chosen": 0.00653076171875, + "logits/rejected": -0.126953125, + "logps/chosen": -404.0, + "logps/rejected": -370.0, + "loss": 0.0581, + "rewards/accuracies": 0.9609375, + "rewards/chosen": 1.6484375, + "rewards/margins": 9.0, + "rewards/rejected": -7.375, + "step": 822 + }, + { + "epoch": 1.0098159509202453, + "grad_norm": 7.2063058089923135, + "learning_rate": 1.0053719008264463e-06, + "logits/chosen": 0.03466796875, + "logits/rejected": -0.1474609375, + "logps/chosen": -364.0, + "logps/rejected": -350.0, + "loss": 0.0354, + "rewards/accuracies": 0.984375, + "rewards/chosen": 1.84375, + "rewards/margins": 9.4375, + "rewards/rejected": -7.5625, + "step": 823 + }, + { + "epoch": 1.011042944785276, + "grad_norm": 10.387457112253443, + "learning_rate": 1.0047520661157024e-06, + "logits/chosen": -0.04833984375, + "logits/rejected": -0.1904296875, + "logps/chosen": -378.0, + "logps/rejected": -362.0, + "loss": 0.0277, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 1.921875, + "rewards/margins": 8.875, + "rewards/rejected": -7.0, + "step": 824 + }, + { + "epoch": 1.0122699386503067, + "grad_norm": 6.9895041600763665, + "learning_rate": 1.0041322314049586e-06, + "logits/chosen": 0.140625, + "logits/rejected": -0.130859375, + "logps/chosen": -396.0, + "logps/rejected": -356.0, + "loss": 0.0298, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 1.5546875, + "rewards/margins": 9.625, + "rewards/rejected": -8.0625, + "step": 825 + }, + { + "epoch": 1.0134969325153373, + "grad_norm": 12.841731144722432, + "learning_rate": 1.003512396694215e-06, + "logits/chosen": -0.064453125, + "logits/rejected": -0.197265625, + "logps/chosen": -372.0, + "logps/rejected": -354.0, + "loss": 0.0693, + "rewards/accuracies": 0.9609375, + "rewards/chosen": 1.1640625, + "rewards/margins": 8.6875, + "rewards/rejected": -7.5, + "step": 826 + }, + { + "epoch": 1.014723926380368, + "grad_norm": 6.4682689307001455, + "learning_rate": 1.002892561983471e-06, + "logits/chosen": 0.01708984375, + "logits/rejected": -0.091796875, + "logps/chosen": -344.0, + "logps/rejected": -338.0, + "loss": 0.0373, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.65625, + "rewards/margins": 8.4375, + "rewards/rejected": -7.75, + "step": 827 + }, + { + "epoch": 1.0159509202453987, + "grad_norm": 4.914698955857516, + "learning_rate": 1.0022727272727274e-06, + "logits/chosen": 0.15625, + "logits/rejected": -0.036865234375, + "logps/chosen": -382.0, + "logps/rejected": -330.0, + "loss": 0.0321, + "rewards/accuracies": 1.0, + "rewards/chosen": 1.15625, + "rewards/margins": 8.9375, + "rewards/rejected": -7.75, + "step": 828 + }, + { + "epoch": 1.0171779141104293, + "grad_norm": 13.706631134085173, + "learning_rate": 1.0016528925619836e-06, + "logits/chosen": -0.1396484375, + "logits/rejected": -0.1748046875, + "logps/chosen": -376.0, + "logps/rejected": -386.0, + "loss": 0.0804, + "rewards/accuracies": 0.96875, + "rewards/chosen": 1.5078125, + "rewards/margins": 9.875, + "rewards/rejected": -8.375, + "step": 829 + }, + { + "epoch": 1.01840490797546, + "grad_norm": 4.7042664986817755, + "learning_rate": 1.0010330578512397e-06, + "logits/chosen": -0.059814453125, + "logits/rejected": -0.25390625, + "logps/chosen": -410.0, + "logps/rejected": -384.0, + "loss": 0.0274, + "rewards/accuracies": 1.0, + "rewards/chosen": 2.03125, + "rewards/margins": 11.125, + "rewards/rejected": -9.125, + "step": 830 + }, + { + "epoch": 1.019631901840491, + "grad_norm": 10.294544241161669, + "learning_rate": 1.0004132231404959e-06, + "logits/chosen": -0.0341796875, + "logits/rejected": -0.0810546875, + "logps/chosen": -404.0, + "logps/rejected": -404.0, + "loss": 0.0445, + "rewards/accuracies": 0.984375, + "rewards/chosen": 1.7109375, + "rewards/margins": 11.1875, + "rewards/rejected": -9.4375, + "step": 831 + }, + { + "epoch": 1.0208588957055216, + "grad_norm": 11.737112567032785, + "learning_rate": 9.99793388429752e-07, + "logits/chosen": -0.037353515625, + "logits/rejected": -0.2412109375, + "logps/chosen": -406.0, + "logps/rejected": -388.0, + "loss": 0.0521, + "rewards/accuracies": 0.9765625, + "rewards/chosen": 1.15625, + "rewards/margins": 10.0, + "rewards/rejected": -8.875, + "step": 832 + }, + { + "epoch": 1.0220858895705522, + "grad_norm": 22.2024732224313, + "learning_rate": 9.991735537190084e-07, + "logits/chosen": -0.111328125, + "logits/rejected": -0.26953125, + "logps/chosen": -450.0, + "logps/rejected": -410.0, + "loss": 0.1041, + "rewards/accuracies": 0.9765625, + "rewards/chosen": 1.09375, + "rewards/margins": 10.5, + "rewards/rejected": -9.4375, + "step": 833 + }, + { + "epoch": 1.023312883435583, + "grad_norm": 10.90677573357752, + "learning_rate": 9.985537190082645e-07, + "logits/chosen": 0.024658203125, + "logits/rejected": 0.00970458984375, + "logps/chosen": -356.0, + "logps/rejected": -366.0, + "loss": 0.0368, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.1435546875, + "rewards/margins": 9.0, + "rewards/rejected": -8.875, + "step": 834 + }, + { + "epoch": 1.0245398773006136, + "grad_norm": 12.900773925818388, + "learning_rate": 9.979338842975206e-07, + "logits/chosen": 0.0712890625, + "logits/rejected": -0.14453125, + "logps/chosen": -414.0, + "logps/rejected": -404.0, + "loss": 0.0634, + "rewards/accuracies": 0.9765625, + "rewards/chosen": 0.859375, + "rewards/margins": 10.1875, + "rewards/rejected": -9.3125, + "step": 835 + }, + { + "epoch": 1.0257668711656442, + "grad_norm": 16.38159113775479, + "learning_rate": 9.97314049586777e-07, + "logits/chosen": -0.0419921875, + "logits/rejected": -0.123046875, + "logps/chosen": -376.0, + "logps/rejected": -378.0, + "loss": 0.0545, + "rewards/accuracies": 0.9765625, + "rewards/chosen": 0.703125, + "rewards/margins": 10.4375, + "rewards/rejected": -9.75, + "step": 836 + }, + { + "epoch": 1.026993865030675, + "grad_norm": 5.759934115077334, + "learning_rate": 9.966942148760331e-07, + "logits/chosen": 0.12353515625, + "logits/rejected": -0.13671875, + "logps/chosen": -444.0, + "logps/rejected": -408.0, + "loss": 0.0291, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 1.15625, + "rewards/margins": 10.5625, + "rewards/rejected": -9.4375, + "step": 837 + }, + { + "epoch": 1.0282208588957056, + "grad_norm": 10.475656005242545, + "learning_rate": 9.960743801652893e-07, + "logits/chosen": -0.043212890625, + "logits/rejected": -0.2265625, + "logps/chosen": -394.0, + "logps/rejected": -386.0, + "loss": 0.0452, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.30078125, + "rewards/margins": 10.4375, + "rewards/rejected": -10.125, + "step": 838 + }, + { + "epoch": 1.0294478527607362, + "grad_norm": 16.273519813105544, + "learning_rate": 9.954545454545456e-07, + "logits/chosen": 0.0142822265625, + "logits/rejected": -0.1328125, + "logps/chosen": -398.0, + "logps/rejected": -388.0, + "loss": 0.052, + "rewards/accuracies": 0.96875, + "rewards/chosen": 0.640625, + "rewards/margins": 10.25, + "rewards/rejected": -9.625, + "step": 839 + }, + { + "epoch": 1.030674846625767, + "grad_norm": 15.221579329063106, + "learning_rate": 9.948347107438016e-07, + "logits/chosen": -0.0167236328125, + "logits/rejected": -0.2470703125, + "logps/chosen": -446.0, + "logps/rejected": -422.0, + "loss": 0.0552, + "rewards/accuracies": 0.96875, + "rewards/chosen": 0.91015625, + "rewards/margins": 11.875, + "rewards/rejected": -11.0, + "step": 840 + }, + { + "epoch": 1.0319018404907976, + "grad_norm": 8.194954853669916, + "learning_rate": 9.94214876033058e-07, + "logits/chosen": 0.054443359375, + "logits/rejected": -0.09716796875, + "logps/chosen": -388.0, + "logps/rejected": -396.0, + "loss": 0.042, + "rewards/accuracies": 0.984375, + "rewards/chosen": -0.0072021484375, + "rewards/margins": 10.125, + "rewards/rejected": -10.125, + "step": 841 + }, + { + "epoch": 1.0331288343558283, + "grad_norm": 6.243363966929884, + "learning_rate": 9.93595041322314e-07, + "logits/chosen": -0.03955078125, + "logits/rejected": -0.16015625, + "logps/chosen": -456.0, + "logps/rejected": -364.0, + "loss": 0.0377, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.39453125, + "rewards/margins": 10.0625, + "rewards/rejected": -9.625, + "step": 842 + }, + { + "epoch": 1.034355828220859, + "grad_norm": 13.025481965324344, + "learning_rate": 9.929752066115702e-07, + "logits/chosen": -0.03955078125, + "logits/rejected": -0.1533203125, + "logps/chosen": -460.0, + "logps/rejected": -416.0, + "loss": 0.0486, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.5390625, + "rewards/margins": 10.6875, + "rewards/rejected": -10.125, + "step": 843 + }, + { + "epoch": 1.0355828220858896, + "grad_norm": 9.843030269655861, + "learning_rate": 9.923553719008265e-07, + "logits/chosen": 0.0703125, + "logits/rejected": -0.15625, + "logps/chosen": -434.0, + "logps/rejected": -380.0, + "loss": 0.0284, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 1.1953125, + "rewards/margins": 11.4375, + "rewards/rejected": -10.25, + "step": 844 + }, + { + "epoch": 1.0368098159509203, + "grad_norm": 6.820305865616003, + "learning_rate": 9.917355371900827e-07, + "logits/chosen": -0.00494384765625, + "logits/rejected": -0.11767578125, + "logps/chosen": -374.0, + "logps/rejected": -352.0, + "loss": 0.0243, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.41015625, + "rewards/margins": 9.1875, + "rewards/rejected": -8.75, + "step": 845 + }, + { + "epoch": 1.038036809815951, + "grad_norm": 10.389123582255607, + "learning_rate": 9.911157024793388e-07, + "logits/chosen": 0.1025390625, + "logits/rejected": -0.11181640625, + "logps/chosen": -422.0, + "logps/rejected": -356.0, + "loss": 0.0361, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 1.109375, + "rewards/margins": 11.4375, + "rewards/rejected": -10.3125, + "step": 846 + }, + { + "epoch": 1.0392638036809816, + "grad_norm": 6.3892793464793165, + "learning_rate": 9.904958677685952e-07, + "logits/chosen": 0.07666015625, + "logits/rejected": -0.0830078125, + "logps/chosen": -374.0, + "logps/rejected": -366.0, + "loss": 0.0319, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.63671875, + "rewards/margins": 10.0625, + "rewards/rejected": -9.375, + "step": 847 + }, + { + "epoch": 1.0404907975460123, + "grad_norm": 6.1623878145786435, + "learning_rate": 9.898760330578511e-07, + "logits/chosen": -0.03955078125, + "logits/rejected": -0.1357421875, + "logps/chosen": -354.0, + "logps/rejected": -362.0, + "loss": 0.0272, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.205078125, + "rewards/margins": 10.625, + "rewards/rejected": -10.4375, + "step": 848 + }, + { + "epoch": 1.041717791411043, + "grad_norm": 9.98252857798321, + "learning_rate": 9.892561983471075e-07, + "logits/chosen": 0.01361083984375, + "logits/rejected": -0.05859375, + "logps/chosen": -384.0, + "logps/rejected": -418.0, + "loss": 0.044, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.06396484375, + "rewards/margins": 10.1875, + "rewards/rejected": -10.125, + "step": 849 + }, + { + "epoch": 1.0429447852760736, + "grad_norm": 6.152234316016775, + "learning_rate": 9.886363636363636e-07, + "logits/chosen": -0.0224609375, + "logits/rejected": -0.26171875, + "logps/chosen": -374.0, + "logps/rejected": -368.0, + "loss": 0.0233, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.7734375, + "rewards/margins": 10.875, + "rewards/rejected": -10.125, + "step": 850 + }, + { + "epoch": 1.0441717791411043, + "grad_norm": 5.157350169674604, + "learning_rate": 9.880165289256197e-07, + "logits/chosen": -0.07666015625, + "logits/rejected": -0.275390625, + "logps/chosen": -382.0, + "logps/rejected": -394.0, + "loss": 0.016, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.6875, + "rewards/margins": 11.0, + "rewards/rejected": -10.3125, + "step": 851 + }, + { + "epoch": 1.045398773006135, + "grad_norm": 4.950192991710521, + "learning_rate": 9.87396694214876e-07, + "logits/chosen": -0.138671875, + "logits/rejected": -0.287109375, + "logps/chosen": -432.0, + "logps/rejected": -418.0, + "loss": 0.0193, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.90234375, + "rewards/margins": 10.875, + "rewards/rejected": -10.0, + "step": 852 + }, + { + "epoch": 1.0466257668711656, + "grad_norm": 6.8291284978183375, + "learning_rate": 9.867768595041322e-07, + "logits/chosen": -0.05615234375, + "logits/rejected": -0.15625, + "logps/chosen": -416.0, + "logps/rejected": -372.0, + "loss": 0.0293, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.71875, + "rewards/margins": 10.1875, + "rewards/rejected": -9.4375, + "step": 853 + }, + { + "epoch": 1.0478527607361963, + "grad_norm": 21.523576031133718, + "learning_rate": 9.861570247933884e-07, + "logits/chosen": 0.037109375, + "logits/rejected": -0.09814453125, + "logps/chosen": -396.0, + "logps/rejected": -388.0, + "loss": 0.128, + "rewards/accuracies": 0.9609375, + "rewards/chosen": 0.2265625, + "rewards/margins": 9.25, + "rewards/rejected": -9.0625, + "step": 854 + }, + { + "epoch": 1.049079754601227, + "grad_norm": 7.615435271592111, + "learning_rate": 9.855371900826447e-07, + "logits/chosen": 0.0947265625, + "logits/rejected": -0.045654296875, + "logps/chosen": -398.0, + "logps/rejected": -412.0, + "loss": 0.0403, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 1.3125, + "rewards/margins": 10.8125, + "rewards/rejected": -9.5, + "step": 855 + }, + { + "epoch": 1.0503067484662576, + "grad_norm": 13.390214785568759, + "learning_rate": 9.849173553719009e-07, + "logits/chosen": -0.020751953125, + "logits/rejected": -0.2021484375, + "logps/chosen": -408.0, + "logps/rejected": -364.0, + "loss": 0.0522, + "rewards/accuracies": 0.96875, + "rewards/chosen": 0.89453125, + "rewards/margins": 10.3125, + "rewards/rejected": -9.375, + "step": 856 + }, + { + "epoch": 1.0515337423312883, + "grad_norm": 7.721662927396619, + "learning_rate": 9.84297520661157e-07, + "logits/chosen": -0.1396484375, + "logits/rejected": -0.31640625, + "logps/chosen": -404.0, + "logps/rejected": -382.0, + "loss": 0.0336, + "rewards/accuracies": 0.9765625, + "rewards/chosen": 0.98828125, + "rewards/margins": 10.4375, + "rewards/rejected": -9.5, + "step": 857 + }, + { + "epoch": 1.052760736196319, + "grad_norm": 5.024633893121141, + "learning_rate": 9.836776859504132e-07, + "logits/chosen": -0.008056640625, + "logits/rejected": -0.06884765625, + "logps/chosen": -322.0, + "logps/rejected": -328.0, + "loss": 0.0242, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.2109375, + "rewards/margins": 10.375, + "rewards/rejected": -10.125, + "step": 858 + }, + { + "epoch": 1.0539877300613496, + "grad_norm": 11.331768445922425, + "learning_rate": 9.830578512396693e-07, + "logits/chosen": -0.05029296875, + "logits/rejected": -0.267578125, + "logps/chosen": -400.0, + "logps/rejected": -352.0, + "loss": 0.0499, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.263671875, + "rewards/margins": 10.1875, + "rewards/rejected": -9.9375, + "step": 859 + }, + { + "epoch": 1.0552147239263803, + "grad_norm": 11.938945819408568, + "learning_rate": 9.824380165289257e-07, + "logits/chosen": 0.042724609375, + "logits/rejected": -0.16796875, + "logps/chosen": -416.0, + "logps/rejected": -384.0, + "loss": 0.0571, + "rewards/accuracies": 0.984375, + "rewards/chosen": -0.267578125, + "rewards/margins": 9.125, + "rewards/rejected": -9.375, + "step": 860 + }, + { + "epoch": 1.056441717791411, + "grad_norm": 10.622500605691952, + "learning_rate": 9.818181818181818e-07, + "logits/chosen": -0.00323486328125, + "logits/rejected": -0.1953125, + "logps/chosen": -404.0, + "logps/rejected": -358.0, + "loss": 0.0358, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.65625, + "rewards/margins": 10.25, + "rewards/rejected": -9.5625, + "step": 861 + }, + { + "epoch": 1.0576687116564418, + "grad_norm": 13.987502566864515, + "learning_rate": 9.811983471074381e-07, + "logits/chosen": -0.123046875, + "logits/rejected": -0.2021484375, + "logps/chosen": -386.0, + "logps/rejected": -378.0, + "loss": 0.053, + "rewards/accuracies": 0.9765625, + "rewards/chosen": -0.5078125, + "rewards/margins": 10.0625, + "rewards/rejected": -10.5625, + "step": 862 + }, + { + "epoch": 1.0588957055214725, + "grad_norm": 5.222070055196097, + "learning_rate": 9.805785123966943e-07, + "logits/chosen": -0.12890625, + "logits/rejected": -0.345703125, + "logps/chosen": -428.0, + "logps/rejected": -404.0, + "loss": 0.0159, + "rewards/accuracies": 1.0, + "rewards/chosen": 1.640625, + "rewards/margins": 11.875, + "rewards/rejected": -10.25, + "step": 863 + }, + { + "epoch": 1.0601226993865032, + "grad_norm": 4.4447902099136645, + "learning_rate": 9.799586776859504e-07, + "logits/chosen": -0.051513671875, + "logits/rejected": -0.248046875, + "logps/chosen": -382.0, + "logps/rejected": -390.0, + "loss": 0.022, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.193359375, + "rewards/margins": 11.0, + "rewards/rejected": -10.75, + "step": 864 + }, + { + "epoch": 1.0613496932515338, + "grad_norm": 6.7260760377615, + "learning_rate": 9.793388429752068e-07, + "logits/chosen": -0.2177734375, + "logits/rejected": -0.30859375, + "logps/chosen": -432.0, + "logps/rejected": -370.0, + "loss": 0.0262, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.1982421875, + "rewards/margins": 10.6875, + "rewards/rejected": -10.5, + "step": 865 + }, + { + "epoch": 1.0625766871165645, + "grad_norm": 7.412871849821934, + "learning_rate": 9.787190082644627e-07, + "logits/chosen": -0.166015625, + "logits/rejected": -0.3515625, + "logps/chosen": -356.0, + "logps/rejected": -364.0, + "loss": 0.0325, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.455078125, + "rewards/margins": 10.8125, + "rewards/rejected": -10.375, + "step": 866 + }, + { + "epoch": 1.0638036809815952, + "grad_norm": 12.621207505635246, + "learning_rate": 9.78099173553719e-07, + "logits/chosen": -0.1025390625, + "logits/rejected": -0.251953125, + "logps/chosen": -434.0, + "logps/rejected": -378.0, + "loss": 0.0623, + "rewards/accuracies": 0.9765625, + "rewards/chosen": 0.1123046875, + "rewards/margins": 9.9375, + "rewards/rejected": -9.8125, + "step": 867 + }, + { + "epoch": 1.0650306748466258, + "grad_norm": 6.401402652027187, + "learning_rate": 9.774793388429752e-07, + "logits/chosen": 0.1396484375, + "logits/rejected": -0.06396484375, + "logps/chosen": -358.0, + "logps/rejected": -326.0, + "loss": 0.0388, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.042236328125, + "rewards/margins": 9.1875, + "rewards/rejected": -9.125, + "step": 868 + }, + { + "epoch": 1.0662576687116565, + "grad_norm": 8.548172524646416, + "learning_rate": 9.768595041322313e-07, + "logits/chosen": -0.10205078125, + "logits/rejected": -0.2158203125, + "logps/chosen": -410.0, + "logps/rejected": -394.0, + "loss": 0.0415, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.5, + "rewards/margins": 10.8125, + "rewards/rejected": -10.3125, + "step": 869 + }, + { + "epoch": 1.0674846625766872, + "grad_norm": 10.984226790709622, + "learning_rate": 9.762396694214877e-07, + "logits/chosen": 0.06396484375, + "logits/rejected": -0.11474609375, + "logps/chosen": -390.0, + "logps/rejected": -374.0, + "loss": 0.0438, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.6015625, + "rewards/margins": 10.625, + "rewards/rejected": -10.0, + "step": 870 + }, + { + "epoch": 1.0687116564417178, + "grad_norm": 17.83709695663358, + "learning_rate": 9.756198347107438e-07, + "logits/chosen": 0.0732421875, + "logits/rejected": -0.09130859375, + "logps/chosen": -410.0, + "logps/rejected": -364.0, + "loss": 0.0866, + "rewards/accuracies": 0.96875, + "rewards/chosen": 0.1845703125, + "rewards/margins": 9.375, + "rewards/rejected": -9.25, + "step": 871 + }, + { + "epoch": 1.0699386503067485, + "grad_norm": 6.167620411597598, + "learning_rate": 9.75e-07, + "logits/chosen": 0.03466796875, + "logits/rejected": -0.0859375, + "logps/chosen": -418.0, + "logps/rejected": -338.0, + "loss": 0.0272, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.53125, + "rewards/margins": 10.1875, + "rewards/rejected": -9.625, + "step": 872 + }, + { + "epoch": 1.0711656441717792, + "grad_norm": 7.7202347168828265, + "learning_rate": 9.743801652892563e-07, + "logits/chosen": -0.056640625, + "logits/rejected": -0.21484375, + "logps/chosen": -380.0, + "logps/rejected": -394.0, + "loss": 0.0402, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.6484375, + "rewards/margins": 11.4375, + "rewards/rejected": -10.8125, + "step": 873 + }, + { + "epoch": 1.0723926380368098, + "grad_norm": 20.769443880639034, + "learning_rate": 9.737603305785125e-07, + "logits/chosen": -0.08837890625, + "logits/rejected": -0.2158203125, + "logps/chosen": -436.0, + "logps/rejected": -412.0, + "loss": 0.0993, + "rewards/accuracies": 0.9609375, + "rewards/chosen": 1.1328125, + "rewards/margins": 11.3125, + "rewards/rejected": -10.1875, + "step": 874 + }, + { + "epoch": 1.0736196319018405, + "grad_norm": 4.947185378698119, + "learning_rate": 9.731404958677686e-07, + "logits/chosen": 0.03564453125, + "logits/rejected": -0.11572265625, + "logps/chosen": -354.0, + "logps/rejected": -388.0, + "loss": 0.0241, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.474609375, + "rewards/margins": 10.375, + "rewards/rejected": -9.9375, + "step": 875 + }, + { + "epoch": 1.0748466257668712, + "grad_norm": 5.995360821265753, + "learning_rate": 9.725206611570248e-07, + "logits/chosen": -0.09814453125, + "logits/rejected": -0.279296875, + "logps/chosen": -420.0, + "logps/rejected": -418.0, + "loss": 0.0269, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.443359375, + "rewards/margins": 11.375, + "rewards/rejected": -10.9375, + "step": 876 + }, + { + "epoch": 1.0760736196319018, + "grad_norm": 2.795180251678899, + "learning_rate": 9.71900826446281e-07, + "logits/chosen": 0.0986328125, + "logits/rejected": -0.050048828125, + "logps/chosen": -408.0, + "logps/rejected": -368.0, + "loss": 0.0117, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.0380859375, + "rewards/margins": 9.9375, + "rewards/rejected": -9.9375, + "step": 877 + }, + { + "epoch": 1.0773006134969325, + "grad_norm": 11.795724961832926, + "learning_rate": 9.712809917355373e-07, + "logits/chosen": -0.08056640625, + "logits/rejected": -0.287109375, + "logps/chosen": -384.0, + "logps/rejected": -380.0, + "loss": 0.0589, + "rewards/accuracies": 0.9609375, + "rewards/chosen": -0.21484375, + "rewards/margins": 10.0, + "rewards/rejected": -10.25, + "step": 878 + }, + { + "epoch": 1.0785276073619632, + "grad_norm": 4.020220366948393, + "learning_rate": 9.706611570247934e-07, + "logits/chosen": 0.040283203125, + "logits/rejected": -0.2216796875, + "logps/chosen": -414.0, + "logps/rejected": -362.0, + "loss": 0.0148, + "rewards/accuracies": 1.0, + "rewards/chosen": 1.09375, + "rewards/margins": 11.0625, + "rewards/rejected": -10.0, + "step": 879 + }, + { + "epoch": 1.0797546012269938, + "grad_norm": 7.945458923967855, + "learning_rate": 9.700413223140495e-07, + "logits/chosen": 0.037109375, + "logits/rejected": -0.205078125, + "logps/chosen": -354.0, + "logps/rejected": -354.0, + "loss": 0.0369, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.421875, + "rewards/margins": 10.625, + "rewards/rejected": -11.0625, + "step": 880 + }, + { + "epoch": 1.0809815950920245, + "grad_norm": 7.655504936651129, + "learning_rate": 9.694214876033059e-07, + "logits/chosen": 0.08447265625, + "logits/rejected": -0.055908203125, + "logps/chosen": -350.0, + "logps/rejected": -320.0, + "loss": 0.0397, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.267578125, + "rewards/margins": 8.8125, + "rewards/rejected": -8.5625, + "step": 881 + }, + { + "epoch": 1.0822085889570552, + "grad_norm": 17.14199925290297, + "learning_rate": 9.68801652892562e-07, + "logits/chosen": -0.06787109375, + "logits/rejected": -0.205078125, + "logps/chosen": -402.0, + "logps/rejected": -398.0, + "loss": 0.092, + "rewards/accuracies": 0.9765625, + "rewards/chosen": 1.296875, + "rewards/margins": 10.625, + "rewards/rejected": -9.375, + "step": 882 + }, + { + "epoch": 1.0834355828220859, + "grad_norm": 6.074505313460856, + "learning_rate": 9.681818181818184e-07, + "logits/chosen": -0.11181640625, + "logits/rejected": -0.232421875, + "logps/chosen": -388.0, + "logps/rejected": -364.0, + "loss": 0.0179, + "rewards/accuracies": 1.0, + "rewards/chosen": 1.4375, + "rewards/margins": 9.9375, + "rewards/rejected": -8.5, + "step": 883 + }, + { + "epoch": 1.0846625766871165, + "grad_norm": 5.184873219214592, + "learning_rate": 9.675619834710743e-07, + "logits/chosen": -0.1015625, + "logits/rejected": -0.263671875, + "logps/chosen": -428.0, + "logps/rejected": -410.0, + "loss": 0.015, + "rewards/accuracies": 1.0, + "rewards/chosen": 1.8515625, + "rewards/margins": 11.625, + "rewards/rejected": -9.75, + "step": 884 + }, + { + "epoch": 1.0858895705521472, + "grad_norm": 5.141932376248221, + "learning_rate": 9.669421487603305e-07, + "logits/chosen": -0.0250244140625, + "logits/rejected": -0.16015625, + "logps/chosen": -392.0, + "logps/rejected": -354.0, + "loss": 0.0325, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 1.5, + "rewards/margins": 10.25, + "rewards/rejected": -8.75, + "step": 885 + }, + { + "epoch": 1.0871165644171779, + "grad_norm": 10.51661040179474, + "learning_rate": 9.663223140495868e-07, + "logits/chosen": 0.07666015625, + "logits/rejected": -0.126953125, + "logps/chosen": -384.0, + "logps/rejected": -384.0, + "loss": 0.0416, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.75390625, + "rewards/margins": 9.3125, + "rewards/rejected": -8.5625, + "step": 886 + }, + { + "epoch": 1.0883435582822085, + "grad_norm": 10.214979385844872, + "learning_rate": 9.65702479338843e-07, + "logits/chosen": -0.11376953125, + "logits/rejected": -0.294921875, + "logps/chosen": -402.0, + "logps/rejected": -390.0, + "loss": 0.0423, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.80859375, + "rewards/margins": 10.9375, + "rewards/rejected": -10.125, + "step": 887 + }, + { + "epoch": 1.0895705521472392, + "grad_norm": 12.886011610183632, + "learning_rate": 9.650826446280993e-07, + "logits/chosen": -0.08544921875, + "logits/rejected": -0.357421875, + "logps/chosen": -400.0, + "logps/rejected": -380.0, + "loss": 0.0493, + "rewards/accuracies": 0.9765625, + "rewards/chosen": 1.0859375, + "rewards/margins": 10.5, + "rewards/rejected": -9.4375, + "step": 888 + }, + { + "epoch": 1.0907975460122699, + "grad_norm": 21.84348705309875, + "learning_rate": 9.644628099173554e-07, + "logits/chosen": -0.044677734375, + "logits/rejected": -0.1611328125, + "logps/chosen": -454.0, + "logps/rejected": -420.0, + "loss": 0.0864, + "rewards/accuracies": 0.9609375, + "rewards/chosen": 1.3828125, + "rewards/margins": 10.875, + "rewards/rejected": -9.5, + "step": 889 + }, + { + "epoch": 1.0920245398773005, + "grad_norm": 7.008527769291815, + "learning_rate": 9.638429752066116e-07, + "logits/chosen": -0.0947265625, + "logits/rejected": -0.2158203125, + "logps/chosen": -366.0, + "logps/rejected": -358.0, + "loss": 0.0303, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 1.140625, + "rewards/margins": 10.3125, + "rewards/rejected": -9.125, + "step": 890 + }, + { + "epoch": 1.0932515337423312, + "grad_norm": 9.38763327259464, + "learning_rate": 9.63223140495868e-07, + "logits/chosen": -0.1416015625, + "logits/rejected": -0.28125, + "logps/chosen": -368.0, + "logps/rejected": -386.0, + "loss": 0.0416, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.435546875, + "rewards/margins": 10.5625, + "rewards/rejected": -10.125, + "step": 891 + }, + { + "epoch": 1.0944785276073619, + "grad_norm": 7.190745784431904, + "learning_rate": 9.62603305785124e-07, + "logits/chosen": -0.04248046875, + "logits/rejected": -0.26953125, + "logps/chosen": -358.0, + "logps/rejected": -364.0, + "loss": 0.0387, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.2099609375, + "rewards/margins": 10.5625, + "rewards/rejected": -10.375, + "step": 892 + }, + { + "epoch": 1.0957055214723925, + "grad_norm": 10.096733031161396, + "learning_rate": 9.619834710743802e-07, + "logits/chosen": -0.0869140625, + "logits/rejected": -0.09326171875, + "logps/chosen": -320.0, + "logps/rejected": -332.0, + "loss": 0.0549, + "rewards/accuracies": 0.953125, + "rewards/chosen": 0.54296875, + "rewards/margins": 9.8125, + "rewards/rejected": -9.3125, + "step": 893 + }, + { + "epoch": 1.0969325153374232, + "grad_norm": 15.853564017698188, + "learning_rate": 9.613636363636364e-07, + "logits/chosen": -0.06689453125, + "logits/rejected": -0.28125, + "logps/chosen": -442.0, + "logps/rejected": -416.0, + "loss": 0.0575, + "rewards/accuracies": 0.984375, + "rewards/chosen": 1.1640625, + "rewards/margins": 11.5, + "rewards/rejected": -10.3125, + "step": 894 + }, + { + "epoch": 1.098159509202454, + "grad_norm": 10.254886625086469, + "learning_rate": 9.607438016528925e-07, + "logits/chosen": -0.045166015625, + "logits/rejected": -0.20703125, + "logps/chosen": -388.0, + "logps/rejected": -376.0, + "loss": 0.0352, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 1.515625, + "rewards/margins": 11.0, + "rewards/rejected": -9.4375, + "step": 895 + }, + { + "epoch": 1.0993865030674848, + "grad_norm": 7.1661826290113755, + "learning_rate": 9.601239669421489e-07, + "logits/chosen": -0.09033203125, + "logits/rejected": -0.248046875, + "logps/chosen": -370.0, + "logps/rejected": -374.0, + "loss": 0.0272, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 1.0546875, + "rewards/margins": 10.5, + "rewards/rejected": -9.4375, + "step": 896 + }, + { + "epoch": 1.1006134969325154, + "grad_norm": 13.118106354556259, + "learning_rate": 9.59504132231405e-07, + "logits/chosen": -0.12060546875, + "logits/rejected": -0.357421875, + "logps/chosen": -442.0, + "logps/rejected": -392.0, + "loss": 0.0388, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.90234375, + "rewards/margins": 11.25, + "rewards/rejected": -10.375, + "step": 897 + }, + { + "epoch": 1.101840490797546, + "grad_norm": 7.692840365219393, + "learning_rate": 9.588842975206611e-07, + "logits/chosen": -0.12353515625, + "logits/rejected": -0.33203125, + "logps/chosen": -420.0, + "logps/rejected": -388.0, + "loss": 0.0326, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 1.0859375, + "rewards/margins": 11.4375, + "rewards/rejected": -10.375, + "step": 898 + }, + { + "epoch": 1.1030674846625768, + "grad_norm": 4.182192778148996, + "learning_rate": 9.582644628099175e-07, + "logits/chosen": -0.053466796875, + "logits/rejected": -0.1806640625, + "logps/chosen": -340.0, + "logps/rejected": -348.0, + "loss": 0.0244, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.06298828125, + "rewards/margins": 10.25, + "rewards/rejected": -10.3125, + "step": 899 + }, + { + "epoch": 1.1042944785276074, + "grad_norm": 17.13756577176759, + "learning_rate": 9.576446280991736e-07, + "logits/chosen": 0.0125732421875, + "logits/rejected": -0.1748046875, + "logps/chosen": -404.0, + "logps/rejected": -398.0, + "loss": 0.0925, + "rewards/accuracies": 0.953125, + "rewards/chosen": 0.447265625, + "rewards/margins": 10.1875, + "rewards/rejected": -9.75, + "step": 900 + }, + { + "epoch": 1.105521472392638, + "grad_norm": 6.771958107284894, + "learning_rate": 9.570247933884298e-07, + "logits/chosen": -0.1455078125, + "logits/rejected": -0.36328125, + "logps/chosen": -490.0, + "logps/rejected": -428.0, + "loss": 0.0284, + "rewards/accuracies": 0.984375, + "rewards/chosen": 1.890625, + "rewards/margins": 13.125, + "rewards/rejected": -11.25, + "step": 901 + }, + { + "epoch": 1.1067484662576688, + "grad_norm": 7.061170750215255, + "learning_rate": 9.56404958677686e-07, + "logits/chosen": -0.18359375, + "logits/rejected": -0.341796875, + "logps/chosen": -376.0, + "logps/rejected": -384.0, + "loss": 0.0349, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.5390625, + "rewards/margins": 11.0625, + "rewards/rejected": -10.5625, + "step": 902 + }, + { + "epoch": 1.1079754601226994, + "grad_norm": 8.245692452787445, + "learning_rate": 9.55785123966942e-07, + "logits/chosen": -0.08642578125, + "logits/rejected": -0.30078125, + "logps/chosen": -364.0, + "logps/rejected": -370.0, + "loss": 0.0341, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.75390625, + "rewards/margins": 11.1875, + "rewards/rejected": -10.4375, + "step": 903 + }, + { + "epoch": 1.10920245398773, + "grad_norm": 7.285502821165709, + "learning_rate": 9.551652892561984e-07, + "logits/chosen": -0.09228515625, + "logits/rejected": -0.22265625, + "logps/chosen": -378.0, + "logps/rejected": -416.0, + "loss": 0.0233, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 1.109375, + "rewards/margins": 11.8125, + "rewards/rejected": -10.6875, + "step": 904 + }, + { + "epoch": 1.1104294478527608, + "grad_norm": 11.766334881942175, + "learning_rate": 9.545454545454546e-07, + "logits/chosen": -0.26171875, + "logits/rejected": -0.458984375, + "logps/chosen": -392.0, + "logps/rejected": -392.0, + "loss": 0.0414, + "rewards/accuracies": 0.984375, + "rewards/chosen": 1.453125, + "rewards/margins": 12.0625, + "rewards/rejected": -10.5625, + "step": 905 + }, + { + "epoch": 1.1116564417177914, + "grad_norm": 11.885836553849193, + "learning_rate": 9.539256198347107e-07, + "logits/chosen": -0.1806640625, + "logits/rejected": -0.26171875, + "logps/chosen": -342.0, + "logps/rejected": -382.0, + "loss": 0.0611, + "rewards/accuracies": 0.984375, + "rewards/chosen": -0.337890625, + "rewards/margins": 9.5, + "rewards/rejected": -9.875, + "step": 906 + }, + { + "epoch": 1.112883435582822, + "grad_norm": 8.496822948479098, + "learning_rate": 9.533057851239669e-07, + "logits/chosen": -0.11376953125, + "logits/rejected": -0.388671875, + "logps/chosen": -386.0, + "logps/rejected": -366.0, + "loss": 0.0392, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.4921875, + "rewards/margins": 11.6875, + "rewards/rejected": -11.1875, + "step": 907 + }, + { + "epoch": 1.1141104294478528, + "grad_norm": 4.416563202380208, + "learning_rate": 9.526859504132232e-07, + "logits/chosen": -0.16796875, + "logits/rejected": -0.23828125, + "logps/chosen": -392.0, + "logps/rejected": -380.0, + "loss": 0.0153, + "rewards/accuracies": 1.0, + "rewards/chosen": 1.4453125, + "rewards/margins": 11.5, + "rewards/rejected": -10.0, + "step": 908 + }, + { + "epoch": 1.1153374233128834, + "grad_norm": 12.914705090980133, + "learning_rate": 9.520661157024794e-07, + "logits/chosen": -0.014892578125, + "logits/rejected": -0.1123046875, + "logps/chosen": -418.0, + "logps/rejected": -524.0, + "loss": 0.0577, + "rewards/accuracies": 0.9765625, + "rewards/chosen": 0.7265625, + "rewards/margins": 19.375, + "rewards/rejected": -18.625, + "step": 909 + }, + { + "epoch": 1.116564417177914, + "grad_norm": 11.193369005763312, + "learning_rate": 9.514462809917357e-07, + "logits/chosen": -0.1591796875, + "logits/rejected": -0.3359375, + "logps/chosen": -392.0, + "logps/rejected": -400.0, + "loss": 0.0381, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.9296875, + "rewards/margins": 11.625, + "rewards/rejected": -10.6875, + "step": 910 + }, + { + "epoch": 1.1177914110429448, + "grad_norm": 10.08348423519109, + "learning_rate": 9.508264462809917e-07, + "logits/chosen": -0.109375, + "logits/rejected": -0.3515625, + "logps/chosen": -472.0, + "logps/rejected": -428.0, + "loss": 0.0619, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 1.6953125, + "rewards/margins": 13.375, + "rewards/rejected": -11.6875, + "step": 911 + }, + { + "epoch": 1.1190184049079754, + "grad_norm": 14.853944201654002, + "learning_rate": 9.502066115702479e-07, + "logits/chosen": 0.01153564453125, + "logits/rejected": -0.0849609375, + "logps/chosen": -362.0, + "logps/rejected": -384.0, + "loss": 0.0499, + "rewards/accuracies": 0.9765625, + "rewards/chosen": 0.07080078125, + "rewards/margins": 10.25, + "rewards/rejected": -10.1875, + "step": 912 + }, + { + "epoch": 1.120245398773006, + "grad_norm": 7.41210069844449, + "learning_rate": 9.495867768595041e-07, + "logits/chosen": -0.1845703125, + "logits/rejected": -0.28125, + "logps/chosen": -328.0, + "logps/rejected": -376.0, + "loss": 0.0393, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.56640625, + "rewards/margins": 11.625, + "rewards/rejected": -11.0, + "step": 913 + }, + { + "epoch": 1.1214723926380368, + "grad_norm": 17.23126040997008, + "learning_rate": 9.489669421487604e-07, + "logits/chosen": 0.0013275146484375, + "logits/rejected": -0.1748046875, + "logps/chosen": -412.0, + "logps/rejected": -376.0, + "loss": 0.0818, + "rewards/accuracies": 0.9375, + "rewards/chosen": -0.037109375, + "rewards/margins": 10.125, + "rewards/rejected": -10.125, + "step": 914 + }, + { + "epoch": 1.1226993865030674, + "grad_norm": 7.728386443202459, + "learning_rate": 9.483471074380166e-07, + "logits/chosen": -0.034912109375, + "logits/rejected": -0.291015625, + "logps/chosen": -408.0, + "logps/rejected": -402.0, + "loss": 0.0253, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.81640625, + "rewards/margins": 11.625, + "rewards/rejected": -10.8125, + "step": 915 + }, + { + "epoch": 1.123926380368098, + "grad_norm": 8.320575830299521, + "learning_rate": 9.477272727272727e-07, + "logits/chosen": -0.2001953125, + "logits/rejected": -0.34765625, + "logps/chosen": -394.0, + "logps/rejected": -376.0, + "loss": 0.0332, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.51171875, + "rewards/margins": 11.375, + "rewards/rejected": -10.875, + "step": 916 + }, + { + "epoch": 1.1251533742331288, + "grad_norm": 10.220447898199359, + "learning_rate": 9.47107438016529e-07, + "logits/chosen": -0.1650390625, + "logits/rejected": -0.392578125, + "logps/chosen": -428.0, + "logps/rejected": -430.0, + "loss": 0.0592, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.291015625, + "rewards/margins": 10.9375, + "rewards/rejected": -10.625, + "step": 917 + }, + { + "epoch": 1.1263803680981594, + "grad_norm": 16.327456489125716, + "learning_rate": 9.464876033057852e-07, + "logits/chosen": -0.040771484375, + "logits/rejected": -0.1806640625, + "logps/chosen": -388.0, + "logps/rejected": -384.0, + "loss": 0.0625, + "rewards/accuracies": 0.9609375, + "rewards/chosen": 0.287109375, + "rewards/margins": 10.625, + "rewards/rejected": -10.375, + "step": 918 + }, + { + "epoch": 1.1276073619631901, + "grad_norm": 9.650345793046794, + "learning_rate": 9.458677685950414e-07, + "logits/chosen": -0.2412109375, + "logits/rejected": -0.294921875, + "logps/chosen": -384.0, + "logps/rejected": -408.0, + "loss": 0.0513, + "rewards/accuracies": 0.984375, + "rewards/chosen": -0.44140625, + "rewards/margins": 10.625, + "rewards/rejected": -11.0625, + "step": 919 + }, + { + "epoch": 1.1288343558282208, + "grad_norm": 5.645777430480824, + "learning_rate": 9.452479338842975e-07, + "logits/chosen": -0.07373046875, + "logits/rejected": -0.287109375, + "logps/chosen": -410.0, + "logps/rejected": -396.0, + "loss": 0.0183, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.0281982421875, + "rewards/margins": 11.0, + "rewards/rejected": -11.0, + "step": 920 + }, + { + "epoch": 1.1300613496932514, + "grad_norm": 6.67262482291788, + "learning_rate": 9.446280991735537e-07, + "logits/chosen": -0.08935546875, + "logits/rejected": -0.2421875, + "logps/chosen": -398.0, + "logps/rejected": -396.0, + "loss": 0.0236, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 1.078125, + "rewards/margins": 11.6875, + "rewards/rejected": -10.5625, + "step": 921 + }, + { + "epoch": 1.1312883435582821, + "grad_norm": 6.984229781332186, + "learning_rate": 9.440082644628099e-07, + "logits/chosen": -0.036865234375, + "logits/rejected": -0.15234375, + "logps/chosen": -392.0, + "logps/rejected": -402.0, + "loss": 0.0334, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.98046875, + "rewards/margins": 10.9375, + "rewards/rejected": -9.9375, + "step": 922 + }, + { + "epoch": 1.132515337423313, + "grad_norm": 12.662651710075986, + "learning_rate": 9.433884297520662e-07, + "logits/chosen": 0.005828857421875, + "logits/rejected": -0.08349609375, + "logps/chosen": -364.0, + "logps/rejected": -378.0, + "loss": 0.0303, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.9609375, + "rewards/margins": 10.5625, + "rewards/rejected": -9.625, + "step": 923 + }, + { + "epoch": 1.1337423312883437, + "grad_norm": 9.002647218488805, + "learning_rate": 9.427685950413223e-07, + "logits/chosen": -0.142578125, + "logits/rejected": -0.2265625, + "logps/chosen": -384.0, + "logps/rejected": -368.0, + "loss": 0.0351, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.91015625, + "rewards/margins": 11.25, + "rewards/rejected": -10.375, + "step": 924 + }, + { + "epoch": 1.1349693251533743, + "grad_norm": 12.813399120951164, + "learning_rate": 9.421487603305785e-07, + "logits/chosen": -0.203125, + "logits/rejected": -0.404296875, + "logps/chosen": -378.0, + "logps/rejected": -364.0, + "loss": 0.0329, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.0306396484375, + "rewards/margins": 10.875, + "rewards/rejected": -10.8125, + "step": 925 + }, + { + "epoch": 1.136196319018405, + "grad_norm": 13.681237077440016, + "learning_rate": 9.415289256198348e-07, + "logits/chosen": -0.2138671875, + "logits/rejected": -0.384765625, + "logps/chosen": -426.0, + "logps/rejected": -410.0, + "loss": 0.0828, + "rewards/accuracies": 0.9765625, + "rewards/chosen": 0.9921875, + "rewards/margins": 11.1875, + "rewards/rejected": -10.1875, + "step": 926 + }, + { + "epoch": 1.1374233128834357, + "grad_norm": 14.007812307598066, + "learning_rate": 9.40909090909091e-07, + "logits/chosen": -0.091796875, + "logits/rejected": -0.2431640625, + "logps/chosen": -426.0, + "logps/rejected": -406.0, + "loss": 0.0454, + "rewards/accuracies": 0.9765625, + "rewards/chosen": 0.921875, + "rewards/margins": 11.0625, + "rewards/rejected": -10.1875, + "step": 927 + }, + { + "epoch": 1.1386503067484663, + "grad_norm": 5.804375472545515, + "learning_rate": 9.402892561983472e-07, + "logits/chosen": -0.06982421875, + "logits/rejected": -0.2734375, + "logps/chosen": -416.0, + "logps/rejected": -398.0, + "loss": 0.0282, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.67578125, + "rewards/margins": 10.8125, + "rewards/rejected": -10.125, + "step": 928 + }, + { + "epoch": 1.139877300613497, + "grad_norm": 7.0706705773655205, + "learning_rate": 9.396694214876032e-07, + "logits/chosen": -0.038330078125, + "logits/rejected": -0.212890625, + "logps/chosen": -430.0, + "logps/rejected": -348.0, + "loss": 0.0216, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.90625, + "rewards/margins": 10.75, + "rewards/rejected": -9.8125, + "step": 929 + }, + { + "epoch": 1.1411042944785277, + "grad_norm": 20.03419906761046, + "learning_rate": 9.390495867768595e-07, + "logits/chosen": -0.0673828125, + "logits/rejected": -0.302734375, + "logps/chosen": -408.0, + "logps/rejected": -392.0, + "loss": 0.0678, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.546875, + "rewards/margins": 11.5, + "rewards/rejected": -10.9375, + "step": 930 + }, + { + "epoch": 1.1423312883435583, + "grad_norm": 8.79614969540993, + "learning_rate": 9.384297520661157e-07, + "logits/chosen": -0.0712890625, + "logits/rejected": -0.154296875, + "logps/chosen": -412.0, + "logps/rejected": -376.0, + "loss": 0.0392, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.041748046875, + "rewards/margins": 9.5, + "rewards/rejected": -9.5625, + "step": 931 + }, + { + "epoch": 1.143558282208589, + "grad_norm": 5.533709744700053, + "learning_rate": 9.37809917355372e-07, + "logits/chosen": -0.1923828125, + "logits/rejected": -0.3671875, + "logps/chosen": -380.0, + "logps/rejected": -380.0, + "loss": 0.0302, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.045166015625, + "rewards/margins": 11.125, + "rewards/rejected": -11.0625, + "step": 932 + }, + { + "epoch": 1.1447852760736197, + "grad_norm": 17.392392128691636, + "learning_rate": 9.371900826446281e-07, + "logits/chosen": -0.04052734375, + "logits/rejected": -0.1318359375, + "logps/chosen": -372.0, + "logps/rejected": -400.0, + "loss": 0.0582, + "rewards/accuracies": 0.9765625, + "rewards/chosen": 0.4296875, + "rewards/margins": 10.3125, + "rewards/rejected": -9.875, + "step": 933 + }, + { + "epoch": 1.1460122699386504, + "grad_norm": 11.220639802991899, + "learning_rate": 9.365702479338843e-07, + "logits/chosen": -0.076171875, + "logits/rejected": -0.23828125, + "logps/chosen": -386.0, + "logps/rejected": -366.0, + "loss": 0.0466, + "rewards/accuracies": 0.9765625, + "rewards/chosen": 0.94140625, + "rewards/margins": 10.25, + "rewards/rejected": -9.3125, + "step": 934 + }, + { + "epoch": 1.147239263803681, + "grad_norm": 3.127170097265917, + "learning_rate": 9.359504132231406e-07, + "logits/chosen": -0.125, + "logits/rejected": -0.26171875, + "logps/chosen": -416.0, + "logps/rejected": -408.0, + "loss": 0.0156, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.703125, + "rewards/margins": 10.6875, + "rewards/rejected": -10.0, + "step": 935 + }, + { + "epoch": 1.1484662576687117, + "grad_norm": 8.797510711759905, + "learning_rate": 9.353305785123968e-07, + "logits/chosen": 0.04833984375, + "logits/rejected": -0.1953125, + "logps/chosen": -366.0, + "logps/rejected": -382.0, + "loss": 0.0399, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.0927734375, + "rewards/margins": 10.5625, + "rewards/rejected": -10.6875, + "step": 936 + }, + { + "epoch": 1.1496932515337424, + "grad_norm": 8.508479799139819, + "learning_rate": 9.347107438016529e-07, + "logits/chosen": -0.032958984375, + "logits/rejected": -0.267578125, + "logps/chosen": -434.0, + "logps/rejected": -406.0, + "loss": 0.0222, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.255859375, + "rewards/margins": 11.0625, + "rewards/rejected": -10.8125, + "step": 937 + }, + { + "epoch": 1.150920245398773, + "grad_norm": 11.734844982011332, + "learning_rate": 9.34090909090909e-07, + "logits/chosen": -0.01031494140625, + "logits/rejected": -0.058837890625, + "logps/chosen": -374.0, + "logps/rejected": -378.0, + "loss": 0.0576, + "rewards/accuracies": 0.9765625, + "rewards/chosen": 0.294921875, + "rewards/margins": 10.6875, + "rewards/rejected": -10.375, + "step": 938 + }, + { + "epoch": 1.1521472392638037, + "grad_norm": 6.686498950384566, + "learning_rate": 9.334710743801653e-07, + "logits/chosen": -0.1796875, + "logits/rejected": -0.318359375, + "logps/chosen": -450.0, + "logps/rejected": -420.0, + "loss": 0.022, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 1.0703125, + "rewards/margins": 12.0625, + "rewards/rejected": -11.0, + "step": 939 + }, + { + "epoch": 1.1533742331288344, + "grad_norm": 9.333043593563687, + "learning_rate": 9.328512396694215e-07, + "logits/chosen": -0.169921875, + "logits/rejected": -0.361328125, + "logps/chosen": -420.0, + "logps/rejected": -388.0, + "loss": 0.0298, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 1.0234375, + "rewards/margins": 12.125, + "rewards/rejected": -11.125, + "step": 940 + }, + { + "epoch": 1.154601226993865, + "grad_norm": 13.328010603903623, + "learning_rate": 9.322314049586778e-07, + "logits/chosen": -0.0150146484375, + "logits/rejected": -0.220703125, + "logps/chosen": -418.0, + "logps/rejected": -390.0, + "loss": 0.058, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 1.25, + "rewards/margins": 10.9375, + "rewards/rejected": -9.6875, + "step": 941 + }, + { + "epoch": 1.1558282208588957, + "grad_norm": 10.748405478390469, + "learning_rate": 9.316115702479339e-07, + "logits/chosen": -0.2431640625, + "logits/rejected": -0.380859375, + "logps/chosen": -418.0, + "logps/rejected": -406.0, + "loss": 0.0373, + "rewards/accuracies": 0.984375, + "rewards/chosen": 1.2109375, + "rewards/margins": 12.25, + "rewards/rejected": -11.0625, + "step": 942 + }, + { + "epoch": 1.1570552147239264, + "grad_norm": 6.66797822267693, + "learning_rate": 9.309917355371901e-07, + "logits/chosen": 0.0140380859375, + "logits/rejected": -0.1474609375, + "logps/chosen": -354.0, + "logps/rejected": -350.0, + "loss": 0.021, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.359375, + "rewards/margins": 10.25, + "rewards/rejected": -9.875, + "step": 943 + }, + { + "epoch": 1.158282208588957, + "grad_norm": 6.910993717404971, + "learning_rate": 9.303719008264464e-07, + "logits/chosen": -0.0791015625, + "logits/rejected": -0.2490234375, + "logps/chosen": -404.0, + "logps/rejected": -412.0, + "loss": 0.0213, + "rewards/accuracies": 1.0, + "rewards/chosen": 1.1171875, + "rewards/margins": 11.5625, + "rewards/rejected": -10.4375, + "step": 944 + }, + { + "epoch": 1.1595092024539877, + "grad_norm": 8.94698116889764, + "learning_rate": 9.297520661157025e-07, + "logits/chosen": -0.1640625, + "logits/rejected": -0.33984375, + "logps/chosen": -416.0, + "logps/rejected": -380.0, + "loss": 0.0401, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 1.390625, + "rewards/margins": 10.75, + "rewards/rejected": -9.375, + "step": 945 + }, + { + "epoch": 1.1607361963190184, + "grad_norm": 5.757411937782045, + "learning_rate": 9.291322314049587e-07, + "logits/chosen": -0.0016937255859375, + "logits/rejected": -0.07177734375, + "logps/chosen": -380.0, + "logps/rejected": -400.0, + "loss": 0.0224, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.291015625, + "rewards/margins": 9.8125, + "rewards/rejected": -9.5625, + "step": 946 + }, + { + "epoch": 1.161963190184049, + "grad_norm": 10.946703026968246, + "learning_rate": 9.285123966942148e-07, + "logits/chosen": -0.2138671875, + "logits/rejected": -0.3984375, + "logps/chosen": -346.0, + "logps/rejected": -368.0, + "loss": 0.0589, + "rewards/accuracies": 0.9765625, + "rewards/chosen": 0.7734375, + "rewards/margins": 10.125, + "rewards/rejected": -9.375, + "step": 947 + }, + { + "epoch": 1.1631901840490797, + "grad_norm": 13.673425770565624, + "learning_rate": 9.278925619834711e-07, + "logits/chosen": -0.1962890625, + "logits/rejected": -0.427734375, + "logps/chosen": -428.0, + "logps/rejected": -380.0, + "loss": 0.0544, + "rewards/accuracies": 0.984375, + "rewards/chosen": 1.5, + "rewards/margins": 10.6875, + "rewards/rejected": -9.1875, + "step": 948 + }, + { + "epoch": 1.1644171779141104, + "grad_norm": 10.764489540205084, + "learning_rate": 9.272727272727273e-07, + "logits/chosen": 0.0032958984375, + "logits/rejected": -0.265625, + "logps/chosen": -414.0, + "logps/rejected": -410.0, + "loss": 0.0627, + "rewards/accuracies": 0.96875, + "rewards/chosen": 1.46875, + "rewards/margins": 11.625, + "rewards/rejected": -10.125, + "step": 949 + }, + { + "epoch": 1.165644171779141, + "grad_norm": 6.163510999316528, + "learning_rate": 9.266528925619834e-07, + "logits/chosen": -0.134765625, + "logits/rejected": -0.23828125, + "logps/chosen": -338.0, + "logps/rejected": -336.0, + "loss": 0.0233, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.265625, + "rewards/margins": 10.5, + "rewards/rejected": -10.25, + "step": 950 + }, + { + "epoch": 1.1668711656441717, + "grad_norm": 7.49094764434787, + "learning_rate": 9.260330578512397e-07, + "logits/chosen": -0.021240234375, + "logits/rejected": -0.146484375, + "logps/chosen": -350.0, + "logps/rejected": -354.0, + "loss": 0.0353, + "rewards/accuracies": 0.9765625, + "rewards/chosen": 0.259765625, + "rewards/margins": 9.9375, + "rewards/rejected": -9.6875, + "step": 951 + }, + { + "epoch": 1.1680981595092024, + "grad_norm": 11.812912807851047, + "learning_rate": 9.254132231404959e-07, + "logits/chosen": -0.2197265625, + "logits/rejected": -0.494140625, + "logps/chosen": -424.0, + "logps/rejected": -386.0, + "loss": 0.0565, + "rewards/accuracies": 0.984375, + "rewards/chosen": 1.09375, + "rewards/margins": 11.1875, + "rewards/rejected": -10.0625, + "step": 952 + }, + { + "epoch": 1.169325153374233, + "grad_norm": 9.899530945112486, + "learning_rate": 9.247933884297522e-07, + "logits/chosen": -0.07080078125, + "logits/rejected": -0.271484375, + "logps/chosen": -388.0, + "logps/rejected": -360.0, + "loss": 0.038, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.96484375, + "rewards/margins": 10.5625, + "rewards/rejected": -9.5625, + "step": 953 + }, + { + "epoch": 1.1705521472392637, + "grad_norm": 14.874148700852642, + "learning_rate": 9.241735537190083e-07, + "logits/chosen": -0.0771484375, + "logits/rejected": -0.283203125, + "logps/chosen": -322.0, + "logps/rejected": -320.0, + "loss": 0.0591, + "rewards/accuracies": 0.96875, + "rewards/chosen": -0.439453125, + "rewards/margins": 9.8125, + "rewards/rejected": -10.25, + "step": 954 + }, + { + "epoch": 1.1717791411042944, + "grad_norm": 14.69776784288544, + "learning_rate": 9.235537190082644e-07, + "logits/chosen": -0.005157470703125, + "logits/rejected": -0.0771484375, + "logps/chosen": -380.0, + "logps/rejected": -378.0, + "loss": 0.0538, + "rewards/accuracies": 0.984375, + "rewards/chosen": 1.015625, + "rewards/margins": 10.8125, + "rewards/rejected": -9.8125, + "step": 955 + }, + { + "epoch": 1.173006134969325, + "grad_norm": 8.526200653171099, + "learning_rate": 9.229338842975206e-07, + "logits/chosen": -0.10009765625, + "logits/rejected": -0.2041015625, + "logps/chosen": -316.0, + "logps/rejected": -344.0, + "loss": 0.0306, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.30859375, + "rewards/margins": 9.75, + "rewards/rejected": -9.4375, + "step": 956 + }, + { + "epoch": 1.1742331288343557, + "grad_norm": 7.973457405245253, + "learning_rate": 9.223140495867769e-07, + "logits/chosen": -0.134765625, + "logits/rejected": -0.357421875, + "logps/chosen": -390.0, + "logps/rejected": -404.0, + "loss": 0.0368, + "rewards/accuracies": 0.984375, + "rewards/chosen": 1.0859375, + "rewards/margins": 10.875, + "rewards/rejected": -9.8125, + "step": 957 + }, + { + "epoch": 1.1754601226993864, + "grad_norm": 14.106346468292243, + "learning_rate": 9.216942148760331e-07, + "logits/chosen": -0.055908203125, + "logits/rejected": -0.15625, + "logps/chosen": -364.0, + "logps/rejected": -404.0, + "loss": 0.0479, + "rewards/accuracies": 0.9765625, + "rewards/chosen": 0.404296875, + "rewards/margins": 11.0, + "rewards/rejected": -10.625, + "step": 958 + }, + { + "epoch": 1.1766871165644173, + "grad_norm": 11.510149735049389, + "learning_rate": 9.210743801652892e-07, + "logits/chosen": -0.095703125, + "logits/rejected": -0.267578125, + "logps/chosen": -396.0, + "logps/rejected": -388.0, + "loss": 0.0431, + "rewards/accuracies": 0.9765625, + "rewards/chosen": 0.63671875, + "rewards/margins": 10.8125, + "rewards/rejected": -10.1875, + "step": 959 + }, + { + "epoch": 1.177914110429448, + "grad_norm": 3.8552096147370842, + "learning_rate": 9.204545454545455e-07, + "logits/chosen": -0.1728515625, + "logits/rejected": -0.330078125, + "logps/chosen": -346.0, + "logps/rejected": -350.0, + "loss": 0.0149, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.546875, + "rewards/margins": 10.25, + "rewards/rejected": -10.75, + "step": 960 + }, + { + "epoch": 1.1791411042944786, + "grad_norm": 13.533375661298507, + "learning_rate": 9.198347107438017e-07, + "logits/chosen": -0.12451171875, + "logits/rejected": -0.2353515625, + "logps/chosen": -404.0, + "logps/rejected": -380.0, + "loss": 0.0407, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.85546875, + "rewards/margins": 11.5, + "rewards/rejected": -10.625, + "step": 961 + }, + { + "epoch": 1.1803680981595093, + "grad_norm": 9.63092706255587, + "learning_rate": 9.192148760330579e-07, + "logits/chosen": 0.00921630859375, + "logits/rejected": -0.16015625, + "logps/chosen": -440.0, + "logps/rejected": -430.0, + "loss": 0.0318, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.609375, + "rewards/margins": 11.0, + "rewards/rejected": -10.375, + "step": 962 + }, + { + "epoch": 1.18159509202454, + "grad_norm": 13.10088160617771, + "learning_rate": 9.185950413223141e-07, + "logits/chosen": -0.11474609375, + "logits/rejected": -0.306640625, + "logps/chosen": -378.0, + "logps/rejected": -352.0, + "loss": 0.0397, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.61328125, + "rewards/margins": 10.875, + "rewards/rejected": -10.25, + "step": 963 + }, + { + "epoch": 1.1828220858895706, + "grad_norm": 7.358279878141561, + "learning_rate": 9.179752066115702e-07, + "logits/chosen": -0.049560546875, + "logits/rejected": -0.2314453125, + "logps/chosen": -380.0, + "logps/rejected": -368.0, + "loss": 0.0233, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.69921875, + "rewards/margins": 11.375, + "rewards/rejected": -10.6875, + "step": 964 + }, + { + "epoch": 1.1840490797546013, + "grad_norm": 10.963522347339602, + "learning_rate": 9.173553719008264e-07, + "logits/chosen": -0.12158203125, + "logits/rejected": -0.255859375, + "logps/chosen": -366.0, + "logps/rejected": -386.0, + "loss": 0.0329, + "rewards/accuracies": 0.984375, + "rewards/chosen": -0.287109375, + "rewards/margins": 10.75, + "rewards/rejected": -11.0625, + "step": 965 + }, + { + "epoch": 1.185276073619632, + "grad_norm": 11.155993123724087, + "learning_rate": 9.167355371900827e-07, + "logits/chosen": -0.1875, + "logits/rejected": -0.322265625, + "logps/chosen": -396.0, + "logps/rejected": -360.0, + "loss": 0.0412, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.00421142578125, + "rewards/margins": 10.6875, + "rewards/rejected": -10.625, + "step": 966 + }, + { + "epoch": 1.1865030674846626, + "grad_norm": 7.534460905603958, + "learning_rate": 9.161157024793388e-07, + "logits/chosen": -0.11279296875, + "logits/rejected": -0.28515625, + "logps/chosen": -412.0, + "logps/rejected": -378.0, + "loss": 0.027, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.95703125, + "rewards/margins": 11.4375, + "rewards/rejected": -10.5, + "step": 967 + }, + { + "epoch": 1.1877300613496933, + "grad_norm": 17.6032434079415, + "learning_rate": 9.15495867768595e-07, + "logits/chosen": -0.041015625, + "logits/rejected": -0.224609375, + "logps/chosen": -424.0, + "logps/rejected": -396.0, + "loss": 0.0462, + "rewards/accuracies": 0.9765625, + "rewards/chosen": 0.875, + "rewards/margins": 12.25, + "rewards/rejected": -11.375, + "step": 968 + }, + { + "epoch": 1.188957055214724, + "grad_norm": 15.694096678580106, + "learning_rate": 9.148760330578513e-07, + "logits/chosen": -0.1455078125, + "logits/rejected": -0.3046875, + "logps/chosen": -388.0, + "logps/rejected": -422.0, + "loss": 0.0629, + "rewards/accuracies": 0.96875, + "rewards/chosen": -1.390625, + "rewards/margins": 10.0, + "rewards/rejected": -11.375, + "step": 969 + }, + { + "epoch": 1.1901840490797546, + "grad_norm": 11.921476909261088, + "learning_rate": 9.142561983471075e-07, + "logits/chosen": -0.10595703125, + "logits/rejected": -0.2451171875, + "logps/chosen": -372.0, + "logps/rejected": -502.0, + "loss": 0.0453, + "rewards/accuracies": 0.984375, + "rewards/chosen": -0.01531982421875, + "rewards/margins": 23.125, + "rewards/rejected": -23.125, + "step": 970 + }, + { + "epoch": 1.1914110429447853, + "grad_norm": 13.62660815818298, + "learning_rate": 9.136363636363637e-07, + "logits/chosen": -0.0091552734375, + "logits/rejected": -0.25, + "logps/chosen": -358.0, + "logps/rejected": -356.0, + "loss": 0.0572, + "rewards/accuracies": 0.984375, + "rewards/chosen": -0.09375, + "rewards/margins": 10.5625, + "rewards/rejected": -10.625, + "step": 971 + }, + { + "epoch": 1.192638036809816, + "grad_norm": 7.237399332254699, + "learning_rate": 9.130165289256199e-07, + "logits/chosen": -0.057373046875, + "logits/rejected": -0.1982421875, + "logps/chosen": -342.0, + "logps/rejected": -312.0, + "loss": 0.0381, + "rewards/accuracies": 0.984375, + "rewards/chosen": -0.3515625, + "rewards/margins": 9.5625, + "rewards/rejected": -9.875, + "step": 972 + }, + { + "epoch": 1.1938650306748466, + "grad_norm": 9.089933594933969, + "learning_rate": 9.12396694214876e-07, + "logits/chosen": -0.1533203125, + "logits/rejected": -0.279296875, + "logps/chosen": -412.0, + "logps/rejected": -432.0, + "loss": 0.029, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.7109375, + "rewards/margins": 11.625, + "rewards/rejected": -10.875, + "step": 973 + }, + { + "epoch": 1.1950920245398773, + "grad_norm": 3.7486390717851217, + "learning_rate": 9.117768595041322e-07, + "logits/chosen": -0.2275390625, + "logits/rejected": -0.41796875, + "logps/chosen": -398.0, + "logps/rejected": -406.0, + "loss": 0.0146, + "rewards/accuracies": 1.0, + "rewards/chosen": 1.09375, + "rewards/margins": 11.9375, + "rewards/rejected": -10.8125, + "step": 974 + }, + { + "epoch": 1.196319018404908, + "grad_norm": 17.692363179531846, + "learning_rate": 9.111570247933885e-07, + "logits/chosen": 0.045654296875, + "logits/rejected": -0.07861328125, + "logps/chosen": -360.0, + "logps/rejected": -332.0, + "loss": 0.0651, + "rewards/accuracies": 0.9765625, + "rewards/chosen": 0.2265625, + "rewards/margins": 9.9375, + "rewards/rejected": -9.6875, + "step": 975 + }, + { + "epoch": 1.1975460122699386, + "grad_norm": 13.042080391360054, + "learning_rate": 9.105371900826446e-07, + "logits/chosen": -0.173828125, + "logits/rejected": -0.392578125, + "logps/chosen": -422.0, + "logps/rejected": -390.0, + "loss": 0.0488, + "rewards/accuracies": 0.984375, + "rewards/chosen": 1.0078125, + "rewards/margins": 10.4375, + "rewards/rejected": -9.4375, + "step": 976 + }, + { + "epoch": 1.1987730061349693, + "grad_norm": 17.283449726274785, + "learning_rate": 9.099173553719009e-07, + "logits/chosen": -0.06787109375, + "logits/rejected": -0.267578125, + "logps/chosen": -388.0, + "logps/rejected": -374.0, + "loss": 0.083, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.91796875, + "rewards/margins": 11.5, + "rewards/rejected": -10.5625, + "step": 977 + }, + { + "epoch": 1.2, + "grad_norm": 10.327665992150298, + "learning_rate": 9.092975206611571e-07, + "logits/chosen": -0.11865234375, + "logits/rejected": -0.1416015625, + "logps/chosen": -364.0, + "logps/rejected": -344.0, + "loss": 0.0616, + "rewards/accuracies": 0.96875, + "rewards/chosen": -0.046630859375, + "rewards/margins": 10.125, + "rewards/rejected": -10.125, + "step": 978 + }, + { + "epoch": 1.2012269938650306, + "grad_norm": 6.639005543404221, + "learning_rate": 9.086776859504132e-07, + "logits/chosen": -0.00439453125, + "logits/rejected": -0.126953125, + "logps/chosen": -410.0, + "logps/rejected": -384.0, + "loss": 0.0294, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.76953125, + "rewards/margins": 11.25, + "rewards/rejected": -10.4375, + "step": 979 + }, + { + "epoch": 1.2024539877300613, + "grad_norm": 3.7048596375075307, + "learning_rate": 9.080578512396695e-07, + "logits/chosen": -0.173828125, + "logits/rejected": -0.35546875, + "logps/chosen": -362.0, + "logps/rejected": -374.0, + "loss": 0.0142, + "rewards/accuracies": 1.0, + "rewards/chosen": 1.6484375, + "rewards/margins": 11.625, + "rewards/rejected": -10.0, + "step": 980 + }, + { + "epoch": 1.203680981595092, + "grad_norm": 9.842884355659873, + "learning_rate": 9.074380165289257e-07, + "logits/chosen": -0.0908203125, + "logits/rejected": -0.1796875, + "logps/chosen": -386.0, + "logps/rejected": -386.0, + "loss": 0.0228, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.8671875, + "rewards/margins": 10.9375, + "rewards/rejected": -10.0625, + "step": 981 + }, + { + "epoch": 1.2049079754601226, + "grad_norm": 8.509249883161877, + "learning_rate": 9.068181818181818e-07, + "logits/chosen": -0.09130859375, + "logits/rejected": -0.267578125, + "logps/chosen": -408.0, + "logps/rejected": -408.0, + "loss": 0.0327, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.734375, + "rewards/margins": 10.75, + "rewards/rejected": -10.0, + "step": 982 + }, + { + "epoch": 1.2061349693251533, + "grad_norm": 17.32331392424542, + "learning_rate": 9.06198347107438e-07, + "logits/chosen": -0.09814453125, + "logits/rejected": -0.28125, + "logps/chosen": -382.0, + "logps/rejected": -346.0, + "loss": 0.0823, + "rewards/accuracies": 0.953125, + "rewards/chosen": 0.671875, + "rewards/margins": 9.625, + "rewards/rejected": -9.0, + "step": 983 + }, + { + "epoch": 1.207361963190184, + "grad_norm": 6.308837245908633, + "learning_rate": 9.055785123966942e-07, + "logits/chosen": -0.1328125, + "logits/rejected": -0.310546875, + "logps/chosen": -362.0, + "logps/rejected": -346.0, + "loss": 0.0285, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.447265625, + "rewards/margins": 10.3125, + "rewards/rejected": -9.875, + "step": 984 + }, + { + "epoch": 1.2085889570552146, + "grad_norm": 11.159178783645137, + "learning_rate": 9.049586776859504e-07, + "logits/chosen": -0.04443359375, + "logits/rejected": -0.212890625, + "logps/chosen": -350.0, + "logps/rejected": -352.0, + "loss": 0.0427, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.76171875, + "rewards/margins": 10.3125, + "rewards/rejected": -9.5625, + "step": 985 + }, + { + "epoch": 1.2098159509202455, + "grad_norm": 9.051373556481083, + "learning_rate": 9.043388429752067e-07, + "logits/chosen": 0.0625, + "logits/rejected": -0.08935546875, + "logps/chosen": -390.0, + "logps/rejected": -356.0, + "loss": 0.0347, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 1.0546875, + "rewards/margins": 10.0, + "rewards/rejected": -8.9375, + "step": 986 + }, + { + "epoch": 1.2110429447852762, + "grad_norm": 6.198302604249075, + "learning_rate": 9.037190082644629e-07, + "logits/chosen": -0.1435546875, + "logits/rejected": -0.390625, + "logps/chosen": -458.0, + "logps/rejected": -406.0, + "loss": 0.0282, + "rewards/accuracies": 0.984375, + "rewards/chosen": 1.28125, + "rewards/margins": 11.1875, + "rewards/rejected": -9.875, + "step": 987 + }, + { + "epoch": 1.2122699386503069, + "grad_norm": 4.102469498417344, + "learning_rate": 9.03099173553719e-07, + "logits/chosen": -0.047607421875, + "logits/rejected": -0.26171875, + "logps/chosen": -372.0, + "logps/rejected": -348.0, + "loss": 0.0143, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.240234375, + "rewards/margins": 10.5625, + "rewards/rejected": -10.3125, + "step": 988 + }, + { + "epoch": 1.2134969325153375, + "grad_norm": 8.46030603702519, + "learning_rate": 9.024793388429753e-07, + "logits/chosen": -0.0023345947265625, + "logits/rejected": -0.16796875, + "logps/chosen": -354.0, + "logps/rejected": -342.0, + "loss": 0.0404, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.21484375, + "rewards/margins": 9.4375, + "rewards/rejected": -9.6875, + "step": 989 + }, + { + "epoch": 1.2147239263803682, + "grad_norm": 10.249236383125274, + "learning_rate": 9.018595041322315e-07, + "logits/chosen": -0.1494140625, + "logits/rejected": -0.326171875, + "logps/chosen": -402.0, + "logps/rejected": -396.0, + "loss": 0.0387, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.58984375, + "rewards/margins": 11.5, + "rewards/rejected": -10.875, + "step": 990 + }, + { + "epoch": 1.2159509202453989, + "grad_norm": 20.10350249363253, + "learning_rate": 9.012396694214876e-07, + "logits/chosen": -0.09130859375, + "logits/rejected": -0.2197265625, + "logps/chosen": -416.0, + "logps/rejected": -408.0, + "loss": 0.0671, + "rewards/accuracies": 0.96875, + "rewards/chosen": -0.263671875, + "rewards/margins": 11.0, + "rewards/rejected": -11.25, + "step": 991 + }, + { + "epoch": 1.2171779141104295, + "grad_norm": 15.353242393368603, + "learning_rate": 9.006198347107438e-07, + "logits/chosen": -0.11572265625, + "logits/rejected": -0.302734375, + "logps/chosen": -388.0, + "logps/rejected": -392.0, + "loss": 0.067, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.11962890625, + "rewards/margins": 10.875, + "rewards/rejected": -10.75, + "step": 992 + }, + { + "epoch": 1.2184049079754602, + "grad_norm": 14.594366015899684, + "learning_rate": 9e-07, + "logits/chosen": -0.1689453125, + "logits/rejected": -0.314453125, + "logps/chosen": -376.0, + "logps/rejected": -376.0, + "loss": 0.0584, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.7265625, + "rewards/margins": 10.8125, + "rewards/rejected": -10.0625, + "step": 993 + }, + { + "epoch": 1.2196319018404909, + "grad_norm": 17.819835095986992, + "learning_rate": 8.993801652892562e-07, + "logits/chosen": -0.169921875, + "logits/rejected": -0.341796875, + "logps/chosen": -428.0, + "logps/rejected": -408.0, + "loss": 0.0555, + "rewards/accuracies": 0.9765625, + "rewards/chosen": 0.251953125, + "rewards/margins": 11.0625, + "rewards/rejected": -10.8125, + "step": 994 + }, + { + "epoch": 1.2208588957055215, + "grad_norm": 15.26610473864711, + "learning_rate": 8.987603305785125e-07, + "logits/chosen": -0.099609375, + "logits/rejected": -0.2412109375, + "logps/chosen": -362.0, + "logps/rejected": -398.0, + "loss": 0.046, + "rewards/accuracies": 0.9765625, + "rewards/chosen": -0.52734375, + "rewards/margins": 10.3125, + "rewards/rejected": -10.8125, + "step": 995 + }, + { + "epoch": 1.2220858895705522, + "grad_norm": 19.263245504901857, + "learning_rate": 8.981404958677686e-07, + "logits/chosen": -0.023193359375, + "logits/rejected": -0.287109375, + "logps/chosen": -388.0, + "logps/rejected": -372.0, + "loss": 0.0498, + "rewards/accuracies": 0.984375, + "rewards/chosen": -0.2021484375, + "rewards/margins": 10.125, + "rewards/rejected": -10.375, + "step": 996 + }, + { + "epoch": 1.2233128834355829, + "grad_norm": 12.681921452742241, + "learning_rate": 8.975206611570248e-07, + "logits/chosen": -0.18359375, + "logits/rejected": -0.37109375, + "logps/chosen": -420.0, + "logps/rejected": -376.0, + "loss": 0.0371, + "rewards/accuracies": 0.984375, + "rewards/chosen": 1.171875, + "rewards/margins": 10.9375, + "rewards/rejected": -9.8125, + "step": 997 + }, + { + "epoch": 1.2245398773006135, + "grad_norm": 6.006248666573442, + "learning_rate": 8.969008264462811e-07, + "logits/chosen": -0.017578125, + "logits/rejected": -0.21484375, + "logps/chosen": -446.0, + "logps/rejected": -414.0, + "loss": 0.0184, + "rewards/accuracies": 1.0, + "rewards/chosen": 1.171875, + "rewards/margins": 11.25, + "rewards/rejected": -10.0625, + "step": 998 + }, + { + "epoch": 1.2257668711656442, + "grad_norm": 11.143450935670302, + "learning_rate": 8.962809917355373e-07, + "logits/chosen": -0.07177734375, + "logits/rejected": -0.240234375, + "logps/chosen": -410.0, + "logps/rejected": -386.0, + "loss": 0.0398, + "rewards/accuracies": 0.984375, + "rewards/chosen": 1.21875, + "rewards/margins": 11.25, + "rewards/rejected": -10.0625, + "step": 999 + }, + { + "epoch": 1.2269938650306749, + "grad_norm": 8.965376937967298, + "learning_rate": 8.956611570247934e-07, + "logits/chosen": -0.01385498046875, + "logits/rejected": -0.1015625, + "logps/chosen": -386.0, + "logps/rejected": -386.0, + "loss": 0.0469, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 1.0234375, + "rewards/margins": 10.8125, + "rewards/rejected": -9.75, + "step": 1000 + }, + { + "epoch": 1.2282208588957055, + "grad_norm": 11.244494297541378, + "learning_rate": 8.950413223140495e-07, + "logits/chosen": -0.185546875, + "logits/rejected": -0.3515625, + "logps/chosen": -410.0, + "logps/rejected": -378.0, + "loss": 0.0447, + "rewards/accuracies": 0.984375, + "rewards/chosen": 1.4453125, + "rewards/margins": 11.75, + "rewards/rejected": -10.3125, + "step": 1001 + }, + { + "epoch": 1.2294478527607362, + "grad_norm": 16.003709589565055, + "learning_rate": 8.944214876033058e-07, + "logits/chosen": -0.059326171875, + "logits/rejected": -0.09375, + "logps/chosen": -378.0, + "logps/rejected": -410.0, + "loss": 0.0566, + "rewards/accuracies": 0.9765625, + "rewards/chosen": 0.875, + "rewards/margins": 10.8125, + "rewards/rejected": -9.9375, + "step": 1002 + }, + { + "epoch": 1.2306748466257669, + "grad_norm": 12.870892763097368, + "learning_rate": 8.93801652892562e-07, + "logits/chosen": -0.1552734375, + "logits/rejected": -0.283203125, + "logps/chosen": -362.0, + "logps/rejected": -376.0, + "loss": 0.0662, + "rewards/accuracies": 0.9765625, + "rewards/chosen": 0.48828125, + "rewards/margins": 10.625, + "rewards/rejected": -10.125, + "step": 1003 + }, + { + "epoch": 1.2319018404907975, + "grad_norm": 5.2378730711968435, + "learning_rate": 8.931818181818183e-07, + "logits/chosen": -0.03076171875, + "logits/rejected": -0.2451171875, + "logps/chosen": -360.0, + "logps/rejected": -328.0, + "loss": 0.0257, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.55859375, + "rewards/margins": 9.875, + "rewards/rejected": -9.3125, + "step": 1004 + }, + { + "epoch": 1.2331288343558282, + "grad_norm": 6.60847796992945, + "learning_rate": 8.925619834710744e-07, + "logits/chosen": -0.0322265625, + "logits/rejected": -0.2177734375, + "logps/chosen": -396.0, + "logps/rejected": -360.0, + "loss": 0.031, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.443359375, + "rewards/margins": 10.5625, + "rewards/rejected": -10.125, + "step": 1005 + }, + { + "epoch": 1.2343558282208589, + "grad_norm": 7.331811572998349, + "learning_rate": 8.919421487603306e-07, + "logits/chosen": -0.0927734375, + "logits/rejected": -0.26171875, + "logps/chosen": -368.0, + "logps/rejected": -360.0, + "loss": 0.03, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.57421875, + "rewards/margins": 10.9375, + "rewards/rejected": -10.375, + "step": 1006 + }, + { + "epoch": 1.2355828220858895, + "grad_norm": 12.728485046758394, + "learning_rate": 8.913223140495869e-07, + "logits/chosen": -0.0595703125, + "logits/rejected": -0.2177734375, + "logps/chosen": -404.0, + "logps/rejected": -404.0, + "loss": 0.0543, + "rewards/accuracies": 0.96875, + "rewards/chosen": 0.8671875, + "rewards/margins": 11.125, + "rewards/rejected": -10.25, + "step": 1007 + }, + { + "epoch": 1.2368098159509202, + "grad_norm": 4.4503501042063816, + "learning_rate": 8.90702479338843e-07, + "logits/chosen": -0.0751953125, + "logits/rejected": -0.1904296875, + "logps/chosen": -358.0, + "logps/rejected": -356.0, + "loss": 0.0251, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.1806640625, + "rewards/margins": 10.125, + "rewards/rejected": -9.9375, + "step": 1008 + }, + { + "epoch": 1.2380368098159509, + "grad_norm": 12.812283743763317, + "learning_rate": 8.900826446280992e-07, + "logits/chosen": 0.0169677734375, + "logits/rejected": -0.185546875, + "logps/chosen": -426.0, + "logps/rejected": -380.0, + "loss": 0.0431, + "rewards/accuracies": 0.984375, + "rewards/chosen": 1.40625, + "rewards/margins": 11.1875, + "rewards/rejected": -9.8125, + "step": 1009 + }, + { + "epoch": 1.2392638036809815, + "grad_norm": 6.899754042678334, + "learning_rate": 8.894628099173553e-07, + "logits/chosen": -0.0181884765625, + "logits/rejected": -0.10205078125, + "logps/chosen": -354.0, + "logps/rejected": -374.0, + "loss": 0.03, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.10595703125, + "rewards/margins": 10.0, + "rewards/rejected": -10.0625, + "step": 1010 + }, + { + "epoch": 1.2404907975460122, + "grad_norm": 6.300076629714054, + "learning_rate": 8.888429752066116e-07, + "logits/chosen": -0.0289306640625, + "logits/rejected": -0.21875, + "logps/chosen": -412.0, + "logps/rejected": -390.0, + "loss": 0.0272, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.4375, + "rewards/margins": 10.6875, + "rewards/rejected": -10.25, + "step": 1011 + }, + { + "epoch": 1.2417177914110429, + "grad_norm": 15.956703887603753, + "learning_rate": 8.882231404958678e-07, + "logits/chosen": -0.10107421875, + "logits/rejected": -0.25390625, + "logps/chosen": -384.0, + "logps/rejected": -372.0, + "loss": 0.0733, + "rewards/accuracies": 0.9609375, + "rewards/chosen": 0.462890625, + "rewards/margins": 10.25, + "rewards/rejected": -9.75, + "step": 1012 + }, + { + "epoch": 1.2429447852760735, + "grad_norm": 7.738683663773001, + "learning_rate": 8.87603305785124e-07, + "logits/chosen": 0.01318359375, + "logits/rejected": -0.049560546875, + "logps/chosen": -398.0, + "logps/rejected": -380.0, + "loss": 0.0336, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.11328125, + "rewards/margins": 9.0625, + "rewards/rejected": -9.1875, + "step": 1013 + }, + { + "epoch": 1.2441717791411042, + "grad_norm": 8.908202927372313, + "learning_rate": 8.869834710743802e-07, + "logits/chosen": -0.09765625, + "logits/rejected": -0.2275390625, + "logps/chosen": -386.0, + "logps/rejected": -406.0, + "loss": 0.0403, + "rewards/accuracies": 0.9765625, + "rewards/chosen": -0.00567626953125, + "rewards/margins": 10.0, + "rewards/rejected": -10.0, + "step": 1014 + }, + { + "epoch": 1.2453987730061349, + "grad_norm": 8.437111565369221, + "learning_rate": 8.863636363636364e-07, + "logits/chosen": -0.06689453125, + "logits/rejected": -0.1826171875, + "logps/chosen": -370.0, + "logps/rejected": -388.0, + "loss": 0.036, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 1.21875, + "rewards/margins": 11.6875, + "rewards/rejected": -10.4375, + "step": 1015 + }, + { + "epoch": 1.2466257668711656, + "grad_norm": 19.806315774782945, + "learning_rate": 8.857438016528927e-07, + "logits/chosen": -0.0859375, + "logits/rejected": -0.1669921875, + "logps/chosen": -364.0, + "logps/rejected": -348.0, + "loss": 0.0999, + "rewards/accuracies": 0.9609375, + "rewards/chosen": -0.06787109375, + "rewards/margins": 9.4375, + "rewards/rejected": -9.5, + "step": 1016 + }, + { + "epoch": 1.2478527607361962, + "grad_norm": 1.9621085105816412, + "learning_rate": 8.851239669421488e-07, + "logits/chosen": -0.08740234375, + "logits/rejected": -0.31640625, + "logps/chosen": -404.0, + "logps/rejected": -362.0, + "loss": 0.0089, + "rewards/accuracies": 1.0, + "rewards/chosen": 1.234375, + "rewards/margins": 11.8125, + "rewards/rejected": -10.5625, + "step": 1017 + }, + { + "epoch": 1.2490797546012269, + "grad_norm": 7.315995365487366, + "learning_rate": 8.845041322314049e-07, + "logits/chosen": -0.11865234375, + "logits/rejected": -0.1962890625, + "logps/chosen": -408.0, + "logps/rejected": -416.0, + "loss": 0.0283, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.11572265625, + "rewards/margins": 11.125, + "rewards/rejected": -11.25, + "step": 1018 + }, + { + "epoch": 1.2503067484662576, + "grad_norm": 6.452661280214959, + "learning_rate": 8.838842975206611e-07, + "logits/chosen": -0.15234375, + "logits/rejected": -0.302734375, + "logps/chosen": -376.0, + "logps/rejected": -378.0, + "loss": 0.0318, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.296875, + "rewards/margins": 11.1875, + "rewards/rejected": -10.875, + "step": 1019 + }, + { + "epoch": 1.2515337423312882, + "grad_norm": 4.8866762625917, + "learning_rate": 8.832644628099174e-07, + "logits/chosen": -0.09033203125, + "logits/rejected": -0.193359375, + "logps/chosen": -414.0, + "logps/rejected": -402.0, + "loss": 0.0215, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.33203125, + "rewards/margins": 11.75, + "rewards/rejected": -11.4375, + "step": 1020 + }, + { + "epoch": 1.252760736196319, + "grad_norm": 7.044566658635158, + "learning_rate": 8.826446280991736e-07, + "logits/chosen": -0.05908203125, + "logits/rejected": -0.13671875, + "logps/chosen": -370.0, + "logps/rejected": -378.0, + "loss": 0.0313, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.171875, + "rewards/margins": 11.4375, + "rewards/rejected": -11.25, + "step": 1021 + }, + { + "epoch": 1.2539877300613496, + "grad_norm": 2.074684049190498, + "learning_rate": 8.820247933884298e-07, + "logits/chosen": -0.1708984375, + "logits/rejected": -0.30859375, + "logps/chosen": -420.0, + "logps/rejected": -396.0, + "loss": 0.0134, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.400390625, + "rewards/margins": 11.3125, + "rewards/rejected": -10.9375, + "step": 1022 + }, + { + "epoch": 1.2552147239263804, + "grad_norm": 11.38544236701507, + "learning_rate": 8.81404958677686e-07, + "logits/chosen": -0.142578125, + "logits/rejected": -0.314453125, + "logps/chosen": -402.0, + "logps/rejected": -398.0, + "loss": 0.0566, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.9453125, + "rewards/margins": 11.8125, + "rewards/rejected": -10.875, + "step": 1023 + }, + { + "epoch": 1.2564417177914111, + "grad_norm": 7.64965729920436, + "learning_rate": 8.807851239669422e-07, + "logits/chosen": -0.028564453125, + "logits/rejected": -0.275390625, + "logps/chosen": -410.0, + "logps/rejected": -420.0, + "loss": 0.0254, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.408203125, + "rewards/margins": 11.9375, + "rewards/rejected": -11.5, + "step": 1024 + }, + { + "epoch": 1.2576687116564418, + "grad_norm": 23.522395476453028, + "learning_rate": 8.801652892561984e-07, + "logits/chosen": -0.03466796875, + "logits/rejected": -0.1962890625, + "logps/chosen": -414.0, + "logps/rejected": -418.0, + "loss": 0.1276, + "rewards/accuracies": 0.96875, + "rewards/chosen": 1.0078125, + "rewards/margins": 11.3125, + "rewards/rejected": -10.3125, + "step": 1025 + }, + { + "epoch": 1.2588957055214725, + "grad_norm": 4.975922218453701, + "learning_rate": 8.795454545454545e-07, + "logits/chosen": -0.1318359375, + "logits/rejected": -0.30078125, + "logps/chosen": -370.0, + "logps/rejected": -376.0, + "loss": 0.0313, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.30859375, + "rewards/margins": 10.9375, + "rewards/rejected": -10.625, + "step": 1026 + }, + { + "epoch": 1.2601226993865031, + "grad_norm": 4.952502473285682, + "learning_rate": 8.789256198347107e-07, + "logits/chosen": -0.17578125, + "logits/rejected": -0.41015625, + "logps/chosen": -448.0, + "logps/rejected": -412.0, + "loss": 0.0181, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.6015625, + "rewards/margins": 11.375, + "rewards/rejected": -10.75, + "step": 1027 + }, + { + "epoch": 1.2613496932515338, + "grad_norm": 7.5494432139799015, + "learning_rate": 8.783057851239669e-07, + "logits/chosen": -0.158203125, + "logits/rejected": -0.3125, + "logps/chosen": -396.0, + "logps/rejected": -396.0, + "loss": 0.0222, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 1.2421875, + "rewards/margins": 12.3125, + "rewards/rejected": -11.0625, + "step": 1028 + }, + { + "epoch": 1.2625766871165645, + "grad_norm": 8.079926823970226, + "learning_rate": 8.776859504132232e-07, + "logits/chosen": -0.1083984375, + "logits/rejected": -0.298828125, + "logps/chosen": -390.0, + "logps/rejected": -384.0, + "loss": 0.0315, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.455078125, + "rewards/margins": 10.5, + "rewards/rejected": -10.0, + "step": 1029 + }, + { + "epoch": 1.2638036809815951, + "grad_norm": 6.946808552773218, + "learning_rate": 8.770661157024793e-07, + "logits/chosen": -0.3046875, + "logits/rejected": -0.38671875, + "logps/chosen": -414.0, + "logps/rejected": -410.0, + "loss": 0.0214, + "rewards/accuracies": 1.0, + "rewards/chosen": 1.1796875, + "rewards/margins": 12.125, + "rewards/rejected": -10.9375, + "step": 1030 + }, + { + "epoch": 1.2650306748466258, + "grad_norm": 13.376232318087313, + "learning_rate": 8.764462809917356e-07, + "logits/chosen": -0.0703125, + "logits/rejected": -0.30859375, + "logps/chosen": -382.0, + "logps/rejected": -360.0, + "loss": 0.0746, + "rewards/accuracies": 0.9765625, + "rewards/chosen": 0.55859375, + "rewards/margins": 10.375, + "rewards/rejected": -9.875, + "step": 1031 + }, + { + "epoch": 1.2662576687116565, + "grad_norm": 6.898643809079758, + "learning_rate": 8.758264462809918e-07, + "logits/chosen": 0.042724609375, + "logits/rejected": -0.1669921875, + "logps/chosen": -376.0, + "logps/rejected": -358.0, + "loss": 0.0263, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.76171875, + "rewards/margins": 10.5625, + "rewards/rejected": -9.8125, + "step": 1032 + }, + { + "epoch": 1.2674846625766871, + "grad_norm": 14.180165703858819, + "learning_rate": 8.75206611570248e-07, + "logits/chosen": -0.123046875, + "logits/rejected": -0.33984375, + "logps/chosen": -438.0, + "logps/rejected": -392.0, + "loss": 0.0718, + "rewards/accuracies": 0.9609375, + "rewards/chosen": -0.004547119140625, + "rewards/margins": 10.0, + "rewards/rejected": -10.0625, + "step": 1033 + }, + { + "epoch": 1.2687116564417178, + "grad_norm": 5.086342064289492, + "learning_rate": 8.745867768595042e-07, + "logits/chosen": -0.1572265625, + "logits/rejected": -0.38671875, + "logps/chosen": -464.0, + "logps/rejected": -420.0, + "loss": 0.0192, + "rewards/accuracies": 1.0, + "rewards/chosen": 1.25, + "rewards/margins": 12.1875, + "rewards/rejected": -10.9375, + "step": 1034 + }, + { + "epoch": 1.2699386503067485, + "grad_norm": 14.981826471029674, + "learning_rate": 8.739669421487602e-07, + "logits/chosen": -0.0712890625, + "logits/rejected": -0.279296875, + "logps/chosen": -408.0, + "logps/rejected": -388.0, + "loss": 0.0436, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.8046875, + "rewards/margins": 11.375, + "rewards/rejected": -10.5625, + "step": 1035 + }, + { + "epoch": 1.2711656441717791, + "grad_norm": 11.555096518453992, + "learning_rate": 8.733471074380165e-07, + "logits/chosen": -0.0693359375, + "logits/rejected": -0.20703125, + "logps/chosen": -380.0, + "logps/rejected": -368.0, + "loss": 0.0478, + "rewards/accuracies": 0.96875, + "rewards/chosen": 1.1640625, + "rewards/margins": 10.875, + "rewards/rejected": -9.6875, + "step": 1036 + }, + { + "epoch": 1.2723926380368098, + "grad_norm": 4.796864490403106, + "learning_rate": 8.727272727272727e-07, + "logits/chosen": -0.27734375, + "logits/rejected": -0.41015625, + "logps/chosen": -416.0, + "logps/rejected": -432.0, + "loss": 0.0167, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.984375, + "rewards/margins": 11.5625, + "rewards/rejected": -10.5625, + "step": 1037 + }, + { + "epoch": 1.2736196319018405, + "grad_norm": 7.4606854944626555, + "learning_rate": 8.72107438016529e-07, + "logits/chosen": -0.1103515625, + "logits/rejected": -0.2734375, + "logps/chosen": -382.0, + "logps/rejected": -392.0, + "loss": 0.0268, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.60546875, + "rewards/margins": 11.6875, + "rewards/rejected": -11.0625, + "step": 1038 + }, + { + "epoch": 1.2748466257668711, + "grad_norm": 11.522250663977559, + "learning_rate": 8.714876033057851e-07, + "logits/chosen": -0.146484375, + "logits/rejected": -0.2392578125, + "logps/chosen": -382.0, + "logps/rejected": -414.0, + "loss": 0.0475, + "rewards/accuracies": 0.9765625, + "rewards/chosen": 0.0751953125, + "rewards/margins": 10.375, + "rewards/rejected": -10.3125, + "step": 1039 + }, + { + "epoch": 1.2760736196319018, + "grad_norm": 13.412392481033551, + "learning_rate": 8.708677685950414e-07, + "logits/chosen": -0.18359375, + "logits/rejected": -0.470703125, + "logps/chosen": -426.0, + "logps/rejected": -366.0, + "loss": 0.0344, + "rewards/accuracies": 0.9765625, + "rewards/chosen": 0.7734375, + "rewards/margins": 11.875, + "rewards/rejected": -11.125, + "step": 1040 + }, + { + "epoch": 1.2773006134969325, + "grad_norm": 4.076327393298983, + "learning_rate": 8.702479338842976e-07, + "logits/chosen": -0.1357421875, + "logits/rejected": -0.2734375, + "logps/chosen": -382.0, + "logps/rejected": -378.0, + "loss": 0.013, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.279296875, + "rewards/margins": 11.3125, + "rewards/rejected": -11.0625, + "step": 1041 + }, + { + "epoch": 1.2785276073619631, + "grad_norm": 12.198994819526218, + "learning_rate": 8.696280991735537e-07, + "logits/chosen": -0.3046875, + "logits/rejected": -0.5, + "logps/chosen": -452.0, + "logps/rejected": -434.0, + "loss": 0.0344, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.68359375, + "rewards/margins": 12.6875, + "rewards/rejected": -12.0, + "step": 1042 + }, + { + "epoch": 1.2797546012269938, + "grad_norm": 15.876129225419682, + "learning_rate": 8.6900826446281e-07, + "logits/chosen": -0.10400390625, + "logits/rejected": -0.287109375, + "logps/chosen": -380.0, + "logps/rejected": -386.0, + "loss": 0.0536, + "rewards/accuracies": 0.9765625, + "rewards/chosen": -0.63671875, + "rewards/margins": 10.125, + "rewards/rejected": -10.8125, + "step": 1043 + }, + { + "epoch": 1.2809815950920245, + "grad_norm": 5.177920548454582, + "learning_rate": 8.68388429752066e-07, + "logits/chosen": -0.2041015625, + "logits/rejected": -0.421875, + "logps/chosen": -432.0, + "logps/rejected": -402.0, + "loss": 0.0149, + "rewards/accuracies": 1.0, + "rewards/chosen": 1.1796875, + "rewards/margins": 12.25, + "rewards/rejected": -11.0625, + "step": 1044 + }, + { + "epoch": 1.2822085889570551, + "grad_norm": 8.9147491979472, + "learning_rate": 8.677685950413223e-07, + "logits/chosen": -0.251953125, + "logits/rejected": -0.52734375, + "logps/chosen": -430.0, + "logps/rejected": -374.0, + "loss": 0.0292, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.734375, + "rewards/margins": 11.75, + "rewards/rejected": -11.0, + "step": 1045 + }, + { + "epoch": 1.2834355828220858, + "grad_norm": 6.882145843965057, + "learning_rate": 8.671487603305785e-07, + "logits/chosen": -0.017578125, + "logits/rejected": -0.1689453125, + "logps/chosen": -396.0, + "logps/rejected": -404.0, + "loss": 0.0288, + "rewards/accuracies": 0.984375, + "rewards/chosen": -0.0260009765625, + "rewards/margins": 10.625, + "rewards/rejected": -10.6875, + "step": 1046 + }, + { + "epoch": 1.2846625766871167, + "grad_norm": 12.443133314198835, + "learning_rate": 8.665289256198347e-07, + "logits/chosen": -0.1826171875, + "logits/rejected": -0.330078125, + "logps/chosen": -404.0, + "logps/rejected": -404.0, + "loss": 0.0533, + "rewards/accuracies": 0.9765625, + "rewards/chosen": 0.68359375, + "rewards/margins": 12.25, + "rewards/rejected": -11.5625, + "step": 1047 + }, + { + "epoch": 1.2858895705521474, + "grad_norm": 4.187308341376978, + "learning_rate": 8.659090909090909e-07, + "logits/chosen": -0.0703125, + "logits/rejected": -0.265625, + "logps/chosen": -342.0, + "logps/rejected": -348.0, + "loss": 0.019, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.1494140625, + "rewards/margins": 10.9375, + "rewards/rejected": -10.8125, + "step": 1048 + }, + { + "epoch": 1.287116564417178, + "grad_norm": 12.339404257771985, + "learning_rate": 8.652892561983472e-07, + "logits/chosen": -0.10693359375, + "logits/rejected": -0.326171875, + "logps/chosen": -376.0, + "logps/rejected": -374.0, + "loss": 0.0446, + "rewards/accuracies": 0.9765625, + "rewards/chosen": -0.419921875, + "rewards/margins": 10.4375, + "rewards/rejected": -10.8125, + "step": 1049 + }, + { + "epoch": 1.2883435582822087, + "grad_norm": 7.946831033807572, + "learning_rate": 8.646694214876034e-07, + "logits/chosen": -0.2451171875, + "logits/rejected": -0.384765625, + "logps/chosen": -398.0, + "logps/rejected": -406.0, + "loss": 0.0372, + "rewards/accuracies": 0.984375, + "rewards/chosen": -0.25390625, + "rewards/margins": 11.0625, + "rewards/rejected": -11.3125, + "step": 1050 + }, + { + "epoch": 1.2895705521472394, + "grad_norm": 12.451016602757491, + "learning_rate": 8.640495867768595e-07, + "logits/chosen": -0.279296875, + "logits/rejected": -0.466796875, + "logps/chosen": -420.0, + "logps/rejected": -424.0, + "loss": 0.037, + "rewards/accuracies": 0.9765625, + "rewards/chosen": 0.71484375, + "rewards/margins": 12.6875, + "rewards/rejected": -12.0, + "step": 1051 + }, + { + "epoch": 1.29079754601227, + "grad_norm": 14.98941632692573, + "learning_rate": 8.634297520661158e-07, + "logits/chosen": -0.1328125, + "logits/rejected": -0.3359375, + "logps/chosen": -432.0, + "logps/rejected": -414.0, + "loss": 0.0593, + "rewards/accuracies": 0.9609375, + "rewards/chosen": -0.12353515625, + "rewards/margins": 10.6875, + "rewards/rejected": -10.8125, + "step": 1052 + }, + { + "epoch": 1.2920245398773007, + "grad_norm": 17.525115785642342, + "learning_rate": 8.628099173553718e-07, + "logits/chosen": -0.06640625, + "logits/rejected": -0.287109375, + "logps/chosen": -366.0, + "logps/rejected": -366.0, + "loss": 0.069, + "rewards/accuracies": 0.96875, + "rewards/chosen": 0.26171875, + "rewards/margins": 11.3125, + "rewards/rejected": -11.0625, + "step": 1053 + }, + { + "epoch": 1.2932515337423314, + "grad_norm": 11.652127715235839, + "learning_rate": 8.621900826446281e-07, + "logits/chosen": -0.11767578125, + "logits/rejected": -0.3359375, + "logps/chosen": -402.0, + "logps/rejected": -394.0, + "loss": 0.0503, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.263671875, + "rewards/margins": 11.9375, + "rewards/rejected": -11.6875, + "step": 1054 + }, + { + "epoch": 1.294478527607362, + "grad_norm": 5.171446811615976, + "learning_rate": 8.615702479338843e-07, + "logits/chosen": -0.11962890625, + "logits/rejected": -0.33203125, + "logps/chosen": -386.0, + "logps/rejected": -386.0, + "loss": 0.0178, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.0439453125, + "rewards/margins": 11.6875, + "rewards/rejected": -11.75, + "step": 1055 + }, + { + "epoch": 1.2957055214723927, + "grad_norm": 3.3720116007928245, + "learning_rate": 8.609504132231405e-07, + "logits/chosen": -0.15625, + "logits/rejected": -0.34375, + "logps/chosen": -368.0, + "logps/rejected": -346.0, + "loss": 0.0109, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.0181884765625, + "rewards/margins": 10.875, + "rewards/rejected": -10.875, + "step": 1056 + }, + { + "epoch": 1.2969325153374234, + "grad_norm": 12.84642536946459, + "learning_rate": 8.603305785123967e-07, + "logits/chosen": -0.1298828125, + "logits/rejected": -0.2734375, + "logps/chosen": -404.0, + "logps/rejected": -390.0, + "loss": 0.0402, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.365234375, + "rewards/margins": 11.25, + "rewards/rejected": -10.9375, + "step": 1057 + }, + { + "epoch": 1.298159509202454, + "grad_norm": 9.137246685227527, + "learning_rate": 8.59710743801653e-07, + "logits/chosen": -0.130859375, + "logits/rejected": -0.365234375, + "logps/chosen": -390.0, + "logps/rejected": -388.0, + "loss": 0.0341, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.84765625, + "rewards/margins": 12.375, + "rewards/rejected": -11.5, + "step": 1058 + }, + { + "epoch": 1.2993865030674847, + "grad_norm": 8.50866186346429, + "learning_rate": 8.590909090909092e-07, + "logits/chosen": -0.1845703125, + "logits/rejected": -0.38671875, + "logps/chosen": -416.0, + "logps/rejected": -384.0, + "loss": 0.0325, + "rewards/accuracies": 0.984375, + "rewards/chosen": -0.341796875, + "rewards/margins": 11.3125, + "rewards/rejected": -11.6875, + "step": 1059 + }, + { + "epoch": 1.3006134969325154, + "grad_norm": 8.4621363068171, + "learning_rate": 8.584710743801653e-07, + "logits/chosen": -0.15234375, + "logits/rejected": -0.2421875, + "logps/chosen": -338.0, + "logps/rejected": -350.0, + "loss": 0.0316, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.392578125, + "rewards/margins": 11.3125, + "rewards/rejected": -10.875, + "step": 1060 + }, + { + "epoch": 1.301840490797546, + "grad_norm": 12.798702675578042, + "learning_rate": 8.578512396694216e-07, + "logits/chosen": -0.059326171875, + "logits/rejected": -0.1630859375, + "logps/chosen": -342.0, + "logps/rejected": -372.0, + "loss": 0.0443, + "rewards/accuracies": 0.984375, + "rewards/chosen": -0.052734375, + "rewards/margins": 10.6875, + "rewards/rejected": -10.6875, + "step": 1061 + }, + { + "epoch": 1.3030674846625767, + "grad_norm": 10.362126702943744, + "learning_rate": 8.572314049586776e-07, + "logits/chosen": -0.1015625, + "logits/rejected": -0.2373046875, + "logps/chosen": -422.0, + "logps/rejected": -416.0, + "loss": 0.0494, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.369140625, + "rewards/margins": 11.0625, + "rewards/rejected": -10.6875, + "step": 1062 + }, + { + "epoch": 1.3042944785276074, + "grad_norm": 13.766378342401254, + "learning_rate": 8.566115702479339e-07, + "logits/chosen": -0.1630859375, + "logits/rejected": -0.296875, + "logps/chosen": -370.0, + "logps/rejected": -366.0, + "loss": 0.0643, + "rewards/accuracies": 0.9765625, + "rewards/chosen": -0.228515625, + "rewards/margins": 10.75, + "rewards/rejected": -10.9375, + "step": 1063 + }, + { + "epoch": 1.305521472392638, + "grad_norm": 8.950574541603398, + "learning_rate": 8.559917355371901e-07, + "logits/chosen": -0.1640625, + "logits/rejected": -0.3046875, + "logps/chosen": -386.0, + "logps/rejected": -392.0, + "loss": 0.034, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.057861328125, + "rewards/margins": 11.5, + "rewards/rejected": -11.4375, + "step": 1064 + }, + { + "epoch": 1.3067484662576687, + "grad_norm": 8.164498598328498, + "learning_rate": 8.553719008264463e-07, + "logits/chosen": -0.11279296875, + "logits/rejected": -0.4296875, + "logps/chosen": -448.0, + "logps/rejected": -408.0, + "loss": 0.0283, + "rewards/accuracies": 0.984375, + "rewards/chosen": 1.2734375, + "rewards/margins": 13.3125, + "rewards/rejected": -12.0, + "step": 1065 + }, + { + "epoch": 1.3079754601226994, + "grad_norm": 9.939864149496472, + "learning_rate": 8.547520661157025e-07, + "logits/chosen": -0.1435546875, + "logits/rejected": -0.435546875, + "logps/chosen": -464.0, + "logps/rejected": -430.0, + "loss": 0.0316, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.328125, + "rewards/margins": 11.5, + "rewards/rejected": -11.1875, + "step": 1066 + }, + { + "epoch": 1.30920245398773, + "grad_norm": 5.026276230946228, + "learning_rate": 8.541322314049588e-07, + "logits/chosen": -0.1640625, + "logits/rejected": -0.27734375, + "logps/chosen": -376.0, + "logps/rejected": -386.0, + "loss": 0.0294, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.359375, + "rewards/margins": 9.75, + "rewards/rejected": -10.125, + "step": 1067 + }, + { + "epoch": 1.3104294478527607, + "grad_norm": 11.031737625122464, + "learning_rate": 8.535123966942149e-07, + "logits/chosen": -0.1767578125, + "logits/rejected": -0.32421875, + "logps/chosen": -388.0, + "logps/rejected": -366.0, + "loss": 0.0537, + "rewards/accuracies": 0.9765625, + "rewards/chosen": 0.283203125, + "rewards/margins": 11.3125, + "rewards/rejected": -11.0625, + "step": 1068 + }, + { + "epoch": 1.3116564417177914, + "grad_norm": 8.905100714168142, + "learning_rate": 8.528925619834711e-07, + "logits/chosen": -0.07666015625, + "logits/rejected": -0.3046875, + "logps/chosen": -406.0, + "logps/rejected": -402.0, + "loss": 0.0429, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.48828125, + "rewards/margins": 10.6875, + "rewards/rejected": -10.1875, + "step": 1069 + }, + { + "epoch": 1.312883435582822, + "grad_norm": 11.752525584916722, + "learning_rate": 8.522727272727274e-07, + "logits/chosen": -0.19140625, + "logits/rejected": -0.259765625, + "logps/chosen": -386.0, + "logps/rejected": -406.0, + "loss": 0.0348, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.96484375, + "rewards/margins": 12.125, + "rewards/rejected": -11.1875, + "step": 1070 + }, + { + "epoch": 1.3141104294478527, + "grad_norm": 12.851082658918303, + "learning_rate": 8.516528925619834e-07, + "logits/chosen": -0.2255859375, + "logits/rejected": -0.408203125, + "logps/chosen": -392.0, + "logps/rejected": -388.0, + "loss": 0.0656, + "rewards/accuracies": 0.9765625, + "rewards/chosen": 0.4375, + "rewards/margins": 11.25, + "rewards/rejected": -10.8125, + "step": 1071 + }, + { + "epoch": 1.3153374233128834, + "grad_norm": 9.860743385521058, + "learning_rate": 8.510330578512397e-07, + "logits/chosen": -0.002685546875, + "logits/rejected": -0.197265625, + "logps/chosen": -408.0, + "logps/rejected": -398.0, + "loss": 0.0258, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.66015625, + "rewards/margins": 11.125, + "rewards/rejected": -10.4375, + "step": 1072 + }, + { + "epoch": 1.316564417177914, + "grad_norm": 8.94951495923183, + "learning_rate": 8.504132231404958e-07, + "logits/chosen": -0.1982421875, + "logits/rejected": -0.30859375, + "logps/chosen": -330.0, + "logps/rejected": -372.0, + "loss": 0.0282, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.0400390625, + "rewards/margins": 10.25, + "rewards/rejected": -10.1875, + "step": 1073 + }, + { + "epoch": 1.3177914110429447, + "grad_norm": 5.703540754764619, + "learning_rate": 8.497933884297521e-07, + "logits/chosen": -0.2099609375, + "logits/rejected": -0.3203125, + "logps/chosen": -390.0, + "logps/rejected": -388.0, + "loss": 0.0261, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.291015625, + "rewards/margins": 11.375, + "rewards/rejected": -11.125, + "step": 1074 + }, + { + "epoch": 1.3190184049079754, + "grad_norm": 9.876644774460352, + "learning_rate": 8.491735537190083e-07, + "logits/chosen": -0.060791015625, + "logits/rejected": -0.224609375, + "logps/chosen": -398.0, + "logps/rejected": -396.0, + "loss": 0.0274, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.3046875, + "rewards/margins": 11.0625, + "rewards/rejected": -10.75, + "step": 1075 + }, + { + "epoch": 1.320245398773006, + "grad_norm": 4.967869121366148, + "learning_rate": 8.485537190082646e-07, + "logits/chosen": -0.037353515625, + "logits/rejected": -0.123046875, + "logps/chosen": -400.0, + "logps/rejected": -398.0, + "loss": 0.0212, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.28125, + "rewards/margins": 11.4375, + "rewards/rejected": -11.1875, + "step": 1076 + }, + { + "epoch": 1.3214723926380367, + "grad_norm": 9.322961880086396, + "learning_rate": 8.479338842975207e-07, + "logits/chosen": -0.2099609375, + "logits/rejected": -0.3359375, + "logps/chosen": -302.0, + "logps/rejected": -312.0, + "loss": 0.0452, + "rewards/accuracies": 0.984375, + "rewards/chosen": -0.3984375, + "rewards/margins": 9.875, + "rewards/rejected": -10.25, + "step": 1077 + }, + { + "epoch": 1.3226993865030674, + "grad_norm": 5.738661119438272, + "learning_rate": 8.473140495867769e-07, + "logits/chosen": -0.049072265625, + "logits/rejected": -0.158203125, + "logps/chosen": -386.0, + "logps/rejected": -396.0, + "loss": 0.0334, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.5078125, + "rewards/margins": 9.3125, + "rewards/rejected": -9.8125, + "step": 1078 + }, + { + "epoch": 1.323926380368098, + "grad_norm": 6.333892508394478, + "learning_rate": 8.466942148760332e-07, + "logits/chosen": -0.1201171875, + "logits/rejected": -0.259765625, + "logps/chosen": -384.0, + "logps/rejected": -388.0, + "loss": 0.0328, + "rewards/accuracies": 0.9765625, + "rewards/chosen": 0.16796875, + "rewards/margins": 11.6875, + "rewards/rejected": -11.5, + "step": 1079 + }, + { + "epoch": 1.3251533742331287, + "grad_norm": 13.790501614554328, + "learning_rate": 8.460743801652892e-07, + "logits/chosen": -0.0654296875, + "logits/rejected": -0.263671875, + "logps/chosen": -366.0, + "logps/rejected": -358.0, + "loss": 0.047, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.0322265625, + "rewards/margins": 10.0625, + "rewards/rejected": -10.0625, + "step": 1080 + }, + { + "epoch": 1.3263803680981594, + "grad_norm": 7.653539387358885, + "learning_rate": 8.454545454545455e-07, + "logits/chosen": -0.1005859375, + "logits/rejected": -0.259765625, + "logps/chosen": -392.0, + "logps/rejected": -396.0, + "loss": 0.0364, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.1640625, + "rewards/margins": 10.9375, + "rewards/rejected": -10.75, + "step": 1081 + }, + { + "epoch": 1.32760736196319, + "grad_norm": 10.410957750314225, + "learning_rate": 8.448347107438016e-07, + "logits/chosen": -0.2021484375, + "logits/rejected": -0.435546875, + "logps/chosen": -412.0, + "logps/rejected": -366.0, + "loss": 0.0374, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.2119140625, + "rewards/margins": 10.375, + "rewards/rejected": -10.125, + "step": 1082 + }, + { + "epoch": 1.3288343558282207, + "grad_norm": 5.66314951524206, + "learning_rate": 8.442148760330579e-07, + "logits/chosen": -0.1025390625, + "logits/rejected": -0.283203125, + "logps/chosen": -372.0, + "logps/rejected": -328.0, + "loss": 0.0214, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.142578125, + "rewards/margins": 9.875, + "rewards/rejected": -10.0, + "step": 1083 + }, + { + "epoch": 1.3300613496932514, + "grad_norm": 11.42793523481754, + "learning_rate": 8.435950413223141e-07, + "logits/chosen": -0.15625, + "logits/rejected": -0.16015625, + "logps/chosen": -354.0, + "logps/rejected": -362.0, + "loss": 0.0449, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.1640625, + "rewards/margins": 10.625, + "rewards/rejected": -10.5, + "step": 1084 + }, + { + "epoch": 1.331288343558282, + "grad_norm": 6.950168336723294, + "learning_rate": 8.429752066115703e-07, + "logits/chosen": -0.033203125, + "logits/rejected": -0.181640625, + "logps/chosen": -386.0, + "logps/rejected": -376.0, + "loss": 0.0221, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.125, + "rewards/margins": 10.125, + "rewards/rejected": -10.25, + "step": 1085 + }, + { + "epoch": 1.332515337423313, + "grad_norm": 7.903285776855893, + "learning_rate": 8.423553719008265e-07, + "logits/chosen": -0.10107421875, + "logits/rejected": -0.228515625, + "logps/chosen": -316.0, + "logps/rejected": -344.0, + "loss": 0.0384, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.365234375, + "rewards/margins": 10.75, + "rewards/rejected": -10.375, + "step": 1086 + }, + { + "epoch": 1.3337423312883436, + "grad_norm": 14.294247050939283, + "learning_rate": 8.417355371900827e-07, + "logits/chosen": -0.2099609375, + "logits/rejected": -0.33984375, + "logps/chosen": -322.0, + "logps/rejected": -376.0, + "loss": 0.0381, + "rewards/accuracies": 0.984375, + "rewards/chosen": -0.251953125, + "rewards/margins": 10.375, + "rewards/rejected": -10.625, + "step": 1087 + }, + { + "epoch": 1.3349693251533743, + "grad_norm": 12.270443455514881, + "learning_rate": 8.41115702479339e-07, + "logits/chosen": -0.1376953125, + "logits/rejected": -0.365234375, + "logps/chosen": -446.0, + "logps/rejected": -378.0, + "loss": 0.0352, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.65234375, + "rewards/margins": 11.25, + "rewards/rejected": -10.625, + "step": 1088 + }, + { + "epoch": 1.336196319018405, + "grad_norm": 9.38631227726379, + "learning_rate": 8.40495867768595e-07, + "logits/chosen": -0.1884765625, + "logits/rejected": -0.34375, + "logps/chosen": -390.0, + "logps/rejected": -360.0, + "loss": 0.039, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.6015625, + "rewards/margins": 9.875, + "rewards/rejected": -10.5, + "step": 1089 + }, + { + "epoch": 1.3374233128834356, + "grad_norm": 13.010546372728042, + "learning_rate": 8.398760330578512e-07, + "logits/chosen": -0.1357421875, + "logits/rejected": -0.33984375, + "logps/chosen": -374.0, + "logps/rejected": -352.0, + "loss": 0.0449, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.1728515625, + "rewards/margins": 10.125, + "rewards/rejected": -10.25, + "step": 1090 + }, + { + "epoch": 1.3386503067484663, + "grad_norm": 17.2682088486596, + "learning_rate": 8.392561983471074e-07, + "logits/chosen": 0.013427734375, + "logits/rejected": -0.1552734375, + "logps/chosen": -424.0, + "logps/rejected": -368.0, + "loss": 0.0874, + "rewards/accuracies": 0.9765625, + "rewards/chosen": 0.1650390625, + "rewards/margins": 10.125, + "rewards/rejected": -9.9375, + "step": 1091 + }, + { + "epoch": 1.339877300613497, + "grad_norm": 6.844436331855588, + "learning_rate": 8.386363636363637e-07, + "logits/chosen": -0.1455078125, + "logits/rejected": -0.31640625, + "logps/chosen": -414.0, + "logps/rejected": -398.0, + "loss": 0.0256, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.484375, + "rewards/margins": 10.75, + "rewards/rejected": -10.25, + "step": 1092 + }, + { + "epoch": 1.3411042944785276, + "grad_norm": 5.767510463883874, + "learning_rate": 8.380165289256199e-07, + "logits/chosen": -0.2041015625, + "logits/rejected": -0.34375, + "logps/chosen": -378.0, + "logps/rejected": -372.0, + "loss": 0.0207, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.328125, + "rewards/margins": 10.25, + "rewards/rejected": -9.9375, + "step": 1093 + }, + { + "epoch": 1.3423312883435583, + "grad_norm": 8.158470602909508, + "learning_rate": 8.37396694214876e-07, + "logits/chosen": -0.08251953125, + "logits/rejected": -0.30859375, + "logps/chosen": -426.0, + "logps/rejected": -420.0, + "loss": 0.0291, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.546875, + "rewards/margins": 10.875, + "rewards/rejected": -10.3125, + "step": 1094 + }, + { + "epoch": 1.343558282208589, + "grad_norm": 5.508962768344179, + "learning_rate": 8.367768595041323e-07, + "logits/chosen": -0.138671875, + "logits/rejected": -0.349609375, + "logps/chosen": -446.0, + "logps/rejected": -430.0, + "loss": 0.0202, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.890625, + "rewards/margins": 12.4375, + "rewards/rejected": -11.5, + "step": 1095 + }, + { + "epoch": 1.3447852760736196, + "grad_norm": 18.396109473702488, + "learning_rate": 8.361570247933885e-07, + "logits/chosen": -0.146484375, + "logits/rejected": -0.279296875, + "logps/chosen": -382.0, + "logps/rejected": -374.0, + "loss": 0.0744, + "rewards/accuracies": 0.9609375, + "rewards/chosen": -0.091796875, + "rewards/margins": 9.8125, + "rewards/rejected": -9.9375, + "step": 1096 + }, + { + "epoch": 1.3460122699386503, + "grad_norm": 4.9008028557309355, + "learning_rate": 8.355371900826447e-07, + "logits/chosen": -0.1015625, + "logits/rejected": -0.279296875, + "logps/chosen": -432.0, + "logps/rejected": -380.0, + "loss": 0.0189, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.66015625, + "rewards/margins": 10.6875, + "rewards/rejected": -10.0625, + "step": 1097 + }, + { + "epoch": 1.347239263803681, + "grad_norm": 17.180940504647815, + "learning_rate": 8.349173553719008e-07, + "logits/chosen": -0.0299072265625, + "logits/rejected": -0.1826171875, + "logps/chosen": -364.0, + "logps/rejected": -384.0, + "loss": 0.0604, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.1845703125, + "rewards/margins": 10.8125, + "rewards/rejected": -10.625, + "step": 1098 + }, + { + "epoch": 1.3484662576687116, + "grad_norm": 15.090380221292753, + "learning_rate": 8.34297520661157e-07, + "logits/chosen": -0.09619140625, + "logits/rejected": -0.2080078125, + "logps/chosen": -354.0, + "logps/rejected": -344.0, + "loss": 0.0831, + "rewards/accuracies": 0.984375, + "rewards/chosen": -0.130859375, + "rewards/margins": 9.8125, + "rewards/rejected": -9.9375, + "step": 1099 + }, + { + "epoch": 1.3496932515337423, + "grad_norm": 8.957606377667311, + "learning_rate": 8.336776859504132e-07, + "logits/chosen": -0.1572265625, + "logits/rejected": -0.37890625, + "logps/chosen": -406.0, + "logps/rejected": -374.0, + "loss": 0.0292, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.92578125, + "rewards/margins": 10.875, + "rewards/rejected": -9.9375, + "step": 1100 + }, + { + "epoch": 1.350920245398773, + "grad_norm": 12.12291867951929, + "learning_rate": 8.330578512396695e-07, + "logits/chosen": 0.0400390625, + "logits/rejected": -0.1171875, + "logps/chosen": -370.0, + "logps/rejected": -348.0, + "loss": 0.0654, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.4921875, + "rewards/margins": 10.75, + "rewards/rejected": -10.25, + "step": 1101 + }, + { + "epoch": 1.3521472392638036, + "grad_norm": 18.332708975287847, + "learning_rate": 8.324380165289256e-07, + "logits/chosen": -0.251953125, + "logits/rejected": -0.431640625, + "logps/chosen": -370.0, + "logps/rejected": -352.0, + "loss": 0.0716, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.671875, + "rewards/margins": 11.75, + "rewards/rejected": -11.0625, + "step": 1102 + }, + { + "epoch": 1.3533742331288343, + "grad_norm": 10.7430697279738, + "learning_rate": 8.318181818181819e-07, + "logits/chosen": -0.119140625, + "logits/rejected": -0.275390625, + "logps/chosen": -416.0, + "logps/rejected": -400.0, + "loss": 0.0442, + "rewards/accuracies": 0.9765625, + "rewards/chosen": 0.3984375, + "rewards/margins": 10.8125, + "rewards/rejected": -10.4375, + "step": 1103 + }, + { + "epoch": 1.354601226993865, + "grad_norm": 12.033995818293517, + "learning_rate": 8.311983471074381e-07, + "logits/chosen": -0.078125, + "logits/rejected": -0.2265625, + "logps/chosen": -364.0, + "logps/rejected": -374.0, + "loss": 0.0536, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.4296875, + "rewards/margins": 10.375, + "rewards/rejected": -9.9375, + "step": 1104 + }, + { + "epoch": 1.3558282208588956, + "grad_norm": 4.495393110896528, + "learning_rate": 8.305785123966943e-07, + "logits/chosen": -0.2177734375, + "logits/rejected": -0.361328125, + "logps/chosen": -408.0, + "logps/rejected": -414.0, + "loss": 0.0158, + "rewards/accuracies": 1.0, + "rewards/chosen": 1.09375, + "rewards/margins": 11.3125, + "rewards/rejected": -10.25, + "step": 1105 + }, + { + "epoch": 1.3570552147239263, + "grad_norm": 10.27676045940915, + "learning_rate": 8.299586776859504e-07, + "logits/chosen": -0.150390625, + "logits/rejected": -0.263671875, + "logps/chosen": -360.0, + "logps/rejected": -374.0, + "loss": 0.0323, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.177734375, + "rewards/margins": 10.6875, + "rewards/rejected": -10.5, + "step": 1106 + }, + { + "epoch": 1.358282208588957, + "grad_norm": 15.415657480989857, + "learning_rate": 8.293388429752065e-07, + "logits/chosen": 0.03759765625, + "logits/rejected": -0.07861328125, + "logps/chosen": -402.0, + "logps/rejected": -374.0, + "loss": 0.0477, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.76953125, + "rewards/margins": 10.3125, + "rewards/rejected": -9.5625, + "step": 1107 + }, + { + "epoch": 1.3595092024539877, + "grad_norm": 5.235763636929729, + "learning_rate": 8.287190082644628e-07, + "logits/chosen": -0.1552734375, + "logits/rejected": -0.3359375, + "logps/chosen": -440.0, + "logps/rejected": -402.0, + "loss": 0.0197, + "rewards/accuracies": 1.0, + "rewards/chosen": 1.1484375, + "rewards/margins": 12.1875, + "rewards/rejected": -11.0, + "step": 1108 + }, + { + "epoch": 1.3607361963190185, + "grad_norm": 10.622377911792848, + "learning_rate": 8.28099173553719e-07, + "logits/chosen": -0.0361328125, + "logits/rejected": -0.1787109375, + "logps/chosen": -400.0, + "logps/rejected": -382.0, + "loss": 0.0409, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.6328125, + "rewards/margins": 10.5, + "rewards/rejected": -9.875, + "step": 1109 + }, + { + "epoch": 1.3619631901840492, + "grad_norm": 15.219013654594981, + "learning_rate": 8.274793388429753e-07, + "logits/chosen": -0.11572265625, + "logits/rejected": -0.330078125, + "logps/chosen": -398.0, + "logps/rejected": -392.0, + "loss": 0.0485, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.482421875, + "rewards/margins": 11.4375, + "rewards/rejected": -10.9375, + "step": 1110 + }, + { + "epoch": 1.3631901840490799, + "grad_norm": 5.052963784815555, + "learning_rate": 8.268595041322314e-07, + "logits/chosen": -0.150390625, + "logits/rejected": -0.333984375, + "logps/chosen": -432.0, + "logps/rejected": -380.0, + "loss": 0.0166, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.6640625, + "rewards/margins": 11.3125, + "rewards/rejected": -10.625, + "step": 1111 + }, + { + "epoch": 1.3644171779141105, + "grad_norm": 13.661073432102611, + "learning_rate": 8.262396694214877e-07, + "logits/chosen": -0.138671875, + "logits/rejected": -0.23046875, + "logps/chosen": -418.0, + "logps/rejected": -406.0, + "loss": 0.0562, + "rewards/accuracies": 0.96875, + "rewards/chosen": 0.298828125, + "rewards/margins": 11.0625, + "rewards/rejected": -10.75, + "step": 1112 + }, + { + "epoch": 1.3656441717791412, + "grad_norm": 11.164562514831367, + "learning_rate": 8.256198347107439e-07, + "logits/chosen": -0.10888671875, + "logits/rejected": -0.16015625, + "logps/chosen": -356.0, + "logps/rejected": -360.0, + "loss": 0.0353, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.263671875, + "rewards/margins": 10.1875, + "rewards/rejected": -9.9375, + "step": 1113 + }, + { + "epoch": 1.3668711656441719, + "grad_norm": 12.65975505597321, + "learning_rate": 8.25e-07, + "logits/chosen": -0.25390625, + "logits/rejected": -0.3671875, + "logps/chosen": -344.0, + "logps/rejected": -358.0, + "loss": 0.0419, + "rewards/accuracies": 0.9765625, + "rewards/chosen": -0.45703125, + "rewards/margins": 10.0625, + "rewards/rejected": -10.5, + "step": 1114 + }, + { + "epoch": 1.3680981595092025, + "grad_norm": 9.32147436119456, + "learning_rate": 8.243801652892562e-07, + "logits/chosen": -0.1396484375, + "logits/rejected": -0.25, + "logps/chosen": -376.0, + "logps/rejected": -368.0, + "loss": 0.0374, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.388671875, + "rewards/margins": 11.1875, + "rewards/rejected": -10.8125, + "step": 1115 + }, + { + "epoch": 1.3693251533742332, + "grad_norm": 7.035037640219457, + "learning_rate": 8.237603305785123e-07, + "logits/chosen": -0.06982421875, + "logits/rejected": -0.2490234375, + "logps/chosen": -430.0, + "logps/rejected": -402.0, + "loss": 0.0362, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.326171875, + "rewards/margins": 12.0, + "rewards/rejected": -11.6875, + "step": 1116 + }, + { + "epoch": 1.3705521472392639, + "grad_norm": 7.072591615437579, + "learning_rate": 8.231404958677686e-07, + "logits/chosen": -0.251953125, + "logits/rejected": -0.47265625, + "logps/chosen": -434.0, + "logps/rejected": -400.0, + "loss": 0.028, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.671875, + "rewards/margins": 11.875, + "rewards/rejected": -11.1875, + "step": 1117 + }, + { + "epoch": 1.3717791411042946, + "grad_norm": 15.059088219830123, + "learning_rate": 8.225206611570248e-07, + "logits/chosen": -0.056884765625, + "logits/rejected": -0.2294921875, + "logps/chosen": -380.0, + "logps/rejected": -382.0, + "loss": 0.0791, + "rewards/accuracies": 0.9609375, + "rewards/chosen": -0.40234375, + "rewards/margins": 10.5, + "rewards/rejected": -10.875, + "step": 1118 + }, + { + "epoch": 1.3730061349693252, + "grad_norm": 7.182852167794888, + "learning_rate": 8.21900826446281e-07, + "logits/chosen": -0.150390625, + "logits/rejected": -0.328125, + "logps/chosen": -424.0, + "logps/rejected": -420.0, + "loss": 0.0365, + "rewards/accuracies": 0.984375, + "rewards/chosen": -0.271484375, + "rewards/margins": 10.6875, + "rewards/rejected": -11.0, + "step": 1119 + }, + { + "epoch": 1.3742331288343559, + "grad_norm": 6.6804769663365855, + "learning_rate": 8.212809917355372e-07, + "logits/chosen": -0.1103515625, + "logits/rejected": -0.400390625, + "logps/chosen": -430.0, + "logps/rejected": -398.0, + "loss": 0.0239, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.58984375, + "rewards/margins": 12.0, + "rewards/rejected": -11.4375, + "step": 1120 + }, + { + "epoch": 1.3754601226993866, + "grad_norm": 16.092490568103823, + "learning_rate": 8.206611570247935e-07, + "logits/chosen": -0.216796875, + "logits/rejected": -0.375, + "logps/chosen": -444.0, + "logps/rejected": -406.0, + "loss": 0.0841, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.1826171875, + "rewards/margins": 11.3125, + "rewards/rejected": -11.125, + "step": 1121 + }, + { + "epoch": 1.3766871165644172, + "grad_norm": 45.62488030196555, + "learning_rate": 8.200413223140497e-07, + "logits/chosen": -0.2314453125, + "logits/rejected": -0.361328125, + "logps/chosen": -372.0, + "logps/rejected": -374.0, + "loss": 0.1206, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -1.2421875, + "rewards/margins": 10.1875, + "rewards/rejected": -11.375, + "step": 1122 + }, + { + "epoch": 1.377914110429448, + "grad_norm": 7.987049229922856, + "learning_rate": 8.194214876033058e-07, + "logits/chosen": -0.23046875, + "logits/rejected": -0.380859375, + "logps/chosen": -446.0, + "logps/rejected": -384.0, + "loss": 0.0291, + "rewards/accuracies": 0.984375, + "rewards/chosen": -0.59765625, + "rewards/margins": 11.0, + "rewards/rejected": -11.625, + "step": 1123 + }, + { + "epoch": 1.3791411042944786, + "grad_norm": 10.84925435010219, + "learning_rate": 8.188016528925619e-07, + "logits/chosen": -0.1494140625, + "logits/rejected": -0.33203125, + "logps/chosen": -398.0, + "logps/rejected": -372.0, + "loss": 0.045, + "rewards/accuracies": 0.9765625, + "rewards/chosen": -0.53515625, + "rewards/margins": 11.4375, + "rewards/rejected": -12.0, + "step": 1124 + }, + { + "epoch": 1.3803680981595092, + "grad_norm": 12.438670127459194, + "learning_rate": 8.181818181818181e-07, + "logits/chosen": -0.2197265625, + "logits/rejected": -0.3671875, + "logps/chosen": -388.0, + "logps/rejected": -382.0, + "loss": 0.0514, + "rewards/accuracies": 0.96875, + "rewards/chosen": -0.70703125, + "rewards/margins": 10.6875, + "rewards/rejected": -11.4375, + "step": 1125 + }, + { + "epoch": 1.38159509202454, + "grad_norm": 7.702209911740994, + "learning_rate": 8.175619834710744e-07, + "logits/chosen": -0.21875, + "logits/rejected": -0.345703125, + "logps/chosen": -378.0, + "logps/rejected": -372.0, + "loss": 0.0244, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.5, + "rewards/margins": 11.4375, + "rewards/rejected": -11.9375, + "step": 1126 + }, + { + "epoch": 1.3828220858895706, + "grad_norm": 9.98545953456932, + "learning_rate": 8.169421487603306e-07, + "logits/chosen": -0.177734375, + "logits/rejected": -0.392578125, + "logps/chosen": -452.0, + "logps/rejected": -438.0, + "loss": 0.0506, + "rewards/accuracies": 0.9765625, + "rewards/chosen": 0.515625, + "rewards/margins": 12.0625, + "rewards/rejected": -11.5625, + "step": 1127 + }, + { + "epoch": 1.3840490797546012, + "grad_norm": 14.966530393829544, + "learning_rate": 8.163223140495868e-07, + "logits/chosen": -0.1396484375, + "logits/rejected": -0.294921875, + "logps/chosen": -398.0, + "logps/rejected": -416.0, + "loss": 0.0573, + "rewards/accuracies": 0.96875, + "rewards/chosen": -0.83203125, + "rewards/margins": 10.0625, + "rewards/rejected": -10.875, + "step": 1128 + }, + { + "epoch": 1.385276073619632, + "grad_norm": 6.960083817877731, + "learning_rate": 8.15702479338843e-07, + "logits/chosen": -0.0244140625, + "logits/rejected": -0.166015625, + "logps/chosen": -400.0, + "logps/rejected": -372.0, + "loss": 0.022, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.29296875, + "rewards/margins": 10.625, + "rewards/rejected": -10.9375, + "step": 1129 + }, + { + "epoch": 1.3865030674846626, + "grad_norm": 15.732715609813184, + "learning_rate": 8.150826446280993e-07, + "logits/chosen": 0.0286865234375, + "logits/rejected": -0.1923828125, + "logps/chosen": -438.0, + "logps/rejected": -408.0, + "loss": 0.0498, + "rewards/accuracies": 0.984375, + "rewards/chosen": -0.361328125, + "rewards/margins": 10.5625, + "rewards/rejected": -10.9375, + "step": 1130 + }, + { + "epoch": 1.3877300613496932, + "grad_norm": 5.221583127777315, + "learning_rate": 8.144628099173554e-07, + "logits/chosen": -0.232421875, + "logits/rejected": -0.5234375, + "logps/chosen": -434.0, + "logps/rejected": -400.0, + "loss": 0.0197, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.04541015625, + "rewards/margins": 10.9375, + "rewards/rejected": -10.9375, + "step": 1131 + }, + { + "epoch": 1.388957055214724, + "grad_norm": 12.813444795393385, + "learning_rate": 8.138429752066116e-07, + "logits/chosen": -0.134765625, + "logits/rejected": -0.36328125, + "logps/chosen": -494.0, + "logps/rejected": -408.0, + "loss": 0.0313, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.890625, + "rewards/margins": 11.3125, + "rewards/rejected": -10.375, + "step": 1132 + }, + { + "epoch": 1.3901840490797546, + "grad_norm": 15.660321100678845, + "learning_rate": 8.132231404958677e-07, + "logits/chosen": -0.2099609375, + "logits/rejected": -0.34765625, + "logps/chosen": -410.0, + "logps/rejected": -410.0, + "loss": 0.0504, + "rewards/accuracies": 0.9765625, + "rewards/chosen": 0.72265625, + "rewards/margins": 11.875, + "rewards/rejected": -11.125, + "step": 1133 + }, + { + "epoch": 1.3914110429447852, + "grad_norm": 7.9073668160942745, + "learning_rate": 8.126033057851239e-07, + "logits/chosen": -0.1962890625, + "logits/rejected": -0.404296875, + "logps/chosen": -324.0, + "logps/rejected": -340.0, + "loss": 0.0393, + "rewards/accuracies": 0.984375, + "rewards/chosen": -0.91015625, + "rewards/margins": 10.0625, + "rewards/rejected": -10.9375, + "step": 1134 + }, + { + "epoch": 1.392638036809816, + "grad_norm": 7.004247584369281, + "learning_rate": 8.119834710743802e-07, + "logits/chosen": -0.146484375, + "logits/rejected": -0.3359375, + "logps/chosen": -410.0, + "logps/rejected": -374.0, + "loss": 0.0317, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.69140625, + "rewards/margins": 11.5625, + "rewards/rejected": -10.875, + "step": 1135 + }, + { + "epoch": 1.3938650306748466, + "grad_norm": 13.652544230864027, + "learning_rate": 8.113636363636363e-07, + "logits/chosen": -0.29296875, + "logits/rejected": -0.4375, + "logps/chosen": -466.0, + "logps/rejected": -408.0, + "loss": 0.0464, + "rewards/accuracies": 0.96875, + "rewards/chosen": 1.1796875, + "rewards/margins": 12.125, + "rewards/rejected": -10.9375, + "step": 1136 + }, + { + "epoch": 1.3950920245398772, + "grad_norm": 13.746874338763654, + "learning_rate": 8.107438016528926e-07, + "logits/chosen": -0.2314453125, + "logits/rejected": -0.48046875, + "logps/chosen": -466.0, + "logps/rejected": -406.0, + "loss": 0.0681, + "rewards/accuracies": 0.9765625, + "rewards/chosen": 0.345703125, + "rewards/margins": 10.9375, + "rewards/rejected": -10.5625, + "step": 1137 + }, + { + "epoch": 1.396319018404908, + "grad_norm": 13.137416168579717, + "learning_rate": 8.101239669421488e-07, + "logits/chosen": -0.05126953125, + "logits/rejected": -0.25, + "logps/chosen": -386.0, + "logps/rejected": -356.0, + "loss": 0.0423, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.051513671875, + "rewards/margins": 9.875, + "rewards/rejected": -9.8125, + "step": 1138 + }, + { + "epoch": 1.3975460122699386, + "grad_norm": 14.313209901423209, + "learning_rate": 8.095041322314051e-07, + "logits/chosen": -0.2001953125, + "logits/rejected": -0.404296875, + "logps/chosen": -360.0, + "logps/rejected": -368.0, + "loss": 0.0629, + "rewards/accuracies": 0.96875, + "rewards/chosen": 0.2392578125, + "rewards/margins": 11.1875, + "rewards/rejected": -10.9375, + "step": 1139 + }, + { + "epoch": 1.3987730061349692, + "grad_norm": 12.684020520423767, + "learning_rate": 8.088842975206612e-07, + "logits/chosen": -0.10498046875, + "logits/rejected": -0.232421875, + "logps/chosen": -352.0, + "logps/rejected": -360.0, + "loss": 0.0614, + "rewards/accuracies": 0.984375, + "rewards/chosen": -0.2294921875, + "rewards/margins": 10.625, + "rewards/rejected": -10.8125, + "step": 1140 + }, + { + "epoch": 1.4, + "grad_norm": 8.366317687764122, + "learning_rate": 8.082644628099174e-07, + "logits/chosen": -0.1689453125, + "logits/rejected": -0.396484375, + "logps/chosen": -354.0, + "logps/rejected": -356.0, + "loss": 0.0322, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.154296875, + "rewards/margins": 10.4375, + "rewards/rejected": -10.25, + "step": 1141 + }, + { + "epoch": 1.4012269938650306, + "grad_norm": 9.78962851865542, + "learning_rate": 8.076446280991735e-07, + "logits/chosen": -0.057861328125, + "logits/rejected": -0.2001953125, + "logps/chosen": -346.0, + "logps/rejected": -382.0, + "loss": 0.0316, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.34765625, + "rewards/margins": 11.0, + "rewards/rejected": -10.625, + "step": 1142 + }, + { + "epoch": 1.4024539877300612, + "grad_norm": 15.82156683988277, + "learning_rate": 8.070247933884297e-07, + "logits/chosen": -0.1328125, + "logits/rejected": -0.248046875, + "logps/chosen": -396.0, + "logps/rejected": -416.0, + "loss": 0.0618, + "rewards/accuracies": 0.9765625, + "rewards/chosen": -0.51953125, + "rewards/margins": 10.75, + "rewards/rejected": -11.25, + "step": 1143 + }, + { + "epoch": 1.403680981595092, + "grad_norm": 17.350320312048925, + "learning_rate": 8.06404958677686e-07, + "logits/chosen": -0.255859375, + "logits/rejected": -0.38671875, + "logps/chosen": -396.0, + "logps/rejected": -388.0, + "loss": 0.0738, + "rewards/accuracies": 0.984375, + "rewards/chosen": -0.169921875, + "rewards/margins": 10.9375, + "rewards/rejected": -11.125, + "step": 1144 + }, + { + "epoch": 1.4049079754601226, + "grad_norm": 13.43693939499879, + "learning_rate": 8.057851239669421e-07, + "logits/chosen": -0.1572265625, + "logits/rejected": -0.333984375, + "logps/chosen": -456.0, + "logps/rejected": -434.0, + "loss": 0.0376, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.86328125, + "rewards/margins": 11.8125, + "rewards/rejected": -10.9375, + "step": 1145 + }, + { + "epoch": 1.4061349693251532, + "grad_norm": 10.84486528809094, + "learning_rate": 8.051652892561984e-07, + "logits/chosen": -0.046875, + "logits/rejected": -0.16796875, + "logps/chosen": -376.0, + "logps/rejected": -400.0, + "loss": 0.0344, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.05322265625, + "rewards/margins": 11.625, + "rewards/rejected": -11.6875, + "step": 1146 + }, + { + "epoch": 1.407361963190184, + "grad_norm": 7.584140894053648, + "learning_rate": 8.045454545454546e-07, + "logits/chosen": -0.1572265625, + "logits/rejected": -0.361328125, + "logps/chosen": -394.0, + "logps/rejected": -386.0, + "loss": 0.0283, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.5, + "rewards/margins": 12.125, + "rewards/rejected": -11.625, + "step": 1147 + }, + { + "epoch": 1.4085889570552146, + "grad_norm": 9.922196775746215, + "learning_rate": 8.039256198347108e-07, + "logits/chosen": -0.0849609375, + "logits/rejected": -0.2353515625, + "logps/chosen": -370.0, + "logps/rejected": -378.0, + "loss": 0.0426, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.052978515625, + "rewards/margins": 11.3125, + "rewards/rejected": -11.25, + "step": 1148 + }, + { + "epoch": 1.4098159509202455, + "grad_norm": 6.148272524003038, + "learning_rate": 8.03305785123967e-07, + "logits/chosen": -0.140625, + "logits/rejected": -0.30859375, + "logps/chosen": -384.0, + "logps/rejected": -398.0, + "loss": 0.0231, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.35546875, + "rewards/margins": 11.8125, + "rewards/rejected": -11.5, + "step": 1149 + }, + { + "epoch": 1.4110429447852761, + "grad_norm": 6.406336886570724, + "learning_rate": 8.026859504132232e-07, + "logits/chosen": -0.06689453125, + "logits/rejected": -0.240234375, + "logps/chosen": -390.0, + "logps/rejected": -386.0, + "loss": 0.0284, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.0020904541015625, + "rewards/margins": 11.125, + "rewards/rejected": -11.125, + "step": 1150 + }, + { + "epoch": 1.4122699386503068, + "grad_norm": 5.500976110615787, + "learning_rate": 8.020661157024793e-07, + "logits/chosen": -0.1259765625, + "logits/rejected": -0.330078125, + "logps/chosen": -388.0, + "logps/rejected": -406.0, + "loss": 0.022, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.138671875, + "rewards/margins": 11.625, + "rewards/rejected": -11.75, + "step": 1151 + }, + { + "epoch": 1.4134969325153375, + "grad_norm": 15.46899721034435, + "learning_rate": 8.014462809917355e-07, + "logits/chosen": -0.2041015625, + "logits/rejected": -0.318359375, + "logps/chosen": -420.0, + "logps/rejected": -436.0, + "loss": 0.0474, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.5859375, + "rewards/margins": 12.25, + "rewards/rejected": -11.6875, + "step": 1152 + }, + { + "epoch": 1.4147239263803681, + "grad_norm": 29.448301297633176, + "learning_rate": 8.008264462809917e-07, + "logits/chosen": -0.12109375, + "logits/rejected": -0.26171875, + "logps/chosen": -406.0, + "logps/rejected": -406.0, + "loss": 0.0778, + "rewards/accuracies": 0.9765625, + "rewards/chosen": 0.349609375, + "rewards/margins": 11.375, + "rewards/rejected": -11.0, + "step": 1153 + }, + { + "epoch": 1.4159509202453988, + "grad_norm": 8.55335933719368, + "learning_rate": 8.002066115702479e-07, + "logits/chosen": -0.2041015625, + "logits/rejected": -0.373046875, + "logps/chosen": -446.0, + "logps/rejected": -424.0, + "loss": 0.0282, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.62109375, + "rewards/margins": 12.6875, + "rewards/rejected": -12.0625, + "step": 1154 + }, + { + "epoch": 1.4171779141104295, + "grad_norm": 11.91106417252758, + "learning_rate": 7.995867768595042e-07, + "logits/chosen": -0.1962890625, + "logits/rejected": -0.310546875, + "logps/chosen": -378.0, + "logps/rejected": -386.0, + "loss": 0.0437, + "rewards/accuracies": 0.984375, + "rewards/chosen": -0.146484375, + "rewards/margins": 10.8125, + "rewards/rejected": -10.9375, + "step": 1155 + }, + { + "epoch": 1.4184049079754601, + "grad_norm": 14.345078375136291, + "learning_rate": 7.989669421487604e-07, + "logits/chosen": -0.1103515625, + "logits/rejected": -0.369140625, + "logps/chosen": -440.0, + "logps/rejected": -396.0, + "loss": 0.0455, + "rewards/accuracies": 0.9765625, + "rewards/chosen": 0.7421875, + "rewards/margins": 12.3125, + "rewards/rejected": -11.625, + "step": 1156 + }, + { + "epoch": 1.4196319018404908, + "grad_norm": 13.702503887945511, + "learning_rate": 7.983471074380166e-07, + "logits/chosen": -0.150390625, + "logits/rejected": -0.28125, + "logps/chosen": -450.0, + "logps/rejected": -424.0, + "loss": 0.0482, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.93359375, + "rewards/margins": 11.9375, + "rewards/rejected": -11.0, + "step": 1157 + }, + { + "epoch": 1.4208588957055215, + "grad_norm": 8.654555323915172, + "learning_rate": 7.977272727272728e-07, + "logits/chosen": -0.1650390625, + "logits/rejected": -0.306640625, + "logps/chosen": -346.0, + "logps/rejected": -354.0, + "loss": 0.0362, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.498046875, + "rewards/margins": 11.375, + "rewards/rejected": -10.875, + "step": 1158 + }, + { + "epoch": 1.4220858895705522, + "grad_norm": 14.010487033573805, + "learning_rate": 7.97107438016529e-07, + "logits/chosen": -0.11376953125, + "logits/rejected": -0.1875, + "logps/chosen": -354.0, + "logps/rejected": -388.0, + "loss": 0.0558, + "rewards/accuracies": 0.9765625, + "rewards/chosen": -0.10986328125, + "rewards/margins": 10.6875, + "rewards/rejected": -10.8125, + "step": 1159 + }, + { + "epoch": 1.4233128834355828, + "grad_norm": 16.133153822756856, + "learning_rate": 7.964876033057851e-07, + "logits/chosen": -0.2373046875, + "logits/rejected": -0.3984375, + "logps/chosen": -414.0, + "logps/rejected": -398.0, + "loss": 0.0599, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.23828125, + "rewards/margins": 11.125, + "rewards/rejected": -10.875, + "step": 1160 + }, + { + "epoch": 1.4245398773006135, + "grad_norm": 9.461066515172439, + "learning_rate": 7.958677685950413e-07, + "logits/chosen": -0.0791015625, + "logits/rejected": -0.1962890625, + "logps/chosen": -378.0, + "logps/rejected": -414.0, + "loss": 0.07, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.07421875, + "rewards/margins": 10.6875, + "rewards/rejected": -10.625, + "step": 1161 + }, + { + "epoch": 1.4257668711656442, + "grad_norm": 13.085163631657917, + "learning_rate": 7.952479338842975e-07, + "logits/chosen": -0.1923828125, + "logits/rejected": -0.314453125, + "logps/chosen": -394.0, + "logps/rejected": -382.0, + "loss": 0.0558, + "rewards/accuracies": 0.9765625, + "rewards/chosen": 0.059814453125, + "rewards/margins": 10.8125, + "rewards/rejected": -10.75, + "step": 1162 + }, + { + "epoch": 1.4269938650306748, + "grad_norm": 9.56097265388881, + "learning_rate": 7.946280991735537e-07, + "logits/chosen": -0.279296875, + "logits/rejected": -0.52734375, + "logps/chosen": -402.0, + "logps/rejected": -378.0, + "loss": 0.0236, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.341796875, + "rewards/margins": 11.5625, + "rewards/rejected": -11.25, + "step": 1163 + }, + { + "epoch": 1.4282208588957055, + "grad_norm": 16.722388625925955, + "learning_rate": 7.9400826446281e-07, + "logits/chosen": -0.091796875, + "logits/rejected": -0.126953125, + "logps/chosen": -336.0, + "logps/rejected": -368.0, + "loss": 0.0682, + "rewards/accuracies": 0.96875, + "rewards/chosen": -0.017822265625, + "rewards/margins": 10.4375, + "rewards/rejected": -10.5, + "step": 1164 + }, + { + "epoch": 1.4294478527607362, + "grad_norm": 7.348322611328401, + "learning_rate": 7.933884297520661e-07, + "logits/chosen": -0.0986328125, + "logits/rejected": -0.2373046875, + "logps/chosen": -388.0, + "logps/rejected": -370.0, + "loss": 0.0227, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.08154296875, + "rewards/margins": 10.75, + "rewards/rejected": -10.625, + "step": 1165 + }, + { + "epoch": 1.4306748466257668, + "grad_norm": 5.34988661629868, + "learning_rate": 7.927685950413224e-07, + "logits/chosen": -0.138671875, + "logits/rejected": -0.369140625, + "logps/chosen": -410.0, + "logps/rejected": -386.0, + "loss": 0.0213, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.2353515625, + "rewards/margins": 10.0625, + "rewards/rejected": -10.3125, + "step": 1166 + }, + { + "epoch": 1.4319018404907975, + "grad_norm": 2.351503468437324, + "learning_rate": 7.921487603305786e-07, + "logits/chosen": -0.1630859375, + "logits/rejected": -0.3515625, + "logps/chosen": -404.0, + "logps/rejected": -410.0, + "loss": 0.0096, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.2734375, + "rewards/margins": 11.6875, + "rewards/rejected": -11.4375, + "step": 1167 + }, + { + "epoch": 1.4331288343558282, + "grad_norm": 10.275425558023933, + "learning_rate": 7.915289256198348e-07, + "logits/chosen": -0.150390625, + "logits/rejected": -0.2734375, + "logps/chosen": -372.0, + "logps/rejected": -374.0, + "loss": 0.0434, + "rewards/accuracies": 0.9765625, + "rewards/chosen": -0.357421875, + "rewards/margins": 11.1875, + "rewards/rejected": -11.5625, + "step": 1168 + }, + { + "epoch": 1.4343558282208588, + "grad_norm": 9.69113419509306, + "learning_rate": 7.909090909090909e-07, + "logits/chosen": -0.07568359375, + "logits/rejected": -0.236328125, + "logps/chosen": -376.0, + "logps/rejected": -398.0, + "loss": 0.0378, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.390625, + "rewards/margins": 10.25, + "rewards/rejected": -9.875, + "step": 1169 + }, + { + "epoch": 1.4355828220858895, + "grad_norm": 11.100318005337881, + "learning_rate": 7.90289256198347e-07, + "logits/chosen": -0.171875, + "logits/rejected": -0.421875, + "logps/chosen": -396.0, + "logps/rejected": -364.0, + "loss": 0.0327, + "rewards/accuracies": 0.984375, + "rewards/chosen": -0.095703125, + "rewards/margins": 10.6875, + "rewards/rejected": -10.75, + "step": 1170 + }, + { + "epoch": 1.4368098159509202, + "grad_norm": 13.42384166771764, + "learning_rate": 7.896694214876033e-07, + "logits/chosen": -0.126953125, + "logits/rejected": -0.28515625, + "logps/chosen": -380.0, + "logps/rejected": -372.0, + "loss": 0.0555, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.400390625, + "rewards/margins": 10.6875, + "rewards/rejected": -10.3125, + "step": 1171 + }, + { + "epoch": 1.438036809815951, + "grad_norm": 16.33374781104785, + "learning_rate": 7.890495867768595e-07, + "logits/chosen": -0.055419921875, + "logits/rejected": -0.173828125, + "logps/chosen": -436.0, + "logps/rejected": -400.0, + "loss": 0.0706, + "rewards/accuracies": 0.96875, + "rewards/chosen": -0.2421875, + "rewards/margins": 9.25, + "rewards/rejected": -9.4375, + "step": 1172 + }, + { + "epoch": 1.4392638036809817, + "grad_norm": 17.67963845713936, + "learning_rate": 7.884297520661158e-07, + "logits/chosen": -0.125, + "logits/rejected": -0.275390625, + "logps/chosen": -428.0, + "logps/rejected": -410.0, + "loss": 0.0608, + "rewards/accuracies": 0.9765625, + "rewards/chosen": 0.275390625, + "rewards/margins": 10.875, + "rewards/rejected": -10.625, + "step": 1173 + }, + { + "epoch": 1.4404907975460124, + "grad_norm": 5.1384201359953, + "learning_rate": 7.878099173553719e-07, + "logits/chosen": -0.138671875, + "logits/rejected": -0.166015625, + "logps/chosen": -362.0, + "logps/rejected": -366.0, + "loss": 0.0129, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.017578125, + "rewards/margins": 10.625, + "rewards/rejected": -10.5625, + "step": 1174 + }, + { + "epoch": 1.441717791411043, + "grad_norm": 5.148940884458012, + "learning_rate": 7.871900826446282e-07, + "logits/chosen": -0.1083984375, + "logits/rejected": -0.173828125, + "logps/chosen": -346.0, + "logps/rejected": -364.0, + "loss": 0.0273, + "rewards/accuracies": 0.984375, + "rewards/chosen": -0.35546875, + "rewards/margins": 10.0, + "rewards/rejected": -10.375, + "step": 1175 + }, + { + "epoch": 1.4429447852760737, + "grad_norm": 11.603523639095071, + "learning_rate": 7.865702479338844e-07, + "logits/chosen": -0.1923828125, + "logits/rejected": -0.310546875, + "logps/chosen": -392.0, + "logps/rejected": -402.0, + "loss": 0.0389, + "rewards/accuracies": 0.984375, + "rewards/chosen": -0.306640625, + "rewards/margins": 9.875, + "rewards/rejected": -10.1875, + "step": 1176 + }, + { + "epoch": 1.4441717791411044, + "grad_norm": 10.040926511598048, + "learning_rate": 7.859504132231406e-07, + "logits/chosen": -0.0517578125, + "logits/rejected": -0.244140625, + "logps/chosen": -384.0, + "logps/rejected": -354.0, + "loss": 0.0383, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.04296875, + "rewards/margins": 9.5, + "rewards/rejected": -9.5, + "step": 1177 + }, + { + "epoch": 1.445398773006135, + "grad_norm": 11.028023088957683, + "learning_rate": 7.853305785123967e-07, + "logits/chosen": -0.1484375, + "logits/rejected": -0.35546875, + "logps/chosen": -396.0, + "logps/rejected": -364.0, + "loss": 0.0425, + "rewards/accuracies": 0.984375, + "rewards/chosen": 1.0859375, + "rewards/margins": 10.9375, + "rewards/rejected": -9.875, + "step": 1178 + }, + { + "epoch": 1.4466257668711657, + "grad_norm": 8.150158411239879, + "learning_rate": 7.847107438016528e-07, + "logits/chosen": -0.2412109375, + "logits/rejected": -0.443359375, + "logps/chosen": -438.0, + "logps/rejected": -394.0, + "loss": 0.0374, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.515625, + "rewards/margins": 10.75, + "rewards/rejected": -10.1875, + "step": 1179 + }, + { + "epoch": 1.4478527607361964, + "grad_norm": 6.894255501689617, + "learning_rate": 7.840909090909091e-07, + "logits/chosen": -0.1279296875, + "logits/rejected": -0.1748046875, + "logps/chosen": -344.0, + "logps/rejected": -386.0, + "loss": 0.0309, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.05029296875, + "rewards/margins": 10.75, + "rewards/rejected": -10.6875, + "step": 1180 + }, + { + "epoch": 1.449079754601227, + "grad_norm": 8.019249451553959, + "learning_rate": 7.834710743801653e-07, + "logits/chosen": -0.162109375, + "logits/rejected": -0.2890625, + "logps/chosen": -434.0, + "logps/rejected": -430.0, + "loss": 0.0239, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.87890625, + "rewards/margins": 10.8125, + "rewards/rejected": -9.9375, + "step": 1181 + }, + { + "epoch": 1.4503067484662577, + "grad_norm": 8.340607467826109, + "learning_rate": 7.828512396694216e-07, + "logits/chosen": -0.1083984375, + "logits/rejected": -0.2490234375, + "logps/chosen": -426.0, + "logps/rejected": -404.0, + "loss": 0.0307, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.427734375, + "rewards/margins": 10.1875, + "rewards/rejected": -10.625, + "step": 1182 + }, + { + "epoch": 1.4515337423312884, + "grad_norm": 11.72673950966423, + "learning_rate": 7.822314049586777e-07, + "logits/chosen": -0.142578125, + "logits/rejected": -0.19921875, + "logps/chosen": -326.0, + "logps/rejected": -372.0, + "loss": 0.0563, + "rewards/accuracies": 0.984375, + "rewards/chosen": -0.06787109375, + "rewards/margins": 9.875, + "rewards/rejected": -9.9375, + "step": 1183 + }, + { + "epoch": 1.452760736196319, + "grad_norm": 9.058514886167021, + "learning_rate": 7.81611570247934e-07, + "logits/chosen": -0.15234375, + "logits/rejected": -0.34765625, + "logps/chosen": -366.0, + "logps/rejected": -352.0, + "loss": 0.0407, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.330078125, + "rewards/margins": 10.9375, + "rewards/rejected": -10.625, + "step": 1184 + }, + { + "epoch": 1.4539877300613497, + "grad_norm": 11.422467086533596, + "learning_rate": 7.809917355371902e-07, + "logits/chosen": -0.041259765625, + "logits/rejected": -0.19140625, + "logps/chosen": -420.0, + "logps/rejected": -402.0, + "loss": 0.0448, + "rewards/accuracies": 0.9765625, + "rewards/chosen": 0.5625, + "rewards/margins": 10.8125, + "rewards/rejected": -10.25, + "step": 1185 + }, + { + "epoch": 1.4552147239263804, + "grad_norm": 9.735079419768947, + "learning_rate": 7.803719008264463e-07, + "logits/chosen": -0.11328125, + "logits/rejected": -0.263671875, + "logps/chosen": -378.0, + "logps/rejected": -390.0, + "loss": 0.0379, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.11328125, + "rewards/margins": 10.375, + "rewards/rejected": -10.25, + "step": 1186 + }, + { + "epoch": 1.456441717791411, + "grad_norm": 6.162772380389009, + "learning_rate": 7.797520661157025e-07, + "logits/chosen": -0.1396484375, + "logits/rejected": -0.3359375, + "logps/chosen": -456.0, + "logps/rejected": -410.0, + "loss": 0.0253, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.6875, + "rewards/margins": 10.375, + "rewards/rejected": -9.6875, + "step": 1187 + }, + { + "epoch": 1.4576687116564417, + "grad_norm": 15.719129543958939, + "learning_rate": 7.791322314049586e-07, + "logits/chosen": -0.0654296875, + "logits/rejected": -0.1787109375, + "logps/chosen": -432.0, + "logps/rejected": -434.0, + "loss": 0.0721, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.08447265625, + "rewards/margins": 10.25, + "rewards/rejected": -10.125, + "step": 1188 + }, + { + "epoch": 1.4588957055214724, + "grad_norm": 12.022319651435222, + "learning_rate": 7.785123966942149e-07, + "logits/chosen": -0.1689453125, + "logits/rejected": -0.349609375, + "logps/chosen": -428.0, + "logps/rejected": -400.0, + "loss": 0.082, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.380859375, + "rewards/margins": 10.875, + "rewards/rejected": -10.5, + "step": 1189 + }, + { + "epoch": 1.460122699386503, + "grad_norm": 10.515256662579745, + "learning_rate": 7.778925619834711e-07, + "logits/chosen": -0.173828125, + "logits/rejected": -0.39453125, + "logps/chosen": -436.0, + "logps/rejected": -410.0, + "loss": 0.0408, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.80078125, + "rewards/margins": 12.125, + "rewards/rejected": -11.3125, + "step": 1190 + }, + { + "epoch": 1.4613496932515337, + "grad_norm": 14.208293904428151, + "learning_rate": 7.772727272727273e-07, + "logits/chosen": 0.002960205078125, + "logits/rejected": -0.13671875, + "logps/chosen": -370.0, + "logps/rejected": -390.0, + "loss": 0.0511, + "rewards/accuracies": 0.96875, + "rewards/chosen": 0.140625, + "rewards/margins": 11.375, + "rewards/rejected": -11.25, + "step": 1191 + }, + { + "epoch": 1.4625766871165644, + "grad_norm": 4.726495134099289, + "learning_rate": 7.766528925619835e-07, + "logits/chosen": -0.1162109375, + "logits/rejected": -0.29296875, + "logps/chosen": -400.0, + "logps/rejected": -406.0, + "loss": 0.0213, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.0208740234375, + "rewards/margins": 11.4375, + "rewards/rejected": -11.375, + "step": 1192 + }, + { + "epoch": 1.463803680981595, + "grad_norm": 7.11137377163575, + "learning_rate": 7.760330578512398e-07, + "logits/chosen": -0.181640625, + "logits/rejected": -0.373046875, + "logps/chosen": -458.0, + "logps/rejected": -420.0, + "loss": 0.0249, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.44921875, + "rewards/margins": 11.4375, + "rewards/rejected": -11.0, + "step": 1193 + }, + { + "epoch": 1.4650306748466257, + "grad_norm": 7.389252182902934, + "learning_rate": 7.75413223140496e-07, + "logits/chosen": -0.15625, + "logits/rejected": -0.275390625, + "logps/chosen": -420.0, + "logps/rejected": -404.0, + "loss": 0.0133, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.326171875, + "rewards/margins": 10.9375, + "rewards/rejected": -10.625, + "step": 1194 + }, + { + "epoch": 1.4662576687116564, + "grad_norm": 10.994717707834955, + "learning_rate": 7.74793388429752e-07, + "logits/chosen": -0.12451171875, + "logits/rejected": -0.2451171875, + "logps/chosen": -402.0, + "logps/rejected": -410.0, + "loss": 0.0418, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.08251953125, + "rewards/margins": 10.6875, + "rewards/rejected": -10.8125, + "step": 1195 + }, + { + "epoch": 1.467484662576687, + "grad_norm": 8.858341862725984, + "learning_rate": 7.741735537190082e-07, + "logits/chosen": -0.04638671875, + "logits/rejected": -0.1708984375, + "logps/chosen": -390.0, + "logps/rejected": -392.0, + "loss": 0.0296, + "rewards/accuracies": 0.984375, + "rewards/chosen": -0.1884765625, + "rewards/margins": 10.125, + "rewards/rejected": -10.3125, + "step": 1196 + }, + { + "epoch": 1.4687116564417177, + "grad_norm": 15.001467453574548, + "learning_rate": 7.735537190082644e-07, + "logits/chosen": -0.0546875, + "logits/rejected": -0.2333984375, + "logps/chosen": -422.0, + "logps/rejected": -382.0, + "loss": 0.0572, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.05810546875, + "rewards/margins": 10.875, + "rewards/rejected": -10.8125, + "step": 1197 + }, + { + "epoch": 1.4699386503067484, + "grad_norm": 11.467996330718492, + "learning_rate": 7.729338842975207e-07, + "logits/chosen": 0.01531982421875, + "logits/rejected": -0.126953125, + "logps/chosen": -384.0, + "logps/rejected": -356.0, + "loss": 0.0766, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.07177734375, + "rewards/margins": 10.8125, + "rewards/rejected": -10.6875, + "step": 1198 + }, + { + "epoch": 1.471165644171779, + "grad_norm": 8.297507128344717, + "learning_rate": 7.723140495867769e-07, + "logits/chosen": -0.2041015625, + "logits/rejected": -0.3671875, + "logps/chosen": -394.0, + "logps/rejected": -376.0, + "loss": 0.0247, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.376953125, + "rewards/margins": 11.625, + "rewards/rejected": -11.1875, + "step": 1199 + }, + { + "epoch": 1.4723926380368098, + "grad_norm": 6.708462240020914, + "learning_rate": 7.716942148760331e-07, + "logits/chosen": -0.13671875, + "logits/rejected": -0.2890625, + "logps/chosen": -330.0, + "logps/rejected": -350.0, + "loss": 0.0204, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.1572265625, + "rewards/margins": 10.125, + "rewards/rejected": -10.25, + "step": 1200 + }, + { + "epoch": 1.4736196319018404, + "grad_norm": 10.300151959699189, + "learning_rate": 7.710743801652893e-07, + "logits/chosen": -0.041015625, + "logits/rejected": -0.18359375, + "logps/chosen": -378.0, + "logps/rejected": -410.0, + "loss": 0.0294, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.1494140625, + "rewards/margins": 10.8125, + "rewards/rejected": -10.6875, + "step": 1201 + }, + { + "epoch": 1.474846625766871, + "grad_norm": 10.130166213309426, + "learning_rate": 7.704545454545456e-07, + "logits/chosen": -0.1337890625, + "logits/rejected": -0.2734375, + "logps/chosen": -428.0, + "logps/rejected": -422.0, + "loss": 0.0315, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.81640625, + "rewards/margins": 11.5625, + "rewards/rejected": -10.75, + "step": 1202 + }, + { + "epoch": 1.4760736196319018, + "grad_norm": 9.219567443331156, + "learning_rate": 7.698347107438017e-07, + "logits/chosen": -0.2314453125, + "logits/rejected": -0.408203125, + "logps/chosen": -410.0, + "logps/rejected": -372.0, + "loss": 0.0433, + "rewards/accuracies": 0.9765625, + "rewards/chosen": 0.2060546875, + "rewards/margins": 11.125, + "rewards/rejected": -10.9375, + "step": 1203 + }, + { + "epoch": 1.4773006134969324, + "grad_norm": 9.51752284318554, + "learning_rate": 7.692148760330578e-07, + "logits/chosen": -0.08203125, + "logits/rejected": -0.30859375, + "logps/chosen": -374.0, + "logps/rejected": -392.0, + "loss": 0.041, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.546875, + "rewards/margins": 11.125, + "rewards/rejected": -10.625, + "step": 1204 + }, + { + "epoch": 1.478527607361963, + "grad_norm": 23.546428660464922, + "learning_rate": 7.68595041322314e-07, + "logits/chosen": 0.02734375, + "logits/rejected": -0.2021484375, + "logps/chosen": -364.0, + "logps/rejected": -378.0, + "loss": 0.1268, + "rewards/accuracies": 0.9453125, + "rewards/chosen": 0.2470703125, + "rewards/margins": 10.4375, + "rewards/rejected": -10.1875, + "step": 1205 + }, + { + "epoch": 1.4797546012269938, + "grad_norm": 6.082107723283603, + "learning_rate": 7.679752066115702e-07, + "logits/chosen": -0.02685546875, + "logits/rejected": -0.3046875, + "logps/chosen": -350.0, + "logps/rejected": -344.0, + "loss": 0.0275, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.0791015625, + "rewards/margins": 10.25, + "rewards/rejected": -10.1875, + "step": 1206 + }, + { + "epoch": 1.4809815950920244, + "grad_norm": 10.622559047787306, + "learning_rate": 7.673553719008265e-07, + "logits/chosen": -0.09375, + "logits/rejected": -0.27734375, + "logps/chosen": -406.0, + "logps/rejected": -348.0, + "loss": 0.0353, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.05859375, + "rewards/margins": 10.9375, + "rewards/rejected": -10.875, + "step": 1207 + }, + { + "epoch": 1.482208588957055, + "grad_norm": 13.683817866460686, + "learning_rate": 7.667355371900826e-07, + "logits/chosen": -0.1123046875, + "logits/rejected": -0.26953125, + "logps/chosen": -376.0, + "logps/rejected": -382.0, + "loss": 0.0619, + "rewards/accuracies": 0.953125, + "rewards/chosen": -0.11474609375, + "rewards/margins": 10.25, + "rewards/rejected": -10.3125, + "step": 1208 + }, + { + "epoch": 1.4834355828220858, + "grad_norm": 8.88934565774754, + "learning_rate": 7.661157024793389e-07, + "logits/chosen": -0.08154296875, + "logits/rejected": -0.26171875, + "logps/chosen": -362.0, + "logps/rejected": -364.0, + "loss": 0.042, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.0166015625, + "rewards/margins": 11.3125, + "rewards/rejected": -11.375, + "step": 1209 + }, + { + "epoch": 1.4846625766871164, + "grad_norm": 13.517803461526613, + "learning_rate": 7.654958677685951e-07, + "logits/chosen": -0.036865234375, + "logits/rejected": -0.19140625, + "logps/chosen": -366.0, + "logps/rejected": -364.0, + "loss": 0.048, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.57421875, + "rewards/margins": 11.4375, + "rewards/rejected": -10.875, + "step": 1210 + }, + { + "epoch": 1.4858895705521473, + "grad_norm": 3.0410283674490257, + "learning_rate": 7.648760330578514e-07, + "logits/chosen": -0.06640625, + "logits/rejected": -0.283203125, + "logps/chosen": -358.0, + "logps/rejected": -382.0, + "loss": 0.0121, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.2021484375, + "rewards/margins": 11.0, + "rewards/rejected": -10.8125, + "step": 1211 + }, + { + "epoch": 1.487116564417178, + "grad_norm": 5.560187356994683, + "learning_rate": 7.642561983471075e-07, + "logits/chosen": -0.1767578125, + "logits/rejected": -0.28125, + "logps/chosen": -356.0, + "logps/rejected": -386.0, + "loss": 0.0259, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.12158203125, + "rewards/margins": 11.25, + "rewards/rejected": -11.375, + "step": 1212 + }, + { + "epoch": 1.4883435582822087, + "grad_norm": 7.567720346296838, + "learning_rate": 7.636363636363635e-07, + "logits/chosen": -0.1669921875, + "logits/rejected": -0.42578125, + "logps/chosen": -432.0, + "logps/rejected": -396.0, + "loss": 0.0322, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.7890625, + "rewards/margins": 11.5625, + "rewards/rejected": -10.75, + "step": 1213 + }, + { + "epoch": 1.4895705521472393, + "grad_norm": 10.044279259721861, + "learning_rate": 7.630165289256198e-07, + "logits/chosen": -0.1259765625, + "logits/rejected": -0.2421875, + "logps/chosen": -348.0, + "logps/rejected": -380.0, + "loss": 0.0415, + "rewards/accuracies": 0.984375, + "rewards/chosen": -0.369140625, + "rewards/margins": 11.0625, + "rewards/rejected": -11.4375, + "step": 1214 + }, + { + "epoch": 1.49079754601227, + "grad_norm": 11.888900857643913, + "learning_rate": 7.62396694214876e-07, + "logits/chosen": -0.181640625, + "logits/rejected": -0.390625, + "logps/chosen": -382.0, + "logps/rejected": -392.0, + "loss": 0.0302, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.31640625, + "rewards/margins": 12.25, + "rewards/rejected": -11.9375, + "step": 1215 + }, + { + "epoch": 1.4920245398773007, + "grad_norm": 11.234520606273014, + "learning_rate": 7.617768595041323e-07, + "logits/chosen": -0.140625, + "logits/rejected": -0.1845703125, + "logps/chosen": -378.0, + "logps/rejected": -370.0, + "loss": 0.0389, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.396484375, + "rewards/margins": 10.625, + "rewards/rejected": -11.0, + "step": 1216 + }, + { + "epoch": 1.4932515337423313, + "grad_norm": 5.600087268075799, + "learning_rate": 7.611570247933884e-07, + "logits/chosen": -0.1484375, + "logits/rejected": -0.26953125, + "logps/chosen": -404.0, + "logps/rejected": -406.0, + "loss": 0.0147, + "rewards/accuracies": 1.0, + "rewards/chosen": 1.078125, + "rewards/margins": 12.0625, + "rewards/rejected": -11.0, + "step": 1217 + }, + { + "epoch": 1.494478527607362, + "grad_norm": 13.512440969064604, + "learning_rate": 7.605371900826447e-07, + "logits/chosen": -0.0634765625, + "logits/rejected": -0.2333984375, + "logps/chosen": -374.0, + "logps/rejected": -386.0, + "loss": 0.0486, + "rewards/accuracies": 0.9765625, + "rewards/chosen": -0.76171875, + "rewards/margins": 10.0625, + "rewards/rejected": -10.8125, + "step": 1218 + }, + { + "epoch": 1.4957055214723927, + "grad_norm": 11.083011010632404, + "learning_rate": 7.599173553719009e-07, + "logits/chosen": -0.1513671875, + "logits/rejected": -0.302734375, + "logps/chosen": -394.0, + "logps/rejected": -408.0, + "loss": 0.0355, + "rewards/accuracies": 0.984375, + "rewards/chosen": -0.1767578125, + "rewards/margins": 10.8125, + "rewards/rejected": -10.9375, + "step": 1219 + }, + { + "epoch": 1.4969325153374233, + "grad_norm": 12.719289884310554, + "learning_rate": 7.59297520661157e-07, + "logits/chosen": -0.181640625, + "logits/rejected": -0.271484375, + "logps/chosen": -404.0, + "logps/rejected": -420.0, + "loss": 0.0511, + "rewards/accuracies": 0.96875, + "rewards/chosen": -0.53515625, + "rewards/margins": 11.0, + "rewards/rejected": -11.5, + "step": 1220 + }, + { + "epoch": 1.498159509202454, + "grad_norm": 11.10401374516513, + "learning_rate": 7.586776859504133e-07, + "logits/chosen": -0.2109375, + "logits/rejected": -0.3671875, + "logps/chosen": -412.0, + "logps/rejected": -382.0, + "loss": 0.038, + "rewards/accuracies": 0.96875, + "rewards/chosen": 0.087890625, + "rewards/margins": 11.5625, + "rewards/rejected": -11.4375, + "step": 1221 + }, + { + "epoch": 1.4993865030674847, + "grad_norm": 10.550256846021844, + "learning_rate": 7.580578512396693e-07, + "logits/chosen": -0.08935546875, + "logits/rejected": -0.263671875, + "logps/chosen": -342.0, + "logps/rejected": -370.0, + "loss": 0.0459, + "rewards/accuracies": 0.984375, + "rewards/chosen": -1.015625, + "rewards/margins": 10.0625, + "rewards/rejected": -11.0625, + "step": 1222 + }, + { + "epoch": 1.5006134969325153, + "grad_norm": 4.907032447202447, + "learning_rate": 7.574380165289256e-07, + "logits/chosen": -0.11328125, + "logits/rejected": -0.2421875, + "logps/chosen": -428.0, + "logps/rejected": -406.0, + "loss": 0.0214, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.4609375, + "rewards/margins": 11.8125, + "rewards/rejected": -11.3125, + "step": 1223 + }, + { + "epoch": 1.501840490797546, + "grad_norm": 7.212415747597125, + "learning_rate": 7.568181818181818e-07, + "logits/chosen": -0.1044921875, + "logits/rejected": -0.2041015625, + "logps/chosen": -360.0, + "logps/rejected": -374.0, + "loss": 0.0386, + "rewards/accuracies": 0.9765625, + "rewards/chosen": -0.96875, + "rewards/margins": 9.8125, + "rewards/rejected": -10.8125, + "step": 1224 + }, + { + "epoch": 1.5030674846625767, + "grad_norm": 5.207118696961701, + "learning_rate": 7.56198347107438e-07, + "logits/chosen": -0.09033203125, + "logits/rejected": -0.220703125, + "logps/chosen": -362.0, + "logps/rejected": -378.0, + "loss": 0.0192, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.35546875, + "rewards/margins": 10.5625, + "rewards/rejected": -10.9375, + "step": 1225 + }, + { + "epoch": 1.5042944785276073, + "grad_norm": 12.469106555425835, + "learning_rate": 7.555785123966942e-07, + "logits/chosen": -0.1669921875, + "logits/rejected": -0.34765625, + "logps/chosen": -394.0, + "logps/rejected": -384.0, + "loss": 0.0364, + "rewards/accuracies": 0.9765625, + "rewards/chosen": -0.58203125, + "rewards/margins": 10.75, + "rewards/rejected": -11.3125, + "step": 1226 + }, + { + "epoch": 1.505521472392638, + "grad_norm": 6.703095964441331, + "learning_rate": 7.549586776859505e-07, + "logits/chosen": -0.31640625, + "logits/rejected": -0.396484375, + "logps/chosen": -398.0, + "logps/rejected": -422.0, + "loss": 0.0161, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.228515625, + "rewards/margins": 11.625, + "rewards/rejected": -11.375, + "step": 1227 + }, + { + "epoch": 1.5067484662576687, + "grad_norm": 20.182875127429853, + "learning_rate": 7.543388429752067e-07, + "logits/chosen": -0.14453125, + "logits/rejected": -0.251953125, + "logps/chosen": -416.0, + "logps/rejected": -426.0, + "loss": 0.1114, + "rewards/accuracies": 0.9609375, + "rewards/chosen": -0.458984375, + "rewards/margins": 10.4375, + "rewards/rejected": -10.875, + "step": 1228 + }, + { + "epoch": 1.5079754601226993, + "grad_norm": 7.168144452671762, + "learning_rate": 7.537190082644629e-07, + "logits/chosen": -0.2314453125, + "logits/rejected": -0.34765625, + "logps/chosen": -382.0, + "logps/rejected": -394.0, + "loss": 0.0249, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.330078125, + "rewards/margins": 11.0, + "rewards/rejected": -11.3125, + "step": 1229 + }, + { + "epoch": 1.50920245398773, + "grad_norm": 8.739084266889561, + "learning_rate": 7.530991735537191e-07, + "logits/chosen": -0.1220703125, + "logits/rejected": -0.29296875, + "logps/chosen": -370.0, + "logps/rejected": -370.0, + "loss": 0.041, + "rewards/accuracies": 0.984375, + "rewards/chosen": -0.0230712890625, + "rewards/margins": 10.5, + "rewards/rejected": -10.5625, + "step": 1230 + }, + { + "epoch": 1.510429447852761, + "grad_norm": 14.93632729231571, + "learning_rate": 7.524793388429751e-07, + "logits/chosen": -0.0654296875, + "logits/rejected": -0.23046875, + "logps/chosen": -356.0, + "logps/rejected": -356.0, + "loss": 0.0501, + "rewards/accuracies": 0.9609375, + "rewards/chosen": -0.1796875, + "rewards/margins": 10.375, + "rewards/rejected": -10.5625, + "step": 1231 + }, + { + "epoch": 1.5116564417177916, + "grad_norm": 6.602158622702926, + "learning_rate": 7.518595041322314e-07, + "logits/chosen": -0.296875, + "logits/rejected": -0.486328125, + "logps/chosen": -400.0, + "logps/rejected": -396.0, + "loss": 0.0195, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.482421875, + "rewards/margins": 12.1875, + "rewards/rejected": -11.75, + "step": 1232 + }, + { + "epoch": 1.5128834355828222, + "grad_norm": 6.2838103080905015, + "learning_rate": 7.512396694214876e-07, + "logits/chosen": -0.193359375, + "logits/rejected": -0.37890625, + "logps/chosen": -384.0, + "logps/rejected": -374.0, + "loss": 0.0189, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.302734375, + "rewards/margins": 10.875, + "rewards/rejected": -11.1875, + "step": 1233 + }, + { + "epoch": 1.514110429447853, + "grad_norm": 18.0136951443616, + "learning_rate": 7.506198347107438e-07, + "logits/chosen": -0.1318359375, + "logits/rejected": -0.328125, + "logps/chosen": -426.0, + "logps/rejected": -394.0, + "loss": 0.0583, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.65234375, + "rewards/margins": 11.375, + "rewards/rejected": -10.6875, + "step": 1234 + }, + { + "epoch": 1.5153374233128836, + "grad_norm": 7.846641320419013, + "learning_rate": 7.5e-07, + "logits/chosen": -0.0546875, + "logits/rejected": -0.19140625, + "logps/chosen": -392.0, + "logps/rejected": -382.0, + "loss": 0.0282, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.9609375, + "rewards/margins": 10.375, + "rewards/rejected": -11.375, + "step": 1235 + }, + { + "epoch": 1.5165644171779142, + "grad_norm": 14.299732747614243, + "learning_rate": 7.493801652892563e-07, + "logits/chosen": -0.138671875, + "logits/rejected": -0.310546875, + "logps/chosen": -392.0, + "logps/rejected": -388.0, + "loss": 0.0557, + "rewards/accuracies": 0.96875, + "rewards/chosen": 0.0291748046875, + "rewards/margins": 10.8125, + "rewards/rejected": -10.8125, + "step": 1236 + }, + { + "epoch": 1.517791411042945, + "grad_norm": 15.326417785804917, + "learning_rate": 7.487603305785124e-07, + "logits/chosen": -0.24609375, + "logits/rejected": -0.40625, + "logps/chosen": -434.0, + "logps/rejected": -402.0, + "loss": 0.0642, + "rewards/accuracies": 0.9609375, + "rewards/chosen": -0.23046875, + "rewards/margins": 12.125, + "rewards/rejected": -12.3125, + "step": 1237 + }, + { + "epoch": 1.5190184049079756, + "grad_norm": 10.442301727950197, + "learning_rate": 7.481404958677685e-07, + "logits/chosen": -0.07373046875, + "logits/rejected": -0.2216796875, + "logps/chosen": -464.0, + "logps/rejected": -396.0, + "loss": 0.0259, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.7734375, + "rewards/margins": 11.625, + "rewards/rejected": -10.875, + "step": 1238 + }, + { + "epoch": 1.5202453987730062, + "grad_norm": 10.812028014205586, + "learning_rate": 7.475206611570248e-07, + "logits/chosen": -0.193359375, + "logits/rejected": -0.40625, + "logps/chosen": -404.0, + "logps/rejected": -382.0, + "loss": 0.0373, + "rewards/accuracies": 0.984375, + "rewards/chosen": -0.10302734375, + "rewards/margins": 11.0625, + "rewards/rejected": -11.1875, + "step": 1239 + }, + { + "epoch": 1.521472392638037, + "grad_norm": 7.399322537588399, + "learning_rate": 7.46900826446281e-07, + "logits/chosen": -0.0830078125, + "logits/rejected": -0.28125, + "logps/chosen": -410.0, + "logps/rejected": -402.0, + "loss": 0.0326, + "rewards/accuracies": 0.9765625, + "rewards/chosen": -0.353515625, + "rewards/margins": 10.9375, + "rewards/rejected": -11.25, + "step": 1240 + }, + { + "epoch": 1.5226993865030676, + "grad_norm": 11.737470905903368, + "learning_rate": 7.462809917355373e-07, + "logits/chosen": -0.033447265625, + "logits/rejected": -0.1806640625, + "logps/chosen": -360.0, + "logps/rejected": -368.0, + "loss": 0.0451, + "rewards/accuracies": 0.984375, + "rewards/chosen": -0.46484375, + "rewards/margins": 10.0625, + "rewards/rejected": -10.5625, + "step": 1241 + }, + { + "epoch": 1.5239263803680982, + "grad_norm": 11.68475946384956, + "learning_rate": 7.456611570247933e-07, + "logits/chosen": -0.263671875, + "logits/rejected": -0.42578125, + "logps/chosen": -388.0, + "logps/rejected": -386.0, + "loss": 0.0484, + "rewards/accuracies": 0.9765625, + "rewards/chosen": -0.5703125, + "rewards/margins": 11.25, + "rewards/rejected": -11.8125, + "step": 1242 + }, + { + "epoch": 1.525153374233129, + "grad_norm": 7.544728556493131, + "learning_rate": 7.450413223140496e-07, + "logits/chosen": -0.059814453125, + "logits/rejected": -0.2158203125, + "logps/chosen": -334.0, + "logps/rejected": -354.0, + "loss": 0.028, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.73828125, + "rewards/margins": 10.125, + "rewards/rejected": -10.875, + "step": 1243 + }, + { + "epoch": 1.5263803680981596, + "grad_norm": 8.662327485339514, + "learning_rate": 7.444214876033058e-07, + "logits/chosen": -0.06787109375, + "logits/rejected": -0.16796875, + "logps/chosen": -342.0, + "logps/rejected": -376.0, + "loss": 0.0232, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.59375, + "rewards/margins": 10.0, + "rewards/rejected": -10.5625, + "step": 1244 + }, + { + "epoch": 1.5276073619631902, + "grad_norm": 14.324574848335546, + "learning_rate": 7.438016528925621e-07, + "logits/chosen": -0.10498046875, + "logits/rejected": -0.302734375, + "logps/chosen": -422.0, + "logps/rejected": -418.0, + "loss": 0.0653, + "rewards/accuracies": 0.96875, + "rewards/chosen": -0.08154296875, + "rewards/margins": 11.25, + "rewards/rejected": -11.3125, + "step": 1245 + }, + { + "epoch": 1.528834355828221, + "grad_norm": 8.939887633566299, + "learning_rate": 7.431818181818182e-07, + "logits/chosen": -0.23046875, + "logits/rejected": -0.396484375, + "logps/chosen": -352.0, + "logps/rejected": -370.0, + "loss": 0.0281, + "rewards/accuracies": 0.984375, + "rewards/chosen": -0.05908203125, + "rewards/margins": 11.125, + "rewards/rejected": -11.1875, + "step": 1246 + }, + { + "epoch": 1.5300613496932516, + "grad_norm": 2.737977434539333, + "learning_rate": 7.425619834710743e-07, + "logits/chosen": -0.189453125, + "logits/rejected": -0.302734375, + "logps/chosen": -434.0, + "logps/rejected": -404.0, + "loss": 0.0103, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.318359375, + "rewards/margins": 12.0625, + "rewards/rejected": -11.75, + "step": 1247 + }, + { + "epoch": 1.5312883435582823, + "grad_norm": 7.955215828013632, + "learning_rate": 7.419421487603306e-07, + "logits/chosen": -0.1494140625, + "logits/rejected": -0.3359375, + "logps/chosen": -376.0, + "logps/rejected": -386.0, + "loss": 0.0229, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.3828125, + "rewards/margins": 11.375, + "rewards/rejected": -11.0, + "step": 1248 + }, + { + "epoch": 1.532515337423313, + "grad_norm": 13.148164884666349, + "learning_rate": 7.413223140495868e-07, + "logits/chosen": 0.036376953125, + "logits/rejected": -0.0673828125, + "logps/chosen": -350.0, + "logps/rejected": -378.0, + "loss": 0.0618, + "rewards/accuracies": 0.984375, + "rewards/chosen": -0.54296875, + "rewards/margins": 10.875, + "rewards/rejected": -11.4375, + "step": 1249 + }, + { + "epoch": 1.5337423312883436, + "grad_norm": 6.7014580907570505, + "learning_rate": 7.40702479338843e-07, + "logits/chosen": -0.2333984375, + "logits/rejected": -0.384765625, + "logps/chosen": -446.0, + "logps/rejected": -472.0, + "loss": 0.028, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.39453125, + "rewards/margins": 13.125, + "rewards/rejected": -12.75, + "step": 1250 + }, + { + "epoch": 1.5349693251533743, + "grad_norm": 9.188465721829859, + "learning_rate": 7.400826446280991e-07, + "logits/chosen": -0.291015625, + "logits/rejected": -0.39453125, + "logps/chosen": -376.0, + "logps/rejected": -414.0, + "loss": 0.0271, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.68359375, + "rewards/margins": 12.625, + "rewards/rejected": -11.9375, + "step": 1251 + }, + { + "epoch": 1.536196319018405, + "grad_norm": 8.458678355370827, + "learning_rate": 7.394628099173554e-07, + "logits/chosen": -0.181640625, + "logits/rejected": -0.34765625, + "logps/chosen": -390.0, + "logps/rejected": -372.0, + "loss": 0.0304, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.158203125, + "rewards/margins": 12.375, + "rewards/rejected": -12.1875, + "step": 1252 + }, + { + "epoch": 1.5374233128834356, + "grad_norm": 8.012255387049295, + "learning_rate": 7.388429752066116e-07, + "logits/chosen": -0.16796875, + "logits/rejected": -0.326171875, + "logps/chosen": -420.0, + "logps/rejected": -410.0, + "loss": 0.0317, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.126953125, + "rewards/margins": 11.3125, + "rewards/rejected": -11.1875, + "step": 1253 + }, + { + "epoch": 1.5386503067484663, + "grad_norm": 9.41643727038603, + "learning_rate": 7.382231404958678e-07, + "logits/chosen": -0.0654296875, + "logits/rejected": -0.19921875, + "logps/chosen": -378.0, + "logps/rejected": -360.0, + "loss": 0.0366, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.267578125, + "rewards/margins": 11.0, + "rewards/rejected": -10.75, + "step": 1254 + }, + { + "epoch": 1.539877300613497, + "grad_norm": 7.434007112296638, + "learning_rate": 7.37603305785124e-07, + "logits/chosen": -0.203125, + "logits/rejected": -0.427734375, + "logps/chosen": -410.0, + "logps/rejected": -432.0, + "loss": 0.0221, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.62890625, + "rewards/margins": 12.6875, + "rewards/rejected": -12.0625, + "step": 1255 + }, + { + "epoch": 1.5411042944785276, + "grad_norm": 8.974145176497167, + "learning_rate": 7.369834710743802e-07, + "logits/chosen": -0.1142578125, + "logits/rejected": -0.328125, + "logps/chosen": -410.0, + "logps/rejected": -398.0, + "loss": 0.0302, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.06494140625, + "rewards/margins": 10.6875, + "rewards/rejected": -10.625, + "step": 1256 + }, + { + "epoch": 1.5423312883435583, + "grad_norm": 13.782445851616703, + "learning_rate": 7.363636363636364e-07, + "logits/chosen": -0.0732421875, + "logits/rejected": -0.26171875, + "logps/chosen": -390.0, + "logps/rejected": -386.0, + "loss": 0.0591, + "rewards/accuracies": 0.9765625, + "rewards/chosen": 0.0693359375, + "rewards/margins": 11.0, + "rewards/rejected": -10.9375, + "step": 1257 + }, + { + "epoch": 1.543558282208589, + "grad_norm": 8.478001588142458, + "learning_rate": 7.357438016528926e-07, + "logits/chosen": -0.10791015625, + "logits/rejected": -0.3125, + "logps/chosen": -404.0, + "logps/rejected": -368.0, + "loss": 0.0315, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.82421875, + "rewards/margins": 11.375, + "rewards/rejected": -10.5625, + "step": 1258 + }, + { + "epoch": 1.5447852760736196, + "grad_norm": 12.343587239938957, + "learning_rate": 7.351239669421488e-07, + "logits/chosen": -0.11181640625, + "logits/rejected": -0.34375, + "logps/chosen": -404.0, + "logps/rejected": -380.0, + "loss": 0.0382, + "rewards/accuracies": 0.96875, + "rewards/chosen": 0.6484375, + "rewards/margins": 11.25, + "rewards/rejected": -10.625, + "step": 1259 + }, + { + "epoch": 1.5460122699386503, + "grad_norm": 10.786647686676162, + "learning_rate": 7.345041322314049e-07, + "logits/chosen": -0.169921875, + "logits/rejected": -0.34765625, + "logps/chosen": -450.0, + "logps/rejected": -396.0, + "loss": 0.0273, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.9453125, + "rewards/margins": 11.6875, + "rewards/rejected": -10.75, + "step": 1260 + }, + { + "epoch": 1.547239263803681, + "grad_norm": 11.69554028398202, + "learning_rate": 7.338842975206612e-07, + "logits/chosen": -0.2236328125, + "logits/rejected": -0.345703125, + "logps/chosen": -406.0, + "logps/rejected": -388.0, + "loss": 0.0333, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.490234375, + "rewards/margins": 10.375, + "rewards/rejected": -9.875, + "step": 1261 + }, + { + "epoch": 1.5484662576687116, + "grad_norm": 7.864292149216722, + "learning_rate": 7.332644628099174e-07, + "logits/chosen": -0.146484375, + "logits/rejected": -0.333984375, + "logps/chosen": -386.0, + "logps/rejected": -362.0, + "loss": 0.038, + "rewards/accuracies": 0.984375, + "rewards/chosen": -0.13671875, + "rewards/margins": 10.9375, + "rewards/rejected": -11.0625, + "step": 1262 + }, + { + "epoch": 1.5496932515337423, + "grad_norm": 13.897917738806647, + "learning_rate": 7.326446280991736e-07, + "logits/chosen": -0.0830078125, + "logits/rejected": -0.298828125, + "logps/chosen": -394.0, + "logps/rejected": -380.0, + "loss": 0.0455, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.69921875, + "rewards/margins": 11.625, + "rewards/rejected": -10.9375, + "step": 1263 + }, + { + "epoch": 1.550920245398773, + "grad_norm": 11.512996209757597, + "learning_rate": 7.320247933884297e-07, + "logits/chosen": -0.0218505859375, + "logits/rejected": -0.1875, + "logps/chosen": -390.0, + "logps/rejected": -398.0, + "loss": 0.0581, + "rewards/accuracies": 0.984375, + "rewards/chosen": -0.1591796875, + "rewards/margins": 10.1875, + "rewards/rejected": -10.3125, + "step": 1264 + }, + { + "epoch": 1.5521472392638036, + "grad_norm": 13.705005476191044, + "learning_rate": 7.31404958677686e-07, + "logits/chosen": -0.1689453125, + "logits/rejected": -0.341796875, + "logps/chosen": -380.0, + "logps/rejected": -390.0, + "loss": 0.0413, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.3359375, + "rewards/margins": 11.125, + "rewards/rejected": -10.8125, + "step": 1265 + }, + { + "epoch": 1.5533742331288343, + "grad_norm": 10.295352755658561, + "learning_rate": 7.307851239669422e-07, + "logits/chosen": -0.2021484375, + "logits/rejected": -0.296875, + "logps/chosen": -352.0, + "logps/rejected": -372.0, + "loss": 0.0297, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.09130859375, + "rewards/margins": 10.5625, + "rewards/rejected": -10.5, + "step": 1266 + }, + { + "epoch": 1.554601226993865, + "grad_norm": 18.45895537867682, + "learning_rate": 7.301652892561983e-07, + "logits/chosen": -0.130859375, + "logits/rejected": -0.271484375, + "logps/chosen": -368.0, + "logps/rejected": -366.0, + "loss": 0.1103, + "rewards/accuracies": 0.96875, + "rewards/chosen": 0.29296875, + "rewards/margins": 10.8125, + "rewards/rejected": -10.5, + "step": 1267 + }, + { + "epoch": 1.5558282208588956, + "grad_norm": 14.03503002507992, + "learning_rate": 7.295454545454546e-07, + "logits/chosen": -0.14453125, + "logits/rejected": -0.255859375, + "logps/chosen": -498.0, + "logps/rejected": -422.0, + "loss": 0.0495, + "rewards/accuracies": 0.9765625, + "rewards/chosen": 0.9609375, + "rewards/margins": 12.125, + "rewards/rejected": -11.1875, + "step": 1268 + }, + { + "epoch": 1.5570552147239263, + "grad_norm": 3.7230278336867006, + "learning_rate": 7.289256198347107e-07, + "logits/chosen": -0.1806640625, + "logits/rejected": -0.314453125, + "logps/chosen": -378.0, + "logps/rejected": -362.0, + "loss": 0.0144, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.8515625, + "rewards/margins": 11.5625, + "rewards/rejected": -10.75, + "step": 1269 + }, + { + "epoch": 1.558282208588957, + "grad_norm": 8.890488407699403, + "learning_rate": 7.28305785123967e-07, + "logits/chosen": -0.1689453125, + "logits/rejected": -0.22265625, + "logps/chosen": -406.0, + "logps/rejected": -406.0, + "loss": 0.026, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.396484375, + "rewards/margins": 10.75, + "rewards/rejected": -10.375, + "step": 1270 + }, + { + "epoch": 1.5595092024539876, + "grad_norm": 8.516198696845315, + "learning_rate": 7.276859504132231e-07, + "logits/chosen": -0.0830078125, + "logits/rejected": -0.1650390625, + "logps/chosen": -356.0, + "logps/rejected": -366.0, + "loss": 0.0302, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.56640625, + "rewards/margins": 10.8125, + "rewards/rejected": -10.25, + "step": 1271 + }, + { + "epoch": 1.5607361963190183, + "grad_norm": 3.950798580884482, + "learning_rate": 7.270661157024794e-07, + "logits/chosen": -0.2578125, + "logits/rejected": -0.412109375, + "logps/chosen": -426.0, + "logps/rejected": -392.0, + "loss": 0.0131, + "rewards/accuracies": 1.0, + "rewards/chosen": 1.46875, + "rewards/margins": 12.25, + "rewards/rejected": -10.75, + "step": 1272 + }, + { + "epoch": 1.561963190184049, + "grad_norm": 12.063277382793855, + "learning_rate": 7.264462809917355e-07, + "logits/chosen": -0.2080078125, + "logits/rejected": -0.3046875, + "logps/chosen": -426.0, + "logps/rejected": -412.0, + "loss": 0.0509, + "rewards/accuracies": 0.9765625, + "rewards/chosen": 0.365234375, + "rewards/margins": 10.6875, + "rewards/rejected": -10.3125, + "step": 1273 + }, + { + "epoch": 1.5631901840490796, + "grad_norm": 5.052885121682078, + "learning_rate": 7.258264462809918e-07, + "logits/chosen": -0.31640625, + "logits/rejected": -0.5078125, + "logps/chosen": -400.0, + "logps/rejected": -388.0, + "loss": 0.0173, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.263671875, + "rewards/margins": 11.0, + "rewards/rejected": -10.75, + "step": 1274 + }, + { + "epoch": 1.5644171779141103, + "grad_norm": 15.33141936648023, + "learning_rate": 7.25206611570248e-07, + "logits/chosen": -0.10791015625, + "logits/rejected": -0.2578125, + "logps/chosen": -388.0, + "logps/rejected": -394.0, + "loss": 0.0744, + "rewards/accuracies": 0.96875, + "rewards/chosen": -0.111328125, + "rewards/margins": 10.6875, + "rewards/rejected": -10.8125, + "step": 1275 + }, + { + "epoch": 1.565644171779141, + "grad_norm": 7.988279468708803, + "learning_rate": 7.245867768595041e-07, + "logits/chosen": -0.025146484375, + "logits/rejected": -0.2138671875, + "logps/chosen": -418.0, + "logps/rejected": -374.0, + "loss": 0.0247, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.48046875, + "rewards/margins": 11.0, + "rewards/rejected": -10.5, + "step": 1276 + }, + { + "epoch": 1.5668711656441716, + "grad_norm": 12.554173231892838, + "learning_rate": 7.239669421487604e-07, + "logits/chosen": -0.185546875, + "logits/rejected": -0.314453125, + "logps/chosen": -358.0, + "logps/rejected": -354.0, + "loss": 0.043, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.034423828125, + "rewards/margins": 10.0, + "rewards/rejected": -10.0625, + "step": 1277 + }, + { + "epoch": 1.5680981595092025, + "grad_norm": 15.815759980577592, + "learning_rate": 7.233471074380165e-07, + "logits/chosen": -0.15625, + "logits/rejected": -0.365234375, + "logps/chosen": -404.0, + "logps/rejected": -396.0, + "loss": 0.028, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.77734375, + "rewards/margins": 12.125, + "rewards/rejected": -11.3125, + "step": 1278 + }, + { + "epoch": 1.5693251533742332, + "grad_norm": 13.88863047530027, + "learning_rate": 7.227272727272728e-07, + "logits/chosen": -0.169921875, + "logits/rejected": -0.310546875, + "logps/chosen": -356.0, + "logps/rejected": -394.0, + "loss": 0.0486, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.1826171875, + "rewards/margins": 10.6875, + "rewards/rejected": -10.5, + "step": 1279 + }, + { + "epoch": 1.5705521472392638, + "grad_norm": 11.958915852962361, + "learning_rate": 7.221074380165289e-07, + "logits/chosen": -0.2294921875, + "logits/rejected": -0.3515625, + "logps/chosen": -434.0, + "logps/rejected": -432.0, + "loss": 0.0382, + "rewards/accuracies": 0.9765625, + "rewards/chosen": 0.470703125, + "rewards/margins": 12.1875, + "rewards/rejected": -11.75, + "step": 1280 + }, + { + "epoch": 1.5717791411042945, + "grad_norm": 13.174861138807707, + "learning_rate": 7.214876033057852e-07, + "logits/chosen": -0.2236328125, + "logits/rejected": -0.384765625, + "logps/chosen": -438.0, + "logps/rejected": -440.0, + "loss": 0.0556, + "rewards/accuracies": 0.96875, + "rewards/chosen": 1.125, + "rewards/margins": 12.6875, + "rewards/rejected": -11.5625, + "step": 1281 + }, + { + "epoch": 1.5730061349693252, + "grad_norm": 9.377186147771738, + "learning_rate": 7.208677685950413e-07, + "logits/chosen": 0.048828125, + "logits/rejected": -0.185546875, + "logps/chosen": -394.0, + "logps/rejected": -374.0, + "loss": 0.0309, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.032470703125, + "rewards/margins": 11.125, + "rewards/rejected": -11.125, + "step": 1282 + }, + { + "epoch": 1.5742331288343558, + "grad_norm": 9.78074406651563, + "learning_rate": 7.202479338842976e-07, + "logits/chosen": -0.0028228759765625, + "logits/rejected": -0.11865234375, + "logps/chosen": -354.0, + "logps/rejected": -380.0, + "loss": 0.0371, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.88671875, + "rewards/margins": 10.8125, + "rewards/rejected": -11.6875, + "step": 1283 + }, + { + "epoch": 1.5754601226993865, + "grad_norm": 15.650011079839588, + "learning_rate": 7.196280991735537e-07, + "logits/chosen": -0.072265625, + "logits/rejected": -0.2490234375, + "logps/chosen": -404.0, + "logps/rejected": -420.0, + "loss": 0.0681, + "rewards/accuracies": 0.9765625, + "rewards/chosen": -0.33203125, + "rewards/margins": 12.1875, + "rewards/rejected": -12.5, + "step": 1284 + }, + { + "epoch": 1.5766871165644172, + "grad_norm": 15.27608982764389, + "learning_rate": 7.190082644628099e-07, + "logits/chosen": -0.130859375, + "logits/rejected": -0.373046875, + "logps/chosen": -422.0, + "logps/rejected": -420.0, + "loss": 0.0468, + "rewards/accuracies": 0.9765625, + "rewards/chosen": -0.109375, + "rewards/margins": 12.1875, + "rewards/rejected": -12.3125, + "step": 1285 + }, + { + "epoch": 1.5779141104294478, + "grad_norm": 15.001506307085636, + "learning_rate": 7.183884297520662e-07, + "logits/chosen": -0.0771484375, + "logits/rejected": -0.271484375, + "logps/chosen": -372.0, + "logps/rejected": -356.0, + "loss": 0.0383, + "rewards/accuracies": 0.984375, + "rewards/chosen": -0.921875, + "rewards/margins": 10.75, + "rewards/rejected": -11.6875, + "step": 1286 + }, + { + "epoch": 1.5791411042944785, + "grad_norm": 5.225854839258308, + "learning_rate": 7.177685950413223e-07, + "logits/chosen": -0.1748046875, + "logits/rejected": -0.251953125, + "logps/chosen": -330.0, + "logps/rejected": -380.0, + "loss": 0.0262, + "rewards/accuracies": 0.984375, + "rewards/chosen": -1.375, + "rewards/margins": 11.0, + "rewards/rejected": -12.375, + "step": 1287 + }, + { + "epoch": 1.5803680981595092, + "grad_norm": 11.803831089679342, + "learning_rate": 7.171487603305785e-07, + "logits/chosen": -0.2255859375, + "logits/rejected": -0.400390625, + "logps/chosen": -394.0, + "logps/rejected": -412.0, + "loss": 0.0419, + "rewards/accuracies": 0.96875, + "rewards/chosen": -0.2236328125, + "rewards/margins": 12.125, + "rewards/rejected": -12.3125, + "step": 1288 + }, + { + "epoch": 1.5815950920245399, + "grad_norm": 9.347363939888405, + "learning_rate": 7.165289256198347e-07, + "logits/chosen": -0.087890625, + "logits/rejected": -0.2353515625, + "logps/chosen": -334.0, + "logps/rejected": -342.0, + "loss": 0.0396, + "rewards/accuracies": 0.984375, + "rewards/chosen": -1.0390625, + "rewards/margins": 10.75, + "rewards/rejected": -11.8125, + "step": 1289 + }, + { + "epoch": 1.5828220858895705, + "grad_norm": 4.3767999812751714, + "learning_rate": 7.15909090909091e-07, + "logits/chosen": -0.2294921875, + "logits/rejected": -0.40234375, + "logps/chosen": -396.0, + "logps/rejected": -416.0, + "loss": 0.0096, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.8984375, + "rewards/margins": 12.375, + "rewards/rejected": -13.25, + "step": 1290 + }, + { + "epoch": 1.5840490797546012, + "grad_norm": 4.228705291026931, + "learning_rate": 7.152892561983471e-07, + "logits/chosen": -0.099609375, + "logits/rejected": -0.21484375, + "logps/chosen": -372.0, + "logps/rejected": -376.0, + "loss": 0.0152, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.875, + "rewards/margins": 11.3125, + "rewards/rejected": -12.1875, + "step": 1291 + }, + { + "epoch": 1.5852760736196319, + "grad_norm": 7.584547777512873, + "learning_rate": 7.146694214876034e-07, + "logits/chosen": -0.25, + "logits/rejected": -0.35546875, + "logps/chosen": -390.0, + "logps/rejected": -428.0, + "loss": 0.0277, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.67578125, + "rewards/margins": 12.1875, + "rewards/rejected": -12.875, + "step": 1292 + }, + { + "epoch": 1.5865030674846625, + "grad_norm": 10.655223291637148, + "learning_rate": 7.140495867768595e-07, + "logits/chosen": -0.0137939453125, + "logits/rejected": -0.28515625, + "logps/chosen": -416.0, + "logps/rejected": -384.0, + "loss": 0.0202, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.25390625, + "rewards/margins": 11.375, + "rewards/rejected": -11.625, + "step": 1293 + }, + { + "epoch": 1.5877300613496934, + "grad_norm": 4.2898047053372155, + "learning_rate": 7.134297520661157e-07, + "logits/chosen": -0.11279296875, + "logits/rejected": -0.30078125, + "logps/chosen": -412.0, + "logps/rejected": -428.0, + "loss": 0.0154, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.51953125, + "rewards/margins": 12.25, + "rewards/rejected": -12.8125, + "step": 1294 + }, + { + "epoch": 1.588957055214724, + "grad_norm": 8.499456654824181, + "learning_rate": 7.12809917355372e-07, + "logits/chosen": -0.2470703125, + "logits/rejected": -0.431640625, + "logps/chosen": -370.0, + "logps/rejected": -378.0, + "loss": 0.0387, + "rewards/accuracies": 0.9765625, + "rewards/chosen": -0.59375, + "rewards/margins": 11.0, + "rewards/rejected": -11.5625, + "step": 1295 + }, + { + "epoch": 1.5901840490797547, + "grad_norm": 16.676691340584743, + "learning_rate": 7.121900826446281e-07, + "logits/chosen": -0.1220703125, + "logits/rejected": -0.2099609375, + "logps/chosen": -394.0, + "logps/rejected": -402.0, + "loss": 0.0806, + "rewards/accuracies": 0.96875, + "rewards/chosen": -1.234375, + "rewards/margins": 10.75, + "rewards/rejected": -12.0, + "step": 1296 + }, + { + "epoch": 1.5914110429447854, + "grad_norm": 9.185721305153498, + "learning_rate": 7.115702479338843e-07, + "logits/chosen": -0.042236328125, + "logits/rejected": -0.2080078125, + "logps/chosen": -414.0, + "logps/rejected": -384.0, + "loss": 0.0365, + "rewards/accuracies": 0.984375, + "rewards/chosen": -0.41015625, + "rewards/margins": 9.875, + "rewards/rejected": -10.25, + "step": 1297 + }, + { + "epoch": 1.592638036809816, + "grad_norm": 9.516915353409866, + "learning_rate": 7.109504132231405e-07, + "logits/chosen": -0.07568359375, + "logits/rejected": -0.197265625, + "logps/chosen": -410.0, + "logps/rejected": -414.0, + "loss": 0.0373, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.279296875, + "rewards/margins": 11.75, + "rewards/rejected": -11.4375, + "step": 1298 + }, + { + "epoch": 1.5938650306748468, + "grad_norm": 12.688329159049632, + "learning_rate": 7.103305785123968e-07, + "logits/chosen": -0.197265625, + "logits/rejected": -0.28515625, + "logps/chosen": -394.0, + "logps/rejected": -402.0, + "loss": 0.0515, + "rewards/accuracies": 0.984375, + "rewards/chosen": -0.294921875, + "rewards/margins": 10.9375, + "rewards/rejected": -11.1875, + "step": 1299 + }, + { + "epoch": 1.5950920245398774, + "grad_norm": 9.480010594113034, + "learning_rate": 7.097107438016529e-07, + "logits/chosen": -0.08447265625, + "logits/rejected": -0.283203125, + "logps/chosen": -406.0, + "logps/rejected": -410.0, + "loss": 0.0238, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.345703125, + "rewards/margins": 11.8125, + "rewards/rejected": -11.4375, + "step": 1300 + }, + { + "epoch": 1.596319018404908, + "grad_norm": 6.743619593496769, + "learning_rate": 7.09090909090909e-07, + "logits/chosen": -0.07958984375, + "logits/rejected": -0.27734375, + "logps/chosen": -384.0, + "logps/rejected": -394.0, + "loss": 0.0206, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.045166015625, + "rewards/margins": 11.1875, + "rewards/rejected": -11.25, + "step": 1301 + }, + { + "epoch": 1.5975460122699388, + "grad_norm": 10.079704999041756, + "learning_rate": 7.084710743801653e-07, + "logits/chosen": -0.283203125, + "logits/rejected": -0.396484375, + "logps/chosen": -404.0, + "logps/rejected": -402.0, + "loss": 0.0333, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.49609375, + "rewards/margins": 11.8125, + "rewards/rejected": -11.3125, + "step": 1302 + }, + { + "epoch": 1.5987730061349694, + "grad_norm": 4.128642305901009, + "learning_rate": 7.078512396694215e-07, + "logits/chosen": -0.15625, + "logits/rejected": -0.369140625, + "logps/chosen": -440.0, + "logps/rejected": -408.0, + "loss": 0.0206, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.6328125, + "rewards/margins": 12.5625, + "rewards/rejected": -11.9375, + "step": 1303 + }, + { + "epoch": 1.6, + "grad_norm": 6.006214161041272, + "learning_rate": 7.072314049586777e-07, + "logits/chosen": -0.2255859375, + "logits/rejected": -0.388671875, + "logps/chosen": -380.0, + "logps/rejected": -408.0, + "loss": 0.0223, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.7421875, + "rewards/margins": 12.3125, + "rewards/rejected": -11.5625, + "step": 1304 + }, + { + "epoch": 1.6012269938650308, + "grad_norm": 15.374011330047086, + "learning_rate": 7.066115702479339e-07, + "logits/chosen": -0.12451171875, + "logits/rejected": -0.259765625, + "logps/chosen": -346.0, + "logps/rejected": -366.0, + "loss": 0.068, + "rewards/accuracies": 0.9765625, + "rewards/chosen": 0.423828125, + "rewards/margins": 10.875, + "rewards/rejected": -10.4375, + "step": 1305 + }, + { + "epoch": 1.6024539877300614, + "grad_norm": 10.751324968165987, + "learning_rate": 7.059917355371901e-07, + "logits/chosen": -0.1953125, + "logits/rejected": -0.294921875, + "logps/chosen": -406.0, + "logps/rejected": -384.0, + "loss": 0.0244, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.44921875, + "rewards/margins": 11.3125, + "rewards/rejected": -10.875, + "step": 1306 + }, + { + "epoch": 1.603680981595092, + "grad_norm": 7.9285821638566905, + "learning_rate": 7.053719008264463e-07, + "logits/chosen": -0.1953125, + "logits/rejected": -0.38671875, + "logps/chosen": -440.0, + "logps/rejected": -402.0, + "loss": 0.0207, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 1.0078125, + "rewards/margins": 11.25, + "rewards/rejected": -10.25, + "step": 1307 + }, + { + "epoch": 1.6049079754601228, + "grad_norm": 4.924145204402143, + "learning_rate": 7.047520661157026e-07, + "logits/chosen": -0.25, + "logits/rejected": -0.392578125, + "logps/chosen": -380.0, + "logps/rejected": -372.0, + "loss": 0.0136, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.875, + "rewards/margins": 11.5625, + "rewards/rejected": -10.6875, + "step": 1308 + }, + { + "epoch": 1.6061349693251534, + "grad_norm": 11.95129952622703, + "learning_rate": 7.041322314049587e-07, + "logits/chosen": -0.185546875, + "logits/rejected": -0.3046875, + "logps/chosen": -392.0, + "logps/rejected": -394.0, + "loss": 0.0371, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.263671875, + "rewards/margins": 11.25, + "rewards/rejected": -11.0, + "step": 1309 + }, + { + "epoch": 1.607361963190184, + "grad_norm": 6.935480593251807, + "learning_rate": 7.035123966942149e-07, + "logits/chosen": -0.09619140625, + "logits/rejected": -0.2578125, + "logps/chosen": -334.0, + "logps/rejected": -356.0, + "loss": 0.0257, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.7109375, + "rewards/margins": 11.125, + "rewards/rejected": -10.4375, + "step": 1310 + }, + { + "epoch": 1.6085889570552148, + "grad_norm": 7.070447805786784, + "learning_rate": 7.028925619834711e-07, + "logits/chosen": -0.046142578125, + "logits/rejected": -0.24609375, + "logps/chosen": -418.0, + "logps/rejected": -380.0, + "loss": 0.0364, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.0751953125, + "rewards/margins": 9.9375, + "rewards/rejected": -9.8125, + "step": 1311 + }, + { + "epoch": 1.6098159509202454, + "grad_norm": 6.5989010041947065, + "learning_rate": 7.022727272727273e-07, + "logits/chosen": -0.2216796875, + "logits/rejected": -0.453125, + "logps/chosen": -404.0, + "logps/rejected": -392.0, + "loss": 0.0213, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.75390625, + "rewards/margins": 12.3125, + "rewards/rejected": -11.5625, + "step": 1312 + }, + { + "epoch": 1.611042944785276, + "grad_norm": 5.053991836327145, + "learning_rate": 7.016528925619835e-07, + "logits/chosen": -0.1376953125, + "logits/rejected": -0.28125, + "logps/chosen": -366.0, + "logps/rejected": -374.0, + "loss": 0.0231, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.2294921875, + "rewards/margins": 10.75, + "rewards/rejected": -10.5, + "step": 1313 + }, + { + "epoch": 1.6122699386503068, + "grad_norm": 7.196334947724689, + "learning_rate": 7.010330578512396e-07, + "logits/chosen": -0.310546875, + "logits/rejected": -0.390625, + "logps/chosen": -392.0, + "logps/rejected": -408.0, + "loss": 0.0192, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.99609375, + "rewards/margins": 12.4375, + "rewards/rejected": -11.4375, + "step": 1314 + }, + { + "epoch": 1.6134969325153374, + "grad_norm": 7.456124142101998, + "learning_rate": 7.004132231404959e-07, + "logits/chosen": -0.158203125, + "logits/rejected": -0.384765625, + "logps/chosen": -452.0, + "logps/rejected": -422.0, + "loss": 0.0295, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.396484375, + "rewards/margins": 11.5625, + "rewards/rejected": -11.1875, + "step": 1315 + }, + { + "epoch": 1.614723926380368, + "grad_norm": 14.771588037378546, + "learning_rate": 6.997933884297521e-07, + "logits/chosen": -0.022705078125, + "logits/rejected": -0.2138671875, + "logps/chosen": -378.0, + "logps/rejected": -366.0, + "loss": 0.0521, + "rewards/accuracies": 0.984375, + "rewards/chosen": -0.09912109375, + "rewards/margins": 10.1875, + "rewards/rejected": -10.3125, + "step": 1316 + }, + { + "epoch": 1.6159509202453988, + "grad_norm": 6.5021730031025555, + "learning_rate": 6.991735537190084e-07, + "logits/chosen": -0.19140625, + "logits/rejected": -0.296875, + "logps/chosen": -388.0, + "logps/rejected": -414.0, + "loss": 0.024, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.146484375, + "rewards/margins": 11.1875, + "rewards/rejected": -11.0625, + "step": 1317 + }, + { + "epoch": 1.6171779141104294, + "grad_norm": 7.036484087537271, + "learning_rate": 6.985537190082644e-07, + "logits/chosen": -0.126953125, + "logits/rejected": -0.333984375, + "logps/chosen": -388.0, + "logps/rejected": -382.0, + "loss": 0.0165, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.037109375, + "rewards/margins": 11.375, + "rewards/rejected": -11.3125, + "step": 1318 + }, + { + "epoch": 1.61840490797546, + "grad_norm": 12.457147867029073, + "learning_rate": 6.979338842975207e-07, + "logits/chosen": -0.1552734375, + "logits/rejected": -0.30859375, + "logps/chosen": -390.0, + "logps/rejected": -424.0, + "loss": 0.0243, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.474609375, + "rewards/margins": 12.0, + "rewards/rejected": -11.5625, + "step": 1319 + }, + { + "epoch": 1.6196319018404908, + "grad_norm": 18.65684903039485, + "learning_rate": 6.973140495867769e-07, + "logits/chosen": -0.123046875, + "logits/rejected": -0.314453125, + "logps/chosen": -438.0, + "logps/rejected": -406.0, + "loss": 0.0705, + "rewards/accuracies": 0.9765625, + "rewards/chosen": 0.4375, + "rewards/margins": 11.625, + "rewards/rejected": -11.1875, + "step": 1320 + }, + { + "epoch": 1.6208588957055214, + "grad_norm": 15.725190139889248, + "learning_rate": 6.966942148760331e-07, + "logits/chosen": -0.16796875, + "logits/rejected": -0.296875, + "logps/chosen": -384.0, + "logps/rejected": -376.0, + "loss": 0.0439, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.17578125, + "rewards/margins": 11.375, + "rewards/rejected": -11.25, + "step": 1321 + }, + { + "epoch": 1.622085889570552, + "grad_norm": 13.624128902680484, + "learning_rate": 6.960743801652893e-07, + "logits/chosen": -0.189453125, + "logits/rejected": -0.345703125, + "logps/chosen": -366.0, + "logps/rejected": -396.0, + "loss": 0.0327, + "rewards/accuracies": 0.984375, + "rewards/chosen": -0.1865234375, + "rewards/margins": 10.9375, + "rewards/rejected": -11.125, + "step": 1322 + }, + { + "epoch": 1.6233128834355828, + "grad_norm": 2.854089790732775, + "learning_rate": 6.954545454545454e-07, + "logits/chosen": -0.0157470703125, + "logits/rejected": -0.208984375, + "logps/chosen": -336.0, + "logps/rejected": -376.0, + "loss": 0.0118, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.03955078125, + "rewards/margins": 11.375, + "rewards/rejected": -11.375, + "step": 1323 + }, + { + "epoch": 1.6245398773006134, + "grad_norm": 13.205081682962938, + "learning_rate": 6.948347107438017e-07, + "logits/chosen": -0.1826171875, + "logits/rejected": -0.404296875, + "logps/chosen": -416.0, + "logps/rejected": -394.0, + "loss": 0.0504, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.265625, + "rewards/margins": 11.875, + "rewards/rejected": -11.625, + "step": 1324 + }, + { + "epoch": 1.6257668711656441, + "grad_norm": 10.512892902569012, + "learning_rate": 6.942148760330579e-07, + "logits/chosen": -0.07275390625, + "logits/rejected": -0.33203125, + "logps/chosen": -388.0, + "logps/rejected": -390.0, + "loss": 0.0409, + "rewards/accuracies": 0.984375, + "rewards/chosen": -0.427734375, + "rewards/margins": 11.3125, + "rewards/rejected": -11.75, + "step": 1325 + }, + { + "epoch": 1.6269938650306748, + "grad_norm": 16.75758409246438, + "learning_rate": 6.935950413223141e-07, + "logits/chosen": -0.19921875, + "logits/rejected": -0.33203125, + "logps/chosen": -412.0, + "logps/rejected": -418.0, + "loss": 0.0433, + "rewards/accuracies": 0.9765625, + "rewards/chosen": -0.076171875, + "rewards/margins": 12.25, + "rewards/rejected": -12.3125, + "step": 1326 + }, + { + "epoch": 1.6282208588957054, + "grad_norm": 17.757254444197297, + "learning_rate": 6.929752066115702e-07, + "logits/chosen": -0.1201171875, + "logits/rejected": -0.2265625, + "logps/chosen": -372.0, + "logps/rejected": -406.0, + "loss": 0.095, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.80859375, + "rewards/margins": 10.875, + "rewards/rejected": -11.6875, + "step": 1327 + }, + { + "epoch": 1.6294478527607361, + "grad_norm": 13.567940362966437, + "learning_rate": 6.923553719008265e-07, + "logits/chosen": -0.1953125, + "logits/rejected": -0.40625, + "logps/chosen": -414.0, + "logps/rejected": -398.0, + "loss": 0.0402, + "rewards/accuracies": 0.9765625, + "rewards/chosen": -0.330078125, + "rewards/margins": 11.75, + "rewards/rejected": -12.0625, + "step": 1328 + }, + { + "epoch": 1.6306748466257668, + "grad_norm": 15.662951369347224, + "learning_rate": 6.917355371900827e-07, + "logits/chosen": -0.2353515625, + "logits/rejected": -0.41015625, + "logps/chosen": -422.0, + "logps/rejected": -414.0, + "loss": 0.0593, + "rewards/accuracies": 0.96875, + "rewards/chosen": -0.6640625, + "rewards/margins": 10.75, + "rewards/rejected": -11.4375, + "step": 1329 + }, + { + "epoch": 1.6319018404907975, + "grad_norm": 10.57654744394464, + "learning_rate": 6.911157024793388e-07, + "logits/chosen": -0.12158203125, + "logits/rejected": -0.33203125, + "logps/chosen": -404.0, + "logps/rejected": -404.0, + "loss": 0.032, + "rewards/accuracies": 0.9765625, + "rewards/chosen": -0.37109375, + "rewards/margins": 11.5625, + "rewards/rejected": -11.875, + "step": 1330 + }, + { + "epoch": 1.6331288343558281, + "grad_norm": 7.812525368210755, + "learning_rate": 6.90495867768595e-07, + "logits/chosen": -0.169921875, + "logits/rejected": -0.2392578125, + "logps/chosen": -366.0, + "logps/rejected": -388.0, + "loss": 0.0254, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -1.0546875, + "rewards/margins": 11.8125, + "rewards/rejected": -12.875, + "step": 1331 + }, + { + "epoch": 1.6343558282208588, + "grad_norm": 12.529370972774968, + "learning_rate": 6.898760330578512e-07, + "logits/chosen": -0.263671875, + "logits/rejected": -0.416015625, + "logps/chosen": -386.0, + "logps/rejected": -390.0, + "loss": 0.0267, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.328125, + "rewards/margins": 11.375, + "rewards/rejected": -11.75, + "step": 1332 + }, + { + "epoch": 1.6355828220858895, + "grad_norm": 15.494182278895307, + "learning_rate": 6.892561983471075e-07, + "logits/chosen": -0.27734375, + "logits/rejected": -0.416015625, + "logps/chosen": -420.0, + "logps/rejected": -440.0, + "loss": 0.0527, + "rewards/accuracies": 0.96875, + "rewards/chosen": -0.6953125, + "rewards/margins": 11.1875, + "rewards/rejected": -11.875, + "step": 1333 + }, + { + "epoch": 1.6368098159509201, + "grad_norm": 8.698893043286152, + "learning_rate": 6.886363636363637e-07, + "logits/chosen": -0.10009765625, + "logits/rejected": -0.244140625, + "logps/chosen": -382.0, + "logps/rejected": -400.0, + "loss": 0.0306, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.5234375, + "rewards/margins": 12.1875, + "rewards/rejected": -11.6875, + "step": 1334 + }, + { + "epoch": 1.6380368098159508, + "grad_norm": 14.987956201874853, + "learning_rate": 6.880165289256199e-07, + "logits/chosen": -0.2451171875, + "logits/rejected": -0.44921875, + "logps/chosen": -432.0, + "logps/rejected": -386.0, + "loss": 0.0336, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 1.109375, + "rewards/margins": 11.625, + "rewards/rejected": -10.5625, + "step": 1335 + }, + { + "epoch": 1.6392638036809815, + "grad_norm": 15.60191378628464, + "learning_rate": 6.87396694214876e-07, + "logits/chosen": -0.361328125, + "logits/rejected": -0.53125, + "logps/chosen": -436.0, + "logps/rejected": -436.0, + "loss": 0.0541, + "rewards/accuracies": 0.9765625, + "rewards/chosen": -0.53515625, + "rewards/margins": 12.0625, + "rewards/rejected": -12.5625, + "step": 1336 + }, + { + "epoch": 1.6404907975460121, + "grad_norm": 6.108403854935857, + "learning_rate": 6.867768595041323e-07, + "logits/chosen": -0.2001953125, + "logits/rejected": -0.376953125, + "logps/chosen": -414.0, + "logps/rejected": -384.0, + "loss": 0.0178, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.1640625, + "rewards/margins": 11.8125, + "rewards/rejected": -11.6875, + "step": 1337 + }, + { + "epoch": 1.6417177914110428, + "grad_norm": 6.235103209750328, + "learning_rate": 6.861570247933885e-07, + "logits/chosen": -0.2734375, + "logits/rejected": -0.400390625, + "logps/chosen": -432.0, + "logps/rejected": -410.0, + "loss": 0.0253, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.435546875, + "rewards/margins": 11.625, + "rewards/rejected": -11.1875, + "step": 1338 + }, + { + "epoch": 1.6429447852760735, + "grad_norm": 13.809775053491462, + "learning_rate": 6.855371900826446e-07, + "logits/chosen": -0.13671875, + "logits/rejected": -0.1923828125, + "logps/chosen": -390.0, + "logps/rejected": -346.0, + "loss": 0.0495, + "rewards/accuracies": 0.984375, + "rewards/chosen": -0.033935546875, + "rewards/margins": 10.375, + "rewards/rejected": -10.375, + "step": 1339 + }, + { + "epoch": 1.6441717791411041, + "grad_norm": 14.649539845278797, + "learning_rate": 6.849173553719008e-07, + "logits/chosen": -0.060791015625, + "logits/rejected": -0.220703125, + "logps/chosen": -412.0, + "logps/rejected": -394.0, + "loss": 0.0626, + "rewards/accuracies": 0.9765625, + "rewards/chosen": 0.271484375, + "rewards/margins": 10.4375, + "rewards/rejected": -10.1875, + "step": 1340 + }, + { + "epoch": 1.645398773006135, + "grad_norm": 6.096103726931023, + "learning_rate": 6.84297520661157e-07, + "logits/chosen": -0.224609375, + "logits/rejected": -0.453125, + "logps/chosen": -438.0, + "logps/rejected": -398.0, + "loss": 0.0186, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.68359375, + "rewards/margins": 12.1875, + "rewards/rejected": -11.5, + "step": 1341 + }, + { + "epoch": 1.6466257668711657, + "grad_norm": 4.043549433090365, + "learning_rate": 6.836776859504133e-07, + "logits/chosen": -0.251953125, + "logits/rejected": -0.470703125, + "logps/chosen": -410.0, + "logps/rejected": -378.0, + "loss": 0.0267, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.60546875, + "rewards/margins": 11.0625, + "rewards/rejected": -10.4375, + "step": 1342 + }, + { + "epoch": 1.6478527607361964, + "grad_norm": 22.56474705026688, + "learning_rate": 6.830578512396694e-07, + "logits/chosen": -0.1708984375, + "logits/rejected": -0.40625, + "logps/chosen": -408.0, + "logps/rejected": -404.0, + "loss": 0.0661, + "rewards/accuracies": 0.9765625, + "rewards/chosen": -0.1650390625, + "rewards/margins": 10.75, + "rewards/rejected": -10.9375, + "step": 1343 + }, + { + "epoch": 1.649079754601227, + "grad_norm": 7.990667626479867, + "learning_rate": 6.824380165289256e-07, + "logits/chosen": -0.158203125, + "logits/rejected": -0.416015625, + "logps/chosen": -434.0, + "logps/rejected": -388.0, + "loss": 0.0183, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 1.203125, + "rewards/margins": 11.9375, + "rewards/rejected": -10.75, + "step": 1344 + }, + { + "epoch": 1.6503067484662577, + "grad_norm": 9.134005678853116, + "learning_rate": 6.818181818181818e-07, + "logits/chosen": -0.1943359375, + "logits/rejected": -0.34375, + "logps/chosen": -422.0, + "logps/rejected": -376.0, + "loss": 0.0366, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.361328125, + "rewards/margins": 10.0, + "rewards/rejected": -9.625, + "step": 1345 + }, + { + "epoch": 1.6515337423312884, + "grad_norm": 19.441190331880875, + "learning_rate": 6.811983471074381e-07, + "logits/chosen": -0.1259765625, + "logits/rejected": -0.32421875, + "logps/chosen": -438.0, + "logps/rejected": -396.0, + "loss": 0.0749, + "rewards/accuracies": 0.96875, + "rewards/chosen": 0.8125, + "rewards/margins": 11.125, + "rewards/rejected": -10.3125, + "step": 1346 + }, + { + "epoch": 1.652760736196319, + "grad_norm": 14.390784632473302, + "learning_rate": 6.805785123966942e-07, + "logits/chosen": -0.1416015625, + "logits/rejected": -0.3046875, + "logps/chosen": -420.0, + "logps/rejected": -420.0, + "loss": 0.0671, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.55078125, + "rewards/margins": 11.0, + "rewards/rejected": -10.4375, + "step": 1347 + }, + { + "epoch": 1.6539877300613497, + "grad_norm": 8.537444327620346, + "learning_rate": 6.799586776859504e-07, + "logits/chosen": -0.1455078125, + "logits/rejected": -0.32421875, + "logps/chosen": -340.0, + "logps/rejected": -382.0, + "loss": 0.0348, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.048095703125, + "rewards/margins": 10.6875, + "rewards/rejected": -10.6875, + "step": 1348 + }, + { + "epoch": 1.6552147239263804, + "grad_norm": 7.326685705649235, + "learning_rate": 6.793388429752066e-07, + "logits/chosen": -0.123046875, + "logits/rejected": -0.337890625, + "logps/chosen": -368.0, + "logps/rejected": -380.0, + "loss": 0.0303, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.80078125, + "rewards/margins": 11.1875, + "rewards/rejected": -10.375, + "step": 1349 + }, + { + "epoch": 1.656441717791411, + "grad_norm": 11.253855573956272, + "learning_rate": 6.787190082644628e-07, + "logits/chosen": -0.1669921875, + "logits/rejected": -0.3515625, + "logps/chosen": -422.0, + "logps/rejected": -432.0, + "loss": 0.0313, + "rewards/accuracies": 0.984375, + "rewards/chosen": 1.375, + "rewards/margins": 12.5, + "rewards/rejected": -11.125, + "step": 1350 + }, + { + "epoch": 1.6576687116564417, + "grad_norm": 20.220907692170936, + "learning_rate": 6.780991735537191e-07, + "logits/chosen": -0.11962890625, + "logits/rejected": -0.26171875, + "logps/chosen": -396.0, + "logps/rejected": -392.0, + "loss": 0.1149, + "rewards/accuracies": 0.953125, + "rewards/chosen": 0.1650390625, + "rewards/margins": 10.875, + "rewards/rejected": -10.75, + "step": 1351 + }, + { + "epoch": 1.6588957055214724, + "grad_norm": 8.156843128459297, + "learning_rate": 6.774793388429752e-07, + "logits/chosen": -0.2119140625, + "logits/rejected": -0.369140625, + "logps/chosen": -388.0, + "logps/rejected": -398.0, + "loss": 0.0315, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.2314453125, + "rewards/margins": 11.4375, + "rewards/rejected": -11.25, + "step": 1352 + }, + { + "epoch": 1.660122699386503, + "grad_norm": 6.558928813231135, + "learning_rate": 6.768595041322314e-07, + "logits/chosen": -0.10986328125, + "logits/rejected": -0.146484375, + "logps/chosen": -376.0, + "logps/rejected": -412.0, + "loss": 0.0286, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.0908203125, + "rewards/margins": 10.75, + "rewards/rejected": -10.8125, + "step": 1353 + }, + { + "epoch": 1.6613496932515337, + "grad_norm": 4.479021963688224, + "learning_rate": 6.762396694214876e-07, + "logits/chosen": -0.1826171875, + "logits/rejected": -0.333984375, + "logps/chosen": -368.0, + "logps/rejected": -394.0, + "loss": 0.025, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.41796875, + "rewards/margins": 11.6875, + "rewards/rejected": -11.25, + "step": 1354 + }, + { + "epoch": 1.6625766871165644, + "grad_norm": 12.69028856164417, + "learning_rate": 6.756198347107439e-07, + "logits/chosen": -0.20703125, + "logits/rejected": -0.345703125, + "logps/chosen": -424.0, + "logps/rejected": -422.0, + "loss": 0.0465, + "rewards/accuracies": 0.9765625, + "rewards/chosen": -0.20703125, + "rewards/margins": 10.875, + "rewards/rejected": -11.0625, + "step": 1355 + }, + { + "epoch": 1.6638036809815953, + "grad_norm": 6.416852390478722, + "learning_rate": 6.75e-07, + "logits/chosen": -0.16015625, + "logits/rejected": -0.322265625, + "logps/chosen": -408.0, + "logps/rejected": -412.0, + "loss": 0.0263, + "rewards/accuracies": 0.984375, + "rewards/chosen": -0.396484375, + "rewards/margins": 11.875, + "rewards/rejected": -12.25, + "step": 1356 + }, + { + "epoch": 1.665030674846626, + "grad_norm": 15.843916290648238, + "learning_rate": 6.743801652892562e-07, + "logits/chosen": -0.1259765625, + "logits/rejected": -0.353515625, + "logps/chosen": -436.0, + "logps/rejected": -410.0, + "loss": 0.0297, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.30859375, + "rewards/margins": 12.0, + "rewards/rejected": -11.6875, + "step": 1357 + }, + { + "epoch": 1.6662576687116566, + "grad_norm": 15.902047502639894, + "learning_rate": 6.737603305785124e-07, + "logits/chosen": -0.2080078125, + "logits/rejected": -0.453125, + "logps/chosen": -400.0, + "logps/rejected": -388.0, + "loss": 0.06, + "rewards/accuracies": 0.9765625, + "rewards/chosen": -0.11328125, + "rewards/margins": 10.8125, + "rewards/rejected": -10.9375, + "step": 1358 + }, + { + "epoch": 1.6674846625766873, + "grad_norm": 3.3586369592233085, + "learning_rate": 6.731404958677686e-07, + "logits/chosen": -0.1201171875, + "logits/rejected": -0.318359375, + "logps/chosen": -410.0, + "logps/rejected": -412.0, + "loss": 0.0118, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.6171875, + "rewards/margins": 12.3125, + "rewards/rejected": -11.6875, + "step": 1359 + }, + { + "epoch": 1.668711656441718, + "grad_norm": 12.937160200064577, + "learning_rate": 6.725206611570248e-07, + "logits/chosen": -0.388671875, + "logits/rejected": -0.6171875, + "logps/chosen": -392.0, + "logps/rejected": -380.0, + "loss": 0.0478, + "rewards/accuracies": 0.96875, + "rewards/chosen": 0.66015625, + "rewards/margins": 11.9375, + "rewards/rejected": -11.3125, + "step": 1360 + }, + { + "epoch": 1.6699386503067486, + "grad_norm": 11.349224605689916, + "learning_rate": 6.71900826446281e-07, + "logits/chosen": -0.0869140625, + "logits/rejected": -0.251953125, + "logps/chosen": -416.0, + "logps/rejected": -422.0, + "loss": 0.0391, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.201171875, + "rewards/margins": 11.0625, + "rewards/rejected": -10.875, + "step": 1361 + }, + { + "epoch": 1.6711656441717793, + "grad_norm": 16.146708757099134, + "learning_rate": 6.712809917355372e-07, + "logits/chosen": -0.119140625, + "logits/rejected": -0.2578125, + "logps/chosen": -402.0, + "logps/rejected": -376.0, + "loss": 0.051, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.474609375, + "rewards/margins": 11.5625, + "rewards/rejected": -11.0625, + "step": 1362 + }, + { + "epoch": 1.67239263803681, + "grad_norm": 3.943632661486204, + "learning_rate": 6.706611570247934e-07, + "logits/chosen": -0.224609375, + "logits/rejected": -0.41796875, + "logps/chosen": -398.0, + "logps/rejected": -366.0, + "loss": 0.0129, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.48046875, + "rewards/margins": 11.5625, + "rewards/rejected": -11.125, + "step": 1363 + }, + { + "epoch": 1.6736196319018406, + "grad_norm": 4.1389200081135655, + "learning_rate": 6.700413223140497e-07, + "logits/chosen": -0.09033203125, + "logits/rejected": -0.279296875, + "logps/chosen": -374.0, + "logps/rejected": -356.0, + "loss": 0.0111, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.453125, + "rewards/margins": 11.5625, + "rewards/rejected": -11.125, + "step": 1364 + }, + { + "epoch": 1.6748466257668713, + "grad_norm": 10.368495733373145, + "learning_rate": 6.694214876033058e-07, + "logits/chosen": -0.0458984375, + "logits/rejected": -0.2275390625, + "logps/chosen": -408.0, + "logps/rejected": -370.0, + "loss": 0.033, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.6640625, + "rewards/margins": 11.5625, + "rewards/rejected": -10.875, + "step": 1365 + }, + { + "epoch": 1.676073619631902, + "grad_norm": 10.90227263887412, + "learning_rate": 6.68801652892562e-07, + "logits/chosen": -0.12353515625, + "logits/rejected": -0.34765625, + "logps/chosen": -398.0, + "logps/rejected": -368.0, + "loss": 0.034, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.7109375, + "rewards/margins": 10.875, + "rewards/rejected": -10.125, + "step": 1366 + }, + { + "epoch": 1.6773006134969326, + "grad_norm": 14.33752123603866, + "learning_rate": 6.681818181818182e-07, + "logits/chosen": -0.1025390625, + "logits/rejected": -0.25390625, + "logps/chosen": -394.0, + "logps/rejected": -400.0, + "loss": 0.0614, + "rewards/accuracies": 0.9609375, + "rewards/chosen": 0.62109375, + "rewards/margins": 11.75, + "rewards/rejected": -11.125, + "step": 1367 + }, + { + "epoch": 1.6785276073619633, + "grad_norm": 10.4754734935581, + "learning_rate": 6.675619834710744e-07, + "logits/chosen": -0.3203125, + "logits/rejected": -0.462890625, + "logps/chosen": -400.0, + "logps/rejected": -390.0, + "loss": 0.0374, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.365234375, + "rewards/margins": 11.5, + "rewards/rejected": -11.125, + "step": 1368 + }, + { + "epoch": 1.679754601226994, + "grad_norm": 13.976696266239935, + "learning_rate": 6.669421487603306e-07, + "logits/chosen": -0.2451171875, + "logits/rejected": -0.4453125, + "logps/chosen": -404.0, + "logps/rejected": -394.0, + "loss": 0.0565, + "rewards/accuracies": 0.9765625, + "rewards/chosen": 0.10986328125, + "rewards/margins": 10.8125, + "rewards/rejected": -10.6875, + "step": 1369 + }, + { + "epoch": 1.6809815950920246, + "grad_norm": 10.93083816897367, + "learning_rate": 6.663223140495868e-07, + "logits/chosen": -0.1865234375, + "logits/rejected": -0.296875, + "logps/chosen": -382.0, + "logps/rejected": -402.0, + "loss": 0.0301, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.66796875, + "rewards/margins": 11.5, + "rewards/rejected": -10.875, + "step": 1370 + }, + { + "epoch": 1.6822085889570553, + "grad_norm": 23.02977282251546, + "learning_rate": 6.65702479338843e-07, + "logits/chosen": -0.1201171875, + "logits/rejected": -0.388671875, + "logps/chosen": -468.0, + "logps/rejected": -428.0, + "loss": 0.0674, + "rewards/accuracies": 0.9765625, + "rewards/chosen": 0.79296875, + "rewards/margins": 11.8125, + "rewards/rejected": -11.0, + "step": 1371 + }, + { + "epoch": 1.683435582822086, + "grad_norm": 6.529767256435788, + "learning_rate": 6.650826446280992e-07, + "logits/chosen": -0.1787109375, + "logits/rejected": -0.404296875, + "logps/chosen": -392.0, + "logps/rejected": -384.0, + "loss": 0.0215, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.6328125, + "rewards/margins": 11.6875, + "rewards/rejected": -11.0625, + "step": 1372 + }, + { + "epoch": 1.6846625766871166, + "grad_norm": 16.218767103827304, + "learning_rate": 6.644628099173554e-07, + "logits/chosen": -0.1318359375, + "logits/rejected": -0.216796875, + "logps/chosen": -376.0, + "logps/rejected": -410.0, + "loss": 0.0695, + "rewards/accuracies": 0.96875, + "rewards/chosen": 0.41796875, + "rewards/margins": 10.9375, + "rewards/rejected": -10.5, + "step": 1373 + }, + { + "epoch": 1.6858895705521473, + "grad_norm": 12.743040998410756, + "learning_rate": 6.638429752066116e-07, + "logits/chosen": -0.056396484375, + "logits/rejected": -0.29296875, + "logps/chosen": -394.0, + "logps/rejected": -368.0, + "loss": 0.0465, + "rewards/accuracies": 0.9765625, + "rewards/chosen": 0.1943359375, + "rewards/margins": 10.625, + "rewards/rejected": -10.4375, + "step": 1374 + }, + { + "epoch": 1.687116564417178, + "grad_norm": 11.631482703134425, + "learning_rate": 6.632231404958678e-07, + "logits/chosen": -0.0947265625, + "logits/rejected": -0.1962890625, + "logps/chosen": -328.0, + "logps/rejected": -358.0, + "loss": 0.0442, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.162109375, + "rewards/margins": 10.4375, + "rewards/rejected": -10.25, + "step": 1375 + }, + { + "epoch": 1.6883435582822086, + "grad_norm": 8.378575051965539, + "learning_rate": 6.62603305785124e-07, + "logits/chosen": -0.25, + "logits/rejected": -0.349609375, + "logps/chosen": -358.0, + "logps/rejected": -402.0, + "loss": 0.0241, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.51953125, + "rewards/margins": 11.875, + "rewards/rejected": -11.3125, + "step": 1376 + }, + { + "epoch": 1.6895705521472393, + "grad_norm": 6.071796287867807, + "learning_rate": 6.619834710743801e-07, + "logits/chosen": -0.2138671875, + "logits/rejected": -0.3984375, + "logps/chosen": -360.0, + "logps/rejected": -380.0, + "loss": 0.0204, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.75390625, + "rewards/margins": 12.3125, + "rewards/rejected": -11.5, + "step": 1377 + }, + { + "epoch": 1.69079754601227, + "grad_norm": 6.709699176048537, + "learning_rate": 6.613636363636364e-07, + "logits/chosen": -0.1767578125, + "logits/rejected": -0.35546875, + "logps/chosen": -366.0, + "logps/rejected": -362.0, + "loss": 0.0212, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.259765625, + "rewards/margins": 10.25, + "rewards/rejected": -10.5, + "step": 1378 + }, + { + "epoch": 1.6920245398773006, + "grad_norm": 9.699757128977033, + "learning_rate": 6.607438016528926e-07, + "logits/chosen": -0.1943359375, + "logits/rejected": -0.33984375, + "logps/chosen": -410.0, + "logps/rejected": -378.0, + "loss": 0.0463, + "rewards/accuracies": 0.9765625, + "rewards/chosen": 0.62890625, + "rewards/margins": 10.5, + "rewards/rejected": -9.875, + "step": 1379 + }, + { + "epoch": 1.6932515337423313, + "grad_norm": 13.747966139148284, + "learning_rate": 6.601239669421488e-07, + "logits/chosen": -0.14453125, + "logits/rejected": -0.37109375, + "logps/chosen": -422.0, + "logps/rejected": -354.0, + "loss": 0.0229, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.68359375, + "rewards/margins": 10.8125, + "rewards/rejected": -10.125, + "step": 1380 + }, + { + "epoch": 1.694478527607362, + "grad_norm": 12.861297927766412, + "learning_rate": 6.59504132231405e-07, + "logits/chosen": -0.0712890625, + "logits/rejected": -0.0966796875, + "logps/chosen": -362.0, + "logps/rejected": -364.0, + "loss": 0.0549, + "rewards/accuracies": 0.9765625, + "rewards/chosen": 0.26171875, + "rewards/margins": 9.9375, + "rewards/rejected": -9.6875, + "step": 1381 + }, + { + "epoch": 1.6957055214723926, + "grad_norm": 17.309330036067298, + "learning_rate": 6.588842975206612e-07, + "logits/chosen": -0.1376953125, + "logits/rejected": -0.25, + "logps/chosen": -348.0, + "logps/rejected": -376.0, + "loss": 0.0564, + "rewards/accuracies": 0.9765625, + "rewards/chosen": -0.046630859375, + "rewards/margins": 10.1875, + "rewards/rejected": -10.25, + "step": 1382 + }, + { + "epoch": 1.6969325153374233, + "grad_norm": 9.03911314369842, + "learning_rate": 6.582644628099174e-07, + "logits/chosen": -0.15234375, + "logits/rejected": -0.33203125, + "logps/chosen": -386.0, + "logps/rejected": -392.0, + "loss": 0.0248, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.95703125, + "rewards/margins": 11.375, + "rewards/rejected": -10.375, + "step": 1383 + }, + { + "epoch": 1.698159509202454, + "grad_norm": 16.803262144198918, + "learning_rate": 6.576446280991736e-07, + "logits/chosen": -0.01507568359375, + "logits/rejected": -0.130859375, + "logps/chosen": -370.0, + "logps/rejected": -352.0, + "loss": 0.0558, + "rewards/accuracies": 0.9765625, + "rewards/chosen": -0.04931640625, + "rewards/margins": 9.875, + "rewards/rejected": -9.875, + "step": 1384 + }, + { + "epoch": 1.6993865030674846, + "grad_norm": 9.539239180399765, + "learning_rate": 6.570247933884298e-07, + "logits/chosen": -0.146484375, + "logits/rejected": -0.265625, + "logps/chosen": -328.0, + "logps/rejected": -342.0, + "loss": 0.0495, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.462890625, + "rewards/margins": 10.375, + "rewards/rejected": -9.875, + "step": 1385 + }, + { + "epoch": 1.7006134969325153, + "grad_norm": 8.811291557763298, + "learning_rate": 6.564049586776859e-07, + "logits/chosen": -0.294921875, + "logits/rejected": -0.50390625, + "logps/chosen": -416.0, + "logps/rejected": -410.0, + "loss": 0.0357, + "rewards/accuracies": 0.984375, + "rewards/chosen": 1.1484375, + "rewards/margins": 12.0625, + "rewards/rejected": -10.9375, + "step": 1386 + }, + { + "epoch": 1.701840490797546, + "grad_norm": 9.305768357984167, + "learning_rate": 6.557851239669422e-07, + "logits/chosen": -0.2294921875, + "logits/rejected": -0.404296875, + "logps/chosen": -428.0, + "logps/rejected": -440.0, + "loss": 0.029, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.68359375, + "rewards/margins": 11.375, + "rewards/rejected": -10.6875, + "step": 1387 + }, + { + "epoch": 1.7030674846625766, + "grad_norm": 7.800969393038809, + "learning_rate": 6.551652892561984e-07, + "logits/chosen": -0.01025390625, + "logits/rejected": -0.1474609375, + "logps/chosen": -372.0, + "logps/rejected": -350.0, + "loss": 0.0219, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.27734375, + "rewards/margins": 9.875, + "rewards/rejected": -9.625, + "step": 1388 + }, + { + "epoch": 1.7042944785276073, + "grad_norm": 7.529656036678668, + "learning_rate": 6.545454545454546e-07, + "logits/chosen": -0.29296875, + "logits/rejected": -0.3984375, + "logps/chosen": -408.0, + "logps/rejected": -412.0, + "loss": 0.0288, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.83203125, + "rewards/margins": 11.75, + "rewards/rejected": -10.9375, + "step": 1389 + }, + { + "epoch": 1.705521472392638, + "grad_norm": 10.948563401414935, + "learning_rate": 6.539256198347107e-07, + "logits/chosen": -0.158203125, + "logits/rejected": -0.29296875, + "logps/chosen": -374.0, + "logps/rejected": -386.0, + "loss": 0.032, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.345703125, + "rewards/margins": 11.375, + "rewards/rejected": -11.0, + "step": 1390 + }, + { + "epoch": 1.7067484662576686, + "grad_norm": 8.365223158634247, + "learning_rate": 6.53305785123967e-07, + "logits/chosen": -0.158203125, + "logits/rejected": -0.40625, + "logps/chosen": -392.0, + "logps/rejected": -386.0, + "loss": 0.031, + "rewards/accuracies": 0.9765625, + "rewards/chosen": 1.125, + "rewards/margins": 11.125, + "rewards/rejected": -10.0, + "step": 1391 + }, + { + "epoch": 1.7079754601226993, + "grad_norm": 7.910192843345908, + "learning_rate": 6.526859504132232e-07, + "logits/chosen": 0.01153564453125, + "logits/rejected": -0.162109375, + "logps/chosen": -360.0, + "logps/rejected": -346.0, + "loss": 0.037, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.1953125, + "rewards/margins": 9.5, + "rewards/rejected": -9.6875, + "step": 1392 + }, + { + "epoch": 1.70920245398773, + "grad_norm": 10.155639345314109, + "learning_rate": 6.520661157024793e-07, + "logits/chosen": -0.19921875, + "logits/rejected": -0.388671875, + "logps/chosen": -372.0, + "logps/rejected": -344.0, + "loss": 0.0319, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.7734375, + "rewards/margins": 10.5, + "rewards/rejected": -9.75, + "step": 1393 + }, + { + "epoch": 1.7104294478527606, + "grad_norm": 11.604341441937358, + "learning_rate": 6.514462809917355e-07, + "logits/chosen": -0.248046875, + "logits/rejected": -0.41015625, + "logps/chosen": -406.0, + "logps/rejected": -382.0, + "loss": 0.0466, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.2890625, + "rewards/margins": 11.125, + "rewards/rejected": -10.8125, + "step": 1394 + }, + { + "epoch": 1.7116564417177913, + "grad_norm": 6.312301480394843, + "learning_rate": 6.508264462809917e-07, + "logits/chosen": -0.15234375, + "logits/rejected": -0.349609375, + "logps/chosen": -408.0, + "logps/rejected": -382.0, + "loss": 0.0145, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.66796875, + "rewards/margins": 11.0, + "rewards/rejected": -10.3125, + "step": 1395 + }, + { + "epoch": 1.712883435582822, + "grad_norm": 2.1889604215084524, + "learning_rate": 6.50206611570248e-07, + "logits/chosen": 0.01226806640625, + "logits/rejected": -0.2021484375, + "logps/chosen": -382.0, + "logps/rejected": -372.0, + "loss": 0.0102, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.60546875, + "rewards/margins": 12.0, + "rewards/rejected": -11.4375, + "step": 1396 + }, + { + "epoch": 1.7141104294478526, + "grad_norm": 8.251001626937791, + "learning_rate": 6.495867768595042e-07, + "logits/chosen": -0.1611328125, + "logits/rejected": -0.306640625, + "logps/chosen": -388.0, + "logps/rejected": -388.0, + "loss": 0.0266, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 1.09375, + "rewards/margins": 12.5625, + "rewards/rejected": -11.4375, + "step": 1397 + }, + { + "epoch": 1.7153374233128833, + "grad_norm": 12.58694243429673, + "learning_rate": 6.489669421487604e-07, + "logits/chosen": -0.1044921875, + "logits/rejected": -0.21484375, + "logps/chosen": -432.0, + "logps/rejected": -426.0, + "loss": 0.0351, + "rewards/accuracies": 0.9765625, + "rewards/chosen": 0.365234375, + "rewards/margins": 11.625, + "rewards/rejected": -11.25, + "step": 1398 + }, + { + "epoch": 1.716564417177914, + "grad_norm": 7.217759877360516, + "learning_rate": 6.483471074380165e-07, + "logits/chosen": -0.07275390625, + "logits/rejected": -0.181640625, + "logps/chosen": -362.0, + "logps/rejected": -402.0, + "loss": 0.0306, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.018798828125, + "rewards/margins": 10.5, + "rewards/rejected": -10.5625, + "step": 1399 + }, + { + "epoch": 1.7177914110429446, + "grad_norm": 21.648162026893676, + "learning_rate": 6.477272727272728e-07, + "logits/chosen": -0.072265625, + "logits/rejected": -0.271484375, + "logps/chosen": -344.0, + "logps/rejected": -352.0, + "loss": 0.0312, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.5, + "rewards/margins": 10.3125, + "rewards/rejected": -10.8125, + "step": 1400 + }, + { + "epoch": 1.7190184049079753, + "grad_norm": 6.36710851995934, + "learning_rate": 6.47107438016529e-07, + "logits/chosen": -0.134765625, + "logits/rejected": -0.36328125, + "logps/chosen": -446.0, + "logps/rejected": -442.0, + "loss": 0.0253, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.6796875, + "rewards/margins": 12.5, + "rewards/rejected": -11.8125, + "step": 1401 + }, + { + "epoch": 1.720245398773006, + "grad_norm": 6.937087144639035, + "learning_rate": 6.464876033057851e-07, + "logits/chosen": -0.1611328125, + "logits/rejected": -0.37890625, + "logps/chosen": -406.0, + "logps/rejected": -372.0, + "loss": 0.0245, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.4453125, + "rewards/margins": 11.875, + "rewards/rejected": -11.4375, + "step": 1402 + }, + { + "epoch": 1.7214723926380369, + "grad_norm": 10.928603486238107, + "learning_rate": 6.458677685950413e-07, + "logits/chosen": -0.0498046875, + "logits/rejected": -0.19921875, + "logps/chosen": -406.0, + "logps/rejected": -386.0, + "loss": 0.036, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.10888671875, + "rewards/margins": 10.9375, + "rewards/rejected": -10.8125, + "step": 1403 + }, + { + "epoch": 1.7226993865030675, + "grad_norm": 15.583870620895178, + "learning_rate": 6.452479338842975e-07, + "logits/chosen": -0.07568359375, + "logits/rejected": -0.228515625, + "logps/chosen": -336.0, + "logps/rejected": -330.0, + "loss": 0.0482, + "rewards/accuracies": 0.9765625, + "rewards/chosen": 0.7734375, + "rewards/margins": 11.75, + "rewards/rejected": -11.0, + "step": 1404 + }, + { + "epoch": 1.7239263803680982, + "grad_norm": 11.143077881808916, + "learning_rate": 6.446280991735538e-07, + "logits/chosen": -0.173828125, + "logits/rejected": -0.34765625, + "logps/chosen": -428.0, + "logps/rejected": -412.0, + "loss": 0.0398, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.240234375, + "rewards/margins": 11.75, + "rewards/rejected": -11.5, + "step": 1405 + }, + { + "epoch": 1.7251533742331289, + "grad_norm": 4.224299099232638, + "learning_rate": 6.440082644628099e-07, + "logits/chosen": -0.302734375, + "logits/rejected": -0.421875, + "logps/chosen": -328.0, + "logps/rejected": -374.0, + "loss": 0.0104, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.39453125, + "rewards/margins": 11.75, + "rewards/rejected": -12.125, + "step": 1406 + }, + { + "epoch": 1.7263803680981595, + "grad_norm": 13.330656715870173, + "learning_rate": 6.433884297520661e-07, + "logits/chosen": -0.111328125, + "logits/rejected": -0.390625, + "logps/chosen": -442.0, + "logps/rejected": -426.0, + "loss": 0.0462, + "rewards/accuracies": 0.9765625, + "rewards/chosen": 0.3671875, + "rewards/margins": 11.75, + "rewards/rejected": -11.375, + "step": 1407 + }, + { + "epoch": 1.7276073619631902, + "grad_norm": 7.332633738244965, + "learning_rate": 6.427685950413223e-07, + "logits/chosen": -0.1328125, + "logits/rejected": -0.30078125, + "logps/chosen": -470.0, + "logps/rejected": -440.0, + "loss": 0.0309, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.765625, + "rewards/margins": 13.0, + "rewards/rejected": -12.1875, + "step": 1408 + }, + { + "epoch": 1.7288343558282209, + "grad_norm": 12.35088317930436, + "learning_rate": 6.421487603305786e-07, + "logits/chosen": -0.0927734375, + "logits/rejected": -0.255859375, + "logps/chosen": -380.0, + "logps/rejected": -376.0, + "loss": 0.0446, + "rewards/accuracies": 0.9765625, + "rewards/chosen": 0.0240478515625, + "rewards/margins": 11.75, + "rewards/rejected": -11.75, + "step": 1409 + }, + { + "epoch": 1.7300613496932515, + "grad_norm": 12.863248116987783, + "learning_rate": 6.415289256198348e-07, + "logits/chosen": -0.1787109375, + "logits/rejected": -0.353515625, + "logps/chosen": -400.0, + "logps/rejected": -388.0, + "loss": 0.0733, + "rewards/accuracies": 0.9609375, + "rewards/chosen": -0.5859375, + "rewards/margins": 10.9375, + "rewards/rejected": -11.5, + "step": 1410 + }, + { + "epoch": 1.7312883435582822, + "grad_norm": 5.395301448670732, + "learning_rate": 6.409090909090908e-07, + "logits/chosen": -0.130859375, + "logits/rejected": -0.279296875, + "logps/chosen": -360.0, + "logps/rejected": -382.0, + "loss": 0.0197, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.451171875, + "rewards/margins": 11.4375, + "rewards/rejected": -11.875, + "step": 1411 + }, + { + "epoch": 1.7325153374233129, + "grad_norm": 11.512023543876147, + "learning_rate": 6.402892561983471e-07, + "logits/chosen": -0.0908203125, + "logits/rejected": -0.330078125, + "logps/chosen": -430.0, + "logps/rejected": -398.0, + "loss": 0.063, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.1904296875, + "rewards/margins": 11.8125, + "rewards/rejected": -11.625, + "step": 1412 + }, + { + "epoch": 1.7337423312883435, + "grad_norm": 15.349502769641875, + "learning_rate": 6.396694214876033e-07, + "logits/chosen": -0.2333984375, + "logits/rejected": -0.4453125, + "logps/chosen": -400.0, + "logps/rejected": -392.0, + "loss": 0.0712, + "rewards/accuracies": 0.96875, + "rewards/chosen": -0.734375, + "rewards/margins": 11.875, + "rewards/rejected": -12.625, + "step": 1413 + }, + { + "epoch": 1.7349693251533742, + "grad_norm": 10.27908988260348, + "learning_rate": 6.390495867768596e-07, + "logits/chosen": -0.10595703125, + "logits/rejected": -0.287109375, + "logps/chosen": -436.0, + "logps/rejected": -390.0, + "loss": 0.0453, + "rewards/accuracies": 0.9765625, + "rewards/chosen": 0.158203125, + "rewards/margins": 11.4375, + "rewards/rejected": -11.3125, + "step": 1414 + }, + { + "epoch": 1.7361963190184049, + "grad_norm": 9.292914335874663, + "learning_rate": 6.384297520661157e-07, + "logits/chosen": -0.169921875, + "logits/rejected": -0.349609375, + "logps/chosen": -350.0, + "logps/rejected": -380.0, + "loss": 0.041, + "rewards/accuracies": 0.9765625, + "rewards/chosen": -0.15625, + "rewards/margins": 11.5, + "rewards/rejected": -11.6875, + "step": 1415 + }, + { + "epoch": 1.7374233128834355, + "grad_norm": 7.447145147683031, + "learning_rate": 6.378099173553719e-07, + "logits/chosen": -0.09521484375, + "logits/rejected": -0.361328125, + "logps/chosen": -420.0, + "logps/rejected": -406.0, + "loss": 0.027, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.609375, + "rewards/margins": 12.625, + "rewards/rejected": -12.0625, + "step": 1416 + }, + { + "epoch": 1.7386503067484662, + "grad_norm": 8.792777557704559, + "learning_rate": 6.371900826446281e-07, + "logits/chosen": -0.2353515625, + "logits/rejected": -0.37109375, + "logps/chosen": -374.0, + "logps/rejected": -386.0, + "loss": 0.0291, + "rewards/accuracies": 0.984375, + "rewards/chosen": -0.162109375, + "rewards/margins": 11.875, + "rewards/rejected": -12.0, + "step": 1417 + }, + { + "epoch": 1.7398773006134969, + "grad_norm": 13.38073206624744, + "learning_rate": 6.365702479338844e-07, + "logits/chosen": -0.12060546875, + "logits/rejected": -0.3671875, + "logps/chosen": -396.0, + "logps/rejected": -408.0, + "loss": 0.0355, + "rewards/accuracies": 0.984375, + "rewards/chosen": -0.02587890625, + "rewards/margins": 12.4375, + "rewards/rejected": -12.4375, + "step": 1418 + }, + { + "epoch": 1.7411042944785278, + "grad_norm": 10.607683414699668, + "learning_rate": 6.359504132231405e-07, + "logits/chosen": -0.0908203125, + "logits/rejected": -0.21875, + "logps/chosen": -346.0, + "logps/rejected": -342.0, + "loss": 0.0453, + "rewards/accuracies": 0.9765625, + "rewards/chosen": -0.63671875, + "rewards/margins": 10.5625, + "rewards/rejected": -11.1875, + "step": 1419 + }, + { + "epoch": 1.7423312883435584, + "grad_norm": 10.995863145556806, + "learning_rate": 6.353305785123966e-07, + "logits/chosen": -0.3203125, + "logits/rejected": -0.466796875, + "logps/chosen": -430.0, + "logps/rejected": -422.0, + "loss": 0.0369, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.00186920166015625, + "rewards/margins": 12.375, + "rewards/rejected": -12.375, + "step": 1420 + }, + { + "epoch": 1.743558282208589, + "grad_norm": 4.993783324754888, + "learning_rate": 6.347107438016529e-07, + "logits/chosen": -0.2119140625, + "logits/rejected": -0.412109375, + "logps/chosen": -378.0, + "logps/rejected": -392.0, + "loss": 0.0189, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.1796875, + "rewards/margins": 12.5625, + "rewards/rejected": -12.375, + "step": 1421 + }, + { + "epoch": 1.7447852760736198, + "grad_norm": 9.402562866512993, + "learning_rate": 6.340909090909091e-07, + "logits/chosen": -0.08935546875, + "logits/rejected": -0.251953125, + "logps/chosen": -398.0, + "logps/rejected": -392.0, + "loss": 0.034, + "rewards/accuracies": 0.984375, + "rewards/chosen": -0.1494140625, + "rewards/margins": 11.125, + "rewards/rejected": -11.3125, + "step": 1422 + }, + { + "epoch": 1.7460122699386504, + "grad_norm": 7.9284649955491595, + "learning_rate": 6.334710743801654e-07, + "logits/chosen": -0.16015625, + "logits/rejected": -0.33984375, + "logps/chosen": -382.0, + "logps/rejected": -422.0, + "loss": 0.0293, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.91796875, + "rewards/margins": 12.3125, + "rewards/rejected": -13.25, + "step": 1423 + }, + { + "epoch": 1.747239263803681, + "grad_norm": 6.1006118198416805, + "learning_rate": 6.328512396694215e-07, + "logits/chosen": -0.255859375, + "logits/rejected": -0.49609375, + "logps/chosen": -424.0, + "logps/rejected": -410.0, + "loss": 0.0223, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.23046875, + "rewards/margins": 12.125, + "rewards/rejected": -11.875, + "step": 1424 + }, + { + "epoch": 1.7484662576687118, + "grad_norm": 12.036819296468309, + "learning_rate": 6.322314049586777e-07, + "logits/chosen": -0.12255859375, + "logits/rejected": -0.2373046875, + "logps/chosen": -386.0, + "logps/rejected": -392.0, + "loss": 0.0329, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.1484375, + "rewards/margins": 11.875, + "rewards/rejected": -11.75, + "step": 1425 + }, + { + "epoch": 1.7496932515337424, + "grad_norm": 18.4014829822398, + "learning_rate": 6.316115702479339e-07, + "logits/chosen": -0.1259765625, + "logits/rejected": -0.30078125, + "logps/chosen": -374.0, + "logps/rejected": -388.0, + "loss": 0.0692, + "rewards/accuracies": 0.9765625, + "rewards/chosen": 0.271484375, + "rewards/margins": 11.4375, + "rewards/rejected": -11.1875, + "step": 1426 + }, + { + "epoch": 1.7509202453987731, + "grad_norm": 12.513859109903631, + "learning_rate": 6.309917355371902e-07, + "logits/chosen": -0.1298828125, + "logits/rejected": -0.265625, + "logps/chosen": -380.0, + "logps/rejected": -386.0, + "loss": 0.0557, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.427734375, + "rewards/margins": 11.3125, + "rewards/rejected": -10.875, + "step": 1427 + }, + { + "epoch": 1.7521472392638038, + "grad_norm": 11.939507817491306, + "learning_rate": 6.303719008264463e-07, + "logits/chosen": -0.1845703125, + "logits/rejected": -0.298828125, + "logps/chosen": -420.0, + "logps/rejected": -410.0, + "loss": 0.0502, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.8671875, + "rewards/margins": 12.375, + "rewards/rejected": -11.5, + "step": 1428 + }, + { + "epoch": 1.7533742331288344, + "grad_norm": 6.914751025127418, + "learning_rate": 6.297520661157024e-07, + "logits/chosen": -0.134765625, + "logits/rejected": -0.296875, + "logps/chosen": -354.0, + "logps/rejected": -360.0, + "loss": 0.0278, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.37109375, + "rewards/margins": 11.5, + "rewards/rejected": -11.125, + "step": 1429 + }, + { + "epoch": 1.7546012269938651, + "grad_norm": 5.616187235790868, + "learning_rate": 6.291322314049587e-07, + "logits/chosen": -0.212890625, + "logits/rejected": -0.298828125, + "logps/chosen": -388.0, + "logps/rejected": -416.0, + "loss": 0.0209, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.22265625, + "rewards/margins": 11.625, + "rewards/rejected": -11.4375, + "step": 1430 + }, + { + "epoch": 1.7558282208588958, + "grad_norm": 9.370787243161711, + "learning_rate": 6.285123966942149e-07, + "logits/chosen": -0.0196533203125, + "logits/rejected": -0.181640625, + "logps/chosen": -396.0, + "logps/rejected": -380.0, + "loss": 0.0259, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.0771484375, + "rewards/margins": 10.4375, + "rewards/rejected": -10.3125, + "step": 1431 + }, + { + "epoch": 1.7570552147239265, + "grad_norm": 4.016040968990093, + "learning_rate": 6.278925619834711e-07, + "logits/chosen": -0.181640625, + "logits/rejected": -0.22265625, + "logps/chosen": -336.0, + "logps/rejected": -380.0, + "loss": 0.012, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.41796875, + "rewards/margins": 11.875, + "rewards/rejected": -11.5, + "step": 1432 + }, + { + "epoch": 1.7582822085889571, + "grad_norm": 12.729230743822482, + "learning_rate": 6.272727272727272e-07, + "logits/chosen": 0.0184326171875, + "logits/rejected": -0.0966796875, + "logps/chosen": -360.0, + "logps/rejected": -354.0, + "loss": 0.0589, + "rewards/accuracies": 0.984375, + "rewards/chosen": -0.134765625, + "rewards/margins": 10.0, + "rewards/rejected": -10.1875, + "step": 1433 + }, + { + "epoch": 1.7595092024539878, + "grad_norm": 18.378377289476965, + "learning_rate": 6.266528925619835e-07, + "logits/chosen": -0.0145263671875, + "logits/rejected": -0.1103515625, + "logps/chosen": -344.0, + "logps/rejected": -410.0, + "loss": 0.0787, + "rewards/accuracies": 0.96875, + "rewards/chosen": 0.2216796875, + "rewards/margins": 10.3125, + "rewards/rejected": -10.125, + "step": 1434 + }, + { + "epoch": 1.7607361963190185, + "grad_norm": 6.317649451474902, + "learning_rate": 6.260330578512397e-07, + "logits/chosen": -0.05810546875, + "logits/rejected": -0.169921875, + "logps/chosen": -352.0, + "logps/rejected": -368.0, + "loss": 0.0262, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.31640625, + "rewards/margins": 10.625, + "rewards/rejected": -10.3125, + "step": 1435 + }, + { + "epoch": 1.7619631901840491, + "grad_norm": 4.69256168204721, + "learning_rate": 6.254132231404959e-07, + "logits/chosen": -0.09619140625, + "logits/rejected": -0.306640625, + "logps/chosen": -388.0, + "logps/rejected": -404.0, + "loss": 0.0142, + "rewards/accuracies": 1.0, + "rewards/chosen": 1.46875, + "rewards/margins": 12.4375, + "rewards/rejected": -10.9375, + "step": 1436 + }, + { + "epoch": 1.7631901840490798, + "grad_norm": 29.257144870088368, + "learning_rate": 6.247933884297521e-07, + "logits/chosen": -0.248046875, + "logits/rejected": -0.453125, + "logps/chosen": -396.0, + "logps/rejected": -416.0, + "loss": 0.0505, + "rewards/accuracies": 0.96875, + "rewards/chosen": 0.53125, + "rewards/margins": 11.9375, + "rewards/rejected": -11.375, + "step": 1437 + }, + { + "epoch": 1.7644171779141105, + "grad_norm": 3.93681972201715, + "learning_rate": 6.241735537190082e-07, + "logits/chosen": -0.12109375, + "logits/rejected": -0.27734375, + "logps/chosen": -382.0, + "logps/rejected": -380.0, + "loss": 0.0156, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.2138671875, + "rewards/margins": 10.625, + "rewards/rejected": -10.375, + "step": 1438 + }, + { + "epoch": 1.7656441717791411, + "grad_norm": 7.385811382795677, + "learning_rate": 6.235537190082645e-07, + "logits/chosen": -0.1728515625, + "logits/rejected": -0.41015625, + "logps/chosen": -368.0, + "logps/rejected": -372.0, + "loss": 0.0188, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.13671875, + "rewards/margins": 10.5625, + "rewards/rejected": -10.6875, + "step": 1439 + }, + { + "epoch": 1.7668711656441718, + "grad_norm": 5.11550436529797, + "learning_rate": 6.229338842975207e-07, + "logits/chosen": -0.138671875, + "logits/rejected": -0.34375, + "logps/chosen": -470.0, + "logps/rejected": -432.0, + "loss": 0.0149, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.81640625, + "rewards/margins": 12.5, + "rewards/rejected": -11.6875, + "step": 1440 + }, + { + "epoch": 1.7680981595092025, + "grad_norm": 8.599376446703577, + "learning_rate": 6.223140495867769e-07, + "logits/chosen": -0.28515625, + "logits/rejected": -0.359375, + "logps/chosen": -360.0, + "logps/rejected": -398.0, + "loss": 0.0371, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.2001953125, + "rewards/margins": 11.8125, + "rewards/rejected": -11.625, + "step": 1441 + }, + { + "epoch": 1.7693251533742331, + "grad_norm": 20.234174712489157, + "learning_rate": 6.21694214876033e-07, + "logits/chosen": -0.21875, + "logits/rejected": -0.43359375, + "logps/chosen": -434.0, + "logps/rejected": -404.0, + "loss": 0.0897, + "rewards/accuracies": 0.9609375, + "rewards/chosen": 0.057373046875, + "rewards/margins": 11.4375, + "rewards/rejected": -11.375, + "step": 1442 + }, + { + "epoch": 1.7705521472392638, + "grad_norm": 8.089818483065772, + "learning_rate": 6.210743801652893e-07, + "logits/chosen": -0.06640625, + "logits/rejected": -0.21484375, + "logps/chosen": -422.0, + "logps/rejected": -416.0, + "loss": 0.0331, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.46875, + "rewards/margins": 12.0625, + "rewards/rejected": -11.5625, + "step": 1443 + }, + { + "epoch": 1.7717791411042945, + "grad_norm": 7.490857640483026, + "learning_rate": 6.204545454545455e-07, + "logits/chosen": -0.1435546875, + "logits/rejected": -0.3046875, + "logps/chosen": -398.0, + "logps/rejected": -398.0, + "loss": 0.023, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.69921875, + "rewards/margins": 11.9375, + "rewards/rejected": -11.25, + "step": 1444 + }, + { + "epoch": 1.7730061349693251, + "grad_norm": 4.507209000487265, + "learning_rate": 6.198347107438017e-07, + "logits/chosen": -0.09375, + "logits/rejected": -0.267578125, + "logps/chosen": -330.0, + "logps/rejected": -350.0, + "loss": 0.0205, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.20703125, + "rewards/margins": 10.75, + "rewards/rejected": -11.0, + "step": 1445 + }, + { + "epoch": 1.7742331288343558, + "grad_norm": 9.506356568202962, + "learning_rate": 6.192148760330579e-07, + "logits/chosen": -0.0830078125, + "logits/rejected": -0.2412109375, + "logps/chosen": -418.0, + "logps/rejected": -416.0, + "loss": 0.0362, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.055908203125, + "rewards/margins": 12.25, + "rewards/rejected": -12.1875, + "step": 1446 + }, + { + "epoch": 1.7754601226993865, + "grad_norm": 10.56612980003067, + "learning_rate": 6.18595041322314e-07, + "logits/chosen": -0.24609375, + "logits/rejected": -0.40625, + "logps/chosen": -386.0, + "logps/rejected": -392.0, + "loss": 0.0424, + "rewards/accuracies": 0.96875, + "rewards/chosen": -0.2314453125, + "rewards/margins": 11.8125, + "rewards/rejected": -12.0625, + "step": 1447 + }, + { + "epoch": 1.7766871165644171, + "grad_norm": 17.112227162070322, + "learning_rate": 6.179752066115703e-07, + "logits/chosen": -0.0888671875, + "logits/rejected": -0.208984375, + "logps/chosen": -346.0, + "logps/rejected": -376.0, + "loss": 0.0438, + "rewards/accuracies": 0.984375, + "rewards/chosen": -0.75390625, + "rewards/margins": 10.8125, + "rewards/rejected": -11.5625, + "step": 1448 + }, + { + "epoch": 1.7779141104294478, + "grad_norm": 7.673926079275721, + "learning_rate": 6.173553719008264e-07, + "logits/chosen": -0.003936767578125, + "logits/rejected": -0.201171875, + "logps/chosen": -430.0, + "logps/rejected": -420.0, + "loss": 0.0294, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.091796875, + "rewards/margins": 11.25, + "rewards/rejected": -11.3125, + "step": 1449 + }, + { + "epoch": 1.7791411042944785, + "grad_norm": 10.143124845251792, + "learning_rate": 6.167355371900827e-07, + "logits/chosen": -0.166015625, + "logits/rejected": -0.333984375, + "logps/chosen": -428.0, + "logps/rejected": -398.0, + "loss": 0.0361, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.099609375, + "rewards/margins": 11.8125, + "rewards/rejected": -11.6875, + "step": 1450 + }, + { + "epoch": 1.7803680981595091, + "grad_norm": 6.237913509142783, + "learning_rate": 6.161157024793388e-07, + "logits/chosen": -0.0751953125, + "logits/rejected": -0.23046875, + "logps/chosen": -422.0, + "logps/rejected": -410.0, + "loss": 0.0221, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.3515625, + "rewards/margins": 12.1875, + "rewards/rejected": -11.8125, + "step": 1451 + }, + { + "epoch": 1.7815950920245398, + "grad_norm": 13.734009163796884, + "learning_rate": 6.154958677685951e-07, + "logits/chosen": -0.29296875, + "logits/rejected": -0.55078125, + "logps/chosen": -458.0, + "logps/rejected": -422.0, + "loss": 0.0286, + "rewards/accuracies": 0.984375, + "rewards/chosen": 1.2265625, + "rewards/margins": 13.5625, + "rewards/rejected": -12.375, + "step": 1452 + }, + { + "epoch": 1.7828220858895705, + "grad_norm": 5.142550278798512, + "learning_rate": 6.148760330578512e-07, + "logits/chosen": -0.166015625, + "logits/rejected": -0.3671875, + "logps/chosen": -412.0, + "logps/rejected": -388.0, + "loss": 0.0161, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.265625, + "rewards/margins": 12.375, + "rewards/rejected": -12.125, + "step": 1453 + }, + { + "epoch": 1.7840490797546011, + "grad_norm": 11.252004249668094, + "learning_rate": 6.142561983471075e-07, + "logits/chosen": -0.053955078125, + "logits/rejected": -0.25390625, + "logps/chosen": -402.0, + "logps/rejected": -396.0, + "loss": 0.0341, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.04833984375, + "rewards/margins": 12.0, + "rewards/rejected": -12.0, + "step": 1454 + }, + { + "epoch": 1.7852760736196318, + "grad_norm": 5.828112394054352, + "learning_rate": 6.136363636363637e-07, + "logits/chosen": -0.1416015625, + "logits/rejected": -0.267578125, + "logps/chosen": -398.0, + "logps/rejected": -418.0, + "loss": 0.024, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.1318359375, + "rewards/margins": 11.5625, + "rewards/rejected": -11.4375, + "step": 1455 + }, + { + "epoch": 1.7865030674846625, + "grad_norm": 8.622121139819185, + "learning_rate": 6.130165289256198e-07, + "logits/chosen": -0.177734375, + "logits/rejected": -0.29296875, + "logps/chosen": -370.0, + "logps/rejected": -366.0, + "loss": 0.0437, + "rewards/accuracies": 0.984375, + "rewards/chosen": -0.232421875, + "rewards/margins": 10.3125, + "rewards/rejected": -10.5625, + "step": 1456 + }, + { + "epoch": 1.7877300613496931, + "grad_norm": 10.346638960597081, + "learning_rate": 6.123966942148761e-07, + "logits/chosen": -0.19140625, + "logits/rejected": -0.41796875, + "logps/chosen": -424.0, + "logps/rejected": -418.0, + "loss": 0.0402, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.09521484375, + "rewards/margins": 11.8125, + "rewards/rejected": -11.75, + "step": 1457 + }, + { + "epoch": 1.7889570552147238, + "grad_norm": 9.510862849605985, + "learning_rate": 6.117768595041322e-07, + "logits/chosen": -0.04736328125, + "logits/rejected": -0.130859375, + "logps/chosen": -360.0, + "logps/rejected": -396.0, + "loss": 0.0352, + "rewards/accuracies": 0.984375, + "rewards/chosen": -0.19921875, + "rewards/margins": 10.3125, + "rewards/rejected": -10.5, + "step": 1458 + }, + { + "epoch": 1.7901840490797545, + "grad_norm": 2985.9302248963727, + "learning_rate": 6.111570247933885e-07, + "logits/chosen": -0.1572265625, + "logits/rejected": -0.376953125, + "logps/chosen": -404.0, + "logps/rejected": -386.0, + "loss": 1.516, + "rewards/accuracies": 0.984375, + "rewards/chosen": -1.1484375, + "rewards/margins": 10.6875, + "rewards/rejected": -11.8125, + "step": 1459 + }, + { + "epoch": 1.7914110429447851, + "grad_norm": 6.456485664548984, + "learning_rate": 6.105371900826446e-07, + "logits/chosen": -0.12060546875, + "logits/rejected": -0.265625, + "logps/chosen": -422.0, + "logps/rejected": -406.0, + "loss": 0.0201, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.5234375, + "rewards/margins": 11.625, + "rewards/rejected": -11.125, + "step": 1460 + }, + { + "epoch": 1.7926380368098158, + "grad_norm": 12.803301057131963, + "learning_rate": 6.099173553719009e-07, + "logits/chosen": -0.11328125, + "logits/rejected": -0.32421875, + "logps/chosen": -430.0, + "logps/rejected": -416.0, + "loss": 0.0432, + "rewards/accuracies": 0.9765625, + "rewards/chosen": 0.52734375, + "rewards/margins": 13.625, + "rewards/rejected": -13.125, + "step": 1461 + }, + { + "epoch": 1.7938650306748465, + "grad_norm": 8.137637416090907, + "learning_rate": 6.09297520661157e-07, + "logits/chosen": -0.13671875, + "logits/rejected": -0.2021484375, + "logps/chosen": -326.0, + "logps/rejected": -372.0, + "loss": 0.0317, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.201171875, + "rewards/margins": 11.6875, + "rewards/rejected": -11.5, + "step": 1462 + }, + { + "epoch": 1.7950920245398772, + "grad_norm": 9.5142759213274, + "learning_rate": 6.086776859504133e-07, + "logits/chosen": -0.058837890625, + "logits/rejected": -0.29296875, + "logps/chosen": -406.0, + "logps/rejected": -392.0, + "loss": 0.0229, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.33203125, + "rewards/margins": 11.375, + "rewards/rejected": -11.0625, + "step": 1463 + }, + { + "epoch": 1.7963190184049078, + "grad_norm": 16.72592677806428, + "learning_rate": 6.080578512396695e-07, + "logits/chosen": -0.1884765625, + "logits/rejected": -0.32421875, + "logps/chosen": -346.0, + "logps/rejected": -364.0, + "loss": 0.0569, + "rewards/accuracies": 0.9765625, + "rewards/chosen": 0.1328125, + "rewards/margins": 11.625, + "rewards/rejected": -11.5, + "step": 1464 + }, + { + "epoch": 1.7975460122699385, + "grad_norm": 8.34969796778253, + "learning_rate": 6.074380165289256e-07, + "logits/chosen": -0.2294921875, + "logits/rejected": -0.3515625, + "logps/chosen": -396.0, + "logps/rejected": -414.0, + "loss": 0.0236, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.52734375, + "rewards/margins": 12.5625, + "rewards/rejected": -12.0, + "step": 1465 + }, + { + "epoch": 1.7987730061349694, + "grad_norm": 6.129551015730449, + "learning_rate": 6.068181818181818e-07, + "logits/chosen": -0.1455078125, + "logits/rejected": -0.359375, + "logps/chosen": -364.0, + "logps/rejected": -378.0, + "loss": 0.0164, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.10205078125, + "rewards/margins": 11.9375, + "rewards/rejected": -12.0, + "step": 1466 + }, + { + "epoch": 1.8, + "grad_norm": 6.8967202729590245, + "learning_rate": 6.06198347107438e-07, + "logits/chosen": -0.2060546875, + "logits/rejected": -0.3046875, + "logps/chosen": -426.0, + "logps/rejected": -420.0, + "loss": 0.0195, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.76171875, + "rewards/margins": 11.6875, + "rewards/rejected": -10.9375, + "step": 1467 + }, + { + "epoch": 1.8012269938650307, + "grad_norm": 8.361065202748346, + "learning_rate": 6.055785123966943e-07, + "logits/chosen": -0.10107421875, + "logits/rejected": -0.29296875, + "logps/chosen": -412.0, + "logps/rejected": -402.0, + "loss": 0.0182, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 1.03125, + "rewards/margins": 12.6875, + "rewards/rejected": -11.625, + "step": 1468 + }, + { + "epoch": 1.8024539877300614, + "grad_norm": 12.455616695494426, + "learning_rate": 6.049586776859504e-07, + "logits/chosen": -0.08203125, + "logits/rejected": -0.224609375, + "logps/chosen": -402.0, + "logps/rejected": -390.0, + "loss": 0.0396, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.06494140625, + "rewards/margins": 11.75, + "rewards/rejected": -11.6875, + "step": 1469 + }, + { + "epoch": 1.803680981595092, + "grad_norm": 9.292975088281931, + "learning_rate": 6.043388429752066e-07, + "logits/chosen": -0.05126953125, + "logits/rejected": -0.2734375, + "logps/chosen": -404.0, + "logps/rejected": -384.0, + "loss": 0.0445, + "rewards/accuracies": 0.984375, + "rewards/chosen": -0.419921875, + "rewards/margins": 10.9375, + "rewards/rejected": -11.375, + "step": 1470 + }, + { + "epoch": 1.8049079754601227, + "grad_norm": 9.047461001821528, + "learning_rate": 6.037190082644628e-07, + "logits/chosen": -0.07177734375, + "logits/rejected": -0.1748046875, + "logps/chosen": -372.0, + "logps/rejected": -396.0, + "loss": 0.0295, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.478515625, + "rewards/margins": 11.125, + "rewards/rejected": -10.6875, + "step": 1471 + }, + { + "epoch": 1.8061349693251534, + "grad_norm": 13.42466742331635, + "learning_rate": 6.030991735537191e-07, + "logits/chosen": -0.1083984375, + "logits/rejected": -0.2470703125, + "logps/chosen": -392.0, + "logps/rejected": -398.0, + "loss": 0.04, + "rewards/accuracies": 0.9765625, + "rewards/chosen": 0.34765625, + "rewards/margins": 12.25, + "rewards/rejected": -11.875, + "step": 1472 + }, + { + "epoch": 1.807361963190184, + "grad_norm": 6.913536686973756, + "learning_rate": 6.024793388429752e-07, + "logits/chosen": -0.14453125, + "logits/rejected": -0.40234375, + "logps/chosen": -384.0, + "logps/rejected": -390.0, + "loss": 0.0335, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.287109375, + "rewards/margins": 12.5625, + "rewards/rejected": -12.25, + "step": 1473 + }, + { + "epoch": 1.8085889570552147, + "grad_norm": 17.128950617994402, + "learning_rate": 6.018595041322314e-07, + "logits/chosen": -0.017822265625, + "logits/rejected": -0.234375, + "logps/chosen": -426.0, + "logps/rejected": -418.0, + "loss": 0.0425, + "rewards/accuracies": 0.9765625, + "rewards/chosen": 0.318359375, + "rewards/margins": 12.25, + "rewards/rejected": -11.9375, + "step": 1474 + }, + { + "epoch": 1.8098159509202454, + "grad_norm": 13.951944722139784, + "learning_rate": 6.012396694214876e-07, + "logits/chosen": -0.154296875, + "logits/rejected": -0.349609375, + "logps/chosen": -402.0, + "logps/rejected": -398.0, + "loss": 0.0357, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.1357421875, + "rewards/margins": 12.0625, + "rewards/rejected": -11.9375, + "step": 1475 + }, + { + "epoch": 1.811042944785276, + "grad_norm": 12.939142286970506, + "learning_rate": 6.006198347107438e-07, + "logits/chosen": -0.181640625, + "logits/rejected": -0.3359375, + "logps/chosen": -368.0, + "logps/rejected": -392.0, + "loss": 0.0343, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.032470703125, + "rewards/margins": 12.5, + "rewards/rejected": -12.4375, + "step": 1476 + }, + { + "epoch": 1.8122699386503067, + "grad_norm": 15.184402213006727, + "learning_rate": 6.000000000000001e-07, + "logits/chosen": -0.057861328125, + "logits/rejected": -0.185546875, + "logps/chosen": -390.0, + "logps/rejected": -404.0, + "loss": 0.0564, + "rewards/accuracies": 0.9765625, + "rewards/chosen": -0.54296875, + "rewards/margins": 11.125, + "rewards/rejected": -11.6875, + "step": 1477 + }, + { + "epoch": 1.8134969325153374, + "grad_norm": 7.953555230394229, + "learning_rate": 5.993801652892562e-07, + "logits/chosen": -0.09619140625, + "logits/rejected": -0.267578125, + "logps/chosen": -402.0, + "logps/rejected": -424.0, + "loss": 0.027, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.8828125, + "rewards/margins": 12.375, + "rewards/rejected": -11.5, + "step": 1478 + }, + { + "epoch": 1.814723926380368, + "grad_norm": 11.656764281029574, + "learning_rate": 5.987603305785124e-07, + "logits/chosen": -0.1884765625, + "logits/rejected": -0.359375, + "logps/chosen": -384.0, + "logps/rejected": -396.0, + "loss": 0.0344, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.25, + "rewards/margins": 12.1875, + "rewards/rejected": -11.9375, + "step": 1479 + }, + { + "epoch": 1.8159509202453987, + "grad_norm": 10.973049014636235, + "learning_rate": 5.981404958677686e-07, + "logits/chosen": -0.267578125, + "logits/rejected": -0.408203125, + "logps/chosen": -396.0, + "logps/rejected": -424.0, + "loss": 0.0437, + "rewards/accuracies": 0.9765625, + "rewards/chosen": 0.72265625, + "rewards/margins": 11.375, + "rewards/rejected": -10.6875, + "step": 1480 + }, + { + "epoch": 1.8171779141104294, + "grad_norm": 16.33917392069797, + "learning_rate": 5.975206611570249e-07, + "logits/chosen": -0.2197265625, + "logits/rejected": -0.333984375, + "logps/chosen": -434.0, + "logps/rejected": -434.0, + "loss": 0.0484, + "rewards/accuracies": 0.9765625, + "rewards/chosen": -0.1962890625, + "rewards/margins": 11.9375, + "rewards/rejected": -12.125, + "step": 1481 + }, + { + "epoch": 1.8184049079754603, + "grad_norm": 13.067184042335276, + "learning_rate": 5.96900826446281e-07, + "logits/chosen": -0.046142578125, + "logits/rejected": -0.203125, + "logps/chosen": -374.0, + "logps/rejected": -376.0, + "loss": 0.0686, + "rewards/accuracies": 0.9765625, + "rewards/chosen": 0.08984375, + "rewards/margins": 11.375, + "rewards/rejected": -11.25, + "step": 1482 + }, + { + "epoch": 1.819631901840491, + "grad_norm": 14.380040955958671, + "learning_rate": 5.962809917355371e-07, + "logits/chosen": 0.0086669921875, + "logits/rejected": -0.146484375, + "logps/chosen": -356.0, + "logps/rejected": -342.0, + "loss": 0.0636, + "rewards/accuracies": 0.96875, + "rewards/chosen": -0.322265625, + "rewards/margins": 11.0, + "rewards/rejected": -11.3125, + "step": 1483 + }, + { + "epoch": 1.8208588957055216, + "grad_norm": 13.600405309118111, + "learning_rate": 5.956611570247934e-07, + "logits/chosen": -0.2158203125, + "logits/rejected": -0.373046875, + "logps/chosen": -354.0, + "logps/rejected": -374.0, + "loss": 0.0529, + "rewards/accuracies": 0.984375, + "rewards/chosen": -0.5078125, + "rewards/margins": 10.9375, + "rewards/rejected": -11.4375, + "step": 1484 + }, + { + "epoch": 1.8220858895705523, + "grad_norm": 19.36891506988035, + "learning_rate": 5.950413223140496e-07, + "logits/chosen": -0.2021484375, + "logits/rejected": -0.23046875, + "logps/chosen": -350.0, + "logps/rejected": -424.0, + "loss": 0.0361, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.5859375, + "rewards/margins": 11.0625, + "rewards/rejected": -11.6875, + "step": 1485 + }, + { + "epoch": 1.823312883435583, + "grad_norm": 13.246751492795608, + "learning_rate": 5.944214876033059e-07, + "logits/chosen": -0.2890625, + "logits/rejected": -0.447265625, + "logps/chosen": -382.0, + "logps/rejected": -402.0, + "loss": 0.0601, + "rewards/accuracies": 0.9765625, + "rewards/chosen": -0.38671875, + "rewards/margins": 11.75, + "rewards/rejected": -12.125, + "step": 1486 + }, + { + "epoch": 1.8245398773006136, + "grad_norm": 3.3945752544087964, + "learning_rate": 5.93801652892562e-07, + "logits/chosen": -0.34375, + "logits/rejected": -0.5078125, + "logps/chosen": -416.0, + "logps/rejected": -388.0, + "loss": 0.013, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.07958984375, + "rewards/margins": 12.5, + "rewards/rejected": -12.4375, + "step": 1487 + }, + { + "epoch": 1.8257668711656443, + "grad_norm": 8.043522667448203, + "learning_rate": 5.931818181818182e-07, + "logits/chosen": -0.201171875, + "logits/rejected": -0.462890625, + "logps/chosen": -416.0, + "logps/rejected": -378.0, + "loss": 0.0227, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.1240234375, + "rewards/margins": 12.1875, + "rewards/rejected": -12.0625, + "step": 1488 + }, + { + "epoch": 1.826993865030675, + "grad_norm": 10.658513612446942, + "learning_rate": 5.925619834710744e-07, + "logits/chosen": -0.333984375, + "logits/rejected": -0.54296875, + "logps/chosen": -410.0, + "logps/rejected": -408.0, + "loss": 0.0417, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.162109375, + "rewards/margins": 12.75, + "rewards/rejected": -12.5625, + "step": 1489 + }, + { + "epoch": 1.8282208588957056, + "grad_norm": 5.907167171601785, + "learning_rate": 5.919421487603307e-07, + "logits/chosen": -0.193359375, + "logits/rejected": -0.416015625, + "logps/chosen": -398.0, + "logps/rejected": -404.0, + "loss": 0.0174, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.71875, + "rewards/margins": 12.875, + "rewards/rejected": -12.125, + "step": 1490 + }, + { + "epoch": 1.8294478527607363, + "grad_norm": 7.708311095918759, + "learning_rate": 5.913223140495868e-07, + "logits/chosen": -0.2109375, + "logits/rejected": -0.466796875, + "logps/chosen": -484.0, + "logps/rejected": -444.0, + "loss": 0.0265, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.9140625, + "rewards/margins": 13.0625, + "rewards/rejected": -12.125, + "step": 1491 + }, + { + "epoch": 1.830674846625767, + "grad_norm": 11.650261479643984, + "learning_rate": 5.907024793388429e-07, + "logits/chosen": -0.349609375, + "logits/rejected": -0.498046875, + "logps/chosen": -390.0, + "logps/rejected": -396.0, + "loss": 0.0365, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.03369140625, + "rewards/margins": 11.75, + "rewards/rejected": -11.8125, + "step": 1492 + }, + { + "epoch": 1.8319018404907976, + "grad_norm": 6.3612166084027235, + "learning_rate": 5.900826446280992e-07, + "logits/chosen": -0.3125, + "logits/rejected": -0.42578125, + "logps/chosen": -330.0, + "logps/rejected": -366.0, + "loss": 0.0247, + "rewards/accuracies": 0.984375, + "rewards/chosen": -0.97265625, + "rewards/margins": 11.375, + "rewards/rejected": -12.375, + "step": 1493 + }, + { + "epoch": 1.8331288343558283, + "grad_norm": 7.617904477738232, + "learning_rate": 5.894628099173554e-07, + "logits/chosen": -0.251953125, + "logits/rejected": -0.38671875, + "logps/chosen": -442.0, + "logps/rejected": -438.0, + "loss": 0.0286, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.1708984375, + "rewards/margins": 12.125, + "rewards/rejected": -12.3125, + "step": 1494 + }, + { + "epoch": 1.834355828220859, + "grad_norm": 14.805530582972839, + "learning_rate": 5.888429752066116e-07, + "logits/chosen": -0.2138671875, + "logits/rejected": -0.369140625, + "logps/chosen": -366.0, + "logps/rejected": -378.0, + "loss": 0.0427, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.051025390625, + "rewards/margins": 11.0625, + "rewards/rejected": -11.0, + "step": 1495 + }, + { + "epoch": 1.8355828220858896, + "grad_norm": 8.603548186700504, + "learning_rate": 5.882231404958677e-07, + "logits/chosen": -0.1533203125, + "logits/rejected": -0.2890625, + "logps/chosen": -374.0, + "logps/rejected": -394.0, + "loss": 0.0246, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.57421875, + "rewards/margins": 11.875, + "rewards/rejected": -11.3125, + "step": 1496 + }, + { + "epoch": 1.8368098159509203, + "grad_norm": 5.119681631341246, + "learning_rate": 5.87603305785124e-07, + "logits/chosen": -0.14453125, + "logits/rejected": -0.375, + "logps/chosen": -414.0, + "logps/rejected": -380.0, + "loss": 0.0172, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.466796875, + "rewards/margins": 11.5625, + "rewards/rejected": -11.0625, + "step": 1497 + }, + { + "epoch": 1.838036809815951, + "grad_norm": 13.256747203318977, + "learning_rate": 5.869834710743802e-07, + "logits/chosen": -0.31640625, + "logits/rejected": -0.49609375, + "logps/chosen": -424.0, + "logps/rejected": -436.0, + "loss": 0.0546, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.328125, + "rewards/margins": 11.625, + "rewards/rejected": -11.3125, + "step": 1498 + }, + { + "epoch": 1.8392638036809816, + "grad_norm": 12.35912732127943, + "learning_rate": 5.863636363636365e-07, + "logits/chosen": -0.1181640625, + "logits/rejected": -0.298828125, + "logps/chosen": -388.0, + "logps/rejected": -396.0, + "loss": 0.0348, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.58984375, + "rewards/margins": 11.9375, + "rewards/rejected": -11.375, + "step": 1499 + }, + { + "epoch": 1.8404907975460123, + "grad_norm": 7.9066533942342945, + "learning_rate": 5.857438016528925e-07, + "logits/chosen": -0.1474609375, + "logits/rejected": -0.361328125, + "logps/chosen": -370.0, + "logps/rejected": -348.0, + "loss": 0.034, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 1.015625, + "rewards/margins": 11.75, + "rewards/rejected": -10.75, + "step": 1500 + }, + { + "epoch": 1.841717791411043, + "grad_norm": 7.790234578512824, + "learning_rate": 5.851239669421487e-07, + "logits/chosen": -0.14453125, + "logits/rejected": -0.27734375, + "logps/chosen": -374.0, + "logps/rejected": -398.0, + "loss": 0.0245, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.61328125, + "rewards/margins": 11.8125, + "rewards/rejected": -11.25, + "step": 1501 + }, + { + "epoch": 1.8429447852760736, + "grad_norm": 8.476340764027082, + "learning_rate": 5.84504132231405e-07, + "logits/chosen": -0.087890625, + "logits/rejected": -0.1953125, + "logps/chosen": -366.0, + "logps/rejected": -394.0, + "loss": 0.0284, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.6953125, + "rewards/margins": 10.875, + "rewards/rejected": -11.5625, + "step": 1502 + }, + { + "epoch": 1.8441717791411043, + "grad_norm": 8.931095134479317, + "learning_rate": 5.838842975206612e-07, + "logits/chosen": -0.201171875, + "logits/rejected": -0.412109375, + "logps/chosen": -430.0, + "logps/rejected": -456.0, + "loss": 0.0258, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.4296875, + "rewards/margins": 12.125, + "rewards/rejected": -11.6875, + "step": 1503 + }, + { + "epoch": 1.845398773006135, + "grad_norm": 9.61981018831853, + "learning_rate": 5.832644628099174e-07, + "logits/chosen": -0.259765625, + "logits/rejected": -0.408203125, + "logps/chosen": -388.0, + "logps/rejected": -408.0, + "loss": 0.0334, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.58203125, + "rewards/margins": 11.5625, + "rewards/rejected": -10.9375, + "step": 1504 + }, + { + "epoch": 1.8466257668711656, + "grad_norm": 8.898568846872344, + "learning_rate": 5.826446280991735e-07, + "logits/chosen": -0.23828125, + "logits/rejected": -0.404296875, + "logps/chosen": -416.0, + "logps/rejected": -420.0, + "loss": 0.0376, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.2099609375, + "rewards/margins": 11.625, + "rewards/rejected": -11.4375, + "step": 1505 + }, + { + "epoch": 1.8478527607361963, + "grad_norm": 8.234391016597, + "learning_rate": 5.820247933884298e-07, + "logits/chosen": -0.1787109375, + "logits/rejected": -0.353515625, + "logps/chosen": -358.0, + "logps/rejected": -398.0, + "loss": 0.0295, + "rewards/accuracies": 0.984375, + "rewards/chosen": -0.48046875, + "rewards/margins": 11.125, + "rewards/rejected": -11.625, + "step": 1506 + }, + { + "epoch": 1.849079754601227, + "grad_norm": 12.615811463348589, + "learning_rate": 5.81404958677686e-07, + "logits/chosen": -0.15625, + "logits/rejected": -0.365234375, + "logps/chosen": -400.0, + "logps/rejected": -390.0, + "loss": 0.0434, + "rewards/accuracies": 0.9765625, + "rewards/chosen": 0.609375, + "rewards/margins": 11.1875, + "rewards/rejected": -10.625, + "step": 1507 + }, + { + "epoch": 1.8503067484662576, + "grad_norm": 11.413630239199696, + "learning_rate": 5.807851239669422e-07, + "logits/chosen": -0.059326171875, + "logits/rejected": -0.2353515625, + "logps/chosen": -392.0, + "logps/rejected": -404.0, + "loss": 0.0325, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.55859375, + "rewards/margins": 10.875, + "rewards/rejected": -11.4375, + "step": 1508 + }, + { + "epoch": 1.8515337423312883, + "grad_norm": 7.238823334535515, + "learning_rate": 5.801652892561983e-07, + "logits/chosen": -0.1611328125, + "logits/rejected": -0.361328125, + "logps/chosen": -444.0, + "logps/rejected": -422.0, + "loss": 0.0186, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.453125, + "rewards/margins": 11.3125, + "rewards/rejected": -10.875, + "step": 1509 + }, + { + "epoch": 1.852760736196319, + "grad_norm": 4.424946915077485, + "learning_rate": 5.795454545454545e-07, + "logits/chosen": -0.126953125, + "logits/rejected": -0.375, + "logps/chosen": -432.0, + "logps/rejected": -398.0, + "loss": 0.0131, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.53515625, + "rewards/margins": 12.1875, + "rewards/rejected": -11.625, + "step": 1510 + }, + { + "epoch": 1.8539877300613496, + "grad_norm": 7.053221824008459, + "learning_rate": 5.789256198347108e-07, + "logits/chosen": -0.203125, + "logits/rejected": -0.267578125, + "logps/chosen": -420.0, + "logps/rejected": -456.0, + "loss": 0.0169, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.322265625, + "rewards/margins": 12.5, + "rewards/rejected": -12.1875, + "step": 1511 + }, + { + "epoch": 1.8552147239263803, + "grad_norm": 3.5547744931254903, + "learning_rate": 5.783057851239669e-07, + "logits/chosen": -0.1494140625, + "logits/rejected": -0.373046875, + "logps/chosen": -358.0, + "logps/rejected": -364.0, + "loss": 0.0135, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.9921875, + "rewards/margins": 13.0, + "rewards/rejected": -12.0, + "step": 1512 + }, + { + "epoch": 1.856441717791411, + "grad_norm": 15.688605244420566, + "learning_rate": 5.776859504132232e-07, + "logits/chosen": -0.150390625, + "logits/rejected": -0.376953125, + "logps/chosen": -404.0, + "logps/rejected": -378.0, + "loss": 0.0586, + "rewards/accuracies": 0.9765625, + "rewards/chosen": 0.12451171875, + "rewards/margins": 11.75, + "rewards/rejected": -11.625, + "step": 1513 + }, + { + "epoch": 1.8576687116564417, + "grad_norm": 5.4110839912069, + "learning_rate": 5.770661157024793e-07, + "logits/chosen": -0.1953125, + "logits/rejected": -0.37109375, + "logps/chosen": -412.0, + "logps/rejected": -426.0, + "loss": 0.0181, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.369140625, + "rewards/margins": 13.125, + "rewards/rejected": -12.75, + "step": 1514 + }, + { + "epoch": 1.8588957055214723, + "grad_norm": 10.478869181509733, + "learning_rate": 5.764462809917356e-07, + "logits/chosen": -0.2265625, + "logits/rejected": -0.4453125, + "logps/chosen": -408.0, + "logps/rejected": -390.0, + "loss": 0.0311, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.94140625, + "rewards/margins": 13.0625, + "rewards/rejected": -12.125, + "step": 1515 + }, + { + "epoch": 1.860122699386503, + "grad_norm": 9.184505008497789, + "learning_rate": 5.758264462809918e-07, + "logits/chosen": -0.1806640625, + "logits/rejected": -0.328125, + "logps/chosen": -420.0, + "logps/rejected": -416.0, + "loss": 0.0329, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.8046875, + "rewards/margins": 11.125, + "rewards/rejected": -11.9375, + "step": 1516 + }, + { + "epoch": 1.8613496932515337, + "grad_norm": 4.539913274268795, + "learning_rate": 5.75206611570248e-07, + "logits/chosen": -0.224609375, + "logits/rejected": -0.42578125, + "logps/chosen": -418.0, + "logps/rejected": -398.0, + "loss": 0.0172, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.69921875, + "rewards/margins": 13.0, + "rewards/rejected": -12.3125, + "step": 1517 + }, + { + "epoch": 1.8625766871165643, + "grad_norm": 14.818256983584963, + "learning_rate": 5.745867768595041e-07, + "logits/chosen": -0.2451171875, + "logits/rejected": -0.443359375, + "logps/chosen": -384.0, + "logps/rejected": -372.0, + "loss": 0.0612, + "rewards/accuracies": 0.96875, + "rewards/chosen": -0.3125, + "rewards/margins": 12.3125, + "rewards/rejected": -12.625, + "step": 1518 + }, + { + "epoch": 1.863803680981595, + "grad_norm": 21.719699854423048, + "learning_rate": 5.739669421487603e-07, + "logits/chosen": -0.2314453125, + "logits/rejected": -0.349609375, + "logps/chosen": -374.0, + "logps/rejected": -412.0, + "loss": 0.0598, + "rewards/accuracies": 0.9765625, + "rewards/chosen": -1.1953125, + "rewards/margins": 10.875, + "rewards/rejected": -12.125, + "step": 1519 + }, + { + "epoch": 1.8650306748466257, + "grad_norm": 4.643580726862002, + "learning_rate": 5.733471074380166e-07, + "logits/chosen": -0.2236328125, + "logits/rejected": -0.30078125, + "logps/chosen": -396.0, + "logps/rejected": -404.0, + "loss": 0.0153, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.41796875, + "rewards/margins": 12.0625, + "rewards/rejected": -12.5, + "step": 1520 + }, + { + "epoch": 1.8662576687116563, + "grad_norm": 9.963670178751276, + "learning_rate": 5.727272727272727e-07, + "logits/chosen": -0.1376953125, + "logits/rejected": -0.3359375, + "logps/chosen": -374.0, + "logps/rejected": -372.0, + "loss": 0.0385, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.28125, + "rewards/margins": 11.5625, + "rewards/rejected": -11.8125, + "step": 1521 + }, + { + "epoch": 1.867484662576687, + "grad_norm": 13.279186623430544, + "learning_rate": 5.721074380165289e-07, + "logits/chosen": -0.1640625, + "logits/rejected": -0.35546875, + "logps/chosen": -370.0, + "logps/rejected": -374.0, + "loss": 0.0478, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.0673828125, + "rewards/margins": 11.5625, + "rewards/rejected": -11.5, + "step": 1522 + }, + { + "epoch": 1.8687116564417177, + "grad_norm": 6.578971949059929, + "learning_rate": 5.714876033057851e-07, + "logits/chosen": -0.2412109375, + "logits/rejected": -0.373046875, + "logps/chosen": -376.0, + "logps/rejected": -376.0, + "loss": 0.0198, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.1591796875, + "rewards/margins": 12.0625, + "rewards/rejected": -11.875, + "step": 1523 + }, + { + "epoch": 1.8699386503067483, + "grad_norm": 13.581236746412541, + "learning_rate": 5.708677685950414e-07, + "logits/chosen": -0.083984375, + "logits/rejected": -0.236328125, + "logps/chosen": -348.0, + "logps/rejected": -400.0, + "loss": 0.04, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.75, + "rewards/margins": 11.4375, + "rewards/rejected": -12.1875, + "step": 1524 + }, + { + "epoch": 1.871165644171779, + "grad_norm": 8.911113085416726, + "learning_rate": 5.702479338842975e-07, + "logits/chosen": -0.2099609375, + "logits/rejected": -0.36328125, + "logps/chosen": -352.0, + "logps/rejected": -390.0, + "loss": 0.0247, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.470703125, + "rewards/margins": 12.375, + "rewards/rejected": -11.9375, + "step": 1525 + }, + { + "epoch": 1.8723926380368097, + "grad_norm": 12.467591280037412, + "learning_rate": 5.696280991735538e-07, + "logits/chosen": -0.07470703125, + "logits/rejected": -0.26953125, + "logps/chosen": -410.0, + "logps/rejected": -390.0, + "loss": 0.0599, + "rewards/accuracies": 0.9765625, + "rewards/chosen": 0.671875, + "rewards/margins": 11.6875, + "rewards/rejected": -11.0625, + "step": 1526 + }, + { + "epoch": 1.8736196319018403, + "grad_norm": 7.9523490622772, + "learning_rate": 5.690082644628099e-07, + "logits/chosen": -0.11279296875, + "logits/rejected": -0.3203125, + "logps/chosen": -412.0, + "logps/rejected": -366.0, + "loss": 0.0304, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.48046875, + "rewards/margins": 11.75, + "rewards/rejected": -11.25, + "step": 1527 + }, + { + "epoch": 1.874846625766871, + "grad_norm": 16.90843634388907, + "learning_rate": 5.683884297520661e-07, + "logits/chosen": -0.07470703125, + "logits/rejected": -0.29296875, + "logps/chosen": -398.0, + "logps/rejected": -362.0, + "loss": 0.0687, + "rewards/accuracies": 0.9609375, + "rewards/chosen": 0.443359375, + "rewards/margins": 11.9375, + "rewards/rejected": -11.5, + "step": 1528 + }, + { + "epoch": 1.876073619631902, + "grad_norm": 5.8636813967032415, + "learning_rate": 5.677685950413223e-07, + "logits/chosen": -0.2021484375, + "logits/rejected": -0.447265625, + "logps/chosen": -422.0, + "logps/rejected": -402.0, + "loss": 0.0195, + "rewards/accuracies": 1.0, + "rewards/chosen": 1.359375, + "rewards/margins": 13.8125, + "rewards/rejected": -12.4375, + "step": 1529 + }, + { + "epoch": 1.8773006134969326, + "grad_norm": 10.180871320115374, + "learning_rate": 5.671487603305785e-07, + "logits/chosen": -0.189453125, + "logits/rejected": -0.326171875, + "logps/chosen": -388.0, + "logps/rejected": -368.0, + "loss": 0.0404, + "rewards/accuracies": 0.9765625, + "rewards/chosen": -0.7265625, + "rewards/margins": 11.8125, + "rewards/rejected": -12.5625, + "step": 1530 + }, + { + "epoch": 1.8785276073619632, + "grad_norm": 9.962751714095937, + "learning_rate": 5.665289256198347e-07, + "logits/chosen": -0.042236328125, + "logits/rejected": -0.1953125, + "logps/chosen": -378.0, + "logps/rejected": -374.0, + "loss": 0.0435, + "rewards/accuracies": 0.9765625, + "rewards/chosen": -0.3984375, + "rewards/margins": 11.5, + "rewards/rejected": -11.875, + "step": 1531 + }, + { + "epoch": 1.879754601226994, + "grad_norm": 13.344235509044795, + "learning_rate": 5.659090909090909e-07, + "logits/chosen": -0.328125, + "logits/rejected": -0.4921875, + "logps/chosen": -436.0, + "logps/rejected": -450.0, + "loss": 0.0678, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.6953125, + "rewards/margins": 13.625, + "rewards/rejected": -12.9375, + "step": 1532 + }, + { + "epoch": 1.8809815950920246, + "grad_norm": 12.313294846355545, + "learning_rate": 5.652892561983472e-07, + "logits/chosen": -0.12060546875, + "logits/rejected": -0.26171875, + "logps/chosen": -382.0, + "logps/rejected": -402.0, + "loss": 0.0423, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.057861328125, + "rewards/margins": 11.0625, + "rewards/rejected": -11.0, + "step": 1533 + }, + { + "epoch": 1.8822085889570552, + "grad_norm": 7.89566619418515, + "learning_rate": 5.646694214876033e-07, + "logits/chosen": -0.205078125, + "logits/rejected": -0.337890625, + "logps/chosen": -392.0, + "logps/rejected": -414.0, + "loss": 0.025, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.59375, + "rewards/margins": 12.0, + "rewards/rejected": -12.5625, + "step": 1534 + }, + { + "epoch": 1.883435582822086, + "grad_norm": 11.645948617484482, + "learning_rate": 5.640495867768596e-07, + "logits/chosen": -0.046875, + "logits/rejected": -0.3046875, + "logps/chosen": -374.0, + "logps/rejected": -360.0, + "loss": 0.0563, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.404296875, + "rewards/margins": 11.9375, + "rewards/rejected": -11.5625, + "step": 1535 + }, + { + "epoch": 1.8846625766871166, + "grad_norm": 12.956631655705353, + "learning_rate": 5.634297520661157e-07, + "logits/chosen": -0.056640625, + "logits/rejected": -0.2412109375, + "logps/chosen": -458.0, + "logps/rejected": -440.0, + "loss": 0.0515, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.34375, + "rewards/margins": 11.875, + "rewards/rejected": -11.5625, + "step": 1536 + }, + { + "epoch": 1.8858895705521472, + "grad_norm": 4.634384436264029, + "learning_rate": 5.628099173553719e-07, + "logits/chosen": -0.1123046875, + "logits/rejected": -0.34375, + "logps/chosen": -380.0, + "logps/rejected": -390.0, + "loss": 0.0124, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.037353515625, + "rewards/margins": 12.1875, + "rewards/rejected": -12.125, + "step": 1537 + }, + { + "epoch": 1.887116564417178, + "grad_norm": 5.6435058037855965, + "learning_rate": 5.621900826446281e-07, + "logits/chosen": -0.10498046875, + "logits/rejected": -0.251953125, + "logps/chosen": -366.0, + "logps/rejected": -394.0, + "loss": 0.0103, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.51953125, + "rewards/margins": 12.375, + "rewards/rejected": -11.875, + "step": 1538 + }, + { + "epoch": 1.8883435582822086, + "grad_norm": 8.930748067631301, + "learning_rate": 5.615702479338843e-07, + "logits/chosen": -0.01275634765625, + "logits/rejected": -0.240234375, + "logps/chosen": -380.0, + "logps/rejected": -390.0, + "loss": 0.0294, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.201171875, + "rewards/margins": 11.5625, + "rewards/rejected": -11.375, + "step": 1539 + }, + { + "epoch": 1.8895705521472392, + "grad_norm": 10.323806155698314, + "learning_rate": 5.609504132231405e-07, + "logits/chosen": -0.302734375, + "logits/rejected": -0.4375, + "logps/chosen": -438.0, + "logps/rejected": -468.0, + "loss": 0.0427, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.486328125, + "rewards/margins": 12.625, + "rewards/rejected": -12.125, + "step": 1540 + }, + { + "epoch": 1.89079754601227, + "grad_norm": 11.351287245365558, + "learning_rate": 5.603305785123967e-07, + "logits/chosen": -0.2421875, + "logits/rejected": -0.396484375, + "logps/chosen": -384.0, + "logps/rejected": -406.0, + "loss": 0.0483, + "rewards/accuracies": 0.9765625, + "rewards/chosen": 0.11669921875, + "rewards/margins": 12.0, + "rewards/rejected": -11.875, + "step": 1541 + }, + { + "epoch": 1.8920245398773006, + "grad_norm": 3.2489476684935736, + "learning_rate": 5.597107438016529e-07, + "logits/chosen": -0.2041015625, + "logits/rejected": -0.375, + "logps/chosen": -404.0, + "logps/rejected": -406.0, + "loss": 0.0088, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.80078125, + "rewards/margins": 13.1875, + "rewards/rejected": -12.375, + "step": 1542 + }, + { + "epoch": 1.8932515337423312, + "grad_norm": 8.504627067419271, + "learning_rate": 5.590909090909091e-07, + "logits/chosen": -0.1884765625, + "logits/rejected": -0.357421875, + "logps/chosen": -378.0, + "logps/rejected": -396.0, + "loss": 0.0291, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.158203125, + "rewards/margins": 12.5, + "rewards/rejected": -12.3125, + "step": 1543 + }, + { + "epoch": 1.8944785276073621, + "grad_norm": 8.19609184280983, + "learning_rate": 5.584710743801654e-07, + "logits/chosen": -0.2451171875, + "logits/rejected": -0.408203125, + "logps/chosen": -422.0, + "logps/rejected": -450.0, + "loss": 0.0282, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.314453125, + "rewards/margins": 12.625, + "rewards/rejected": -12.3125, + "step": 1544 + }, + { + "epoch": 1.8957055214723928, + "grad_norm": 8.691100794961908, + "learning_rate": 5.578512396694215e-07, + "logits/chosen": -0.12451171875, + "logits/rejected": -0.326171875, + "logps/chosen": -392.0, + "logps/rejected": -358.0, + "loss": 0.0275, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.294921875, + "rewards/margins": 10.75, + "rewards/rejected": -11.0625, + "step": 1545 + }, + { + "epoch": 1.8969325153374235, + "grad_norm": 8.614624685931496, + "learning_rate": 5.572314049586777e-07, + "logits/chosen": -0.1640625, + "logits/rejected": -0.3359375, + "logps/chosen": -376.0, + "logps/rejected": -372.0, + "loss": 0.024, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.1435546875, + "rewards/margins": 11.5, + "rewards/rejected": -11.375, + "step": 1546 + }, + { + "epoch": 1.8981595092024541, + "grad_norm": 15.386153310990915, + "learning_rate": 5.566115702479339e-07, + "logits/chosen": -0.1494140625, + "logits/rejected": -0.484375, + "logps/chosen": -490.0, + "logps/rejected": -418.0, + "loss": 0.0277, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.361328125, + "rewards/margins": 13.1875, + "rewards/rejected": -12.8125, + "step": 1547 + }, + { + "epoch": 1.8993865030674848, + "grad_norm": 21.182540544956645, + "learning_rate": 5.559917355371901e-07, + "logits/chosen": -0.1572265625, + "logits/rejected": -0.2001953125, + "logps/chosen": -400.0, + "logps/rejected": -430.0, + "loss": 0.0662, + "rewards/accuracies": 0.984375, + "rewards/chosen": -0.11376953125, + "rewards/margins": 11.9375, + "rewards/rejected": -12.0625, + "step": 1548 + }, + { + "epoch": 1.9006134969325155, + "grad_norm": 9.056705219686792, + "learning_rate": 5.553719008264463e-07, + "logits/chosen": -0.046875, + "logits/rejected": -0.26953125, + "logps/chosen": -388.0, + "logps/rejected": -374.0, + "loss": 0.032, + "rewards/accuracies": 0.9765625, + "rewards/chosen": -0.0185546875, + "rewards/margins": 11.4375, + "rewards/rejected": -11.4375, + "step": 1549 + }, + { + "epoch": 1.9018404907975461, + "grad_norm": 12.174899827400916, + "learning_rate": 5.547520661157025e-07, + "logits/chosen": -0.1923828125, + "logits/rejected": -0.2890625, + "logps/chosen": -394.0, + "logps/rejected": -396.0, + "loss": 0.0361, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.0478515625, + "rewards/margins": 12.6875, + "rewards/rejected": -12.6875, + "step": 1550 + }, + { + "epoch": 1.9030674846625768, + "grad_norm": 11.776464119336328, + "learning_rate": 5.541322314049587e-07, + "logits/chosen": -0.2177734375, + "logits/rejected": -0.42578125, + "logps/chosen": -434.0, + "logps/rejected": -416.0, + "loss": 0.0399, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.466796875, + "rewards/margins": 11.375, + "rewards/rejected": -11.8125, + "step": 1551 + }, + { + "epoch": 1.9042944785276075, + "grad_norm": 15.301018527355753, + "learning_rate": 5.535123966942149e-07, + "logits/chosen": -0.1416015625, + "logits/rejected": -0.26953125, + "logps/chosen": -358.0, + "logps/rejected": -362.0, + "loss": 0.071, + "rewards/accuracies": 0.96875, + "rewards/chosen": -0.19140625, + "rewards/margins": 10.375, + "rewards/rejected": -10.5625, + "step": 1552 + }, + { + "epoch": 1.9055214723926381, + "grad_norm": 13.713477834316057, + "learning_rate": 5.528925619834712e-07, + "logits/chosen": -0.078125, + "logits/rejected": -0.2041015625, + "logps/chosen": -392.0, + "logps/rejected": -356.0, + "loss": 0.0288, + "rewards/accuracies": 0.984375, + "rewards/chosen": -0.015625, + "rewards/margins": 11.5, + "rewards/rejected": -11.5, + "step": 1553 + }, + { + "epoch": 1.9067484662576688, + "grad_norm": 13.970935293353676, + "learning_rate": 5.522727272727273e-07, + "logits/chosen": -0.29296875, + "logits/rejected": -0.55078125, + "logps/chosen": -426.0, + "logps/rejected": -406.0, + "loss": 0.0417, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.05810546875, + "rewards/margins": 12.1875, + "rewards/rejected": -12.125, + "step": 1554 + }, + { + "epoch": 1.9079754601226995, + "grad_norm": 17.08022297387505, + "learning_rate": 5.516528925619834e-07, + "logits/chosen": -0.306640625, + "logits/rejected": -0.318359375, + "logps/chosen": -388.0, + "logps/rejected": -414.0, + "loss": 0.0317, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.18359375, + "rewards/margins": 11.875, + "rewards/rejected": -12.0625, + "step": 1555 + }, + { + "epoch": 1.9092024539877301, + "grad_norm": 5.512042362005046, + "learning_rate": 5.510330578512397e-07, + "logits/chosen": -0.099609375, + "logits/rejected": -0.3359375, + "logps/chosen": -410.0, + "logps/rejected": -392.0, + "loss": 0.0214, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.23828125, + "rewards/margins": 11.5625, + "rewards/rejected": -11.375, + "step": 1556 + }, + { + "epoch": 1.9104294478527608, + "grad_norm": 9.862650604155213, + "learning_rate": 5.504132231404959e-07, + "logits/chosen": -0.01397705078125, + "logits/rejected": -0.263671875, + "logps/chosen": -404.0, + "logps/rejected": -414.0, + "loss": 0.0364, + "rewards/accuracies": 0.9765625, + "rewards/chosen": 0.041748046875, + "rewards/margins": 11.25, + "rewards/rejected": -11.1875, + "step": 1557 + }, + { + "epoch": 1.9116564417177915, + "grad_norm": 9.10335392738119, + "learning_rate": 5.497933884297521e-07, + "logits/chosen": -0.173828125, + "logits/rejected": -0.337890625, + "logps/chosen": -394.0, + "logps/rejected": -432.0, + "loss": 0.0229, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.296875, + "rewards/margins": 12.375, + "rewards/rejected": -12.0625, + "step": 1558 + }, + { + "epoch": 1.9128834355828221, + "grad_norm": 10.053624701482729, + "learning_rate": 5.491735537190082e-07, + "logits/chosen": -0.032958984375, + "logits/rejected": -0.1162109375, + "logps/chosen": -396.0, + "logps/rejected": -354.0, + "loss": 0.0385, + "rewards/accuracies": 0.984375, + "rewards/chosen": -0.1142578125, + "rewards/margins": 10.9375, + "rewards/rejected": -11.0625, + "step": 1559 + }, + { + "epoch": 1.9141104294478528, + "grad_norm": 6.811810662950314, + "learning_rate": 5.485537190082645e-07, + "logits/chosen": -0.09228515625, + "logits/rejected": -0.287109375, + "logps/chosen": -376.0, + "logps/rejected": -370.0, + "loss": 0.0255, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.91796875, + "rewards/margins": 11.1875, + "rewards/rejected": -12.125, + "step": 1560 + }, + { + "epoch": 1.9153374233128835, + "grad_norm": 10.499380910825229, + "learning_rate": 5.479338842975207e-07, + "logits/chosen": -0.1923828125, + "logits/rejected": -0.470703125, + "logps/chosen": -456.0, + "logps/rejected": -410.0, + "loss": 0.0363, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.734375, + "rewards/margins": 12.9375, + "rewards/rejected": -12.1875, + "step": 1561 + }, + { + "epoch": 1.9165644171779141, + "grad_norm": 12.232131041485115, + "learning_rate": 5.473140495867769e-07, + "logits/chosen": -0.21875, + "logits/rejected": -0.4296875, + "logps/chosen": -438.0, + "logps/rejected": -414.0, + "loss": 0.03, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 1.34375, + "rewards/margins": 12.625, + "rewards/rejected": -11.25, + "step": 1562 + }, + { + "epoch": 1.9177914110429448, + "grad_norm": 14.943326703058906, + "learning_rate": 5.466942148760331e-07, + "logits/chosen": -0.173828125, + "logits/rejected": -0.408203125, + "logps/chosen": -404.0, + "logps/rejected": -404.0, + "loss": 0.0469, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.7421875, + "rewards/margins": 12.5, + "rewards/rejected": -11.75, + "step": 1563 + }, + { + "epoch": 1.9190184049079755, + "grad_norm": 2.974048087039378, + "learning_rate": 5.460743801652892e-07, + "logits/chosen": -0.1650390625, + "logits/rejected": -0.361328125, + "logps/chosen": -388.0, + "logps/rejected": -402.0, + "loss": 0.0101, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.10302734375, + "rewards/margins": 11.1875, + "rewards/rejected": -11.25, + "step": 1564 + }, + { + "epoch": 1.9202453987730062, + "grad_norm": 6.834929386098197, + "learning_rate": 5.454545454545455e-07, + "logits/chosen": -0.298828125, + "logits/rejected": -0.490234375, + "logps/chosen": -322.0, + "logps/rejected": -342.0, + "loss": 0.0246, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.474609375, + "rewards/margins": 12.0625, + "rewards/rejected": -11.5625, + "step": 1565 + }, + { + "epoch": 1.9214723926380368, + "grad_norm": 19.709537367801847, + "learning_rate": 5.448347107438017e-07, + "logits/chosen": -0.1279296875, + "logits/rejected": -0.267578125, + "logps/chosen": -406.0, + "logps/rejected": -422.0, + "loss": 0.0777, + "rewards/accuracies": 0.9609375, + "rewards/chosen": 0.890625, + "rewards/margins": 12.3125, + "rewards/rejected": -11.4375, + "step": 1566 + }, + { + "epoch": 1.9226993865030675, + "grad_norm": 10.759644336799223, + "learning_rate": 5.442148760330579e-07, + "logits/chosen": -0.15234375, + "logits/rejected": -0.302734375, + "logps/chosen": -376.0, + "logps/rejected": -388.0, + "loss": 0.0364, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.220703125, + "rewards/margins": 12.0, + "rewards/rejected": -11.75, + "step": 1567 + }, + { + "epoch": 1.9239263803680982, + "grad_norm": 20.138956412255148, + "learning_rate": 5.43595041322314e-07, + "logits/chosen": -0.10107421875, + "logits/rejected": -0.328125, + "logps/chosen": -392.0, + "logps/rejected": -404.0, + "loss": 0.0545, + "rewards/accuracies": 0.96875, + "rewards/chosen": 1.2890625, + "rewards/margins": 12.75, + "rewards/rejected": -11.4375, + "step": 1568 + }, + { + "epoch": 1.9251533742331288, + "grad_norm": 4.621297903485169, + "learning_rate": 5.429752066115703e-07, + "logits/chosen": 0.022705078125, + "logits/rejected": -0.1220703125, + "logps/chosen": -390.0, + "logps/rejected": -370.0, + "loss": 0.0098, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.396484375, + "rewards/margins": 12.1875, + "rewards/rejected": -11.8125, + "step": 1569 + }, + { + "epoch": 1.9263803680981595, + "grad_norm": 12.208310607847823, + "learning_rate": 5.423553719008265e-07, + "logits/chosen": -0.109375, + "logits/rejected": -0.33984375, + "logps/chosen": -384.0, + "logps/rejected": -388.0, + "loss": 0.0424, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.12158203125, + "rewards/margins": 11.1875, + "rewards/rejected": -11.3125, + "step": 1570 + }, + { + "epoch": 1.9276073619631902, + "grad_norm": 13.248924763237904, + "learning_rate": 5.417355371900827e-07, + "logits/chosen": -0.1416015625, + "logits/rejected": -0.294921875, + "logps/chosen": -426.0, + "logps/rejected": -430.0, + "loss": 0.0522, + "rewards/accuracies": 0.9609375, + "rewards/chosen": 0.76171875, + "rewards/margins": 12.3125, + "rewards/rejected": -11.5, + "step": 1571 + }, + { + "epoch": 1.9288343558282208, + "grad_norm": 7.855413156125594, + "learning_rate": 5.411157024793388e-07, + "logits/chosen": -0.185546875, + "logits/rejected": -0.392578125, + "logps/chosen": -414.0, + "logps/rejected": -406.0, + "loss": 0.0196, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.1494140625, + "rewards/margins": 12.6875, + "rewards/rejected": -12.5625, + "step": 1572 + }, + { + "epoch": 1.9300613496932515, + "grad_norm": 11.846684123578562, + "learning_rate": 5.40495867768595e-07, + "logits/chosen": -0.17578125, + "logits/rejected": -0.3046875, + "logps/chosen": -402.0, + "logps/rejected": -396.0, + "loss": 0.0502, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.279296875, + "rewards/margins": 11.375, + "rewards/rejected": -11.625, + "step": 1573 + }, + { + "epoch": 1.9312883435582822, + "grad_norm": 15.381676115792011, + "learning_rate": 5.398760330578513e-07, + "logits/chosen": -0.02880859375, + "logits/rejected": -0.162109375, + "logps/chosen": -366.0, + "logps/rejected": -404.0, + "loss": 0.0444, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.078125, + "rewards/margins": 11.3125, + "rewards/rejected": -11.1875, + "step": 1574 + }, + { + "epoch": 1.9325153374233128, + "grad_norm": 16.45905252471069, + "learning_rate": 5.392561983471075e-07, + "logits/chosen": -0.1669921875, + "logits/rejected": -0.30078125, + "logps/chosen": -398.0, + "logps/rejected": -384.0, + "loss": 0.0753, + "rewards/accuracies": 0.9765625, + "rewards/chosen": 0.57421875, + "rewards/margins": 12.5, + "rewards/rejected": -11.9375, + "step": 1575 + }, + { + "epoch": 1.9337423312883435, + "grad_norm": 5.91372888725709, + "learning_rate": 5.386363636363636e-07, + "logits/chosen": -0.1630859375, + "logits/rejected": -0.37890625, + "logps/chosen": -378.0, + "logps/rejected": -384.0, + "loss": 0.0169, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.69921875, + "rewards/margins": 11.5625, + "rewards/rejected": -12.3125, + "step": 1576 + }, + { + "epoch": 1.9349693251533742, + "grad_norm": 13.068737835207726, + "learning_rate": 5.380165289256198e-07, + "logits/chosen": -0.10693359375, + "logits/rejected": -0.294921875, + "logps/chosen": -388.0, + "logps/rejected": -346.0, + "loss": 0.0301, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.82421875, + "rewards/margins": 12.4375, + "rewards/rejected": -11.625, + "step": 1577 + }, + { + "epoch": 1.9361963190184048, + "grad_norm": 12.574938700274403, + "learning_rate": 5.373966942148761e-07, + "logits/chosen": -0.12255859375, + "logits/rejected": -0.296875, + "logps/chosen": -382.0, + "logps/rejected": -420.0, + "loss": 0.0401, + "rewards/accuracies": 0.984375, + "rewards/chosen": -0.0927734375, + "rewards/margins": 12.125, + "rewards/rejected": -12.1875, + "step": 1578 + }, + { + "epoch": 1.9374233128834355, + "grad_norm": 6.1456398779932035, + "learning_rate": 5.367768595041323e-07, + "logits/chosen": -0.1494140625, + "logits/rejected": -0.255859375, + "logps/chosen": -384.0, + "logps/rejected": -386.0, + "loss": 0.0222, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.19140625, + "rewards/margins": 12.1875, + "rewards/rejected": -12.0, + "step": 1579 + }, + { + "epoch": 1.9386503067484662, + "grad_norm": 7.144532600637711, + "learning_rate": 5.361570247933885e-07, + "logits/chosen": -0.15625, + "logits/rejected": -0.376953125, + "logps/chosen": -444.0, + "logps/rejected": -416.0, + "loss": 0.0192, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.78515625, + "rewards/margins": 12.4375, + "rewards/rejected": -11.6875, + "step": 1580 + }, + { + "epoch": 1.9398773006134968, + "grad_norm": 5.065784309723247, + "learning_rate": 5.355371900826446e-07, + "logits/chosen": -0.09765625, + "logits/rejected": -0.2470703125, + "logps/chosen": -416.0, + "logps/rejected": -408.0, + "loss": 0.0165, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.6015625, + "rewards/margins": 12.375, + "rewards/rejected": -11.75, + "step": 1581 + }, + { + "epoch": 1.9411042944785275, + "grad_norm": 8.432710268891027, + "learning_rate": 5.349173553719008e-07, + "logits/chosen": -0.0242919921875, + "logits/rejected": -0.142578125, + "logps/chosen": -332.0, + "logps/rejected": -352.0, + "loss": 0.0494, + "rewards/accuracies": 0.9765625, + "rewards/chosen": -0.31640625, + "rewards/margins": 10.6875, + "rewards/rejected": -11.0, + "step": 1582 + }, + { + "epoch": 1.9423312883435582, + "grad_norm": 14.414561246083233, + "learning_rate": 5.342975206611571e-07, + "logits/chosen": -0.12255859375, + "logits/rejected": -0.30078125, + "logps/chosen": -418.0, + "logps/rejected": -400.0, + "loss": 0.0637, + "rewards/accuracies": 0.9765625, + "rewards/chosen": 0.7734375, + "rewards/margins": 12.1875, + "rewards/rejected": -11.4375, + "step": 1583 + }, + { + "epoch": 1.9435582822085888, + "grad_norm": 13.517917688736201, + "learning_rate": 5.336776859504132e-07, + "logits/chosen": -0.08740234375, + "logits/rejected": -0.271484375, + "logps/chosen": -436.0, + "logps/rejected": -394.0, + "loss": 0.064, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.37890625, + "rewards/margins": 11.0625, + "rewards/rejected": -10.6875, + "step": 1584 + }, + { + "epoch": 1.9447852760736195, + "grad_norm": 18.795243167760674, + "learning_rate": 5.330578512396694e-07, + "logits/chosen": -0.181640625, + "logits/rejected": -0.326171875, + "logps/chosen": -418.0, + "logps/rejected": -434.0, + "loss": 0.073, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.82421875, + "rewards/margins": 13.1875, + "rewards/rejected": -12.375, + "step": 1585 + }, + { + "epoch": 1.9460122699386502, + "grad_norm": 9.671395753436496, + "learning_rate": 5.324380165289256e-07, + "logits/chosen": -0.203125, + "logits/rejected": -0.37890625, + "logps/chosen": -404.0, + "logps/rejected": -386.0, + "loss": 0.0373, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.169921875, + "rewards/margins": 11.125, + "rewards/rejected": -10.9375, + "step": 1586 + }, + { + "epoch": 1.9472392638036808, + "grad_norm": 10.924001936653777, + "learning_rate": 5.318181818181819e-07, + "logits/chosen": -0.1669921875, + "logits/rejected": -0.34765625, + "logps/chosen": -358.0, + "logps/rejected": -366.0, + "loss": 0.0462, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.13671875, + "rewards/margins": 11.1875, + "rewards/rejected": -11.3125, + "step": 1587 + }, + { + "epoch": 1.9484662576687115, + "grad_norm": 10.287013784424305, + "learning_rate": 5.31198347107438e-07, + "logits/chosen": -0.0869140625, + "logits/rejected": -0.251953125, + "logps/chosen": -374.0, + "logps/rejected": -378.0, + "loss": 0.0323, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.35546875, + "rewards/margins": 11.375, + "rewards/rejected": -11.0, + "step": 1588 + }, + { + "epoch": 1.9496932515337422, + "grad_norm": 11.345438684638955, + "learning_rate": 5.305785123966941e-07, + "logits/chosen": -0.1748046875, + "logits/rejected": -0.33984375, + "logps/chosen": -360.0, + "logps/rejected": -388.0, + "loss": 0.028, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.26953125, + "rewards/margins": 12.5, + "rewards/rejected": -12.1875, + "step": 1589 + }, + { + "epoch": 1.9509202453987728, + "grad_norm": 12.380473368422328, + "learning_rate": 5.299586776859504e-07, + "logits/chosen": -0.091796875, + "logits/rejected": -0.232421875, + "logps/chosen": -398.0, + "logps/rejected": -420.0, + "loss": 0.0306, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.4765625, + "rewards/margins": 12.375, + "rewards/rejected": -11.875, + "step": 1590 + }, + { + "epoch": 1.9521472392638037, + "grad_norm": 6.3846910410496855, + "learning_rate": 5.293388429752066e-07, + "logits/chosen": -0.1796875, + "logits/rejected": -0.328125, + "logps/chosen": -364.0, + "logps/rejected": -368.0, + "loss": 0.0163, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.02490234375, + "rewards/margins": 11.875, + "rewards/rejected": -11.875, + "step": 1591 + }, + { + "epoch": 1.9533742331288344, + "grad_norm": 12.161899209714043, + "learning_rate": 5.287190082644629e-07, + "logits/chosen": -0.07666015625, + "logits/rejected": -0.2158203125, + "logps/chosen": -378.0, + "logps/rejected": -402.0, + "loss": 0.0384, + "rewards/accuracies": 0.9765625, + "rewards/chosen": -0.6015625, + "rewards/margins": 11.5, + "rewards/rejected": -12.0625, + "step": 1592 + }, + { + "epoch": 1.954601226993865, + "grad_norm": 16.096985565614556, + "learning_rate": 5.28099173553719e-07, + "logits/chosen": -0.1083984375, + "logits/rejected": -0.263671875, + "logps/chosen": -406.0, + "logps/rejected": -396.0, + "loss": 0.0467, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.478515625, + "rewards/margins": 11.75, + "rewards/rejected": -12.25, + "step": 1593 + }, + { + "epoch": 1.9558282208588957, + "grad_norm": 8.516823030452967, + "learning_rate": 5.274793388429752e-07, + "logits/chosen": -0.2001953125, + "logits/rejected": -0.34765625, + "logps/chosen": -332.0, + "logps/rejected": -344.0, + "loss": 0.0568, + "rewards/accuracies": 0.984375, + "rewards/chosen": -0.1103515625, + "rewards/margins": 11.3125, + "rewards/rejected": -11.4375, + "step": 1594 + }, + { + "epoch": 1.9570552147239264, + "grad_norm": 20.676589588590403, + "learning_rate": 5.268595041322314e-07, + "logits/chosen": -0.146484375, + "logits/rejected": -0.21875, + "logps/chosen": -382.0, + "logps/rejected": -404.0, + "loss": 0.0608, + "rewards/accuracies": 0.984375, + "rewards/chosen": -0.953125, + "rewards/margins": 11.8125, + "rewards/rejected": -12.8125, + "step": 1595 + }, + { + "epoch": 1.958282208588957, + "grad_norm": 11.402306615191995, + "learning_rate": 5.262396694214877e-07, + "logits/chosen": -0.2119140625, + "logits/rejected": -0.25390625, + "logps/chosen": -362.0, + "logps/rejected": -370.0, + "loss": 0.0339, + "rewards/accuracies": 0.9765625, + "rewards/chosen": -0.5234375, + "rewards/margins": 11.8125, + "rewards/rejected": -12.375, + "step": 1596 + }, + { + "epoch": 1.9595092024539877, + "grad_norm": 20.832383714357192, + "learning_rate": 5.256198347107438e-07, + "logits/chosen": -0.12353515625, + "logits/rejected": -0.330078125, + "logps/chosen": -402.0, + "logps/rejected": -394.0, + "loss": 0.0374, + "rewards/accuracies": 0.96875, + "rewards/chosen": -0.25390625, + "rewards/margins": 12.5, + "rewards/rejected": -12.75, + "step": 1597 + }, + { + "epoch": 1.9607361963190184, + "grad_norm": 17.67799399509982, + "learning_rate": 5.25e-07, + "logits/chosen": -0.2216796875, + "logits/rejected": -0.30078125, + "logps/chosen": -396.0, + "logps/rejected": -432.0, + "loss": 0.0746, + "rewards/accuracies": 0.9765625, + "rewards/chosen": -1.0078125, + "rewards/margins": 11.625, + "rewards/rejected": -12.6875, + "step": 1598 + }, + { + "epoch": 1.961963190184049, + "grad_norm": 2.4424485737982744, + "learning_rate": 5.243801652892562e-07, + "logits/chosen": -0.123046875, + "logits/rejected": -0.36328125, + "logps/chosen": -402.0, + "logps/rejected": -388.0, + "loss": 0.011, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.27734375, + "rewards/margins": 12.0, + "rewards/rejected": -12.25, + "step": 1599 + }, + { + "epoch": 1.9631901840490797, + "grad_norm": 13.907963965939771, + "learning_rate": 5.237603305785124e-07, + "logits/chosen": -0.06884765625, + "logits/rejected": -0.2392578125, + "logps/chosen": -416.0, + "logps/rejected": -410.0, + "loss": 0.0418, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.5078125, + "rewards/margins": 11.625, + "rewards/rejected": -12.125, + "step": 1600 + }, + { + "epoch": 1.9644171779141104, + "grad_norm": 13.814689419274158, + "learning_rate": 5.231404958677686e-07, + "logits/chosen": -0.033447265625, + "logits/rejected": -0.1435546875, + "logps/chosen": -346.0, + "logps/rejected": -372.0, + "loss": 0.0472, + "rewards/accuracies": 0.9765625, + "rewards/chosen": -0.98046875, + "rewards/margins": 10.6875, + "rewards/rejected": -11.6875, + "step": 1601 + }, + { + "epoch": 1.965644171779141, + "grad_norm": 11.545621891190732, + "learning_rate": 5.225206611570248e-07, + "logits/chosen": -0.11962890625, + "logits/rejected": -0.326171875, + "logps/chosen": -442.0, + "logps/rejected": -380.0, + "loss": 0.0297, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.2578125, + "rewards/margins": 12.1875, + "rewards/rejected": -11.9375, + "step": 1602 + }, + { + "epoch": 1.9668711656441717, + "grad_norm": 8.879321275799807, + "learning_rate": 5.21900826446281e-07, + "logits/chosen": -0.146484375, + "logits/rejected": -0.341796875, + "logps/chosen": -472.0, + "logps/rejected": -436.0, + "loss": 0.0337, + "rewards/accuracies": 0.984375, + "rewards/chosen": -0.373046875, + "rewards/margins": 13.0, + "rewards/rejected": -13.3125, + "step": 1603 + }, + { + "epoch": 1.9680981595092024, + "grad_norm": 10.570209103673362, + "learning_rate": 5.212809917355372e-07, + "logits/chosen": -0.080078125, + "logits/rejected": -0.2734375, + "logps/chosen": -394.0, + "logps/rejected": -392.0, + "loss": 0.0386, + "rewards/accuracies": 0.984375, + "rewards/chosen": -0.07861328125, + "rewards/margins": 11.1875, + "rewards/rejected": -11.3125, + "step": 1604 + }, + { + "epoch": 1.969325153374233, + "grad_norm": 11.659167158794315, + "learning_rate": 5.206611570247935e-07, + "logits/chosen": -0.08837890625, + "logits/rejected": -0.232421875, + "logps/chosen": -434.0, + "logps/rejected": -436.0, + "loss": 0.0439, + "rewards/accuracies": 0.984375, + "rewards/chosen": -0.875, + "rewards/margins": 11.625, + "rewards/rejected": -12.5, + "step": 1605 + }, + { + "epoch": 1.9705521472392638, + "grad_norm": 4.504606116838098, + "learning_rate": 5.200413223140496e-07, + "logits/chosen": -0.2138671875, + "logits/rejected": -0.427734375, + "logps/chosen": -418.0, + "logps/rejected": -434.0, + "loss": 0.0155, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.2470703125, + "rewards/margins": 12.5625, + "rewards/rejected": -12.3125, + "step": 1606 + }, + { + "epoch": 1.9717791411042946, + "grad_norm": 3.9125260909977104, + "learning_rate": 5.194214876033058e-07, + "logits/chosen": -0.16015625, + "logits/rejected": -0.396484375, + "logps/chosen": -384.0, + "logps/rejected": -370.0, + "loss": 0.0139, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.2578125, + "rewards/margins": 12.75, + "rewards/rejected": -12.4375, + "step": 1607 + }, + { + "epoch": 1.9730061349693253, + "grad_norm": 12.198720753980819, + "learning_rate": 5.18801652892562e-07, + "logits/chosen": -0.166015625, + "logits/rejected": -0.369140625, + "logps/chosen": -392.0, + "logps/rejected": -414.0, + "loss": 0.0264, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.2236328125, + "rewards/margins": 12.75, + "rewards/rejected": -12.5625, + "step": 1608 + }, + { + "epoch": 1.974233128834356, + "grad_norm": 13.187714303157518, + "learning_rate": 5.181818181818182e-07, + "logits/chosen": -0.1806640625, + "logits/rejected": -0.3125, + "logps/chosen": -378.0, + "logps/rejected": -398.0, + "loss": 0.0497, + "rewards/accuracies": 0.984375, + "rewards/chosen": -0.57421875, + "rewards/margins": 12.0625, + "rewards/rejected": -12.625, + "step": 1609 + }, + { + "epoch": 1.9754601226993866, + "grad_norm": 13.077595266689482, + "learning_rate": 5.175619834710744e-07, + "logits/chosen": -0.22265625, + "logits/rejected": -0.3125, + "logps/chosen": -366.0, + "logps/rejected": -362.0, + "loss": 0.0473, + "rewards/accuracies": 0.984375, + "rewards/chosen": -0.06982421875, + "rewards/margins": 12.0, + "rewards/rejected": -12.0625, + "step": 1610 + }, + { + "epoch": 1.9766871165644173, + "grad_norm": 11.034197753449089, + "learning_rate": 5.169421487603305e-07, + "logits/chosen": -0.0849609375, + "logits/rejected": -0.216796875, + "logps/chosen": -328.0, + "logps/rejected": -342.0, + "loss": 0.0548, + "rewards/accuracies": 0.9765625, + "rewards/chosen": -0.78125, + "rewards/margins": 10.5, + "rewards/rejected": -11.25, + "step": 1611 + }, + { + "epoch": 1.977914110429448, + "grad_norm": 19.25493001940352, + "learning_rate": 5.163223140495868e-07, + "logits/chosen": -0.07275390625, + "logits/rejected": -0.275390625, + "logps/chosen": -436.0, + "logps/rejected": -462.0, + "loss": 0.0788, + "rewards/accuracies": 0.96875, + "rewards/chosen": 0.478515625, + "rewards/margins": 12.75, + "rewards/rejected": -12.25, + "step": 1612 + }, + { + "epoch": 1.9791411042944786, + "grad_norm": 2.50058852484782, + "learning_rate": 5.15702479338843e-07, + "logits/chosen": -0.041259765625, + "logits/rejected": -0.2119140625, + "logps/chosen": -442.0, + "logps/rejected": -420.0, + "loss": 0.0116, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.1943359375, + "rewards/margins": 13.5, + "rewards/rejected": -13.3125, + "step": 1613 + }, + { + "epoch": 1.9803680981595093, + "grad_norm": 9.641767117339855, + "learning_rate": 5.150826446280992e-07, + "logits/chosen": -0.259765625, + "logits/rejected": -0.470703125, + "logps/chosen": -436.0, + "logps/rejected": -432.0, + "loss": 0.0409, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.1953125, + "rewards/margins": 13.0, + "rewards/rejected": -12.8125, + "step": 1614 + }, + { + "epoch": 1.98159509202454, + "grad_norm": 13.7846054101889, + "learning_rate": 5.144628099173554e-07, + "logits/chosen": -0.283203125, + "logits/rejected": -0.37890625, + "logps/chosen": -376.0, + "logps/rejected": -414.0, + "loss": 0.0534, + "rewards/accuracies": 0.96875, + "rewards/chosen": 0.126953125, + "rewards/margins": 12.6875, + "rewards/rejected": -12.5625, + "step": 1615 + }, + { + "epoch": 1.9828220858895707, + "grad_norm": 10.800122873462064, + "learning_rate": 5.138429752066116e-07, + "logits/chosen": -0.0927734375, + "logits/rejected": -0.2470703125, + "logps/chosen": -352.0, + "logps/rejected": -348.0, + "loss": 0.0361, + "rewards/accuracies": 0.984375, + "rewards/chosen": -0.28515625, + "rewards/margins": 11.3125, + "rewards/rejected": -11.625, + "step": 1616 + }, + { + "epoch": 1.9840490797546013, + "grad_norm": 10.77210403419169, + "learning_rate": 5.132231404958678e-07, + "logits/chosen": -0.033935546875, + "logits/rejected": -0.2216796875, + "logps/chosen": -360.0, + "logps/rejected": -372.0, + "loss": 0.0398, + "rewards/accuracies": 0.9765625, + "rewards/chosen": -0.0654296875, + "rewards/margins": 10.875, + "rewards/rejected": -10.9375, + "step": 1617 + }, + { + "epoch": 1.985276073619632, + "grad_norm": 9.529792792248653, + "learning_rate": 5.126033057851239e-07, + "logits/chosen": -0.2080078125, + "logits/rejected": -0.3203125, + "logps/chosen": -370.0, + "logps/rejected": -378.0, + "loss": 0.037, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.2373046875, + "rewards/margins": 11.875, + "rewards/rejected": -11.6875, + "step": 1618 + }, + { + "epoch": 1.9865030674846627, + "grad_norm": 8.741040335110851, + "learning_rate": 5.119834710743802e-07, + "logits/chosen": -0.2314453125, + "logits/rejected": -0.384765625, + "logps/chosen": -416.0, + "logps/rejected": -388.0, + "loss": 0.0321, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.0927734375, + "rewards/margins": 11.75, + "rewards/rejected": -11.875, + "step": 1619 + }, + { + "epoch": 1.9877300613496933, + "grad_norm": 7.785518167433976, + "learning_rate": 5.113636363636363e-07, + "logits/chosen": -0.134765625, + "logits/rejected": -0.3203125, + "logps/chosen": -372.0, + "logps/rejected": -372.0, + "loss": 0.0185, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.33984375, + "rewards/margins": 11.75, + "rewards/rejected": -11.375, + "step": 1620 + }, + { + "epoch": 1.988957055214724, + "grad_norm": 5.464211487320141, + "learning_rate": 5.107438016528926e-07, + "logits/chosen": -0.10888671875, + "logits/rejected": -0.2353515625, + "logps/chosen": -408.0, + "logps/rejected": -422.0, + "loss": 0.0187, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.39453125, + "rewards/margins": 12.4375, + "rewards/rejected": -12.0625, + "step": 1621 + }, + { + "epoch": 1.9901840490797547, + "grad_norm": 13.071939990292023, + "learning_rate": 5.101239669421488e-07, + "logits/chosen": -0.3828125, + "logits/rejected": -0.5078125, + "logps/chosen": -408.0, + "logps/rejected": -442.0, + "loss": 0.0443, + "rewards/accuracies": 0.9765625, + "rewards/chosen": -0.10546875, + "rewards/margins": 11.9375, + "rewards/rejected": -12.0625, + "step": 1622 + }, + { + "epoch": 1.9914110429447853, + "grad_norm": 3.434637787021447, + "learning_rate": 5.09504132231405e-07, + "logits/chosen": -0.0947265625, + "logits/rejected": -0.30078125, + "logps/chosen": -402.0, + "logps/rejected": -402.0, + "loss": 0.0095, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.61328125, + "rewards/margins": 12.5, + "rewards/rejected": -11.875, + "step": 1623 + }, + { + "epoch": 1.992638036809816, + "grad_norm": 3.315305852103551, + "learning_rate": 5.088842975206612e-07, + "logits/chosen": -0.212890625, + "logits/rejected": -0.474609375, + "logps/chosen": -376.0, + "logps/rejected": -394.0, + "loss": 0.0142, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.75390625, + "rewards/margins": 12.75, + "rewards/rejected": -12.0, + "step": 1624 + }, + { + "epoch": 1.9938650306748467, + "grad_norm": 22.7278950349499, + "learning_rate": 5.082644628099174e-07, + "logits/chosen": -0.043212890625, + "logits/rejected": -0.2578125, + "logps/chosen": -440.0, + "logps/rejected": -408.0, + "loss": 0.0652, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.7734375, + "rewards/margins": 12.375, + "rewards/rejected": -11.625, + "step": 1625 + }, + { + "epoch": 1.9950920245398773, + "grad_norm": 7.501231527411867, + "learning_rate": 5.076446280991736e-07, + "logits/chosen": -0.17578125, + "logits/rejected": -0.28125, + "logps/chosen": -432.0, + "logps/rejected": -426.0, + "loss": 0.0195, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.251953125, + "rewards/margins": 12.0, + "rewards/rejected": -11.75, + "step": 1626 + }, + { + "epoch": 1.996319018404908, + "grad_norm": 8.75791968734159, + "learning_rate": 5.070247933884297e-07, + "logits/chosen": -0.24609375, + "logits/rejected": -0.345703125, + "logps/chosen": -372.0, + "logps/rejected": -368.0, + "loss": 0.0302, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.4765625, + "rewards/margins": 12.1875, + "rewards/rejected": -11.6875, + "step": 1627 + }, + { + "epoch": 1.9975460122699387, + "grad_norm": 12.320799734184225, + "learning_rate": 5.06404958677686e-07, + "logits/chosen": -0.1689453125, + "logits/rejected": -0.296875, + "logps/chosen": -368.0, + "logps/rejected": -406.0, + "loss": 0.0386, + "rewards/accuracies": 0.9765625, + "rewards/chosen": 0.65625, + "rewards/margins": 12.6875, + "rewards/rejected": -12.0, + "step": 1628 + }, + { + "epoch": 1.9987730061349693, + "grad_norm": 16.231968818300622, + "learning_rate": 5.057851239669421e-07, + "logits/chosen": -0.05126953125, + "logits/rejected": -0.216796875, + "logps/chosen": -398.0, + "logps/rejected": -408.0, + "loss": 0.0634, + "rewards/accuracies": 0.9765625, + "rewards/chosen": -0.11181640625, + "rewards/margins": 11.5, + "rewards/rejected": -11.625, + "step": 1629 + }, + { + "epoch": 2.0, + "grad_norm": 7.251135077046663, + "learning_rate": 5.051652892561984e-07, + "logits/chosen": -0.1787109375, + "logits/rejected": -0.341796875, + "logps/chosen": -428.0, + "logps/rejected": -424.0, + "loss": 0.0224, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.1337890625, + "rewards/margins": 12.75, + "rewards/rejected": -12.625, + "step": 1630 + }, + { + "epoch": 2.0012269938650307, + "grad_norm": 0.5204524366827928, + "learning_rate": 5.045454545454545e-07, + "logits/chosen": -0.1337890625, + "logits/rejected": -0.36328125, + "logps/chosen": -434.0, + "logps/rejected": -400.0, + "loss": 0.002, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.91796875, + "rewards/margins": 13.4375, + "rewards/rejected": -12.5, + "step": 1631 + }, + { + "epoch": 2.0024539877300613, + "grad_norm": 0.5662039341048904, + "learning_rate": 5.039256198347108e-07, + "logits/chosen": -0.103515625, + "logits/rejected": -0.2216796875, + "logps/chosen": -378.0, + "logps/rejected": -412.0, + "loss": 0.0016, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.93359375, + "rewards/margins": 13.375, + "rewards/rejected": -12.4375, + "step": 1632 + }, + { + "epoch": 2.003680981595092, + "grad_norm": 2.1326080177048703, + "learning_rate": 5.03305785123967e-07, + "logits/chosen": -0.0011444091796875, + "logits/rejected": -0.1875, + "logps/chosen": -462.0, + "logps/rejected": -406.0, + "loss": 0.0063, + "rewards/accuracies": 1.0, + "rewards/chosen": 1.0234375, + "rewards/margins": 13.1875, + "rewards/rejected": -12.125, + "step": 1633 + }, + { + "epoch": 2.0049079754601227, + "grad_norm": 1.1003722866419343, + "learning_rate": 5.026859504132232e-07, + "logits/chosen": -0.18359375, + "logits/rejected": -0.3515625, + "logps/chosen": -394.0, + "logps/rejected": -412.0, + "loss": 0.0042, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.031005859375, + "rewards/margins": 12.75, + "rewards/rejected": -12.6875, + "step": 1634 + }, + { + "epoch": 2.0061349693251533, + "grad_norm": 1.057258934353059, + "learning_rate": 5.020661157024793e-07, + "logits/chosen": -0.130859375, + "logits/rejected": -0.283203125, + "logps/chosen": -406.0, + "logps/rejected": -406.0, + "loss": 0.0038, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.79296875, + "rewards/margins": 14.0, + "rewards/rejected": -13.25, + "step": 1635 + }, + { + "epoch": 2.007361963190184, + "grad_norm": 0.5163494249205891, + "learning_rate": 5.014462809917355e-07, + "logits/chosen": -0.169921875, + "logits/rejected": -0.341796875, + "logps/chosen": -390.0, + "logps/rejected": -404.0, + "loss": 0.0019, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.55078125, + "rewards/margins": 13.9375, + "rewards/rejected": -13.375, + "step": 1636 + }, + { + "epoch": 2.0085889570552147, + "grad_norm": 2.9368729995631018, + "learning_rate": 5.008264462809918e-07, + "logits/chosen": -0.12890625, + "logits/rejected": -0.265625, + "logps/chosen": -338.0, + "logps/rejected": -368.0, + "loss": 0.0053, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.30078125, + "rewards/margins": 12.0625, + "rewards/rejected": -12.3125, + "step": 1637 + }, + { + "epoch": 2.0098159509202453, + "grad_norm": 0.5837971177243837, + "learning_rate": 5.002066115702479e-07, + "logits/chosen": -0.232421875, + "logits/rejected": -0.310546875, + "logps/chosen": -386.0, + "logps/rejected": -412.0, + "loss": 0.0023, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.5625, + "rewards/margins": 14.0625, + "rewards/rejected": -13.5, + "step": 1638 + }, + { + "epoch": 2.011042944785276, + "grad_norm": 0.9496637317471645, + "learning_rate": 4.995867768595042e-07, + "logits/chosen": -0.09912109375, + "logits/rejected": -0.2236328125, + "logps/chosen": -348.0, + "logps/rejected": -392.0, + "loss": 0.0027, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.78515625, + "rewards/margins": 13.1875, + "rewards/rejected": -12.4375, + "step": 1639 + }, + { + "epoch": 2.0122699386503067, + "grad_norm": 0.633005358240855, + "learning_rate": 4.989669421487603e-07, + "logits/chosen": -0.0888671875, + "logits/rejected": -0.26953125, + "logps/chosen": -390.0, + "logps/rejected": -424.0, + "loss": 0.0023, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.140625, + "rewards/margins": 12.75, + "rewards/rejected": -12.625, + "step": 1640 + }, + { + "epoch": 2.0134969325153373, + "grad_norm": 1.1803266889891297, + "learning_rate": 4.983471074380166e-07, + "logits/chosen": -0.150390625, + "logits/rejected": -0.330078125, + "logps/chosen": -388.0, + "logps/rejected": -416.0, + "loss": 0.0028, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.400390625, + "rewards/margins": 12.9375, + "rewards/rejected": -12.5, + "step": 1641 + }, + { + "epoch": 2.014723926380368, + "grad_norm": 1.945429507712606, + "learning_rate": 4.977272727272728e-07, + "logits/chosen": -0.15625, + "logits/rejected": -0.326171875, + "logps/chosen": -420.0, + "logps/rejected": -448.0, + "loss": 0.0084, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.220703125, + "rewards/margins": 13.6875, + "rewards/rejected": -13.5, + "step": 1642 + }, + { + "epoch": 2.0159509202453987, + "grad_norm": 1.1746863030218866, + "learning_rate": 4.97107438016529e-07, + "logits/chosen": -0.10205078125, + "logits/rejected": -0.2412109375, + "logps/chosen": -340.0, + "logps/rejected": -366.0, + "loss": 0.0037, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.474609375, + "rewards/margins": 12.3125, + "rewards/rejected": -12.8125, + "step": 1643 + }, + { + "epoch": 2.0171779141104293, + "grad_norm": 0.9932806825245308, + "learning_rate": 4.964876033057851e-07, + "logits/chosen": -0.12255859375, + "logits/rejected": -0.353515625, + "logps/chosen": -432.0, + "logps/rejected": -458.0, + "loss": 0.0034, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.8125, + "rewards/margins": 13.75, + "rewards/rejected": -12.9375, + "step": 1644 + }, + { + "epoch": 2.01840490797546, + "grad_norm": 3.876583681015391, + "learning_rate": 4.958677685950413e-07, + "logits/chosen": -0.1875, + "logits/rejected": -0.369140625, + "logps/chosen": -398.0, + "logps/rejected": -398.0, + "loss": 0.0059, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.89453125, + "rewards/margins": 13.625, + "rewards/rejected": -12.6875, + "step": 1645 + }, + { + "epoch": 2.0196319018404907, + "grad_norm": 3.2664101521962605, + "learning_rate": 4.952479338842976e-07, + "logits/chosen": -0.099609375, + "logits/rejected": -0.162109375, + "logps/chosen": -352.0, + "logps/rejected": -380.0, + "loss": 0.0057, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.07861328125, + "rewards/margins": 11.9375, + "rewards/rejected": -11.875, + "step": 1646 + }, + { + "epoch": 2.0208588957055214, + "grad_norm": 1.912106504346034, + "learning_rate": 4.946280991735537e-07, + "logits/chosen": -0.263671875, + "logits/rejected": -0.53125, + "logps/chosen": -442.0, + "logps/rejected": -406.0, + "loss": 0.0037, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.0830078125, + "rewards/margins": 13.5, + "rewards/rejected": -13.4375, + "step": 1647 + }, + { + "epoch": 2.022085889570552, + "grad_norm": 1.5787823377673078, + "learning_rate": 4.940082644628099e-07, + "logits/chosen": -0.2119140625, + "logits/rejected": -0.392578125, + "logps/chosen": -390.0, + "logps/rejected": -426.0, + "loss": 0.0109, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.146484375, + "rewards/margins": 13.75, + "rewards/rejected": -13.625, + "step": 1648 + }, + { + "epoch": 2.0233128834355827, + "grad_norm": 0.9800480609128162, + "learning_rate": 4.933884297520661e-07, + "logits/chosen": -0.224609375, + "logits/rejected": -0.435546875, + "logps/chosen": -384.0, + "logps/rejected": -390.0, + "loss": 0.003, + "rewards/accuracies": 1.0, + "rewards/chosen": 1.171875, + "rewards/margins": 13.375, + "rewards/rejected": -12.1875, + "step": 1649 + }, + { + "epoch": 2.0245398773006134, + "grad_norm": 2.037740299728709, + "learning_rate": 4.927685950413224e-07, + "logits/chosen": -0.064453125, + "logits/rejected": -0.333984375, + "logps/chosen": -392.0, + "logps/rejected": -362.0, + "loss": 0.0106, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.17578125, + "rewards/margins": 12.0, + "rewards/rejected": -11.8125, + "step": 1650 + }, + { + "epoch": 2.025766871165644, + "grad_norm": 0.7557706445618647, + "learning_rate": 4.921487603305785e-07, + "logits/chosen": -0.1865234375, + "logits/rejected": -0.361328125, + "logps/chosen": -430.0, + "logps/rejected": -436.0, + "loss": 0.0018, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.63671875, + "rewards/margins": 13.375, + "rewards/rejected": -12.75, + "step": 1651 + }, + { + "epoch": 2.0269938650306747, + "grad_norm": 1.6028274145794823, + "learning_rate": 4.915289256198346e-07, + "logits/chosen": -0.142578125, + "logits/rejected": -0.263671875, + "logps/chosen": -344.0, + "logps/rejected": -380.0, + "loss": 0.0041, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.2890625, + "rewards/margins": 12.5, + "rewards/rejected": -12.1875, + "step": 1652 + }, + { + "epoch": 2.0282208588957054, + "grad_norm": 1.1879405185647063, + "learning_rate": 4.909090909090909e-07, + "logits/chosen": -0.140625, + "logits/rejected": -0.294921875, + "logps/chosen": -380.0, + "logps/rejected": -378.0, + "loss": 0.008, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.404296875, + "rewards/margins": 12.9375, + "rewards/rejected": -12.5, + "step": 1653 + }, + { + "epoch": 2.029447852760736, + "grad_norm": 2.437511154657876, + "learning_rate": 4.902892561983471e-07, + "logits/chosen": 0.042724609375, + "logits/rejected": -0.1572265625, + "logps/chosen": -406.0, + "logps/rejected": -400.0, + "loss": 0.0092, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.396484375, + "rewards/margins": 12.5, + "rewards/rejected": -12.0625, + "step": 1654 + }, + { + "epoch": 2.0306748466257667, + "grad_norm": 0.5322032894812679, + "learning_rate": 4.896694214876034e-07, + "logits/chosen": -0.267578125, + "logits/rejected": -0.35546875, + "logps/chosen": -406.0, + "logps/rejected": -434.0, + "loss": 0.0019, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.451171875, + "rewards/margins": 13.8125, + "rewards/rejected": -13.375, + "step": 1655 + }, + { + "epoch": 2.0319018404907974, + "grad_norm": 0.7420031304997047, + "learning_rate": 4.890495867768595e-07, + "logits/chosen": -0.1357421875, + "logits/rejected": -0.255859375, + "logps/chosen": -450.0, + "logps/rejected": -444.0, + "loss": 0.0021, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.0303955078125, + "rewards/margins": 13.125, + "rewards/rejected": -13.125, + "step": 1656 + }, + { + "epoch": 2.033128834355828, + "grad_norm": 1.2111846320782014, + "learning_rate": 4.884297520661157e-07, + "logits/chosen": -0.0966796875, + "logits/rejected": -0.166015625, + "logps/chosen": -380.0, + "logps/rejected": -424.0, + "loss": 0.0039, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.53515625, + "rewards/margins": 13.25, + "rewards/rejected": -12.75, + "step": 1657 + }, + { + "epoch": 2.0343558282208587, + "grad_norm": 8.556637736674327, + "learning_rate": 4.878099173553719e-07, + "logits/chosen": -0.15625, + "logits/rejected": -0.279296875, + "logps/chosen": -388.0, + "logps/rejected": -432.0, + "loss": 0.0243, + "rewards/accuracies": 0.984375, + "rewards/chosen": -0.158203125, + "rewards/margins": 12.875, + "rewards/rejected": -13.0625, + "step": 1658 + }, + { + "epoch": 2.0355828220858894, + "grad_norm": 0.8344861667694685, + "learning_rate": 4.871900826446282e-07, + "logits/chosen": -0.2578125, + "logits/rejected": -0.427734375, + "logps/chosen": -418.0, + "logps/rejected": -458.0, + "loss": 0.0021, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.62109375, + "rewards/margins": 13.8125, + "rewards/rejected": -14.4375, + "step": 1659 + }, + { + "epoch": 2.03680981595092, + "grad_norm": 9.39912696757979, + "learning_rate": 4.865702479338843e-07, + "logits/chosen": -0.19140625, + "logits/rejected": -0.30078125, + "logps/chosen": -376.0, + "logps/rejected": -392.0, + "loss": 0.0134, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.2890625, + "rewards/margins": 12.5, + "rewards/rejected": -12.75, + "step": 1660 + }, + { + "epoch": 2.038036809815951, + "grad_norm": 0.7643626806797574, + "learning_rate": 4.859504132231405e-07, + "logits/chosen": -0.30859375, + "logits/rejected": -0.3984375, + "logps/chosen": -370.0, + "logps/rejected": -396.0, + "loss": 0.0018, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.578125, + "rewards/margins": 13.5625, + "rewards/rejected": -14.125, + "step": 1661 + }, + { + "epoch": 2.039263803680982, + "grad_norm": 1.7895261435958536, + "learning_rate": 4.853305785123967e-07, + "logits/chosen": -0.3125, + "logits/rejected": -0.5234375, + "logps/chosen": -468.0, + "logps/rejected": -456.0, + "loss": 0.0036, + "rewards/accuracies": 1.0, + "rewards/chosen": 1.1875, + "rewards/margins": 15.8125, + "rewards/rejected": -14.625, + "step": 1662 + }, + { + "epoch": 2.0404907975460125, + "grad_norm": 1.2489450937032716, + "learning_rate": 4.847107438016529e-07, + "logits/chosen": -0.11767578125, + "logits/rejected": -0.396484375, + "logps/chosen": -430.0, + "logps/rejected": -436.0, + "loss": 0.0025, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.0218505859375, + "rewards/margins": 14.125, + "rewards/rejected": -14.125, + "step": 1663 + }, + { + "epoch": 2.041717791411043, + "grad_norm": 0.45184451446095947, + "learning_rate": 4.840909090909092e-07, + "logits/chosen": -0.1953125, + "logits/rejected": -0.314453125, + "logps/chosen": -392.0, + "logps/rejected": -396.0, + "loss": 0.0014, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.09912109375, + "rewards/margins": 14.25, + "rewards/rejected": -14.375, + "step": 1664 + }, + { + "epoch": 2.042944785276074, + "grad_norm": 2.24452873979522, + "learning_rate": 4.834710743801652e-07, + "logits/chosen": -0.19140625, + "logits/rejected": -0.328125, + "logps/chosen": -396.0, + "logps/rejected": -382.0, + "loss": 0.0046, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.56640625, + "rewards/margins": 12.5, + "rewards/rejected": -13.0625, + "step": 1665 + }, + { + "epoch": 2.0441717791411045, + "grad_norm": 1.8162003295866649, + "learning_rate": 4.828512396694215e-07, + "logits/chosen": -0.154296875, + "logits/rejected": -0.337890625, + "logps/chosen": -372.0, + "logps/rejected": -392.0, + "loss": 0.0033, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.423828125, + "rewards/margins": 13.8125, + "rewards/rejected": -13.375, + "step": 1666 + }, + { + "epoch": 2.045398773006135, + "grad_norm": 1.1278980834058954, + "learning_rate": 4.822314049586777e-07, + "logits/chosen": -0.275390625, + "logits/rejected": -0.439453125, + "logps/chosen": -422.0, + "logps/rejected": -426.0, + "loss": 0.0083, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -1.3046875, + "rewards/margins": 12.6875, + "rewards/rejected": -14.0, + "step": 1667 + }, + { + "epoch": 2.046625766871166, + "grad_norm": 0.8260259607817262, + "learning_rate": 4.81611570247934e-07, + "logits/chosen": -0.271484375, + "logits/rejected": -0.43359375, + "logps/chosen": -372.0, + "logps/rejected": -402.0, + "loss": 0.0018, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.228515625, + "rewards/margins": 13.6875, + "rewards/rejected": -13.4375, + "step": 1668 + }, + { + "epoch": 2.0478527607361965, + "grad_norm": 1.938836721194188, + "learning_rate": 4.809917355371901e-07, + "logits/chosen": -0.1259765625, + "logits/rejected": -0.287109375, + "logps/chosen": -382.0, + "logps/rejected": -404.0, + "loss": 0.0112, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.185546875, + "rewards/margins": 12.8125, + "rewards/rejected": -13.0, + "step": 1669 + }, + { + "epoch": 2.049079754601227, + "grad_norm": 2.44071014344612, + "learning_rate": 4.803719008264463e-07, + "logits/chosen": -0.197265625, + "logits/rejected": -0.435546875, + "logps/chosen": -418.0, + "logps/rejected": -432.0, + "loss": 0.0033, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.55078125, + "rewards/margins": 14.8125, + "rewards/rejected": -14.25, + "step": 1670 + }, + { + "epoch": 2.050306748466258, + "grad_norm": 0.4394178857267706, + "learning_rate": 4.797520661157025e-07, + "logits/chosen": -0.216796875, + "logits/rejected": -0.3828125, + "logps/chosen": -414.0, + "logps/rejected": -424.0, + "loss": 0.0013, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.59765625, + "rewards/margins": 14.3125, + "rewards/rejected": -13.6875, + "step": 1671 + }, + { + "epoch": 2.0515337423312885, + "grad_norm": 0.6202467405653835, + "learning_rate": 4.791322314049587e-07, + "logits/chosen": -0.1455078125, + "logits/rejected": -0.3828125, + "logps/chosen": -400.0, + "logps/rejected": -420.0, + "loss": 0.0015, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.6484375, + "rewards/margins": 14.1875, + "rewards/rejected": -13.5, + "step": 1672 + }, + { + "epoch": 2.052760736196319, + "grad_norm": 1.3279336913281012, + "learning_rate": 4.785123966942149e-07, + "logits/chosen": -0.09130859375, + "logits/rejected": -0.1591796875, + "logps/chosen": -388.0, + "logps/rejected": -396.0, + "loss": 0.0037, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.26171875, + "rewards/margins": 12.625, + "rewards/rejected": -12.375, + "step": 1673 + }, + { + "epoch": 2.05398773006135, + "grad_norm": 2.0669645541712027, + "learning_rate": 4.77892561983471e-07, + "logits/chosen": -0.2236328125, + "logits/rejected": -0.421875, + "logps/chosen": -386.0, + "logps/rejected": -408.0, + "loss": 0.004, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.470703125, + "rewards/margins": 14.125, + "rewards/rejected": -13.6875, + "step": 1674 + }, + { + "epoch": 2.0552147239263805, + "grad_norm": 1.1449941039533331, + "learning_rate": 4.772727272727273e-07, + "logits/chosen": -0.2412109375, + "logits/rejected": -0.423828125, + "logps/chosen": -440.0, + "logps/rejected": -432.0, + "loss": 0.0027, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.8125, + "rewards/margins": 13.9375, + "rewards/rejected": -13.125, + "step": 1675 + }, + { + "epoch": 2.056441717791411, + "grad_norm": 3.193374979191636, + "learning_rate": 4.7665289256198347e-07, + "logits/chosen": -0.11962890625, + "logits/rejected": -0.306640625, + "logps/chosen": -406.0, + "logps/rejected": -410.0, + "loss": 0.0049, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.43359375, + "rewards/margins": 13.5625, + "rewards/rejected": -13.125, + "step": 1676 + }, + { + "epoch": 2.057668711656442, + "grad_norm": 0.5713027756444637, + "learning_rate": 4.760330578512397e-07, + "logits/chosen": -0.3359375, + "logits/rejected": -0.515625, + "logps/chosen": -352.0, + "logps/rejected": -380.0, + "loss": 0.0069, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.890625, + "rewards/margins": 13.3125, + "rewards/rejected": -14.1875, + "step": 1677 + }, + { + "epoch": 2.0588957055214725, + "grad_norm": 2.0338861912705397, + "learning_rate": 4.7541322314049586e-07, + "logits/chosen": -0.1689453125, + "logits/rejected": -0.2412109375, + "logps/chosen": -380.0, + "logps/rejected": -444.0, + "loss": 0.0043, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.6796875, + "rewards/margins": 13.1875, + "rewards/rejected": -13.875, + "step": 1678 + }, + { + "epoch": 2.060122699386503, + "grad_norm": 0.5108998012960351, + "learning_rate": 4.7479338842975205e-07, + "logits/chosen": -0.1669921875, + "logits/rejected": -0.263671875, + "logps/chosen": -306.0, + "logps/rejected": -358.0, + "loss": 0.0019, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.69921875, + "rewards/margins": 12.1875, + "rewards/rejected": -12.875, + "step": 1679 + }, + { + "epoch": 2.061349693251534, + "grad_norm": 5.274929902335514, + "learning_rate": 4.741735537190083e-07, + "logits/chosen": -0.1630859375, + "logits/rejected": -0.30078125, + "logps/chosen": -386.0, + "logps/rejected": -386.0, + "loss": 0.0076, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.3828125, + "rewards/margins": 13.5, + "rewards/rejected": -13.0625, + "step": 1680 + }, + { + "epoch": 2.0625766871165645, + "grad_norm": 0.8115927920808855, + "learning_rate": 4.735537190082645e-07, + "logits/chosen": -0.28515625, + "logits/rejected": -0.447265625, + "logps/chosen": -406.0, + "logps/rejected": -438.0, + "loss": 0.0018, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.796875, + "rewards/margins": 14.3125, + "rewards/rejected": -13.5, + "step": 1681 + }, + { + "epoch": 2.063803680981595, + "grad_norm": 0.5982216569864724, + "learning_rate": 4.729338842975207e-07, + "logits/chosen": -0.26171875, + "logits/rejected": -0.400390625, + "logps/chosen": -406.0, + "logps/rejected": -414.0, + "loss": 0.0017, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.07666015625, + "rewards/margins": 13.5625, + "rewards/rejected": -13.4375, + "step": 1682 + }, + { + "epoch": 2.065030674846626, + "grad_norm": 1.320607883291853, + "learning_rate": 4.7231404958677683e-07, + "logits/chosen": -0.09130859375, + "logits/rejected": -0.279296875, + "logps/chosen": -412.0, + "logps/rejected": -388.0, + "loss": 0.003, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.341796875, + "rewards/margins": 12.6875, + "rewards/rejected": -13.0, + "step": 1683 + }, + { + "epoch": 2.0662576687116565, + "grad_norm": 3.0245591010331396, + "learning_rate": 4.716942148760331e-07, + "logits/chosen": -0.2333984375, + "logits/rejected": -0.3125, + "logps/chosen": -354.0, + "logps/rejected": -378.0, + "loss": 0.0059, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.7421875, + "rewards/margins": 12.125, + "rewards/rejected": -12.875, + "step": 1684 + }, + { + "epoch": 2.067484662576687, + "grad_norm": 2.8809996321511804, + "learning_rate": 4.7107438016528927e-07, + "logits/chosen": -0.034912109375, + "logits/rejected": -0.310546875, + "logps/chosen": -400.0, + "logps/rejected": -406.0, + "loss": 0.0059, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.46484375, + "rewards/margins": 12.5, + "rewards/rejected": -12.9375, + "step": 1685 + }, + { + "epoch": 2.068711656441718, + "grad_norm": 1.777848442091093, + "learning_rate": 4.704545454545455e-07, + "logits/chosen": -0.32421875, + "logits/rejected": -0.5078125, + "logps/chosen": -398.0, + "logps/rejected": -372.0, + "loss": 0.0028, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.37890625, + "rewards/margins": 12.9375, + "rewards/rejected": -13.3125, + "step": 1686 + }, + { + "epoch": 2.0699386503067485, + "grad_norm": 1.5798495879850163, + "learning_rate": 4.698347107438016e-07, + "logits/chosen": -0.09130859375, + "logits/rejected": -0.404296875, + "logps/chosen": -424.0, + "logps/rejected": -414.0, + "loss": 0.0033, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.271484375, + "rewards/margins": 14.25, + "rewards/rejected": -14.0, + "step": 1687 + }, + { + "epoch": 2.071165644171779, + "grad_norm": 0.8060147988496915, + "learning_rate": 4.6921487603305785e-07, + "logits/chosen": -0.13671875, + "logits/rejected": -0.30859375, + "logps/chosen": -386.0, + "logps/rejected": -422.0, + "loss": 0.0015, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.0654296875, + "rewards/margins": 13.625, + "rewards/rejected": -13.6875, + "step": 1688 + }, + { + "epoch": 2.07239263803681, + "grad_norm": 3.974520476908473, + "learning_rate": 4.6859504132231405e-07, + "logits/chosen": -0.12890625, + "logits/rejected": -0.29296875, + "logps/chosen": -402.0, + "logps/rejected": -404.0, + "loss": 0.0067, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.22265625, + "rewards/margins": 13.5625, + "rewards/rejected": -13.8125, + "step": 1689 + }, + { + "epoch": 2.0736196319018405, + "grad_norm": 0.9018134285427851, + "learning_rate": 4.679752066115703e-07, + "logits/chosen": -0.271484375, + "logits/rejected": -0.3984375, + "logps/chosen": -390.0, + "logps/rejected": -430.0, + "loss": 0.002, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.86328125, + "rewards/margins": 13.5625, + "rewards/rejected": -14.4375, + "step": 1690 + }, + { + "epoch": 2.074846625766871, + "grad_norm": 2.3862075229590083, + "learning_rate": 4.6735537190082644e-07, + "logits/chosen": -0.1728515625, + "logits/rejected": -0.400390625, + "logps/chosen": -368.0, + "logps/rejected": -392.0, + "loss": 0.0033, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.0177001953125, + "rewards/margins": 14.3125, + "rewards/rejected": -14.375, + "step": 1691 + }, + { + "epoch": 2.076073619631902, + "grad_norm": 2.6937495306047454, + "learning_rate": 4.6673553719008263e-07, + "logits/chosen": -0.2451171875, + "logits/rejected": -0.376953125, + "logps/chosen": -414.0, + "logps/rejected": -458.0, + "loss": 0.0083, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.1484375, + "rewards/margins": 13.875, + "rewards/rejected": -14.0, + "step": 1692 + }, + { + "epoch": 2.0773006134969325, + "grad_norm": 0.3743121793546649, + "learning_rate": 4.661157024793389e-07, + "logits/chosen": -0.2734375, + "logits/rejected": -0.390625, + "logps/chosen": -404.0, + "logps/rejected": -468.0, + "loss": 0.0009, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.61328125, + "rewards/margins": 14.875, + "rewards/rejected": -15.5, + "step": 1693 + }, + { + "epoch": 2.078527607361963, + "grad_norm": 0.6887982974666251, + "learning_rate": 4.6549586776859507e-07, + "logits/chosen": -0.322265625, + "logits/rejected": -0.494140625, + "logps/chosen": -410.0, + "logps/rejected": -452.0, + "loss": 0.007, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.4140625, + "rewards/margins": 13.5, + "rewards/rejected": -13.9375, + "step": 1694 + }, + { + "epoch": 2.079754601226994, + "grad_norm": 1.001290542863563, + "learning_rate": 4.6487603305785127e-07, + "logits/chosen": -0.2177734375, + "logits/rejected": -0.419921875, + "logps/chosen": -370.0, + "logps/rejected": -428.0, + "loss": 0.0079, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.06005859375, + "rewards/margins": 14.625, + "rewards/rejected": -14.5625, + "step": 1695 + }, + { + "epoch": 2.0809815950920245, + "grad_norm": 0.5867814747813861, + "learning_rate": 4.642561983471074e-07, + "logits/chosen": -0.33203125, + "logits/rejected": -0.5546875, + "logps/chosen": -400.0, + "logps/rejected": -442.0, + "loss": 0.0016, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.578125, + "rewards/margins": 15.1875, + "rewards/rejected": -15.75, + "step": 1696 + }, + { + "epoch": 2.082208588957055, + "grad_norm": 2.9558174051344626, + "learning_rate": 4.6363636363636365e-07, + "logits/chosen": -0.232421875, + "logits/rejected": -0.423828125, + "logps/chosen": -382.0, + "logps/rejected": -436.0, + "loss": 0.0045, + "rewards/accuracies": 1.0, + "rewards/chosen": -1.1328125, + "rewards/margins": 13.4375, + "rewards/rejected": -14.5625, + "step": 1697 + }, + { + "epoch": 2.083435582822086, + "grad_norm": 3.3154080606006016, + "learning_rate": 4.6301652892561985e-07, + "logits/chosen": -0.25390625, + "logits/rejected": -0.33984375, + "logps/chosen": -394.0, + "logps/rejected": -448.0, + "loss": 0.0043, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.1953125, + "rewards/margins": 14.5625, + "rewards/rejected": -14.375, + "step": 1698 + }, + { + "epoch": 2.0846625766871165, + "grad_norm": 0.7109584737737233, + "learning_rate": 4.623966942148761e-07, + "logits/chosen": -0.1923828125, + "logits/rejected": -0.357421875, + "logps/chosen": -470.0, + "logps/rejected": -444.0, + "loss": 0.0018, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.007354736328125, + "rewards/margins": 14.4375, + "rewards/rejected": -14.4375, + "step": 1699 + }, + { + "epoch": 2.085889570552147, + "grad_norm": 1.0188007782115722, + "learning_rate": 4.617768595041322e-07, + "logits/chosen": -0.0859375, + "logits/rejected": -0.232421875, + "logps/chosen": -416.0, + "logps/rejected": -406.0, + "loss": 0.003, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.008056640625, + "rewards/margins": 13.5, + "rewards/rejected": -13.5, + "step": 1700 + }, + { + "epoch": 2.087116564417178, + "grad_norm": 1.7938810148902795, + "learning_rate": 4.6115702479338843e-07, + "logits/chosen": -0.1298828125, + "logits/rejected": -0.306640625, + "logps/chosen": -390.0, + "logps/rejected": -432.0, + "loss": 0.0029, + "rewards/accuracies": 1.0, + "rewards/chosen": -1.3359375, + "rewards/margins": 13.1875, + "rewards/rejected": -14.5, + "step": 1701 + }, + { + "epoch": 2.0883435582822085, + "grad_norm": 0.551743879281472, + "learning_rate": 4.605371900826446e-07, + "logits/chosen": -0.201171875, + "logits/rejected": -0.4140625, + "logps/chosen": -438.0, + "logps/rejected": -440.0, + "loss": 0.0069, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.07080078125, + "rewards/margins": 15.625, + "rewards/rejected": -15.6875, + "step": 1702 + }, + { + "epoch": 2.089570552147239, + "grad_norm": 0.9878076667908823, + "learning_rate": 4.5991735537190087e-07, + "logits/chosen": -0.181640625, + "logits/rejected": -0.431640625, + "logps/chosen": -404.0, + "logps/rejected": -434.0, + "loss": 0.0025, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.291015625, + "rewards/margins": 14.8125, + "rewards/rejected": -15.125, + "step": 1703 + }, + { + "epoch": 2.09079754601227, + "grad_norm": 1.1670687882296444, + "learning_rate": 4.5929752066115707e-07, + "logits/chosen": -0.203125, + "logits/rejected": -0.44140625, + "logps/chosen": -466.0, + "logps/rejected": -462.0, + "loss": 0.002, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.671875, + "rewards/margins": 16.625, + "rewards/rejected": -15.9375, + "step": 1704 + }, + { + "epoch": 2.0920245398773005, + "grad_norm": 1.519643759736808, + "learning_rate": 4.586776859504132e-07, + "logits/chosen": -0.2490234375, + "logits/rejected": -0.40234375, + "logps/chosen": -384.0, + "logps/rejected": -394.0, + "loss": 0.0035, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.13671875, + "rewards/margins": 14.25, + "rewards/rejected": -14.375, + "step": 1705 + }, + { + "epoch": 2.093251533742331, + "grad_norm": 0.9738983109009536, + "learning_rate": 4.580578512396694e-07, + "logits/chosen": -0.287109375, + "logits/rejected": -0.5234375, + "logps/chosen": -436.0, + "logps/rejected": -424.0, + "loss": 0.002, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.037353515625, + "rewards/margins": 14.3125, + "rewards/rejected": -14.25, + "step": 1706 + }, + { + "epoch": 2.094478527607362, + "grad_norm": 1.1334725923361157, + "learning_rate": 4.5743801652892565e-07, + "logits/chosen": -0.2578125, + "logits/rejected": -0.416015625, + "logps/chosen": -358.0, + "logps/rejected": -404.0, + "loss": 0.0027, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.546875, + "rewards/margins": 13.5, + "rewards/rejected": -14.0625, + "step": 1707 + }, + { + "epoch": 2.0957055214723925, + "grad_norm": 0.4564424724274855, + "learning_rate": 4.5681818181818184e-07, + "logits/chosen": -0.3046875, + "logits/rejected": -0.5078125, + "logps/chosen": -404.0, + "logps/rejected": -418.0, + "loss": 0.0009, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.212890625, + "rewards/margins": 14.75, + "rewards/rejected": -14.9375, + "step": 1708 + }, + { + "epoch": 2.096932515337423, + "grad_norm": 1.0291369976636726, + "learning_rate": 4.56198347107438e-07, + "logits/chosen": -0.23046875, + "logits/rejected": -0.36328125, + "logps/chosen": -404.0, + "logps/rejected": -454.0, + "loss": 0.0037, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.81640625, + "rewards/margins": 13.25, + "rewards/rejected": -14.0625, + "step": 1709 + }, + { + "epoch": 2.098159509202454, + "grad_norm": 0.8980956611520831, + "learning_rate": 4.5557851239669423e-07, + "logits/chosen": -0.33984375, + "logits/rejected": -0.55078125, + "logps/chosen": -478.0, + "logps/rejected": -430.0, + "loss": 0.0029, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.248046875, + "rewards/margins": 14.625, + "rewards/rejected": -14.875, + "step": 1710 + }, + { + "epoch": 2.0993865030674845, + "grad_norm": 3.1416199705405456, + "learning_rate": 4.549586776859504e-07, + "logits/chosen": -0.271484375, + "logits/rejected": -0.35546875, + "logps/chosen": -408.0, + "logps/rejected": -460.0, + "loss": 0.0048, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.431640625, + "rewards/margins": 13.9375, + "rewards/rejected": -14.375, + "step": 1711 + }, + { + "epoch": 2.100613496932515, + "grad_norm": 0.4916465112202383, + "learning_rate": 4.543388429752066e-07, + "logits/chosen": -0.2578125, + "logits/rejected": -0.453125, + "logps/chosen": -406.0, + "logps/rejected": -422.0, + "loss": 0.0015, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.69140625, + "rewards/margins": 13.8125, + "rewards/rejected": -14.5, + "step": 1712 + }, + { + "epoch": 2.101840490797546, + "grad_norm": 2.335447389732648, + "learning_rate": 4.5371900826446287e-07, + "logits/chosen": -0.27734375, + "logits/rejected": -0.48046875, + "logps/chosen": -386.0, + "logps/rejected": -432.0, + "loss": 0.0143, + "rewards/accuracies": 0.984375, + "rewards/chosen": -0.271484375, + "rewards/margins": 14.375, + "rewards/rejected": -14.625, + "step": 1713 + }, + { + "epoch": 2.1030674846625765, + "grad_norm": 0.4916412570161682, + "learning_rate": 4.53099173553719e-07, + "logits/chosen": -0.1767578125, + "logits/rejected": -0.396484375, + "logps/chosen": -374.0, + "logps/rejected": -378.0, + "loss": 0.0014, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.04443359375, + "rewards/margins": 13.8125, + "rewards/rejected": -13.8125, + "step": 1714 + }, + { + "epoch": 2.104294478527607, + "grad_norm": 2.0479433042327093, + "learning_rate": 4.524793388429752e-07, + "logits/chosen": -0.2119140625, + "logits/rejected": -0.419921875, + "logps/chosen": -386.0, + "logps/rejected": -424.0, + "loss": 0.003, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.546875, + "rewards/margins": 14.0, + "rewards/rejected": -14.5625, + "step": 1715 + }, + { + "epoch": 2.105521472392638, + "grad_norm": 1.873701093866083, + "learning_rate": 4.5185950413223145e-07, + "logits/chosen": -0.1396484375, + "logits/rejected": -0.328125, + "logps/chosen": -390.0, + "logps/rejected": -424.0, + "loss": 0.004, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.2412109375, + "rewards/margins": 13.5625, + "rewards/rejected": -13.8125, + "step": 1716 + }, + { + "epoch": 2.1067484662576685, + "grad_norm": 0.8111965799519333, + "learning_rate": 4.5123966942148764e-07, + "logits/chosen": -0.25390625, + "logits/rejected": -0.451171875, + "logps/chosen": -400.0, + "logps/rejected": -420.0, + "loss": 0.002, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.453125, + "rewards/margins": 14.4375, + "rewards/rejected": -14.9375, + "step": 1717 + }, + { + "epoch": 2.107975460122699, + "grad_norm": 0.8447380889380348, + "learning_rate": 4.506198347107438e-07, + "logits/chosen": -0.1259765625, + "logits/rejected": -0.30859375, + "logps/chosen": -382.0, + "logps/rejected": -412.0, + "loss": 0.0076, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.421875, + "rewards/margins": 13.9375, + "rewards/rejected": -14.375, + "step": 1718 + }, + { + "epoch": 2.10920245398773, + "grad_norm": 0.4604925939049009, + "learning_rate": 4.5e-07, + "logits/chosen": -0.1806640625, + "logits/rejected": -0.380859375, + "logps/chosen": -444.0, + "logps/rejected": -452.0, + "loss": 0.012, + "rewards/accuracies": 0.984375, + "rewards/chosen": 0.8828125, + "rewards/margins": 16.125, + "rewards/rejected": -15.1875, + "step": 1719 + }, + { + "epoch": 2.1104294478527605, + "grad_norm": 0.9684416944064264, + "learning_rate": 4.4938016528925623e-07, + "logits/chosen": -0.193359375, + "logits/rejected": -0.32421875, + "logps/chosen": -344.0, + "logps/rejected": -382.0, + "loss": 0.0026, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.77734375, + "rewards/margins": 13.25, + "rewards/rejected": -14.0, + "step": 1720 + }, + { + "epoch": 2.111656441717791, + "grad_norm": 0.8760601710673682, + "learning_rate": 4.487603305785124e-07, + "logits/chosen": -0.1923828125, + "logits/rejected": -0.451171875, + "logps/chosen": -440.0, + "logps/rejected": -466.0, + "loss": 0.002, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.047119140625, + "rewards/margins": 16.0, + "rewards/rejected": -15.9375, + "step": 1721 + }, + { + "epoch": 2.112883435582822, + "grad_norm": 2.185469603546492, + "learning_rate": 4.4814049586776867e-07, + "logits/chosen": -0.197265625, + "logits/rejected": -0.302734375, + "logps/chosen": -374.0, + "logps/rejected": -416.0, + "loss": 0.0068, + "rewards/accuracies": 1.0, + "rewards/chosen": -1.4609375, + "rewards/margins": 13.8125, + "rewards/rejected": -15.25, + "step": 1722 + }, + { + "epoch": 2.1141104294478525, + "grad_norm": 0.7963254119108496, + "learning_rate": 4.4752066115702476e-07, + "logits/chosen": -0.2392578125, + "logits/rejected": -0.451171875, + "logps/chosen": -416.0, + "logps/rejected": -422.0, + "loss": 0.0017, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.80078125, + "rewards/margins": 13.4375, + "rewards/rejected": -14.25, + "step": 1723 + }, + { + "epoch": 2.1153374233128837, + "grad_norm": 2.0844646263561613, + "learning_rate": 4.46900826446281e-07, + "logits/chosen": -0.2734375, + "logits/rejected": -0.412109375, + "logps/chosen": -350.0, + "logps/rejected": -386.0, + "loss": 0.0032, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.1796875, + "rewards/margins": 13.9375, + "rewards/rejected": -14.125, + "step": 1724 + }, + { + "epoch": 2.116564417177914, + "grad_norm": 1.688705032270998, + "learning_rate": 4.462809917355372e-07, + "logits/chosen": -0.255859375, + "logits/rejected": -0.43359375, + "logps/chosen": -402.0, + "logps/rejected": -452.0, + "loss": 0.0026, + "rewards/accuracies": 1.0, + "rewards/chosen": -1.03125, + "rewards/margins": 14.25, + "rewards/rejected": -15.25, + "step": 1725 + }, + { + "epoch": 2.117791411042945, + "grad_norm": 1.045438531237962, + "learning_rate": 4.4566115702479345e-07, + "logits/chosen": -0.357421875, + "logits/rejected": -0.53125, + "logps/chosen": -440.0, + "logps/rejected": -472.0, + "loss": 0.0075, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.0966796875, + "rewards/margins": 15.75, + "rewards/rejected": -15.8125, + "step": 1726 + }, + { + "epoch": 2.1190184049079757, + "grad_norm": 1.0869974716503041, + "learning_rate": 4.450413223140496e-07, + "logits/chosen": -0.2255859375, + "logits/rejected": -0.314453125, + "logps/chosen": -428.0, + "logps/rejected": -442.0, + "loss": 0.0016, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.578125, + "rewards/margins": 16.25, + "rewards/rejected": -15.6875, + "step": 1727 + }, + { + "epoch": 2.1202453987730063, + "grad_norm": 1.565532034994413, + "learning_rate": 4.444214876033058e-07, + "logits/chosen": -0.1748046875, + "logits/rejected": -0.353515625, + "logps/chosen": -440.0, + "logps/rejected": -416.0, + "loss": 0.0031, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.4453125, + "rewards/margins": 14.5, + "rewards/rejected": -14.9375, + "step": 1728 + }, + { + "epoch": 2.121472392638037, + "grad_norm": 0.8565888027932397, + "learning_rate": 4.43801652892562e-07, + "logits/chosen": -0.040771484375, + "logits/rejected": -0.2255859375, + "logps/chosen": -386.0, + "logps/rejected": -408.0, + "loss": 0.0017, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.96875, + "rewards/margins": 13.625, + "rewards/rejected": -14.5625, + "step": 1729 + }, + { + "epoch": 2.1226993865030677, + "grad_norm": 1.8224165245692583, + "learning_rate": 4.431818181818182e-07, + "logits/chosen": -0.234375, + "logits/rejected": -0.34765625, + "logps/chosen": -392.0, + "logps/rejected": -434.0, + "loss": 0.0038, + "rewards/accuracies": 1.0, + "rewards/chosen": -1.421875, + "rewards/margins": 13.5625, + "rewards/rejected": -15.0, + "step": 1730 + }, + { + "epoch": 2.1239263803680983, + "grad_norm": 1.3192398501715272, + "learning_rate": 4.425619834710744e-07, + "logits/chosen": -0.263671875, + "logits/rejected": -0.439453125, + "logps/chosen": -372.0, + "logps/rejected": -418.0, + "loss": 0.0021, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.6015625, + "rewards/margins": 14.3125, + "rewards/rejected": -14.875, + "step": 1731 + }, + { + "epoch": 2.125153374233129, + "grad_norm": 0.32249951938537536, + "learning_rate": 4.4194214876033056e-07, + "logits/chosen": -0.1865234375, + "logits/rejected": -0.330078125, + "logps/chosen": -422.0, + "logps/rejected": -458.0, + "loss": 0.0009, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.1728515625, + "rewards/margins": 15.5, + "rewards/rejected": -15.3125, + "step": 1732 + }, + { + "epoch": 2.1263803680981597, + "grad_norm": 1.9304157394045243, + "learning_rate": 4.413223140495868e-07, + "logits/chosen": -0.12109375, + "logits/rejected": -0.337890625, + "logps/chosen": -426.0, + "logps/rejected": -428.0, + "loss": 0.0042, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.205078125, + "rewards/margins": 14.6875, + "rewards/rejected": -14.4375, + "step": 1733 + }, + { + "epoch": 2.1276073619631903, + "grad_norm": 0.5358887722872807, + "learning_rate": 4.40702479338843e-07, + "logits/chosen": -0.2490234375, + "logits/rejected": -0.4375, + "logps/chosen": -432.0, + "logps/rejected": -460.0, + "loss": 0.0014, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.419921875, + "rewards/margins": 15.375, + "rewards/rejected": -15.75, + "step": 1734 + }, + { + "epoch": 2.128834355828221, + "grad_norm": 1.4925488730788283, + "learning_rate": 4.400826446280992e-07, + "logits/chosen": -0.1708984375, + "logits/rejected": -0.314453125, + "logps/chosen": -398.0, + "logps/rejected": -408.0, + "loss": 0.0031, + "rewards/accuracies": 1.0, + "rewards/chosen": -1.09375, + "rewards/margins": 13.4375, + "rewards/rejected": -14.5625, + "step": 1735 + }, + { + "epoch": 2.1300613496932517, + "grad_norm": 7.7862545748599485, + "learning_rate": 4.3946280991735533e-07, + "logits/chosen": -0.1796875, + "logits/rejected": -0.32421875, + "logps/chosen": -428.0, + "logps/rejected": -448.0, + "loss": 0.0407, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -1.0078125, + "rewards/margins": 13.375, + "rewards/rejected": -14.375, + "step": 1736 + }, + { + "epoch": 2.1312883435582823, + "grad_norm": 1.9585076033026356, + "learning_rate": 4.388429752066116e-07, + "logits/chosen": -0.1943359375, + "logits/rejected": -0.353515625, + "logps/chosen": -402.0, + "logps/rejected": -434.0, + "loss": 0.0041, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.59765625, + "rewards/margins": 13.5625, + "rewards/rejected": -14.125, + "step": 1737 + }, + { + "epoch": 2.132515337423313, + "grad_norm": 0.551276787198678, + "learning_rate": 4.382231404958678e-07, + "logits/chosen": -0.19921875, + "logits/rejected": -0.40234375, + "logps/chosen": -402.0, + "logps/rejected": -466.0, + "loss": 0.0067, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.5703125, + "rewards/margins": 15.1875, + "rewards/rejected": -15.75, + "step": 1738 + }, + { + "epoch": 2.1337423312883437, + "grad_norm": 1.7253072285581432, + "learning_rate": 4.37603305785124e-07, + "logits/chosen": -0.2333984375, + "logits/rejected": -0.333984375, + "logps/chosen": -384.0, + "logps/rejected": -458.0, + "loss": 0.0053, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.6796875, + "rewards/margins": 13.875, + "rewards/rejected": -14.5625, + "step": 1739 + }, + { + "epoch": 2.1349693251533743, + "grad_norm": 1.2872132990664973, + "learning_rate": 4.369834710743801e-07, + "logits/chosen": -0.1943359375, + "logits/rejected": -0.30078125, + "logps/chosen": -358.0, + "logps/rejected": -408.0, + "loss": 0.0081, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.412109375, + "rewards/margins": 13.9375, + "rewards/rejected": -14.375, + "step": 1740 + }, + { + "epoch": 2.136196319018405, + "grad_norm": 3.2998898284673777, + "learning_rate": 4.3636363636363636e-07, + "logits/chosen": -0.2421875, + "logits/rejected": -0.30859375, + "logps/chosen": -368.0, + "logps/rejected": -430.0, + "loss": 0.0045, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.546875, + "rewards/margins": 14.125, + "rewards/rejected": -14.625, + "step": 1741 + }, + { + "epoch": 2.1374233128834357, + "grad_norm": 1.0516563953062639, + "learning_rate": 4.3574380165289255e-07, + "logits/chosen": -0.240234375, + "logits/rejected": -0.388671875, + "logps/chosen": -408.0, + "logps/rejected": -378.0, + "loss": 0.002, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.46875, + "rewards/margins": 14.0, + "rewards/rejected": -14.5, + "step": 1742 + }, + { + "epoch": 2.1386503067484663, + "grad_norm": 0.9262983256007776, + "learning_rate": 4.351239669421488e-07, + "logits/chosen": -0.2392578125, + "logits/rejected": -0.4296875, + "logps/chosen": -400.0, + "logps/rejected": -424.0, + "loss": 0.0075, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.1279296875, + "rewards/margins": 14.1875, + "rewards/rejected": -14.0625, + "step": 1743 + }, + { + "epoch": 2.139877300613497, + "grad_norm": 0.4092707461888209, + "learning_rate": 4.34504132231405e-07, + "logits/chosen": -0.3359375, + "logits/rejected": -0.5859375, + "logps/chosen": -402.0, + "logps/rejected": -442.0, + "loss": 0.0009, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.67578125, + "rewards/margins": 14.75, + "rewards/rejected": -15.4375, + "step": 1744 + }, + { + "epoch": 2.1411042944785277, + "grad_norm": 2.9298647198606305, + "learning_rate": 4.3388429752066114e-07, + "logits/chosen": -0.271484375, + "logits/rejected": -0.46484375, + "logps/chosen": -430.0, + "logps/rejected": -452.0, + "loss": 0.0031, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.47265625, + "rewards/margins": 14.75, + "rewards/rejected": -15.25, + "step": 1745 + }, + { + "epoch": 2.1423312883435583, + "grad_norm": 0.9055696138333951, + "learning_rate": 4.3326446280991733e-07, + "logits/chosen": -0.1279296875, + "logits/rejected": -0.294921875, + "logps/chosen": -442.0, + "logps/rejected": -418.0, + "loss": 0.0072, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.11767578125, + "rewards/margins": 13.6875, + "rewards/rejected": -13.5625, + "step": 1746 + }, + { + "epoch": 2.143558282208589, + "grad_norm": 0.9706771460706346, + "learning_rate": 4.326446280991736e-07, + "logits/chosen": -0.1943359375, + "logits/rejected": -0.345703125, + "logps/chosen": -406.0, + "logps/rejected": -432.0, + "loss": 0.0017, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.47265625, + "rewards/margins": 14.8125, + "rewards/rejected": -15.25, + "step": 1747 + }, + { + "epoch": 2.1447852760736197, + "grad_norm": 2.6379068584485657, + "learning_rate": 4.3202479338842977e-07, + "logits/chosen": -0.26171875, + "logits/rejected": -0.49609375, + "logps/chosen": -432.0, + "logps/rejected": -428.0, + "loss": 0.0053, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.000370025634765625, + "rewards/margins": 14.125, + "rewards/rejected": -14.125, + "step": 1748 + }, + { + "epoch": 2.1460122699386504, + "grad_norm": 3.407143614439923, + "learning_rate": 4.314049586776859e-07, + "logits/chosen": -0.0830078125, + "logits/rejected": -0.34375, + "logps/chosen": -428.0, + "logps/rejected": -446.0, + "loss": 0.0047, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.09619140625, + "rewards/margins": 13.75, + "rewards/rejected": -13.625, + "step": 1749 + }, + { + "epoch": 2.147239263803681, + "grad_norm": 0.9927053668266275, + "learning_rate": 4.3078512396694216e-07, + "logits/chosen": -0.2578125, + "logits/rejected": -0.474609375, + "logps/chosen": -414.0, + "logps/rejected": -414.0, + "loss": 0.0019, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.9296875, + "rewards/margins": 13.75, + "rewards/rejected": -14.6875, + "step": 1750 + }, + { + "epoch": 2.1484662576687117, + "grad_norm": 4.932208691717357, + "learning_rate": 4.3016528925619835e-07, + "logits/chosen": -0.12890625, + "logits/rejected": -0.33203125, + "logps/chosen": -324.0, + "logps/rejected": -382.0, + "loss": 0.007, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.5390625, + "rewards/margins": 13.125, + "rewards/rejected": -13.625, + "step": 1751 + }, + { + "epoch": 2.1496932515337424, + "grad_norm": 1.3960199135945588, + "learning_rate": 4.295454545454546e-07, + "logits/chosen": -0.1455078125, + "logits/rejected": -0.349609375, + "logps/chosen": -458.0, + "logps/rejected": -438.0, + "loss": 0.0022, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.5078125, + "rewards/margins": 13.875, + "rewards/rejected": -14.4375, + "step": 1752 + }, + { + "epoch": 2.150920245398773, + "grad_norm": 2.8443586010916175, + "learning_rate": 4.289256198347108e-07, + "logits/chosen": -0.1650390625, + "logits/rejected": -0.35546875, + "logps/chosen": -400.0, + "logps/rejected": -402.0, + "loss": 0.0033, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.4296875, + "rewards/margins": 13.9375, + "rewards/rejected": -13.5, + "step": 1753 + }, + { + "epoch": 2.1521472392638037, + "grad_norm": 3.9576243606153954, + "learning_rate": 4.2830578512396694e-07, + "logits/chosen": -0.1435546875, + "logits/rejected": -0.328125, + "logps/chosen": -394.0, + "logps/rejected": -400.0, + "loss": 0.0046, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.79296875, + "rewards/margins": 13.125, + "rewards/rejected": -13.9375, + "step": 1754 + }, + { + "epoch": 2.1533742331288344, + "grad_norm": 3.623358375007302, + "learning_rate": 4.2768595041322313e-07, + "logits/chosen": -0.1103515625, + "logits/rejected": -0.318359375, + "logps/chosen": -378.0, + "logps/rejected": -402.0, + "loss": 0.007, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.140625, + "rewards/margins": 13.5625, + "rewards/rejected": -13.4375, + "step": 1755 + }, + { + "epoch": 2.154601226993865, + "grad_norm": 0.9128310199193256, + "learning_rate": 4.270661157024794e-07, + "logits/chosen": -0.14453125, + "logits/rejected": -0.3515625, + "logps/chosen": -428.0, + "logps/rejected": -448.0, + "loss": 0.002, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.0966796875, + "rewards/margins": 13.75, + "rewards/rejected": -13.6875, + "step": 1756 + }, + { + "epoch": 2.1558282208588957, + "grad_norm": 1.120555541800032, + "learning_rate": 4.2644628099173557e-07, + "logits/chosen": -0.18359375, + "logits/rejected": -0.31640625, + "logps/chosen": -380.0, + "logps/rejected": -428.0, + "loss": 0.0027, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.333984375, + "rewards/margins": 12.9375, + "rewards/rejected": -13.25, + "step": 1757 + }, + { + "epoch": 2.1570552147239264, + "grad_norm": 1.276172717089728, + "learning_rate": 4.258264462809917e-07, + "logits/chosen": -0.185546875, + "logits/rejected": -0.2421875, + "logps/chosen": -334.0, + "logps/rejected": -392.0, + "loss": 0.0042, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.578125, + "rewards/margins": 13.25, + "rewards/rejected": -13.8125, + "step": 1758 + }, + { + "epoch": 2.158282208588957, + "grad_norm": 1.8439077969289037, + "learning_rate": 4.252066115702479e-07, + "logits/chosen": -0.357421875, + "logits/rejected": -0.53125, + "logps/chosen": -384.0, + "logps/rejected": -420.0, + "loss": 0.0041, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.027587890625, + "rewards/margins": 14.3125, + "rewards/rejected": -14.3125, + "step": 1759 + }, + { + "epoch": 2.1595092024539877, + "grad_norm": 0.4759080122609661, + "learning_rate": 4.2458677685950415e-07, + "logits/chosen": -0.2109375, + "logits/rejected": -0.453125, + "logps/chosen": -432.0, + "logps/rejected": -426.0, + "loss": 0.001, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.77734375, + "rewards/margins": 14.8125, + "rewards/rejected": -14.0, + "step": 1760 + }, + { + "epoch": 2.1607361963190184, + "grad_norm": 3.770504079662017, + "learning_rate": 4.2396694214876035e-07, + "logits/chosen": -0.236328125, + "logits/rejected": -0.43359375, + "logps/chosen": -400.0, + "logps/rejected": -452.0, + "loss": 0.0063, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.287109375, + "rewards/margins": 14.9375, + "rewards/rejected": -14.6875, + "step": 1761 + }, + { + "epoch": 2.161963190184049, + "grad_norm": 5.98876557005412, + "learning_rate": 4.233471074380166e-07, + "logits/chosen": -0.31640625, + "logits/rejected": -0.41015625, + "logps/chosen": -372.0, + "logps/rejected": -422.0, + "loss": 0.0075, + "rewards/accuracies": 1.0, + "rewards/chosen": -1.15625, + "rewards/margins": 12.9375, + "rewards/rejected": -14.0625, + "step": 1762 + }, + { + "epoch": 2.1631901840490797, + "grad_norm": 10.937814739385605, + "learning_rate": 4.2272727272727274e-07, + "logits/chosen": -0.2421875, + "logits/rejected": -0.44140625, + "logps/chosen": -436.0, + "logps/rejected": -448.0, + "loss": 0.0163, + "rewards/accuracies": 0.984375, + "rewards/chosen": -1.34375, + "rewards/margins": 13.4375, + "rewards/rejected": -14.8125, + "step": 1763 + }, + { + "epoch": 2.1644171779141104, + "grad_norm": 2.8142648655990663, + "learning_rate": 4.2210743801652893e-07, + "logits/chosen": -0.1474609375, + "logits/rejected": -0.28515625, + "logps/chosen": -406.0, + "logps/rejected": -408.0, + "loss": 0.0201, + "rewards/accuracies": 0.9765625, + "rewards/chosen": -0.30859375, + "rewards/margins": 13.0, + "rewards/rejected": -13.3125, + "step": 1764 + }, + { + "epoch": 2.165644171779141, + "grad_norm": 0.6932972658815392, + "learning_rate": 4.214876033057851e-07, + "logits/chosen": -0.380859375, + "logits/rejected": -0.59375, + "logps/chosen": -460.0, + "logps/rejected": -446.0, + "loss": 0.0067, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.73046875, + "rewards/margins": 15.3125, + "rewards/rejected": -14.5625, + "step": 1765 + }, + { + "epoch": 2.1668711656441717, + "grad_norm": 1.533856732837065, + "learning_rate": 4.2086776859504137e-07, + "logits/chosen": -0.08544921875, + "logits/rejected": -0.291015625, + "logps/chosen": -348.0, + "logps/rejected": -388.0, + "loss": 0.003, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.2421875, + "rewards/margins": 12.8125, + "rewards/rejected": -13.0625, + "step": 1766 + }, + { + "epoch": 2.1680981595092024, + "grad_norm": 0.4975377708011512, + "learning_rate": 4.202479338842975e-07, + "logits/chosen": -0.10888671875, + "logits/rejected": -0.22265625, + "logps/chosen": -396.0, + "logps/rejected": -404.0, + "loss": 0.0011, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.31640625, + "rewards/margins": 14.1875, + "rewards/rejected": -13.875, + "step": 1767 + }, + { + "epoch": 2.169325153374233, + "grad_norm": 1.321225143064983, + "learning_rate": 4.196280991735537e-07, + "logits/chosen": -0.1396484375, + "logits/rejected": -0.32421875, + "logps/chosen": -356.0, + "logps/rejected": -376.0, + "loss": 0.0025, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.2265625, + "rewards/margins": 13.625, + "rewards/rejected": -13.8125, + "step": 1768 + }, + { + "epoch": 2.1705521472392637, + "grad_norm": 1.9948005657442378, + "learning_rate": 4.1900826446280996e-07, + "logits/chosen": -0.201171875, + "logits/rejected": -0.43359375, + "logps/chosen": -396.0, + "logps/rejected": -434.0, + "loss": 0.0036, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.458984375, + "rewards/margins": 14.125, + "rewards/rejected": -14.625, + "step": 1769 + }, + { + "epoch": 2.1717791411042944, + "grad_norm": 1.2093570691567386, + "learning_rate": 4.1838842975206615e-07, + "logits/chosen": -0.40234375, + "logits/rejected": -0.53515625, + "logps/chosen": -390.0, + "logps/rejected": -414.0, + "loss": 0.0028, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.58203125, + "rewards/margins": 13.875, + "rewards/rejected": -14.4375, + "step": 1770 + }, + { + "epoch": 2.173006134969325, + "grad_norm": 0.8558480758921507, + "learning_rate": 4.1776859504132234e-07, + "logits/chosen": -0.08544921875, + "logits/rejected": -0.296875, + "logps/chosen": -390.0, + "logps/rejected": -420.0, + "loss": 0.0017, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.34375, + "rewards/margins": 14.0, + "rewards/rejected": -14.375, + "step": 1771 + }, + { + "epoch": 2.1742331288343557, + "grad_norm": 0.6342313939000331, + "learning_rate": 4.171487603305785e-07, + "logits/chosen": -0.115234375, + "logits/rejected": -0.1943359375, + "logps/chosen": -366.0, + "logps/rejected": -398.0, + "loss": 0.0013, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.10302734375, + "rewards/margins": 13.75, + "rewards/rejected": -13.8125, + "step": 1772 + }, + { + "epoch": 2.1754601226993864, + "grad_norm": 1.2547150146341541, + "learning_rate": 4.1652892561983473e-07, + "logits/chosen": -0.189453125, + "logits/rejected": -0.326171875, + "logps/chosen": -384.0, + "logps/rejected": -436.0, + "loss": 0.0028, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.04052734375, + "rewards/margins": 13.875, + "rewards/rejected": -13.9375, + "step": 1773 + }, + { + "epoch": 2.176687116564417, + "grad_norm": 1.7171893074863533, + "learning_rate": 4.159090909090909e-07, + "logits/chosen": -0.134765625, + "logits/rejected": -0.2294921875, + "logps/chosen": -388.0, + "logps/rejected": -432.0, + "loss": 0.006, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.01458740234375, + "rewards/margins": 13.6875, + "rewards/rejected": -13.6875, + "step": 1774 + }, + { + "epoch": 2.1779141104294477, + "grad_norm": 3.4575516469313903, + "learning_rate": 4.1528925619834717e-07, + "logits/chosen": -0.248046875, + "logits/rejected": -0.333984375, + "logps/chosen": -408.0, + "logps/rejected": -416.0, + "loss": 0.0045, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.259765625, + "rewards/margins": 14.0, + "rewards/rejected": -14.25, + "step": 1775 + }, + { + "epoch": 2.1791411042944784, + "grad_norm": 2.068057201179283, + "learning_rate": 4.1466942148760326e-07, + "logits/chosen": -0.2236328125, + "logits/rejected": -0.423828125, + "logps/chosen": -412.0, + "logps/rejected": -436.0, + "loss": 0.0033, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.185546875, + "rewards/margins": 14.625, + "rewards/rejected": -14.4375, + "step": 1776 + }, + { + "epoch": 2.180368098159509, + "grad_norm": 0.8937570254770851, + "learning_rate": 4.140495867768595e-07, + "logits/chosen": -0.146484375, + "logits/rejected": -0.267578125, + "logps/chosen": -402.0, + "logps/rejected": -424.0, + "loss": 0.0019, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.1337890625, + "rewards/margins": 14.3125, + "rewards/rejected": -14.1875, + "step": 1777 + }, + { + "epoch": 2.1815950920245397, + "grad_norm": 1.3610233610561817, + "learning_rate": 4.134297520661157e-07, + "logits/chosen": -0.1591796875, + "logits/rejected": -0.419921875, + "logps/chosen": -412.0, + "logps/rejected": -404.0, + "loss": 0.0078, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.16015625, + "rewards/margins": 13.9375, + "rewards/rejected": -13.8125, + "step": 1778 + }, + { + "epoch": 2.1828220858895704, + "grad_norm": 1.1031603422660603, + "learning_rate": 4.1280991735537195e-07, + "logits/chosen": -0.255859375, + "logits/rejected": -0.474609375, + "logps/chosen": -446.0, + "logps/rejected": -428.0, + "loss": 0.0022, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.703125, + "rewards/margins": 15.1875, + "rewards/rejected": -14.5, + "step": 1779 + }, + { + "epoch": 2.184049079754601, + "grad_norm": 4.52582915184338, + "learning_rate": 4.121900826446281e-07, + "logits/chosen": -0.330078125, + "logits/rejected": -0.51953125, + "logps/chosen": -402.0, + "logps/rejected": -452.0, + "loss": 0.0124, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -1.328125, + "rewards/margins": 13.5, + "rewards/rejected": -14.8125, + "step": 1780 + }, + { + "epoch": 2.1852760736196317, + "grad_norm": 0.39317900071873063, + "learning_rate": 4.115702479338843e-07, + "logits/chosen": -0.314453125, + "logits/rejected": -0.5625, + "logps/chosen": -414.0, + "logps/rejected": -458.0, + "loss": 0.0065, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.27734375, + "rewards/margins": 15.0625, + "rewards/rejected": -14.8125, + "step": 1781 + }, + { + "epoch": 2.1865030674846624, + "grad_norm": 1.382307328776299, + "learning_rate": 4.109504132231405e-07, + "logits/chosen": -0.2890625, + "logits/rejected": -0.390625, + "logps/chosen": -368.0, + "logps/rejected": -430.0, + "loss": 0.0034, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.49609375, + "rewards/margins": 14.0625, + "rewards/rejected": -14.5625, + "step": 1782 + }, + { + "epoch": 2.187730061349693, + "grad_norm": 1.9660290325036802, + "learning_rate": 4.1033057851239673e-07, + "logits/chosen": -0.193359375, + "logits/rejected": -0.34375, + "logps/chosen": -364.0, + "logps/rejected": -406.0, + "loss": 0.003, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.427734375, + "rewards/margins": 13.375, + "rewards/rejected": -13.75, + "step": 1783 + }, + { + "epoch": 2.1889570552147237, + "grad_norm": 0.9782965659752122, + "learning_rate": 4.097107438016529e-07, + "logits/chosen": -0.4140625, + "logits/rejected": -0.6328125, + "logps/chosen": -420.0, + "logps/rejected": -416.0, + "loss": 0.0023, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.173828125, + "rewards/margins": 14.75, + "rewards/rejected": -14.9375, + "step": 1784 + }, + { + "epoch": 2.190184049079755, + "grad_norm": 4.496421603453417, + "learning_rate": 4.0909090909090906e-07, + "logits/chosen": -0.22265625, + "logits/rejected": -0.400390625, + "logps/chosen": -396.0, + "logps/rejected": -448.0, + "loss": 0.0045, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.90234375, + "rewards/margins": 14.125, + "rewards/rejected": -15.0625, + "step": 1785 + }, + { + "epoch": 2.191411042944785, + "grad_norm": 1.3648663691300442, + "learning_rate": 4.084710743801653e-07, + "logits/chosen": -0.208984375, + "logits/rejected": -0.33984375, + "logps/chosen": -366.0, + "logps/rejected": -394.0, + "loss": 0.0021, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.5859375, + "rewards/margins": 13.3125, + "rewards/rejected": -13.875, + "step": 1786 + }, + { + "epoch": 2.192638036809816, + "grad_norm": 1.9541719257929342, + "learning_rate": 4.078512396694215e-07, + "logits/chosen": -0.291015625, + "logits/rejected": -0.3671875, + "logps/chosen": -408.0, + "logps/rejected": -472.0, + "loss": 0.0141, + "rewards/accuracies": 0.984375, + "rewards/chosen": -0.37109375, + "rewards/margins": 13.625, + "rewards/rejected": -14.0, + "step": 1787 + }, + { + "epoch": 2.1938650306748464, + "grad_norm": 3.604340390957139, + "learning_rate": 4.072314049586777e-07, + "logits/chosen": -0.02587890625, + "logits/rejected": -0.322265625, + "logps/chosen": -408.0, + "logps/rejected": -412.0, + "loss": 0.005, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.5859375, + "rewards/margins": 13.5625, + "rewards/rejected": -14.125, + "step": 1788 + }, + { + "epoch": 2.1950920245398775, + "grad_norm": 1.792615808262299, + "learning_rate": 4.0661157024793384e-07, + "logits/chosen": -0.275390625, + "logits/rejected": -0.44921875, + "logps/chosen": -410.0, + "logps/rejected": -442.0, + "loss": 0.0031, + "rewards/accuracies": 1.0, + "rewards/chosen": -1.25, + "rewards/margins": 13.375, + "rewards/rejected": -14.625, + "step": 1789 + }, + { + "epoch": 2.196319018404908, + "grad_norm": 1.2057588198432694, + "learning_rate": 4.059917355371901e-07, + "logits/chosen": -0.09814453125, + "logits/rejected": -0.390625, + "logps/chosen": -420.0, + "logps/rejected": -406.0, + "loss": 0.0085, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.06396484375, + "rewards/margins": 14.0, + "rewards/rejected": -13.9375, + "step": 1790 + }, + { + "epoch": 2.197546012269939, + "grad_norm": 6.332707596009572, + "learning_rate": 4.053719008264463e-07, + "logits/chosen": -0.287109375, + "logits/rejected": -0.453125, + "logps/chosen": -374.0, + "logps/rejected": -428.0, + "loss": 0.0088, + "rewards/accuracies": 1.0, + "rewards/chosen": -1.390625, + "rewards/margins": 12.875, + "rewards/rejected": -14.3125, + "step": 1791 + }, + { + "epoch": 2.1987730061349695, + "grad_norm": 2.271912358968024, + "learning_rate": 4.0475206611570253e-07, + "logits/chosen": -0.2119140625, + "logits/rejected": -0.359375, + "logps/chosen": -372.0, + "logps/rejected": -422.0, + "loss": 0.0039, + "rewards/accuracies": 1.0, + "rewards/chosen": -1.6328125, + "rewards/margins": 12.875, + "rewards/rejected": -14.5625, + "step": 1792 + }, + { + "epoch": 2.2, + "grad_norm": 1.9456164525015973, + "learning_rate": 4.041322314049587e-07, + "logits/chosen": -0.28515625, + "logits/rejected": -0.361328125, + "logps/chosen": -360.0, + "logps/rejected": -426.0, + "loss": 0.0033, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.94140625, + "rewards/margins": 13.8125, + "rewards/rejected": -14.75, + "step": 1793 + }, + { + "epoch": 2.201226993865031, + "grad_norm": 0.5696120331539009, + "learning_rate": 4.0351239669421486e-07, + "logits/chosen": -0.2099609375, + "logits/rejected": -0.40625, + "logps/chosen": -424.0, + "logps/rejected": -444.0, + "loss": 0.0016, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.96484375, + "rewards/margins": 14.4375, + "rewards/rejected": -15.375, + "step": 1794 + }, + { + "epoch": 2.2024539877300615, + "grad_norm": 0.9454302859196569, + "learning_rate": 4.0289256198347106e-07, + "logits/chosen": -0.27734375, + "logits/rejected": -0.4140625, + "logps/chosen": -398.0, + "logps/rejected": -436.0, + "loss": 0.0071, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.66015625, + "rewards/margins": 13.875, + "rewards/rejected": -14.5, + "step": 1795 + }, + { + "epoch": 2.203680981595092, + "grad_norm": 1.0860843658066783, + "learning_rate": 4.022727272727273e-07, + "logits/chosen": -0.30078125, + "logits/rejected": -0.3984375, + "logps/chosen": -402.0, + "logps/rejected": -432.0, + "loss": 0.0029, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.062255859375, + "rewards/margins": 14.5625, + "rewards/rejected": -14.625, + "step": 1796 + }, + { + "epoch": 2.204907975460123, + "grad_norm": 1.9455110658309405, + "learning_rate": 4.016528925619835e-07, + "logits/chosen": -0.3046875, + "logits/rejected": -0.486328125, + "logps/chosen": -406.0, + "logps/rejected": -434.0, + "loss": 0.0026, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.58203125, + "rewards/margins": 14.375, + "rewards/rejected": -14.9375, + "step": 1797 + }, + { + "epoch": 2.2061349693251535, + "grad_norm": 2.768227784396514, + "learning_rate": 4.0103305785123964e-07, + "logits/chosen": -0.2373046875, + "logits/rejected": -0.421875, + "logps/chosen": -406.0, + "logps/rejected": -454.0, + "loss": 0.005, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.21484375, + "rewards/margins": 15.125, + "rewards/rejected": -15.3125, + "step": 1798 + }, + { + "epoch": 2.207361963190184, + "grad_norm": 1.3541770355269183, + "learning_rate": 4.0041322314049583e-07, + "logits/chosen": -0.197265625, + "logits/rejected": -0.3671875, + "logps/chosen": -400.0, + "logps/rejected": -452.0, + "loss": 0.003, + "rewards/accuracies": 1.0, + "rewards/chosen": -1.0, + "rewards/margins": 14.5625, + "rewards/rejected": -15.5625, + "step": 1799 + }, + { + "epoch": 2.208588957055215, + "grad_norm": 2.073405560530031, + "learning_rate": 3.997933884297521e-07, + "logits/chosen": -0.2392578125, + "logits/rejected": -0.37890625, + "logps/chosen": -388.0, + "logps/rejected": -428.0, + "loss": 0.0035, + "rewards/accuracies": 1.0, + "rewards/chosen": -1.296875, + "rewards/margins": 14.5, + "rewards/rejected": -15.75, + "step": 1800 + }, + { + "epoch": 2.2098159509202455, + "grad_norm": 2.5946800349252523, + "learning_rate": 3.991735537190083e-07, + "logits/chosen": -0.2158203125, + "logits/rejected": -0.375, + "logps/chosen": -360.0, + "logps/rejected": -426.0, + "loss": 0.0057, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.94921875, + "rewards/margins": 14.375, + "rewards/rejected": -15.3125, + "step": 1801 + }, + { + "epoch": 2.211042944785276, + "grad_norm": 5.122696905978627, + "learning_rate": 3.985537190082645e-07, + "logits/chosen": -0.322265625, + "logits/rejected": -0.451171875, + "logps/chosen": -400.0, + "logps/rejected": -434.0, + "loss": 0.0079, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.88671875, + "rewards/margins": 13.5, + "rewards/rejected": -14.375, + "step": 1802 + }, + { + "epoch": 2.212269938650307, + "grad_norm": 1.256491751369328, + "learning_rate": 3.9793388429752066e-07, + "logits/chosen": -0.384765625, + "logits/rejected": -0.59765625, + "logps/chosen": -404.0, + "logps/rejected": -434.0, + "loss": 0.0021, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.5546875, + "rewards/margins": 15.8125, + "rewards/rejected": -16.375, + "step": 1803 + }, + { + "epoch": 2.2134969325153375, + "grad_norm": 0.8850306694106227, + "learning_rate": 3.9731404958677686e-07, + "logits/chosen": -0.2216796875, + "logits/rejected": -0.4453125, + "logps/chosen": -406.0, + "logps/rejected": -430.0, + "loss": 0.0021, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.28125, + "rewards/margins": 15.4375, + "rewards/rejected": -15.6875, + "step": 1804 + }, + { + "epoch": 2.214723926380368, + "grad_norm": 0.6506125506189719, + "learning_rate": 3.9669421487603305e-07, + "logits/chosen": -0.1826171875, + "logits/rejected": -0.3984375, + "logps/chosen": -454.0, + "logps/rejected": -394.0, + "loss": 0.0017, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.4140625, + "rewards/margins": 14.6875, + "rewards/rejected": -15.125, + "step": 1805 + }, + { + "epoch": 2.215950920245399, + "grad_norm": 1.1202035891414293, + "learning_rate": 3.960743801652893e-07, + "logits/chosen": -0.2275390625, + "logits/rejected": -0.39453125, + "logps/chosen": -378.0, + "logps/rejected": -416.0, + "loss": 0.0026, + "rewards/accuracies": 1.0, + "rewards/chosen": -1.640625, + "rewards/margins": 13.4375, + "rewards/rejected": -15.0625, + "step": 1806 + }, + { + "epoch": 2.2171779141104295, + "grad_norm": 3.545014490080869, + "learning_rate": 3.9545454545454544e-07, + "logits/chosen": -0.1337890625, + "logits/rejected": -0.326171875, + "logps/chosen": -396.0, + "logps/rejected": -432.0, + "loss": 0.006, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.42578125, + "rewards/margins": 13.75, + "rewards/rejected": -14.1875, + "step": 1807 + }, + { + "epoch": 2.21840490797546, + "grad_norm": 0.2558505200432013, + "learning_rate": 3.9483471074380164e-07, + "logits/chosen": -0.333984375, + "logits/rejected": -0.49609375, + "logps/chosen": -364.0, + "logps/rejected": -448.0, + "loss": 0.0007, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.384765625, + "rewards/margins": 15.25, + "rewards/rejected": -15.625, + "step": 1808 + }, + { + "epoch": 2.219631901840491, + "grad_norm": 0.501745054178594, + "learning_rate": 3.942148760330579e-07, + "logits/chosen": -0.271484375, + "logits/rejected": -0.466796875, + "logps/chosen": -430.0, + "logps/rejected": -414.0, + "loss": 0.001, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.7265625, + "rewards/margins": 14.8125, + "rewards/rejected": -15.5625, + "step": 1809 + }, + { + "epoch": 2.2208588957055215, + "grad_norm": 2.328192985755221, + "learning_rate": 3.935950413223141e-07, + "logits/chosen": -0.1416015625, + "logits/rejected": -0.2890625, + "logps/chosen": -422.0, + "logps/rejected": -432.0, + "loss": 0.0035, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.70703125, + "rewards/margins": 13.5, + "rewards/rejected": -14.1875, + "step": 1810 + }, + { + "epoch": 2.222085889570552, + "grad_norm": 0.899790219472086, + "learning_rate": 3.929752066115703e-07, + "logits/chosen": -0.2314453125, + "logits/rejected": -0.376953125, + "logps/chosen": -370.0, + "logps/rejected": -392.0, + "loss": 0.0022, + "rewards/accuracies": 1.0, + "rewards/chosen": -1.3125, + "rewards/margins": 13.625, + "rewards/rejected": -14.9375, + "step": 1811 + }, + { + "epoch": 2.223312883435583, + "grad_norm": 0.7175401098394532, + "learning_rate": 3.923553719008264e-07, + "logits/chosen": -0.1962890625, + "logits/rejected": -0.361328125, + "logps/chosen": -384.0, + "logps/rejected": -400.0, + "loss": 0.0126, + "rewards/accuracies": 0.984375, + "rewards/chosen": -0.12255859375, + "rewards/margins": 15.0625, + "rewards/rejected": -15.1875, + "step": 1812 + }, + { + "epoch": 2.2245398773006135, + "grad_norm": 0.44058118443052646, + "learning_rate": 3.9173553719008266e-07, + "logits/chosen": -0.3359375, + "logits/rejected": -0.5390625, + "logps/chosen": -408.0, + "logps/rejected": -426.0, + "loss": 0.001, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.326171875, + "rewards/margins": 14.0625, + "rewards/rejected": -14.375, + "step": 1813 + }, + { + "epoch": 2.225766871165644, + "grad_norm": 1.394145444681074, + "learning_rate": 3.9111570247933885e-07, + "logits/chosen": -0.287109375, + "logits/rejected": -0.498046875, + "logps/chosen": -400.0, + "logps/rejected": -412.0, + "loss": 0.0077, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.546875, + "rewards/margins": 14.0, + "rewards/rejected": -14.5625, + "step": 1814 + }, + { + "epoch": 2.226993865030675, + "grad_norm": 0.9603756704079264, + "learning_rate": 3.904958677685951e-07, + "logits/chosen": -0.177734375, + "logits/rejected": -0.419921875, + "logps/chosen": -416.0, + "logps/rejected": -450.0, + "loss": 0.0017, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.439453125, + "rewards/margins": 14.6875, + "rewards/rejected": -15.125, + "step": 1815 + }, + { + "epoch": 2.2282208588957055, + "grad_norm": 1.7991246579030078, + "learning_rate": 3.8987603305785124e-07, + "logits/chosen": -0.2421875, + "logits/rejected": -0.419921875, + "logps/chosen": -430.0, + "logps/rejected": -448.0, + "loss": 0.0088, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.41015625, + "rewards/margins": 13.9375, + "rewards/rejected": -14.375, + "step": 1816 + }, + { + "epoch": 2.229447852760736, + "grad_norm": 0.9305690210458416, + "learning_rate": 3.8925619834710744e-07, + "logits/chosen": -0.193359375, + "logits/rejected": -0.439453125, + "logps/chosen": -442.0, + "logps/rejected": -430.0, + "loss": 0.0074, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.0908203125, + "rewards/margins": 15.125, + "rewards/rejected": -15.0625, + "step": 1817 + }, + { + "epoch": 2.230674846625767, + "grad_norm": 1.0250630003567436, + "learning_rate": 3.8863636363636363e-07, + "logits/chosen": -0.2275390625, + "logits/rejected": -0.47265625, + "logps/chosen": -402.0, + "logps/rejected": -430.0, + "loss": 0.0077, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.484375, + "rewards/margins": 14.5625, + "rewards/rejected": -15.0625, + "step": 1818 + }, + { + "epoch": 2.2319018404907975, + "grad_norm": 1.967115006721641, + "learning_rate": 3.880165289256199e-07, + "logits/chosen": -0.1826171875, + "logits/rejected": -0.36328125, + "logps/chosen": -378.0, + "logps/rejected": -390.0, + "loss": 0.0044, + "rewards/accuracies": 1.0, + "rewards/chosen": -1.3125, + "rewards/margins": 13.25, + "rewards/rejected": -14.5625, + "step": 1819 + }, + { + "epoch": 2.233128834355828, + "grad_norm": 0.7085803511028309, + "learning_rate": 3.87396694214876e-07, + "logits/chosen": -0.1025390625, + "logits/rejected": -0.2158203125, + "logps/chosen": -340.0, + "logps/rejected": -388.0, + "loss": 0.0014, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.05859375, + "rewards/margins": 14.0, + "rewards/rejected": -13.9375, + "step": 1820 + }, + { + "epoch": 2.234355828220859, + "grad_norm": 2.131272143912189, + "learning_rate": 3.867768595041322e-07, + "logits/chosen": -0.30078125, + "logits/rejected": -0.474609375, + "logps/chosen": -392.0, + "logps/rejected": -420.0, + "loss": 0.0029, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.3515625, + "rewards/margins": 14.5625, + "rewards/rejected": -14.9375, + "step": 1821 + }, + { + "epoch": 2.2355828220858895, + "grad_norm": 1.4723056773510292, + "learning_rate": 3.8615702479338846e-07, + "logits/chosen": -0.232421875, + "logits/rejected": -0.34375, + "logps/chosen": -394.0, + "logps/rejected": -410.0, + "loss": 0.0028, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.298828125, + "rewards/margins": 13.625, + "rewards/rejected": -13.875, + "step": 1822 + }, + { + "epoch": 2.23680981595092, + "grad_norm": 1.319023680883535, + "learning_rate": 3.8553719008264465e-07, + "logits/chosen": -0.08837890625, + "logits/rejected": -0.30859375, + "logps/chosen": -454.0, + "logps/rejected": -456.0, + "loss": 0.0037, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.02099609375, + "rewards/margins": 14.3125, + "rewards/rejected": -14.25, + "step": 1823 + }, + { + "epoch": 2.238036809815951, + "grad_norm": 0.47692754400802073, + "learning_rate": 3.8491735537190085e-07, + "logits/chosen": -0.0145263671875, + "logits/rejected": -0.203125, + "logps/chosen": -362.0, + "logps/rejected": -404.0, + "loss": 0.0015, + "rewards/accuracies": 1.0, + "rewards/chosen": -3.814697265625e-06, + "rewards/margins": 13.5625, + "rewards/rejected": -13.5625, + "step": 1824 + }, + { + "epoch": 2.2392638036809815, + "grad_norm": 2.0518515009155203, + "learning_rate": 3.84297520661157e-07, + "logits/chosen": -0.09765625, + "logits/rejected": -0.263671875, + "logps/chosen": -414.0, + "logps/rejected": -442.0, + "loss": 0.0032, + "rewards/accuracies": 1.0, + "rewards/chosen": -1.1171875, + "rewards/margins": 13.5625, + "rewards/rejected": -14.6875, + "step": 1825 + }, + { + "epoch": 2.240490797546012, + "grad_norm": 3.272072603340873, + "learning_rate": 3.8367768595041324e-07, + "logits/chosen": -0.1943359375, + "logits/rejected": -0.2890625, + "logps/chosen": -380.0, + "logps/rejected": -440.0, + "loss": 0.0083, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -1.03125, + "rewards/margins": 13.5, + "rewards/rejected": -14.5, + "step": 1826 + }, + { + "epoch": 2.241717791411043, + "grad_norm": 4.073640637902823, + "learning_rate": 3.8305785123966943e-07, + "logits/chosen": -0.193359375, + "logits/rejected": -0.4609375, + "logps/chosen": -398.0, + "logps/rejected": -418.0, + "loss": 0.0047, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.412109375, + "rewards/margins": 14.0625, + "rewards/rejected": -14.5, + "step": 1827 + }, + { + "epoch": 2.2429447852760735, + "grad_norm": 0.5802301531847157, + "learning_rate": 3.824380165289257e-07, + "logits/chosen": -0.193359375, + "logits/rejected": -0.33984375, + "logps/chosen": -374.0, + "logps/rejected": -430.0, + "loss": 0.0014, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.3671875, + "rewards/margins": 15.0, + "rewards/rejected": -15.375, + "step": 1828 + }, + { + "epoch": 2.244171779141104, + "grad_norm": 6.797984686094458, + "learning_rate": 3.8181818181818177e-07, + "logits/chosen": -0.21875, + "logits/rejected": -0.40234375, + "logps/chosen": -402.0, + "logps/rejected": -426.0, + "loss": 0.0076, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.7734375, + "rewards/margins": 12.625, + "rewards/rejected": -13.375, + "step": 1829 + }, + { + "epoch": 2.245398773006135, + "grad_norm": 1.5015878821386786, + "learning_rate": 3.81198347107438e-07, + "logits/chosen": -0.1357421875, + "logits/rejected": -0.314453125, + "logps/chosen": -374.0, + "logps/rejected": -396.0, + "loss": 0.0055, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.87890625, + "rewards/margins": 13.4375, + "rewards/rejected": -14.375, + "step": 1830 + }, + { + "epoch": 2.2466257668711656, + "grad_norm": 1.0450585891442645, + "learning_rate": 3.805785123966942e-07, + "logits/chosen": -0.234375, + "logits/rejected": -0.37109375, + "logps/chosen": -358.0, + "logps/rejected": -446.0, + "loss": 0.0023, + "rewards/accuracies": 1.0, + "rewards/chosen": -1.2890625, + "rewards/margins": 14.625, + "rewards/rejected": -15.875, + "step": 1831 + }, + { + "epoch": 2.247852760736196, + "grad_norm": 0.9970032472391912, + "learning_rate": 3.7995867768595046e-07, + "logits/chosen": -0.154296875, + "logits/rejected": -0.283203125, + "logps/chosen": -488.0, + "logps/rejected": -470.0, + "loss": 0.0025, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.34375, + "rewards/margins": 14.5625, + "rewards/rejected": -14.875, + "step": 1832 + }, + { + "epoch": 2.249079754601227, + "grad_norm": 1.411100974980419, + "learning_rate": 3.7933884297520665e-07, + "logits/chosen": -0.2216796875, + "logits/rejected": -0.421875, + "logps/chosen": -382.0, + "logps/rejected": -438.0, + "loss": 0.0084, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.6953125, + "rewards/margins": 14.0, + "rewards/rejected": -14.6875, + "step": 1833 + }, + { + "epoch": 2.2503067484662576, + "grad_norm": 0.47083334128050147, + "learning_rate": 3.787190082644628e-07, + "logits/chosen": -0.26171875, + "logits/rejected": -0.470703125, + "logps/chosen": -400.0, + "logps/rejected": -406.0, + "loss": 0.0011, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.68359375, + "rewards/margins": 13.25, + "rewards/rejected": -13.9375, + "step": 1834 + }, + { + "epoch": 2.2515337423312882, + "grad_norm": 0.8971630436095729, + "learning_rate": 3.78099173553719e-07, + "logits/chosen": -0.1416015625, + "logits/rejected": -0.34375, + "logps/chosen": -392.0, + "logps/rejected": -420.0, + "loss": 0.0021, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.64453125, + "rewards/margins": 14.0625, + "rewards/rejected": -14.6875, + "step": 1835 + }, + { + "epoch": 2.252760736196319, + "grad_norm": 0.5840464341150459, + "learning_rate": 3.7747933884297523e-07, + "logits/chosen": -0.1376953125, + "logits/rejected": -0.318359375, + "logps/chosen": -346.0, + "logps/rejected": -356.0, + "loss": 0.0015, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.349609375, + "rewards/margins": 13.625, + "rewards/rejected": -14.0, + "step": 1836 + }, + { + "epoch": 2.2539877300613496, + "grad_norm": 1.6126855018568913, + "learning_rate": 3.7685950413223143e-07, + "logits/chosen": -0.29296875, + "logits/rejected": -0.357421875, + "logps/chosen": -396.0, + "logps/rejected": -450.0, + "loss": 0.0021, + "rewards/accuracies": 1.0, + "rewards/chosen": -1.109375, + "rewards/margins": 13.75, + "rewards/rejected": -14.875, + "step": 1837 + }, + { + "epoch": 2.2552147239263802, + "grad_norm": 0.7135317817239797, + "learning_rate": 3.7623966942148757e-07, + "logits/chosen": -0.27734375, + "logits/rejected": -0.455078125, + "logps/chosen": -434.0, + "logps/rejected": -444.0, + "loss": 0.0016, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.56640625, + "rewards/margins": 13.875, + "rewards/rejected": -14.4375, + "step": 1838 + }, + { + "epoch": 2.256441717791411, + "grad_norm": 1.1734308986910358, + "learning_rate": 3.756198347107438e-07, + "logits/chosen": -0.10009765625, + "logits/rejected": -0.27734375, + "logps/chosen": -358.0, + "logps/rejected": -384.0, + "loss": 0.0023, + "rewards/accuracies": 1.0, + "rewards/chosen": -1.2265625, + "rewards/margins": 12.875, + "rewards/rejected": -14.125, + "step": 1839 + }, + { + "epoch": 2.2576687116564416, + "grad_norm": 1.543535337057203, + "learning_rate": 3.75e-07, + "logits/chosen": -0.19140625, + "logits/rejected": -0.373046875, + "logps/chosen": -436.0, + "logps/rejected": -432.0, + "loss": 0.0037, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.059814453125, + "rewards/margins": 14.5625, + "rewards/rejected": -14.5625, + "step": 1840 + }, + { + "epoch": 2.2588957055214722, + "grad_norm": 1.572629642854639, + "learning_rate": 3.743801652892562e-07, + "logits/chosen": -0.205078125, + "logits/rejected": -0.462890625, + "logps/chosen": -444.0, + "logps/rejected": -430.0, + "loss": 0.0032, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.1923828125, + "rewards/margins": 13.75, + "rewards/rejected": -13.9375, + "step": 1841 + }, + { + "epoch": 2.260122699386503, + "grad_norm": 0.5685676331892763, + "learning_rate": 3.737603305785124e-07, + "logits/chosen": -0.255859375, + "logits/rejected": -0.484375, + "logps/chosen": -374.0, + "logps/rejected": -424.0, + "loss": 0.0067, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -1.125, + "rewards/margins": 14.625, + "rewards/rejected": -15.75, + "step": 1842 + }, + { + "epoch": 2.2613496932515336, + "grad_norm": 7.737812617575026, + "learning_rate": 3.7314049586776865e-07, + "logits/chosen": -0.2275390625, + "logits/rejected": -0.365234375, + "logps/chosen": -376.0, + "logps/rejected": -418.0, + "loss": 0.0121, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -1.09375, + "rewards/margins": 13.875, + "rewards/rejected": -14.9375, + "step": 1843 + }, + { + "epoch": 2.2625766871165642, + "grad_norm": 1.0292113024553593, + "learning_rate": 3.725206611570248e-07, + "logits/chosen": -0.2578125, + "logits/rejected": -0.416015625, + "logps/chosen": -418.0, + "logps/rejected": -472.0, + "loss": 0.0017, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.4453125, + "rewards/margins": 15.375, + "rewards/rejected": -15.8125, + "step": 1844 + }, + { + "epoch": 2.263803680981595, + "grad_norm": 0.726798636284396, + "learning_rate": 3.7190082644628103e-07, + "logits/chosen": -0.1376953125, + "logits/rejected": -0.296875, + "logps/chosen": -444.0, + "logps/rejected": -432.0, + "loss": 0.0016, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.1357421875, + "rewards/margins": 14.0, + "rewards/rejected": -14.125, + "step": 1845 + }, + { + "epoch": 2.265030674846626, + "grad_norm": 1.830994682110566, + "learning_rate": 3.712809917355372e-07, + "logits/chosen": -0.283203125, + "logits/rejected": -0.515625, + "logps/chosen": -452.0, + "logps/rejected": -448.0, + "loss": 0.0034, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.875, + "rewards/margins": 14.6875, + "rewards/rejected": -15.5625, + "step": 1846 + }, + { + "epoch": 2.2662576687116562, + "grad_norm": 0.554314623383904, + "learning_rate": 3.706611570247934e-07, + "logits/chosen": -0.435546875, + "logits/rejected": -0.55078125, + "logps/chosen": -406.0, + "logps/rejected": -470.0, + "loss": 0.0069, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.052734375, + "rewards/margins": 15.75, + "rewards/rejected": -15.6875, + "step": 1847 + }, + { + "epoch": 2.2674846625766873, + "grad_norm": 1.285661358075267, + "learning_rate": 3.7004132231404956e-07, + "logits/chosen": -0.123046875, + "logits/rejected": -0.33203125, + "logps/chosen": -432.0, + "logps/rejected": -432.0, + "loss": 0.0077, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.353515625, + "rewards/margins": 14.5, + "rewards/rejected": -14.1875, + "step": 1848 + }, + { + "epoch": 2.2687116564417176, + "grad_norm": 0.659622242358561, + "learning_rate": 3.694214876033058e-07, + "logits/chosen": -0.2890625, + "logits/rejected": -0.546875, + "logps/chosen": -472.0, + "logps/rejected": -452.0, + "loss": 0.0014, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.01171875, + "rewards/margins": 15.8125, + "rewards/rejected": -15.8125, + "step": 1849 + }, + { + "epoch": 2.2699386503067487, + "grad_norm": 6.408848927369335, + "learning_rate": 3.68801652892562e-07, + "logits/chosen": -0.2890625, + "logits/rejected": -0.5234375, + "logps/chosen": -434.0, + "logps/rejected": -432.0, + "loss": 0.0165, + "rewards/accuracies": 0.984375, + "rewards/chosen": -0.255859375, + "rewards/margins": 14.5625, + "rewards/rejected": -14.8125, + "step": 1850 + }, + { + "epoch": 2.271165644171779, + "grad_norm": 1.611704571810828, + "learning_rate": 3.681818181818182e-07, + "logits/chosen": -0.10009765625, + "logits/rejected": -0.3125, + "logps/chosen": -380.0, + "logps/rejected": -446.0, + "loss": 0.0029, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.85546875, + "rewards/margins": 14.9375, + "rewards/rejected": -15.75, + "step": 1851 + }, + { + "epoch": 2.27239263803681, + "grad_norm": 1.8979393045619397, + "learning_rate": 3.675619834710744e-07, + "logits/chosen": -0.1044921875, + "logits/rejected": -0.328125, + "logps/chosen": -458.0, + "logps/rejected": -474.0, + "loss": 0.002, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.30859375, + "rewards/margins": 15.3125, + "rewards/rejected": -15.625, + "step": 1852 + }, + { + "epoch": 2.2736196319018402, + "grad_norm": 6.969581413943249, + "learning_rate": 3.669421487603306e-07, + "logits/chosen": 0.01275634765625, + "logits/rejected": -0.1669921875, + "logps/chosen": -394.0, + "logps/rejected": -388.0, + "loss": 0.0101, + "rewards/accuracies": 1.0, + "rewards/chosen": -1.4609375, + "rewards/margins": 13.1875, + "rewards/rejected": -14.625, + "step": 1853 + }, + { + "epoch": 2.2748466257668714, + "grad_norm": 0.8220031684651788, + "learning_rate": 3.663223140495868e-07, + "logits/chosen": -0.10107421875, + "logits/rejected": -0.251953125, + "logps/chosen": -368.0, + "logps/rejected": -402.0, + "loss": 0.0019, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.34375, + "rewards/margins": 14.6875, + "rewards/rejected": -15.0, + "step": 1854 + }, + { + "epoch": 2.276073619631902, + "grad_norm": 1.6293458498904758, + "learning_rate": 3.65702479338843e-07, + "logits/chosen": -0.23046875, + "logits/rejected": -0.44140625, + "logps/chosen": -452.0, + "logps/rejected": -438.0, + "loss": 0.0038, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.369140625, + "rewards/margins": 15.5, + "rewards/rejected": -15.8125, + "step": 1855 + }, + { + "epoch": 2.2773006134969327, + "grad_norm": 1.228892488562491, + "learning_rate": 3.6508264462809917e-07, + "logits/chosen": -0.2353515625, + "logits/rejected": -0.443359375, + "logps/chosen": -406.0, + "logps/rejected": -402.0, + "loss": 0.0022, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.455078125, + "rewards/margins": 14.6875, + "rewards/rejected": -15.1875, + "step": 1856 + }, + { + "epoch": 2.2785276073619634, + "grad_norm": 1.347146154203732, + "learning_rate": 3.6446280991735536e-07, + "logits/chosen": -0.203125, + "logits/rejected": -0.404296875, + "logps/chosen": -372.0, + "logps/rejected": -380.0, + "loss": 0.0022, + "rewards/accuracies": 1.0, + "rewards/chosen": -1.2890625, + "rewards/margins": 13.6875, + "rewards/rejected": -15.0, + "step": 1857 + }, + { + "epoch": 2.279754601226994, + "grad_norm": 0.4254637728476363, + "learning_rate": 3.6384297520661156e-07, + "logits/chosen": -0.322265625, + "logits/rejected": -0.455078125, + "logps/chosen": -394.0, + "logps/rejected": -452.0, + "loss": 0.0063, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -1.0390625, + "rewards/margins": 14.8125, + "rewards/rejected": -15.875, + "step": 1858 + }, + { + "epoch": 2.2809815950920247, + "grad_norm": 1.9925368374499628, + "learning_rate": 3.6322314049586775e-07, + "logits/chosen": -0.380859375, + "logits/rejected": -0.53515625, + "logps/chosen": -396.0, + "logps/rejected": -424.0, + "loss": 0.0034, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.5, + "rewards/margins": 14.25, + "rewards/rejected": -14.75, + "step": 1859 + }, + { + "epoch": 2.2822085889570554, + "grad_norm": 2.056567045112219, + "learning_rate": 3.62603305785124e-07, + "logits/chosen": -0.26953125, + "logits/rejected": -0.470703125, + "logps/chosen": -424.0, + "logps/rejected": -442.0, + "loss": 0.0034, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.625, + "rewards/margins": 13.5625, + "rewards/rejected": -14.1875, + "step": 1860 + }, + { + "epoch": 2.283435582822086, + "grad_norm": 2.8142356385846314, + "learning_rate": 3.619834710743802e-07, + "logits/chosen": -0.32421875, + "logits/rejected": -0.478515625, + "logps/chosen": -418.0, + "logps/rejected": -444.0, + "loss": 0.0053, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.953125, + "rewards/margins": 13.9375, + "rewards/rejected": -14.9375, + "step": 1861 + }, + { + "epoch": 2.2846625766871167, + "grad_norm": 0.759260631978291, + "learning_rate": 3.613636363636364e-07, + "logits/chosen": -0.2138671875, + "logits/rejected": -0.37890625, + "logps/chosen": -364.0, + "logps/rejected": -402.0, + "loss": 0.0019, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.58203125, + "rewards/margins": 13.625, + "rewards/rejected": -14.1875, + "step": 1862 + }, + { + "epoch": 2.2858895705521474, + "grad_norm": 2.731584291697469, + "learning_rate": 3.607438016528926e-07, + "logits/chosen": -0.21875, + "logits/rejected": -0.302734375, + "logps/chosen": -440.0, + "logps/rejected": -440.0, + "loss": 0.004, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.474609375, + "rewards/margins": 14.3125, + "rewards/rejected": -14.8125, + "step": 1863 + }, + { + "epoch": 2.287116564417178, + "grad_norm": 1.3642254394105224, + "learning_rate": 3.601239669421488e-07, + "logits/chosen": -0.2734375, + "logits/rejected": -0.470703125, + "logps/chosen": -438.0, + "logps/rejected": -468.0, + "loss": 0.0076, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.86328125, + "rewards/margins": 15.1875, + "rewards/rejected": -16.0, + "step": 1864 + }, + { + "epoch": 2.2883435582822087, + "grad_norm": 1.2319362408571326, + "learning_rate": 3.5950413223140497e-07, + "logits/chosen": -0.265625, + "logits/rejected": -0.369140625, + "logps/chosen": -402.0, + "logps/rejected": -434.0, + "loss": 0.0025, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.4765625, + "rewards/margins": 14.0625, + "rewards/rejected": -14.5625, + "step": 1865 + }, + { + "epoch": 2.2895705521472394, + "grad_norm": 1.2108912803681235, + "learning_rate": 3.5888429752066117e-07, + "logits/chosen": -0.21875, + "logits/rejected": -0.4609375, + "logps/chosen": -396.0, + "logps/rejected": -434.0, + "loss": 0.0024, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.94140625, + "rewards/margins": 15.0, + "rewards/rejected": -15.9375, + "step": 1866 + }, + { + "epoch": 2.29079754601227, + "grad_norm": 0.681863992392613, + "learning_rate": 3.5826446280991736e-07, + "logits/chosen": -0.041748046875, + "logits/rejected": -0.1962890625, + "logps/chosen": -352.0, + "logps/rejected": -408.0, + "loss": 0.0068, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.26953125, + "rewards/margins": 15.125, + "rewards/rejected": -15.375, + "step": 1867 + }, + { + "epoch": 2.2920245398773007, + "grad_norm": 1.7677144051946254, + "learning_rate": 3.5764462809917355e-07, + "logits/chosen": -0.265625, + "logits/rejected": -0.369140625, + "logps/chosen": -384.0, + "logps/rejected": -416.0, + "loss": 0.0034, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.83203125, + "rewards/margins": 14.0, + "rewards/rejected": -14.8125, + "step": 1868 + }, + { + "epoch": 2.2932515337423314, + "grad_norm": 0.5954790801095706, + "learning_rate": 3.5702479338842975e-07, + "logits/chosen": -0.236328125, + "logits/rejected": -0.419921875, + "logps/chosen": -420.0, + "logps/rejected": -426.0, + "loss": 0.0013, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.5703125, + "rewards/margins": 13.75, + "rewards/rejected": -14.3125, + "step": 1869 + }, + { + "epoch": 2.294478527607362, + "grad_norm": 0.5791521538225278, + "learning_rate": 3.56404958677686e-07, + "logits/chosen": -0.166015625, + "logits/rejected": -0.32421875, + "logps/chosen": -360.0, + "logps/rejected": -384.0, + "loss": 0.0013, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.8125, + "rewards/margins": 14.0, + "rewards/rejected": -14.8125, + "step": 1870 + }, + { + "epoch": 2.2957055214723927, + "grad_norm": 1.2850139977026085, + "learning_rate": 3.5578512396694214e-07, + "logits/chosen": -0.1806640625, + "logits/rejected": -0.314453125, + "logps/chosen": -396.0, + "logps/rejected": -408.0, + "loss": 0.0024, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.458984375, + "rewards/margins": 14.6875, + "rewards/rejected": -15.125, + "step": 1871 + }, + { + "epoch": 2.2969325153374234, + "grad_norm": 0.9713733882952269, + "learning_rate": 3.551652892561984e-07, + "logits/chosen": -0.318359375, + "logits/rejected": -0.38671875, + "logps/chosen": -424.0, + "logps/rejected": -446.0, + "loss": 0.0021, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.1767578125, + "rewards/margins": 14.5, + "rewards/rejected": -14.6875, + "step": 1872 + }, + { + "epoch": 2.298159509202454, + "grad_norm": 0.8570866903482623, + "learning_rate": 3.545454545454545e-07, + "logits/chosen": -0.1298828125, + "logits/rejected": -0.2060546875, + "logps/chosen": -380.0, + "logps/rejected": -404.0, + "loss": 0.0024, + "rewards/accuracies": 1.0, + "rewards/chosen": -1.1328125, + "rewards/margins": 12.8125, + "rewards/rejected": -13.9375, + "step": 1873 + }, + { + "epoch": 2.2993865030674847, + "grad_norm": 0.877243495599524, + "learning_rate": 3.5392561983471077e-07, + "logits/chosen": -0.2470703125, + "logits/rejected": -0.51171875, + "logps/chosen": -446.0, + "logps/rejected": -454.0, + "loss": 0.0018, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.5, + "rewards/margins": 16.375, + "rewards/rejected": -15.875, + "step": 1874 + }, + { + "epoch": 2.3006134969325154, + "grad_norm": 0.6142695878171099, + "learning_rate": 3.5330578512396697e-07, + "logits/chosen": -0.2216796875, + "logits/rejected": -0.435546875, + "logps/chosen": -402.0, + "logps/rejected": -432.0, + "loss": 0.0015, + "rewards/accuracies": 1.0, + "rewards/chosen": -1.0625, + "rewards/margins": 15.5, + "rewards/rejected": -16.625, + "step": 1875 + }, + { + "epoch": 2.301840490797546, + "grad_norm": 1.7038425887893978, + "learning_rate": 3.5268595041322316e-07, + "logits/chosen": -0.19921875, + "logits/rejected": -0.392578125, + "logps/chosen": -468.0, + "logps/rejected": -448.0, + "loss": 0.0035, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.115234375, + "rewards/margins": 15.3125, + "rewards/rejected": -15.4375, + "step": 1876 + }, + { + "epoch": 2.3030674846625767, + "grad_norm": 1.1729173763686422, + "learning_rate": 3.5206611570247935e-07, + "logits/chosen": -0.0595703125, + "logits/rejected": -0.2578125, + "logps/chosen": -392.0, + "logps/rejected": -402.0, + "loss": 0.0026, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.05517578125, + "rewards/margins": 12.9375, + "rewards/rejected": -13.0, + "step": 1877 + }, + { + "epoch": 2.3042944785276074, + "grad_norm": 2.7245370546235055, + "learning_rate": 3.5144628099173555e-07, + "logits/chosen": -0.271484375, + "logits/rejected": -0.52734375, + "logps/chosen": -388.0, + "logps/rejected": -422.0, + "loss": 0.0058, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.3828125, + "rewards/margins": 13.75, + "rewards/rejected": -14.1875, + "step": 1878 + }, + { + "epoch": 2.305521472392638, + "grad_norm": 0.6800209428189053, + "learning_rate": 3.5082644628099174e-07, + "logits/chosen": -0.1982421875, + "logits/rejected": -0.462890625, + "logps/chosen": -428.0, + "logps/rejected": -420.0, + "loss": 0.0069, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.734375, + "rewards/margins": 13.8125, + "rewards/rejected": -14.5625, + "step": 1879 + }, + { + "epoch": 2.3067484662576687, + "grad_norm": 1.0458743836235773, + "learning_rate": 3.5020661157024794e-07, + "logits/chosen": -0.2138671875, + "logits/rejected": -0.4453125, + "logps/chosen": -440.0, + "logps/rejected": -436.0, + "loss": 0.0028, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.10888671875, + "rewards/margins": 13.9375, + "rewards/rejected": -13.8125, + "step": 1880 + }, + { + "epoch": 2.3079754601226994, + "grad_norm": 2.0525796408749195, + "learning_rate": 3.495867768595042e-07, + "logits/chosen": -0.2119140625, + "logits/rejected": -0.375, + "logps/chosen": -440.0, + "logps/rejected": -448.0, + "loss": 0.0028, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.46484375, + "rewards/margins": 13.0, + "rewards/rejected": -13.4375, + "step": 1881 + }, + { + "epoch": 2.30920245398773, + "grad_norm": 0.7484011374378289, + "learning_rate": 3.489669421487603e-07, + "logits/chosen": -0.279296875, + "logits/rejected": -0.384765625, + "logps/chosen": -422.0, + "logps/rejected": -432.0, + "loss": 0.0068, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.54296875, + "rewards/margins": 13.8125, + "rewards/rejected": -14.3125, + "step": 1882 + }, + { + "epoch": 2.3104294478527607, + "grad_norm": 1.710519086785903, + "learning_rate": 3.4834710743801657e-07, + "logits/chosen": -0.1298828125, + "logits/rejected": -0.3125, + "logps/chosen": -364.0, + "logps/rejected": -354.0, + "loss": 0.0033, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.83203125, + "rewards/margins": 12.5, + "rewards/rejected": -13.375, + "step": 1883 + }, + { + "epoch": 2.3116564417177914, + "grad_norm": 0.9138575267778296, + "learning_rate": 3.477272727272727e-07, + "logits/chosen": -0.1279296875, + "logits/rejected": -0.275390625, + "logps/chosen": -390.0, + "logps/rejected": -416.0, + "loss": 0.0025, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.5, + "rewards/margins": 13.8125, + "rewards/rejected": -14.3125, + "step": 1884 + }, + { + "epoch": 2.312883435582822, + "grad_norm": 4.324288149496707, + "learning_rate": 3.4710743801652896e-07, + "logits/chosen": -0.2216796875, + "logits/rejected": -0.3984375, + "logps/chosen": -372.0, + "logps/rejected": -420.0, + "loss": 0.0073, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.7109375, + "rewards/margins": 14.375, + "rewards/rejected": -15.125, + "step": 1885 + }, + { + "epoch": 2.3141104294478527, + "grad_norm": 2.7532208503200293, + "learning_rate": 3.464876033057851e-07, + "logits/chosen": -0.1630859375, + "logits/rejected": -0.33203125, + "logps/chosen": -392.0, + "logps/rejected": -426.0, + "loss": 0.0032, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.93359375, + "rewards/margins": 14.0, + "rewards/rejected": -14.9375, + "step": 1886 + }, + { + "epoch": 2.3153374233128834, + "grad_norm": 2.121063217175678, + "learning_rate": 3.4586776859504135e-07, + "logits/chosen": -0.2001953125, + "logits/rejected": -0.361328125, + "logps/chosen": -450.0, + "logps/rejected": -486.0, + "loss": 0.003, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.404296875, + "rewards/margins": 14.5, + "rewards/rejected": -14.875, + "step": 1887 + }, + { + "epoch": 2.316564417177914, + "grad_norm": 0.4658535116192763, + "learning_rate": 3.452479338842975e-07, + "logits/chosen": -0.234375, + "logits/rejected": -0.48828125, + "logps/chosen": -438.0, + "logps/rejected": -446.0, + "loss": 0.0012, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.80859375, + "rewards/margins": 14.6875, + "rewards/rejected": -15.5, + "step": 1888 + }, + { + "epoch": 2.3177914110429447, + "grad_norm": 0.9806083563971221, + "learning_rate": 3.4462809917355374e-07, + "logits/chosen": -0.265625, + "logits/rejected": -0.439453125, + "logps/chosen": -396.0, + "logps/rejected": -424.0, + "loss": 0.0023, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.59375, + "rewards/margins": 13.9375, + "rewards/rejected": -14.5625, + "step": 1889 + }, + { + "epoch": 2.3190184049079754, + "grad_norm": 2.2270988598464982, + "learning_rate": 3.4400826446280993e-07, + "logits/chosen": -0.1962890625, + "logits/rejected": -0.482421875, + "logps/chosen": -436.0, + "logps/rejected": -416.0, + "loss": 0.003, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.1318359375, + "rewards/margins": 13.6875, + "rewards/rejected": -13.875, + "step": 1890 + }, + { + "epoch": 2.320245398773006, + "grad_norm": 4.592674575248516, + "learning_rate": 3.433884297520661e-07, + "logits/chosen": -0.32421875, + "logits/rejected": -0.470703125, + "logps/chosen": -364.0, + "logps/rejected": -414.0, + "loss": 0.0063, + "rewards/accuracies": 1.0, + "rewards/chosen": -1.546875, + "rewards/margins": 13.75, + "rewards/rejected": -15.3125, + "step": 1891 + }, + { + "epoch": 2.3214723926380367, + "grad_norm": 2.489776248965657, + "learning_rate": 3.427685950413223e-07, + "logits/chosen": -0.2734375, + "logits/rejected": -0.46484375, + "logps/chosen": -434.0, + "logps/rejected": -438.0, + "loss": 0.0119, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.5859375, + "rewards/margins": 14.4375, + "rewards/rejected": -15.0, + "step": 1892 + }, + { + "epoch": 2.3226993865030674, + "grad_norm": 2.152754358936679, + "learning_rate": 3.421487603305785e-07, + "logits/chosen": -0.208984375, + "logits/rejected": -0.421875, + "logps/chosen": -398.0, + "logps/rejected": -404.0, + "loss": 0.0031, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.50390625, + "rewards/margins": 13.625, + "rewards/rejected": -14.125, + "step": 1893 + }, + { + "epoch": 2.323926380368098, + "grad_norm": 1.146355611803609, + "learning_rate": 3.415289256198347e-07, + "logits/chosen": -0.2353515625, + "logits/rejected": -0.4140625, + "logps/chosen": -420.0, + "logps/rejected": -440.0, + "loss": 0.0023, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.53125, + "rewards/margins": 13.3125, + "rewards/rejected": -13.875, + "step": 1894 + }, + { + "epoch": 2.3251533742331287, + "grad_norm": 0.6707710990221305, + "learning_rate": 3.409090909090909e-07, + "logits/chosen": -0.2255859375, + "logits/rejected": -0.3046875, + "logps/chosen": -406.0, + "logps/rejected": -414.0, + "loss": 0.0014, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.78515625, + "rewards/margins": 14.125, + "rewards/rejected": -14.875, + "step": 1895 + }, + { + "epoch": 2.3263803680981594, + "grad_norm": 2.445018309979311, + "learning_rate": 3.402892561983471e-07, + "logits/chosen": -0.2021484375, + "logits/rejected": -0.341796875, + "logps/chosen": -374.0, + "logps/rejected": -424.0, + "loss": 0.0089, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.08154296875, + "rewards/margins": 14.375, + "rewards/rejected": -14.4375, + "step": 1896 + }, + { + "epoch": 2.32760736196319, + "grad_norm": 0.9549808009623026, + "learning_rate": 3.396694214876033e-07, + "logits/chosen": -0.34765625, + "logits/rejected": -0.498046875, + "logps/chosen": -402.0, + "logps/rejected": -424.0, + "loss": 0.0019, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.00555419921875, + "rewards/margins": 14.25, + "rewards/rejected": -14.25, + "step": 1897 + }, + { + "epoch": 2.3288343558282207, + "grad_norm": 2.190249668442515, + "learning_rate": 3.3904958677685954e-07, + "logits/chosen": -0.271484375, + "logits/rejected": -0.3359375, + "logps/chosen": -380.0, + "logps/rejected": -420.0, + "loss": 0.0042, + "rewards/accuracies": 1.0, + "rewards/chosen": -1.359375, + "rewards/margins": 13.75, + "rewards/rejected": -15.125, + "step": 1898 + }, + { + "epoch": 2.3300613496932514, + "grad_norm": 0.7061460369831141, + "learning_rate": 3.384297520661157e-07, + "logits/chosen": -0.0869140625, + "logits/rejected": -0.2431640625, + "logps/chosen": -412.0, + "logps/rejected": -424.0, + "loss": 0.0072, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -2.203125, + "rewards/margins": 13.25, + "rewards/rejected": -15.4375, + "step": 1899 + }, + { + "epoch": 2.331288343558282, + "grad_norm": 2.742155067958955, + "learning_rate": 3.3780991735537193e-07, + "logits/chosen": -0.189453125, + "logits/rejected": -0.345703125, + "logps/chosen": -416.0, + "logps/rejected": -404.0, + "loss": 0.0037, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.66015625, + "rewards/margins": 13.125, + "rewards/rejected": -13.8125, + "step": 1900 + }, + { + "epoch": 2.3325153374233127, + "grad_norm": 1.3292102754711506, + "learning_rate": 3.371900826446281e-07, + "logits/chosen": -0.375, + "logits/rejected": -0.51171875, + "logps/chosen": -384.0, + "logps/rejected": -434.0, + "loss": 0.0024, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.734375, + "rewards/margins": 14.1875, + "rewards/rejected": -14.9375, + "step": 1901 + }, + { + "epoch": 2.3337423312883434, + "grad_norm": 0.8779816436731417, + "learning_rate": 3.365702479338843e-07, + "logits/chosen": -0.087890625, + "logits/rejected": -0.30078125, + "logps/chosen": -448.0, + "logps/rejected": -464.0, + "loss": 0.0024, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.1513671875, + "rewards/margins": 13.8125, + "rewards/rejected": -13.9375, + "step": 1902 + }, + { + "epoch": 2.334969325153374, + "grad_norm": 3.9533572626817413, + "learning_rate": 3.359504132231405e-07, + "logits/chosen": -0.27734375, + "logits/rejected": -0.478515625, + "logps/chosen": -408.0, + "logps/rejected": -454.0, + "loss": 0.0056, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.05029296875, + "rewards/margins": 15.6875, + "rewards/rejected": -15.75, + "step": 1903 + }, + { + "epoch": 2.3361963190184047, + "grad_norm": 1.6036517619838904, + "learning_rate": 3.353305785123967e-07, + "logits/chosen": -0.162109375, + "logits/rejected": -0.37109375, + "logps/chosen": -404.0, + "logps/rejected": -424.0, + "loss": 0.0031, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.039794921875, + "rewards/margins": 14.0625, + "rewards/rejected": -14.0625, + "step": 1904 + }, + { + "epoch": 2.3374233128834354, + "grad_norm": 1.4760802020266122, + "learning_rate": 3.347107438016529e-07, + "logits/chosen": -0.26953125, + "logits/rejected": -0.4375, + "logps/chosen": -428.0, + "logps/rejected": -480.0, + "loss": 0.0024, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.061279296875, + "rewards/margins": 15.6875, + "rewards/rejected": -15.625, + "step": 1905 + }, + { + "epoch": 2.338650306748466, + "grad_norm": 1.6158233357595397, + "learning_rate": 3.340909090909091e-07, + "logits/chosen": -0.265625, + "logits/rejected": -0.416015625, + "logps/chosen": -354.0, + "logps/rejected": -426.0, + "loss": 0.0033, + "rewards/accuracies": 1.0, + "rewards/chosen": -1.3984375, + "rewards/margins": 13.3125, + "rewards/rejected": -14.6875, + "step": 1906 + }, + { + "epoch": 2.3398773006134967, + "grad_norm": 1.3000792388953792, + "learning_rate": 3.334710743801653e-07, + "logits/chosen": -0.37109375, + "logits/rejected": -0.58984375, + "logps/chosen": -400.0, + "logps/rejected": -426.0, + "loss": 0.0037, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.0869140625, + "rewards/margins": 14.375, + "rewards/rejected": -14.5, + "step": 1907 + }, + { + "epoch": 2.3411042944785274, + "grad_norm": 0.5392221622650879, + "learning_rate": 3.328512396694215e-07, + "logits/chosen": -0.267578125, + "logits/rejected": -0.443359375, + "logps/chosen": -406.0, + "logps/rejected": -422.0, + "loss": 0.0011, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.703125, + "rewards/margins": 14.25, + "rewards/rejected": -15.0, + "step": 1908 + }, + { + "epoch": 2.3423312883435585, + "grad_norm": 1.126221146289658, + "learning_rate": 3.322314049586777e-07, + "logits/chosen": -0.337890625, + "logits/rejected": -0.51953125, + "logps/chosen": -434.0, + "logps/rejected": -430.0, + "loss": 0.0024, + "rewards/accuracies": 1.0, + "rewards/chosen": 1.25, + "rewards/margins": 15.0, + "rewards/rejected": -13.75, + "step": 1909 + }, + { + "epoch": 2.3435582822085887, + "grad_norm": 0.7055865722405307, + "learning_rate": 3.316115702479339e-07, + "logits/chosen": -0.240234375, + "logits/rejected": -0.38671875, + "logps/chosen": -406.0, + "logps/rejected": -436.0, + "loss": 0.0017, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.34375, + "rewards/margins": 14.0625, + "rewards/rejected": -13.6875, + "step": 1910 + }, + { + "epoch": 2.34478527607362, + "grad_norm": 2.302191850290034, + "learning_rate": 3.3099173553719006e-07, + "logits/chosen": -0.2490234375, + "logits/rejected": -0.4609375, + "logps/chosen": -404.0, + "logps/rejected": -416.0, + "loss": 0.0143, + "rewards/accuracies": 0.984375, + "rewards/chosen": -0.5625, + "rewards/margins": 14.0625, + "rewards/rejected": -14.625, + "step": 1911 + }, + { + "epoch": 2.34601226993865, + "grad_norm": 0.754133813333508, + "learning_rate": 3.303719008264463e-07, + "logits/chosen": -0.32421875, + "logits/rejected": -0.408203125, + "logps/chosen": -410.0, + "logps/rejected": -436.0, + "loss": 0.0018, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.60546875, + "rewards/margins": 13.5, + "rewards/rejected": -14.125, + "step": 1912 + }, + { + "epoch": 2.347239263803681, + "grad_norm": 1.7069831017924897, + "learning_rate": 3.297520661157025e-07, + "logits/chosen": -0.2890625, + "logits/rejected": -0.5625, + "logps/chosen": -386.0, + "logps/rejected": -382.0, + "loss": 0.0022, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.515625, + "rewards/margins": 13.8125, + "rewards/rejected": -14.3125, + "step": 1913 + }, + { + "epoch": 2.3484662576687114, + "grad_norm": 0.5099642162228847, + "learning_rate": 3.291322314049587e-07, + "logits/chosen": -0.3359375, + "logits/rejected": -0.515625, + "logps/chosen": -362.0, + "logps/rejected": -392.0, + "loss": 0.0013, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.279296875, + "rewards/margins": 14.125, + "rewards/rejected": -14.4375, + "step": 1914 + }, + { + "epoch": 2.3496932515337425, + "grad_norm": 1.5684554016711487, + "learning_rate": 3.285123966942149e-07, + "logits/chosen": -0.287109375, + "logits/rejected": -0.46875, + "logps/chosen": -422.0, + "logps/rejected": -446.0, + "loss": 0.0016, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.498046875, + "rewards/margins": 15.5625, + "rewards/rejected": -15.0625, + "step": 1915 + }, + { + "epoch": 2.3509202453987728, + "grad_norm": 1.55997230989347, + "learning_rate": 3.278925619834711e-07, + "logits/chosen": -0.19140625, + "logits/rejected": -0.345703125, + "logps/chosen": -440.0, + "logps/rejected": -436.0, + "loss": 0.0024, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.28125, + "rewards/margins": 15.125, + "rewards/rejected": -15.4375, + "step": 1916 + }, + { + "epoch": 2.352147239263804, + "grad_norm": 0.7047902620839831, + "learning_rate": 3.272727272727273e-07, + "logits/chosen": -0.287109375, + "logits/rejected": -0.5390625, + "logps/chosen": -404.0, + "logps/rejected": -422.0, + "loss": 0.0016, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.2353515625, + "rewards/margins": 14.4375, + "rewards/rejected": -14.1875, + "step": 1917 + }, + { + "epoch": 2.3533742331288345, + "grad_norm": 1.4078396812648037, + "learning_rate": 3.266528925619835e-07, + "logits/chosen": -0.2314453125, + "logits/rejected": -0.49609375, + "logps/chosen": -418.0, + "logps/rejected": -448.0, + "loss": 0.0039, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.423828125, + "rewards/margins": 15.8125, + "rewards/rejected": -16.25, + "step": 1918 + }, + { + "epoch": 2.354601226993865, + "grad_norm": 1.4557324329157222, + "learning_rate": 3.2603305785123967e-07, + "logits/chosen": -0.208984375, + "logits/rejected": -0.412109375, + "logps/chosen": -394.0, + "logps/rejected": -432.0, + "loss": 0.0084, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.7265625, + "rewards/margins": 14.0625, + "rewards/rejected": -14.8125, + "step": 1919 + }, + { + "epoch": 2.355828220858896, + "grad_norm": 2.410866652825109, + "learning_rate": 3.2541322314049586e-07, + "logits/chosen": -0.3515625, + "logits/rejected": -0.51953125, + "logps/chosen": -400.0, + "logps/rejected": -436.0, + "loss": 0.006, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.5234375, + "rewards/margins": 14.0, + "rewards/rejected": -14.5625, + "step": 1920 + }, + { + "epoch": 2.3570552147239265, + "grad_norm": 0.8309362283856945, + "learning_rate": 3.247933884297521e-07, + "logits/chosen": -0.2734375, + "logits/rejected": -0.45703125, + "logps/chosen": -386.0, + "logps/rejected": -428.0, + "loss": 0.002, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.298828125, + "rewards/margins": 14.6875, + "rewards/rejected": -15.0, + "step": 1921 + }, + { + "epoch": 2.358282208588957, + "grad_norm": 0.7461628610325094, + "learning_rate": 3.2417355371900825e-07, + "logits/chosen": -0.203125, + "logits/rejected": -0.333984375, + "logps/chosen": -358.0, + "logps/rejected": -388.0, + "loss": 0.0021, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.70703125, + "rewards/margins": 13.25, + "rewards/rejected": -14.0, + "step": 1922 + }, + { + "epoch": 2.359509202453988, + "grad_norm": 1.4434409839271034, + "learning_rate": 3.235537190082645e-07, + "logits/chosen": -0.345703125, + "logits/rejected": -0.6328125, + "logps/chosen": -392.0, + "logps/rejected": -430.0, + "loss": 0.0022, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.322265625, + "rewards/margins": 14.875, + "rewards/rejected": -15.1875, + "step": 1923 + }, + { + "epoch": 2.3607361963190185, + "grad_norm": 0.5681651031083352, + "learning_rate": 3.2293388429752064e-07, + "logits/chosen": -0.185546875, + "logits/rejected": -0.396484375, + "logps/chosen": -408.0, + "logps/rejected": -430.0, + "loss": 0.0013, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.72265625, + "rewards/margins": 14.625, + "rewards/rejected": -13.875, + "step": 1924 + }, + { + "epoch": 2.361963190184049, + "grad_norm": 2.8639234150839723, + "learning_rate": 3.223140495867769e-07, + "logits/chosen": -0.09814453125, + "logits/rejected": -0.26171875, + "logps/chosen": -396.0, + "logps/rejected": -418.0, + "loss": 0.0037, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.91796875, + "rewards/margins": 13.0625, + "rewards/rejected": -14.0, + "step": 1925 + }, + { + "epoch": 2.36319018404908, + "grad_norm": 7.46778734641622, + "learning_rate": 3.2169421487603303e-07, + "logits/chosen": -0.357421875, + "logits/rejected": -0.56640625, + "logps/chosen": -408.0, + "logps/rejected": -462.0, + "loss": 0.0073, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.255859375, + "rewards/margins": 15.125, + "rewards/rejected": -15.375, + "step": 1926 + }, + { + "epoch": 2.3644171779141105, + "grad_norm": 0.875788459298664, + "learning_rate": 3.210743801652893e-07, + "logits/chosen": -0.2177734375, + "logits/rejected": -0.392578125, + "logps/chosen": -412.0, + "logps/rejected": -420.0, + "loss": 0.0017, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.0810546875, + "rewards/margins": 14.6875, + "rewards/rejected": -14.5625, + "step": 1927 + }, + { + "epoch": 2.365644171779141, + "grad_norm": 7.096467597027557, + "learning_rate": 3.204545454545454e-07, + "logits/chosen": -0.201171875, + "logits/rejected": -0.5234375, + "logps/chosen": -402.0, + "logps/rejected": -420.0, + "loss": 0.008, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.1513671875, + "rewards/margins": 14.6875, + "rewards/rejected": -14.5625, + "step": 1928 + }, + { + "epoch": 2.366871165644172, + "grad_norm": 2.2219881195246485, + "learning_rate": 3.1983471074380167e-07, + "logits/chosen": -0.205078125, + "logits/rejected": -0.44140625, + "logps/chosen": -414.0, + "logps/rejected": -436.0, + "loss": 0.0098, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.93359375, + "rewards/margins": 14.125, + "rewards/rejected": -15.0625, + "step": 1929 + }, + { + "epoch": 2.3680981595092025, + "grad_norm": 0.3825582992298532, + "learning_rate": 3.1921487603305786e-07, + "logits/chosen": -0.20703125, + "logits/rejected": -0.451171875, + "logps/chosen": -396.0, + "logps/rejected": -426.0, + "loss": 0.0008, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.224609375, + "rewards/margins": 14.625, + "rewards/rejected": -14.8125, + "step": 1930 + }, + { + "epoch": 2.369325153374233, + "grad_norm": 0.668744763611061, + "learning_rate": 3.1859504132231405e-07, + "logits/chosen": -0.2431640625, + "logits/rejected": -0.38671875, + "logps/chosen": -384.0, + "logps/rejected": -422.0, + "loss": 0.0015, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.66015625, + "rewards/margins": 14.25, + "rewards/rejected": -14.875, + "step": 1931 + }, + { + "epoch": 2.370552147239264, + "grad_norm": 0.9715291686066304, + "learning_rate": 3.1797520661157025e-07, + "logits/chosen": -0.17578125, + "logits/rejected": -0.306640625, + "logps/chosen": -386.0, + "logps/rejected": -462.0, + "loss": 0.0025, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.390625, + "rewards/margins": 13.8125, + "rewards/rejected": -14.1875, + "step": 1932 + }, + { + "epoch": 2.3717791411042946, + "grad_norm": 1.7678412908118337, + "learning_rate": 3.1735537190082644e-07, + "logits/chosen": -0.04833984375, + "logits/rejected": -0.291015625, + "logps/chosen": -402.0, + "logps/rejected": -390.0, + "loss": 0.003, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.04248046875, + "rewards/margins": 13.0625, + "rewards/rejected": -13.125, + "step": 1933 + }, + { + "epoch": 2.373006134969325, + "grad_norm": 1.752767561980237, + "learning_rate": 3.167355371900827e-07, + "logits/chosen": -0.185546875, + "logits/rejected": -0.384765625, + "logps/chosen": -442.0, + "logps/rejected": -476.0, + "loss": 0.0029, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.255859375, + "rewards/margins": 15.5, + "rewards/rejected": -15.75, + "step": 1934 + }, + { + "epoch": 2.374233128834356, + "grad_norm": 1.4713836570162997, + "learning_rate": 3.1611570247933883e-07, + "logits/chosen": -0.2119140625, + "logits/rejected": -0.419921875, + "logps/chosen": -404.0, + "logps/rejected": -426.0, + "loss": 0.0022, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.0595703125, + "rewards/margins": 15.6875, + "rewards/rejected": -15.625, + "step": 1935 + }, + { + "epoch": 2.3754601226993866, + "grad_norm": 0.4838355216666658, + "learning_rate": 3.154958677685951e-07, + "logits/chosen": -0.2890625, + "logits/rejected": -0.38671875, + "logps/chosen": -352.0, + "logps/rejected": -418.0, + "loss": 0.0066, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -1.109375, + "rewards/margins": 13.5625, + "rewards/rejected": -14.6875, + "step": 1936 + }, + { + "epoch": 2.3766871165644172, + "grad_norm": 1.6720937161714895, + "learning_rate": 3.148760330578512e-07, + "logits/chosen": -0.1806640625, + "logits/rejected": -0.345703125, + "logps/chosen": -414.0, + "logps/rejected": -410.0, + "loss": 0.003, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.61328125, + "rewards/margins": 13.6875, + "rewards/rejected": -14.25, + "step": 1937 + }, + { + "epoch": 2.377914110429448, + "grad_norm": 1.3353602251002044, + "learning_rate": 3.1425619834710747e-07, + "logits/chosen": -0.197265625, + "logits/rejected": -0.296875, + "logps/chosen": -408.0, + "logps/rejected": -464.0, + "loss": 0.0021, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.3203125, + "rewards/margins": 15.1875, + "rewards/rejected": -15.5, + "step": 1938 + }, + { + "epoch": 2.3791411042944786, + "grad_norm": 2.0160440052534137, + "learning_rate": 3.136363636363636e-07, + "logits/chosen": -0.369140625, + "logits/rejected": -0.55078125, + "logps/chosen": -398.0, + "logps/rejected": -438.0, + "loss": 0.003, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.06640625, + "rewards/margins": 15.75, + "rewards/rejected": -15.6875, + "step": 1939 + }, + { + "epoch": 2.3803680981595092, + "grad_norm": 0.7320868666683136, + "learning_rate": 3.1301652892561985e-07, + "logits/chosen": -0.1767578125, + "logits/rejected": -0.287109375, + "logps/chosen": -392.0, + "logps/rejected": -412.0, + "loss": 0.0019, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.63671875, + "rewards/margins": 13.125, + "rewards/rejected": -13.75, + "step": 1940 + }, + { + "epoch": 2.38159509202454, + "grad_norm": 1.2023531305817914, + "learning_rate": 3.1239669421487605e-07, + "logits/chosen": -0.26171875, + "logits/rejected": -0.435546875, + "logps/chosen": -424.0, + "logps/rejected": -422.0, + "loss": 0.0027, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.57421875, + "rewards/margins": 13.75, + "rewards/rejected": -14.375, + "step": 1941 + }, + { + "epoch": 2.3828220858895706, + "grad_norm": 1.4976939338912971, + "learning_rate": 3.1177685950413224e-07, + "logits/chosen": -0.07568359375, + "logits/rejected": -0.296875, + "logps/chosen": -386.0, + "logps/rejected": -398.0, + "loss": 0.004, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.90625, + "rewards/margins": 13.125, + "rewards/rejected": -14.0625, + "step": 1942 + }, + { + "epoch": 2.3840490797546012, + "grad_norm": 1.639539473533953, + "learning_rate": 3.1115702479338844e-07, + "logits/chosen": -0.294921875, + "logits/rejected": -0.5, + "logps/chosen": -452.0, + "logps/rejected": -450.0, + "loss": 0.0018, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.87890625, + "rewards/margins": 15.625, + "rewards/rejected": -14.75, + "step": 1943 + }, + { + "epoch": 2.385276073619632, + "grad_norm": 2.084180250827855, + "learning_rate": 3.1053719008264463e-07, + "logits/chosen": -0.2890625, + "logits/rejected": -0.39453125, + "logps/chosen": -404.0, + "logps/rejected": -440.0, + "loss": 0.0058, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.890625, + "rewards/margins": 14.0625, + "rewards/rejected": -14.9375, + "step": 1944 + }, + { + "epoch": 2.3865030674846626, + "grad_norm": 1.5235592282513357, + "learning_rate": 3.099173553719008e-07, + "logits/chosen": -0.30078125, + "logits/rejected": -0.54296875, + "logps/chosen": -428.0, + "logps/rejected": -442.0, + "loss": 0.0018, + "rewards/accuracies": 1.0, + "rewards/chosen": 1.3671875, + "rewards/margins": 16.375, + "rewards/rejected": -15.0, + "step": 1945 + }, + { + "epoch": 2.3877300613496932, + "grad_norm": 1.157851205840618, + "learning_rate": 3.09297520661157e-07, + "logits/chosen": -0.0172119140625, + "logits/rejected": -0.2314453125, + "logps/chosen": -400.0, + "logps/rejected": -424.0, + "loss": 0.0075, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.73046875, + "rewards/margins": 13.625, + "rewards/rejected": -14.375, + "step": 1946 + }, + { + "epoch": 2.388957055214724, + "grad_norm": 1.3732060276308538, + "learning_rate": 3.086776859504132e-07, + "logits/chosen": -0.318359375, + "logits/rejected": -0.54296875, + "logps/chosen": -420.0, + "logps/rejected": -414.0, + "loss": 0.0024, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.216796875, + "rewards/margins": 14.75, + "rewards/rejected": -15.0, + "step": 1947 + }, + { + "epoch": 2.3901840490797546, + "grad_norm": 4.8280339618051515, + "learning_rate": 3.080578512396694e-07, + "logits/chosen": -0.1611328125, + "logits/rejected": -0.369140625, + "logps/chosen": -414.0, + "logps/rejected": -406.0, + "loss": 0.007, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.00970458984375, + "rewards/margins": 13.9375, + "rewards/rejected": -13.9375, + "step": 1948 + }, + { + "epoch": 2.3914110429447852, + "grad_norm": 1.3571589198753051, + "learning_rate": 3.074380165289256e-07, + "logits/chosen": -0.1826171875, + "logits/rejected": -0.4375, + "logps/chosen": -428.0, + "logps/rejected": -460.0, + "loss": 0.0022, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.77734375, + "rewards/margins": 15.1875, + "rewards/rejected": -16.0, + "step": 1949 + }, + { + "epoch": 2.392638036809816, + "grad_norm": 1.580900843285348, + "learning_rate": 3.0681818181818185e-07, + "logits/chosen": -0.1630859375, + "logits/rejected": -0.310546875, + "logps/chosen": -386.0, + "logps/rejected": -552.0, + "loss": 0.0052, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.65234375, + "rewards/margins": 26.875, + "rewards/rejected": -27.5, + "step": 1950 + }, + { + "epoch": 2.3938650306748466, + "grad_norm": 1.7706821147516993, + "learning_rate": 3.0619834710743804e-07, + "logits/chosen": -0.0166015625, + "logits/rejected": -0.23046875, + "logps/chosen": -396.0, + "logps/rejected": -366.0, + "loss": 0.0084, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -1.078125, + "rewards/margins": 12.1875, + "rewards/rejected": -13.3125, + "step": 1951 + }, + { + "epoch": 2.3950920245398772, + "grad_norm": 1.4789095463546456, + "learning_rate": 3.0557851239669424e-07, + "logits/chosen": -0.1982421875, + "logits/rejected": -0.3984375, + "logps/chosen": -446.0, + "logps/rejected": -466.0, + "loss": 0.0034, + "rewards/accuracies": 1.0, + "rewards/chosen": -1.203125, + "rewards/margins": 14.6875, + "rewards/rejected": -15.875, + "step": 1952 + }, + { + "epoch": 2.396319018404908, + "grad_norm": 0.9487876379983259, + "learning_rate": 3.0495867768595043e-07, + "logits/chosen": -0.123046875, + "logits/rejected": -0.388671875, + "logps/chosen": -396.0, + "logps/rejected": -438.0, + "loss": 0.0015, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.71875, + "rewards/margins": 14.8125, + "rewards/rejected": -15.5, + "step": 1953 + }, + { + "epoch": 2.3975460122699386, + "grad_norm": 0.7725711719339576, + "learning_rate": 3.0433884297520663e-07, + "logits/chosen": -0.263671875, + "logits/rejected": -0.3515625, + "logps/chosen": -396.0, + "logps/rejected": -450.0, + "loss": 0.0017, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.470703125, + "rewards/margins": 15.125, + "rewards/rejected": -15.625, + "step": 1954 + }, + { + "epoch": 2.3987730061349692, + "grad_norm": 0.4862146856103838, + "learning_rate": 3.037190082644628e-07, + "logits/chosen": -0.21875, + "logits/rejected": -0.42578125, + "logps/chosen": -456.0, + "logps/rejected": -446.0, + "loss": 0.0067, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.33203125, + "rewards/margins": 14.6875, + "rewards/rejected": -15.0, + "step": 1955 + }, + { + "epoch": 2.4, + "grad_norm": 2.2594957735655226, + "learning_rate": 3.03099173553719e-07, + "logits/chosen": -0.2275390625, + "logits/rejected": -0.32421875, + "logps/chosen": -410.0, + "logps/rejected": -490.0, + "loss": 0.0039, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.69921875, + "rewards/margins": 15.75, + "rewards/rejected": -16.5, + "step": 1956 + }, + { + "epoch": 2.4012269938650306, + "grad_norm": 1.1736544713806183, + "learning_rate": 3.024793388429752e-07, + "logits/chosen": -0.201171875, + "logits/rejected": -0.38671875, + "logps/chosen": -392.0, + "logps/rejected": -432.0, + "loss": 0.0027, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.62890625, + "rewards/margins": 14.3125, + "rewards/rejected": -14.9375, + "step": 1957 + }, + { + "epoch": 2.4024539877300612, + "grad_norm": 1.8801304080399017, + "learning_rate": 3.018595041322314e-07, + "logits/chosen": -0.23828125, + "logits/rejected": -0.41796875, + "logps/chosen": -418.0, + "logps/rejected": -456.0, + "loss": 0.0044, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.46484375, + "rewards/margins": 14.1875, + "rewards/rejected": -14.625, + "step": 1958 + }, + { + "epoch": 2.403680981595092, + "grad_norm": 1.1381884472514572, + "learning_rate": 3.012396694214876e-07, + "logits/chosen": -0.1982421875, + "logits/rejected": -0.34765625, + "logps/chosen": -362.0, + "logps/rejected": -406.0, + "loss": 0.0023, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.68359375, + "rewards/margins": 14.375, + "rewards/rejected": -15.0625, + "step": 1959 + }, + { + "epoch": 2.4049079754601226, + "grad_norm": 7.504238565996868, + "learning_rate": 3.006198347107438e-07, + "logits/chosen": -0.283203125, + "logits/rejected": -0.46484375, + "logps/chosen": -398.0, + "logps/rejected": -418.0, + "loss": 0.0141, + "rewards/accuracies": 0.984375, + "rewards/chosen": -1.03125, + "rewards/margins": 14.625, + "rewards/rejected": -15.625, + "step": 1960 + }, + { + "epoch": 2.4061349693251532, + "grad_norm": 0.9284716393585608, + "learning_rate": 3.0000000000000004e-07, + "logits/chosen": -0.3046875, + "logits/rejected": -0.34375, + "logps/chosen": -422.0, + "logps/rejected": -432.0, + "loss": 0.007, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.671875, + "rewards/margins": 15.25, + "rewards/rejected": -15.9375, + "step": 1961 + }, + { + "epoch": 2.407361963190184, + "grad_norm": 0.8816543500250089, + "learning_rate": 2.993801652892562e-07, + "logits/chosen": -0.16015625, + "logits/rejected": -0.33203125, + "logps/chosen": -382.0, + "logps/rejected": -430.0, + "loss": 0.0015, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.64453125, + "rewards/margins": 15.0625, + "rewards/rejected": -15.6875, + "step": 1962 + }, + { + "epoch": 2.4085889570552146, + "grad_norm": 12.815101787728196, + "learning_rate": 2.9876033057851243e-07, + "logits/chosen": -0.107421875, + "logits/rejected": -0.4296875, + "logps/chosen": -436.0, + "logps/rejected": -426.0, + "loss": 0.011, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.61328125, + "rewards/margins": 15.0, + "rewards/rejected": -14.375, + "step": 1963 + }, + { + "epoch": 2.4098159509202453, + "grad_norm": 1.8385191749147407, + "learning_rate": 2.9814049586776857e-07, + "logits/chosen": -0.318359375, + "logits/rejected": -0.5390625, + "logps/chosen": -438.0, + "logps/rejected": -468.0, + "loss": 0.0019, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.361328125, + "rewards/margins": 15.875, + "rewards/rejected": -16.25, + "step": 1964 + }, + { + "epoch": 2.411042944785276, + "grad_norm": 0.736650496296312, + "learning_rate": 2.975206611570248e-07, + "logits/chosen": -0.1298828125, + "logits/rejected": -0.1875, + "logps/chosen": -354.0, + "logps/rejected": -412.0, + "loss": 0.0016, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.3125, + "rewards/margins": 13.9375, + "rewards/rejected": -14.25, + "step": 1965 + }, + { + "epoch": 2.4122699386503066, + "grad_norm": 0.6806551225371621, + "learning_rate": 2.96900826446281e-07, + "logits/chosen": -0.279296875, + "logits/rejected": -0.5, + "logps/chosen": -382.0, + "logps/rejected": -412.0, + "loss": 0.0014, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.1591796875, + "rewards/margins": 15.0, + "rewards/rejected": -14.875, + "step": 1966 + }, + { + "epoch": 2.4134969325153373, + "grad_norm": 1.790516928878384, + "learning_rate": 2.962809917355372e-07, + "logits/chosen": -0.275390625, + "logits/rejected": -0.455078125, + "logps/chosen": -394.0, + "logps/rejected": -406.0, + "loss": 0.0082, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -1.046875, + "rewards/margins": 14.6875, + "rewards/rejected": -15.75, + "step": 1967 + }, + { + "epoch": 2.414723926380368, + "grad_norm": 1.4054278646534835, + "learning_rate": 2.956611570247934e-07, + "logits/chosen": -0.279296875, + "logits/rejected": -0.5078125, + "logps/chosen": -388.0, + "logps/rejected": -412.0, + "loss": 0.0029, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.83203125, + "rewards/margins": 14.0625, + "rewards/rejected": -14.9375, + "step": 1968 + }, + { + "epoch": 2.4159509202453986, + "grad_norm": 0.9848806716913605, + "learning_rate": 2.950413223140496e-07, + "logits/chosen": -0.236328125, + "logits/rejected": -0.435546875, + "logps/chosen": -418.0, + "logps/rejected": -424.0, + "loss": 0.002, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.005279541015625, + "rewards/margins": 14.375, + "rewards/rejected": -14.375, + "step": 1969 + }, + { + "epoch": 2.4171779141104293, + "grad_norm": 1.8131938186785144, + "learning_rate": 2.944214876033058e-07, + "logits/chosen": -0.1259765625, + "logits/rejected": -0.25390625, + "logps/chosen": -392.0, + "logps/rejected": -436.0, + "loss": 0.0085, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.58203125, + "rewards/margins": 14.8125, + "rewards/rejected": -15.375, + "step": 1970 + }, + { + "epoch": 2.41840490797546, + "grad_norm": 4.9515423580873135, + "learning_rate": 2.93801652892562e-07, + "logits/chosen": -0.16796875, + "logits/rejected": -0.330078125, + "logps/chosen": -364.0, + "logps/rejected": -400.0, + "loss": 0.0068, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.72265625, + "rewards/margins": 12.75, + "rewards/rejected": -13.5, + "step": 1971 + }, + { + "epoch": 2.419631901840491, + "grad_norm": 0.4703596116272296, + "learning_rate": 2.9318181818181823e-07, + "logits/chosen": -0.279296875, + "logits/rejected": -0.419921875, + "logps/chosen": -410.0, + "logps/rejected": -426.0, + "loss": 0.0065, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.08056640625, + "rewards/margins": 15.1875, + "rewards/rejected": -15.3125, + "step": 1972 + }, + { + "epoch": 2.4208588957055213, + "grad_norm": 0.4695810291596011, + "learning_rate": 2.9256198347107437e-07, + "logits/chosen": -0.2041015625, + "logits/rejected": -0.263671875, + "logps/chosen": -398.0, + "logps/rejected": -422.0, + "loss": 0.0009, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.75390625, + "rewards/margins": 14.3125, + "rewards/rejected": -15.0625, + "step": 1973 + }, + { + "epoch": 2.4220858895705524, + "grad_norm": 1.9210187853733611, + "learning_rate": 2.919421487603306e-07, + "logits/chosen": -0.236328125, + "logits/rejected": -0.43359375, + "logps/chosen": -446.0, + "logps/rejected": -454.0, + "loss": 0.0036, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.0732421875, + "rewards/margins": 15.5625, + "rewards/rejected": -15.625, + "step": 1974 + }, + { + "epoch": 2.4233128834355826, + "grad_norm": 11.30754843320625, + "learning_rate": 2.9132231404958676e-07, + "logits/chosen": -0.30859375, + "logits/rejected": -0.53515625, + "logps/chosen": -442.0, + "logps/rejected": -458.0, + "loss": 0.0111, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.029541015625, + "rewards/margins": 14.5625, + "rewards/rejected": -14.5625, + "step": 1975 + }, + { + "epoch": 2.4245398773006137, + "grad_norm": 2.5124120113261106, + "learning_rate": 2.90702479338843e-07, + "logits/chosen": -0.2470703125, + "logits/rejected": -0.34765625, + "logps/chosen": -370.0, + "logps/rejected": -410.0, + "loss": 0.0085, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.6796875, + "rewards/margins": 13.9375, + "rewards/rejected": -14.625, + "step": 1976 + }, + { + "epoch": 2.425766871165644, + "grad_norm": 1.3444304267954004, + "learning_rate": 2.9008264462809915e-07, + "logits/chosen": -0.072265625, + "logits/rejected": -0.2421875, + "logps/chosen": -370.0, + "logps/rejected": -400.0, + "loss": 0.0019, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.439453125, + "rewards/margins": 13.3125, + "rewards/rejected": -13.75, + "step": 1977 + }, + { + "epoch": 2.426993865030675, + "grad_norm": 0.27740764610834023, + "learning_rate": 2.894628099173554e-07, + "logits/chosen": -0.2734375, + "logits/rejected": -0.5078125, + "logps/chosen": -470.0, + "logps/rejected": -496.0, + "loss": 0.0006, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.66015625, + "rewards/margins": 15.8125, + "rewards/rejected": -16.5, + "step": 1978 + }, + { + "epoch": 2.4282208588957057, + "grad_norm": 7.304769063519445, + "learning_rate": 2.888429752066116e-07, + "logits/chosen": -0.1689453125, + "logits/rejected": -0.453125, + "logps/chosen": -422.0, + "logps/rejected": -430.0, + "loss": 0.0092, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.76171875, + "rewards/margins": 14.9375, + "rewards/rejected": -15.6875, + "step": 1979 + }, + { + "epoch": 2.4294478527607364, + "grad_norm": 3.058491087110553, + "learning_rate": 2.882231404958678e-07, + "logits/chosen": -0.25390625, + "logits/rejected": -0.46875, + "logps/chosen": -388.0, + "logps/rejected": -428.0, + "loss": 0.0035, + "rewards/accuracies": 1.0, + "rewards/chosen": -1.1875, + "rewards/margins": 14.3125, + "rewards/rejected": -15.5, + "step": 1980 + }, + { + "epoch": 2.430674846625767, + "grad_norm": 5.174599557000674, + "learning_rate": 2.87603305785124e-07, + "logits/chosen": -0.173828125, + "logits/rejected": -0.294921875, + "logps/chosen": -428.0, + "logps/rejected": -454.0, + "loss": 0.0049, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.71875, + "rewards/margins": 13.875, + "rewards/rejected": -14.5625, + "step": 1981 + }, + { + "epoch": 2.4319018404907977, + "grad_norm": 4.729409247624802, + "learning_rate": 2.8698347107438017e-07, + "logits/chosen": -0.265625, + "logits/rejected": -0.421875, + "logps/chosen": -398.0, + "logps/rejected": -442.0, + "loss": 0.0043, + "rewards/accuracies": 1.0, + "rewards/chosen": -1.03125, + "rewards/margins": 14.125, + "rewards/rejected": -15.1875, + "step": 1982 + }, + { + "epoch": 2.4331288343558284, + "grad_norm": 1.4248004355619839, + "learning_rate": 2.8636363636363637e-07, + "logits/chosen": -0.169921875, + "logits/rejected": -0.40625, + "logps/chosen": -460.0, + "logps/rejected": -486.0, + "loss": 0.0022, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.052978515625, + "rewards/margins": 15.8125, + "rewards/rejected": -15.875, + "step": 1983 + }, + { + "epoch": 2.434355828220859, + "grad_norm": 1.883193707596539, + "learning_rate": 2.8574380165289256e-07, + "logits/chosen": -0.2021484375, + "logits/rejected": -0.349609375, + "logps/chosen": -410.0, + "logps/rejected": -468.0, + "loss": 0.0034, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.484375, + "rewards/margins": 14.75, + "rewards/rejected": -15.1875, + "step": 1984 + }, + { + "epoch": 2.4355828220858897, + "grad_norm": 0.9787240890391893, + "learning_rate": 2.8512396694214875e-07, + "logits/chosen": -0.25, + "logits/rejected": -0.57421875, + "logps/chosen": -456.0, + "logps/rejected": -426.0, + "loss": 0.0017, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.115234375, + "rewards/margins": 15.5, + "rewards/rejected": -15.375, + "step": 1985 + }, + { + "epoch": 2.4368098159509204, + "grad_norm": 1.59723611476287, + "learning_rate": 2.8450413223140495e-07, + "logits/chosen": -0.37109375, + "logits/rejected": -0.578125, + "logps/chosen": -452.0, + "logps/rejected": -454.0, + "loss": 0.002, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.060791015625, + "rewards/margins": 15.625, + "rewards/rejected": -15.5625, + "step": 1986 + }, + { + "epoch": 2.438036809815951, + "grad_norm": 1.2822330375215538, + "learning_rate": 2.8388429752066114e-07, + "logits/chosen": -0.1806640625, + "logits/rejected": -0.27734375, + "logps/chosen": -364.0, + "logps/rejected": -398.0, + "loss": 0.0032, + "rewards/accuracies": 1.0, + "rewards/chosen": -1.140625, + "rewards/margins": 14.3125, + "rewards/rejected": -15.5, + "step": 1987 + }, + { + "epoch": 2.4392638036809817, + "grad_norm": 0.5467876144186856, + "learning_rate": 2.8326446280991734e-07, + "logits/chosen": -0.203125, + "logits/rejected": -0.306640625, + "logps/chosen": -372.0, + "logps/rejected": -428.0, + "loss": 0.001, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.5234375, + "rewards/margins": 14.75, + "rewards/rejected": -15.25, + "step": 1988 + }, + { + "epoch": 2.4404907975460124, + "grad_norm": 0.6747474333913942, + "learning_rate": 2.826446280991736e-07, + "logits/chosen": -0.240234375, + "logits/rejected": -0.52734375, + "logps/chosen": -416.0, + "logps/rejected": -444.0, + "loss": 0.0011, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.09375, + "rewards/margins": 15.875, + "rewards/rejected": -16.0, + "step": 1989 + }, + { + "epoch": 2.441717791411043, + "grad_norm": 0.6215752360767873, + "learning_rate": 2.820247933884298e-07, + "logits/chosen": -0.2158203125, + "logits/rejected": -0.42578125, + "logps/chosen": -410.0, + "logps/rejected": -420.0, + "loss": 0.0013, + "rewards/accuracies": 1.0, + "rewards/chosen": -1.109375, + "rewards/margins": 14.5, + "rewards/rejected": -15.625, + "step": 1990 + }, + { + "epoch": 2.4429447852760737, + "grad_norm": 0.8538306307702517, + "learning_rate": 2.8140495867768597e-07, + "logits/chosen": -0.1396484375, + "logits/rejected": -0.341796875, + "logps/chosen": -354.0, + "logps/rejected": -412.0, + "loss": 0.0019, + "rewards/accuracies": 1.0, + "rewards/chosen": -1.0234375, + "rewards/margins": 14.5, + "rewards/rejected": -15.5, + "step": 1991 + }, + { + "epoch": 2.4441717791411044, + "grad_norm": 4.111258032415621, + "learning_rate": 2.8078512396694217e-07, + "logits/chosen": -0.169921875, + "logits/rejected": -0.3984375, + "logps/chosen": -420.0, + "logps/rejected": -404.0, + "loss": 0.007, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.1435546875, + "rewards/margins": 15.0, + "rewards/rejected": -15.1875, + "step": 1992 + }, + { + "epoch": 2.445398773006135, + "grad_norm": 0.6697009083922223, + "learning_rate": 2.8016528925619836e-07, + "logits/chosen": -0.087890625, + "logits/rejected": -0.267578125, + "logps/chosen": -366.0, + "logps/rejected": -404.0, + "loss": 0.0017, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.74609375, + "rewards/margins": 14.125, + "rewards/rejected": -14.875, + "step": 1993 + }, + { + "epoch": 2.4466257668711657, + "grad_norm": 1.8112304031279607, + "learning_rate": 2.7954545454545455e-07, + "logits/chosen": -0.2353515625, + "logits/rejected": -0.34375, + "logps/chosen": -384.0, + "logps/rejected": -436.0, + "loss": 0.0026, + "rewards/accuracies": 1.0, + "rewards/chosen": -1.1796875, + "rewards/margins": 14.625, + "rewards/rejected": -15.8125, + "step": 1994 + }, + { + "epoch": 2.4478527607361964, + "grad_norm": 1.5752098541297537, + "learning_rate": 2.7892561983471075e-07, + "logits/chosen": -0.2275390625, + "logits/rejected": -0.48828125, + "logps/chosen": -464.0, + "logps/rejected": -446.0, + "loss": 0.003, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.205078125, + "rewards/margins": 15.375, + "rewards/rejected": -15.1875, + "step": 1995 + }, + { + "epoch": 2.449079754601227, + "grad_norm": 1.9894113620252913, + "learning_rate": 2.7830578512396694e-07, + "logits/chosen": -0.23828125, + "logits/rejected": -0.3984375, + "logps/chosen": -356.0, + "logps/rejected": -424.0, + "loss": 0.0034, + "rewards/accuracies": 1.0, + "rewards/chosen": -1.171875, + "rewards/margins": 14.0, + "rewards/rejected": -15.1875, + "step": 1996 + }, + { + "epoch": 2.4503067484662577, + "grad_norm": 1.4320841926513423, + "learning_rate": 2.7768595041322314e-07, + "logits/chosen": -0.2255859375, + "logits/rejected": -0.3828125, + "logps/chosen": -428.0, + "logps/rejected": -442.0, + "loss": 0.0031, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.140625, + "rewards/margins": 15.0, + "rewards/rejected": -15.125, + "step": 1997 + }, + { + "epoch": 2.4515337423312884, + "grad_norm": 1.3953183068592212, + "learning_rate": 2.7706611570247933e-07, + "logits/chosen": -0.1669921875, + "logits/rejected": -0.314453125, + "logps/chosen": -376.0, + "logps/rejected": -428.0, + "loss": 0.0019, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.2392578125, + "rewards/margins": 14.875, + "rewards/rejected": -15.125, + "step": 1998 + }, + { + "epoch": 2.452760736196319, + "grad_norm": 2.574772412655261, + "learning_rate": 2.764462809917356e-07, + "logits/chosen": -0.23828125, + "logits/rejected": -0.404296875, + "logps/chosen": -366.0, + "logps/rejected": -392.0, + "loss": 0.0079, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.326171875, + "rewards/margins": 13.5, + "rewards/rejected": -13.8125, + "step": 1999 + }, + { + "epoch": 2.4539877300613497, + "grad_norm": 1.5846974270863767, + "learning_rate": 2.758264462809917e-07, + "logits/chosen": -0.26171875, + "logits/rejected": -0.439453125, + "logps/chosen": -382.0, + "logps/rejected": -412.0, + "loss": 0.0037, + "rewards/accuracies": 1.0, + "rewards/chosen": -1.171875, + "rewards/margins": 13.9375, + "rewards/rejected": -15.125, + "step": 2000 + }, + { + "epoch": 2.4552147239263804, + "grad_norm": 0.7523782989958074, + "learning_rate": 2.7520661157024797e-07, + "logits/chosen": -0.1689453125, + "logits/rejected": -0.4609375, + "logps/chosen": -418.0, + "logps/rejected": -390.0, + "loss": 0.0018, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.51171875, + "rewards/margins": 14.5, + "rewards/rejected": -15.0, + "step": 2001 + }, + { + "epoch": 2.456441717791411, + "grad_norm": 2.5149188293502993, + "learning_rate": 2.745867768595041e-07, + "logits/chosen": -0.1689453125, + "logits/rejected": -0.33984375, + "logps/chosen": -392.0, + "logps/rejected": -400.0, + "loss": 0.0033, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.38671875, + "rewards/margins": 14.1875, + "rewards/rejected": -14.625, + "step": 2002 + }, + { + "epoch": 2.4576687116564417, + "grad_norm": 1.5231745953388507, + "learning_rate": 2.7396694214876036e-07, + "logits/chosen": -0.1826171875, + "logits/rejected": -0.3828125, + "logps/chosen": -388.0, + "logps/rejected": -446.0, + "loss": 0.0045, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.9609375, + "rewards/margins": 13.8125, + "rewards/rejected": -14.8125, + "step": 2003 + }, + { + "epoch": 2.4588957055214724, + "grad_norm": 1.235269981754148, + "learning_rate": 2.7334710743801655e-07, + "logits/chosen": -0.220703125, + "logits/rejected": -0.31640625, + "logps/chosen": -416.0, + "logps/rejected": -462.0, + "loss": 0.0035, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.53515625, + "rewards/margins": 13.5625, + "rewards/rejected": -14.125, + "step": 2004 + }, + { + "epoch": 2.460122699386503, + "grad_norm": 0.717451116297991, + "learning_rate": 2.7272727272727274e-07, + "logits/chosen": -0.3359375, + "logits/rejected": -0.42578125, + "logps/chosen": -374.0, + "logps/rejected": -426.0, + "loss": 0.0017, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.404296875, + "rewards/margins": 14.5625, + "rewards/rejected": -14.9375, + "step": 2005 + }, + { + "epoch": 2.4613496932515337, + "grad_norm": 1.126459905689613, + "learning_rate": 2.7210743801652894e-07, + "logits/chosen": -0.1181640625, + "logits/rejected": -0.2431640625, + "logps/chosen": -382.0, + "logps/rejected": -420.0, + "loss": 0.0019, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.197265625, + "rewards/margins": 14.4375, + "rewards/rejected": -14.25, + "step": 2006 + }, + { + "epoch": 2.4625766871165644, + "grad_norm": 1.1444414537455012, + "learning_rate": 2.7148760330578513e-07, + "logits/chosen": -0.259765625, + "logits/rejected": -0.4296875, + "logps/chosen": -406.0, + "logps/rejected": -448.0, + "loss": 0.0026, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.72265625, + "rewards/margins": 14.375, + "rewards/rejected": -15.0625, + "step": 2007 + }, + { + "epoch": 2.463803680981595, + "grad_norm": 1.5243527341169045, + "learning_rate": 2.708677685950413e-07, + "logits/chosen": -0.310546875, + "logits/rejected": -0.431640625, + "logps/chosen": -362.0, + "logps/rejected": -380.0, + "loss": 0.0041, + "rewards/accuracies": 1.0, + "rewards/chosen": -1.234375, + "rewards/margins": 13.5, + "rewards/rejected": -14.75, + "step": 2008 + }, + { + "epoch": 2.4650306748466257, + "grad_norm": 0.7303614810508574, + "learning_rate": 2.702479338842975e-07, + "logits/chosen": -0.34765625, + "logits/rejected": -0.5859375, + "logps/chosen": -368.0, + "logps/rejected": -412.0, + "loss": 0.0011, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.6953125, + "rewards/margins": 15.3125, + "rewards/rejected": -16.0, + "step": 2009 + }, + { + "epoch": 2.4662576687116564, + "grad_norm": 2.563268337915507, + "learning_rate": 2.6962809917355377e-07, + "logits/chosen": -0.2080078125, + "logits/rejected": -0.3984375, + "logps/chosen": -404.0, + "logps/rejected": -410.0, + "loss": 0.0064, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.62109375, + "rewards/margins": 14.125, + "rewards/rejected": -14.75, + "step": 2010 + }, + { + "epoch": 2.467484662576687, + "grad_norm": 0.9480540833382409, + "learning_rate": 2.690082644628099e-07, + "logits/chosen": -0.16796875, + "logits/rejected": -0.287109375, + "logps/chosen": -368.0, + "logps/rejected": -402.0, + "loss": 0.0018, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.478515625, + "rewards/margins": 13.625, + "rewards/rejected": -14.125, + "step": 2011 + }, + { + "epoch": 2.4687116564417177, + "grad_norm": 9.64198444729727, + "learning_rate": 2.6838842975206616e-07, + "logits/chosen": -0.171875, + "logits/rejected": -0.384765625, + "logps/chosen": -394.0, + "logps/rejected": -452.0, + "loss": 0.0113, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.7265625, + "rewards/margins": 14.75, + "rewards/rejected": -15.4375, + "step": 2012 + }, + { + "epoch": 2.4699386503067484, + "grad_norm": 0.4583096256892898, + "learning_rate": 2.677685950413223e-07, + "logits/chosen": -0.384765625, + "logits/rejected": -0.53125, + "logps/chosen": -430.0, + "logps/rejected": -452.0, + "loss": 0.0012, + "rewards/accuracies": 1.0, + "rewards/chosen": -1.0859375, + "rewards/margins": 15.6875, + "rewards/rejected": -16.75, + "step": 2013 + }, + { + "epoch": 2.471165644171779, + "grad_norm": 1.0527189432725, + "learning_rate": 2.6714876033057854e-07, + "logits/chosen": -0.28515625, + "logits/rejected": -0.458984375, + "logps/chosen": -422.0, + "logps/rejected": -468.0, + "loss": 0.0022, + "rewards/accuracies": 1.0, + "rewards/chosen": -1.1953125, + "rewards/margins": 15.0625, + "rewards/rejected": -16.25, + "step": 2014 + }, + { + "epoch": 2.4723926380368098, + "grad_norm": 0.2536993017828806, + "learning_rate": 2.665289256198347e-07, + "logits/chosen": -0.330078125, + "logits/rejected": -0.578125, + "logps/chosen": -392.0, + "logps/rejected": -412.0, + "loss": 0.0007, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.024658203125, + "rewards/margins": 14.5, + "rewards/rejected": -14.5, + "step": 2015 + }, + { + "epoch": 2.4736196319018404, + "grad_norm": 2.1872047343298986, + "learning_rate": 2.6590909090909093e-07, + "logits/chosen": -0.314453125, + "logits/rejected": -0.51171875, + "logps/chosen": -454.0, + "logps/rejected": -424.0, + "loss": 0.0101, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.1669921875, + "rewards/margins": 14.5625, + "rewards/rejected": -14.75, + "step": 2016 + }, + { + "epoch": 2.474846625766871, + "grad_norm": 1.2415112289700299, + "learning_rate": 2.652892561983471e-07, + "logits/chosen": -0.2578125, + "logits/rejected": -0.486328125, + "logps/chosen": -406.0, + "logps/rejected": -434.0, + "loss": 0.0028, + "rewards/accuracies": 1.0, + "rewards/chosen": -1.046875, + "rewards/margins": 14.875, + "rewards/rejected": -15.9375, + "step": 2017 + }, + { + "epoch": 2.4760736196319018, + "grad_norm": 0.7639345040393662, + "learning_rate": 2.646694214876033e-07, + "logits/chosen": -0.208984375, + "logits/rejected": -0.376953125, + "logps/chosen": -408.0, + "logps/rejected": -418.0, + "loss": 0.0017, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.5625, + "rewards/margins": 15.125, + "rewards/rejected": -15.6875, + "step": 2018 + }, + { + "epoch": 2.4773006134969324, + "grad_norm": 1.261896420221238, + "learning_rate": 2.640495867768595e-07, + "logits/chosen": -0.279296875, + "logits/rejected": -0.384765625, + "logps/chosen": -374.0, + "logps/rejected": -422.0, + "loss": 0.002, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.9375, + "rewards/margins": 15.125, + "rewards/rejected": -16.0, + "step": 2019 + }, + { + "epoch": 2.478527607361963, + "grad_norm": 0.4525423700655121, + "learning_rate": 2.634297520661157e-07, + "logits/chosen": -0.353515625, + "logits/rejected": -0.5625, + "logps/chosen": -388.0, + "logps/rejected": -436.0, + "loss": 0.0011, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.75, + "rewards/margins": 14.875, + "rewards/rejected": -15.625, + "step": 2020 + }, + { + "epoch": 2.4797546012269938, + "grad_norm": 1.5511243340957956, + "learning_rate": 2.628099173553719e-07, + "logits/chosen": -0.1171875, + "logits/rejected": -0.33203125, + "logps/chosen": -420.0, + "logps/rejected": -436.0, + "loss": 0.0026, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.62109375, + "rewards/margins": 14.9375, + "rewards/rejected": -15.5625, + "step": 2021 + }, + { + "epoch": 2.4809815950920244, + "grad_norm": 1.1231814728778773, + "learning_rate": 2.621900826446281e-07, + "logits/chosen": -0.18359375, + "logits/rejected": -0.265625, + "logps/chosen": -400.0, + "logps/rejected": -426.0, + "loss": 0.0031, + "rewards/accuracies": 1.0, + "rewards/chosen": -1.671875, + "rewards/margins": 13.3125, + "rewards/rejected": -15.0, + "step": 2022 + }, + { + "epoch": 2.482208588957055, + "grad_norm": 0.8587184320408277, + "learning_rate": 2.615702479338843e-07, + "logits/chosen": -0.259765625, + "logits/rejected": -0.3984375, + "logps/chosen": -424.0, + "logps/rejected": -446.0, + "loss": 0.0021, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.765625, + "rewards/margins": 13.5625, + "rewards/rejected": -14.3125, + "step": 2023 + }, + { + "epoch": 2.4834355828220858, + "grad_norm": 4.320792284496451, + "learning_rate": 2.609504132231405e-07, + "logits/chosen": -0.1796875, + "logits/rejected": -0.384765625, + "logps/chosen": -386.0, + "logps/rejected": -476.0, + "loss": 0.0063, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.59375, + "rewards/margins": 15.1875, + "rewards/rejected": -15.75, + "step": 2024 + }, + { + "epoch": 2.4846625766871164, + "grad_norm": 1.5287067726408572, + "learning_rate": 2.6033057851239673e-07, + "logits/chosen": -0.365234375, + "logits/rejected": -0.51171875, + "logps/chosen": -364.0, + "logps/rejected": -412.0, + "loss": 0.0076, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -1.625, + "rewards/margins": 13.6875, + "rewards/rejected": -15.3125, + "step": 2025 + }, + { + "epoch": 2.485889570552147, + "grad_norm": 0.7720479692737491, + "learning_rate": 2.597107438016529e-07, + "logits/chosen": -0.228515625, + "logits/rejected": -0.447265625, + "logps/chosen": -412.0, + "logps/rejected": -460.0, + "loss": 0.0018, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.8671875, + "rewards/margins": 16.125, + "rewards/rejected": -17.0, + "step": 2026 + }, + { + "epoch": 2.4871165644171778, + "grad_norm": 0.5192675103765902, + "learning_rate": 2.590909090909091e-07, + "logits/chosen": -0.2158203125, + "logits/rejected": -0.373046875, + "logps/chosen": -388.0, + "logps/rejected": -410.0, + "loss": 0.0012, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.671875, + "rewards/margins": 15.5, + "rewards/rejected": -16.25, + "step": 2027 + }, + { + "epoch": 2.4883435582822084, + "grad_norm": 1.303250774548485, + "learning_rate": 2.5847107438016526e-07, + "logits/chosen": -0.24609375, + "logits/rejected": -0.470703125, + "logps/chosen": -406.0, + "logps/rejected": -452.0, + "loss": 0.0026, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.84375, + "rewards/margins": 15.8125, + "rewards/rejected": -16.625, + "step": 2028 + }, + { + "epoch": 2.489570552147239, + "grad_norm": 0.7484743520297892, + "learning_rate": 2.578512396694215e-07, + "logits/chosen": -0.296875, + "logits/rejected": -0.470703125, + "logps/chosen": -432.0, + "logps/rejected": -438.0, + "loss": 0.0015, + "rewards/accuracies": 1.0, + "rewards/chosen": -1.6640625, + "rewards/margins": 13.8125, + "rewards/rejected": -15.5, + "step": 2029 + }, + { + "epoch": 2.4907975460122698, + "grad_norm": 0.9222395400754245, + "learning_rate": 2.572314049586777e-07, + "logits/chosen": -0.333984375, + "logits/rejected": -0.51171875, + "logps/chosen": -410.0, + "logps/rejected": -442.0, + "loss": 0.0011, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.67578125, + "rewards/margins": 15.5625, + "rewards/rejected": -16.25, + "step": 2030 + }, + { + "epoch": 2.4920245398773004, + "grad_norm": 1.9312656839256892, + "learning_rate": 2.566115702479339e-07, + "logits/chosen": -0.291015625, + "logits/rejected": -0.416015625, + "logps/chosen": -386.0, + "logps/rejected": -452.0, + "loss": 0.0038, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.97265625, + "rewards/margins": 14.4375, + "rewards/rejected": -15.4375, + "step": 2031 + }, + { + "epoch": 2.493251533742331, + "grad_norm": 1.0973412879077933, + "learning_rate": 2.559917355371901e-07, + "logits/chosen": -0.2158203125, + "logits/rejected": -0.412109375, + "logps/chosen": -382.0, + "logps/rejected": -438.0, + "loss": 0.0018, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.69921875, + "rewards/margins": 15.25, + "rewards/rejected": -15.9375, + "step": 2032 + }, + { + "epoch": 2.4944785276073618, + "grad_norm": 2.809903242187741, + "learning_rate": 2.553719008264463e-07, + "logits/chosen": -0.1337890625, + "logits/rejected": -0.322265625, + "logps/chosen": -410.0, + "logps/rejected": -414.0, + "loss": 0.0055, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.431640625, + "rewards/margins": 14.875, + "rewards/rejected": -15.3125, + "step": 2033 + }, + { + "epoch": 2.4957055214723924, + "grad_norm": 0.4762785473594232, + "learning_rate": 2.547520661157025e-07, + "logits/chosen": -0.296875, + "logits/rejected": -0.478515625, + "logps/chosen": -442.0, + "logps/rejected": -488.0, + "loss": 0.0011, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.65234375, + "rewards/margins": 16.125, + "rewards/rejected": -16.875, + "step": 2034 + }, + { + "epoch": 2.4969325153374236, + "grad_norm": 3.033694169625623, + "learning_rate": 2.541322314049587e-07, + "logits/chosen": -0.248046875, + "logits/rejected": -0.400390625, + "logps/chosen": -398.0, + "logps/rejected": -430.0, + "loss": 0.0038, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.65625, + "rewards/margins": 14.625, + "rewards/rejected": -15.25, + "step": 2035 + }, + { + "epoch": 2.4981595092024538, + "grad_norm": 0.6189834188868035, + "learning_rate": 2.5351239669421487e-07, + "logits/chosen": -0.05615234375, + "logits/rejected": -0.30078125, + "logps/chosen": -386.0, + "logps/rejected": -420.0, + "loss": 0.0011, + "rewards/accuracies": 1.0, + "rewards/chosen": -1.1796875, + "rewards/margins": 15.0625, + "rewards/rejected": -16.25, + "step": 2036 + }, + { + "epoch": 2.499386503067485, + "grad_norm": 1.1268737442904941, + "learning_rate": 2.5289256198347106e-07, + "logits/chosen": -0.267578125, + "logits/rejected": -0.447265625, + "logps/chosen": -424.0, + "logps/rejected": -444.0, + "loss": 0.0079, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.6796875, + "rewards/margins": 15.1875, + "rewards/rejected": -15.875, + "step": 2037 + }, + { + "epoch": 2.500613496932515, + "grad_norm": 5.934947942890086, + "learning_rate": 2.5227272727272726e-07, + "logits/chosen": -0.259765625, + "logits/rejected": -0.41015625, + "logps/chosen": -390.0, + "logps/rejected": -440.0, + "loss": 0.0139, + "rewards/accuracies": 0.984375, + "rewards/chosen": -1.7734375, + "rewards/margins": 14.4375, + "rewards/rejected": -16.25, + "step": 2038 + }, + { + "epoch": 2.5018404907975462, + "grad_norm": 0.6678617209905979, + "learning_rate": 2.516528925619835e-07, + "logits/chosen": -0.189453125, + "logits/rejected": -0.298828125, + "logps/chosen": -412.0, + "logps/rejected": -446.0, + "loss": 0.0011, + "rewards/accuracies": 1.0, + "rewards/chosen": -1.1953125, + "rewards/margins": 15.8125, + "rewards/rejected": -17.0, + "step": 2039 + }, + { + "epoch": 2.5030674846625764, + "grad_norm": 3.6160395292343024, + "learning_rate": 2.5103305785123965e-07, + "logits/chosen": -0.1875, + "logits/rejected": -0.44140625, + "logps/chosen": -360.0, + "logps/rejected": -400.0, + "loss": 0.0032, + "rewards/accuracies": 1.0, + "rewards/chosen": -1.25, + "rewards/margins": 14.5625, + "rewards/rejected": -15.8125, + "step": 2040 + }, + { + "epoch": 2.5042944785276076, + "grad_norm": 0.6425409931405636, + "learning_rate": 2.504132231404959e-07, + "logits/chosen": -0.375, + "logits/rejected": -0.5234375, + "logps/chosen": -402.0, + "logps/rejected": -440.0, + "loss": 0.0015, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.33203125, + "rewards/margins": 15.3125, + "rewards/rejected": -15.6875, + "step": 2041 + }, + { + "epoch": 2.505521472392638, + "grad_norm": 1.0585478800882246, + "learning_rate": 2.497933884297521e-07, + "logits/chosen": -0.28125, + "logits/rejected": -0.4609375, + "logps/chosen": -402.0, + "logps/rejected": -420.0, + "loss": 0.0021, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.3046875, + "rewards/margins": 16.5, + "rewards/rejected": -16.25, + "step": 2042 + }, + { + "epoch": 2.506748466257669, + "grad_norm": 3.251487006476127, + "learning_rate": 2.491735537190083e-07, + "logits/chosen": -0.255859375, + "logits/rejected": -0.42578125, + "logps/chosen": -414.0, + "logps/rejected": -458.0, + "loss": 0.0054, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.63671875, + "rewards/margins": 15.3125, + "rewards/rejected": -15.9375, + "step": 2043 + }, + { + "epoch": 2.507975460122699, + "grad_norm": 1.806492957858242, + "learning_rate": 2.485537190082645e-07, + "logits/chosen": -0.28125, + "logits/rejected": -0.412109375, + "logps/chosen": -320.0, + "logps/rejected": -374.0, + "loss": 0.0081, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -1.15625, + "rewards/margins": 14.25, + "rewards/rejected": -15.375, + "step": 2044 + }, + { + "epoch": 2.5092024539877302, + "grad_norm": 5.499259304303003, + "learning_rate": 2.4793388429752067e-07, + "logits/chosen": -0.2109375, + "logits/rejected": -0.419921875, + "logps/chosen": -420.0, + "logps/rejected": -418.0, + "loss": 0.0052, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.6484375, + "rewards/margins": 14.625, + "rewards/rejected": -15.3125, + "step": 2045 + }, + { + "epoch": 2.510429447852761, + "grad_norm": 0.8216551437277856, + "learning_rate": 2.4731404958677687e-07, + "logits/chosen": -0.2353515625, + "logits/rejected": -0.3828125, + "logps/chosen": -380.0, + "logps/rejected": -408.0, + "loss": 0.0017, + "rewards/accuracies": 1.0, + "rewards/chosen": -1.1328125, + "rewards/margins": 14.375, + "rewards/rejected": -15.5625, + "step": 2046 + }, + { + "epoch": 2.5116564417177916, + "grad_norm": 6.239738305908644, + "learning_rate": 2.4669421487603306e-07, + "logits/chosen": -0.265625, + "logits/rejected": -0.400390625, + "logps/chosen": -402.0, + "logps/rejected": -410.0, + "loss": 0.0144, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.9765625, + "rewards/margins": 15.25, + "rewards/rejected": -16.25, + "step": 2047 + }, + { + "epoch": 2.5128834355828222, + "grad_norm": 1.9107906837373008, + "learning_rate": 2.4607438016528925e-07, + "logits/chosen": -0.1904296875, + "logits/rejected": -0.359375, + "logps/chosen": -416.0, + "logps/rejected": -444.0, + "loss": 0.0027, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.97265625, + "rewards/margins": 14.5, + "rewards/rejected": -15.4375, + "step": 2048 + }, + { + "epoch": 2.514110429447853, + "grad_norm": 3.527988183713008, + "learning_rate": 2.4545454545454545e-07, + "logits/chosen": -0.26953125, + "logits/rejected": -0.431640625, + "logps/chosen": -462.0, + "logps/rejected": -446.0, + "loss": 0.0063, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.0654296875, + "rewards/margins": 15.875, + "rewards/rejected": -15.9375, + "step": 2049 + }, + { + "epoch": 2.5153374233128836, + "grad_norm": 2.475934748711706, + "learning_rate": 2.448347107438017e-07, + "logits/chosen": -0.259765625, + "logits/rejected": -0.50390625, + "logps/chosen": -462.0, + "logps/rejected": -472.0, + "loss": 0.0047, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.00104522705078125, + "rewards/margins": 15.875, + "rewards/rejected": -15.8125, + "step": 2050 + }, + { + "epoch": 2.5165644171779142, + "grad_norm": 1.086471909264626, + "learning_rate": 2.4421487603305784e-07, + "logits/chosen": -0.265625, + "logits/rejected": -0.43359375, + "logps/chosen": -426.0, + "logps/rejected": -418.0, + "loss": 0.0025, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.328125, + "rewards/margins": 14.0, + "rewards/rejected": -14.3125, + "step": 2051 + }, + { + "epoch": 2.517791411042945, + "grad_norm": 2.3903554974007086, + "learning_rate": 2.435950413223141e-07, + "logits/chosen": -0.2060546875, + "logits/rejected": -0.41796875, + "logps/chosen": -378.0, + "logps/rejected": -400.0, + "loss": 0.0089, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.93359375, + "rewards/margins": 14.0, + "rewards/rejected": -14.9375, + "step": 2052 + }, + { + "epoch": 2.5190184049079756, + "grad_norm": 1.7310573167836287, + "learning_rate": 2.429752066115702e-07, + "logits/chosen": -0.3359375, + "logits/rejected": -0.49609375, + "logps/chosen": -406.0, + "logps/rejected": -418.0, + "loss": 0.0018, + "rewards/accuracies": 1.0, + "rewards/chosen": -1.0625, + "rewards/margins": 14.0, + "rewards/rejected": -15.0625, + "step": 2053 + }, + { + "epoch": 2.5202453987730062, + "grad_norm": 1.436828182180415, + "learning_rate": 2.4235537190082647e-07, + "logits/chosen": -0.2138671875, + "logits/rejected": -0.44140625, + "logps/chosen": -460.0, + "logps/rejected": -454.0, + "loss": 0.0025, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.52734375, + "rewards/margins": 15.625, + "rewards/rejected": -16.125, + "step": 2054 + }, + { + "epoch": 2.521472392638037, + "grad_norm": 1.227356158494044, + "learning_rate": 2.417355371900826e-07, + "logits/chosen": -0.02392578125, + "logits/rejected": -0.2001953125, + "logps/chosen": -364.0, + "logps/rejected": -402.0, + "loss": 0.0026, + "rewards/accuracies": 1.0, + "rewards/chosen": -1.0703125, + "rewards/margins": 13.25, + "rewards/rejected": -14.3125, + "step": 2055 + }, + { + "epoch": 2.5226993865030676, + "grad_norm": 4.939722222576737, + "learning_rate": 2.4111570247933886e-07, + "logits/chosen": -0.248046875, + "logits/rejected": -0.5234375, + "logps/chosen": -454.0, + "logps/rejected": -468.0, + "loss": 0.0064, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.765625, + "rewards/margins": 15.1875, + "rewards/rejected": -15.9375, + "step": 2056 + }, + { + "epoch": 2.5239263803680982, + "grad_norm": 1.0017398634656944, + "learning_rate": 2.4049586776859505e-07, + "logits/chosen": -0.2490234375, + "logits/rejected": -0.373046875, + "logps/chosen": -422.0, + "logps/rejected": -468.0, + "loss": 0.0017, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.158203125, + "rewards/margins": 15.3125, + "rewards/rejected": -15.4375, + "step": 2057 + }, + { + "epoch": 2.525153374233129, + "grad_norm": 1.8625127743472178, + "learning_rate": 2.3987603305785125e-07, + "logits/chosen": -0.2734375, + "logits/rejected": -0.41015625, + "logps/chosen": -442.0, + "logps/rejected": -476.0, + "loss": 0.0031, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.6171875, + "rewards/margins": 14.375, + "rewards/rejected": -15.0, + "step": 2058 + }, + { + "epoch": 2.5263803680981596, + "grad_norm": 0.9268798698200328, + "learning_rate": 2.3925619834710744e-07, + "logits/chosen": -0.263671875, + "logits/rejected": -0.470703125, + "logps/chosen": -400.0, + "logps/rejected": -452.0, + "loss": 0.0016, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.447265625, + "rewards/margins": 14.8125, + "rewards/rejected": -15.25, + "step": 2059 + }, + { + "epoch": 2.5276073619631902, + "grad_norm": 1.124116669821742, + "learning_rate": 2.3863636363636364e-07, + "logits/chosen": -0.21484375, + "logits/rejected": -0.47265625, + "logps/chosen": -408.0, + "logps/rejected": -412.0, + "loss": 0.0026, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.5, + "rewards/margins": 14.1875, + "rewards/rejected": -14.6875, + "step": 2060 + }, + { + "epoch": 2.528834355828221, + "grad_norm": 1.407904367387, + "learning_rate": 2.3801652892561986e-07, + "logits/chosen": -0.2236328125, + "logits/rejected": -0.33984375, + "logps/chosen": -382.0, + "logps/rejected": -462.0, + "loss": 0.0024, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.08935546875, + "rewards/margins": 15.5625, + "rewards/rejected": -15.625, + "step": 2061 + }, + { + "epoch": 2.5300613496932516, + "grad_norm": 1.4960572299080206, + "learning_rate": 2.3739669421487603e-07, + "logits/chosen": -0.2080078125, + "logits/rejected": -0.35546875, + "logps/chosen": -402.0, + "logps/rejected": -444.0, + "loss": 0.0023, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.0498046875, + "rewards/margins": 14.75, + "rewards/rejected": -14.8125, + "step": 2062 + }, + { + "epoch": 2.5312883435582823, + "grad_norm": 1.26807284891066, + "learning_rate": 2.3677685950413225e-07, + "logits/chosen": -0.39453125, + "logits/rejected": -0.5390625, + "logps/chosen": -400.0, + "logps/rejected": -434.0, + "loss": 0.002, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.82421875, + "rewards/margins": 14.5625, + "rewards/rejected": -15.375, + "step": 2063 + }, + { + "epoch": 2.532515337423313, + "grad_norm": 0.9591862800358381, + "learning_rate": 2.3615702479338841e-07, + "logits/chosen": -0.25390625, + "logits/rejected": -0.578125, + "logps/chosen": -440.0, + "logps/rejected": -452.0, + "loss": 0.002, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.03369140625, + "rewards/margins": 14.75, + "rewards/rejected": -14.75, + "step": 2064 + }, + { + "epoch": 2.5337423312883436, + "grad_norm": 0.9656243591079007, + "learning_rate": 2.3553719008264464e-07, + "logits/chosen": -0.10693359375, + "logits/rejected": -0.279296875, + "logps/chosen": -414.0, + "logps/rejected": -430.0, + "loss": 0.0017, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.474609375, + "rewards/margins": 13.875, + "rewards/rejected": -14.375, + "step": 2065 + }, + { + "epoch": 2.5349693251533743, + "grad_norm": 2.7232815381036652, + "learning_rate": 2.349173553719008e-07, + "logits/chosen": -0.2373046875, + "logits/rejected": -0.380859375, + "logps/chosen": -368.0, + "logps/rejected": -412.0, + "loss": 0.007, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.83203125, + "rewards/margins": 13.3125, + "rewards/rejected": -14.1875, + "step": 2066 + }, + { + "epoch": 2.536196319018405, + "grad_norm": 4.132028689699233, + "learning_rate": 2.3429752066115702e-07, + "logits/chosen": -0.296875, + "logits/rejected": -0.431640625, + "logps/chosen": -378.0, + "logps/rejected": -428.0, + "loss": 0.0062, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.8046875, + "rewards/margins": 14.1875, + "rewards/rejected": -14.9375, + "step": 2067 + }, + { + "epoch": 2.5374233128834356, + "grad_norm": 0.44715082457543454, + "learning_rate": 2.3367768595041322e-07, + "logits/chosen": -0.2236328125, + "logits/rejected": -0.51171875, + "logps/chosen": -458.0, + "logps/rejected": -448.0, + "loss": 0.0011, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.71484375, + "rewards/margins": 16.375, + "rewards/rejected": -15.75, + "step": 2068 + }, + { + "epoch": 2.5386503067484663, + "grad_norm": 1.7554282050111156, + "learning_rate": 2.3305785123966944e-07, + "logits/chosen": -0.138671875, + "logits/rejected": -0.421875, + "logps/chosen": -386.0, + "logps/rejected": -398.0, + "loss": 0.0036, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.09228515625, + "rewards/margins": 14.8125, + "rewards/rejected": -14.9375, + "step": 2069 + }, + { + "epoch": 2.539877300613497, + "grad_norm": 1.8567752726531759, + "learning_rate": 2.3243801652892563e-07, + "logits/chosen": -0.3125, + "logits/rejected": -0.41796875, + "logps/chosen": -350.0, + "logps/rejected": -390.0, + "loss": 0.0021, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.4921875, + "rewards/margins": 14.3125, + "rewards/rejected": -14.8125, + "step": 2070 + }, + { + "epoch": 2.5411042944785276, + "grad_norm": 1.2386537530675938, + "learning_rate": 2.3181818181818183e-07, + "logits/chosen": -0.30078125, + "logits/rejected": -0.5234375, + "logps/chosen": -434.0, + "logps/rejected": -430.0, + "loss": 0.0023, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.44140625, + "rewards/margins": 15.3125, + "rewards/rejected": -14.875, + "step": 2071 + }, + { + "epoch": 2.5423312883435583, + "grad_norm": 0.614490185071627, + "learning_rate": 2.3119834710743805e-07, + "logits/chosen": -0.287109375, + "logits/rejected": -0.52734375, + "logps/chosen": -418.0, + "logps/rejected": -426.0, + "loss": 0.0012, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.1240234375, + "rewards/margins": 14.3125, + "rewards/rejected": -14.4375, + "step": 2072 + }, + { + "epoch": 2.543558282208589, + "grad_norm": 2.6340408655030023, + "learning_rate": 2.3057851239669422e-07, + "logits/chosen": -0.357421875, + "logits/rejected": -0.55859375, + "logps/chosen": -414.0, + "logps/rejected": -446.0, + "loss": 0.0029, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.189453125, + "rewards/margins": 15.3125, + "rewards/rejected": -15.5, + "step": 2073 + }, + { + "epoch": 2.5447852760736196, + "grad_norm": 2.0139396141339363, + "learning_rate": 2.2995867768595044e-07, + "logits/chosen": -0.318359375, + "logits/rejected": -0.62890625, + "logps/chosen": -430.0, + "logps/rejected": -438.0, + "loss": 0.0072, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.09228515625, + "rewards/margins": 15.1875, + "rewards/rejected": -15.0625, + "step": 2074 + }, + { + "epoch": 2.5460122699386503, + "grad_norm": 2.586127761094063, + "learning_rate": 2.293388429752066e-07, + "logits/chosen": -0.1875, + "logits/rejected": -0.474609375, + "logps/chosen": -406.0, + "logps/rejected": -422.0, + "loss": 0.0115, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.205078125, + "rewards/margins": 15.0, + "rewards/rejected": -14.75, + "step": 2075 + }, + { + "epoch": 2.547239263803681, + "grad_norm": 0.8496979919213687, + "learning_rate": 2.2871900826446282e-07, + "logits/chosen": -0.2333984375, + "logits/rejected": -0.435546875, + "logps/chosen": -360.0, + "logps/rejected": -418.0, + "loss": 0.0017, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.447265625, + "rewards/margins": 14.625, + "rewards/rejected": -15.0625, + "step": 2076 + }, + { + "epoch": 2.5484662576687116, + "grad_norm": 0.277401604687148, + "learning_rate": 2.28099173553719e-07, + "logits/chosen": -0.3671875, + "logits/rejected": -0.478515625, + "logps/chosen": -462.0, + "logps/rejected": -478.0, + "loss": 0.0006, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.35546875, + "rewards/margins": 16.125, + "rewards/rejected": -16.5, + "step": 2077 + }, + { + "epoch": 2.5496932515337423, + "grad_norm": 0.4298920917043149, + "learning_rate": 2.274793388429752e-07, + "logits/chosen": -0.248046875, + "logits/rejected": -0.4453125, + "logps/chosen": -364.0, + "logps/rejected": -418.0, + "loss": 0.0066, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -1.09375, + "rewards/margins": 13.5625, + "rewards/rejected": -14.6875, + "step": 2078 + }, + { + "epoch": 2.550920245398773, + "grad_norm": 3.397416803747359, + "learning_rate": 2.2685950413223143e-07, + "logits/chosen": -0.2265625, + "logits/rejected": -0.51171875, + "logps/chosen": -440.0, + "logps/rejected": -442.0, + "loss": 0.0035, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.006744384765625, + "rewards/margins": 14.875, + "rewards/rejected": -14.875, + "step": 2079 + }, + { + "epoch": 2.5521472392638036, + "grad_norm": 1.2690312466199192, + "learning_rate": 2.262396694214876e-07, + "logits/chosen": -0.19921875, + "logits/rejected": -0.3828125, + "logps/chosen": -428.0, + "logps/rejected": -432.0, + "loss": 0.0029, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.5, + "rewards/margins": 15.0625, + "rewards/rejected": -15.5625, + "step": 2080 + }, + { + "epoch": 2.5533742331288343, + "grad_norm": 1.1584978556419345, + "learning_rate": 2.2561983471074382e-07, + "logits/chosen": -0.263671875, + "logits/rejected": -0.388671875, + "logps/chosen": -374.0, + "logps/rejected": -432.0, + "loss": 0.0017, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.7578125, + "rewards/margins": 15.3125, + "rewards/rejected": -16.125, + "step": 2081 + }, + { + "epoch": 2.554601226993865, + "grad_norm": 3.403067528925662, + "learning_rate": 2.25e-07, + "logits/chosen": -0.21484375, + "logits/rejected": -0.41015625, + "logps/chosen": -398.0, + "logps/rejected": -410.0, + "loss": 0.0058, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.0230712890625, + "rewards/margins": 15.0, + "rewards/rejected": -15.0, + "step": 2082 + }, + { + "epoch": 2.5558282208588956, + "grad_norm": 2.6414390492451774, + "learning_rate": 2.243801652892562e-07, + "logits/chosen": -0.11962890625, + "logits/rejected": -0.30078125, + "logps/chosen": -384.0, + "logps/rejected": -410.0, + "loss": 0.0101, + "rewards/accuracies": 1.0, + "rewards/chosen": -1.3046875, + "rewards/margins": 13.1875, + "rewards/rejected": -14.5, + "step": 2083 + }, + { + "epoch": 2.5570552147239263, + "grad_norm": 0.7317248386639725, + "learning_rate": 2.2376033057851238e-07, + "logits/chosen": -0.2734375, + "logits/rejected": -0.421875, + "logps/chosen": -420.0, + "logps/rejected": -458.0, + "loss": 0.0071, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.365234375, + "rewards/margins": 15.0, + "rewards/rejected": -15.375, + "step": 2084 + }, + { + "epoch": 2.558282208588957, + "grad_norm": 6.825675950042904, + "learning_rate": 2.231404958677686e-07, + "logits/chosen": -0.11279296875, + "logits/rejected": -0.2890625, + "logps/chosen": -410.0, + "logps/rejected": -404.0, + "loss": 0.0054, + "rewards/accuracies": 1.0, + "rewards/chosen": -1.046875, + "rewards/margins": 13.25, + "rewards/rejected": -14.3125, + "step": 2085 + }, + { + "epoch": 2.5595092024539876, + "grad_norm": 0.7326407550186762, + "learning_rate": 2.225206611570248e-07, + "logits/chosen": -0.28515625, + "logits/rejected": -0.451171875, + "logps/chosen": -386.0, + "logps/rejected": -434.0, + "loss": 0.0019, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.5859375, + "rewards/margins": 14.375, + "rewards/rejected": -15.0, + "step": 2086 + }, + { + "epoch": 2.5607361963190183, + "grad_norm": 0.4562009950826071, + "learning_rate": 2.21900826446281e-07, + "logits/chosen": -0.1884765625, + "logits/rejected": -0.431640625, + "logps/chosen": -406.0, + "logps/rejected": -420.0, + "loss": 0.0009, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.65625, + "rewards/margins": 15.1875, + "rewards/rejected": -15.8125, + "step": 2087 + }, + { + "epoch": 2.561963190184049, + "grad_norm": 0.9508869811422345, + "learning_rate": 2.212809917355372e-07, + "logits/chosen": -0.353515625, + "logits/rejected": -0.3828125, + "logps/chosen": -364.0, + "logps/rejected": -510.0, + "loss": 0.0012, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.6171875, + "rewards/margins": 21.125, + "rewards/rejected": -21.75, + "step": 2088 + }, + { + "epoch": 2.5631901840490796, + "grad_norm": 1.064724545143142, + "learning_rate": 2.206611570247934e-07, + "logits/chosen": -0.33984375, + "logits/rejected": -0.546875, + "logps/chosen": -372.0, + "logps/rejected": -440.0, + "loss": 0.0024, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.96484375, + "rewards/margins": 15.5, + "rewards/rejected": -16.5, + "step": 2089 + }, + { + "epoch": 2.5644171779141103, + "grad_norm": 2.519871948532995, + "learning_rate": 2.200413223140496e-07, + "logits/chosen": -0.39453125, + "logits/rejected": -0.46484375, + "logps/chosen": -386.0, + "logps/rejected": -442.0, + "loss": 0.0041, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.828125, + "rewards/margins": 15.5, + "rewards/rejected": -16.375, + "step": 2090 + }, + { + "epoch": 2.565644171779141, + "grad_norm": 1.8165543515839644, + "learning_rate": 2.194214876033058e-07, + "logits/chosen": -0.388671875, + "logits/rejected": -0.45703125, + "logps/chosen": -372.0, + "logps/rejected": -418.0, + "loss": 0.0022, + "rewards/accuracies": 1.0, + "rewards/chosen": -1.0078125, + "rewards/margins": 14.125, + "rewards/rejected": -15.125, + "step": 2091 + }, + { + "epoch": 2.5668711656441716, + "grad_norm": 1.9172916006096838, + "learning_rate": 2.18801652892562e-07, + "logits/chosen": -0.26953125, + "logits/rejected": -0.5390625, + "logps/chosen": -460.0, + "logps/rejected": -462.0, + "loss": 0.0029, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.0966796875, + "rewards/margins": 15.9375, + "rewards/rejected": -16.0, + "step": 2092 + }, + { + "epoch": 2.5680981595092023, + "grad_norm": 0.6258994711744279, + "learning_rate": 2.1818181818181818e-07, + "logits/chosen": -0.193359375, + "logits/rejected": -0.359375, + "logps/chosen": -402.0, + "logps/rejected": -434.0, + "loss": 0.001, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.09033203125, + "rewards/margins": 15.1875, + "rewards/rejected": -15.25, + "step": 2093 + }, + { + "epoch": 2.5693251533742334, + "grad_norm": 6.268907139243496, + "learning_rate": 2.175619834710744e-07, + "logits/chosen": -0.2578125, + "logits/rejected": -0.462890625, + "logps/chosen": -402.0, + "logps/rejected": -424.0, + "loss": 0.0062, + "rewards/accuracies": 1.0, + "rewards/chosen": -1.03125, + "rewards/margins": 14.0, + "rewards/rejected": -15.0625, + "step": 2094 + }, + { + "epoch": 2.5705521472392636, + "grad_norm": 0.8612300552981011, + "learning_rate": 2.1694214876033057e-07, + "logits/chosen": -0.318359375, + "logits/rejected": -0.419921875, + "logps/chosen": -430.0, + "logps/rejected": -444.0, + "loss": 0.0015, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.1181640625, + "rewards/margins": 14.8125, + "rewards/rejected": -14.9375, + "step": 2095 + }, + { + "epoch": 2.5717791411042947, + "grad_norm": 0.7455085415238701, + "learning_rate": 2.163223140495868e-07, + "logits/chosen": -0.2001953125, + "logits/rejected": -0.357421875, + "logps/chosen": -396.0, + "logps/rejected": -414.0, + "loss": 0.0017, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.462890625, + "rewards/margins": 15.375, + "rewards/rejected": -15.8125, + "step": 2096 + }, + { + "epoch": 2.573006134969325, + "grad_norm": 1.5242663256778535, + "learning_rate": 2.1570247933884296e-07, + "logits/chosen": -0.158203125, + "logits/rejected": -0.322265625, + "logps/chosen": -418.0, + "logps/rejected": -460.0, + "loss": 0.0037, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.58984375, + "rewards/margins": 14.625, + "rewards/rejected": -15.25, + "step": 2097 + }, + { + "epoch": 2.574233128834356, + "grad_norm": 1.48450822204834, + "learning_rate": 2.1508264462809918e-07, + "logits/chosen": -0.1962890625, + "logits/rejected": -0.380859375, + "logps/chosen": -394.0, + "logps/rejected": -410.0, + "loss": 0.0034, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.9453125, + "rewards/margins": 13.9375, + "rewards/rejected": -14.9375, + "step": 2098 + }, + { + "epoch": 2.5754601226993863, + "grad_norm": 1.5177042356856036, + "learning_rate": 2.144628099173554e-07, + "logits/chosen": -0.1767578125, + "logits/rejected": -0.359375, + "logps/chosen": -412.0, + "logps/rejected": -444.0, + "loss": 0.0026, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.5546875, + "rewards/margins": 14.6875, + "rewards/rejected": -15.25, + "step": 2099 + }, + { + "epoch": 2.5766871165644174, + "grad_norm": 1.0685901241967959, + "learning_rate": 2.1384297520661157e-07, + "logits/chosen": -0.2138671875, + "logits/rejected": -0.4140625, + "logps/chosen": -402.0, + "logps/rejected": -442.0, + "loss": 0.0016, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.050537109375, + "rewards/margins": 15.3125, + "rewards/rejected": -15.375, + "step": 2100 + }, + { + "epoch": 2.5779141104294476, + "grad_norm": 0.7301064691557566, + "learning_rate": 2.1322314049586779e-07, + "logits/chosen": -0.259765625, + "logits/rejected": -0.439453125, + "logps/chosen": -424.0, + "logps/rejected": -414.0, + "loss": 0.0012, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.06689453125, + "rewards/margins": 15.0625, + "rewards/rejected": -15.0, + "step": 2101 + }, + { + "epoch": 2.5791411042944787, + "grad_norm": 1.2004708840936749, + "learning_rate": 2.1260330578512395e-07, + "logits/chosen": -0.1474609375, + "logits/rejected": -0.32421875, + "logps/chosen": -362.0, + "logps/rejected": -412.0, + "loss": 0.002, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.34765625, + "rewards/margins": 14.9375, + "rewards/rejected": -15.3125, + "step": 2102 + }, + { + "epoch": 2.580368098159509, + "grad_norm": 0.8394165989809456, + "learning_rate": 2.1198347107438017e-07, + "logits/chosen": -0.1328125, + "logits/rejected": -0.275390625, + "logps/chosen": -386.0, + "logps/rejected": -440.0, + "loss": 0.0015, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.3125, + "rewards/margins": 14.625, + "rewards/rejected": -14.9375, + "step": 2103 + }, + { + "epoch": 2.58159509202454, + "grad_norm": 6.671908142693215, + "learning_rate": 2.1136363636363637e-07, + "logits/chosen": 0.015869140625, + "logits/rejected": -0.2421875, + "logps/chosen": -356.0, + "logps/rejected": -382.0, + "loss": 0.0426, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -1.2890625, + "rewards/margins": 12.4375, + "rewards/rejected": -13.6875, + "step": 2104 + }, + { + "epoch": 2.5828220858895703, + "grad_norm": 1.4777679274710984, + "learning_rate": 2.1074380165289256e-07, + "logits/chosen": -0.162109375, + "logits/rejected": -0.34765625, + "logps/chosen": -382.0, + "logps/rejected": -426.0, + "loss": 0.0024, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.0810546875, + "rewards/margins": 14.8125, + "rewards/rejected": -14.9375, + "step": 2105 + }, + { + "epoch": 2.5840490797546014, + "grad_norm": 1.7863137034643577, + "learning_rate": 2.1012396694214876e-07, + "logits/chosen": -0.3515625, + "logits/rejected": -0.5390625, + "logps/chosen": -408.0, + "logps/rejected": -426.0, + "loss": 0.0028, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.90234375, + "rewards/margins": 15.0, + "rewards/rejected": -15.875, + "step": 2106 + }, + { + "epoch": 2.5852760736196316, + "grad_norm": 0.40853730199141597, + "learning_rate": 2.0950413223140498e-07, + "logits/chosen": -0.361328125, + "logits/rejected": -0.51953125, + "logps/chosen": -392.0, + "logps/rejected": -442.0, + "loss": 0.0012, + "rewards/accuracies": 1.0, + "rewards/chosen": -1.4140625, + "rewards/margins": 14.5, + "rewards/rejected": -15.9375, + "step": 2107 + }, + { + "epoch": 2.5865030674846627, + "grad_norm": 1.7892639851164942, + "learning_rate": 2.0888429752066117e-07, + "logits/chosen": -0.3203125, + "logits/rejected": -0.5078125, + "logps/chosen": -432.0, + "logps/rejected": -434.0, + "loss": 0.004, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.671875, + "rewards/margins": 13.9375, + "rewards/rejected": -14.625, + "step": 2108 + }, + { + "epoch": 2.5877300613496934, + "grad_norm": 1.5545161716611164, + "learning_rate": 2.0826446280991737e-07, + "logits/chosen": -0.1513671875, + "logits/rejected": -0.365234375, + "logps/chosen": -404.0, + "logps/rejected": -430.0, + "loss": 0.0021, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.75390625, + "rewards/margins": 15.5, + "rewards/rejected": -14.75, + "step": 2109 + }, + { + "epoch": 2.588957055214724, + "grad_norm": 2.060535768784622, + "learning_rate": 2.0764462809917359e-07, + "logits/chosen": -0.1923828125, + "logits/rejected": -0.2490234375, + "logps/chosen": -414.0, + "logps/rejected": -436.0, + "loss": 0.0038, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.1328125, + "rewards/margins": 14.625, + "rewards/rejected": -14.75, + "step": 2110 + }, + { + "epoch": 2.5901840490797547, + "grad_norm": 1.5544798624919849, + "learning_rate": 2.0702479338842975e-07, + "logits/chosen": -0.11328125, + "logits/rejected": -0.220703125, + "logps/chosen": -326.0, + "logps/rejected": -368.0, + "loss": 0.0036, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.8828125, + "rewards/margins": 12.75, + "rewards/rejected": -13.6875, + "step": 2111 + }, + { + "epoch": 2.5914110429447854, + "grad_norm": 0.31463181844289, + "learning_rate": 2.0640495867768598e-07, + "logits/chosen": -0.037353515625, + "logits/rejected": -0.255859375, + "logps/chosen": -418.0, + "logps/rejected": -418.0, + "loss": 0.0008, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.45703125, + "rewards/margins": 15.0, + "rewards/rejected": -15.4375, + "step": 2112 + }, + { + "epoch": 2.592638036809816, + "grad_norm": 0.6868403707094698, + "learning_rate": 2.0578512396694214e-07, + "logits/chosen": -0.177734375, + "logits/rejected": -0.333984375, + "logps/chosen": -376.0, + "logps/rejected": -398.0, + "loss": 0.0026, + "rewards/accuracies": 1.0, + "rewards/chosen": -1.1484375, + "rewards/margins": 14.125, + "rewards/rejected": -15.25, + "step": 2113 + }, + { + "epoch": 2.5938650306748468, + "grad_norm": 1.6384125880745102, + "learning_rate": 2.0516528925619836e-07, + "logits/chosen": -0.208984375, + "logits/rejected": -0.42578125, + "logps/chosen": -386.0, + "logps/rejected": -400.0, + "loss": 0.0027, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.166015625, + "rewards/margins": 14.125, + "rewards/rejected": -14.25, + "step": 2114 + }, + { + "epoch": 2.5950920245398774, + "grad_norm": 1.6077690777046378, + "learning_rate": 2.0454545454545453e-07, + "logits/chosen": -0.1611328125, + "logits/rejected": -0.361328125, + "logps/chosen": -396.0, + "logps/rejected": -394.0, + "loss": 0.0094, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.02490234375, + "rewards/margins": 14.375, + "rewards/rejected": -14.375, + "step": 2115 + }, + { + "epoch": 2.596319018404908, + "grad_norm": 3.418277992559877, + "learning_rate": 2.0392561983471075e-07, + "logits/chosen": -0.05322265625, + "logits/rejected": -0.193359375, + "logps/chosen": -362.0, + "logps/rejected": -406.0, + "loss": 0.0055, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.0546875, + "rewards/margins": 13.5625, + "rewards/rejected": -13.5, + "step": 2116 + }, + { + "epoch": 2.5975460122699388, + "grad_norm": 0.6655651386550483, + "learning_rate": 2.0330578512396692e-07, + "logits/chosen": -0.287109375, + "logits/rejected": -0.51953125, + "logps/chosen": -378.0, + "logps/rejected": -410.0, + "loss": 0.0016, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.035400390625, + "rewards/margins": 14.5, + "rewards/rejected": -14.5, + "step": 2117 + }, + { + "epoch": 2.5987730061349694, + "grad_norm": 0.6647081658529351, + "learning_rate": 2.0268595041322314e-07, + "logits/chosen": -0.1669921875, + "logits/rejected": -0.28515625, + "logps/chosen": -396.0, + "logps/rejected": -406.0, + "loss": 0.0016, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.87890625, + "rewards/margins": 15.375, + "rewards/rejected": -16.25, + "step": 2118 + }, + { + "epoch": 2.6, + "grad_norm": 5.516095249308202, + "learning_rate": 2.0206611570247936e-07, + "logits/chosen": -0.2060546875, + "logits/rejected": -0.37109375, + "logps/chosen": -390.0, + "logps/rejected": -406.0, + "loss": 0.0106, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.1513671875, + "rewards/margins": 14.9375, + "rewards/rejected": -14.75, + "step": 2119 + }, + { + "epoch": 2.6012269938650308, + "grad_norm": 2.603928084691748, + "learning_rate": 2.0144628099173553e-07, + "logits/chosen": -0.2021484375, + "logits/rejected": -0.45703125, + "logps/chosen": -428.0, + "logps/rejected": -464.0, + "loss": 0.0037, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.2490234375, + "rewards/margins": 14.375, + "rewards/rejected": -14.625, + "step": 2120 + }, + { + "epoch": 2.6024539877300614, + "grad_norm": 1.3105264643702497, + "learning_rate": 2.0082644628099175e-07, + "logits/chosen": -0.2734375, + "logits/rejected": -0.455078125, + "logps/chosen": -390.0, + "logps/rejected": -444.0, + "loss": 0.0029, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.279296875, + "rewards/margins": 15.3125, + "rewards/rejected": -15.5625, + "step": 2121 + }, + { + "epoch": 2.603680981595092, + "grad_norm": 1.4848588559145735, + "learning_rate": 2.0020661157024792e-07, + "logits/chosen": -0.28515625, + "logits/rejected": -0.51171875, + "logps/chosen": -374.0, + "logps/rejected": -422.0, + "loss": 0.0079, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.5, + "rewards/margins": 14.0, + "rewards/rejected": -14.5, + "step": 2122 + }, + { + "epoch": 2.6049079754601228, + "grad_norm": 1.7887048442278464, + "learning_rate": 1.9958677685950414e-07, + "logits/chosen": -0.1220703125, + "logits/rejected": -0.333984375, + "logps/chosen": -388.0, + "logps/rejected": -400.0, + "loss": 0.0054, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.384765625, + "rewards/margins": 14.0625, + "rewards/rejected": -14.4375, + "step": 2123 + }, + { + "epoch": 2.6061349693251534, + "grad_norm": 1.5538314177030113, + "learning_rate": 1.9896694214876033e-07, + "logits/chosen": -0.251953125, + "logits/rejected": -0.5, + "logps/chosen": -422.0, + "logps/rejected": -404.0, + "loss": 0.002, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.58203125, + "rewards/margins": 15.5, + "rewards/rejected": -14.875, + "step": 2124 + }, + { + "epoch": 2.607361963190184, + "grad_norm": 0.822757502345369, + "learning_rate": 1.9834710743801653e-07, + "logits/chosen": -0.2255859375, + "logits/rejected": -0.46875, + "logps/chosen": -402.0, + "logps/rejected": -488.0, + "loss": 0.0067, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.220703125, + "rewards/margins": 15.8125, + "rewards/rejected": -15.5625, + "step": 2125 + }, + { + "epoch": 2.6085889570552148, + "grad_norm": 2.8687436013685574, + "learning_rate": 1.9772727272727272e-07, + "logits/chosen": -0.2119140625, + "logits/rejected": -0.4140625, + "logps/chosen": -408.0, + "logps/rejected": -428.0, + "loss": 0.0069, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.0291748046875, + "rewards/margins": 14.4375, + "rewards/rejected": -14.4375, + "step": 2126 + }, + { + "epoch": 2.6098159509202454, + "grad_norm": 0.4807583259304409, + "learning_rate": 1.9710743801652894e-07, + "logits/chosen": -0.359375, + "logits/rejected": -0.5390625, + "logps/chosen": -416.0, + "logps/rejected": -440.0, + "loss": 0.0011, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.7421875, + "rewards/margins": 15.8125, + "rewards/rejected": -15.0625, + "step": 2127 + }, + { + "epoch": 2.611042944785276, + "grad_norm": 1.3442518168408093, + "learning_rate": 1.9648760330578516e-07, + "logits/chosen": -0.279296875, + "logits/rejected": -0.48828125, + "logps/chosen": -468.0, + "logps/rejected": -450.0, + "loss": 0.002, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.50390625, + "rewards/margins": 15.8125, + "rewards/rejected": -16.375, + "step": 2128 + }, + { + "epoch": 2.6122699386503068, + "grad_norm": 1.2015798492883039, + "learning_rate": 1.9586776859504133e-07, + "logits/chosen": -0.275390625, + "logits/rejected": -0.515625, + "logps/chosen": -442.0, + "logps/rejected": -462.0, + "loss": 0.0028, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.328125, + "rewards/margins": 15.125, + "rewards/rejected": -15.4375, + "step": 2129 + }, + { + "epoch": 2.6134969325153374, + "grad_norm": 5.474069733874836, + "learning_rate": 1.9524793388429755e-07, + "logits/chosen": -0.349609375, + "logits/rejected": -0.58984375, + "logps/chosen": -394.0, + "logps/rejected": -432.0, + "loss": 0.0256, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.88671875, + "rewards/margins": 14.625, + "rewards/rejected": -15.5, + "step": 2130 + }, + { + "epoch": 2.614723926380368, + "grad_norm": 0.7211948566855909, + "learning_rate": 1.9462809917355372e-07, + "logits/chosen": -0.255859375, + "logits/rejected": -0.4140625, + "logps/chosen": -390.0, + "logps/rejected": -460.0, + "loss": 0.0016, + "rewards/accuracies": 1.0, + "rewards/chosen": -1.890625, + "rewards/margins": 14.5, + "rewards/rejected": -16.375, + "step": 2131 + }, + { + "epoch": 2.6159509202453988, + "grad_norm": 4.2114951551430595, + "learning_rate": 1.9400826446280994e-07, + "logits/chosen": -0.166015625, + "logits/rejected": -0.255859375, + "logps/chosen": -414.0, + "logps/rejected": -472.0, + "loss": 0.0025, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.138671875, + "rewards/margins": 15.125, + "rewards/rejected": -15.0, + "step": 2132 + }, + { + "epoch": 2.6171779141104294, + "grad_norm": 1.619004995244786, + "learning_rate": 1.933884297520661e-07, + "logits/chosen": -0.2353515625, + "logits/rejected": -0.44921875, + "logps/chosen": -420.0, + "logps/rejected": -436.0, + "loss": 0.0028, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.279296875, + "rewards/margins": 15.75, + "rewards/rejected": -15.4375, + "step": 2133 + }, + { + "epoch": 2.61840490797546, + "grad_norm": 1.0509561548110093, + "learning_rate": 1.9276859504132233e-07, + "logits/chosen": -0.2353515625, + "logits/rejected": -0.51953125, + "logps/chosen": -442.0, + "logps/rejected": -470.0, + "loss": 0.0022, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.359375, + "rewards/margins": 14.8125, + "rewards/rejected": -15.125, + "step": 2134 + }, + { + "epoch": 2.6196319018404908, + "grad_norm": 1.4320999858513912, + "learning_rate": 1.921487603305785e-07, + "logits/chosen": -0.212890625, + "logits/rejected": -0.3671875, + "logps/chosen": -410.0, + "logps/rejected": -424.0, + "loss": 0.0031, + "rewards/accuracies": 1.0, + "rewards/chosen": -1.21875, + "rewards/margins": 13.75, + "rewards/rejected": -14.9375, + "step": 2135 + }, + { + "epoch": 2.6208588957055214, + "grad_norm": 0.6695158743568701, + "learning_rate": 1.9152892561983472e-07, + "logits/chosen": -0.275390625, + "logits/rejected": -0.61328125, + "logps/chosen": -478.0, + "logps/rejected": -458.0, + "loss": 0.0016, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.228515625, + "rewards/margins": 15.75, + "rewards/rejected": -15.5, + "step": 2136 + }, + { + "epoch": 2.622085889570552, + "grad_norm": 0.7587451147798011, + "learning_rate": 1.9090909090909088e-07, + "logits/chosen": -0.1064453125, + "logits/rejected": -0.34375, + "logps/chosen": -354.0, + "logps/rejected": -394.0, + "loss": 0.0015, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.498046875, + "rewards/margins": 14.75, + "rewards/rejected": -15.25, + "step": 2137 + }, + { + "epoch": 2.6233128834355828, + "grad_norm": 0.7018949671639728, + "learning_rate": 1.902892561983471e-07, + "logits/chosen": -0.384765625, + "logits/rejected": -0.53515625, + "logps/chosen": -382.0, + "logps/rejected": -464.0, + "loss": 0.0014, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.033935546875, + "rewards/margins": 15.875, + "rewards/rejected": -15.875, + "step": 2138 + }, + { + "epoch": 2.6245398773006134, + "grad_norm": 0.6094003259590463, + "learning_rate": 1.8966942148760332e-07, + "logits/chosen": -0.248046875, + "logits/rejected": -0.44140625, + "logps/chosen": -428.0, + "logps/rejected": -464.0, + "loss": 0.001, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.3203125, + "rewards/margins": 15.9375, + "rewards/rejected": -16.25, + "step": 2139 + }, + { + "epoch": 2.625766871165644, + "grad_norm": 1.6426952200387588, + "learning_rate": 1.890495867768595e-07, + "logits/chosen": -0.25390625, + "logits/rejected": -0.419921875, + "logps/chosen": -360.0, + "logps/rejected": -388.0, + "loss": 0.0033, + "rewards/accuracies": 1.0, + "rewards/chosen": -1.171875, + "rewards/margins": 14.4375, + "rewards/rejected": -15.5625, + "step": 2140 + }, + { + "epoch": 2.626993865030675, + "grad_norm": 1.3318242043075208, + "learning_rate": 1.8842975206611571e-07, + "logits/chosen": -0.173828125, + "logits/rejected": -0.271484375, + "logps/chosen": -408.0, + "logps/rejected": -414.0, + "loss": 0.0024, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.51953125, + "rewards/margins": 14.625, + "rewards/rejected": -15.125, + "step": 2141 + }, + { + "epoch": 2.6282208588957054, + "grad_norm": 9.19832380180926, + "learning_rate": 1.878099173553719e-07, + "logits/chosen": -0.27734375, + "logits/rejected": -0.490234375, + "logps/chosen": -396.0, + "logps/rejected": -436.0, + "loss": 0.0461, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.9140625, + "rewards/margins": 14.8125, + "rewards/rejected": -15.75, + "step": 2142 + }, + { + "epoch": 2.629447852760736, + "grad_norm": 2.3000501207289394, + "learning_rate": 1.871900826446281e-07, + "logits/chosen": -0.09716796875, + "logits/rejected": -0.26171875, + "logps/chosen": -424.0, + "logps/rejected": -432.0, + "loss": 0.0034, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.53125, + "rewards/margins": 14.1875, + "rewards/rejected": -14.75, + "step": 2143 + }, + { + "epoch": 2.630674846625767, + "grad_norm": 0.8976658361047792, + "learning_rate": 1.8657024793388432e-07, + "logits/chosen": -0.31640625, + "logits/rejected": -0.3984375, + "logps/chosen": -386.0, + "logps/rejected": -438.0, + "loss": 0.002, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.921875, + "rewards/margins": 14.3125, + "rewards/rejected": -15.25, + "step": 2144 + }, + { + "epoch": 2.6319018404907975, + "grad_norm": 5.590951921682611, + "learning_rate": 1.8595041322314052e-07, + "logits/chosen": -0.06005859375, + "logits/rejected": -0.19921875, + "logps/chosen": -382.0, + "logps/rejected": -406.0, + "loss": 0.0059, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.294921875, + "rewards/margins": 14.75, + "rewards/rejected": -15.0625, + "step": 2145 + }, + { + "epoch": 2.633128834355828, + "grad_norm": 1.2681221568897652, + "learning_rate": 1.853305785123967e-07, + "logits/chosen": -0.130859375, + "logits/rejected": -0.287109375, + "logps/chosen": -400.0, + "logps/rejected": -456.0, + "loss": 0.0031, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.578125, + "rewards/margins": 14.9375, + "rewards/rejected": -15.5, + "step": 2146 + }, + { + "epoch": 2.634355828220859, + "grad_norm": 3.0809105239571455, + "learning_rate": 1.847107438016529e-07, + "logits/chosen": -0.2001953125, + "logits/rejected": -0.369140625, + "logps/chosen": -380.0, + "logps/rejected": -402.0, + "loss": 0.0025, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.80078125, + "rewards/margins": 14.6875, + "rewards/rejected": -15.5, + "step": 2147 + }, + { + "epoch": 2.6355828220858895, + "grad_norm": 1.0892232741563885, + "learning_rate": 1.840909090909091e-07, + "logits/chosen": -0.197265625, + "logits/rejected": -0.4140625, + "logps/chosen": -372.0, + "logps/rejected": -424.0, + "loss": 0.0027, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.91796875, + "rewards/margins": 13.875, + "rewards/rejected": -14.8125, + "step": 2148 + }, + { + "epoch": 2.63680981595092, + "grad_norm": 1.2533240530284258, + "learning_rate": 1.834710743801653e-07, + "logits/chosen": -0.251953125, + "logits/rejected": -0.451171875, + "logps/chosen": -394.0, + "logps/rejected": -436.0, + "loss": 0.0023, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.69140625, + "rewards/margins": 14.5, + "rewards/rejected": -15.1875, + "step": 2149 + }, + { + "epoch": 2.638036809815951, + "grad_norm": 1.0024871888641584, + "learning_rate": 1.828512396694215e-07, + "logits/chosen": -0.25, + "logits/rejected": -0.3984375, + "logps/chosen": -356.0, + "logps/rejected": -412.0, + "loss": 0.0021, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.439453125, + "rewards/margins": 13.75, + "rewards/rejected": -14.1875, + "step": 2150 + }, + { + "epoch": 2.6392638036809815, + "grad_norm": 1.0087281280983416, + "learning_rate": 1.8223140495867768e-07, + "logits/chosen": -0.3046875, + "logits/rejected": -0.55078125, + "logps/chosen": -442.0, + "logps/rejected": -442.0, + "loss": 0.0021, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.474609375, + "rewards/margins": 15.375, + "rewards/rejected": -15.8125, + "step": 2151 + }, + { + "epoch": 2.640490797546012, + "grad_norm": 0.8857858233400364, + "learning_rate": 1.8161157024793388e-07, + "logits/chosen": -0.26953125, + "logits/rejected": -0.546875, + "logps/chosen": -438.0, + "logps/rejected": -398.0, + "loss": 0.0015, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.302734375, + "rewards/margins": 14.375, + "rewards/rejected": -14.6875, + "step": 2152 + }, + { + "epoch": 2.641717791411043, + "grad_norm": 1.0258382714509484, + "learning_rate": 1.809917355371901e-07, + "logits/chosen": -0.14453125, + "logits/rejected": -0.32421875, + "logps/chosen": -366.0, + "logps/rejected": -406.0, + "loss": 0.002, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.734375, + "rewards/margins": 14.375, + "rewards/rejected": -15.125, + "step": 2153 + }, + { + "epoch": 2.6429447852760735, + "grad_norm": 1.2167258433154844, + "learning_rate": 1.803719008264463e-07, + "logits/chosen": -0.203125, + "logits/rejected": -0.435546875, + "logps/chosen": -436.0, + "logps/rejected": -422.0, + "loss": 0.0019, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.4453125, + "rewards/margins": 15.375, + "rewards/rejected": -15.8125, + "step": 2154 + }, + { + "epoch": 2.644171779141104, + "grad_norm": 1.7969457064112921, + "learning_rate": 1.7975206611570249e-07, + "logits/chosen": -0.1162109375, + "logits/rejected": -0.314453125, + "logps/chosen": -362.0, + "logps/rejected": -400.0, + "loss": 0.0037, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.296875, + "rewards/margins": 14.0625, + "rewards/rejected": -14.375, + "step": 2155 + }, + { + "epoch": 2.645398773006135, + "grad_norm": 6.46967954500202, + "learning_rate": 1.7913223140495868e-07, + "logits/chosen": -0.1845703125, + "logits/rejected": -0.3671875, + "logps/chosen": -410.0, + "logps/rejected": -446.0, + "loss": 0.0066, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.11181640625, + "rewards/margins": 15.0625, + "rewards/rejected": -14.9375, + "step": 2156 + }, + { + "epoch": 2.646625766871166, + "grad_norm": 0.44023882083468124, + "learning_rate": 1.7851239669421487e-07, + "logits/chosen": -0.224609375, + "logits/rejected": -0.35546875, + "logps/chosen": -418.0, + "logps/rejected": -462.0, + "loss": 0.001, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.6484375, + "rewards/margins": 16.375, + "rewards/rejected": -15.75, + "step": 2157 + }, + { + "epoch": 2.647852760736196, + "grad_norm": 1.023514481236037, + "learning_rate": 1.7789256198347107e-07, + "logits/chosen": -0.3046875, + "logits/rejected": -0.4921875, + "logps/chosen": -416.0, + "logps/rejected": -418.0, + "loss": 0.0025, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.546875, + "rewards/margins": 14.375, + "rewards/rejected": -14.9375, + "step": 2158 + }, + { + "epoch": 2.6490797546012272, + "grad_norm": 0.45089491181266805, + "learning_rate": 1.7727272727272726e-07, + "logits/chosen": -0.421875, + "logits/rejected": -0.578125, + "logps/chosen": -370.0, + "logps/rejected": -422.0, + "loss": 0.001, + "rewards/accuracies": 1.0, + "rewards/chosen": -1.484375, + "rewards/margins": 15.125, + "rewards/rejected": -16.625, + "step": 2159 + }, + { + "epoch": 2.6503067484662575, + "grad_norm": 0.7705737808323231, + "learning_rate": 1.7665289256198348e-07, + "logits/chosen": -0.173828125, + "logits/rejected": -0.388671875, + "logps/chosen": -406.0, + "logps/rejected": -402.0, + "loss": 0.0017, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.1494140625, + "rewards/margins": 14.375, + "rewards/rejected": -14.5625, + "step": 2160 + }, + { + "epoch": 2.6515337423312886, + "grad_norm": 2.498562119721359, + "learning_rate": 1.7603305785123968e-07, + "logits/chosen": -0.1904296875, + "logits/rejected": -0.306640625, + "logps/chosen": -368.0, + "logps/rejected": -404.0, + "loss": 0.0086, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.9765625, + "rewards/margins": 14.75, + "rewards/rejected": -15.75, + "step": 2161 + }, + { + "epoch": 2.652760736196319, + "grad_norm": 1.113008184363733, + "learning_rate": 1.7541322314049587e-07, + "logits/chosen": -0.33984375, + "logits/rejected": -0.5234375, + "logps/chosen": -404.0, + "logps/rejected": -444.0, + "loss": 0.0024, + "rewards/accuracies": 1.0, + "rewards/chosen": -1.5, + "rewards/margins": 14.625, + "rewards/rejected": -16.125, + "step": 2162 + }, + { + "epoch": 2.65398773006135, + "grad_norm": 0.6805998854848132, + "learning_rate": 1.747933884297521e-07, + "logits/chosen": -0.1259765625, + "logits/rejected": -0.318359375, + "logps/chosen": -418.0, + "logps/rejected": -440.0, + "loss": 0.0015, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.43359375, + "rewards/margins": 15.5, + "rewards/rejected": -15.0625, + "step": 2163 + }, + { + "epoch": 2.65521472392638, + "grad_norm": 1.466456637565294, + "learning_rate": 1.7417355371900829e-07, + "logits/chosen": -0.345703125, + "logits/rejected": -0.53125, + "logps/chosen": -412.0, + "logps/rejected": -426.0, + "loss": 0.0028, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.1201171875, + "rewards/margins": 15.5, + "rewards/rejected": -15.625, + "step": 2164 + }, + { + "epoch": 2.6564417177914113, + "grad_norm": 2.7305149106161943, + "learning_rate": 1.7355371900826448e-07, + "logits/chosen": -0.17578125, + "logits/rejected": -0.3203125, + "logps/chosen": -400.0, + "logps/rejected": -438.0, + "loss": 0.0031, + "rewards/accuracies": 1.0, + "rewards/chosen": -1.015625, + "rewards/margins": 14.8125, + "rewards/rejected": -15.875, + "step": 2165 + }, + { + "epoch": 2.6576687116564415, + "grad_norm": 1.6345222243099997, + "learning_rate": 1.7293388429752067e-07, + "logits/chosen": -0.3125, + "logits/rejected": -0.5234375, + "logps/chosen": -436.0, + "logps/rejected": -470.0, + "loss": 0.0085, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.8359375, + "rewards/margins": 15.1875, + "rewards/rejected": -16.0, + "step": 2166 + }, + { + "epoch": 2.6588957055214726, + "grad_norm": 2.42064907366646, + "learning_rate": 1.7231404958677687e-07, + "logits/chosen": -0.109375, + "logits/rejected": -0.244140625, + "logps/chosen": -408.0, + "logps/rejected": -436.0, + "loss": 0.0039, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.54296875, + "rewards/margins": 14.625, + "rewards/rejected": -15.125, + "step": 2167 + }, + { + "epoch": 2.660122699386503, + "grad_norm": 0.5456277154332506, + "learning_rate": 1.7169421487603306e-07, + "logits/chosen": -0.1806640625, + "logits/rejected": -0.353515625, + "logps/chosen": -372.0, + "logps/rejected": -396.0, + "loss": 0.0013, + "rewards/accuracies": 1.0, + "rewards/chosen": -1.0390625, + "rewards/margins": 14.5, + "rewards/rejected": -15.5, + "step": 2168 + }, + { + "epoch": 2.661349693251534, + "grad_norm": 1.4888987259042152, + "learning_rate": 1.7107438016528926e-07, + "logits/chosen": -0.328125, + "logits/rejected": -0.56640625, + "logps/chosen": -420.0, + "logps/rejected": -472.0, + "loss": 0.0072, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.63671875, + "rewards/margins": 15.75, + "rewards/rejected": -16.375, + "step": 2169 + }, + { + "epoch": 2.662576687116564, + "grad_norm": 1.7469943525704466, + "learning_rate": 1.7045454545454545e-07, + "logits/chosen": -0.25390625, + "logits/rejected": -0.3203125, + "logps/chosen": -376.0, + "logps/rejected": -440.0, + "loss": 0.0083, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.55859375, + "rewards/margins": 14.3125, + "rewards/rejected": -14.875, + "step": 2170 + }, + { + "epoch": 2.6638036809815953, + "grad_norm": 1.6593577347025876, + "learning_rate": 1.6983471074380165e-07, + "logits/chosen": -0.2265625, + "logits/rejected": -0.451171875, + "logps/chosen": -404.0, + "logps/rejected": -424.0, + "loss": 0.003, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.2080078125, + "rewards/margins": 15.3125, + "rewards/rejected": -15.125, + "step": 2171 + }, + { + "epoch": 2.665030674846626, + "grad_norm": 1.1143791799627931, + "learning_rate": 1.6921487603305784e-07, + "logits/chosen": -0.26171875, + "logits/rejected": -0.458984375, + "logps/chosen": -332.0, + "logps/rejected": -410.0, + "loss": 0.0081, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.6796875, + "rewards/margins": 14.8125, + "rewards/rejected": -15.5, + "step": 2172 + }, + { + "epoch": 2.6662576687116566, + "grad_norm": 1.2704208530512522, + "learning_rate": 1.6859504132231406e-07, + "logits/chosen": -0.283203125, + "logits/rejected": -0.380859375, + "logps/chosen": -398.0, + "logps/rejected": -456.0, + "loss": 0.0016, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.30859375, + "rewards/margins": 15.625, + "rewards/rejected": -15.9375, + "step": 2173 + }, + { + "epoch": 2.6674846625766873, + "grad_norm": 1.5957472578769754, + "learning_rate": 1.6797520661157025e-07, + "logits/chosen": -0.2431640625, + "logits/rejected": -0.40625, + "logps/chosen": -380.0, + "logps/rejected": -416.0, + "loss": 0.0037, + "rewards/accuracies": 1.0, + "rewards/chosen": -1.2578125, + "rewards/margins": 14.0625, + "rewards/rejected": -15.3125, + "step": 2174 + }, + { + "epoch": 2.668711656441718, + "grad_norm": 1.3800596380625676, + "learning_rate": 1.6735537190082645e-07, + "logits/chosen": -0.28515625, + "logits/rejected": -0.494140625, + "logps/chosen": -390.0, + "logps/rejected": -412.0, + "loss": 0.0025, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.78125, + "rewards/margins": 14.3125, + "rewards/rejected": -15.0625, + "step": 2175 + }, + { + "epoch": 2.6699386503067486, + "grad_norm": 2.378262752481987, + "learning_rate": 1.6673553719008264e-07, + "logits/chosen": -0.34375, + "logits/rejected": -0.51171875, + "logps/chosen": -416.0, + "logps/rejected": -442.0, + "loss": 0.0093, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -1.1875, + "rewards/margins": 14.25, + "rewards/rejected": -15.4375, + "step": 2176 + }, + { + "epoch": 2.6711656441717793, + "grad_norm": 1.0313745613615635, + "learning_rate": 1.6611570247933884e-07, + "logits/chosen": -0.345703125, + "logits/rejected": -0.458984375, + "logps/chosen": -386.0, + "logps/rejected": -436.0, + "loss": 0.0017, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.66015625, + "rewards/margins": 15.375, + "rewards/rejected": -16.0, + "step": 2177 + }, + { + "epoch": 2.67239263803681, + "grad_norm": 1.5064891234198339, + "learning_rate": 1.6549586776859503e-07, + "logits/chosen": -0.1708984375, + "logits/rejected": -0.298828125, + "logps/chosen": -368.0, + "logps/rejected": -386.0, + "loss": 0.0035, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.8046875, + "rewards/margins": 13.6875, + "rewards/rejected": -14.5, + "step": 2178 + }, + { + "epoch": 2.6736196319018406, + "grad_norm": 1.3298264521589045, + "learning_rate": 1.6487603305785125e-07, + "logits/chosen": -0.185546875, + "logits/rejected": -0.337890625, + "logps/chosen": -374.0, + "logps/rejected": -416.0, + "loss": 0.0029, + "rewards/accuracies": 1.0, + "rewards/chosen": -1.1484375, + "rewards/margins": 14.1875, + "rewards/rejected": -15.375, + "step": 2179 + }, + { + "epoch": 2.6748466257668713, + "grad_norm": 3.3100130185868517, + "learning_rate": 1.6425619834710745e-07, + "logits/chosen": -0.234375, + "logits/rejected": -0.31640625, + "logps/chosen": -360.0, + "logps/rejected": -440.0, + "loss": 0.0037, + "rewards/accuracies": 1.0, + "rewards/chosen": -1.1640625, + "rewards/margins": 14.1875, + "rewards/rejected": -15.375, + "step": 2180 + }, + { + "epoch": 2.676073619631902, + "grad_norm": 2.450599915459223, + "learning_rate": 1.6363636363636364e-07, + "logits/chosen": -0.1416015625, + "logits/rejected": -0.345703125, + "logps/chosen": -410.0, + "logps/rejected": -406.0, + "loss": 0.004, + "rewards/accuracies": 1.0, + "rewards/chosen": -1.7734375, + "rewards/margins": 12.875, + "rewards/rejected": -14.6875, + "step": 2181 + }, + { + "epoch": 2.6773006134969326, + "grad_norm": 0.8099925838463904, + "learning_rate": 1.6301652892561984e-07, + "logits/chosen": -0.287109375, + "logits/rejected": -0.4140625, + "logps/chosen": -360.0, + "logps/rejected": -408.0, + "loss": 0.0016, + "rewards/accuracies": 1.0, + "rewards/chosen": -1.109375, + "rewards/margins": 13.75, + "rewards/rejected": -14.875, + "step": 2182 + }, + { + "epoch": 2.6785276073619633, + "grad_norm": 0.7861241737405547, + "learning_rate": 1.6239669421487606e-07, + "logits/chosen": -0.361328125, + "logits/rejected": -0.5625, + "logps/chosen": -366.0, + "logps/rejected": -396.0, + "loss": 0.0014, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.69140625, + "rewards/margins": 15.125, + "rewards/rejected": -15.8125, + "step": 2183 + }, + { + "epoch": 2.679754601226994, + "grad_norm": 0.8167802533876892, + "learning_rate": 1.6177685950413225e-07, + "logits/chosen": -0.2734375, + "logits/rejected": -0.36328125, + "logps/chosen": -380.0, + "logps/rejected": -430.0, + "loss": 0.0019, + "rewards/accuracies": 1.0, + "rewards/chosen": -1.171875, + "rewards/margins": 14.5625, + "rewards/rejected": -15.6875, + "step": 2184 + }, + { + "epoch": 2.6809815950920246, + "grad_norm": 2.4270749640269407, + "learning_rate": 1.6115702479338844e-07, + "logits/chosen": -0.2490234375, + "logits/rejected": -0.49609375, + "logps/chosen": -398.0, + "logps/rejected": -428.0, + "loss": 0.0039, + "rewards/accuracies": 1.0, + "rewards/chosen": -1.625, + "rewards/margins": 14.25, + "rewards/rejected": -15.875, + "step": 2185 + }, + { + "epoch": 2.6822085889570553, + "grad_norm": 2.0587937044268902, + "learning_rate": 1.6053719008264464e-07, + "logits/chosen": -0.10986328125, + "logits/rejected": -0.1767578125, + "logps/chosen": -342.0, + "logps/rejected": -424.0, + "loss": 0.0031, + "rewards/accuracies": 1.0, + "rewards/chosen": -1.5, + "rewards/margins": 14.0, + "rewards/rejected": -15.5, + "step": 2186 + }, + { + "epoch": 2.683435582822086, + "grad_norm": 0.6156838480931863, + "learning_rate": 1.5991735537190083e-07, + "logits/chosen": -0.35546875, + "logits/rejected": -0.455078125, + "logps/chosen": -374.0, + "logps/rejected": -432.0, + "loss": 0.0014, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.96875, + "rewards/margins": 15.5, + "rewards/rejected": -16.375, + "step": 2187 + }, + { + "epoch": 2.6846625766871166, + "grad_norm": 0.7092275272478239, + "learning_rate": 1.5929752066115703e-07, + "logits/chosen": -0.1962890625, + "logits/rejected": -0.4453125, + "logps/chosen": -416.0, + "logps/rejected": -422.0, + "loss": 0.0015, + "rewards/accuracies": 1.0, + "rewards/chosen": -1.1953125, + "rewards/margins": 15.0, + "rewards/rejected": -16.25, + "step": 2188 + }, + { + "epoch": 2.6858895705521473, + "grad_norm": 0.7957329182536127, + "learning_rate": 1.5867768595041322e-07, + "logits/chosen": -0.3359375, + "logits/rejected": -0.53515625, + "logps/chosen": -408.0, + "logps/rejected": -418.0, + "loss": 0.0018, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.82421875, + "rewards/margins": 14.75, + "rewards/rejected": -15.5625, + "step": 2189 + }, + { + "epoch": 2.687116564417178, + "grad_norm": 1.2182524077732229, + "learning_rate": 1.5805785123966942e-07, + "logits/chosen": -0.34765625, + "logits/rejected": -0.54296875, + "logps/chosen": -438.0, + "logps/rejected": -446.0, + "loss": 0.0026, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.474609375, + "rewards/margins": 15.3125, + "rewards/rejected": -15.8125, + "step": 2190 + }, + { + "epoch": 2.6883435582822086, + "grad_norm": 0.7951004826395361, + "learning_rate": 1.574380165289256e-07, + "logits/chosen": -0.244140625, + "logits/rejected": -0.478515625, + "logps/chosen": -414.0, + "logps/rejected": -414.0, + "loss": 0.0069, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.0159912109375, + "rewards/margins": 15.5, + "rewards/rejected": -15.5, + "step": 2191 + }, + { + "epoch": 2.6895705521472393, + "grad_norm": 4.88704946265994, + "learning_rate": 1.568181818181818e-07, + "logits/chosen": -0.1884765625, + "logits/rejected": -0.32421875, + "logps/chosen": -384.0, + "logps/rejected": -454.0, + "loss": 0.0076, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.78515625, + "rewards/margins": 15.1875, + "rewards/rejected": -16.0, + "step": 2192 + }, + { + "epoch": 2.69079754601227, + "grad_norm": 1.1054065326840035, + "learning_rate": 1.5619834710743802e-07, + "logits/chosen": -0.1806640625, + "logits/rejected": -0.25, + "logps/chosen": -378.0, + "logps/rejected": -412.0, + "loss": 0.0013, + "rewards/accuracies": 1.0, + "rewards/chosen": -1.234375, + "rewards/margins": 14.4375, + "rewards/rejected": -15.6875, + "step": 2193 + }, + { + "epoch": 2.6920245398773006, + "grad_norm": 1.1211056763054352, + "learning_rate": 1.5557851239669422e-07, + "logits/chosen": -0.279296875, + "logits/rejected": -0.376953125, + "logps/chosen": -416.0, + "logps/rejected": -464.0, + "loss": 0.0018, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.75390625, + "rewards/margins": 14.9375, + "rewards/rejected": -15.6875, + "step": 2194 + }, + { + "epoch": 2.6932515337423313, + "grad_norm": 0.9483782178641457, + "learning_rate": 1.549586776859504e-07, + "logits/chosen": -0.1982421875, + "logits/rejected": -0.36328125, + "logps/chosen": -494.0, + "logps/rejected": -496.0, + "loss": 0.002, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.27734375, + "rewards/margins": 16.125, + "rewards/rejected": -16.375, + "step": 2195 + }, + { + "epoch": 2.694478527607362, + "grad_norm": 0.6385562606771361, + "learning_rate": 1.543388429752066e-07, + "logits/chosen": -0.380859375, + "logits/rejected": -0.6171875, + "logps/chosen": -404.0, + "logps/rejected": -430.0, + "loss": 0.0012, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.11669921875, + "rewards/margins": 15.3125, + "rewards/rejected": -15.4375, + "step": 2196 + }, + { + "epoch": 2.6957055214723926, + "grad_norm": 0.43978310625841105, + "learning_rate": 1.537190082644628e-07, + "logits/chosen": -0.2890625, + "logits/rejected": -0.578125, + "logps/chosen": -450.0, + "logps/rejected": -458.0, + "loss": 0.001, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.1533203125, + "rewards/margins": 16.75, + "rewards/rejected": -16.5, + "step": 2197 + }, + { + "epoch": 2.6969325153374233, + "grad_norm": 1.9062157270749442, + "learning_rate": 1.5309917355371902e-07, + "logits/chosen": -0.279296875, + "logits/rejected": -0.35546875, + "logps/chosen": -414.0, + "logps/rejected": -432.0, + "loss": 0.0037, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.74609375, + "rewards/margins": 14.75, + "rewards/rejected": -15.5, + "step": 2198 + }, + { + "epoch": 2.698159509202454, + "grad_norm": 1.6948805129620825, + "learning_rate": 1.5247933884297522e-07, + "logits/chosen": -0.2001953125, + "logits/rejected": -0.458984375, + "logps/chosen": -456.0, + "logps/rejected": -448.0, + "loss": 0.0036, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.83203125, + "rewards/margins": 14.875, + "rewards/rejected": -15.6875, + "step": 2199 + }, + { + "epoch": 2.6993865030674846, + "grad_norm": 0.7245042260216192, + "learning_rate": 1.518595041322314e-07, + "logits/chosen": -0.1015625, + "logits/rejected": -0.275390625, + "logps/chosen": -398.0, + "logps/rejected": -434.0, + "loss": 0.0016, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.609375, + "rewards/margins": 14.625, + "rewards/rejected": -15.25, + "step": 2200 + }, + { + "epoch": 2.7006134969325153, + "grad_norm": 0.3692209282967543, + "learning_rate": 1.512396694214876e-07, + "logits/chosen": -0.291015625, + "logits/rejected": -0.54296875, + "logps/chosen": -464.0, + "logps/rejected": -482.0, + "loss": 0.0009, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.7265625, + "rewards/margins": 15.9375, + "rewards/rejected": -15.1875, + "step": 2201 + }, + { + "epoch": 2.701840490797546, + "grad_norm": 0.9234659035205631, + "learning_rate": 1.506198347107438e-07, + "logits/chosen": -0.21875, + "logits/rejected": -0.30859375, + "logps/chosen": -398.0, + "logps/rejected": -434.0, + "loss": 0.0019, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.84375, + "rewards/margins": 14.5625, + "rewards/rejected": -15.375, + "step": 2202 + }, + { + "epoch": 2.7030674846625766, + "grad_norm": 2.775318951926902, + "learning_rate": 1.5000000000000002e-07, + "logits/chosen": -0.14453125, + "logits/rejected": -0.423828125, + "logps/chosen": -402.0, + "logps/rejected": -414.0, + "loss": 0.0027, + "rewards/accuracies": 1.0, + "rewards/chosen": -1.1015625, + "rewards/margins": 15.125, + "rewards/rejected": -16.25, + "step": 2203 + }, + { + "epoch": 2.7042944785276073, + "grad_norm": 0.6824885029219661, + "learning_rate": 1.4938016528925621e-07, + "logits/chosen": -0.26171875, + "logits/rejected": -0.44921875, + "logps/chosen": -414.0, + "logps/rejected": -436.0, + "loss": 0.0011, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.5625, + "rewards/margins": 15.9375, + "rewards/rejected": -16.5, + "step": 2204 + }, + { + "epoch": 2.705521472392638, + "grad_norm": 2.1037046875425247, + "learning_rate": 1.487603305785124e-07, + "logits/chosen": -0.228515625, + "logits/rejected": -0.39453125, + "logps/chosen": -388.0, + "logps/rejected": -408.0, + "loss": 0.0031, + "rewards/accuracies": 1.0, + "rewards/chosen": -1.0859375, + "rewards/margins": 14.0625, + "rewards/rejected": -15.1875, + "step": 2205 + }, + { + "epoch": 2.7067484662576686, + "grad_norm": 3.5218765738077478, + "learning_rate": 1.481404958677686e-07, + "logits/chosen": -0.28125, + "logits/rejected": -0.328125, + "logps/chosen": -354.0, + "logps/rejected": -426.0, + "loss": 0.0097, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.7578125, + "rewards/margins": 15.0, + "rewards/rejected": -15.8125, + "step": 2206 + }, + { + "epoch": 2.7079754601226993, + "grad_norm": 1.3710371628237512, + "learning_rate": 1.475206611570248e-07, + "logits/chosen": -0.26953125, + "logits/rejected": -0.435546875, + "logps/chosen": -396.0, + "logps/rejected": -452.0, + "loss": 0.0022, + "rewards/accuracies": 1.0, + "rewards/chosen": -1.0, + "rewards/margins": 14.4375, + "rewards/rejected": -15.4375, + "step": 2207 + }, + { + "epoch": 2.70920245398773, + "grad_norm": 0.38063457248893257, + "learning_rate": 1.46900826446281e-07, + "logits/chosen": -0.328125, + "logits/rejected": -0.578125, + "logps/chosen": -424.0, + "logps/rejected": -462.0, + "loss": 0.0005, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.1865234375, + "rewards/margins": 16.625, + "rewards/rejected": -16.75, + "step": 2208 + }, + { + "epoch": 2.7104294478527606, + "grad_norm": 0.8534657125579109, + "learning_rate": 1.4628099173553718e-07, + "logits/chosen": -0.3359375, + "logits/rejected": -0.52734375, + "logps/chosen": -420.0, + "logps/rejected": -440.0, + "loss": 0.0075, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.54296875, + "rewards/margins": 15.25, + "rewards/rejected": -15.8125, + "step": 2209 + }, + { + "epoch": 2.7116564417177913, + "grad_norm": 3.9770202245718482, + "learning_rate": 1.4566115702479338e-07, + "logits/chosen": -0.2138671875, + "logits/rejected": -0.431640625, + "logps/chosen": -398.0, + "logps/rejected": -404.0, + "loss": 0.005, + "rewards/accuracies": 1.0, + "rewards/chosen": -1.0078125, + "rewards/margins": 14.4375, + "rewards/rejected": -15.4375, + "step": 2210 + }, + { + "epoch": 2.712883435582822, + "grad_norm": 0.532344800365596, + "learning_rate": 1.4504132231404957e-07, + "logits/chosen": -0.1962890625, + "logits/rejected": -0.384765625, + "logps/chosen": -376.0, + "logps/rejected": -434.0, + "loss": 0.0012, + "rewards/accuracies": 1.0, + "rewards/chosen": -1.625, + "rewards/margins": 14.375, + "rewards/rejected": -16.0, + "step": 2211 + }, + { + "epoch": 2.7141104294478526, + "grad_norm": 1.0208607803311958, + "learning_rate": 1.444214876033058e-07, + "logits/chosen": -0.287109375, + "logits/rejected": -0.46484375, + "logps/chosen": -434.0, + "logps/rejected": -438.0, + "loss": 0.0072, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.9609375, + "rewards/margins": 15.3125, + "rewards/rejected": -16.25, + "step": 2212 + }, + { + "epoch": 2.7153374233128833, + "grad_norm": 1.2327568417038834, + "learning_rate": 1.43801652892562e-07, + "logits/chosen": -0.2177734375, + "logits/rejected": -0.455078125, + "logps/chosen": -424.0, + "logps/rejected": -460.0, + "loss": 0.0021, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.228515625, + "rewards/margins": 15.875, + "rewards/rejected": -15.6875, + "step": 2213 + }, + { + "epoch": 2.716564417177914, + "grad_norm": 0.9780631553268927, + "learning_rate": 1.4318181818181818e-07, + "logits/chosen": -0.130859375, + "logits/rejected": -0.37109375, + "logps/chosen": -378.0, + "logps/rejected": -410.0, + "loss": 0.0022, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.875, + "rewards/margins": 14.4375, + "rewards/rejected": -15.3125, + "step": 2214 + }, + { + "epoch": 2.7177914110429446, + "grad_norm": 2.9164658608180196, + "learning_rate": 1.4256198347107438e-07, + "logits/chosen": -0.2060546875, + "logits/rejected": -0.357421875, + "logps/chosen": -394.0, + "logps/rejected": -458.0, + "loss": 0.0019, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.859375, + "rewards/margins": 16.0, + "rewards/rejected": -16.875, + "step": 2215 + }, + { + "epoch": 2.7190184049079753, + "grad_norm": 1.7460599389400162, + "learning_rate": 1.4194214876033057e-07, + "logits/chosen": -0.08740234375, + "logits/rejected": -0.318359375, + "logps/chosen": -376.0, + "logps/rejected": -424.0, + "loss": 0.0092, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.83203125, + "rewards/margins": 14.6875, + "rewards/rejected": -15.5625, + "step": 2216 + }, + { + "epoch": 2.720245398773006, + "grad_norm": 1.1318085685165589, + "learning_rate": 1.413223140495868e-07, + "logits/chosen": -0.193359375, + "logits/rejected": -0.404296875, + "logps/chosen": -428.0, + "logps/rejected": -428.0, + "loss": 0.0076, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.0869140625, + "rewards/margins": 14.6875, + "rewards/rejected": -14.75, + "step": 2217 + }, + { + "epoch": 2.721472392638037, + "grad_norm": 0.6058311831684835, + "learning_rate": 1.4070247933884299e-07, + "logits/chosen": -0.2353515625, + "logits/rejected": -0.41796875, + "logps/chosen": -392.0, + "logps/rejected": -426.0, + "loss": 0.0069, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.59375, + "rewards/margins": 15.5625, + "rewards/rejected": -16.125, + "step": 2218 + }, + { + "epoch": 2.7226993865030673, + "grad_norm": 1.5054108552400738, + "learning_rate": 1.4008264462809918e-07, + "logits/chosen": -0.1875, + "logits/rejected": -0.390625, + "logps/chosen": -408.0, + "logps/rejected": -460.0, + "loss": 0.0019, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.953125, + "rewards/margins": 14.9375, + "rewards/rejected": -15.875, + "step": 2219 + }, + { + "epoch": 2.7239263803680984, + "grad_norm": 1.6211663118248798, + "learning_rate": 1.3946280991735537e-07, + "logits/chosen": -0.2490234375, + "logits/rejected": -0.482421875, + "logps/chosen": -444.0, + "logps/rejected": -426.0, + "loss": 0.0039, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.2490234375, + "rewards/margins": 15.375, + "rewards/rejected": -15.125, + "step": 2220 + }, + { + "epoch": 2.7251533742331286, + "grad_norm": 1.264677290107995, + "learning_rate": 1.3884297520661157e-07, + "logits/chosen": -0.1435546875, + "logits/rejected": -0.26171875, + "logps/chosen": -388.0, + "logps/rejected": -452.0, + "loss": 0.0098, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.80078125, + "rewards/margins": 15.25, + "rewards/rejected": -16.125, + "step": 2221 + }, + { + "epoch": 2.7263803680981598, + "grad_norm": 1.0517464829055336, + "learning_rate": 1.382231404958678e-07, + "logits/chosen": -0.19921875, + "logits/rejected": -0.42578125, + "logps/chosen": -420.0, + "logps/rejected": -456.0, + "loss": 0.0022, + "rewards/accuracies": 1.0, + "rewards/chosen": -1.3515625, + "rewards/margins": 14.5625, + "rewards/rejected": -15.9375, + "step": 2222 + }, + { + "epoch": 2.72760736196319, + "grad_norm": 0.5811109452820935, + "learning_rate": 1.3760330578512398e-07, + "logits/chosen": -0.142578125, + "logits/rejected": -0.37109375, + "logps/chosen": -426.0, + "logps/rejected": -430.0, + "loss": 0.0014, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.609375, + "rewards/margins": 15.4375, + "rewards/rejected": -16.0, + "step": 2223 + }, + { + "epoch": 2.728834355828221, + "grad_norm": 3.5703426018515954, + "learning_rate": 1.3698347107438018e-07, + "logits/chosen": -0.1865234375, + "logits/rejected": -0.357421875, + "logps/chosen": -416.0, + "logps/rejected": -420.0, + "loss": 0.0098, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -1.359375, + "rewards/margins": 14.375, + "rewards/rejected": -15.75, + "step": 2224 + }, + { + "epoch": 2.7300613496932513, + "grad_norm": 2.5490546292661733, + "learning_rate": 1.3636363636363637e-07, + "logits/chosen": -0.275390625, + "logits/rejected": -0.375, + "logps/chosen": -386.0, + "logps/rejected": -442.0, + "loss": 0.0091, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -1.078125, + "rewards/margins": 14.4375, + "rewards/rejected": -15.5625, + "step": 2225 + }, + { + "epoch": 2.7312883435582824, + "grad_norm": 0.4039530508685643, + "learning_rate": 1.3574380165289257e-07, + "logits/chosen": -0.1787109375, + "logits/rejected": -0.453125, + "logps/chosen": -450.0, + "logps/rejected": -454.0, + "loss": 0.001, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.8671875, + "rewards/margins": 15.5, + "rewards/rejected": -16.375, + "step": 2226 + }, + { + "epoch": 2.7325153374233127, + "grad_norm": 1.328727910981409, + "learning_rate": 1.3512396694214876e-07, + "logits/chosen": -0.2021484375, + "logits/rejected": -0.283203125, + "logps/chosen": -356.0, + "logps/rejected": -454.0, + "loss": 0.0029, + "rewards/accuracies": 1.0, + "rewards/chosen": -1.6484375, + "rewards/margins": 13.8125, + "rewards/rejected": -15.4375, + "step": 2227 + }, + { + "epoch": 2.7337423312883438, + "grad_norm": 11.008845822379248, + "learning_rate": 1.3450413223140495e-07, + "logits/chosen": -0.21484375, + "logits/rejected": -0.484375, + "logps/chosen": -378.0, + "logps/rejected": -400.0, + "loss": 0.0093, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.87109375, + "rewards/margins": 15.4375, + "rewards/rejected": -16.25, + "step": 2228 + }, + { + "epoch": 2.734969325153374, + "grad_norm": 1.0498886069892437, + "learning_rate": 1.3388429752066115e-07, + "logits/chosen": -0.07763671875, + "logits/rejected": -0.353515625, + "logps/chosen": -394.0, + "logps/rejected": -410.0, + "loss": 0.0028, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.193359375, + "rewards/margins": 15.0, + "rewards/rejected": -15.1875, + "step": 2229 + }, + { + "epoch": 2.736196319018405, + "grad_norm": 11.032197627658368, + "learning_rate": 1.3326446280991734e-07, + "logits/chosen": -0.37109375, + "logits/rejected": -0.58984375, + "logps/chosen": -424.0, + "logps/rejected": -442.0, + "loss": 0.0089, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.349609375, + "rewards/margins": 16.375, + "rewards/rejected": -16.75, + "step": 2230 + }, + { + "epoch": 2.7374233128834353, + "grad_norm": 2.1137206797437713, + "learning_rate": 1.3264462809917354e-07, + "logits/chosen": -0.2890625, + "logits/rejected": -0.55859375, + "logps/chosen": -392.0, + "logps/rejected": -412.0, + "loss": 0.0034, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.296875, + "rewards/margins": 15.875, + "rewards/rejected": -16.25, + "step": 2231 + }, + { + "epoch": 2.7386503067484664, + "grad_norm": 2.11943927996083, + "learning_rate": 1.3202479338842976e-07, + "logits/chosen": -0.263671875, + "logits/rejected": -0.384765625, + "logps/chosen": -354.0, + "logps/rejected": -430.0, + "loss": 0.0031, + "rewards/accuracies": 1.0, + "rewards/chosen": -1.71875, + "rewards/margins": 14.3125, + "rewards/rejected": -16.0, + "step": 2232 + }, + { + "epoch": 2.7398773006134967, + "grad_norm": 1.4623466429202086, + "learning_rate": 1.3140495867768595e-07, + "logits/chosen": -0.326171875, + "logits/rejected": -0.5625, + "logps/chosen": -440.0, + "logps/rejected": -466.0, + "loss": 0.0042, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.796875, + "rewards/margins": 14.9375, + "rewards/rejected": -15.75, + "step": 2233 + }, + { + "epoch": 2.7411042944785278, + "grad_norm": 0.4057453588172053, + "learning_rate": 1.3078512396694215e-07, + "logits/chosen": -0.330078125, + "logits/rejected": -0.5546875, + "logps/chosen": -436.0, + "logps/rejected": -472.0, + "loss": 0.0064, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.62890625, + "rewards/margins": 16.0, + "rewards/rejected": -16.625, + "step": 2234 + }, + { + "epoch": 2.7423312883435584, + "grad_norm": 1.7078411465513808, + "learning_rate": 1.3016528925619837e-07, + "logits/chosen": -0.142578125, + "logits/rejected": -0.2392578125, + "logps/chosen": -336.0, + "logps/rejected": -408.0, + "loss": 0.0087, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -1.0546875, + "rewards/margins": 13.6875, + "rewards/rejected": -14.75, + "step": 2235 + }, + { + "epoch": 2.743558282208589, + "grad_norm": 1.6075097639435214, + "learning_rate": 1.2954545454545456e-07, + "logits/chosen": -0.1748046875, + "logits/rejected": -0.322265625, + "logps/chosen": -448.0, + "logps/rejected": -504.0, + "loss": 0.0022, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.2060546875, + "rewards/margins": 16.25, + "rewards/rejected": -16.375, + "step": 2236 + }, + { + "epoch": 2.7447852760736198, + "grad_norm": 4.859470374706864, + "learning_rate": 1.2892561983471076e-07, + "logits/chosen": -0.283203125, + "logits/rejected": -0.515625, + "logps/chosen": -418.0, + "logps/rejected": -418.0, + "loss": 0.0104, + "rewards/accuracies": 1.0, + "rewards/chosen": -1.015625, + "rewards/margins": 14.375, + "rewards/rejected": -15.375, + "step": 2237 + }, + { + "epoch": 2.7460122699386504, + "grad_norm": 0.8132857713828272, + "learning_rate": 1.2830578512396695e-07, + "logits/chosen": -0.2314453125, + "logits/rejected": -0.365234375, + "logps/chosen": -372.0, + "logps/rejected": -460.0, + "loss": 0.0016, + "rewards/accuracies": 1.0, + "rewards/chosen": -1.0078125, + "rewards/margins": 15.6875, + "rewards/rejected": -16.75, + "step": 2238 + }, + { + "epoch": 2.747239263803681, + "grad_norm": 1.5644656509825712, + "learning_rate": 1.2768595041322314e-07, + "logits/chosen": -0.29296875, + "logits/rejected": -0.451171875, + "logps/chosen": -368.0, + "logps/rejected": -408.0, + "loss": 0.0078, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.80859375, + "rewards/margins": 14.5625, + "rewards/rejected": -15.375, + "step": 2239 + }, + { + "epoch": 2.7484662576687118, + "grad_norm": 0.5243725906218296, + "learning_rate": 1.2706611570247934e-07, + "logits/chosen": -0.248046875, + "logits/rejected": -0.34765625, + "logps/chosen": -428.0, + "logps/rejected": -458.0, + "loss": 0.0119, + "rewards/accuracies": 0.984375, + "rewards/chosen": -0.88671875, + "rewards/margins": 14.9375, + "rewards/rejected": -15.8125, + "step": 2240 + }, + { + "epoch": 2.7496932515337424, + "grad_norm": 1.5428893428421564, + "learning_rate": 1.2644628099173553e-07, + "logits/chosen": -0.251953125, + "logits/rejected": -0.333984375, + "logps/chosen": -376.0, + "logps/rejected": -436.0, + "loss": 0.0025, + "rewards/accuracies": 1.0, + "rewards/chosen": -1.3125, + "rewards/margins": 14.625, + "rewards/rejected": -15.9375, + "step": 2241 + }, + { + "epoch": 2.750920245398773, + "grad_norm": 1.0946317624157251, + "learning_rate": 1.2582644628099175e-07, + "logits/chosen": -0.271484375, + "logits/rejected": -0.44921875, + "logps/chosen": -354.0, + "logps/rejected": -398.0, + "loss": 0.0031, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.671875, + "rewards/margins": 14.375, + "rewards/rejected": -15.0625, + "step": 2242 + }, + { + "epoch": 2.752147239263804, + "grad_norm": 1.3647294507982635, + "learning_rate": 1.2520661157024795e-07, + "logits/chosen": -0.306640625, + "logits/rejected": -0.396484375, + "logps/chosen": -406.0, + "logps/rejected": -444.0, + "loss": 0.0077, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -1.2265625, + "rewards/margins": 15.0, + "rewards/rejected": -16.25, + "step": 2243 + }, + { + "epoch": 2.7533742331288344, + "grad_norm": 2.5145615886260337, + "learning_rate": 1.2458677685950414e-07, + "logits/chosen": -0.3125, + "logits/rejected": -0.486328125, + "logps/chosen": -402.0, + "logps/rejected": -428.0, + "loss": 0.0097, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -1.546875, + "rewards/margins": 15.125, + "rewards/rejected": -16.625, + "step": 2244 + }, + { + "epoch": 2.754601226993865, + "grad_norm": 0.6452020514295409, + "learning_rate": 1.2396694214876034e-07, + "logits/chosen": -0.1640625, + "logits/rejected": -0.396484375, + "logps/chosen": -400.0, + "logps/rejected": -432.0, + "loss": 0.0013, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.00067901611328125, + "rewards/margins": 14.875, + "rewards/rejected": -14.875, + "step": 2245 + }, + { + "epoch": 2.755828220858896, + "grad_norm": 1.6525267253749958, + "learning_rate": 1.2334710743801653e-07, + "logits/chosen": -0.318359375, + "logits/rejected": -0.52734375, + "logps/chosen": -360.0, + "logps/rejected": -382.0, + "loss": 0.0027, + "rewards/accuracies": 1.0, + "rewards/chosen": -1.1328125, + "rewards/margins": 14.375, + "rewards/rejected": -15.5, + "step": 2246 + }, + { + "epoch": 2.7570552147239265, + "grad_norm": 2.5046551673504007, + "learning_rate": 1.2272727272727272e-07, + "logits/chosen": -0.23828125, + "logits/rejected": -0.44921875, + "logps/chosen": -454.0, + "logps/rejected": -470.0, + "loss": 0.003, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.37890625, + "rewards/margins": 16.625, + "rewards/rejected": -17.0, + "step": 2247 + }, + { + "epoch": 2.758282208588957, + "grad_norm": 2.8992847450921495, + "learning_rate": 1.2210743801652892e-07, + "logits/chosen": -0.2314453125, + "logits/rejected": -0.3359375, + "logps/chosen": -418.0, + "logps/rejected": -456.0, + "loss": 0.0044, + "rewards/accuracies": 1.0, + "rewards/chosen": -1.3125, + "rewards/margins": 14.9375, + "rewards/rejected": -16.25, + "step": 2248 + }, + { + "epoch": 2.759509202453988, + "grad_norm": 0.6877665833056033, + "learning_rate": 1.214876033057851e-07, + "logits/chosen": -0.22265625, + "logits/rejected": -0.388671875, + "logps/chosen": -440.0, + "logps/rejected": -412.0, + "loss": 0.0013, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.34375, + "rewards/margins": 14.1875, + "rewards/rejected": -14.5, + "step": 2249 + }, + { + "epoch": 2.7607361963190185, + "grad_norm": 2.2213518369836134, + "learning_rate": 1.208677685950413e-07, + "logits/chosen": -0.1787109375, + "logits/rejected": -0.34765625, + "logps/chosen": -372.0, + "logps/rejected": -422.0, + "loss": 0.0037, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.82421875, + "rewards/margins": 13.3125, + "rewards/rejected": -14.1875, + "step": 2250 + }, + { + "epoch": 2.761963190184049, + "grad_norm": 2.2293080927446187, + "learning_rate": 1.2024793388429753e-07, + "logits/chosen": -0.185546875, + "logits/rejected": -0.3359375, + "logps/chosen": -384.0, + "logps/rejected": -414.0, + "loss": 0.0044, + "rewards/accuracies": 1.0, + "rewards/chosen": -1.5546875, + "rewards/margins": 13.0625, + "rewards/rejected": -14.625, + "step": 2251 + }, + { + "epoch": 2.76319018404908, + "grad_norm": 0.734365168671495, + "learning_rate": 1.1962809917355372e-07, + "logits/chosen": -0.2109375, + "logits/rejected": -0.3515625, + "logps/chosen": -404.0, + "logps/rejected": -494.0, + "loss": 0.0014, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.9765625, + "rewards/margins": 15.4375, + "rewards/rejected": -16.375, + "step": 2252 + }, + { + "epoch": 2.7644171779141105, + "grad_norm": 1.1985213795877028, + "learning_rate": 1.1900826446280993e-07, + "logits/chosen": -0.361328125, + "logits/rejected": -0.50390625, + "logps/chosen": -410.0, + "logps/rejected": -446.0, + "loss": 0.0018, + "rewards/accuracies": 1.0, + "rewards/chosen": -1.1015625, + "rewards/margins": 14.75, + "rewards/rejected": -15.875, + "step": 2253 + }, + { + "epoch": 2.765644171779141, + "grad_norm": 0.8662752321202715, + "learning_rate": 1.1838842975206612e-07, + "logits/chosen": -0.251953125, + "logits/rejected": -0.435546875, + "logps/chosen": -368.0, + "logps/rejected": -408.0, + "loss": 0.0073, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.166015625, + "rewards/margins": 14.9375, + "rewards/rejected": -15.125, + "step": 2254 + }, + { + "epoch": 2.766871165644172, + "grad_norm": 1.7234132790889785, + "learning_rate": 1.1776859504132232e-07, + "logits/chosen": -0.1943359375, + "logits/rejected": -0.384765625, + "logps/chosen": -398.0, + "logps/rejected": -444.0, + "loss": 0.0026, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.77734375, + "rewards/margins": 15.375, + "rewards/rejected": -16.125, + "step": 2255 + }, + { + "epoch": 2.7680981595092025, + "grad_norm": 4.306073064451764, + "learning_rate": 1.1714876033057851e-07, + "logits/chosen": -0.216796875, + "logits/rejected": -0.40625, + "logps/chosen": -418.0, + "logps/rejected": -422.0, + "loss": 0.0058, + "rewards/accuracies": 1.0, + "rewards/chosen": -1.171875, + "rewards/margins": 14.75, + "rewards/rejected": -15.875, + "step": 2256 + }, + { + "epoch": 2.769325153374233, + "grad_norm": 0.7567216732455713, + "learning_rate": 1.1652892561983472e-07, + "logits/chosen": -0.25390625, + "logits/rejected": -0.484375, + "logps/chosen": -430.0, + "logps/rejected": -468.0, + "loss": 0.0013, + "rewards/accuracies": 1.0, + "rewards/chosen": -1.140625, + "rewards/margins": 14.75, + "rewards/rejected": -15.9375, + "step": 2257 + }, + { + "epoch": 2.770552147239264, + "grad_norm": 1.022394093982203, + "learning_rate": 1.1590909090909091e-07, + "logits/chosen": -0.2373046875, + "logits/rejected": -0.40625, + "logps/chosen": -388.0, + "logps/rejected": -438.0, + "loss": 0.0017, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.99609375, + "rewards/margins": 15.1875, + "rewards/rejected": -16.25, + "step": 2258 + }, + { + "epoch": 2.7717791411042945, + "grad_norm": 0.9849611181620934, + "learning_rate": 1.1528925619834711e-07, + "logits/chosen": -0.314453125, + "logits/rejected": -0.43359375, + "logps/chosen": -444.0, + "logps/rejected": -476.0, + "loss": 0.0019, + "rewards/accuracies": 1.0, + "rewards/chosen": -1.1171875, + "rewards/margins": 14.375, + "rewards/rejected": -15.5, + "step": 2259 + }, + { + "epoch": 2.773006134969325, + "grad_norm": 1.0865208660675305, + "learning_rate": 1.146694214876033e-07, + "logits/chosen": -0.353515625, + "logits/rejected": -0.5546875, + "logps/chosen": -384.0, + "logps/rejected": -424.0, + "loss": 0.0018, + "rewards/accuracies": 1.0, + "rewards/chosen": -1.3671875, + "rewards/margins": 15.25, + "rewards/rejected": -16.625, + "step": 2260 + }, + { + "epoch": 2.774233128834356, + "grad_norm": 0.8588683551083438, + "learning_rate": 1.140495867768595e-07, + "logits/chosen": -0.203125, + "logits/rejected": -0.41796875, + "logps/chosen": -418.0, + "logps/rejected": -476.0, + "loss": 0.0021, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.1640625, + "rewards/margins": 15.125, + "rewards/rejected": -15.25, + "step": 2261 + }, + { + "epoch": 2.7754601226993865, + "grad_norm": 1.835965819425323, + "learning_rate": 1.1342975206611572e-07, + "logits/chosen": -0.3515625, + "logits/rejected": -0.52734375, + "logps/chosen": -444.0, + "logps/rejected": -440.0, + "loss": 0.0021, + "rewards/accuracies": 1.0, + "rewards/chosen": -1.1328125, + "rewards/margins": 15.5, + "rewards/rejected": -16.625, + "step": 2262 + }, + { + "epoch": 2.776687116564417, + "grad_norm": 0.6246036044533957, + "learning_rate": 1.1280991735537191e-07, + "logits/chosen": -0.26171875, + "logits/rejected": -0.4140625, + "logps/chosen": -408.0, + "logps/rejected": -434.0, + "loss": 0.0063, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -1.15625, + "rewards/margins": 16.0, + "rewards/rejected": -17.125, + "step": 2263 + }, + { + "epoch": 2.777914110429448, + "grad_norm": 2.695525064739411, + "learning_rate": 1.121900826446281e-07, + "logits/chosen": -0.1923828125, + "logits/rejected": -0.34765625, + "logps/chosen": -386.0, + "logps/rejected": -440.0, + "loss": 0.0032, + "rewards/accuracies": 1.0, + "rewards/chosen": -1.09375, + "rewards/margins": 14.625, + "rewards/rejected": -15.75, + "step": 2264 + }, + { + "epoch": 2.7791411042944785, + "grad_norm": 0.4748152420657213, + "learning_rate": 1.115702479338843e-07, + "logits/chosen": -0.259765625, + "logits/rejected": -0.3984375, + "logps/chosen": -420.0, + "logps/rejected": -442.0, + "loss": 0.0009, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.53515625, + "rewards/margins": 15.5, + "rewards/rejected": -16.0, + "step": 2265 + }, + { + "epoch": 2.780368098159509, + "grad_norm": 3.1818743096823527, + "learning_rate": 1.109504132231405e-07, + "logits/chosen": -0.25, + "logits/rejected": -0.49609375, + "logps/chosen": -454.0, + "logps/rejected": -450.0, + "loss": 0.0033, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.515625, + "rewards/margins": 15.4375, + "rewards/rejected": -15.9375, + "step": 2266 + }, + { + "epoch": 2.78159509202454, + "grad_norm": 1.3492746133428128, + "learning_rate": 1.103305785123967e-07, + "logits/chosen": -0.310546875, + "logits/rejected": -0.52734375, + "logps/chosen": -478.0, + "logps/rejected": -470.0, + "loss": 0.0031, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.031494140625, + "rewards/margins": 15.9375, + "rewards/rejected": -15.9375, + "step": 2267 + }, + { + "epoch": 2.7828220858895705, + "grad_norm": 2.565501427573599, + "learning_rate": 1.097107438016529e-07, + "logits/chosen": -0.0498046875, + "logits/rejected": -0.1708984375, + "logps/chosen": -392.0, + "logps/rejected": -440.0, + "loss": 0.0033, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.625, + "rewards/margins": 14.125, + "rewards/rejected": -14.8125, + "step": 2268 + }, + { + "epoch": 2.784049079754601, + "grad_norm": 1.7377255307493438, + "learning_rate": 1.0909090909090909e-07, + "logits/chosen": -0.283203125, + "logits/rejected": -0.431640625, + "logps/chosen": -394.0, + "logps/rejected": -430.0, + "loss": 0.0035, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.3046875, + "rewards/margins": 14.875, + "rewards/rejected": -15.1875, + "step": 2269 + }, + { + "epoch": 2.785276073619632, + "grad_norm": 2.4987933510387257, + "learning_rate": 1.0847107438016528e-07, + "logits/chosen": -0.35546875, + "logits/rejected": -0.46875, + "logps/chosen": -426.0, + "logps/rejected": -458.0, + "loss": 0.0041, + "rewards/accuracies": 1.0, + "rewards/chosen": -1.125, + "rewards/margins": 14.9375, + "rewards/rejected": -16.125, + "step": 2270 + }, + { + "epoch": 2.7865030674846625, + "grad_norm": 0.4770281293782162, + "learning_rate": 1.0785123966942148e-07, + "logits/chosen": -0.1201171875, + "logits/rejected": -0.328125, + "logps/chosen": -398.0, + "logps/rejected": -438.0, + "loss": 0.0067, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -1.4453125, + "rewards/margins": 14.1875, + "rewards/rejected": -15.625, + "step": 2271 + }, + { + "epoch": 2.787730061349693, + "grad_norm": 1.6135760135672685, + "learning_rate": 1.072314049586777e-07, + "logits/chosen": -0.265625, + "logits/rejected": -0.48046875, + "logps/chosen": -424.0, + "logps/rejected": -442.0, + "loss": 0.0089, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -1.21875, + "rewards/margins": 14.5, + "rewards/rejected": -15.6875, + "step": 2272 + }, + { + "epoch": 2.788957055214724, + "grad_norm": 1.2135461327941435, + "learning_rate": 1.0661157024793389e-07, + "logits/chosen": -0.09814453125, + "logits/rejected": -0.294921875, + "logps/chosen": -452.0, + "logps/rejected": -446.0, + "loss": 0.0019, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.001007080078125, + "rewards/margins": 16.0, + "rewards/rejected": -16.0, + "step": 2273 + }, + { + "epoch": 2.7901840490797545, + "grad_norm": 1.3804715900570843, + "learning_rate": 1.0599173553719009e-07, + "logits/chosen": -0.2216796875, + "logits/rejected": -0.455078125, + "logps/chosen": -444.0, + "logps/rejected": -416.0, + "loss": 0.003, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.232421875, + "rewards/margins": 14.1875, + "rewards/rejected": -14.4375, + "step": 2274 + }, + { + "epoch": 2.791411042944785, + "grad_norm": 2.2958207178728687, + "learning_rate": 1.0537190082644628e-07, + "logits/chosen": -0.330078125, + "logits/rejected": -0.53515625, + "logps/chosen": -458.0, + "logps/rejected": -456.0, + "loss": 0.0044, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.49609375, + "rewards/margins": 15.9375, + "rewards/rejected": -15.4375, + "step": 2275 + }, + { + "epoch": 2.792638036809816, + "grad_norm": 0.5704313116557475, + "learning_rate": 1.0475206611570249e-07, + "logits/chosen": -0.2109375, + "logits/rejected": -0.36328125, + "logps/chosen": -414.0, + "logps/rejected": -464.0, + "loss": 0.0066, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.25, + "rewards/margins": 16.125, + "rewards/rejected": -15.875, + "step": 2276 + }, + { + "epoch": 2.7938650306748465, + "grad_norm": 0.3587396741687113, + "learning_rate": 1.0413223140495868e-07, + "logits/chosen": -0.255859375, + "logits/rejected": -0.326171875, + "logps/chosen": -366.0, + "logps/rejected": -444.0, + "loss": 0.0009, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.61328125, + "rewards/margins": 14.625, + "rewards/rejected": -15.25, + "step": 2277 + }, + { + "epoch": 2.795092024539877, + "grad_norm": 0.9592155303436893, + "learning_rate": 1.0351239669421488e-07, + "logits/chosen": -0.234375, + "logits/rejected": -0.46484375, + "logps/chosen": -422.0, + "logps/rejected": -450.0, + "loss": 0.0024, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.28125, + "rewards/margins": 15.25, + "rewards/rejected": -15.5625, + "step": 2278 + }, + { + "epoch": 2.796319018404908, + "grad_norm": 2.230844944330533, + "learning_rate": 1.0289256198347107e-07, + "logits/chosen": -0.263671875, + "logits/rejected": -0.427734375, + "logps/chosen": -428.0, + "logps/rejected": -432.0, + "loss": 0.0029, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.435546875, + "rewards/margins": 14.9375, + "rewards/rejected": -15.375, + "step": 2279 + }, + { + "epoch": 2.7975460122699385, + "grad_norm": 0.6545569974009644, + "learning_rate": 1.0227272727272727e-07, + "logits/chosen": -0.1962890625, + "logits/rejected": -0.41796875, + "logps/chosen": -458.0, + "logps/rejected": -446.0, + "loss": 0.0014, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.703125, + "rewards/margins": 15.0, + "rewards/rejected": -15.6875, + "step": 2280 + }, + { + "epoch": 2.7987730061349696, + "grad_norm": 1.0045554292035184, + "learning_rate": 1.0165289256198346e-07, + "logits/chosen": -0.3203125, + "logits/rejected": -0.5078125, + "logps/chosen": -462.0, + "logps/rejected": -478.0, + "loss": 0.0011, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.388671875, + "rewards/margins": 16.25, + "rewards/rejected": -16.625, + "step": 2281 + }, + { + "epoch": 2.8, + "grad_norm": 1.1061589479027565, + "learning_rate": 1.0103305785123968e-07, + "logits/chosen": -0.29296875, + "logits/rejected": -0.53125, + "logps/chosen": -494.0, + "logps/rejected": -488.0, + "loss": 0.0017, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.06396484375, + "rewards/margins": 15.8125, + "rewards/rejected": -15.875, + "step": 2282 + }, + { + "epoch": 2.801226993865031, + "grad_norm": 0.5762467009400392, + "learning_rate": 1.0041322314049587e-07, + "logits/chosen": -0.216796875, + "logits/rejected": -0.431640625, + "logps/chosen": -394.0, + "logps/rejected": -448.0, + "loss": 0.0013, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.81640625, + "rewards/margins": 15.125, + "rewards/rejected": -15.9375, + "step": 2283 + }, + { + "epoch": 2.802453987730061, + "grad_norm": 0.29883131568411625, + "learning_rate": 9.979338842975207e-08, + "logits/chosen": -0.28515625, + "logits/rejected": -0.474609375, + "logps/chosen": -362.0, + "logps/rejected": -410.0, + "loss": 0.0007, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.64453125, + "rewards/margins": 15.125, + "rewards/rejected": -15.75, + "step": 2284 + }, + { + "epoch": 2.8036809815950923, + "grad_norm": 1.0969112760505852, + "learning_rate": 9.917355371900826e-08, + "logits/chosen": -0.2421875, + "logits/rejected": -0.4921875, + "logps/chosen": -410.0, + "logps/rejected": -416.0, + "loss": 0.0019, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.58984375, + "rewards/margins": 15.3125, + "rewards/rejected": -15.9375, + "step": 2285 + }, + { + "epoch": 2.8049079754601225, + "grad_norm": 0.7942142192591548, + "learning_rate": 9.855371900826447e-08, + "logits/chosen": -0.294921875, + "logits/rejected": -0.421875, + "logps/chosen": -394.0, + "logps/rejected": -476.0, + "loss": 0.0016, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.341796875, + "rewards/margins": 15.125, + "rewards/rejected": -15.5, + "step": 2286 + }, + { + "epoch": 2.8061349693251536, + "grad_norm": 1.6644953284420183, + "learning_rate": 9.793388429752066e-08, + "logits/chosen": -0.2294921875, + "logits/rejected": -0.4375, + "logps/chosen": -390.0, + "logps/rejected": -412.0, + "loss": 0.0029, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.248046875, + "rewards/margins": 14.6875, + "rewards/rejected": -14.9375, + "step": 2287 + }, + { + "epoch": 2.807361963190184, + "grad_norm": 0.7321338810437472, + "learning_rate": 9.731404958677686e-08, + "logits/chosen": -0.330078125, + "logits/rejected": -0.47265625, + "logps/chosen": -388.0, + "logps/rejected": -442.0, + "loss": 0.0069, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -1.25, + "rewards/margins": 15.4375, + "rewards/rejected": -16.75, + "step": 2288 + }, + { + "epoch": 2.808588957055215, + "grad_norm": 1.5333119254885192, + "learning_rate": 9.669421487603305e-08, + "logits/chosen": -0.203125, + "logits/rejected": -0.41796875, + "logps/chosen": -432.0, + "logps/rejected": -434.0, + "loss": 0.0039, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.76953125, + "rewards/margins": 15.625, + "rewards/rejected": -16.375, + "step": 2289 + }, + { + "epoch": 2.809815950920245, + "grad_norm": 0.88964813518976, + "learning_rate": 9.607438016528925e-08, + "logits/chosen": -0.2412109375, + "logits/rejected": -0.443359375, + "logps/chosen": -424.0, + "logps/rejected": -464.0, + "loss": 0.0016, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.212890625, + "rewards/margins": 14.8125, + "rewards/rejected": -14.625, + "step": 2290 + }, + { + "epoch": 2.8110429447852763, + "grad_norm": 1.1230227150855807, + "learning_rate": 9.545454545454544e-08, + "logits/chosen": -0.291015625, + "logits/rejected": -0.412109375, + "logps/chosen": -392.0, + "logps/rejected": -450.0, + "loss": 0.002, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.470703125, + "rewards/margins": 15.875, + "rewards/rejected": -16.375, + "step": 2291 + }, + { + "epoch": 2.8122699386503065, + "grad_norm": 0.8451515479179217, + "learning_rate": 9.483471074380166e-08, + "logits/chosen": -0.341796875, + "logits/rejected": -0.421875, + "logps/chosen": -420.0, + "logps/rejected": -448.0, + "loss": 0.0013, + "rewards/accuracies": 1.0, + "rewards/chosen": -1.4296875, + "rewards/margins": 15.8125, + "rewards/rejected": -17.25, + "step": 2292 + }, + { + "epoch": 2.8134969325153376, + "grad_norm": 0.4551915966853917, + "learning_rate": 9.421487603305786e-08, + "logits/chosen": -0.16796875, + "logits/rejected": -0.421875, + "logps/chosen": -516.0, + "logps/rejected": -496.0, + "loss": 0.0064, + "rewards/accuracies": 0.9921875, + "rewards/chosen": 0.92578125, + "rewards/margins": 16.25, + "rewards/rejected": -15.375, + "step": 2293 + }, + { + "epoch": 2.814723926380368, + "grad_norm": 2.2249998837464964, + "learning_rate": 9.359504132231405e-08, + "logits/chosen": -0.36328125, + "logits/rejected": -0.490234375, + "logps/chosen": -428.0, + "logps/rejected": -470.0, + "loss": 0.0074, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.9921875, + "rewards/margins": 15.875, + "rewards/rejected": -16.875, + "step": 2294 + }, + { + "epoch": 2.815950920245399, + "grad_norm": 3.0883191465144684, + "learning_rate": 9.297520661157026e-08, + "logits/chosen": -0.1474609375, + "logits/rejected": -0.416015625, + "logps/chosen": -410.0, + "logps/rejected": -448.0, + "loss": 0.01, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -1.1640625, + "rewards/margins": 15.25, + "rewards/rejected": -16.5, + "step": 2295 + }, + { + "epoch": 2.817177914110429, + "grad_norm": 2.3338329719027358, + "learning_rate": 9.235537190082645e-08, + "logits/chosen": -0.322265625, + "logits/rejected": -0.546875, + "logps/chosen": -456.0, + "logps/rejected": -472.0, + "loss": 0.0041, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.65234375, + "rewards/margins": 15.1875, + "rewards/rejected": -15.875, + "step": 2296 + }, + { + "epoch": 2.8184049079754603, + "grad_norm": 0.7388070422646501, + "learning_rate": 9.173553719008265e-08, + "logits/chosen": -0.314453125, + "logits/rejected": -0.431640625, + "logps/chosen": -396.0, + "logps/rejected": -468.0, + "loss": 0.0014, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.80859375, + "rewards/margins": 15.9375, + "rewards/rejected": -16.75, + "step": 2297 + }, + { + "epoch": 2.819631901840491, + "grad_norm": 2.3398442790763356, + "learning_rate": 9.111570247933884e-08, + "logits/chosen": -0.220703125, + "logits/rejected": -0.384765625, + "logps/chosen": -358.0, + "logps/rejected": -398.0, + "loss": 0.0095, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.98046875, + "rewards/margins": 14.5, + "rewards/rejected": -15.5, + "step": 2298 + }, + { + "epoch": 2.8208588957055216, + "grad_norm": 2.419522489582249, + "learning_rate": 9.049586776859505e-08, + "logits/chosen": -0.1826171875, + "logits/rejected": -0.357421875, + "logps/chosen": -366.0, + "logps/rejected": -422.0, + "loss": 0.0047, + "rewards/accuracies": 1.0, + "rewards/chosen": -1.171875, + "rewards/margins": 14.375, + "rewards/rejected": -15.5625, + "step": 2299 + }, + { + "epoch": 2.8220858895705523, + "grad_norm": 0.5353583448894994, + "learning_rate": 8.987603305785124e-08, + "logits/chosen": -0.162109375, + "logits/rejected": -0.369140625, + "logps/chosen": -442.0, + "logps/rejected": -436.0, + "loss": 0.0014, + "rewards/accuracies": 1.0, + "rewards/chosen": -1.9921875, + "rewards/margins": 14.9375, + "rewards/rejected": -16.875, + "step": 2300 + }, + { + "epoch": 2.823312883435583, + "grad_norm": 1.2898802061740815, + "learning_rate": 8.925619834710744e-08, + "logits/chosen": -0.318359375, + "logits/rejected": -0.50390625, + "logps/chosen": -356.0, + "logps/rejected": -384.0, + "loss": 0.0042, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.90234375, + "rewards/margins": 14.8125, + "rewards/rejected": -15.75, + "step": 2301 + }, + { + "epoch": 2.8245398773006136, + "grad_norm": 0.8769168863248077, + "learning_rate": 8.863636363636363e-08, + "logits/chosen": -0.43359375, + "logits/rejected": -0.640625, + "logps/chosen": -414.0, + "logps/rejected": -434.0, + "loss": 0.0018, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.875, + "rewards/margins": 15.625, + "rewards/rejected": -16.5, + "step": 2302 + }, + { + "epoch": 2.8257668711656443, + "grad_norm": 2.38274554271401, + "learning_rate": 8.801652892561984e-08, + "logits/chosen": -0.1533203125, + "logits/rejected": -0.265625, + "logps/chosen": -376.0, + "logps/rejected": -430.0, + "loss": 0.0025, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.984375, + "rewards/margins": 14.1875, + "rewards/rejected": -15.125, + "step": 2303 + }, + { + "epoch": 2.826993865030675, + "grad_norm": 2.333465264592718, + "learning_rate": 8.739669421487605e-08, + "logits/chosen": -0.189453125, + "logits/rejected": -0.455078125, + "logps/chosen": -396.0, + "logps/rejected": -414.0, + "loss": 0.0046, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.31640625, + "rewards/margins": 15.4375, + "rewards/rejected": -15.75, + "step": 2304 + }, + { + "epoch": 2.8282208588957056, + "grad_norm": 0.4390239879976031, + "learning_rate": 8.677685950413224e-08, + "logits/chosen": -0.2578125, + "logits/rejected": -0.412109375, + "logps/chosen": -320.0, + "logps/rejected": -396.0, + "loss": 0.0011, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.73828125, + "rewards/margins": 15.0, + "rewards/rejected": -15.75, + "step": 2305 + }, + { + "epoch": 2.8294478527607363, + "grad_norm": 1.0877025940069465, + "learning_rate": 8.615702479338843e-08, + "logits/chosen": -0.31640625, + "logits/rejected": -0.470703125, + "logps/chosen": -414.0, + "logps/rejected": -446.0, + "loss": 0.0078, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -1.5234375, + "rewards/margins": 14.625, + "rewards/rejected": -16.125, + "step": 2306 + }, + { + "epoch": 2.830674846625767, + "grad_norm": 0.825209595226472, + "learning_rate": 8.553719008264463e-08, + "logits/chosen": -0.2265625, + "logits/rejected": -0.326171875, + "logps/chosen": -384.0, + "logps/rejected": -412.0, + "loss": 0.0077, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.4765625, + "rewards/margins": 14.375, + "rewards/rejected": -14.8125, + "step": 2307 + }, + { + "epoch": 2.8319018404907976, + "grad_norm": 1.5454128798823934, + "learning_rate": 8.491735537190082e-08, + "logits/chosen": -0.35546875, + "logits/rejected": -0.55078125, + "logps/chosen": -412.0, + "logps/rejected": -464.0, + "loss": 0.0034, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.6796875, + "rewards/margins": 16.25, + "rewards/rejected": -16.875, + "step": 2308 + }, + { + "epoch": 2.8331288343558283, + "grad_norm": 0.8009715448405459, + "learning_rate": 8.429752066115703e-08, + "logits/chosen": -0.224609375, + "logits/rejected": -0.365234375, + "logps/chosen": -356.0, + "logps/rejected": -422.0, + "loss": 0.0027, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.8359375, + "rewards/margins": 15.0625, + "rewards/rejected": -15.875, + "step": 2309 + }, + { + "epoch": 2.834355828220859, + "grad_norm": 0.8904740221416118, + "learning_rate": 8.367768595041322e-08, + "logits/chosen": -0.314453125, + "logits/rejected": -0.498046875, + "logps/chosen": -422.0, + "logps/rejected": -436.0, + "loss": 0.0021, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.7421875, + "rewards/margins": 15.5625, + "rewards/rejected": -16.25, + "step": 2310 + }, + { + "epoch": 2.8355828220858896, + "grad_norm": 0.35488072481003313, + "learning_rate": 8.305785123966942e-08, + "logits/chosen": -0.357421875, + "logits/rejected": -0.53125, + "logps/chosen": -398.0, + "logps/rejected": -450.0, + "loss": 0.0006, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.2431640625, + "rewards/margins": 16.5, + "rewards/rejected": -16.75, + "step": 2311 + }, + { + "epoch": 2.8368098159509203, + "grad_norm": 3.0136855594669676, + "learning_rate": 8.243801652892563e-08, + "logits/chosen": -0.2431640625, + "logits/rejected": -0.34765625, + "logps/chosen": -384.0, + "logps/rejected": -424.0, + "loss": 0.0032, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.57421875, + "rewards/margins": 15.6875, + "rewards/rejected": -16.25, + "step": 2312 + }, + { + "epoch": 2.838036809815951, + "grad_norm": 19.060215902254743, + "learning_rate": 8.181818181818182e-08, + "logits/chosen": -0.32421875, + "logits/rejected": -0.5546875, + "logps/chosen": -432.0, + "logps/rejected": -446.0, + "loss": 0.0349, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -1.2578125, + "rewards/margins": 14.6875, + "rewards/rejected": -15.9375, + "step": 2313 + }, + { + "epoch": 2.8392638036809816, + "grad_norm": 0.7561268449331441, + "learning_rate": 8.119834710743803e-08, + "logits/chosen": -0.283203125, + "logits/rejected": -0.447265625, + "logps/chosen": -378.0, + "logps/rejected": -448.0, + "loss": 0.0015, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.17578125, + "rewards/margins": 15.5625, + "rewards/rejected": -15.6875, + "step": 2314 + }, + { + "epoch": 2.8404907975460123, + "grad_norm": 5.852740131739244, + "learning_rate": 8.057851239669422e-08, + "logits/chosen": -0.291015625, + "logits/rejected": -0.4140625, + "logps/chosen": -382.0, + "logps/rejected": -466.0, + "loss": 0.0061, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.30859375, + "rewards/margins": 16.0, + "rewards/rejected": -16.375, + "step": 2315 + }, + { + "epoch": 2.841717791411043, + "grad_norm": 1.7271422300263546, + "learning_rate": 7.995867768595042e-08, + "logits/chosen": -0.408203125, + "logits/rejected": -0.58984375, + "logps/chosen": -336.0, + "logps/rejected": -398.0, + "loss": 0.0076, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -1.03125, + "rewards/margins": 15.0, + "rewards/rejected": -16.0, + "step": 2316 + }, + { + "epoch": 2.8429447852760736, + "grad_norm": 2.943595740253166, + "learning_rate": 7.933884297520661e-08, + "logits/chosen": -0.25, + "logits/rejected": -0.408203125, + "logps/chosen": -422.0, + "logps/rejected": -464.0, + "loss": 0.0077, + "rewards/accuracies": 1.0, + "rewards/chosen": -1.2421875, + "rewards/margins": 14.875, + "rewards/rejected": -16.125, + "step": 2317 + }, + { + "epoch": 2.8441717791411043, + "grad_norm": 1.469457294349218, + "learning_rate": 7.87190082644628e-08, + "logits/chosen": -0.2158203125, + "logits/rejected": -0.32421875, + "logps/chosen": -404.0, + "logps/rejected": -438.0, + "loss": 0.0025, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.6953125, + "rewards/margins": 14.4375, + "rewards/rejected": -15.125, + "step": 2318 + }, + { + "epoch": 2.845398773006135, + "grad_norm": 1.048026677580622, + "learning_rate": 7.809917355371901e-08, + "logits/chosen": -0.259765625, + "logits/rejected": -0.298828125, + "logps/chosen": -420.0, + "logps/rejected": -444.0, + "loss": 0.0072, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.06494140625, + "rewards/margins": 14.8125, + "rewards/rejected": -14.875, + "step": 2319 + }, + { + "epoch": 2.8466257668711656, + "grad_norm": 0.7658022684118194, + "learning_rate": 7.74793388429752e-08, + "logits/chosen": -0.265625, + "logits/rejected": -0.400390625, + "logps/chosen": -384.0, + "logps/rejected": -416.0, + "loss": 0.0015, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.43359375, + "rewards/margins": 14.4375, + "rewards/rejected": -14.875, + "step": 2320 + }, + { + "epoch": 2.8478527607361963, + "grad_norm": 1.7444393516001366, + "learning_rate": 7.68595041322314e-08, + "logits/chosen": -0.330078125, + "logits/rejected": -0.486328125, + "logps/chosen": -384.0, + "logps/rejected": -418.0, + "loss": 0.0026, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.46875, + "rewards/margins": 14.375, + "rewards/rejected": -14.8125, + "step": 2321 + }, + { + "epoch": 2.849079754601227, + "grad_norm": 0.5692165278958462, + "learning_rate": 7.623966942148761e-08, + "logits/chosen": -0.228515625, + "logits/rejected": -0.396484375, + "logps/chosen": -428.0, + "logps/rejected": -452.0, + "loss": 0.0013, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.3359375, + "rewards/margins": 14.75, + "rewards/rejected": -15.0625, + "step": 2322 + }, + { + "epoch": 2.8503067484662576, + "grad_norm": 5.741641415819842, + "learning_rate": 7.56198347107438e-08, + "logits/chosen": -0.2255859375, + "logits/rejected": -0.294921875, + "logps/chosen": -360.0, + "logps/rejected": -418.0, + "loss": 0.0073, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.8359375, + "rewards/margins": 13.75, + "rewards/rejected": -14.625, + "step": 2323 + }, + { + "epoch": 2.8515337423312883, + "grad_norm": 0.894350470843216, + "learning_rate": 7.500000000000001e-08, + "logits/chosen": -0.275390625, + "logits/rejected": -0.412109375, + "logps/chosen": -400.0, + "logps/rejected": -418.0, + "loss": 0.0015, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.58984375, + "rewards/margins": 14.875, + "rewards/rejected": -15.4375, + "step": 2324 + }, + { + "epoch": 2.852760736196319, + "grad_norm": 1.5487468531182036, + "learning_rate": 7.43801652892562e-08, + "logits/chosen": -0.197265625, + "logits/rejected": -0.369140625, + "logps/chosen": -408.0, + "logps/rejected": -448.0, + "loss": 0.0136, + "rewards/accuracies": 0.984375, + "rewards/chosen": -1.1796875, + "rewards/margins": 14.3125, + "rewards/rejected": -15.5, + "step": 2325 + }, + { + "epoch": 2.8539877300613496, + "grad_norm": 0.4089023423427577, + "learning_rate": 7.37603305785124e-08, + "logits/chosen": -0.2197265625, + "logits/rejected": -0.4453125, + "logps/chosen": -394.0, + "logps/rejected": -418.0, + "loss": 0.0009, + "rewards/accuracies": 1.0, + "rewards/chosen": -1.515625, + "rewards/margins": 13.9375, + "rewards/rejected": -15.4375, + "step": 2326 + }, + { + "epoch": 2.8552147239263803, + "grad_norm": 1.0532761161584236, + "learning_rate": 7.314049586776859e-08, + "logits/chosen": -0.271484375, + "logits/rejected": -0.458984375, + "logps/chosen": -382.0, + "logps/rejected": -430.0, + "loss": 0.0015, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.8046875, + "rewards/margins": 14.375, + "rewards/rejected": -15.1875, + "step": 2327 + }, + { + "epoch": 2.856441717791411, + "grad_norm": 1.4837906430698262, + "learning_rate": 7.252066115702479e-08, + "logits/chosen": -0.2353515625, + "logits/rejected": -0.392578125, + "logps/chosen": -394.0, + "logps/rejected": -402.0, + "loss": 0.0024, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.12109375, + "rewards/margins": 14.5, + "rewards/rejected": -14.625, + "step": 2328 + }, + { + "epoch": 2.8576687116564417, + "grad_norm": 1.7768166942923738, + "learning_rate": 7.1900826446281e-08, + "logits/chosen": -0.333984375, + "logits/rejected": -0.5703125, + "logps/chosen": -400.0, + "logps/rejected": -450.0, + "loss": 0.0055, + "rewards/accuracies": 1.0, + "rewards/chosen": -1.796875, + "rewards/margins": 14.9375, + "rewards/rejected": -16.75, + "step": 2329 + }, + { + "epoch": 2.8588957055214723, + "grad_norm": 3.619514325996752, + "learning_rate": 7.128099173553719e-08, + "logits/chosen": -0.267578125, + "logits/rejected": -0.400390625, + "logps/chosen": -458.0, + "logps/rejected": -436.0, + "loss": 0.0043, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.78125, + "rewards/margins": 15.1875, + "rewards/rejected": -15.9375, + "step": 2330 + }, + { + "epoch": 2.860122699386503, + "grad_norm": 0.7900977910433934, + "learning_rate": 7.06611570247934e-08, + "logits/chosen": -0.287109375, + "logits/rejected": -0.4453125, + "logps/chosen": -402.0, + "logps/rejected": -448.0, + "loss": 0.0014, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.83984375, + "rewards/margins": 15.1875, + "rewards/rejected": -16.0, + "step": 2331 + }, + { + "epoch": 2.8613496932515337, + "grad_norm": 13.63891032889161, + "learning_rate": 7.004132231404959e-08, + "logits/chosen": -0.3125, + "logits/rejected": -0.380859375, + "logps/chosen": -376.0, + "logps/rejected": -454.0, + "loss": 0.005, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.11767578125, + "rewards/margins": 16.125, + "rewards/rejected": -16.0, + "step": 2332 + }, + { + "epoch": 2.8625766871165643, + "grad_norm": 0.7240530685919171, + "learning_rate": 6.942148760330578e-08, + "logits/chosen": -0.248046875, + "logits/rejected": -0.453125, + "logps/chosen": -388.0, + "logps/rejected": -440.0, + "loss": 0.0014, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.54296875, + "rewards/margins": 14.9375, + "rewards/rejected": -15.5, + "step": 2333 + }, + { + "epoch": 2.863803680981595, + "grad_norm": 0.8070639580036532, + "learning_rate": 6.880165289256199e-08, + "logits/chosen": -0.322265625, + "logits/rejected": -0.419921875, + "logps/chosen": -376.0, + "logps/rejected": -408.0, + "loss": 0.0019, + "rewards/accuracies": 1.0, + "rewards/chosen": -1.3125, + "rewards/margins": 14.8125, + "rewards/rejected": -16.125, + "step": 2334 + }, + { + "epoch": 2.8650306748466257, + "grad_norm": 0.9943069140433298, + "learning_rate": 6.818181818181819e-08, + "logits/chosen": -0.3984375, + "logits/rejected": -0.5390625, + "logps/chosen": -460.0, + "logps/rejected": -452.0, + "loss": 0.0021, + "rewards/accuracies": 1.0, + "rewards/chosen": -1.015625, + "rewards/margins": 15.5625, + "rewards/rejected": -16.5, + "step": 2335 + }, + { + "epoch": 2.8662576687116563, + "grad_norm": 0.23701051805381054, + "learning_rate": 6.756198347107438e-08, + "logits/chosen": -0.400390625, + "logits/rejected": -0.5859375, + "logps/chosen": -406.0, + "logps/rejected": -452.0, + "loss": 0.0005, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.69140625, + "rewards/margins": 15.375, + "rewards/rejected": -16.0, + "step": 2336 + }, + { + "epoch": 2.867484662576687, + "grad_norm": 0.4832460561673461, + "learning_rate": 6.694214876033057e-08, + "logits/chosen": -0.22265625, + "logits/rejected": -0.390625, + "logps/chosen": -356.0, + "logps/rejected": -450.0, + "loss": 0.0009, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.890625, + "rewards/margins": 14.9375, + "rewards/rejected": -15.875, + "step": 2337 + }, + { + "epoch": 2.8687116564417177, + "grad_norm": 1.582220694083379, + "learning_rate": 6.632231404958677e-08, + "logits/chosen": -0.2099609375, + "logits/rejected": -0.310546875, + "logps/chosen": -370.0, + "logps/rejected": -422.0, + "loss": 0.0028, + "rewards/accuracies": 1.0, + "rewards/chosen": -1.5078125, + "rewards/margins": 13.5, + "rewards/rejected": -15.0, + "step": 2338 + }, + { + "epoch": 2.8699386503067483, + "grad_norm": 0.9090980520121215, + "learning_rate": 6.570247933884298e-08, + "logits/chosen": -0.15625, + "logits/rejected": -0.396484375, + "logps/chosen": -362.0, + "logps/rejected": -386.0, + "loss": 0.0019, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.43359375, + "rewards/margins": 13.625, + "rewards/rejected": -14.0625, + "step": 2339 + }, + { + "epoch": 2.871165644171779, + "grad_norm": 0.5878762185245545, + "learning_rate": 6.508264462809918e-08, + "logits/chosen": -0.2734375, + "logits/rejected": -0.423828125, + "logps/chosen": -428.0, + "logps/rejected": -476.0, + "loss": 0.0013, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.90625, + "rewards/margins": 15.875, + "rewards/rejected": -16.75, + "step": 2340 + }, + { + "epoch": 2.8723926380368097, + "grad_norm": 1.1295573049923202, + "learning_rate": 6.446280991735538e-08, + "logits/chosen": -0.322265625, + "logits/rejected": -0.44921875, + "logps/chosen": -368.0, + "logps/rejected": -458.0, + "loss": 0.0017, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.021240234375, + "rewards/margins": 15.4375, + "rewards/rejected": -15.4375, + "step": 2341 + }, + { + "epoch": 2.8736196319018403, + "grad_norm": 2.890035879542409, + "learning_rate": 6.384297520661157e-08, + "logits/chosen": -0.408203125, + "logits/rejected": -0.65625, + "logps/chosen": -442.0, + "logps/rejected": -458.0, + "loss": 0.0052, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.47265625, + "rewards/margins": 15.9375, + "rewards/rejected": -16.375, + "step": 2342 + }, + { + "epoch": 2.874846625766871, + "grad_norm": 0.8307527095974809, + "learning_rate": 6.322314049586777e-08, + "logits/chosen": -0.326171875, + "logits/rejected": -0.46875, + "logps/chosen": -394.0, + "logps/rejected": -438.0, + "loss": 0.0014, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.86328125, + "rewards/margins": 14.25, + "rewards/rejected": -15.125, + "step": 2343 + }, + { + "epoch": 2.876073619631902, + "grad_norm": 8.561457249691601, + "learning_rate": 6.260330578512397e-08, + "logits/chosen": -0.1962890625, + "logits/rejected": -0.353515625, + "logps/chosen": -406.0, + "logps/rejected": -452.0, + "loss": 0.0076, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.30078125, + "rewards/margins": 15.6875, + "rewards/rejected": -16.0, + "step": 2344 + }, + { + "epoch": 2.8773006134969323, + "grad_norm": 1.0974014587577448, + "learning_rate": 6.198347107438017e-08, + "logits/chosen": -0.373046875, + "logits/rejected": -0.474609375, + "logps/chosen": -378.0, + "logps/rejected": -428.0, + "loss": 0.0019, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.3515625, + "rewards/margins": 15.875, + "rewards/rejected": -16.25, + "step": 2345 + }, + { + "epoch": 2.8785276073619634, + "grad_norm": 0.791933080021046, + "learning_rate": 6.136363636363636e-08, + "logits/chosen": -0.216796875, + "logits/rejected": -0.439453125, + "logps/chosen": -456.0, + "logps/rejected": -462.0, + "loss": 0.007, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.263671875, + "rewards/margins": 15.8125, + "rewards/rejected": -16.0, + "step": 2346 + }, + { + "epoch": 2.8797546012269937, + "grad_norm": 6.731493316101986, + "learning_rate": 6.074380165289256e-08, + "logits/chosen": -0.2734375, + "logits/rejected": -0.41796875, + "logps/chosen": -366.0, + "logps/rejected": -404.0, + "loss": 0.0065, + "rewards/accuracies": 1.0, + "rewards/chosen": -1.03125, + "rewards/margins": 15.0, + "rewards/rejected": -16.0, + "step": 2347 + }, + { + "epoch": 2.880981595092025, + "grad_norm": 1.3363437244882383, + "learning_rate": 6.012396694214876e-08, + "logits/chosen": -0.1865234375, + "logits/rejected": -0.34375, + "logps/chosen": -414.0, + "logps/rejected": -426.0, + "loss": 0.0024, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.69921875, + "rewards/margins": 14.75, + "rewards/rejected": -15.4375, + "step": 2348 + }, + { + "epoch": 2.882208588957055, + "grad_norm": 1.22420374509171, + "learning_rate": 5.9504132231404965e-08, + "logits/chosen": -0.32421875, + "logits/rejected": -0.48046875, + "logps/chosen": -412.0, + "logps/rejected": -422.0, + "loss": 0.0018, + "rewards/accuracies": 1.0, + "rewards/chosen": -1.125, + "rewards/margins": 15.375, + "rewards/rejected": -16.5, + "step": 2349 + }, + { + "epoch": 2.883435582822086, + "grad_norm": 0.4470194453570177, + "learning_rate": 5.888429752066116e-08, + "logits/chosen": -0.37890625, + "logits/rejected": -0.640625, + "logps/chosen": -454.0, + "logps/rejected": -476.0, + "loss": 0.0009, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.54296875, + "rewards/margins": 16.625, + "rewards/rejected": -16.125, + "step": 2350 + }, + { + "epoch": 2.8846625766871163, + "grad_norm": 0.4286588636550065, + "learning_rate": 5.826446280991736e-08, + "logits/chosen": -0.349609375, + "logits/rejected": -0.5234375, + "logps/chosen": -406.0, + "logps/rejected": -452.0, + "loss": 0.0175, + "rewards/accuracies": 0.9765625, + "rewards/chosen": -1.1328125, + "rewards/margins": 15.0, + "rewards/rejected": -16.125, + "step": 2351 + }, + { + "epoch": 2.8858895705521475, + "grad_norm": 0.37690193045206194, + "learning_rate": 5.7644628099173554e-08, + "logits/chosen": -0.240234375, + "logits/rejected": -0.376953125, + "logps/chosen": -362.0, + "logps/rejected": -436.0, + "loss": 0.0008, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.82421875, + "rewards/margins": 15.3125, + "rewards/rejected": -16.125, + "step": 2352 + }, + { + "epoch": 2.8871165644171777, + "grad_norm": 0.9716536389208864, + "learning_rate": 5.702479338842975e-08, + "logits/chosen": -0.232421875, + "logits/rejected": -0.412109375, + "logps/chosen": -422.0, + "logps/rejected": -460.0, + "loss": 0.0017, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.396484375, + "rewards/margins": 16.5, + "rewards/rejected": -17.0, + "step": 2353 + }, + { + "epoch": 2.888343558282209, + "grad_norm": 0.6720100363232397, + "learning_rate": 5.6404958677685956e-08, + "logits/chosen": -0.345703125, + "logits/rejected": -0.47265625, + "logps/chosen": -438.0, + "logps/rejected": -436.0, + "loss": 0.0016, + "rewards/accuracies": 1.0, + "rewards/chosen": -1.0703125, + "rewards/margins": 14.0, + "rewards/rejected": -15.0625, + "step": 2354 + }, + { + "epoch": 2.889570552147239, + "grad_norm": 1.3006350972464606, + "learning_rate": 5.578512396694215e-08, + "logits/chosen": -0.2236328125, + "logits/rejected": -0.25, + "logps/chosen": -366.0, + "logps/rejected": -442.0, + "loss": 0.0023, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.765625, + "rewards/margins": 14.3125, + "rewards/rejected": -15.125, + "step": 2355 + }, + { + "epoch": 2.89079754601227, + "grad_norm": 1.5577258340384186, + "learning_rate": 5.516528925619835e-08, + "logits/chosen": -0.1298828125, + "logits/rejected": -0.3203125, + "logps/chosen": -394.0, + "logps/rejected": -444.0, + "loss": 0.0017, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.29296875, + "rewards/margins": 15.25, + "rewards/rejected": -15.5625, + "step": 2356 + }, + { + "epoch": 2.8920245398773003, + "grad_norm": 8.239596797294748, + "learning_rate": 5.4545454545454545e-08, + "logits/chosen": -0.447265625, + "logits/rejected": -0.5546875, + "logps/chosen": -398.0, + "logps/rejected": -470.0, + "loss": 0.008, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -1.25, + "rewards/margins": 15.1875, + "rewards/rejected": -16.375, + "step": 2357 + }, + { + "epoch": 2.8932515337423315, + "grad_norm": 3.817437809929131, + "learning_rate": 5.392561983471074e-08, + "logits/chosen": -0.2099609375, + "logits/rejected": -0.416015625, + "logps/chosen": -418.0, + "logps/rejected": -456.0, + "loss": 0.0045, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.48828125, + "rewards/margins": 14.5, + "rewards/rejected": -15.0, + "step": 2358 + }, + { + "epoch": 2.894478527607362, + "grad_norm": 0.9424852781702008, + "learning_rate": 5.3305785123966946e-08, + "logits/chosen": -0.27734375, + "logits/rejected": -0.427734375, + "logps/chosen": -482.0, + "logps/rejected": -486.0, + "loss": 0.0014, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.1796875, + "rewards/margins": 16.25, + "rewards/rejected": -16.375, + "step": 2359 + }, + { + "epoch": 2.895705521472393, + "grad_norm": 4.656155915570481, + "learning_rate": 5.268595041322314e-08, + "logits/chosen": -0.279296875, + "logits/rejected": -0.3984375, + "logps/chosen": -370.0, + "logps/rejected": -416.0, + "loss": 0.0071, + "rewards/accuracies": 1.0, + "rewards/chosen": -1.1640625, + "rewards/margins": 13.9375, + "rewards/rejected": -15.125, + "step": 2360 + }, + { + "epoch": 2.8969325153374235, + "grad_norm": 0.8461122398314118, + "learning_rate": 5.206611570247934e-08, + "logits/chosen": -0.1826171875, + "logits/rejected": -0.369140625, + "logps/chosen": -392.0, + "logps/rejected": -446.0, + "loss": 0.0016, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.2294921875, + "rewards/margins": 15.3125, + "rewards/rejected": -15.5625, + "step": 2361 + }, + { + "epoch": 2.898159509202454, + "grad_norm": 0.5229693711482295, + "learning_rate": 5.1446280991735536e-08, + "logits/chosen": -0.232421875, + "logits/rejected": -0.4453125, + "logps/chosen": -404.0, + "logps/rejected": -438.0, + "loss": 0.001, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.40625, + "rewards/margins": 15.625, + "rewards/rejected": -16.0, + "step": 2362 + }, + { + "epoch": 2.899386503067485, + "grad_norm": 1.8508807018314375, + "learning_rate": 5.082644628099173e-08, + "logits/chosen": -0.375, + "logits/rejected": -0.51171875, + "logps/chosen": -400.0, + "logps/rejected": -466.0, + "loss": 0.0026, + "rewards/accuracies": 1.0, + "rewards/chosen": -1.1015625, + "rewards/margins": 15.3125, + "rewards/rejected": -16.375, + "step": 2363 + }, + { + "epoch": 2.9006134969325155, + "grad_norm": 1.2112958022415172, + "learning_rate": 5.020661157024794e-08, + "logits/chosen": -0.251953125, + "logits/rejected": -0.44140625, + "logps/chosen": -440.0, + "logps/rejected": -480.0, + "loss": 0.0016, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.1689453125, + "rewards/margins": 16.5, + "rewards/rejected": -16.625, + "step": 2364 + }, + { + "epoch": 2.901840490797546, + "grad_norm": 1.1421932409583224, + "learning_rate": 4.958677685950413e-08, + "logits/chosen": -0.181640625, + "logits/rejected": -0.287109375, + "logps/chosen": -408.0, + "logps/rejected": -430.0, + "loss": 0.008, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -1.4375, + "rewards/margins": 14.75, + "rewards/rejected": -16.125, + "step": 2365 + }, + { + "epoch": 2.903067484662577, + "grad_norm": 0.35460788616803024, + "learning_rate": 4.896694214876033e-08, + "logits/chosen": -0.26953125, + "logits/rejected": -0.43359375, + "logps/chosen": -422.0, + "logps/rejected": -458.0, + "loss": 0.0061, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.953125, + "rewards/margins": 14.5625, + "rewards/rejected": -15.5, + "step": 2366 + }, + { + "epoch": 2.9042944785276075, + "grad_norm": 0.8332554194820654, + "learning_rate": 4.8347107438016527e-08, + "logits/chosen": -0.2578125, + "logits/rejected": -0.431640625, + "logps/chosen": -344.0, + "logps/rejected": -398.0, + "loss": 0.0014, + "rewards/accuracies": 1.0, + "rewards/chosen": -1.359375, + "rewards/margins": 14.25, + "rewards/rejected": -15.625, + "step": 2367 + }, + { + "epoch": 2.905521472392638, + "grad_norm": 1.9316340753164032, + "learning_rate": 4.772727272727272e-08, + "logits/chosen": -0.35546875, + "logits/rejected": -0.53515625, + "logps/chosen": -416.0, + "logps/rejected": -474.0, + "loss": 0.0027, + "rewards/accuracies": 1.0, + "rewards/chosen": -1.1328125, + "rewards/margins": 15.6875, + "rewards/rejected": -16.875, + "step": 2368 + }, + { + "epoch": 2.906748466257669, + "grad_norm": 1.901430832098615, + "learning_rate": 4.710743801652893e-08, + "logits/chosen": -0.26953125, + "logits/rejected": -0.482421875, + "logps/chosen": -438.0, + "logps/rejected": -454.0, + "loss": 0.0023, + "rewards/accuracies": 1.0, + "rewards/chosen": -1.0, + "rewards/margins": 15.0625, + "rewards/rejected": -16.125, + "step": 2369 + }, + { + "epoch": 2.9079754601226995, + "grad_norm": 2.900061305075675, + "learning_rate": 4.648760330578513e-08, + "logits/chosen": -0.185546875, + "logits/rejected": -0.326171875, + "logps/chosen": -422.0, + "logps/rejected": -446.0, + "loss": 0.0029, + "rewards/accuracies": 1.0, + "rewards/chosen": -1.0078125, + "rewards/margins": 14.625, + "rewards/rejected": -15.625, + "step": 2370 + }, + { + "epoch": 2.90920245398773, + "grad_norm": 1.9229348416819452, + "learning_rate": 4.5867768595041323e-08, + "logits/chosen": -0.2470703125, + "logits/rejected": -0.41796875, + "logps/chosen": -402.0, + "logps/rejected": -426.0, + "loss": 0.0044, + "rewards/accuracies": 1.0, + "rewards/chosen": -1.625, + "rewards/margins": 14.3125, + "rewards/rejected": -15.9375, + "step": 2371 + }, + { + "epoch": 2.910429447852761, + "grad_norm": 0.6438475304429977, + "learning_rate": 4.5247933884297524e-08, + "logits/chosen": -0.259765625, + "logits/rejected": -0.57421875, + "logps/chosen": -424.0, + "logps/rejected": -418.0, + "loss": 0.0011, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.4765625, + "rewards/margins": 16.375, + "rewards/rejected": -16.875, + "step": 2372 + }, + { + "epoch": 2.9116564417177915, + "grad_norm": 2.5367119329823224, + "learning_rate": 4.462809917355372e-08, + "logits/chosen": -0.06005859375, + "logits/rejected": -0.3125, + "logps/chosen": -442.0, + "logps/rejected": -446.0, + "loss": 0.0034, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.05908203125, + "rewards/margins": 15.3125, + "rewards/rejected": -15.375, + "step": 2373 + }, + { + "epoch": 2.912883435582822, + "grad_norm": 0.8950723734258177, + "learning_rate": 4.400826446280992e-08, + "logits/chosen": -0.26953125, + "logits/rejected": -0.462890625, + "logps/chosen": -426.0, + "logps/rejected": -432.0, + "loss": 0.0017, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.396484375, + "rewards/margins": 15.1875, + "rewards/rejected": -15.5625, + "step": 2374 + }, + { + "epoch": 2.914110429447853, + "grad_norm": 1.0070853164949929, + "learning_rate": 4.338842975206612e-08, + "logits/chosen": -0.2373046875, + "logits/rejected": -0.3828125, + "logps/chosen": -398.0, + "logps/rejected": -452.0, + "loss": 0.0017, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.9375, + "rewards/margins": 14.25, + "rewards/rejected": -15.1875, + "step": 2375 + }, + { + "epoch": 2.9153374233128835, + "grad_norm": 5.607432613117638, + "learning_rate": 4.2768595041322314e-08, + "logits/chosen": -0.1328125, + "logits/rejected": -0.29296875, + "logps/chosen": -414.0, + "logps/rejected": -434.0, + "loss": 0.007, + "rewards/accuracies": 1.0, + "rewards/chosen": -1.0703125, + "rewards/margins": 13.9375, + "rewards/rejected": -15.0, + "step": 2376 + }, + { + "epoch": 2.916564417177914, + "grad_norm": 1.4268224397259108, + "learning_rate": 4.2148760330578515e-08, + "logits/chosen": -0.06103515625, + "logits/rejected": -0.2314453125, + "logps/chosen": -384.0, + "logps/rejected": -434.0, + "loss": 0.0026, + "rewards/accuracies": 1.0, + "rewards/chosen": -1.1796875, + "rewards/margins": 14.1875, + "rewards/rejected": -15.375, + "step": 2377 + }, + { + "epoch": 2.917791411042945, + "grad_norm": 1.068691442613502, + "learning_rate": 4.152892561983471e-08, + "logits/chosen": -0.1962890625, + "logits/rejected": -0.419921875, + "logps/chosen": -406.0, + "logps/rejected": -456.0, + "loss": 0.007, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -1.5, + "rewards/margins": 14.25, + "rewards/rejected": -15.75, + "step": 2378 + }, + { + "epoch": 2.9190184049079755, + "grad_norm": 1.1361160049467256, + "learning_rate": 4.090909090909091e-08, + "logits/chosen": -0.2490234375, + "logits/rejected": -0.38671875, + "logps/chosen": -380.0, + "logps/rejected": -434.0, + "loss": 0.0022, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.609375, + "rewards/margins": 14.875, + "rewards/rejected": -15.5, + "step": 2379 + }, + { + "epoch": 2.920245398773006, + "grad_norm": 2.171941714978488, + "learning_rate": 4.028925619834711e-08, + "logits/chosen": -0.24609375, + "logits/rejected": -0.443359375, + "logps/chosen": -416.0, + "logps/rejected": -460.0, + "loss": 0.003, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.9765625, + "rewards/margins": 15.3125, + "rewards/rejected": -16.25, + "step": 2380 + }, + { + "epoch": 2.921472392638037, + "grad_norm": 0.6594243507456602, + "learning_rate": 3.9669421487603305e-08, + "logits/chosen": -0.271484375, + "logits/rejected": -0.482421875, + "logps/chosen": -404.0, + "logps/rejected": -442.0, + "loss": 0.0013, + "rewards/accuracies": 1.0, + "rewards/chosen": -1.3359375, + "rewards/margins": 15.0, + "rewards/rejected": -16.25, + "step": 2381 + }, + { + "epoch": 2.9226993865030675, + "grad_norm": 0.7803526320138159, + "learning_rate": 3.9049586776859506e-08, + "logits/chosen": -0.2333984375, + "logits/rejected": -0.408203125, + "logps/chosen": -390.0, + "logps/rejected": -426.0, + "loss": 0.0071, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.90234375, + "rewards/margins": 14.9375, + "rewards/rejected": -15.8125, + "step": 2382 + }, + { + "epoch": 2.923926380368098, + "grad_norm": 1.6864870070689804, + "learning_rate": 3.84297520661157e-08, + "logits/chosen": -0.1796875, + "logits/rejected": -0.37890625, + "logps/chosen": -426.0, + "logps/rejected": -430.0, + "loss": 0.0021, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.6640625, + "rewards/margins": 15.375, + "rewards/rejected": -16.0, + "step": 2383 + }, + { + "epoch": 2.925153374233129, + "grad_norm": 1.1425414963381448, + "learning_rate": 3.78099173553719e-08, + "logits/chosen": -0.2353515625, + "logits/rejected": -0.45703125, + "logps/chosen": -412.0, + "logps/rejected": -454.0, + "loss": 0.0073, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.2138671875, + "rewards/margins": 15.5625, + "rewards/rejected": -15.75, + "step": 2384 + }, + { + "epoch": 2.9263803680981595, + "grad_norm": 1.5688979551791058, + "learning_rate": 3.71900826446281e-08, + "logits/chosen": -0.283203125, + "logits/rejected": -0.490234375, + "logps/chosen": -410.0, + "logps/rejected": -426.0, + "loss": 0.0029, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.85546875, + "rewards/margins": 14.5625, + "rewards/rejected": -15.4375, + "step": 2385 + }, + { + "epoch": 2.92760736196319, + "grad_norm": 3.1225588850268537, + "learning_rate": 3.6570247933884296e-08, + "logits/chosen": -0.232421875, + "logits/rejected": -0.400390625, + "logps/chosen": -424.0, + "logps/rejected": -468.0, + "loss": 0.0037, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.15625, + "rewards/margins": 14.6875, + "rewards/rejected": -14.5625, + "step": 2386 + }, + { + "epoch": 2.928834355828221, + "grad_norm": 0.8531405887295871, + "learning_rate": 3.59504132231405e-08, + "logits/chosen": -0.29296875, + "logits/rejected": -0.486328125, + "logps/chosen": -388.0, + "logps/rejected": -454.0, + "loss": 0.0014, + "rewards/accuracies": 1.0, + "rewards/chosen": -1.1796875, + "rewards/margins": 15.3125, + "rewards/rejected": -16.5, + "step": 2387 + }, + { + "epoch": 2.9300613496932515, + "grad_norm": 0.763591376701033, + "learning_rate": 3.53305785123967e-08, + "logits/chosen": -0.25390625, + "logits/rejected": -0.53515625, + "logps/chosen": -424.0, + "logps/rejected": -446.0, + "loss": 0.0013, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.2275390625, + "rewards/margins": 16.25, + "rewards/rejected": -16.5, + "step": 2388 + }, + { + "epoch": 2.931288343558282, + "grad_norm": 0.9364337453907985, + "learning_rate": 3.471074380165289e-08, + "logits/chosen": -0.17578125, + "logits/rejected": -0.2490234375, + "logps/chosen": -376.0, + "logps/rejected": -446.0, + "loss": 0.0019, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.59375, + "rewards/margins": 14.25, + "rewards/rejected": -14.875, + "step": 2389 + }, + { + "epoch": 2.932515337423313, + "grad_norm": 1.9880954726896056, + "learning_rate": 3.409090909090909e-08, + "logits/chosen": -0.18359375, + "logits/rejected": -0.443359375, + "logps/chosen": -432.0, + "logps/rejected": -434.0, + "loss": 0.0042, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.287109375, + "rewards/margins": 15.25, + "rewards/rejected": -15.5625, + "step": 2390 + }, + { + "epoch": 2.9337423312883435, + "grad_norm": 1.5054421577470123, + "learning_rate": 3.347107438016529e-08, + "logits/chosen": -0.1748046875, + "logits/rejected": -0.4765625, + "logps/chosen": -384.0, + "logps/rejected": -408.0, + "loss": 0.0081, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -1.015625, + "rewards/margins": 13.875, + "rewards/rejected": -14.875, + "step": 2391 + }, + { + "epoch": 2.934969325153374, + "grad_norm": 0.4162938013743738, + "learning_rate": 3.285123966942149e-08, + "logits/chosen": -0.228515625, + "logits/rejected": -0.45703125, + "logps/chosen": -420.0, + "logps/rejected": -456.0, + "loss": 0.0008, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.69140625, + "rewards/margins": 15.6875, + "rewards/rejected": -16.375, + "step": 2392 + }, + { + "epoch": 2.936196319018405, + "grad_norm": 1.1563498071279603, + "learning_rate": 3.223140495867769e-08, + "logits/chosen": -0.146484375, + "logits/rejected": -0.29296875, + "logps/chosen": -356.0, + "logps/rejected": -416.0, + "loss": 0.0047, + "rewards/accuracies": 1.0, + "rewards/chosen": -1.28125, + "rewards/margins": 14.3125, + "rewards/rejected": -15.5625, + "step": 2393 + }, + { + "epoch": 2.9374233128834355, + "grad_norm": 13.931472025317438, + "learning_rate": 3.161157024793388e-08, + "logits/chosen": -0.287109375, + "logits/rejected": -0.453125, + "logps/chosen": -406.0, + "logps/rejected": -456.0, + "loss": 0.0162, + "rewards/accuracies": 0.984375, + "rewards/chosen": -0.8359375, + "rewards/margins": 14.875, + "rewards/rejected": -15.75, + "step": 2394 + }, + { + "epoch": 2.938650306748466, + "grad_norm": 0.6376330005746446, + "learning_rate": 3.0991735537190084e-08, + "logits/chosen": -0.30859375, + "logits/rejected": -0.45703125, + "logps/chosen": -378.0, + "logps/rejected": -426.0, + "loss": 0.0011, + "rewards/accuracies": 1.0, + "rewards/chosen": -1.1328125, + "rewards/margins": 14.9375, + "rewards/rejected": -16.125, + "step": 2395 + }, + { + "epoch": 2.939877300613497, + "grad_norm": 0.36159944995317783, + "learning_rate": 3.037190082644628e-08, + "logits/chosen": -0.3203125, + "logits/rejected": -0.494140625, + "logps/chosen": -424.0, + "logps/rejected": -464.0, + "loss": 0.0009, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.21875, + "rewards/margins": 15.4375, + "rewards/rejected": -15.25, + "step": 2396 + }, + { + "epoch": 2.9411042944785275, + "grad_norm": 2.3252455508312804, + "learning_rate": 2.9752066115702482e-08, + "logits/chosen": -0.2392578125, + "logits/rejected": -0.396484375, + "logps/chosen": -366.0, + "logps/rejected": -440.0, + "loss": 0.0032, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.76953125, + "rewards/margins": 15.125, + "rewards/rejected": -15.875, + "step": 2397 + }, + { + "epoch": 2.942331288343558, + "grad_norm": 6.928930893785217, + "learning_rate": 2.913223140495868e-08, + "logits/chosen": -0.0859375, + "logits/rejected": -0.2275390625, + "logps/chosen": -396.0, + "logps/rejected": -436.0, + "loss": 0.0043, + "rewards/accuracies": 1.0, + "rewards/chosen": -1.6328125, + "rewards/margins": 14.375, + "rewards/rejected": -16.0, + "step": 2398 + }, + { + "epoch": 2.943558282208589, + "grad_norm": 2.3097575550392713, + "learning_rate": 2.8512396694214874e-08, + "logits/chosen": -0.1826171875, + "logits/rejected": -0.369140625, + "logps/chosen": -428.0, + "logps/rejected": -430.0, + "loss": 0.0032, + "rewards/accuracies": 1.0, + "rewards/chosen": -1.15625, + "rewards/margins": 14.9375, + "rewards/rejected": -16.125, + "step": 2399 + }, + { + "epoch": 2.9447852760736195, + "grad_norm": 2.5601897595533187, + "learning_rate": 2.7892561983471075e-08, + "logits/chosen": -0.265625, + "logits/rejected": -0.431640625, + "logps/chosen": -436.0, + "logps/rejected": -464.0, + "loss": 0.0029, + "rewards/accuracies": 1.0, + "rewards/chosen": -1.1015625, + "rewards/margins": 15.6875, + "rewards/rejected": -16.75, + "step": 2400 + }, + { + "epoch": 2.94601226993865, + "grad_norm": 1.0396246433831653, + "learning_rate": 2.7272727272727272e-08, + "logits/chosen": -0.26171875, + "logits/rejected": -0.431640625, + "logps/chosen": -366.0, + "logps/rejected": -416.0, + "loss": 0.0068, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.80078125, + "rewards/margins": 14.625, + "rewards/rejected": -15.4375, + "step": 2401 + }, + { + "epoch": 2.947239263803681, + "grad_norm": 1.4154196488039863, + "learning_rate": 2.6652892561983473e-08, + "logits/chosen": -0.244140625, + "logits/rejected": -0.46484375, + "logps/chosen": -420.0, + "logps/rejected": -450.0, + "loss": 0.002, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.482421875, + "rewards/margins": 15.5, + "rewards/rejected": -16.0, + "step": 2402 + }, + { + "epoch": 2.9484662576687115, + "grad_norm": 3.6306989229811024, + "learning_rate": 2.603305785123967e-08, + "logits/chosen": -0.28125, + "logits/rejected": -0.470703125, + "logps/chosen": -380.0, + "logps/rejected": -426.0, + "loss": 0.0032, + "rewards/accuracies": 1.0, + "rewards/chosen": -1.2109375, + "rewards/margins": 14.75, + "rewards/rejected": -15.9375, + "step": 2403 + }, + { + "epoch": 2.949693251533742, + "grad_norm": 7.2147901612073175, + "learning_rate": 2.5413223140495865e-08, + "logits/chosen": -0.294921875, + "logits/rejected": -0.47265625, + "logps/chosen": -446.0, + "logps/rejected": -458.0, + "loss": 0.0071, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.5234375, + "rewards/margins": 15.0, + "rewards/rejected": -15.5625, + "step": 2404 + }, + { + "epoch": 2.950920245398773, + "grad_norm": 3.2170783615311365, + "learning_rate": 2.4793388429752066e-08, + "logits/chosen": -0.1708984375, + "logits/rejected": -0.439453125, + "logps/chosen": -444.0, + "logps/rejected": -422.0, + "loss": 0.0023, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.5703125, + "rewards/margins": 14.0, + "rewards/rejected": -14.5625, + "step": 2405 + }, + { + "epoch": 2.9521472392638035, + "grad_norm": 1.873956429822679, + "learning_rate": 2.4173553719008263e-08, + "logits/chosen": -0.1826171875, + "logits/rejected": -0.416015625, + "logps/chosen": -392.0, + "logps/rejected": -426.0, + "loss": 0.0029, + "rewards/accuracies": 1.0, + "rewards/chosen": -1.546875, + "rewards/margins": 14.3125, + "rewards/rejected": -15.875, + "step": 2406 + }, + { + "epoch": 2.9533742331288346, + "grad_norm": 0.2741250362285918, + "learning_rate": 2.3553719008264464e-08, + "logits/chosen": -0.13671875, + "logits/rejected": -0.291015625, + "logps/chosen": -376.0, + "logps/rejected": -432.0, + "loss": 0.0007, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.59375, + "rewards/margins": 14.6875, + "rewards/rejected": -15.3125, + "step": 2407 + }, + { + "epoch": 2.954601226993865, + "grad_norm": 1.3567995130180175, + "learning_rate": 2.2933884297520662e-08, + "logits/chosen": -0.126953125, + "logits/rejected": -0.3359375, + "logps/chosen": -410.0, + "logps/rejected": -462.0, + "loss": 0.0033, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.0546875, + "rewards/margins": 15.9375, + "rewards/rejected": -15.875, + "step": 2408 + }, + { + "epoch": 2.955828220858896, + "grad_norm": 1.0094189001662057, + "learning_rate": 2.231404958677686e-08, + "logits/chosen": -0.447265625, + "logits/rejected": -0.6015625, + "logps/chosen": -406.0, + "logps/rejected": -452.0, + "loss": 0.0016, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.1494140625, + "rewards/margins": 16.0, + "rewards/rejected": -16.125, + "step": 2409 + }, + { + "epoch": 2.957055214723926, + "grad_norm": 0.5210393968183529, + "learning_rate": 2.169421487603306e-08, + "logits/chosen": -0.2138671875, + "logits/rejected": -0.396484375, + "logps/chosen": -404.0, + "logps/rejected": -424.0, + "loss": 0.0015, + "rewards/accuracies": 1.0, + "rewards/chosen": -1.125, + "rewards/margins": 13.75, + "rewards/rejected": -14.875, + "step": 2410 + }, + { + "epoch": 2.9582822085889573, + "grad_norm": 3.9294043193350383, + "learning_rate": 2.1074380165289258e-08, + "logits/chosen": -0.2099609375, + "logits/rejected": -0.44921875, + "logps/chosen": -408.0, + "logps/rejected": -438.0, + "loss": 0.0056, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.1552734375, + "rewards/margins": 15.5, + "rewards/rejected": -15.375, + "step": 2411 + }, + { + "epoch": 2.9595092024539875, + "grad_norm": 3.0104151963551002, + "learning_rate": 2.0454545454545455e-08, + "logits/chosen": -0.29296875, + "logits/rejected": -0.43359375, + "logps/chosen": -428.0, + "logps/rejected": -468.0, + "loss": 0.0045, + "rewards/accuracies": 1.0, + "rewards/chosen": -1.1171875, + "rewards/margins": 14.625, + "rewards/rejected": -15.75, + "step": 2412 + }, + { + "epoch": 2.9607361963190186, + "grad_norm": 1.346094636847527, + "learning_rate": 1.9834710743801653e-08, + "logits/chosen": -0.34765625, + "logits/rejected": -0.546875, + "logps/chosen": -468.0, + "logps/rejected": -476.0, + "loss": 0.0016, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.4375, + "rewards/margins": 16.75, + "rewards/rejected": -17.125, + "step": 2413 + }, + { + "epoch": 2.961963190184049, + "grad_norm": 2.045174380043189, + "learning_rate": 1.921487603305785e-08, + "logits/chosen": -0.134765625, + "logits/rejected": -0.2080078125, + "logps/chosen": -324.0, + "logps/rejected": -374.0, + "loss": 0.0026, + "rewards/accuracies": 1.0, + "rewards/chosen": -1.9140625, + "rewards/margins": 12.875, + "rewards/rejected": -14.75, + "step": 2414 + }, + { + "epoch": 2.96319018404908, + "grad_norm": 0.4727424528273002, + "learning_rate": 1.859504132231405e-08, + "logits/chosen": -0.400390625, + "logits/rejected": -0.6171875, + "logps/chosen": -422.0, + "logps/rejected": -444.0, + "loss": 0.0012, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.921875, + "rewards/margins": 16.0, + "rewards/rejected": -16.875, + "step": 2415 + }, + { + "epoch": 2.96441717791411, + "grad_norm": 0.4680059685080328, + "learning_rate": 1.797520661157025e-08, + "logits/chosen": -0.322265625, + "logits/rejected": -0.40625, + "logps/chosen": -400.0, + "logps/rejected": -482.0, + "loss": 0.0007, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.56640625, + "rewards/margins": 15.9375, + "rewards/rejected": -16.5, + "step": 2416 + }, + { + "epoch": 2.9656441717791413, + "grad_norm": 2.1538069783062066, + "learning_rate": 1.7355371900826446e-08, + "logits/chosen": -0.30859375, + "logits/rejected": -0.5, + "logps/chosen": -406.0, + "logps/rejected": -448.0, + "loss": 0.0041, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.5546875, + "rewards/margins": 15.5, + "rewards/rejected": -16.0, + "step": 2417 + }, + { + "epoch": 2.9668711656441715, + "grad_norm": 1.3407524781049733, + "learning_rate": 1.6735537190082644e-08, + "logits/chosen": -0.294921875, + "logits/rejected": -0.470703125, + "logps/chosen": -422.0, + "logps/rejected": -448.0, + "loss": 0.0022, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.66796875, + "rewards/margins": 14.9375, + "rewards/rejected": -15.5625, + "step": 2418 + }, + { + "epoch": 2.9680981595092026, + "grad_norm": 4.776978584337765, + "learning_rate": 1.6115702479338844e-08, + "logits/chosen": -0.263671875, + "logits/rejected": -0.51953125, + "logps/chosen": -416.0, + "logps/rejected": -440.0, + "loss": 0.0065, + "rewards/accuracies": 1.0, + "rewards/chosen": -1.46875, + "rewards/margins": 14.375, + "rewards/rejected": -15.8125, + "step": 2419 + }, + { + "epoch": 2.969325153374233, + "grad_norm": 2.121414733196059, + "learning_rate": 1.5495867768595042e-08, + "logits/chosen": -0.1357421875, + "logits/rejected": -0.291015625, + "logps/chosen": -392.0, + "logps/rejected": -388.0, + "loss": 0.0039, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.75390625, + "rewards/margins": 14.5625, + "rewards/rejected": -15.3125, + "step": 2420 + }, + { + "epoch": 2.970552147239264, + "grad_norm": 2.280301212201905, + "learning_rate": 1.4876033057851241e-08, + "logits/chosen": -0.255859375, + "logits/rejected": -0.384765625, + "logps/chosen": -374.0, + "logps/rejected": -434.0, + "loss": 0.0028, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.41015625, + "rewards/margins": 14.4375, + "rewards/rejected": -14.8125, + "step": 2421 + }, + { + "epoch": 2.9717791411042946, + "grad_norm": 1.2621800078667251, + "learning_rate": 1.4256198347107437e-08, + "logits/chosen": -0.361328125, + "logits/rejected": -0.51171875, + "logps/chosen": -396.0, + "logps/rejected": -456.0, + "loss": 0.0025, + "rewards/accuracies": 1.0, + "rewards/chosen": -1.34375, + "rewards/margins": 14.75, + "rewards/rejected": -16.125, + "step": 2422 + }, + { + "epoch": 2.9730061349693253, + "grad_norm": 1.0835168373685027, + "learning_rate": 1.3636363636363636e-08, + "logits/chosen": -0.24609375, + "logits/rejected": -0.4453125, + "logps/chosen": -382.0, + "logps/rejected": -404.0, + "loss": 0.002, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.9140625, + "rewards/margins": 14.6875, + "rewards/rejected": -15.5625, + "step": 2423 + }, + { + "epoch": 2.974233128834356, + "grad_norm": 0.6935455594489254, + "learning_rate": 1.3016528925619835e-08, + "logits/chosen": -0.236328125, + "logits/rejected": -0.349609375, + "logps/chosen": -382.0, + "logps/rejected": -428.0, + "loss": 0.0015, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.154296875, + "rewards/margins": 15.9375, + "rewards/rejected": -15.8125, + "step": 2424 + }, + { + "epoch": 2.9754601226993866, + "grad_norm": 0.6625403782713548, + "learning_rate": 1.2396694214876033e-08, + "logits/chosen": -0.2138671875, + "logits/rejected": -0.412109375, + "logps/chosen": -398.0, + "logps/rejected": -472.0, + "loss": 0.0014, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.08837890625, + "rewards/margins": 16.625, + "rewards/rejected": -16.625, + "step": 2425 + }, + { + "epoch": 2.9766871165644173, + "grad_norm": 1.3169148000211959, + "learning_rate": 1.1776859504132232e-08, + "logits/chosen": -0.388671875, + "logits/rejected": -0.59375, + "logps/chosen": -420.0, + "logps/rejected": -430.0, + "loss": 0.0015, + "rewards/accuracies": 1.0, + "rewards/chosen": -1.09375, + "rewards/margins": 16.0, + "rewards/rejected": -17.125, + "step": 2426 + }, + { + "epoch": 2.977914110429448, + "grad_norm": 0.5494631934561397, + "learning_rate": 1.115702479338843e-08, + "logits/chosen": -0.22265625, + "logits/rejected": -0.412109375, + "logps/chosen": -378.0, + "logps/rejected": -412.0, + "loss": 0.0014, + "rewards/accuracies": 1.0, + "rewards/chosen": -1.5234375, + "rewards/margins": 14.5, + "rewards/rejected": -16.0, + "step": 2427 + }, + { + "epoch": 2.9791411042944786, + "grad_norm": 0.9496619925669106, + "learning_rate": 1.0537190082644629e-08, + "logits/chosen": -0.248046875, + "logits/rejected": -0.37890625, + "logps/chosen": -382.0, + "logps/rejected": -414.0, + "loss": 0.0015, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.08349609375, + "rewards/margins": 15.125, + "rewards/rejected": -15.0, + "step": 2428 + }, + { + "epoch": 2.9803680981595093, + "grad_norm": 1.383768588171115, + "learning_rate": 9.917355371900826e-09, + "logits/chosen": -0.244140625, + "logits/rejected": -0.474609375, + "logps/chosen": -434.0, + "logps/rejected": -442.0, + "loss": 0.003, + "rewards/accuracies": 1.0, + "rewards/chosen": -1.0078125, + "rewards/margins": 15.4375, + "rewards/rejected": -16.375, + "step": 2429 + }, + { + "epoch": 2.98159509202454, + "grad_norm": 0.9185079716506662, + "learning_rate": 9.297520661157026e-09, + "logits/chosen": -0.1787109375, + "logits/rejected": -0.369140625, + "logps/chosen": -384.0, + "logps/rejected": -424.0, + "loss": 0.0015, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.87109375, + "rewards/margins": 15.1875, + "rewards/rejected": -16.0, + "step": 2430 + }, + { + "epoch": 2.9828220858895707, + "grad_norm": 0.9191252563608965, + "learning_rate": 8.677685950413223e-09, + "logits/chosen": -0.267578125, + "logits/rejected": -0.44140625, + "logps/chosen": -414.0, + "logps/rejected": -460.0, + "loss": 0.0014, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.55859375, + "rewards/margins": 15.8125, + "rewards/rejected": -16.375, + "step": 2431 + }, + { + "epoch": 2.9840490797546013, + "grad_norm": 0.9066542746349595, + "learning_rate": 8.057851239669422e-09, + "logits/chosen": -0.283203125, + "logits/rejected": -0.439453125, + "logps/chosen": -404.0, + "logps/rejected": -468.0, + "loss": 0.0069, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -2.21875, + "rewards/margins": 14.3125, + "rewards/rejected": -16.5, + "step": 2432 + }, + { + "epoch": 2.985276073619632, + "grad_norm": 0.3501339409428375, + "learning_rate": 7.4380165289256206e-09, + "logits/chosen": -0.2373046875, + "logits/rejected": -0.4921875, + "logps/chosen": -392.0, + "logps/rejected": -426.0, + "loss": 0.006, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.388671875, + "rewards/margins": 15.4375, + "rewards/rejected": -15.8125, + "step": 2433 + }, + { + "epoch": 2.9865030674846627, + "grad_norm": 0.9974134266287437, + "learning_rate": 6.818181818181818e-09, + "logits/chosen": -0.2578125, + "logits/rejected": -0.5703125, + "logps/chosen": -414.0, + "logps/rejected": -422.0, + "loss": 0.0018, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.419921875, + "rewards/margins": 15.25, + "rewards/rejected": -14.875, + "step": 2434 + }, + { + "epoch": 2.9877300613496933, + "grad_norm": 0.36821306186725417, + "learning_rate": 6.1983471074380165e-09, + "logits/chosen": -0.1201171875, + "logits/rejected": -0.2890625, + "logps/chosen": -388.0, + "logps/rejected": -402.0, + "loss": 0.0009, + "rewards/accuracies": 1.0, + "rewards/chosen": -1.0390625, + "rewards/margins": 13.6875, + "rewards/rejected": -14.6875, + "step": 2435 + }, + { + "epoch": 2.988957055214724, + "grad_norm": 1.765461266392412, + "learning_rate": 5.578512396694215e-09, + "logits/chosen": -0.22265625, + "logits/rejected": -0.4375, + "logps/chosen": -486.0, + "logps/rejected": -448.0, + "loss": 0.0025, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.5859375, + "rewards/margins": 15.875, + "rewards/rejected": -15.3125, + "step": 2436 + }, + { + "epoch": 2.9901840490797547, + "grad_norm": 0.49820358242622914, + "learning_rate": 4.958677685950413e-09, + "logits/chosen": -0.298828125, + "logits/rejected": -0.50390625, + "logps/chosen": -438.0, + "logps/rejected": -474.0, + "loss": 0.0012, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.453125, + "rewards/margins": 15.5625, + "rewards/rejected": -16.0, + "step": 2437 + }, + { + "epoch": 2.9914110429447853, + "grad_norm": 1.5672869641609464, + "learning_rate": 4.3388429752066115e-09, + "logits/chosen": -0.201171875, + "logits/rejected": -0.42578125, + "logps/chosen": -402.0, + "logps/rejected": -408.0, + "loss": 0.0018, + "rewards/accuracies": 1.0, + "rewards/chosen": 0.033935546875, + "rewards/margins": 15.1875, + "rewards/rejected": -15.1875, + "step": 2438 + }, + { + "epoch": 2.992638036809816, + "grad_norm": 1.3796295692726182, + "learning_rate": 3.7190082644628103e-09, + "logits/chosen": -0.2451171875, + "logits/rejected": -0.484375, + "logps/chosen": -406.0, + "logps/rejected": -396.0, + "loss": 0.0035, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.314453125, + "rewards/margins": 14.4375, + "rewards/rejected": -14.75, + "step": 2439 + }, + { + "epoch": 2.9938650306748467, + "grad_norm": 0.9122859418995007, + "learning_rate": 3.0991735537190082e-09, + "logits/chosen": -0.443359375, + "logits/rejected": -0.51171875, + "logps/chosen": -416.0, + "logps/rejected": -490.0, + "loss": 0.0071, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -0.345703125, + "rewards/margins": 16.5, + "rewards/rejected": -16.875, + "step": 2440 + }, + { + "epoch": 2.9950920245398773, + "grad_norm": 0.664952517399126, + "learning_rate": 2.4793388429752066e-09, + "logits/chosen": -0.1748046875, + "logits/rejected": -0.30078125, + "logps/chosen": -386.0, + "logps/rejected": -410.0, + "loss": 0.0072, + "rewards/accuracies": 0.9921875, + "rewards/chosen": -2.0625, + "rewards/margins": 12.875, + "rewards/rejected": -14.9375, + "step": 2441 + }, + { + "epoch": 2.996319018404908, + "grad_norm": 0.43047953310665543, + "learning_rate": 1.8595041322314051e-09, + "logits/chosen": -0.1982421875, + "logits/rejected": -0.283203125, + "logps/chosen": -386.0, + "logps/rejected": -442.0, + "loss": 0.001, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.25390625, + "rewards/margins": 15.375, + "rewards/rejected": -15.6875, + "step": 2442 + }, + { + "epoch": 2.9975460122699387, + "grad_norm": 0.8462623631296795, + "learning_rate": 1.2396694214876033e-09, + "logits/chosen": -0.2373046875, + "logits/rejected": -0.296875, + "logps/chosen": -384.0, + "logps/rejected": -410.0, + "loss": 0.0016, + "rewards/accuracies": 1.0, + "rewards/chosen": -1.8203125, + "rewards/margins": 13.9375, + "rewards/rejected": -15.75, + "step": 2443 + }, + { + "epoch": 2.9987730061349693, + "grad_norm": 0.3208084970187418, + "learning_rate": 6.198347107438016e-10, + "logits/chosen": -0.3671875, + "logits/rejected": -0.55078125, + "logps/chosen": -428.0, + "logps/rejected": -454.0, + "loss": 0.0008, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.6171875, + "rewards/margins": 15.3125, + "rewards/rejected": -15.875, + "step": 2444 + }, + { + "epoch": 3.0, + "grad_norm": 0.8655208544730418, + "learning_rate": 0.0, + "logits/chosen": -0.3671875, + "logits/rejected": -0.48828125, + "logps/chosen": -358.0, + "logps/rejected": -452.0, + "loss": 0.0015, + "rewards/accuracies": 1.0, + "rewards/chosen": -0.318359375, + "rewards/margins": 15.1875, + "rewards/rejected": -15.5, + "step": 2445 + } + ], + "logging_steps": 1, + "max_steps": 2445, + "num_input_tokens_seen": 0, + "num_train_epochs": 3, + "save_steps": 500, + "stateful_callbacks": { + "TrainerControl": { + "args": { + "should_epoch_stop": false, + "should_evaluate": false, + "should_log": false, + "should_save": true, + "should_training_stop": true + }, + "attributes": {} + } + }, + "total_flos": 0.0, + "train_batch_size": 4, + "trial_name": null, + "trial_params": null +} diff --git a/checkpoint-2445/training_args.bin b/checkpoint-2445/training_args.bin new file mode 100644 index 0000000..a88cbd3 --- /dev/null +++ b/checkpoint-2445/training_args.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:e0cffaa6f04a26499dd6171106fabe250a7e89bd923590f4fdeb897bc28870a9 +size 9528 diff --git a/checkpoint-2445/zero_to_fp32.py b/checkpoint-2445/zero_to_fp32.py new file mode 100644 index 0000000..24cc342 --- /dev/null +++ b/checkpoint-2445/zero_to_fp32.py @@ -0,0 +1,604 @@ +#!/usr/bin/env python + +# Copyright (c) Microsoft Corporation. +# SPDX-License-Identifier: Apache-2.0 + +# DeepSpeed Team + +# This script extracts fp32 consolidated weights from a zero 1, 2 and 3 DeepSpeed checkpoints. It gets +# copied into the top level checkpoint dir, so the user can easily do the conversion at any point in +# the future. Once extracted, the weights don't require DeepSpeed and can be used in any +# application. +# +# example: python zero_to_fp32.py . pytorch_model.bin + +import argparse +import torch +import glob +import math +import os +import re +from collections import OrderedDict +from dataclasses import dataclass + +# while this script doesn't use deepspeed to recover data, since the checkpoints are pickled with +# DeepSpeed data structures it has to be available in the current python environment. +from deepspeed.utils import logger +from deepspeed.checkpoint.constants import (DS_VERSION, OPTIMIZER_STATE_DICT, SINGLE_PARTITION_OF_FP32_GROUPS, + FP32_FLAT_GROUPS, ZERO_STAGE, PARTITION_COUNT, PARAM_SHAPES, BUFFER_NAMES, + FROZEN_PARAM_SHAPES, FROZEN_PARAM_FRAGMENTS) + + +@dataclass +class zero_model_state: + buffers: dict() + param_shapes: dict() + shared_params: list + ds_version: int + frozen_param_shapes: dict() + frozen_param_fragments: dict() + + +debug = 0 + +# load to cpu +device = torch.device('cpu') + + +def atoi(text): + return int(text) if text.isdigit() else text + + +def natural_keys(text): + ''' + alist.sort(key=natural_keys) sorts in human order + http://nedbatchelder.com/blog/200712/human_sorting.html + (See Toothy's implementation in the comments) + ''' + return [atoi(c) for c in re.split(r'(\d+)', text)] + + +def get_model_state_file(checkpoint_dir, zero_stage): + if not os.path.isdir(checkpoint_dir): + raise FileNotFoundError(f"Directory '{checkpoint_dir}' doesn't exist") + + # there should be only one file + if zero_stage <= 2: + file = os.path.join(checkpoint_dir, "mp_rank_00_model_states.pt") + elif zero_stage == 3: + file = os.path.join(checkpoint_dir, "zero_pp_rank_0_mp_rank_00_model_states.pt") + + if not os.path.exists(file): + raise FileNotFoundError(f"can't find model states file at '{file}'") + + return file + + +def get_checkpoint_files(checkpoint_dir, glob_pattern): + # XXX: need to test that this simple glob rule works for multi-node setup too + ckpt_files = sorted(glob.glob(os.path.join(checkpoint_dir, glob_pattern)), key=natural_keys) + + if len(ckpt_files) == 0: + raise FileNotFoundError(f"can't find {glob_pattern} files in directory '{checkpoint_dir}'") + + return ckpt_files + + +def get_optim_files(checkpoint_dir): + return get_checkpoint_files(checkpoint_dir, "*_optim_states.pt") + + +def get_model_state_files(checkpoint_dir): + return get_checkpoint_files(checkpoint_dir, "*_model_states.pt") + + +def parse_model_states(files): + zero_model_states = [] + for file in files: + state_dict = torch.load(file, map_location=device) + + if BUFFER_NAMES not in state_dict: + raise ValueError(f"{file} is not a model state checkpoint") + buffer_names = state_dict[BUFFER_NAMES] + if debug: + print("Found buffers:", buffer_names) + + # recover just the buffers while restoring them to fp32 if they were saved in fp16 + buffers = {k: v.float() for k, v in state_dict["module"].items() if k in buffer_names} + param_shapes = state_dict[PARAM_SHAPES] + + # collect parameters that are included in param_shapes + param_names = [] + for s in param_shapes: + for name in s.keys(): + param_names.append(name) + + # update with frozen parameters + frozen_param_shapes = state_dict.get(FROZEN_PARAM_SHAPES, None) + if frozen_param_shapes is not None: + if debug: + print(f"Found frozen_param_shapes: {frozen_param_shapes}") + param_names += list(frozen_param_shapes.keys()) + + # handle shared params + shared_params = [[k, v] for k, v in state_dict["shared_params"].items()] + + ds_version = state_dict.get(DS_VERSION, None) + + frozen_param_fragments = state_dict.get(FROZEN_PARAM_FRAGMENTS, None) + + z_model_state = zero_model_state(buffers=buffers, + param_shapes=param_shapes, + shared_params=shared_params, + ds_version=ds_version, + frozen_param_shapes=frozen_param_shapes, + frozen_param_fragments=frozen_param_fragments) + zero_model_states.append(z_model_state) + + return zero_model_states + + +def parse_optim_states(files, ds_checkpoint_dir): + + total_files = len(files) + state_dicts = [] + for f in files: + state_dict = torch.load(f, map_location=device) + # immediately discard the potentially huge 2 optimizer states as we only care for fp32 master weights + # and also handle the case where it was already removed by another helper script + state_dict["optimizer_state_dict"].pop("optimizer_state_dict", None) + state_dicts.append(state_dict) + + if not ZERO_STAGE in state_dicts[0][OPTIMIZER_STATE_DICT]: + raise ValueError(f"{files[0]} is not a zero checkpoint") + zero_stage = state_dicts[0][OPTIMIZER_STATE_DICT][ZERO_STAGE] + world_size = state_dicts[0][OPTIMIZER_STATE_DICT][PARTITION_COUNT] + + # For ZeRO-2 each param group can have different partition_count as data parallelism for expert + # parameters can be different from data parallelism for non-expert parameters. So we can just + # use the max of the partition_count to get the dp world_size. + + if type(world_size) is list: + world_size = max(world_size) + + if world_size != total_files: + raise ValueError( + f"Expected {world_size} of '*_optim_states.pt' under '{ds_checkpoint_dir}' but found {total_files} files. " + "Possibly due to an overwrite of an old checkpoint, or a checkpoint didn't get saved by one or more processes." + ) + + # the groups are named differently in each stage + if zero_stage <= 2: + fp32_groups_key = SINGLE_PARTITION_OF_FP32_GROUPS + elif zero_stage == 3: + fp32_groups_key = FP32_FLAT_GROUPS + else: + raise ValueError(f"unknown zero stage {zero_stage}") + + if zero_stage <= 2: + fp32_flat_groups = [state_dicts[i][OPTIMIZER_STATE_DICT][fp32_groups_key] for i in range(len(state_dicts))] + elif zero_stage == 3: + # if there is more than one param group, there will be multiple flattened tensors - one + # flattened tensor per group - for simplicity merge them into a single tensor + # + # XXX: could make the script more memory efficient for when there are multiple groups - it + # will require matching the sub-lists of param_shapes for each param group flattened tensor + + fp32_flat_groups = [ + torch.cat(state_dicts[i][OPTIMIZER_STATE_DICT][fp32_groups_key], 0) for i in range(len(state_dicts)) + ] + + return zero_stage, world_size, fp32_flat_groups + + +def _get_fp32_state_dict_from_zero_checkpoint(ds_checkpoint_dir, exclude_frozen_parameters): + """ + Returns fp32 state_dict reconstructed from ds checkpoint + + Args: + - ``ds_checkpoint_dir``: path to the deepspeed checkpoint folder (where the optimizer files are) + + """ + print(f"Processing zero checkpoint '{ds_checkpoint_dir}'") + + optim_files = get_optim_files(ds_checkpoint_dir) + zero_stage, world_size, fp32_flat_groups = parse_optim_states(optim_files, ds_checkpoint_dir) + print(f"Detected checkpoint of type zero stage {zero_stage}, world_size: {world_size}") + + model_files = get_model_state_files(ds_checkpoint_dir) + + zero_model_states = parse_model_states(model_files) + print(f'Parsing checkpoint created by deepspeed=={zero_model_states[0].ds_version}') + + if zero_stage <= 2: + return _get_fp32_state_dict_from_zero2_checkpoint(world_size, fp32_flat_groups, zero_model_states, + exclude_frozen_parameters) + elif zero_stage == 3: + return _get_fp32_state_dict_from_zero3_checkpoint(world_size, fp32_flat_groups, zero_model_states, + exclude_frozen_parameters) + + +def _zero2_merge_frozen_params(state_dict, zero_model_states): + if zero_model_states[0].frozen_param_shapes is None or len(zero_model_states[0].frozen_param_shapes) == 0: + return + + frozen_param_shapes = zero_model_states[0].frozen_param_shapes + frozen_param_fragments = zero_model_states[0].frozen_param_fragments + + if debug: + num_elem = sum(s.numel() for s in frozen_param_shapes.values()) + print(f'rank 0: {FROZEN_PARAM_SHAPES}.numel = {num_elem}') + + wanted_params = len(frozen_param_shapes) + wanted_numel = sum(s.numel() for s in frozen_param_shapes.values()) + avail_numel = sum([p.numel() for p in frozen_param_fragments.values()]) + print(f'Frozen params: Have {avail_numel} numels to process.') + print(f'Frozen params: Need {wanted_numel} numels in {wanted_params} params') + + total_params = 0 + total_numel = 0 + for name, shape in frozen_param_shapes.items(): + total_params += 1 + unpartitioned_numel = shape.numel() + total_numel += unpartitioned_numel + + state_dict[name] = frozen_param_fragments[name] + + if debug: + print(f"{name} full shape: {shape} unpartitioned numel {unpartitioned_numel} ") + + print(f"Reconstructed Frozen fp32 state dict with {total_params} params {total_numel} elements") + + +def _has_callable(obj, fn): + attr = getattr(obj, fn, None) + return callable(attr) + + +def _zero2_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states): + param_shapes = zero_model_states[0].param_shapes + + # Reconstruction protocol: + # + # XXX: document this + + if debug: + for i in range(world_size): + for j in range(len(fp32_flat_groups[0])): + print(f"{FP32_FLAT_GROUPS}[{i}][{j}].shape={fp32_flat_groups[i][j].shape}") + + # XXX: memory usage doubles here (zero2) + num_param_groups = len(fp32_flat_groups[0]) + merged_single_partition_of_fp32_groups = [] + for i in range(num_param_groups): + merged_partitions = [sd[i] for sd in fp32_flat_groups] + full_single_fp32_vector = torch.cat(merged_partitions, 0) + merged_single_partition_of_fp32_groups.append(full_single_fp32_vector) + avail_numel = sum( + [full_single_fp32_vector.numel() for full_single_fp32_vector in merged_single_partition_of_fp32_groups]) + + if debug: + wanted_params = sum([len(shapes) for shapes in param_shapes]) + wanted_numel = sum([sum(shape.numel() for shape in shapes.values()) for shapes in param_shapes]) + # not asserting if there is a mismatch due to possible padding + print(f"Have {avail_numel} numels to process.") + print(f"Need {wanted_numel} numels in {wanted_params} params.") + + # params + # XXX: for huge models that can't fit into the host's RAM we will have to recode this to support + # out-of-core computing solution + total_numel = 0 + total_params = 0 + for shapes, full_single_fp32_vector in zip(param_shapes, merged_single_partition_of_fp32_groups): + offset = 0 + avail_numel = full_single_fp32_vector.numel() + for name, shape in shapes.items(): + + unpartitioned_numel = shape.numel() if _has_callable(shape, 'numel') else math.prod(shape) + total_numel += unpartitioned_numel + total_params += 1 + + if debug: + print(f"{name} full shape: {shape} unpartitioned numel {unpartitioned_numel} ") + state_dict[name] = full_single_fp32_vector.narrow(0, offset, unpartitioned_numel).view(shape) + offset += unpartitioned_numel + + # Z2 started to align to 2*world_size to improve nccl performance. Therefore both offset and + # avail_numel can differ by anywhere between 0..2*world_size. Due to two unrelated complex + # paddings performed in the code it's almost impossible to predict the exact numbers w/o the + # live optimizer object, so we are checking that the numbers are within the right range + align_to = 2 * world_size + + def zero2_align(x): + return align_to * math.ceil(x / align_to) + + if debug: + print(f"original offset={offset}, avail_numel={avail_numel}") + + offset = zero2_align(offset) + avail_numel = zero2_align(avail_numel) + + if debug: + print(f"aligned offset={offset}, avail_numel={avail_numel}") + + # Sanity check + if offset != avail_numel: + raise ValueError(f"consumed {offset} numels out of {avail_numel} - something is wrong") + + print(f"Reconstructed fp32 state dict with {total_params} params {total_numel} elements") + + +def _get_fp32_state_dict_from_zero2_checkpoint(world_size, fp32_flat_groups, zero_model_states, + exclude_frozen_parameters): + state_dict = OrderedDict() + + # buffers + buffers = zero_model_states[0].buffers + state_dict.update(buffers) + if debug: + print(f"added {len(buffers)} buffers") + + if not exclude_frozen_parameters: + _zero2_merge_frozen_params(state_dict, zero_model_states) + + _zero2_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states) + + # recover shared parameters + for pair in zero_model_states[0].shared_params: + if pair[1] in state_dict: + state_dict[pair[0]] = state_dict[pair[1]] + + return state_dict + + +def zero3_partitioned_param_info(unpartitioned_numel, world_size): + remainder = unpartitioned_numel % world_size + padding_numel = (world_size - remainder) if remainder else 0 + partitioned_numel = math.ceil(unpartitioned_numel / world_size) + return partitioned_numel, padding_numel + + +def _zero3_merge_frozen_params(state_dict, world_size, zero_model_states): + if zero_model_states[0].frozen_param_shapes is None or len(zero_model_states[0].frozen_param_shapes) == 0: + return + + if debug: + for i in range(world_size): + num_elem = sum(s.numel() for s in zero_model_states[i].frozen_param_fragments.values()) + print(f'rank {i}: {FROZEN_PARAM_SHAPES}.numel = {num_elem}') + + frozen_param_shapes = zero_model_states[0].frozen_param_shapes + wanted_params = len(frozen_param_shapes) + wanted_numel = sum(s.numel() for s in frozen_param_shapes.values()) + avail_numel = sum([p.numel() for p in zero_model_states[0].frozen_param_fragments.values()]) * world_size + print(f'Frozen params: Have {avail_numel} numels to process.') + print(f'Frozen params: Need {wanted_numel} numels in {wanted_params} params') + + total_params = 0 + total_numel = 0 + for name, shape in zero_model_states[0].frozen_param_shapes.items(): + total_params += 1 + unpartitioned_numel = shape.numel() + total_numel += unpartitioned_numel + + param_frags = tuple(model_state.frozen_param_fragments[name] for model_state in zero_model_states) + state_dict[name] = torch.cat(param_frags, 0).narrow(0, 0, unpartitioned_numel).view(shape) + + partitioned_numel, partitioned_padding_numel = zero3_partitioned_param_info(unpartitioned_numel, world_size) + + if debug: + print( + f"Frozen params: {total_params} {name} full shape: {shape} partition0 numel={partitioned_numel} partitioned_padding_numel={partitioned_padding_numel}" + ) + + print(f"Reconstructed Frozen fp32 state dict with {total_params} params {total_numel} elements") + + +def _zero3_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states): + param_shapes = zero_model_states[0].param_shapes + avail_numel = fp32_flat_groups[0].numel() * world_size + # Reconstruction protocol: For zero3 we need to zip the partitions together at boundary of each + # param, re-consolidating each param, while dealing with padding if any + + # merge list of dicts, preserving order + param_shapes = {k: v for d in param_shapes for k, v in d.items()} + + if debug: + for i in range(world_size): + print(f"{FP32_FLAT_GROUPS}[{i}].shape={fp32_flat_groups[i].shape}") + + wanted_params = len(param_shapes) + wanted_numel = sum(shape.numel() for shape in param_shapes.values()) + # not asserting if there is a mismatch due to possible padding + avail_numel = fp32_flat_groups[0].numel() * world_size + print(f"Trainable params: Have {avail_numel} numels to process.") + print(f"Trainable params: Need {wanted_numel} numels in {wanted_params} params.") + + # params + # XXX: for huge models that can't fit into the host's RAM we will have to recode this to support + # out-of-core computing solution + offset = 0 + total_numel = 0 + total_params = 0 + for name, shape in param_shapes.items(): + + unpartitioned_numel = shape.numel() + total_numel += unpartitioned_numel + total_params += 1 + + partitioned_numel, partitioned_padding_numel = zero3_partitioned_param_info(unpartitioned_numel, world_size) + + if debug: + print( + f"Trainable params: {total_params} {name} full shape: {shape} partition0 numel={partitioned_numel} partitioned_padding_numel={partitioned_padding_numel}" + ) + + # XXX: memory usage doubles here + state_dict[name] = torch.cat( + tuple(fp32_flat_groups[i].narrow(0, offset, partitioned_numel) for i in range(world_size)), + 0).narrow(0, 0, unpartitioned_numel).view(shape) + offset += partitioned_numel + + offset *= world_size + + # Sanity check + if offset != avail_numel: + raise ValueError(f"consumed {offset} numels out of {avail_numel} - something is wrong") + + print(f"Reconstructed Trainable fp32 state dict with {total_params} params {total_numel} elements") + + +def _get_fp32_state_dict_from_zero3_checkpoint(world_size, fp32_flat_groups, zero_model_states, + exclude_frozen_parameters): + state_dict = OrderedDict() + + # buffers + buffers = zero_model_states[0].buffers + state_dict.update(buffers) + if debug: + print(f"added {len(buffers)} buffers") + + if not exclude_frozen_parameters: + _zero3_merge_frozen_params(state_dict, world_size, zero_model_states) + + _zero3_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states) + + # recover shared parameters + for pair in zero_model_states[0].shared_params: + if pair[1] in state_dict: + state_dict[pair[0]] = state_dict[pair[1]] + + return state_dict + + +def get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag=None, exclude_frozen_parameters=False): + """ + Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated state_dict that can be loaded with + ``load_state_dict()`` and used for training without DeepSpeed or shared with others, for example + via a model hub. + + Args: + - ``checkpoint_dir``: path to the desired checkpoint folder + - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in 'latest' file. e.g., ``global_step14`` + - ``exclude_frozen_parameters``: exclude frozen parameters + + Returns: + - pytorch ``state_dict`` + + Note: this approach may not work if your application doesn't have sufficient free CPU memory and + you may need to use the offline approach using the ``zero_to_fp32.py`` script that is saved with + the checkpoint. + + A typical usage might be :: + + from deepspeed.utils.zero_to_fp32 import get_fp32_state_dict_from_zero_checkpoint + # do the training and checkpoint saving + state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir) # already on cpu + model = model.cpu() # move to cpu + model.load_state_dict(state_dict) + # submit to model hub or save the model to share with others + + In this example the ``model`` will no longer be usable in the deepspeed context of the same + application. i.e. you will need to re-initialize the deepspeed engine, since + ``model.load_state_dict(state_dict)`` will remove all the deepspeed magic from it. + + If you want it all done for you, use ``load_state_dict_from_zero_checkpoint`` instead. + + """ + if tag is None: + latest_path = os.path.join(checkpoint_dir, 'latest') + if os.path.isfile(latest_path): + with open(latest_path, 'r') as fd: + tag = fd.read().strip() + else: + raise ValueError(f"Unable to find 'latest' file at {latest_path}") + + ds_checkpoint_dir = os.path.join(checkpoint_dir, tag) + + if not os.path.isdir(ds_checkpoint_dir): + raise FileNotFoundError(f"Directory '{ds_checkpoint_dir}' doesn't exist") + + return _get_fp32_state_dict_from_zero_checkpoint(ds_checkpoint_dir, exclude_frozen_parameters) + + +def convert_zero_checkpoint_to_fp32_state_dict(checkpoint_dir, output_file, tag=None, exclude_frozen_parameters=False): + """ + Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated ``state_dict`` file that can be + loaded with ``torch.load(file)`` + ``load_state_dict()`` and used for training without DeepSpeed. + + Args: + - ``checkpoint_dir``: path to the desired checkpoint folder. (one that contains the tag-folder, like ``global_step14``) + - ``output_file``: path to the pytorch fp32 state_dict output file (e.g. path/pytorch_model.bin) + - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in the file named ``latest`` in the checkpoint folder, e.g., ``global_step14`` + - ``exclude_frozen_parameters``: exclude frozen parameters + """ + + state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag, exclude_frozen_parameters) + print(f"Saving fp32 state dict to {output_file}") + torch.save(state_dict, output_file) + + +def load_state_dict_from_zero_checkpoint(model, checkpoint_dir, tag=None): + """ + 1. Put the provided model to cpu + 2. Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated ``state_dict`` + 3. Load it into the provided model + + Args: + - ``model``: the model object to update + - ``checkpoint_dir``: path to the desired checkpoint folder. (one that contains the tag-folder, like ``global_step14``) + - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in the file named ``latest`` in the checkpoint folder, e.g., ``global_step14`` + + Returns: + - ``model`: modified model + + Make sure you have plenty of CPU memory available before you call this function. If you don't + have enough use the ``zero_to_fp32.py`` utility to do the conversion. You will find it + conveniently placed for you in the checkpoint folder. + + A typical usage might be :: + + from deepspeed.utils.zero_to_fp32 import load_state_dict_from_zero_checkpoint + model = load_state_dict_from_zero_checkpoint(trainer.model, checkpoint_dir) + # submit to model hub or save the model to share with others + + Note, that once this was run, the ``model`` will no longer be usable in the deepspeed context + of the same application. i.e. you will need to re-initialize the deepspeed engine, since + ``model.load_state_dict(state_dict)`` will remove all the deepspeed magic from it. + + """ + logger.info(f"Extracting fp32 weights") + state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag) + + logger.info(f"Overwriting model with fp32 weights") + model = model.cpu() + model.load_state_dict(state_dict, strict=False) + + return model + + +if __name__ == "__main__": + + parser = argparse.ArgumentParser() + parser.add_argument("checkpoint_dir", + type=str, + help="path to the desired checkpoint folder, e.g., path/checkpoint-12") + parser.add_argument( + "output_file", + type=str, + help="path to the pytorch fp32 state_dict output file (e.g. path/checkpoint-12/pytorch_model.bin)") + parser.add_argument("-t", + "--tag", + type=str, + default=None, + help="checkpoint tag used as a unique identifier for checkpoint. e.g., global_step1") + parser.add_argument("--exclude_frozen_parameters", action='store_true', help="exclude frozen parameters") + parser.add_argument("-d", "--debug", action='store_true', help="enable debug") + args = parser.parse_args() + + debug = args.debug + + convert_zero_checkpoint_to_fp32_state_dict(args.checkpoint_dir, + args.output_file, + tag=args.tag, + exclude_frozen_parameters=args.exclude_frozen_parameters) diff --git a/checkpoint-815/config.json b/checkpoint-815/config.json new file mode 100644 index 0000000..ded5a8b --- /dev/null +++ b/checkpoint-815/config.json @@ -0,0 +1,35 @@ +{ + "architectures": [ + "LlamaForCausalLM" + ], + "attention_bias": false, + "attention_dropout": 0.0, + "bos_token_id": 128000, + "eos_token_id": 128009, + "head_dim": 128, + "hidden_act": "silu", + "hidden_size": 4096, + "initializer_range": 0.02, + "intermediate_size": 14336, + "max_position_embeddings": 131072, + "mlp_bias": false, + "model_type": "llama", + "num_attention_heads": 32, + "num_hidden_layers": 32, + "num_key_value_heads": 8, + "pretraining_tp": 1, + "rms_norm_eps": 1e-05, + "rope_scaling": { + "factor": 8.0, + "high_freq_factor": 4.0, + "low_freq_factor": 1.0, + "original_max_position_embeddings": 8192, + "rope_type": "llama3" + }, + "rope_theta": 500000.0, + "tie_word_embeddings": false, + "torch_dtype": "bfloat16", + "transformers_version": "4.50.0", + "use_cache": false, + "vocab_size": 128256 +} diff --git a/checkpoint-815/generation_config.json b/checkpoint-815/generation_config.json new file mode 100644 index 0000000..0acdb5a --- /dev/null +++ b/checkpoint-815/generation_config.json @@ -0,0 +1,12 @@ +{ + "bos_token_id": 128000, + "do_sample": true, + "eos_token_id": [ + 128001, + 128008, + 128009 + ], + "temperature": 0.6, + "top_p": 0.9, + "transformers_version": "4.50.0" +} diff --git a/checkpoint-815/global_step815/bf16_zero_pp_rank_0_mp_rank_00_optim_states.pt b/checkpoint-815/global_step815/bf16_zero_pp_rank_0_mp_rank_00_optim_states.pt new file mode 100644 index 0000000..8fcec72 --- /dev/null +++ b/checkpoint-815/global_step815/bf16_zero_pp_rank_0_mp_rank_00_optim_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:eee579a84150fb48fa7bd8b161e29b4d8fe846fe6fe103a92b0abd73f2a54c4f +size 1514015111 diff --git a/checkpoint-815/global_step815/bf16_zero_pp_rank_10_mp_rank_00_optim_states.pt b/checkpoint-815/global_step815/bf16_zero_pp_rank_10_mp_rank_00_optim_states.pt new file mode 100644 index 0000000..6df3796 --- /dev/null +++ b/checkpoint-815/global_step815/bf16_zero_pp_rank_10_mp_rank_00_optim_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:a700a5211043cd1663415cf002eaa2148653e20ed9d213d0774137ff4b9cb554 +size 1514016442 diff --git a/checkpoint-815/global_step815/bf16_zero_pp_rank_11_mp_rank_00_optim_states.pt b/checkpoint-815/global_step815/bf16_zero_pp_rank_11_mp_rank_00_optim_states.pt new file mode 100644 index 0000000..74d8640 --- /dev/null +++ b/checkpoint-815/global_step815/bf16_zero_pp_rank_11_mp_rank_00_optim_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:28922e4735a816b84599e3c04daba6b5dd2ae1972ee4fe3aabb89de5e2a76d66 +size 1514016442 diff --git a/checkpoint-815/global_step815/bf16_zero_pp_rank_12_mp_rank_00_optim_states.pt b/checkpoint-815/global_step815/bf16_zero_pp_rank_12_mp_rank_00_optim_states.pt new file mode 100644 index 0000000..4128a06 --- /dev/null +++ b/checkpoint-815/global_step815/bf16_zero_pp_rank_12_mp_rank_00_optim_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:1dd4568a9479dc47e5bdff792b90303c990d26247bdacc4a1ec0d044775f6d99 +size 1514016442 diff --git a/checkpoint-815/global_step815/bf16_zero_pp_rank_13_mp_rank_00_optim_states.pt b/checkpoint-815/global_step815/bf16_zero_pp_rank_13_mp_rank_00_optim_states.pt new file mode 100644 index 0000000..76f4fe4 --- /dev/null +++ b/checkpoint-815/global_step815/bf16_zero_pp_rank_13_mp_rank_00_optim_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:44a524fa3ae925e9a530d0760707d0a96a03ae7e75cfca81b82291cb2dc650c5 +size 1514016442 diff --git a/checkpoint-815/global_step815/bf16_zero_pp_rank_14_mp_rank_00_optim_states.pt b/checkpoint-815/global_step815/bf16_zero_pp_rank_14_mp_rank_00_optim_states.pt new file mode 100644 index 0000000..36c71da --- /dev/null +++ b/checkpoint-815/global_step815/bf16_zero_pp_rank_14_mp_rank_00_optim_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:87ed25ec6c88da372be088288e529c1ad8eba009f932d9a2ae0ba07d57c60d77 +size 1514016442 diff --git a/checkpoint-815/global_step815/bf16_zero_pp_rank_15_mp_rank_00_optim_states.pt b/checkpoint-815/global_step815/bf16_zero_pp_rank_15_mp_rank_00_optim_states.pt new file mode 100644 index 0000000..c8e0807 --- /dev/null +++ b/checkpoint-815/global_step815/bf16_zero_pp_rank_15_mp_rank_00_optim_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:43d1c11052aecadefa3fd7daca67eac6b37765a3e76d4e652b17a32685a4a94d +size 1514016442 diff --git a/checkpoint-815/global_step815/bf16_zero_pp_rank_16_mp_rank_00_optim_states.pt b/checkpoint-815/global_step815/bf16_zero_pp_rank_16_mp_rank_00_optim_states.pt new file mode 100644 index 0000000..a8409f6 --- /dev/null +++ b/checkpoint-815/global_step815/bf16_zero_pp_rank_16_mp_rank_00_optim_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:8610b14209a2255320d2eb0d3e86fdaaa5cd1c062406a791ddedcc72e3bca0f0 +size 1514016442 diff --git a/checkpoint-815/global_step815/bf16_zero_pp_rank_17_mp_rank_00_optim_states.pt b/checkpoint-815/global_step815/bf16_zero_pp_rank_17_mp_rank_00_optim_states.pt new file mode 100644 index 0000000..6dd5f26 --- /dev/null +++ b/checkpoint-815/global_step815/bf16_zero_pp_rank_17_mp_rank_00_optim_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:e3b9073fee615a9c40083200bf0d94eced4fe46862055d1be719d121dcb85235 +size 1514016442 diff --git a/checkpoint-815/global_step815/bf16_zero_pp_rank_18_mp_rank_00_optim_states.pt b/checkpoint-815/global_step815/bf16_zero_pp_rank_18_mp_rank_00_optim_states.pt new file mode 100644 index 0000000..4abcfee --- /dev/null +++ b/checkpoint-815/global_step815/bf16_zero_pp_rank_18_mp_rank_00_optim_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:b0352e527e6f785986514cb279ae2c5d60b5db6e7ba9d93755f7e34f9e48e92b +size 1514016442 diff --git a/checkpoint-815/global_step815/bf16_zero_pp_rank_19_mp_rank_00_optim_states.pt b/checkpoint-815/global_step815/bf16_zero_pp_rank_19_mp_rank_00_optim_states.pt new file mode 100644 index 0000000..196cebb --- /dev/null +++ b/checkpoint-815/global_step815/bf16_zero_pp_rank_19_mp_rank_00_optim_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:25bb5a4dad546479d7e61d26404b38ea19d82e5ad002ab81d0b74680029b7cb9 +size 1514016442 diff --git a/checkpoint-815/global_step815/bf16_zero_pp_rank_1_mp_rank_00_optim_states.pt b/checkpoint-815/global_step815/bf16_zero_pp_rank_1_mp_rank_00_optim_states.pt new file mode 100644 index 0000000..ed28d06 --- /dev/null +++ b/checkpoint-815/global_step815/bf16_zero_pp_rank_1_mp_rank_00_optim_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:fe1cb25d307d34ded7710dc0d4842ab7700b47098d0d23225a2be06f8b3cf4c3 +size 1514015111 diff --git a/checkpoint-815/global_step815/bf16_zero_pp_rank_20_mp_rank_00_optim_states.pt b/checkpoint-815/global_step815/bf16_zero_pp_rank_20_mp_rank_00_optim_states.pt new file mode 100644 index 0000000..8526488 --- /dev/null +++ b/checkpoint-815/global_step815/bf16_zero_pp_rank_20_mp_rank_00_optim_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:6494e6beea58a5693df5d6d0fed95e83985fec399b323f57247a67d7381d432e +size 1514016442 diff --git a/checkpoint-815/global_step815/bf16_zero_pp_rank_21_mp_rank_00_optim_states.pt b/checkpoint-815/global_step815/bf16_zero_pp_rank_21_mp_rank_00_optim_states.pt new file mode 100644 index 0000000..88ec70d --- /dev/null +++ b/checkpoint-815/global_step815/bf16_zero_pp_rank_21_mp_rank_00_optim_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:635fdf4f2763e44391efcb64bedc871faa5a4effe6b1586af7130ead34fdfc7f +size 1514016442 diff --git a/checkpoint-815/global_step815/bf16_zero_pp_rank_22_mp_rank_00_optim_states.pt b/checkpoint-815/global_step815/bf16_zero_pp_rank_22_mp_rank_00_optim_states.pt new file mode 100644 index 0000000..99c997c --- /dev/null +++ b/checkpoint-815/global_step815/bf16_zero_pp_rank_22_mp_rank_00_optim_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:c4070ac71fe9b6b5b38d4a2044cc49b156074ebe8d524c520e937ad8bf8c50e6 +size 1514016442 diff --git a/checkpoint-815/global_step815/bf16_zero_pp_rank_23_mp_rank_00_optim_states.pt b/checkpoint-815/global_step815/bf16_zero_pp_rank_23_mp_rank_00_optim_states.pt new file mode 100644 index 0000000..0b1e8f9 --- /dev/null +++ b/checkpoint-815/global_step815/bf16_zero_pp_rank_23_mp_rank_00_optim_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:53dfab80959a0f28d1ad90a7fd3fda9b6f3208dcb999eb98a574b704a525b8f7 +size 1514016442 diff --git a/checkpoint-815/global_step815/bf16_zero_pp_rank_24_mp_rank_00_optim_states.pt b/checkpoint-815/global_step815/bf16_zero_pp_rank_24_mp_rank_00_optim_states.pt new file mode 100644 index 0000000..aa594a6 --- /dev/null +++ b/checkpoint-815/global_step815/bf16_zero_pp_rank_24_mp_rank_00_optim_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:f6a3ea9914e69d7c10ce68cd4ad80d04f7c9e8ec0ddb59ebb760db07aea33e1f +size 1514016442 diff --git a/checkpoint-815/global_step815/bf16_zero_pp_rank_25_mp_rank_00_optim_states.pt b/checkpoint-815/global_step815/bf16_zero_pp_rank_25_mp_rank_00_optim_states.pt new file mode 100644 index 0000000..02a9309 --- /dev/null +++ b/checkpoint-815/global_step815/bf16_zero_pp_rank_25_mp_rank_00_optim_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:b8660179db79d800c285f1755ac741c76c94d2d14f5456075f66a2d128515ddc +size 1514016442 diff --git a/checkpoint-815/global_step815/bf16_zero_pp_rank_26_mp_rank_00_optim_states.pt b/checkpoint-815/global_step815/bf16_zero_pp_rank_26_mp_rank_00_optim_states.pt new file mode 100644 index 0000000..6f6f7c9 --- /dev/null +++ b/checkpoint-815/global_step815/bf16_zero_pp_rank_26_mp_rank_00_optim_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:fbef494abdc8854b358e51eb5981b949ddd0b263f1b14c358155c890765d962d +size 1514016442 diff --git a/checkpoint-815/global_step815/bf16_zero_pp_rank_27_mp_rank_00_optim_states.pt b/checkpoint-815/global_step815/bf16_zero_pp_rank_27_mp_rank_00_optim_states.pt new file mode 100644 index 0000000..a25eb1f --- /dev/null +++ b/checkpoint-815/global_step815/bf16_zero_pp_rank_27_mp_rank_00_optim_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:7497a37f9d017780d1df1f4850fabbe73564c5fe98ed48055bc0eabe98d64016 +size 1514016442 diff --git a/checkpoint-815/global_step815/bf16_zero_pp_rank_28_mp_rank_00_optim_states.pt b/checkpoint-815/global_step815/bf16_zero_pp_rank_28_mp_rank_00_optim_states.pt new file mode 100644 index 0000000..eac3df0 --- /dev/null +++ b/checkpoint-815/global_step815/bf16_zero_pp_rank_28_mp_rank_00_optim_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:94388fd663cd2529e6cec125a2f5f2d186d8a7b56e20c59a9437321fc5e1f6bf +size 1514016442 diff --git a/checkpoint-815/global_step815/bf16_zero_pp_rank_29_mp_rank_00_optim_states.pt b/checkpoint-815/global_step815/bf16_zero_pp_rank_29_mp_rank_00_optim_states.pt new file mode 100644 index 0000000..5ef9db9 --- /dev/null +++ b/checkpoint-815/global_step815/bf16_zero_pp_rank_29_mp_rank_00_optim_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:13e16cb68b2ef90a52f3adfa2fb40db57b4a74be1712c76a897d97ecb202267d +size 1514016442 diff --git a/checkpoint-815/global_step815/bf16_zero_pp_rank_2_mp_rank_00_optim_states.pt b/checkpoint-815/global_step815/bf16_zero_pp_rank_2_mp_rank_00_optim_states.pt new file mode 100644 index 0000000..68fcacd --- /dev/null +++ b/checkpoint-815/global_step815/bf16_zero_pp_rank_2_mp_rank_00_optim_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:b6e0f4f20f711aec67bfad31100bf5578882ca3afa80bcfa2b55ea34f0a0c7d5 +size 1514015111 diff --git a/checkpoint-815/global_step815/bf16_zero_pp_rank_30_mp_rank_00_optim_states.pt b/checkpoint-815/global_step815/bf16_zero_pp_rank_30_mp_rank_00_optim_states.pt new file mode 100644 index 0000000..08a2421 --- /dev/null +++ b/checkpoint-815/global_step815/bf16_zero_pp_rank_30_mp_rank_00_optim_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:799f029e90fe5203153a19e3eb6d91e2883d06de1e9999f929a25a6bb380aa27 +size 1514016442 diff --git a/checkpoint-815/global_step815/bf16_zero_pp_rank_31_mp_rank_00_optim_states.pt b/checkpoint-815/global_step815/bf16_zero_pp_rank_31_mp_rank_00_optim_states.pt new file mode 100644 index 0000000..79362e8 --- /dev/null +++ b/checkpoint-815/global_step815/bf16_zero_pp_rank_31_mp_rank_00_optim_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:152f12c520a01a827f4452cf984a7555ce7b3b5af9fa14147146572105bc37ca +size 1514016442 diff --git a/checkpoint-815/global_step815/bf16_zero_pp_rank_3_mp_rank_00_optim_states.pt b/checkpoint-815/global_step815/bf16_zero_pp_rank_3_mp_rank_00_optim_states.pt new file mode 100644 index 0000000..88c1e9b --- /dev/null +++ b/checkpoint-815/global_step815/bf16_zero_pp_rank_3_mp_rank_00_optim_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:610bd93dec3f4395daf6036ac58aba6ba825cbc2d751c0ace4e1e7d570e39a44 +size 1514015111 diff --git a/checkpoint-815/global_step815/bf16_zero_pp_rank_4_mp_rank_00_optim_states.pt b/checkpoint-815/global_step815/bf16_zero_pp_rank_4_mp_rank_00_optim_states.pt new file mode 100644 index 0000000..824a31d --- /dev/null +++ b/checkpoint-815/global_step815/bf16_zero_pp_rank_4_mp_rank_00_optim_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:864cf6466f76d8509701e7a8709b73862d1f2cf1bb31f302c8c4fe7335068b60 +size 1514015111 diff --git a/checkpoint-815/global_step815/bf16_zero_pp_rank_5_mp_rank_00_optim_states.pt b/checkpoint-815/global_step815/bf16_zero_pp_rank_5_mp_rank_00_optim_states.pt new file mode 100644 index 0000000..6b25a21 --- /dev/null +++ b/checkpoint-815/global_step815/bf16_zero_pp_rank_5_mp_rank_00_optim_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:593fee477855272451bddc863c5f98015a5628a0b06ad5756b37320a96d92eeb +size 1514015111 diff --git a/checkpoint-815/global_step815/bf16_zero_pp_rank_6_mp_rank_00_optim_states.pt b/checkpoint-815/global_step815/bf16_zero_pp_rank_6_mp_rank_00_optim_states.pt new file mode 100644 index 0000000..ef3a415 --- /dev/null +++ b/checkpoint-815/global_step815/bf16_zero_pp_rank_6_mp_rank_00_optim_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:2fde8ae861b05628e847ce25ba5400604464240ef98932d8a5570f3cdf893c1d +size 1514015111 diff --git a/checkpoint-815/global_step815/bf16_zero_pp_rank_7_mp_rank_00_optim_states.pt b/checkpoint-815/global_step815/bf16_zero_pp_rank_7_mp_rank_00_optim_states.pt new file mode 100644 index 0000000..e333e16 --- /dev/null +++ b/checkpoint-815/global_step815/bf16_zero_pp_rank_7_mp_rank_00_optim_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:5e74470dc8c6eb416ccb6a8b2b8f7fef2b49631d716d14a1f15fef8c53f9bda8 +size 1514015111 diff --git a/checkpoint-815/global_step815/bf16_zero_pp_rank_8_mp_rank_00_optim_states.pt b/checkpoint-815/global_step815/bf16_zero_pp_rank_8_mp_rank_00_optim_states.pt new file mode 100644 index 0000000..7cd386c --- /dev/null +++ b/checkpoint-815/global_step815/bf16_zero_pp_rank_8_mp_rank_00_optim_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:8a72db4bd9fb61cfe704a944968731427981046c416fb8c4e1377d71c98e194f +size 1514015111 diff --git a/checkpoint-815/global_step815/bf16_zero_pp_rank_9_mp_rank_00_optim_states.pt b/checkpoint-815/global_step815/bf16_zero_pp_rank_9_mp_rank_00_optim_states.pt new file mode 100644 index 0000000..0cc46ec --- /dev/null +++ b/checkpoint-815/global_step815/bf16_zero_pp_rank_9_mp_rank_00_optim_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:ed83c739697afebdd959ca7db8835142f0ae34e07fc7095ac2ef371768b609b9 +size 1514015111 diff --git a/checkpoint-815/global_step815/zero_pp_rank_0_mp_rank_00_model_states.pt b/checkpoint-815/global_step815/zero_pp_rank_0_mp_rank_00_model_states.pt new file mode 100644 index 0000000..8948f42 --- /dev/null +++ b/checkpoint-815/global_step815/zero_pp_rank_0_mp_rank_00_model_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:40ba6c0d314a71a8a038eeffa9046553c73f5e76e4b9d8e856629862855dfcfb +size 150245 diff --git a/checkpoint-815/global_step815/zero_pp_rank_10_mp_rank_00_model_states.pt b/checkpoint-815/global_step815/zero_pp_rank_10_mp_rank_00_model_states.pt new file mode 100644 index 0000000..2212482 --- /dev/null +++ b/checkpoint-815/global_step815/zero_pp_rank_10_mp_rank_00_model_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:024eef49228ce027257c81680e51700d64008ec55789eb59dcde70a36817558a +size 150540 diff --git a/checkpoint-815/global_step815/zero_pp_rank_11_mp_rank_00_model_states.pt b/checkpoint-815/global_step815/zero_pp_rank_11_mp_rank_00_model_states.pt new file mode 100644 index 0000000..40bd1bd --- /dev/null +++ b/checkpoint-815/global_step815/zero_pp_rank_11_mp_rank_00_model_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:5c00f3b3e366b7cfacd447b4434da21ef301448603fb55d4335db39299b66575 +size 150540 diff --git a/checkpoint-815/global_step815/zero_pp_rank_12_mp_rank_00_model_states.pt b/checkpoint-815/global_step815/zero_pp_rank_12_mp_rank_00_model_states.pt new file mode 100644 index 0000000..9915c4d --- /dev/null +++ b/checkpoint-815/global_step815/zero_pp_rank_12_mp_rank_00_model_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:6cf1da6bf30a5b8f93f325e72d167effb11dfdf29766364fef6b9d4506dfb81e +size 150540 diff --git a/checkpoint-815/global_step815/zero_pp_rank_13_mp_rank_00_model_states.pt b/checkpoint-815/global_step815/zero_pp_rank_13_mp_rank_00_model_states.pt new file mode 100644 index 0000000..f03df1f --- /dev/null +++ b/checkpoint-815/global_step815/zero_pp_rank_13_mp_rank_00_model_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:9f94229da5713f438ab5c4189e38eadbc168fc643fd01a5b6eb930bd2f722efd +size 150540 diff --git a/checkpoint-815/global_step815/zero_pp_rank_14_mp_rank_00_model_states.pt b/checkpoint-815/global_step815/zero_pp_rank_14_mp_rank_00_model_states.pt new file mode 100644 index 0000000..058ff88 --- /dev/null +++ b/checkpoint-815/global_step815/zero_pp_rank_14_mp_rank_00_model_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:b5b97915295b3f6c7eca69f6541ba5833b8919fc0cb698d51f36e8f4b9fdf1d5 +size 150540 diff --git a/checkpoint-815/global_step815/zero_pp_rank_15_mp_rank_00_model_states.pt b/checkpoint-815/global_step815/zero_pp_rank_15_mp_rank_00_model_states.pt new file mode 100644 index 0000000..b950999 --- /dev/null +++ b/checkpoint-815/global_step815/zero_pp_rank_15_mp_rank_00_model_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:b99f4e99e6ffd910f61d39cd0667ad6f7cf15a6179452f54cd0ab9eb3be0f7c9 +size 150540 diff --git a/checkpoint-815/global_step815/zero_pp_rank_16_mp_rank_00_model_states.pt b/checkpoint-815/global_step815/zero_pp_rank_16_mp_rank_00_model_states.pt new file mode 100644 index 0000000..8974561 --- /dev/null +++ b/checkpoint-815/global_step815/zero_pp_rank_16_mp_rank_00_model_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:61fef944ea931de77863bf7ca7b70cda46266397ec948e9282e6930a74bf2ed0 +size 150540 diff --git a/checkpoint-815/global_step815/zero_pp_rank_17_mp_rank_00_model_states.pt b/checkpoint-815/global_step815/zero_pp_rank_17_mp_rank_00_model_states.pt new file mode 100644 index 0000000..05f05a1 --- /dev/null +++ b/checkpoint-815/global_step815/zero_pp_rank_17_mp_rank_00_model_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:8c65e902223780bcaa882920de3575f35c5deb86a74decf182ff60de992138c6 +size 150540 diff --git a/checkpoint-815/global_step815/zero_pp_rank_18_mp_rank_00_model_states.pt b/checkpoint-815/global_step815/zero_pp_rank_18_mp_rank_00_model_states.pt new file mode 100644 index 0000000..238e639 --- /dev/null +++ b/checkpoint-815/global_step815/zero_pp_rank_18_mp_rank_00_model_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:a133f1cb598eed313d7415e6de2861d456036a9823785dafc9b52a26fa533f4c +size 150540 diff --git a/checkpoint-815/global_step815/zero_pp_rank_19_mp_rank_00_model_states.pt b/checkpoint-815/global_step815/zero_pp_rank_19_mp_rank_00_model_states.pt new file mode 100644 index 0000000..2c4dc0d --- /dev/null +++ b/checkpoint-815/global_step815/zero_pp_rank_19_mp_rank_00_model_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:55591d7a7ad8cacac5cf43918a7ad42f693db1b850f57a260cf2778aee083cef +size 150540 diff --git a/checkpoint-815/global_step815/zero_pp_rank_1_mp_rank_00_model_states.pt b/checkpoint-815/global_step815/zero_pp_rank_1_mp_rank_00_model_states.pt new file mode 100644 index 0000000..17bb96a --- /dev/null +++ b/checkpoint-815/global_step815/zero_pp_rank_1_mp_rank_00_model_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:a5947cccfabfeb8fc303b07ca84c4d4146bb568b41e8dee350a89b8dbb72981c +size 150245 diff --git a/checkpoint-815/global_step815/zero_pp_rank_20_mp_rank_00_model_states.pt b/checkpoint-815/global_step815/zero_pp_rank_20_mp_rank_00_model_states.pt new file mode 100644 index 0000000..1604c21 --- /dev/null +++ b/checkpoint-815/global_step815/zero_pp_rank_20_mp_rank_00_model_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:e7c17e090edef85f94359ec12ac733fd9394d7a36675b4cf63dda67fe6da8f24 +size 150540 diff --git a/checkpoint-815/global_step815/zero_pp_rank_21_mp_rank_00_model_states.pt b/checkpoint-815/global_step815/zero_pp_rank_21_mp_rank_00_model_states.pt new file mode 100644 index 0000000..98d5fce --- /dev/null +++ b/checkpoint-815/global_step815/zero_pp_rank_21_mp_rank_00_model_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:9d35bbd41a45c012aa1db3445f9b72b3dd567ebc41b00dce74bd1e4d82e2ebfd +size 150540 diff --git a/checkpoint-815/global_step815/zero_pp_rank_22_mp_rank_00_model_states.pt b/checkpoint-815/global_step815/zero_pp_rank_22_mp_rank_00_model_states.pt new file mode 100644 index 0000000..91cd1fe --- /dev/null +++ b/checkpoint-815/global_step815/zero_pp_rank_22_mp_rank_00_model_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:e38f525185b44ebf31ebd0ac8f0d8a32e0fb510d3f674311c087e3508b40c993 +size 150540 diff --git a/checkpoint-815/global_step815/zero_pp_rank_23_mp_rank_00_model_states.pt b/checkpoint-815/global_step815/zero_pp_rank_23_mp_rank_00_model_states.pt new file mode 100644 index 0000000..168e223 --- /dev/null +++ b/checkpoint-815/global_step815/zero_pp_rank_23_mp_rank_00_model_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:6f7cce2fd60981a15a0f3f33cc90ce2537fcaab35671f44f3c49fcea6fca1891 +size 150540 diff --git a/checkpoint-815/global_step815/zero_pp_rank_24_mp_rank_00_model_states.pt b/checkpoint-815/global_step815/zero_pp_rank_24_mp_rank_00_model_states.pt new file mode 100644 index 0000000..ab22b84 --- /dev/null +++ b/checkpoint-815/global_step815/zero_pp_rank_24_mp_rank_00_model_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:4f8d1850a9fd50d9ad858ec306dadb3409f83a743868167753d2ceb1ee0b8782 +size 150540 diff --git a/checkpoint-815/global_step815/zero_pp_rank_25_mp_rank_00_model_states.pt b/checkpoint-815/global_step815/zero_pp_rank_25_mp_rank_00_model_states.pt new file mode 100644 index 0000000..3d155ea --- /dev/null +++ b/checkpoint-815/global_step815/zero_pp_rank_25_mp_rank_00_model_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:55bc11c638b6fadef2d46875c4a0e4a788f2b78ae95eacdff3a036f1cbfc8b9e +size 150540 diff --git a/checkpoint-815/global_step815/zero_pp_rank_26_mp_rank_00_model_states.pt b/checkpoint-815/global_step815/zero_pp_rank_26_mp_rank_00_model_states.pt new file mode 100644 index 0000000..32b8482 --- /dev/null +++ b/checkpoint-815/global_step815/zero_pp_rank_26_mp_rank_00_model_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:ff219377c7fb54bb953c6c570cf957b0e25d7bb9d2326d87143bc2db6e131b02 +size 150540 diff --git a/checkpoint-815/global_step815/zero_pp_rank_27_mp_rank_00_model_states.pt b/checkpoint-815/global_step815/zero_pp_rank_27_mp_rank_00_model_states.pt new file mode 100644 index 0000000..ad8f112 --- /dev/null +++ b/checkpoint-815/global_step815/zero_pp_rank_27_mp_rank_00_model_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:18ed23baa1f8cd3b5be593df62b46bd925963eae398fb856c46af8bca0bdc706 +size 150540 diff --git a/checkpoint-815/global_step815/zero_pp_rank_28_mp_rank_00_model_states.pt b/checkpoint-815/global_step815/zero_pp_rank_28_mp_rank_00_model_states.pt new file mode 100644 index 0000000..572add1 --- /dev/null +++ b/checkpoint-815/global_step815/zero_pp_rank_28_mp_rank_00_model_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:7e4fe2622f3c42b1d9b6319829c12a40ba41eed4b5165643ba1e541bbd9d6773 +size 150540 diff --git a/checkpoint-815/global_step815/zero_pp_rank_29_mp_rank_00_model_states.pt b/checkpoint-815/global_step815/zero_pp_rank_29_mp_rank_00_model_states.pt new file mode 100644 index 0000000..8f66a39 --- /dev/null +++ b/checkpoint-815/global_step815/zero_pp_rank_29_mp_rank_00_model_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:3aaab1637c614043c25c7321822568b135d719fb8ee06fdcd666152646a036f8 +size 150540 diff --git a/checkpoint-815/global_step815/zero_pp_rank_2_mp_rank_00_model_states.pt b/checkpoint-815/global_step815/zero_pp_rank_2_mp_rank_00_model_states.pt new file mode 100644 index 0000000..bce022e --- /dev/null +++ b/checkpoint-815/global_step815/zero_pp_rank_2_mp_rank_00_model_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:955cdff9e07f053cbc74b4a713e4d8def17fa2e19b7e9802bad227f51d4b61af +size 150245 diff --git a/checkpoint-815/global_step815/zero_pp_rank_30_mp_rank_00_model_states.pt b/checkpoint-815/global_step815/zero_pp_rank_30_mp_rank_00_model_states.pt new file mode 100644 index 0000000..d78e9d3 --- /dev/null +++ b/checkpoint-815/global_step815/zero_pp_rank_30_mp_rank_00_model_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:cc85012ee362480a408246d70a9f1fce990efdf194b19f0a18f7f3aac1eb1ab3 +size 150540 diff --git a/checkpoint-815/global_step815/zero_pp_rank_31_mp_rank_00_model_states.pt b/checkpoint-815/global_step815/zero_pp_rank_31_mp_rank_00_model_states.pt new file mode 100644 index 0000000..025216a --- /dev/null +++ b/checkpoint-815/global_step815/zero_pp_rank_31_mp_rank_00_model_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:2fbdfe073710344bded771325bbd465f5b8a32a32bdf561a674751b41cf6e111 +size 150540 diff --git a/checkpoint-815/global_step815/zero_pp_rank_3_mp_rank_00_model_states.pt b/checkpoint-815/global_step815/zero_pp_rank_3_mp_rank_00_model_states.pt new file mode 100644 index 0000000..45a8ff1 --- /dev/null +++ b/checkpoint-815/global_step815/zero_pp_rank_3_mp_rank_00_model_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:1bbe876d4201139b9ab1ecc5cc62f74e0eb73ef120e1cff4c5f84dca3059dad7 +size 150245 diff --git a/checkpoint-815/global_step815/zero_pp_rank_4_mp_rank_00_model_states.pt b/checkpoint-815/global_step815/zero_pp_rank_4_mp_rank_00_model_states.pt new file mode 100644 index 0000000..51ec98a --- /dev/null +++ b/checkpoint-815/global_step815/zero_pp_rank_4_mp_rank_00_model_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:bfae4502edaad044ae8c1a527900b97fd20fea07eaa512af7273532d23b47618 +size 150245 diff --git a/checkpoint-815/global_step815/zero_pp_rank_5_mp_rank_00_model_states.pt b/checkpoint-815/global_step815/zero_pp_rank_5_mp_rank_00_model_states.pt new file mode 100644 index 0000000..32e7ce4 --- /dev/null +++ b/checkpoint-815/global_step815/zero_pp_rank_5_mp_rank_00_model_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:73170415f2c2bef296c8730bf78811e31654ac55c87c2d71d77bd88d30b5a270 +size 150245 diff --git a/checkpoint-815/global_step815/zero_pp_rank_6_mp_rank_00_model_states.pt b/checkpoint-815/global_step815/zero_pp_rank_6_mp_rank_00_model_states.pt new file mode 100644 index 0000000..82d10da --- /dev/null +++ b/checkpoint-815/global_step815/zero_pp_rank_6_mp_rank_00_model_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:859c2393bcde923cc30a036b27c836d6eac277f575f022b4cbb0b541237091dd +size 150245 diff --git a/checkpoint-815/global_step815/zero_pp_rank_7_mp_rank_00_model_states.pt b/checkpoint-815/global_step815/zero_pp_rank_7_mp_rank_00_model_states.pt new file mode 100644 index 0000000..00c0875 --- /dev/null +++ b/checkpoint-815/global_step815/zero_pp_rank_7_mp_rank_00_model_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:4494a1f9263497b860dafc470670372cabcf493bd54adb561d84e9c57be3dac7 +size 150245 diff --git a/checkpoint-815/global_step815/zero_pp_rank_8_mp_rank_00_model_states.pt b/checkpoint-815/global_step815/zero_pp_rank_8_mp_rank_00_model_states.pt new file mode 100644 index 0000000..aee6ef3 --- /dev/null +++ b/checkpoint-815/global_step815/zero_pp_rank_8_mp_rank_00_model_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:edb26fdb729d3f1fea2b31878462cdc9057312c888b9c8ef8bae28e1002dfdf4 +size 150245 diff --git a/checkpoint-815/global_step815/zero_pp_rank_9_mp_rank_00_model_states.pt b/checkpoint-815/global_step815/zero_pp_rank_9_mp_rank_00_model_states.pt new file mode 100644 index 0000000..077564d --- /dev/null +++ b/checkpoint-815/global_step815/zero_pp_rank_9_mp_rank_00_model_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:22843c3e9c8f1d5116a8f88a79ac3c5368fdf132311c803b3b1b2d75ea9cb1b6 +size 150245 diff --git a/checkpoint-815/latest b/checkpoint-815/latest new file mode 100644 index 0000000..c057bc6 --- /dev/null +++ b/checkpoint-815/latest @@ -0,0 +1 @@ +global_step815 \ No newline at end of file diff --git a/checkpoint-815/model-00001-of-00004.safetensors b/checkpoint-815/model-00001-of-00004.safetensors new file mode 100644 index 0000000..f4e7b55 --- /dev/null +++ b/checkpoint-815/model-00001-of-00004.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:927fa798cf67aff394e20a462d2386ab41a2f410ea81e9d694960a127b5da9a2 +size 4976698672 diff --git a/checkpoint-815/model-00002-of-00004.safetensors b/checkpoint-815/model-00002-of-00004.safetensors new file mode 100644 index 0000000..9e9c13f --- /dev/null +++ b/checkpoint-815/model-00002-of-00004.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:69feb78c8a4549a7c776ec232e447bc6b862bb7931ca7bfde20524a639b1dafd +size 4999802720 diff --git a/checkpoint-815/model-00003-of-00004.safetensors b/checkpoint-815/model-00003-of-00004.safetensors new file mode 100644 index 0000000..1022d55 --- /dev/null +++ b/checkpoint-815/model-00003-of-00004.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:383bb20be792b2a52e36f9060a12fdfbf3645f18095b0a6a06d1e5884c754fc0 +size 4915916176 diff --git a/checkpoint-815/model-00004-of-00004.safetensors b/checkpoint-815/model-00004-of-00004.safetensors new file mode 100644 index 0000000..3a729b5 --- /dev/null +++ b/checkpoint-815/model-00004-of-00004.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:89d5a870e45cf8bbad056d29a521d1918dee6f1d4fd53f3b4fb503790aec26b0 +size 1168138808 diff --git a/checkpoint-815/model.safetensors.index.json b/checkpoint-815/model.safetensors.index.json new file mode 100644 index 0000000..0fd8120 --- /dev/null +++ b/checkpoint-815/model.safetensors.index.json @@ -0,0 +1,298 @@ +{ + "metadata": { + "total_size": 16060522496 + }, + "weight_map": { + "lm_head.weight": "model-00004-of-00004.safetensors", + "model.embed_tokens.weight": "model-00001-of-00004.safetensors", + "model.layers.0.input_layernorm.weight": "model-00001-of-00004.safetensors", + "model.layers.0.mlp.down_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.0.mlp.gate_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.0.mlp.up_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.0.post_attention_layernorm.weight": "model-00001-of-00004.safetensors", + "model.layers.0.self_attn.k_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.0.self_attn.o_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.0.self_attn.q_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.0.self_attn.v_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.1.input_layernorm.weight": "model-00001-of-00004.safetensors", + "model.layers.1.mlp.down_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.1.mlp.gate_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.1.mlp.up_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.1.post_attention_layernorm.weight": "model-00001-of-00004.safetensors", + "model.layers.1.self_attn.k_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.1.self_attn.o_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.1.self_attn.q_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.1.self_attn.v_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.10.input_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.10.mlp.down_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.10.mlp.gate_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.10.mlp.up_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.10.post_attention_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.10.self_attn.k_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.10.self_attn.o_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.10.self_attn.q_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.10.self_attn.v_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.11.input_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.11.mlp.down_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.11.mlp.gate_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.11.mlp.up_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.11.post_attention_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.11.self_attn.k_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.11.self_attn.o_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.11.self_attn.q_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.11.self_attn.v_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.12.input_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.12.mlp.down_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.12.mlp.gate_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.12.mlp.up_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.12.post_attention_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.12.self_attn.k_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.12.self_attn.o_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.12.self_attn.q_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.12.self_attn.v_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.13.input_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.13.mlp.down_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.13.mlp.gate_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.13.mlp.up_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.13.post_attention_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.13.self_attn.k_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.13.self_attn.o_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.13.self_attn.q_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.13.self_attn.v_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.14.input_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.14.mlp.down_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.14.mlp.gate_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.14.mlp.up_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.14.post_attention_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.14.self_attn.k_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.14.self_attn.o_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.14.self_attn.q_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.14.self_attn.v_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.15.input_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.15.mlp.down_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.15.mlp.gate_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.15.mlp.up_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.15.post_attention_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.15.self_attn.k_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.15.self_attn.o_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.15.self_attn.q_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.15.self_attn.v_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.16.input_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.16.mlp.down_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.16.mlp.gate_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.16.mlp.up_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.16.post_attention_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.16.self_attn.k_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.16.self_attn.o_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.16.self_attn.q_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.16.self_attn.v_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.17.input_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.17.mlp.down_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.17.mlp.gate_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.17.mlp.up_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.17.post_attention_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.17.self_attn.k_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.17.self_attn.o_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.17.self_attn.q_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.17.self_attn.v_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.18.input_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.18.mlp.down_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.18.mlp.gate_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.18.mlp.up_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.18.post_attention_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.18.self_attn.k_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.18.self_attn.o_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.18.self_attn.q_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.18.self_attn.v_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.19.input_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.19.mlp.down_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.19.mlp.gate_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.19.mlp.up_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.19.post_attention_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.19.self_attn.k_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.19.self_attn.o_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.19.self_attn.q_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.19.self_attn.v_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.2.input_layernorm.weight": "model-00001-of-00004.safetensors", + "model.layers.2.mlp.down_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.2.mlp.gate_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.2.mlp.up_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.2.post_attention_layernorm.weight": "model-00001-of-00004.safetensors", + "model.layers.2.self_attn.k_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.2.self_attn.o_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.2.self_attn.q_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.2.self_attn.v_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.20.input_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.20.mlp.down_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.20.mlp.gate_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.20.mlp.up_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.20.post_attention_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.20.self_attn.k_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.20.self_attn.o_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.20.self_attn.q_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.20.self_attn.v_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.21.input_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.21.mlp.down_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.21.mlp.gate_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.21.mlp.up_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.21.post_attention_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.21.self_attn.k_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.21.self_attn.o_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.21.self_attn.q_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.21.self_attn.v_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.22.input_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.22.mlp.down_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.22.mlp.gate_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.22.mlp.up_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.22.post_attention_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.22.self_attn.k_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.22.self_attn.o_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.22.self_attn.q_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.22.self_attn.v_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.23.input_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.23.mlp.down_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.23.mlp.gate_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.23.mlp.up_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.23.post_attention_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.23.self_attn.k_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.23.self_attn.o_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.23.self_attn.q_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.23.self_attn.v_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.24.input_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.24.mlp.down_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.24.mlp.gate_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.24.mlp.up_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.24.post_attention_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.24.self_attn.k_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.24.self_attn.o_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.24.self_attn.q_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.24.self_attn.v_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.25.input_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.25.mlp.down_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.25.mlp.gate_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.25.mlp.up_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.25.post_attention_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.25.self_attn.k_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.25.self_attn.o_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.25.self_attn.q_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.25.self_attn.v_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.26.input_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.26.mlp.down_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.26.mlp.gate_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.26.mlp.up_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.26.post_attention_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.26.self_attn.k_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.26.self_attn.o_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.26.self_attn.q_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.26.self_attn.v_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.27.input_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.27.mlp.down_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.27.mlp.gate_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.27.mlp.up_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.27.post_attention_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.27.self_attn.k_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.27.self_attn.o_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.27.self_attn.q_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.27.self_attn.v_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.28.input_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.28.mlp.down_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.28.mlp.gate_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.28.mlp.up_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.28.post_attention_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.28.self_attn.k_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.28.self_attn.o_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.28.self_attn.q_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.28.self_attn.v_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.29.input_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.29.mlp.down_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.29.mlp.gate_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.29.mlp.up_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.29.post_attention_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.29.self_attn.k_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.29.self_attn.o_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.29.self_attn.q_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.29.self_attn.v_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.3.input_layernorm.weight": "model-00001-of-00004.safetensors", + "model.layers.3.mlp.down_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.3.mlp.gate_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.3.mlp.up_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.3.post_attention_layernorm.weight": "model-00001-of-00004.safetensors", + "model.layers.3.self_attn.k_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.3.self_attn.o_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.3.self_attn.q_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.3.self_attn.v_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.30.input_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.30.mlp.down_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.30.mlp.gate_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.30.mlp.up_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.30.post_attention_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.30.self_attn.k_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.30.self_attn.o_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.30.self_attn.q_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.30.self_attn.v_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.31.input_layernorm.weight": "model-00004-of-00004.safetensors", + "model.layers.31.mlp.down_proj.weight": "model-00004-of-00004.safetensors", + "model.layers.31.mlp.gate_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.31.mlp.up_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.31.post_attention_layernorm.weight": "model-00004-of-00004.safetensors", + "model.layers.31.self_attn.k_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.31.self_attn.o_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.31.self_attn.q_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.31.self_attn.v_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.4.input_layernorm.weight": "model-00001-of-00004.safetensors", + "model.layers.4.mlp.down_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.4.mlp.gate_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.4.mlp.up_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.4.post_attention_layernorm.weight": "model-00001-of-00004.safetensors", + "model.layers.4.self_attn.k_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.4.self_attn.o_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.4.self_attn.q_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.4.self_attn.v_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.5.input_layernorm.weight": "model-00001-of-00004.safetensors", + "model.layers.5.mlp.down_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.5.mlp.gate_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.5.mlp.up_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.5.post_attention_layernorm.weight": "model-00001-of-00004.safetensors", + "model.layers.5.self_attn.k_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.5.self_attn.o_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.5.self_attn.q_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.5.self_attn.v_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.6.input_layernorm.weight": "model-00001-of-00004.safetensors", + "model.layers.6.mlp.down_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.6.mlp.gate_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.6.mlp.up_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.6.post_attention_layernorm.weight": "model-00001-of-00004.safetensors", + "model.layers.6.self_attn.k_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.6.self_attn.o_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.6.self_attn.q_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.6.self_attn.v_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.7.input_layernorm.weight": "model-00001-of-00004.safetensors", + "model.layers.7.mlp.down_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.7.mlp.gate_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.7.mlp.up_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.7.post_attention_layernorm.weight": "model-00001-of-00004.safetensors", + "model.layers.7.self_attn.k_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.7.self_attn.o_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.7.self_attn.q_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.7.self_attn.v_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.8.input_layernorm.weight": "model-00001-of-00004.safetensors", + "model.layers.8.mlp.down_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.8.mlp.gate_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.8.mlp.up_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.8.post_attention_layernorm.weight": "model-00001-of-00004.safetensors", + "model.layers.8.self_attn.k_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.8.self_attn.o_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.8.self_attn.q_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.8.self_attn.v_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.9.input_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.9.mlp.down_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.9.mlp.gate_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.9.mlp.up_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.9.post_attention_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.9.self_attn.k_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.9.self_attn.o_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.9.self_attn.q_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.9.self_attn.v_proj.weight": "model-00002-of-00004.safetensors", + "model.norm.weight": "model-00004-of-00004.safetensors" + } +} diff --git a/checkpoint-815/rng_state_0.pth b/checkpoint-815/rng_state_0.pth new file mode 100644 index 0000000..b647361 --- /dev/null +++ b/checkpoint-815/rng_state_0.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:575119a228f98110923ffa2dedcb50e3317251b26054355d015e0b2240d566f2 +size 15984 diff --git a/checkpoint-815/rng_state_1.pth b/checkpoint-815/rng_state_1.pth new file mode 100644 index 0000000..8506e00 --- /dev/null +++ b/checkpoint-815/rng_state_1.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:0728b56dab7abb5ef8a0d4bae3519c5767c97467bdd886d26bf19cc8599d0312 +size 15984 diff --git a/checkpoint-815/rng_state_10.pth b/checkpoint-815/rng_state_10.pth new file mode 100644 index 0000000..4a20fa5 --- /dev/null +++ b/checkpoint-815/rng_state_10.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:532d1ab071177c2e9d0cc3435d6314c1777a0356f9ce1a5d8f10fbe5f8fd8926 +size 15997 diff --git a/checkpoint-815/rng_state_11.pth b/checkpoint-815/rng_state_11.pth new file mode 100644 index 0000000..a0cf51d --- /dev/null +++ b/checkpoint-815/rng_state_11.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:5a089513416d3ed447265187782ec1194ac2303155b941284463976c62bf3d9b +size 15997 diff --git a/checkpoint-815/rng_state_12.pth b/checkpoint-815/rng_state_12.pth new file mode 100644 index 0000000..97aa200 --- /dev/null +++ b/checkpoint-815/rng_state_12.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:2a87db6b16c9956496c45bf5c81750dc2349e157e60e544e5d256bb957be1243 +size 15997 diff --git a/checkpoint-815/rng_state_13.pth b/checkpoint-815/rng_state_13.pth new file mode 100644 index 0000000..c9f4510 --- /dev/null +++ b/checkpoint-815/rng_state_13.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:36d5fe5d1b92643a3c3dd35cf2f0cbfa4a92599fe3ccbd8941395c1550f088aa +size 15997 diff --git a/checkpoint-815/rng_state_14.pth b/checkpoint-815/rng_state_14.pth new file mode 100644 index 0000000..3c9daf6 --- /dev/null +++ b/checkpoint-815/rng_state_14.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:d4a08278e808af796415aac256bf0c1e8619db1f0f025ff15775cd3b8d02bd59 +size 15997 diff --git a/checkpoint-815/rng_state_15.pth b/checkpoint-815/rng_state_15.pth new file mode 100644 index 0000000..1697bc6 --- /dev/null +++ b/checkpoint-815/rng_state_15.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:88600b3520001a5a7e82c0e184d278b46492fda5424ea22ea243af15b40b2e82 +size 15997 diff --git a/checkpoint-815/rng_state_16.pth b/checkpoint-815/rng_state_16.pth new file mode 100644 index 0000000..480705c --- /dev/null +++ b/checkpoint-815/rng_state_16.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:d9a5825e92eb7518856e88ce648529dc0597c32a5f52764f0361a119217a4dfb +size 15997 diff --git a/checkpoint-815/rng_state_17.pth b/checkpoint-815/rng_state_17.pth new file mode 100644 index 0000000..d425906 --- /dev/null +++ b/checkpoint-815/rng_state_17.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:b30bd2cfa327fb55a93d390d852db109e0a2e1f102ecabf520870cbc79ca95ad +size 15997 diff --git a/checkpoint-815/rng_state_18.pth b/checkpoint-815/rng_state_18.pth new file mode 100644 index 0000000..a8bc0d6 --- /dev/null +++ b/checkpoint-815/rng_state_18.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:77c95f2456fa4473e7231ab94b79ecbc3c126d31d3a94c61f86ee6125b547456 +size 15997 diff --git a/checkpoint-815/rng_state_19.pth b/checkpoint-815/rng_state_19.pth new file mode 100644 index 0000000..40fe188 --- /dev/null +++ b/checkpoint-815/rng_state_19.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:301f380300f7ae9a581c304ac407794ecff9d66e6c9c55ff6cffcfc8d94690a4 +size 15997 diff --git a/checkpoint-815/rng_state_2.pth b/checkpoint-815/rng_state_2.pth new file mode 100644 index 0000000..ea499e2 --- /dev/null +++ b/checkpoint-815/rng_state_2.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:f4e481d4ef1546694da7337f6bb6c658b866dcb79b85deeb477da0d27ebe851e +size 15984 diff --git a/checkpoint-815/rng_state_20.pth b/checkpoint-815/rng_state_20.pth new file mode 100644 index 0000000..4b90c8b --- /dev/null +++ b/checkpoint-815/rng_state_20.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:8b2e90afb8f9481396fa1f82913f7fb908e83e63bf2a38a6f6d2b869ca1d06db +size 15997 diff --git a/checkpoint-815/rng_state_21.pth b/checkpoint-815/rng_state_21.pth new file mode 100644 index 0000000..fb87d85 --- /dev/null +++ b/checkpoint-815/rng_state_21.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:35fafcbe672a0f62668f219b7ae7b88ddc86330086a2eca82505408964d61498 +size 15997 diff --git a/checkpoint-815/rng_state_22.pth b/checkpoint-815/rng_state_22.pth new file mode 100644 index 0000000..ac9beaf --- /dev/null +++ b/checkpoint-815/rng_state_22.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:9b8f851ca87ff60251262afa155c3ab50fc55615fb72c857120fd7fa3fe3177c +size 15997 diff --git a/checkpoint-815/rng_state_23.pth b/checkpoint-815/rng_state_23.pth new file mode 100644 index 0000000..59e574d --- /dev/null +++ b/checkpoint-815/rng_state_23.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:27d8cd100c48bfcfbf5c8a5164b02d5879f69b5a6276dae705a9c60de9a079c1 +size 15997 diff --git a/checkpoint-815/rng_state_24.pth b/checkpoint-815/rng_state_24.pth new file mode 100644 index 0000000..5064859 --- /dev/null +++ b/checkpoint-815/rng_state_24.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:78d1a58c82846308cc3df216d63b5388e6910766e8fbc6b4e8b28255bfff9f0b +size 15997 diff --git a/checkpoint-815/rng_state_25.pth b/checkpoint-815/rng_state_25.pth new file mode 100644 index 0000000..c0ea34f --- /dev/null +++ b/checkpoint-815/rng_state_25.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:bdee831f657ca735b9fcf1c4034be6dcc25cc224594a3b2c7488ea6ec175be96 +size 15997 diff --git a/checkpoint-815/rng_state_26.pth b/checkpoint-815/rng_state_26.pth new file mode 100644 index 0000000..f7f04e9 --- /dev/null +++ b/checkpoint-815/rng_state_26.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:1e922469f65021a6179e6ecc7c2713db0a41dead9dc08d91e98ed34e65afd14e +size 15997 diff --git a/checkpoint-815/rng_state_27.pth b/checkpoint-815/rng_state_27.pth new file mode 100644 index 0000000..f354e1d --- /dev/null +++ b/checkpoint-815/rng_state_27.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:11867e8beaaaacbefb60c912ca48d2026f0222b2247dd97d78a6ddd227e75bcf +size 15997 diff --git a/checkpoint-815/rng_state_28.pth b/checkpoint-815/rng_state_28.pth new file mode 100644 index 0000000..12c799a --- /dev/null +++ b/checkpoint-815/rng_state_28.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:7664653b5278a30fcc7bd77ee73888988040ecbb6a9cb71fd8a5bb5ccd2c9a31 +size 15997 diff --git a/checkpoint-815/rng_state_29.pth b/checkpoint-815/rng_state_29.pth new file mode 100644 index 0000000..6feba08 --- /dev/null +++ b/checkpoint-815/rng_state_29.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:5007396f0c3c95678e1e62a10e31552d6bfe2303e6e78ab6882752474b2efdc7 +size 15997 diff --git a/checkpoint-815/rng_state_3.pth b/checkpoint-815/rng_state_3.pth new file mode 100644 index 0000000..aeb38f9 --- /dev/null +++ b/checkpoint-815/rng_state_3.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:353c60be37ea56fc992fca446598ceca5d1fd002aa3bd6dbb9ad740e6f47ebb3 +size 15984 diff --git a/checkpoint-815/rng_state_30.pth b/checkpoint-815/rng_state_30.pth new file mode 100644 index 0000000..52254ce --- /dev/null +++ b/checkpoint-815/rng_state_30.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:f01c58f06cc2a38c9093fadd85def1f922a5e8910e17a0d885f0df5570b13a98 +size 15997 diff --git a/checkpoint-815/rng_state_31.pth b/checkpoint-815/rng_state_31.pth new file mode 100644 index 0000000..f6010f3 --- /dev/null +++ b/checkpoint-815/rng_state_31.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:35e075e15ab066bd7c9405a66e387f2ebf6b302fbd19ba7905da3f8cc8f71009 +size 15997 diff --git a/checkpoint-815/rng_state_4.pth b/checkpoint-815/rng_state_4.pth new file mode 100644 index 0000000..9d5856c --- /dev/null +++ b/checkpoint-815/rng_state_4.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:e9107fe964ba7205e354084b85210e5a5ea1c98cfd4d38adb9cd3926945dcae4 +size 15984 diff --git a/checkpoint-815/rng_state_5.pth b/checkpoint-815/rng_state_5.pth new file mode 100644 index 0000000..b824ee2 --- /dev/null +++ b/checkpoint-815/rng_state_5.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:69d1bb1abee38b92e53f3f23549b642ce0f1edcdccf7b6129847ac61636e96d5 +size 15984 diff --git a/checkpoint-815/rng_state_6.pth b/checkpoint-815/rng_state_6.pth new file mode 100644 index 0000000..a9fd036 --- /dev/null +++ b/checkpoint-815/rng_state_6.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:afd5516048e20f36959601574e29e40106085a7d3cdc7bf425ce5e84633490e6 +size 15984 diff --git a/checkpoint-815/rng_state_7.pth b/checkpoint-815/rng_state_7.pth new file mode 100644 index 0000000..4e80125 --- /dev/null +++ b/checkpoint-815/rng_state_7.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:8e2c46927fc06939b4c976a01e4b95dec1f8b98ceaea86d31a5d756fc30ff006 +size 15984 diff --git a/checkpoint-815/rng_state_8.pth b/checkpoint-815/rng_state_8.pth new file mode 100644 index 0000000..b95c8ee --- /dev/null +++ b/checkpoint-815/rng_state_8.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:87544a1ecb93bba51fc38c978d8a027be1635c93e80f7f0679c1154bc76e16d3 +size 15984 diff --git a/checkpoint-815/rng_state_9.pth b/checkpoint-815/rng_state_9.pth new file mode 100644 index 0000000..8ff6eb6 --- /dev/null +++ b/checkpoint-815/rng_state_9.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:3f51e1e81143bcfd2cfd36874b3b7606bb4a14aac945154b167b31a816d9f450 +size 15984 diff --git a/checkpoint-815/scheduler.pt b/checkpoint-815/scheduler.pt new file mode 100644 index 0000000..7da08a0 --- /dev/null +++ b/checkpoint-815/scheduler.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:c0e36c989c2706e3716f57653cb08c5a663bd94882237c06685061e30b2fffbe +size 1064 diff --git a/checkpoint-815/special_tokens_map.json b/checkpoint-815/special_tokens_map.json new file mode 100644 index 0000000..278b7f0 --- /dev/null +++ b/checkpoint-815/special_tokens_map.json @@ -0,0 +1,23 @@ +{ + "bos_token": { + "content": "<|begin_of_text|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false + }, + "eos_token": { + "content": "<|eot_id|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false + }, + "pad_token": { + "content": "<|end_of_text|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false + } +} diff --git a/checkpoint-815/tokenizer.json b/checkpoint-815/tokenizer.json new file mode 100644 index 0000000..1c1d8d5 --- /dev/null +++ b/checkpoint-815/tokenizer.json @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:6b9e4e7fb171f92fd137b777cc2714bf87d11576700a1dcd7a399e7bbe39537b +size 17209920 diff --git a/checkpoint-815/tokenizer_config.json b/checkpoint-815/tokenizer_config.json new file mode 100644 index 0000000..8098161 --- /dev/null +++ b/checkpoint-815/tokenizer_config.json @@ -0,0 +1,2064 @@ +{ + "added_tokens_decoder": { + "128000": { + "content": "<|begin_of_text|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128001": { + "content": "<|end_of_text|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128002": { + "content": "<|reserved_special_token_0|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128003": { + "content": "<|reserved_special_token_1|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128004": { + "content": "<|finetune_right_pad_id|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128005": { + "content": "<|reserved_special_token_2|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128006": { + "content": "<|start_header_id|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128007": { + "content": "<|end_header_id|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128008": { + "content": "<|eom_id|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128009": { + "content": "<|eot_id|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128010": { + "content": "<|python_tag|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128011": { + "content": "<|reserved_special_token_3|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128012": { + "content": "<|reserved_special_token_4|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128013": { + "content": "<|reserved_special_token_5|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128014": { + "content": "<|reserved_special_token_6|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128015": { + "content": "<|reserved_special_token_7|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128016": { + "content": "<|reserved_special_token_8|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128017": { + "content": "<|reserved_special_token_9|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128018": { + "content": "<|reserved_special_token_10|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128019": { + "content": "<|reserved_special_token_11|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128020": { + "content": "<|reserved_special_token_12|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128021": { + "content": "<|reserved_special_token_13|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128022": { + "content": "<|reserved_special_token_14|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128023": { + "content": "<|reserved_special_token_15|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128024": { + "content": "<|reserved_special_token_16|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128025": { + "content": "<|reserved_special_token_17|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128026": { + "content": "<|reserved_special_token_18|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128027": { + "content": "<|reserved_special_token_19|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128028": { + "content": "<|reserved_special_token_20|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128029": { + "content": "<|reserved_special_token_21|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128030": { + "content": "<|reserved_special_token_22|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128031": { + "content": "<|reserved_special_token_23|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128032": { + "content": "<|reserved_special_token_24|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128033": { + "content": "<|reserved_special_token_25|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128034": { + "content": "<|reserved_special_token_26|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128035": { + "content": "<|reserved_special_token_27|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128036": { + "content": "<|reserved_special_token_28|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128037": { + "content": "<|reserved_special_token_29|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128038": { + "content": "<|reserved_special_token_30|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128039": { + "content": "<|reserved_special_token_31|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128040": { + "content": "<|reserved_special_token_32|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128041": { + "content": "<|reserved_special_token_33|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128042": { + "content": "<|reserved_special_token_34|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128043": { + "content": "<|reserved_special_token_35|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128044": { + "content": "<|reserved_special_token_36|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128045": { + "content": "<|reserved_special_token_37|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128046": { + "content": "<|reserved_special_token_38|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128047": { + "content": "<|reserved_special_token_39|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128048": { + "content": "<|reserved_special_token_40|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128049": { + "content": "<|reserved_special_token_41|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128050": { + "content": "<|reserved_special_token_42|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128051": { + "content": "<|reserved_special_token_43|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128052": { + "content": "<|reserved_special_token_44|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128053": { + "content": "<|reserved_special_token_45|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128054": { + "content": "<|reserved_special_token_46|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128055": { + "content": "<|reserved_special_token_47|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128056": { + "content": "<|reserved_special_token_48|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128057": { + "content": "<|reserved_special_token_49|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128058": { + "content": "<|reserved_special_token_50|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128059": { + "content": "<|reserved_special_token_51|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128060": { + "content": "<|reserved_special_token_52|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128061": { + "content": "<|reserved_special_token_53|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128062": { + "content": "<|reserved_special_token_54|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128063": { + "content": "<|reserved_special_token_55|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128064": { + "content": "<|reserved_special_token_56|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128065": { + "content": "<|reserved_special_token_57|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128066": { + "content": "<|reserved_special_token_58|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128067": { + "content": "<|reserved_special_token_59|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128068": { + "content": "<|reserved_special_token_60|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128069": { + "content": "<|reserved_special_token_61|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128070": { + "content": "<|reserved_special_token_62|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128071": { + "content": "<|reserved_special_token_63|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128072": { + "content": "<|reserved_special_token_64|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128073": { + "content": "<|reserved_special_token_65|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128074": { + "content": "<|reserved_special_token_66|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128075": { + "content": "<|reserved_special_token_67|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128076": { + "content": "<|reserved_special_token_68|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128077": { + "content": "<|reserved_special_token_69|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128078": { + "content": "<|reserved_special_token_70|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128079": { + "content": "<|reserved_special_token_71|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128080": { + "content": "<|reserved_special_token_72|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128081": { + "content": "<|reserved_special_token_73|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128082": { + "content": "<|reserved_special_token_74|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128083": { + "content": "<|reserved_special_token_75|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128084": { + "content": "<|reserved_special_token_76|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128085": { + "content": "<|reserved_special_token_77|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128086": { + "content": "<|reserved_special_token_78|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128087": { + "content": "<|reserved_special_token_79|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128088": { + "content": "<|reserved_special_token_80|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128089": { + "content": "<|reserved_special_token_81|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128090": { + "content": "<|reserved_special_token_82|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128091": { + "content": "<|reserved_special_token_83|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128092": { + "content": "<|reserved_special_token_84|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128093": { + "content": "<|reserved_special_token_85|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128094": { + "content": "<|reserved_special_token_86|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128095": { + "content": "<|reserved_special_token_87|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128096": { + "content": "<|reserved_special_token_88|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128097": { + "content": "<|reserved_special_token_89|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128098": { + "content": "<|reserved_special_token_90|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128099": { + "content": "<|reserved_special_token_91|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128100": { + "content": "<|reserved_special_token_92|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128101": { + "content": "<|reserved_special_token_93|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128102": { + "content": "<|reserved_special_token_94|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128103": { + "content": "<|reserved_special_token_95|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128104": { + "content": "<|reserved_special_token_96|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128105": { + "content": "<|reserved_special_token_97|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128106": { + "content": "<|reserved_special_token_98|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128107": { + "content": "<|reserved_special_token_99|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128108": { + "content": "<|reserved_special_token_100|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128109": { + "content": "<|reserved_special_token_101|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128110": { + "content": "<|reserved_special_token_102|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128111": { + "content": "<|reserved_special_token_103|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128112": { + "content": "<|reserved_special_token_104|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128113": { + "content": "<|reserved_special_token_105|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128114": { + "content": "<|reserved_special_token_106|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128115": { + "content": "<|reserved_special_token_107|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128116": { + "content": "<|reserved_special_token_108|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128117": { + "content": "<|reserved_special_token_109|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128118": { + "content": "<|reserved_special_token_110|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128119": { + "content": "<|reserved_special_token_111|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128120": { + "content": "<|reserved_special_token_112|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128121": { + "content": "<|reserved_special_token_113|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128122": { + "content": "<|reserved_special_token_114|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128123": { + "content": "<|reserved_special_token_115|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128124": { + "content": "<|reserved_special_token_116|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128125": { + "content": "<|reserved_special_token_117|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128126": { + "content": "<|reserved_special_token_118|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128127": { + "content": "<|reserved_special_token_119|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128128": { + "content": "<|reserved_special_token_120|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128129": { + "content": "<|reserved_special_token_121|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128130": { + "content": "<|reserved_special_token_122|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128131": { + "content": "<|reserved_special_token_123|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128132": { + "content": "<|reserved_special_token_124|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128133": { + "content": "<|reserved_special_token_125|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128134": { + "content": "<|reserved_special_token_126|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128135": { + "content": "<|reserved_special_token_127|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128136": { + "content": "<|reserved_special_token_128|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128137": { + "content": "<|reserved_special_token_129|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128138": { + "content": "<|reserved_special_token_130|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128139": { + "content": "<|reserved_special_token_131|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128140": { + "content": "<|reserved_special_token_132|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128141": { + "content": "<|reserved_special_token_133|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128142": { + "content": "<|reserved_special_token_134|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128143": { + "content": "<|reserved_special_token_135|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128144": { + "content": "<|reserved_special_token_136|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128145": { + "content": "<|reserved_special_token_137|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128146": { + "content": "<|reserved_special_token_138|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128147": { + "content": "<|reserved_special_token_139|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128148": { + "content": "<|reserved_special_token_140|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128149": { + "content": "<|reserved_special_token_141|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128150": { + "content": "<|reserved_special_token_142|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128151": { + "content": "<|reserved_special_token_143|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128152": { + "content": "<|reserved_special_token_144|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128153": { + "content": "<|reserved_special_token_145|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128154": { + "content": "<|reserved_special_token_146|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128155": { + "content": "<|reserved_special_token_147|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128156": { + "content": "<|reserved_special_token_148|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128157": { + "content": "<|reserved_special_token_149|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128158": { + "content": "<|reserved_special_token_150|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128159": { + "content": "<|reserved_special_token_151|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128160": { + "content": "<|reserved_special_token_152|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128161": { + "content": "<|reserved_special_token_153|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128162": { + "content": "<|reserved_special_token_154|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128163": { + "content": "<|reserved_special_token_155|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128164": { + "content": "<|reserved_special_token_156|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128165": { + "content": "<|reserved_special_token_157|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128166": { + "content": "<|reserved_special_token_158|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128167": { + "content": "<|reserved_special_token_159|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128168": { + "content": "<|reserved_special_token_160|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128169": { + "content": "<|reserved_special_token_161|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128170": { + "content": "<|reserved_special_token_162|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128171": { + "content": "<|reserved_special_token_163|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128172": { + "content": "<|reserved_special_token_164|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128173": { + "content": "<|reserved_special_token_165|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128174": { + "content": "<|reserved_special_token_166|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128175": { + "content": "<|reserved_special_token_167|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128176": { + "content": "<|reserved_special_token_168|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128177": { + "content": "<|reserved_special_token_169|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128178": { + "content": "<|reserved_special_token_170|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128179": { + "content": "<|reserved_special_token_171|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128180": { + "content": "<|reserved_special_token_172|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128181": { + "content": "<|reserved_special_token_173|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128182": { + "content": "<|reserved_special_token_174|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128183": { + "content": "<|reserved_special_token_175|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128184": { + "content": "<|reserved_special_token_176|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128185": { + "content": "<|reserved_special_token_177|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128186": { + "content": "<|reserved_special_token_178|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128187": { + "content": "<|reserved_special_token_179|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128188": { + "content": "<|reserved_special_token_180|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128189": { + "content": "<|reserved_special_token_181|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128190": { + "content": "<|reserved_special_token_182|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128191": { + "content": "<|reserved_special_token_183|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128192": { + "content": "<|reserved_special_token_184|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128193": { + "content": "<|reserved_special_token_185|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128194": { + "content": "<|reserved_special_token_186|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128195": { + "content": "<|reserved_special_token_187|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128196": { + "content": "<|reserved_special_token_188|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128197": { + "content": "<|reserved_special_token_189|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128198": { + "content": "<|reserved_special_token_190|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128199": { + "content": "<|reserved_special_token_191|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128200": { + "content": "<|reserved_special_token_192|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128201": { + "content": "<|reserved_special_token_193|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128202": { + "content": "<|reserved_special_token_194|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128203": { + "content": "<|reserved_special_token_195|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128204": { + "content": "<|reserved_special_token_196|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128205": { + "content": "<|reserved_special_token_197|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128206": { + "content": "<|reserved_special_token_198|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128207": { + "content": "<|reserved_special_token_199|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128208": { + "content": "<|reserved_special_token_200|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128209": { + "content": "<|reserved_special_token_201|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128210": { + "content": "<|reserved_special_token_202|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128211": { + "content": "<|reserved_special_token_203|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128212": { + "content": "<|reserved_special_token_204|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128213": { + "content": "<|reserved_special_token_205|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128214": { + "content": "<|reserved_special_token_206|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128215": { + "content": "<|reserved_special_token_207|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128216": { + "content": "<|reserved_special_token_208|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128217": { + "content": "<|reserved_special_token_209|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128218": { + "content": "<|reserved_special_token_210|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128219": { + "content": "<|reserved_special_token_211|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128220": { + "content": "<|reserved_special_token_212|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128221": { + "content": "<|reserved_special_token_213|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128222": { + "content": "<|reserved_special_token_214|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128223": { + "content": "<|reserved_special_token_215|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128224": { + "content": "<|reserved_special_token_216|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128225": { + "content": "<|reserved_special_token_217|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128226": { + "content": "<|reserved_special_token_218|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128227": { + "content": "<|reserved_special_token_219|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128228": { + "content": "<|reserved_special_token_220|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128229": { + "content": "<|reserved_special_token_221|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128230": { + "content": "<|reserved_special_token_222|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128231": { + "content": "<|reserved_special_token_223|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128232": { + "content": "<|reserved_special_token_224|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128233": { + "content": "<|reserved_special_token_225|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128234": { + "content": "<|reserved_special_token_226|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128235": { + "content": "<|reserved_special_token_227|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128236": { + "content": "<|reserved_special_token_228|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128237": { + "content": "<|reserved_special_token_229|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128238": { + "content": "<|reserved_special_token_230|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128239": { + "content": "<|reserved_special_token_231|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128240": { + "content": "<|reserved_special_token_232|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128241": { + "content": "<|reserved_special_token_233|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128242": { + "content": "<|reserved_special_token_234|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128243": { + "content": "<|reserved_special_token_235|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128244": { + "content": "<|reserved_special_token_236|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128245": { + "content": "<|reserved_special_token_237|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128246": { + "content": "<|reserved_special_token_238|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128247": { + "content": "<|reserved_special_token_239|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128248": { + "content": "<|reserved_special_token_240|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128249": { + "content": "<|reserved_special_token_241|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128250": { + "content": "<|reserved_special_token_242|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128251": { + "content": "<|reserved_special_token_243|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128252": { + "content": "<|reserved_special_token_244|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128253": { + "content": "<|reserved_special_token_245|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128254": { + "content": "<|reserved_special_token_246|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128255": { + "content": "<|reserved_special_token_247|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + } + }, + "bos_token": "<|begin_of_text|>", + "chat_template": "{% if not add_generation_prompt is defined %}{% set add_generation_prompt = false %}{% endif %}{% set loop_messages = messages %}{% for message in loop_messages %}{% set content = '<|start_header_id|>' + message['role'] + '<|end_header_id|>\n\n'+ message['content'] | trim + '<|eot_id|>' %}{% if loop.index0 == 0 %}{% set content = bos_token + content %}{% endif %}{{ content }}{% endfor %}{% if add_generation_prompt %}{{ '<|start_header_id|>assistant<|end_header_id|>\n\n' }}{% endif %}", + "clean_up_tokenization_spaces": true, + "eos_token": "<|eot_id|>", + "extra_special_tokens": {}, + "model_input_names": [ + "input_ids", + "attention_mask" + ], + "model_max_length": 131072, + "pad_token": "<|end_of_text|>", + "tokenizer_class": "PreTrainedTokenizer" +} diff --git a/checkpoint-815/trainer_state.json b/checkpoint-815/trainer_state.json new file mode 100644 index 0000000..7482f6e --- /dev/null +++ b/checkpoint-815/trainer_state.json @@ -0,0 +1,12259 @@ +{ + "best_global_step": null, + "best_metric": null, + "best_model_checkpoint": null, + "epoch": 1.0, + "eval_steps": 500, + "global_step": 815, + "is_hyper_param_search": false, + "is_local_process_zero": true, + "is_world_process_zero": true, + "log_history": [ + { + "epoch": 0.001226993865030675, + "grad_norm": 151.20719434086217, + "learning_rate": 6.000000000000001e-08, + "logits/chosen": -0.32421875, + "logits/rejected": -0.515625, + "logps/chosen": -474.0, + "logps/rejected": -316.0, + "loss": 0.6914, + "rewards/accuracies": 0.0, + "rewards/chosen": 0.0, + "rewards/margins": 0.0, + "rewards/rejected": 0.0, + "step": 1 + }, + { + "epoch": 0.00245398773006135, + "grad_norm": 162.171350331327, + "learning_rate": 1.2000000000000002e-07, + "logits/chosen": -0.22265625, + "logits/rejected": -0.439453125, + "logps/chosen": -436.0, + "logps/rejected": -292.0, + "loss": 0.6914, + "rewards/accuracies": 0.0, + "rewards/chosen": 0.0, + "rewards/margins": 0.0, + "rewards/rejected": 0.0, + "step": 2 + }, + { + "epoch": 0.0036809815950920245, + "grad_norm": 121.91181472288706, + "learning_rate": 1.8e-07, + "logits/chosen": -0.1298828125, + "logits/rejected": -0.2578125, + "logps/chosen": -368.0, + "logps/rejected": -272.0, + "loss": 0.7012, + "rewards/accuracies": 0.3359375, + "rewards/chosen": -0.0006103515625, + "rewards/margins": -0.00506591796875, + "rewards/rejected": 0.00445556640625, + "step": 3 + }, + { + "epoch": 0.0049079754601227, + "grad_norm": 138.61232520616198, + "learning_rate": 2.4000000000000003e-07, + "logits/chosen": -0.166015625, + "logits/rejected": -0.326171875, + "logps/chosen": -422.0, + "logps/rejected": -242.0, + "loss": 0.7003, + "rewards/accuracies": 0.359375, + "rewards/chosen": 0.0057373046875, + "rewards/margins": -0.00762939453125, + "rewards/rejected": 0.01336669921875, + "step": 4 + }, + { + "epoch": 0.006134969325153374, + "grad_norm": 318.4843338984337, + "learning_rate": 3.0000000000000004e-07, + "logits/chosen": -0.1826171875, + "logits/rejected": -0.32421875, + "logps/chosen": -380.0, + "logps/rejected": -270.0, + "loss": 0.7002, + "rewards/accuracies": 0.296875, + "rewards/chosen": -0.0037994384765625, + "rewards/margins": -0.01123046875, + "rewards/rejected": 0.007415771484375, + "step": 5 + }, + { + "epoch": 0.007361963190184049, + "grad_norm": 356.6573592678301, + "learning_rate": 3.6e-07, + "logits/chosen": -0.255859375, + "logits/rejected": -0.345703125, + "logps/chosen": -412.0, + "logps/rejected": -306.0, + "loss": 0.692, + "rewards/accuracies": 0.359375, + "rewards/chosen": 0.0208740234375, + "rewards/margins": 0.01153564453125, + "rewards/rejected": 0.00927734375, + "step": 6 + }, + { + "epoch": 0.008588957055214725, + "grad_norm": 228.0843622624859, + "learning_rate": 4.2000000000000006e-07, + "logits/chosen": -0.2060546875, + "logits/rejected": -0.38671875, + "logps/chosen": -378.0, + "logps/rejected": -266.0, + "loss": 0.6962, + "rewards/accuracies": 0.3671875, + "rewards/chosen": 0.006683349609375, + "rewards/margins": -0.001617431640625, + "rewards/rejected": 0.00830078125, + "step": 7 + }, + { + "epoch": 0.0098159509202454, + "grad_norm": 139.9677750400214, + "learning_rate": 4.800000000000001e-07, + "logits/chosen": -0.111328125, + "logits/rejected": -0.3203125, + "logps/chosen": -432.0, + "logps/rejected": -292.0, + "loss": 0.6667, + "rewards/accuracies": 0.453125, + "rewards/chosen": 0.054443359375, + "rewards/margins": 0.06201171875, + "rewards/rejected": -0.00762939453125, + "step": 8 + }, + { + "epoch": 0.011042944785276074, + "grad_norm": 114.26865689975604, + "learning_rate": 5.4e-07, + "logits/chosen": -0.1318359375, + "logits/rejected": -0.359375, + "logps/chosen": -402.0, + "logps/rejected": -294.0, + "loss": 0.6492, + "rewards/accuracies": 0.4921875, + "rewards/chosen": 0.0654296875, + "rewards/margins": 0.10986328125, + "rewards/rejected": -0.044677734375, + "step": 9 + }, + { + "epoch": 0.012269938650306749, + "grad_norm": 150.56729793626556, + "learning_rate": 6.000000000000001e-07, + "logits/chosen": -0.1953125, + "logits/rejected": -0.349609375, + "logps/chosen": -402.0, + "logps/rejected": -278.0, + "loss": 0.6146, + "rewards/accuracies": 0.59375, + "rewards/chosen": 0.158203125, + "rewards/margins": 0.1943359375, + "rewards/rejected": -0.0361328125, + "step": 10 + }, + { + "epoch": 0.013496932515337423, + "grad_norm": 85.94823037139018, + "learning_rate": 6.6e-07, + "logits/chosen": -0.375, + "logits/rejected": -0.515625, + "logps/chosen": -420.0, + "logps/rejected": -298.0, + "loss": 0.5535, + "rewards/accuracies": 0.6875, + "rewards/chosen": 0.2451171875, + "rewards/margins": 0.3984375, + "rewards/rejected": -0.1533203125, + "step": 11 + }, + { + "epoch": 0.014723926380368098, + "grad_norm": 60.96373680958018, + "learning_rate": 7.2e-07, + "logits/chosen": -0.1337890625, + "logits/rejected": -0.244140625, + "logps/chosen": -372.0, + "logps/rejected": -262.0, + "loss": 0.548, + "rewards/accuracies": 0.6953125, + "rewards/chosen": 0.255859375, + "rewards/margins": 0.435546875, + "rewards/rejected": -0.1796875, + "step": 12 + }, + { + "epoch": 0.015950920245398775, + "grad_norm": 92.70871898741228, + "learning_rate": 7.8e-07, + "logits/chosen": -0.240234375, + "logits/rejected": -0.408203125, + "logps/chosen": -390.0, + "logps/rejected": -251.0, + "loss": 0.5255, + "rewards/accuracies": 0.703125, + "rewards/chosen": 0.302734375, + "rewards/margins": 0.56640625, + "rewards/rejected": -0.26171875, + "step": 13 + }, + { + "epoch": 0.01717791411042945, + "grad_norm": 48.59299098972598, + "learning_rate": 8.400000000000001e-07, + "logits/chosen": -0.2138671875, + "logits/rejected": -0.3984375, + "logps/chosen": -466.0, + "logps/rejected": -314.0, + "loss": 0.4612, + "rewards/accuracies": 0.7265625, + "rewards/chosen": 0.765625, + "rewards/margins": 1.078125, + "rewards/rejected": -0.31640625, + "step": 14 + }, + { + "epoch": 0.018404907975460124, + "grad_norm": 59.699183777930486, + "learning_rate": 9e-07, + "logits/chosen": -0.10400390625, + "logits/rejected": -0.177734375, + "logps/chosen": -414.0, + "logps/rejected": -292.0, + "loss": 0.5014, + "rewards/accuracies": 0.734375, + "rewards/chosen": 0.6484375, + "rewards/margins": 0.96875, + "rewards/rejected": -0.31640625, + "step": 15 + }, + { + "epoch": 0.0196319018404908, + "grad_norm": 52.10075632980986, + "learning_rate": 9.600000000000001e-07, + "logits/chosen": -0.25, + "logits/rejected": -0.42578125, + "logps/chosen": -400.0, + "logps/rejected": -310.0, + "loss": 0.4424, + "rewards/accuracies": 0.7265625, + "rewards/chosen": 0.84375, + "rewards/margins": 1.390625, + "rewards/rejected": -0.54296875, + "step": 16 + }, + { + "epoch": 0.020858895705521473, + "grad_norm": 42.16097538458702, + "learning_rate": 1.0200000000000002e-06, + "logits/chosen": -0.1572265625, + "logits/rejected": -0.31640625, + "logps/chosen": -426.0, + "logps/rejected": -260.0, + "loss": 0.4308, + "rewards/accuracies": 0.796875, + "rewards/chosen": 0.984375, + "rewards/margins": 1.578125, + "rewards/rejected": -0.58984375, + "step": 17 + }, + { + "epoch": 0.022085889570552148, + "grad_norm": 46.56295761289983, + "learning_rate": 1.08e-06, + "logits/chosen": -0.1435546875, + "logits/rejected": -0.330078125, + "logps/chosen": -398.0, + "logps/rejected": -282.0, + "loss": 0.432, + "rewards/accuracies": 0.828125, + "rewards/chosen": 0.875, + "rewards/margins": 1.6875, + "rewards/rejected": -0.8125, + "step": 18 + }, + { + "epoch": 0.023312883435582823, + "grad_norm": 49.57615257325465, + "learning_rate": 1.14e-06, + "logits/chosen": -0.185546875, + "logits/rejected": -0.3671875, + "logps/chosen": -382.0, + "logps/rejected": -296.0, + "loss": 0.4802, + "rewards/accuracies": 0.7734375, + "rewards/chosen": 0.76171875, + "rewards/margins": 1.8125, + "rewards/rejected": -1.0546875, + "step": 19 + }, + { + "epoch": 0.024539877300613498, + "grad_norm": 42.63506109619643, + "learning_rate": 1.2000000000000002e-06, + "logits/chosen": -0.263671875, + "logits/rejected": -0.439453125, + "logps/chosen": -456.0, + "logps/rejected": -324.0, + "loss": 0.3801, + "rewards/accuracies": 0.8125, + "rewards/chosen": 1.4375, + "rewards/margins": 3.140625, + "rewards/rejected": -1.703125, + "step": 20 + }, + { + "epoch": 0.025766871165644172, + "grad_norm": 34.456600308133275, + "learning_rate": 1.26e-06, + "logits/chosen": -0.1328125, + "logits/rejected": -0.3203125, + "logps/chosen": -404.0, + "logps/rejected": -272.0, + "loss": 0.3497, + "rewards/accuracies": 0.8515625, + "rewards/chosen": 1.1015625, + "rewards/margins": 2.640625, + "rewards/rejected": -1.53125, + "step": 21 + }, + { + "epoch": 0.026993865030674847, + "grad_norm": 44.983878720566445, + "learning_rate": 1.32e-06, + "logits/chosen": -0.1298828125, + "logits/rejected": -0.271484375, + "logps/chosen": -374.0, + "logps/rejected": -306.0, + "loss": 0.4156, + "rewards/accuracies": 0.7890625, + "rewards/chosen": 1.0546875, + "rewards/margins": 2.84375, + "rewards/rejected": -1.7890625, + "step": 22 + }, + { + "epoch": 0.02822085889570552, + "grad_norm": 56.72818428668396, + "learning_rate": 1.3800000000000001e-06, + "logits/chosen": -0.251953125, + "logits/rejected": -0.314453125, + "logps/chosen": -370.0, + "logps/rejected": -318.0, + "loss": 0.4566, + "rewards/accuracies": 0.7890625, + "rewards/chosen": 1.46875, + "rewards/margins": 3.1875, + "rewards/rejected": -1.7265625, + "step": 23 + }, + { + "epoch": 0.029447852760736196, + "grad_norm": 55.662509241103734, + "learning_rate": 1.44e-06, + "logits/chosen": -0.1884765625, + "logits/rejected": -0.375, + "logps/chosen": -376.0, + "logps/rejected": -288.0, + "loss": 0.4763, + "rewards/accuracies": 0.7421875, + "rewards/chosen": 0.64453125, + "rewards/margins": 2.265625, + "rewards/rejected": -1.625, + "step": 24 + }, + { + "epoch": 0.03067484662576687, + "grad_norm": 35.38065334196675, + "learning_rate": 1.5e-06, + "logits/chosen": -0.2578125, + "logits/rejected": -0.4453125, + "logps/chosen": -430.0, + "logps/rejected": -312.0, + "loss": 0.3378, + "rewards/accuracies": 0.8984375, + "rewards/chosen": 1.2890625, + "rewards/margins": 3.515625, + "rewards/rejected": -2.21875, + "step": 25 + }, + { + "epoch": 0.03190184049079755, + "grad_norm": 42.887450752414956, + "learning_rate": 1.4993801652892562e-06, + "logits/chosen": -0.15625, + "logits/rejected": -0.31640625, + "logps/chosen": -380.0, + "logps/rejected": -288.0, + "loss": 0.4058, + "rewards/accuracies": 0.8046875, + "rewards/chosen": 1.640625, + "rewards/margins": 3.59375, + "rewards/rejected": -1.9609375, + "step": 26 + }, + { + "epoch": 0.033128834355828224, + "grad_norm": 37.07120513953294, + "learning_rate": 1.4987603305785125e-06, + "logits/chosen": -0.095703125, + "logits/rejected": -0.2734375, + "logps/chosen": -354.0, + "logps/rejected": -292.0, + "loss": 0.3683, + "rewards/accuracies": 0.765625, + "rewards/chosen": 1.7578125, + "rewards/margins": 4.09375, + "rewards/rejected": -2.34375, + "step": 27 + }, + { + "epoch": 0.0343558282208589, + "grad_norm": 35.87136073029194, + "learning_rate": 1.4981404958677687e-06, + "logits/chosen": -0.036376953125, + "logits/rejected": -0.1484375, + "logps/chosen": -374.0, + "logps/rejected": -274.0, + "loss": 0.3615, + "rewards/accuracies": 0.8125, + "rewards/chosen": 1.6171875, + "rewards/margins": 3.421875, + "rewards/rejected": -1.796875, + "step": 28 + }, + { + "epoch": 0.03558282208588957, + "grad_norm": 38.17662218776474, + "learning_rate": 1.4975206611570248e-06, + "logits/chosen": -0.263671875, + "logits/rejected": -0.4140625, + "logps/chosen": -414.0, + "logps/rejected": -334.0, + "loss": 0.3367, + "rewards/accuracies": 0.8046875, + "rewards/chosen": 1.765625, + "rewards/margins": 4.0625, + "rewards/rejected": -2.296875, + "step": 29 + }, + { + "epoch": 0.03680981595092025, + "grad_norm": 39.994932388621756, + "learning_rate": 1.496900826446281e-06, + "logits/chosen": -0.1337890625, + "logits/rejected": -0.248046875, + "logps/chosen": -332.0, + "logps/rejected": -278.0, + "loss": 0.3648, + "rewards/accuracies": 0.8203125, + "rewards/chosen": 1.203125, + "rewards/margins": 3.375, + "rewards/rejected": -2.171875, + "step": 30 + }, + { + "epoch": 0.03803680981595092, + "grad_norm": 42.39396872212231, + "learning_rate": 1.496280991735537e-06, + "logits/chosen": -0.10986328125, + "logits/rejected": -0.2275390625, + "logps/chosen": -378.0, + "logps/rejected": -286.0, + "loss": 0.364, + "rewards/accuracies": 0.7890625, + "rewards/chosen": 1.9765625, + "rewards/margins": 4.25, + "rewards/rejected": -2.265625, + "step": 31 + }, + { + "epoch": 0.0392638036809816, + "grad_norm": 43.84495796047598, + "learning_rate": 1.4956611570247935e-06, + "logits/chosen": -0.15625, + "logits/rejected": -0.333984375, + "logps/chosen": -388.0, + "logps/rejected": -298.0, + "loss": 0.3861, + "rewards/accuracies": 0.78125, + "rewards/chosen": 1.71875, + "rewards/margins": 4.0, + "rewards/rejected": -2.28125, + "step": 32 + }, + { + "epoch": 0.04049079754601227, + "grad_norm": 38.440307864705964, + "learning_rate": 1.4950413223140496e-06, + "logits/chosen": -0.04443359375, + "logits/rejected": -0.177734375, + "logps/chosen": -328.0, + "logps/rejected": -248.0, + "loss": 0.3456, + "rewards/accuracies": 0.8515625, + "rewards/chosen": 1.078125, + "rewards/margins": 3.5625, + "rewards/rejected": -2.484375, + "step": 33 + }, + { + "epoch": 0.04171779141104295, + "grad_norm": 42.669938783190766, + "learning_rate": 1.4944214876033057e-06, + "logits/chosen": -0.1796875, + "logits/rejected": -0.296875, + "logps/chosen": -384.0, + "logps/rejected": -316.0, + "loss": 0.3991, + "rewards/accuracies": 0.7890625, + "rewards/chosen": 1.5859375, + "rewards/margins": 4.40625, + "rewards/rejected": -2.8125, + "step": 34 + }, + { + "epoch": 0.04294478527607362, + "grad_norm": 39.82238013261856, + "learning_rate": 1.493801652892562e-06, + "logits/chosen": -0.0810546875, + "logits/rejected": -0.2197265625, + "logps/chosen": -356.0, + "logps/rejected": -278.0, + "loss": 0.3621, + "rewards/accuracies": 0.828125, + "rewards/chosen": 1.1796875, + "rewards/margins": 4.125, + "rewards/rejected": -2.9375, + "step": 35 + }, + { + "epoch": 0.044171779141104296, + "grad_norm": 34.79756258065961, + "learning_rate": 1.4931818181818182e-06, + "logits/chosen": -0.1376953125, + "logits/rejected": -0.20703125, + "logps/chosen": -398.0, + "logps/rejected": -310.0, + "loss": 0.3113, + "rewards/accuracies": 0.8515625, + "rewards/chosen": 1.6640625, + "rewards/margins": 4.78125, + "rewards/rejected": -3.125, + "step": 36 + }, + { + "epoch": 0.04539877300613497, + "grad_norm": 40.21812818356396, + "learning_rate": 1.4925619834710746e-06, + "logits/chosen": -0.091796875, + "logits/rejected": -0.1943359375, + "logps/chosen": -384.0, + "logps/rejected": -322.0, + "loss": 0.3512, + "rewards/accuracies": 0.796875, + "rewards/chosen": 2.3125, + "rewards/margins": 4.625, + "rewards/rejected": -2.328125, + "step": 37 + }, + { + "epoch": 0.046625766871165646, + "grad_norm": 38.815201724599156, + "learning_rate": 1.4919421487603307e-06, + "logits/chosen": -0.032470703125, + "logits/rejected": -0.1484375, + "logps/chosen": -390.0, + "logps/rejected": -312.0, + "loss": 0.36, + "rewards/accuracies": 0.828125, + "rewards/chosen": 2.546875, + "rewards/margins": 5.09375, + "rewards/rejected": -2.546875, + "step": 38 + }, + { + "epoch": 0.04785276073619632, + "grad_norm": 36.01668225984001, + "learning_rate": 1.4913223140495867e-06, + "logits/chosen": -0.142578125, + "logits/rejected": -0.2890625, + "logps/chosen": -418.0, + "logps/rejected": -318.0, + "loss": 0.3347, + "rewards/accuracies": 0.8359375, + "rewards/chosen": 2.296875, + "rewards/margins": 5.21875, + "rewards/rejected": -2.921875, + "step": 39 + }, + { + "epoch": 0.049079754601226995, + "grad_norm": 30.447104323671855, + "learning_rate": 1.490702479338843e-06, + "logits/chosen": -0.0712890625, + "logits/rejected": -0.1865234375, + "logps/chosen": -370.0, + "logps/rejected": -296.0, + "loss": 0.2729, + "rewards/accuracies": 0.8203125, + "rewards/chosen": 2.09375, + "rewards/margins": 4.90625, + "rewards/rejected": -2.796875, + "step": 40 + }, + { + "epoch": 0.05030674846625767, + "grad_norm": 36.96390592719149, + "learning_rate": 1.4900826446280991e-06, + "logits/chosen": -0.0751953125, + "logits/rejected": -0.2490234375, + "logps/chosen": -374.0, + "logps/rejected": -286.0, + "loss": 0.3822, + "rewards/accuracies": 0.8046875, + "rewards/chosen": 1.875, + "rewards/margins": 4.5625, + "rewards/rejected": -2.671875, + "step": 41 + }, + { + "epoch": 0.051533742331288344, + "grad_norm": 33.41932525699284, + "learning_rate": 1.4894628099173555e-06, + "logits/chosen": -0.022705078125, + "logits/rejected": -0.1650390625, + "logps/chosen": -360.0, + "logps/rejected": -294.0, + "loss": 0.3014, + "rewards/accuracies": 0.84375, + "rewards/chosen": 1.75, + "rewards/margins": 4.3125, + "rewards/rejected": -2.5625, + "step": 42 + }, + { + "epoch": 0.05276073619631902, + "grad_norm": 38.41247549359973, + "learning_rate": 1.4888429752066116e-06, + "logits/chosen": -0.09033203125, + "logits/rejected": -0.248046875, + "logps/chosen": -412.0, + "logps/rejected": -336.0, + "loss": 0.2954, + "rewards/accuracies": 0.84375, + "rewards/chosen": 2.09375, + "rewards/margins": 5.09375, + "rewards/rejected": -3.0, + "step": 43 + }, + { + "epoch": 0.053987730061349694, + "grad_norm": 43.07116536752478, + "learning_rate": 1.4882231404958678e-06, + "logits/chosen": -0.1484375, + "logits/rejected": -0.2490234375, + "logps/chosen": -384.0, + "logps/rejected": -304.0, + "loss": 0.3938, + "rewards/accuracies": 0.8046875, + "rewards/chosen": 1.578125, + "rewards/margins": 4.53125, + "rewards/rejected": -2.96875, + "step": 44 + }, + { + "epoch": 0.05521472392638037, + "grad_norm": 39.15973652461684, + "learning_rate": 1.4876033057851241e-06, + "logits/chosen": -0.2197265625, + "logits/rejected": -0.294921875, + "logps/chosen": -350.0, + "logps/rejected": -300.0, + "loss": 0.3339, + "rewards/accuracies": 0.796875, + "rewards/chosen": 1.828125, + "rewards/margins": 4.9375, + "rewards/rejected": -3.109375, + "step": 45 + }, + { + "epoch": 0.05644171779141104, + "grad_norm": 37.261369579802675, + "learning_rate": 1.4869834710743803e-06, + "logits/chosen": -0.09716796875, + "logits/rejected": -0.259765625, + "logps/chosen": -416.0, + "logps/rejected": -302.0, + "loss": 0.2861, + "rewards/accuracies": 0.828125, + "rewards/chosen": 2.46875, + "rewards/margins": 5.375, + "rewards/rejected": -2.890625, + "step": 46 + }, + { + "epoch": 0.05766871165644172, + "grad_norm": 54.78531831058221, + "learning_rate": 1.4863636363636364e-06, + "logits/chosen": -0.15625, + "logits/rejected": -0.216796875, + "logps/chosen": -390.0, + "logps/rejected": -372.0, + "loss": 0.5073, + "rewards/accuracies": 0.7109375, + "rewards/chosen": 1.5234375, + "rewards/margins": 4.1875, + "rewards/rejected": -2.6875, + "step": 47 + }, + { + "epoch": 0.05889570552147239, + "grad_norm": 43.69804279742604, + "learning_rate": 1.4857438016528926e-06, + "logits/chosen": -0.1494140625, + "logits/rejected": -0.26171875, + "logps/chosen": -352.0, + "logps/rejected": -288.0, + "loss": 0.4242, + "rewards/accuracies": 0.7734375, + "rewards/chosen": 1.1015625, + "rewards/margins": 3.84375, + "rewards/rejected": -2.734375, + "step": 48 + }, + { + "epoch": 0.06012269938650307, + "grad_norm": 34.89468290879934, + "learning_rate": 1.4851239669421487e-06, + "logits/chosen": -0.146484375, + "logits/rejected": -0.328125, + "logps/chosen": -400.0, + "logps/rejected": -300.0, + "loss": 0.2721, + "rewards/accuracies": 0.859375, + "rewards/chosen": 2.25, + "rewards/margins": 5.46875, + "rewards/rejected": -3.21875, + "step": 49 + }, + { + "epoch": 0.06134969325153374, + "grad_norm": 54.15414768657872, + "learning_rate": 1.484504132231405e-06, + "logits/chosen": -0.14453125, + "logits/rejected": -0.296875, + "logps/chosen": -404.0, + "logps/rejected": -324.0, + "loss": 0.4387, + "rewards/accuracies": 0.765625, + "rewards/chosen": 1.515625, + "rewards/margins": 4.6875, + "rewards/rejected": -3.171875, + "step": 50 + }, + { + "epoch": 0.06257668711656442, + "grad_norm": 34.25072100129564, + "learning_rate": 1.4838842975206612e-06, + "logits/chosen": -0.08251953125, + "logits/rejected": -0.2177734375, + "logps/chosen": -440.0, + "logps/rejected": -364.0, + "loss": 0.2872, + "rewards/accuracies": 0.8203125, + "rewards/chosen": 1.984375, + "rewards/margins": 5.09375, + "rewards/rejected": -3.125, + "step": 51 + }, + { + "epoch": 0.0638036809815951, + "grad_norm": 38.20166040651749, + "learning_rate": 1.4832644628099173e-06, + "logits/chosen": -0.10693359375, + "logits/rejected": -0.265625, + "logps/chosen": -396.0, + "logps/rejected": -338.0, + "loss": 0.3424, + "rewards/accuracies": 0.78125, + "rewards/chosen": 1.65625, + "rewards/margins": 4.84375, + "rewards/rejected": -3.171875, + "step": 52 + }, + { + "epoch": 0.06503067484662577, + "grad_norm": 37.089420431960214, + "learning_rate": 1.4826446280991737e-06, + "logits/chosen": -0.055908203125, + "logits/rejected": -0.189453125, + "logps/chosen": -398.0, + "logps/rejected": -326.0, + "loss": 0.3026, + "rewards/accuracies": 0.8359375, + "rewards/chosen": 1.953125, + "rewards/margins": 4.71875, + "rewards/rejected": -2.75, + "step": 53 + }, + { + "epoch": 0.06625766871165645, + "grad_norm": 35.67709547242796, + "learning_rate": 1.4820247933884298e-06, + "logits/chosen": 0.045654296875, + "logits/rejected": -0.1162109375, + "logps/chosen": -336.0, + "logps/rejected": -266.0, + "loss": 0.2813, + "rewards/accuracies": 0.8671875, + "rewards/chosen": 1.3046875, + "rewards/margins": 4.625, + "rewards/rejected": -3.296875, + "step": 54 + }, + { + "epoch": 0.06748466257668712, + "grad_norm": 40.58362122970572, + "learning_rate": 1.481404958677686e-06, + "logits/chosen": -0.068359375, + "logits/rejected": -0.2060546875, + "logps/chosen": -376.0, + "logps/rejected": -302.0, + "loss": 0.3515, + "rewards/accuracies": 0.828125, + "rewards/chosen": 1.8046875, + "rewards/margins": 5.03125, + "rewards/rejected": -3.234375, + "step": 55 + }, + { + "epoch": 0.0687116564417178, + "grad_norm": 43.576103099929334, + "learning_rate": 1.4807851239669423e-06, + "logits/chosen": -0.034912109375, + "logits/rejected": -0.1787109375, + "logps/chosen": -384.0, + "logps/rejected": -334.0, + "loss": 0.4139, + "rewards/accuracies": 0.78125, + "rewards/chosen": 2.015625, + "rewards/margins": 4.8125, + "rewards/rejected": -2.796875, + "step": 56 + }, + { + "epoch": 0.06993865030674846, + "grad_norm": 51.562196128944976, + "learning_rate": 1.4801652892561983e-06, + "logits/chosen": -0.042724609375, + "logits/rejected": -0.0966796875, + "logps/chosen": -352.0, + "logps/rejected": -326.0, + "loss": 0.4667, + "rewards/accuracies": 0.78125, + "rewards/chosen": 1.6015625, + "rewards/margins": 4.59375, + "rewards/rejected": -2.984375, + "step": 57 + }, + { + "epoch": 0.07116564417177915, + "grad_norm": 49.38086728368548, + "learning_rate": 1.4795454545454546e-06, + "logits/chosen": -0.046630859375, + "logits/rejected": -0.10009765625, + "logps/chosen": -376.0, + "logps/rejected": -334.0, + "loss": 0.3869, + "rewards/accuracies": 0.765625, + "rewards/chosen": 1.640625, + "rewards/margins": 4.875, + "rewards/rejected": -3.234375, + "step": 58 + }, + { + "epoch": 0.07239263803680981, + "grad_norm": 43.27222220894176, + "learning_rate": 1.4789256198347107e-06, + "logits/chosen": 0.0240478515625, + "logits/rejected": -0.051025390625, + "logps/chosen": -378.0, + "logps/rejected": -302.0, + "loss": 0.3983, + "rewards/accuracies": 0.765625, + "rewards/chosen": 1.3125, + "rewards/margins": 4.25, + "rewards/rejected": -2.953125, + "step": 59 + }, + { + "epoch": 0.0736196319018405, + "grad_norm": 39.673404465674906, + "learning_rate": 1.4783057851239669e-06, + "logits/chosen": -0.049072265625, + "logits/rejected": -0.1669921875, + "logps/chosen": -386.0, + "logps/rejected": -294.0, + "loss": 0.3927, + "rewards/accuracies": 0.8046875, + "rewards/chosen": 0.99609375, + "rewards/margins": 4.1875, + "rewards/rejected": -3.1875, + "step": 60 + }, + { + "epoch": 0.07484662576687116, + "grad_norm": 40.98812217259802, + "learning_rate": 1.4776859504132232e-06, + "logits/chosen": -0.064453125, + "logits/rejected": -0.173828125, + "logps/chosen": -418.0, + "logps/rejected": -328.0, + "loss": 0.3496, + "rewards/accuracies": 0.8203125, + "rewards/chosen": 1.890625, + "rewards/margins": 5.65625, + "rewards/rejected": -3.765625, + "step": 61 + }, + { + "epoch": 0.07607361963190185, + "grad_norm": 40.09146702222396, + "learning_rate": 1.4770661157024794e-06, + "logits/chosen": -0.051513671875, + "logits/rejected": -0.2275390625, + "logps/chosen": -426.0, + "logps/rejected": -316.0, + "loss": 0.3416, + "rewards/accuracies": 0.7890625, + "rewards/chosen": 2.546875, + "rewards/margins": 6.1875, + "rewards/rejected": -3.640625, + "step": 62 + }, + { + "epoch": 0.07730061349693251, + "grad_norm": 46.30805404508898, + "learning_rate": 1.4764462809917355e-06, + "logits/chosen": 0.031982421875, + "logits/rejected": -0.11767578125, + "logps/chosen": -418.0, + "logps/rejected": -314.0, + "loss": 0.3323, + "rewards/accuracies": 0.8203125, + "rewards/chosen": 1.515625, + "rewards/margins": 4.84375, + "rewards/rejected": -3.328125, + "step": 63 + }, + { + "epoch": 0.0785276073619632, + "grad_norm": 47.96974437982722, + "learning_rate": 1.4758264462809919e-06, + "logits/chosen": -0.123046875, + "logits/rejected": -0.27734375, + "logps/chosen": -402.0, + "logps/rejected": -336.0, + "loss": 0.4046, + "rewards/accuracies": 0.78125, + "rewards/chosen": 2.046875, + "rewards/margins": 5.46875, + "rewards/rejected": -3.4375, + "step": 64 + }, + { + "epoch": 0.07975460122699386, + "grad_norm": 35.57381547040094, + "learning_rate": 1.475206611570248e-06, + "logits/chosen": -0.18359375, + "logits/rejected": -0.25390625, + "logps/chosen": -390.0, + "logps/rejected": -310.0, + "loss": 0.2899, + "rewards/accuracies": 0.84375, + "rewards/chosen": 2.046875, + "rewards/margins": 5.53125, + "rewards/rejected": -3.484375, + "step": 65 + }, + { + "epoch": 0.08098159509202454, + "grad_norm": 33.414611619044024, + "learning_rate": 1.4745867768595042e-06, + "logits/chosen": -0.162109375, + "logits/rejected": -0.287109375, + "logps/chosen": -314.0, + "logps/rejected": -270.0, + "loss": 0.2704, + "rewards/accuracies": 0.8515625, + "rewards/chosen": 2.328125, + "rewards/margins": 6.125, + "rewards/rejected": -3.78125, + "step": 66 + }, + { + "epoch": 0.08220858895705521, + "grad_norm": 42.668537708095855, + "learning_rate": 1.4739669421487603e-06, + "logits/chosen": -0.068359375, + "logits/rejected": -0.1767578125, + "logps/chosen": -332.0, + "logps/rejected": -302.0, + "loss": 0.3919, + "rewards/accuracies": 0.8046875, + "rewards/chosen": 1.2578125, + "rewards/margins": 4.84375, + "rewards/rejected": -3.578125, + "step": 67 + }, + { + "epoch": 0.0834355828220859, + "grad_norm": 40.70660622767979, + "learning_rate": 1.4733471074380164e-06, + "logits/chosen": -0.1455078125, + "logits/rejected": -0.193359375, + "logps/chosen": -358.0, + "logps/rejected": -338.0, + "loss": 0.3949, + "rewards/accuracies": 0.7890625, + "rewards/chosen": 1.625, + "rewards/margins": 4.96875, + "rewards/rejected": -3.359375, + "step": 68 + }, + { + "epoch": 0.08466257668711656, + "grad_norm": 42.29365611407333, + "learning_rate": 1.4727272727272728e-06, + "logits/chosen": -0.2216796875, + "logits/rejected": -0.36328125, + "logps/chosen": -406.0, + "logps/rejected": -290.0, + "loss": 0.3163, + "rewards/accuracies": 0.8515625, + "rewards/chosen": 2.484375, + "rewards/margins": 6.0, + "rewards/rejected": -3.515625, + "step": 69 + }, + { + "epoch": 0.08588957055214724, + "grad_norm": 45.16433653526153, + "learning_rate": 1.472107438016529e-06, + "logits/chosen": -0.08447265625, + "logits/rejected": -0.267578125, + "logps/chosen": -406.0, + "logps/rejected": -346.0, + "loss": 0.3817, + "rewards/accuracies": 0.7890625, + "rewards/chosen": 2.515625, + "rewards/margins": 5.6875, + "rewards/rejected": -3.15625, + "step": 70 + }, + { + "epoch": 0.08711656441717791, + "grad_norm": 37.819952615726415, + "learning_rate": 1.4714876033057853e-06, + "logits/chosen": -0.0341796875, + "logits/rejected": -0.1005859375, + "logps/chosen": -344.0, + "logps/rejected": -286.0, + "loss": 0.3666, + "rewards/accuracies": 0.828125, + "rewards/chosen": 1.46875, + "rewards/margins": 4.625, + "rewards/rejected": -3.15625, + "step": 71 + }, + { + "epoch": 0.08834355828220859, + "grad_norm": 41.96140458025261, + "learning_rate": 1.4708677685950414e-06, + "logits/chosen": -0.19921875, + "logits/rejected": -0.33203125, + "logps/chosen": -428.0, + "logps/rejected": -354.0, + "loss": 0.3187, + "rewards/accuracies": 0.8359375, + "rewards/chosen": 2.125, + "rewards/margins": 5.1875, + "rewards/rejected": -3.0625, + "step": 72 + }, + { + "epoch": 0.08957055214723926, + "grad_norm": 45.003627256411235, + "learning_rate": 1.4702479338842976e-06, + "logits/chosen": -0.14453125, + "logits/rejected": -0.220703125, + "logps/chosen": -346.0, + "logps/rejected": -308.0, + "loss": 0.4099, + "rewards/accuracies": 0.78125, + "rewards/chosen": 1.6171875, + "rewards/margins": 5.25, + "rewards/rejected": -3.625, + "step": 73 + }, + { + "epoch": 0.09079754601226994, + "grad_norm": 37.795821367857755, + "learning_rate": 1.4696280991735537e-06, + "logits/chosen": -0.1435546875, + "logits/rejected": -0.212890625, + "logps/chosen": -352.0, + "logps/rejected": -338.0, + "loss": 0.2989, + "rewards/accuracies": 0.8515625, + "rewards/chosen": 1.4296875, + "rewards/margins": 5.0625, + "rewards/rejected": -3.625, + "step": 74 + }, + { + "epoch": 0.09202453987730061, + "grad_norm": 38.6325533255864, + "learning_rate": 1.4690082644628099e-06, + "logits/chosen": -0.1083984375, + "logits/rejected": -0.2578125, + "logps/chosen": -376.0, + "logps/rejected": -326.0, + "loss": 0.3138, + "rewards/accuracies": 0.8125, + "rewards/chosen": 2.265625, + "rewards/margins": 5.9375, + "rewards/rejected": -3.671875, + "step": 75 + }, + { + "epoch": 0.09325153374233129, + "grad_norm": 41.500214372239896, + "learning_rate": 1.4683884297520662e-06, + "logits/chosen": -0.11376953125, + "logits/rejected": -0.283203125, + "logps/chosen": -394.0, + "logps/rejected": -296.0, + "loss": 0.3196, + "rewards/accuracies": 0.8203125, + "rewards/chosen": 2.28125, + "rewards/margins": 5.625, + "rewards/rejected": -3.359375, + "step": 76 + }, + { + "epoch": 0.09447852760736196, + "grad_norm": 37.661301568651126, + "learning_rate": 1.4677685950413223e-06, + "logits/chosen": -0.1337890625, + "logits/rejected": -0.267578125, + "logps/chosen": -350.0, + "logps/rejected": -308.0, + "loss": 0.3298, + "rewards/accuracies": 0.8125, + "rewards/chosen": 1.4453125, + "rewards/margins": 5.5, + "rewards/rejected": -4.0625, + "step": 77 + }, + { + "epoch": 0.09570552147239264, + "grad_norm": 37.401749661903814, + "learning_rate": 1.4671487603305785e-06, + "logits/chosen": -0.08837890625, + "logits/rejected": -0.2412109375, + "logps/chosen": -414.0, + "logps/rejected": -324.0, + "loss": 0.3295, + "rewards/accuracies": 0.8125, + "rewards/chosen": 2.015625, + "rewards/margins": 5.90625, + "rewards/rejected": -3.90625, + "step": 78 + }, + { + "epoch": 0.09693251533742331, + "grad_norm": 41.34926796893681, + "learning_rate": 1.4665289256198348e-06, + "logits/chosen": -0.078125, + "logits/rejected": -0.2109375, + "logps/chosen": -354.0, + "logps/rejected": -300.0, + "loss": 0.3855, + "rewards/accuracies": 0.8046875, + "rewards/chosen": 1.25, + "rewards/margins": 5.53125, + "rewards/rejected": -4.28125, + "step": 79 + }, + { + "epoch": 0.09815950920245399, + "grad_norm": 40.52807889159641, + "learning_rate": 1.465909090909091e-06, + "logits/chosen": -0.1396484375, + "logits/rejected": -0.33984375, + "logps/chosen": -404.0, + "logps/rejected": -340.0, + "loss": 0.3387, + "rewards/accuracies": 0.8515625, + "rewards/chosen": 2.484375, + "rewards/margins": 7.21875, + "rewards/rejected": -4.75, + "step": 80 + }, + { + "epoch": 0.09938650306748466, + "grad_norm": 35.605812619698284, + "learning_rate": 1.4652892561983471e-06, + "logits/chosen": -0.208984375, + "logits/rejected": -0.365234375, + "logps/chosen": -436.0, + "logps/rejected": -356.0, + "loss": 0.3191, + "rewards/accuracies": 0.828125, + "rewards/chosen": 1.8828125, + "rewards/margins": 6.65625, + "rewards/rejected": -4.78125, + "step": 81 + }, + { + "epoch": 0.10061349693251534, + "grad_norm": 47.4617133344908, + "learning_rate": 1.4646694214876035e-06, + "logits/chosen": -0.1171875, + "logits/rejected": -0.228515625, + "logps/chosen": -430.0, + "logps/rejected": -348.0, + "loss": 0.3954, + "rewards/accuracies": 0.8203125, + "rewards/chosen": 1.03125, + "rewards/margins": 5.875, + "rewards/rejected": -4.84375, + "step": 82 + }, + { + "epoch": 0.10184049079754601, + "grad_norm": 42.72902928212931, + "learning_rate": 1.4640495867768594e-06, + "logits/chosen": -0.1025390625, + "logits/rejected": -0.255859375, + "logps/chosen": -352.0, + "logps/rejected": -320.0, + "loss": 0.3735, + "rewards/accuracies": 0.8125, + "rewards/chosen": 1.109375, + "rewards/margins": 5.84375, + "rewards/rejected": -4.71875, + "step": 83 + }, + { + "epoch": 0.10306748466257669, + "grad_norm": 59.351571106495314, + "learning_rate": 1.4634297520661158e-06, + "logits/chosen": -0.0869140625, + "logits/rejected": -0.310546875, + "logps/chosen": -448.0, + "logps/rejected": -360.0, + "loss": 0.2887, + "rewards/accuracies": 0.859375, + "rewards/chosen": 1.75, + "rewards/margins": 6.8125, + "rewards/rejected": -5.0625, + "step": 84 + }, + { + "epoch": 0.10429447852760736, + "grad_norm": 40.67504361329859, + "learning_rate": 1.462809917355372e-06, + "logits/chosen": -0.189453125, + "logits/rejected": -0.3359375, + "logps/chosen": -376.0, + "logps/rejected": -338.0, + "loss": 0.2933, + "rewards/accuracies": 0.875, + "rewards/chosen": 1.03125, + "rewards/margins": 6.03125, + "rewards/rejected": -4.96875, + "step": 85 + }, + { + "epoch": 0.10552147239263804, + "grad_norm": 44.45491286199163, + "learning_rate": 1.462190082644628e-06, + "logits/chosen": -0.12109375, + "logits/rejected": -0.224609375, + "logps/chosen": -424.0, + "logps/rejected": -314.0, + "loss": 0.4003, + "rewards/accuracies": 0.78125, + "rewards/chosen": 1.625, + "rewards/margins": 5.875, + "rewards/rejected": -4.25, + "step": 86 + }, + { + "epoch": 0.1067484662576687, + "grad_norm": 37.230987939956755, + "learning_rate": 1.4615702479338844e-06, + "logits/chosen": -0.1728515625, + "logits/rejected": -0.3046875, + "logps/chosen": -354.0, + "logps/rejected": -274.0, + "loss": 0.3433, + "rewards/accuracies": 0.8203125, + "rewards/chosen": 0.6484375, + "rewards/margins": 5.65625, + "rewards/rejected": -5.0, + "step": 87 + }, + { + "epoch": 0.10797546012269939, + "grad_norm": 35.435756690174514, + "learning_rate": 1.4609504132231405e-06, + "logits/chosen": -0.142578125, + "logits/rejected": -0.302734375, + "logps/chosen": -362.0, + "logps/rejected": -286.0, + "loss": 0.2796, + "rewards/accuracies": 0.859375, + "rewards/chosen": 1.2890625, + "rewards/margins": 5.78125, + "rewards/rejected": -4.5, + "step": 88 + }, + { + "epoch": 0.10920245398773006, + "grad_norm": 39.218068339324205, + "learning_rate": 1.4603305785123967e-06, + "logits/chosen": -0.1708984375, + "logits/rejected": -0.2734375, + "logps/chosen": -392.0, + "logps/rejected": -340.0, + "loss": 0.3661, + "rewards/accuracies": 0.8046875, + "rewards/chosen": 1.609375, + "rewards/margins": 5.875, + "rewards/rejected": -4.28125, + "step": 89 + }, + { + "epoch": 0.11042944785276074, + "grad_norm": 35.86351693081174, + "learning_rate": 1.459710743801653e-06, + "logits/chosen": -0.2021484375, + "logits/rejected": -0.375, + "logps/chosen": -378.0, + "logps/rejected": -334.0, + "loss": 0.3305, + "rewards/accuracies": 0.828125, + "rewards/chosen": 1.0625, + "rewards/margins": 5.6875, + "rewards/rejected": -4.625, + "step": 90 + }, + { + "epoch": 0.1116564417177914, + "grad_norm": 39.260039435042906, + "learning_rate": 1.4590909090909092e-06, + "logits/chosen": -0.028076171875, + "logits/rejected": -0.28125, + "logps/chosen": -392.0, + "logps/rejected": -308.0, + "loss": 0.3104, + "rewards/accuracies": 0.8203125, + "rewards/chosen": 1.5234375, + "rewards/margins": 5.9375, + "rewards/rejected": -4.40625, + "step": 91 + }, + { + "epoch": 0.11288343558282209, + "grad_norm": 35.99771043781269, + "learning_rate": 1.4584710743801653e-06, + "logits/chosen": -0.1259765625, + "logits/rejected": -0.236328125, + "logps/chosen": -390.0, + "logps/rejected": -316.0, + "loss": 0.304, + "rewards/accuracies": 0.84375, + "rewards/chosen": 1.484375, + "rewards/margins": 6.3125, + "rewards/rejected": -4.8125, + "step": 92 + }, + { + "epoch": 0.11411042944785275, + "grad_norm": 38.41867053252873, + "learning_rate": 1.4578512396694215e-06, + "logits/chosen": -0.15234375, + "logits/rejected": -0.36328125, + "logps/chosen": -400.0, + "logps/rejected": -310.0, + "loss": 0.3311, + "rewards/accuracies": 0.8515625, + "rewards/chosen": 2.328125, + "rewards/margins": 6.625, + "rewards/rejected": -4.3125, + "step": 93 + }, + { + "epoch": 0.11533742331288344, + "grad_norm": 37.099757310606805, + "learning_rate": 1.4572314049586776e-06, + "logits/chosen": -0.171875, + "logits/rejected": -0.390625, + "logps/chosen": -426.0, + "logps/rejected": -354.0, + "loss": 0.2709, + "rewards/accuracies": 0.84375, + "rewards/chosen": 2.765625, + "rewards/margins": 7.09375, + "rewards/rejected": -4.3125, + "step": 94 + }, + { + "epoch": 0.1165644171779141, + "grad_norm": 36.042450420967036, + "learning_rate": 1.456611570247934e-06, + "logits/chosen": -0.2490234375, + "logits/rejected": -0.392578125, + "logps/chosen": -378.0, + "logps/rejected": -320.0, + "loss": 0.3138, + "rewards/accuracies": 0.859375, + "rewards/chosen": 2.0625, + "rewards/margins": 6.6875, + "rewards/rejected": -4.625, + "step": 95 + }, + { + "epoch": 0.11779141104294479, + "grad_norm": 37.51605196372948, + "learning_rate": 1.45599173553719e-06, + "logits/chosen": -0.2392578125, + "logits/rejected": -0.400390625, + "logps/chosen": -366.0, + "logps/rejected": -320.0, + "loss": 0.2944, + "rewards/accuracies": 0.84375, + "rewards/chosen": 2.625, + "rewards/margins": 7.125, + "rewards/rejected": -4.5, + "step": 96 + }, + { + "epoch": 0.11901840490797547, + "grad_norm": 47.87658843319023, + "learning_rate": 1.4553719008264462e-06, + "logits/chosen": -0.234375, + "logits/rejected": -0.373046875, + "logps/chosen": -422.0, + "logps/rejected": -362.0, + "loss": 0.4367, + "rewards/accuracies": 0.796875, + "rewards/chosen": 2.25, + "rewards/margins": 6.125, + "rewards/rejected": -3.859375, + "step": 97 + }, + { + "epoch": 0.12024539877300613, + "grad_norm": 37.61598969810632, + "learning_rate": 1.4547520661157026e-06, + "logits/chosen": -0.03759765625, + "logits/rejected": -0.236328125, + "logps/chosen": -354.0, + "logps/rejected": -296.0, + "loss": 0.3071, + "rewards/accuracies": 0.84375, + "rewards/chosen": 1.984375, + "rewards/margins": 6.09375, + "rewards/rejected": -4.09375, + "step": 98 + }, + { + "epoch": 0.12147239263803682, + "grad_norm": 35.33546250741774, + "learning_rate": 1.4541322314049587e-06, + "logits/chosen": -0.189453125, + "logits/rejected": -0.287109375, + "logps/chosen": -338.0, + "logps/rejected": -258.0, + "loss": 0.2669, + "rewards/accuracies": 0.8515625, + "rewards/chosen": 1.359375, + "rewards/margins": 5.5, + "rewards/rejected": -4.15625, + "step": 99 + }, + { + "epoch": 0.12269938650306748, + "grad_norm": 32.14334216613253, + "learning_rate": 1.453512396694215e-06, + "logits/chosen": -0.2197265625, + "logits/rejected": -0.404296875, + "logps/chosen": -394.0, + "logps/rejected": -332.0, + "loss": 0.2858, + "rewards/accuracies": 0.8359375, + "rewards/chosen": 1.8359375, + "rewards/margins": 6.46875, + "rewards/rejected": -4.625, + "step": 100 + }, + { + "epoch": 0.12392638036809817, + "grad_norm": 45.60783475518701, + "learning_rate": 1.452892561983471e-06, + "logits/chosen": -0.220703125, + "logits/rejected": -0.400390625, + "logps/chosen": -412.0, + "logps/rejected": -352.0, + "loss": 0.4076, + "rewards/accuracies": 0.796875, + "rewards/chosen": 1.9140625, + "rewards/margins": 6.3125, + "rewards/rejected": -4.375, + "step": 101 + }, + { + "epoch": 0.12515337423312883, + "grad_norm": 39.740396682339004, + "learning_rate": 1.4522727272727272e-06, + "logits/chosen": -0.1865234375, + "logits/rejected": -0.298828125, + "logps/chosen": -392.0, + "logps/rejected": -330.0, + "loss": 0.3534, + "rewards/accuracies": 0.8046875, + "rewards/chosen": 2.078125, + "rewards/margins": 6.0, + "rewards/rejected": -3.9375, + "step": 102 + }, + { + "epoch": 0.1263803680981595, + "grad_norm": 35.16122653572647, + "learning_rate": 1.4516528925619835e-06, + "logits/chosen": -0.17578125, + "logits/rejected": -0.326171875, + "logps/chosen": -350.0, + "logps/rejected": -310.0, + "loss": 0.3153, + "rewards/accuracies": 0.84375, + "rewards/chosen": 1.4765625, + "rewards/margins": 5.75, + "rewards/rejected": -4.28125, + "step": 103 + }, + { + "epoch": 0.1276073619631902, + "grad_norm": 46.63941850097279, + "learning_rate": 1.4510330578512396e-06, + "logits/chosen": -0.091796875, + "logits/rejected": -0.2255859375, + "logps/chosen": -414.0, + "logps/rejected": -348.0, + "loss": 0.381, + "rewards/accuracies": 0.7890625, + "rewards/chosen": 1.6484375, + "rewards/margins": 5.78125, + "rewards/rejected": -4.125, + "step": 104 + }, + { + "epoch": 0.12883435582822086, + "grad_norm": 35.04246201772006, + "learning_rate": 1.450413223140496e-06, + "logits/chosen": -0.181640625, + "logits/rejected": -0.3828125, + "logps/chosen": -374.0, + "logps/rejected": -322.0, + "loss": 0.2864, + "rewards/accuracies": 0.8515625, + "rewards/chosen": 2.03125, + "rewards/margins": 6.8125, + "rewards/rejected": -4.78125, + "step": 105 + }, + { + "epoch": 0.13006134969325153, + "grad_norm": 39.36563074587024, + "learning_rate": 1.4497933884297521e-06, + "logits/chosen": -0.1591796875, + "logits/rejected": -0.3046875, + "logps/chosen": -392.0, + "logps/rejected": -322.0, + "loss": 0.3462, + "rewards/accuracies": 0.7890625, + "rewards/chosen": 1.7890625, + "rewards/margins": 6.03125, + "rewards/rejected": -4.25, + "step": 106 + }, + { + "epoch": 0.1312883435582822, + "grad_norm": 35.07229699022507, + "learning_rate": 1.4491735537190083e-06, + "logits/chosen": -0.1865234375, + "logits/rejected": -0.26953125, + "logps/chosen": -352.0, + "logps/rejected": -312.0, + "loss": 0.2577, + "rewards/accuracies": 0.8671875, + "rewards/chosen": 1.3515625, + "rewards/margins": 5.84375, + "rewards/rejected": -4.5, + "step": 107 + }, + { + "epoch": 0.1325153374233129, + "grad_norm": 28.019513861884207, + "learning_rate": 1.4485537190082646e-06, + "logits/chosen": -0.19921875, + "logits/rejected": -0.408203125, + "logps/chosen": -382.0, + "logps/rejected": -320.0, + "loss": 0.2438, + "rewards/accuracies": 0.8828125, + "rewards/chosen": 1.3046875, + "rewards/margins": 6.5625, + "rewards/rejected": -5.25, + "step": 108 + }, + { + "epoch": 0.13374233128834356, + "grad_norm": 35.40605672013622, + "learning_rate": 1.4479338842975208e-06, + "logits/chosen": -0.18359375, + "logits/rejected": -0.359375, + "logps/chosen": -412.0, + "logps/rejected": -326.0, + "loss": 0.3584, + "rewards/accuracies": 0.8359375, + "rewards/chosen": 2.5625, + "rewards/margins": 7.1875, + "rewards/rejected": -4.625, + "step": 109 + }, + { + "epoch": 0.13496932515337423, + "grad_norm": 38.14439771999889, + "learning_rate": 1.447314049586777e-06, + "logits/chosen": -0.1728515625, + "logits/rejected": -0.310546875, + "logps/chosen": -406.0, + "logps/rejected": -348.0, + "loss": 0.3481, + "rewards/accuracies": 0.8125, + "rewards/chosen": 1.9140625, + "rewards/margins": 6.4375, + "rewards/rejected": -4.53125, + "step": 110 + }, + { + "epoch": 0.1361963190184049, + "grad_norm": 36.779042619503116, + "learning_rate": 1.446694214876033e-06, + "logits/chosen": -0.09716796875, + "logits/rejected": -0.251953125, + "logps/chosen": -334.0, + "logps/rejected": -282.0, + "loss": 0.3363, + "rewards/accuracies": 0.859375, + "rewards/chosen": 1.3671875, + "rewards/margins": 5.4375, + "rewards/rejected": -4.0625, + "step": 111 + }, + { + "epoch": 0.1374233128834356, + "grad_norm": 36.86435500620325, + "learning_rate": 1.4460743801652892e-06, + "logits/chosen": -0.177734375, + "logits/rejected": -0.29296875, + "logps/chosen": -342.0, + "logps/rejected": -292.0, + "loss": 0.343, + "rewards/accuracies": 0.859375, + "rewards/chosen": 1.3046875, + "rewards/margins": 5.75, + "rewards/rejected": -4.4375, + "step": 112 + }, + { + "epoch": 0.13865030674846626, + "grad_norm": 38.139474814179394, + "learning_rate": 1.4454545454545456e-06, + "logits/chosen": -0.0849609375, + "logits/rejected": -0.169921875, + "logps/chosen": -374.0, + "logps/rejected": -312.0, + "loss": 0.3952, + "rewards/accuracies": 0.828125, + "rewards/chosen": 1.859375, + "rewards/margins": 5.75, + "rewards/rejected": -3.90625, + "step": 113 + }, + { + "epoch": 0.13987730061349693, + "grad_norm": 38.81539599706057, + "learning_rate": 1.4448347107438017e-06, + "logits/chosen": -0.1220703125, + "logits/rejected": -0.275390625, + "logps/chosen": -372.0, + "logps/rejected": -316.0, + "loss": 0.3854, + "rewards/accuracies": 0.859375, + "rewards/chosen": 1.5078125, + "rewards/margins": 5.375, + "rewards/rejected": -3.875, + "step": 114 + }, + { + "epoch": 0.1411042944785276, + "grad_norm": 45.65683573404254, + "learning_rate": 1.4442148760330578e-06, + "logits/chosen": -0.1298828125, + "logits/rejected": -0.2490234375, + "logps/chosen": -338.0, + "logps/rejected": -320.0, + "loss": 0.4779, + "rewards/accuracies": 0.75, + "rewards/chosen": 1.046875, + "rewards/margins": 4.71875, + "rewards/rejected": -3.65625, + "step": 115 + }, + { + "epoch": 0.1423312883435583, + "grad_norm": 32.97644733528144, + "learning_rate": 1.4435950413223142e-06, + "logits/chosen": -0.185546875, + "logits/rejected": -0.408203125, + "logps/chosen": -376.0, + "logps/rejected": -324.0, + "loss": 0.2978, + "rewards/accuracies": 0.8359375, + "rewards/chosen": 2.40625, + "rewards/margins": 6.03125, + "rewards/rejected": -3.625, + "step": 116 + }, + { + "epoch": 0.14355828220858896, + "grad_norm": 38.235071091905795, + "learning_rate": 1.4429752066115703e-06, + "logits/chosen": -0.130859375, + "logits/rejected": -0.26953125, + "logps/chosen": -324.0, + "logps/rejected": -270.0, + "loss": 0.4211, + "rewards/accuracies": 0.796875, + "rewards/chosen": 1.3125, + "rewards/margins": 5.53125, + "rewards/rejected": -4.21875, + "step": 117 + }, + { + "epoch": 0.14478527607361963, + "grad_norm": 41.12118912071453, + "learning_rate": 1.4423553719008265e-06, + "logits/chosen": -0.15234375, + "logits/rejected": -0.298828125, + "logps/chosen": -418.0, + "logps/rejected": -360.0, + "loss": 0.3658, + "rewards/accuracies": 0.8046875, + "rewards/chosen": 2.28125, + "rewards/margins": 6.125, + "rewards/rejected": -3.84375, + "step": 118 + }, + { + "epoch": 0.1460122699386503, + "grad_norm": 35.626041219957244, + "learning_rate": 1.4417355371900826e-06, + "logits/chosen": -0.220703125, + "logits/rejected": -0.30859375, + "logps/chosen": -352.0, + "logps/rejected": -324.0, + "loss": 0.3003, + "rewards/accuracies": 0.8125, + "rewards/chosen": 1.6953125, + "rewards/margins": 5.59375, + "rewards/rejected": -3.90625, + "step": 119 + }, + { + "epoch": 0.147239263803681, + "grad_norm": 40.42176001046704, + "learning_rate": 1.4411157024793388e-06, + "logits/chosen": -0.201171875, + "logits/rejected": -0.37890625, + "logps/chosen": -418.0, + "logps/rejected": -354.0, + "loss": 0.3506, + "rewards/accuracies": 0.8359375, + "rewards/chosen": 2.6875, + "rewards/margins": 6.625, + "rewards/rejected": -3.9375, + "step": 120 + }, + { + "epoch": 0.14846625766871166, + "grad_norm": 33.17627892781642, + "learning_rate": 1.4404958677685951e-06, + "logits/chosen": -0.12060546875, + "logits/rejected": -0.27734375, + "logps/chosen": -404.0, + "logps/rejected": -336.0, + "loss": 0.2824, + "rewards/accuracies": 0.8203125, + "rewards/chosen": 2.625, + "rewards/margins": 6.75, + "rewards/rejected": -4.125, + "step": 121 + }, + { + "epoch": 0.14969325153374233, + "grad_norm": 41.361282319998544, + "learning_rate": 1.4398760330578512e-06, + "logits/chosen": -0.130859375, + "logits/rejected": -0.283203125, + "logps/chosen": -378.0, + "logps/rejected": -298.0, + "loss": 0.3454, + "rewards/accuracies": 0.828125, + "rewards/chosen": 1.7421875, + "rewards/margins": 5.75, + "rewards/rejected": -4.03125, + "step": 122 + }, + { + "epoch": 0.150920245398773, + "grad_norm": 39.0679541541933, + "learning_rate": 1.4392561983471074e-06, + "logits/chosen": -0.0191650390625, + "logits/rejected": -0.1884765625, + "logps/chosen": -380.0, + "logps/rejected": -290.0, + "loss": 0.384, + "rewards/accuracies": 0.7890625, + "rewards/chosen": 1.2734375, + "rewards/margins": 5.625, + "rewards/rejected": -4.375, + "step": 123 + }, + { + "epoch": 0.1521472392638037, + "grad_norm": 30.841344702577548, + "learning_rate": 1.4386363636363637e-06, + "logits/chosen": -0.193359375, + "logits/rejected": -0.255859375, + "logps/chosen": -348.0, + "logps/rejected": -294.0, + "loss": 0.2792, + "rewards/accuracies": 0.890625, + "rewards/chosen": 2.203125, + "rewards/margins": 7.0, + "rewards/rejected": -4.8125, + "step": 124 + }, + { + "epoch": 0.15337423312883436, + "grad_norm": 40.57809799655024, + "learning_rate": 1.4380165289256199e-06, + "logits/chosen": -0.259765625, + "logits/rejected": -0.431640625, + "logps/chosen": -434.0, + "logps/rejected": -340.0, + "loss": 0.3257, + "rewards/accuracies": 0.84375, + "rewards/chosen": 1.6015625, + "rewards/margins": 6.625, + "rewards/rejected": -5.0, + "step": 125 + }, + { + "epoch": 0.15460122699386503, + "grad_norm": 49.176910835221456, + "learning_rate": 1.437396694214876e-06, + "logits/chosen": -0.11279296875, + "logits/rejected": -0.30078125, + "logps/chosen": -368.0, + "logps/rejected": -320.0, + "loss": 0.4071, + "rewards/accuracies": 0.8359375, + "rewards/chosen": 0.6328125, + "rewards/margins": 5.59375, + "rewards/rejected": -4.96875, + "step": 126 + }, + { + "epoch": 0.1558282208588957, + "grad_norm": 42.06253723411312, + "learning_rate": 1.4367768595041324e-06, + "logits/chosen": -0.064453125, + "logits/rejected": -0.2431640625, + "logps/chosen": -388.0, + "logps/rejected": -340.0, + "loss": 0.4543, + "rewards/accuracies": 0.7578125, + "rewards/chosen": 1.75, + "rewards/margins": 6.40625, + "rewards/rejected": -4.65625, + "step": 127 + }, + { + "epoch": 0.1570552147239264, + "grad_norm": 49.94695723772631, + "learning_rate": 1.4361570247933883e-06, + "logits/chosen": -0.1044921875, + "logits/rejected": -0.2890625, + "logps/chosen": -374.0, + "logps/rejected": -344.0, + "loss": 0.3851, + "rewards/accuracies": 0.78125, + "rewards/chosen": 1.5625, + "rewards/margins": 6.6875, + "rewards/rejected": -5.09375, + "step": 128 + }, + { + "epoch": 0.15828220858895706, + "grad_norm": 31.69804344527413, + "learning_rate": 1.4355371900826447e-06, + "logits/chosen": -0.00439453125, + "logits/rejected": -0.1279296875, + "logps/chosen": -342.0, + "logps/rejected": -302.0, + "loss": 0.3139, + "rewards/accuracies": 0.8359375, + "rewards/chosen": 0.71484375, + "rewards/margins": 5.5, + "rewards/rejected": -4.78125, + "step": 129 + }, + { + "epoch": 0.15950920245398773, + "grad_norm": 39.02664413781982, + "learning_rate": 1.4349173553719008e-06, + "logits/chosen": -0.11328125, + "logits/rejected": -0.251953125, + "logps/chosen": -388.0, + "logps/rejected": -324.0, + "loss": 0.3918, + "rewards/accuracies": 0.75, + "rewards/chosen": 0.66015625, + "rewards/margins": 6.03125, + "rewards/rejected": -5.375, + "step": 130 + }, + { + "epoch": 0.1607361963190184, + "grad_norm": 33.65136979573152, + "learning_rate": 1.434297520661157e-06, + "logits/chosen": -0.2255859375, + "logits/rejected": -0.361328125, + "logps/chosen": -396.0, + "logps/rejected": -356.0, + "loss": 0.2429, + "rewards/accuracies": 0.84375, + "rewards/chosen": 1.3671875, + "rewards/margins": 6.96875, + "rewards/rejected": -5.59375, + "step": 131 + }, + { + "epoch": 0.1619631901840491, + "grad_norm": 40.70444309722836, + "learning_rate": 1.4336776859504133e-06, + "logits/chosen": -0.1845703125, + "logits/rejected": -0.3671875, + "logps/chosen": -392.0, + "logps/rejected": -336.0, + "loss": 0.3917, + "rewards/accuracies": 0.796875, + "rewards/chosen": 1.1796875, + "rewards/margins": 6.34375, + "rewards/rejected": -5.15625, + "step": 132 + }, + { + "epoch": 0.16319018404907976, + "grad_norm": 38.39829631153826, + "learning_rate": 1.4330578512396694e-06, + "logits/chosen": -0.220703125, + "logits/rejected": -0.375, + "logps/chosen": -388.0, + "logps/rejected": -342.0, + "loss": 0.3065, + "rewards/accuracies": 0.84375, + "rewards/chosen": 0.94921875, + "rewards/margins": 6.40625, + "rewards/rejected": -5.46875, + "step": 133 + }, + { + "epoch": 0.16441717791411042, + "grad_norm": 54.071557291057594, + "learning_rate": 1.4324380165289258e-06, + "logits/chosen": -0.1416015625, + "logits/rejected": -0.34765625, + "logps/chosen": -458.0, + "logps/rejected": -374.0, + "loss": 0.4193, + "rewards/accuracies": 0.796875, + "rewards/chosen": 0.81640625, + "rewards/margins": 5.4375, + "rewards/rejected": -4.625, + "step": 134 + }, + { + "epoch": 0.1656441717791411, + "grad_norm": 42.09400341898127, + "learning_rate": 1.431818181818182e-06, + "logits/chosen": -0.0140380859375, + "logits/rejected": -0.11767578125, + "logps/chosen": -378.0, + "logps/rejected": -330.0, + "loss": 0.4491, + "rewards/accuracies": 0.8046875, + "rewards/chosen": 0.16015625, + "rewards/margins": 5.21875, + "rewards/rejected": -5.0625, + "step": 135 + }, + { + "epoch": 0.1668711656441718, + "grad_norm": 34.790669543589566, + "learning_rate": 1.431198347107438e-06, + "logits/chosen": -0.259765625, + "logits/rejected": -0.3203125, + "logps/chosen": -392.0, + "logps/rejected": -322.0, + "loss": 0.3098, + "rewards/accuracies": 0.828125, + "rewards/chosen": 0.87890625, + "rewards/margins": 6.34375, + "rewards/rejected": -5.46875, + "step": 136 + }, + { + "epoch": 0.16809815950920245, + "grad_norm": 46.051678091055635, + "learning_rate": 1.4305785123966942e-06, + "logits/chosen": -0.020751953125, + "logits/rejected": -0.1669921875, + "logps/chosen": -398.0, + "logps/rejected": -320.0, + "loss": 0.4495, + "rewards/accuracies": 0.7890625, + "rewards/chosen": 0.609375, + "rewards/margins": 5.21875, + "rewards/rejected": -4.625, + "step": 137 + }, + { + "epoch": 0.16932515337423312, + "grad_norm": 39.60892351140069, + "learning_rate": 1.4299586776859504e-06, + "logits/chosen": -0.10546875, + "logits/rejected": -0.2421875, + "logps/chosen": -406.0, + "logps/rejected": -360.0, + "loss": 0.3662, + "rewards/accuracies": 0.796875, + "rewards/chosen": 0.56640625, + "rewards/margins": 5.59375, + "rewards/rejected": -5.0, + "step": 138 + }, + { + "epoch": 0.1705521472392638, + "grad_norm": 44.067268759137896, + "learning_rate": 1.4293388429752067e-06, + "logits/chosen": -0.10546875, + "logits/rejected": -0.2734375, + "logps/chosen": -418.0, + "logps/rejected": -380.0, + "loss": 0.4261, + "rewards/accuracies": 0.78125, + "rewards/chosen": 1.2109375, + "rewards/margins": 6.8125, + "rewards/rejected": -5.59375, + "step": 139 + }, + { + "epoch": 0.17177914110429449, + "grad_norm": 46.60040949853179, + "learning_rate": 1.4287190082644628e-06, + "logits/chosen": -0.07177734375, + "logits/rejected": -0.251953125, + "logps/chosen": -426.0, + "logps/rejected": -376.0, + "loss": 0.4219, + "rewards/accuracies": 0.765625, + "rewards/chosen": 0.9453125, + "rewards/margins": 6.125, + "rewards/rejected": -5.1875, + "step": 140 + }, + { + "epoch": 0.17300613496932515, + "grad_norm": 34.208928830023055, + "learning_rate": 1.428099173553719e-06, + "logits/chosen": -0.05126953125, + "logits/rejected": -0.173828125, + "logps/chosen": -358.0, + "logps/rejected": -338.0, + "loss": 0.3019, + "rewards/accuracies": 0.8515625, + "rewards/chosen": 0.99609375, + "rewards/margins": 6.28125, + "rewards/rejected": -5.28125, + "step": 141 + }, + { + "epoch": 0.17423312883435582, + "grad_norm": 41.96948447879753, + "learning_rate": 1.4274793388429753e-06, + "logits/chosen": -0.0927734375, + "logits/rejected": -0.27734375, + "logps/chosen": -400.0, + "logps/rejected": -354.0, + "loss": 0.4327, + "rewards/accuracies": 0.796875, + "rewards/chosen": 0.98828125, + "rewards/margins": 6.375, + "rewards/rejected": -5.375, + "step": 142 + }, + { + "epoch": 0.1754601226993865, + "grad_norm": 39.736356252739405, + "learning_rate": 1.4268595041322315e-06, + "logits/chosen": -0.12451171875, + "logits/rejected": -0.33984375, + "logps/chosen": -384.0, + "logps/rejected": -328.0, + "loss": 0.3314, + "rewards/accuracies": 0.828125, + "rewards/chosen": 0.64453125, + "rewards/margins": 6.28125, + "rewards/rejected": -5.625, + "step": 143 + }, + { + "epoch": 0.17668711656441718, + "grad_norm": 40.38397393664774, + "learning_rate": 1.4262396694214876e-06, + "logits/chosen": -0.224609375, + "logits/rejected": -0.3984375, + "logps/chosen": -420.0, + "logps/rejected": -356.0, + "loss": 0.3711, + "rewards/accuracies": 0.828125, + "rewards/chosen": 1.5625, + "rewards/margins": 6.75, + "rewards/rejected": -5.1875, + "step": 144 + }, + { + "epoch": 0.17791411042944785, + "grad_norm": 36.78367346174047, + "learning_rate": 1.425619834710744e-06, + "logits/chosen": -0.20703125, + "logits/rejected": -0.349609375, + "logps/chosen": -380.0, + "logps/rejected": -348.0, + "loss": 0.2992, + "rewards/accuracies": 0.859375, + "rewards/chosen": 1.5625, + "rewards/margins": 6.78125, + "rewards/rejected": -5.21875, + "step": 145 + }, + { + "epoch": 0.17914110429447852, + "grad_norm": 30.78038936070514, + "learning_rate": 1.425e-06, + "logits/chosen": -0.2021484375, + "logits/rejected": -0.30859375, + "logps/chosen": -388.0, + "logps/rejected": -334.0, + "loss": 0.2674, + "rewards/accuracies": 0.875, + "rewards/chosen": 0.53125, + "rewards/margins": 6.03125, + "rewards/rejected": -5.5, + "step": 146 + }, + { + "epoch": 0.18036809815950922, + "grad_norm": 33.88526702598482, + "learning_rate": 1.4243801652892563e-06, + "logits/chosen": -0.2001953125, + "logits/rejected": -0.349609375, + "logps/chosen": -388.0, + "logps/rejected": -332.0, + "loss": 0.3151, + "rewards/accuracies": 0.828125, + "rewards/chosen": 1.5625, + "rewards/margins": 6.875, + "rewards/rejected": -5.3125, + "step": 147 + }, + { + "epoch": 0.18159509202453988, + "grad_norm": 49.54475790087749, + "learning_rate": 1.4237603305785124e-06, + "logits/chosen": -0.2197265625, + "logits/rejected": -0.34765625, + "logps/chosen": -416.0, + "logps/rejected": -368.0, + "loss": 0.4146, + "rewards/accuracies": 0.7734375, + "rewards/chosen": 0.89453125, + "rewards/margins": 6.25, + "rewards/rejected": -5.34375, + "step": 148 + }, + { + "epoch": 0.18282208588957055, + "grad_norm": 43.24095049009772, + "learning_rate": 1.4231404958677685e-06, + "logits/chosen": -0.09521484375, + "logits/rejected": -0.244140625, + "logps/chosen": -406.0, + "logps/rejected": -384.0, + "loss": 0.3173, + "rewards/accuracies": 0.859375, + "rewards/chosen": 1.5859375, + "rewards/margins": 6.625, + "rewards/rejected": -5.03125, + "step": 149 + }, + { + "epoch": 0.18404907975460122, + "grad_norm": 37.7299116363035, + "learning_rate": 1.422520661157025e-06, + "logits/chosen": -0.1181640625, + "logits/rejected": -0.291015625, + "logps/chosen": -424.0, + "logps/rejected": -346.0, + "loss": 0.344, + "rewards/accuracies": 0.8359375, + "rewards/chosen": 1.296875, + "rewards/margins": 5.84375, + "rewards/rejected": -4.53125, + "step": 150 + }, + { + "epoch": 0.18527607361963191, + "grad_norm": 36.98087652686202, + "learning_rate": 1.421900826446281e-06, + "logits/chosen": -0.1376953125, + "logits/rejected": -0.263671875, + "logps/chosen": -382.0, + "logps/rejected": -326.0, + "loss": 0.3467, + "rewards/accuracies": 0.8046875, + "rewards/chosen": 1.1796875, + "rewards/margins": 5.8125, + "rewards/rejected": -4.625, + "step": 151 + }, + { + "epoch": 0.18650306748466258, + "grad_norm": 33.96526048608977, + "learning_rate": 1.4212809917355372e-06, + "logits/chosen": -0.2255859375, + "logits/rejected": -0.462890625, + "logps/chosen": -402.0, + "logps/rejected": -316.0, + "loss": 0.322, + "rewards/accuracies": 0.8515625, + "rewards/chosen": 0.83984375, + "rewards/margins": 6.5625, + "rewards/rejected": -5.71875, + "step": 152 + }, + { + "epoch": 0.18773006134969325, + "grad_norm": 34.59925431133825, + "learning_rate": 1.4206611570247935e-06, + "logits/chosen": -0.1875, + "logits/rejected": -0.298828125, + "logps/chosen": -350.0, + "logps/rejected": -320.0, + "loss": 0.3333, + "rewards/accuracies": 0.84375, + "rewards/chosen": 1.390625, + "rewards/margins": 6.28125, + "rewards/rejected": -4.875, + "step": 153 + }, + { + "epoch": 0.18895705521472392, + "grad_norm": 45.111496339674666, + "learning_rate": 1.4200413223140497e-06, + "logits/chosen": -0.060546875, + "logits/rejected": -0.2431640625, + "logps/chosen": -424.0, + "logps/rejected": -340.0, + "loss": 0.3836, + "rewards/accuracies": 0.7734375, + "rewards/chosen": 1.703125, + "rewards/margins": 6.09375, + "rewards/rejected": -4.375, + "step": 154 + }, + { + "epoch": 0.1901840490797546, + "grad_norm": 28.559326947209748, + "learning_rate": 1.4194214876033058e-06, + "logits/chosen": -0.1748046875, + "logits/rejected": -0.41796875, + "logps/chosen": -382.0, + "logps/rejected": -320.0, + "loss": 0.2127, + "rewards/accuracies": 0.90625, + "rewards/chosen": 2.359375, + "rewards/margins": 7.375, + "rewards/rejected": -5.03125, + "step": 155 + }, + { + "epoch": 0.19141104294478528, + "grad_norm": 41.969901464316806, + "learning_rate": 1.418801652892562e-06, + "logits/chosen": -0.06787109375, + "logits/rejected": -0.1748046875, + "logps/chosen": -370.0, + "logps/rejected": -350.0, + "loss": 0.3849, + "rewards/accuracies": 0.8203125, + "rewards/chosen": 1.1171875, + "rewards/margins": 5.59375, + "rewards/rejected": -4.46875, + "step": 156 + }, + { + "epoch": 0.19263803680981595, + "grad_norm": 35.115688822084685, + "learning_rate": 1.418181818181818e-06, + "logits/chosen": -0.05810546875, + "logits/rejected": -0.1796875, + "logps/chosen": -394.0, + "logps/rejected": -310.0, + "loss": 0.3364, + "rewards/accuracies": 0.828125, + "rewards/chosen": 1.2421875, + "rewards/margins": 5.875, + "rewards/rejected": -4.625, + "step": 157 + }, + { + "epoch": 0.19386503067484662, + "grad_norm": 29.216146857412586, + "learning_rate": 1.4175619834710745e-06, + "logits/chosen": -0.2734375, + "logits/rejected": -0.44921875, + "logps/chosen": -388.0, + "logps/rejected": -324.0, + "loss": 0.2283, + "rewards/accuracies": 0.8515625, + "rewards/chosen": 1.390625, + "rewards/margins": 6.59375, + "rewards/rejected": -5.21875, + "step": 158 + }, + { + "epoch": 0.1950920245398773, + "grad_norm": 39.305920966554595, + "learning_rate": 1.4169421487603306e-06, + "logits/chosen": -0.1630859375, + "logits/rejected": -0.28125, + "logps/chosen": -360.0, + "logps/rejected": -328.0, + "loss": 0.3649, + "rewards/accuracies": 0.84375, + "rewards/chosen": 1.359375, + "rewards/margins": 5.71875, + "rewards/rejected": -4.34375, + "step": 159 + }, + { + "epoch": 0.19631901840490798, + "grad_norm": 26.74081618489141, + "learning_rate": 1.416322314049587e-06, + "logits/chosen": -0.047607421875, + "logits/rejected": -0.26953125, + "logps/chosen": -354.0, + "logps/rejected": -310.0, + "loss": 0.2471, + "rewards/accuracies": 0.890625, + "rewards/chosen": 0.94921875, + "rewards/margins": 6.125, + "rewards/rejected": -5.15625, + "step": 160 + }, + { + "epoch": 0.19754601226993865, + "grad_norm": 32.344289601166764, + "learning_rate": 1.415702479338843e-06, + "logits/chosen": -0.2294921875, + "logits/rejected": -0.4296875, + "logps/chosen": -386.0, + "logps/rejected": -302.0, + "loss": 0.2831, + "rewards/accuracies": 0.8515625, + "rewards/chosen": 1.9453125, + "rewards/margins": 6.6875, + "rewards/rejected": -4.71875, + "step": 161 + }, + { + "epoch": 0.19877300613496932, + "grad_norm": 39.80838608230206, + "learning_rate": 1.4150826446280992e-06, + "logits/chosen": -0.1357421875, + "logits/rejected": -0.2060546875, + "logps/chosen": -350.0, + "logps/rejected": -340.0, + "loss": 0.4018, + "rewards/accuracies": 0.8046875, + "rewards/chosen": 1.4453125, + "rewards/margins": 5.8125, + "rewards/rejected": -4.375, + "step": 162 + }, + { + "epoch": 0.2, + "grad_norm": 38.00229232317159, + "learning_rate": 1.4144628099173554e-06, + "logits/chosen": -0.07373046875, + "logits/rejected": -0.220703125, + "logps/chosen": -392.0, + "logps/rejected": -344.0, + "loss": 0.2926, + "rewards/accuracies": 0.8671875, + "rewards/chosen": 1.421875, + "rewards/margins": 6.25, + "rewards/rejected": -4.8125, + "step": 163 + }, + { + "epoch": 0.20122699386503068, + "grad_norm": 31.02665679742822, + "learning_rate": 1.4138429752066115e-06, + "logits/chosen": -0.236328125, + "logits/rejected": -0.466796875, + "logps/chosen": -392.0, + "logps/rejected": -310.0, + "loss": 0.2813, + "rewards/accuracies": 0.859375, + "rewards/chosen": 2.15625, + "rewards/margins": 7.34375, + "rewards/rejected": -5.1875, + "step": 164 + }, + { + "epoch": 0.20245398773006135, + "grad_norm": 43.699969414714815, + "learning_rate": 1.4132231404958679e-06, + "logits/chosen": -0.021728515625, + "logits/rejected": -0.212890625, + "logps/chosen": -444.0, + "logps/rejected": -326.0, + "loss": 0.377, + "rewards/accuracies": 0.828125, + "rewards/chosen": 1.5078125, + "rewards/margins": 6.3125, + "rewards/rejected": -4.78125, + "step": 165 + }, + { + "epoch": 0.20368098159509201, + "grad_norm": 38.30377721360019, + "learning_rate": 1.412603305785124e-06, + "logits/chosen": -0.09375, + "logits/rejected": -0.314453125, + "logps/chosen": -388.0, + "logps/rejected": -336.0, + "loss": 0.3851, + "rewards/accuracies": 0.828125, + "rewards/chosen": 1.9140625, + "rewards/margins": 6.53125, + "rewards/rejected": -4.625, + "step": 166 + }, + { + "epoch": 0.2049079754601227, + "grad_norm": 38.75636778509963, + "learning_rate": 1.4119834710743801e-06, + "logits/chosen": -0.06591796875, + "logits/rejected": -0.2080078125, + "logps/chosen": -398.0, + "logps/rejected": -318.0, + "loss": 0.3749, + "rewards/accuracies": 0.7890625, + "rewards/chosen": 2.21875, + "rewards/margins": 6.875, + "rewards/rejected": -4.65625, + "step": 167 + }, + { + "epoch": 0.20613496932515338, + "grad_norm": 37.779783793801464, + "learning_rate": 1.4113636363636365e-06, + "logits/chosen": -0.049560546875, + "logits/rejected": -0.1494140625, + "logps/chosen": -348.0, + "logps/rejected": -292.0, + "loss": 0.3337, + "rewards/accuracies": 0.84375, + "rewards/chosen": 0.99609375, + "rewards/margins": 6.46875, + "rewards/rejected": -5.46875, + "step": 168 + }, + { + "epoch": 0.20736196319018405, + "grad_norm": 34.571770475148426, + "learning_rate": 1.4107438016528926e-06, + "logits/chosen": -0.1259765625, + "logits/rejected": -0.27734375, + "logps/chosen": -372.0, + "logps/rejected": -366.0, + "loss": 0.3006, + "rewards/accuracies": 0.8671875, + "rewards/chosen": 1.0078125, + "rewards/margins": 6.25, + "rewards/rejected": -5.25, + "step": 169 + }, + { + "epoch": 0.2085889570552147, + "grad_norm": 42.205138483103795, + "learning_rate": 1.4101239669421488e-06, + "logits/chosen": -0.1123046875, + "logits/rejected": -0.349609375, + "logps/chosen": -450.0, + "logps/rejected": -346.0, + "loss": 0.338, + "rewards/accuracies": 0.8125, + "rewards/chosen": 1.875, + "rewards/margins": 7.34375, + "rewards/rejected": -5.46875, + "step": 170 + }, + { + "epoch": 0.2098159509202454, + "grad_norm": 29.828895289514126, + "learning_rate": 1.4095041322314051e-06, + "logits/chosen": -0.10888671875, + "logits/rejected": -0.279296875, + "logps/chosen": -384.0, + "logps/rejected": -322.0, + "loss": 0.3086, + "rewards/accuracies": 0.859375, + "rewards/chosen": 1.5546875, + "rewards/margins": 7.46875, + "rewards/rejected": -5.90625, + "step": 171 + }, + { + "epoch": 0.21104294478527608, + "grad_norm": 30.980573058674597, + "learning_rate": 1.408884297520661e-06, + "logits/chosen": -0.062255859375, + "logits/rejected": -0.2216796875, + "logps/chosen": -438.0, + "logps/rejected": -354.0, + "loss": 0.2599, + "rewards/accuracies": 0.859375, + "rewards/chosen": 1.84375, + "rewards/margins": 7.84375, + "rewards/rejected": -6.0, + "step": 172 + }, + { + "epoch": 0.21226993865030674, + "grad_norm": 25.22446362010358, + "learning_rate": 1.4082644628099174e-06, + "logits/chosen": -0.123046875, + "logits/rejected": -0.353515625, + "logps/chosen": -428.0, + "logps/rejected": -362.0, + "loss": 0.1997, + "rewards/accuracies": 0.890625, + "rewards/chosen": 1.703125, + "rewards/margins": 8.25, + "rewards/rejected": -6.5625, + "step": 173 + }, + { + "epoch": 0.2134969325153374, + "grad_norm": 43.79333405262764, + "learning_rate": 1.4076446280991736e-06, + "logits/chosen": 0.05126953125, + "logits/rejected": -0.158203125, + "logps/chosen": -384.0, + "logps/rejected": -348.0, + "loss": 0.5204, + "rewards/accuracies": 0.765625, + "rewards/chosen": 0.78125, + "rewards/margins": 6.8125, + "rewards/rejected": -6.03125, + "step": 174 + }, + { + "epoch": 0.2147239263803681, + "grad_norm": 45.27680571467687, + "learning_rate": 1.4070247933884297e-06, + "logits/chosen": -0.10546875, + "logits/rejected": -0.220703125, + "logps/chosen": -402.0, + "logps/rejected": -366.0, + "loss": 0.4454, + "rewards/accuracies": 0.8125, + "rewards/chosen": 1.09375, + "rewards/margins": 7.34375, + "rewards/rejected": -6.25, + "step": 175 + }, + { + "epoch": 0.21595092024539878, + "grad_norm": 35.16824036294685, + "learning_rate": 1.406404958677686e-06, + "logits/chosen": -0.1298828125, + "logits/rejected": -0.30859375, + "logps/chosen": -434.0, + "logps/rejected": -378.0, + "loss": 0.3443, + "rewards/accuracies": 0.796875, + "rewards/chosen": 1.3359375, + "rewards/margins": 7.65625, + "rewards/rejected": -6.3125, + "step": 176 + }, + { + "epoch": 0.21717791411042944, + "grad_norm": 27.30910780814845, + "learning_rate": 1.4057851239669422e-06, + "logits/chosen": -0.08251953125, + "logits/rejected": -0.30859375, + "logps/chosen": -416.0, + "logps/rejected": -356.0, + "loss": 0.2873, + "rewards/accuracies": 0.875, + "rewards/chosen": 2.0625, + "rewards/margins": 8.8125, + "rewards/rejected": -6.71875, + "step": 177 + }, + { + "epoch": 0.2184049079754601, + "grad_norm": 37.84132418939681, + "learning_rate": 1.4051652892561983e-06, + "logits/chosen": -0.07275390625, + "logits/rejected": -0.216796875, + "logps/chosen": -330.0, + "logps/rejected": -318.0, + "loss": 0.4137, + "rewards/accuracies": 0.828125, + "rewards/chosen": 0.1806640625, + "rewards/margins": 6.75, + "rewards/rejected": -6.59375, + "step": 178 + }, + { + "epoch": 0.2196319018404908, + "grad_norm": 32.98852515847511, + "learning_rate": 1.4045454545454547e-06, + "logits/chosen": -0.1064453125, + "logits/rejected": -0.31640625, + "logps/chosen": -372.0, + "logps/rejected": -324.0, + "loss": 0.3252, + "rewards/accuracies": 0.859375, + "rewards/chosen": 0.671875, + "rewards/margins": 6.875, + "rewards/rejected": -6.1875, + "step": 179 + }, + { + "epoch": 0.22085889570552147, + "grad_norm": 38.78908549807395, + "learning_rate": 1.4039256198347108e-06, + "logits/chosen": -0.099609375, + "logits/rejected": -0.21484375, + "logps/chosen": -356.0, + "logps/rejected": -314.0, + "loss": 0.3669, + "rewards/accuracies": 0.828125, + "rewards/chosen": 1.234375, + "rewards/margins": 6.46875, + "rewards/rejected": -5.21875, + "step": 180 + }, + { + "epoch": 0.22208588957055214, + "grad_norm": 35.32530991965649, + "learning_rate": 1.403305785123967e-06, + "logits/chosen": -0.09423828125, + "logits/rejected": -0.2412109375, + "logps/chosen": -410.0, + "logps/rejected": -386.0, + "loss": 0.3373, + "rewards/accuracies": 0.859375, + "rewards/chosen": 1.6171875, + "rewards/margins": 6.875, + "rewards/rejected": -5.25, + "step": 181 + }, + { + "epoch": 0.2233128834355828, + "grad_norm": 36.9736807793252, + "learning_rate": 1.4026859504132231e-06, + "logits/chosen": -0.12890625, + "logits/rejected": -0.2470703125, + "logps/chosen": -370.0, + "logps/rejected": -322.0, + "loss": 0.3661, + "rewards/accuracies": 0.8515625, + "rewards/chosen": 1.109375, + "rewards/margins": 5.875, + "rewards/rejected": -4.75, + "step": 182 + }, + { + "epoch": 0.2245398773006135, + "grad_norm": 40.8878376560248, + "learning_rate": 1.4020661157024793e-06, + "logits/chosen": 0.00762939453125, + "logits/rejected": -0.1669921875, + "logps/chosen": -412.0, + "logps/rejected": -378.0, + "loss": 0.3866, + "rewards/accuracies": 0.8203125, + "rewards/chosen": 0.95703125, + "rewards/margins": 4.90625, + "rewards/rejected": -3.953125, + "step": 183 + }, + { + "epoch": 0.22576687116564417, + "grad_norm": 42.156241310731595, + "learning_rate": 1.4014462809917356e-06, + "logits/chosen": -0.0255126953125, + "logits/rejected": -0.1328125, + "logps/chosen": -390.0, + "logps/rejected": -340.0, + "loss": 0.4166, + "rewards/accuracies": 0.7890625, + "rewards/chosen": 0.87109375, + "rewards/margins": 5.1875, + "rewards/rejected": -4.3125, + "step": 184 + }, + { + "epoch": 0.22699386503067484, + "grad_norm": 44.943258114663124, + "learning_rate": 1.4008264462809917e-06, + "logits/chosen": -0.06787109375, + "logits/rejected": -0.1220703125, + "logps/chosen": -382.0, + "logps/rejected": -350.0, + "loss": 0.4011, + "rewards/accuracies": 0.78125, + "rewards/chosen": 1.9765625, + "rewards/margins": 5.8125, + "rewards/rejected": -3.828125, + "step": 185 + }, + { + "epoch": 0.2282208588957055, + "grad_norm": 33.47267300169336, + "learning_rate": 1.4002066115702479e-06, + "logits/chosen": -0.08203125, + "logits/rejected": -0.1982421875, + "logps/chosen": -336.0, + "logps/rejected": -312.0, + "loss": 0.3163, + "rewards/accuracies": 0.859375, + "rewards/chosen": 1.4453125, + "rewards/margins": 5.78125, + "rewards/rejected": -4.3125, + "step": 186 + }, + { + "epoch": 0.2294478527607362, + "grad_norm": 32.03885108074069, + "learning_rate": 1.3995867768595042e-06, + "logits/chosen": -0.07080078125, + "logits/rejected": -0.2314453125, + "logps/chosen": -370.0, + "logps/rejected": -340.0, + "loss": 0.258, + "rewards/accuracies": 0.8671875, + "rewards/chosen": 1.6328125, + "rewards/margins": 6.4375, + "rewards/rejected": -4.78125, + "step": 187 + }, + { + "epoch": 0.23067484662576687, + "grad_norm": 35.158897439200416, + "learning_rate": 1.3989669421487604e-06, + "logits/chosen": 0.0228271484375, + "logits/rejected": -0.0615234375, + "logps/chosen": -330.0, + "logps/rejected": -292.0, + "loss": 0.3609, + "rewards/accuracies": 0.8046875, + "rewards/chosen": 1.40625, + "rewards/margins": 5.875, + "rewards/rejected": -4.46875, + "step": 188 + }, + { + "epoch": 0.23190184049079754, + "grad_norm": 32.15934523587848, + "learning_rate": 1.3983471074380167e-06, + "logits/chosen": -0.080078125, + "logits/rejected": -0.263671875, + "logps/chosen": -396.0, + "logps/rejected": -322.0, + "loss": 0.2271, + "rewards/accuracies": 0.890625, + "rewards/chosen": 2.046875, + "rewards/margins": 6.78125, + "rewards/rejected": -4.71875, + "step": 189 + }, + { + "epoch": 0.2331288343558282, + "grad_norm": 32.11586724143221, + "learning_rate": 1.3977272727272727e-06, + "logits/chosen": 0.041748046875, + "logits/rejected": -0.10693359375, + "logps/chosen": -340.0, + "logps/rejected": -316.0, + "loss": 0.3271, + "rewards/accuracies": 0.8515625, + "rewards/chosen": 1.4140625, + "rewards/margins": 6.0, + "rewards/rejected": -4.59375, + "step": 190 + }, + { + "epoch": 0.2343558282208589, + "grad_norm": 46.26761231273777, + "learning_rate": 1.3971074380165288e-06, + "logits/chosen": -0.0390625, + "logits/rejected": -0.220703125, + "logps/chosen": -400.0, + "logps/rejected": -306.0, + "loss": 0.3906, + "rewards/accuracies": 0.8046875, + "rewards/chosen": 1.828125, + "rewards/margins": 6.90625, + "rewards/rejected": -5.0625, + "step": 191 + }, + { + "epoch": 0.23558282208588957, + "grad_norm": 39.15535099753523, + "learning_rate": 1.3964876033057852e-06, + "logits/chosen": -0.052490234375, + "logits/rejected": -0.2490234375, + "logps/chosen": -388.0, + "logps/rejected": -302.0, + "loss": 0.3379, + "rewards/accuracies": 0.8046875, + "rewards/chosen": 1.1328125, + "rewards/margins": 5.65625, + "rewards/rejected": -4.5, + "step": 192 + }, + { + "epoch": 0.23680981595092024, + "grad_norm": 31.3365855799527, + "learning_rate": 1.3958677685950413e-06, + "logits/chosen": -0.1767578125, + "logits/rejected": -0.423828125, + "logps/chosen": -410.0, + "logps/rejected": -346.0, + "loss": 0.2485, + "rewards/accuracies": 0.8828125, + "rewards/chosen": 1.7578125, + "rewards/margins": 7.375, + "rewards/rejected": -5.625, + "step": 193 + }, + { + "epoch": 0.23803680981595093, + "grad_norm": 38.5490731428379, + "learning_rate": 1.3952479338842977e-06, + "logits/chosen": -0.21875, + "logits/rejected": -0.326171875, + "logps/chosen": -394.0, + "logps/rejected": -336.0, + "loss": 0.296, + "rewards/accuracies": 0.8203125, + "rewards/chosen": 0.89453125, + "rewards/margins": 6.875, + "rewards/rejected": -5.96875, + "step": 194 + }, + { + "epoch": 0.2392638036809816, + "grad_norm": 35.180225705298604, + "learning_rate": 1.3946280991735538e-06, + "logits/chosen": -0.031982421875, + "logits/rejected": -0.2890625, + "logps/chosen": -408.0, + "logps/rejected": -348.0, + "loss": 0.2727, + "rewards/accuracies": 0.8671875, + "rewards/chosen": 2.046875, + "rewards/margins": 8.1875, + "rewards/rejected": -6.125, + "step": 195 + }, + { + "epoch": 0.24049079754601227, + "grad_norm": 37.22166797602668, + "learning_rate": 1.39400826446281e-06, + "logits/chosen": 0.0157470703125, + "logits/rejected": -0.1845703125, + "logps/chosen": -394.0, + "logps/rejected": -308.0, + "loss": 0.3333, + "rewards/accuracies": 0.8359375, + "rewards/chosen": 1.1171875, + "rewards/margins": 6.65625, + "rewards/rejected": -5.53125, + "step": 196 + }, + { + "epoch": 0.24171779141104294, + "grad_norm": 28.562917988218086, + "learning_rate": 1.3933884297520663e-06, + "logits/chosen": -0.115234375, + "logits/rejected": -0.34765625, + "logps/chosen": -380.0, + "logps/rejected": -334.0, + "loss": 0.3009, + "rewards/accuracies": 0.8828125, + "rewards/chosen": 1.3359375, + "rewards/margins": 7.125, + "rewards/rejected": -5.78125, + "step": 197 + }, + { + "epoch": 0.24294478527607363, + "grad_norm": 36.54466818998776, + "learning_rate": 1.3927685950413224e-06, + "logits/chosen": -0.0174560546875, + "logits/rejected": -0.10205078125, + "logps/chosen": -346.0, + "logps/rejected": -276.0, + "loss": 0.4465, + "rewards/accuracies": 0.8046875, + "rewards/chosen": 0.625, + "rewards/margins": 6.0625, + "rewards/rejected": -5.4375, + "step": 198 + }, + { + "epoch": 0.2441717791411043, + "grad_norm": 25.79094354827764, + "learning_rate": 1.3921487603305786e-06, + "logits/chosen": -0.05615234375, + "logits/rejected": -0.26171875, + "logps/chosen": -434.0, + "logps/rejected": -330.0, + "loss": 0.1969, + "rewards/accuracies": 0.8984375, + "rewards/chosen": 1.7265625, + "rewards/margins": 7.84375, + "rewards/rejected": -6.125, + "step": 199 + }, + { + "epoch": 0.24539877300613497, + "grad_norm": 34.674086597448465, + "learning_rate": 1.3915289256198347e-06, + "logits/chosen": -0.037353515625, + "logits/rejected": -0.25390625, + "logps/chosen": -442.0, + "logps/rejected": -392.0, + "loss": 0.3066, + "rewards/accuracies": 0.8515625, + "rewards/chosen": 1.2734375, + "rewards/margins": 7.4375, + "rewards/rejected": -6.15625, + "step": 200 + }, + { + "epoch": 0.24662576687116564, + "grad_norm": 29.903922318940893, + "learning_rate": 1.3909090909090909e-06, + "logits/chosen": -0.03662109375, + "logits/rejected": -0.259765625, + "logps/chosen": -404.0, + "logps/rejected": -330.0, + "loss": 0.2629, + "rewards/accuracies": 0.8828125, + "rewards/chosen": 1.109375, + "rewards/margins": 7.0625, + "rewards/rejected": -5.9375, + "step": 201 + }, + { + "epoch": 0.24785276073619633, + "grad_norm": 33.833661458600005, + "learning_rate": 1.3902892561983472e-06, + "logits/chosen": -0.02001953125, + "logits/rejected": -0.2333984375, + "logps/chosen": -432.0, + "logps/rejected": -340.0, + "loss": 0.2977, + "rewards/accuracies": 0.8203125, + "rewards/chosen": 1.03125, + "rewards/margins": 6.90625, + "rewards/rejected": -5.875, + "step": 202 + }, + { + "epoch": 0.249079754601227, + "grad_norm": 37.642944458655066, + "learning_rate": 1.3896694214876034e-06, + "logits/chosen": 0.0272216796875, + "logits/rejected": -0.1474609375, + "logps/chosen": -392.0, + "logps/rejected": -354.0, + "loss": 0.482, + "rewards/accuracies": 0.796875, + "rewards/chosen": 0.96875, + "rewards/margins": 6.03125, + "rewards/rejected": -5.0625, + "step": 203 + }, + { + "epoch": 0.25030674846625767, + "grad_norm": 35.44282157848665, + "learning_rate": 1.3890495867768595e-06, + "logits/chosen": 0.0390625, + "logits/rejected": -0.1005859375, + "logps/chosen": -356.0, + "logps/rejected": -290.0, + "loss": 0.4211, + "rewards/accuracies": 0.8359375, + "rewards/chosen": 0.6484375, + "rewards/margins": 6.4375, + "rewards/rejected": -5.78125, + "step": 204 + }, + { + "epoch": 0.25153374233128833, + "grad_norm": 38.48812661842802, + "learning_rate": 1.3884297520661158e-06, + "logits/chosen": 0.08984375, + "logits/rejected": -0.04833984375, + "logps/chosen": -390.0, + "logps/rejected": -316.0, + "loss": 0.4435, + "rewards/accuracies": 0.8046875, + "rewards/chosen": 0.8671875, + "rewards/margins": 6.0, + "rewards/rejected": -5.125, + "step": 205 + }, + { + "epoch": 0.252760736196319, + "grad_norm": 26.640868179250116, + "learning_rate": 1.387809917355372e-06, + "logits/chosen": -0.036376953125, + "logits/rejected": -0.3125, + "logps/chosen": -434.0, + "logps/rejected": -350.0, + "loss": 0.2375, + "rewards/accuracies": 0.9140625, + "rewards/chosen": 1.203125, + "rewards/margins": 7.46875, + "rewards/rejected": -6.28125, + "step": 206 + }, + { + "epoch": 0.25398773006134967, + "grad_norm": 37.72822731879787, + "learning_rate": 1.3871900826446281e-06, + "logits/chosen": -0.16015625, + "logits/rejected": -0.353515625, + "logps/chosen": -434.0, + "logps/rejected": -376.0, + "loss": 0.4045, + "rewards/accuracies": 0.828125, + "rewards/chosen": 1.21875, + "rewards/margins": 6.75, + "rewards/rejected": -5.53125, + "step": 207 + }, + { + "epoch": 0.2552147239263804, + "grad_norm": 39.99822195882057, + "learning_rate": 1.3865702479338843e-06, + "logits/chosen": -0.12255859375, + "logits/rejected": -0.2470703125, + "logps/chosen": -396.0, + "logps/rejected": -340.0, + "loss": 0.4402, + "rewards/accuracies": 0.765625, + "rewards/chosen": 0.83984375, + "rewards/margins": 6.53125, + "rewards/rejected": -5.6875, + "step": 208 + }, + { + "epoch": 0.25644171779141106, + "grad_norm": 31.995520167951547, + "learning_rate": 1.3859504132231404e-06, + "logits/chosen": -0.125, + "logits/rejected": -0.2314453125, + "logps/chosen": -398.0, + "logps/rejected": -352.0, + "loss": 0.3203, + "rewards/accuracies": 0.84375, + "rewards/chosen": 0.76171875, + "rewards/margins": 6.34375, + "rewards/rejected": -5.5625, + "step": 209 + }, + { + "epoch": 0.25766871165644173, + "grad_norm": 42.71588448177596, + "learning_rate": 1.3853305785123968e-06, + "logits/chosen": -0.0732421875, + "logits/rejected": -0.1806640625, + "logps/chosen": -400.0, + "logps/rejected": -382.0, + "loss": 0.4413, + "rewards/accuracies": 0.8203125, + "rewards/chosen": 0.314453125, + "rewards/margins": 5.90625, + "rewards/rejected": -5.59375, + "step": 210 + }, + { + "epoch": 0.2588957055214724, + "grad_norm": 35.38452290000429, + "learning_rate": 1.384710743801653e-06, + "logits/chosen": -0.1806640625, + "logits/rejected": -0.375, + "logps/chosen": -438.0, + "logps/rejected": -362.0, + "loss": 0.3649, + "rewards/accuracies": 0.8125, + "rewards/chosen": 0.9375, + "rewards/margins": 6.84375, + "rewards/rejected": -5.90625, + "step": 211 + }, + { + "epoch": 0.26012269938650306, + "grad_norm": 35.00437697538439, + "learning_rate": 1.384090909090909e-06, + "logits/chosen": -0.099609375, + "logits/rejected": -0.251953125, + "logps/chosen": -404.0, + "logps/rejected": -328.0, + "loss": 0.3606, + "rewards/accuracies": 0.8203125, + "rewards/chosen": 1.125, + "rewards/margins": 6.6875, + "rewards/rejected": -5.5625, + "step": 212 + }, + { + "epoch": 0.26134969325153373, + "grad_norm": 40.33706711227177, + "learning_rate": 1.3834710743801654e-06, + "logits/chosen": -0.031982421875, + "logits/rejected": -0.0888671875, + "logps/chosen": -352.0, + "logps/rejected": -296.0, + "loss": 0.3034, + "rewards/accuracies": 0.828125, + "rewards/chosen": 0.4765625, + "rewards/margins": 5.78125, + "rewards/rejected": -5.3125, + "step": 213 + }, + { + "epoch": 0.2625766871165644, + "grad_norm": 36.278922089821435, + "learning_rate": 1.3828512396694215e-06, + "logits/chosen": -0.0152587890625, + "logits/rejected": -0.171875, + "logps/chosen": -376.0, + "logps/rejected": -326.0, + "loss": 0.3538, + "rewards/accuracies": 0.796875, + "rewards/chosen": 1.1328125, + "rewards/margins": 6.375, + "rewards/rejected": -5.25, + "step": 214 + }, + { + "epoch": 0.26380368098159507, + "grad_norm": 39.61373791294869, + "learning_rate": 1.3822314049586777e-06, + "logits/chosen": 0.0224609375, + "logits/rejected": -0.06689453125, + "logps/chosen": -356.0, + "logps/rejected": -316.0, + "loss": 0.3393, + "rewards/accuracies": 0.8203125, + "rewards/chosen": 0.55859375, + "rewards/margins": 5.5625, + "rewards/rejected": -5.0, + "step": 215 + }, + { + "epoch": 0.2650306748466258, + "grad_norm": 36.52111248304287, + "learning_rate": 1.381611570247934e-06, + "logits/chosen": -0.04443359375, + "logits/rejected": -0.1572265625, + "logps/chosen": -334.0, + "logps/rejected": -302.0, + "loss": 0.3612, + "rewards/accuracies": 0.8203125, + "rewards/chosen": 0.42578125, + "rewards/margins": 5.375, + "rewards/rejected": -4.96875, + "step": 216 + }, + { + "epoch": 0.26625766871165646, + "grad_norm": 35.893159477690816, + "learning_rate": 1.38099173553719e-06, + "logits/chosen": 0.05517578125, + "logits/rejected": -0.08154296875, + "logps/chosen": -402.0, + "logps/rejected": -316.0, + "loss": 0.4174, + "rewards/accuracies": 0.7734375, + "rewards/chosen": 1.2890625, + "rewards/margins": 6.0625, + "rewards/rejected": -4.78125, + "step": 217 + }, + { + "epoch": 0.2674846625766871, + "grad_norm": 43.28791050351139, + "learning_rate": 1.3803719008264463e-06, + "logits/chosen": 0.0108642578125, + "logits/rejected": -0.0908203125, + "logps/chosen": -384.0, + "logps/rejected": -344.0, + "loss": 0.514, + "rewards/accuracies": 0.8125, + "rewards/chosen": 0.79296875, + "rewards/margins": 5.625, + "rewards/rejected": -4.8125, + "step": 218 + }, + { + "epoch": 0.2687116564417178, + "grad_norm": 30.96899305966375, + "learning_rate": 1.3797520661157025e-06, + "logits/chosen": -0.091796875, + "logits/rejected": -0.189453125, + "logps/chosen": -354.0, + "logps/rejected": -310.0, + "loss": 0.3339, + "rewards/accuracies": 0.828125, + "rewards/chosen": 0.81640625, + "rewards/margins": 6.09375, + "rewards/rejected": -5.28125, + "step": 219 + }, + { + "epoch": 0.26993865030674846, + "grad_norm": 49.60415252304962, + "learning_rate": 1.3791322314049586e-06, + "logits/chosen": 0.07373046875, + "logits/rejected": 0.0047607421875, + "logps/chosen": -366.0, + "logps/rejected": -372.0, + "loss": 0.4563, + "rewards/accuracies": 0.765625, + "rewards/chosen": 0.416015625, + "rewards/margins": 5.0625, + "rewards/rejected": -4.65625, + "step": 220 + }, + { + "epoch": 0.27116564417177913, + "grad_norm": 35.817382292638634, + "learning_rate": 1.378512396694215e-06, + "logits/chosen": -0.049560546875, + "logits/rejected": -0.25390625, + "logps/chosen": -364.0, + "logps/rejected": -306.0, + "loss": 0.4136, + "rewards/accuracies": 0.796875, + "rewards/chosen": 0.46484375, + "rewards/margins": 5.75, + "rewards/rejected": -5.28125, + "step": 221 + }, + { + "epoch": 0.2723926380368098, + "grad_norm": 33.842292905676736, + "learning_rate": 1.377892561983471e-06, + "logits/chosen": -0.240234375, + "logits/rejected": -0.3671875, + "logps/chosen": -388.0, + "logps/rejected": -318.0, + "loss": 0.3321, + "rewards/accuracies": 0.859375, + "rewards/chosen": 1.4296875, + "rewards/margins": 7.375, + "rewards/rejected": -5.9375, + "step": 222 + }, + { + "epoch": 0.27361963190184047, + "grad_norm": 34.24344157812324, + "learning_rate": 1.3772727272727274e-06, + "logits/chosen": -0.0262451171875, + "logits/rejected": -0.2060546875, + "logps/chosen": -390.0, + "logps/rejected": -338.0, + "loss": 0.3389, + "rewards/accuracies": 0.84375, + "rewards/chosen": 0.58984375, + "rewards/margins": 6.1875, + "rewards/rejected": -5.59375, + "step": 223 + }, + { + "epoch": 0.2748466257668712, + "grad_norm": 36.32910023908209, + "learning_rate": 1.3766528925619836e-06, + "logits/chosen": -0.1591796875, + "logits/rejected": -0.33203125, + "logps/chosen": -452.0, + "logps/rejected": -352.0, + "loss": 0.3172, + "rewards/accuracies": 0.8046875, + "rewards/chosen": 1.28125, + "rewards/margins": 7.1875, + "rewards/rejected": -5.90625, + "step": 224 + }, + { + "epoch": 0.27607361963190186, + "grad_norm": 33.96451254717418, + "learning_rate": 1.3760330578512397e-06, + "logits/chosen": -0.021484375, + "logits/rejected": -0.2119140625, + "logps/chosen": -344.0, + "logps/rejected": -322.0, + "loss": 0.2991, + "rewards/accuracies": 0.84375, + "rewards/chosen": 0.4609375, + "rewards/margins": 6.5, + "rewards/rejected": -6.0625, + "step": 225 + }, + { + "epoch": 0.2773006134969325, + "grad_norm": 37.753965128366424, + "learning_rate": 1.3754132231404959e-06, + "logits/chosen": -0.035888671875, + "logits/rejected": -0.1689453125, + "logps/chosen": -378.0, + "logps/rejected": -328.0, + "loss": 0.4435, + "rewards/accuracies": 0.8046875, + "rewards/chosen": 0.82421875, + "rewards/margins": 6.4375, + "rewards/rejected": -5.625, + "step": 226 + }, + { + "epoch": 0.2785276073619632, + "grad_norm": 37.97162675437385, + "learning_rate": 1.374793388429752e-06, + "logits/chosen": -0.078125, + "logits/rejected": -0.1416015625, + "logps/chosen": -362.0, + "logps/rejected": -332.0, + "loss": 0.3808, + "rewards/accuracies": 0.859375, + "rewards/chosen": 0.625, + "rewards/margins": 6.625, + "rewards/rejected": -6.0, + "step": 227 + }, + { + "epoch": 0.27975460122699386, + "grad_norm": 42.90390749582614, + "learning_rate": 1.3741735537190084e-06, + "logits/chosen": -0.0908203125, + "logits/rejected": -0.2734375, + "logps/chosen": -366.0, + "logps/rejected": -338.0, + "loss": 0.3568, + "rewards/accuracies": 0.8125, + "rewards/chosen": 0.57421875, + "rewards/margins": 6.03125, + "rewards/rejected": -5.46875, + "step": 228 + }, + { + "epoch": 0.2809815950920245, + "grad_norm": 35.198650610347336, + "learning_rate": 1.3735537190082645e-06, + "logits/chosen": 0.0174560546875, + "logits/rejected": -0.1845703125, + "logps/chosen": -422.0, + "logps/rejected": -344.0, + "loss": 0.3812, + "rewards/accuracies": 0.84375, + "rewards/chosen": 1.640625, + "rewards/margins": 6.03125, + "rewards/rejected": -4.375, + "step": 229 + }, + { + "epoch": 0.2822085889570552, + "grad_norm": 31.76904887089288, + "learning_rate": 1.3729338842975206e-06, + "logits/chosen": 0.0027618408203125, + "logits/rejected": -0.1806640625, + "logps/chosen": -376.0, + "logps/rejected": -334.0, + "loss": 0.2964, + "rewards/accuracies": 0.859375, + "rewards/chosen": 0.99609375, + "rewards/margins": 6.375, + "rewards/rejected": -5.375, + "step": 230 + }, + { + "epoch": 0.28343558282208586, + "grad_norm": 38.2475164063325, + "learning_rate": 1.372314049586777e-06, + "logits/chosen": -0.0615234375, + "logits/rejected": -0.177734375, + "logps/chosen": -370.0, + "logps/rejected": -320.0, + "loss": 0.4068, + "rewards/accuracies": 0.8203125, + "rewards/chosen": 1.03125, + "rewards/margins": 5.6875, + "rewards/rejected": -4.65625, + "step": 231 + }, + { + "epoch": 0.2846625766871166, + "grad_norm": 34.785350120825576, + "learning_rate": 1.3716942148760331e-06, + "logits/chosen": -0.1474609375, + "logits/rejected": -0.27734375, + "logps/chosen": -384.0, + "logps/rejected": -342.0, + "loss": 0.3319, + "rewards/accuracies": 0.8203125, + "rewards/chosen": 0.6796875, + "rewards/margins": 5.5625, + "rewards/rejected": -4.875, + "step": 232 + }, + { + "epoch": 0.28588957055214725, + "grad_norm": 33.915962784771814, + "learning_rate": 1.3710743801652893e-06, + "logits/chosen": -0.04052734375, + "logits/rejected": -0.255859375, + "logps/chosen": -372.0, + "logps/rejected": -296.0, + "loss": 0.3377, + "rewards/accuracies": 0.8359375, + "rewards/chosen": 1.4296875, + "rewards/margins": 6.28125, + "rewards/rejected": -4.84375, + "step": 233 + }, + { + "epoch": 0.2871165644171779, + "grad_norm": 36.944834789970336, + "learning_rate": 1.3704545454545456e-06, + "logits/chosen": -0.189453125, + "logits/rejected": -0.302734375, + "logps/chosen": -374.0, + "logps/rejected": -318.0, + "loss": 0.3474, + "rewards/accuracies": 0.8359375, + "rewards/chosen": 1.203125, + "rewards/margins": 5.625, + "rewards/rejected": -4.40625, + "step": 234 + }, + { + "epoch": 0.2883435582822086, + "grad_norm": 36.36585753307406, + "learning_rate": 1.3698347107438016e-06, + "logits/chosen": -0.201171875, + "logits/rejected": -0.34765625, + "logps/chosen": -362.0, + "logps/rejected": -326.0, + "loss": 0.3987, + "rewards/accuracies": 0.8359375, + "rewards/chosen": 1.1171875, + "rewards/margins": 5.90625, + "rewards/rejected": -4.78125, + "step": 235 + }, + { + "epoch": 0.28957055214723926, + "grad_norm": 39.75324625491103, + "learning_rate": 1.369214876033058e-06, + "logits/chosen": 0.058837890625, + "logits/rejected": -0.1123046875, + "logps/chosen": -396.0, + "logps/rejected": -374.0, + "loss": 0.4644, + "rewards/accuracies": 0.7890625, + "rewards/chosen": 1.7421875, + "rewards/margins": 5.625, + "rewards/rejected": -3.890625, + "step": 236 + }, + { + "epoch": 0.2907975460122699, + "grad_norm": 33.93837750645907, + "learning_rate": 1.368595041322314e-06, + "logits/chosen": -0.1025390625, + "logits/rejected": -0.248046875, + "logps/chosen": -368.0, + "logps/rejected": -314.0, + "loss": 0.3597, + "rewards/accuracies": 0.8515625, + "rewards/chosen": 1.5, + "rewards/margins": 6.1875, + "rewards/rejected": -4.6875, + "step": 237 + }, + { + "epoch": 0.2920245398773006, + "grad_norm": 40.82126535762802, + "learning_rate": 1.3679752066115702e-06, + "logits/chosen": -0.11376953125, + "logits/rejected": -0.259765625, + "logps/chosen": -426.0, + "logps/rejected": -356.0, + "loss": 0.5098, + "rewards/accuracies": 0.8046875, + "rewards/chosen": 1.4453125, + "rewards/margins": 5.6875, + "rewards/rejected": -4.25, + "step": 238 + }, + { + "epoch": 0.29325153374233126, + "grad_norm": 37.96435040796285, + "learning_rate": 1.3673553719008266e-06, + "logits/chosen": -0.130859375, + "logits/rejected": -0.298828125, + "logps/chosen": -382.0, + "logps/rejected": -314.0, + "loss": 0.371, + "rewards/accuracies": 0.796875, + "rewards/chosen": 2.078125, + "rewards/margins": 6.46875, + "rewards/rejected": -4.375, + "step": 239 + }, + { + "epoch": 0.294478527607362, + "grad_norm": 34.11140908926358, + "learning_rate": 1.3667355371900827e-06, + "logits/chosen": -0.15234375, + "logits/rejected": -0.28515625, + "logps/chosen": -418.0, + "logps/rejected": -364.0, + "loss": 0.2283, + "rewards/accuracies": 0.890625, + "rewards/chosen": 1.890625, + "rewards/margins": 6.875, + "rewards/rejected": -5.0, + "step": 240 + }, + { + "epoch": 0.29570552147239265, + "grad_norm": 33.461868421672946, + "learning_rate": 1.3661157024793388e-06, + "logits/chosen": -0.076171875, + "logits/rejected": -0.2001953125, + "logps/chosen": -366.0, + "logps/rejected": -336.0, + "loss": 0.2828, + "rewards/accuracies": 0.84375, + "rewards/chosen": 1.8359375, + "rewards/margins": 6.4375, + "rewards/rejected": -4.59375, + "step": 241 + }, + { + "epoch": 0.2969325153374233, + "grad_norm": 32.77289894542164, + "learning_rate": 1.3654958677685952e-06, + "logits/chosen": -0.10009765625, + "logits/rejected": -0.2490234375, + "logps/chosen": -350.0, + "logps/rejected": -324.0, + "loss": 0.2975, + "rewards/accuracies": 0.84375, + "rewards/chosen": 1.2578125, + "rewards/margins": 6.59375, + "rewards/rejected": -5.34375, + "step": 242 + }, + { + "epoch": 0.298159509202454, + "grad_norm": 38.07813065801456, + "learning_rate": 1.3648760330578511e-06, + "logits/chosen": -0.1923828125, + "logits/rejected": -0.328125, + "logps/chosen": -390.0, + "logps/rejected": -364.0, + "loss": 0.4242, + "rewards/accuracies": 0.8203125, + "rewards/chosen": 1.328125, + "rewards/margins": 5.625, + "rewards/rejected": -4.3125, + "step": 243 + }, + { + "epoch": 0.29938650306748466, + "grad_norm": 39.871497711059654, + "learning_rate": 1.3642561983471075e-06, + "logits/chosen": -0.1806640625, + "logits/rejected": -0.39453125, + "logps/chosen": -386.0, + "logps/rejected": -346.0, + "loss": 0.5523, + "rewards/accuracies": 0.7734375, + "rewards/chosen": 1.265625, + "rewards/margins": 5.46875, + "rewards/rejected": -4.1875, + "step": 244 + }, + { + "epoch": 0.3006134969325153, + "grad_norm": 32.221505039188294, + "learning_rate": 1.3636363636363636e-06, + "logits/chosen": -0.15625, + "logits/rejected": -0.2890625, + "logps/chosen": -398.0, + "logps/rejected": -354.0, + "loss": 0.2967, + "rewards/accuracies": 0.8515625, + "rewards/chosen": 1.9140625, + "rewards/margins": 6.46875, + "rewards/rejected": -4.53125, + "step": 245 + }, + { + "epoch": 0.301840490797546, + "grad_norm": 34.1015902082894, + "learning_rate": 1.3630165289256198e-06, + "logits/chosen": -0.166015625, + "logits/rejected": -0.31640625, + "logps/chosen": -374.0, + "logps/rejected": -346.0, + "loss": 0.3765, + "rewards/accuracies": 0.828125, + "rewards/chosen": 2.171875, + "rewards/margins": 6.46875, + "rewards/rejected": -4.3125, + "step": 246 + }, + { + "epoch": 0.3030674846625767, + "grad_norm": 32.83309359401401, + "learning_rate": 1.3623966942148761e-06, + "logits/chosen": -0.1572265625, + "logits/rejected": -0.337890625, + "logps/chosen": -372.0, + "logps/rejected": -308.0, + "loss": 0.3534, + "rewards/accuracies": 0.8125, + "rewards/chosen": 1.1640625, + "rewards/margins": 5.96875, + "rewards/rejected": -4.78125, + "step": 247 + }, + { + "epoch": 0.3042944785276074, + "grad_norm": 33.089465845458236, + "learning_rate": 1.3617768595041322e-06, + "logits/chosen": -0.076171875, + "logits/rejected": -0.3203125, + "logps/chosen": -388.0, + "logps/rejected": -332.0, + "loss": 0.2682, + "rewards/accuracies": 0.8671875, + "rewards/chosen": 1.703125, + "rewards/margins": 6.03125, + "rewards/rejected": -4.3125, + "step": 248 + }, + { + "epoch": 0.30552147239263805, + "grad_norm": 39.03476607944447, + "learning_rate": 1.3611570247933884e-06, + "logits/chosen": -0.20703125, + "logits/rejected": -0.396484375, + "logps/chosen": -474.0, + "logps/rejected": -368.0, + "loss": 0.3241, + "rewards/accuracies": 0.875, + "rewards/chosen": 2.640625, + "rewards/margins": 6.78125, + "rewards/rejected": -4.15625, + "step": 249 + }, + { + "epoch": 0.3067484662576687, + "grad_norm": 38.29235099515543, + "learning_rate": 1.3605371900826447e-06, + "logits/chosen": -0.30078125, + "logits/rejected": -0.349609375, + "logps/chosen": -340.0, + "logps/rejected": -324.0, + "loss": 0.3804, + "rewards/accuracies": 0.84375, + "rewards/chosen": 1.34375, + "rewards/margins": 5.78125, + "rewards/rejected": -4.4375, + "step": 250 + }, + { + "epoch": 0.3079754601226994, + "grad_norm": 37.34357400280288, + "learning_rate": 1.3599173553719009e-06, + "logits/chosen": 0.0289306640625, + "logits/rejected": -0.134765625, + "logps/chosen": -376.0, + "logps/rejected": -308.0, + "loss": 0.4118, + "rewards/accuracies": 0.7890625, + "rewards/chosen": 1.421875, + "rewards/margins": 5.375, + "rewards/rejected": -3.9375, + "step": 251 + }, + { + "epoch": 0.30920245398773005, + "grad_norm": 37.841749969537275, + "learning_rate": 1.359297520661157e-06, + "logits/chosen": -0.1494140625, + "logits/rejected": -0.29296875, + "logps/chosen": -392.0, + "logps/rejected": -310.0, + "loss": 0.4506, + "rewards/accuracies": 0.8359375, + "rewards/chosen": 1.890625, + "rewards/margins": 5.78125, + "rewards/rejected": -3.875, + "step": 252 + }, + { + "epoch": 0.3104294478527607, + "grad_norm": 25.74895782175542, + "learning_rate": 1.3586776859504132e-06, + "logits/chosen": -0.1630859375, + "logits/rejected": -0.392578125, + "logps/chosen": -336.0, + "logps/rejected": -298.0, + "loss": 0.2694, + "rewards/accuracies": 0.8671875, + "rewards/chosen": 1.46875, + "rewards/margins": 6.0625, + "rewards/rejected": -4.59375, + "step": 253 + }, + { + "epoch": 0.3116564417177914, + "grad_norm": 36.56021325822385, + "learning_rate": 1.3580578512396693e-06, + "logits/chosen": -0.10302734375, + "logits/rejected": -0.25, + "logps/chosen": -352.0, + "logps/rejected": -320.0, + "loss": 0.4183, + "rewards/accuracies": 0.8046875, + "rewards/chosen": 1.8671875, + "rewards/margins": 6.1875, + "rewards/rejected": -4.3125, + "step": 254 + }, + { + "epoch": 0.3128834355828221, + "grad_norm": 37.45936986772829, + "learning_rate": 1.3574380165289257e-06, + "logits/chosen": -0.07958984375, + "logits/rejected": -0.228515625, + "logps/chosen": -384.0, + "logps/rejected": -312.0, + "loss": 0.439, + "rewards/accuracies": 0.796875, + "rewards/chosen": 1.8515625, + "rewards/margins": 5.71875, + "rewards/rejected": -3.859375, + "step": 255 + }, + { + "epoch": 0.3141104294478528, + "grad_norm": 33.113238785557265, + "learning_rate": 1.3568181818181818e-06, + "logits/chosen": -0.1640625, + "logits/rejected": -0.328125, + "logps/chosen": -348.0, + "logps/rejected": -316.0, + "loss": 0.3148, + "rewards/accuracies": 0.875, + "rewards/chosen": 2.3125, + "rewards/margins": 7.0, + "rewards/rejected": -4.6875, + "step": 256 + }, + { + "epoch": 0.31533742331288345, + "grad_norm": 34.825667619250744, + "learning_rate": 1.3561983471074382e-06, + "logits/chosen": -0.1494140625, + "logits/rejected": -0.2890625, + "logps/chosen": -426.0, + "logps/rejected": -356.0, + "loss": 0.3455, + "rewards/accuracies": 0.828125, + "rewards/chosen": 2.25, + "rewards/margins": 6.5625, + "rewards/rejected": -4.3125, + "step": 257 + }, + { + "epoch": 0.3165644171779141, + "grad_norm": 35.08263037322888, + "learning_rate": 1.3555785123966943e-06, + "logits/chosen": -0.0908203125, + "logits/rejected": -0.28125, + "logps/chosen": -396.0, + "logps/rejected": -342.0, + "loss": 0.319, + "rewards/accuracies": 0.875, + "rewards/chosen": 1.9140625, + "rewards/margins": 6.65625, + "rewards/rejected": -4.75, + "step": 258 + }, + { + "epoch": 0.3177914110429448, + "grad_norm": 33.547677568046986, + "learning_rate": 1.3549586776859504e-06, + "logits/chosen": -0.1689453125, + "logits/rejected": -0.349609375, + "logps/chosen": -344.0, + "logps/rejected": -356.0, + "loss": 0.3553, + "rewards/accuracies": 0.84375, + "rewards/chosen": 1.296875, + "rewards/margins": 6.5, + "rewards/rejected": -5.1875, + "step": 259 + }, + { + "epoch": 0.31901840490797545, + "grad_norm": 35.38305104419636, + "learning_rate": 1.3543388429752068e-06, + "logits/chosen": -0.1904296875, + "logits/rejected": -0.26953125, + "logps/chosen": -390.0, + "logps/rejected": -326.0, + "loss": 0.3657, + "rewards/accuracies": 0.8359375, + "rewards/chosen": 1.296875, + "rewards/margins": 6.4375, + "rewards/rejected": -5.15625, + "step": 260 + }, + { + "epoch": 0.3202453987730061, + "grad_norm": 37.62925536561278, + "learning_rate": 1.3537190082644627e-06, + "logits/chosen": -0.11181640625, + "logits/rejected": -0.263671875, + "logps/chosen": -380.0, + "logps/rejected": -330.0, + "loss": 0.3941, + "rewards/accuracies": 0.8125, + "rewards/chosen": 1.296875, + "rewards/margins": 6.8125, + "rewards/rejected": -5.53125, + "step": 261 + }, + { + "epoch": 0.3214723926380368, + "grad_norm": 45.47620412654006, + "learning_rate": 1.353099173553719e-06, + "logits/chosen": -0.1591796875, + "logits/rejected": -0.3359375, + "logps/chosen": -386.0, + "logps/rejected": -340.0, + "loss": 0.4019, + "rewards/accuracies": 0.8125, + "rewards/chosen": 1.875, + "rewards/margins": 6.9375, + "rewards/rejected": -5.0625, + "step": 262 + }, + { + "epoch": 0.3226993865030675, + "grad_norm": 43.246204937096174, + "learning_rate": 1.3524793388429752e-06, + "logits/chosen": -0.1494140625, + "logits/rejected": -0.2734375, + "logps/chosen": -370.0, + "logps/rejected": -322.0, + "loss": 0.4521, + "rewards/accuracies": 0.765625, + "rewards/chosen": 0.6484375, + "rewards/margins": 5.65625, + "rewards/rejected": -5.0, + "step": 263 + }, + { + "epoch": 0.3239263803680982, + "grad_norm": 37.571005698421835, + "learning_rate": 1.3518595041322314e-06, + "logits/chosen": -0.09521484375, + "logits/rejected": -0.20703125, + "logps/chosen": -354.0, + "logps/rejected": -308.0, + "loss": 0.3748, + "rewards/accuracies": 0.8125, + "rewards/chosen": 1.0078125, + "rewards/margins": 6.1875, + "rewards/rejected": -5.15625, + "step": 264 + }, + { + "epoch": 0.32515337423312884, + "grad_norm": 45.351424976255885, + "learning_rate": 1.3512396694214877e-06, + "logits/chosen": -0.1064453125, + "logits/rejected": -0.265625, + "logps/chosen": -360.0, + "logps/rejected": -288.0, + "loss": 0.3745, + "rewards/accuracies": 0.8203125, + "rewards/chosen": 0.87109375, + "rewards/margins": 6.75, + "rewards/rejected": -5.875, + "step": 265 + }, + { + "epoch": 0.3263803680981595, + "grad_norm": 44.31098246879602, + "learning_rate": 1.3506198347107439e-06, + "logits/chosen": -0.07861328125, + "logits/rejected": -0.228515625, + "logps/chosen": -326.0, + "logps/rejected": -316.0, + "loss": 0.5312, + "rewards/accuracies": 0.78125, + "rewards/chosen": -0.27734375, + "rewards/margins": 5.25, + "rewards/rejected": -5.53125, + "step": 266 + }, + { + "epoch": 0.3276073619631902, + "grad_norm": 43.337206726154484, + "learning_rate": 1.35e-06, + "logits/chosen": -0.0966796875, + "logits/rejected": -0.31640625, + "logps/chosen": -406.0, + "logps/rejected": -322.0, + "loss": 0.3998, + "rewards/accuracies": 0.8359375, + "rewards/chosen": 1.2421875, + "rewards/margins": 6.875, + "rewards/rejected": -5.625, + "step": 267 + }, + { + "epoch": 0.32883435582822085, + "grad_norm": 28.822383795146568, + "learning_rate": 1.3493801652892563e-06, + "logits/chosen": -0.25390625, + "logits/rejected": -0.365234375, + "logps/chosen": -392.0, + "logps/rejected": -344.0, + "loss": 0.209, + "rewards/accuracies": 0.90625, + "rewards/chosen": 0.75390625, + "rewards/margins": 7.40625, + "rewards/rejected": -6.65625, + "step": 268 + }, + { + "epoch": 0.3300613496932515, + "grad_norm": 30.291051549986957, + "learning_rate": 1.3487603305785125e-06, + "logits/chosen": -0.1240234375, + "logits/rejected": -0.43359375, + "logps/chosen": -412.0, + "logps/rejected": -322.0, + "loss": 0.237, + "rewards/accuracies": 0.890625, + "rewards/chosen": 0.83203125, + "rewards/margins": 7.125, + "rewards/rejected": -6.3125, + "step": 269 + }, + { + "epoch": 0.3312883435582822, + "grad_norm": 38.1594801225088, + "learning_rate": 1.3481404958677686e-06, + "logits/chosen": -0.09912109375, + "logits/rejected": -0.2138671875, + "logps/chosen": -334.0, + "logps/rejected": -308.0, + "loss": 0.4766, + "rewards/accuracies": 0.8046875, + "rewards/chosen": 0.462890625, + "rewards/margins": 5.625, + "rewards/rejected": -5.15625, + "step": 270 + }, + { + "epoch": 0.3325153374233129, + "grad_norm": 28.374821367974555, + "learning_rate": 1.3475206611570248e-06, + "logits/chosen": -0.12451171875, + "logits/rejected": -0.240234375, + "logps/chosen": -390.0, + "logps/rejected": -344.0, + "loss": 0.3004, + "rewards/accuracies": 0.875, + "rewards/chosen": 1.3984375, + "rewards/margins": 7.5, + "rewards/rejected": -6.09375, + "step": 271 + }, + { + "epoch": 0.3337423312883436, + "grad_norm": 34.17489484053231, + "learning_rate": 1.346900826446281e-06, + "logits/chosen": -0.00823974609375, + "logits/rejected": -0.171875, + "logps/chosen": -366.0, + "logps/rejected": -318.0, + "loss": 0.2861, + "rewards/accuracies": 0.8671875, + "rewards/chosen": 0.423828125, + "rewards/margins": 6.75, + "rewards/rejected": -6.34375, + "step": 272 + }, + { + "epoch": 0.33496932515337424, + "grad_norm": 41.21053534408477, + "learning_rate": 1.3462809917355373e-06, + "logits/chosen": 0.06640625, + "logits/rejected": -0.007110595703125, + "logps/chosen": -344.0, + "logps/rejected": -314.0, + "loss": 0.5165, + "rewards/accuracies": 0.7578125, + "rewards/chosen": 0.166015625, + "rewards/margins": 5.75, + "rewards/rejected": -5.5625, + "step": 273 + }, + { + "epoch": 0.3361963190184049, + "grad_norm": 34.571787806623156, + "learning_rate": 1.3456611570247934e-06, + "logits/chosen": 0.0419921875, + "logits/rejected": -0.1025390625, + "logps/chosen": -410.0, + "logps/rejected": -346.0, + "loss": 0.3839, + "rewards/accuracies": 0.8515625, + "rewards/chosen": 0.83203125, + "rewards/margins": 6.40625, + "rewards/rejected": -5.5625, + "step": 274 + }, + { + "epoch": 0.3374233128834356, + "grad_norm": 31.979565052575825, + "learning_rate": 1.3450413223140495e-06, + "logits/chosen": -0.0302734375, + "logits/rejected": -0.119140625, + "logps/chosen": -326.0, + "logps/rejected": -326.0, + "loss": 0.4175, + "rewards/accuracies": 0.8046875, + "rewards/chosen": 0.412109375, + "rewards/margins": 6.09375, + "rewards/rejected": -5.6875, + "step": 275 + }, + { + "epoch": 0.33865030674846625, + "grad_norm": 28.189301031425032, + "learning_rate": 1.344421487603306e-06, + "logits/chosen": 0.0498046875, + "logits/rejected": -0.162109375, + "logps/chosen": -390.0, + "logps/rejected": -330.0, + "loss": 0.3505, + "rewards/accuracies": 0.8125, + "rewards/chosen": 0.2470703125, + "rewards/margins": 6.21875, + "rewards/rejected": -5.96875, + "step": 276 + }, + { + "epoch": 0.3398773006134969, + "grad_norm": 29.004541701793432, + "learning_rate": 1.343801652892562e-06, + "logits/chosen": -0.15234375, + "logits/rejected": -0.337890625, + "logps/chosen": -414.0, + "logps/rejected": -344.0, + "loss": 0.2468, + "rewards/accuracies": 0.8671875, + "rewards/chosen": 1.1640625, + "rewards/margins": 7.53125, + "rewards/rejected": -6.375, + "step": 277 + }, + { + "epoch": 0.3411042944785276, + "grad_norm": 40.163399091060114, + "learning_rate": 1.3431818181818184e-06, + "logits/chosen": -0.03759765625, + "logits/rejected": -0.08447265625, + "logps/chosen": -384.0, + "logps/rejected": -356.0, + "loss": 0.4008, + "rewards/accuracies": 0.8203125, + "rewards/chosen": 0.69921875, + "rewards/margins": 6.4375, + "rewards/rejected": -5.75, + "step": 278 + }, + { + "epoch": 0.3423312883435583, + "grad_norm": 21.8515947233153, + "learning_rate": 1.3425619834710743e-06, + "logits/chosen": -0.10888671875, + "logits/rejected": -0.2890625, + "logps/chosen": -392.0, + "logps/rejected": -326.0, + "loss": 0.1765, + "rewards/accuracies": 0.921875, + "rewards/chosen": 0.65625, + "rewards/margins": 7.46875, + "rewards/rejected": -6.8125, + "step": 279 + }, + { + "epoch": 0.34355828220858897, + "grad_norm": 34.753521492334585, + "learning_rate": 1.3419421487603305e-06, + "logits/chosen": -0.09228515625, + "logits/rejected": -0.234375, + "logps/chosen": -408.0, + "logps/rejected": -372.0, + "loss": 0.4097, + "rewards/accuracies": 0.8125, + "rewards/chosen": 0.83984375, + "rewards/margins": 6.96875, + "rewards/rejected": -6.15625, + "step": 280 + }, + { + "epoch": 0.34478527607361964, + "grad_norm": 35.589195433139196, + "learning_rate": 1.3413223140495868e-06, + "logits/chosen": -0.00921630859375, + "logits/rejected": -0.201171875, + "logps/chosen": -382.0, + "logps/rejected": -330.0, + "loss": 0.3917, + "rewards/accuracies": 0.8046875, + "rewards/chosen": 0.609375, + "rewards/margins": 5.65625, + "rewards/rejected": -5.0625, + "step": 281 + }, + { + "epoch": 0.3460122699386503, + "grad_norm": 29.302596470256102, + "learning_rate": 1.340702479338843e-06, + "logits/chosen": -0.2021484375, + "logits/rejected": -0.341796875, + "logps/chosen": -344.0, + "logps/rejected": -324.0, + "loss": 0.2481, + "rewards/accuracies": 0.875, + "rewards/chosen": 1.0859375, + "rewards/margins": 7.1875, + "rewards/rejected": -6.09375, + "step": 282 + }, + { + "epoch": 0.347239263803681, + "grad_norm": 33.7375126879184, + "learning_rate": 1.3400826446280993e-06, + "logits/chosen": -0.2265625, + "logits/rejected": -0.3125, + "logps/chosen": -354.0, + "logps/rejected": -298.0, + "loss": 0.3519, + "rewards/accuracies": 0.8203125, + "rewards/chosen": 0.59765625, + "rewards/margins": 6.15625, + "rewards/rejected": -5.5625, + "step": 283 + }, + { + "epoch": 0.34846625766871164, + "grad_norm": 35.97047400633379, + "learning_rate": 1.3394628099173555e-06, + "logits/chosen": -0.034423828125, + "logits/rejected": -0.166015625, + "logps/chosen": -370.0, + "logps/rejected": -320.0, + "loss": 0.2954, + "rewards/accuracies": 0.8515625, + "rewards/chosen": 0.859375, + "rewards/margins": 6.6875, + "rewards/rejected": -5.84375, + "step": 284 + }, + { + "epoch": 0.3496932515337423, + "grad_norm": 29.423070534885003, + "learning_rate": 1.3388429752066116e-06, + "logits/chosen": -0.130859375, + "logits/rejected": -0.322265625, + "logps/chosen": -360.0, + "logps/rejected": -328.0, + "loss": 0.2845, + "rewards/accuracies": 0.84375, + "rewards/chosen": 0.80078125, + "rewards/margins": 6.65625, + "rewards/rejected": -5.875, + "step": 285 + }, + { + "epoch": 0.350920245398773, + "grad_norm": 35.376791661357736, + "learning_rate": 1.338223140495868e-06, + "logits/chosen": -0.04736328125, + "logits/rejected": -0.2412109375, + "logps/chosen": -422.0, + "logps/rejected": -356.0, + "loss": 0.3799, + "rewards/accuracies": 0.796875, + "rewards/chosen": 0.8515625, + "rewards/margins": 6.9375, + "rewards/rejected": -6.0625, + "step": 286 + }, + { + "epoch": 0.3521472392638037, + "grad_norm": 38.51049635159653, + "learning_rate": 1.337603305785124e-06, + "logits/chosen": -0.150390625, + "logits/rejected": -0.283203125, + "logps/chosen": -382.0, + "logps/rejected": -330.0, + "loss": 0.3715, + "rewards/accuracies": 0.8125, + "rewards/chosen": 0.375, + "rewards/margins": 6.1875, + "rewards/rejected": -5.8125, + "step": 287 + }, + { + "epoch": 0.35337423312883437, + "grad_norm": 31.395092155644413, + "learning_rate": 1.3369834710743802e-06, + "logits/chosen": 0.057373046875, + "logits/rejected": -0.0888671875, + "logps/chosen": -386.0, + "logps/rejected": -336.0, + "loss": 0.3837, + "rewards/accuracies": 0.8359375, + "rewards/chosen": 0.298828125, + "rewards/margins": 5.9375, + "rewards/rejected": -5.625, + "step": 288 + }, + { + "epoch": 0.35460122699386504, + "grad_norm": 25.699158986974325, + "learning_rate": 1.3363636363636364e-06, + "logits/chosen": -0.11083984375, + "logits/rejected": -0.248046875, + "logps/chosen": -404.0, + "logps/rejected": -330.0, + "loss": 0.1963, + "rewards/accuracies": 0.921875, + "rewards/chosen": 0.94140625, + "rewards/margins": 7.65625, + "rewards/rejected": -6.71875, + "step": 289 + }, + { + "epoch": 0.3558282208588957, + "grad_norm": 26.172712857085628, + "learning_rate": 1.3357438016528925e-06, + "logits/chosen": -0.283203125, + "logits/rejected": -0.41796875, + "logps/chosen": -380.0, + "logps/rejected": -350.0, + "loss": 0.2207, + "rewards/accuracies": 0.9296875, + "rewards/chosen": 0.65625, + "rewards/margins": 7.65625, + "rewards/rejected": -7.0, + "step": 290 + }, + { + "epoch": 0.3570552147239264, + "grad_norm": 35.706855965144605, + "learning_rate": 1.3351239669421489e-06, + "logits/chosen": -0.2275390625, + "logits/rejected": -0.3359375, + "logps/chosen": -394.0, + "logps/rejected": -340.0, + "loss": 0.4375, + "rewards/accuracies": 0.8046875, + "rewards/chosen": 1.5625, + "rewards/margins": 7.09375, + "rewards/rejected": -5.53125, + "step": 291 + }, + { + "epoch": 0.35828220858895704, + "grad_norm": 32.40631116792141, + "learning_rate": 1.334504132231405e-06, + "logits/chosen": -0.169921875, + "logits/rejected": -0.390625, + "logps/chosen": -406.0, + "logps/rejected": -348.0, + "loss": 0.2821, + "rewards/accuracies": 0.8671875, + "rewards/chosen": 0.7734375, + "rewards/margins": 7.15625, + "rewards/rejected": -6.40625, + "step": 292 + }, + { + "epoch": 0.3595092024539877, + "grad_norm": 27.724555696310663, + "learning_rate": 1.3338842975206611e-06, + "logits/chosen": -0.08251953125, + "logits/rejected": -0.275390625, + "logps/chosen": -382.0, + "logps/rejected": -332.0, + "loss": 0.2357, + "rewards/accuracies": 0.8828125, + "rewards/chosen": 0.033203125, + "rewards/margins": 6.78125, + "rewards/rejected": -6.71875, + "step": 293 + }, + { + "epoch": 0.36073619631901843, + "grad_norm": 26.756162333522596, + "learning_rate": 1.3332644628099175e-06, + "logits/chosen": -0.359375, + "logits/rejected": -0.4921875, + "logps/chosen": -472.0, + "logps/rejected": -368.0, + "loss": 0.3094, + "rewards/accuracies": 0.8515625, + "rewards/chosen": 0.66796875, + "rewards/margins": 7.65625, + "rewards/rejected": -7.0, + "step": 294 + }, + { + "epoch": 0.3619631901840491, + "grad_norm": 27.599547481570436, + "learning_rate": 1.3326446280991736e-06, + "logits/chosen": -0.04345703125, + "logits/rejected": -0.1884765625, + "logps/chosen": -356.0, + "logps/rejected": -332.0, + "loss": 0.2495, + "rewards/accuracies": 0.8828125, + "rewards/chosen": 0.2734375, + "rewards/margins": 6.875, + "rewards/rejected": -6.59375, + "step": 295 + }, + { + "epoch": 0.36319018404907977, + "grad_norm": 35.1081345543023, + "learning_rate": 1.3320247933884298e-06, + "logits/chosen": -0.007598876953125, + "logits/rejected": -0.154296875, + "logps/chosen": -420.0, + "logps/rejected": -338.0, + "loss": 0.3852, + "rewards/accuracies": 0.859375, + "rewards/chosen": 1.1640625, + "rewards/margins": 6.6875, + "rewards/rejected": -5.5, + "step": 296 + }, + { + "epoch": 0.36441717791411044, + "grad_norm": 31.495501539039623, + "learning_rate": 1.331404958677686e-06, + "logits/chosen": -0.154296875, + "logits/rejected": -0.30859375, + "logps/chosen": -360.0, + "logps/rejected": -322.0, + "loss": 0.2553, + "rewards/accuracies": 0.859375, + "rewards/chosen": 0.79296875, + "rewards/margins": 7.0625, + "rewards/rejected": -6.25, + "step": 297 + }, + { + "epoch": 0.3656441717791411, + "grad_norm": 37.56929706779854, + "learning_rate": 1.330785123966942e-06, + "logits/chosen": -0.009521484375, + "logits/rejected": -0.09716796875, + "logps/chosen": -384.0, + "logps/rejected": -336.0, + "loss": 0.3993, + "rewards/accuracies": 0.8203125, + "rewards/chosen": 0.498046875, + "rewards/margins": 6.03125, + "rewards/rejected": -5.53125, + "step": 298 + }, + { + "epoch": 0.36687116564417177, + "grad_norm": 36.46646666012725, + "learning_rate": 1.3301652892561984e-06, + "logits/chosen": -0.058837890625, + "logits/rejected": -0.228515625, + "logps/chosen": -400.0, + "logps/rejected": -334.0, + "loss": 0.3895, + "rewards/accuracies": 0.84375, + "rewards/chosen": -0.1650390625, + "rewards/margins": 5.78125, + "rewards/rejected": -5.9375, + "step": 299 + }, + { + "epoch": 0.36809815950920244, + "grad_norm": 37.01196046043553, + "learning_rate": 1.3295454545454546e-06, + "logits/chosen": -0.173828125, + "logits/rejected": -0.3203125, + "logps/chosen": -372.0, + "logps/rejected": -328.0, + "loss": 0.4644, + "rewards/accuracies": 0.8203125, + "rewards/chosen": 0.482421875, + "rewards/margins": 6.65625, + "rewards/rejected": -6.15625, + "step": 300 + }, + { + "epoch": 0.3693251533742331, + "grad_norm": 32.17417313364543, + "learning_rate": 1.3289256198347107e-06, + "logits/chosen": -0.11865234375, + "logits/rejected": -0.298828125, + "logps/chosen": -460.0, + "logps/rejected": -386.0, + "loss": 0.3417, + "rewards/accuracies": 0.828125, + "rewards/chosen": 0.8515625, + "rewards/margins": 7.1875, + "rewards/rejected": -6.34375, + "step": 301 + }, + { + "epoch": 0.37055214723926383, + "grad_norm": 25.798891470824255, + "learning_rate": 1.328305785123967e-06, + "logits/chosen": -0.150390625, + "logits/rejected": -0.34765625, + "logps/chosen": -382.0, + "logps/rejected": -316.0, + "loss": 0.2452, + "rewards/accuracies": 0.8984375, + "rewards/chosen": 0.287109375, + "rewards/margins": 6.6875, + "rewards/rejected": -6.40625, + "step": 302 + }, + { + "epoch": 0.3717791411042945, + "grad_norm": 40.25697610802204, + "learning_rate": 1.3276859504132232e-06, + "logits/chosen": -0.1474609375, + "logits/rejected": -0.2578125, + "logps/chosen": -388.0, + "logps/rejected": -376.0, + "loss": 0.4569, + "rewards/accuracies": 0.78125, + "rewards/chosen": -0.466796875, + "rewards/margins": 6.1875, + "rewards/rejected": -6.65625, + "step": 303 + }, + { + "epoch": 0.37300613496932516, + "grad_norm": 30.538056754282714, + "learning_rate": 1.3270661157024793e-06, + "logits/chosen": -0.0264892578125, + "logits/rejected": -0.287109375, + "logps/chosen": -396.0, + "logps/rejected": -334.0, + "loss": 0.2974, + "rewards/accuracies": 0.875, + "rewards/chosen": -0.140625, + "rewards/margins": 6.875, + "rewards/rejected": -7.0, + "step": 304 + }, + { + "epoch": 0.37423312883435583, + "grad_norm": 27.228440712710828, + "learning_rate": 1.3264462809917357e-06, + "logits/chosen": -0.109375, + "logits/rejected": -0.2490234375, + "logps/chosen": -416.0, + "logps/rejected": -344.0, + "loss": 0.2623, + "rewards/accuracies": 0.90625, + "rewards/chosen": 0.341796875, + "rewards/margins": 7.71875, + "rewards/rejected": -7.375, + "step": 305 + }, + { + "epoch": 0.3754601226993865, + "grad_norm": 34.02132219179046, + "learning_rate": 1.3258264462809916e-06, + "logits/chosen": -0.2177734375, + "logits/rejected": -0.37890625, + "logps/chosen": -430.0, + "logps/rejected": -364.0, + "loss": 0.2805, + "rewards/accuracies": 0.8828125, + "rewards/chosen": -0.33203125, + "rewards/margins": 6.75, + "rewards/rejected": -7.0625, + "step": 306 + }, + { + "epoch": 0.37668711656441717, + "grad_norm": 34.60907155238812, + "learning_rate": 1.325206611570248e-06, + "logits/chosen": -0.06103515625, + "logits/rejected": -0.19140625, + "logps/chosen": -382.0, + "logps/rejected": -344.0, + "loss": 0.4791, + "rewards/accuracies": 0.796875, + "rewards/chosen": -0.18359375, + "rewards/margins": 6.90625, + "rewards/rejected": -7.09375, + "step": 307 + }, + { + "epoch": 0.37791411042944784, + "grad_norm": 34.523248358542865, + "learning_rate": 1.3245867768595041e-06, + "logits/chosen": -0.00148773193359375, + "logits/rejected": -0.162109375, + "logps/chosen": -398.0, + "logps/rejected": -360.0, + "loss": 0.3857, + "rewards/accuracies": 0.828125, + "rewards/chosen": -0.53515625, + "rewards/margins": 6.0, + "rewards/rejected": -6.5625, + "step": 308 + }, + { + "epoch": 0.3791411042944785, + "grad_norm": 29.962650665710807, + "learning_rate": 1.3239669421487603e-06, + "logits/chosen": -0.208984375, + "logits/rejected": -0.31640625, + "logps/chosen": -388.0, + "logps/rejected": -354.0, + "loss": 0.2543, + "rewards/accuracies": 0.875, + "rewards/chosen": 0.609375, + "rewards/margins": 7.5625, + "rewards/rejected": -6.9375, + "step": 309 + }, + { + "epoch": 0.3803680981595092, + "grad_norm": 33.20481159292772, + "learning_rate": 1.3233471074380166e-06, + "logits/chosen": 0.051513671875, + "logits/rejected": -0.037353515625, + "logps/chosen": -322.0, + "logps/rejected": -312.0, + "loss": 0.4174, + "rewards/accuracies": 0.8359375, + "rewards/chosen": -0.875, + "rewards/margins": 5.5, + "rewards/rejected": -6.375, + "step": 310 + }, + { + "epoch": 0.3815950920245399, + "grad_norm": 32.36453885140526, + "learning_rate": 1.3227272727272727e-06, + "logits/chosen": -0.158203125, + "logits/rejected": -0.3359375, + "logps/chosen": -392.0, + "logps/rejected": -348.0, + "loss": 0.3525, + "rewards/accuracies": 0.8203125, + "rewards/chosen": 0.486328125, + "rewards/margins": 7.78125, + "rewards/rejected": -7.28125, + "step": 311 + }, + { + "epoch": 0.38282208588957056, + "grad_norm": 39.08052846369179, + "learning_rate": 1.322107438016529e-06, + "logits/chosen": 0.0162353515625, + "logits/rejected": -0.1142578125, + "logps/chosen": -392.0, + "logps/rejected": -378.0, + "loss": 0.5608, + "rewards/accuracies": 0.7421875, + "rewards/chosen": 0.0712890625, + "rewards/margins": 6.34375, + "rewards/rejected": -6.28125, + "step": 312 + }, + { + "epoch": 0.38404907975460123, + "grad_norm": 35.15009453427041, + "learning_rate": 1.3214876033057852e-06, + "logits/chosen": -0.0235595703125, + "logits/rejected": -0.298828125, + "logps/chosen": -410.0, + "logps/rejected": -372.0, + "loss": 0.383, + "rewards/accuracies": 0.84375, + "rewards/chosen": 1.1640625, + "rewards/margins": 8.0625, + "rewards/rejected": -6.9375, + "step": 313 + }, + { + "epoch": 0.3852760736196319, + "grad_norm": 28.038472021341548, + "learning_rate": 1.3208677685950414e-06, + "logits/chosen": -0.05517578125, + "logits/rejected": -0.2421875, + "logps/chosen": -458.0, + "logps/rejected": -360.0, + "loss": 0.2315, + "rewards/accuracies": 0.8828125, + "rewards/chosen": 2.03125, + "rewards/margins": 9.375, + "rewards/rejected": -7.34375, + "step": 314 + }, + { + "epoch": 0.38650306748466257, + "grad_norm": 31.702508477036716, + "learning_rate": 1.3202479338842975e-06, + "logits/chosen": 0.076171875, + "logits/rejected": -0.11865234375, + "logps/chosen": -350.0, + "logps/rejected": -300.0, + "loss": 0.3293, + "rewards/accuracies": 0.8203125, + "rewards/chosen": 0.2431640625, + "rewards/margins": 6.84375, + "rewards/rejected": -6.59375, + "step": 315 + }, + { + "epoch": 0.38773006134969323, + "grad_norm": 27.443273044054905, + "learning_rate": 1.3196280991735537e-06, + "logits/chosen": 0.05419921875, + "logits/rejected": -0.08544921875, + "logps/chosen": -360.0, + "logps/rejected": -308.0, + "loss": 0.3129, + "rewards/accuracies": 0.875, + "rewards/chosen": 0.75390625, + "rewards/margins": 7.5625, + "rewards/rejected": -6.8125, + "step": 316 + }, + { + "epoch": 0.3889570552147239, + "grad_norm": 28.339536052474916, + "learning_rate": 1.31900826446281e-06, + "logits/chosen": -0.09326171875, + "logits/rejected": -0.3046875, + "logps/chosen": -400.0, + "logps/rejected": -368.0, + "loss": 0.2873, + "rewards/accuracies": 0.859375, + "rewards/chosen": 0.94140625, + "rewards/margins": 8.125, + "rewards/rejected": -7.15625, + "step": 317 + }, + { + "epoch": 0.3901840490797546, + "grad_norm": 38.792002326692455, + "learning_rate": 1.3183884297520662e-06, + "logits/chosen": -0.142578125, + "logits/rejected": -0.296875, + "logps/chosen": -402.0, + "logps/rejected": -366.0, + "loss": 0.4379, + "rewards/accuracies": 0.828125, + "rewards/chosen": 0.91796875, + "rewards/margins": 7.0625, + "rewards/rejected": -6.125, + "step": 318 + }, + { + "epoch": 0.3914110429447853, + "grad_norm": 17.992013049840406, + "learning_rate": 1.3177685950413223e-06, + "logits/chosen": -0.068359375, + "logits/rejected": -0.255859375, + "logps/chosen": -414.0, + "logps/rejected": -320.0, + "loss": 0.1471, + "rewards/accuracies": 0.9296875, + "rewards/chosen": 1.1640625, + "rewards/margins": 8.8125, + "rewards/rejected": -7.65625, + "step": 319 + }, + { + "epoch": 0.39263803680981596, + "grad_norm": 31.193757820598883, + "learning_rate": 1.3171487603305787e-06, + "logits/chosen": -0.056884765625, + "logits/rejected": -0.166015625, + "logps/chosen": -320.0, + "logps/rejected": -286.0, + "loss": 0.325, + "rewards/accuracies": 0.859375, + "rewards/chosen": 0.31640625, + "rewards/margins": 6.84375, + "rewards/rejected": -6.53125, + "step": 320 + }, + { + "epoch": 0.39386503067484663, + "grad_norm": 39.71075166734815, + "learning_rate": 1.3165289256198348e-06, + "logits/chosen": -0.201171875, + "logits/rejected": -0.35546875, + "logps/chosen": -410.0, + "logps/rejected": -398.0, + "loss": 0.3784, + "rewards/accuracies": 0.8515625, + "rewards/chosen": 0.439453125, + "rewards/margins": 7.84375, + "rewards/rejected": -7.375, + "step": 321 + }, + { + "epoch": 0.3950920245398773, + "grad_norm": 26.7824487712344, + "learning_rate": 1.315909090909091e-06, + "logits/chosen": -0.107421875, + "logits/rejected": -0.2578125, + "logps/chosen": -394.0, + "logps/rejected": -330.0, + "loss": 0.2893, + "rewards/accuracies": 0.8984375, + "rewards/chosen": 1.203125, + "rewards/margins": 7.84375, + "rewards/rejected": -6.65625, + "step": 322 + }, + { + "epoch": 0.39631901840490796, + "grad_norm": 36.45438287117724, + "learning_rate": 1.3152892561983473e-06, + "logits/chosen": -0.1318359375, + "logits/rejected": -0.265625, + "logps/chosen": -354.0, + "logps/rejected": -334.0, + "loss": 0.4474, + "rewards/accuracies": 0.8203125, + "rewards/chosen": 0.6875, + "rewards/margins": 6.15625, + "rewards/rejected": -5.46875, + "step": 323 + }, + { + "epoch": 0.39754601226993863, + "grad_norm": 31.15880722522094, + "learning_rate": 1.3146694214876032e-06, + "logits/chosen": -0.193359375, + "logits/rejected": -0.361328125, + "logps/chosen": -356.0, + "logps/rejected": -328.0, + "loss": 0.3486, + "rewards/accuracies": 0.8515625, + "rewards/chosen": 0.62890625, + "rewards/margins": 7.4375, + "rewards/rejected": -6.8125, + "step": 324 + }, + { + "epoch": 0.3987730061349693, + "grad_norm": 36.47850348671778, + "learning_rate": 1.3140495867768596e-06, + "logits/chosen": -0.0272216796875, + "logits/rejected": -0.09326171875, + "logps/chosen": -330.0, + "logps/rejected": -304.0, + "loss": 0.5542, + "rewards/accuracies": 0.8046875, + "rewards/chosen": 0.33203125, + "rewards/margins": 6.15625, + "rewards/rejected": -5.84375, + "step": 325 + }, + { + "epoch": 0.4, + "grad_norm": 26.62936328977598, + "learning_rate": 1.3134297520661157e-06, + "logits/chosen": -0.2294921875, + "logits/rejected": -0.369140625, + "logps/chosen": -398.0, + "logps/rejected": -374.0, + "loss": 0.1993, + "rewards/accuracies": 0.8671875, + "rewards/chosen": 1.5546875, + "rewards/margins": 8.375, + "rewards/rejected": -6.84375, + "step": 326 + }, + { + "epoch": 0.4012269938650307, + "grad_norm": 39.76264342514329, + "learning_rate": 1.3128099173553719e-06, + "logits/chosen": -0.1015625, + "logits/rejected": -0.271484375, + "logps/chosen": -430.0, + "logps/rejected": -352.0, + "loss": 0.4557, + "rewards/accuracies": 0.8046875, + "rewards/chosen": 0.921875, + "rewards/margins": 6.125, + "rewards/rejected": -5.1875, + "step": 327 + }, + { + "epoch": 0.40245398773006136, + "grad_norm": 33.62855666124103, + "learning_rate": 1.3121900826446282e-06, + "logits/chosen": -0.1328125, + "logits/rejected": -0.259765625, + "logps/chosen": -362.0, + "logps/rejected": -288.0, + "loss": 0.363, + "rewards/accuracies": 0.84375, + "rewards/chosen": 0.228515625, + "rewards/margins": 6.15625, + "rewards/rejected": -5.9375, + "step": 328 + }, + { + "epoch": 0.403680981595092, + "grad_norm": 36.1359556726495, + "learning_rate": 1.3115702479338844e-06, + "logits/chosen": -0.08203125, + "logits/rejected": -0.283203125, + "logps/chosen": -466.0, + "logps/rejected": -386.0, + "loss": 0.3754, + "rewards/accuracies": 0.796875, + "rewards/chosen": 1.078125, + "rewards/margins": 8.3125, + "rewards/rejected": -7.25, + "step": 329 + }, + { + "epoch": 0.4049079754601227, + "grad_norm": 27.601952208856858, + "learning_rate": 1.3109504132231405e-06, + "logits/chosen": -0.22265625, + "logits/rejected": -0.423828125, + "logps/chosen": -418.0, + "logps/rejected": -352.0, + "loss": 0.2524, + "rewards/accuracies": 0.8515625, + "rewards/chosen": 1.0078125, + "rewards/margins": 8.25, + "rewards/rejected": -7.25, + "step": 330 + }, + { + "epoch": 0.40613496932515336, + "grad_norm": 26.526762211128467, + "learning_rate": 1.3103305785123968e-06, + "logits/chosen": -0.1767578125, + "logits/rejected": -0.283203125, + "logps/chosen": -408.0, + "logps/rejected": -356.0, + "loss": 0.2755, + "rewards/accuracies": 0.859375, + "rewards/chosen": 1.09375, + "rewards/margins": 7.625, + "rewards/rejected": -6.53125, + "step": 331 + }, + { + "epoch": 0.40736196319018403, + "grad_norm": 24.05844405102659, + "learning_rate": 1.3097107438016528e-06, + "logits/chosen": -0.1826171875, + "logits/rejected": -0.37109375, + "logps/chosen": -430.0, + "logps/rejected": -332.0, + "loss": 0.2196, + "rewards/accuracies": 0.8984375, + "rewards/chosen": 1.109375, + "rewards/margins": 7.96875, + "rewards/rejected": -6.84375, + "step": 332 + }, + { + "epoch": 0.4085889570552147, + "grad_norm": 36.41468851723906, + "learning_rate": 1.3090909090909091e-06, + "logits/chosen": 0.017822265625, + "logits/rejected": -0.1044921875, + "logps/chosen": -372.0, + "logps/rejected": -346.0, + "loss": 0.4229, + "rewards/accuracies": 0.7734375, + "rewards/chosen": 0.48046875, + "rewards/margins": 6.40625, + "rewards/rejected": -5.90625, + "step": 333 + }, + { + "epoch": 0.4098159509202454, + "grad_norm": 30.33603786358444, + "learning_rate": 1.3084710743801653e-06, + "logits/chosen": -0.1181640625, + "logits/rejected": -0.349609375, + "logps/chosen": -382.0, + "logps/rejected": -366.0, + "loss": 0.3109, + "rewards/accuracies": 0.8671875, + "rewards/chosen": 0.458984375, + "rewards/margins": 7.75, + "rewards/rejected": -7.28125, + "step": 334 + }, + { + "epoch": 0.4110429447852761, + "grad_norm": 33.01078113652063, + "learning_rate": 1.3078512396694214e-06, + "logits/chosen": -0.0213623046875, + "logits/rejected": -0.28125, + "logps/chosen": -432.0, + "logps/rejected": -332.0, + "loss": 0.2319, + "rewards/accuracies": 0.859375, + "rewards/chosen": 0.34375, + "rewards/margins": 7.0625, + "rewards/rejected": -6.71875, + "step": 335 + }, + { + "epoch": 0.41226993865030676, + "grad_norm": 31.965240510538873, + "learning_rate": 1.3072314049586778e-06, + "logits/chosen": -0.177734375, + "logits/rejected": -0.326171875, + "logps/chosen": -402.0, + "logps/rejected": -352.0, + "loss": 0.3352, + "rewards/accuracies": 0.859375, + "rewards/chosen": 0.3515625, + "rewards/margins": 7.15625, + "rewards/rejected": -6.8125, + "step": 336 + }, + { + "epoch": 0.4134969325153374, + "grad_norm": 36.98349242868308, + "learning_rate": 1.306611570247934e-06, + "logits/chosen": 0.0264892578125, + "logits/rejected": -0.1318359375, + "logps/chosen": -358.0, + "logps/rejected": -352.0, + "loss": 0.4393, + "rewards/accuracies": 0.796875, + "rewards/chosen": 0.1748046875, + "rewards/margins": 6.90625, + "rewards/rejected": -6.75, + "step": 337 + }, + { + "epoch": 0.4147239263803681, + "grad_norm": 31.08894150655132, + "learning_rate": 1.30599173553719e-06, + "logits/chosen": -0.005157470703125, + "logits/rejected": -0.1484375, + "logps/chosen": -392.0, + "logps/rejected": -354.0, + "loss": 0.3686, + "rewards/accuracies": 0.8515625, + "rewards/chosen": 0.2216796875, + "rewards/margins": 6.96875, + "rewards/rejected": -6.75, + "step": 338 + }, + { + "epoch": 0.41595092024539876, + "grad_norm": 34.40525022311832, + "learning_rate": 1.3053719008264464e-06, + "logits/chosen": 0.0654296875, + "logits/rejected": -0.06103515625, + "logps/chosen": -378.0, + "logps/rejected": -346.0, + "loss": 0.3699, + "rewards/accuracies": 0.796875, + "rewards/chosen": -0.02880859375, + "rewards/margins": 6.53125, + "rewards/rejected": -6.5625, + "step": 339 + }, + { + "epoch": 0.4171779141104294, + "grad_norm": 35.90111385254676, + "learning_rate": 1.3047520661157025e-06, + "logits/chosen": -0.03759765625, + "logits/rejected": -0.267578125, + "logps/chosen": -366.0, + "logps/rejected": -350.0, + "loss": 0.3448, + "rewards/accuracies": 0.828125, + "rewards/chosen": -0.058349609375, + "rewards/margins": 6.125, + "rewards/rejected": -6.1875, + "step": 340 + }, + { + "epoch": 0.41840490797546015, + "grad_norm": 27.984626902771293, + "learning_rate": 1.3041322314049587e-06, + "logits/chosen": -0.09521484375, + "logits/rejected": -0.3125, + "logps/chosen": -398.0, + "logps/rejected": -334.0, + "loss": 0.344, + "rewards/accuracies": 0.8515625, + "rewards/chosen": 0.56640625, + "rewards/margins": 7.3125, + "rewards/rejected": -6.75, + "step": 341 + }, + { + "epoch": 0.4196319018404908, + "grad_norm": 34.66561631382139, + "learning_rate": 1.3035123966942148e-06, + "logits/chosen": -0.021240234375, + "logits/rejected": -0.119140625, + "logps/chosen": -340.0, + "logps/rejected": -318.0, + "loss": 0.3745, + "rewards/accuracies": 0.78125, + "rewards/chosen": 0.049072265625, + "rewards/margins": 6.09375, + "rewards/rejected": -6.03125, + "step": 342 + }, + { + "epoch": 0.4208588957055215, + "grad_norm": 34.35806991693024, + "learning_rate": 1.302892561983471e-06, + "logits/chosen": -0.06396484375, + "logits/rejected": -0.1513671875, + "logps/chosen": -386.0, + "logps/rejected": -356.0, + "loss": 0.4138, + "rewards/accuracies": 0.8046875, + "rewards/chosen": -0.251953125, + "rewards/margins": 5.53125, + "rewards/rejected": -5.8125, + "step": 343 + }, + { + "epoch": 0.42208588957055215, + "grad_norm": 36.57762286698119, + "learning_rate": 1.3022727272727273e-06, + "logits/chosen": -0.0654296875, + "logits/rejected": -0.2041015625, + "logps/chosen": -408.0, + "logps/rejected": -340.0, + "loss": 0.4341, + "rewards/accuracies": 0.8125, + "rewards/chosen": 0.036376953125, + "rewards/margins": 5.875, + "rewards/rejected": -5.84375, + "step": 344 + }, + { + "epoch": 0.4233128834355828, + "grad_norm": 32.698706365667604, + "learning_rate": 1.3016528925619835e-06, + "logits/chosen": -0.04931640625, + "logits/rejected": -0.2265625, + "logps/chosen": -414.0, + "logps/rejected": -340.0, + "loss": 0.3793, + "rewards/accuracies": 0.84375, + "rewards/chosen": 0.2177734375, + "rewards/margins": 6.4375, + "rewards/rejected": -6.21875, + "step": 345 + }, + { + "epoch": 0.4245398773006135, + "grad_norm": 34.95248105852957, + "learning_rate": 1.3010330578512398e-06, + "logits/chosen": -0.036376953125, + "logits/rejected": -0.23046875, + "logps/chosen": -344.0, + "logps/rejected": -306.0, + "loss": 0.4481, + "rewards/accuracies": 0.828125, + "rewards/chosen": -0.19921875, + "rewards/margins": 5.84375, + "rewards/rejected": -6.03125, + "step": 346 + }, + { + "epoch": 0.42576687116564416, + "grad_norm": 27.628512071923325, + "learning_rate": 1.300413223140496e-06, + "logits/chosen": -0.06884765625, + "logits/rejected": -0.1875, + "logps/chosen": -370.0, + "logps/rejected": -356.0, + "loss": 0.317, + "rewards/accuracies": 0.8515625, + "rewards/chosen": 0.20703125, + "rewards/margins": 6.125, + "rewards/rejected": -5.90625, + "step": 347 + }, + { + "epoch": 0.4269938650306748, + "grad_norm": 24.80952384589789, + "learning_rate": 1.299793388429752e-06, + "logits/chosen": -0.12451171875, + "logits/rejected": -0.318359375, + "logps/chosen": -418.0, + "logps/rejected": -328.0, + "loss": 0.2563, + "rewards/accuracies": 0.8671875, + "rewards/chosen": 0.126953125, + "rewards/margins": 7.4375, + "rewards/rejected": -7.3125, + "step": 348 + }, + { + "epoch": 0.42822085889570555, + "grad_norm": 38.377577445863466, + "learning_rate": 1.2991735537190084e-06, + "logits/chosen": -0.0986328125, + "logits/rejected": -0.2412109375, + "logps/chosen": -406.0, + "logps/rejected": -322.0, + "loss": 0.4136, + "rewards/accuracies": 0.8046875, + "rewards/chosen": 0.046875, + "rewards/margins": 6.90625, + "rewards/rejected": -6.84375, + "step": 349 + }, + { + "epoch": 0.4294478527607362, + "grad_norm": 37.75213306544608, + "learning_rate": 1.2985537190082644e-06, + "logits/chosen": 0.06689453125, + "logits/rejected": -0.12890625, + "logps/chosen": -412.0, + "logps/rejected": -318.0, + "loss": 0.3697, + "rewards/accuracies": 0.796875, + "rewards/chosen": 0.045654296875, + "rewards/margins": 6.03125, + "rewards/rejected": -6.0, + "step": 350 + }, + { + "epoch": 0.4306748466257669, + "grad_norm": 35.53586893463626, + "learning_rate": 1.2979338842975207e-06, + "logits/chosen": -0.11279296875, + "logits/rejected": -0.251953125, + "logps/chosen": -342.0, + "logps/rejected": -326.0, + "loss": 0.4086, + "rewards/accuracies": 0.8125, + "rewards/chosen": -0.1337890625, + "rewards/margins": 6.03125, + "rewards/rejected": -6.1875, + "step": 351 + }, + { + "epoch": 0.43190184049079755, + "grad_norm": 30.59296136721169, + "learning_rate": 1.2973140495867769e-06, + "logits/chosen": -0.0230712890625, + "logits/rejected": -0.2177734375, + "logps/chosen": -386.0, + "logps/rejected": -334.0, + "loss": 0.4358, + "rewards/accuracies": 0.828125, + "rewards/chosen": 1.0625, + "rewards/margins": 6.34375, + "rewards/rejected": -5.28125, + "step": 352 + }, + { + "epoch": 0.4331288343558282, + "grad_norm": 24.76896622328251, + "learning_rate": 1.296694214876033e-06, + "logits/chosen": -0.07666015625, + "logits/rejected": -0.244140625, + "logps/chosen": -360.0, + "logps/rejected": -330.0, + "loss": 0.2411, + "rewards/accuracies": 0.859375, + "rewards/chosen": 0.921875, + "rewards/margins": 7.46875, + "rewards/rejected": -6.5625, + "step": 353 + }, + { + "epoch": 0.4343558282208589, + "grad_norm": 31.01854459387497, + "learning_rate": 1.2960743801652894e-06, + "logits/chosen": 0.01043701171875, + "logits/rejected": -0.162109375, + "logps/chosen": -444.0, + "logps/rejected": -378.0, + "loss": 0.3525, + "rewards/accuracies": 0.8359375, + "rewards/chosen": 0.99609375, + "rewards/margins": 6.53125, + "rewards/rejected": -5.53125, + "step": 354 + }, + { + "epoch": 0.43558282208588955, + "grad_norm": 32.456178590454805, + "learning_rate": 1.2954545454545455e-06, + "logits/chosen": -0.1572265625, + "logits/rejected": -0.24609375, + "logps/chosen": -348.0, + "logps/rejected": -356.0, + "loss": 0.4542, + "rewards/accuracies": 0.8671875, + "rewards/chosen": 0.296875, + "rewards/margins": 6.3125, + "rewards/rejected": -6.0, + "step": 355 + }, + { + "epoch": 0.4368098159509202, + "grad_norm": 31.595656958351967, + "learning_rate": 1.2948347107438016e-06, + "logits/chosen": 0.06396484375, + "logits/rejected": -0.0859375, + "logps/chosen": -396.0, + "logps/rejected": -332.0, + "loss": 0.4115, + "rewards/accuracies": 0.8046875, + "rewards/chosen": 0.46484375, + "rewards/margins": 6.0625, + "rewards/rejected": -5.59375, + "step": 356 + }, + { + "epoch": 0.43803680981595094, + "grad_norm": 28.72400215925028, + "learning_rate": 1.294214876033058e-06, + "logits/chosen": -0.23046875, + "logits/rejected": -0.384765625, + "logps/chosen": -402.0, + "logps/rejected": -340.0, + "loss": 0.3662, + "rewards/accuracies": 0.875, + "rewards/chosen": 0.63671875, + "rewards/margins": 7.40625, + "rewards/rejected": -6.78125, + "step": 357 + }, + { + "epoch": 0.4392638036809816, + "grad_norm": 29.5841283121426, + "learning_rate": 1.2935950413223141e-06, + "logits/chosen": -0.1708984375, + "logits/rejected": -0.328125, + "logps/chosen": -446.0, + "logps/rejected": -330.0, + "loss": 0.3109, + "rewards/accuracies": 0.8671875, + "rewards/chosen": 1.2109375, + "rewards/margins": 7.125, + "rewards/rejected": -5.90625, + "step": 358 + }, + { + "epoch": 0.4404907975460123, + "grad_norm": 33.37778751879116, + "learning_rate": 1.2929752066115703e-06, + "logits/chosen": -0.043212890625, + "logits/rejected": -0.154296875, + "logps/chosen": -424.0, + "logps/rejected": -334.0, + "loss": 0.3631, + "rewards/accuracies": 0.8515625, + "rewards/chosen": 0.89453125, + "rewards/margins": 6.71875, + "rewards/rejected": -5.8125, + "step": 359 + }, + { + "epoch": 0.44171779141104295, + "grad_norm": 32.72250588219352, + "learning_rate": 1.2923553719008264e-06, + "logits/chosen": -0.02587890625, + "logits/rejected": -0.1484375, + "logps/chosen": -396.0, + "logps/rejected": -358.0, + "loss": 0.2778, + "rewards/accuracies": 0.8515625, + "rewards/chosen": 0.33984375, + "rewards/margins": 6.9375, + "rewards/rejected": -6.59375, + "step": 360 + }, + { + "epoch": 0.4429447852760736, + "grad_norm": 29.99311979639803, + "learning_rate": 1.2917355371900826e-06, + "logits/chosen": 0.010498046875, + "logits/rejected": -0.177734375, + "logps/chosen": -438.0, + "logps/rejected": -382.0, + "loss": 0.3143, + "rewards/accuracies": 0.859375, + "rewards/chosen": 0.6015625, + "rewards/margins": 7.5, + "rewards/rejected": -6.875, + "step": 361 + }, + { + "epoch": 0.4441717791411043, + "grad_norm": 30.82537911112007, + "learning_rate": 1.291115702479339e-06, + "logits/chosen": -0.11083984375, + "logits/rejected": -0.267578125, + "logps/chosen": -418.0, + "logps/rejected": -344.0, + "loss": 0.3606, + "rewards/accuracies": 0.8203125, + "rewards/chosen": 0.62109375, + "rewards/margins": 7.21875, + "rewards/rejected": -6.59375, + "step": 362 + }, + { + "epoch": 0.44539877300613495, + "grad_norm": 26.541431319509446, + "learning_rate": 1.290495867768595e-06, + "logits/chosen": -0.018798828125, + "logits/rejected": -0.171875, + "logps/chosen": -392.0, + "logps/rejected": -344.0, + "loss": 0.2712, + "rewards/accuracies": 0.8984375, + "rewards/chosen": 0.77734375, + "rewards/margins": 6.8125, + "rewards/rejected": -6.03125, + "step": 363 + }, + { + "epoch": 0.4466257668711656, + "grad_norm": 36.761947246861894, + "learning_rate": 1.2898760330578512e-06, + "logits/chosen": 0.06591796875, + "logits/rejected": -0.109375, + "logps/chosen": -392.0, + "logps/rejected": -314.0, + "loss": 0.3953, + "rewards/accuracies": 0.8671875, + "rewards/chosen": 0.765625, + "rewards/margins": 6.75, + "rewards/rejected": -6.0, + "step": 364 + }, + { + "epoch": 0.44785276073619634, + "grad_norm": 27.87437703363369, + "learning_rate": 1.2892561983471076e-06, + "logits/chosen": -0.0018463134765625, + "logits/rejected": -0.1435546875, + "logps/chosen": -352.0, + "logps/rejected": -314.0, + "loss": 0.3086, + "rewards/accuracies": 0.8828125, + "rewards/chosen": 0.53125, + "rewards/margins": 6.65625, + "rewards/rejected": -6.125, + "step": 365 + }, + { + "epoch": 0.449079754601227, + "grad_norm": 32.787747807253844, + "learning_rate": 1.2886363636363637e-06, + "logits/chosen": 0.0291748046875, + "logits/rejected": -0.06787109375, + "logps/chosen": -366.0, + "logps/rejected": -332.0, + "loss": 0.447, + "rewards/accuracies": 0.8203125, + "rewards/chosen": 0.87890625, + "rewards/margins": 6.0, + "rewards/rejected": -5.125, + "step": 366 + }, + { + "epoch": 0.4503067484662577, + "grad_norm": 26.662012319210124, + "learning_rate": 1.2880165289256198e-06, + "logits/chosen": -0.142578125, + "logits/rejected": -0.326171875, + "logps/chosen": -394.0, + "logps/rejected": -336.0, + "loss": 0.2357, + "rewards/accuracies": 0.8984375, + "rewards/chosen": 0.9453125, + "rewards/margins": 6.65625, + "rewards/rejected": -5.71875, + "step": 367 + }, + { + "epoch": 0.45153374233128835, + "grad_norm": 36.83996177643189, + "learning_rate": 1.287396694214876e-06, + "logits/chosen": 0.06787109375, + "logits/rejected": 0.0234375, + "logps/chosen": -368.0, + "logps/rejected": -328.0, + "loss": 0.3909, + "rewards/accuracies": 0.84375, + "rewards/chosen": 0.6796875, + "rewards/margins": 6.0, + "rewards/rejected": -5.3125, + "step": 368 + }, + { + "epoch": 0.452760736196319, + "grad_norm": 35.726856908761434, + "learning_rate": 1.2867768595041321e-06, + "logits/chosen": 0.05322265625, + "logits/rejected": -0.06591796875, + "logps/chosen": -344.0, + "logps/rejected": -298.0, + "loss": 0.3738, + "rewards/accuracies": 0.8359375, + "rewards/chosen": 0.169921875, + "rewards/margins": 5.71875, + "rewards/rejected": -5.5625, + "step": 369 + }, + { + "epoch": 0.4539877300613497, + "grad_norm": 31.108323501043053, + "learning_rate": 1.2861570247933885e-06, + "logits/chosen": -0.1484375, + "logits/rejected": -0.328125, + "logps/chosen": -402.0, + "logps/rejected": -348.0, + "loss": 0.369, + "rewards/accuracies": 0.84375, + "rewards/chosen": 1.75, + "rewards/margins": 6.65625, + "rewards/rejected": -4.90625, + "step": 370 + }, + { + "epoch": 0.45521472392638035, + "grad_norm": 31.06340590217051, + "learning_rate": 1.2855371900826446e-06, + "logits/chosen": -0.130859375, + "logits/rejected": -0.298828125, + "logps/chosen": -460.0, + "logps/rejected": -364.0, + "loss": 0.3181, + "rewards/accuracies": 0.8984375, + "rewards/chosen": 1.625, + "rewards/margins": 6.0, + "rewards/rejected": -4.375, + "step": 371 + }, + { + "epoch": 0.456441717791411, + "grad_norm": 28.189166139003635, + "learning_rate": 1.2849173553719008e-06, + "logits/chosen": -0.0301513671875, + "logits/rejected": -0.1875, + "logps/chosen": -410.0, + "logps/rejected": -324.0, + "loss": 0.2419, + "rewards/accuracies": 0.890625, + "rewards/chosen": 1.5078125, + "rewards/margins": 6.4375, + "rewards/rejected": -4.9375, + "step": 372 + }, + { + "epoch": 0.45766871165644174, + "grad_norm": 33.21287873989905, + "learning_rate": 1.2842975206611571e-06, + "logits/chosen": -0.053466796875, + "logits/rejected": -0.1591796875, + "logps/chosen": -362.0, + "logps/rejected": -322.0, + "loss": 0.4071, + "rewards/accuracies": 0.8359375, + "rewards/chosen": 0.58203125, + "rewards/margins": 5.71875, + "rewards/rejected": -5.125, + "step": 373 + }, + { + "epoch": 0.4588957055214724, + "grad_norm": 26.151862774833457, + "learning_rate": 1.2836776859504132e-06, + "logits/chosen": -0.0908203125, + "logits/rejected": -0.2197265625, + "logps/chosen": -382.0, + "logps/rejected": -316.0, + "loss": 0.2534, + "rewards/accuracies": 0.8984375, + "rewards/chosen": 2.0, + "rewards/margins": 6.78125, + "rewards/rejected": -4.78125, + "step": 374 + }, + { + "epoch": 0.4601226993865031, + "grad_norm": 32.35488575008133, + "learning_rate": 1.2830578512396696e-06, + "logits/chosen": -0.14453125, + "logits/rejected": -0.2177734375, + "logps/chosen": -392.0, + "logps/rejected": -350.0, + "loss": 0.2771, + "rewards/accuracies": 0.859375, + "rewards/chosen": 1.7109375, + "rewards/margins": 5.8125, + "rewards/rejected": -4.09375, + "step": 375 + }, + { + "epoch": 0.46134969325153374, + "grad_norm": 34.001922340864844, + "learning_rate": 1.2824380165289257e-06, + "logits/chosen": -0.00946044921875, + "logits/rejected": -0.15625, + "logps/chosen": -376.0, + "logps/rejected": -344.0, + "loss": 0.3408, + "rewards/accuracies": 0.8515625, + "rewards/chosen": 1.6953125, + "rewards/margins": 5.75, + "rewards/rejected": -4.03125, + "step": 376 + }, + { + "epoch": 0.4625766871165644, + "grad_norm": 35.73955039729585, + "learning_rate": 1.2818181818181817e-06, + "logits/chosen": -0.0927734375, + "logits/rejected": -0.150390625, + "logps/chosen": -352.0, + "logps/rejected": -318.0, + "loss": 0.4247, + "rewards/accuracies": 0.7734375, + "rewards/chosen": 1.5625, + "rewards/margins": 6.65625, + "rewards/rejected": -5.0625, + "step": 377 + }, + { + "epoch": 0.4638036809815951, + "grad_norm": 33.82458737973786, + "learning_rate": 1.281198347107438e-06, + "logits/chosen": -0.1396484375, + "logits/rejected": -0.376953125, + "logps/chosen": -376.0, + "logps/rejected": -332.0, + "loss": 0.3987, + "rewards/accuracies": 0.796875, + "rewards/chosen": 1.625, + "rewards/margins": 6.625, + "rewards/rejected": -5.0, + "step": 378 + }, + { + "epoch": 0.46503067484662575, + "grad_norm": 31.86429872362659, + "learning_rate": 1.2805785123966942e-06, + "logits/chosen": -0.1953125, + "logits/rejected": -0.359375, + "logps/chosen": -408.0, + "logps/rejected": -344.0, + "loss": 0.3028, + "rewards/accuracies": 0.8671875, + "rewards/chosen": 1.71875, + "rewards/margins": 7.1875, + "rewards/rejected": -5.46875, + "step": 379 + }, + { + "epoch": 0.4662576687116564, + "grad_norm": 31.446849732786838, + "learning_rate": 1.2799586776859505e-06, + "logits/chosen": -0.09326171875, + "logits/rejected": -0.2314453125, + "logps/chosen": -382.0, + "logps/rejected": -326.0, + "loss": 0.3068, + "rewards/accuracies": 0.859375, + "rewards/chosen": 1.625, + "rewards/margins": 6.78125, + "rewards/rejected": -5.15625, + "step": 380 + }, + { + "epoch": 0.46748466257668714, + "grad_norm": 31.679856008810177, + "learning_rate": 1.2793388429752067e-06, + "logits/chosen": -0.189453125, + "logits/rejected": -0.3671875, + "logps/chosen": -372.0, + "logps/rejected": -328.0, + "loss": 0.374, + "rewards/accuracies": 0.8515625, + "rewards/chosen": 1.046875, + "rewards/margins": 7.0, + "rewards/rejected": -5.9375, + "step": 381 + }, + { + "epoch": 0.4687116564417178, + "grad_norm": 31.436122523167693, + "learning_rate": 1.2787190082644628e-06, + "logits/chosen": -0.185546875, + "logits/rejected": -0.279296875, + "logps/chosen": -380.0, + "logps/rejected": -340.0, + "loss": 0.3158, + "rewards/accuracies": 0.8515625, + "rewards/chosen": 1.078125, + "rewards/margins": 6.875, + "rewards/rejected": -5.8125, + "step": 382 + }, + { + "epoch": 0.4699386503067485, + "grad_norm": 35.065632595795535, + "learning_rate": 1.2780991735537192e-06, + "logits/chosen": -0.09912109375, + "logits/rejected": -0.2578125, + "logps/chosen": -402.0, + "logps/rejected": -378.0, + "loss": 0.4456, + "rewards/accuracies": 0.8359375, + "rewards/chosen": 1.1015625, + "rewards/margins": 6.78125, + "rewards/rejected": -5.6875, + "step": 383 + }, + { + "epoch": 0.47116564417177914, + "grad_norm": 33.79101177267741, + "learning_rate": 1.2774793388429753e-06, + "logits/chosen": -0.2001953125, + "logits/rejected": -0.435546875, + "logps/chosen": -398.0, + "logps/rejected": -342.0, + "loss": 0.2961, + "rewards/accuracies": 0.890625, + "rewards/chosen": 0.640625, + "rewards/margins": 7.6875, + "rewards/rejected": -7.03125, + "step": 384 + }, + { + "epoch": 0.4723926380368098, + "grad_norm": 36.518171637968166, + "learning_rate": 1.2768595041322314e-06, + "logits/chosen": -0.10302734375, + "logits/rejected": -0.283203125, + "logps/chosen": -394.0, + "logps/rejected": -330.0, + "loss": 0.3771, + "rewards/accuracies": 0.8046875, + "rewards/chosen": 0.283203125, + "rewards/margins": 6.34375, + "rewards/rejected": -6.0625, + "step": 385 + }, + { + "epoch": 0.4736196319018405, + "grad_norm": 34.95682432591222, + "learning_rate": 1.2762396694214876e-06, + "logits/chosen": -0.043701171875, + "logits/rejected": -0.2041015625, + "logps/chosen": -342.0, + "logps/rejected": -310.0, + "loss": 0.369, + "rewards/accuracies": 0.875, + "rewards/chosen": -0.3203125, + "rewards/margins": 6.09375, + "rewards/rejected": -6.40625, + "step": 386 + }, + { + "epoch": 0.47484662576687114, + "grad_norm": 64.7195376952873, + "learning_rate": 1.2756198347107437e-06, + "logits/chosen": -0.068359375, + "logits/rejected": -0.271484375, + "logps/chosen": -422.0, + "logps/rejected": -350.0, + "loss": 0.4035, + "rewards/accuracies": 0.8359375, + "rewards/chosen": 0.6171875, + "rewards/margins": 7.34375, + "rewards/rejected": -6.75, + "step": 387 + }, + { + "epoch": 0.47607361963190187, + "grad_norm": 42.357455920572896, + "learning_rate": 1.275e-06, + "logits/chosen": -0.08349609375, + "logits/rejected": -0.1826171875, + "logps/chosen": -406.0, + "logps/rejected": -326.0, + "loss": 0.5163, + "rewards/accuracies": 0.7890625, + "rewards/chosen": -0.04736328125, + "rewards/margins": 5.53125, + "rewards/rejected": -5.59375, + "step": 388 + }, + { + "epoch": 0.47730061349693254, + "grad_norm": 38.953026072404185, + "learning_rate": 1.2743801652892562e-06, + "logits/chosen": -0.1435546875, + "logits/rejected": -0.2236328125, + "logps/chosen": -412.0, + "logps/rejected": -340.0, + "loss": 0.4455, + "rewards/accuracies": 0.8203125, + "rewards/chosen": 0.099609375, + "rewards/margins": 6.3125, + "rewards/rejected": -6.1875, + "step": 389 + }, + { + "epoch": 0.4785276073619632, + "grad_norm": 38.5573431700193, + "learning_rate": 1.2737603305785124e-06, + "logits/chosen": -0.2216796875, + "logits/rejected": -0.34375, + "logps/chosen": -372.0, + "logps/rejected": -344.0, + "loss": 0.492, + "rewards/accuracies": 0.828125, + "rewards/chosen": -0.62890625, + "rewards/margins": 6.0, + "rewards/rejected": -6.65625, + "step": 390 + }, + { + "epoch": 0.47975460122699387, + "grad_norm": 37.845045364551986, + "learning_rate": 1.2731404958677687e-06, + "logits/chosen": -0.1201171875, + "logits/rejected": -0.23828125, + "logps/chosen": -390.0, + "logps/rejected": -344.0, + "loss": 0.4807, + "rewards/accuracies": 0.8203125, + "rewards/chosen": -0.181640625, + "rewards/margins": 6.4375, + "rewards/rejected": -6.625, + "step": 391 + }, + { + "epoch": 0.48098159509202454, + "grad_norm": 29.781954844950388, + "learning_rate": 1.2725206611570249e-06, + "logits/chosen": -0.060791015625, + "logits/rejected": -0.208984375, + "logps/chosen": -362.0, + "logps/rejected": -330.0, + "loss": 0.3103, + "rewards/accuracies": 0.875, + "rewards/chosen": -0.06982421875, + "rewards/margins": 6.59375, + "rewards/rejected": -6.65625, + "step": 392 + }, + { + "epoch": 0.4822085889570552, + "grad_norm": 33.5926238632541, + "learning_rate": 1.271900826446281e-06, + "logits/chosen": -0.0286865234375, + "logits/rejected": -0.166015625, + "logps/chosen": -416.0, + "logps/rejected": -374.0, + "loss": 0.3323, + "rewards/accuracies": 0.8359375, + "rewards/chosen": -0.263671875, + "rewards/margins": 6.4375, + "rewards/rejected": -6.6875, + "step": 393 + }, + { + "epoch": 0.4834355828220859, + "grad_norm": 32.087849602750985, + "learning_rate": 1.2712809917355373e-06, + "logits/chosen": -0.09521484375, + "logits/rejected": -0.298828125, + "logps/chosen": -398.0, + "logps/rejected": -342.0, + "loss": 0.262, + "rewards/accuracies": 0.84375, + "rewards/chosen": -0.166015625, + "rewards/margins": 6.1875, + "rewards/rejected": -6.375, + "step": 394 + }, + { + "epoch": 0.48466257668711654, + "grad_norm": 34.842476262528464, + "learning_rate": 1.2706611570247933e-06, + "logits/chosen": -0.1025390625, + "logits/rejected": -0.216796875, + "logps/chosen": -448.0, + "logps/rejected": -356.0, + "loss": 0.3764, + "rewards/accuracies": 0.8046875, + "rewards/chosen": -0.09375, + "rewards/margins": 6.125, + "rewards/rejected": -6.21875, + "step": 395 + }, + { + "epoch": 0.48588957055214727, + "grad_norm": 22.42844686295793, + "learning_rate": 1.2700413223140496e-06, + "logits/chosen": -0.1806640625, + "logits/rejected": -0.388671875, + "logps/chosen": -348.0, + "logps/rejected": -332.0, + "loss": 0.2277, + "rewards/accuracies": 0.90625, + "rewards/chosen": -0.54296875, + "rewards/margins": 6.3125, + "rewards/rejected": -6.875, + "step": 396 + }, + { + "epoch": 0.48711656441717793, + "grad_norm": 33.97555781703141, + "learning_rate": 1.2694214876033058e-06, + "logits/chosen": -0.1513671875, + "logits/rejected": -0.314453125, + "logps/chosen": -388.0, + "logps/rejected": -350.0, + "loss": 0.3848, + "rewards/accuracies": 0.828125, + "rewards/chosen": -0.625, + "rewards/margins": 6.78125, + "rewards/rejected": -7.40625, + "step": 397 + }, + { + "epoch": 0.4883435582822086, + "grad_norm": 39.08030474072834, + "learning_rate": 1.268801652892562e-06, + "logits/chosen": -0.1328125, + "logits/rejected": -0.27734375, + "logps/chosen": -428.0, + "logps/rejected": -346.0, + "loss": 0.4427, + "rewards/accuracies": 0.796875, + "rewards/chosen": -0.08740234375, + "rewards/margins": 6.375, + "rewards/rejected": -6.46875, + "step": 398 + }, + { + "epoch": 0.48957055214723927, + "grad_norm": 33.83840388728663, + "learning_rate": 1.2681818181818183e-06, + "logits/chosen": -0.1611328125, + "logits/rejected": -0.291015625, + "logps/chosen": -394.0, + "logps/rejected": -346.0, + "loss": 0.2963, + "rewards/accuracies": 0.859375, + "rewards/chosen": -0.0035858154296875, + "rewards/margins": 6.53125, + "rewards/rejected": -6.53125, + "step": 399 + }, + { + "epoch": 0.49079754601226994, + "grad_norm": 41.49480537946823, + "learning_rate": 1.2675619834710744e-06, + "logits/chosen": -0.09130859375, + "logits/rejected": -0.1650390625, + "logps/chosen": -384.0, + "logps/rejected": -368.0, + "loss": 0.5062, + "rewards/accuracies": 0.796875, + "rewards/chosen": -0.5390625, + "rewards/margins": 5.59375, + "rewards/rejected": -6.15625, + "step": 400 + }, + { + "epoch": 0.4920245398773006, + "grad_norm": 27.800803034272004, + "learning_rate": 1.2669421487603308e-06, + "logits/chosen": -0.23828125, + "logits/rejected": -0.373046875, + "logps/chosen": -456.0, + "logps/rejected": -390.0, + "loss": 0.2938, + "rewards/accuracies": 0.8671875, + "rewards/chosen": 0.40625, + "rewards/margins": 7.59375, + "rewards/rejected": -7.1875, + "step": 401 + }, + { + "epoch": 0.49325153374233127, + "grad_norm": 32.09882908595564, + "learning_rate": 1.266322314049587e-06, + "logits/chosen": -0.0791015625, + "logits/rejected": -0.1796875, + "logps/chosen": -394.0, + "logps/rejected": -350.0, + "loss": 0.349, + "rewards/accuracies": 0.8359375, + "rewards/chosen": -0.04443359375, + "rewards/margins": 6.5, + "rewards/rejected": -6.53125, + "step": 402 + }, + { + "epoch": 0.49447852760736194, + "grad_norm": 29.930227321008882, + "learning_rate": 1.265702479338843e-06, + "logits/chosen": -0.07958984375, + "logits/rejected": -0.15625, + "logps/chosen": -396.0, + "logps/rejected": -374.0, + "loss": 0.341, + "rewards/accuracies": 0.8671875, + "rewards/chosen": -0.033447265625, + "rewards/margins": 6.78125, + "rewards/rejected": -6.8125, + "step": 403 + }, + { + "epoch": 0.49570552147239266, + "grad_norm": 25.987833840695732, + "learning_rate": 1.2650826446280992e-06, + "logits/chosen": -0.021728515625, + "logits/rejected": -0.150390625, + "logps/chosen": -372.0, + "logps/rejected": -348.0, + "loss": 0.2606, + "rewards/accuracies": 0.8984375, + "rewards/chosen": 0.1669921875, + "rewards/margins": 7.0625, + "rewards/rejected": -6.90625, + "step": 404 + }, + { + "epoch": 0.49693251533742333, + "grad_norm": 38.14679307536266, + "learning_rate": 1.2644628099173553e-06, + "logits/chosen": -0.1064453125, + "logits/rejected": -0.328125, + "logps/chosen": -370.0, + "logps/rejected": -322.0, + "loss": 0.4214, + "rewards/accuracies": 0.8125, + "rewards/chosen": 0.6484375, + "rewards/margins": 6.65625, + "rewards/rejected": -6.03125, + "step": 405 + }, + { + "epoch": 0.498159509202454, + "grad_norm": 37.82975021877776, + "learning_rate": 1.2638429752066117e-06, + "logits/chosen": 0.0166015625, + "logits/rejected": -0.0625, + "logps/chosen": -378.0, + "logps/rejected": -340.0, + "loss": 0.5427, + "rewards/accuracies": 0.8046875, + "rewards/chosen": 0.3515625, + "rewards/margins": 6.09375, + "rewards/rejected": -5.75, + "step": 406 + }, + { + "epoch": 0.49938650306748467, + "grad_norm": 38.332851373691135, + "learning_rate": 1.2632231404958678e-06, + "logits/chosen": 0.0115966796875, + "logits/rejected": -0.134765625, + "logps/chosen": -408.0, + "logps/rejected": -344.0, + "loss": 0.4382, + "rewards/accuracies": 0.8046875, + "rewards/chosen": 1.34375, + "rewards/margins": 7.40625, + "rewards/rejected": -6.0625, + "step": 407 + }, + { + "epoch": 0.5006134969325153, + "grad_norm": 41.6208959235071, + "learning_rate": 1.262603305785124e-06, + "logits/chosen": -0.083984375, + "logits/rejected": -0.1474609375, + "logps/chosen": -426.0, + "logps/rejected": -360.0, + "loss": 0.5081, + "rewards/accuracies": 0.7734375, + "rewards/chosen": -0.091796875, + "rewards/margins": 5.1875, + "rewards/rejected": -5.28125, + "step": 408 + }, + { + "epoch": 0.501840490797546, + "grad_norm": 42.88666645295333, + "learning_rate": 1.2619834710743803e-06, + "logits/chosen": -0.00860595703125, + "logits/rejected": -0.205078125, + "logps/chosen": -426.0, + "logps/rejected": -342.0, + "loss": 0.4119, + "rewards/accuracies": 0.8203125, + "rewards/chosen": 0.56640625, + "rewards/margins": 6.71875, + "rewards/rejected": -6.15625, + "step": 409 + }, + { + "epoch": 0.5030674846625767, + "grad_norm": 36.428858153344706, + "learning_rate": 1.2613636363636365e-06, + "logits/chosen": -0.00095367431640625, + "logits/rejected": -0.1943359375, + "logps/chosen": -388.0, + "logps/rejected": -360.0, + "loss": 0.4038, + "rewards/accuracies": 0.796875, + "rewards/chosen": 0.3828125, + "rewards/margins": 6.125, + "rewards/rejected": -5.75, + "step": 410 + }, + { + "epoch": 0.5042944785276073, + "grad_norm": 33.21094894864319, + "learning_rate": 1.2607438016528926e-06, + "logits/chosen": 0.0140380859375, + "logits/rejected": -0.12890625, + "logps/chosen": -342.0, + "logps/rejected": -320.0, + "loss": 0.3058, + "rewards/accuracies": 0.8359375, + "rewards/chosen": 0.1298828125, + "rewards/margins": 6.46875, + "rewards/rejected": -6.34375, + "step": 411 + }, + { + "epoch": 0.505521472392638, + "grad_norm": 42.15288062503197, + "learning_rate": 1.260123966942149e-06, + "logits/chosen": 0.007080078125, + "logits/rejected": -0.1259765625, + "logps/chosen": -428.0, + "logps/rejected": -368.0, + "loss": 0.4444, + "rewards/accuracies": 0.796875, + "rewards/chosen": 0.30078125, + "rewards/margins": 5.96875, + "rewards/rejected": -5.6875, + "step": 412 + }, + { + "epoch": 0.5067484662576687, + "grad_norm": 37.795734491735246, + "learning_rate": 1.2595041322314049e-06, + "logits/chosen": -0.042236328125, + "logits/rejected": -0.26171875, + "logps/chosen": -408.0, + "logps/rejected": -332.0, + "loss": 0.3878, + "rewards/accuracies": 0.8125, + "rewards/chosen": 0.9140625, + "rewards/margins": 6.96875, + "rewards/rejected": -6.0625, + "step": 413 + }, + { + "epoch": 0.5079754601226993, + "grad_norm": 37.27154344231617, + "learning_rate": 1.2588842975206612e-06, + "logits/chosen": -0.0859375, + "logits/rejected": -0.2265625, + "logps/chosen": -406.0, + "logps/rejected": -354.0, + "loss": 0.3559, + "rewards/accuracies": 0.84375, + "rewards/chosen": 0.85546875, + "rewards/margins": 6.71875, + "rewards/rejected": -5.84375, + "step": 414 + }, + { + "epoch": 0.50920245398773, + "grad_norm": 30.678885373774655, + "learning_rate": 1.2582644628099174e-06, + "logits/chosen": -0.041748046875, + "logits/rejected": -0.212890625, + "logps/chosen": -350.0, + "logps/rejected": -332.0, + "loss": 0.3048, + "rewards/accuracies": 0.828125, + "rewards/chosen": 0.52734375, + "rewards/margins": 6.75, + "rewards/rejected": -6.21875, + "step": 415 + }, + { + "epoch": 0.5104294478527608, + "grad_norm": 26.882010464050307, + "learning_rate": 1.2576446280991735e-06, + "logits/chosen": -0.004730224609375, + "logits/rejected": -0.2197265625, + "logps/chosen": -360.0, + "logps/rejected": -324.0, + "loss": 0.1899, + "rewards/accuracies": 0.890625, + "rewards/chosen": 0.458984375, + "rewards/margins": 6.75, + "rewards/rejected": -6.28125, + "step": 416 + }, + { + "epoch": 0.5116564417177915, + "grad_norm": 30.354406082685717, + "learning_rate": 1.2570247933884299e-06, + "logits/chosen": -0.2197265625, + "logits/rejected": -0.384765625, + "logps/chosen": -438.0, + "logps/rejected": -382.0, + "loss": 0.3453, + "rewards/accuracies": 0.84375, + "rewards/chosen": 1.4765625, + "rewards/margins": 8.125, + "rewards/rejected": -6.65625, + "step": 417 + }, + { + "epoch": 0.5128834355828221, + "grad_norm": 33.53763392215037, + "learning_rate": 1.256404958677686e-06, + "logits/chosen": -0.126953125, + "logits/rejected": -0.2099609375, + "logps/chosen": -368.0, + "logps/rejected": -306.0, + "loss": 0.3744, + "rewards/accuracies": 0.859375, + "rewards/chosen": 0.7890625, + "rewards/margins": 7.1875, + "rewards/rejected": -6.375, + "step": 418 + }, + { + "epoch": 0.5141104294478528, + "grad_norm": 28.10440726715766, + "learning_rate": 1.2557851239669421e-06, + "logits/chosen": -0.0546875, + "logits/rejected": -0.21875, + "logps/chosen": -380.0, + "logps/rejected": -342.0, + "loss": 0.3206, + "rewards/accuracies": 0.8671875, + "rewards/chosen": 0.99609375, + "rewards/margins": 7.78125, + "rewards/rejected": -6.78125, + "step": 419 + }, + { + "epoch": 0.5153374233128835, + "grad_norm": 30.627621824355007, + "learning_rate": 1.2551652892561985e-06, + "logits/chosen": 0.0262451171875, + "logits/rejected": -0.134765625, + "logps/chosen": -346.0, + "logps/rejected": -312.0, + "loss": 0.3514, + "rewards/accuracies": 0.828125, + "rewards/chosen": 0.796875, + "rewards/margins": 6.4375, + "rewards/rejected": -5.65625, + "step": 420 + }, + { + "epoch": 0.5165644171779141, + "grad_norm": 31.844076287237254, + "learning_rate": 1.2545454545454544e-06, + "logits/chosen": -0.01544189453125, + "logits/rejected": -0.2158203125, + "logps/chosen": -364.0, + "logps/rejected": -320.0, + "loss": 0.4123, + "rewards/accuracies": 0.8125, + "rewards/chosen": 1.1875, + "rewards/margins": 6.8125, + "rewards/rejected": -5.625, + "step": 421 + }, + { + "epoch": 0.5177914110429448, + "grad_norm": 33.522822368907185, + "learning_rate": 1.2539256198347108e-06, + "logits/chosen": 0.0595703125, + "logits/rejected": -0.0986328125, + "logps/chosen": -360.0, + "logps/rejected": -302.0, + "loss": 0.3724, + "rewards/accuracies": 0.84375, + "rewards/chosen": 0.52734375, + "rewards/margins": 6.6875, + "rewards/rejected": -6.15625, + "step": 422 + }, + { + "epoch": 0.5190184049079755, + "grad_norm": 37.40841671544669, + "learning_rate": 1.253305785123967e-06, + "logits/chosen": -0.0595703125, + "logits/rejected": -0.20703125, + "logps/chosen": -398.0, + "logps/rejected": -362.0, + "loss": 0.4749, + "rewards/accuracies": 0.8125, + "rewards/chosen": 0.578125, + "rewards/margins": 6.0, + "rewards/rejected": -5.4375, + "step": 423 + }, + { + "epoch": 0.5202453987730061, + "grad_norm": 34.546118881756456, + "learning_rate": 1.252685950413223e-06, + "logits/chosen": 0.0235595703125, + "logits/rejected": -0.166015625, + "logps/chosen": -378.0, + "logps/rejected": -342.0, + "loss": 0.4503, + "rewards/accuracies": 0.8359375, + "rewards/chosen": 1.2109375, + "rewards/margins": 7.09375, + "rewards/rejected": -5.875, + "step": 424 + }, + { + "epoch": 0.5214723926380368, + "grad_norm": 37.84213132068162, + "learning_rate": 1.2520661157024794e-06, + "logits/chosen": -0.0123291015625, + "logits/rejected": -0.11572265625, + "logps/chosen": -396.0, + "logps/rejected": -348.0, + "loss": 0.5144, + "rewards/accuracies": 0.796875, + "rewards/chosen": 1.3203125, + "rewards/margins": 6.84375, + "rewards/rejected": -5.53125, + "step": 425 + }, + { + "epoch": 0.5226993865030675, + "grad_norm": 36.174280199739684, + "learning_rate": 1.2514462809917356e-06, + "logits/chosen": -0.06396484375, + "logits/rejected": -0.166015625, + "logps/chosen": -370.0, + "logps/rejected": -348.0, + "loss": 0.4214, + "rewards/accuracies": 0.875, + "rewards/chosen": 1.015625, + "rewards/margins": 7.0625, + "rewards/rejected": -6.0625, + "step": 426 + }, + { + "epoch": 0.5239263803680981, + "grad_norm": 45.37106746020955, + "learning_rate": 1.2508264462809917e-06, + "logits/chosen": -0.01483154296875, + "logits/rejected": -0.1279296875, + "logps/chosen": -378.0, + "logps/rejected": -334.0, + "loss": 0.4318, + "rewards/accuracies": 0.8203125, + "rewards/chosen": 1.1171875, + "rewards/margins": 6.75, + "rewards/rejected": -5.625, + "step": 427 + }, + { + "epoch": 0.5251533742331288, + "grad_norm": 33.589023647082186, + "learning_rate": 1.250206611570248e-06, + "logits/chosen": -0.0419921875, + "logits/rejected": -0.1533203125, + "logps/chosen": -364.0, + "logps/rejected": -324.0, + "loss": 0.4488, + "rewards/accuracies": 0.8046875, + "rewards/chosen": 0.462890625, + "rewards/margins": 6.46875, + "rewards/rejected": -6.0, + "step": 428 + }, + { + "epoch": 0.5263803680981595, + "grad_norm": 30.11424815064351, + "learning_rate": 1.2495867768595042e-06, + "logits/chosen": -0.044677734375, + "logits/rejected": -0.1494140625, + "logps/chosen": -360.0, + "logps/rejected": -376.0, + "loss": 0.2639, + "rewards/accuracies": 0.8671875, + "rewards/chosen": 0.373046875, + "rewards/margins": 7.96875, + "rewards/rejected": -7.59375, + "step": 429 + }, + { + "epoch": 0.5276073619631901, + "grad_norm": 23.36575464672381, + "learning_rate": 1.2489669421487603e-06, + "logits/chosen": -0.11767578125, + "logits/rejected": -0.28125, + "logps/chosen": -394.0, + "logps/rejected": -350.0, + "loss": 0.2208, + "rewards/accuracies": 0.875, + "rewards/chosen": 0.921875, + "rewards/margins": 7.5625, + "rewards/rejected": -6.625, + "step": 430 + }, + { + "epoch": 0.5288343558282208, + "grad_norm": 39.63938191210347, + "learning_rate": 1.2483471074380165e-06, + "logits/chosen": 0.01806640625, + "logits/rejected": -0.150390625, + "logps/chosen": -386.0, + "logps/rejected": -346.0, + "loss": 0.4684, + "rewards/accuracies": 0.8203125, + "rewards/chosen": 0.12255859375, + "rewards/margins": 6.4375, + "rewards/rejected": -6.3125, + "step": 431 + }, + { + "epoch": 0.5300613496932516, + "grad_norm": 35.90106246867024, + "learning_rate": 1.2477272727272726e-06, + "logits/chosen": 0.06689453125, + "logits/rejected": 0.035888671875, + "logps/chosen": -382.0, + "logps/rejected": -354.0, + "loss": 0.4356, + "rewards/accuracies": 0.84375, + "rewards/chosen": 0.78515625, + "rewards/margins": 6.71875, + "rewards/rejected": -5.9375, + "step": 432 + }, + { + "epoch": 0.5312883435582823, + "grad_norm": 43.12860297142818, + "learning_rate": 1.247107438016529e-06, + "logits/chosen": 0.08740234375, + "logits/rejected": -0.1328125, + "logps/chosen": -388.0, + "logps/rejected": -320.0, + "loss": 0.4803, + "rewards/accuracies": 0.8046875, + "rewards/chosen": -0.12353515625, + "rewards/margins": 6.4375, + "rewards/rejected": -6.5625, + "step": 433 + }, + { + "epoch": 0.5325153374233129, + "grad_norm": 37.083326074229944, + "learning_rate": 1.2464876033057851e-06, + "logits/chosen": -0.0869140625, + "logits/rejected": -0.1611328125, + "logps/chosen": -390.0, + "logps/rejected": -356.0, + "loss": 0.3206, + "rewards/accuracies": 0.859375, + "rewards/chosen": 0.07666015625, + "rewards/margins": 6.375, + "rewards/rejected": -6.28125, + "step": 434 + }, + { + "epoch": 0.5337423312883436, + "grad_norm": 37.88064967580844, + "learning_rate": 1.2458677685950415e-06, + "logits/chosen": -0.01458740234375, + "logits/rejected": -0.208984375, + "logps/chosen": -394.0, + "logps/rejected": -362.0, + "loss": 0.3939, + "rewards/accuracies": 0.8359375, + "rewards/chosen": 0.51953125, + "rewards/margins": 7.15625, + "rewards/rejected": -6.65625, + "step": 435 + }, + { + "epoch": 0.5349693251533743, + "grad_norm": 30.102357096787667, + "learning_rate": 1.2452479338842976e-06, + "logits/chosen": -0.0250244140625, + "logits/rejected": -0.2255859375, + "logps/chosen": -422.0, + "logps/rejected": -358.0, + "loss": 0.2562, + "rewards/accuracies": 0.8828125, + "rewards/chosen": 1.1171875, + "rewards/margins": 7.75, + "rewards/rejected": -6.625, + "step": 436 + }, + { + "epoch": 0.5361963190184049, + "grad_norm": 25.03840429478704, + "learning_rate": 1.2446280991735538e-06, + "logits/chosen": -0.080078125, + "logits/rejected": -0.25, + "logps/chosen": -388.0, + "logps/rejected": -350.0, + "loss": 0.241, + "rewards/accuracies": 0.8671875, + "rewards/chosen": 0.41796875, + "rewards/margins": 7.28125, + "rewards/rejected": -6.875, + "step": 437 + }, + { + "epoch": 0.5374233128834356, + "grad_norm": 36.15225304248156, + "learning_rate": 1.24400826446281e-06, + "logits/chosen": 0.02001953125, + "logits/rejected": -0.06982421875, + "logps/chosen": -366.0, + "logps/rejected": -334.0, + "loss": 0.5276, + "rewards/accuracies": 0.8125, + "rewards/chosen": -0.0250244140625, + "rewards/margins": 5.65625, + "rewards/rejected": -5.6875, + "step": 438 + }, + { + "epoch": 0.5386503067484663, + "grad_norm": 34.58997013752512, + "learning_rate": 1.243388429752066e-06, + "logits/chosen": 0.0625, + "logits/rejected": -0.126953125, + "logps/chosen": -372.0, + "logps/rejected": -348.0, + "loss": 0.3671, + "rewards/accuracies": 0.84375, + "rewards/chosen": -0.0703125, + "rewards/margins": 6.65625, + "rewards/rejected": -6.71875, + "step": 439 + }, + { + "epoch": 0.5398773006134969, + "grad_norm": 43.86154514789445, + "learning_rate": 1.2427685950413224e-06, + "logits/chosen": 0.0218505859375, + "logits/rejected": -0.0751953125, + "logps/chosen": -392.0, + "logps/rejected": -344.0, + "loss": 0.4825, + "rewards/accuracies": 0.8046875, + "rewards/chosen": -0.515625, + "rewards/margins": 5.4375, + "rewards/rejected": -5.96875, + "step": 440 + }, + { + "epoch": 0.5411042944785276, + "grad_norm": 32.02976017157544, + "learning_rate": 1.2421487603305785e-06, + "logits/chosen": 0.09619140625, + "logits/rejected": -0.0108642578125, + "logps/chosen": -340.0, + "logps/rejected": -314.0, + "loss": 0.4408, + "rewards/accuracies": 0.828125, + "rewards/chosen": -0.0869140625, + "rewards/margins": 6.15625, + "rewards/rejected": -6.21875, + "step": 441 + }, + { + "epoch": 0.5423312883435583, + "grad_norm": 31.122024927447573, + "learning_rate": 1.2415289256198347e-06, + "logits/chosen": -0.064453125, + "logits/rejected": -0.1982421875, + "logps/chosen": -390.0, + "logps/rejected": -352.0, + "loss": 0.3842, + "rewards/accuracies": 0.859375, + "rewards/chosen": 0.51171875, + "rewards/margins": 6.5625, + "rewards/rejected": -6.0625, + "step": 442 + }, + { + "epoch": 0.5435582822085889, + "grad_norm": 35.35206800809496, + "learning_rate": 1.240909090909091e-06, + "logits/chosen": -0.024169921875, + "logits/rejected": -0.1435546875, + "logps/chosen": -372.0, + "logps/rejected": -332.0, + "loss": 0.3707, + "rewards/accuracies": 0.828125, + "rewards/chosen": 0.69140625, + "rewards/margins": 6.8125, + "rewards/rejected": -6.125, + "step": 443 + }, + { + "epoch": 0.5447852760736196, + "grad_norm": 35.692369145811284, + "learning_rate": 1.2402892561983472e-06, + "logits/chosen": -0.06201171875, + "logits/rejected": -0.298828125, + "logps/chosen": -454.0, + "logps/rejected": -382.0, + "loss": 0.3655, + "rewards/accuracies": 0.859375, + "rewards/chosen": 1.2890625, + "rewards/margins": 8.0, + "rewards/rejected": -6.71875, + "step": 444 + }, + { + "epoch": 0.5460122699386503, + "grad_norm": 32.62912831625137, + "learning_rate": 1.2396694214876033e-06, + "logits/chosen": -0.016357421875, + "logits/rejected": -0.208984375, + "logps/chosen": -358.0, + "logps/rejected": -314.0, + "loss": 0.4199, + "rewards/accuracies": 0.796875, + "rewards/chosen": 0.66796875, + "rewards/margins": 6.21875, + "rewards/rejected": -5.5625, + "step": 445 + }, + { + "epoch": 0.5472392638036809, + "grad_norm": 25.78605234898617, + "learning_rate": 1.2390495867768597e-06, + "logits/chosen": -0.040283203125, + "logits/rejected": -0.203125, + "logps/chosen": -438.0, + "logps/rejected": -370.0, + "loss": 0.2432, + "rewards/accuracies": 0.890625, + "rewards/chosen": 0.81640625, + "rewards/margins": 6.46875, + "rewards/rejected": -5.65625, + "step": 446 + }, + { + "epoch": 0.5484662576687117, + "grad_norm": 28.526549681854004, + "learning_rate": 1.2384297520661158e-06, + "logits/chosen": -0.0341796875, + "logits/rejected": -0.23828125, + "logps/chosen": -414.0, + "logps/rejected": -360.0, + "loss": 0.3087, + "rewards/accuracies": 0.8671875, + "rewards/chosen": 1.359375, + "rewards/margins": 8.1875, + "rewards/rejected": -6.84375, + "step": 447 + }, + { + "epoch": 0.5496932515337424, + "grad_norm": 33.26832004918433, + "learning_rate": 1.237809917355372e-06, + "logits/chosen": 0.042724609375, + "logits/rejected": -0.10009765625, + "logps/chosen": -346.0, + "logps/rejected": -314.0, + "loss": 0.3362, + "rewards/accuracies": 0.84375, + "rewards/chosen": 0.1796875, + "rewards/margins": 6.03125, + "rewards/rejected": -5.84375, + "step": 448 + }, + { + "epoch": 0.550920245398773, + "grad_norm": 29.824788916298026, + "learning_rate": 1.237190082644628e-06, + "logits/chosen": -0.0341796875, + "logits/rejected": -0.1650390625, + "logps/chosen": -400.0, + "logps/rejected": -346.0, + "loss": 0.3203, + "rewards/accuracies": 0.8515625, + "rewards/chosen": 1.234375, + "rewards/margins": 7.34375, + "rewards/rejected": -6.09375, + "step": 449 + }, + { + "epoch": 0.5521472392638037, + "grad_norm": 29.646416826102993, + "learning_rate": 1.2365702479338842e-06, + "logits/chosen": -0.1103515625, + "logits/rejected": -0.1494140625, + "logps/chosen": -350.0, + "logps/rejected": -302.0, + "loss": 0.2642, + "rewards/accuracies": 0.8359375, + "rewards/chosen": 0.66015625, + "rewards/margins": 6.78125, + "rewards/rejected": -6.125, + "step": 450 + }, + { + "epoch": 0.5533742331288344, + "grad_norm": 31.583737211583365, + "learning_rate": 1.2359504132231406e-06, + "logits/chosen": 0.03173828125, + "logits/rejected": -0.1396484375, + "logps/chosen": -348.0, + "logps/rejected": -294.0, + "loss": 0.3883, + "rewards/accuracies": 0.84375, + "rewards/chosen": 0.54296875, + "rewards/margins": 6.40625, + "rewards/rejected": -5.875, + "step": 451 + }, + { + "epoch": 0.554601226993865, + "grad_norm": 27.43380959875228, + "learning_rate": 1.2353305785123967e-06, + "logits/chosen": -0.0439453125, + "logits/rejected": -0.28125, + "logps/chosen": -460.0, + "logps/rejected": -342.0, + "loss": 0.2556, + "rewards/accuracies": 0.859375, + "rewards/chosen": 1.9453125, + "rewards/margins": 7.53125, + "rewards/rejected": -5.59375, + "step": 452 + }, + { + "epoch": 0.5558282208588957, + "grad_norm": 24.613642812642805, + "learning_rate": 1.2347107438016529e-06, + "logits/chosen": -0.130859375, + "logits/rejected": -0.33984375, + "logps/chosen": -440.0, + "logps/rejected": -378.0, + "loss": 0.2702, + "rewards/accuracies": 0.90625, + "rewards/chosen": 2.125, + "rewards/margins": 8.1875, + "rewards/rejected": -6.0625, + "step": 453 + }, + { + "epoch": 0.5570552147239264, + "grad_norm": 31.985648649098213, + "learning_rate": 1.2340909090909092e-06, + "logits/chosen": 0.046875, + "logits/rejected": -0.0206298828125, + "logps/chosen": -346.0, + "logps/rejected": -320.0, + "loss": 0.3305, + "rewards/accuracies": 0.8359375, + "rewards/chosen": 0.0859375, + "rewards/margins": 6.0625, + "rewards/rejected": -6.0, + "step": 454 + }, + { + "epoch": 0.558282208588957, + "grad_norm": 26.277923249806037, + "learning_rate": 1.2334710743801654e-06, + "logits/chosen": -0.0771484375, + "logits/rejected": -0.201171875, + "logps/chosen": -372.0, + "logps/rejected": -362.0, + "loss": 0.2299, + "rewards/accuracies": 0.8828125, + "rewards/chosen": 1.5, + "rewards/margins": 7.59375, + "rewards/rejected": -6.09375, + "step": 455 + }, + { + "epoch": 0.5595092024539877, + "grad_norm": 35.777155080443045, + "learning_rate": 1.2328512396694215e-06, + "logits/chosen": 0.0419921875, + "logits/rejected": -0.0576171875, + "logps/chosen": -318.0, + "logps/rejected": -314.0, + "loss": 0.4054, + "rewards/accuracies": 0.8125, + "rewards/chosen": 0.412109375, + "rewards/margins": 6.8125, + "rewards/rejected": -6.375, + "step": 456 + }, + { + "epoch": 0.5607361963190184, + "grad_norm": 33.41411882209408, + "learning_rate": 1.2322314049586776e-06, + "logits/chosen": 0.0126953125, + "logits/rejected": -0.1142578125, + "logps/chosen": -360.0, + "logps/rejected": -330.0, + "loss": 0.3362, + "rewards/accuracies": 0.8359375, + "rewards/chosen": 0.63671875, + "rewards/margins": 6.21875, + "rewards/rejected": -5.59375, + "step": 457 + }, + { + "epoch": 0.561963190184049, + "grad_norm": 27.510163645115238, + "learning_rate": 1.2316115702479338e-06, + "logits/chosen": -0.0264892578125, + "logits/rejected": -0.1845703125, + "logps/chosen": -376.0, + "logps/rejected": -302.0, + "loss": 0.2733, + "rewards/accuracies": 0.8828125, + "rewards/chosen": 1.109375, + "rewards/margins": 7.40625, + "rewards/rejected": -6.3125, + "step": 458 + }, + { + "epoch": 0.5631901840490797, + "grad_norm": 32.20829218573613, + "learning_rate": 1.2309917355371901e-06, + "logits/chosen": -0.0172119140625, + "logits/rejected": -0.1796875, + "logps/chosen": -396.0, + "logps/rejected": -356.0, + "loss": 0.3698, + "rewards/accuracies": 0.8515625, + "rewards/chosen": 1.171875, + "rewards/margins": 7.375, + "rewards/rejected": -6.1875, + "step": 459 + }, + { + "epoch": 0.5644171779141104, + "grad_norm": 28.463767397560563, + "learning_rate": 1.2303719008264463e-06, + "logits/chosen": -0.0159912109375, + "logits/rejected": -0.11572265625, + "logps/chosen": -354.0, + "logps/rejected": -342.0, + "loss": 0.2835, + "rewards/accuracies": 0.859375, + "rewards/chosen": 0.146484375, + "rewards/margins": 7.46875, + "rewards/rejected": -7.3125, + "step": 460 + }, + { + "epoch": 0.5656441717791411, + "grad_norm": 30.885477551125014, + "learning_rate": 1.2297520661157024e-06, + "logits/chosen": -0.2158203125, + "logits/rejected": -0.330078125, + "logps/chosen": -380.0, + "logps/rejected": -352.0, + "loss": 0.2845, + "rewards/accuracies": 0.8671875, + "rewards/chosen": 0.1884765625, + "rewards/margins": 6.5625, + "rewards/rejected": -6.375, + "step": 461 + }, + { + "epoch": 0.5668711656441717, + "grad_norm": 44.21195903099332, + "learning_rate": 1.2291322314049588e-06, + "logits/chosen": 0.047607421875, + "logits/rejected": -0.1552734375, + "logps/chosen": -370.0, + "logps/rejected": -310.0, + "loss": 0.4151, + "rewards/accuracies": 0.8203125, + "rewards/chosen": -0.1337890625, + "rewards/margins": 5.875, + "rewards/rejected": -6.03125, + "step": 462 + }, + { + "epoch": 0.5680981595092025, + "grad_norm": 33.208183287178784, + "learning_rate": 1.228512396694215e-06, + "logits/chosen": -0.0849609375, + "logits/rejected": -0.18359375, + "logps/chosen": -370.0, + "logps/rejected": -372.0, + "loss": 0.3252, + "rewards/accuracies": 0.8203125, + "rewards/chosen": 0.4140625, + "rewards/margins": 7.09375, + "rewards/rejected": -6.6875, + "step": 463 + }, + { + "epoch": 0.5693251533742332, + "grad_norm": 37.13310236043479, + "learning_rate": 1.2278925619834713e-06, + "logits/chosen": 0.05224609375, + "logits/rejected": -0.119140625, + "logps/chosen": -398.0, + "logps/rejected": -334.0, + "loss": 0.3492, + "rewards/accuracies": 0.859375, + "rewards/chosen": 0.3359375, + "rewards/margins": 6.90625, + "rewards/rejected": -6.59375, + "step": 464 + }, + { + "epoch": 0.5705521472392638, + "grad_norm": 21.80723319666503, + "learning_rate": 1.2272727272727274e-06, + "logits/chosen": -0.06396484375, + "logits/rejected": -0.302734375, + "logps/chosen": -404.0, + "logps/rejected": -362.0, + "loss": 0.1834, + "rewards/accuracies": 0.9140625, + "rewards/chosen": 0.6171875, + "rewards/margins": 8.25, + "rewards/rejected": -7.625, + "step": 465 + }, + { + "epoch": 0.5717791411042945, + "grad_norm": 32.116293843093004, + "learning_rate": 1.2266528925619833e-06, + "logits/chosen": -0.016845703125, + "logits/rejected": -0.1416015625, + "logps/chosen": -424.0, + "logps/rejected": -356.0, + "loss": 0.3118, + "rewards/accuracies": 0.8203125, + "rewards/chosen": 0.0458984375, + "rewards/margins": 7.5, + "rewards/rejected": -7.46875, + "step": 466 + }, + { + "epoch": 0.5730061349693252, + "grad_norm": 37.63110345633159, + "learning_rate": 1.2260330578512397e-06, + "logits/chosen": -0.0279541015625, + "logits/rejected": -0.2373046875, + "logps/chosen": -388.0, + "logps/rejected": -364.0, + "loss": 0.3871, + "rewards/accuracies": 0.828125, + "rewards/chosen": 0.8046875, + "rewards/margins": 7.21875, + "rewards/rejected": -6.40625, + "step": 467 + }, + { + "epoch": 0.5742331288343558, + "grad_norm": 33.08460200253717, + "learning_rate": 1.2254132231404958e-06, + "logits/chosen": -0.0419921875, + "logits/rejected": -0.1806640625, + "logps/chosen": -372.0, + "logps/rejected": -358.0, + "loss": 0.3099, + "rewards/accuracies": 0.8671875, + "rewards/chosen": 0.5234375, + "rewards/margins": 7.625, + "rewards/rejected": -7.09375, + "step": 468 + }, + { + "epoch": 0.5754601226993865, + "grad_norm": 22.648898471514507, + "learning_rate": 1.2247933884297522e-06, + "logits/chosen": 0.0230712890625, + "logits/rejected": -0.15234375, + "logps/chosen": -390.0, + "logps/rejected": -320.0, + "loss": 0.2207, + "rewards/accuracies": 0.90625, + "rewards/chosen": 0.796875, + "rewards/margins": 7.84375, + "rewards/rejected": -7.0625, + "step": 469 + }, + { + "epoch": 0.5766871165644172, + "grad_norm": 33.108021332496286, + "learning_rate": 1.2241735537190083e-06, + "logits/chosen": -0.010498046875, + "logits/rejected": -0.07373046875, + "logps/chosen": -356.0, + "logps/rejected": -338.0, + "loss": 0.367, + "rewards/accuracies": 0.828125, + "rewards/chosen": 0.69921875, + "rewards/margins": 7.5, + "rewards/rejected": -6.8125, + "step": 470 + }, + { + "epoch": 0.5779141104294478, + "grad_norm": 30.281427499200404, + "learning_rate": 1.2235537190082645e-06, + "logits/chosen": 0.1171875, + "logits/rejected": -0.0458984375, + "logps/chosen": -408.0, + "logps/rejected": -350.0, + "loss": 0.3469, + "rewards/accuracies": 0.8125, + "rewards/chosen": 0.328125, + "rewards/margins": 6.90625, + "rewards/rejected": -6.5625, + "step": 471 + }, + { + "epoch": 0.5791411042944785, + "grad_norm": 30.766178072522326, + "learning_rate": 1.2229338842975208e-06, + "logits/chosen": 0.0654296875, + "logits/rejected": -0.15234375, + "logps/chosen": -414.0, + "logps/rejected": -336.0, + "loss": 0.341, + "rewards/accuracies": 0.84375, + "rewards/chosen": 0.400390625, + "rewards/margins": 7.125, + "rewards/rejected": -6.71875, + "step": 472 + }, + { + "epoch": 0.5803680981595092, + "grad_norm": 35.82872232913366, + "learning_rate": 1.222314049586777e-06, + "logits/chosen": -0.10986328125, + "logits/rejected": -0.2138671875, + "logps/chosen": -406.0, + "logps/rejected": -394.0, + "loss": 0.3929, + "rewards/accuracies": 0.796875, + "rewards/chosen": 0.267578125, + "rewards/margins": 6.78125, + "rewards/rejected": -6.5, + "step": 473 + }, + { + "epoch": 0.5815950920245399, + "grad_norm": 38.235476934077774, + "learning_rate": 1.221694214876033e-06, + "logits/chosen": -0.06591796875, + "logits/rejected": -0.271484375, + "logps/chosen": -440.0, + "logps/rejected": -362.0, + "loss": 0.3842, + "rewards/accuracies": 0.8359375, + "rewards/chosen": 0.0673828125, + "rewards/margins": 6.875, + "rewards/rejected": -6.8125, + "step": 474 + }, + { + "epoch": 0.5828220858895705, + "grad_norm": 28.310340512331347, + "learning_rate": 1.2210743801652892e-06, + "logits/chosen": -0.06689453125, + "logits/rejected": -0.2734375, + "logps/chosen": -434.0, + "logps/rejected": -362.0, + "loss": 0.2596, + "rewards/accuracies": 0.8828125, + "rewards/chosen": 1.140625, + "rewards/margins": 8.4375, + "rewards/rejected": -7.3125, + "step": 475 + }, + { + "epoch": 0.5840490797546012, + "grad_norm": 31.21556050331037, + "learning_rate": 1.2204545454545454e-06, + "logits/chosen": 0.00927734375, + "logits/rejected": -0.1806640625, + "logps/chosen": -420.0, + "logps/rejected": -360.0, + "loss": 0.3346, + "rewards/accuracies": 0.8359375, + "rewards/chosen": -1.078125, + "rewards/margins": 6.53125, + "rewards/rejected": -7.625, + "step": 476 + }, + { + "epoch": 0.5852760736196319, + "grad_norm": 31.773096981894017, + "learning_rate": 1.2198347107438017e-06, + "logits/chosen": -0.11474609375, + "logits/rejected": -0.255859375, + "logps/chosen": -432.0, + "logps/rejected": -372.0, + "loss": 0.3232, + "rewards/accuracies": 0.890625, + "rewards/chosen": 0.1484375, + "rewards/margins": 7.8125, + "rewards/rejected": -7.65625, + "step": 477 + }, + { + "epoch": 0.5865030674846625, + "grad_norm": 30.507836762025494, + "learning_rate": 1.2192148760330579e-06, + "logits/chosen": -0.04150390625, + "logits/rejected": -0.2080078125, + "logps/chosen": -430.0, + "logps/rejected": -368.0, + "loss": 0.3843, + "rewards/accuracies": 0.8671875, + "rewards/chosen": -0.024658203125, + "rewards/margins": 7.625, + "rewards/rejected": -7.65625, + "step": 478 + }, + { + "epoch": 0.5877300613496933, + "grad_norm": 32.261887607989564, + "learning_rate": 1.218595041322314e-06, + "logits/chosen": -0.03759765625, + "logits/rejected": -0.298828125, + "logps/chosen": -370.0, + "logps/rejected": -352.0, + "loss": 0.322, + "rewards/accuracies": 0.84375, + "rewards/chosen": -1.1796875, + "rewards/margins": 6.625, + "rewards/rejected": -7.78125, + "step": 479 + }, + { + "epoch": 0.588957055214724, + "grad_norm": 28.188949742159306, + "learning_rate": 1.2179752066115704e-06, + "logits/chosen": -0.0791015625, + "logits/rejected": -0.208984375, + "logps/chosen": -372.0, + "logps/rejected": -348.0, + "loss": 0.2558, + "rewards/accuracies": 0.8828125, + "rewards/chosen": -0.8125, + "rewards/margins": 7.21875, + "rewards/rejected": -8.0, + "step": 480 + }, + { + "epoch": 0.5901840490797546, + "grad_norm": 37.17733184183342, + "learning_rate": 1.2173553719008265e-06, + "logits/chosen": -0.07568359375, + "logits/rejected": -0.1826171875, + "logps/chosen": -438.0, + "logps/rejected": -398.0, + "loss": 0.4973, + "rewards/accuracies": 0.796875, + "rewards/chosen": -0.578125, + "rewards/margins": 6.71875, + "rewards/rejected": -7.3125, + "step": 481 + }, + { + "epoch": 0.5914110429447853, + "grad_norm": 34.97683920834337, + "learning_rate": 1.2167355371900826e-06, + "logits/chosen": -0.07373046875, + "logits/rejected": -0.244140625, + "logps/chosen": -424.0, + "logps/rejected": -352.0, + "loss": 0.3731, + "rewards/accuracies": 0.8125, + "rewards/chosen": -0.255859375, + "rewards/margins": 6.875, + "rewards/rejected": -7.125, + "step": 482 + }, + { + "epoch": 0.592638036809816, + "grad_norm": 32.698615963886375, + "learning_rate": 1.216115702479339e-06, + "logits/chosen": 0.07763671875, + "logits/rejected": -0.0556640625, + "logps/chosen": -358.0, + "logps/rejected": -330.0, + "loss": 0.3304, + "rewards/accuracies": 0.8046875, + "rewards/chosen": -0.55859375, + "rewards/margins": 6.78125, + "rewards/rejected": -7.34375, + "step": 483 + }, + { + "epoch": 0.5938650306748466, + "grad_norm": 31.73235234166197, + "learning_rate": 1.215495867768595e-06, + "logits/chosen": 0.00390625, + "logits/rejected": -0.142578125, + "logps/chosen": -380.0, + "logps/rejected": -340.0, + "loss": 0.3937, + "rewards/accuracies": 0.828125, + "rewards/chosen": 0.1611328125, + "rewards/margins": 6.46875, + "rewards/rejected": -6.3125, + "step": 484 + }, + { + "epoch": 0.5950920245398773, + "grad_norm": 28.55956265398308, + "learning_rate": 1.2148760330578513e-06, + "logits/chosen": 0.023681640625, + "logits/rejected": -0.12109375, + "logps/chosen": -370.0, + "logps/rejected": -324.0, + "loss": 0.3173, + "rewards/accuracies": 0.859375, + "rewards/chosen": -0.36328125, + "rewards/margins": 6.9375, + "rewards/rejected": -7.3125, + "step": 485 + }, + { + "epoch": 0.596319018404908, + "grad_norm": 33.81986931883939, + "learning_rate": 1.2142561983471074e-06, + "logits/chosen": -0.0050048828125, + "logits/rejected": -0.14453125, + "logps/chosen": -414.0, + "logps/rejected": -352.0, + "loss": 0.3778, + "rewards/accuracies": 0.875, + "rewards/chosen": 0.455078125, + "rewards/margins": 7.0, + "rewards/rejected": -6.53125, + "step": 486 + }, + { + "epoch": 0.5975460122699386, + "grad_norm": 30.028596310368773, + "learning_rate": 1.2136363636363636e-06, + "logits/chosen": -0.08251953125, + "logits/rejected": -0.27734375, + "logps/chosen": -410.0, + "logps/rejected": -366.0, + "loss": 0.2781, + "rewards/accuracies": 0.875, + "rewards/chosen": 0.33984375, + "rewards/margins": 7.375, + "rewards/rejected": -7.03125, + "step": 487 + }, + { + "epoch": 0.5987730061349693, + "grad_norm": 35.71061050092003, + "learning_rate": 1.21301652892562e-06, + "logits/chosen": -0.0693359375, + "logits/rejected": -0.138671875, + "logps/chosen": -424.0, + "logps/rejected": -370.0, + "loss": 0.3544, + "rewards/accuracies": 0.84375, + "rewards/chosen": 0.5078125, + "rewards/margins": 6.96875, + "rewards/rejected": -6.46875, + "step": 488 + }, + { + "epoch": 0.6, + "grad_norm": 26.138379124419743, + "learning_rate": 1.212396694214876e-06, + "logits/chosen": 0.01324462890625, + "logits/rejected": -0.224609375, + "logps/chosen": -384.0, + "logps/rejected": -340.0, + "loss": 0.3008, + "rewards/accuracies": 0.859375, + "rewards/chosen": 0.92578125, + "rewards/margins": 7.875, + "rewards/rejected": -6.9375, + "step": 489 + }, + { + "epoch": 0.6012269938650306, + "grad_norm": 38.68817736099095, + "learning_rate": 1.2117768595041322e-06, + "logits/chosen": 0.055419921875, + "logits/rejected": -0.1689453125, + "logps/chosen": -432.0, + "logps/rejected": -364.0, + "loss": 0.404, + "rewards/accuracies": 0.8125, + "rewards/chosen": 0.263671875, + "rewards/margins": 6.84375, + "rewards/rejected": -6.5625, + "step": 490 + }, + { + "epoch": 0.6024539877300613, + "grad_norm": 30.239394074669093, + "learning_rate": 1.2111570247933886e-06, + "logits/chosen": 0.07666015625, + "logits/rejected": -0.134765625, + "logps/chosen": -372.0, + "logps/rejected": -328.0, + "loss": 0.3068, + "rewards/accuracies": 0.8828125, + "rewards/chosen": 0.443359375, + "rewards/margins": 6.96875, + "rewards/rejected": -6.53125, + "step": 491 + }, + { + "epoch": 0.603680981595092, + "grad_norm": 33.80787533439265, + "learning_rate": 1.2105371900826447e-06, + "logits/chosen": -0.013427734375, + "logits/rejected": -0.1533203125, + "logps/chosen": -398.0, + "logps/rejected": -390.0, + "loss": 0.3711, + "rewards/accuracies": 0.8515625, + "rewards/chosen": 0.5625, + "rewards/margins": 6.6875, + "rewards/rejected": -6.125, + "step": 492 + }, + { + "epoch": 0.6049079754601226, + "grad_norm": 41.137878058973136, + "learning_rate": 1.2099173553719008e-06, + "logits/chosen": 0.14453125, + "logits/rejected": 0.043701171875, + "logps/chosen": -392.0, + "logps/rejected": -340.0, + "loss": 0.5049, + "rewards/accuracies": 0.796875, + "rewards/chosen": 0.19921875, + "rewards/margins": 5.65625, + "rewards/rejected": -5.4375, + "step": 493 + }, + { + "epoch": 0.6061349693251534, + "grad_norm": 28.550910744803765, + "learning_rate": 1.209297520661157e-06, + "logits/chosen": 0.044921875, + "logits/rejected": -0.1435546875, + "logps/chosen": -428.0, + "logps/rejected": -340.0, + "loss": 0.3041, + "rewards/accuracies": 0.875, + "rewards/chosen": 1.1796875, + "rewards/margins": 7.21875, + "rewards/rejected": -6.0625, + "step": 494 + }, + { + "epoch": 0.6073619631901841, + "grad_norm": 32.18887165372733, + "learning_rate": 1.2086776859504131e-06, + "logits/chosen": 0.031494140625, + "logits/rejected": -0.1064453125, + "logps/chosen": -402.0, + "logps/rejected": -350.0, + "loss": 0.3067, + "rewards/accuracies": 0.859375, + "rewards/chosen": 1.0859375, + "rewards/margins": 7.25, + "rewards/rejected": -6.15625, + "step": 495 + }, + { + "epoch": 0.6085889570552148, + "grad_norm": 34.7968597037722, + "learning_rate": 1.2080578512396695e-06, + "logits/chosen": -0.0025634765625, + "logits/rejected": -0.21484375, + "logps/chosen": -382.0, + "logps/rejected": -322.0, + "loss": 0.3605, + "rewards/accuracies": 0.859375, + "rewards/chosen": 0.1708984375, + "rewards/margins": 6.25, + "rewards/rejected": -6.09375, + "step": 496 + }, + { + "epoch": 0.6098159509202454, + "grad_norm": 36.914820937247434, + "learning_rate": 1.2074380165289256e-06, + "logits/chosen": 0.0361328125, + "logits/rejected": -0.125, + "logps/chosen": -404.0, + "logps/rejected": -366.0, + "loss": 0.4718, + "rewards/accuracies": 0.78125, + "rewards/chosen": 0.1337890625, + "rewards/margins": 6.5, + "rewards/rejected": -6.375, + "step": 497 + }, + { + "epoch": 0.6110429447852761, + "grad_norm": 27.898760612078714, + "learning_rate": 1.206818181818182e-06, + "logits/chosen": -0.10302734375, + "logits/rejected": -0.265625, + "logps/chosen": -400.0, + "logps/rejected": -366.0, + "loss": 0.2974, + "rewards/accuracies": 0.8359375, + "rewards/chosen": 0.9765625, + "rewards/margins": 7.46875, + "rewards/rejected": -6.5, + "step": 498 + }, + { + "epoch": 0.6122699386503068, + "grad_norm": 33.514362873417944, + "learning_rate": 1.2061983471074381e-06, + "logits/chosen": 0.041015625, + "logits/rejected": -0.177734375, + "logps/chosen": -426.0, + "logps/rejected": -350.0, + "loss": 0.3147, + "rewards/accuracies": 0.8671875, + "rewards/chosen": 0.984375, + "rewards/margins": 6.96875, + "rewards/rejected": -6.0, + "step": 499 + }, + { + "epoch": 0.6134969325153374, + "grad_norm": 30.177881260758163, + "learning_rate": 1.2055785123966943e-06, + "logits/chosen": 0.0224609375, + "logits/rejected": -0.09375, + "logps/chosen": -418.0, + "logps/rejected": -338.0, + "loss": 0.3575, + "rewards/accuracies": 0.796875, + "rewards/chosen": 1.0390625, + "rewards/margins": 7.3125, + "rewards/rejected": -6.28125, + "step": 500 + }, + { + "epoch": 0.6147239263803681, + "grad_norm": 34.255808614960706, + "learning_rate": 1.2049586776859504e-06, + "logits/chosen": -0.0390625, + "logits/rejected": -0.171875, + "logps/chosen": -410.0, + "logps/rejected": -352.0, + "loss": 0.4018, + "rewards/accuracies": 0.8359375, + "rewards/chosen": 1.234375, + "rewards/margins": 7.59375, + "rewards/rejected": -6.34375, + "step": 501 + }, + { + "epoch": 0.6159509202453988, + "grad_norm": 28.29113723996875, + "learning_rate": 1.2043388429752065e-06, + "logits/chosen": -0.01165771484375, + "logits/rejected": -0.28125, + "logps/chosen": -386.0, + "logps/rejected": -340.0, + "loss": 0.3467, + "rewards/accuracies": 0.8828125, + "rewards/chosen": 0.6796875, + "rewards/margins": 7.71875, + "rewards/rejected": -7.03125, + "step": 502 + }, + { + "epoch": 0.6171779141104294, + "grad_norm": 29.752413229458206, + "learning_rate": 1.2037190082644629e-06, + "logits/chosen": 0.1904296875, + "logits/rejected": -0.0013275146484375, + "logps/chosen": -334.0, + "logps/rejected": -324.0, + "loss": 0.2996, + "rewards/accuracies": 0.8046875, + "rewards/chosen": 0.28515625, + "rewards/margins": 6.6875, + "rewards/rejected": -6.40625, + "step": 503 + }, + { + "epoch": 0.6184049079754601, + "grad_norm": 24.506669229979327, + "learning_rate": 1.203099173553719e-06, + "logits/chosen": -0.06201171875, + "logits/rejected": -0.234375, + "logps/chosen": -388.0, + "logps/rejected": -340.0, + "loss": 0.2718, + "rewards/accuracies": 0.8984375, + "rewards/chosen": -0.1025390625, + "rewards/margins": 6.21875, + "rewards/rejected": -6.3125, + "step": 504 + }, + { + "epoch": 0.6196319018404908, + "grad_norm": 38.20086418972538, + "learning_rate": 1.2024793388429752e-06, + "logits/chosen": -0.07666015625, + "logits/rejected": -0.298828125, + "logps/chosen": -434.0, + "logps/rejected": -406.0, + "loss": 0.4423, + "rewards/accuracies": 0.828125, + "rewards/chosen": -0.609375, + "rewards/margins": 6.78125, + "rewards/rejected": -7.375, + "step": 505 + }, + { + "epoch": 0.6208588957055214, + "grad_norm": 67.15816433272404, + "learning_rate": 1.2018595041322315e-06, + "logits/chosen": 0.06982421875, + "logits/rejected": -0.2373046875, + "logps/chosen": -428.0, + "logps/rejected": -340.0, + "loss": 0.3369, + "rewards/accuracies": 0.859375, + "rewards/chosen": -0.0081787109375, + "rewards/margins": 7.03125, + "rewards/rejected": -7.03125, + "step": 506 + }, + { + "epoch": 0.6220858895705521, + "grad_norm": 25.2127178538785, + "learning_rate": 1.2012396694214877e-06, + "logits/chosen": -0.04296875, + "logits/rejected": -0.150390625, + "logps/chosen": -394.0, + "logps/rejected": -318.0, + "loss": 0.1989, + "rewards/accuracies": 0.8984375, + "rewards/chosen": 0.11474609375, + "rewards/margins": 7.34375, + "rewards/rejected": -7.21875, + "step": 507 + }, + { + "epoch": 0.6233128834355828, + "grad_norm": 31.70942946416519, + "learning_rate": 1.2006198347107438e-06, + "logits/chosen": 0.14453125, + "logits/rejected": -0.033935546875, + "logps/chosen": -396.0, + "logps/rejected": -320.0, + "loss": 0.3214, + "rewards/accuracies": 0.84375, + "rewards/chosen": -0.1494140625, + "rewards/margins": 6.40625, + "rewards/rejected": -6.5625, + "step": 508 + }, + { + "epoch": 0.6245398773006134, + "grad_norm": 39.83330773441419, + "learning_rate": 1.2000000000000002e-06, + "logits/chosen": -0.0225830078125, + "logits/rejected": -0.2578125, + "logps/chosen": -474.0, + "logps/rejected": -396.0, + "loss": 0.569, + "rewards/accuracies": 0.8125, + "rewards/chosen": 1.0390625, + "rewards/margins": 8.25, + "rewards/rejected": -7.21875, + "step": 509 + }, + { + "epoch": 0.6257668711656442, + "grad_norm": 29.863730506340346, + "learning_rate": 1.199380165289256e-06, + "logits/chosen": -0.11181640625, + "logits/rejected": -0.291015625, + "logps/chosen": -368.0, + "logps/rejected": -314.0, + "loss": 0.2926, + "rewards/accuracies": 0.875, + "rewards/chosen": -0.375, + "rewards/margins": 6.3125, + "rewards/rejected": -6.6875, + "step": 510 + }, + { + "epoch": 0.6269938650306749, + "grad_norm": 35.188221718693796, + "learning_rate": 1.1987603305785124e-06, + "logits/chosen": -0.1328125, + "logits/rejected": -0.259765625, + "logps/chosen": -402.0, + "logps/rejected": -366.0, + "loss": 0.4071, + "rewards/accuracies": 0.8359375, + "rewards/chosen": 0.0859375, + "rewards/margins": 7.3125, + "rewards/rejected": -7.21875, + "step": 511 + }, + { + "epoch": 0.6282208588957056, + "grad_norm": 35.23659404136466, + "learning_rate": 1.1981404958677686e-06, + "logits/chosen": 0.0791015625, + "logits/rejected": -0.0654296875, + "logps/chosen": -388.0, + "logps/rejected": -398.0, + "loss": 0.4354, + "rewards/accuracies": 0.8203125, + "rewards/chosen": -0.388671875, + "rewards/margins": 5.90625, + "rewards/rejected": -6.28125, + "step": 512 + }, + { + "epoch": 0.6294478527607362, + "grad_norm": 33.94396537230905, + "learning_rate": 1.1975206611570247e-06, + "logits/chosen": 0.07958984375, + "logits/rejected": -0.0517578125, + "logps/chosen": -384.0, + "logps/rejected": -326.0, + "loss": 0.4151, + "rewards/accuracies": 0.796875, + "rewards/chosen": 0.1328125, + "rewards/margins": 6.96875, + "rewards/rejected": -6.8125, + "step": 513 + }, + { + "epoch": 0.6306748466257669, + "grad_norm": 27.33280439239978, + "learning_rate": 1.196900826446281e-06, + "logits/chosen": 0.09765625, + "logits/rejected": -0.0537109375, + "logps/chosen": -340.0, + "logps/rejected": -302.0, + "loss": 0.2677, + "rewards/accuracies": 0.8828125, + "rewards/chosen": -1.015625, + "rewards/margins": 6.5, + "rewards/rejected": -7.53125, + "step": 514 + }, + { + "epoch": 0.6319018404907976, + "grad_norm": 32.16724360242484, + "learning_rate": 1.1962809917355372e-06, + "logits/chosen": 0.07275390625, + "logits/rejected": -0.07763671875, + "logps/chosen": -358.0, + "logps/rejected": -362.0, + "loss": 0.3029, + "rewards/accuracies": 0.8515625, + "rewards/chosen": -1.0703125, + "rewards/margins": 6.5, + "rewards/rejected": -7.5625, + "step": 515 + }, + { + "epoch": 0.6331288343558282, + "grad_norm": 26.23429258203521, + "learning_rate": 1.1956611570247934e-06, + "logits/chosen": 0.0279541015625, + "logits/rejected": -0.07861328125, + "logps/chosen": -378.0, + "logps/rejected": -346.0, + "loss": 0.2424, + "rewards/accuracies": 0.8828125, + "rewards/chosen": -0.8828125, + "rewards/margins": 6.65625, + "rewards/rejected": -7.53125, + "step": 516 + }, + { + "epoch": 0.6343558282208589, + "grad_norm": 37.00771401511282, + "learning_rate": 1.1950413223140497e-06, + "logits/chosen": 0.0703125, + "logits/rejected": -0.08642578125, + "logps/chosen": -400.0, + "logps/rejected": -352.0, + "loss": 0.4735, + "rewards/accuracies": 0.78125, + "rewards/chosen": -0.48828125, + "rewards/margins": 6.28125, + "rewards/rejected": -6.78125, + "step": 517 + }, + { + "epoch": 0.6355828220858896, + "grad_norm": 28.05157133110184, + "learning_rate": 1.1944214876033059e-06, + "logits/chosen": -0.04541015625, + "logits/rejected": -0.150390625, + "logps/chosen": -424.0, + "logps/rejected": -360.0, + "loss": 0.2475, + "rewards/accuracies": 0.8671875, + "rewards/chosen": 0.396484375, + "rewards/margins": 7.625, + "rewards/rejected": -7.21875, + "step": 518 + }, + { + "epoch": 0.6368098159509202, + "grad_norm": 38.286039724245256, + "learning_rate": 1.193801652892562e-06, + "logits/chosen": -0.026123046875, + "logits/rejected": -0.1630859375, + "logps/chosen": -396.0, + "logps/rejected": -354.0, + "loss": 0.3562, + "rewards/accuracies": 0.828125, + "rewards/chosen": -0.60546875, + "rewards/margins": 6.9375, + "rewards/rejected": -7.53125, + "step": 519 + }, + { + "epoch": 0.6380368098159509, + "grad_norm": 34.57537616564543, + "learning_rate": 1.1931818181818181e-06, + "logits/chosen": -0.0517578125, + "logits/rejected": -0.279296875, + "logps/chosen": -380.0, + "logps/rejected": -350.0, + "loss": 0.37, + "rewards/accuracies": 0.84375, + "rewards/chosen": -0.271484375, + "rewards/margins": 7.71875, + "rewards/rejected": -8.0, + "step": 520 + }, + { + "epoch": 0.6392638036809816, + "grad_norm": 29.132838683315903, + "learning_rate": 1.1925619834710743e-06, + "logits/chosen": 0.158203125, + "logits/rejected": 0.01312255859375, + "logps/chosen": -428.0, + "logps/rejected": -378.0, + "loss": 0.2975, + "rewards/accuracies": 0.859375, + "rewards/chosen": 0.01263427734375, + "rewards/margins": 7.8125, + "rewards/rejected": -7.8125, + "step": 521 + }, + { + "epoch": 0.6404907975460122, + "grad_norm": 30.38859952385869, + "learning_rate": 1.1919421487603306e-06, + "logits/chosen": -0.0546875, + "logits/rejected": -0.2060546875, + "logps/chosen": -438.0, + "logps/rejected": -374.0, + "loss": 0.3466, + "rewards/accuracies": 0.8359375, + "rewards/chosen": -0.439453125, + "rewards/margins": 8.1875, + "rewards/rejected": -8.625, + "step": 522 + }, + { + "epoch": 0.6417177914110429, + "grad_norm": 36.440630501558566, + "learning_rate": 1.1913223140495868e-06, + "logits/chosen": 0.0238037109375, + "logits/rejected": -0.12890625, + "logps/chosen": -380.0, + "logps/rejected": -346.0, + "loss": 0.4467, + "rewards/accuracies": 0.8125, + "rewards/chosen": -1.09375, + "rewards/margins": 6.3125, + "rewards/rejected": -7.40625, + "step": 523 + }, + { + "epoch": 0.6429447852760736, + "grad_norm": 39.231543678462224, + "learning_rate": 1.1907024793388431e-06, + "logits/chosen": 0.05615234375, + "logits/rejected": -0.1474609375, + "logps/chosen": -440.0, + "logps/rejected": -380.0, + "loss": 0.4136, + "rewards/accuracies": 0.859375, + "rewards/chosen": -1.1953125, + "rewards/margins": 6.40625, + "rewards/rejected": -7.59375, + "step": 524 + }, + { + "epoch": 0.6441717791411042, + "grad_norm": 47.6608831234207, + "learning_rate": 1.1900826446280993e-06, + "logits/chosen": 0.07666015625, + "logits/rejected": 0.02587890625, + "logps/chosen": -406.0, + "logps/rejected": -382.0, + "loss": 0.5676, + "rewards/accuracies": 0.7421875, + "rewards/chosen": -1.3125, + "rewards/margins": 5.53125, + "rewards/rejected": -6.84375, + "step": 525 + }, + { + "epoch": 0.645398773006135, + "grad_norm": 31.7891594700196, + "learning_rate": 1.1894628099173554e-06, + "logits/chosen": 0.0458984375, + "logits/rejected": -0.10107421875, + "logps/chosen": -402.0, + "logps/rejected": -356.0, + "loss": 0.3316, + "rewards/accuracies": 0.8203125, + "rewards/chosen": -0.6171875, + "rewards/margins": 7.21875, + "rewards/rejected": -7.84375, + "step": 526 + }, + { + "epoch": 0.6466257668711657, + "grad_norm": 32.21728261594982, + "learning_rate": 1.1888429752066118e-06, + "logits/chosen": 0.1201171875, + "logits/rejected": -0.103515625, + "logps/chosen": -438.0, + "logps/rejected": -348.0, + "loss": 0.3529, + "rewards/accuracies": 0.8359375, + "rewards/chosen": -0.7265625, + "rewards/margins": 7.03125, + "rewards/rejected": -7.75, + "step": 527 + }, + { + "epoch": 0.6478527607361964, + "grad_norm": 26.60182420517136, + "learning_rate": 1.1882231404958677e-06, + "logits/chosen": -0.0380859375, + "logits/rejected": -0.220703125, + "logps/chosen": -368.0, + "logps/rejected": -344.0, + "loss": 0.2928, + "rewards/accuracies": 0.859375, + "rewards/chosen": -0.671875, + "rewards/margins": 7.375, + "rewards/rejected": -8.0625, + "step": 528 + }, + { + "epoch": 0.649079754601227, + "grad_norm": 32.109755070015275, + "learning_rate": 1.187603305785124e-06, + "logits/chosen": 0.030029296875, + "logits/rejected": -0.14453125, + "logps/chosen": -352.0, + "logps/rejected": -352.0, + "loss": 0.418, + "rewards/accuracies": 0.8203125, + "rewards/chosen": -0.3984375, + "rewards/margins": 6.9375, + "rewards/rejected": -7.34375, + "step": 529 + }, + { + "epoch": 0.6503067484662577, + "grad_norm": 30.441331665820965, + "learning_rate": 1.1869834710743802e-06, + "logits/chosen": 0.07373046875, + "logits/rejected": -0.0517578125, + "logps/chosen": -370.0, + "logps/rejected": -366.0, + "loss": 0.2962, + "rewards/accuracies": 0.8515625, + "rewards/chosen": -0.1103515625, + "rewards/margins": 7.5, + "rewards/rejected": -7.59375, + "step": 530 + }, + { + "epoch": 0.6515337423312884, + "grad_norm": 25.97427159310545, + "learning_rate": 1.1863636363636363e-06, + "logits/chosen": -0.0005950927734375, + "logits/rejected": -0.1767578125, + "logps/chosen": -394.0, + "logps/rejected": -338.0, + "loss": 0.3013, + "rewards/accuracies": 0.8671875, + "rewards/chosen": -0.2001953125, + "rewards/margins": 7.0625, + "rewards/rejected": -7.25, + "step": 531 + }, + { + "epoch": 0.652760736196319, + "grad_norm": 29.45586726470446, + "learning_rate": 1.1857438016528927e-06, + "logits/chosen": 0.12060546875, + "logits/rejected": -0.01611328125, + "logps/chosen": -394.0, + "logps/rejected": -342.0, + "loss": 0.3256, + "rewards/accuracies": 0.8671875, + "rewards/chosen": -0.1650390625, + "rewards/margins": 6.5625, + "rewards/rejected": -6.71875, + "step": 532 + }, + { + "epoch": 0.6539877300613497, + "grad_norm": 31.510396061786174, + "learning_rate": 1.1851239669421488e-06, + "logits/chosen": -0.0634765625, + "logits/rejected": -0.265625, + "logps/chosen": -414.0, + "logps/rejected": -360.0, + "loss": 0.322, + "rewards/accuracies": 0.8359375, + "rewards/chosen": 0.71484375, + "rewards/margins": 7.78125, + "rewards/rejected": -7.09375, + "step": 533 + }, + { + "epoch": 0.6552147239263804, + "grad_norm": 29.028681159770606, + "learning_rate": 1.184504132231405e-06, + "logits/chosen": 0.0908203125, + "logits/rejected": -0.02685546875, + "logps/chosen": -374.0, + "logps/rejected": -338.0, + "loss": 0.3243, + "rewards/accuracies": 0.8203125, + "rewards/chosen": 0.0341796875, + "rewards/margins": 6.40625, + "rewards/rejected": -6.375, + "step": 534 + }, + { + "epoch": 0.656441717791411, + "grad_norm": 33.385986566673, + "learning_rate": 1.1838842975206613e-06, + "logits/chosen": -0.022705078125, + "logits/rejected": -0.125, + "logps/chosen": -378.0, + "logps/rejected": -348.0, + "loss": 0.3502, + "rewards/accuracies": 0.84375, + "rewards/chosen": -0.7421875, + "rewards/margins": 6.40625, + "rewards/rejected": -7.125, + "step": 535 + }, + { + "epoch": 0.6576687116564417, + "grad_norm": 27.982276879832543, + "learning_rate": 1.1832644628099175e-06, + "logits/chosen": 0.0201416015625, + "logits/rejected": -0.07568359375, + "logps/chosen": -394.0, + "logps/rejected": -330.0, + "loss": 0.2519, + "rewards/accuracies": 0.890625, + "rewards/chosen": 0.6015625, + "rewards/margins": 7.40625, + "rewards/rejected": -6.78125, + "step": 536 + }, + { + "epoch": 0.6588957055214724, + "grad_norm": 40.68442319965527, + "learning_rate": 1.1826446280991736e-06, + "logits/chosen": 0.06201171875, + "logits/rejected": -0.0927734375, + "logps/chosen": -386.0, + "logps/rejected": -318.0, + "loss": 0.5685, + "rewards/accuracies": 0.7734375, + "rewards/chosen": -0.50390625, + "rewards/margins": 5.28125, + "rewards/rejected": -5.78125, + "step": 537 + }, + { + "epoch": 0.660122699386503, + "grad_norm": 35.32326255751808, + "learning_rate": 1.1820247933884297e-06, + "logits/chosen": 0.068359375, + "logits/rejected": -0.09912109375, + "logps/chosen": -402.0, + "logps/rejected": -360.0, + "loss": 0.4164, + "rewards/accuracies": 0.84375, + "rewards/chosen": 0.78515625, + "rewards/margins": 6.75, + "rewards/rejected": -5.96875, + "step": 538 + }, + { + "epoch": 0.6613496932515337, + "grad_norm": 32.525023600229474, + "learning_rate": 1.1814049586776859e-06, + "logits/chosen": 0.07666015625, + "logits/rejected": -0.04541015625, + "logps/chosen": -400.0, + "logps/rejected": -346.0, + "loss": 0.4519, + "rewards/accuracies": 0.8125, + "rewards/chosen": 0.80859375, + "rewards/margins": 5.84375, + "rewards/rejected": -5.03125, + "step": 539 + }, + { + "epoch": 0.6625766871165644, + "grad_norm": 35.156155067546834, + "learning_rate": 1.1807851239669422e-06, + "logits/chosen": 0.0986328125, + "logits/rejected": -0.06787109375, + "logps/chosen": -396.0, + "logps/rejected": -332.0, + "loss": 0.432, + "rewards/accuracies": 0.8515625, + "rewards/chosen": 0.80859375, + "rewards/margins": 6.71875, + "rewards/rejected": -5.9375, + "step": 540 + }, + { + "epoch": 0.6638036809815951, + "grad_norm": 45.099896848619814, + "learning_rate": 1.1801652892561984e-06, + "logits/chosen": 0.08447265625, + "logits/rejected": -0.0064697265625, + "logps/chosen": -390.0, + "logps/rejected": -354.0, + "loss": 0.3808, + "rewards/accuracies": 0.828125, + "rewards/chosen": 0.279296875, + "rewards/margins": 5.84375, + "rewards/rejected": -5.5625, + "step": 541 + }, + { + "epoch": 0.6650306748466258, + "grad_norm": 30.118145811004524, + "learning_rate": 1.1795454545454545e-06, + "logits/chosen": 0.031494140625, + "logits/rejected": -0.053466796875, + "logps/chosen": -356.0, + "logps/rejected": -320.0, + "loss": 0.3659, + "rewards/accuracies": 0.859375, + "rewards/chosen": 0.87109375, + "rewards/margins": 6.03125, + "rewards/rejected": -5.15625, + "step": 542 + }, + { + "epoch": 0.6662576687116565, + "grad_norm": 32.83770736751292, + "learning_rate": 1.1789256198347109e-06, + "logits/chosen": -0.02294921875, + "logits/rejected": -0.201171875, + "logps/chosen": -450.0, + "logps/rejected": -354.0, + "loss": 0.3396, + "rewards/accuracies": 0.8359375, + "rewards/chosen": 1.5078125, + "rewards/margins": 6.875, + "rewards/rejected": -5.375, + "step": 543 + }, + { + "epoch": 0.6674846625766871, + "grad_norm": 35.490916209773005, + "learning_rate": 1.178305785123967e-06, + "logits/chosen": 0.007598876953125, + "logits/rejected": -0.1318359375, + "logps/chosen": -378.0, + "logps/rejected": -318.0, + "loss": 0.434, + "rewards/accuracies": 0.8359375, + "rewards/chosen": 0.376953125, + "rewards/margins": 6.15625, + "rewards/rejected": -5.78125, + "step": 544 + }, + { + "epoch": 0.6687116564417178, + "grad_norm": 33.60719342971152, + "learning_rate": 1.1776859504132231e-06, + "logits/chosen": -0.054443359375, + "logits/rejected": -0.21875, + "logps/chosen": -458.0, + "logps/rejected": -368.0, + "loss": 0.3855, + "rewards/accuracies": 0.8125, + "rewards/chosen": 0.94140625, + "rewards/margins": 6.1875, + "rewards/rejected": -5.25, + "step": 545 + }, + { + "epoch": 0.6699386503067485, + "grad_norm": 33.972456048977826, + "learning_rate": 1.1770661157024793e-06, + "logits/chosen": 0.0693359375, + "logits/rejected": -0.028564453125, + "logps/chosen": -374.0, + "logps/rejected": -338.0, + "loss": 0.3815, + "rewards/accuracies": 0.828125, + "rewards/chosen": 0.6953125, + "rewards/margins": 6.40625, + "rewards/rejected": -5.71875, + "step": 546 + }, + { + "epoch": 0.6711656441717792, + "grad_norm": 28.303815223441166, + "learning_rate": 1.1764462809917354e-06, + "logits/chosen": 0.0133056640625, + "logits/rejected": -0.166015625, + "logps/chosen": -386.0, + "logps/rejected": -312.0, + "loss": 0.2992, + "rewards/accuracies": 0.84375, + "rewards/chosen": 0.88671875, + "rewards/margins": 6.4375, + "rewards/rejected": -5.53125, + "step": 547 + }, + { + "epoch": 0.6723926380368098, + "grad_norm": 38.31133772907115, + "learning_rate": 1.1758264462809918e-06, + "logits/chosen": -0.06494140625, + "logits/rejected": -0.265625, + "logps/chosen": -402.0, + "logps/rejected": -358.0, + "loss": 0.3532, + "rewards/accuracies": 0.8515625, + "rewards/chosen": 1.015625, + "rewards/margins": 6.125, + "rewards/rejected": -5.125, + "step": 548 + }, + { + "epoch": 0.6736196319018405, + "grad_norm": 33.292222723297044, + "learning_rate": 1.175206611570248e-06, + "logits/chosen": 0.047607421875, + "logits/rejected": -0.1572265625, + "logps/chosen": -414.0, + "logps/rejected": -344.0, + "loss": 0.3292, + "rewards/accuracies": 0.8359375, + "rewards/chosen": 0.5625, + "rewards/margins": 5.96875, + "rewards/rejected": -5.40625, + "step": 549 + }, + { + "epoch": 0.6748466257668712, + "grad_norm": 30.807870745155437, + "learning_rate": 1.174586776859504e-06, + "logits/chosen": -0.024658203125, + "logits/rejected": -0.08642578125, + "logps/chosen": -402.0, + "logps/rejected": -330.0, + "loss": 0.3926, + "rewards/accuracies": 0.8125, + "rewards/chosen": 0.12890625, + "rewards/margins": 5.71875, + "rewards/rejected": -5.5625, + "step": 550 + }, + { + "epoch": 0.6760736196319018, + "grad_norm": 28.384977078060157, + "learning_rate": 1.1739669421487604e-06, + "logits/chosen": 0.06787109375, + "logits/rejected": -0.04541015625, + "logps/chosen": -410.0, + "logps/rejected": -346.0, + "loss": 0.3704, + "rewards/accuracies": 0.859375, + "rewards/chosen": 1.2109375, + "rewards/margins": 6.75, + "rewards/rejected": -5.53125, + "step": 551 + }, + { + "epoch": 0.6773006134969325, + "grad_norm": 29.63861597355486, + "learning_rate": 1.1733471074380166e-06, + "logits/chosen": -0.022705078125, + "logits/rejected": -0.158203125, + "logps/chosen": -416.0, + "logps/rejected": -360.0, + "loss": 0.3331, + "rewards/accuracies": 0.828125, + "rewards/chosen": 0.71875, + "rewards/margins": 6.25, + "rewards/rejected": -5.53125, + "step": 552 + }, + { + "epoch": 0.6785276073619632, + "grad_norm": 32.698792932077886, + "learning_rate": 1.172727272727273e-06, + "logits/chosen": 0.01806640625, + "logits/rejected": -0.1259765625, + "logps/chosen": -366.0, + "logps/rejected": -342.0, + "loss": 0.4129, + "rewards/accuracies": 0.8359375, + "rewards/chosen": -0.1474609375, + "rewards/margins": 5.625, + "rewards/rejected": -5.78125, + "step": 553 + }, + { + "epoch": 0.6797546012269938, + "grad_norm": 32.72020008420759, + "learning_rate": 1.172107438016529e-06, + "logits/chosen": 0.0791015625, + "logits/rejected": -0.1171875, + "logps/chosen": -434.0, + "logps/rejected": -334.0, + "loss": 0.3464, + "rewards/accuracies": 0.84375, + "rewards/chosen": 0.83984375, + "rewards/margins": 6.4375, + "rewards/rejected": -5.59375, + "step": 554 + }, + { + "epoch": 0.6809815950920245, + "grad_norm": 30.608371906092913, + "learning_rate": 1.171487603305785e-06, + "logits/chosen": 0.2021484375, + "logits/rejected": 0.09130859375, + "logps/chosen": -364.0, + "logps/rejected": -318.0, + "loss": 0.3666, + "rewards/accuracies": 0.8046875, + "rewards/chosen": 0.3515625, + "rewards/margins": 6.34375, + "rewards/rejected": -6.0, + "step": 555 + }, + { + "epoch": 0.6822085889570552, + "grad_norm": 26.18493705591653, + "learning_rate": 1.1708677685950413e-06, + "logits/chosen": 0.053466796875, + "logits/rejected": -0.162109375, + "logps/chosen": -424.0, + "logps/rejected": -360.0, + "loss": 0.2487, + "rewards/accuracies": 0.890625, + "rewards/chosen": 0.96875, + "rewards/margins": 7.5625, + "rewards/rejected": -6.59375, + "step": 556 + }, + { + "epoch": 0.6834355828220859, + "grad_norm": 27.830299010779843, + "learning_rate": 1.1702479338842975e-06, + "logits/chosen": 0.05029296875, + "logits/rejected": -0.16796875, + "logps/chosen": -398.0, + "logps/rejected": -346.0, + "loss": 0.3343, + "rewards/accuracies": 0.828125, + "rewards/chosen": 0.53515625, + "rewards/margins": 6.6875, + "rewards/rejected": -6.15625, + "step": 557 + }, + { + "epoch": 0.6846625766871166, + "grad_norm": 25.854408165941038, + "learning_rate": 1.1696280991735538e-06, + "logits/chosen": 0.087890625, + "logits/rejected": -0.078125, + "logps/chosen": -378.0, + "logps/rejected": -340.0, + "loss": 0.2782, + "rewards/accuracies": 0.8671875, + "rewards/chosen": -0.3125, + "rewards/margins": 6.9375, + "rewards/rejected": -7.25, + "step": 558 + }, + { + "epoch": 0.6858895705521473, + "grad_norm": 33.3339739767785, + "learning_rate": 1.16900826446281e-06, + "logits/chosen": -0.072265625, + "logits/rejected": -0.1591796875, + "logps/chosen": -388.0, + "logps/rejected": -354.0, + "loss": 0.3848, + "rewards/accuracies": 0.828125, + "rewards/chosen": -0.255859375, + "rewards/margins": 6.5, + "rewards/rejected": -6.78125, + "step": 559 + }, + { + "epoch": 0.6871165644171779, + "grad_norm": 21.74474553758103, + "learning_rate": 1.1683884297520661e-06, + "logits/chosen": -0.1669921875, + "logits/rejected": -0.302734375, + "logps/chosen": -394.0, + "logps/rejected": -344.0, + "loss": 0.2247, + "rewards/accuracies": 0.9140625, + "rewards/chosen": 0.29296875, + "rewards/margins": 7.96875, + "rewards/rejected": -7.6875, + "step": 560 + }, + { + "epoch": 0.6883435582822086, + "grad_norm": 24.604934270788146, + "learning_rate": 1.1677685950413225e-06, + "logits/chosen": 0.00164031982421875, + "logits/rejected": -0.177734375, + "logps/chosen": -360.0, + "logps/rejected": -314.0, + "loss": 0.2132, + "rewards/accuracies": 0.890625, + "rewards/chosen": 0.171875, + "rewards/margins": 7.25, + "rewards/rejected": -7.09375, + "step": 561 + }, + { + "epoch": 0.6895705521472393, + "grad_norm": 25.33619721263671, + "learning_rate": 1.1671487603305786e-06, + "logits/chosen": 0.142578125, + "logits/rejected": -0.064453125, + "logps/chosen": -382.0, + "logps/rejected": -336.0, + "loss": 0.2353, + "rewards/accuracies": 0.875, + "rewards/chosen": -0.55078125, + "rewards/margins": 6.75, + "rewards/rejected": -7.3125, + "step": 562 + }, + { + "epoch": 0.69079754601227, + "grad_norm": 28.2123146430669, + "learning_rate": 1.1665289256198348e-06, + "logits/chosen": -0.055908203125, + "logits/rejected": -0.2041015625, + "logps/chosen": -398.0, + "logps/rejected": -386.0, + "loss": 0.3157, + "rewards/accuracies": 0.8828125, + "rewards/chosen": 0.71484375, + "rewards/margins": 8.0, + "rewards/rejected": -7.3125, + "step": 563 + }, + { + "epoch": 0.6920245398773006, + "grad_norm": 24.610651917154996, + "learning_rate": 1.1659090909090909e-06, + "logits/chosen": -0.040283203125, + "logits/rejected": -0.1552734375, + "logps/chosen": -396.0, + "logps/rejected": -350.0, + "loss": 0.3066, + "rewards/accuracies": 0.859375, + "rewards/chosen": -0.21875, + "rewards/margins": 6.625, + "rewards/rejected": -6.84375, + "step": 564 + }, + { + "epoch": 0.6932515337423313, + "grad_norm": 30.241252760906917, + "learning_rate": 1.165289256198347e-06, + "logits/chosen": 0.0081787109375, + "logits/rejected": -0.1796875, + "logps/chosen": -404.0, + "logps/rejected": -354.0, + "loss": 0.4933, + "rewards/accuracies": 0.828125, + "rewards/chosen": -0.396484375, + "rewards/margins": 6.03125, + "rewards/rejected": -6.40625, + "step": 565 + }, + { + "epoch": 0.694478527607362, + "grad_norm": 30.063355062368128, + "learning_rate": 1.1646694214876034e-06, + "logits/chosen": 0.004852294921875, + "logits/rejected": -0.08544921875, + "logps/chosen": -420.0, + "logps/rejected": -348.0, + "loss": 0.2711, + "rewards/accuracies": 0.84375, + "rewards/chosen": 0.765625, + "rewards/margins": 7.3125, + "rewards/rejected": -6.53125, + "step": 566 + }, + { + "epoch": 0.6957055214723926, + "grad_norm": 23.27581640569879, + "learning_rate": 1.1640495867768595e-06, + "logits/chosen": 0.11865234375, + "logits/rejected": -0.0096435546875, + "logps/chosen": -322.0, + "logps/rejected": -288.0, + "loss": 0.2144, + "rewards/accuracies": 0.890625, + "rewards/chosen": -0.30859375, + "rewards/margins": 6.03125, + "rewards/rejected": -6.34375, + "step": 567 + }, + { + "epoch": 0.6969325153374233, + "grad_norm": 30.55160170642105, + "learning_rate": 1.1634297520661157e-06, + "logits/chosen": -0.0390625, + "logits/rejected": -0.0869140625, + "logps/chosen": -388.0, + "logps/rejected": -334.0, + "loss": 0.3414, + "rewards/accuracies": 0.8203125, + "rewards/chosen": 0.357421875, + "rewards/margins": 6.71875, + "rewards/rejected": -6.375, + "step": 568 + }, + { + "epoch": 0.698159509202454, + "grad_norm": 29.293298679377934, + "learning_rate": 1.162809917355372e-06, + "logits/chosen": 0.04736328125, + "logits/rejected": -0.0986328125, + "logps/chosen": -388.0, + "logps/rejected": -316.0, + "loss": 0.2723, + "rewards/accuracies": 0.84375, + "rewards/chosen": -0.0361328125, + "rewards/margins": 6.28125, + "rewards/rejected": -6.3125, + "step": 569 + }, + { + "epoch": 0.6993865030674846, + "grad_norm": 29.217886094805987, + "learning_rate": 1.1621900826446282e-06, + "logits/chosen": -0.1298828125, + "logits/rejected": -0.2490234375, + "logps/chosen": -406.0, + "logps/rejected": -348.0, + "loss": 0.2078, + "rewards/accuracies": 0.8671875, + "rewards/chosen": 0.9375, + "rewards/margins": 7.40625, + "rewards/rejected": -6.46875, + "step": 570 + }, + { + "epoch": 0.7006134969325153, + "grad_norm": 29.496729411168463, + "learning_rate": 1.1615702479338843e-06, + "logits/chosen": 0.0179443359375, + "logits/rejected": -0.10205078125, + "logps/chosen": -332.0, + "logps/rejected": -314.0, + "loss": 0.3188, + "rewards/accuracies": 0.8359375, + "rewards/chosen": 0.0294189453125, + "rewards/margins": 5.875, + "rewards/rejected": -5.84375, + "step": 571 + }, + { + "epoch": 0.701840490797546, + "grad_norm": 28.740254642158334, + "learning_rate": 1.1609504132231407e-06, + "logits/chosen": 0.03173828125, + "logits/rejected": -0.11669921875, + "logps/chosen": -388.0, + "logps/rejected": -378.0, + "loss": 0.2732, + "rewards/accuracies": 0.890625, + "rewards/chosen": 1.1484375, + "rewards/margins": 7.40625, + "rewards/rejected": -6.25, + "step": 572 + }, + { + "epoch": 0.7030674846625767, + "grad_norm": 35.19506105279461, + "learning_rate": 1.1603305785123966e-06, + "logits/chosen": 0.078125, + "logits/rejected": -0.050048828125, + "logps/chosen": -394.0, + "logps/rejected": -346.0, + "loss": 0.5289, + "rewards/accuracies": 0.8125, + "rewards/chosen": -0.150390625, + "rewards/margins": 6.46875, + "rewards/rejected": -6.59375, + "step": 573 + }, + { + "epoch": 0.7042944785276074, + "grad_norm": 30.319920980400575, + "learning_rate": 1.159710743801653e-06, + "logits/chosen": -0.06201171875, + "logits/rejected": -0.1865234375, + "logps/chosen": -328.0, + "logps/rejected": -320.0, + "loss": 0.34, + "rewards/accuracies": 0.8515625, + "rewards/chosen": 0.28125, + "rewards/margins": 6.78125, + "rewards/rejected": -6.5, + "step": 574 + }, + { + "epoch": 0.7055214723926381, + "grad_norm": 29.615184087034994, + "learning_rate": 1.159090909090909e-06, + "logits/chosen": -0.119140625, + "logits/rejected": -0.2734375, + "logps/chosen": -412.0, + "logps/rejected": -388.0, + "loss": 0.2626, + "rewards/accuracies": 0.8828125, + "rewards/chosen": 1.03125, + "rewards/margins": 8.375, + "rewards/rejected": -7.34375, + "step": 575 + }, + { + "epoch": 0.7067484662576687, + "grad_norm": 22.847983936498952, + "learning_rate": 1.1584710743801652e-06, + "logits/chosen": 0.03271484375, + "logits/rejected": -0.1787109375, + "logps/chosen": -386.0, + "logps/rejected": -344.0, + "loss": 0.1933, + "rewards/accuracies": 0.90625, + "rewards/chosen": 1.0625, + "rewards/margins": 8.5, + "rewards/rejected": -7.46875, + "step": 576 + }, + { + "epoch": 0.7079754601226994, + "grad_norm": 33.54348915562207, + "learning_rate": 1.1578512396694216e-06, + "logits/chosen": 0.08935546875, + "logits/rejected": -0.0296630859375, + "logps/chosen": -368.0, + "logps/rejected": -334.0, + "loss": 0.4071, + "rewards/accuracies": 0.875, + "rewards/chosen": 0.828125, + "rewards/margins": 8.1875, + "rewards/rejected": -7.34375, + "step": 577 + }, + { + "epoch": 0.7092024539877301, + "grad_norm": 31.17463611360663, + "learning_rate": 1.1572314049586777e-06, + "logits/chosen": -0.06591796875, + "logits/rejected": -0.208984375, + "logps/chosen": -420.0, + "logps/rejected": -354.0, + "loss": 0.3672, + "rewards/accuracies": 0.8515625, + "rewards/chosen": 0.921875, + "rewards/margins": 8.6875, + "rewards/rejected": -7.75, + "step": 578 + }, + { + "epoch": 0.7104294478527607, + "grad_norm": 37.58239380773895, + "learning_rate": 1.1566115702479339e-06, + "logits/chosen": -0.025390625, + "logits/rejected": -0.1181640625, + "logps/chosen": -440.0, + "logps/rejected": -352.0, + "loss": 0.4655, + "rewards/accuracies": 0.765625, + "rewards/chosen": -0.6953125, + "rewards/margins": 6.28125, + "rewards/rejected": -7.0, + "step": 579 + }, + { + "epoch": 0.7116564417177914, + "grad_norm": 34.56148267033386, + "learning_rate": 1.1559917355371902e-06, + "logits/chosen": 0.0673828125, + "logits/rejected": -0.10791015625, + "logps/chosen": -390.0, + "logps/rejected": -374.0, + "loss": 0.2813, + "rewards/accuracies": 0.859375, + "rewards/chosen": -0.21484375, + "rewards/margins": 8.0625, + "rewards/rejected": -8.3125, + "step": 580 + }, + { + "epoch": 0.7128834355828221, + "grad_norm": 26.418081399921064, + "learning_rate": 1.1553719008264464e-06, + "logits/chosen": -0.0498046875, + "logits/rejected": -0.244140625, + "logps/chosen": -380.0, + "logps/rejected": -338.0, + "loss": 0.2267, + "rewards/accuracies": 0.8515625, + "rewards/chosen": 0.271484375, + "rewards/margins": 8.3125, + "rewards/rejected": -8.0625, + "step": 581 + }, + { + "epoch": 0.7141104294478527, + "grad_norm": 30.496748663730113, + "learning_rate": 1.1547520661157025e-06, + "logits/chosen": 0.197265625, + "logits/rejected": -0.02587890625, + "logps/chosen": -402.0, + "logps/rejected": -350.0, + "loss": 0.3169, + "rewards/accuracies": 0.859375, + "rewards/chosen": -0.01220703125, + "rewards/margins": 6.96875, + "rewards/rejected": -7.0, + "step": 582 + }, + { + "epoch": 0.7153374233128834, + "grad_norm": 31.96559757214927, + "learning_rate": 1.1541322314049586e-06, + "logits/chosen": -0.02880859375, + "logits/rejected": -0.1748046875, + "logps/chosen": -392.0, + "logps/rejected": -364.0, + "loss": 0.3362, + "rewards/accuracies": 0.8359375, + "rewards/chosen": 0.3671875, + "rewards/margins": 7.5, + "rewards/rejected": -7.125, + "step": 583 + }, + { + "epoch": 0.7165644171779141, + "grad_norm": 36.407577582473756, + "learning_rate": 1.1535123966942148e-06, + "logits/chosen": 0.08154296875, + "logits/rejected": 0.0218505859375, + "logps/chosen": -380.0, + "logps/rejected": -392.0, + "loss": 0.431, + "rewards/accuracies": 0.8125, + "rewards/chosen": -0.58203125, + "rewards/margins": 5.8125, + "rewards/rejected": -6.375, + "step": 584 + }, + { + "epoch": 0.7177914110429447, + "grad_norm": 41.799396220975545, + "learning_rate": 1.1528925619834711e-06, + "logits/chosen": 0.1796875, + "logits/rejected": 0.0634765625, + "logps/chosen": -410.0, + "logps/rejected": -344.0, + "loss": 0.5833, + "rewards/accuracies": 0.7578125, + "rewards/chosen": 0.2255859375, + "rewards/margins": 6.4375, + "rewards/rejected": -6.21875, + "step": 585 + }, + { + "epoch": 0.7190184049079754, + "grad_norm": 32.85377702391839, + "learning_rate": 1.1522727272727273e-06, + "logits/chosen": 0.035888671875, + "logits/rejected": -0.07373046875, + "logps/chosen": -356.0, + "logps/rejected": -324.0, + "loss": 0.3732, + "rewards/accuracies": 0.8515625, + "rewards/chosen": 0.00927734375, + "rewards/margins": 6.9375, + "rewards/rejected": -6.90625, + "step": 586 + }, + { + "epoch": 0.7202453987730061, + "grad_norm": 28.585914695021586, + "learning_rate": 1.1516528925619836e-06, + "logits/chosen": 0.01116943359375, + "logits/rejected": -0.09912109375, + "logps/chosen": -342.0, + "logps/rejected": -332.0, + "loss": 0.3594, + "rewards/accuracies": 0.859375, + "rewards/chosen": 0.08935546875, + "rewards/margins": 7.0625, + "rewards/rejected": -6.96875, + "step": 587 + }, + { + "epoch": 0.7214723926380369, + "grad_norm": 24.10336561473994, + "learning_rate": 1.1510330578512398e-06, + "logits/chosen": -0.0262451171875, + "logits/rejected": -0.1826171875, + "logps/chosen": -388.0, + "logps/rejected": -342.0, + "loss": 0.2892, + "rewards/accuracies": 0.8671875, + "rewards/chosen": 1.2109375, + "rewards/margins": 8.0625, + "rewards/rejected": -6.875, + "step": 588 + }, + { + "epoch": 0.7226993865030675, + "grad_norm": 28.27312313115512, + "learning_rate": 1.150413223140496e-06, + "logits/chosen": -0.0986328125, + "logits/rejected": -0.2041015625, + "logps/chosen": -372.0, + "logps/rejected": -354.0, + "loss": 0.3598, + "rewards/accuracies": 0.8359375, + "rewards/chosen": -0.37890625, + "rewards/margins": 6.28125, + "rewards/rejected": -6.65625, + "step": 589 + }, + { + "epoch": 0.7239263803680982, + "grad_norm": 30.343626428171632, + "learning_rate": 1.149793388429752e-06, + "logits/chosen": 0.00927734375, + "logits/rejected": -0.103515625, + "logps/chosen": -374.0, + "logps/rejected": -330.0, + "loss": 0.3324, + "rewards/accuracies": 0.8359375, + "rewards/chosen": 0.0986328125, + "rewards/margins": 6.875, + "rewards/rejected": -6.78125, + "step": 590 + }, + { + "epoch": 0.7251533742331289, + "grad_norm": 36.86660943740023, + "learning_rate": 1.1491735537190082e-06, + "logits/chosen": -0.169921875, + "logits/rejected": -0.373046875, + "logps/chosen": -414.0, + "logps/rejected": -354.0, + "loss": 0.28, + "rewards/accuracies": 0.8828125, + "rewards/chosen": 0.71484375, + "rewards/margins": 8.0, + "rewards/rejected": -7.3125, + "step": 591 + }, + { + "epoch": 0.7263803680981595, + "grad_norm": 27.76314433981488, + "learning_rate": 1.1485537190082645e-06, + "logits/chosen": -0.02490234375, + "logits/rejected": -0.2138671875, + "logps/chosen": -378.0, + "logps/rejected": -328.0, + "loss": 0.2629, + "rewards/accuracies": 0.8828125, + "rewards/chosen": 0.64453125, + "rewards/margins": 8.4375, + "rewards/rejected": -7.78125, + "step": 592 + }, + { + "epoch": 0.7276073619631902, + "grad_norm": 36.82046387528082, + "learning_rate": 1.1479338842975207e-06, + "logits/chosen": 0.022705078125, + "logits/rejected": -0.2333984375, + "logps/chosen": -454.0, + "logps/rejected": -392.0, + "loss": 0.4117, + "rewards/accuracies": 0.859375, + "rewards/chosen": 0.76953125, + "rewards/margins": 7.84375, + "rewards/rejected": -7.09375, + "step": 593 + }, + { + "epoch": 0.7288343558282209, + "grad_norm": 30.96499084941041, + "learning_rate": 1.1473140495867768e-06, + "logits/chosen": -0.036865234375, + "logits/rejected": -0.197265625, + "logps/chosen": -366.0, + "logps/rejected": -334.0, + "loss": 0.3336, + "rewards/accuracies": 0.84375, + "rewards/chosen": -0.294921875, + "rewards/margins": 7.375, + "rewards/rejected": -7.65625, + "step": 594 + }, + { + "epoch": 0.7300613496932515, + "grad_norm": 33.881983550723454, + "learning_rate": 1.1466942148760332e-06, + "logits/chosen": -0.059814453125, + "logits/rejected": -0.26953125, + "logps/chosen": -454.0, + "logps/rejected": -372.0, + "loss": 0.2673, + "rewards/accuracies": 0.890625, + "rewards/chosen": 0.5234375, + "rewards/margins": 8.25, + "rewards/rejected": -7.71875, + "step": 595 + }, + { + "epoch": 0.7312883435582822, + "grad_norm": 34.4461650357885, + "learning_rate": 1.1460743801652893e-06, + "logits/chosen": -0.0286865234375, + "logits/rejected": -0.18359375, + "logps/chosen": -422.0, + "logps/rejected": -368.0, + "loss": 0.3909, + "rewards/accuracies": 0.8125, + "rewards/chosen": -0.004608154296875, + "rewards/margins": 7.25, + "rewards/rejected": -7.25, + "step": 596 + }, + { + "epoch": 0.7325153374233129, + "grad_norm": 32.96451484117692, + "learning_rate": 1.1454545454545455e-06, + "logits/chosen": 0.0830078125, + "logits/rejected": -0.09228515625, + "logps/chosen": -388.0, + "logps/rejected": -306.0, + "loss": 0.4178, + "rewards/accuracies": 0.8203125, + "rewards/chosen": 0.271484375, + "rewards/margins": 6.8125, + "rewards/rejected": -6.53125, + "step": 597 + }, + { + "epoch": 0.7337423312883435, + "grad_norm": 25.225306353625232, + "learning_rate": 1.1448347107438018e-06, + "logits/chosen": -0.0211181640625, + "logits/rejected": -0.216796875, + "logps/chosen": -446.0, + "logps/rejected": -338.0, + "loss": 0.2285, + "rewards/accuracies": 0.875, + "rewards/chosen": 1.296875, + "rewards/margins": 8.375, + "rewards/rejected": -7.09375, + "step": 598 + }, + { + "epoch": 0.7349693251533742, + "grad_norm": 30.51091893610521, + "learning_rate": 1.1442148760330577e-06, + "logits/chosen": -0.169921875, + "logits/rejected": -0.3125, + "logps/chosen": -456.0, + "logps/rejected": -386.0, + "loss": 0.3423, + "rewards/accuracies": 0.8359375, + "rewards/chosen": 0.82421875, + "rewards/margins": 8.375, + "rewards/rejected": -7.53125, + "step": 599 + }, + { + "epoch": 0.7361963190184049, + "grad_norm": 31.65374536927, + "learning_rate": 1.143595041322314e-06, + "logits/chosen": -0.1796875, + "logits/rejected": -0.25390625, + "logps/chosen": -364.0, + "logps/rejected": -334.0, + "loss": 0.437, + "rewards/accuracies": 0.84375, + "rewards/chosen": 0.3515625, + "rewards/margins": 7.5625, + "rewards/rejected": -7.1875, + "step": 600 + }, + { + "epoch": 0.7374233128834355, + "grad_norm": 29.257508369320092, + "learning_rate": 1.1429752066115702e-06, + "logits/chosen": -0.03759765625, + "logits/rejected": -0.224609375, + "logps/chosen": -348.0, + "logps/rejected": -324.0, + "loss": 0.3337, + "rewards/accuracies": 0.84375, + "rewards/chosen": -0.51953125, + "rewards/margins": 7.21875, + "rewards/rejected": -7.71875, + "step": 601 + }, + { + "epoch": 0.7386503067484662, + "grad_norm": 23.637703488224467, + "learning_rate": 1.1423553719008264e-06, + "logits/chosen": 0.1103515625, + "logits/rejected": -0.0517578125, + "logps/chosen": -406.0, + "logps/rejected": -350.0, + "loss": 0.2192, + "rewards/accuracies": 0.9296875, + "rewards/chosen": 0.64453125, + "rewards/margins": 8.4375, + "rewards/rejected": -7.78125, + "step": 602 + }, + { + "epoch": 0.7398773006134969, + "grad_norm": 24.798455930230393, + "learning_rate": 1.1417355371900827e-06, + "logits/chosen": -0.0615234375, + "logits/rejected": -0.1943359375, + "logps/chosen": -382.0, + "logps/rejected": -336.0, + "loss": 0.3279, + "rewards/accuracies": 0.8671875, + "rewards/chosen": 0.82421875, + "rewards/margins": 8.3125, + "rewards/rejected": -7.5, + "step": 603 + }, + { + "epoch": 0.7411042944785277, + "grad_norm": 28.661443673376716, + "learning_rate": 1.1411157024793389e-06, + "logits/chosen": -0.11376953125, + "logits/rejected": -0.275390625, + "logps/chosen": -420.0, + "logps/rejected": -374.0, + "loss": 0.2818, + "rewards/accuracies": 0.8828125, + "rewards/chosen": 0.267578125, + "rewards/margins": 8.125, + "rewards/rejected": -7.8125, + "step": 604 + }, + { + "epoch": 0.7423312883435583, + "grad_norm": 28.682789953669307, + "learning_rate": 1.140495867768595e-06, + "logits/chosen": -0.057373046875, + "logits/rejected": -0.24609375, + "logps/chosen": -380.0, + "logps/rejected": -348.0, + "loss": 0.2877, + "rewards/accuracies": 0.8828125, + "rewards/chosen": 0.349609375, + "rewards/margins": 8.5, + "rewards/rejected": -8.125, + "step": 605 + }, + { + "epoch": 0.743558282208589, + "grad_norm": 31.866662459492442, + "learning_rate": 1.1398760330578514e-06, + "logits/chosen": -0.138671875, + "logits/rejected": -0.263671875, + "logps/chosen": -390.0, + "logps/rejected": -348.0, + "loss": 0.3751, + "rewards/accuracies": 0.828125, + "rewards/chosen": 0.443359375, + "rewards/margins": 7.65625, + "rewards/rejected": -7.21875, + "step": 606 + }, + { + "epoch": 0.7447852760736197, + "grad_norm": 29.660350580126263, + "learning_rate": 1.1392561983471075e-06, + "logits/chosen": -0.0289306640625, + "logits/rejected": -0.25, + "logps/chosen": -404.0, + "logps/rejected": -386.0, + "loss": 0.3196, + "rewards/accuracies": 0.890625, + "rewards/chosen": -0.265625, + "rewards/margins": 8.0625, + "rewards/rejected": -8.3125, + "step": 607 + }, + { + "epoch": 0.7460122699386503, + "grad_norm": 31.21931912524627, + "learning_rate": 1.1386363636363636e-06, + "logits/chosen": 0.10498046875, + "logits/rejected": -0.1416015625, + "logps/chosen": -404.0, + "logps/rejected": -360.0, + "loss": 0.3793, + "rewards/accuracies": 0.8671875, + "rewards/chosen": -0.6953125, + "rewards/margins": 6.8125, + "rewards/rejected": -7.5, + "step": 608 + }, + { + "epoch": 0.747239263803681, + "grad_norm": 35.278030303126314, + "learning_rate": 1.1380165289256198e-06, + "logits/chosen": -0.02978515625, + "logits/rejected": -0.1953125, + "logps/chosen": -412.0, + "logps/rejected": -342.0, + "loss": 0.4025, + "rewards/accuracies": 0.8515625, + "rewards/chosen": -0.255859375, + "rewards/margins": 7.9375, + "rewards/rejected": -8.1875, + "step": 609 + }, + { + "epoch": 0.7484662576687117, + "grad_norm": 29.748859331814412, + "learning_rate": 1.137396694214876e-06, + "logits/chosen": -0.10302734375, + "logits/rejected": -0.3203125, + "logps/chosen": -452.0, + "logps/rejected": -396.0, + "loss": 0.2827, + "rewards/accuracies": 0.859375, + "rewards/chosen": 0.2216796875, + "rewards/margins": 8.625, + "rewards/rejected": -8.375, + "step": 610 + }, + { + "epoch": 0.7496932515337423, + "grad_norm": 42.36265371910942, + "learning_rate": 1.1367768595041323e-06, + "logits/chosen": -0.050537109375, + "logits/rejected": -0.2216796875, + "logps/chosen": -442.0, + "logps/rejected": -384.0, + "loss": 0.493, + "rewards/accuracies": 0.828125, + "rewards/chosen": -1.234375, + "rewards/margins": 6.84375, + "rewards/rejected": -8.0625, + "step": 611 + }, + { + "epoch": 0.750920245398773, + "grad_norm": 36.03030744089288, + "learning_rate": 1.1361570247933884e-06, + "logits/chosen": -0.042236328125, + "logits/rejected": -0.1796875, + "logps/chosen": -402.0, + "logps/rejected": -376.0, + "loss": 0.4916, + "rewards/accuracies": 0.8203125, + "rewards/chosen": -0.462890625, + "rewards/margins": 6.78125, + "rewards/rejected": -7.25, + "step": 612 + }, + { + "epoch": 0.7521472392638037, + "grad_norm": 35.23308877108843, + "learning_rate": 1.1355371900826446e-06, + "logits/chosen": -0.04052734375, + "logits/rejected": -0.08447265625, + "logps/chosen": -388.0, + "logps/rejected": -348.0, + "loss": 0.4598, + "rewards/accuracies": 0.8359375, + "rewards/chosen": -1.0078125, + "rewards/margins": 6.84375, + "rewards/rejected": -7.875, + "step": 613 + }, + { + "epoch": 0.7533742331288343, + "grad_norm": 35.827377930375896, + "learning_rate": 1.134917355371901e-06, + "logits/chosen": -0.095703125, + "logits/rejected": -0.2216796875, + "logps/chosen": -430.0, + "logps/rejected": -364.0, + "loss": 0.3803, + "rewards/accuracies": 0.828125, + "rewards/chosen": -0.333984375, + "rewards/margins": 7.375, + "rewards/rejected": -7.71875, + "step": 614 + }, + { + "epoch": 0.754601226993865, + "grad_norm": 31.70255072714647, + "learning_rate": 1.134297520661157e-06, + "logits/chosen": -0.07373046875, + "logits/rejected": -0.166015625, + "logps/chosen": -376.0, + "logps/rejected": -364.0, + "loss": 0.3435, + "rewards/accuracies": 0.84375, + "rewards/chosen": 0.037841796875, + "rewards/margins": 7.84375, + "rewards/rejected": -7.8125, + "step": 615 + }, + { + "epoch": 0.7558282208588957, + "grad_norm": 30.125086244839487, + "learning_rate": 1.1336776859504134e-06, + "logits/chosen": -0.004241943359375, + "logits/rejected": -0.15625, + "logps/chosen": -332.0, + "logps/rejected": -310.0, + "loss": 0.3151, + "rewards/accuracies": 0.8359375, + "rewards/chosen": -0.6328125, + "rewards/margins": 6.53125, + "rewards/rejected": -7.15625, + "step": 616 + }, + { + "epoch": 0.7570552147239263, + "grad_norm": 32.7474160937551, + "learning_rate": 1.1330578512396693e-06, + "logits/chosen": 0.030517578125, + "logits/rejected": -0.0390625, + "logps/chosen": -378.0, + "logps/rejected": -336.0, + "loss": 0.3705, + "rewards/accuracies": 0.828125, + "rewards/chosen": -0.419921875, + "rewards/margins": 6.75, + "rewards/rejected": -7.15625, + "step": 617 + }, + { + "epoch": 0.758282208588957, + "grad_norm": 45.056199177526004, + "learning_rate": 1.1324380165289255e-06, + "logits/chosen": 0.109375, + "logits/rejected": -0.056396484375, + "logps/chosen": -426.0, + "logps/rejected": -332.0, + "loss": 0.5324, + "rewards/accuracies": 0.78125, + "rewards/chosen": -0.427734375, + "rewards/margins": 6.1875, + "rewards/rejected": -6.625, + "step": 618 + }, + { + "epoch": 0.7595092024539877, + "grad_norm": 29.95096958444537, + "learning_rate": 1.1318181818181818e-06, + "logits/chosen": 0.08056640625, + "logits/rejected": -0.12060546875, + "logps/chosen": -432.0, + "logps/rejected": -350.0, + "loss": 0.2823, + "rewards/accuracies": 0.8828125, + "rewards/chosen": 0.36328125, + "rewards/margins": 7.71875, + "rewards/rejected": -7.34375, + "step": 619 + }, + { + "epoch": 0.7607361963190185, + "grad_norm": 37.035466575479155, + "learning_rate": 1.131198347107438e-06, + "logits/chosen": 0.08349609375, + "logits/rejected": -0.11181640625, + "logps/chosen": -410.0, + "logps/rejected": -346.0, + "loss": 0.3202, + "rewards/accuracies": 0.828125, + "rewards/chosen": -0.03271484375, + "rewards/margins": 6.65625, + "rewards/rejected": -6.71875, + "step": 620 + }, + { + "epoch": 0.7619631901840491, + "grad_norm": 30.546031931386054, + "learning_rate": 1.1305785123966943e-06, + "logits/chosen": -0.052490234375, + "logits/rejected": -0.2060546875, + "logps/chosen": -482.0, + "logps/rejected": -380.0, + "loss": 0.3488, + "rewards/accuracies": 0.8828125, + "rewards/chosen": 0.7109375, + "rewards/margins": 8.125, + "rewards/rejected": -7.40625, + "step": 621 + }, + { + "epoch": 0.7631901840490798, + "grad_norm": 30.102626290429743, + "learning_rate": 1.1299586776859505e-06, + "logits/chosen": 0.07666015625, + "logits/rejected": -0.0791015625, + "logps/chosen": -382.0, + "logps/rejected": -354.0, + "loss": 0.3683, + "rewards/accuracies": 0.8671875, + "rewards/chosen": 0.2060546875, + "rewards/margins": 7.375, + "rewards/rejected": -7.15625, + "step": 622 + }, + { + "epoch": 0.7644171779141105, + "grad_norm": 31.272882227536257, + "learning_rate": 1.1293388429752066e-06, + "logits/chosen": 0.041259765625, + "logits/rejected": -0.1318359375, + "logps/chosen": -456.0, + "logps/rejected": -390.0, + "loss": 0.338, + "rewards/accuracies": 0.8125, + "rewards/chosen": 1.0078125, + "rewards/margins": 7.78125, + "rewards/rejected": -6.78125, + "step": 623 + }, + { + "epoch": 0.7656441717791411, + "grad_norm": 29.25638333606671, + "learning_rate": 1.128719008264463e-06, + "logits/chosen": -0.036865234375, + "logits/rejected": -0.208984375, + "logps/chosen": -382.0, + "logps/rejected": -334.0, + "loss": 0.302, + "rewards/accuracies": 0.890625, + "rewards/chosen": 0.8671875, + "rewards/margins": 8.5, + "rewards/rejected": -7.59375, + "step": 624 + }, + { + "epoch": 0.7668711656441718, + "grad_norm": 27.76458750020888, + "learning_rate": 1.1280991735537191e-06, + "logits/chosen": 0.01214599609375, + "logits/rejected": -0.123046875, + "logps/chosen": -408.0, + "logps/rejected": -334.0, + "loss": 0.2518, + "rewards/accuracies": 0.859375, + "rewards/chosen": 0.69140625, + "rewards/margins": 7.15625, + "rewards/rejected": -6.46875, + "step": 625 + }, + { + "epoch": 0.7680981595092025, + "grad_norm": 34.03667353357896, + "learning_rate": 1.1274793388429753e-06, + "logits/chosen": -0.119140625, + "logits/rejected": -0.2451171875, + "logps/chosen": -416.0, + "logps/rejected": -358.0, + "loss": 0.4042, + "rewards/accuracies": 0.859375, + "rewards/chosen": 1.140625, + "rewards/margins": 7.03125, + "rewards/rejected": -5.875, + "step": 626 + }, + { + "epoch": 0.7693251533742331, + "grad_norm": 30.346756983376576, + "learning_rate": 1.1268595041322314e-06, + "logits/chosen": 0.095703125, + "logits/rejected": 0.0223388671875, + "logps/chosen": -314.0, + "logps/rejected": -332.0, + "loss": 0.3854, + "rewards/accuracies": 0.8046875, + "rewards/chosen": -0.65625, + "rewards/margins": 5.28125, + "rewards/rejected": -5.9375, + "step": 627 + }, + { + "epoch": 0.7705521472392638, + "grad_norm": 31.196230115400443, + "learning_rate": 1.1262396694214875e-06, + "logits/chosen": 0.0732421875, + "logits/rejected": -0.06884765625, + "logps/chosen": -380.0, + "logps/rejected": -304.0, + "loss": 0.2959, + "rewards/accuracies": 0.8515625, + "rewards/chosen": 0.70703125, + "rewards/margins": 6.40625, + "rewards/rejected": -5.71875, + "step": 628 + }, + { + "epoch": 0.7717791411042945, + "grad_norm": 27.109929558619076, + "learning_rate": 1.1256198347107439e-06, + "logits/chosen": 0.0262451171875, + "logits/rejected": -0.09033203125, + "logps/chosen": -352.0, + "logps/rejected": -292.0, + "loss": 0.2072, + "rewards/accuracies": 0.90625, + "rewards/chosen": 0.353515625, + "rewards/margins": 6.5, + "rewards/rejected": -6.15625, + "step": 629 + }, + { + "epoch": 0.7730061349693251, + "grad_norm": 27.79699333558513, + "learning_rate": 1.125e-06, + "logits/chosen": 0.0947265625, + "logits/rejected": -0.1318359375, + "logps/chosen": -374.0, + "logps/rejected": -372.0, + "loss": 0.3309, + "rewards/accuracies": 0.8671875, + "rewards/chosen": 1.21875, + "rewards/margins": 7.09375, + "rewards/rejected": -5.875, + "step": 630 + }, + { + "epoch": 0.7742331288343558, + "grad_norm": 32.55842929304895, + "learning_rate": 1.1243801652892562e-06, + "logits/chosen": -0.11474609375, + "logits/rejected": -0.2001953125, + "logps/chosen": -394.0, + "logps/rejected": -326.0, + "loss": 0.4115, + "rewards/accuracies": 0.8046875, + "rewards/chosen": 1.03125, + "rewards/margins": 6.71875, + "rewards/rejected": -5.6875, + "step": 631 + }, + { + "epoch": 0.7754601226993865, + "grad_norm": 30.882292248466843, + "learning_rate": 1.1237603305785125e-06, + "logits/chosen": -0.0111083984375, + "logits/rejected": -0.134765625, + "logps/chosen": -406.0, + "logps/rejected": -344.0, + "loss": 0.3282, + "rewards/accuracies": 0.828125, + "rewards/chosen": 1.125, + "rewards/margins": 6.21875, + "rewards/rejected": -5.09375, + "step": 632 + }, + { + "epoch": 0.7766871165644171, + "grad_norm": 32.30019011289937, + "learning_rate": 1.1231404958677687e-06, + "logits/chosen": 0.072265625, + "logits/rejected": -0.03857421875, + "logps/chosen": -342.0, + "logps/rejected": -338.0, + "loss": 0.3515, + "rewards/accuracies": 0.8046875, + "rewards/chosen": 0.39453125, + "rewards/margins": 5.71875, + "rewards/rejected": -5.34375, + "step": 633 + }, + { + "epoch": 0.7779141104294478, + "grad_norm": 26.84260857948927, + "learning_rate": 1.1225206611570248e-06, + "logits/chosen": -0.052490234375, + "logits/rejected": -0.248046875, + "logps/chosen": -416.0, + "logps/rejected": -328.0, + "loss": 0.2366, + "rewards/accuracies": 0.9140625, + "rewards/chosen": 2.546875, + "rewards/margins": 8.3125, + "rewards/rejected": -5.78125, + "step": 634 + }, + { + "epoch": 0.7791411042944786, + "grad_norm": 23.744218806065337, + "learning_rate": 1.121900826446281e-06, + "logits/chosen": -0.0380859375, + "logits/rejected": -0.09912109375, + "logps/chosen": -362.0, + "logps/rejected": -302.0, + "loss": 0.2194, + "rewards/accuracies": 0.90625, + "rewards/chosen": 0.6484375, + "rewards/margins": 6.1875, + "rewards/rejected": -5.5625, + "step": 635 + }, + { + "epoch": 0.7803680981595092, + "grad_norm": 34.66381046216169, + "learning_rate": 1.121280991735537e-06, + "logits/chosen": -0.06201171875, + "logits/rejected": -0.166015625, + "logps/chosen": -414.0, + "logps/rejected": -368.0, + "loss": 0.4205, + "rewards/accuracies": 0.84375, + "rewards/chosen": 1.46875, + "rewards/margins": 6.65625, + "rewards/rejected": -5.1875, + "step": 636 + }, + { + "epoch": 0.7815950920245399, + "grad_norm": 41.24072249858344, + "learning_rate": 1.1206611570247934e-06, + "logits/chosen": 0.11962890625, + "logits/rejected": -0.0615234375, + "logps/chosen": -382.0, + "logps/rejected": -306.0, + "loss": 0.4805, + "rewards/accuracies": 0.8125, + "rewards/chosen": 0.439453125, + "rewards/margins": 6.125, + "rewards/rejected": -5.6875, + "step": 637 + }, + { + "epoch": 0.7828220858895706, + "grad_norm": 31.604533876936294, + "learning_rate": 1.1200413223140496e-06, + "logits/chosen": -0.044677734375, + "logits/rejected": -0.2451171875, + "logps/chosen": -406.0, + "logps/rejected": -340.0, + "loss": 0.2858, + "rewards/accuracies": 0.859375, + "rewards/chosen": 1.3046875, + "rewards/margins": 6.9375, + "rewards/rejected": -5.625, + "step": 638 + }, + { + "epoch": 0.7840490797546013, + "grad_norm": 34.862589865848754, + "learning_rate": 1.1194214876033057e-06, + "logits/chosen": 0.0869140625, + "logits/rejected": -0.080078125, + "logps/chosen": -408.0, + "logps/rejected": -342.0, + "loss": 0.4829, + "rewards/accuracies": 0.828125, + "rewards/chosen": 1.046875, + "rewards/margins": 6.34375, + "rewards/rejected": -5.28125, + "step": 639 + }, + { + "epoch": 0.7852760736196319, + "grad_norm": 29.526720756728196, + "learning_rate": 1.118801652892562e-06, + "logits/chosen": -0.08837890625, + "logits/rejected": -0.2119140625, + "logps/chosen": -446.0, + "logps/rejected": -394.0, + "loss": 0.2982, + "rewards/accuracies": 0.890625, + "rewards/chosen": 1.6328125, + "rewards/margins": 7.5, + "rewards/rejected": -5.875, + "step": 640 + }, + { + "epoch": 0.7865030674846626, + "grad_norm": 38.16790195191018, + "learning_rate": 1.1181818181818182e-06, + "logits/chosen": -0.1259765625, + "logits/rejected": -0.251953125, + "logps/chosen": -398.0, + "logps/rejected": -360.0, + "loss": 0.5882, + "rewards/accuracies": 0.8046875, + "rewards/chosen": 0.90625, + "rewards/margins": 6.15625, + "rewards/rejected": -5.25, + "step": 641 + }, + { + "epoch": 0.7877300613496933, + "grad_norm": 25.77138922359556, + "learning_rate": 1.1175619834710746e-06, + "logits/chosen": -0.01422119140625, + "logits/rejected": -0.2333984375, + "logps/chosen": -352.0, + "logps/rejected": -304.0, + "loss": 0.2926, + "rewards/accuracies": 0.875, + "rewards/chosen": 0.9140625, + "rewards/margins": 6.65625, + "rewards/rejected": -5.75, + "step": 642 + }, + { + "epoch": 0.7889570552147239, + "grad_norm": 35.220692662985385, + "learning_rate": 1.1169421487603307e-06, + "logits/chosen": -0.1748046875, + "logits/rejected": -0.333984375, + "logps/chosen": -406.0, + "logps/rejected": -386.0, + "loss": 0.327, + "rewards/accuracies": 0.8671875, + "rewards/chosen": 0.92578125, + "rewards/margins": 6.84375, + "rewards/rejected": -5.9375, + "step": 643 + }, + { + "epoch": 0.7901840490797546, + "grad_norm": 30.676171080381682, + "learning_rate": 1.1163223140495866e-06, + "logits/chosen": -0.07666015625, + "logits/rejected": -0.21875, + "logps/chosen": -392.0, + "logps/rejected": -338.0, + "loss": 0.3619, + "rewards/accuracies": 0.8515625, + "rewards/chosen": 0.75390625, + "rewards/margins": 7.25, + "rewards/rejected": -6.5, + "step": 644 + }, + { + "epoch": 0.7914110429447853, + "grad_norm": 29.15888390878358, + "learning_rate": 1.115702479338843e-06, + "logits/chosen": -0.03955078125, + "logits/rejected": -0.234375, + "logps/chosen": -336.0, + "logps/rejected": -346.0, + "loss": 0.2816, + "rewards/accuracies": 0.859375, + "rewards/chosen": 0.98046875, + "rewards/margins": 7.53125, + "rewards/rejected": -6.5625, + "step": 645 + }, + { + "epoch": 0.7926380368098159, + "grad_norm": 28.593454046054227, + "learning_rate": 1.1150826446280991e-06, + "logits/chosen": 0.047119140625, + "logits/rejected": -0.134765625, + "logps/chosen": -386.0, + "logps/rejected": -340.0, + "loss": 0.2765, + "rewards/accuracies": 0.8671875, + "rewards/chosen": 0.68359375, + "rewards/margins": 7.28125, + "rewards/rejected": -6.59375, + "step": 646 + }, + { + "epoch": 0.7938650306748466, + "grad_norm": 32.656529890149606, + "learning_rate": 1.1144628099173555e-06, + "logits/chosen": -0.033935546875, + "logits/rejected": -0.1708984375, + "logps/chosen": -386.0, + "logps/rejected": -334.0, + "loss": 0.3573, + "rewards/accuracies": 0.84375, + "rewards/chosen": 0.9296875, + "rewards/margins": 7.78125, + "rewards/rejected": -6.875, + "step": 647 + }, + { + "epoch": 0.7950920245398773, + "grad_norm": 36.18303049965862, + "learning_rate": 1.1138429752066116e-06, + "logits/chosen": -0.0908203125, + "logits/rejected": -0.08837890625, + "logps/chosen": -334.0, + "logps/rejected": -336.0, + "loss": 0.438, + "rewards/accuracies": 0.8125, + "rewards/chosen": -0.0986328125, + "rewards/margins": 6.375, + "rewards/rejected": -6.46875, + "step": 648 + }, + { + "epoch": 0.7963190184049079, + "grad_norm": 33.937125436202074, + "learning_rate": 1.1132231404958678e-06, + "logits/chosen": 0.061279296875, + "logits/rejected": -0.027587890625, + "logps/chosen": -372.0, + "logps/rejected": -342.0, + "loss": 0.5003, + "rewards/accuracies": 0.8125, + "rewards/chosen": 0.0732421875, + "rewards/margins": 6.0, + "rewards/rejected": -5.9375, + "step": 649 + }, + { + "epoch": 0.7975460122699386, + "grad_norm": 36.01649609906715, + "learning_rate": 1.1126033057851241e-06, + "logits/chosen": 0.0264892578125, + "logits/rejected": -0.1298828125, + "logps/chosen": -432.0, + "logps/rejected": -364.0, + "loss": 0.4749, + "rewards/accuracies": 0.8203125, + "rewards/chosen": 0.71875, + "rewards/margins": 6.5, + "rewards/rejected": -5.78125, + "step": 650 + }, + { + "epoch": 0.7987730061349694, + "grad_norm": 37.059100709512336, + "learning_rate": 1.1119834710743803e-06, + "logits/chosen": -0.12109375, + "logits/rejected": -0.1845703125, + "logps/chosen": -376.0, + "logps/rejected": -362.0, + "loss": 0.4208, + "rewards/accuracies": 0.8359375, + "rewards/chosen": 0.9609375, + "rewards/margins": 8.0625, + "rewards/rejected": -7.125, + "step": 651 + }, + { + "epoch": 0.8, + "grad_norm": 40.43433181753938, + "learning_rate": 1.1113636363636364e-06, + "logits/chosen": -0.031494140625, + "logits/rejected": -0.1455078125, + "logps/chosen": -376.0, + "logps/rejected": -360.0, + "loss": 0.4169, + "rewards/accuracies": 0.8515625, + "rewards/chosen": -0.5078125, + "rewards/margins": 6.625, + "rewards/rejected": -7.125, + "step": 652 + }, + { + "epoch": 0.8012269938650307, + "grad_norm": 32.30744337627645, + "learning_rate": 1.1107438016528925e-06, + "logits/chosen": 0.032958984375, + "logits/rejected": -0.142578125, + "logps/chosen": -358.0, + "logps/rejected": -346.0, + "loss": 0.2863, + "rewards/accuracies": 0.8359375, + "rewards/chosen": 0.0498046875, + "rewards/margins": 7.625, + "rewards/rejected": -7.59375, + "step": 653 + }, + { + "epoch": 0.8024539877300614, + "grad_norm": 23.943395122229603, + "learning_rate": 1.1101239669421487e-06, + "logits/chosen": 0.050537109375, + "logits/rejected": -0.15234375, + "logps/chosen": -400.0, + "logps/rejected": -336.0, + "loss": 0.226, + "rewards/accuracies": 0.890625, + "rewards/chosen": 0.333984375, + "rewards/margins": 7.96875, + "rewards/rejected": -7.625, + "step": 654 + }, + { + "epoch": 0.803680981595092, + "grad_norm": 31.423267487900187, + "learning_rate": 1.109504132231405e-06, + "logits/chosen": -0.007659912109375, + "logits/rejected": -0.1787109375, + "logps/chosen": -396.0, + "logps/rejected": -366.0, + "loss": 0.3332, + "rewards/accuracies": 0.8359375, + "rewards/chosen": -0.1943359375, + "rewards/margins": 7.125, + "rewards/rejected": -7.3125, + "step": 655 + }, + { + "epoch": 0.8049079754601227, + "grad_norm": 35.483670047647635, + "learning_rate": 1.1088842975206612e-06, + "logits/chosen": 0.055419921875, + "logits/rejected": -0.08740234375, + "logps/chosen": -380.0, + "logps/rejected": -344.0, + "loss": 0.4163, + "rewards/accuracies": 0.828125, + "rewards/chosen": -0.419921875, + "rewards/margins": 7.09375, + "rewards/rejected": -7.5, + "step": 656 + }, + { + "epoch": 0.8061349693251534, + "grad_norm": 27.21341640433181, + "learning_rate": 1.1082644628099173e-06, + "logits/chosen": -0.10107421875, + "logits/rejected": -0.2431640625, + "logps/chosen": -372.0, + "logps/rejected": -354.0, + "loss": 0.3384, + "rewards/accuracies": 0.8359375, + "rewards/chosen": -0.388671875, + "rewards/margins": 7.1875, + "rewards/rejected": -7.5625, + "step": 657 + }, + { + "epoch": 0.807361963190184, + "grad_norm": 31.34176371809714, + "learning_rate": 1.1076446280991737e-06, + "logits/chosen": -0.08349609375, + "logits/rejected": -0.212890625, + "logps/chosen": -402.0, + "logps/rejected": -366.0, + "loss": 0.3063, + "rewards/accuracies": 0.875, + "rewards/chosen": -0.412109375, + "rewards/margins": 7.21875, + "rewards/rejected": -7.625, + "step": 658 + }, + { + "epoch": 0.8085889570552147, + "grad_norm": 32.62830203002507, + "learning_rate": 1.1070247933884298e-06, + "logits/chosen": -0.0291748046875, + "logits/rejected": -0.16796875, + "logps/chosen": -374.0, + "logps/rejected": -348.0, + "loss": 0.3798, + "rewards/accuracies": 0.84375, + "rewards/chosen": -1.1171875, + "rewards/margins": 6.59375, + "rewards/rejected": -7.71875, + "step": 659 + }, + { + "epoch": 0.8098159509202454, + "grad_norm": 29.335771689820604, + "learning_rate": 1.106404958677686e-06, + "logits/chosen": -0.0380859375, + "logits/rejected": -0.220703125, + "logps/chosen": -412.0, + "logps/rejected": -374.0, + "loss": 0.3215, + "rewards/accuracies": 0.8359375, + "rewards/chosen": 0.053466796875, + "rewards/margins": 7.8125, + "rewards/rejected": -7.75, + "step": 660 + }, + { + "epoch": 0.811042944785276, + "grad_norm": 39.19555957381655, + "learning_rate": 1.1057851239669423e-06, + "logits/chosen": 0.054443359375, + "logits/rejected": -0.052734375, + "logps/chosen": -374.0, + "logps/rejected": -318.0, + "loss": 0.5135, + "rewards/accuracies": 0.8046875, + "rewards/chosen": -1.09375, + "rewards/margins": 6.46875, + "rewards/rejected": -7.5625, + "step": 661 + }, + { + "epoch": 0.8122699386503067, + "grad_norm": 25.59951308246572, + "learning_rate": 1.1051652892561982e-06, + "logits/chosen": 0.05224609375, + "logits/rejected": -0.06884765625, + "logps/chosen": -372.0, + "logps/rejected": -342.0, + "loss": 0.218, + "rewards/accuracies": 0.890625, + "rewards/chosen": 0.3671875, + "rewards/margins": 8.1875, + "rewards/rejected": -7.8125, + "step": 662 + }, + { + "epoch": 0.8134969325153374, + "grad_norm": 29.103568092770104, + "learning_rate": 1.1045454545454546e-06, + "logits/chosen": 0.1142578125, + "logits/rejected": -0.080078125, + "logps/chosen": -362.0, + "logps/rejected": -340.0, + "loss": 0.3666, + "rewards/accuracies": 0.8125, + "rewards/chosen": 0.54296875, + "rewards/margins": 7.375, + "rewards/rejected": -6.8125, + "step": 663 + }, + { + "epoch": 0.8147239263803681, + "grad_norm": 24.88151445188299, + "learning_rate": 1.1039256198347107e-06, + "logits/chosen": -0.2255859375, + "logits/rejected": -0.35546875, + "logps/chosen": -388.0, + "logps/rejected": -376.0, + "loss": 0.213, + "rewards/accuracies": 0.9140625, + "rewards/chosen": 0.7578125, + "rewards/margins": 9.0625, + "rewards/rejected": -8.3125, + "step": 664 + }, + { + "epoch": 0.8159509202453987, + "grad_norm": 34.48345990804285, + "learning_rate": 1.1033057851239669e-06, + "logits/chosen": 0.1259765625, + "logits/rejected": -0.0419921875, + "logps/chosen": -342.0, + "logps/rejected": -308.0, + "loss": 0.4057, + "rewards/accuracies": 0.8125, + "rewards/chosen": -0.2470703125, + "rewards/margins": 7.125, + "rewards/rejected": -7.375, + "step": 665 + }, + { + "epoch": 0.8171779141104294, + "grad_norm": 35.79692152845337, + "learning_rate": 1.1026859504132232e-06, + "logits/chosen": 0.0791015625, + "logits/rejected": -0.0927734375, + "logps/chosen": -422.0, + "logps/rejected": -362.0, + "loss": 0.4703, + "rewards/accuracies": 0.8125, + "rewards/chosen": 0.68359375, + "rewards/margins": 7.28125, + "rewards/rejected": -6.59375, + "step": 666 + }, + { + "epoch": 0.8184049079754602, + "grad_norm": 29.967347606106397, + "learning_rate": 1.1020661157024794e-06, + "logits/chosen": -0.010986328125, + "logits/rejected": -0.2109375, + "logps/chosen": -376.0, + "logps/rejected": -344.0, + "loss": 0.2918, + "rewards/accuracies": 0.890625, + "rewards/chosen": 0.7421875, + "rewards/margins": 7.53125, + "rewards/rejected": -6.78125, + "step": 667 + }, + { + "epoch": 0.8196319018404908, + "grad_norm": 30.862796893024488, + "learning_rate": 1.1014462809917355e-06, + "logits/chosen": 0.0123291015625, + "logits/rejected": -0.119140625, + "logps/chosen": -392.0, + "logps/rejected": -360.0, + "loss": 0.3721, + "rewards/accuracies": 0.8359375, + "rewards/chosen": 0.65625, + "rewards/margins": 7.09375, + "rewards/rejected": -6.4375, + "step": 668 + }, + { + "epoch": 0.8208588957055215, + "grad_norm": 35.17919160040694, + "learning_rate": 1.1008264462809919e-06, + "logits/chosen": -0.052490234375, + "logits/rejected": -0.2060546875, + "logps/chosen": -340.0, + "logps/rejected": -338.0, + "loss": 0.3167, + "rewards/accuracies": 0.875, + "rewards/chosen": 0.6953125, + "rewards/margins": 7.09375, + "rewards/rejected": -6.40625, + "step": 669 + }, + { + "epoch": 0.8220858895705522, + "grad_norm": 25.88170844832394, + "learning_rate": 1.100206611570248e-06, + "logits/chosen": 0.17578125, + "logits/rejected": 0.0250244140625, + "logps/chosen": -324.0, + "logps/rejected": -302.0, + "loss": 0.2514, + "rewards/accuracies": 0.8984375, + "rewards/chosen": 1.0390625, + "rewards/margins": 6.8125, + "rewards/rejected": -5.78125, + "step": 670 + }, + { + "epoch": 0.8233128834355828, + "grad_norm": 30.762328731871737, + "learning_rate": 1.0995867768595042e-06, + "logits/chosen": 0.0390625, + "logits/rejected": -0.12109375, + "logps/chosen": -372.0, + "logps/rejected": -332.0, + "loss": 0.2995, + "rewards/accuracies": 0.890625, + "rewards/chosen": 1.7421875, + "rewards/margins": 7.5, + "rewards/rejected": -5.78125, + "step": 671 + }, + { + "epoch": 0.8245398773006135, + "grad_norm": 30.48600462690235, + "learning_rate": 1.0989669421487603e-06, + "logits/chosen": 0.054931640625, + "logits/rejected": -0.2197265625, + "logps/chosen": -446.0, + "logps/rejected": -332.0, + "loss": 0.3199, + "rewards/accuracies": 0.8515625, + "rewards/chosen": 2.015625, + "rewards/margins": 7.71875, + "rewards/rejected": -5.6875, + "step": 672 + }, + { + "epoch": 0.8257668711656442, + "grad_norm": 32.03977574623459, + "learning_rate": 1.0983471074380164e-06, + "logits/chosen": 0.1416015625, + "logits/rejected": -0.0986328125, + "logps/chosen": -390.0, + "logps/rejected": -328.0, + "loss": 0.3096, + "rewards/accuracies": 0.859375, + "rewards/chosen": 1.109375, + "rewards/margins": 6.21875, + "rewards/rejected": -5.125, + "step": 673 + }, + { + "epoch": 0.8269938650306748, + "grad_norm": 32.93059543816188, + "learning_rate": 1.0977272727272728e-06, + "logits/chosen": 0.162109375, + "logits/rejected": 0.00286865234375, + "logps/chosen": -358.0, + "logps/rejected": -330.0, + "loss": 0.3985, + "rewards/accuracies": 0.8359375, + "rewards/chosen": 1.0546875, + "rewards/margins": 6.59375, + "rewards/rejected": -5.53125, + "step": 674 + }, + { + "epoch": 0.8282208588957055, + "grad_norm": 34.971531039669564, + "learning_rate": 1.097107438016529e-06, + "logits/chosen": 0.08203125, + "logits/rejected": -0.130859375, + "logps/chosen": -426.0, + "logps/rejected": -368.0, + "loss": 0.4407, + "rewards/accuracies": 0.84375, + "rewards/chosen": 1.3828125, + "rewards/margins": 7.40625, + "rewards/rejected": -6.0, + "step": 675 + }, + { + "epoch": 0.8294478527607362, + "grad_norm": 34.2224078373622, + "learning_rate": 1.0964876033057853e-06, + "logits/chosen": -0.037353515625, + "logits/rejected": -0.14453125, + "logps/chosen": -394.0, + "logps/rejected": -352.0, + "loss": 0.3813, + "rewards/accuracies": 0.859375, + "rewards/chosen": 0.4375, + "rewards/margins": 6.8125, + "rewards/rejected": -6.375, + "step": 676 + }, + { + "epoch": 0.8306748466257668, + "grad_norm": 22.049121537748125, + "learning_rate": 1.0958677685950414e-06, + "logits/chosen": 0.09423828125, + "logits/rejected": -0.08447265625, + "logps/chosen": -338.0, + "logps/rejected": -316.0, + "loss": 0.2974, + "rewards/accuracies": 0.875, + "rewards/chosen": 0.2099609375, + "rewards/margins": 7.1875, + "rewards/rejected": -6.96875, + "step": 677 + }, + { + "epoch": 0.8319018404907975, + "grad_norm": 31.042587162348006, + "learning_rate": 1.0952479338842976e-06, + "logits/chosen": 0.07421875, + "logits/rejected": -0.1376953125, + "logps/chosen": -422.0, + "logps/rejected": -362.0, + "loss": 0.4105, + "rewards/accuracies": 0.84375, + "rewards/chosen": 0.439453125, + "rewards/margins": 6.96875, + "rewards/rejected": -6.53125, + "step": 678 + }, + { + "epoch": 0.8331288343558282, + "grad_norm": 32.125565947605324, + "learning_rate": 1.0946280991735537e-06, + "logits/chosen": 0.1689453125, + "logits/rejected": 0.06201171875, + "logps/chosen": -344.0, + "logps/rejected": -328.0, + "loss": 0.3816, + "rewards/accuracies": 0.859375, + "rewards/chosen": 0.44921875, + "rewards/margins": 6.53125, + "rewards/rejected": -6.0625, + "step": 679 + }, + { + "epoch": 0.8343558282208589, + "grad_norm": 30.54691503767555, + "learning_rate": 1.0940082644628098e-06, + "logits/chosen": 0.025390625, + "logits/rejected": -0.150390625, + "logps/chosen": -454.0, + "logps/rejected": -374.0, + "loss": 0.3146, + "rewards/accuracies": 0.84375, + "rewards/chosen": 1.25, + "rewards/margins": 8.125, + "rewards/rejected": -6.875, + "step": 680 + }, + { + "epoch": 0.8355828220858895, + "grad_norm": 25.35697096492706, + "learning_rate": 1.0933884297520662e-06, + "logits/chosen": -0.0262451171875, + "logits/rejected": -0.1796875, + "logps/chosen": -348.0, + "logps/rejected": -338.0, + "loss": 0.2471, + "rewards/accuracies": 0.890625, + "rewards/chosen": 0.72265625, + "rewards/margins": 8.4375, + "rewards/rejected": -7.71875, + "step": 681 + }, + { + "epoch": 0.8368098159509203, + "grad_norm": 29.372252652247152, + "learning_rate": 1.0927685950413223e-06, + "logits/chosen": -0.000934600830078125, + "logits/rejected": -0.115234375, + "logps/chosen": -352.0, + "logps/rejected": -334.0, + "loss": 0.3998, + "rewards/accuracies": 0.84375, + "rewards/chosen": 0.5546875, + "rewards/margins": 7.4375, + "rewards/rejected": -6.90625, + "step": 682 + }, + { + "epoch": 0.838036809815951, + "grad_norm": 28.48014927993986, + "learning_rate": 1.0921487603305785e-06, + "logits/chosen": -0.0264892578125, + "logits/rejected": -0.1328125, + "logps/chosen": -374.0, + "logps/rejected": -330.0, + "loss": 0.3192, + "rewards/accuracies": 0.8359375, + "rewards/chosen": 0.67578125, + "rewards/margins": 8.0625, + "rewards/rejected": -7.375, + "step": 683 + }, + { + "epoch": 0.8392638036809816, + "grad_norm": 30.9956365019602, + "learning_rate": 1.0915289256198348e-06, + "logits/chosen": 0.2001953125, + "logits/rejected": 0.08544921875, + "logps/chosen": -322.0, + "logps/rejected": -352.0, + "loss": 0.3553, + "rewards/accuracies": 0.8515625, + "rewards/chosen": -0.6953125, + "rewards/margins": 6.3125, + "rewards/rejected": -7.03125, + "step": 684 + }, + { + "epoch": 0.8404907975460123, + "grad_norm": 31.024167516184594, + "learning_rate": 1.090909090909091e-06, + "logits/chosen": 0.040283203125, + "logits/rejected": -0.12255859375, + "logps/chosen": -404.0, + "logps/rejected": -380.0, + "loss": 0.3423, + "rewards/accuracies": 0.875, + "rewards/chosen": 0.400390625, + "rewards/margins": 8.625, + "rewards/rejected": -8.25, + "step": 685 + }, + { + "epoch": 0.841717791411043, + "grad_norm": 25.880225574187453, + "learning_rate": 1.0902892561983471e-06, + "logits/chosen": 0.031982421875, + "logits/rejected": -0.095703125, + "logps/chosen": -400.0, + "logps/rejected": -348.0, + "loss": 0.2926, + "rewards/accuracies": 0.8515625, + "rewards/chosen": 0.27734375, + "rewards/margins": 7.9375, + "rewards/rejected": -7.65625, + "step": 686 + }, + { + "epoch": 0.8429447852760736, + "grad_norm": 28.80339614616542, + "learning_rate": 1.0896694214876035e-06, + "logits/chosen": 0.03759765625, + "logits/rejected": -0.1044921875, + "logps/chosen": -380.0, + "logps/rejected": -364.0, + "loss": 0.2973, + "rewards/accuracies": 0.8359375, + "rewards/chosen": 0.484375, + "rewards/margins": 7.90625, + "rewards/rejected": -7.40625, + "step": 687 + }, + { + "epoch": 0.8441717791411043, + "grad_norm": 30.11912577456189, + "learning_rate": 1.0890495867768594e-06, + "logits/chosen": -0.000240325927734375, + "logits/rejected": -0.158203125, + "logps/chosen": -384.0, + "logps/rejected": -328.0, + "loss": 0.327, + "rewards/accuracies": 0.8203125, + "rewards/chosen": -0.5625, + "rewards/margins": 7.5, + "rewards/rejected": -8.0625, + "step": 688 + }, + { + "epoch": 0.845398773006135, + "grad_norm": 30.49529821286284, + "learning_rate": 1.0884297520661158e-06, + "logits/chosen": 0.027099609375, + "logits/rejected": -0.058837890625, + "logps/chosen": -382.0, + "logps/rejected": -348.0, + "loss": 0.344, + "rewards/accuracies": 0.8359375, + "rewards/chosen": 0.271484375, + "rewards/margins": 7.65625, + "rewards/rejected": -7.40625, + "step": 689 + }, + { + "epoch": 0.8466257668711656, + "grad_norm": 41.86773593768432, + "learning_rate": 1.0878099173553719e-06, + "logits/chosen": 0.171875, + "logits/rejected": -0.0174560546875, + "logps/chosen": -414.0, + "logps/rejected": -370.0, + "loss": 0.5772, + "rewards/accuracies": 0.7890625, + "rewards/chosen": 0.408203125, + "rewards/margins": 7.625, + "rewards/rejected": -7.21875, + "step": 690 + }, + { + "epoch": 0.8478527607361963, + "grad_norm": 27.177968327720407, + "learning_rate": 1.087190082644628e-06, + "logits/chosen": -0.162109375, + "logits/rejected": -0.2734375, + "logps/chosen": -366.0, + "logps/rejected": -368.0, + "loss": 0.3098, + "rewards/accuracies": 0.859375, + "rewards/chosen": 0.29296875, + "rewards/margins": 8.1875, + "rewards/rejected": -7.90625, + "step": 691 + }, + { + "epoch": 0.849079754601227, + "grad_norm": 38.42433716881971, + "learning_rate": 1.0865702479338844e-06, + "logits/chosen": 0.0179443359375, + "logits/rejected": -0.142578125, + "logps/chosen": -416.0, + "logps/rejected": -358.0, + "loss": 0.3292, + "rewards/accuracies": 0.859375, + "rewards/chosen": 0.8203125, + "rewards/margins": 8.375, + "rewards/rejected": -7.53125, + "step": 692 + }, + { + "epoch": 0.8503067484662576, + "grad_norm": 31.708672912333903, + "learning_rate": 1.0859504132231405e-06, + "logits/chosen": -0.00286865234375, + "logits/rejected": -0.1904296875, + "logps/chosen": -452.0, + "logps/rejected": -386.0, + "loss": 0.3787, + "rewards/accuracies": 0.84375, + "rewards/chosen": 1.0625, + "rewards/margins": 8.875, + "rewards/rejected": -7.8125, + "step": 693 + }, + { + "epoch": 0.8515337423312883, + "grad_norm": 39.65580173728978, + "learning_rate": 1.0853305785123967e-06, + "logits/chosen": 0.024658203125, + "logits/rejected": -0.162109375, + "logps/chosen": -384.0, + "logps/rejected": -352.0, + "loss": 0.5333, + "rewards/accuracies": 0.8203125, + "rewards/chosen": -0.671875, + "rewards/margins": 6.09375, + "rewards/rejected": -6.75, + "step": 694 + }, + { + "epoch": 0.852760736196319, + "grad_norm": 38.14890644660088, + "learning_rate": 1.084710743801653e-06, + "logits/chosen": 0.06103515625, + "logits/rejected": -0.0179443359375, + "logps/chosen": -338.0, + "logps/rejected": -322.0, + "loss": 0.5352, + "rewards/accuracies": 0.75, + "rewards/chosen": -0.4609375, + "rewards/margins": 5.375, + "rewards/rejected": -5.8125, + "step": 695 + }, + { + "epoch": 0.8539877300613496, + "grad_norm": 32.915799758521196, + "learning_rate": 1.0840909090909092e-06, + "logits/chosen": 0.0947265625, + "logits/rejected": -0.08154296875, + "logps/chosen": -390.0, + "logps/rejected": -336.0, + "loss": 0.3241, + "rewards/accuracies": 0.84375, + "rewards/chosen": 0.365234375, + "rewards/margins": 7.125, + "rewards/rejected": -6.78125, + "step": 696 + }, + { + "epoch": 0.8552147239263803, + "grad_norm": 35.9913417584187, + "learning_rate": 1.0834710743801653e-06, + "logits/chosen": -0.0771484375, + "logits/rejected": -0.10009765625, + "logps/chosen": -398.0, + "logps/rejected": -384.0, + "loss": 0.4041, + "rewards/accuracies": 0.84375, + "rewards/chosen": 0.9375, + "rewards/margins": 8.0, + "rewards/rejected": -7.0625, + "step": 697 + }, + { + "epoch": 0.8564417177914111, + "grad_norm": 31.064185654999054, + "learning_rate": 1.0828512396694214e-06, + "logits/chosen": 0.08984375, + "logits/rejected": -0.027587890625, + "logps/chosen": -404.0, + "logps/rejected": -366.0, + "loss": 0.3837, + "rewards/accuracies": 0.875, + "rewards/chosen": 0.392578125, + "rewards/margins": 7.5, + "rewards/rejected": -7.09375, + "step": 698 + }, + { + "epoch": 0.8576687116564418, + "grad_norm": 26.06263909227069, + "learning_rate": 1.0822314049586776e-06, + "logits/chosen": 0.0260009765625, + "logits/rejected": -0.053955078125, + "logps/chosen": -346.0, + "logps/rejected": -340.0, + "loss": 0.2587, + "rewards/accuracies": 0.8671875, + "rewards/chosen": -0.228515625, + "rewards/margins": 6.46875, + "rewards/rejected": -6.6875, + "step": 699 + }, + { + "epoch": 0.8588957055214724, + "grad_norm": 33.7346004258113, + "learning_rate": 1.081611570247934e-06, + "logits/chosen": 0.1484375, + "logits/rejected": 0.06005859375, + "logps/chosen": -380.0, + "logps/rejected": -342.0, + "loss": 0.398, + "rewards/accuracies": 0.8359375, + "rewards/chosen": -0.0072021484375, + "rewards/margins": 6.28125, + "rewards/rejected": -6.28125, + "step": 700 + }, + { + "epoch": 0.8601226993865031, + "grad_norm": 36.508608167851875, + "learning_rate": 1.08099173553719e-06, + "logits/chosen": 0.1630859375, + "logits/rejected": 0.05810546875, + "logps/chosen": -390.0, + "logps/rejected": -342.0, + "loss": 0.4454, + "rewards/accuracies": 0.78125, + "rewards/chosen": -0.267578125, + "rewards/margins": 6.5625, + "rewards/rejected": -6.8125, + "step": 701 + }, + { + "epoch": 0.8613496932515338, + "grad_norm": 29.752428717784724, + "learning_rate": 1.0803719008264462e-06, + "logits/chosen": -0.005767822265625, + "logits/rejected": -0.30078125, + "logps/chosen": -406.0, + "logps/rejected": -348.0, + "loss": 0.2347, + "rewards/accuracies": 0.859375, + "rewards/chosen": 0.91015625, + "rewards/margins": 8.0, + "rewards/rejected": -7.09375, + "step": 702 + }, + { + "epoch": 0.8625766871165644, + "grad_norm": 28.46276315635888, + "learning_rate": 1.0797520661157026e-06, + "logits/chosen": 0.1513671875, + "logits/rejected": 0.0250244140625, + "logps/chosen": -424.0, + "logps/rejected": -358.0, + "loss": 0.3098, + "rewards/accuracies": 0.859375, + "rewards/chosen": 0.1396484375, + "rewards/margins": 6.53125, + "rewards/rejected": -6.375, + "step": 703 + }, + { + "epoch": 0.8638036809815951, + "grad_norm": 28.331719307949253, + "learning_rate": 1.0791322314049587e-06, + "logits/chosen": 0.06689453125, + "logits/rejected": -0.09033203125, + "logps/chosen": -332.0, + "logps/rejected": -316.0, + "loss": 0.3639, + "rewards/accuracies": 0.859375, + "rewards/chosen": 0.07861328125, + "rewards/margins": 6.96875, + "rewards/rejected": -6.875, + "step": 704 + }, + { + "epoch": 0.8650306748466258, + "grad_norm": 33.73975373967079, + "learning_rate": 1.078512396694215e-06, + "logits/chosen": -0.0615234375, + "logits/rejected": -0.177734375, + "logps/chosen": -424.0, + "logps/rejected": -376.0, + "loss": 0.3238, + "rewards/accuracies": 0.8515625, + "rewards/chosen": 0.75, + "rewards/margins": 8.0, + "rewards/rejected": -7.25, + "step": 705 + }, + { + "epoch": 0.8662576687116564, + "grad_norm": 32.06391249106975, + "learning_rate": 1.077892561983471e-06, + "logits/chosen": 0.01544189453125, + "logits/rejected": -0.1796875, + "logps/chosen": -410.0, + "logps/rejected": -316.0, + "loss": 0.3915, + "rewards/accuracies": 0.8515625, + "rewards/chosen": 0.1748046875, + "rewards/margins": 6.84375, + "rewards/rejected": -6.6875, + "step": 706 + }, + { + "epoch": 0.8674846625766871, + "grad_norm": 32.50030520028076, + "learning_rate": 1.0772727272727271e-06, + "logits/chosen": -0.00185394287109375, + "logits/rejected": -0.1650390625, + "logps/chosen": -394.0, + "logps/rejected": -326.0, + "loss": 0.3089, + "rewards/accuracies": 0.8515625, + "rewards/chosen": -0.0947265625, + "rewards/margins": 7.28125, + "rewards/rejected": -7.375, + "step": 707 + }, + { + "epoch": 0.8687116564417178, + "grad_norm": 27.35860386441147, + "learning_rate": 1.0766528925619835e-06, + "logits/chosen": 0.0245361328125, + "logits/rejected": -0.193359375, + "logps/chosen": -380.0, + "logps/rejected": -328.0, + "loss": 0.2856, + "rewards/accuracies": 0.8359375, + "rewards/chosen": 0.65234375, + "rewards/margins": 7.28125, + "rewards/rejected": -6.625, + "step": 708 + }, + { + "epoch": 0.8699386503067484, + "grad_norm": 22.967542887627133, + "learning_rate": 1.0760330578512396e-06, + "logits/chosen": 0.11376953125, + "logits/rejected": -0.031982421875, + "logps/chosen": -444.0, + "logps/rejected": -360.0, + "loss": 0.2311, + "rewards/accuracies": 0.875, + "rewards/chosen": 0.984375, + "rewards/margins": 8.4375, + "rewards/rejected": -7.46875, + "step": 709 + }, + { + "epoch": 0.8711656441717791, + "grad_norm": 30.012370581713046, + "learning_rate": 1.075413223140496e-06, + "logits/chosen": 0.1845703125, + "logits/rejected": 0.06103515625, + "logps/chosen": -372.0, + "logps/rejected": -330.0, + "loss": 0.3445, + "rewards/accuracies": 0.8125, + "rewards/chosen": 0.1064453125, + "rewards/margins": 6.46875, + "rewards/rejected": -6.375, + "step": 710 + }, + { + "epoch": 0.8723926380368098, + "grad_norm": 26.59602404250697, + "learning_rate": 1.0747933884297521e-06, + "logits/chosen": 0.15625, + "logits/rejected": -0.039306640625, + "logps/chosen": -390.0, + "logps/rejected": -332.0, + "loss": 0.3562, + "rewards/accuracies": 0.8515625, + "rewards/chosen": 0.5703125, + "rewards/margins": 7.0, + "rewards/rejected": -6.4375, + "step": 711 + }, + { + "epoch": 0.8736196319018404, + "grad_norm": 34.082753455253695, + "learning_rate": 1.0741735537190083e-06, + "logits/chosen": 0.1484375, + "logits/rejected": 0.00482177734375, + "logps/chosen": -410.0, + "logps/rejected": -356.0, + "loss": 0.4269, + "rewards/accuracies": 0.8125, + "rewards/chosen": 1.328125, + "rewards/margins": 7.03125, + "rewards/rejected": -5.71875, + "step": 712 + }, + { + "epoch": 0.8748466257668711, + "grad_norm": 24.8779582799026, + "learning_rate": 1.0735537190082646e-06, + "logits/chosen": 0.05810546875, + "logits/rejected": -0.1484375, + "logps/chosen": -386.0, + "logps/rejected": -336.0, + "loss": 0.2908, + "rewards/accuracies": 0.875, + "rewards/chosen": 0.625, + "rewards/margins": 7.28125, + "rewards/rejected": -6.65625, + "step": 713 + }, + { + "epoch": 0.8760736196319019, + "grad_norm": 27.823114674287325, + "learning_rate": 1.0729338842975208e-06, + "logits/chosen": 0.181640625, + "logits/rejected": -0.038818359375, + "logps/chosen": -370.0, + "logps/rejected": -328.0, + "loss": 0.3057, + "rewards/accuracies": 0.890625, + "rewards/chosen": 0.87109375, + "rewards/margins": 6.84375, + "rewards/rejected": -6.0, + "step": 714 + }, + { + "epoch": 0.8773006134969326, + "grad_norm": 29.027125190845624, + "learning_rate": 1.072314049586777e-06, + "logits/chosen": -0.06494140625, + "logits/rejected": -0.1318359375, + "logps/chosen": -384.0, + "logps/rejected": -328.0, + "loss": 0.3391, + "rewards/accuracies": 0.8203125, + "rewards/chosen": 1.015625, + "rewards/margins": 7.75, + "rewards/rejected": -6.71875, + "step": 715 + }, + { + "epoch": 0.8785276073619632, + "grad_norm": 26.494530473617345, + "learning_rate": 1.071694214876033e-06, + "logits/chosen": -0.018798828125, + "logits/rejected": -0.1865234375, + "logps/chosen": -384.0, + "logps/rejected": -324.0, + "loss": 0.3073, + "rewards/accuracies": 0.8515625, + "rewards/chosen": 1.1640625, + "rewards/margins": 6.84375, + "rewards/rejected": -5.65625, + "step": 716 + }, + { + "epoch": 0.8797546012269939, + "grad_norm": 29.02718718589355, + "learning_rate": 1.0710743801652892e-06, + "logits/chosen": 0.07470703125, + "logits/rejected": -0.08056640625, + "logps/chosen": -344.0, + "logps/rejected": -284.0, + "loss": 0.3344, + "rewards/accuracies": 0.84375, + "rewards/chosen": 0.6875, + "rewards/margins": 6.90625, + "rewards/rejected": -6.21875, + "step": 717 + }, + { + "epoch": 0.8809815950920246, + "grad_norm": 23.193149611746794, + "learning_rate": 1.0704545454545455e-06, + "logits/chosen": -0.034912109375, + "logits/rejected": -0.1640625, + "logps/chosen": -402.0, + "logps/rejected": -356.0, + "loss": 0.2308, + "rewards/accuracies": 0.8828125, + "rewards/chosen": 1.5859375, + "rewards/margins": 7.84375, + "rewards/rejected": -6.25, + "step": 718 + }, + { + "epoch": 0.8822085889570552, + "grad_norm": 36.13316553444024, + "learning_rate": 1.0698347107438017e-06, + "logits/chosen": -0.07763671875, + "logits/rejected": -0.19921875, + "logps/chosen": -410.0, + "logps/rejected": -368.0, + "loss": 0.3568, + "rewards/accuracies": 0.8203125, + "rewards/chosen": 1.8046875, + "rewards/margins": 8.0, + "rewards/rejected": -6.1875, + "step": 719 + }, + { + "epoch": 0.8834355828220859, + "grad_norm": 25.470910809931194, + "learning_rate": 1.0692148760330578e-06, + "logits/chosen": -0.07421875, + "logits/rejected": -0.279296875, + "logps/chosen": -412.0, + "logps/rejected": -350.0, + "loss": 0.2706, + "rewards/accuracies": 0.8828125, + "rewards/chosen": 1.9140625, + "rewards/margins": 7.9375, + "rewards/rejected": -6.03125, + "step": 720 + }, + { + "epoch": 0.8846625766871166, + "grad_norm": 24.45362594207483, + "learning_rate": 1.0685950413223142e-06, + "logits/chosen": -0.0810546875, + "logits/rejected": -0.337890625, + "logps/chosen": -414.0, + "logps/rejected": -364.0, + "loss": 0.2805, + "rewards/accuracies": 0.921875, + "rewards/chosen": 1.5625, + "rewards/margins": 8.8125, + "rewards/rejected": -7.25, + "step": 721 + }, + { + "epoch": 0.8858895705521472, + "grad_norm": 26.653798150160988, + "learning_rate": 1.0679752066115703e-06, + "logits/chosen": -0.049072265625, + "logits/rejected": -0.1376953125, + "logps/chosen": -380.0, + "logps/rejected": -358.0, + "loss": 0.2495, + "rewards/accuracies": 0.890625, + "rewards/chosen": 1.609375, + "rewards/margins": 7.875, + "rewards/rejected": -6.28125, + "step": 722 + }, + { + "epoch": 0.8871165644171779, + "grad_norm": 29.259615066962372, + "learning_rate": 1.0673553719008265e-06, + "logits/chosen": -0.043212890625, + "logits/rejected": -0.14453125, + "logps/chosen": -376.0, + "logps/rejected": -334.0, + "loss": 0.3287, + "rewards/accuracies": 0.859375, + "rewards/chosen": 0.94921875, + "rewards/margins": 7.4375, + "rewards/rejected": -6.5, + "step": 723 + }, + { + "epoch": 0.8883435582822086, + "grad_norm": 30.45569089414315, + "learning_rate": 1.0667355371900826e-06, + "logits/chosen": -0.0308837890625, + "logits/rejected": -0.251953125, + "logps/chosen": -382.0, + "logps/rejected": -320.0, + "loss": 0.3219, + "rewards/accuracies": 0.859375, + "rewards/chosen": 1.3359375, + "rewards/margins": 7.8125, + "rewards/rejected": -6.5, + "step": 724 + }, + { + "epoch": 0.8895705521472392, + "grad_norm": 30.8228410953136, + "learning_rate": 1.0661157024793387e-06, + "logits/chosen": 0.12255859375, + "logits/rejected": -0.05419921875, + "logps/chosen": -404.0, + "logps/rejected": -340.0, + "loss": 0.3833, + "rewards/accuracies": 0.8359375, + "rewards/chosen": 0.984375, + "rewards/margins": 7.375, + "rewards/rejected": -6.375, + "step": 725 + }, + { + "epoch": 0.8907975460122699, + "grad_norm": 32.44931458712784, + "learning_rate": 1.065495867768595e-06, + "logits/chosen": -0.030029296875, + "logits/rejected": -0.154296875, + "logps/chosen": -340.0, + "logps/rejected": -302.0, + "loss": 0.2852, + "rewards/accuracies": 0.8515625, + "rewards/chosen": 0.5703125, + "rewards/margins": 7.53125, + "rewards/rejected": -6.96875, + "step": 726 + }, + { + "epoch": 0.8920245398773006, + "grad_norm": 30.00382649809932, + "learning_rate": 1.0648760330578512e-06, + "logits/chosen": -0.0013427734375, + "logits/rejected": -0.1474609375, + "logps/chosen": -374.0, + "logps/rejected": -338.0, + "loss": 0.2567, + "rewards/accuracies": 0.890625, + "rewards/chosen": 0.84765625, + "rewards/margins": 7.5625, + "rewards/rejected": -6.71875, + "step": 727 + }, + { + "epoch": 0.8932515337423312, + "grad_norm": 32.084882438300085, + "learning_rate": 1.0642561983471074e-06, + "logits/chosen": -0.0244140625, + "logits/rejected": -0.1611328125, + "logps/chosen": -406.0, + "logps/rejected": -384.0, + "loss": 0.3182, + "rewards/accuracies": 0.8515625, + "rewards/chosen": 1.234375, + "rewards/margins": 8.875, + "rewards/rejected": -7.625, + "step": 728 + }, + { + "epoch": 0.894478527607362, + "grad_norm": 24.427532348622456, + "learning_rate": 1.0636363636363637e-06, + "logits/chosen": -0.040771484375, + "logits/rejected": -0.1708984375, + "logps/chosen": -396.0, + "logps/rejected": -372.0, + "loss": 0.2099, + "rewards/accuracies": 0.8984375, + "rewards/chosen": 0.9765625, + "rewards/margins": 8.875, + "rewards/rejected": -7.90625, + "step": 729 + }, + { + "epoch": 0.8957055214723927, + "grad_norm": 24.535854490724105, + "learning_rate": 1.0630165289256199e-06, + "logits/chosen": -0.103515625, + "logits/rejected": -0.2060546875, + "logps/chosen": -404.0, + "logps/rejected": -380.0, + "loss": 0.2127, + "rewards/accuracies": 0.9140625, + "rewards/chosen": 1.515625, + "rewards/margins": 8.8125, + "rewards/rejected": -7.3125, + "step": 730 + }, + { + "epoch": 0.8969325153374234, + "grad_norm": 34.14852950381922, + "learning_rate": 1.062396694214876e-06, + "logits/chosen": 0.1171875, + "logits/rejected": 0.002349853515625, + "logps/chosen": -438.0, + "logps/rejected": -366.0, + "loss": 0.3465, + "rewards/accuracies": 0.859375, + "rewards/chosen": 0.62890625, + "rewards/margins": 8.1875, + "rewards/rejected": -7.5625, + "step": 731 + }, + { + "epoch": 0.898159509202454, + "grad_norm": 34.434393814707754, + "learning_rate": 1.0617768595041324e-06, + "logits/chosen": -0.058837890625, + "logits/rejected": -0.2099609375, + "logps/chosen": -380.0, + "logps/rejected": -372.0, + "loss": 0.4237, + "rewards/accuracies": 0.8046875, + "rewards/chosen": 0.6875, + "rewards/margins": 7.71875, + "rewards/rejected": -7.03125, + "step": 732 + }, + { + "epoch": 0.8993865030674847, + "grad_norm": 28.68332287996042, + "learning_rate": 1.0611570247933883e-06, + "logits/chosen": 0.0673828125, + "logits/rejected": -0.083984375, + "logps/chosen": -366.0, + "logps/rejected": -362.0, + "loss": 0.2815, + "rewards/accuracies": 0.859375, + "rewards/chosen": 0.1015625, + "rewards/margins": 7.65625, + "rewards/rejected": -7.5625, + "step": 733 + }, + { + "epoch": 0.9006134969325154, + "grad_norm": 32.670441266137765, + "learning_rate": 1.0605371900826447e-06, + "logits/chosen": 0.06884765625, + "logits/rejected": -0.1044921875, + "logps/chosen": -384.0, + "logps/rejected": -354.0, + "loss": 0.3319, + "rewards/accuracies": 0.859375, + "rewards/chosen": 0.408203125, + "rewards/margins": 8.3125, + "rewards/rejected": -7.90625, + "step": 734 + }, + { + "epoch": 0.901840490797546, + "grad_norm": 28.766622871975752, + "learning_rate": 1.0599173553719008e-06, + "logits/chosen": -0.037841796875, + "logits/rejected": -0.259765625, + "logps/chosen": -434.0, + "logps/rejected": -390.0, + "loss": 0.2447, + "rewards/accuracies": 0.875, + "rewards/chosen": 0.2490234375, + "rewards/margins": 8.875, + "rewards/rejected": -8.625, + "step": 735 + }, + { + "epoch": 0.9030674846625767, + "grad_norm": 31.10013526298491, + "learning_rate": 1.059297520661157e-06, + "logits/chosen": 0.1533203125, + "logits/rejected": 0.029541015625, + "logps/chosen": -362.0, + "logps/rejected": -340.0, + "loss": 0.4211, + "rewards/accuracies": 0.8359375, + "rewards/chosen": -0.53515625, + "rewards/margins": 7.03125, + "rewards/rejected": -7.5625, + "step": 736 + }, + { + "epoch": 0.9042944785276074, + "grad_norm": 38.03816635846729, + "learning_rate": 1.0586776859504133e-06, + "logits/chosen": 0.04541015625, + "logits/rejected": -0.051513671875, + "logps/chosen": -360.0, + "logps/rejected": -348.0, + "loss": 0.4641, + "rewards/accuracies": 0.8046875, + "rewards/chosen": -0.478515625, + "rewards/margins": 7.09375, + "rewards/rejected": -7.5625, + "step": 737 + }, + { + "epoch": 0.905521472392638, + "grad_norm": 32.716309617319716, + "learning_rate": 1.0580578512396694e-06, + "logits/chosen": -0.0380859375, + "logits/rejected": -0.1474609375, + "logps/chosen": -450.0, + "logps/rejected": -364.0, + "loss": 0.3437, + "rewards/accuracies": 0.8359375, + "rewards/chosen": 0.60546875, + "rewards/margins": 8.1875, + "rewards/rejected": -7.59375, + "step": 738 + }, + { + "epoch": 0.9067484662576687, + "grad_norm": 29.627927045089628, + "learning_rate": 1.0574380165289258e-06, + "logits/chosen": 0.060302734375, + "logits/rejected": -0.0654296875, + "logps/chosen": -346.0, + "logps/rejected": -340.0, + "loss": 0.2827, + "rewards/accuracies": 0.84375, + "rewards/chosen": -0.06640625, + "rewards/margins": 7.1875, + "rewards/rejected": -7.25, + "step": 739 + }, + { + "epoch": 0.9079754601226994, + "grad_norm": 29.046321525927297, + "learning_rate": 1.056818181818182e-06, + "logits/chosen": -0.0308837890625, + "logits/rejected": -0.18359375, + "logps/chosen": -344.0, + "logps/rejected": -326.0, + "loss": 0.3577, + "rewards/accuracies": 0.8515625, + "rewards/chosen": 0.02392578125, + "rewards/margins": 7.5625, + "rewards/rejected": -7.53125, + "step": 740 + }, + { + "epoch": 0.90920245398773, + "grad_norm": 29.826427625052336, + "learning_rate": 1.056198347107438e-06, + "logits/chosen": -0.00982666015625, + "logits/rejected": -0.162109375, + "logps/chosen": -356.0, + "logps/rejected": -328.0, + "loss": 0.3514, + "rewards/accuracies": 0.84375, + "rewards/chosen": 0.8359375, + "rewards/margins": 8.75, + "rewards/rejected": -7.875, + "step": 741 + }, + { + "epoch": 0.9104294478527607, + "grad_norm": 28.054610141633944, + "learning_rate": 1.0555785123966942e-06, + "logits/chosen": -0.033935546875, + "logits/rejected": -0.1982421875, + "logps/chosen": -368.0, + "logps/rejected": -372.0, + "loss": 0.2813, + "rewards/accuracies": 0.8828125, + "rewards/chosen": 0.6640625, + "rewards/margins": 8.5625, + "rewards/rejected": -7.90625, + "step": 742 + }, + { + "epoch": 0.9116564417177914, + "grad_norm": 35.487753263657986, + "learning_rate": 1.0549586776859503e-06, + "logits/chosen": 0.01055908203125, + "logits/rejected": -0.040771484375, + "logps/chosen": -424.0, + "logps/rejected": -392.0, + "loss": 0.3821, + "rewards/accuracies": 0.8203125, + "rewards/chosen": 0.2265625, + "rewards/margins": 7.21875, + "rewards/rejected": -7.0, + "step": 743 + }, + { + "epoch": 0.912883435582822, + "grad_norm": 28.496366117513464, + "learning_rate": 1.0543388429752067e-06, + "logits/chosen": 0.035400390625, + "logits/rejected": -0.14453125, + "logps/chosen": -380.0, + "logps/rejected": -340.0, + "loss": 0.2775, + "rewards/accuracies": 0.859375, + "rewards/chosen": 0.9609375, + "rewards/margins": 8.6875, + "rewards/rejected": -7.71875, + "step": 744 + }, + { + "epoch": 0.9141104294478528, + "grad_norm": 30.005979043165016, + "learning_rate": 1.0537190082644628e-06, + "logits/chosen": 0.08642578125, + "logits/rejected": -0.0003509521484375, + "logps/chosen": -346.0, + "logps/rejected": -318.0, + "loss": 0.3381, + "rewards/accuracies": 0.8515625, + "rewards/chosen": 0.1044921875, + "rewards/margins": 6.84375, + "rewards/rejected": -6.75, + "step": 745 + }, + { + "epoch": 0.9153374233128835, + "grad_norm": 36.365949373279655, + "learning_rate": 1.053099173553719e-06, + "logits/chosen": 0.09033203125, + "logits/rejected": -0.07958984375, + "logps/chosen": -408.0, + "logps/rejected": -356.0, + "loss": 0.4283, + "rewards/accuracies": 0.84375, + "rewards/chosen": 0.4375, + "rewards/margins": 7.4375, + "rewards/rejected": -7.0, + "step": 746 + }, + { + "epoch": 0.9165644171779141, + "grad_norm": 30.46892734256714, + "learning_rate": 1.0524793388429753e-06, + "logits/chosen": -0.1123046875, + "logits/rejected": -0.21875, + "logps/chosen": -342.0, + "logps/rejected": -344.0, + "loss": 0.2905, + "rewards/accuracies": 0.8515625, + "rewards/chosen": -0.0234375, + "rewards/margins": 6.75, + "rewards/rejected": -6.75, + "step": 747 + }, + { + "epoch": 0.9177914110429448, + "grad_norm": 37.226894144134455, + "learning_rate": 1.0518595041322315e-06, + "logits/chosen": 0.125, + "logits/rejected": 0.01458740234375, + "logps/chosen": -380.0, + "logps/rejected": -354.0, + "loss": 0.4992, + "rewards/accuracies": 0.78125, + "rewards/chosen": -0.50390625, + "rewards/margins": 6.28125, + "rewards/rejected": -6.78125, + "step": 748 + }, + { + "epoch": 0.9190184049079755, + "grad_norm": 34.24194812875213, + "learning_rate": 1.0512396694214876e-06, + "logits/chosen": -0.04541015625, + "logits/rejected": -0.2431640625, + "logps/chosen": -424.0, + "logps/rejected": -358.0, + "loss": 0.4049, + "rewards/accuracies": 0.8125, + "rewards/chosen": 0.2470703125, + "rewards/margins": 7.59375, + "rewards/rejected": -7.34375, + "step": 749 + }, + { + "epoch": 0.9202453987730062, + "grad_norm": 35.848596312284045, + "learning_rate": 1.050619834710744e-06, + "logits/chosen": 0.01190185546875, + "logits/rejected": -0.119140625, + "logps/chosen": -412.0, + "logps/rejected": -348.0, + "loss": 0.4472, + "rewards/accuracies": 0.828125, + "rewards/chosen": 0.09326171875, + "rewards/margins": 7.375, + "rewards/rejected": -7.28125, + "step": 750 + }, + { + "epoch": 0.9214723926380368, + "grad_norm": 36.49684940562663, + "learning_rate": 1.05e-06, + "logits/chosen": -0.0064697265625, + "logits/rejected": -0.169921875, + "logps/chosen": -428.0, + "logps/rejected": -366.0, + "loss": 0.3626, + "rewards/accuracies": 0.84375, + "rewards/chosen": 0.5234375, + "rewards/margins": 8.3125, + "rewards/rejected": -7.8125, + "step": 751 + }, + { + "epoch": 0.9226993865030675, + "grad_norm": 31.9902592360704, + "learning_rate": 1.0493801652892563e-06, + "logits/chosen": -0.028564453125, + "logits/rejected": -0.275390625, + "logps/chosen": -424.0, + "logps/rejected": -368.0, + "loss": 0.3427, + "rewards/accuracies": 0.8515625, + "rewards/chosen": 0.640625, + "rewards/margins": 8.375, + "rewards/rejected": -7.71875, + "step": 752 + }, + { + "epoch": 0.9239263803680982, + "grad_norm": 30.210419690529548, + "learning_rate": 1.0487603305785124e-06, + "logits/chosen": 0.142578125, + "logits/rejected": 0.01361083984375, + "logps/chosen": -408.0, + "logps/rejected": -362.0, + "loss": 0.3032, + "rewards/accuracies": 0.859375, + "rewards/chosen": 0.126953125, + "rewards/margins": 7.375, + "rewards/rejected": -7.25, + "step": 753 + }, + { + "epoch": 0.9251533742331288, + "grad_norm": 33.70924536763799, + "learning_rate": 1.0481404958677685e-06, + "logits/chosen": 0.06640625, + "logits/rejected": -0.1044921875, + "logps/chosen": -394.0, + "logps/rejected": -348.0, + "loss": 0.4405, + "rewards/accuracies": 0.796875, + "rewards/chosen": -0.1416015625, + "rewards/margins": 7.09375, + "rewards/rejected": -7.21875, + "step": 754 + }, + { + "epoch": 0.9263803680981595, + "grad_norm": 28.923057351049156, + "learning_rate": 1.0475206611570249e-06, + "logits/chosen": 0.01611328125, + "logits/rejected": -0.1064453125, + "logps/chosen": -388.0, + "logps/rejected": -366.0, + "loss": 0.313, + "rewards/accuracies": 0.875, + "rewards/chosen": -0.291015625, + "rewards/margins": 7.21875, + "rewards/rejected": -7.5, + "step": 755 + }, + { + "epoch": 0.9276073619631902, + "grad_norm": 26.956563960777117, + "learning_rate": 1.046900826446281e-06, + "logits/chosen": 0.115234375, + "logits/rejected": 0.02978515625, + "logps/chosen": -342.0, + "logps/rejected": -326.0, + "loss": 0.2891, + "rewards/accuracies": 0.8828125, + "rewards/chosen": -0.640625, + "rewards/margins": 6.8125, + "rewards/rejected": -7.46875, + "step": 756 + }, + { + "epoch": 0.9288343558282208, + "grad_norm": 25.54044780724446, + "learning_rate": 1.0462809917355372e-06, + "logits/chosen": 0.15625, + "logits/rejected": -0.0595703125, + "logps/chosen": -404.0, + "logps/rejected": -344.0, + "loss": 0.27, + "rewards/accuracies": 0.875, + "rewards/chosen": 0.7578125, + "rewards/margins": 8.0625, + "rewards/rejected": -7.3125, + "step": 757 + }, + { + "epoch": 0.9300613496932515, + "grad_norm": 35.658646153584634, + "learning_rate": 1.0456611570247935e-06, + "logits/chosen": 0.08740234375, + "logits/rejected": -0.080078125, + "logps/chosen": -376.0, + "logps/rejected": -320.0, + "loss": 0.3867, + "rewards/accuracies": 0.8125, + "rewards/chosen": -0.55859375, + "rewards/margins": 6.71875, + "rewards/rejected": -7.28125, + "step": 758 + }, + { + "epoch": 0.9312883435582822, + "grad_norm": 29.003566311314795, + "learning_rate": 1.0450413223140497e-06, + "logits/chosen": 0.1123046875, + "logits/rejected": -0.11572265625, + "logps/chosen": -396.0, + "logps/rejected": -340.0, + "loss": 0.2888, + "rewards/accuracies": 0.84375, + "rewards/chosen": 0.16015625, + "rewards/margins": 6.9375, + "rewards/rejected": -6.78125, + "step": 759 + }, + { + "epoch": 0.9325153374233128, + "grad_norm": 25.36330419002366, + "learning_rate": 1.0444214876033058e-06, + "logits/chosen": -0.0986328125, + "logits/rejected": -0.35546875, + "logps/chosen": -442.0, + "logps/rejected": -366.0, + "loss": 0.2887, + "rewards/accuracies": 0.875, + "rewards/chosen": 1.1796875, + "rewards/margins": 8.625, + "rewards/rejected": -7.4375, + "step": 760 + }, + { + "epoch": 0.9337423312883436, + "grad_norm": 31.72216949313869, + "learning_rate": 1.043801652892562e-06, + "logits/chosen": 0.1484375, + "logits/rejected": 0.0277099609375, + "logps/chosen": -382.0, + "logps/rejected": -354.0, + "loss": 0.3707, + "rewards/accuracies": 0.8671875, + "rewards/chosen": 0.3984375, + "rewards/margins": 7.21875, + "rewards/rejected": -6.8125, + "step": 761 + }, + { + "epoch": 0.9349693251533743, + "grad_norm": 31.01314079949462, + "learning_rate": 1.043181818181818e-06, + "logits/chosen": 0.1591796875, + "logits/rejected": -0.020751953125, + "logps/chosen": -428.0, + "logps/rejected": -364.0, + "loss": 0.3149, + "rewards/accuracies": 0.8984375, + "rewards/chosen": 0.2373046875, + "rewards/margins": 7.53125, + "rewards/rejected": -7.28125, + "step": 762 + }, + { + "epoch": 0.9361963190184049, + "grad_norm": 28.119069224585317, + "learning_rate": 1.0425619834710744e-06, + "logits/chosen": 0.0986328125, + "logits/rejected": 0.0164794921875, + "logps/chosen": -400.0, + "logps/rejected": -364.0, + "loss": 0.3398, + "rewards/accuracies": 0.84375, + "rewards/chosen": 0.4375, + "rewards/margins": 7.5625, + "rewards/rejected": -7.125, + "step": 763 + }, + { + "epoch": 0.9374233128834356, + "grad_norm": 26.42805050753053, + "learning_rate": 1.0419421487603306e-06, + "logits/chosen": -0.1328125, + "logits/rejected": -0.287109375, + "logps/chosen": -330.0, + "logps/rejected": -316.0, + "loss": 0.2955, + "rewards/accuracies": 0.90625, + "rewards/chosen": -0.494140625, + "rewards/margins": 7.0, + "rewards/rejected": -7.46875, + "step": 764 + }, + { + "epoch": 0.9386503067484663, + "grad_norm": 35.659703021872964, + "learning_rate": 1.041322314049587e-06, + "logits/chosen": 0.1298828125, + "logits/rejected": 0.005828857421875, + "logps/chosen": -404.0, + "logps/rejected": -320.0, + "loss": 0.3833, + "rewards/accuracies": 0.796875, + "rewards/chosen": -0.2041015625, + "rewards/margins": 6.125, + "rewards/rejected": -6.34375, + "step": 765 + }, + { + "epoch": 0.939877300613497, + "grad_norm": 27.867245312963597, + "learning_rate": 1.040702479338843e-06, + "logits/chosen": 0.04345703125, + "logits/rejected": -0.1494140625, + "logps/chosen": -446.0, + "logps/rejected": -340.0, + "loss": 0.3188, + "rewards/accuracies": 0.8671875, + "rewards/chosen": 0.515625, + "rewards/margins": 7.5, + "rewards/rejected": -7.0, + "step": 766 + }, + { + "epoch": 0.9411042944785276, + "grad_norm": 33.0423901915957, + "learning_rate": 1.0400826446280992e-06, + "logits/chosen": 0.0390625, + "logits/rejected": -0.10498046875, + "logps/chosen": -386.0, + "logps/rejected": -338.0, + "loss": 0.355, + "rewards/accuracies": 0.796875, + "rewards/chosen": 0.1689453125, + "rewards/margins": 7.5625, + "rewards/rejected": -7.375, + "step": 767 + }, + { + "epoch": 0.9423312883435583, + "grad_norm": 31.83881144027911, + "learning_rate": 1.0394628099173554e-06, + "logits/chosen": 0.003570556640625, + "logits/rejected": -0.1611328125, + "logps/chosen": -410.0, + "logps/rejected": -384.0, + "loss": 0.3826, + "rewards/accuracies": 0.875, + "rewards/chosen": 0.6484375, + "rewards/margins": 7.59375, + "rewards/rejected": -6.96875, + "step": 768 + }, + { + "epoch": 0.943558282208589, + "grad_norm": 24.722102647279602, + "learning_rate": 1.0388429752066115e-06, + "logits/chosen": 0.00079345703125, + "logits/rejected": -0.138671875, + "logps/chosen": -378.0, + "logps/rejected": -344.0, + "loss": 0.2777, + "rewards/accuracies": 0.859375, + "rewards/chosen": -0.068359375, + "rewards/margins": 7.53125, + "rewards/rejected": -7.59375, + "step": 769 + }, + { + "epoch": 0.9447852760736196, + "grad_norm": 33.3608331180557, + "learning_rate": 1.0382231404958679e-06, + "logits/chosen": -0.021728515625, + "logits/rejected": -0.265625, + "logps/chosen": -416.0, + "logps/rejected": -372.0, + "loss": 0.321, + "rewards/accuracies": 0.8515625, + "rewards/chosen": 0.92578125, + "rewards/margins": 7.5, + "rewards/rejected": -6.59375, + "step": 770 + }, + { + "epoch": 0.9460122699386503, + "grad_norm": 36.53871156772584, + "learning_rate": 1.037603305785124e-06, + "logits/chosen": -0.00537109375, + "logits/rejected": -0.01507568359375, + "logps/chosen": -430.0, + "logps/rejected": -360.0, + "loss": 0.5217, + "rewards/accuracies": 0.78125, + "rewards/chosen": -0.6484375, + "rewards/margins": 6.21875, + "rewards/rejected": -6.875, + "step": 771 + }, + { + "epoch": 0.947239263803681, + "grad_norm": 30.200058873190507, + "learning_rate": 1.0369834710743801e-06, + "logits/chosen": 0.1328125, + "logits/rejected": -0.0869140625, + "logps/chosen": -424.0, + "logps/rejected": -362.0, + "loss": 0.3536, + "rewards/accuracies": 0.84375, + "rewards/chosen": 0.38671875, + "rewards/margins": 7.28125, + "rewards/rejected": -6.90625, + "step": 772 + }, + { + "epoch": 0.9484662576687116, + "grad_norm": 31.95788929467944, + "learning_rate": 1.0363636363636365e-06, + "logits/chosen": 0.1484375, + "logits/rejected": -0.037841796875, + "logps/chosen": -374.0, + "logps/rejected": -324.0, + "loss": 0.3406, + "rewards/accuracies": 0.828125, + "rewards/chosen": -0.65234375, + "rewards/margins": 6.5, + "rewards/rejected": -7.15625, + "step": 773 + }, + { + "epoch": 0.9496932515337423, + "grad_norm": 34.11224513726983, + "learning_rate": 1.0357438016528926e-06, + "logits/chosen": 0.1396484375, + "logits/rejected": -0.0037078857421875, + "logps/chosen": -362.0, + "logps/rejected": -326.0, + "loss": 0.4295, + "rewards/accuracies": 0.7890625, + "rewards/chosen": -1.359375, + "rewards/margins": 5.1875, + "rewards/rejected": -6.5625, + "step": 774 + }, + { + "epoch": 0.950920245398773, + "grad_norm": 31.83399815065193, + "learning_rate": 1.0351239669421488e-06, + "logits/chosen": 0.2197265625, + "logits/rejected": 0.04296875, + "logps/chosen": -418.0, + "logps/rejected": -362.0, + "loss": 0.3519, + "rewards/accuracies": 0.84375, + "rewards/chosen": -0.32421875, + "rewards/margins": 6.21875, + "rewards/rejected": -6.53125, + "step": 775 + }, + { + "epoch": 0.9521472392638037, + "grad_norm": 29.278356141500073, + "learning_rate": 1.0345041322314051e-06, + "logits/chosen": -0.01409912109375, + "logits/rejected": -0.1708984375, + "logps/chosen": -414.0, + "logps/rejected": -366.0, + "loss": 0.3978, + "rewards/accuracies": 0.84375, + "rewards/chosen": 0.0888671875, + "rewards/margins": 6.8125, + "rewards/rejected": -6.75, + "step": 776 + }, + { + "epoch": 0.9533742331288344, + "grad_norm": 33.3868265617977, + "learning_rate": 1.033884297520661e-06, + "logits/chosen": 0.0267333984375, + "logits/rejected": -0.1015625, + "logps/chosen": -408.0, + "logps/rejected": -346.0, + "loss": 0.3944, + "rewards/accuracies": 0.8046875, + "rewards/chosen": -0.58203125, + "rewards/margins": 6.71875, + "rewards/rejected": -7.3125, + "step": 777 + }, + { + "epoch": 0.9546012269938651, + "grad_norm": 30.60810394473529, + "learning_rate": 1.0332644628099174e-06, + "logits/chosen": 0.0703125, + "logits/rejected": -0.130859375, + "logps/chosen": -400.0, + "logps/rejected": -328.0, + "loss": 0.2893, + "rewards/accuracies": 0.8671875, + "rewards/chosen": -0.74609375, + "rewards/margins": 6.0, + "rewards/rejected": -6.75, + "step": 778 + }, + { + "epoch": 0.9558282208588957, + "grad_norm": 31.819603506171298, + "learning_rate": 1.0326446280991735e-06, + "logits/chosen": 0.00994873046875, + "logits/rejected": -0.1533203125, + "logps/chosen": -442.0, + "logps/rejected": -362.0, + "loss": 0.331, + "rewards/accuracies": 0.8671875, + "rewards/chosen": 0.16796875, + "rewards/margins": 7.59375, + "rewards/rejected": -7.40625, + "step": 779 + }, + { + "epoch": 0.9570552147239264, + "grad_norm": 33.50608635895235, + "learning_rate": 1.0320247933884297e-06, + "logits/chosen": 0.13671875, + "logits/rejected": -0.037353515625, + "logps/chosen": -386.0, + "logps/rejected": -360.0, + "loss": 0.4421, + "rewards/accuracies": 0.7890625, + "rewards/chosen": -0.91015625, + "rewards/margins": 5.96875, + "rewards/rejected": -6.875, + "step": 780 + }, + { + "epoch": 0.9582822085889571, + "grad_norm": 27.175413708110298, + "learning_rate": 1.031404958677686e-06, + "logits/chosen": -0.0087890625, + "logits/rejected": -0.2392578125, + "logps/chosen": -426.0, + "logps/rejected": -356.0, + "loss": 0.2289, + "rewards/accuracies": 0.8828125, + "rewards/chosen": 0.9140625, + "rewards/margins": 8.75, + "rewards/rejected": -7.8125, + "step": 781 + }, + { + "epoch": 0.9595092024539877, + "grad_norm": 27.961517528382746, + "learning_rate": 1.0307851239669422e-06, + "logits/chosen": 0.0341796875, + "logits/rejected": -0.055419921875, + "logps/chosen": -370.0, + "logps/rejected": -354.0, + "loss": 0.2727, + "rewards/accuracies": 0.890625, + "rewards/chosen": 0.0152587890625, + "rewards/margins": 7.21875, + "rewards/rejected": -7.1875, + "step": 782 + }, + { + "epoch": 0.9607361963190184, + "grad_norm": 36.53707624363368, + "learning_rate": 1.0301652892561983e-06, + "logits/chosen": 0.0115966796875, + "logits/rejected": -0.12353515625, + "logps/chosen": -420.0, + "logps/rejected": -370.0, + "loss": 0.3333, + "rewards/accuracies": 0.859375, + "rewards/chosen": 0.408203125, + "rewards/margins": 7.9375, + "rewards/rejected": -7.5, + "step": 783 + }, + { + "epoch": 0.9619631901840491, + "grad_norm": 33.01812506650061, + "learning_rate": 1.0295454545454547e-06, + "logits/chosen": 0.09716796875, + "logits/rejected": 0.056640625, + "logps/chosen": -386.0, + "logps/rejected": -354.0, + "loss": 0.3887, + "rewards/accuracies": 0.828125, + "rewards/chosen": -0.69140625, + "rewards/margins": 6.4375, + "rewards/rejected": -7.125, + "step": 784 + }, + { + "epoch": 0.9631901840490797, + "grad_norm": 25.64548904648115, + "learning_rate": 1.0289256198347108e-06, + "logits/chosen": 0.126953125, + "logits/rejected": -0.0400390625, + "logps/chosen": -384.0, + "logps/rejected": -324.0, + "loss": 0.2131, + "rewards/accuracies": 0.875, + "rewards/chosen": 0.0047607421875, + "rewards/margins": 7.40625, + "rewards/rejected": -7.375, + "step": 785 + }, + { + "epoch": 0.9644171779141104, + "grad_norm": 24.085836304747, + "learning_rate": 1.028305785123967e-06, + "logits/chosen": 0.11376953125, + "logits/rejected": -0.09033203125, + "logps/chosen": -372.0, + "logps/rejected": -334.0, + "loss": 0.2197, + "rewards/accuracies": 0.8984375, + "rewards/chosen": -0.427734375, + "rewards/margins": 7.375, + "rewards/rejected": -7.8125, + "step": 786 + }, + { + "epoch": 0.9656441717791411, + "grad_norm": 32.97798807698007, + "learning_rate": 1.027685950413223e-06, + "logits/chosen": 0.0269775390625, + "logits/rejected": -0.08740234375, + "logps/chosen": -412.0, + "logps/rejected": -348.0, + "loss": 0.3407, + "rewards/accuracies": 0.859375, + "rewards/chosen": -0.271484375, + "rewards/margins": 7.65625, + "rewards/rejected": -7.9375, + "step": 787 + }, + { + "epoch": 0.9668711656441717, + "grad_norm": 24.19044914351049, + "learning_rate": 1.0270661157024792e-06, + "logits/chosen": -0.00860595703125, + "logits/rejected": -0.1787109375, + "logps/chosen": -390.0, + "logps/rejected": -366.0, + "loss": 0.2123, + "rewards/accuracies": 0.9140625, + "rewards/chosen": 0.486328125, + "rewards/margins": 8.1875, + "rewards/rejected": -7.6875, + "step": 788 + }, + { + "epoch": 0.9680981595092024, + "grad_norm": 30.061433994827276, + "learning_rate": 1.0264462809917356e-06, + "logits/chosen": 0.12060546875, + "logits/rejected": -0.06396484375, + "logps/chosen": -384.0, + "logps/rejected": -332.0, + "loss": 0.383, + "rewards/accuracies": 0.859375, + "rewards/chosen": -0.52734375, + "rewards/margins": 6.71875, + "rewards/rejected": -7.25, + "step": 789 + }, + { + "epoch": 0.9693251533742331, + "grad_norm": 27.089223909733843, + "learning_rate": 1.0258264462809917e-06, + "logits/chosen": -0.028564453125, + "logits/rejected": -0.2373046875, + "logps/chosen": -388.0, + "logps/rejected": -330.0, + "loss": 0.1804, + "rewards/accuracies": 0.8984375, + "rewards/chosen": -0.1533203125, + "rewards/margins": 7.3125, + "rewards/rejected": -7.46875, + "step": 790 + }, + { + "epoch": 0.9705521472392638, + "grad_norm": 34.35477577844613, + "learning_rate": 1.0252066115702479e-06, + "logits/chosen": -0.0303955078125, + "logits/rejected": -0.23046875, + "logps/chosen": -398.0, + "logps/rejected": -346.0, + "loss": 0.3642, + "rewards/accuracies": 0.8125, + "rewards/chosen": 0.349609375, + "rewards/margins": 8.25, + "rewards/rejected": -7.90625, + "step": 791 + }, + { + "epoch": 0.9717791411042945, + "grad_norm": 33.05641026053793, + "learning_rate": 1.0245867768595042e-06, + "logits/chosen": 0.0006256103515625, + "logits/rejected": -0.1279296875, + "logps/chosen": -438.0, + "logps/rejected": -374.0, + "loss": 0.406, + "rewards/accuracies": 0.8125, + "rewards/chosen": 0.380859375, + "rewards/margins": 6.8125, + "rewards/rejected": -6.4375, + "step": 792 + }, + { + "epoch": 0.9730061349693252, + "grad_norm": 32.69405943160363, + "learning_rate": 1.0239669421487604e-06, + "logits/chosen": 0.0284423828125, + "logits/rejected": -0.09765625, + "logps/chosen": -392.0, + "logps/rejected": -362.0, + "loss": 0.4018, + "rewards/accuracies": 0.8671875, + "rewards/chosen": -0.10546875, + "rewards/margins": 6.75, + "rewards/rejected": -6.84375, + "step": 793 + }, + { + "epoch": 0.9742331288343559, + "grad_norm": 33.74613547302303, + "learning_rate": 1.0233471074380167e-06, + "logits/chosen": 0.0830078125, + "logits/rejected": -0.043701171875, + "logps/chosen": -358.0, + "logps/rejected": -336.0, + "loss": 0.4877, + "rewards/accuracies": 0.8046875, + "rewards/chosen": -0.6015625, + "rewards/margins": 6.59375, + "rewards/rejected": -7.1875, + "step": 794 + }, + { + "epoch": 0.9754601226993865, + "grad_norm": 22.619679414305867, + "learning_rate": 1.0227272727272727e-06, + "logits/chosen": -0.1513671875, + "logits/rejected": -0.296875, + "logps/chosen": -392.0, + "logps/rejected": -372.0, + "loss": 0.281, + "rewards/accuracies": 0.875, + "rewards/chosen": 0.80859375, + "rewards/margins": 9.25, + "rewards/rejected": -8.4375, + "step": 795 + }, + { + "epoch": 0.9766871165644172, + "grad_norm": 35.85933901249144, + "learning_rate": 1.0221074380165288e-06, + "logits/chosen": -0.0260009765625, + "logits/rejected": -0.2353515625, + "logps/chosen": -410.0, + "logps/rejected": -338.0, + "loss": 0.3447, + "rewards/accuracies": 0.8828125, + "rewards/chosen": -0.78515625, + "rewards/margins": 7.25, + "rewards/rejected": -8.0625, + "step": 796 + }, + { + "epoch": 0.9779141104294479, + "grad_norm": 24.484140263550707, + "learning_rate": 1.0214876033057852e-06, + "logits/chosen": 0.059814453125, + "logits/rejected": -0.11962890625, + "logps/chosen": -404.0, + "logps/rejected": -324.0, + "loss": 0.243, + "rewards/accuracies": 0.8671875, + "rewards/chosen": -0.0966796875, + "rewards/margins": 7.15625, + "rewards/rejected": -7.25, + "step": 797 + }, + { + "epoch": 0.9791411042944785, + "grad_norm": 35.333603529709684, + "learning_rate": 1.0208677685950413e-06, + "logits/chosen": 0.00506591796875, + "logits/rejected": -0.173828125, + "logps/chosen": -430.0, + "logps/rejected": -378.0, + "loss": 0.3546, + "rewards/accuracies": 0.8203125, + "rewards/chosen": 0.1455078125, + "rewards/margins": 7.78125, + "rewards/rejected": -7.625, + "step": 798 + }, + { + "epoch": 0.9803680981595092, + "grad_norm": 32.17706869798208, + "learning_rate": 1.0202479338842976e-06, + "logits/chosen": -0.05029296875, + "logits/rejected": -0.201171875, + "logps/chosen": -400.0, + "logps/rejected": -360.0, + "loss": 0.3323, + "rewards/accuracies": 0.8203125, + "rewards/chosen": 0.4296875, + "rewards/margins": 7.71875, + "rewards/rejected": -7.28125, + "step": 799 + }, + { + "epoch": 0.9815950920245399, + "grad_norm": 32.44337864899576, + "learning_rate": 1.0196280991735538e-06, + "logits/chosen": -0.0224609375, + "logits/rejected": -0.208984375, + "logps/chosen": -386.0, + "logps/rejected": -344.0, + "loss": 0.3141, + "rewards/accuracies": 0.8828125, + "rewards/chosen": -0.0654296875, + "rewards/margins": 7.28125, + "rewards/rejected": -7.375, + "step": 800 + }, + { + "epoch": 0.9828220858895705, + "grad_norm": 37.380385495976945, + "learning_rate": 1.01900826446281e-06, + "logits/chosen": -0.024658203125, + "logits/rejected": -0.17578125, + "logps/chosen": -436.0, + "logps/rejected": -382.0, + "loss": 0.4546, + "rewards/accuracies": 0.828125, + "rewards/chosen": 0.0001010894775390625, + "rewards/margins": 6.6875, + "rewards/rejected": -6.6875, + "step": 801 + }, + { + "epoch": 0.9840490797546012, + "grad_norm": 35.755684948235334, + "learning_rate": 1.0183884297520663e-06, + "logits/chosen": -0.07421875, + "logits/rejected": -0.21484375, + "logps/chosen": -386.0, + "logps/rejected": -348.0, + "loss": 0.472, + "rewards/accuracies": 0.84375, + "rewards/chosen": -0.1552734375, + "rewards/margins": 7.0, + "rewards/rejected": -7.15625, + "step": 802 + }, + { + "epoch": 0.9852760736196319, + "grad_norm": 29.08898198089266, + "learning_rate": 1.0177685950413224e-06, + "logits/chosen": -0.1513671875, + "logits/rejected": -0.302734375, + "logps/chosen": -398.0, + "logps/rejected": -348.0, + "loss": 0.3346, + "rewards/accuracies": 0.84375, + "rewards/chosen": 0.54296875, + "rewards/margins": 7.75, + "rewards/rejected": -7.21875, + "step": 803 + }, + { + "epoch": 0.9865030674846625, + "grad_norm": 30.37947277261302, + "learning_rate": 1.0171487603305786e-06, + "logits/chosen": 0.04150390625, + "logits/rejected": -0.06689453125, + "logps/chosen": -358.0, + "logps/rejected": -346.0, + "loss": 0.3638, + "rewards/accuracies": 0.859375, + "rewards/chosen": -0.392578125, + "rewards/margins": 6.9375, + "rewards/rejected": -7.34375, + "step": 804 + }, + { + "epoch": 0.9877300613496932, + "grad_norm": 34.04488703971709, + "learning_rate": 1.0165289256198347e-06, + "logits/chosen": -0.078125, + "logits/rejected": -0.2119140625, + "logps/chosen": -434.0, + "logps/rejected": -394.0, + "loss": 0.4569, + "rewards/accuracies": 0.828125, + "rewards/chosen": 0.4296875, + "rewards/margins": 7.78125, + "rewards/rejected": -7.375, + "step": 805 + }, + { + "epoch": 0.9889570552147239, + "grad_norm": 35.48502129676359, + "learning_rate": 1.0159090909090908e-06, + "logits/chosen": 0.16015625, + "logits/rejected": 0.03564453125, + "logps/chosen": -412.0, + "logps/rejected": -384.0, + "loss": 0.4253, + "rewards/accuracies": 0.8046875, + "rewards/chosen": 0.31640625, + "rewards/margins": 7.15625, + "rewards/rejected": -6.84375, + "step": 806 + }, + { + "epoch": 0.9901840490797545, + "grad_norm": 33.61353577378301, + "learning_rate": 1.0152892561983472e-06, + "logits/chosen": 0.061767578125, + "logits/rejected": -0.0908203125, + "logps/chosen": -400.0, + "logps/rejected": -350.0, + "loss": 0.3144, + "rewards/accuracies": 0.8828125, + "rewards/chosen": 0.484375, + "rewards/margins": 8.25, + "rewards/rejected": -7.75, + "step": 807 + }, + { + "epoch": 0.9914110429447853, + "grad_norm": 31.091169744856, + "learning_rate": 1.0146694214876033e-06, + "logits/chosen": -0.083984375, + "logits/rejected": -0.0966796875, + "logps/chosen": -364.0, + "logps/rejected": -352.0, + "loss": 0.3745, + "rewards/accuracies": 0.84375, + "rewards/chosen": -0.640625, + "rewards/margins": 7.5625, + "rewards/rejected": -8.1875, + "step": 808 + }, + { + "epoch": 0.992638036809816, + "grad_norm": 34.39338760048118, + "learning_rate": 1.0140495867768595e-06, + "logits/chosen": -0.00872802734375, + "logits/rejected": -0.2314453125, + "logps/chosen": -438.0, + "logps/rejected": -360.0, + "loss": 0.319, + "rewards/accuracies": 0.859375, + "rewards/chosen": 0.8515625, + "rewards/margins": 8.5625, + "rewards/rejected": -7.71875, + "step": 809 + }, + { + "epoch": 0.9938650306748467, + "grad_norm": 26.950235054794184, + "learning_rate": 1.0134297520661158e-06, + "logits/chosen": -0.11572265625, + "logits/rejected": -0.28515625, + "logps/chosen": -446.0, + "logps/rejected": -384.0, + "loss": 0.2346, + "rewards/accuracies": 0.8828125, + "rewards/chosen": 0.44140625, + "rewards/margins": 7.9375, + "rewards/rejected": -7.5, + "step": 810 + }, + { + "epoch": 0.9950920245398773, + "grad_norm": 26.578554179921486, + "learning_rate": 1.012809917355372e-06, + "logits/chosen": -0.08203125, + "logits/rejected": -0.21484375, + "logps/chosen": -394.0, + "logps/rejected": -380.0, + "loss": 0.3194, + "rewards/accuracies": 0.84375, + "rewards/chosen": 0.44921875, + "rewards/margins": 7.46875, + "rewards/rejected": -7.03125, + "step": 811 + }, + { + "epoch": 0.996319018404908, + "grad_norm": 34.32887556980461, + "learning_rate": 1.0121900826446281e-06, + "logits/chosen": -0.05908203125, + "logits/rejected": -0.125, + "logps/chosen": -364.0, + "logps/rejected": -340.0, + "loss": 0.3746, + "rewards/accuracies": 0.875, + "rewards/chosen": -0.09619140625, + "rewards/margins": 7.625, + "rewards/rejected": -7.71875, + "step": 812 + }, + { + "epoch": 0.9975460122699387, + "grad_norm": 25.82868308838759, + "learning_rate": 1.0115702479338843e-06, + "logits/chosen": 0.01611328125, + "logits/rejected": -0.1123046875, + "logps/chosen": -342.0, + "logps/rejected": -336.0, + "loss": 0.2067, + "rewards/accuracies": 0.8984375, + "rewards/chosen": 0.46484375, + "rewards/margins": 7.84375, + "rewards/rejected": -7.375, + "step": 813 + }, + { + "epoch": 0.9987730061349693, + "grad_norm": 35.210336167923685, + "learning_rate": 1.0109504132231404e-06, + "logits/chosen": 0.054443359375, + "logits/rejected": -0.0206298828125, + "logps/chosen": -388.0, + "logps/rejected": -370.0, + "loss": 0.3789, + "rewards/accuracies": 0.78125, + "rewards/chosen": -0.80859375, + "rewards/margins": 6.125, + "rewards/rejected": -6.90625, + "step": 814 + }, + { + "epoch": 1.0, + "grad_norm": 31.93988486773496, + "learning_rate": 1.0103305785123968e-06, + "logits/chosen": 0.1279296875, + "logits/rejected": -0.061279296875, + "logps/chosen": -388.0, + "logps/rejected": -364.0, + "loss": 0.2533, + "rewards/accuracies": 0.8991597294807434, + "rewards/chosen": -0.095703125, + "rewards/margins": 7.53125, + "rewards/rejected": -7.625, + "step": 815 + } + ], + "logging_steps": 1, + "max_steps": 2445, + "num_input_tokens_seen": 0, + "num_train_epochs": 3, + "save_steps": 500, + "stateful_callbacks": { + "TrainerControl": { + "args": { + "should_epoch_stop": false, + "should_evaluate": false, + "should_log": false, + "should_save": true, + "should_training_stop": false + }, + "attributes": {} + } + }, + "total_flos": 0.0, + "train_batch_size": 4, + "trial_name": null, + "trial_params": null +} diff --git a/checkpoint-815/training_args.bin b/checkpoint-815/training_args.bin new file mode 100644 index 0000000..a88cbd3 --- /dev/null +++ b/checkpoint-815/training_args.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:e0cffaa6f04a26499dd6171106fabe250a7e89bd923590f4fdeb897bc28870a9 +size 9528 diff --git a/checkpoint-815/zero_to_fp32.py b/checkpoint-815/zero_to_fp32.py new file mode 100644 index 0000000..24cc342 --- /dev/null +++ b/checkpoint-815/zero_to_fp32.py @@ -0,0 +1,604 @@ +#!/usr/bin/env python + +# Copyright (c) Microsoft Corporation. +# SPDX-License-Identifier: Apache-2.0 + +# DeepSpeed Team + +# This script extracts fp32 consolidated weights from a zero 1, 2 and 3 DeepSpeed checkpoints. It gets +# copied into the top level checkpoint dir, so the user can easily do the conversion at any point in +# the future. Once extracted, the weights don't require DeepSpeed and can be used in any +# application. +# +# example: python zero_to_fp32.py . pytorch_model.bin + +import argparse +import torch +import glob +import math +import os +import re +from collections import OrderedDict +from dataclasses import dataclass + +# while this script doesn't use deepspeed to recover data, since the checkpoints are pickled with +# DeepSpeed data structures it has to be available in the current python environment. +from deepspeed.utils import logger +from deepspeed.checkpoint.constants import (DS_VERSION, OPTIMIZER_STATE_DICT, SINGLE_PARTITION_OF_FP32_GROUPS, + FP32_FLAT_GROUPS, ZERO_STAGE, PARTITION_COUNT, PARAM_SHAPES, BUFFER_NAMES, + FROZEN_PARAM_SHAPES, FROZEN_PARAM_FRAGMENTS) + + +@dataclass +class zero_model_state: + buffers: dict() + param_shapes: dict() + shared_params: list + ds_version: int + frozen_param_shapes: dict() + frozen_param_fragments: dict() + + +debug = 0 + +# load to cpu +device = torch.device('cpu') + + +def atoi(text): + return int(text) if text.isdigit() else text + + +def natural_keys(text): + ''' + alist.sort(key=natural_keys) sorts in human order + http://nedbatchelder.com/blog/200712/human_sorting.html + (See Toothy's implementation in the comments) + ''' + return [atoi(c) for c in re.split(r'(\d+)', text)] + + +def get_model_state_file(checkpoint_dir, zero_stage): + if not os.path.isdir(checkpoint_dir): + raise FileNotFoundError(f"Directory '{checkpoint_dir}' doesn't exist") + + # there should be only one file + if zero_stage <= 2: + file = os.path.join(checkpoint_dir, "mp_rank_00_model_states.pt") + elif zero_stage == 3: + file = os.path.join(checkpoint_dir, "zero_pp_rank_0_mp_rank_00_model_states.pt") + + if not os.path.exists(file): + raise FileNotFoundError(f"can't find model states file at '{file}'") + + return file + + +def get_checkpoint_files(checkpoint_dir, glob_pattern): + # XXX: need to test that this simple glob rule works for multi-node setup too + ckpt_files = sorted(glob.glob(os.path.join(checkpoint_dir, glob_pattern)), key=natural_keys) + + if len(ckpt_files) == 0: + raise FileNotFoundError(f"can't find {glob_pattern} files in directory '{checkpoint_dir}'") + + return ckpt_files + + +def get_optim_files(checkpoint_dir): + return get_checkpoint_files(checkpoint_dir, "*_optim_states.pt") + + +def get_model_state_files(checkpoint_dir): + return get_checkpoint_files(checkpoint_dir, "*_model_states.pt") + + +def parse_model_states(files): + zero_model_states = [] + for file in files: + state_dict = torch.load(file, map_location=device) + + if BUFFER_NAMES not in state_dict: + raise ValueError(f"{file} is not a model state checkpoint") + buffer_names = state_dict[BUFFER_NAMES] + if debug: + print("Found buffers:", buffer_names) + + # recover just the buffers while restoring them to fp32 if they were saved in fp16 + buffers = {k: v.float() for k, v in state_dict["module"].items() if k in buffer_names} + param_shapes = state_dict[PARAM_SHAPES] + + # collect parameters that are included in param_shapes + param_names = [] + for s in param_shapes: + for name in s.keys(): + param_names.append(name) + + # update with frozen parameters + frozen_param_shapes = state_dict.get(FROZEN_PARAM_SHAPES, None) + if frozen_param_shapes is not None: + if debug: + print(f"Found frozen_param_shapes: {frozen_param_shapes}") + param_names += list(frozen_param_shapes.keys()) + + # handle shared params + shared_params = [[k, v] for k, v in state_dict["shared_params"].items()] + + ds_version = state_dict.get(DS_VERSION, None) + + frozen_param_fragments = state_dict.get(FROZEN_PARAM_FRAGMENTS, None) + + z_model_state = zero_model_state(buffers=buffers, + param_shapes=param_shapes, + shared_params=shared_params, + ds_version=ds_version, + frozen_param_shapes=frozen_param_shapes, + frozen_param_fragments=frozen_param_fragments) + zero_model_states.append(z_model_state) + + return zero_model_states + + +def parse_optim_states(files, ds_checkpoint_dir): + + total_files = len(files) + state_dicts = [] + for f in files: + state_dict = torch.load(f, map_location=device) + # immediately discard the potentially huge 2 optimizer states as we only care for fp32 master weights + # and also handle the case where it was already removed by another helper script + state_dict["optimizer_state_dict"].pop("optimizer_state_dict", None) + state_dicts.append(state_dict) + + if not ZERO_STAGE in state_dicts[0][OPTIMIZER_STATE_DICT]: + raise ValueError(f"{files[0]} is not a zero checkpoint") + zero_stage = state_dicts[0][OPTIMIZER_STATE_DICT][ZERO_STAGE] + world_size = state_dicts[0][OPTIMIZER_STATE_DICT][PARTITION_COUNT] + + # For ZeRO-2 each param group can have different partition_count as data parallelism for expert + # parameters can be different from data parallelism for non-expert parameters. So we can just + # use the max of the partition_count to get the dp world_size. + + if type(world_size) is list: + world_size = max(world_size) + + if world_size != total_files: + raise ValueError( + f"Expected {world_size} of '*_optim_states.pt' under '{ds_checkpoint_dir}' but found {total_files} files. " + "Possibly due to an overwrite of an old checkpoint, or a checkpoint didn't get saved by one or more processes." + ) + + # the groups are named differently in each stage + if zero_stage <= 2: + fp32_groups_key = SINGLE_PARTITION_OF_FP32_GROUPS + elif zero_stage == 3: + fp32_groups_key = FP32_FLAT_GROUPS + else: + raise ValueError(f"unknown zero stage {zero_stage}") + + if zero_stage <= 2: + fp32_flat_groups = [state_dicts[i][OPTIMIZER_STATE_DICT][fp32_groups_key] for i in range(len(state_dicts))] + elif zero_stage == 3: + # if there is more than one param group, there will be multiple flattened tensors - one + # flattened tensor per group - for simplicity merge them into a single tensor + # + # XXX: could make the script more memory efficient for when there are multiple groups - it + # will require matching the sub-lists of param_shapes for each param group flattened tensor + + fp32_flat_groups = [ + torch.cat(state_dicts[i][OPTIMIZER_STATE_DICT][fp32_groups_key], 0) for i in range(len(state_dicts)) + ] + + return zero_stage, world_size, fp32_flat_groups + + +def _get_fp32_state_dict_from_zero_checkpoint(ds_checkpoint_dir, exclude_frozen_parameters): + """ + Returns fp32 state_dict reconstructed from ds checkpoint + + Args: + - ``ds_checkpoint_dir``: path to the deepspeed checkpoint folder (where the optimizer files are) + + """ + print(f"Processing zero checkpoint '{ds_checkpoint_dir}'") + + optim_files = get_optim_files(ds_checkpoint_dir) + zero_stage, world_size, fp32_flat_groups = parse_optim_states(optim_files, ds_checkpoint_dir) + print(f"Detected checkpoint of type zero stage {zero_stage}, world_size: {world_size}") + + model_files = get_model_state_files(ds_checkpoint_dir) + + zero_model_states = parse_model_states(model_files) + print(f'Parsing checkpoint created by deepspeed=={zero_model_states[0].ds_version}') + + if zero_stage <= 2: + return _get_fp32_state_dict_from_zero2_checkpoint(world_size, fp32_flat_groups, zero_model_states, + exclude_frozen_parameters) + elif zero_stage == 3: + return _get_fp32_state_dict_from_zero3_checkpoint(world_size, fp32_flat_groups, zero_model_states, + exclude_frozen_parameters) + + +def _zero2_merge_frozen_params(state_dict, zero_model_states): + if zero_model_states[0].frozen_param_shapes is None or len(zero_model_states[0].frozen_param_shapes) == 0: + return + + frozen_param_shapes = zero_model_states[0].frozen_param_shapes + frozen_param_fragments = zero_model_states[0].frozen_param_fragments + + if debug: + num_elem = sum(s.numel() for s in frozen_param_shapes.values()) + print(f'rank 0: {FROZEN_PARAM_SHAPES}.numel = {num_elem}') + + wanted_params = len(frozen_param_shapes) + wanted_numel = sum(s.numel() for s in frozen_param_shapes.values()) + avail_numel = sum([p.numel() for p in frozen_param_fragments.values()]) + print(f'Frozen params: Have {avail_numel} numels to process.') + print(f'Frozen params: Need {wanted_numel} numels in {wanted_params} params') + + total_params = 0 + total_numel = 0 + for name, shape in frozen_param_shapes.items(): + total_params += 1 + unpartitioned_numel = shape.numel() + total_numel += unpartitioned_numel + + state_dict[name] = frozen_param_fragments[name] + + if debug: + print(f"{name} full shape: {shape} unpartitioned numel {unpartitioned_numel} ") + + print(f"Reconstructed Frozen fp32 state dict with {total_params} params {total_numel} elements") + + +def _has_callable(obj, fn): + attr = getattr(obj, fn, None) + return callable(attr) + + +def _zero2_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states): + param_shapes = zero_model_states[0].param_shapes + + # Reconstruction protocol: + # + # XXX: document this + + if debug: + for i in range(world_size): + for j in range(len(fp32_flat_groups[0])): + print(f"{FP32_FLAT_GROUPS}[{i}][{j}].shape={fp32_flat_groups[i][j].shape}") + + # XXX: memory usage doubles here (zero2) + num_param_groups = len(fp32_flat_groups[0]) + merged_single_partition_of_fp32_groups = [] + for i in range(num_param_groups): + merged_partitions = [sd[i] for sd in fp32_flat_groups] + full_single_fp32_vector = torch.cat(merged_partitions, 0) + merged_single_partition_of_fp32_groups.append(full_single_fp32_vector) + avail_numel = sum( + [full_single_fp32_vector.numel() for full_single_fp32_vector in merged_single_partition_of_fp32_groups]) + + if debug: + wanted_params = sum([len(shapes) for shapes in param_shapes]) + wanted_numel = sum([sum(shape.numel() for shape in shapes.values()) for shapes in param_shapes]) + # not asserting if there is a mismatch due to possible padding + print(f"Have {avail_numel} numels to process.") + print(f"Need {wanted_numel} numels in {wanted_params} params.") + + # params + # XXX: for huge models that can't fit into the host's RAM we will have to recode this to support + # out-of-core computing solution + total_numel = 0 + total_params = 0 + for shapes, full_single_fp32_vector in zip(param_shapes, merged_single_partition_of_fp32_groups): + offset = 0 + avail_numel = full_single_fp32_vector.numel() + for name, shape in shapes.items(): + + unpartitioned_numel = shape.numel() if _has_callable(shape, 'numel') else math.prod(shape) + total_numel += unpartitioned_numel + total_params += 1 + + if debug: + print(f"{name} full shape: {shape} unpartitioned numel {unpartitioned_numel} ") + state_dict[name] = full_single_fp32_vector.narrow(0, offset, unpartitioned_numel).view(shape) + offset += unpartitioned_numel + + # Z2 started to align to 2*world_size to improve nccl performance. Therefore both offset and + # avail_numel can differ by anywhere between 0..2*world_size. Due to two unrelated complex + # paddings performed in the code it's almost impossible to predict the exact numbers w/o the + # live optimizer object, so we are checking that the numbers are within the right range + align_to = 2 * world_size + + def zero2_align(x): + return align_to * math.ceil(x / align_to) + + if debug: + print(f"original offset={offset}, avail_numel={avail_numel}") + + offset = zero2_align(offset) + avail_numel = zero2_align(avail_numel) + + if debug: + print(f"aligned offset={offset}, avail_numel={avail_numel}") + + # Sanity check + if offset != avail_numel: + raise ValueError(f"consumed {offset} numels out of {avail_numel} - something is wrong") + + print(f"Reconstructed fp32 state dict with {total_params} params {total_numel} elements") + + +def _get_fp32_state_dict_from_zero2_checkpoint(world_size, fp32_flat_groups, zero_model_states, + exclude_frozen_parameters): + state_dict = OrderedDict() + + # buffers + buffers = zero_model_states[0].buffers + state_dict.update(buffers) + if debug: + print(f"added {len(buffers)} buffers") + + if not exclude_frozen_parameters: + _zero2_merge_frozen_params(state_dict, zero_model_states) + + _zero2_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states) + + # recover shared parameters + for pair in zero_model_states[0].shared_params: + if pair[1] in state_dict: + state_dict[pair[0]] = state_dict[pair[1]] + + return state_dict + + +def zero3_partitioned_param_info(unpartitioned_numel, world_size): + remainder = unpartitioned_numel % world_size + padding_numel = (world_size - remainder) if remainder else 0 + partitioned_numel = math.ceil(unpartitioned_numel / world_size) + return partitioned_numel, padding_numel + + +def _zero3_merge_frozen_params(state_dict, world_size, zero_model_states): + if zero_model_states[0].frozen_param_shapes is None or len(zero_model_states[0].frozen_param_shapes) == 0: + return + + if debug: + for i in range(world_size): + num_elem = sum(s.numel() for s in zero_model_states[i].frozen_param_fragments.values()) + print(f'rank {i}: {FROZEN_PARAM_SHAPES}.numel = {num_elem}') + + frozen_param_shapes = zero_model_states[0].frozen_param_shapes + wanted_params = len(frozen_param_shapes) + wanted_numel = sum(s.numel() for s in frozen_param_shapes.values()) + avail_numel = sum([p.numel() for p in zero_model_states[0].frozen_param_fragments.values()]) * world_size + print(f'Frozen params: Have {avail_numel} numels to process.') + print(f'Frozen params: Need {wanted_numel} numels in {wanted_params} params') + + total_params = 0 + total_numel = 0 + for name, shape in zero_model_states[0].frozen_param_shapes.items(): + total_params += 1 + unpartitioned_numel = shape.numel() + total_numel += unpartitioned_numel + + param_frags = tuple(model_state.frozen_param_fragments[name] for model_state in zero_model_states) + state_dict[name] = torch.cat(param_frags, 0).narrow(0, 0, unpartitioned_numel).view(shape) + + partitioned_numel, partitioned_padding_numel = zero3_partitioned_param_info(unpartitioned_numel, world_size) + + if debug: + print( + f"Frozen params: {total_params} {name} full shape: {shape} partition0 numel={partitioned_numel} partitioned_padding_numel={partitioned_padding_numel}" + ) + + print(f"Reconstructed Frozen fp32 state dict with {total_params} params {total_numel} elements") + + +def _zero3_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states): + param_shapes = zero_model_states[0].param_shapes + avail_numel = fp32_flat_groups[0].numel() * world_size + # Reconstruction protocol: For zero3 we need to zip the partitions together at boundary of each + # param, re-consolidating each param, while dealing with padding if any + + # merge list of dicts, preserving order + param_shapes = {k: v for d in param_shapes for k, v in d.items()} + + if debug: + for i in range(world_size): + print(f"{FP32_FLAT_GROUPS}[{i}].shape={fp32_flat_groups[i].shape}") + + wanted_params = len(param_shapes) + wanted_numel = sum(shape.numel() for shape in param_shapes.values()) + # not asserting if there is a mismatch due to possible padding + avail_numel = fp32_flat_groups[0].numel() * world_size + print(f"Trainable params: Have {avail_numel} numels to process.") + print(f"Trainable params: Need {wanted_numel} numels in {wanted_params} params.") + + # params + # XXX: for huge models that can't fit into the host's RAM we will have to recode this to support + # out-of-core computing solution + offset = 0 + total_numel = 0 + total_params = 0 + for name, shape in param_shapes.items(): + + unpartitioned_numel = shape.numel() + total_numel += unpartitioned_numel + total_params += 1 + + partitioned_numel, partitioned_padding_numel = zero3_partitioned_param_info(unpartitioned_numel, world_size) + + if debug: + print( + f"Trainable params: {total_params} {name} full shape: {shape} partition0 numel={partitioned_numel} partitioned_padding_numel={partitioned_padding_numel}" + ) + + # XXX: memory usage doubles here + state_dict[name] = torch.cat( + tuple(fp32_flat_groups[i].narrow(0, offset, partitioned_numel) for i in range(world_size)), + 0).narrow(0, 0, unpartitioned_numel).view(shape) + offset += partitioned_numel + + offset *= world_size + + # Sanity check + if offset != avail_numel: + raise ValueError(f"consumed {offset} numels out of {avail_numel} - something is wrong") + + print(f"Reconstructed Trainable fp32 state dict with {total_params} params {total_numel} elements") + + +def _get_fp32_state_dict_from_zero3_checkpoint(world_size, fp32_flat_groups, zero_model_states, + exclude_frozen_parameters): + state_dict = OrderedDict() + + # buffers + buffers = zero_model_states[0].buffers + state_dict.update(buffers) + if debug: + print(f"added {len(buffers)} buffers") + + if not exclude_frozen_parameters: + _zero3_merge_frozen_params(state_dict, world_size, zero_model_states) + + _zero3_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states) + + # recover shared parameters + for pair in zero_model_states[0].shared_params: + if pair[1] in state_dict: + state_dict[pair[0]] = state_dict[pair[1]] + + return state_dict + + +def get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag=None, exclude_frozen_parameters=False): + """ + Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated state_dict that can be loaded with + ``load_state_dict()`` and used for training without DeepSpeed or shared with others, for example + via a model hub. + + Args: + - ``checkpoint_dir``: path to the desired checkpoint folder + - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in 'latest' file. e.g., ``global_step14`` + - ``exclude_frozen_parameters``: exclude frozen parameters + + Returns: + - pytorch ``state_dict`` + + Note: this approach may not work if your application doesn't have sufficient free CPU memory and + you may need to use the offline approach using the ``zero_to_fp32.py`` script that is saved with + the checkpoint. + + A typical usage might be :: + + from deepspeed.utils.zero_to_fp32 import get_fp32_state_dict_from_zero_checkpoint + # do the training and checkpoint saving + state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir) # already on cpu + model = model.cpu() # move to cpu + model.load_state_dict(state_dict) + # submit to model hub or save the model to share with others + + In this example the ``model`` will no longer be usable in the deepspeed context of the same + application. i.e. you will need to re-initialize the deepspeed engine, since + ``model.load_state_dict(state_dict)`` will remove all the deepspeed magic from it. + + If you want it all done for you, use ``load_state_dict_from_zero_checkpoint`` instead. + + """ + if tag is None: + latest_path = os.path.join(checkpoint_dir, 'latest') + if os.path.isfile(latest_path): + with open(latest_path, 'r') as fd: + tag = fd.read().strip() + else: + raise ValueError(f"Unable to find 'latest' file at {latest_path}") + + ds_checkpoint_dir = os.path.join(checkpoint_dir, tag) + + if not os.path.isdir(ds_checkpoint_dir): + raise FileNotFoundError(f"Directory '{ds_checkpoint_dir}' doesn't exist") + + return _get_fp32_state_dict_from_zero_checkpoint(ds_checkpoint_dir, exclude_frozen_parameters) + + +def convert_zero_checkpoint_to_fp32_state_dict(checkpoint_dir, output_file, tag=None, exclude_frozen_parameters=False): + """ + Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated ``state_dict`` file that can be + loaded with ``torch.load(file)`` + ``load_state_dict()`` and used for training without DeepSpeed. + + Args: + - ``checkpoint_dir``: path to the desired checkpoint folder. (one that contains the tag-folder, like ``global_step14``) + - ``output_file``: path to the pytorch fp32 state_dict output file (e.g. path/pytorch_model.bin) + - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in the file named ``latest`` in the checkpoint folder, e.g., ``global_step14`` + - ``exclude_frozen_parameters``: exclude frozen parameters + """ + + state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag, exclude_frozen_parameters) + print(f"Saving fp32 state dict to {output_file}") + torch.save(state_dict, output_file) + + +def load_state_dict_from_zero_checkpoint(model, checkpoint_dir, tag=None): + """ + 1. Put the provided model to cpu + 2. Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated ``state_dict`` + 3. Load it into the provided model + + Args: + - ``model``: the model object to update + - ``checkpoint_dir``: path to the desired checkpoint folder. (one that contains the tag-folder, like ``global_step14``) + - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in the file named ``latest`` in the checkpoint folder, e.g., ``global_step14`` + + Returns: + - ``model`: modified model + + Make sure you have plenty of CPU memory available before you call this function. If you don't + have enough use the ``zero_to_fp32.py`` utility to do the conversion. You will find it + conveniently placed for you in the checkpoint folder. + + A typical usage might be :: + + from deepspeed.utils.zero_to_fp32 import load_state_dict_from_zero_checkpoint + model = load_state_dict_from_zero_checkpoint(trainer.model, checkpoint_dir) + # submit to model hub or save the model to share with others + + Note, that once this was run, the ``model`` will no longer be usable in the deepspeed context + of the same application. i.e. you will need to re-initialize the deepspeed engine, since + ``model.load_state_dict(state_dict)`` will remove all the deepspeed magic from it. + + """ + logger.info(f"Extracting fp32 weights") + state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag) + + logger.info(f"Overwriting model with fp32 weights") + model = model.cpu() + model.load_state_dict(state_dict, strict=False) + + return model + + +if __name__ == "__main__": + + parser = argparse.ArgumentParser() + parser.add_argument("checkpoint_dir", + type=str, + help="path to the desired checkpoint folder, e.g., path/checkpoint-12") + parser.add_argument( + "output_file", + type=str, + help="path to the pytorch fp32 state_dict output file (e.g. path/checkpoint-12/pytorch_model.bin)") + parser.add_argument("-t", + "--tag", + type=str, + default=None, + help="checkpoint tag used as a unique identifier for checkpoint. e.g., global_step1") + parser.add_argument("--exclude_frozen_parameters", action='store_true', help="exclude frozen parameters") + parser.add_argument("-d", "--debug", action='store_true', help="enable debug") + args = parser.parse_args() + + debug = args.debug + + convert_zero_checkpoint_to_fp32_state_dict(args.checkpoint_dir, + args.output_file, + tag=args.tag, + exclude_frozen_parameters=args.exclude_frozen_parameters) diff --git a/config.json b/config.json new file mode 100644 index 0000000..ded5a8b --- /dev/null +++ b/config.json @@ -0,0 +1,35 @@ +{ + "architectures": [ + "LlamaForCausalLM" + ], + "attention_bias": false, + "attention_dropout": 0.0, + "bos_token_id": 128000, + "eos_token_id": 128009, + "head_dim": 128, + "hidden_act": "silu", + "hidden_size": 4096, + "initializer_range": 0.02, + "intermediate_size": 14336, + "max_position_embeddings": 131072, + "mlp_bias": false, + "model_type": "llama", + "num_attention_heads": 32, + "num_hidden_layers": 32, + "num_key_value_heads": 8, + "pretraining_tp": 1, + "rms_norm_eps": 1e-05, + "rope_scaling": { + "factor": 8.0, + "high_freq_factor": 4.0, + "low_freq_factor": 1.0, + "original_max_position_embeddings": 8192, + "rope_type": "llama3" + }, + "rope_theta": 500000.0, + "tie_word_embeddings": false, + "torch_dtype": "bfloat16", + "transformers_version": "4.50.0", + "use_cache": false, + "vocab_size": 128256 +} diff --git a/configuration.json b/configuration.json new file mode 100644 index 0000000..bbeeda1 --- /dev/null +++ b/configuration.json @@ -0,0 +1 @@ +{"framework": "pytorch", "task": "text-generation", "allow_remote": true} \ No newline at end of file diff --git a/generation_config.json b/generation_config.json new file mode 100644 index 0000000..0acdb5a --- /dev/null +++ b/generation_config.json @@ -0,0 +1,12 @@ +{ + "bos_token_id": 128000, + "do_sample": true, + "eos_token_id": [ + 128001, + 128008, + 128009 + ], + "temperature": 0.6, + "top_p": 0.9, + "transformers_version": "4.50.0" +} diff --git a/model-00001-of-00004.safetensors b/model-00001-of-00004.safetensors new file mode 100644 index 0000000..e954b8d --- /dev/null +++ b/model-00001-of-00004.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:0bd99094aa0d9de1c56cec4a5cfe8c2278dd5fa844cd5eb17a98f4eb663b9a9d +size 4976698672 diff --git a/model-00002-of-00004.safetensors b/model-00002-of-00004.safetensors new file mode 100644 index 0000000..479cac7 --- /dev/null +++ b/model-00002-of-00004.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:26f9f2f17af25c67509c30d8c79c3e9bca63682789638a940b9ad429db428593 +size 4999802720 diff --git a/model-00003-of-00004.safetensors b/model-00003-of-00004.safetensors new file mode 100644 index 0000000..725f564 --- /dev/null +++ b/model-00003-of-00004.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:6207337975b4b7d327d6135e93cde49007c00a5c630eadb2a8ce685f83b431a2 +size 4915916176 diff --git a/model-00004-of-00004.safetensors b/model-00004-of-00004.safetensors new file mode 100644 index 0000000..b4c25ca --- /dev/null +++ b/model-00004-of-00004.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:c2e5860979ea435ace5595020a69b0f0a26d819b061ba8455b510448b4226ba1 +size 1168138808 diff --git a/model.safetensors.index.json b/model.safetensors.index.json new file mode 100644 index 0000000..0fd8120 --- /dev/null +++ b/model.safetensors.index.json @@ -0,0 +1,298 @@ +{ + "metadata": { + "total_size": 16060522496 + }, + "weight_map": { + "lm_head.weight": "model-00004-of-00004.safetensors", + "model.embed_tokens.weight": "model-00001-of-00004.safetensors", + "model.layers.0.input_layernorm.weight": "model-00001-of-00004.safetensors", + "model.layers.0.mlp.down_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.0.mlp.gate_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.0.mlp.up_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.0.post_attention_layernorm.weight": "model-00001-of-00004.safetensors", + "model.layers.0.self_attn.k_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.0.self_attn.o_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.0.self_attn.q_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.0.self_attn.v_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.1.input_layernorm.weight": "model-00001-of-00004.safetensors", + "model.layers.1.mlp.down_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.1.mlp.gate_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.1.mlp.up_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.1.post_attention_layernorm.weight": "model-00001-of-00004.safetensors", + "model.layers.1.self_attn.k_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.1.self_attn.o_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.1.self_attn.q_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.1.self_attn.v_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.10.input_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.10.mlp.down_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.10.mlp.gate_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.10.mlp.up_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.10.post_attention_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.10.self_attn.k_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.10.self_attn.o_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.10.self_attn.q_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.10.self_attn.v_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.11.input_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.11.mlp.down_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.11.mlp.gate_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.11.mlp.up_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.11.post_attention_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.11.self_attn.k_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.11.self_attn.o_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.11.self_attn.q_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.11.self_attn.v_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.12.input_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.12.mlp.down_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.12.mlp.gate_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.12.mlp.up_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.12.post_attention_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.12.self_attn.k_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.12.self_attn.o_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.12.self_attn.q_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.12.self_attn.v_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.13.input_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.13.mlp.down_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.13.mlp.gate_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.13.mlp.up_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.13.post_attention_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.13.self_attn.k_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.13.self_attn.o_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.13.self_attn.q_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.13.self_attn.v_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.14.input_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.14.mlp.down_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.14.mlp.gate_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.14.mlp.up_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.14.post_attention_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.14.self_attn.k_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.14.self_attn.o_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.14.self_attn.q_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.14.self_attn.v_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.15.input_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.15.mlp.down_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.15.mlp.gate_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.15.mlp.up_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.15.post_attention_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.15.self_attn.k_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.15.self_attn.o_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.15.self_attn.q_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.15.self_attn.v_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.16.input_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.16.mlp.down_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.16.mlp.gate_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.16.mlp.up_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.16.post_attention_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.16.self_attn.k_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.16.self_attn.o_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.16.self_attn.q_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.16.self_attn.v_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.17.input_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.17.mlp.down_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.17.mlp.gate_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.17.mlp.up_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.17.post_attention_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.17.self_attn.k_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.17.self_attn.o_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.17.self_attn.q_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.17.self_attn.v_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.18.input_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.18.mlp.down_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.18.mlp.gate_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.18.mlp.up_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.18.post_attention_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.18.self_attn.k_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.18.self_attn.o_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.18.self_attn.q_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.18.self_attn.v_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.19.input_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.19.mlp.down_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.19.mlp.gate_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.19.mlp.up_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.19.post_attention_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.19.self_attn.k_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.19.self_attn.o_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.19.self_attn.q_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.19.self_attn.v_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.2.input_layernorm.weight": "model-00001-of-00004.safetensors", + "model.layers.2.mlp.down_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.2.mlp.gate_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.2.mlp.up_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.2.post_attention_layernorm.weight": "model-00001-of-00004.safetensors", + "model.layers.2.self_attn.k_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.2.self_attn.o_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.2.self_attn.q_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.2.self_attn.v_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.20.input_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.20.mlp.down_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.20.mlp.gate_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.20.mlp.up_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.20.post_attention_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.20.self_attn.k_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.20.self_attn.o_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.20.self_attn.q_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.20.self_attn.v_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.21.input_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.21.mlp.down_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.21.mlp.gate_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.21.mlp.up_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.21.post_attention_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.21.self_attn.k_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.21.self_attn.o_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.21.self_attn.q_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.21.self_attn.v_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.22.input_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.22.mlp.down_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.22.mlp.gate_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.22.mlp.up_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.22.post_attention_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.22.self_attn.k_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.22.self_attn.o_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.22.self_attn.q_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.22.self_attn.v_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.23.input_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.23.mlp.down_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.23.mlp.gate_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.23.mlp.up_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.23.post_attention_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.23.self_attn.k_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.23.self_attn.o_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.23.self_attn.q_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.23.self_attn.v_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.24.input_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.24.mlp.down_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.24.mlp.gate_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.24.mlp.up_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.24.post_attention_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.24.self_attn.k_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.24.self_attn.o_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.24.self_attn.q_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.24.self_attn.v_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.25.input_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.25.mlp.down_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.25.mlp.gate_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.25.mlp.up_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.25.post_attention_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.25.self_attn.k_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.25.self_attn.o_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.25.self_attn.q_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.25.self_attn.v_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.26.input_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.26.mlp.down_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.26.mlp.gate_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.26.mlp.up_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.26.post_attention_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.26.self_attn.k_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.26.self_attn.o_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.26.self_attn.q_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.26.self_attn.v_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.27.input_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.27.mlp.down_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.27.mlp.gate_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.27.mlp.up_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.27.post_attention_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.27.self_attn.k_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.27.self_attn.o_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.27.self_attn.q_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.27.self_attn.v_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.28.input_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.28.mlp.down_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.28.mlp.gate_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.28.mlp.up_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.28.post_attention_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.28.self_attn.k_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.28.self_attn.o_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.28.self_attn.q_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.28.self_attn.v_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.29.input_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.29.mlp.down_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.29.mlp.gate_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.29.mlp.up_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.29.post_attention_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.29.self_attn.k_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.29.self_attn.o_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.29.self_attn.q_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.29.self_attn.v_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.3.input_layernorm.weight": "model-00001-of-00004.safetensors", + "model.layers.3.mlp.down_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.3.mlp.gate_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.3.mlp.up_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.3.post_attention_layernorm.weight": "model-00001-of-00004.safetensors", + "model.layers.3.self_attn.k_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.3.self_attn.o_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.3.self_attn.q_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.3.self_attn.v_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.30.input_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.30.mlp.down_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.30.mlp.gate_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.30.mlp.up_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.30.post_attention_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.30.self_attn.k_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.30.self_attn.o_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.30.self_attn.q_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.30.self_attn.v_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.31.input_layernorm.weight": "model-00004-of-00004.safetensors", + "model.layers.31.mlp.down_proj.weight": "model-00004-of-00004.safetensors", + "model.layers.31.mlp.gate_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.31.mlp.up_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.31.post_attention_layernorm.weight": "model-00004-of-00004.safetensors", + "model.layers.31.self_attn.k_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.31.self_attn.o_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.31.self_attn.q_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.31.self_attn.v_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.4.input_layernorm.weight": "model-00001-of-00004.safetensors", + "model.layers.4.mlp.down_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.4.mlp.gate_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.4.mlp.up_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.4.post_attention_layernorm.weight": "model-00001-of-00004.safetensors", + "model.layers.4.self_attn.k_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.4.self_attn.o_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.4.self_attn.q_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.4.self_attn.v_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.5.input_layernorm.weight": "model-00001-of-00004.safetensors", + "model.layers.5.mlp.down_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.5.mlp.gate_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.5.mlp.up_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.5.post_attention_layernorm.weight": "model-00001-of-00004.safetensors", + "model.layers.5.self_attn.k_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.5.self_attn.o_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.5.self_attn.q_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.5.self_attn.v_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.6.input_layernorm.weight": "model-00001-of-00004.safetensors", + "model.layers.6.mlp.down_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.6.mlp.gate_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.6.mlp.up_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.6.post_attention_layernorm.weight": "model-00001-of-00004.safetensors", + "model.layers.6.self_attn.k_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.6.self_attn.o_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.6.self_attn.q_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.6.self_attn.v_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.7.input_layernorm.weight": "model-00001-of-00004.safetensors", + "model.layers.7.mlp.down_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.7.mlp.gate_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.7.mlp.up_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.7.post_attention_layernorm.weight": "model-00001-of-00004.safetensors", + "model.layers.7.self_attn.k_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.7.self_attn.o_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.7.self_attn.q_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.7.self_attn.v_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.8.input_layernorm.weight": "model-00001-of-00004.safetensors", + "model.layers.8.mlp.down_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.8.mlp.gate_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.8.mlp.up_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.8.post_attention_layernorm.weight": "model-00001-of-00004.safetensors", + "model.layers.8.self_attn.k_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.8.self_attn.o_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.8.self_attn.q_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.8.self_attn.v_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.9.input_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.9.mlp.down_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.9.mlp.gate_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.9.mlp.up_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.9.post_attention_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.9.self_attn.k_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.9.self_attn.o_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.9.self_attn.q_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.9.self_attn.v_proj.weight": "model-00002-of-00004.safetensors", + "model.norm.weight": "model-00004-of-00004.safetensors" + } +} diff --git a/runs/Apr03_06-15-34_ip-10-1-79-202/events.out.tfevents.1743660961.ip-10-1-79-202.1157702.0 b/runs/Apr03_06-15-34_ip-10-1-79-202/events.out.tfevents.1743660961.ip-10-1-79-202.1157702.0 new file mode 100644 index 0000000..f722592 --- /dev/null +++ b/runs/Apr03_06-15-34_ip-10-1-79-202/events.out.tfevents.1743660961.ip-10-1-79-202.1157702.0 @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:e8a99802f9ae8ccbccec2f6b8ae207d98c779d4aa31b1279454a887165a9c098 +size 1688675 diff --git a/special_tokens_map.json b/special_tokens_map.json new file mode 100644 index 0000000..278b7f0 --- /dev/null +++ b/special_tokens_map.json @@ -0,0 +1,23 @@ +{ + "bos_token": { + "content": "<|begin_of_text|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false + }, + "eos_token": { + "content": "<|eot_id|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false + }, + "pad_token": { + "content": "<|end_of_text|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false + } +} diff --git a/tokenizer.json b/tokenizer.json new file mode 100644 index 0000000..1c1d8d5 --- /dev/null +++ b/tokenizer.json @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:6b9e4e7fb171f92fd137b777cc2714bf87d11576700a1dcd7a399e7bbe39537b +size 17209920 diff --git a/tokenizer_config.json b/tokenizer_config.json new file mode 100644 index 0000000..8098161 --- /dev/null +++ b/tokenizer_config.json @@ -0,0 +1,2064 @@ +{ + "added_tokens_decoder": { + "128000": { + "content": "<|begin_of_text|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128001": { + "content": "<|end_of_text|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128002": { + "content": "<|reserved_special_token_0|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128003": { + "content": "<|reserved_special_token_1|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128004": { + "content": "<|finetune_right_pad_id|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128005": { + "content": "<|reserved_special_token_2|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128006": { + "content": "<|start_header_id|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128007": { + "content": "<|end_header_id|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128008": { + "content": "<|eom_id|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128009": { + "content": "<|eot_id|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128010": { + "content": "<|python_tag|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128011": { + "content": "<|reserved_special_token_3|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128012": { + "content": "<|reserved_special_token_4|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128013": { + "content": "<|reserved_special_token_5|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128014": { + "content": "<|reserved_special_token_6|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128015": { + "content": "<|reserved_special_token_7|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128016": { + "content": "<|reserved_special_token_8|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128017": { + "content": "<|reserved_special_token_9|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128018": { + "content": "<|reserved_special_token_10|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128019": { + "content": "<|reserved_special_token_11|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128020": { + "content": "<|reserved_special_token_12|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128021": { + "content": "<|reserved_special_token_13|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128022": { + "content": "<|reserved_special_token_14|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128023": { + "content": "<|reserved_special_token_15|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128024": { + "content": "<|reserved_special_token_16|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128025": { + "content": "<|reserved_special_token_17|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128026": { + "content": "<|reserved_special_token_18|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128027": { + "content": "<|reserved_special_token_19|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128028": { + "content": "<|reserved_special_token_20|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128029": { + "content": "<|reserved_special_token_21|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128030": { + "content": "<|reserved_special_token_22|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128031": { + "content": "<|reserved_special_token_23|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128032": { + "content": "<|reserved_special_token_24|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128033": { + "content": "<|reserved_special_token_25|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128034": { + "content": "<|reserved_special_token_26|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128035": { + "content": "<|reserved_special_token_27|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128036": { + "content": "<|reserved_special_token_28|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128037": { + "content": "<|reserved_special_token_29|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128038": { + "content": "<|reserved_special_token_30|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128039": { + "content": "<|reserved_special_token_31|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128040": { + "content": "<|reserved_special_token_32|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128041": { + "content": "<|reserved_special_token_33|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128042": { + "content": "<|reserved_special_token_34|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128043": { + "content": "<|reserved_special_token_35|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128044": { + "content": "<|reserved_special_token_36|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128045": { + "content": "<|reserved_special_token_37|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128046": { + "content": "<|reserved_special_token_38|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128047": { + "content": "<|reserved_special_token_39|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128048": { + "content": "<|reserved_special_token_40|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128049": { + "content": "<|reserved_special_token_41|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128050": { + "content": "<|reserved_special_token_42|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128051": { + "content": "<|reserved_special_token_43|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128052": { + "content": "<|reserved_special_token_44|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128053": { + "content": "<|reserved_special_token_45|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128054": { + "content": "<|reserved_special_token_46|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128055": { + "content": "<|reserved_special_token_47|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128056": { + "content": "<|reserved_special_token_48|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128057": { + "content": "<|reserved_special_token_49|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128058": { + "content": "<|reserved_special_token_50|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128059": { + "content": "<|reserved_special_token_51|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128060": { + "content": "<|reserved_special_token_52|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128061": { + "content": "<|reserved_special_token_53|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128062": { + "content": "<|reserved_special_token_54|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128063": { + "content": "<|reserved_special_token_55|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128064": { + "content": "<|reserved_special_token_56|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128065": { + "content": "<|reserved_special_token_57|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128066": { + "content": "<|reserved_special_token_58|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128067": { + "content": "<|reserved_special_token_59|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128068": { + "content": "<|reserved_special_token_60|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128069": { + "content": "<|reserved_special_token_61|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128070": { + "content": "<|reserved_special_token_62|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128071": { + "content": "<|reserved_special_token_63|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128072": { + "content": "<|reserved_special_token_64|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128073": { + "content": "<|reserved_special_token_65|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128074": { + "content": "<|reserved_special_token_66|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128075": { + "content": "<|reserved_special_token_67|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128076": { + "content": "<|reserved_special_token_68|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128077": { + "content": "<|reserved_special_token_69|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128078": { + "content": "<|reserved_special_token_70|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128079": { + "content": "<|reserved_special_token_71|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128080": { + "content": "<|reserved_special_token_72|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128081": { + "content": "<|reserved_special_token_73|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128082": { + "content": "<|reserved_special_token_74|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128083": { + "content": "<|reserved_special_token_75|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128084": { + "content": "<|reserved_special_token_76|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128085": { + "content": "<|reserved_special_token_77|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128086": { + "content": "<|reserved_special_token_78|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128087": { + "content": "<|reserved_special_token_79|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128088": { + "content": "<|reserved_special_token_80|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128089": { + "content": "<|reserved_special_token_81|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128090": { + "content": "<|reserved_special_token_82|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128091": { + "content": "<|reserved_special_token_83|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128092": { + "content": "<|reserved_special_token_84|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128093": { + "content": "<|reserved_special_token_85|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128094": { + "content": "<|reserved_special_token_86|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128095": { + "content": "<|reserved_special_token_87|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128096": { + "content": "<|reserved_special_token_88|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128097": { + "content": "<|reserved_special_token_89|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128098": { + "content": "<|reserved_special_token_90|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128099": { + "content": "<|reserved_special_token_91|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128100": { + "content": "<|reserved_special_token_92|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128101": { + "content": "<|reserved_special_token_93|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128102": { + "content": "<|reserved_special_token_94|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128103": { + "content": "<|reserved_special_token_95|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128104": { + "content": "<|reserved_special_token_96|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128105": { + "content": "<|reserved_special_token_97|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128106": { + "content": "<|reserved_special_token_98|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128107": { + "content": "<|reserved_special_token_99|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128108": { + "content": "<|reserved_special_token_100|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128109": { + "content": "<|reserved_special_token_101|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128110": { + "content": "<|reserved_special_token_102|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128111": { + "content": "<|reserved_special_token_103|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128112": { + "content": "<|reserved_special_token_104|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128113": { + "content": "<|reserved_special_token_105|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128114": { + "content": "<|reserved_special_token_106|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128115": { + "content": "<|reserved_special_token_107|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128116": { + "content": "<|reserved_special_token_108|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128117": { + "content": "<|reserved_special_token_109|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128118": { + "content": "<|reserved_special_token_110|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128119": { + "content": "<|reserved_special_token_111|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128120": { + "content": "<|reserved_special_token_112|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128121": { + "content": "<|reserved_special_token_113|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128122": { + "content": "<|reserved_special_token_114|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128123": { + "content": "<|reserved_special_token_115|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128124": { + "content": "<|reserved_special_token_116|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128125": { + "content": "<|reserved_special_token_117|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128126": { + "content": "<|reserved_special_token_118|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128127": { + "content": "<|reserved_special_token_119|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128128": { + "content": "<|reserved_special_token_120|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128129": { + "content": "<|reserved_special_token_121|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128130": { + "content": "<|reserved_special_token_122|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128131": { + "content": "<|reserved_special_token_123|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128132": { + "content": "<|reserved_special_token_124|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128133": { + "content": "<|reserved_special_token_125|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128134": { + "content": "<|reserved_special_token_126|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128135": { + "content": "<|reserved_special_token_127|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128136": { + "content": "<|reserved_special_token_128|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128137": { + "content": "<|reserved_special_token_129|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128138": { + "content": "<|reserved_special_token_130|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128139": { + "content": "<|reserved_special_token_131|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128140": { + "content": "<|reserved_special_token_132|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128141": { + "content": "<|reserved_special_token_133|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128142": { + "content": "<|reserved_special_token_134|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128143": { + "content": "<|reserved_special_token_135|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128144": { + "content": "<|reserved_special_token_136|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128145": { + "content": "<|reserved_special_token_137|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128146": { + "content": "<|reserved_special_token_138|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128147": { + "content": "<|reserved_special_token_139|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128148": { + "content": "<|reserved_special_token_140|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128149": { + "content": "<|reserved_special_token_141|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128150": { + "content": "<|reserved_special_token_142|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128151": { + "content": "<|reserved_special_token_143|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128152": { + "content": "<|reserved_special_token_144|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128153": { + "content": "<|reserved_special_token_145|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128154": { + "content": "<|reserved_special_token_146|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128155": { + "content": "<|reserved_special_token_147|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128156": { + "content": "<|reserved_special_token_148|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128157": { + "content": "<|reserved_special_token_149|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128158": { + "content": "<|reserved_special_token_150|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128159": { + "content": "<|reserved_special_token_151|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128160": { + "content": "<|reserved_special_token_152|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128161": { + "content": "<|reserved_special_token_153|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128162": { + "content": "<|reserved_special_token_154|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128163": { + "content": "<|reserved_special_token_155|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128164": { + "content": "<|reserved_special_token_156|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128165": { + "content": "<|reserved_special_token_157|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128166": { + "content": "<|reserved_special_token_158|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128167": { + "content": "<|reserved_special_token_159|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128168": { + "content": "<|reserved_special_token_160|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128169": { + "content": "<|reserved_special_token_161|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128170": { + "content": "<|reserved_special_token_162|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128171": { + "content": "<|reserved_special_token_163|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128172": { + "content": "<|reserved_special_token_164|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128173": { + "content": "<|reserved_special_token_165|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128174": { + "content": "<|reserved_special_token_166|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128175": { + "content": "<|reserved_special_token_167|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128176": { + "content": "<|reserved_special_token_168|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128177": { + "content": "<|reserved_special_token_169|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128178": { + "content": "<|reserved_special_token_170|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128179": { + "content": "<|reserved_special_token_171|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128180": { + "content": "<|reserved_special_token_172|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128181": { + "content": "<|reserved_special_token_173|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128182": { + "content": "<|reserved_special_token_174|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128183": { + "content": "<|reserved_special_token_175|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128184": { + "content": "<|reserved_special_token_176|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128185": { + "content": "<|reserved_special_token_177|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128186": { + "content": "<|reserved_special_token_178|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128187": { + "content": "<|reserved_special_token_179|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128188": { + "content": "<|reserved_special_token_180|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128189": { + "content": "<|reserved_special_token_181|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128190": { + "content": "<|reserved_special_token_182|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128191": { + "content": "<|reserved_special_token_183|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128192": { + "content": "<|reserved_special_token_184|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128193": { + "content": "<|reserved_special_token_185|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128194": { + "content": "<|reserved_special_token_186|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128195": { + "content": "<|reserved_special_token_187|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128196": { + "content": "<|reserved_special_token_188|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128197": { + "content": "<|reserved_special_token_189|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128198": { + "content": "<|reserved_special_token_190|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128199": { + "content": "<|reserved_special_token_191|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128200": { + "content": "<|reserved_special_token_192|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128201": { + "content": "<|reserved_special_token_193|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128202": { + "content": "<|reserved_special_token_194|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128203": { + "content": "<|reserved_special_token_195|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128204": { + "content": "<|reserved_special_token_196|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128205": { + "content": "<|reserved_special_token_197|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128206": { + "content": "<|reserved_special_token_198|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128207": { + "content": "<|reserved_special_token_199|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128208": { + "content": "<|reserved_special_token_200|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128209": { + "content": "<|reserved_special_token_201|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128210": { + "content": "<|reserved_special_token_202|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128211": { + "content": "<|reserved_special_token_203|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128212": { + "content": "<|reserved_special_token_204|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128213": { + "content": "<|reserved_special_token_205|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128214": { + "content": "<|reserved_special_token_206|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128215": { + "content": "<|reserved_special_token_207|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128216": { + "content": "<|reserved_special_token_208|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128217": { + "content": "<|reserved_special_token_209|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128218": { + "content": "<|reserved_special_token_210|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128219": { + "content": "<|reserved_special_token_211|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128220": { + "content": "<|reserved_special_token_212|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128221": { + "content": "<|reserved_special_token_213|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128222": { + "content": "<|reserved_special_token_214|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128223": { + "content": "<|reserved_special_token_215|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128224": { + "content": "<|reserved_special_token_216|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128225": { + "content": "<|reserved_special_token_217|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128226": { + "content": "<|reserved_special_token_218|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128227": { + "content": "<|reserved_special_token_219|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128228": { + "content": "<|reserved_special_token_220|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128229": { + "content": "<|reserved_special_token_221|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128230": { + "content": "<|reserved_special_token_222|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128231": { + "content": "<|reserved_special_token_223|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128232": { + "content": "<|reserved_special_token_224|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128233": { + "content": "<|reserved_special_token_225|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128234": { + "content": "<|reserved_special_token_226|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128235": { + "content": "<|reserved_special_token_227|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128236": { + "content": "<|reserved_special_token_228|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128237": { + "content": "<|reserved_special_token_229|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128238": { + "content": "<|reserved_special_token_230|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128239": { + "content": "<|reserved_special_token_231|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128240": { + "content": "<|reserved_special_token_232|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128241": { + "content": "<|reserved_special_token_233|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128242": { + "content": "<|reserved_special_token_234|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128243": { + "content": "<|reserved_special_token_235|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128244": { + "content": "<|reserved_special_token_236|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128245": { + "content": "<|reserved_special_token_237|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128246": { + "content": "<|reserved_special_token_238|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128247": { + "content": "<|reserved_special_token_239|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128248": { + "content": "<|reserved_special_token_240|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128249": { + "content": "<|reserved_special_token_241|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128250": { + "content": "<|reserved_special_token_242|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128251": { + "content": "<|reserved_special_token_243|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128252": { + "content": "<|reserved_special_token_244|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128253": { + "content": "<|reserved_special_token_245|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128254": { + "content": "<|reserved_special_token_246|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "128255": { + "content": "<|reserved_special_token_247|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + } + }, + "bos_token": "<|begin_of_text|>", + "chat_template": "{% if not add_generation_prompt is defined %}{% set add_generation_prompt = false %}{% endif %}{% set loop_messages = messages %}{% for message in loop_messages %}{% set content = '<|start_header_id|>' + message['role'] + '<|end_header_id|>\n\n'+ message['content'] | trim + '<|eot_id|>' %}{% if loop.index0 == 0 %}{% set content = bos_token + content %}{% endif %}{{ content }}{% endfor %}{% if add_generation_prompt %}{{ '<|start_header_id|>assistant<|end_header_id|>\n\n' }}{% endif %}", + "clean_up_tokenization_spaces": true, + "eos_token": "<|eot_id|>", + "extra_special_tokens": {}, + "model_input_names": [ + "input_ids", + "attention_mask" + ], + "model_max_length": 131072, + "pad_token": "<|end_of_text|>", + "tokenizer_class": "PreTrainedTokenizer" +} diff --git a/training_args.bin b/training_args.bin new file mode 100644 index 0000000..a88cbd3 --- /dev/null +++ b/training_args.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:e0cffaa6f04a26499dd6171106fabe250a7e89bd923590f4fdeb897bc28870a9 +size 9528