commit 35468fdd9f557bd5367c1bf31a4b6cd239740983 Author: ModelHub XC Date: Wed Jul 8 00:19:13 2026 +0800 初始化项目，由ModelHub XC社区提供模型 Model: kaist-ai/janus-7b Source: Original Platform diff --git a/.gitattributes b/.gitattributes new file mode 100644 index 0000000..a6344aa --- /dev/null +++ b/.gitattributes @@ -0,0 +1,35 @@ +*.7z filter=lfs diff=lfs merge=lfs -text +*.arrow filter=lfs diff=lfs merge=lfs -text +*.bin filter=lfs diff=lfs merge=lfs -text +*.bz2 filter=lfs diff=lfs merge=lfs -text +*.ckpt 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 +*.mlmodel filter=lfs diff=lfs merge=lfs -text +*.model filter=lfs diff=lfs merge=lfs -text +*.msgpack filter=lfs diff=lfs merge=lfs -text +*.npy filter=lfs diff=lfs merge=lfs -text +*.npz 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 +*.pickle filter=lfs diff=lfs merge=lfs -text +*.pkl 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 +*.safetensors filter=lfs diff=lfs merge=lfs -text +saved_model/**/* filter=lfs diff=lfs merge=lfs -text +*.tar.* 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 +*.wasm filter=lfs diff=lfs merge=lfs -text +*.xz filter=lfs diff=lfs merge=lfs -text +*.zip filter=lfs diff=lfs merge=lfs -text +*.zst filter=lfs diff=lfs merge=lfs -text +*tfevents* filter=lfs diff=lfs merge=lfs -text diff --git a/README.md b/README.md new file mode 100644 index 0000000..9bb5cc9 --- /dev/null +++ b/README.md @@ -0,0 +1,129 @@ +--- +base_model: alpindale/Mistral-7B-v0.2-hf +tags: +- axolotl +- generated_from_trainer +model-index: +- name: janus-7b + results: [] +license: apache-2.0 +language: +- en +datasets: +- kaist-ai/Multifaceted-Collection-SFT +library_name: transformers +pipeline_tag: text-generation +--- +## Links for Reference + +- **Homepage: https://lklab.kaist.ac.kr/Janus/** +- **Repository: https://github.com/kaistAI/Janus** +- **Paper: https://arxiv.org/abs/2405.17977** +- **Point of Contact: seongyun@kaist.ac.kr** + +# TL; DR +![image/png](https://cdn-uploads.huggingface.co/production/uploads/6550c4f27bbfce1878f5f280/vrQl8D8FV3vqUJYbPgsiG.png) + +Janus is a model trained using [Mistral-7B-v0.2](https://huggingface.co/mistral-community/Mistral-7B-v0.2) as its base model. Janus has been trained on [Multifaceted Collection](https://huggingface.co/datasets/kaist-ai/Multifaceted-Collection-SFT), a preference dataset containing 196k unique system messages for aligning LLMs to diverse human preferences. Janus not only excels at generating personalized responses that cater to various human preferences but is also adept at producing responses that are generally preferred for being helpful and harmless. + +# Model Details +Janus-7B is a model created by supervised fine-tuning using all 196k entries of the training data from the Multifaceted-Collection. + +## Model Description + +- **Model type:** Language model +- **Language(s) (NLP):** English +- **License:** Apache 2.0 +- **Related Models:** [Janus-DPO-7B](https://huggingface.co/kaist-ai/janus-dpo-7b), [Janus-ORPO-7B](https://huggingface.co/kaist-ai/janus-orpo-7b), [Janus-RM-7B](https://huggingface.co/kaist-ai/janus-rm-7b) +- **Training Datasets**: [Multifaceted-Collection-SFT](https://huggingface.co/datasets/kaist-ai/Multifaceted-Collection-SFT) +- **Resources for more information:** + - [Research paper](https://arxiv.org/abs/2405.17977) + - [GitHub Repo](https://github.com/kaistAI/Janus) + +# Usage +Janus is a model generalized for various system messages, allowing users to control the model's response by inputting the desired system message. The input prompt format is as follows: +``` +[INST]{system_message}\n{instruction}[/INST] +``` +Additionally, an example of the inference code applying this is as follows: +```python +from transformers import AutoTokenizer, AutoModelForCausalLM +import torch + +model_name = "kaist-ai/janus-7b" +device = "cuda:0" + +# Load the model and tokenizer +tokenizer = AutoTokenizer.from_pretrained(model_name) + +dtype = "float16" +if torch.cuda.is_bf16_supported(): + dtype = "bfloat16" + +model = AutoModelForCausalLM.from_pretrained(model_name, torch_dtype=getattr(torch, dtype)) +model.eval() +model.to(device) + +# Prepare inputs +system = "As a financial news headline writer with a flair for the dramatic, you have taken on the role of crafting compelling headlines about the integration of AI into the financial sector. Your expertise allows you to weave industry-specific terminology seamlessly into each headline, striking a balance between capturing attention and providing meaningful insights into the transformative benefits of AI in finance. With each headline, you focus on elucidating the key advantages AI brings to financial operations, making complex information accessible and immediately impactful. While your headlines are designed to engage and inform an audience of finance and technology professionals, you navigate the fine line of excitement and accuracy with care, ensuring that the promises made are grounded in reality, thus avoiding any form of sensationalism. Your mission is to distill the essence of AI's impact on finance into a single, powerful line that speaks volumes to the informed reader." +prompt = "Write a headline for an article about the benefits of using AI in the finance sector." + +def apply_template_mistral_instruct(system_message, content): + prompt = f"{system_message}\n{content}".strip() + return f"[INST] {prompt} [/INST] " + +input_str = apply_template_mistral_instruct(system, prompt) +input_ids = tokenizer.encode(input_str, return_tensors="pt") +print(input_str) + +model_inputs = input_ids.to(device) + +# Generate text +output_ids = model.generate(model_inputs, max_new_tokens=1024) +decoded = tokenizer.batch_decode(output_ids, skip_special_tokens=True) +print(decoded[0][len(input_str):]) +# Revolutionary Trends: How AI Is Redefining Efficiency and Accuracy in the Financial Realm +``` +To train Janus and evaluate the responses it generates, please refer to the [GitHub Repo](https://github.com/kaistAI/Janus). +Additionally, refer to the [Multifaceted Bench](https://huggingface.co/datasets/kaist-ai/Multifaceted-Bench), which evaluates how well LLM generates personalized responses. +# Training Details +## Training hyperparameters + +The following hyperparameters were used during training: +- learning_rate: 5e-06 +- train_batch_size: 2 +- eval_batch_size: 2 +- seed: 42 +- distributed_type: multi-GPU +- num_devices: 4 +- gradient_accumulation_steps: 4 +- total_train_batch_size: 32 +- total_eval_batch_size: 8 +- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 +- lr_scheduler_type: cosine +- lr_scheduler_warmup_steps: 10 +- num_epochs: 4 + +## Framework versions + +- Transformers 4.40.0.dev0 +- Pytorch 2.2.2 +- Datasets 2.18.0 +- Tokenizers 0.15.0 + +# Citation + +If you find the following model helpful, please consider citing our paper! + +**BibTeX:** + +```bibtex +@misc{lee2024aligning, + title={Aligning to Thousands of Preferences via System Message Generalization}, + author={Seongyun Lee and Sue Hyun Park and Seungone Kim and Minjoon Seo}, + year={2024}, + eprint={2405.17977}, + archivePrefix={arXiv}, + primaryClass={cs.CL} +} +``` \ No newline at end of file diff --git a/config.json b/config.json new file mode 100644 index 0000000..8bab8d3 --- /dev/null +++ b/config.json @@ -0,0 +1,26 @@ +{ + "_name_or_path": "alpindale/Mistral-7B-v0.2-hf", + "architectures": [ + "MistralForCausalLM" + ], + "attention_dropout": 0.0, + "bos_token_id": 1, + "eos_token_id": 2, + "hidden_act": "silu", + "hidden_size": 4096, + "initializer_range": 0.02, + "intermediate_size": 14336, + "max_position_embeddings": 32768, + "model_type": "mistral", + "num_attention_heads": 32, + "num_hidden_layers": 32, + "num_key_value_heads": 8, + "rms_norm_eps": 1e-05, + "rope_theta": 1000000.0, + "sliding_window": null, + "tie_word_embeddings": false, + "torch_dtype": "bfloat16", + "transformers_version": "4.40.0.dev0", + "use_cache": true, + "vocab_size": 32000 +} diff --git a/generation_config.json b/generation_config.json new file mode 100644 index 0000000..16dd90a --- /dev/null +++ b/generation_config.json @@ -0,0 +1,7 @@ +{ + "_from_model_config": true, + "bos_token_id": 1, + "do_sample": true, + "eos_token_id": 2, + "transformers_version": "4.40.0.dev0" +} diff --git a/last-checkpoint/config.json b/last-checkpoint/config.json new file mode 100644 index 0000000..06d325a --- /dev/null +++ b/last-checkpoint/config.json @@ -0,0 +1,26 @@ +{ + "_name_or_path": "alpindale/Mistral-7B-v0.2-hf", + "architectures": [ + "MistralForCausalLM" + ], + "attention_dropout": 0.0, + "bos_token_id": 1, + "eos_token_id": 2, + "hidden_act": "silu", + "hidden_size": 4096, + "initializer_range": 0.02, + "intermediate_size": 14336, + "max_position_embeddings": 32768, + "model_type": "mistral", + "num_attention_heads": 32, + "num_hidden_layers": 32, + "num_key_value_heads": 8, + "rms_norm_eps": 1e-05, + "rope_theta": 1000000.0, + "sliding_window": null, + "tie_word_embeddings": false, + "torch_dtype": "bfloat16", + "transformers_version": "4.40.0.dev0", + "use_cache": false, + "vocab_size": 32000 +} diff --git a/last-checkpoint/generation_config.json b/last-checkpoint/generation_config.json new file mode 100644 index 0000000..16dd90a --- /dev/null +++ b/last-checkpoint/generation_config.json @@ -0,0 +1,7 @@ +{ + "_from_model_config": true, + "bos_token_id": 1, + "do_sample": true, + "eos_token_id": 2, + "transformers_version": "4.40.0.dev0" +} diff --git a/last-checkpoint/model-00001-of-00003.safetensors b/last-checkpoint/model-00001-of-00003.safetensors new file mode 100644 index 0000000..7a7a35a --- /dev/null +++ b/last-checkpoint/model-00001-of-00003.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:1918667d6c8ea19c91894aed25da7c6c3a707bbca0e11dcfba9225a78109ed5c +size 4943162336 diff --git a/last-checkpoint/model-00002-of-00003.safetensors b/last-checkpoint/model-00002-of-00003.safetensors new file mode 100644 index 0000000..9178ff6 --- /dev/null +++ b/last-checkpoint/model-00002-of-00003.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:7abfc057a799175c2b8cb179b3f28455cb1f2ded40c9ffab62d44b51b4b54db7 +size 4999819336 diff --git a/last-checkpoint/model-00003-of-00003.safetensors b/last-checkpoint/model-00003-of-00003.safetensors new file mode 100644 index 0000000..0c7ecde --- /dev/null +++ b/last-checkpoint/model-00003-of-00003.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:82c09b650cdaaf22e170b97b68130c35b65aa74418387c100a25980f525e1835 +size 4540516344 diff --git a/last-checkpoint/model.safetensors.index.json b/last-checkpoint/model.safetensors.index.json new file mode 100644 index 0000000..b349bc0 --- /dev/null +++ b/last-checkpoint/model.safetensors.index.json @@ -0,0 +1,298 @@ +{ + "metadata": { + "total_size": 14483464192 + }, + "weight_map": { + "lm_head.weight": "model-00003-of-00003.safetensors", + "model.embed_tokens.weight": "model-00001-of-00003.safetensors", + "model.layers.0.input_layernorm.weight": "model-00001-of-00003.safetensors", + "model.layers.0.mlp.down_proj.weight": "model-00001-of-00003.safetensors", + "model.layers.0.mlp.gate_proj.weight": "model-00001-of-00003.safetensors", + "model.layers.0.mlp.up_proj.weight": "model-00001-of-00003.safetensors", + "model.layers.0.post_attention_layernorm.weight": "model-00001-of-00003.safetensors", + "model.layers.0.self_attn.k_proj.weight": "model-00001-of-00003.safetensors", + "model.layers.0.self_attn.o_proj.weight": "model-00001-of-00003.safetensors", + "model.layers.0.self_attn.q_proj.weight": "model-00001-of-00003.safetensors", + "model.layers.0.self_attn.v_proj.weight": "model-00001-of-00003.safetensors", + "model.layers.1.input_layernorm.weight": "model-00001-of-00003.safetensors", + "model.layers.1.mlp.down_proj.weight": "model-00001-of-00003.safetensors", + "model.layers.1.mlp.gate_proj.weight": "model-00001-of-00003.safetensors", + "model.layers.1.mlp.up_proj.weight": "model-00001-of-00003.safetensors", + "model.layers.1.post_attention_layernorm.weight": "model-00001-of-00003.safetensors", + "model.layers.1.self_attn.k_proj.weight": "model-00001-of-00003.safetensors", + "model.layers.1.self_attn.o_proj.weight": "model-00001-of-00003.safetensors", + "model.layers.1.self_attn.q_proj.weight": "model-00001-of-00003.safetensors", + "model.layers.1.self_attn.v_proj.weight": "model-00001-of-00003.safetensors", + "model.layers.10.input_layernorm.weight": "model-00002-of-00003.safetensors", + "model.layers.10.mlp.down_proj.weight": "model-00002-of-00003.safetensors", + "model.layers.10.mlp.gate_proj.weight": "model-00001-of-00003.safetensors", + "model.layers.10.mlp.up_proj.weight": "model-00001-of-00003.safetensors", + "model.layers.10.post_attention_layernorm.weight": "model-00002-of-00003.safetensors", + "model.layers.10.self_attn.k_proj.weight": "model-00001-of-00003.safetensors", + "model.layers.10.self_attn.o_proj.weight": "model-00001-of-00003.safetensors", + "model.layers.10.self_attn.q_proj.weight": "model-00001-of-00003.safetensors", + "model.layers.10.self_attn.v_proj.weight": "model-00001-of-00003.safetensors", + "model.layers.11.input_layernorm.weight": "model-00002-of-00003.safetensors", + "model.layers.11.mlp.down_proj.weight": "model-00002-of-00003.safetensors", + "model.layers.11.mlp.gate_proj.weight": "model-00002-of-00003.safetensors", + "model.layers.11.mlp.up_proj.weight": "model-00002-of-00003.safetensors", + "model.layers.11.post_attention_layernorm.weight": "model-00002-of-00003.safetensors", + "model.layers.11.self_attn.k_proj.weight": "model-00002-of-00003.safetensors", + "model.layers.11.self_attn.o_proj.weight": "model-00002-of-00003.safetensors", + "model.layers.11.self_attn.q_proj.weight": "model-00002-of-00003.safetensors", + "model.layers.11.self_attn.v_proj.weight": "model-00002-of-00003.safetensors", + "model.layers.12.input_layernorm.weight": "model-00002-of-00003.safetensors", + "model.layers.12.mlp.down_proj.weight": "model-00002-of-00003.safetensors", + "model.layers.12.mlp.gate_proj.weight": "model-00002-of-00003.safetensors", + "model.layers.12.mlp.up_proj.weight": "model-00002-of-00003.safetensors", + "model.layers.12.post_attention_layernorm.weight": "model-00002-of-00003.safetensors", + "model.layers.12.self_attn.k_proj.weight": "model-00002-of-00003.safetensors", + "model.layers.12.self_attn.o_proj.weight": "model-00002-of-00003.safetensors", + "model.layers.12.self_attn.q_proj.weight": "model-00002-of-00003.safetensors", + "model.layers.12.self_attn.v_proj.weight": "model-00002-of-00003.safetensors", + "model.layers.13.input_layernorm.weight": "model-00002-of-00003.safetensors", + "model.layers.13.mlp.down_proj.weight": "model-00002-of-00003.safetensors", + "model.layers.13.mlp.gate_proj.weight": "model-00002-of-00003.safetensors", + "model.layers.13.mlp.up_proj.weight": "model-00002-of-00003.safetensors", + "model.layers.13.post_attention_layernorm.weight": "model-00002-of-00003.safetensors", + "model.layers.13.self_attn.k_proj.weight": "model-00002-of-00003.safetensors", + "model.layers.13.self_attn.o_proj.weight": "model-00002-of-00003.safetensors", + "model.layers.13.self_attn.q_proj.weight": "model-00002-of-00003.safetensors", + "model.layers.13.self_attn.v_proj.weight": "model-00002-of-00003.safetensors", + "model.layers.14.input_layernorm.weight": "model-00002-of-00003.safetensors", + "model.layers.14.mlp.down_proj.weight": "model-00002-of-00003.safetensors", + "model.layers.14.mlp.gate_proj.weight": "model-00002-of-00003.safetensors", + "model.layers.14.mlp.up_proj.weight": "model-00002-of-00003.safetensors", + "model.layers.14.post_attention_layernorm.weight": "model-00002-of-00003.safetensors", + "model.layers.14.self_attn.k_proj.weight": "model-00002-of-00003.safetensors", + "model.layers.14.self_attn.o_proj.weight": "model-00002-of-00003.safetensors", + "model.layers.14.self_attn.q_proj.weight": "model-00002-of-00003.safetensors", + "model.layers.14.self_attn.v_proj.weight": "model-00002-of-00003.safetensors", + "model.layers.15.input_layernorm.weight": "model-00002-of-00003.safetensors", + "model.layers.15.mlp.down_proj.weight": "model-00002-of-00003.safetensors", + "model.layers.15.mlp.gate_proj.weight": "model-00002-of-00003.safetensors", + "model.layers.15.mlp.up_proj.weight": "model-00002-of-00003.safetensors", + "model.layers.15.post_attention_layernorm.weight": "model-00002-of-00003.safetensors", + "model.layers.15.self_attn.k_proj.weight": "model-00002-of-00003.safetensors", + "model.layers.15.self_attn.o_proj.weight": "model-00002-of-00003.safetensors", + "model.layers.15.self_attn.q_proj.weight": "model-00002-of-00003.safetensors", + "model.layers.15.self_attn.v_proj.weight": "model-00002-of-00003.safetensors", + "model.layers.16.input_layernorm.weight": "model-00002-of-00003.safetensors", + "model.layers.16.mlp.down_proj.weight": "model-00002-of-00003.safetensors", + "model.layers.16.mlp.gate_proj.weight": "model-00002-of-00003.safetensors", + "model.layers.16.mlp.up_proj.weight": "model-00002-of-00003.safetensors", + "model.layers.16.post_attention_layernorm.weight": "model-00002-of-00003.safetensors", + "model.layers.16.self_attn.k_proj.weight": "model-00002-of-00003.safetensors", + "model.layers.16.self_attn.o_proj.weight": "model-00002-of-00003.safetensors", + "model.layers.16.self_attn.q_proj.weight": "model-00002-of-00003.safetensors", + "model.layers.16.self_attn.v_proj.weight": "model-00002-of-00003.safetensors", + "model.layers.17.input_layernorm.weight": "model-00002-of-00003.safetensors", + "model.layers.17.mlp.down_proj.weight": "model-00002-of-00003.safetensors", + "model.layers.17.mlp.gate_proj.weight": "model-00002-of-00003.safetensors", + "model.layers.17.mlp.up_proj.weight": "model-00002-of-00003.safetensors", + "model.layers.17.post_attention_layernorm.weight": "model-00002-of-00003.safetensors", + "model.layers.17.self_attn.k_proj.weight": "model-00002-of-00003.safetensors", + "model.layers.17.self_attn.o_proj.weight": "model-00002-of-00003.safetensors", + "model.layers.17.self_attn.q_proj.weight": "model-00002-of-00003.safetensors", + "model.layers.17.self_attn.v_proj.weight": "model-00002-of-00003.safetensors", + "model.layers.18.input_layernorm.weight": "model-00002-of-00003.safetensors", + "model.layers.18.mlp.down_proj.weight": "model-00002-of-00003.safetensors", + "model.layers.18.mlp.gate_proj.weight": "model-00002-of-00003.safetensors", + "model.layers.18.mlp.up_proj.weight": "model-00002-of-00003.safetensors", + "model.layers.18.post_attention_layernorm.weight": "model-00002-of-00003.safetensors", + "model.layers.18.self_attn.k_proj.weight": "model-00002-of-00003.safetensors", + "model.layers.18.self_attn.o_proj.weight": "model-00002-of-00003.safetensors", + "model.layers.18.self_attn.q_proj.weight": "model-00002-of-00003.safetensors", + "model.layers.18.self_attn.v_proj.weight": "model-00002-of-00003.safetensors", + "model.layers.19.input_layernorm.weight": "model-00002-of-00003.safetensors", + "model.layers.19.mlp.down_proj.weight": "model-00002-of-00003.safetensors", + "model.layers.19.mlp.gate_proj.weight": "model-00002-of-00003.safetensors", + "model.layers.19.mlp.up_proj.weight": "model-00002-of-00003.safetensors", + "model.layers.19.post_attention_layernorm.weight": "model-00002-of-00003.safetensors", + "model.layers.19.self_attn.k_proj.weight": "model-00002-of-00003.safetensors", + "model.layers.19.self_attn.o_proj.weight": "model-00002-of-00003.safetensors", + "model.layers.19.self_attn.q_proj.weight": "model-00002-of-00003.safetensors", + "model.layers.19.self_attn.v_proj.weight": "model-00002-of-00003.safetensors", + "model.layers.2.input_layernorm.weight": "model-00001-of-00003.safetensors", + "model.layers.2.mlp.down_proj.weight": "model-00001-of-00003.safetensors", + "model.layers.2.mlp.gate_proj.weight": "model-00001-of-00003.safetensors", + "model.layers.2.mlp.up_proj.weight": "model-00001-of-00003.safetensors", + "model.layers.2.post_attention_layernorm.weight": "model-00001-of-00003.safetensors", + "model.layers.2.self_attn.k_proj.weight": "model-00001-of-00003.safetensors", + "model.layers.2.self_attn.o_proj.weight": "model-00001-of-00003.safetensors", + "model.layers.2.self_attn.q_proj.weight": "model-00001-of-00003.safetensors", + "model.layers.2.self_attn.v_proj.weight": "model-00001-of-00003.safetensors", + "model.layers.20.input_layernorm.weight": "model-00002-of-00003.safetensors", + "model.layers.20.mlp.down_proj.weight": "model-00002-of-00003.safetensors", + "model.layers.20.mlp.gate_proj.weight": "model-00002-of-00003.safetensors", + "model.layers.20.mlp.up_proj.weight": "model-00002-of-00003.safetensors", + "model.layers.20.post_attention_layernorm.weight": "model-00002-of-00003.safetensors", + "model.layers.20.self_attn.k_proj.weight": "model-00002-of-00003.safetensors", + "model.layers.20.self_attn.o_proj.weight": "model-00002-of-00003.safetensors", + "model.layers.20.self_attn.q_proj.weight": "model-00002-of-00003.safetensors", + "model.layers.20.self_attn.v_proj.weight": "model-00002-of-00003.safetensors", + "model.layers.21.input_layernorm.weight": "model-00002-of-00003.safetensors", + "model.layers.21.mlp.down_proj.weight": "model-00002-of-00003.safetensors", + "model.layers.21.mlp.gate_proj.weight": "model-00002-of-00003.safetensors", + "model.layers.21.mlp.up_proj.weight": "model-00002-of-00003.safetensors", + "model.layers.21.post_attention_layernorm.weight": "model-00002-of-00003.safetensors", + "model.layers.21.self_attn.k_proj.weight": "model-00002-of-00003.safetensors", + "model.layers.21.self_attn.o_proj.weight": "model-00002-of-00003.safetensors", + "model.layers.21.self_attn.q_proj.weight": "model-00002-of-00003.safetensors", + "model.layers.21.self_attn.v_proj.weight": "model-00002-of-00003.safetensors", + "model.layers.22.input_layernorm.weight": "model-00003-of-00003.safetensors", + "model.layers.22.mlp.down_proj.weight": "model-00003-of-00003.safetensors", + "model.layers.22.mlp.gate_proj.weight": "model-00003-of-00003.safetensors", + "model.layers.22.mlp.up_proj.weight": "model-00003-of-00003.safetensors", + "model.layers.22.post_attention_layernorm.weight": "model-00003-of-00003.safetensors", + "model.layers.22.self_attn.k_proj.weight": "model-00002-of-00003.safetensors", + "model.layers.22.self_attn.o_proj.weight": "model-00002-of-00003.safetensors", + "model.layers.22.self_attn.q_proj.weight": "model-00002-of-00003.safetensors", + "model.layers.22.self_attn.v_proj.weight": "model-00002-of-00003.safetensors", + "model.layers.23.input_layernorm.weight": "model-00003-of-00003.safetensors", + "model.layers.23.mlp.down_proj.weight": "model-00003-of-00003.safetensors", + "model.layers.23.mlp.gate_proj.weight": "model-00003-of-00003.safetensors", + "model.layers.23.mlp.up_proj.weight": "model-00003-of-00003.safetensors", + "model.layers.23.post_attention_layernorm.weight": "model-00003-of-00003.safetensors", + "model.layers.23.self_attn.k_proj.weight": "model-00003-of-00003.safetensors", + "model.layers.23.self_attn.o_proj.weight": "model-00003-of-00003.safetensors", + "model.layers.23.self_attn.q_proj.weight": "model-00003-of-00003.safetensors", + "model.layers.23.self_attn.v_proj.weight": "model-00003-of-00003.safetensors", + "model.layers.24.input_layernorm.weight": "model-00003-of-00003.safetensors", + "model.layers.24.mlp.down_proj.weight": "model-00003-of-00003.safetensors", + "model.layers.24.mlp.gate_proj.weight": "model-00003-of-00003.safetensors", + "model.layers.24.mlp.up_proj.weight": "model-00003-of-00003.safetensors", + "model.layers.24.post_attention_layernorm.weight": "model-00003-of-00003.safetensors", + "model.layers.24.self_attn.k_proj.weight": "model-00003-of-00003.safetensors", + "model.layers.24.self_attn.o_proj.weight": "model-00003-of-00003.safetensors", + "model.layers.24.self_attn.q_proj.weight": "model-00003-of-00003.safetensors", + "model.layers.24.self_attn.v_proj.weight": "model-00003-of-00003.safetensors", + "model.layers.25.input_layernorm.weight": "model-00003-of-00003.safetensors", + "model.layers.25.mlp.down_proj.weight": "model-00003-of-00003.safetensors", + "model.layers.25.mlp.gate_proj.weight": "model-00003-of-00003.safetensors", + "model.layers.25.mlp.up_proj.weight": "model-00003-of-00003.safetensors", + "model.layers.25.post_attention_layernorm.weight": "model-00003-of-00003.safetensors", + "model.layers.25.self_attn.k_proj.weight": "model-00003-of-00003.safetensors", + "model.layers.25.self_attn.o_proj.weight": "model-00003-of-00003.safetensors", + "model.layers.25.self_attn.q_proj.weight": "model-00003-of-00003.safetensors", + "model.layers.25.self_attn.v_proj.weight": "model-00003-of-00003.safetensors", + "model.layers.26.input_layernorm.weight": "model-00003-of-00003.safetensors", + "model.layers.26.mlp.down_proj.weight": "model-00003-of-00003.safetensors", + "model.layers.26.mlp.gate_proj.weight": "model-00003-of-00003.safetensors", + "model.layers.26.mlp.up_proj.weight": "model-00003-of-00003.safetensors", + "model.layers.26.post_attention_layernorm.weight": "model-00003-of-00003.safetensors", + "model.layers.26.self_attn.k_proj.weight": "model-00003-of-00003.safetensors", + "model.layers.26.self_attn.o_proj.weight": "model-00003-of-00003.safetensors", + "model.layers.26.self_attn.q_proj.weight": "model-00003-of-00003.safetensors", + "model.layers.26.self_attn.v_proj.weight": "model-00003-of-00003.safetensors", + "model.layers.27.input_layernorm.weight": "model-00003-of-00003.safetensors", + "model.layers.27.mlp.down_proj.weight": "model-00003-of-00003.safetensors", + "model.layers.27.mlp.gate_proj.weight": "model-00003-of-00003.safetensors", + "model.layers.27.mlp.up_proj.weight": "model-00003-of-00003.safetensors", + "model.layers.27.post_attention_layernorm.weight": "model-00003-of-00003.safetensors", + "model.layers.27.self_attn.k_proj.weight": "model-00003-of-00003.safetensors", + "model.layers.27.self_attn.o_proj.weight": "model-00003-of-00003.safetensors", + "model.layers.27.self_attn.q_proj.weight": "model-00003-of-00003.safetensors", + "model.layers.27.self_attn.v_proj.weight": "model-00003-of-00003.safetensors", + "model.layers.28.input_layernorm.weight": "model-00003-of-00003.safetensors", + "model.layers.28.mlp.down_proj.weight": "model-00003-of-00003.safetensors", + "model.layers.28.mlp.gate_proj.weight": "model-00003-of-00003.safetensors", + "model.layers.28.mlp.up_proj.weight": "model-00003-of-00003.safetensors", + "model.layers.28.post_attention_layernorm.weight": "model-00003-of-00003.safetensors", + "model.layers.28.self_attn.k_proj.weight": "model-00003-of-00003.safetensors", + "model.layers.28.self_attn.o_proj.weight": "model-00003-of-00003.safetensors", + "model.layers.28.self_attn.q_proj.weight": "model-00003-of-00003.safetensors", + "model.layers.28.self_attn.v_proj.weight": "model-00003-of-00003.safetensors", + "model.layers.29.input_layernorm.weight": "model-00003-of-00003.safetensors", + "model.layers.29.mlp.down_proj.weight": "model-00003-of-00003.safetensors", + "model.layers.29.mlp.gate_proj.weight": "model-00003-of-00003.safetensors", + "model.layers.29.mlp.up_proj.weight": "model-00003-of-00003.safetensors", + "model.layers.29.post_attention_layernorm.weight": "model-00003-of-00003.safetensors", + "model.layers.29.self_attn.k_proj.weight": "model-00003-of-00003.safetensors", + "model.layers.29.self_attn.o_proj.weight": "model-00003-of-00003.safetensors", + "model.layers.29.self_attn.q_proj.weight": "model-00003-of-00003.safetensors", + "model.layers.29.self_attn.v_proj.weight": "model-00003-of-00003.safetensors", + "model.layers.3.input_layernorm.weight": "model-00001-of-00003.safetensors", + "model.layers.3.mlp.down_proj.weight": "model-00001-of-00003.safetensors", + "model.layers.3.mlp.gate_proj.weight": "model-00001-of-00003.safetensors", + "model.layers.3.mlp.up_proj.weight": "model-00001-of-00003.safetensors", + "model.layers.3.post_attention_layernorm.weight": "model-00001-of-00003.safetensors", + "model.layers.3.self_attn.k_proj.weight": "model-00001-of-00003.safetensors", + "model.layers.3.self_attn.o_proj.weight": "model-00001-of-00003.safetensors", + "model.layers.3.self_attn.q_proj.weight": "model-00001-of-00003.safetensors", + "model.layers.3.self_attn.v_proj.weight": "model-00001-of-00003.safetensors", + "model.layers.30.input_layernorm.weight": "model-00003-of-00003.safetensors", + "model.layers.30.mlp.down_proj.weight": "model-00003-of-00003.safetensors", + "model.layers.30.mlp.gate_proj.weight": "model-00003-of-00003.safetensors", + "model.layers.30.mlp.up_proj.weight": "model-00003-of-00003.safetensors", + "model.layers.30.post_attention_layernorm.weight": "model-00003-of-00003.safetensors", + "model.layers.30.self_attn.k_proj.weight": "model-00003-of-00003.safetensors", + "model.layers.30.self_attn.o_proj.weight": "model-00003-of-00003.safetensors", + "model.layers.30.self_attn.q_proj.weight": "model-00003-of-00003.safetensors", + "model.layers.30.self_attn.v_proj.weight": "model-00003-of-00003.safetensors", + "model.layers.31.input_layernorm.weight": "model-00003-of-00003.safetensors", + "model.layers.31.mlp.down_proj.weight": "model-00003-of-00003.safetensors", + "model.layers.31.mlp.gate_proj.weight": "model-00003-of-00003.safetensors", + "model.layers.31.mlp.up_proj.weight": "model-00003-of-00003.safetensors", + "model.layers.31.post_attention_layernorm.weight": "model-00003-of-00003.safetensors", + "model.layers.31.self_attn.k_proj.weight": "model-00003-of-00003.safetensors", + "model.layers.31.self_attn.o_proj.weight": "model-00003-of-00003.safetensors", + "model.layers.31.self_attn.q_proj.weight": "model-00003-of-00003.safetensors", + "model.layers.31.self_attn.v_proj.weight": "model-00003-of-00003.safetensors", + "model.layers.4.input_layernorm.weight": "model-00001-of-00003.safetensors", + "model.layers.4.mlp.down_proj.weight": "model-00001-of-00003.safetensors", + "model.layers.4.mlp.gate_proj.weight": "model-00001-of-00003.safetensors", + "model.layers.4.mlp.up_proj.weight": "model-00001-of-00003.safetensors", + "model.layers.4.post_attention_layernorm.weight": "model-00001-of-00003.safetensors", + "model.layers.4.self_attn.k_proj.weight": "model-00001-of-00003.safetensors", + "model.layers.4.self_attn.o_proj.weight": "model-00001-of-00003.safetensors", + "model.layers.4.self_attn.q_proj.weight": "model-00001-of-00003.safetensors", + "model.layers.4.self_attn.v_proj.weight": "model-00001-of-00003.safetensors", + "model.layers.5.input_layernorm.weight": "model-00001-of-00003.safetensors", + "model.layers.5.mlp.down_proj.weight": "model-00001-of-00003.safetensors", + "model.layers.5.mlp.gate_proj.weight": "model-00001-of-00003.safetensors", + "model.layers.5.mlp.up_proj.weight": "model-00001-of-00003.safetensors", + "model.layers.5.post_attention_layernorm.weight": "model-00001-of-00003.safetensors", + "model.layers.5.self_attn.k_proj.weight": "model-00001-of-00003.safetensors", + "model.layers.5.self_attn.o_proj.weight": "model-00001-of-00003.safetensors", + "model.layers.5.self_attn.q_proj.weight": "model-00001-of-00003.safetensors", + "model.layers.5.self_attn.v_proj.weight": "model-00001-of-00003.safetensors", + "model.layers.6.input_layernorm.weight": "model-00001-of-00003.safetensors", + "model.layers.6.mlp.down_proj.weight": "model-00001-of-00003.safetensors", + "model.layers.6.mlp.gate_proj.weight": "model-00001-of-00003.safetensors", + "model.layers.6.mlp.up_proj.weight": "model-00001-of-00003.safetensors", + "model.layers.6.post_attention_layernorm.weight": "model-00001-of-00003.safetensors", + "model.layers.6.self_attn.k_proj.weight": "model-00001-of-00003.safetensors", + "model.layers.6.self_attn.o_proj.weight": "model-00001-of-00003.safetensors", + "model.layers.6.self_attn.q_proj.weight": "model-00001-of-00003.safetensors", + "model.layers.6.self_attn.v_proj.weight": "model-00001-of-00003.safetensors", + "model.layers.7.input_layernorm.weight": "model-00001-of-00003.safetensors", + "model.layers.7.mlp.down_proj.weight": "model-00001-of-00003.safetensors", + "model.layers.7.mlp.gate_proj.weight": "model-00001-of-00003.safetensors", + "model.layers.7.mlp.up_proj.weight": "model-00001-of-00003.safetensors", + "model.layers.7.post_attention_layernorm.weight": "model-00001-of-00003.safetensors", + "model.layers.7.self_attn.k_proj.weight": "model-00001-of-00003.safetensors", + "model.layers.7.self_attn.o_proj.weight": "model-00001-of-00003.safetensors", + "model.layers.7.self_attn.q_proj.weight": "model-00001-of-00003.safetensors", + "model.layers.7.self_attn.v_proj.weight": "model-00001-of-00003.safetensors", + "model.layers.8.input_layernorm.weight": "model-00001-of-00003.safetensors", + "model.layers.8.mlp.down_proj.weight": "model-00001-of-00003.safetensors", + "model.layers.8.mlp.gate_proj.weight": "model-00001-of-00003.safetensors", + "model.layers.8.mlp.up_proj.weight": "model-00001-of-00003.safetensors", + "model.layers.8.post_attention_layernorm.weight": "model-00001-of-00003.safetensors", + "model.layers.8.self_attn.k_proj.weight": "model-00001-of-00003.safetensors", + "model.layers.8.self_attn.o_proj.weight": "model-00001-of-00003.safetensors", + "model.layers.8.self_attn.q_proj.weight": "model-00001-of-00003.safetensors", + "model.layers.8.self_attn.v_proj.weight": "model-00001-of-00003.safetensors", + "model.layers.9.input_layernorm.weight": "model-00001-of-00003.safetensors", + "model.layers.9.mlp.down_proj.weight": "model-00001-of-00003.safetensors", + "model.layers.9.mlp.gate_proj.weight": "model-00001-of-00003.safetensors", + "model.layers.9.mlp.up_proj.weight": "model-00001-of-00003.safetensors", + "model.layers.9.post_attention_layernorm.weight": "model-00001-of-00003.safetensors", + "model.layers.9.self_attn.k_proj.weight": "model-00001-of-00003.safetensors", + "model.layers.9.self_attn.o_proj.weight": "model-00001-of-00003.safetensors", + "model.layers.9.self_attn.q_proj.weight": "model-00001-of-00003.safetensors", + "model.layers.9.self_attn.v_proj.weight": "model-00001-of-00003.safetensors", + "model.norm.weight": "model-00003-of-00003.safetensors" + } +} diff --git a/last-checkpoint/optimizer.pt b/last-checkpoint/optimizer.pt new file mode 100644 index 0000000..41779f5 --- /dev/null +++ b/last-checkpoint/optimizer.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:bc105a8b21b3310ff846e651e39c8c88a9e538c13f944e35129c3878ab8049de +size 14512103560 diff --git a/last-checkpoint/rng_state_0.pth b/last-checkpoint/rng_state_0.pth new file mode 100644 index 0000000..89e481f --- /dev/null +++ b/last-checkpoint/rng_state_0.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:7f2439da621f14c22b4f733e91bfc9de6b506d28d7b8d6f3eaca2e0b4f24c078 +size 15024 diff --git a/last-checkpoint/rng_state_1.pth b/last-checkpoint/rng_state_1.pth new file mode 100644 index 0000000..18841fb --- /dev/null +++ b/last-checkpoint/rng_state_1.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:c9e3fb386557f376b8946af5b8c91f9418f374dddb2ad9da4868b1ef16778c32 +size 15024 diff --git a/last-checkpoint/rng_state_2.pth b/last-checkpoint/rng_state_2.pth new file mode 100644 index 0000000..6737644 --- /dev/null +++ b/last-checkpoint/rng_state_2.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:dc7774d06045635bece9e960378fdc6913bf7bbbc903444cc570d1ca6ac25645 +size 15024 diff --git a/last-checkpoint/rng_state_3.pth b/last-checkpoint/rng_state_3.pth new file mode 100644 index 0000000..8b55550 --- /dev/null +++ b/last-checkpoint/rng_state_3.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:d98c54a80a914fecf43d06ea81432499f46e70664f1d04651bf339163e30fa9e +size 15024 diff --git a/last-checkpoint/scheduler.pt b/last-checkpoint/scheduler.pt new file mode 100644 index 0000000..a630323 --- /dev/null +++ b/last-checkpoint/scheduler.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:c038798f1d87908096bd3ccaf07dbb30becf03a6abd2d487e4aa0f6c591d7c66 +size 1064 diff --git a/last-checkpoint/special_tokens_map.json b/last-checkpoint/special_tokens_map.json new file mode 100644 index 0000000..72ecfee --- /dev/null +++ b/last-checkpoint/special_tokens_map.json @@ -0,0 +1,24 @@ +{ + "bos_token": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false + }, + "eos_token": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false + }, + "pad_token": "", + "unk_token": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false + } +} diff --git a/last-checkpoint/tokenizer.model b/last-checkpoint/tokenizer.model new file mode 100644 index 0000000..8b443ef --- /dev/null +++ b/last-checkpoint/tokenizer.model @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:dadfd56d766715c61d2ef780a525ab43b8e6da4de6865bda3d95fdef5e134055 +size 493443 diff --git a/last-checkpoint/tokenizer_config.json b/last-checkpoint/tokenizer_config.json new file mode 100644 index 0000000..c25b6d0 --- /dev/null +++ b/last-checkpoint/tokenizer_config.json @@ -0,0 +1,43 @@ +{ + "add_bos_token": true, + "add_eos_token": false, + "add_prefix_space": true, + "added_tokens_decoder": { + "0": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "1": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "2": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + } + }, + "bos_token": "", + "clean_up_tokenization_spaces": false, + "eos_token": "", + "legacy": true, + "model_max_length": 1000000000000000019884624838656, + "pad_token": "", + "sp_model_kwargs": {}, + "spaces_between_special_tokens": false, + "tokenizer_class": "LlamaTokenizer", + "unk_token": "", + "use_default_system_prompt": false, + "use_fast": true +} diff --git a/last-checkpoint/trainer_state.json b/last-checkpoint/trainer_state.json new file mode 100644 index 0000000..42818ca --- /dev/null +++ b/last-checkpoint/trainer_state.json @@ -0,0 +1,21049 @@ +{ + "best_metric": null, + "best_model_checkpoint": null, + "epoch": 3.952428476380572, + "eval_steps": 500, + "global_step": 3004, + "is_hyper_param_search": false, + "is_local_process_zero": true, + "is_world_process_zero": true, + "log_history": [ + { + "epoch": 0.0, + "grad_norm": 8.0, + "learning_rate": 5.000000000000001e-07, + "loss": 1.1463, + "step": 1 + }, + { + "epoch": 0.0, + "grad_norm": 8.0, + "learning_rate": 1.0000000000000002e-06, + "loss": 1.2151, + "step": 2 + }, + { + "epoch": 0.0, + "grad_norm": 7.3125, + "learning_rate": 1.5e-06, + "loss": 1.1345, + "step": 3 + }, + { + "epoch": 0.01, + "grad_norm": 6.15625, + "learning_rate": 2.0000000000000003e-06, + "loss": 1.1139, + "step": 4 + }, + { + "epoch": 0.01, + "grad_norm": 5.375, + "learning_rate": 2.5e-06, + "loss": 1.0874, + "step": 5 + }, + { + "epoch": 0.01, + "grad_norm": 4.28125, + "learning_rate": 3e-06, + "loss": 1.1087, + "step": 6 + }, + { + "epoch": 0.01, + "grad_norm": 6.34375, + "learning_rate": 3.5e-06, + "loss": 1.0746, + "step": 7 + }, + { + "epoch": 0.01, + "grad_norm": 5.09375, + "learning_rate": 4.000000000000001e-06, + "loss": 1.0507, + "step": 8 + }, + { + "epoch": 0.01, + "grad_norm": 4.375, + "learning_rate": 4.5e-06, + "loss": 1.0417, + "step": 9 + }, + { + "epoch": 0.01, + "grad_norm": 3.5, + "learning_rate": 5e-06, + "loss": 1.0395, + "step": 10 + }, + { + "epoch": 0.01, + "grad_norm": 3.75, + "learning_rate": 4.999998623722131e-06, + "loss": 1.0696, + "step": 11 + }, + { + "epoch": 0.02, + "grad_norm": 3.390625, + "learning_rate": 4.999994494890037e-06, + "loss": 1.0175, + "step": 12 + }, + { + "epoch": 0.02, + "grad_norm": 3.234375, + "learning_rate": 4.999987613508266e-06, + "loss": 1.0273, + "step": 13 + }, + { + "epoch": 0.02, + "grad_norm": 3.265625, + "learning_rate": 4.999977979584393e-06, + "loss": 1.0296, + "step": 14 + }, + { + "epoch": 0.02, + "grad_norm": 3.1875, + "learning_rate": 4.999965593129025e-06, + "loss": 1.0156, + "step": 15 + }, + { + "epoch": 0.02, + "grad_norm": 3.140625, + "learning_rate": 4.999950454155801e-06, + "loss": 1.0328, + "step": 16 + }, + { + "epoch": 0.02, + "grad_norm": 2.96875, + "learning_rate": 4.999932562681389e-06, + "loss": 0.9897, + "step": 17 + }, + { + "epoch": 0.02, + "grad_norm": 2.984375, + "learning_rate": 4.9999119187254875e-06, + "loss": 1.008, + "step": 18 + }, + { + "epoch": 0.03, + "grad_norm": 3.03125, + "learning_rate": 4.999888522310826e-06, + "loss": 1.0089, + "step": 19 + }, + { + "epoch": 0.03, + "grad_norm": 3.015625, + "learning_rate": 4.999862373463166e-06, + "loss": 1.0183, + "step": 20 + }, + { + "epoch": 0.03, + "grad_norm": 2.75, + "learning_rate": 4.999833472211295e-06, + "loss": 1.0398, + "step": 21 + }, + { + "epoch": 0.03, + "grad_norm": 2.765625, + "learning_rate": 4.999801818587036e-06, + "loss": 1.0246, + "step": 22 + }, + { + "epoch": 0.03, + "grad_norm": 2.84375, + "learning_rate": 4.99976741262524e-06, + "loss": 1.0319, + "step": 23 + }, + { + "epoch": 0.03, + "grad_norm": 2.6875, + "learning_rate": 4.999730254363788e-06, + "loss": 0.9883, + "step": 24 + }, + { + "epoch": 0.03, + "grad_norm": 2.46875, + "learning_rate": 4.999690343843594e-06, + "loss": 1.0165, + "step": 25 + }, + { + "epoch": 0.03, + "grad_norm": 2.578125, + "learning_rate": 4.999647681108598e-06, + "loss": 0.9859, + "step": 26 + }, + { + "epoch": 0.04, + "grad_norm": 2.671875, + "learning_rate": 4.999602266205774e-06, + "loss": 0.9986, + "step": 27 + }, + { + "epoch": 0.04, + "grad_norm": 2.59375, + "learning_rate": 4.999554099185124e-06, + "loss": 1.0136, + "step": 28 + }, + { + "epoch": 0.04, + "grad_norm": 2.71875, + "learning_rate": 4.999503180099683e-06, + "loss": 0.9967, + "step": 29 + }, + { + "epoch": 0.04, + "grad_norm": 2.5, + "learning_rate": 4.999449509005511e-06, + "loss": 0.9814, + "step": 30 + }, + { + "epoch": 0.04, + "grad_norm": 2.484375, + "learning_rate": 4.999393085961704e-06, + "loss": 1.0176, + "step": 31 + }, + { + "epoch": 0.04, + "grad_norm": 2.625, + "learning_rate": 4.999333911030382e-06, + "loss": 0.9849, + "step": 32 + }, + { + "epoch": 0.04, + "grad_norm": 2.46875, + "learning_rate": 4.9992719842767e-06, + "loss": 0.9645, + "step": 33 + }, + { + "epoch": 0.05, + "grad_norm": 2.5625, + "learning_rate": 4.999207305768841e-06, + "loss": 0.9765, + "step": 34 + }, + { + "epoch": 0.05, + "grad_norm": 2.421875, + "learning_rate": 4.999139875578015e-06, + "loss": 0.9966, + "step": 35 + }, + { + "epoch": 0.05, + "grad_norm": 2.671875, + "learning_rate": 4.9990696937784675e-06, + "loss": 0.9941, + "step": 36 + }, + { + "epoch": 0.05, + "grad_norm": 2.65625, + "learning_rate": 4.998996760447469e-06, + "loss": 1.0011, + "step": 37 + }, + { + "epoch": 0.05, + "grad_norm": 2.4375, + "learning_rate": 4.99892107566532e-06, + "loss": 0.9851, + "step": 38 + }, + { + "epoch": 0.05, + "grad_norm": 2.796875, + "learning_rate": 4.99884263951535e-06, + "loss": 0.9769, + "step": 39 + }, + { + "epoch": 0.05, + "grad_norm": 2.578125, + "learning_rate": 4.998761452083922e-06, + "loss": 0.9735, + "step": 40 + }, + { + "epoch": 0.05, + "grad_norm": 2.59375, + "learning_rate": 4.998677513460423e-06, + "loss": 0.9647, + "step": 41 + }, + { + "epoch": 0.06, + "grad_norm": 2.484375, + "learning_rate": 4.998590823737272e-06, + "loss": 0.9532, + "step": 42 + }, + { + "epoch": 0.06, + "grad_norm": 2.546875, + "learning_rate": 4.998501383009916e-06, + "loss": 0.9697, + "step": 43 + }, + { + "epoch": 0.06, + "grad_norm": 2.625, + "learning_rate": 4.998409191376832e-06, + "loss": 0.9847, + "step": 44 + }, + { + "epoch": 0.06, + "grad_norm": 2.796875, + "learning_rate": 4.998314248939524e-06, + "loss": 0.9446, + "step": 45 + }, + { + "epoch": 0.06, + "grad_norm": 2.4375, + "learning_rate": 4.998216555802526e-06, + "loss": 0.9734, + "step": 46 + }, + { + "epoch": 0.06, + "grad_norm": 2.625, + "learning_rate": 4.998116112073401e-06, + "loss": 0.9812, + "step": 47 + }, + { + "epoch": 0.06, + "grad_norm": 2.546875, + "learning_rate": 4.99801291786274e-06, + "loss": 0.9718, + "step": 48 + }, + { + "epoch": 0.07, + "grad_norm": 2.453125, + "learning_rate": 4.99790697328416e-06, + "loss": 0.9361, + "step": 49 + }, + { + "epoch": 0.07, + "grad_norm": 2.671875, + "learning_rate": 4.997798278454311e-06, + "loss": 0.9479, + "step": 50 + }, + { + "epoch": 0.07, + "grad_norm": 2.671875, + "learning_rate": 4.997686833492866e-06, + "loss": 0.9559, + "step": 51 + }, + { + "epoch": 0.07, + "grad_norm": 2.546875, + "learning_rate": 4.997572638522531e-06, + "loss": 0.9877, + "step": 52 + }, + { + "epoch": 0.07, + "grad_norm": 2.6875, + "learning_rate": 4.997455693669036e-06, + "loss": 0.9724, + "step": 53 + }, + { + "epoch": 0.07, + "grad_norm": 2.578125, + "learning_rate": 4.997335999061138e-06, + "loss": 0.9715, + "step": 54 + }, + { + "epoch": 0.07, + "grad_norm": 2.484375, + "learning_rate": 4.997213554830627e-06, + "loss": 0.9499, + "step": 55 + }, + { + "epoch": 0.07, + "grad_norm": 2.671875, + "learning_rate": 4.997088361112314e-06, + "loss": 0.9779, + "step": 56 + }, + { + "epoch": 0.08, + "grad_norm": 2.578125, + "learning_rate": 4.9969604180440414e-06, + "loss": 1.0073, + "step": 57 + }, + { + "epoch": 0.08, + "grad_norm": 2.546875, + "learning_rate": 4.996829725766676e-06, + "loss": 0.9339, + "step": 58 + }, + { + "epoch": 0.08, + "grad_norm": 2.6875, + "learning_rate": 4.996696284424115e-06, + "loss": 0.9566, + "step": 59 + }, + { + "epoch": 0.08, + "grad_norm": 2.703125, + "learning_rate": 4.996560094163278e-06, + "loss": 0.9728, + "step": 60 + }, + { + "epoch": 0.08, + "grad_norm": 2.703125, + "learning_rate": 4.9964211551341166e-06, + "loss": 0.9584, + "step": 61 + }, + { + "epoch": 0.08, + "grad_norm": 2.453125, + "learning_rate": 4.996279467489603e-06, + "loss": 0.9692, + "step": 62 + }, + { + "epoch": 0.08, + "grad_norm": 2.515625, + "learning_rate": 4.99613503138574e-06, + "loss": 0.9662, + "step": 63 + }, + { + "epoch": 0.09, + "grad_norm": 2.578125, + "learning_rate": 4.995987846981554e-06, + "loss": 0.9427, + "step": 64 + }, + { + "epoch": 0.09, + "grad_norm": 2.5, + "learning_rate": 4.995837914439097e-06, + "loss": 0.9443, + "step": 65 + }, + { + "epoch": 0.09, + "grad_norm": 2.453125, + "learning_rate": 4.995685233923452e-06, + "loss": 0.9491, + "step": 66 + }, + { + "epoch": 0.09, + "grad_norm": 2.453125, + "learning_rate": 4.99552980560272e-06, + "loss": 0.9641, + "step": 67 + }, + { + "epoch": 0.09, + "grad_norm": 2.484375, + "learning_rate": 4.9953716296480335e-06, + "loss": 0.9572, + "step": 68 + }, + { + "epoch": 0.09, + "grad_norm": 2.453125, + "learning_rate": 4.995210706233546e-06, + "loss": 0.9752, + "step": 69 + }, + { + "epoch": 0.09, + "grad_norm": 2.5625, + "learning_rate": 4.995047035536439e-06, + "loss": 0.9664, + "step": 70 + }, + { + "epoch": 0.09, + "grad_norm": 2.671875, + "learning_rate": 4.994880617736917e-06, + "loss": 0.9298, + "step": 71 + }, + { + "epoch": 0.1, + "grad_norm": 2.53125, + "learning_rate": 4.994711453018208e-06, + "loss": 0.9727, + "step": 72 + }, + { + "epoch": 0.1, + "grad_norm": 2.671875, + "learning_rate": 4.99453954156657e-06, + "loss": 0.9432, + "step": 73 + }, + { + "epoch": 0.1, + "grad_norm": 2.515625, + "learning_rate": 4.994364883571279e-06, + "loss": 0.9561, + "step": 74 + }, + { + "epoch": 0.1, + "grad_norm": 2.46875, + "learning_rate": 4.994187479224636e-06, + "loss": 0.962, + "step": 75 + }, + { + "epoch": 0.1, + "grad_norm": 2.53125, + "learning_rate": 4.9940073287219705e-06, + "loss": 0.9644, + "step": 76 + }, + { + "epoch": 0.1, + "grad_norm": 2.765625, + "learning_rate": 4.993824432261629e-06, + "loss": 0.9744, + "step": 77 + }, + { + "epoch": 0.1, + "grad_norm": 2.765625, + "learning_rate": 4.993638790044987e-06, + "loss": 0.9822, + "step": 78 + }, + { + "epoch": 0.11, + "grad_norm": 2.953125, + "learning_rate": 4.993450402276439e-06, + "loss": 0.9704, + "step": 79 + }, + { + "epoch": 0.11, + "grad_norm": 2.5, + "learning_rate": 4.993259269163405e-06, + "loss": 0.9615, + "step": 80 + }, + { + "epoch": 0.11, + "grad_norm": 2.734375, + "learning_rate": 4.9930653909163276e-06, + "loss": 0.946, + "step": 81 + }, + { + "epoch": 0.11, + "grad_norm": 2.984375, + "learning_rate": 4.992868767748669e-06, + "loss": 0.9507, + "step": 82 + }, + { + "epoch": 0.11, + "grad_norm": 2.375, + "learning_rate": 4.992669399876917e-06, + "loss": 0.9424, + "step": 83 + }, + { + "epoch": 0.11, + "grad_norm": 2.625, + "learning_rate": 4.99246728752058e-06, + "loss": 0.9538, + "step": 84 + }, + { + "epoch": 0.11, + "grad_norm": 2.84375, + "learning_rate": 4.992262430902188e-06, + "loss": 0.9844, + "step": 85 + }, + { + "epoch": 0.11, + "grad_norm": 2.421875, + "learning_rate": 4.992054830247294e-06, + "loss": 0.9446, + "step": 86 + }, + { + "epoch": 0.12, + "grad_norm": 2.859375, + "learning_rate": 4.991844485784468e-06, + "loss": 0.9878, + "step": 87 + }, + { + "epoch": 0.12, + "grad_norm": 2.78125, + "learning_rate": 4.991631397745307e-06, + "loss": 0.9299, + "step": 88 + }, + { + "epoch": 0.12, + "grad_norm": 2.453125, + "learning_rate": 4.991415566364423e-06, + "loss": 0.9384, + "step": 89 + }, + { + "epoch": 0.12, + "grad_norm": 2.90625, + "learning_rate": 4.991196991879454e-06, + "loss": 0.931, + "step": 90 + }, + { + "epoch": 0.12, + "grad_norm": 2.9375, + "learning_rate": 4.990975674531053e-06, + "loss": 0.9589, + "step": 91 + }, + { + "epoch": 0.12, + "grad_norm": 2.5, + "learning_rate": 4.990751614562897e-06, + "loss": 0.9807, + "step": 92 + }, + { + "epoch": 0.12, + "grad_norm": 2.734375, + "learning_rate": 4.99052481222168e-06, + "loss": 0.942, + "step": 93 + }, + { + "epoch": 0.13, + "grad_norm": 2.59375, + "learning_rate": 4.990295267757117e-06, + "loss": 0.9683, + "step": 94 + }, + { + "epoch": 0.13, + "grad_norm": 2.515625, + "learning_rate": 4.990062981421941e-06, + "loss": 0.9652, + "step": 95 + }, + { + "epoch": 0.13, + "grad_norm": 2.625, + "learning_rate": 4.989827953471905e-06, + "loss": 0.9643, + "step": 96 + }, + { + "epoch": 0.13, + "grad_norm": 2.625, + "learning_rate": 4.989590184165779e-06, + "loss": 0.9868, + "step": 97 + }, + { + "epoch": 0.13, + "grad_norm": 2.71875, + "learning_rate": 4.989349673765354e-06, + "loss": 0.9581, + "step": 98 + }, + { + "epoch": 0.13, + "grad_norm": 2.671875, + "learning_rate": 4.989106422535436e-06, + "loss": 0.9451, + "step": 99 + }, + { + "epoch": 0.13, + "grad_norm": 2.5, + "learning_rate": 4.98886043074385e-06, + "loss": 0.9653, + "step": 100 + }, + { + "epoch": 0.13, + "grad_norm": 2.546875, + "learning_rate": 4.988611698661439e-06, + "loss": 0.959, + "step": 101 + }, + { + "epoch": 0.14, + "grad_norm": 2.703125, + "learning_rate": 4.988360226562063e-06, + "loss": 0.9543, + "step": 102 + }, + { + "epoch": 0.14, + "grad_norm": 2.484375, + "learning_rate": 4.988106014722598e-06, + "loss": 0.9911, + "step": 103 + }, + { + "epoch": 0.14, + "grad_norm": 2.671875, + "learning_rate": 4.987849063422935e-06, + "loss": 0.9606, + "step": 104 + }, + { + "epoch": 0.14, + "grad_norm": 2.640625, + "learning_rate": 4.987589372945987e-06, + "loss": 0.9378, + "step": 105 + }, + { + "epoch": 0.14, + "grad_norm": 2.609375, + "learning_rate": 4.987326943577675e-06, + "loss": 0.9392, + "step": 106 + }, + { + "epoch": 0.14, + "grad_norm": 2.609375, + "learning_rate": 4.987061775606942e-06, + "loss": 0.9475, + "step": 107 + }, + { + "epoch": 0.14, + "grad_norm": 2.65625, + "learning_rate": 4.986793869325743e-06, + "loss": 0.9331, + "step": 108 + }, + { + "epoch": 0.15, + "grad_norm": 2.640625, + "learning_rate": 4.98652322502905e-06, + "loss": 0.9634, + "step": 109 + }, + { + "epoch": 0.15, + "grad_norm": 2.421875, + "learning_rate": 4.986249843014847e-06, + "loss": 0.9308, + "step": 110 + }, + { + "epoch": 0.15, + "grad_norm": 2.78125, + "learning_rate": 4.985973723584134e-06, + "loss": 1.0091, + "step": 111 + }, + { + "epoch": 0.15, + "grad_norm": 2.578125, + "learning_rate": 4.985694867040924e-06, + "loss": 0.9875, + "step": 112 + }, + { + "epoch": 0.15, + "grad_norm": 2.625, + "learning_rate": 4.985413273692247e-06, + "loss": 0.939, + "step": 113 + }, + { + "epoch": 0.15, + "grad_norm": 2.734375, + "learning_rate": 4.98512894384814e-06, + "loss": 0.9955, + "step": 114 + }, + { + "epoch": 0.15, + "grad_norm": 2.515625, + "learning_rate": 4.984841877821659e-06, + "loss": 0.9223, + "step": 115 + }, + { + "epoch": 0.15, + "grad_norm": 2.703125, + "learning_rate": 4.984552075928869e-06, + "loss": 0.9479, + "step": 116 + }, + { + "epoch": 0.16, + "grad_norm": 2.71875, + "learning_rate": 4.984259538488848e-06, + "loss": 0.9505, + "step": 117 + }, + { + "epoch": 0.16, + "grad_norm": 2.453125, + "learning_rate": 4.983964265823687e-06, + "loss": 0.9247, + "step": 118 + }, + { + "epoch": 0.16, + "grad_norm": 2.53125, + "learning_rate": 4.983666258258488e-06, + "loss": 0.9277, + "step": 119 + }, + { + "epoch": 0.16, + "grad_norm": 2.578125, + "learning_rate": 4.983365516121363e-06, + "loss": 0.9235, + "step": 120 + }, + { + "epoch": 0.16, + "grad_norm": 2.421875, + "learning_rate": 4.983062039743436e-06, + "loss": 0.9546, + "step": 121 + }, + { + "epoch": 0.16, + "grad_norm": 2.421875, + "learning_rate": 4.9827558294588425e-06, + "loss": 0.9584, + "step": 122 + }, + { + "epoch": 0.16, + "grad_norm": 2.53125, + "learning_rate": 4.982446885604726e-06, + "loss": 0.9023, + "step": 123 + }, + { + "epoch": 0.17, + "grad_norm": 2.578125, + "learning_rate": 4.98213520852124e-06, + "loss": 0.9454, + "step": 124 + }, + { + "epoch": 0.17, + "grad_norm": 2.78125, + "learning_rate": 4.981820798551549e-06, + "loss": 0.9794, + "step": 125 + }, + { + "epoch": 0.17, + "grad_norm": 2.53125, + "learning_rate": 4.981503656041824e-06, + "loss": 0.9458, + "step": 126 + }, + { + "epoch": 0.17, + "grad_norm": 2.59375, + "learning_rate": 4.981183781341248e-06, + "loss": 0.9751, + "step": 127 + }, + { + "epoch": 0.17, + "grad_norm": 2.28125, + "learning_rate": 4.980861174802009e-06, + "loss": 0.9478, + "step": 128 + }, + { + "epoch": 0.17, + "grad_norm": 2.46875, + "learning_rate": 4.9805358367793025e-06, + "loss": 0.9768, + "step": 129 + }, + { + "epoch": 0.17, + "grad_norm": 2.453125, + "learning_rate": 4.980207767631335e-06, + "loss": 0.9539, + "step": 130 + }, + { + "epoch": 0.17, + "grad_norm": 2.609375, + "learning_rate": 4.9798769677193185e-06, + "loss": 0.9388, + "step": 131 + }, + { + "epoch": 0.18, + "grad_norm": 2.5, + "learning_rate": 4.979543437407469e-06, + "loss": 0.9674, + "step": 132 + }, + { + "epoch": 0.18, + "grad_norm": 2.765625, + "learning_rate": 4.979207177063011e-06, + "loss": 0.9331, + "step": 133 + }, + { + "epoch": 0.18, + "grad_norm": 2.59375, + "learning_rate": 4.978868187056176e-06, + "loss": 0.9582, + "step": 134 + }, + { + "epoch": 0.18, + "grad_norm": 2.515625, + "learning_rate": 4.9785264677602e-06, + "loss": 0.9441, + "step": 135 + }, + { + "epoch": 0.18, + "grad_norm": 2.5, + "learning_rate": 4.978182019551321e-06, + "loss": 0.9655, + "step": 136 + }, + { + "epoch": 0.18, + "grad_norm": 2.34375, + "learning_rate": 4.977834842808786e-06, + "loss": 0.9449, + "step": 137 + }, + { + "epoch": 0.18, + "grad_norm": 2.640625, + "learning_rate": 4.977484937914844e-06, + "loss": 0.9333, + "step": 138 + }, + { + "epoch": 0.18, + "grad_norm": 2.40625, + "learning_rate": 4.9771323052547485e-06, + "loss": 0.9374, + "step": 139 + }, + { + "epoch": 0.19, + "grad_norm": 2.84375, + "learning_rate": 4.976776945216755e-06, + "loss": 0.9386, + "step": 140 + }, + { + "epoch": 0.19, + "grad_norm": 2.484375, + "learning_rate": 4.976418858192124e-06, + "loss": 0.9615, + "step": 141 + }, + { + "epoch": 0.19, + "grad_norm": 2.5625, + "learning_rate": 4.976058044575116e-06, + "loss": 0.9246, + "step": 142 + }, + { + "epoch": 0.19, + "grad_norm": 2.375, + "learning_rate": 4.9756945047629955e-06, + "loss": 0.9142, + "step": 143 + }, + { + "epoch": 0.19, + "grad_norm": 2.359375, + "learning_rate": 4.975328239156029e-06, + "loss": 0.9302, + "step": 144 + }, + { + "epoch": 0.19, + "grad_norm": 2.546875, + "learning_rate": 4.9749592481574806e-06, + "loss": 0.915, + "step": 145 + }, + { + "epoch": 0.19, + "grad_norm": 2.578125, + "learning_rate": 4.97458753217362e-06, + "loss": 0.9329, + "step": 146 + }, + { + "epoch": 0.2, + "grad_norm": 2.578125, + "learning_rate": 4.9742130916137125e-06, + "loss": 0.9297, + "step": 147 + }, + { + "epoch": 0.2, + "grad_norm": 2.78125, + "learning_rate": 4.973835926890027e-06, + "loss": 0.9239, + "step": 148 + }, + { + "epoch": 0.2, + "grad_norm": 2.4375, + "learning_rate": 4.973456038417831e-06, + "loss": 0.9158, + "step": 149 + }, + { + "epoch": 0.2, + "grad_norm": 2.515625, + "learning_rate": 4.973073426615389e-06, + "loss": 0.968, + "step": 150 + }, + { + "epoch": 0.2, + "grad_norm": 2.546875, + "learning_rate": 4.972688091903965e-06, + "loss": 0.9523, + "step": 151 + }, + { + "epoch": 0.2, + "grad_norm": 2.390625, + "learning_rate": 4.972300034707822e-06, + "loss": 0.9265, + "step": 152 + }, + { + "epoch": 0.2, + "grad_norm": 3.328125, + "learning_rate": 4.971909255454219e-06, + "loss": 0.9421, + "step": 153 + }, + { + "epoch": 0.2, + "grad_norm": 2.75, + "learning_rate": 4.9715157545734124e-06, + "loss": 0.9445, + "step": 154 + }, + { + "epoch": 0.21, + "grad_norm": 2.53125, + "learning_rate": 4.971119532498656e-06, + "loss": 0.9364, + "step": 155 + }, + { + "epoch": 0.21, + "grad_norm": 2.53125, + "learning_rate": 4.970720589666199e-06, + "loss": 0.9755, + "step": 156 + }, + { + "epoch": 0.21, + "grad_norm": 2.609375, + "learning_rate": 4.9703189265152865e-06, + "loss": 0.9627, + "step": 157 + }, + { + "epoch": 0.21, + "grad_norm": 2.703125, + "learning_rate": 4.969914543488158e-06, + "loss": 0.9484, + "step": 158 + }, + { + "epoch": 0.21, + "grad_norm": 2.84375, + "learning_rate": 4.969507441030048e-06, + "loss": 0.9873, + "step": 159 + }, + { + "epoch": 0.21, + "grad_norm": 2.515625, + "learning_rate": 4.969097619589187e-06, + "loss": 0.9709, + "step": 160 + }, + { + "epoch": 0.21, + "grad_norm": 2.796875, + "learning_rate": 4.968685079616796e-06, + "loss": 0.9415, + "step": 161 + }, + { + "epoch": 0.22, + "grad_norm": 2.796875, + "learning_rate": 4.968269821567092e-06, + "loss": 0.9673, + "step": 162 + }, + { + "epoch": 0.22, + "grad_norm": 2.46875, + "learning_rate": 4.967851845897282e-06, + "loss": 0.9395, + "step": 163 + }, + { + "epoch": 0.22, + "grad_norm": 2.828125, + "learning_rate": 4.967431153067568e-06, + "loss": 0.912, + "step": 164 + }, + { + "epoch": 0.22, + "grad_norm": 2.59375, + "learning_rate": 4.96700774354114e-06, + "loss": 0.933, + "step": 165 + }, + { + "epoch": 0.22, + "grad_norm": 2.59375, + "learning_rate": 4.9665816177841845e-06, + "loss": 0.941, + "step": 166 + }, + { + "epoch": 0.22, + "grad_norm": 2.75, + "learning_rate": 4.966152776265872e-06, + "loss": 0.929, + "step": 167 + }, + { + "epoch": 0.22, + "grad_norm": 2.625, + "learning_rate": 4.965721219458369e-06, + "loss": 0.9401, + "step": 168 + }, + { + "epoch": 0.22, + "grad_norm": 2.640625, + "learning_rate": 4.965286947836828e-06, + "loss": 0.9366, + "step": 169 + }, + { + "epoch": 0.23, + "grad_norm": 2.546875, + "learning_rate": 4.964849961879392e-06, + "loss": 0.9318, + "step": 170 + }, + { + "epoch": 0.23, + "grad_norm": 2.578125, + "learning_rate": 4.964410262067193e-06, + "loss": 0.9184, + "step": 171 + }, + { + "epoch": 0.23, + "grad_norm": 2.5625, + "learning_rate": 4.963967848884349e-06, + "loss": 0.9049, + "step": 172 + }, + { + "epoch": 0.23, + "grad_norm": 2.640625, + "learning_rate": 4.963522722817968e-06, + "loss": 0.9519, + "step": 173 + }, + { + "epoch": 0.23, + "grad_norm": 2.515625, + "learning_rate": 4.963074884358143e-06, + "loss": 0.9106, + "step": 174 + }, + { + "epoch": 0.23, + "grad_norm": 2.65625, + "learning_rate": 4.962624333997954e-06, + "loss": 0.9483, + "step": 175 + }, + { + "epoch": 0.23, + "grad_norm": 2.515625, + "learning_rate": 4.962171072233468e-06, + "loss": 0.9023, + "step": 176 + }, + { + "epoch": 0.24, + "grad_norm": 2.84375, + "learning_rate": 4.961715099563735e-06, + "loss": 0.9313, + "step": 177 + }, + { + "epoch": 0.24, + "grad_norm": 2.4375, + "learning_rate": 4.961256416490793e-06, + "loss": 0.903, + "step": 178 + }, + { + "epoch": 0.24, + "grad_norm": 2.578125, + "learning_rate": 4.960795023519659e-06, + "loss": 0.9346, + "step": 179 + }, + { + "epoch": 0.24, + "grad_norm": 2.65625, + "learning_rate": 4.96033092115834e-06, + "loss": 0.9435, + "step": 180 + }, + { + "epoch": 0.24, + "grad_norm": 2.375, + "learning_rate": 4.959864109917822e-06, + "loss": 0.919, + "step": 181 + }, + { + "epoch": 0.24, + "grad_norm": 2.546875, + "learning_rate": 4.9593945903120744e-06, + "loss": 0.9309, + "step": 182 + }, + { + "epoch": 0.24, + "grad_norm": 2.40625, + "learning_rate": 4.958922362858048e-06, + "loss": 0.9322, + "step": 183 + }, + { + "epoch": 0.24, + "grad_norm": 2.46875, + "learning_rate": 4.9584474280756785e-06, + "loss": 0.9052, + "step": 184 + }, + { + "epoch": 0.25, + "grad_norm": 2.65625, + "learning_rate": 4.957969786487877e-06, + "loss": 0.9712, + "step": 185 + }, + { + "epoch": 0.25, + "grad_norm": 2.453125, + "learning_rate": 4.9574894386205384e-06, + "loss": 0.9478, + "step": 186 + }, + { + "epoch": 0.25, + "grad_norm": 2.625, + "learning_rate": 4.957006385002536e-06, + "loss": 0.9682, + "step": 187 + }, + { + "epoch": 0.25, + "grad_norm": 2.609375, + "learning_rate": 4.956520626165724e-06, + "loss": 0.9612, + "step": 188 + }, + { + "epoch": 0.25, + "grad_norm": 2.59375, + "learning_rate": 4.956032162644932e-06, + "loss": 0.9473, + "step": 189 + }, + { + "epoch": 0.25, + "grad_norm": 2.625, + "learning_rate": 4.9555409949779695e-06, + "loss": 0.9271, + "step": 190 + }, + { + "epoch": 0.25, + "grad_norm": 2.78125, + "learning_rate": 4.955047123705624e-06, + "loss": 0.9448, + "step": 191 + }, + { + "epoch": 0.26, + "grad_norm": 2.703125, + "learning_rate": 4.954550549371659e-06, + "loss": 0.9144, + "step": 192 + }, + { + "epoch": 0.26, + "grad_norm": 2.421875, + "learning_rate": 4.9540512725228124e-06, + "loss": 0.9308, + "step": 193 + }, + { + "epoch": 0.26, + "grad_norm": 2.84375, + "learning_rate": 4.953549293708801e-06, + "loss": 0.9361, + "step": 194 + }, + { + "epoch": 0.26, + "grad_norm": 2.453125, + "learning_rate": 4.953044613482312e-06, + "loss": 0.9268, + "step": 195 + }, + { + "epoch": 0.26, + "grad_norm": 2.6875, + "learning_rate": 4.952537232399012e-06, + "loss": 0.9317, + "step": 196 + }, + { + "epoch": 0.26, + "grad_norm": 2.734375, + "learning_rate": 4.952027151017538e-06, + "loss": 0.9393, + "step": 197 + }, + { + "epoch": 0.26, + "grad_norm": 2.578125, + "learning_rate": 4.9515143698995006e-06, + "loss": 0.9591, + "step": 198 + }, + { + "epoch": 0.26, + "grad_norm": 2.796875, + "learning_rate": 4.9509988896094845e-06, + "loss": 0.9402, + "step": 199 + }, + { + "epoch": 0.27, + "grad_norm": 2.640625, + "learning_rate": 4.950480710715043e-06, + "loss": 0.9749, + "step": 200 + }, + { + "epoch": 0.27, + "grad_norm": 2.546875, + "learning_rate": 4.949959833786705e-06, + "loss": 0.945, + "step": 201 + }, + { + "epoch": 0.27, + "grad_norm": 2.703125, + "learning_rate": 4.9494362593979665e-06, + "loss": 0.9298, + "step": 202 + }, + { + "epoch": 0.27, + "grad_norm": 2.625, + "learning_rate": 4.9489099881252935e-06, + "loss": 0.9616, + "step": 203 + }, + { + "epoch": 0.27, + "grad_norm": 2.65625, + "learning_rate": 4.9483810205481244e-06, + "loss": 0.9294, + "step": 204 + }, + { + "epoch": 0.27, + "grad_norm": 2.5, + "learning_rate": 4.947849357248862e-06, + "loss": 0.9418, + "step": 205 + }, + { + "epoch": 0.27, + "grad_norm": 2.703125, + "learning_rate": 4.947314998812882e-06, + "loss": 0.9473, + "step": 206 + }, + { + "epoch": 0.28, + "grad_norm": 2.640625, + "learning_rate": 4.946777945828523e-06, + "loss": 0.938, + "step": 207 + }, + { + "epoch": 0.28, + "grad_norm": 2.40625, + "learning_rate": 4.946238198887093e-06, + "loss": 0.9142, + "step": 208 + }, + { + "epoch": 0.28, + "grad_norm": 2.390625, + "learning_rate": 4.945695758582865e-06, + "loss": 0.9513, + "step": 209 + }, + { + "epoch": 0.28, + "grad_norm": 2.828125, + "learning_rate": 4.94515062551308e-06, + "loss": 0.9449, + "step": 210 + }, + { + "epoch": 0.28, + "grad_norm": 2.5, + "learning_rate": 4.944602800277939e-06, + "loss": 0.9239, + "step": 211 + }, + { + "epoch": 0.28, + "grad_norm": 2.484375, + "learning_rate": 4.94405228348061e-06, + "loss": 0.9238, + "step": 212 + }, + { + "epoch": 0.28, + "grad_norm": 2.625, + "learning_rate": 4.9434990757272255e-06, + "loss": 0.9205, + "step": 213 + }, + { + "epoch": 0.28, + "grad_norm": 2.46875, + "learning_rate": 4.942943177626879e-06, + "loss": 0.9312, + "step": 214 + }, + { + "epoch": 0.29, + "grad_norm": 2.53125, + "learning_rate": 4.942384589791627e-06, + "loss": 0.9689, + "step": 215 + }, + { + "epoch": 0.29, + "grad_norm": 2.609375, + "learning_rate": 4.941823312836487e-06, + "loss": 0.9055, + "step": 216 + }, + { + "epoch": 0.29, + "grad_norm": 2.515625, + "learning_rate": 4.941259347379437e-06, + "loss": 0.9175, + "step": 217 + }, + { + "epoch": 0.29, + "grad_norm": 2.640625, + "learning_rate": 4.9406926940414165e-06, + "loss": 0.9503, + "step": 218 + }, + { + "epoch": 0.29, + "grad_norm": 2.609375, + "learning_rate": 4.940123353446322e-06, + "loss": 0.934, + "step": 219 + }, + { + "epoch": 0.29, + "grad_norm": 2.625, + "learning_rate": 4.939551326221012e-06, + "loss": 0.9551, + "step": 220 + }, + { + "epoch": 0.29, + "grad_norm": 2.53125, + "learning_rate": 4.9389766129953e-06, + "loss": 0.9428, + "step": 221 + }, + { + "epoch": 0.3, + "grad_norm": 2.53125, + "learning_rate": 4.9383992144019586e-06, + "loss": 0.9544, + "step": 222 + }, + { + "epoch": 0.3, + "grad_norm": 2.609375, + "learning_rate": 4.937819131076716e-06, + "loss": 0.9369, + "step": 223 + }, + { + "epoch": 0.3, + "grad_norm": 2.671875, + "learning_rate": 4.937236363658257e-06, + "loss": 0.9151, + "step": 224 + }, + { + "epoch": 0.3, + "grad_norm": 2.390625, + "learning_rate": 4.936650912788222e-06, + "loss": 0.9303, + "step": 225 + }, + { + "epoch": 0.3, + "grad_norm": 2.5, + "learning_rate": 4.936062779111205e-06, + "loss": 0.9535, + "step": 226 + }, + { + "epoch": 0.3, + "grad_norm": 2.703125, + "learning_rate": 4.935471963274755e-06, + "loss": 0.9413, + "step": 227 + }, + { + "epoch": 0.3, + "grad_norm": 2.46875, + "learning_rate": 4.934878465929372e-06, + "loss": 0.9249, + "step": 228 + }, + { + "epoch": 0.3, + "grad_norm": 2.46875, + "learning_rate": 4.934282287728511e-06, + "loss": 0.9139, + "step": 229 + }, + { + "epoch": 0.31, + "grad_norm": 2.484375, + "learning_rate": 4.9336834293285774e-06, + "loss": 0.9246, + "step": 230 + }, + { + "epoch": 0.31, + "grad_norm": 2.40625, + "learning_rate": 4.933081891388927e-06, + "loss": 0.9264, + "step": 231 + }, + { + "epoch": 0.31, + "grad_norm": 2.4375, + "learning_rate": 4.932477674571867e-06, + "loss": 0.9087, + "step": 232 + }, + { + "epoch": 0.31, + "grad_norm": 2.5625, + "learning_rate": 4.931870779542654e-06, + "loss": 0.9263, + "step": 233 + }, + { + "epoch": 0.31, + "grad_norm": 2.625, + "learning_rate": 4.931261206969492e-06, + "loss": 0.9539, + "step": 234 + }, + { + "epoch": 0.31, + "grad_norm": 2.578125, + "learning_rate": 4.9306489575235344e-06, + "loss": 0.9408, + "step": 235 + }, + { + "epoch": 0.31, + "grad_norm": 2.4375, + "learning_rate": 4.930034031878882e-06, + "loss": 0.9176, + "step": 236 + }, + { + "epoch": 0.32, + "grad_norm": 2.546875, + "learning_rate": 4.92941643071258e-06, + "loss": 0.9464, + "step": 237 + }, + { + "epoch": 0.32, + "grad_norm": 2.34375, + "learning_rate": 4.928796154704623e-06, + "loss": 0.9155, + "step": 238 + }, + { + "epoch": 0.32, + "grad_norm": 2.578125, + "learning_rate": 4.928173204537948e-06, + "loss": 0.9032, + "step": 239 + }, + { + "epoch": 0.32, + "grad_norm": 2.546875, + "learning_rate": 4.927547580898438e-06, + "loss": 0.9316, + "step": 240 + }, + { + "epoch": 0.32, + "grad_norm": 2.578125, + "learning_rate": 4.9269192844749156e-06, + "loss": 0.9308, + "step": 241 + }, + { + "epoch": 0.32, + "grad_norm": 2.5625, + "learning_rate": 4.926288315959152e-06, + "loss": 0.9319, + "step": 242 + }, + { + "epoch": 0.32, + "grad_norm": 2.609375, + "learning_rate": 4.925654676045857e-06, + "loss": 0.9296, + "step": 243 + }, + { + "epoch": 0.32, + "grad_norm": 2.65625, + "learning_rate": 4.925018365432681e-06, + "loss": 0.9086, + "step": 244 + }, + { + "epoch": 0.33, + "grad_norm": 2.5625, + "learning_rate": 4.924379384820218e-06, + "loss": 0.9198, + "step": 245 + }, + { + "epoch": 0.33, + "grad_norm": 2.703125, + "learning_rate": 4.923737734911999e-06, + "loss": 0.9411, + "step": 246 + }, + { + "epoch": 0.33, + "grad_norm": 2.640625, + "learning_rate": 4.923093416414495e-06, + "loss": 0.9412, + "step": 247 + }, + { + "epoch": 0.33, + "grad_norm": 2.484375, + "learning_rate": 4.922446430037115e-06, + "loss": 0.9282, + "step": 248 + }, + { + "epoch": 0.33, + "grad_norm": 2.46875, + "learning_rate": 4.921796776492205e-06, + "loss": 0.9155, + "step": 249 + }, + { + "epoch": 0.33, + "grad_norm": 3.90625, + "learning_rate": 4.921144456495048e-06, + "loss": 0.9376, + "step": 250 + }, + { + "epoch": 0.33, + "grad_norm": 2.5, + "learning_rate": 4.920489470763864e-06, + "loss": 0.9182, + "step": 251 + }, + { + "epoch": 0.34, + "grad_norm": 2.453125, + "learning_rate": 4.919831820019806e-06, + "loss": 0.9316, + "step": 252 + }, + { + "epoch": 0.34, + "grad_norm": 2.53125, + "learning_rate": 4.919171504986963e-06, + "loss": 0.9456, + "step": 253 + }, + { + "epoch": 0.34, + "grad_norm": 2.4375, + "learning_rate": 4.918508526392355e-06, + "loss": 0.9165, + "step": 254 + }, + { + "epoch": 0.34, + "grad_norm": 2.578125, + "learning_rate": 4.917842884965937e-06, + "loss": 0.9365, + "step": 255 + }, + { + "epoch": 0.34, + "grad_norm": 2.515625, + "learning_rate": 4.9171745814405945e-06, + "loss": 0.9493, + "step": 256 + }, + { + "epoch": 0.34, + "grad_norm": 2.375, + "learning_rate": 4.916503616552146e-06, + "loss": 0.9179, + "step": 257 + }, + { + "epoch": 0.34, + "grad_norm": 2.34375, + "learning_rate": 4.915829991039338e-06, + "loss": 0.9334, + "step": 258 + }, + { + "epoch": 0.34, + "grad_norm": 2.4375, + "learning_rate": 4.9151537056438484e-06, + "loss": 0.952, + "step": 259 + }, + { + "epoch": 0.35, + "grad_norm": 2.53125, + "learning_rate": 4.91447476111028e-06, + "loss": 0.9441, + "step": 260 + }, + { + "epoch": 0.35, + "grad_norm": 2.625, + "learning_rate": 4.913793158186167e-06, + "loss": 0.92, + "step": 261 + }, + { + "epoch": 0.35, + "grad_norm": 2.53125, + "learning_rate": 4.9131088976219695e-06, + "loss": 0.9439, + "step": 262 + }, + { + "epoch": 0.35, + "grad_norm": 2.484375, + "learning_rate": 4.912421980171075e-06, + "loss": 0.9368, + "step": 263 + }, + { + "epoch": 0.35, + "grad_norm": 2.4375, + "learning_rate": 4.911732406589792e-06, + "loss": 0.9176, + "step": 264 + }, + { + "epoch": 0.35, + "grad_norm": 2.546875, + "learning_rate": 4.91104017763736e-06, + "loss": 0.9065, + "step": 265 + }, + { + "epoch": 0.35, + "grad_norm": 2.5, + "learning_rate": 4.910345294075935e-06, + "loss": 0.9171, + "step": 266 + }, + { + "epoch": 0.36, + "grad_norm": 2.40625, + "learning_rate": 4.9096477566706015e-06, + "loss": 0.9234, + "step": 267 + }, + { + "epoch": 0.36, + "grad_norm": 2.484375, + "learning_rate": 4.908947566189362e-06, + "loss": 0.8894, + "step": 268 + }, + { + "epoch": 0.36, + "grad_norm": 2.453125, + "learning_rate": 4.908244723403145e-06, + "loss": 0.9232, + "step": 269 + }, + { + "epoch": 0.36, + "grad_norm": 2.546875, + "learning_rate": 4.9075392290857925e-06, + "loss": 0.9025, + "step": 270 + }, + { + "epoch": 0.36, + "grad_norm": 2.453125, + "learning_rate": 4.9068310840140725e-06, + "loss": 0.9252, + "step": 271 + }, + { + "epoch": 0.36, + "grad_norm": 2.5, + "learning_rate": 4.906120288967665e-06, + "loss": 0.9382, + "step": 272 + }, + { + "epoch": 0.36, + "grad_norm": 2.640625, + "learning_rate": 4.905406844729175e-06, + "loss": 0.9253, + "step": 273 + }, + { + "epoch": 0.36, + "grad_norm": 2.5, + "learning_rate": 4.904690752084117e-06, + "loss": 0.9071, + "step": 274 + }, + { + "epoch": 0.37, + "grad_norm": 2.546875, + "learning_rate": 4.903972011820928e-06, + "loss": 0.9108, + "step": 275 + }, + { + "epoch": 0.37, + "grad_norm": 2.5625, + "learning_rate": 4.9032506247309545e-06, + "loss": 0.9391, + "step": 276 + }, + { + "epoch": 0.37, + "grad_norm": 2.65625, + "learning_rate": 4.9025265916084615e-06, + "loss": 0.9208, + "step": 277 + }, + { + "epoch": 0.37, + "grad_norm": 2.484375, + "learning_rate": 4.9017999132506254e-06, + "loss": 0.9219, + "step": 278 + }, + { + "epoch": 0.37, + "grad_norm": 2.453125, + "learning_rate": 4.901070590457535e-06, + "loss": 0.8975, + "step": 279 + }, + { + "epoch": 0.37, + "grad_norm": 2.6875, + "learning_rate": 4.900338624032191e-06, + "loss": 0.9385, + "step": 280 + }, + { + "epoch": 0.37, + "grad_norm": 2.40625, + "learning_rate": 4.8996040147805045e-06, + "loss": 0.9384, + "step": 281 + }, + { + "epoch": 0.38, + "grad_norm": 2.40625, + "learning_rate": 4.898866763511297e-06, + "loss": 0.8997, + "step": 282 + }, + { + "epoch": 0.38, + "grad_norm": 2.484375, + "learning_rate": 4.898126871036298e-06, + "loss": 0.9267, + "step": 283 + }, + { + "epoch": 0.38, + "grad_norm": 2.46875, + "learning_rate": 4.8973843381701465e-06, + "loss": 0.9006, + "step": 284 + }, + { + "epoch": 0.38, + "grad_norm": 2.5, + "learning_rate": 4.8966391657303875e-06, + "loss": 0.9391, + "step": 285 + }, + { + "epoch": 0.38, + "grad_norm": 2.53125, + "learning_rate": 4.895891354537472e-06, + "loss": 0.9408, + "step": 286 + }, + { + "epoch": 0.38, + "grad_norm": 2.5625, + "learning_rate": 4.895140905414757e-06, + "loss": 0.9378, + "step": 287 + }, + { + "epoch": 0.38, + "grad_norm": 2.5, + "learning_rate": 4.894387819188504e-06, + "loss": 0.9567, + "step": 288 + }, + { + "epoch": 0.38, + "grad_norm": 2.390625, + "learning_rate": 4.893632096687877e-06, + "loss": 0.958, + "step": 289 + }, + { + "epoch": 0.39, + "grad_norm": 2.5625, + "learning_rate": 4.892873738744944e-06, + "loss": 0.9024, + "step": 290 + }, + { + "epoch": 0.39, + "grad_norm": 2.625, + "learning_rate": 4.892112746194674e-06, + "loss": 0.9276, + "step": 291 + }, + { + "epoch": 0.39, + "grad_norm": 2.484375, + "learning_rate": 4.891349119874936e-06, + "loss": 0.9269, + "step": 292 + }, + { + "epoch": 0.39, + "grad_norm": 2.640625, + "learning_rate": 4.890582860626501e-06, + "loss": 0.9452, + "step": 293 + }, + { + "epoch": 0.39, + "grad_norm": 2.640625, + "learning_rate": 4.8898139692930365e-06, + "loss": 0.9617, + "step": 294 + }, + { + "epoch": 0.39, + "grad_norm": 2.53125, + "learning_rate": 4.889042446721109e-06, + "loss": 0.9369, + "step": 295 + }, + { + "epoch": 0.39, + "grad_norm": 2.4375, + "learning_rate": 4.888268293760182e-06, + "loss": 0.911, + "step": 296 + }, + { + "epoch": 0.4, + "grad_norm": 2.46875, + "learning_rate": 4.887491511262616e-06, + "loss": 0.9277, + "step": 297 + }, + { + "epoch": 0.4, + "grad_norm": 2.296875, + "learning_rate": 4.886712100083664e-06, + "loss": 0.9266, + "step": 298 + }, + { + "epoch": 0.4, + "grad_norm": 2.390625, + "learning_rate": 4.885930061081478e-06, + "loss": 0.9333, + "step": 299 + }, + { + "epoch": 0.4, + "grad_norm": 2.46875, + "learning_rate": 4.885145395117099e-06, + "loss": 0.9078, + "step": 300 + }, + { + "epoch": 0.4, + "grad_norm": 2.484375, + "learning_rate": 4.88435810305446e-06, + "loss": 0.9177, + "step": 301 + }, + { + "epoch": 0.4, + "grad_norm": 2.46875, + "learning_rate": 4.88356818576039e-06, + "loss": 0.9046, + "step": 302 + }, + { + "epoch": 0.4, + "grad_norm": 2.5625, + "learning_rate": 4.882775644104604e-06, + "loss": 0.9371, + "step": 303 + }, + { + "epoch": 0.4, + "grad_norm": 2.4375, + "learning_rate": 4.881980478959707e-06, + "loss": 0.9134, + "step": 304 + }, + { + "epoch": 0.41, + "grad_norm": 2.453125, + "learning_rate": 4.881182691201195e-06, + "loss": 0.9257, + "step": 305 + }, + { + "epoch": 0.41, + "grad_norm": 2.46875, + "learning_rate": 4.880382281707451e-06, + "loss": 0.9289, + "step": 306 + }, + { + "epoch": 0.41, + "grad_norm": 2.4375, + "learning_rate": 4.879579251359741e-06, + "loss": 0.9026, + "step": 307 + }, + { + "epoch": 0.41, + "grad_norm": 2.453125, + "learning_rate": 4.878773601042221e-06, + "loss": 0.8994, + "step": 308 + }, + { + "epoch": 0.41, + "grad_norm": 2.46875, + "learning_rate": 4.877965331641931e-06, + "loss": 0.9357, + "step": 309 + }, + { + "epoch": 0.41, + "grad_norm": 2.34375, + "learning_rate": 4.877154444048792e-06, + "loss": 0.9267, + "step": 310 + }, + { + "epoch": 0.41, + "grad_norm": 2.640625, + "learning_rate": 4.876340939155609e-06, + "loss": 0.9506, + "step": 311 + }, + { + "epoch": 0.42, + "grad_norm": 2.40625, + "learning_rate": 4.8755248178580705e-06, + "loss": 0.9568, + "step": 312 + }, + { + "epoch": 0.42, + "grad_norm": 2.40625, + "learning_rate": 4.874706081054743e-06, + "loss": 0.9095, + "step": 313 + }, + { + "epoch": 0.42, + "grad_norm": 2.40625, + "learning_rate": 4.873884729647075e-06, + "loss": 0.9262, + "step": 314 + }, + { + "epoch": 0.42, + "grad_norm": 2.578125, + "learning_rate": 4.873060764539393e-06, + "loss": 0.9181, + "step": 315 + }, + { + "epoch": 0.42, + "grad_norm": 2.515625, + "learning_rate": 4.872234186638898e-06, + "loss": 0.9279, + "step": 316 + }, + { + "epoch": 0.42, + "grad_norm": 2.296875, + "learning_rate": 4.871404996855674e-06, + "loss": 0.9206, + "step": 317 + }, + { + "epoch": 0.42, + "grad_norm": 2.609375, + "learning_rate": 4.870573196102677e-06, + "loss": 0.9132, + "step": 318 + }, + { + "epoch": 0.42, + "grad_norm": 2.59375, + "learning_rate": 4.869738785295737e-06, + "loss": 0.9195, + "step": 319 + }, + { + "epoch": 0.43, + "grad_norm": 2.53125, + "learning_rate": 4.868901765353559e-06, + "loss": 0.9508, + "step": 320 + }, + { + "epoch": 0.43, + "grad_norm": 2.5625, + "learning_rate": 4.8680621371977214e-06, + "loss": 0.9327, + "step": 321 + }, + { + "epoch": 0.43, + "grad_norm": 2.390625, + "learning_rate": 4.8672199017526725e-06, + "loss": 0.8962, + "step": 322 + }, + { + "epoch": 0.43, + "grad_norm": 2.453125, + "learning_rate": 4.866375059945733e-06, + "loss": 0.915, + "step": 323 + }, + { + "epoch": 0.43, + "grad_norm": 2.578125, + "learning_rate": 4.865527612707094e-06, + "loss": 0.9145, + "step": 324 + }, + { + "epoch": 0.43, + "grad_norm": 2.4375, + "learning_rate": 4.8646775609698105e-06, + "loss": 0.9195, + "step": 325 + }, + { + "epoch": 0.43, + "grad_norm": 2.453125, + "learning_rate": 4.863824905669811e-06, + "loss": 0.9137, + "step": 326 + }, + { + "epoch": 0.44, + "grad_norm": 2.515625, + "learning_rate": 4.8629696477458874e-06, + "loss": 0.9165, + "step": 327 + }, + { + "epoch": 0.44, + "grad_norm": 2.375, + "learning_rate": 4.862111788139697e-06, + "loss": 0.9248, + "step": 328 + }, + { + "epoch": 0.44, + "grad_norm": 2.40625, + "learning_rate": 4.8612513277957626e-06, + "loss": 0.9469, + "step": 329 + }, + { + "epoch": 0.44, + "grad_norm": 2.59375, + "learning_rate": 4.860388267661472e-06, + "loss": 0.9309, + "step": 330 + }, + { + "epoch": 0.44, + "grad_norm": 2.46875, + "learning_rate": 4.859522608687071e-06, + "loss": 0.907, + "step": 331 + }, + { + "epoch": 0.44, + "grad_norm": 2.5, + "learning_rate": 4.8586543518256715e-06, + "loss": 0.9261, + "step": 332 + }, + { + "epoch": 0.44, + "grad_norm": 2.625, + "learning_rate": 4.857783498033241e-06, + "loss": 0.9309, + "step": 333 + }, + { + "epoch": 0.44, + "grad_norm": 2.421875, + "learning_rate": 4.856910048268613e-06, + "loss": 0.917, + "step": 334 + }, + { + "epoch": 0.45, + "grad_norm": 2.390625, + "learning_rate": 4.856034003493471e-06, + "loss": 0.9196, + "step": 335 + }, + { + "epoch": 0.45, + "grad_norm": 2.421875, + "learning_rate": 4.855155364672363e-06, + "loss": 0.9213, + "step": 336 + }, + { + "epoch": 0.45, + "grad_norm": 2.5, + "learning_rate": 4.854274132772688e-06, + "loss": 0.9183, + "step": 337 + }, + { + "epoch": 0.45, + "grad_norm": 2.5, + "learning_rate": 4.853390308764703e-06, + "loss": 0.9113, + "step": 338 + }, + { + "epoch": 0.45, + "grad_norm": 2.703125, + "learning_rate": 4.8525038936215165e-06, + "loss": 0.9596, + "step": 339 + }, + { + "epoch": 0.45, + "grad_norm": 2.53125, + "learning_rate": 4.851614888319093e-06, + "loss": 0.9302, + "step": 340 + }, + { + "epoch": 0.45, + "grad_norm": 2.40625, + "learning_rate": 4.850723293836247e-06, + "loss": 0.9169, + "step": 341 + }, + { + "epoch": 0.46, + "grad_norm": 2.46875, + "learning_rate": 4.849829111154643e-06, + "loss": 0.933, + "step": 342 + }, + { + "epoch": 0.46, + "grad_norm": 2.578125, + "learning_rate": 4.848932341258796e-06, + "loss": 0.9312, + "step": 343 + }, + { + "epoch": 0.46, + "grad_norm": 2.4375, + "learning_rate": 4.848032985136071e-06, + "loss": 0.9061, + "step": 344 + }, + { + "epoch": 0.46, + "grad_norm": 2.5625, + "learning_rate": 4.847131043776678e-06, + "loss": 0.916, + "step": 345 + }, + { + "epoch": 0.46, + "grad_norm": 2.5625, + "learning_rate": 4.846226518173676e-06, + "loss": 0.9375, + "step": 346 + }, + { + "epoch": 0.46, + "grad_norm": 2.625, + "learning_rate": 4.845319409322965e-06, + "loss": 0.917, + "step": 347 + }, + { + "epoch": 0.46, + "grad_norm": 2.328125, + "learning_rate": 4.8444097182232944e-06, + "loss": 0.9018, + "step": 348 + }, + { + "epoch": 0.46, + "grad_norm": 2.6875, + "learning_rate": 4.843497445876254e-06, + "loss": 0.9017, + "step": 349 + }, + { + "epoch": 0.47, + "grad_norm": 2.53125, + "learning_rate": 4.842582593286275e-06, + "loss": 0.9268, + "step": 350 + }, + { + "epoch": 0.47, + "grad_norm": 2.515625, + "learning_rate": 4.841665161460632e-06, + "loss": 0.9069, + "step": 351 + }, + { + "epoch": 0.47, + "grad_norm": 2.765625, + "learning_rate": 4.840745151409437e-06, + "loss": 0.896, + "step": 352 + }, + { + "epoch": 0.47, + "grad_norm": 2.34375, + "learning_rate": 4.839822564145642e-06, + "loss": 0.9167, + "step": 353 + }, + { + "epoch": 0.47, + "grad_norm": 2.546875, + "learning_rate": 4.838897400685036e-06, + "loss": 0.9099, + "step": 354 + }, + { + "epoch": 0.47, + "grad_norm": 2.515625, + "learning_rate": 4.837969662046243e-06, + "loss": 0.9308, + "step": 355 + }, + { + "epoch": 0.47, + "grad_norm": 2.484375, + "learning_rate": 4.837039349250726e-06, + "loss": 0.9106, + "step": 356 + }, + { + "epoch": 0.48, + "grad_norm": 2.453125, + "learning_rate": 4.836106463322781e-06, + "loss": 0.887, + "step": 357 + }, + { + "epoch": 0.48, + "grad_norm": 2.4375, + "learning_rate": 4.835171005289533e-06, + "loss": 0.9053, + "step": 358 + }, + { + "epoch": 0.48, + "grad_norm": 2.4375, + "learning_rate": 4.8342329761809444e-06, + "loss": 0.9342, + "step": 359 + }, + { + "epoch": 0.48, + "grad_norm": 2.53125, + "learning_rate": 4.833292377029806e-06, + "loss": 0.9259, + "step": 360 + }, + { + "epoch": 0.48, + "grad_norm": 2.53125, + "learning_rate": 4.832349208871738e-06, + "loss": 0.9105, + "step": 361 + }, + { + "epoch": 0.48, + "grad_norm": 2.4375, + "learning_rate": 4.83140347274519e-06, + "loss": 0.9321, + "step": 362 + }, + { + "epoch": 0.48, + "grad_norm": 2.578125, + "learning_rate": 4.830455169691437e-06, + "loss": 0.8922, + "step": 363 + }, + { + "epoch": 0.48, + "grad_norm": 2.484375, + "learning_rate": 4.8295043007545836e-06, + "loss": 0.9138, + "step": 364 + }, + { + "epoch": 0.49, + "grad_norm": 2.453125, + "learning_rate": 4.828550866981557e-06, + "loss": 0.9181, + "step": 365 + }, + { + "epoch": 0.49, + "grad_norm": 2.625, + "learning_rate": 4.82759486942211e-06, + "loss": 0.9139, + "step": 366 + }, + { + "epoch": 0.49, + "grad_norm": 2.484375, + "learning_rate": 4.826636309128816e-06, + "loss": 0.9029, + "step": 367 + }, + { + "epoch": 0.49, + "grad_norm": 2.5, + "learning_rate": 4.825675187157071e-06, + "loss": 0.8999, + "step": 368 + }, + { + "epoch": 0.49, + "grad_norm": 2.65625, + "learning_rate": 4.824711504565093e-06, + "loss": 0.9349, + "step": 369 + }, + { + "epoch": 0.49, + "grad_norm": 2.421875, + "learning_rate": 4.823745262413917e-06, + "loss": 0.9283, + "step": 370 + }, + { + "epoch": 0.49, + "grad_norm": 2.453125, + "learning_rate": 4.822776461767398e-06, + "loss": 0.9031, + "step": 371 + }, + { + "epoch": 0.5, + "grad_norm": 2.359375, + "learning_rate": 4.821805103692206e-06, + "loss": 0.9443, + "step": 372 + }, + { + "epoch": 0.5, + "grad_norm": 2.53125, + "learning_rate": 4.82083118925783e-06, + "loss": 0.9348, + "step": 373 + }, + { + "epoch": 0.5, + "grad_norm": 2.4375, + "learning_rate": 4.819854719536568e-06, + "loss": 0.9355, + "step": 374 + }, + { + "epoch": 0.5, + "grad_norm": 2.453125, + "learning_rate": 4.818875695603539e-06, + "loss": 0.9098, + "step": 375 + }, + { + "epoch": 0.5, + "grad_norm": 2.359375, + "learning_rate": 4.817894118536667e-06, + "loss": 0.881, + "step": 376 + }, + { + "epoch": 0.5, + "grad_norm": 2.3125, + "learning_rate": 4.816909989416691e-06, + "loss": 0.8945, + "step": 377 + }, + { + "epoch": 0.5, + "grad_norm": 2.40625, + "learning_rate": 4.81592330932716e-06, + "loss": 0.9028, + "step": 378 + }, + { + "epoch": 0.5, + "grad_norm": 2.671875, + "learning_rate": 4.814934079354431e-06, + "loss": 0.9018, + "step": 379 + }, + { + "epoch": 0.51, + "grad_norm": 2.5, + "learning_rate": 4.813942300587666e-06, + "loss": 0.9128, + "step": 380 + }, + { + "epoch": 0.51, + "grad_norm": 2.453125, + "learning_rate": 4.812947974118839e-06, + "loss": 0.9131, + "step": 381 + }, + { + "epoch": 0.51, + "grad_norm": 2.484375, + "learning_rate": 4.811951101042722e-06, + "loss": 0.9352, + "step": 382 + }, + { + "epoch": 0.51, + "grad_norm": 2.625, + "learning_rate": 4.810951682456897e-06, + "loss": 0.9102, + "step": 383 + }, + { + "epoch": 0.51, + "grad_norm": 2.625, + "learning_rate": 4.809949719461745e-06, + "loss": 0.9172, + "step": 384 + }, + { + "epoch": 0.51, + "grad_norm": 2.546875, + "learning_rate": 4.8089452131604495e-06, + "loss": 0.9358, + "step": 385 + }, + { + "epoch": 0.51, + "grad_norm": 2.53125, + "learning_rate": 4.807938164658996e-06, + "loss": 0.8908, + "step": 386 + }, + { + "epoch": 0.51, + "grad_norm": 2.515625, + "learning_rate": 4.806928575066165e-06, + "loss": 0.9267, + "step": 387 + }, + { + "epoch": 0.52, + "grad_norm": 2.515625, + "learning_rate": 4.805916445493538e-06, + "loss": 0.9408, + "step": 388 + }, + { + "epoch": 0.52, + "grad_norm": 2.53125, + "learning_rate": 4.804901777055494e-06, + "loss": 0.9034, + "step": 389 + }, + { + "epoch": 0.52, + "grad_norm": 2.46875, + "learning_rate": 4.803884570869203e-06, + "loss": 0.9193, + "step": 390 + }, + { + "epoch": 0.52, + "grad_norm": 2.5, + "learning_rate": 4.802864828054633e-06, + "loss": 0.929, + "step": 391 + }, + { + "epoch": 0.52, + "grad_norm": 2.5625, + "learning_rate": 4.801842549734542e-06, + "loss": 0.9027, + "step": 392 + }, + { + "epoch": 0.52, + "grad_norm": 2.453125, + "learning_rate": 4.8008177370344845e-06, + "loss": 0.9434, + "step": 393 + }, + { + "epoch": 0.52, + "grad_norm": 2.4375, + "learning_rate": 4.799790391082799e-06, + "loss": 0.9008, + "step": 394 + }, + { + "epoch": 0.53, + "grad_norm": 2.53125, + "learning_rate": 4.7987605130106175e-06, + "loss": 0.8791, + "step": 395 + }, + { + "epoch": 0.53, + "grad_norm": 2.421875, + "learning_rate": 4.797728103951859e-06, + "loss": 0.8861, + "step": 396 + }, + { + "epoch": 0.53, + "grad_norm": 2.390625, + "learning_rate": 4.796693165043229e-06, + "loss": 0.9106, + "step": 397 + }, + { + "epoch": 0.53, + "grad_norm": 2.484375, + "learning_rate": 4.795655697424217e-06, + "loss": 0.9297, + "step": 398 + }, + { + "epoch": 0.53, + "grad_norm": 2.546875, + "learning_rate": 4.7946157022371e-06, + "loss": 0.9272, + "step": 399 + }, + { + "epoch": 0.53, + "grad_norm": 2.421875, + "learning_rate": 4.793573180626934e-06, + "loss": 0.9147, + "step": 400 + }, + { + "epoch": 0.53, + "grad_norm": 2.40625, + "learning_rate": 4.7925281337415584e-06, + "loss": 0.9053, + "step": 401 + }, + { + "epoch": 0.53, + "grad_norm": 2.484375, + "learning_rate": 4.791480562731595e-06, + "loss": 0.9212, + "step": 402 + }, + { + "epoch": 0.54, + "grad_norm": 2.546875, + "learning_rate": 4.790430468750442e-06, + "loss": 0.8995, + "step": 403 + }, + { + "epoch": 0.54, + "grad_norm": 2.640625, + "learning_rate": 4.789377852954276e-06, + "loss": 0.9043, + "step": 404 + }, + { + "epoch": 0.54, + "grad_norm": 2.5625, + "learning_rate": 4.788322716502051e-06, + "loss": 0.9284, + "step": 405 + }, + { + "epoch": 0.54, + "grad_norm": 2.328125, + "learning_rate": 4.787265060555495e-06, + "loss": 0.8956, + "step": 406 + }, + { + "epoch": 0.54, + "grad_norm": 2.59375, + "learning_rate": 4.78620488627911e-06, + "loss": 0.8896, + "step": 407 + }, + { + "epoch": 0.54, + "grad_norm": 2.53125, + "learning_rate": 4.785142194840174e-06, + "loss": 0.9349, + "step": 408 + }, + { + "epoch": 0.54, + "grad_norm": 2.5, + "learning_rate": 4.7840769874087325e-06, + "loss": 0.934, + "step": 409 + }, + { + "epoch": 0.55, + "grad_norm": 2.625, + "learning_rate": 4.783009265157602e-06, + "loss": 0.9395, + "step": 410 + }, + { + "epoch": 0.55, + "grad_norm": 2.59375, + "learning_rate": 4.78193902926237e-06, + "loss": 0.9383, + "step": 411 + }, + { + "epoch": 0.55, + "grad_norm": 2.390625, + "learning_rate": 4.7808662809013895e-06, + "loss": 0.8927, + "step": 412 + }, + { + "epoch": 0.55, + "grad_norm": 2.796875, + "learning_rate": 4.779791021255779e-06, + "loss": 0.927, + "step": 413 + }, + { + "epoch": 0.55, + "grad_norm": 2.578125, + "learning_rate": 4.7787132515094255e-06, + "loss": 0.8992, + "step": 414 + }, + { + "epoch": 0.55, + "grad_norm": 2.625, + "learning_rate": 4.7776329728489764e-06, + "loss": 0.9315, + "step": 415 + }, + { + "epoch": 0.55, + "grad_norm": 2.578125, + "learning_rate": 4.776550186463843e-06, + "loss": 0.9187, + "step": 416 + }, + { + "epoch": 0.55, + "grad_norm": 2.59375, + "learning_rate": 4.775464893546196e-06, + "loss": 0.906, + "step": 417 + }, + { + "epoch": 0.56, + "grad_norm": 2.515625, + "learning_rate": 4.774377095290969e-06, + "loss": 0.9046, + "step": 418 + }, + { + "epoch": 0.56, + "grad_norm": 2.515625, + "learning_rate": 4.773286792895852e-06, + "loss": 0.8962, + "step": 419 + }, + { + "epoch": 0.56, + "grad_norm": 2.421875, + "learning_rate": 4.772193987561291e-06, + "loss": 0.889, + "step": 420 + }, + { + "epoch": 0.56, + "grad_norm": 2.984375, + "learning_rate": 4.771098680490489e-06, + "loss": 0.9171, + "step": 421 + }, + { + "epoch": 0.56, + "grad_norm": 2.609375, + "learning_rate": 4.770000872889404e-06, + "loss": 0.9073, + "step": 422 + }, + { + "epoch": 0.56, + "grad_norm": 2.515625, + "learning_rate": 4.768900565966746e-06, + "loss": 0.9485, + "step": 423 + }, + { + "epoch": 0.56, + "grad_norm": 2.5625, + "learning_rate": 4.76779776093398e-06, + "loss": 0.9146, + "step": 424 + }, + { + "epoch": 0.57, + "grad_norm": 2.90625, + "learning_rate": 4.7666924590053154e-06, + "loss": 0.9302, + "step": 425 + }, + { + "epoch": 0.57, + "grad_norm": 2.5625, + "learning_rate": 4.765584661397717e-06, + "loss": 0.9182, + "step": 426 + }, + { + "epoch": 0.57, + "grad_norm": 2.515625, + "learning_rate": 4.764474369330893e-06, + "loss": 0.9143, + "step": 427 + }, + { + "epoch": 0.57, + "grad_norm": 2.5625, + "learning_rate": 4.7633615840273e-06, + "loss": 0.8779, + "step": 428 + }, + { + "epoch": 0.57, + "grad_norm": 2.59375, + "learning_rate": 4.76224630671214e-06, + "loss": 0.9096, + "step": 429 + }, + { + "epoch": 0.57, + "grad_norm": 2.4375, + "learning_rate": 4.761128538613359e-06, + "loss": 0.9154, + "step": 430 + }, + { + "epoch": 0.57, + "grad_norm": 2.640625, + "learning_rate": 4.760008280961642e-06, + "loss": 0.9437, + "step": 431 + }, + { + "epoch": 0.57, + "grad_norm": 2.625, + "learning_rate": 4.758885534990419e-06, + "loss": 0.8816, + "step": 432 + }, + { + "epoch": 0.58, + "grad_norm": 2.46875, + "learning_rate": 4.75776030193586e-06, + "loss": 0.9069, + "step": 433 + }, + { + "epoch": 0.58, + "grad_norm": 2.421875, + "learning_rate": 4.75663258303687e-06, + "loss": 0.9146, + "step": 434 + }, + { + "epoch": 0.58, + "grad_norm": 2.640625, + "learning_rate": 4.755502379535092e-06, + "loss": 0.9347, + "step": 435 + }, + { + "epoch": 0.58, + "grad_norm": 2.4375, + "learning_rate": 4.754369692674906e-06, + "loss": 0.9009, + "step": 436 + }, + { + "epoch": 0.58, + "grad_norm": 2.53125, + "learning_rate": 4.753234523703426e-06, + "loss": 0.9114, + "step": 437 + }, + { + "epoch": 0.58, + "grad_norm": 2.484375, + "learning_rate": 4.752096873870499e-06, + "loss": 0.879, + "step": 438 + }, + { + "epoch": 0.58, + "grad_norm": 2.328125, + "learning_rate": 4.7509567444287e-06, + "loss": 0.8999, + "step": 439 + }, + { + "epoch": 0.59, + "grad_norm": 2.4375, + "learning_rate": 4.7498141366333395e-06, + "loss": 0.9447, + "step": 440 + }, + { + "epoch": 0.59, + "grad_norm": 2.765625, + "learning_rate": 4.748669051742454e-06, + "loss": 0.8989, + "step": 441 + }, + { + "epoch": 0.59, + "grad_norm": 2.625, + "learning_rate": 4.747521491016805e-06, + "loss": 0.8934, + "step": 442 + }, + { + "epoch": 0.59, + "grad_norm": 2.578125, + "learning_rate": 4.746371455719886e-06, + "loss": 0.9109, + "step": 443 + }, + { + "epoch": 0.59, + "grad_norm": 2.46875, + "learning_rate": 4.74521894711791e-06, + "loss": 0.8961, + "step": 444 + }, + { + "epoch": 0.59, + "grad_norm": 2.59375, + "learning_rate": 4.744063966479813e-06, + "loss": 0.9153, + "step": 445 + }, + { + "epoch": 0.59, + "grad_norm": 2.421875, + "learning_rate": 4.742906515077257e-06, + "loss": 0.9015, + "step": 446 + }, + { + "epoch": 0.59, + "grad_norm": 2.390625, + "learning_rate": 4.74174659418462e-06, + "loss": 0.8945, + "step": 447 + }, + { + "epoch": 0.6, + "grad_norm": 2.5, + "learning_rate": 4.740584205079002e-06, + "loss": 0.898, + "step": 448 + }, + { + "epoch": 0.6, + "grad_norm": 2.546875, + "learning_rate": 4.7394193490402185e-06, + "loss": 0.8939, + "step": 449 + }, + { + "epoch": 0.6, + "grad_norm": 2.375, + "learning_rate": 4.738252027350803e-06, + "loss": 0.9275, + "step": 450 + }, + { + "epoch": 0.6, + "grad_norm": 2.875, + "learning_rate": 4.737082241296001e-06, + "loss": 0.8992, + "step": 451 + }, + { + "epoch": 0.6, + "grad_norm": 2.46875, + "learning_rate": 4.735909992163774e-06, + "loss": 0.9385, + "step": 452 + }, + { + "epoch": 0.6, + "grad_norm": 2.59375, + "learning_rate": 4.734735281244795e-06, + "loss": 0.933, + "step": 453 + }, + { + "epoch": 0.6, + "grad_norm": 2.765625, + "learning_rate": 4.7335581098324465e-06, + "loss": 0.9252, + "step": 454 + }, + { + "epoch": 0.61, + "grad_norm": 2.71875, + "learning_rate": 4.73237847922282e-06, + "loss": 0.9365, + "step": 455 + }, + { + "epoch": 0.61, + "grad_norm": 2.53125, + "learning_rate": 4.731196390714715e-06, + "loss": 0.8771, + "step": 456 + }, + { + "epoch": 0.61, + "grad_norm": 2.5625, + "learning_rate": 4.7300118456096385e-06, + "loss": 0.9012, + "step": 457 + }, + { + "epoch": 0.61, + "grad_norm": 2.453125, + "learning_rate": 4.7288248452118e-06, + "loss": 0.9167, + "step": 458 + }, + { + "epoch": 0.61, + "grad_norm": 2.4375, + "learning_rate": 4.727635390828114e-06, + "loss": 0.9254, + "step": 459 + }, + { + "epoch": 0.61, + "grad_norm": 2.703125, + "learning_rate": 4.726443483768195e-06, + "loss": 0.9191, + "step": 460 + }, + { + "epoch": 0.61, + "grad_norm": 2.546875, + "learning_rate": 4.725249125344361e-06, + "loss": 0.9075, + "step": 461 + }, + { + "epoch": 0.61, + "grad_norm": 2.375, + "learning_rate": 4.724052316871626e-06, + "loss": 0.9084, + "step": 462 + }, + { + "epoch": 0.62, + "grad_norm": 2.671875, + "learning_rate": 4.722853059667704e-06, + "loss": 0.893, + "step": 463 + }, + { + "epoch": 0.62, + "grad_norm": 2.640625, + "learning_rate": 4.721651355053003e-06, + "loss": 0.9535, + "step": 464 + }, + { + "epoch": 0.62, + "grad_norm": 2.46875, + "learning_rate": 4.720447204350627e-06, + "loss": 0.9241, + "step": 465 + }, + { + "epoch": 0.62, + "grad_norm": 2.90625, + "learning_rate": 4.719240608886372e-06, + "loss": 0.9065, + "step": 466 + }, + { + "epoch": 0.62, + "grad_norm": 2.65625, + "learning_rate": 4.718031569988727e-06, + "loss": 0.9284, + "step": 467 + }, + { + "epoch": 0.62, + "grad_norm": 2.53125, + "learning_rate": 4.716820088988871e-06, + "loss": 0.8855, + "step": 468 + }, + { + "epoch": 0.62, + "grad_norm": 3.0625, + "learning_rate": 4.715606167220672e-06, + "loss": 0.9286, + "step": 469 + }, + { + "epoch": 0.63, + "grad_norm": 2.90625, + "learning_rate": 4.7143898060206835e-06, + "loss": 0.8848, + "step": 470 + }, + { + "epoch": 0.63, + "grad_norm": 2.53125, + "learning_rate": 4.713171006728148e-06, + "loss": 0.8955, + "step": 471 + }, + { + "epoch": 0.63, + "grad_norm": 2.625, + "learning_rate": 4.711949770684989e-06, + "loss": 0.9493, + "step": 472 + }, + { + "epoch": 0.63, + "grad_norm": 2.859375, + "learning_rate": 4.710726099235816e-06, + "loss": 0.8894, + "step": 473 + }, + { + "epoch": 0.63, + "grad_norm": 2.515625, + "learning_rate": 4.709499993727919e-06, + "loss": 0.8956, + "step": 474 + }, + { + "epoch": 0.63, + "grad_norm": 2.59375, + "learning_rate": 4.7082714555112655e-06, + "loss": 0.9407, + "step": 475 + }, + { + "epoch": 0.63, + "grad_norm": 2.875, + "learning_rate": 4.707040485938503e-06, + "loss": 0.9198, + "step": 476 + }, + { + "epoch": 0.63, + "grad_norm": 2.453125, + "learning_rate": 4.705807086364961e-06, + "loss": 0.9193, + "step": 477 + }, + { + "epoch": 0.64, + "grad_norm": 2.578125, + "learning_rate": 4.704571258148634e-06, + "loss": 0.9127, + "step": 478 + }, + { + "epoch": 0.64, + "grad_norm": 2.765625, + "learning_rate": 4.7033330026502015e-06, + "loss": 0.8927, + "step": 479 + }, + { + "epoch": 0.64, + "grad_norm": 2.421875, + "learning_rate": 4.702092321233008e-06, + "loss": 0.8932, + "step": 480 + }, + { + "epoch": 0.64, + "grad_norm": 2.453125, + "learning_rate": 4.700849215263071e-06, + "loss": 0.8915, + "step": 481 + }, + { + "epoch": 0.64, + "grad_norm": 2.59375, + "learning_rate": 4.699603686109078e-06, + "loss": 0.9166, + "step": 482 + }, + { + "epoch": 0.64, + "grad_norm": 2.65625, + "learning_rate": 4.698355735142386e-06, + "loss": 0.8853, + "step": 483 + }, + { + "epoch": 0.64, + "grad_norm": 2.53125, + "learning_rate": 4.697105363737015e-06, + "loss": 0.9071, + "step": 484 + }, + { + "epoch": 0.65, + "grad_norm": 2.546875, + "learning_rate": 4.695852573269652e-06, + "loss": 0.9292, + "step": 485 + }, + { + "epoch": 0.65, + "grad_norm": 2.625, + "learning_rate": 4.694597365119649e-06, + "loss": 0.9058, + "step": 486 + }, + { + "epoch": 0.65, + "grad_norm": 2.4375, + "learning_rate": 4.6933397406690165e-06, + "loss": 0.8937, + "step": 487 + }, + { + "epoch": 0.65, + "grad_norm": 2.515625, + "learning_rate": 4.692079701302427e-06, + "loss": 0.9029, + "step": 488 + }, + { + "epoch": 0.65, + "grad_norm": 2.546875, + "learning_rate": 4.690817248407213e-06, + "loss": 0.9258, + "step": 489 + }, + { + "epoch": 0.65, + "grad_norm": 2.453125, + "learning_rate": 4.689552383373362e-06, + "loss": 0.9275, + "step": 490 + }, + { + "epoch": 0.65, + "grad_norm": 2.4375, + "learning_rate": 4.688285107593519e-06, + "loss": 0.8913, + "step": 491 + }, + { + "epoch": 0.65, + "grad_norm": 2.421875, + "learning_rate": 4.6870154224629836e-06, + "loss": 0.9133, + "step": 492 + }, + { + "epoch": 0.66, + "grad_norm": 2.390625, + "learning_rate": 4.685743329379708e-06, + "loss": 0.8976, + "step": 493 + }, + { + "epoch": 0.66, + "grad_norm": 2.328125, + "learning_rate": 4.684468829744293e-06, + "loss": 0.8927, + "step": 494 + }, + { + "epoch": 0.66, + "grad_norm": 2.34375, + "learning_rate": 4.683191924959992e-06, + "loss": 0.8796, + "step": 495 + }, + { + "epoch": 0.66, + "grad_norm": 2.4375, + "learning_rate": 4.681912616432707e-06, + "loss": 0.9008, + "step": 496 + }, + { + "epoch": 0.66, + "grad_norm": 2.34375, + "learning_rate": 4.680630905570983e-06, + "loss": 0.8976, + "step": 497 + }, + { + "epoch": 0.66, + "grad_norm": 2.4375, + "learning_rate": 4.679346793786014e-06, + "loss": 0.8863, + "step": 498 + }, + { + "epoch": 0.66, + "grad_norm": 2.390625, + "learning_rate": 4.678060282491633e-06, + "loss": 0.9357, + "step": 499 + }, + { + "epoch": 0.67, + "grad_norm": 2.53125, + "learning_rate": 4.6767713731043215e-06, + "loss": 0.9006, + "step": 500 + }, + { + "epoch": 0.67, + "grad_norm": 2.328125, + "learning_rate": 4.6754800670431955e-06, + "loss": 0.9202, + "step": 501 + }, + { + "epoch": 0.67, + "grad_norm": 2.515625, + "learning_rate": 4.674186365730012e-06, + "loss": 0.9398, + "step": 502 + }, + { + "epoch": 0.67, + "grad_norm": 2.40625, + "learning_rate": 4.672890270589163e-06, + "loss": 0.9239, + "step": 503 + }, + { + "epoch": 0.67, + "grad_norm": 2.390625, + "learning_rate": 4.671591783047682e-06, + "loss": 0.8904, + "step": 504 + }, + { + "epoch": 0.67, + "grad_norm": 2.421875, + "learning_rate": 4.67029090453523e-06, + "loss": 0.8968, + "step": 505 + }, + { + "epoch": 0.67, + "grad_norm": 2.40625, + "learning_rate": 4.668987636484103e-06, + "loss": 0.8959, + "step": 506 + }, + { + "epoch": 0.67, + "grad_norm": 2.4375, + "learning_rate": 4.66768198032923e-06, + "loss": 0.8865, + "step": 507 + }, + { + "epoch": 0.68, + "grad_norm": 2.421875, + "learning_rate": 4.666373937508166e-06, + "loss": 0.9079, + "step": 508 + }, + { + "epoch": 0.68, + "grad_norm": 2.40625, + "learning_rate": 4.665063509461098e-06, + "loss": 0.9102, + "step": 509 + }, + { + "epoch": 0.68, + "grad_norm": 2.484375, + "learning_rate": 4.663750697630832e-06, + "loss": 0.9287, + "step": 510 + }, + { + "epoch": 0.68, + "grad_norm": 2.53125, + "learning_rate": 4.662435503462806e-06, + "loss": 0.8943, + "step": 511 + }, + { + "epoch": 0.68, + "grad_norm": 2.53125, + "learning_rate": 4.661117928405079e-06, + "loss": 0.9043, + "step": 512 + }, + { + "epoch": 0.68, + "grad_norm": 2.46875, + "learning_rate": 4.659797973908328e-06, + "loss": 0.9049, + "step": 513 + }, + { + "epoch": 0.68, + "grad_norm": 2.65625, + "learning_rate": 4.658475641425854e-06, + "loss": 0.9117, + "step": 514 + }, + { + "epoch": 0.69, + "grad_norm": 2.5625, + "learning_rate": 4.6571509324135735e-06, + "loss": 0.9125, + "step": 515 + }, + { + "epoch": 0.69, + "grad_norm": 2.4375, + "learning_rate": 4.6558238483300215e-06, + "loss": 0.8802, + "step": 516 + }, + { + "epoch": 0.69, + "grad_norm": 2.5, + "learning_rate": 4.654494390636347e-06, + "loss": 0.9083, + "step": 517 + }, + { + "epoch": 0.69, + "grad_norm": 2.578125, + "learning_rate": 4.653162560796313e-06, + "loss": 0.9107, + "step": 518 + }, + { + "epoch": 0.69, + "grad_norm": 2.359375, + "learning_rate": 4.6518283602762925e-06, + "loss": 0.8955, + "step": 519 + }, + { + "epoch": 0.69, + "grad_norm": 2.5625, + "learning_rate": 4.6504917905452705e-06, + "loss": 0.8993, + "step": 520 + }, + { + "epoch": 0.69, + "grad_norm": 2.453125, + "learning_rate": 4.649152853074841e-06, + "loss": 0.9319, + "step": 521 + }, + { + "epoch": 0.69, + "grad_norm": 2.359375, + "learning_rate": 4.647811549339203e-06, + "loss": 0.9099, + "step": 522 + }, + { + "epoch": 0.7, + "grad_norm": 2.59375, + "learning_rate": 4.646467880815163e-06, + "loss": 0.9443, + "step": 523 + }, + { + "epoch": 0.7, + "grad_norm": 2.46875, + "learning_rate": 4.645121848982128e-06, + "loss": 0.9181, + "step": 524 + }, + { + "epoch": 0.7, + "grad_norm": 2.4375, + "learning_rate": 4.643773455322112e-06, + "loss": 0.9079, + "step": 525 + }, + { + "epoch": 0.7, + "grad_norm": 2.53125, + "learning_rate": 4.6424227013197235e-06, + "loss": 0.8995, + "step": 526 + }, + { + "epoch": 0.7, + "grad_norm": 2.65625, + "learning_rate": 4.641069588462174e-06, + "loss": 0.884, + "step": 527 + }, + { + "epoch": 0.7, + "grad_norm": 2.421875, + "learning_rate": 4.639714118239271e-06, + "loss": 0.9278, + "step": 528 + }, + { + "epoch": 0.7, + "grad_norm": 2.75, + "learning_rate": 4.638356292143417e-06, + "loss": 0.906, + "step": 529 + }, + { + "epoch": 0.71, + "grad_norm": 2.640625, + "learning_rate": 4.636996111669609e-06, + "loss": 0.9004, + "step": 530 + }, + { + "epoch": 0.71, + "grad_norm": 2.296875, + "learning_rate": 4.635633578315437e-06, + "loss": 0.9069, + "step": 531 + }, + { + "epoch": 0.71, + "grad_norm": 2.390625, + "learning_rate": 4.6342686935810795e-06, + "loss": 0.8993, + "step": 532 + }, + { + "epoch": 0.71, + "grad_norm": 2.625, + "learning_rate": 4.632901458969306e-06, + "loss": 0.9308, + "step": 533 + }, + { + "epoch": 0.71, + "grad_norm": 2.578125, + "learning_rate": 4.631531875985471e-06, + "loss": 0.9474, + "step": 534 + }, + { + "epoch": 0.71, + "grad_norm": 2.390625, + "learning_rate": 4.630159946137517e-06, + "loss": 0.9289, + "step": 535 + }, + { + "epoch": 0.71, + "grad_norm": 2.546875, + "learning_rate": 4.628785670935968e-06, + "loss": 0.8797, + "step": 536 + }, + { + "epoch": 0.71, + "grad_norm": 2.5, + "learning_rate": 4.627409051893934e-06, + "loss": 0.9047, + "step": 537 + }, + { + "epoch": 0.72, + "grad_norm": 2.5, + "learning_rate": 4.6260300905271e-06, + "loss": 0.9052, + "step": 538 + }, + { + "epoch": 0.72, + "grad_norm": 2.578125, + "learning_rate": 4.624648788353736e-06, + "loss": 0.9036, + "step": 539 + }, + { + "epoch": 0.72, + "grad_norm": 2.453125, + "learning_rate": 4.623265146894685e-06, + "loss": 0.9168, + "step": 540 + }, + { + "epoch": 0.72, + "grad_norm": 2.59375, + "learning_rate": 4.6218791676733676e-06, + "loss": 0.8978, + "step": 541 + }, + { + "epoch": 0.72, + "grad_norm": 2.59375, + "learning_rate": 4.620490852215779e-06, + "loss": 0.9158, + "step": 542 + }, + { + "epoch": 0.72, + "grad_norm": 2.5, + "learning_rate": 4.619100202050482e-06, + "loss": 0.8922, + "step": 543 + }, + { + "epoch": 0.72, + "grad_norm": 2.34375, + "learning_rate": 4.617707218708617e-06, + "loss": 0.8691, + "step": 544 + }, + { + "epoch": 0.73, + "grad_norm": 2.40625, + "learning_rate": 4.6163119037238894e-06, + "loss": 0.9018, + "step": 545 + }, + { + "epoch": 0.73, + "grad_norm": 2.734375, + "learning_rate": 4.6149142586325705e-06, + "loss": 0.9409, + "step": 546 + }, + { + "epoch": 0.73, + "grad_norm": 2.453125, + "learning_rate": 4.613514284973498e-06, + "loss": 0.908, + "step": 547 + }, + { + "epoch": 0.73, + "grad_norm": 2.40625, + "learning_rate": 4.6121119842880765e-06, + "loss": 0.929, + "step": 548 + }, + { + "epoch": 0.73, + "grad_norm": 2.515625, + "learning_rate": 4.610707358120269e-06, + "loss": 0.9262, + "step": 549 + }, + { + "epoch": 0.73, + "grad_norm": 2.609375, + "learning_rate": 4.6093004080166e-06, + "loss": 0.936, + "step": 550 + }, + { + "epoch": 0.73, + "grad_norm": 2.515625, + "learning_rate": 4.607891135526155e-06, + "loss": 0.9007, + "step": 551 + }, + { + "epoch": 0.73, + "grad_norm": 2.484375, + "learning_rate": 4.606479542200572e-06, + "loss": 0.9127, + "step": 552 + }, + { + "epoch": 0.74, + "grad_norm": 2.5, + "learning_rate": 4.605065629594048e-06, + "loss": 0.8838, + "step": 553 + }, + { + "epoch": 0.74, + "grad_norm": 2.5625, + "learning_rate": 4.603649399263331e-06, + "loss": 0.9013, + "step": 554 + }, + { + "epoch": 0.74, + "grad_norm": 2.453125, + "learning_rate": 4.602230852767724e-06, + "loss": 0.9162, + "step": 555 + }, + { + "epoch": 0.74, + "grad_norm": 2.46875, + "learning_rate": 4.600809991669076e-06, + "loss": 0.8751, + "step": 556 + }, + { + "epoch": 0.74, + "grad_norm": 2.609375, + "learning_rate": 4.599386817531789e-06, + "loss": 0.8815, + "step": 557 + }, + { + "epoch": 0.74, + "grad_norm": 2.453125, + "learning_rate": 4.597961331922809e-06, + "loss": 0.8804, + "step": 558 + }, + { + "epoch": 0.74, + "grad_norm": 2.375, + "learning_rate": 4.596533536411626e-06, + "loss": 0.8846, + "step": 559 + }, + { + "epoch": 0.75, + "grad_norm": 2.484375, + "learning_rate": 4.595103432570277e-06, + "loss": 0.876, + "step": 560 + }, + { + "epoch": 0.75, + "grad_norm": 2.453125, + "learning_rate": 4.593671021973336e-06, + "loss": 0.8982, + "step": 561 + }, + { + "epoch": 0.75, + "grad_norm": 2.40625, + "learning_rate": 4.59223630619792e-06, + "loss": 0.9138, + "step": 562 + }, + { + "epoch": 0.75, + "grad_norm": 2.453125, + "learning_rate": 4.590799286823683e-06, + "loss": 0.9065, + "step": 563 + }, + { + "epoch": 0.75, + "grad_norm": 2.5, + "learning_rate": 4.589359965432816e-06, + "loss": 0.8892, + "step": 564 + }, + { + "epoch": 0.75, + "grad_norm": 2.40625, + "learning_rate": 4.587918343610042e-06, + "loss": 0.9372, + "step": 565 + }, + { + "epoch": 0.75, + "grad_norm": 2.46875, + "learning_rate": 4.58647442294262e-06, + "loss": 0.8768, + "step": 566 + }, + { + "epoch": 0.75, + "grad_norm": 2.40625, + "learning_rate": 4.585028205020341e-06, + "loss": 0.8886, + "step": 567 + }, + { + "epoch": 0.76, + "grad_norm": 2.484375, + "learning_rate": 4.5835796914355195e-06, + "loss": 0.9188, + "step": 568 + }, + { + "epoch": 0.76, + "grad_norm": 2.484375, + "learning_rate": 4.582128883783004e-06, + "loss": 0.8921, + "step": 569 + }, + { + "epoch": 0.76, + "grad_norm": 2.578125, + "learning_rate": 4.580675783660163e-06, + "loss": 0.9031, + "step": 570 + }, + { + "epoch": 0.76, + "grad_norm": 2.484375, + "learning_rate": 4.579220392666896e-06, + "loss": 0.8932, + "step": 571 + }, + { + "epoch": 0.76, + "grad_norm": 2.453125, + "learning_rate": 4.577762712405618e-06, + "loss": 0.9398, + "step": 572 + }, + { + "epoch": 0.76, + "grad_norm": 2.578125, + "learning_rate": 4.576302744481268e-06, + "loss": 0.9324, + "step": 573 + }, + { + "epoch": 0.76, + "grad_norm": 2.421875, + "learning_rate": 4.5748404905013045e-06, + "loss": 0.8908, + "step": 574 + }, + { + "epoch": 0.77, + "grad_norm": 2.515625, + "learning_rate": 4.5733759520757e-06, + "loss": 0.9067, + "step": 575 + }, + { + "epoch": 0.77, + "grad_norm": 2.515625, + "learning_rate": 4.5719091308169464e-06, + "loss": 0.8771, + "step": 576 + }, + { + "epoch": 0.77, + "grad_norm": 2.265625, + "learning_rate": 4.570440028340044e-06, + "loss": 0.8843, + "step": 577 + }, + { + "epoch": 0.77, + "grad_norm": 2.578125, + "learning_rate": 4.568968646262509e-06, + "loss": 0.8934, + "step": 578 + }, + { + "epoch": 0.77, + "grad_norm": 2.53125, + "learning_rate": 4.567494986204365e-06, + "loss": 0.945, + "step": 579 + }, + { + "epoch": 0.77, + "grad_norm": 2.453125, + "learning_rate": 4.5660190497881455e-06, + "loss": 0.8931, + "step": 580 + }, + { + "epoch": 0.77, + "grad_norm": 2.421875, + "learning_rate": 4.5645408386388875e-06, + "loss": 0.9008, + "step": 581 + }, + { + "epoch": 0.77, + "grad_norm": 2.390625, + "learning_rate": 4.563060354384137e-06, + "loss": 0.8965, + "step": 582 + }, + { + "epoch": 0.78, + "grad_norm": 2.328125, + "learning_rate": 4.561577598653938e-06, + "loss": 0.8701, + "step": 583 + }, + { + "epoch": 0.78, + "grad_norm": 2.359375, + "learning_rate": 4.560092573080839e-06, + "loss": 0.8996, + "step": 584 + }, + { + "epoch": 0.78, + "grad_norm": 2.40625, + "learning_rate": 4.558605279299886e-06, + "loss": 0.8921, + "step": 585 + }, + { + "epoch": 0.78, + "grad_norm": 2.375, + "learning_rate": 4.557115718948622e-06, + "loss": 0.8921, + "step": 586 + }, + { + "epoch": 0.78, + "grad_norm": 2.5625, + "learning_rate": 4.555623893667087e-06, + "loss": 0.892, + "step": 587 + }, + { + "epoch": 0.78, + "grad_norm": 2.375, + "learning_rate": 4.554129805097813e-06, + "loss": 0.9122, + "step": 588 + }, + { + "epoch": 0.78, + "grad_norm": 2.421875, + "learning_rate": 4.552633454885826e-06, + "loss": 0.9063, + "step": 589 + }, + { + "epoch": 0.79, + "grad_norm": 2.59375, + "learning_rate": 4.55113484467864e-06, + "loss": 0.9161, + "step": 590 + }, + { + "epoch": 0.79, + "grad_norm": 2.375, + "learning_rate": 4.549633976126259e-06, + "loss": 0.8789, + "step": 591 + }, + { + "epoch": 0.79, + "grad_norm": 2.3125, + "learning_rate": 4.548130850881171e-06, + "loss": 0.883, + "step": 592 + }, + { + "epoch": 0.79, + "grad_norm": 2.53125, + "learning_rate": 4.546625470598353e-06, + "loss": 0.8527, + "step": 593 + }, + { + "epoch": 0.79, + "grad_norm": 2.53125, + "learning_rate": 4.54511783693526e-06, + "loss": 0.9305, + "step": 594 + }, + { + "epoch": 0.79, + "grad_norm": 2.40625, + "learning_rate": 4.5436079515518315e-06, + "loss": 0.9114, + "step": 595 + }, + { + "epoch": 0.79, + "grad_norm": 2.359375, + "learning_rate": 4.542095816110485e-06, + "loss": 0.9278, + "step": 596 + }, + { + "epoch": 0.79, + "grad_norm": 2.421875, + "learning_rate": 4.540581432276114e-06, + "loss": 0.8865, + "step": 597 + }, + { + "epoch": 0.8, + "grad_norm": 2.40625, + "learning_rate": 4.53906480171609e-06, + "loss": 0.8821, + "step": 598 + }, + { + "epoch": 0.8, + "grad_norm": 2.265625, + "learning_rate": 4.5375459261002575e-06, + "loss": 0.856, + "step": 599 + }, + { + "epoch": 0.8, + "grad_norm": 2.5625, + "learning_rate": 4.536024807100932e-06, + "loss": 0.9159, + "step": 600 + }, + { + "epoch": 0.8, + "grad_norm": 2.421875, + "learning_rate": 4.534501446392898e-06, + "loss": 0.8963, + "step": 601 + }, + { + "epoch": 0.8, + "grad_norm": 2.421875, + "learning_rate": 4.532975845653412e-06, + "loss": 0.8925, + "step": 602 + }, + { + "epoch": 0.8, + "grad_norm": 2.375, + "learning_rate": 4.531448006562194e-06, + "loss": 0.9035, + "step": 603 + }, + { + "epoch": 0.8, + "grad_norm": 2.59375, + "learning_rate": 4.529917930801427e-06, + "loss": 0.9121, + "step": 604 + }, + { + "epoch": 0.81, + "grad_norm": 2.484375, + "learning_rate": 4.52838562005576e-06, + "loss": 0.9022, + "step": 605 + }, + { + "epoch": 0.81, + "grad_norm": 2.4375, + "learning_rate": 4.5268510760123e-06, + "loss": 0.8698, + "step": 606 + }, + { + "epoch": 0.81, + "grad_norm": 2.59375, + "learning_rate": 4.525314300360617e-06, + "loss": 0.9267, + "step": 607 + }, + { + "epoch": 0.81, + "grad_norm": 2.5, + "learning_rate": 4.523775294792732e-06, + "loss": 0.916, + "step": 608 + }, + { + "epoch": 0.81, + "grad_norm": 2.390625, + "learning_rate": 4.5222340610031275e-06, + "loss": 0.8951, + "step": 609 + }, + { + "epoch": 0.81, + "grad_norm": 2.40625, + "learning_rate": 4.520690600688734e-06, + "loss": 0.9054, + "step": 610 + }, + { + "epoch": 0.81, + "grad_norm": 2.46875, + "learning_rate": 4.519144915548937e-06, + "loss": 0.8976, + "step": 611 + }, + { + "epoch": 0.81, + "grad_norm": 2.421875, + "learning_rate": 4.517597007285568e-06, + "loss": 0.9012, + "step": 612 + }, + { + "epoch": 0.82, + "grad_norm": 2.71875, + "learning_rate": 4.5160468776029114e-06, + "loss": 0.9214, + "step": 613 + }, + { + "epoch": 0.82, + "grad_norm": 2.65625, + "learning_rate": 4.5144945282076935e-06, + "loss": 0.907, + "step": 614 + }, + { + "epoch": 0.82, + "grad_norm": 2.4375, + "learning_rate": 4.512939960809085e-06, + "loss": 0.9226, + "step": 615 + }, + { + "epoch": 0.82, + "grad_norm": 3.65625, + "learning_rate": 4.5113831771187e-06, + "loss": 0.9008, + "step": 616 + }, + { + "epoch": 0.82, + "grad_norm": 2.75, + "learning_rate": 4.509824178850591e-06, + "loss": 0.8953, + "step": 617 + }, + { + "epoch": 0.82, + "grad_norm": 2.3125, + "learning_rate": 4.508262967721252e-06, + "loss": 0.8876, + "step": 618 + }, + { + "epoch": 0.82, + "grad_norm": 2.578125, + "learning_rate": 4.506699545449608e-06, + "loss": 0.9004, + "step": 619 + }, + { + "epoch": 0.83, + "grad_norm": 2.484375, + "learning_rate": 4.505133913757026e-06, + "loss": 0.9125, + "step": 620 + }, + { + "epoch": 0.83, + "grad_norm": 2.390625, + "learning_rate": 4.503566074367298e-06, + "loss": 0.9274, + "step": 621 + }, + { + "epoch": 0.83, + "grad_norm": 2.578125, + "learning_rate": 4.501996029006651e-06, + "loss": 0.911, + "step": 622 + }, + { + "epoch": 0.83, + "grad_norm": 2.53125, + "learning_rate": 4.50042377940374e-06, + "loss": 0.9026, + "step": 623 + }, + { + "epoch": 0.83, + "grad_norm": 2.390625, + "learning_rate": 4.498849327289647e-06, + "loss": 0.9238, + "step": 624 + }, + { + "epoch": 0.83, + "grad_norm": 2.71875, + "learning_rate": 4.497272674397879e-06, + "loss": 0.8924, + "step": 625 + }, + { + "epoch": 0.83, + "grad_norm": 2.453125, + "learning_rate": 4.495693822464366e-06, + "loss": 0.9082, + "step": 626 + }, + { + "epoch": 0.83, + "grad_norm": 2.40625, + "learning_rate": 4.49411277322746e-06, + "loss": 0.9023, + "step": 627 + }, + { + "epoch": 0.84, + "grad_norm": 2.53125, + "learning_rate": 4.492529528427929e-06, + "loss": 0.9072, + "step": 628 + }, + { + "epoch": 0.84, + "grad_norm": 2.34375, + "learning_rate": 4.490944089808964e-06, + "loss": 0.8998, + "step": 629 + }, + { + "epoch": 0.84, + "grad_norm": 2.40625, + "learning_rate": 4.489356459116165e-06, + "loss": 0.9152, + "step": 630 + }, + { + "epoch": 0.84, + "grad_norm": 2.46875, + "learning_rate": 4.487766638097551e-06, + "loss": 0.8784, + "step": 631 + }, + { + "epoch": 0.84, + "grad_norm": 2.484375, + "learning_rate": 4.48617462850355e-06, + "loss": 0.8893, + "step": 632 + }, + { + "epoch": 0.84, + "grad_norm": 2.65625, + "learning_rate": 4.484580432086999e-06, + "loss": 0.8815, + "step": 633 + }, + { + "epoch": 0.84, + "grad_norm": 2.5, + "learning_rate": 4.4829840506031455e-06, + "loss": 0.8862, + "step": 634 + }, + { + "epoch": 0.84, + "grad_norm": 2.4375, + "learning_rate": 4.48138548580964e-06, + "loss": 0.9293, + "step": 635 + }, + { + "epoch": 0.85, + "grad_norm": 2.421875, + "learning_rate": 4.479784739466539e-06, + "loss": 0.9128, + "step": 636 + }, + { + "epoch": 0.85, + "grad_norm": 2.421875, + "learning_rate": 4.478181813336297e-06, + "loss": 0.8722, + "step": 637 + }, + { + "epoch": 0.85, + "grad_norm": 2.375, + "learning_rate": 4.476576709183775e-06, + "loss": 0.9293, + "step": 638 + }, + { + "epoch": 0.85, + "grad_norm": 2.40625, + "learning_rate": 4.474969428776228e-06, + "loss": 0.9013, + "step": 639 + }, + { + "epoch": 0.85, + "grad_norm": 2.390625, + "learning_rate": 4.473359973883305e-06, + "loss": 0.929, + "step": 640 + }, + { + "epoch": 0.85, + "grad_norm": 2.296875, + "learning_rate": 4.471748346277054e-06, + "loss": 0.8905, + "step": 641 + }, + { + "epoch": 0.85, + "grad_norm": 2.3125, + "learning_rate": 4.470134547731912e-06, + "loss": 0.8828, + "step": 642 + }, + { + "epoch": 0.86, + "grad_norm": 2.578125, + "learning_rate": 4.468518580024708e-06, + "loss": 0.9458, + "step": 643 + }, + { + "epoch": 0.86, + "grad_norm": 2.359375, + "learning_rate": 4.466900444934657e-06, + "loss": 0.9242, + "step": 644 + }, + { + "epoch": 0.86, + "grad_norm": 2.53125, + "learning_rate": 4.465280144243363e-06, + "loss": 0.8903, + "step": 645 + }, + { + "epoch": 0.86, + "grad_norm": 2.4375, + "learning_rate": 4.463657679734813e-06, + "loss": 0.8833, + "step": 646 + }, + { + "epoch": 0.86, + "grad_norm": 2.5625, + "learning_rate": 4.462033053195376e-06, + "loss": 0.908, + "step": 647 + }, + { + "epoch": 0.86, + "grad_norm": 2.546875, + "learning_rate": 4.460406266413803e-06, + "loss": 0.8984, + "step": 648 + }, + { + "epoch": 0.86, + "grad_norm": 2.46875, + "learning_rate": 4.458777321181222e-06, + "loss": 0.9178, + "step": 649 + }, + { + "epoch": 0.86, + "grad_norm": 2.390625, + "learning_rate": 4.457146219291137e-06, + "loss": 0.8878, + "step": 650 + }, + { + "epoch": 0.87, + "grad_norm": 2.484375, + "learning_rate": 4.45551296253943e-06, + "loss": 0.9174, + "step": 651 + }, + { + "epoch": 0.87, + "grad_norm": 2.453125, + "learning_rate": 4.453877552724352e-06, + "loss": 0.9021, + "step": 652 + }, + { + "epoch": 0.87, + "grad_norm": 2.40625, + "learning_rate": 4.452239991646524e-06, + "loss": 0.8848, + "step": 653 + }, + { + "epoch": 0.87, + "grad_norm": 2.40625, + "learning_rate": 4.450600281108939e-06, + "loss": 0.883, + "step": 654 + }, + { + "epoch": 0.87, + "grad_norm": 2.5, + "learning_rate": 4.448958422916955e-06, + "loss": 0.9285, + "step": 655 + }, + { + "epoch": 0.87, + "grad_norm": 2.546875, + "learning_rate": 4.447314418878294e-06, + "loss": 0.9376, + "step": 656 + }, + { + "epoch": 0.87, + "grad_norm": 2.453125, + "learning_rate": 4.4456682708030405e-06, + "loss": 0.9148, + "step": 657 + }, + { + "epoch": 0.88, + "grad_norm": 2.421875, + "learning_rate": 4.444019980503641e-06, + "loss": 0.9486, + "step": 658 + }, + { + "epoch": 0.88, + "grad_norm": 2.359375, + "learning_rate": 4.4423695497949e-06, + "loss": 0.8893, + "step": 659 + }, + { + "epoch": 0.88, + "grad_norm": 2.421875, + "learning_rate": 4.440716980493978e-06, + "loss": 0.9185, + "step": 660 + }, + { + "epoch": 0.88, + "grad_norm": 2.421875, + "learning_rate": 4.439062274420391e-06, + "loss": 0.9164, + "step": 661 + }, + { + "epoch": 0.88, + "grad_norm": 2.40625, + "learning_rate": 4.437405433396006e-06, + "loss": 0.9074, + "step": 662 + }, + { + "epoch": 0.88, + "grad_norm": 2.421875, + "learning_rate": 4.435746459245044e-06, + "loss": 0.9056, + "step": 663 + }, + { + "epoch": 0.88, + "grad_norm": 2.5, + "learning_rate": 4.4340853537940715e-06, + "loss": 0.919, + "step": 664 + }, + { + "epoch": 0.88, + "grad_norm": 2.5, + "learning_rate": 4.432422118872003e-06, + "loss": 0.9421, + "step": 665 + }, + { + "epoch": 0.89, + "grad_norm": 2.40625, + "learning_rate": 4.430756756310096e-06, + "loss": 0.8764, + "step": 666 + }, + { + "epoch": 0.89, + "grad_norm": 2.453125, + "learning_rate": 4.429089267941953e-06, + "loss": 0.8854, + "step": 667 + }, + { + "epoch": 0.89, + "grad_norm": 2.453125, + "learning_rate": 4.427419655603515e-06, + "loss": 0.9039, + "step": 668 + }, + { + "epoch": 0.89, + "grad_norm": 2.421875, + "learning_rate": 4.425747921133064e-06, + "loss": 0.8807, + "step": 669 + }, + { + "epoch": 0.89, + "grad_norm": 2.296875, + "learning_rate": 4.424074066371216e-06, + "loss": 0.8806, + "step": 670 + }, + { + "epoch": 0.89, + "grad_norm": 2.53125, + "learning_rate": 4.422398093160921e-06, + "loss": 0.896, + "step": 671 + }, + { + "epoch": 0.89, + "grad_norm": 2.453125, + "learning_rate": 4.420720003347465e-06, + "loss": 0.9087, + "step": 672 + }, + { + "epoch": 0.9, + "grad_norm": 2.40625, + "learning_rate": 4.419039798778463e-06, + "loss": 0.8918, + "step": 673 + }, + { + "epoch": 0.9, + "grad_norm": 2.40625, + "learning_rate": 4.417357481303854e-06, + "loss": 0.8981, + "step": 674 + }, + { + "epoch": 0.9, + "grad_norm": 2.5, + "learning_rate": 4.41567305277591e-06, + "loss": 0.89, + "step": 675 + }, + { + "epoch": 0.9, + "grad_norm": 2.390625, + "learning_rate": 4.4139865150492235e-06, + "loss": 0.9133, + "step": 676 + }, + { + "epoch": 0.9, + "grad_norm": 2.484375, + "learning_rate": 4.412297869980711e-06, + "loss": 0.905, + "step": 677 + }, + { + "epoch": 0.9, + "grad_norm": 2.359375, + "learning_rate": 4.4106071194296065e-06, + "loss": 0.9136, + "step": 678 + }, + { + "epoch": 0.9, + "grad_norm": 2.453125, + "learning_rate": 4.408914265257466e-06, + "loss": 0.9163, + "step": 679 + }, + { + "epoch": 0.9, + "grad_norm": 2.46875, + "learning_rate": 4.407219309328159e-06, + "loss": 0.8835, + "step": 680 + }, + { + "epoch": 0.91, + "grad_norm": 2.421875, + "learning_rate": 4.405522253507869e-06, + "loss": 0.883, + "step": 681 + }, + { + "epoch": 0.91, + "grad_norm": 2.4375, + "learning_rate": 4.403823099665093e-06, + "loss": 0.9026, + "step": 682 + }, + { + "epoch": 0.91, + "grad_norm": 2.6875, + "learning_rate": 4.402121849670637e-06, + "loss": 0.8959, + "step": 683 + }, + { + "epoch": 0.91, + "grad_norm": 2.359375, + "learning_rate": 4.400418505397616e-06, + "loss": 0.8907, + "step": 684 + }, + { + "epoch": 0.91, + "grad_norm": 2.390625, + "learning_rate": 4.398713068721449e-06, + "loss": 0.9037, + "step": 685 + }, + { + "epoch": 0.91, + "grad_norm": 2.4375, + "learning_rate": 4.39700554151986e-06, + "loss": 0.8921, + "step": 686 + }, + { + "epoch": 0.91, + "grad_norm": 2.375, + "learning_rate": 4.395295925672874e-06, + "loss": 0.9063, + "step": 687 + }, + { + "epoch": 0.92, + "grad_norm": 2.453125, + "learning_rate": 4.393584223062819e-06, + "loss": 0.8844, + "step": 688 + }, + { + "epoch": 0.92, + "grad_norm": 2.421875, + "learning_rate": 4.3918704355743144e-06, + "loss": 0.8801, + "step": 689 + }, + { + "epoch": 0.92, + "grad_norm": 2.375, + "learning_rate": 4.390154565094279e-06, + "loss": 0.8886, + "step": 690 + }, + { + "epoch": 0.92, + "grad_norm": 2.40625, + "learning_rate": 4.388436613511926e-06, + "loss": 0.8975, + "step": 691 + }, + { + "epoch": 0.92, + "grad_norm": 2.484375, + "learning_rate": 4.3867165827187575e-06, + "loss": 0.899, + "step": 692 + }, + { + "epoch": 0.92, + "grad_norm": 2.390625, + "learning_rate": 4.384994474608565e-06, + "loss": 0.8694, + "step": 693 + }, + { + "epoch": 0.92, + "grad_norm": 2.375, + "learning_rate": 4.38327029107743e-06, + "loss": 0.8856, + "step": 694 + }, + { + "epoch": 0.92, + "grad_norm": 2.28125, + "learning_rate": 4.381544034023713e-06, + "loss": 0.8712, + "step": 695 + }, + { + "epoch": 0.93, + "grad_norm": 2.59375, + "learning_rate": 4.3798157053480675e-06, + "loss": 0.9077, + "step": 696 + }, + { + "epoch": 0.93, + "grad_norm": 2.46875, + "learning_rate": 4.378085306953417e-06, + "loss": 0.8955, + "step": 697 + }, + { + "epoch": 0.93, + "grad_norm": 2.390625, + "learning_rate": 4.37635284074497e-06, + "loss": 0.9213, + "step": 698 + }, + { + "epoch": 0.93, + "grad_norm": 2.46875, + "learning_rate": 4.374618308630211e-06, + "loss": 0.8937, + "step": 699 + }, + { + "epoch": 0.93, + "grad_norm": 2.453125, + "learning_rate": 4.372881712518898e-06, + "loss": 0.8883, + "step": 700 + }, + { + "epoch": 0.93, + "grad_norm": 2.390625, + "learning_rate": 4.371143054323063e-06, + "loss": 0.8791, + "step": 701 + }, + { + "epoch": 0.93, + "grad_norm": 2.578125, + "learning_rate": 4.369402335957006e-06, + "loss": 0.9023, + "step": 702 + }, + { + "epoch": 0.94, + "grad_norm": 2.40625, + "learning_rate": 4.367659559337299e-06, + "loss": 0.9068, + "step": 703 + }, + { + "epoch": 0.94, + "grad_norm": 2.40625, + "learning_rate": 4.365914726382774e-06, + "loss": 0.902, + "step": 704 + }, + { + "epoch": 0.94, + "grad_norm": 2.421875, + "learning_rate": 4.364167839014535e-06, + "loss": 0.9101, + "step": 705 + }, + { + "epoch": 0.94, + "grad_norm": 2.546875, + "learning_rate": 4.362418899155941e-06, + "loss": 0.9329, + "step": 706 + }, + { + "epoch": 0.94, + "grad_norm": 2.359375, + "learning_rate": 4.360667908732616e-06, + "loss": 0.9207, + "step": 707 + }, + { + "epoch": 0.94, + "grad_norm": 2.4375, + "learning_rate": 4.3589148696724395e-06, + "loss": 0.9197, + "step": 708 + }, + { + "epoch": 0.94, + "grad_norm": 2.5, + "learning_rate": 4.357159783905545e-06, + "loss": 0.9076, + "step": 709 + }, + { + "epoch": 0.94, + "grad_norm": 2.546875, + "learning_rate": 4.355402653364321e-06, + "loss": 0.9011, + "step": 710 + }, + { + "epoch": 0.95, + "grad_norm": 2.390625, + "learning_rate": 4.353643479983408e-06, + "loss": 0.8802, + "step": 711 + }, + { + "epoch": 0.95, + "grad_norm": 2.34375, + "learning_rate": 4.351882265699696e-06, + "loss": 0.888, + "step": 712 + }, + { + "epoch": 0.95, + "grad_norm": 2.453125, + "learning_rate": 4.350119012452321e-06, + "loss": 0.8981, + "step": 713 + }, + { + "epoch": 0.95, + "grad_norm": 2.53125, + "learning_rate": 4.348353722182662e-06, + "loss": 0.8978, + "step": 714 + }, + { + "epoch": 0.95, + "grad_norm": 2.53125, + "learning_rate": 4.3465863968343455e-06, + "loss": 0.949, + "step": 715 + }, + { + "epoch": 0.95, + "grad_norm": 2.46875, + "learning_rate": 4.344817038353235e-06, + "loss": 0.9404, + "step": 716 + }, + { + "epoch": 0.95, + "grad_norm": 2.40625, + "learning_rate": 4.343045648687434e-06, + "loss": 0.8982, + "step": 717 + }, + { + "epoch": 0.96, + "grad_norm": 2.5, + "learning_rate": 4.341272229787281e-06, + "loss": 0.8961, + "step": 718 + }, + { + "epoch": 0.96, + "grad_norm": 2.328125, + "learning_rate": 4.339496783605351e-06, + "loss": 0.8729, + "step": 719 + }, + { + "epoch": 0.96, + "grad_norm": 2.484375, + "learning_rate": 4.33771931209645e-06, + "loss": 0.9091, + "step": 720 + }, + { + "epoch": 0.96, + "grad_norm": 2.484375, + "learning_rate": 4.335939817217612e-06, + "loss": 0.9266, + "step": 721 + }, + { + "epoch": 0.96, + "grad_norm": 2.453125, + "learning_rate": 4.334158300928102e-06, + "loss": 0.8649, + "step": 722 + }, + { + "epoch": 0.96, + "grad_norm": 2.453125, + "learning_rate": 4.332374765189409e-06, + "loss": 0.8755, + "step": 723 + }, + { + "epoch": 0.96, + "grad_norm": 2.46875, + "learning_rate": 4.330589211965246e-06, + "loss": 0.9158, + "step": 724 + }, + { + "epoch": 0.96, + "grad_norm": 2.4375, + "learning_rate": 4.328801643221546e-06, + "loss": 0.9189, + "step": 725 + }, + { + "epoch": 0.97, + "grad_norm": 2.4375, + "learning_rate": 4.327012060926462e-06, + "loss": 0.8881, + "step": 726 + }, + { + "epoch": 0.97, + "grad_norm": 2.4375, + "learning_rate": 4.325220467050364e-06, + "loss": 0.8659, + "step": 727 + }, + { + "epoch": 0.97, + "grad_norm": 2.625, + "learning_rate": 4.323426863565838e-06, + "loss": 0.8801, + "step": 728 + }, + { + "epoch": 0.97, + "grad_norm": 2.46875, + "learning_rate": 4.32163125244768e-06, + "loss": 0.9503, + "step": 729 + }, + { + "epoch": 0.97, + "grad_norm": 2.515625, + "learning_rate": 4.319833635672899e-06, + "loss": 0.8773, + "step": 730 + }, + { + "epoch": 0.97, + "grad_norm": 2.296875, + "learning_rate": 4.318034015220711e-06, + "loss": 0.9003, + "step": 731 + }, + { + "epoch": 0.97, + "grad_norm": 2.421875, + "learning_rate": 4.316232393072538e-06, + "loss": 0.9126, + "step": 732 + }, + { + "epoch": 0.98, + "grad_norm": 2.484375, + "learning_rate": 4.314428771212005e-06, + "loss": 0.9077, + "step": 733 + }, + { + "epoch": 0.98, + "grad_norm": 2.4375, + "learning_rate": 4.3126231516249416e-06, + "loss": 0.9059, + "step": 734 + }, + { + "epoch": 0.98, + "grad_norm": 2.25, + "learning_rate": 4.310815536299374e-06, + "loss": 0.9006, + "step": 735 + }, + { + "epoch": 0.98, + "grad_norm": 2.6875, + "learning_rate": 4.309005927225528e-06, + "loss": 0.9217, + "step": 736 + }, + { + "epoch": 0.98, + "grad_norm": 2.46875, + "learning_rate": 4.307194326395823e-06, + "loss": 0.9118, + "step": 737 + }, + { + "epoch": 0.98, + "grad_norm": 2.453125, + "learning_rate": 4.3053807358048715e-06, + "loss": 0.9227, + "step": 738 + }, + { + "epoch": 0.98, + "grad_norm": 2.5, + "learning_rate": 4.303565157449478e-06, + "loss": 0.8995, + "step": 739 + }, + { + "epoch": 0.98, + "grad_norm": 2.5, + "learning_rate": 4.301747593328634e-06, + "loss": 0.9246, + "step": 740 + }, + { + "epoch": 0.99, + "grad_norm": 2.359375, + "learning_rate": 4.299928045443518e-06, + "loss": 0.9216, + "step": 741 + }, + { + "epoch": 0.99, + "grad_norm": 2.3125, + "learning_rate": 4.2981065157974955e-06, + "loss": 0.892, + "step": 742 + }, + { + "epoch": 0.99, + "grad_norm": 2.375, + "learning_rate": 4.296283006396108e-06, + "loss": 0.875, + "step": 743 + }, + { + "epoch": 0.99, + "grad_norm": 2.421875, + "learning_rate": 4.29445751924708e-06, + "loss": 0.9035, + "step": 744 + }, + { + "epoch": 0.99, + "grad_norm": 2.390625, + "learning_rate": 4.292630056360316e-06, + "loss": 0.9139, + "step": 745 + }, + { + "epoch": 0.99, + "grad_norm": 2.328125, + "learning_rate": 4.2908006197478914e-06, + "loss": 0.9201, + "step": 746 + }, + { + "epoch": 0.99, + "grad_norm": 2.3125, + "learning_rate": 4.288969211424057e-06, + "loss": 0.9056, + "step": 747 + }, + { + "epoch": 1.0, + "grad_norm": 2.328125, + "learning_rate": 4.287135833405235e-06, + "loss": 0.8909, + "step": 748 + }, + { + "epoch": 1.0, + "grad_norm": 2.4375, + "learning_rate": 4.285300487710014e-06, + "loss": 0.8985, + "step": 749 + }, + { + "epoch": 1.0, + "grad_norm": 2.375, + "learning_rate": 4.283463176359153e-06, + "loss": 0.8911, + "step": 750 + }, + { + "epoch": 1.0, + "grad_norm": 2.34375, + "learning_rate": 4.28162390137557e-06, + "loss": 0.8813, + "step": 751 + }, + { + "epoch": 1.0, + "grad_norm": 2.359375, + "learning_rate": 4.27978266478435e-06, + "loss": 0.9032, + "step": 752 + }, + { + "epoch": 1.0, + "grad_norm": 2.375, + "learning_rate": 4.277939468612733e-06, + "loss": 0.8798, + "step": 753 + }, + { + "epoch": 1.0, + "grad_norm": 2.375, + "learning_rate": 4.276094314890122e-06, + "loss": 0.8991, + "step": 754 + }, + { + "epoch": 1.0, + "grad_norm": 2.328125, + "learning_rate": 4.274247205648071e-06, + "loss": 0.8888, + "step": 755 + }, + { + "epoch": 1.01, + "grad_norm": 2.46875, + "learning_rate": 4.272398142920288e-06, + "loss": 0.9216, + "step": 756 + }, + { + "epoch": 1.01, + "grad_norm": 2.34375, + "learning_rate": 4.2705471287426335e-06, + "loss": 0.8979, + "step": 757 + }, + { + "epoch": 1.01, + "grad_norm": 2.4375, + "learning_rate": 4.268694165153113e-06, + "loss": 0.9152, + "step": 758 + }, + { + "epoch": 1.01, + "grad_norm": 2.375, + "learning_rate": 4.266839254191884e-06, + "loss": 0.8924, + "step": 759 + }, + { + "epoch": 1.01, + "grad_norm": 2.4375, + "learning_rate": 4.2649823979012424e-06, + "loss": 0.8965, + "step": 760 + }, + { + "epoch": 1.01, + "grad_norm": 2.359375, + "learning_rate": 4.26312359832563e-06, + "loss": 0.8747, + "step": 761 + }, + { + "epoch": 1.01, + "grad_norm": 2.46875, + "learning_rate": 4.261262857511625e-06, + "loss": 0.8674, + "step": 762 + }, + { + "epoch": 1.0, + "grad_norm": 2.53125, + "learning_rate": 4.259400177507946e-06, + "loss": 0.8954, + "step": 763 + }, + { + "epoch": 1.0, + "grad_norm": 3.546875, + "learning_rate": 4.257535560365444e-06, + "loss": 0.8222, + "step": 764 + }, + { + "epoch": 1.0, + "grad_norm": 2.921875, + "learning_rate": 4.255669008137104e-06, + "loss": 0.822, + "step": 765 + }, + { + "epoch": 1.0, + "grad_norm": 2.640625, + "learning_rate": 4.253800522878043e-06, + "loss": 0.8229, + "step": 766 + }, + { + "epoch": 1.01, + "grad_norm": 2.953125, + "learning_rate": 4.251930106645503e-06, + "loss": 0.8289, + "step": 767 + }, + { + "epoch": 1.01, + "grad_norm": 3.125, + "learning_rate": 4.2500577614988556e-06, + "loss": 0.816, + "step": 768 + }, + { + "epoch": 1.01, + "grad_norm": 3.28125, + "learning_rate": 4.248183489499595e-06, + "loss": 0.823, + "step": 769 + }, + { + "epoch": 1.01, + "grad_norm": 2.890625, + "learning_rate": 4.246307292711335e-06, + "loss": 0.841, + "step": 770 + }, + { + "epoch": 1.01, + "grad_norm": 2.65625, + "learning_rate": 4.24442917319981e-06, + "loss": 0.8112, + "step": 771 + }, + { + "epoch": 1.01, + "grad_norm": 2.953125, + "learning_rate": 4.242549133032872e-06, + "loss": 0.8366, + "step": 772 + }, + { + "epoch": 1.01, + "grad_norm": 2.765625, + "learning_rate": 4.240667174280487e-06, + "loss": 0.8047, + "step": 773 + }, + { + "epoch": 1.01, + "grad_norm": 2.765625, + "learning_rate": 4.238783299014735e-06, + "loss": 0.8208, + "step": 774 + }, + { + "epoch": 1.02, + "grad_norm": 2.6875, + "learning_rate": 4.2368975093098025e-06, + "loss": 0.8348, + "step": 775 + }, + { + "epoch": 1.02, + "grad_norm": 2.65625, + "learning_rate": 4.235009807241986e-06, + "loss": 0.8111, + "step": 776 + }, + { + "epoch": 1.02, + "grad_norm": 2.90625, + "learning_rate": 4.23312019488969e-06, + "loss": 0.7845, + "step": 777 + }, + { + "epoch": 1.02, + "grad_norm": 3.140625, + "learning_rate": 4.2312286743334174e-06, + "loss": 0.7992, + "step": 778 + }, + { + "epoch": 1.02, + "grad_norm": 2.796875, + "learning_rate": 4.229335247655776e-06, + "loss": 0.8103, + "step": 779 + }, + { + "epoch": 1.02, + "grad_norm": 2.46875, + "learning_rate": 4.227439916941469e-06, + "loss": 0.8441, + "step": 780 + }, + { + "epoch": 1.02, + "grad_norm": 2.609375, + "learning_rate": 4.225542684277301e-06, + "loss": 0.8202, + "step": 781 + }, + { + "epoch": 1.03, + "grad_norm": 2.796875, + "learning_rate": 4.223643551752163e-06, + "loss": 0.8045, + "step": 782 + }, + { + "epoch": 1.03, + "grad_norm": 2.71875, + "learning_rate": 4.221742521457046e-06, + "loss": 0.8551, + "step": 783 + }, + { + "epoch": 1.03, + "grad_norm": 2.578125, + "learning_rate": 4.219839595485026e-06, + "loss": 0.8282, + "step": 784 + }, + { + "epoch": 1.03, + "grad_norm": 2.6875, + "learning_rate": 4.217934775931265e-06, + "loss": 0.8071, + "step": 785 + }, + { + "epoch": 1.03, + "grad_norm": 2.59375, + "learning_rate": 4.216028064893014e-06, + "loss": 0.7794, + "step": 786 + }, + { + "epoch": 1.03, + "grad_norm": 2.71875, + "learning_rate": 4.214119464469604e-06, + "loss": 0.806, + "step": 787 + }, + { + "epoch": 1.03, + "grad_norm": 2.640625, + "learning_rate": 4.2122089767624445e-06, + "loss": 0.8096, + "step": 788 + }, + { + "epoch": 1.03, + "grad_norm": 2.59375, + "learning_rate": 4.210296603875027e-06, + "loss": 0.8227, + "step": 789 + }, + { + "epoch": 1.04, + "grad_norm": 2.578125, + "learning_rate": 4.2083823479129175e-06, + "loss": 0.8047, + "step": 790 + }, + { + "epoch": 1.04, + "grad_norm": 2.5, + "learning_rate": 4.206466210983753e-06, + "loss": 0.8075, + "step": 791 + }, + { + "epoch": 1.04, + "grad_norm": 2.625, + "learning_rate": 4.204548195197242e-06, + "loss": 0.8026, + "step": 792 + }, + { + "epoch": 1.04, + "grad_norm": 2.671875, + "learning_rate": 4.202628302665166e-06, + "loss": 0.8334, + "step": 793 + }, + { + "epoch": 1.04, + "grad_norm": 2.609375, + "learning_rate": 4.200706535501366e-06, + "loss": 0.8273, + "step": 794 + }, + { + "epoch": 1.04, + "grad_norm": 2.546875, + "learning_rate": 4.198782895821752e-06, + "loss": 0.8258, + "step": 795 + }, + { + "epoch": 1.04, + "grad_norm": 2.671875, + "learning_rate": 4.196857385744295e-06, + "loss": 0.8185, + "step": 796 + }, + { + "epoch": 1.05, + "grad_norm": 2.59375, + "learning_rate": 4.194930007389023e-06, + "loss": 0.8018, + "step": 797 + }, + { + "epoch": 1.05, + "grad_norm": 2.4375, + "learning_rate": 4.193000762878024e-06, + "loss": 0.8059, + "step": 798 + }, + { + "epoch": 1.05, + "grad_norm": 2.59375, + "learning_rate": 4.1910696543354376e-06, + "loss": 0.7944, + "step": 799 + }, + { + "epoch": 1.05, + "grad_norm": 2.59375, + "learning_rate": 4.189136683887458e-06, + "loss": 0.8324, + "step": 800 + }, + { + "epoch": 1.05, + "grad_norm": 2.484375, + "learning_rate": 4.187201853662329e-06, + "loss": 0.7853, + "step": 801 + }, + { + "epoch": 1.05, + "grad_norm": 2.703125, + "learning_rate": 4.185265165790343e-06, + "loss": 0.7916, + "step": 802 + }, + { + "epoch": 1.05, + "grad_norm": 2.546875, + "learning_rate": 4.183326622403834e-06, + "loss": 0.8006, + "step": 803 + }, + { + "epoch": 1.05, + "grad_norm": 2.671875, + "learning_rate": 4.181386225637184e-06, + "loss": 0.8456, + "step": 804 + }, + { + "epoch": 1.06, + "grad_norm": 2.515625, + "learning_rate": 4.17944397762681e-06, + "loss": 0.8082, + "step": 805 + }, + { + "epoch": 1.06, + "grad_norm": 2.625, + "learning_rate": 4.177499880511174e-06, + "loss": 0.8155, + "step": 806 + }, + { + "epoch": 1.06, + "grad_norm": 2.65625, + "learning_rate": 4.175553936430767e-06, + "loss": 0.8252, + "step": 807 + }, + { + "epoch": 1.06, + "grad_norm": 2.6875, + "learning_rate": 4.17360614752812e-06, + "loss": 0.8191, + "step": 808 + }, + { + "epoch": 1.06, + "grad_norm": 2.5625, + "learning_rate": 4.171656515947789e-06, + "loss": 0.8209, + "step": 809 + }, + { + "epoch": 1.06, + "grad_norm": 2.703125, + "learning_rate": 4.1697050438363636e-06, + "loss": 0.8352, + "step": 810 + }, + { + "epoch": 1.06, + "grad_norm": 2.671875, + "learning_rate": 4.167751733342457e-06, + "loss": 0.8071, + "step": 811 + }, + { + "epoch": 1.07, + "grad_norm": 2.515625, + "learning_rate": 4.1657965866167095e-06, + "loss": 0.8218, + "step": 812 + }, + { + "epoch": 1.07, + "grad_norm": 2.78125, + "learning_rate": 4.16383960581178e-06, + "loss": 0.8485, + "step": 813 + }, + { + "epoch": 1.07, + "grad_norm": 2.734375, + "learning_rate": 4.161880793082348e-06, + "loss": 0.8104, + "step": 814 + }, + { + "epoch": 1.07, + "grad_norm": 2.5625, + "learning_rate": 4.1599201505851086e-06, + "loss": 0.8344, + "step": 815 + }, + { + "epoch": 1.07, + "grad_norm": 2.5625, + "learning_rate": 4.157957680478776e-06, + "loss": 0.8074, + "step": 816 + }, + { + "epoch": 1.07, + "grad_norm": 2.71875, + "learning_rate": 4.155993384924071e-06, + "loss": 0.8307, + "step": 817 + }, + { + "epoch": 1.07, + "grad_norm": 2.8125, + "learning_rate": 4.1540272660837275e-06, + "loss": 0.8108, + "step": 818 + }, + { + "epoch": 1.07, + "grad_norm": 2.65625, + "learning_rate": 4.1520593261224865e-06, + "loss": 0.8325, + "step": 819 + }, + { + "epoch": 1.08, + "grad_norm": 2.546875, + "learning_rate": 4.150089567207094e-06, + "loss": 0.8165, + "step": 820 + }, + { + "epoch": 1.08, + "grad_norm": 2.78125, + "learning_rate": 4.148117991506298e-06, + "loss": 0.8331, + "step": 821 + }, + { + "epoch": 1.08, + "grad_norm": 2.4375, + "learning_rate": 4.146144601190847e-06, + "loss": 0.7845, + "step": 822 + }, + { + "epoch": 1.08, + "grad_norm": 2.625, + "learning_rate": 4.144169398433489e-06, + "loss": 0.8131, + "step": 823 + }, + { + "epoch": 1.08, + "grad_norm": 2.859375, + "learning_rate": 4.142192385408964e-06, + "loss": 0.8046, + "step": 824 + }, + { + "epoch": 1.08, + "grad_norm": 2.515625, + "learning_rate": 4.14021356429401e-06, + "loss": 0.8096, + "step": 825 + }, + { + "epoch": 1.08, + "grad_norm": 2.46875, + "learning_rate": 4.138232937267351e-06, + "loss": 0.8206, + "step": 826 + }, + { + "epoch": 1.09, + "grad_norm": 2.765625, + "learning_rate": 4.1362505065097024e-06, + "loss": 0.8366, + "step": 827 + }, + { + "epoch": 1.09, + "grad_norm": 2.71875, + "learning_rate": 4.134266274203766e-06, + "loss": 0.811, + "step": 828 + }, + { + "epoch": 1.09, + "grad_norm": 2.53125, + "learning_rate": 4.132280242534223e-06, + "loss": 0.8074, + "step": 829 + }, + { + "epoch": 1.09, + "grad_norm": 2.625, + "learning_rate": 4.130292413687742e-06, + "loss": 0.8448, + "step": 830 + }, + { + "epoch": 1.09, + "grad_norm": 2.703125, + "learning_rate": 4.128302789852964e-06, + "loss": 0.8046, + "step": 831 + }, + { + "epoch": 1.09, + "grad_norm": 2.671875, + "learning_rate": 4.126311373220511e-06, + "loss": 0.8352, + "step": 832 + }, + { + "epoch": 1.09, + "grad_norm": 2.65625, + "learning_rate": 4.124318165982974e-06, + "loss": 0.8383, + "step": 833 + }, + { + "epoch": 1.09, + "grad_norm": 2.671875, + "learning_rate": 4.122323170334924e-06, + "loss": 0.8219, + "step": 834 + }, + { + "epoch": 1.1, + "grad_norm": 2.78125, + "learning_rate": 4.120326388472891e-06, + "loss": 0.7904, + "step": 835 + }, + { + "epoch": 1.1, + "grad_norm": 2.609375, + "learning_rate": 4.118327822595377e-06, + "loss": 0.8112, + "step": 836 + }, + { + "epoch": 1.1, + "grad_norm": 2.625, + "learning_rate": 4.11632747490285e-06, + "loss": 0.822, + "step": 837 + }, + { + "epoch": 1.1, + "grad_norm": 2.703125, + "learning_rate": 4.114325347597736e-06, + "loss": 0.8211, + "step": 838 + }, + { + "epoch": 1.1, + "grad_norm": 2.546875, + "learning_rate": 4.112321442884422e-06, + "loss": 0.8075, + "step": 839 + }, + { + "epoch": 1.1, + "grad_norm": 2.59375, + "learning_rate": 4.11031576296925e-06, + "loss": 0.8188, + "step": 840 + }, + { + "epoch": 1.1, + "grad_norm": 2.703125, + "learning_rate": 4.108308310060521e-06, + "loss": 0.8271, + "step": 841 + }, + { + "epoch": 1.11, + "grad_norm": 2.75, + "learning_rate": 4.106299086368484e-06, + "loss": 0.8386, + "step": 842 + }, + { + "epoch": 1.11, + "grad_norm": 2.4375, + "learning_rate": 4.10428809410534e-06, + "loss": 0.7986, + "step": 843 + }, + { + "epoch": 1.11, + "grad_norm": 2.578125, + "learning_rate": 4.102275335485234e-06, + "loss": 0.8094, + "step": 844 + }, + { + "epoch": 1.11, + "grad_norm": 2.8125, + "learning_rate": 4.1002608127242615e-06, + "loss": 0.8326, + "step": 845 + }, + { + "epoch": 1.11, + "grad_norm": 2.875, + "learning_rate": 4.098244528040455e-06, + "loss": 0.8351, + "step": 846 + }, + { + "epoch": 1.11, + "grad_norm": 2.65625, + "learning_rate": 4.096226483653788e-06, + "loss": 0.8132, + "step": 847 + }, + { + "epoch": 1.11, + "grad_norm": 2.75, + "learning_rate": 4.094206681786175e-06, + "loss": 0.8515, + "step": 848 + }, + { + "epoch": 1.11, + "grad_norm": 2.640625, + "learning_rate": 4.09218512466146e-06, + "loss": 0.8254, + "step": 849 + }, + { + "epoch": 1.12, + "grad_norm": 2.6875, + "learning_rate": 4.0901618145054246e-06, + "loss": 0.8385, + "step": 850 + }, + { + "epoch": 1.12, + "grad_norm": 2.6875, + "learning_rate": 4.088136753545777e-06, + "loss": 0.8125, + "step": 851 + }, + { + "epoch": 1.12, + "grad_norm": 2.765625, + "learning_rate": 4.086109944012156e-06, + "loss": 0.8051, + "step": 852 + }, + { + "epoch": 1.12, + "grad_norm": 2.703125, + "learning_rate": 4.084081388136123e-06, + "loss": 0.8303, + "step": 853 + }, + { + "epoch": 1.12, + "grad_norm": 2.578125, + "learning_rate": 4.082051088151163e-06, + "loss": 0.831, + "step": 854 + }, + { + "epoch": 1.12, + "grad_norm": 2.640625, + "learning_rate": 4.080019046292682e-06, + "loss": 0.8068, + "step": 855 + }, + { + "epoch": 1.12, + "grad_norm": 2.765625, + "learning_rate": 4.077985264798004e-06, + "loss": 0.8205, + "step": 856 + }, + { + "epoch": 1.13, + "grad_norm": 2.5, + "learning_rate": 4.0759497459063665e-06, + "loss": 0.8215, + "step": 857 + }, + { + "epoch": 1.13, + "grad_norm": 2.765625, + "learning_rate": 4.0739124918589225e-06, + "loss": 0.83, + "step": 858 + }, + { + "epoch": 1.13, + "grad_norm": 2.703125, + "learning_rate": 4.071873504898732e-06, + "loss": 0.8253, + "step": 859 + }, + { + "epoch": 1.13, + "grad_norm": 2.640625, + "learning_rate": 4.069832787270768e-06, + "loss": 0.8225, + "step": 860 + }, + { + "epoch": 1.13, + "grad_norm": 2.546875, + "learning_rate": 4.067790341221905e-06, + "loss": 0.8296, + "step": 861 + }, + { + "epoch": 1.13, + "grad_norm": 2.65625, + "learning_rate": 4.06574616900092e-06, + "loss": 0.8301, + "step": 862 + }, + { + "epoch": 1.13, + "grad_norm": 2.71875, + "learning_rate": 4.063700272858494e-06, + "loss": 0.8498, + "step": 863 + }, + { + "epoch": 1.13, + "grad_norm": 2.671875, + "learning_rate": 4.061652655047204e-06, + "loss": 0.8344, + "step": 864 + }, + { + "epoch": 1.14, + "grad_norm": 2.578125, + "learning_rate": 4.059603317821522e-06, + "loss": 0.7988, + "step": 865 + }, + { + "epoch": 1.14, + "grad_norm": 2.671875, + "learning_rate": 4.057552263437815e-06, + "loss": 0.8276, + "step": 866 + }, + { + "epoch": 1.14, + "grad_norm": 2.703125, + "learning_rate": 4.05549949415434e-06, + "loss": 0.8078, + "step": 867 + }, + { + "epoch": 1.14, + "grad_norm": 2.75, + "learning_rate": 4.053445012231241e-06, + "loss": 0.8088, + "step": 868 + }, + { + "epoch": 1.14, + "grad_norm": 2.75, + "learning_rate": 4.051388819930547e-06, + "loss": 0.8324, + "step": 869 + }, + { + "epoch": 1.14, + "grad_norm": 2.609375, + "learning_rate": 4.049330919516175e-06, + "loss": 0.8133, + "step": 870 + }, + { + "epoch": 1.14, + "grad_norm": 2.65625, + "learning_rate": 4.047271313253914e-06, + "loss": 0.8345, + "step": 871 + }, + { + "epoch": 1.15, + "grad_norm": 2.6875, + "learning_rate": 4.045210003411443e-06, + "loss": 0.8373, + "step": 872 + }, + { + "epoch": 1.15, + "grad_norm": 2.546875, + "learning_rate": 4.043146992258304e-06, + "loss": 0.779, + "step": 873 + }, + { + "epoch": 1.15, + "grad_norm": 2.6875, + "learning_rate": 4.041082282065922e-06, + "loss": 0.8118, + "step": 874 + }, + { + "epoch": 1.15, + "grad_norm": 2.609375, + "learning_rate": 4.039015875107588e-06, + "loss": 0.813, + "step": 875 + }, + { + "epoch": 1.15, + "grad_norm": 2.59375, + "learning_rate": 4.036947773658463e-06, + "loss": 0.8267, + "step": 876 + }, + { + "epoch": 1.15, + "grad_norm": 2.546875, + "learning_rate": 4.03487797999557e-06, + "loss": 0.8251, + "step": 877 + }, + { + "epoch": 1.15, + "grad_norm": 2.640625, + "learning_rate": 4.032806496397802e-06, + "loss": 0.7909, + "step": 878 + }, + { + "epoch": 1.15, + "grad_norm": 2.71875, + "learning_rate": 4.030733325145905e-06, + "loss": 0.8173, + "step": 879 + }, + { + "epoch": 1.16, + "grad_norm": 2.6875, + "learning_rate": 4.028658468522489e-06, + "loss": 0.8406, + "step": 880 + }, + { + "epoch": 1.16, + "grad_norm": 2.515625, + "learning_rate": 4.026581928812016e-06, + "loss": 0.8209, + "step": 881 + }, + { + "epoch": 1.16, + "grad_norm": 2.640625, + "learning_rate": 4.024503708300803e-06, + "loss": 0.8381, + "step": 882 + }, + { + "epoch": 1.16, + "grad_norm": 2.578125, + "learning_rate": 4.022423809277017e-06, + "loss": 0.8126, + "step": 883 + }, + { + "epoch": 1.16, + "grad_norm": 2.609375, + "learning_rate": 4.020342234030674e-06, + "loss": 0.812, + "step": 884 + }, + { + "epoch": 1.16, + "grad_norm": 2.765625, + "learning_rate": 4.018258984853634e-06, + "loss": 0.8324, + "step": 885 + }, + { + "epoch": 1.16, + "grad_norm": 2.703125, + "learning_rate": 4.016174064039602e-06, + "loss": 0.828, + "step": 886 + }, + { + "epoch": 1.17, + "grad_norm": 2.515625, + "learning_rate": 4.01408747388412e-06, + "loss": 0.7903, + "step": 887 + }, + { + "epoch": 1.17, + "grad_norm": 2.625, + "learning_rate": 4.0119992166845724e-06, + "loss": 0.8173, + "step": 888 + }, + { + "epoch": 1.17, + "grad_norm": 2.828125, + "learning_rate": 4.009909294740176e-06, + "loss": 0.8524, + "step": 889 + }, + { + "epoch": 1.17, + "grad_norm": 2.65625, + "learning_rate": 4.007817710351982e-06, + "loss": 0.8216, + "step": 890 + }, + { + "epoch": 1.17, + "grad_norm": 2.65625, + "learning_rate": 4.00572446582287e-06, + "loss": 0.8252, + "step": 891 + }, + { + "epoch": 1.17, + "grad_norm": 2.703125, + "learning_rate": 4.003629563457551e-06, + "loss": 0.822, + "step": 892 + }, + { + "epoch": 1.17, + "grad_norm": 2.640625, + "learning_rate": 4.001533005562558e-06, + "loss": 0.8319, + "step": 893 + }, + { + "epoch": 1.17, + "grad_norm": 2.734375, + "learning_rate": 3.999434794446247e-06, + "loss": 0.8211, + "step": 894 + }, + { + "epoch": 1.18, + "grad_norm": 2.671875, + "learning_rate": 3.997334932418797e-06, + "loss": 0.8363, + "step": 895 + }, + { + "epoch": 1.18, + "grad_norm": 2.46875, + "learning_rate": 3.995233421792203e-06, + "loss": 0.8366, + "step": 896 + }, + { + "epoch": 1.18, + "grad_norm": 2.546875, + "learning_rate": 3.993130264880273e-06, + "loss": 0.7918, + "step": 897 + }, + { + "epoch": 1.18, + "grad_norm": 2.640625, + "learning_rate": 3.991025463998632e-06, + "loss": 0.8216, + "step": 898 + }, + { + "epoch": 1.18, + "grad_norm": 2.671875, + "learning_rate": 3.988919021464711e-06, + "loss": 0.8454, + "step": 899 + }, + { + "epoch": 1.18, + "grad_norm": 2.5625, + "learning_rate": 3.986810939597752e-06, + "loss": 0.8262, + "step": 900 + }, + { + "epoch": 1.18, + "grad_norm": 2.71875, + "learning_rate": 3.984701220718797e-06, + "loss": 0.8388, + "step": 901 + }, + { + "epoch": 1.19, + "grad_norm": 2.59375, + "learning_rate": 3.982589867150698e-06, + "loss": 0.8195, + "step": 902 + }, + { + "epoch": 1.19, + "grad_norm": 2.609375, + "learning_rate": 3.980476881218098e-06, + "loss": 0.7981, + "step": 903 + }, + { + "epoch": 1.19, + "grad_norm": 2.59375, + "learning_rate": 3.978362265247444e-06, + "loss": 0.7859, + "step": 904 + }, + { + "epoch": 1.19, + "grad_norm": 2.453125, + "learning_rate": 3.976246021566975e-06, + "loss": 0.7747, + "step": 905 + }, + { + "epoch": 1.19, + "grad_norm": 2.640625, + "learning_rate": 3.9741281525067225e-06, + "loss": 0.7928, + "step": 906 + }, + { + "epoch": 1.19, + "grad_norm": 2.640625, + "learning_rate": 3.972008660398506e-06, + "loss": 0.8004, + "step": 907 + }, + { + "epoch": 1.19, + "grad_norm": 2.53125, + "learning_rate": 3.969887547575937e-06, + "loss": 0.8092, + "step": 908 + }, + { + "epoch": 1.19, + "grad_norm": 2.703125, + "learning_rate": 3.9677648163744044e-06, + "loss": 0.8324, + "step": 909 + }, + { + "epoch": 1.2, + "grad_norm": 2.625, + "learning_rate": 3.965640469131084e-06, + "loss": 0.7923, + "step": 910 + }, + { + "epoch": 1.2, + "grad_norm": 2.484375, + "learning_rate": 3.963514508184931e-06, + "loss": 0.8241, + "step": 911 + }, + { + "epoch": 1.2, + "grad_norm": 2.546875, + "learning_rate": 3.961386935876674e-06, + "loss": 0.8342, + "step": 912 + }, + { + "epoch": 1.2, + "grad_norm": 2.546875, + "learning_rate": 3.959257754548817e-06, + "loss": 0.8178, + "step": 913 + }, + { + "epoch": 1.2, + "grad_norm": 2.625, + "learning_rate": 3.957126966545638e-06, + "loss": 0.8262, + "step": 914 + }, + { + "epoch": 1.2, + "grad_norm": 2.609375, + "learning_rate": 3.954994574213181e-06, + "loss": 0.8078, + "step": 915 + }, + { + "epoch": 1.2, + "grad_norm": 2.609375, + "learning_rate": 3.952860579899257e-06, + "loss": 0.8074, + "step": 916 + }, + { + "epoch": 1.21, + "grad_norm": 2.578125, + "learning_rate": 3.950724985953443e-06, + "loss": 0.7923, + "step": 917 + }, + { + "epoch": 1.21, + "grad_norm": 2.5625, + "learning_rate": 3.948587794727074e-06, + "loss": 0.8333, + "step": 918 + }, + { + "epoch": 1.21, + "grad_norm": 2.75, + "learning_rate": 3.946449008573245e-06, + "loss": 0.8512, + "step": 919 + }, + { + "epoch": 1.21, + "grad_norm": 2.796875, + "learning_rate": 3.944308629846809e-06, + "loss": 0.8309, + "step": 920 + }, + { + "epoch": 1.21, + "grad_norm": 2.625, + "learning_rate": 3.942166660904369e-06, + "loss": 0.8321, + "step": 921 + }, + { + "epoch": 1.21, + "grad_norm": 2.640625, + "learning_rate": 3.940023104104281e-06, + "loss": 0.818, + "step": 922 + }, + { + "epoch": 1.21, + "grad_norm": 2.46875, + "learning_rate": 3.93787796180665e-06, + "loss": 0.8129, + "step": 923 + }, + { + "epoch": 1.21, + "grad_norm": 2.5625, + "learning_rate": 3.935731236373324e-06, + "loss": 0.7839, + "step": 924 + }, + { + "epoch": 1.22, + "grad_norm": 2.625, + "learning_rate": 3.933582930167896e-06, + "loss": 0.7893, + "step": 925 + }, + { + "epoch": 1.22, + "grad_norm": 2.578125, + "learning_rate": 3.9314330455557e-06, + "loss": 0.8291, + "step": 926 + }, + { + "epoch": 1.22, + "grad_norm": 2.421875, + "learning_rate": 3.929281584903805e-06, + "loss": 0.7965, + "step": 927 + }, + { + "epoch": 1.22, + "grad_norm": 2.609375, + "learning_rate": 3.9271285505810185e-06, + "loss": 0.8217, + "step": 928 + }, + { + "epoch": 1.22, + "grad_norm": 2.625, + "learning_rate": 3.924973944957879e-06, + "loss": 0.8154, + "step": 929 + }, + { + "epoch": 1.22, + "grad_norm": 2.5, + "learning_rate": 3.922817770406655e-06, + "loss": 0.819, + "step": 930 + }, + { + "epoch": 1.22, + "grad_norm": 2.578125, + "learning_rate": 3.920660029301344e-06, + "loss": 0.811, + "step": 931 + }, + { + "epoch": 1.23, + "grad_norm": 2.546875, + "learning_rate": 3.918500724017665e-06, + "loss": 0.8087, + "step": 932 + }, + { + "epoch": 1.23, + "grad_norm": 2.609375, + "learning_rate": 3.916339856933063e-06, + "loss": 0.8138, + "step": 933 + }, + { + "epoch": 1.23, + "grad_norm": 2.625, + "learning_rate": 3.9141774304267e-06, + "loss": 0.8171, + "step": 934 + }, + { + "epoch": 1.23, + "grad_norm": 2.578125, + "learning_rate": 3.912013446879456e-06, + "loss": 0.8116, + "step": 935 + }, + { + "epoch": 1.23, + "grad_norm": 2.6875, + "learning_rate": 3.909847908673925e-06, + "loss": 0.8272, + "step": 936 + }, + { + "epoch": 1.23, + "grad_norm": 2.609375, + "learning_rate": 3.907680818194414e-06, + "loss": 0.8325, + "step": 937 + }, + { + "epoch": 1.23, + "grad_norm": 2.625, + "learning_rate": 3.905512177826936e-06, + "loss": 0.8242, + "step": 938 + }, + { + "epoch": 1.23, + "grad_norm": 2.625, + "learning_rate": 3.903341989959213e-06, + "loss": 0.826, + "step": 939 + }, + { + "epoch": 1.24, + "grad_norm": 2.734375, + "learning_rate": 3.9011702569806716e-06, + "loss": 0.8128, + "step": 940 + }, + { + "epoch": 1.24, + "grad_norm": 2.578125, + "learning_rate": 3.898996981282437e-06, + "loss": 0.8206, + "step": 941 + }, + { + "epoch": 1.24, + "grad_norm": 2.671875, + "learning_rate": 3.896822165257333e-06, + "loss": 0.8, + "step": 942 + }, + { + "epoch": 1.24, + "grad_norm": 2.90625, + "learning_rate": 3.894645811299883e-06, + "loss": 0.8314, + "step": 943 + }, + { + "epoch": 1.24, + "grad_norm": 2.75, + "learning_rate": 3.8924679218062996e-06, + "loss": 0.833, + "step": 944 + }, + { + "epoch": 1.24, + "grad_norm": 2.9375, + "learning_rate": 3.890288499174489e-06, + "loss": 0.8214, + "step": 945 + }, + { + "epoch": 1.24, + "grad_norm": 2.6875, + "learning_rate": 3.888107545804043e-06, + "loss": 0.8005, + "step": 946 + }, + { + "epoch": 1.25, + "grad_norm": 2.6875, + "learning_rate": 3.8859250640962405e-06, + "loss": 0.7992, + "step": 947 + }, + { + "epoch": 1.25, + "grad_norm": 2.609375, + "learning_rate": 3.8837410564540415e-06, + "loss": 0.8503, + "step": 948 + }, + { + "epoch": 1.25, + "grad_norm": 2.890625, + "learning_rate": 3.881555525282089e-06, + "loss": 0.8427, + "step": 949 + }, + { + "epoch": 1.25, + "grad_norm": 2.734375, + "learning_rate": 3.8793684729867e-06, + "loss": 0.8024, + "step": 950 + }, + { + "epoch": 1.25, + "grad_norm": 2.703125, + "learning_rate": 3.877179901975867e-06, + "loss": 0.8241, + "step": 951 + }, + { + "epoch": 1.25, + "grad_norm": 2.71875, + "learning_rate": 3.874989814659258e-06, + "loss": 0.7807, + "step": 952 + }, + { + "epoch": 1.25, + "grad_norm": 2.75, + "learning_rate": 3.872798213448206e-06, + "loss": 0.8308, + "step": 953 + }, + { + "epoch": 1.25, + "grad_norm": 2.640625, + "learning_rate": 3.8706051007557145e-06, + "loss": 0.8326, + "step": 954 + }, + { + "epoch": 1.26, + "grad_norm": 2.71875, + "learning_rate": 3.868410478996448e-06, + "loss": 0.8609, + "step": 955 + }, + { + "epoch": 1.26, + "grad_norm": 2.765625, + "learning_rate": 3.866214350586734e-06, + "loss": 0.8083, + "step": 956 + }, + { + "epoch": 1.26, + "grad_norm": 2.515625, + "learning_rate": 3.86401671794456e-06, + "loss": 0.8167, + "step": 957 + }, + { + "epoch": 1.26, + "grad_norm": 2.6875, + "learning_rate": 3.861817583489566e-06, + "loss": 0.8056, + "step": 958 + }, + { + "epoch": 1.26, + "grad_norm": 2.53125, + "learning_rate": 3.859616949643053e-06, + "loss": 0.8398, + "step": 959 + }, + { + "epoch": 1.26, + "grad_norm": 2.703125, + "learning_rate": 3.857414818827962e-06, + "loss": 0.853, + "step": 960 + }, + { + "epoch": 1.26, + "grad_norm": 2.5625, + "learning_rate": 3.855211193468892e-06, + "loss": 0.8132, + "step": 961 + }, + { + "epoch": 1.27, + "grad_norm": 2.5625, + "learning_rate": 3.853006075992083e-06, + "loss": 0.8086, + "step": 962 + }, + { + "epoch": 1.27, + "grad_norm": 2.46875, + "learning_rate": 3.850799468825417e-06, + "loss": 0.8358, + "step": 963 + }, + { + "epoch": 1.27, + "grad_norm": 2.625, + "learning_rate": 3.848591374398421e-06, + "loss": 0.8309, + "step": 964 + }, + { + "epoch": 1.27, + "grad_norm": 2.578125, + "learning_rate": 3.846381795142251e-06, + "loss": 0.8213, + "step": 965 + }, + { + "epoch": 1.27, + "grad_norm": 2.5625, + "learning_rate": 3.844170733489708e-06, + "loss": 0.843, + "step": 966 + }, + { + "epoch": 1.27, + "grad_norm": 2.515625, + "learning_rate": 3.841958191875218e-06, + "loss": 0.8042, + "step": 967 + }, + { + "epoch": 1.27, + "grad_norm": 2.578125, + "learning_rate": 3.839744172734839e-06, + "loss": 0.7924, + "step": 968 + }, + { + "epoch": 1.27, + "grad_norm": 2.546875, + "learning_rate": 3.837528678506254e-06, + "loss": 0.8327, + "step": 969 + }, + { + "epoch": 1.28, + "grad_norm": 2.734375, + "learning_rate": 3.835311711628774e-06, + "loss": 0.8036, + "step": 970 + }, + { + "epoch": 1.28, + "grad_norm": 2.703125, + "learning_rate": 3.833093274543327e-06, + "loss": 0.8395, + "step": 971 + }, + { + "epoch": 1.28, + "grad_norm": 2.671875, + "learning_rate": 3.8308733696924624e-06, + "loss": 0.8254, + "step": 972 + }, + { + "epoch": 1.28, + "grad_norm": 2.671875, + "learning_rate": 3.828651999520346e-06, + "loss": 0.8127, + "step": 973 + }, + { + "epoch": 1.28, + "grad_norm": 2.625, + "learning_rate": 3.826429166472755e-06, + "loss": 0.8083, + "step": 974 + }, + { + "epoch": 1.28, + "grad_norm": 2.734375, + "learning_rate": 3.824204872997077e-06, + "loss": 0.8192, + "step": 975 + }, + { + "epoch": 1.28, + "grad_norm": 2.5625, + "learning_rate": 3.82197912154231e-06, + "loss": 0.8254, + "step": 976 + }, + { + "epoch": 1.29, + "grad_norm": 2.578125, + "learning_rate": 3.819751914559056e-06, + "loss": 0.83, + "step": 977 + }, + { + "epoch": 1.29, + "grad_norm": 2.703125, + "learning_rate": 3.817523254499519e-06, + "loss": 0.8629, + "step": 978 + }, + { + "epoch": 1.29, + "grad_norm": 2.65625, + "learning_rate": 3.815293143817504e-06, + "loss": 0.8154, + "step": 979 + }, + { + "epoch": 1.29, + "grad_norm": 2.71875, + "learning_rate": 3.8130615849684114e-06, + "loss": 0.8196, + "step": 980 + }, + { + "epoch": 1.29, + "grad_norm": 2.5625, + "learning_rate": 3.81082858040924e-06, + "loss": 0.8432, + "step": 981 + }, + { + "epoch": 1.29, + "grad_norm": 2.59375, + "learning_rate": 3.808594132598574e-06, + "loss": 0.8229, + "step": 982 + }, + { + "epoch": 1.29, + "grad_norm": 2.65625, + "learning_rate": 3.806358243996593e-06, + "loss": 0.8174, + "step": 983 + }, + { + "epoch": 1.29, + "grad_norm": 2.640625, + "learning_rate": 3.8041209170650586e-06, + "loss": 0.8287, + "step": 984 + }, + { + "epoch": 1.3, + "grad_norm": 2.625, + "learning_rate": 3.801882154267318e-06, + "loss": 0.8004, + "step": 985 + }, + { + "epoch": 1.3, + "grad_norm": 2.59375, + "learning_rate": 3.7996419580682996e-06, + "loss": 0.816, + "step": 986 + }, + { + "epoch": 1.3, + "grad_norm": 2.59375, + "learning_rate": 3.7974003309345086e-06, + "loss": 0.8128, + "step": 987 + }, + { + "epoch": 1.3, + "grad_norm": 2.6875, + "learning_rate": 3.7951572753340273e-06, + "loss": 0.8186, + "step": 988 + }, + { + "epoch": 1.3, + "grad_norm": 2.625, + "learning_rate": 3.7929127937365086e-06, + "loss": 0.826, + "step": 989 + }, + { + "epoch": 1.3, + "grad_norm": 2.609375, + "learning_rate": 3.7906668886131783e-06, + "loss": 0.8414, + "step": 990 + }, + { + "epoch": 1.3, + "grad_norm": 2.59375, + "learning_rate": 3.7884195624368268e-06, + "loss": 0.8154, + "step": 991 + }, + { + "epoch": 1.31, + "grad_norm": 2.671875, + "learning_rate": 3.7861708176818113e-06, + "loss": 0.8312, + "step": 992 + }, + { + "epoch": 1.31, + "grad_norm": 2.5625, + "learning_rate": 3.7839206568240494e-06, + "loss": 0.7868, + "step": 993 + }, + { + "epoch": 1.31, + "grad_norm": 2.625, + "learning_rate": 3.781669082341018e-06, + "loss": 0.8451, + "step": 994 + }, + { + "epoch": 1.31, + "grad_norm": 2.5625, + "learning_rate": 3.779416096711751e-06, + "loss": 0.8159, + "step": 995 + }, + { + "epoch": 1.31, + "grad_norm": 2.5625, + "learning_rate": 3.7771617024168374e-06, + "loss": 0.8022, + "step": 996 + }, + { + "epoch": 1.31, + "grad_norm": 2.65625, + "learning_rate": 3.7749059019384127e-06, + "loss": 0.8012, + "step": 997 + }, + { + "epoch": 1.31, + "grad_norm": 2.5625, + "learning_rate": 3.7726486977601655e-06, + "loss": 0.8251, + "step": 998 + }, + { + "epoch": 1.31, + "grad_norm": 2.515625, + "learning_rate": 3.770390092367328e-06, + "loss": 0.8232, + "step": 999 + }, + { + "epoch": 1.32, + "grad_norm": 2.546875, + "learning_rate": 3.768130088246674e-06, + "loss": 0.8299, + "step": 1000 + }, + { + "epoch": 1.32, + "grad_norm": 2.578125, + "learning_rate": 3.7658686878865193e-06, + "loss": 0.8278, + "step": 1001 + }, + { + "epoch": 1.32, + "grad_norm": 2.625, + "learning_rate": 3.7636058937767156e-06, + "loss": 0.8401, + "step": 1002 + }, + { + "epoch": 1.32, + "grad_norm": 2.578125, + "learning_rate": 3.76134170840865e-06, + "loss": 0.8021, + "step": 1003 + }, + { + "epoch": 1.32, + "grad_norm": 2.578125, + "learning_rate": 3.7590761342752406e-06, + "loss": 0.8258, + "step": 1004 + }, + { + "epoch": 1.32, + "grad_norm": 2.640625, + "learning_rate": 3.756809173870936e-06, + "loss": 0.8252, + "step": 1005 + }, + { + "epoch": 1.32, + "grad_norm": 2.515625, + "learning_rate": 3.7545408296917087e-06, + "loss": 0.7926, + "step": 1006 + }, + { + "epoch": 1.33, + "grad_norm": 2.59375, + "learning_rate": 3.7522711042350574e-06, + "loss": 0.8258, + "step": 1007 + }, + { + "epoch": 1.33, + "grad_norm": 2.65625, + "learning_rate": 3.7500000000000005e-06, + "loss": 0.8165, + "step": 1008 + }, + { + "epoch": 1.33, + "grad_norm": 2.515625, + "learning_rate": 3.7477275194870733e-06, + "loss": 0.8155, + "step": 1009 + }, + { + "epoch": 1.33, + "grad_norm": 2.5625, + "learning_rate": 3.7454536651983284e-06, + "loss": 0.8214, + "step": 1010 + }, + { + "epoch": 1.33, + "grad_norm": 2.640625, + "learning_rate": 3.74317843963733e-06, + "loss": 0.8361, + "step": 1011 + }, + { + "epoch": 1.33, + "grad_norm": 2.515625, + "learning_rate": 3.740901845309152e-06, + "loss": 0.788, + "step": 1012 + }, + { + "epoch": 1.33, + "grad_norm": 2.59375, + "learning_rate": 3.7386238847203756e-06, + "loss": 0.8023, + "step": 1013 + }, + { + "epoch": 1.33, + "grad_norm": 2.609375, + "learning_rate": 3.736344560379086e-06, + "loss": 0.8444, + "step": 1014 + }, + { + "epoch": 1.34, + "grad_norm": 2.5, + "learning_rate": 3.7340638747948697e-06, + "loss": 0.7757, + "step": 1015 + }, + { + "epoch": 1.34, + "grad_norm": 2.609375, + "learning_rate": 3.7317818304788133e-06, + "loss": 0.806, + "step": 1016 + }, + { + "epoch": 1.34, + "grad_norm": 2.65625, + "learning_rate": 3.7294984299434987e-06, + "loss": 0.8603, + "step": 1017 + }, + { + "epoch": 1.34, + "grad_norm": 2.546875, + "learning_rate": 3.727213675703e-06, + "loss": 0.8062, + "step": 1018 + }, + { + "epoch": 1.34, + "grad_norm": 2.6875, + "learning_rate": 3.7249275702728826e-06, + "loss": 0.8077, + "step": 1019 + }, + { + "epoch": 1.34, + "grad_norm": 2.609375, + "learning_rate": 3.722640116170199e-06, + "loss": 0.8152, + "step": 1020 + }, + { + "epoch": 1.34, + "grad_norm": 2.65625, + "learning_rate": 3.7203513159134887e-06, + "loss": 0.8436, + "step": 1021 + }, + { + "epoch": 1.34, + "grad_norm": 2.640625, + "learning_rate": 3.718061172022771e-06, + "loss": 0.8261, + "step": 1022 + }, + { + "epoch": 1.35, + "grad_norm": 2.59375, + "learning_rate": 3.715769687019545e-06, + "loss": 0.7974, + "step": 1023 + }, + { + "epoch": 1.35, + "grad_norm": 2.609375, + "learning_rate": 3.713476863426787e-06, + "loss": 0.8161, + "step": 1024 + }, + { + "epoch": 1.35, + "grad_norm": 2.65625, + "learning_rate": 3.7111827037689473e-06, + "loss": 0.8042, + "step": 1025 + }, + { + "epoch": 1.35, + "grad_norm": 2.65625, + "learning_rate": 3.708887210571947e-06, + "loss": 0.8124, + "step": 1026 + }, + { + "epoch": 1.35, + "grad_norm": 2.609375, + "learning_rate": 3.706590386363174e-06, + "loss": 0.8208, + "step": 1027 + }, + { + "epoch": 1.35, + "grad_norm": 2.625, + "learning_rate": 3.704292233671485e-06, + "loss": 0.8293, + "step": 1028 + }, + { + "epoch": 1.35, + "grad_norm": 2.578125, + "learning_rate": 3.7019927550271952e-06, + "loss": 0.8126, + "step": 1029 + }, + { + "epoch": 1.36, + "grad_norm": 2.65625, + "learning_rate": 3.699691952962083e-06, + "loss": 0.839, + "step": 1030 + }, + { + "epoch": 1.36, + "grad_norm": 2.640625, + "learning_rate": 3.6973898300093823e-06, + "loss": 0.8598, + "step": 1031 + }, + { + "epoch": 1.36, + "grad_norm": 2.609375, + "learning_rate": 3.695086388703783e-06, + "loss": 0.812, + "step": 1032 + }, + { + "epoch": 1.36, + "grad_norm": 2.59375, + "learning_rate": 3.6927816315814237e-06, + "loss": 0.84, + "step": 1033 + }, + { + "epoch": 1.36, + "grad_norm": 2.578125, + "learning_rate": 3.6904755611798935e-06, + "loss": 0.8635, + "step": 1034 + }, + { + "epoch": 1.36, + "grad_norm": 2.6875, + "learning_rate": 3.6881681800382294e-06, + "loss": 0.8342, + "step": 1035 + }, + { + "epoch": 1.36, + "grad_norm": 2.734375, + "learning_rate": 3.6858594906969073e-06, + "loss": 0.8571, + "step": 1036 + }, + { + "epoch": 1.36, + "grad_norm": 2.5625, + "learning_rate": 3.6835494956978468e-06, + "loss": 0.8232, + "step": 1037 + }, + { + "epoch": 1.37, + "grad_norm": 2.453125, + "learning_rate": 3.681238197584403e-06, + "loss": 0.8003, + "step": 1038 + }, + { + "epoch": 1.37, + "grad_norm": 2.671875, + "learning_rate": 3.678925598901367e-06, + "loss": 0.8293, + "step": 1039 + }, + { + "epoch": 1.37, + "grad_norm": 2.671875, + "learning_rate": 3.676611702194963e-06, + "loss": 0.8236, + "step": 1040 + }, + { + "epoch": 1.37, + "grad_norm": 2.53125, + "learning_rate": 3.6742965100128397e-06, + "loss": 0.8254, + "step": 1041 + }, + { + "epoch": 1.37, + "grad_norm": 2.703125, + "learning_rate": 3.6719800249040778e-06, + "loss": 0.7983, + "step": 1042 + }, + { + "epoch": 1.37, + "grad_norm": 2.546875, + "learning_rate": 3.669662249419178e-06, + "loss": 0.8113, + "step": 1043 + }, + { + "epoch": 1.37, + "grad_norm": 2.703125, + "learning_rate": 3.667343186110064e-06, + "loss": 0.8193, + "step": 1044 + }, + { + "epoch": 1.38, + "grad_norm": 2.59375, + "learning_rate": 3.6650228375300743e-06, + "loss": 0.8252, + "step": 1045 + }, + { + "epoch": 1.38, + "grad_norm": 2.578125, + "learning_rate": 3.6627012062339666e-06, + "loss": 0.8064, + "step": 1046 + }, + { + "epoch": 1.38, + "grad_norm": 2.546875, + "learning_rate": 3.6603782947779066e-06, + "loss": 0.8377, + "step": 1047 + }, + { + "epoch": 1.38, + "grad_norm": 2.59375, + "learning_rate": 3.6580541057194728e-06, + "loss": 0.8196, + "step": 1048 + }, + { + "epoch": 1.38, + "grad_norm": 2.671875, + "learning_rate": 3.6557286416176497e-06, + "loss": 0.8145, + "step": 1049 + }, + { + "epoch": 1.38, + "grad_norm": 2.609375, + "learning_rate": 3.6534019050328234e-06, + "loss": 0.8201, + "step": 1050 + }, + { + "epoch": 1.38, + "grad_norm": 2.453125, + "learning_rate": 3.6510738985267847e-06, + "loss": 0.7855, + "step": 1051 + }, + { + "epoch": 1.38, + "grad_norm": 2.515625, + "learning_rate": 3.648744624662719e-06, + "loss": 0.7946, + "step": 1052 + }, + { + "epoch": 1.39, + "grad_norm": 2.703125, + "learning_rate": 3.6464140860052093e-06, + "loss": 0.8377, + "step": 1053 + }, + { + "epoch": 1.39, + "grad_norm": 2.65625, + "learning_rate": 3.6440822851202312e-06, + "loss": 0.8362, + "step": 1054 + }, + { + "epoch": 1.39, + "grad_norm": 2.65625, + "learning_rate": 3.6417492245751494e-06, + "loss": 0.8321, + "step": 1055 + }, + { + "epoch": 1.39, + "grad_norm": 2.640625, + "learning_rate": 3.639414906938714e-06, + "loss": 0.8266, + "step": 1056 + }, + { + "epoch": 1.39, + "grad_norm": 2.71875, + "learning_rate": 3.6370793347810627e-06, + "loss": 0.8628, + "step": 1057 + }, + { + "epoch": 1.39, + "grad_norm": 2.640625, + "learning_rate": 3.6347425106737126e-06, + "loss": 0.8473, + "step": 1058 + }, + { + "epoch": 1.39, + "grad_norm": 2.609375, + "learning_rate": 3.632404437189558e-06, + "loss": 0.8426, + "step": 1059 + }, + { + "epoch": 1.4, + "grad_norm": 2.578125, + "learning_rate": 3.63006511690287e-06, + "loss": 0.8222, + "step": 1060 + }, + { + "epoch": 1.4, + "grad_norm": 2.609375, + "learning_rate": 3.6277245523892922e-06, + "loss": 0.7863, + "step": 1061 + }, + { + "epoch": 1.4, + "grad_norm": 2.640625, + "learning_rate": 3.6253827462258397e-06, + "loss": 0.8217, + "step": 1062 + }, + { + "epoch": 1.4, + "grad_norm": 2.59375, + "learning_rate": 3.623039700990892e-06, + "loss": 0.8081, + "step": 1063 + }, + { + "epoch": 1.4, + "grad_norm": 2.671875, + "learning_rate": 3.6206954192641946e-06, + "loss": 0.794, + "step": 1064 + }, + { + "epoch": 1.4, + "grad_norm": 2.5, + "learning_rate": 3.618349903626853e-06, + "loss": 0.8101, + "step": 1065 + }, + { + "epoch": 1.4, + "grad_norm": 2.5625, + "learning_rate": 3.616003156661334e-06, + "loss": 0.8273, + "step": 1066 + }, + { + "epoch": 1.4, + "grad_norm": 2.59375, + "learning_rate": 3.613655180951458e-06, + "loss": 0.8351, + "step": 1067 + }, + { + "epoch": 1.41, + "grad_norm": 2.625, + "learning_rate": 3.6113059790823975e-06, + "loss": 0.8318, + "step": 1068 + }, + { + "epoch": 1.41, + "grad_norm": 2.65625, + "learning_rate": 3.6089555536406764e-06, + "loss": 0.8246, + "step": 1069 + }, + { + "epoch": 1.41, + "grad_norm": 2.75, + "learning_rate": 3.6066039072141657e-06, + "loss": 0.8399, + "step": 1070 + }, + { + "epoch": 1.41, + "grad_norm": 2.6875, + "learning_rate": 3.6042510423920814e-06, + "loss": 0.8482, + "step": 1071 + }, + { + "epoch": 1.41, + "grad_norm": 2.59375, + "learning_rate": 3.6018969617649784e-06, + "loss": 0.7631, + "step": 1072 + }, + { + "epoch": 1.41, + "grad_norm": 2.609375, + "learning_rate": 3.5995416679247542e-06, + "loss": 0.8339, + "step": 1073 + }, + { + "epoch": 1.41, + "grad_norm": 2.625, + "learning_rate": 3.597185163464637e-06, + "loss": 0.8312, + "step": 1074 + }, + { + "epoch": 1.42, + "grad_norm": 2.453125, + "learning_rate": 3.594827450979193e-06, + "loss": 0.789, + "step": 1075 + }, + { + "epoch": 1.42, + "grad_norm": 2.59375, + "learning_rate": 3.592468533064316e-06, + "loss": 0.7893, + "step": 1076 + }, + { + "epoch": 1.42, + "grad_norm": 2.609375, + "learning_rate": 3.5901084123172263e-06, + "loss": 0.8225, + "step": 1077 + }, + { + "epoch": 1.42, + "grad_norm": 2.546875, + "learning_rate": 3.5877470913364697e-06, + "loss": 0.8248, + "step": 1078 + }, + { + "epoch": 1.42, + "grad_norm": 2.5, + "learning_rate": 3.585384572721914e-06, + "loss": 0.7855, + "step": 1079 + }, + { + "epoch": 1.42, + "grad_norm": 2.484375, + "learning_rate": 3.583020859074744e-06, + "loss": 0.7833, + "step": 1080 + }, + { + "epoch": 1.42, + "grad_norm": 2.5, + "learning_rate": 3.580655952997461e-06, + "loss": 0.7909, + "step": 1081 + }, + { + "epoch": 1.42, + "grad_norm": 2.515625, + "learning_rate": 3.5782898570938805e-06, + "loss": 0.7868, + "step": 1082 + }, + { + "epoch": 1.43, + "grad_norm": 2.546875, + "learning_rate": 3.5759225739691255e-06, + "loss": 0.7936, + "step": 1083 + }, + { + "epoch": 1.43, + "grad_norm": 2.671875, + "learning_rate": 3.5735541062296287e-06, + "loss": 0.8491, + "step": 1084 + }, + { + "epoch": 1.43, + "grad_norm": 2.640625, + "learning_rate": 3.571184456483125e-06, + "loss": 0.8256, + "step": 1085 + }, + { + "epoch": 1.43, + "grad_norm": 2.625, + "learning_rate": 3.568813627338652e-06, + "loss": 0.8156, + "step": 1086 + }, + { + "epoch": 1.43, + "grad_norm": 2.640625, + "learning_rate": 3.566441621406545e-06, + "loss": 0.8262, + "step": 1087 + }, + { + "epoch": 1.43, + "grad_norm": 2.625, + "learning_rate": 3.564068441298436e-06, + "loss": 0.8217, + "step": 1088 + }, + { + "epoch": 1.43, + "grad_norm": 2.6875, + "learning_rate": 3.561694089627249e-06, + "loss": 0.8125, + "step": 1089 + }, + { + "epoch": 1.44, + "grad_norm": 2.671875, + "learning_rate": 3.559318569007198e-06, + "loss": 0.8435, + "step": 1090 + }, + { + "epoch": 1.44, + "grad_norm": 2.65625, + "learning_rate": 3.5569418820537848e-06, + "loss": 0.8503, + "step": 1091 + }, + { + "epoch": 1.44, + "grad_norm": 2.75, + "learning_rate": 3.5545640313837935e-06, + "loss": 0.8532, + "step": 1092 + }, + { + "epoch": 1.44, + "grad_norm": 2.75, + "learning_rate": 3.552185019615292e-06, + "loss": 0.8364, + "step": 1093 + }, + { + "epoch": 1.44, + "grad_norm": 2.5625, + "learning_rate": 3.549804849367624e-06, + "loss": 0.8168, + "step": 1094 + }, + { + "epoch": 1.44, + "grad_norm": 2.53125, + "learning_rate": 3.547423523261411e-06, + "loss": 0.8155, + "step": 1095 + }, + { + "epoch": 1.44, + "grad_norm": 2.78125, + "learning_rate": 3.545041043918546e-06, + "loss": 0.8625, + "step": 1096 + }, + { + "epoch": 1.44, + "grad_norm": 2.84375, + "learning_rate": 3.5426574139621913e-06, + "loss": 0.8318, + "step": 1097 + }, + { + "epoch": 1.45, + "grad_norm": 2.59375, + "learning_rate": 3.5402726360167772e-06, + "loss": 0.8254, + "step": 1098 + }, + { + "epoch": 1.45, + "grad_norm": 2.515625, + "learning_rate": 3.5378867127079973e-06, + "loss": 0.8127, + "step": 1099 + }, + { + "epoch": 1.45, + "grad_norm": 2.703125, + "learning_rate": 3.5354996466628067e-06, + "loss": 0.8187, + "step": 1100 + }, + { + "epoch": 1.45, + "grad_norm": 2.71875, + "learning_rate": 3.5331114405094167e-06, + "loss": 0.845, + "step": 1101 + }, + { + "epoch": 1.45, + "grad_norm": 2.71875, + "learning_rate": 3.5307220968772983e-06, + "loss": 0.8534, + "step": 1102 + }, + { + "epoch": 1.45, + "grad_norm": 2.578125, + "learning_rate": 3.5283316183971694e-06, + "loss": 0.8038, + "step": 1103 + }, + { + "epoch": 1.45, + "grad_norm": 2.671875, + "learning_rate": 3.525940007701002e-06, + "loss": 0.8262, + "step": 1104 + }, + { + "epoch": 1.46, + "grad_norm": 2.734375, + "learning_rate": 3.5235472674220116e-06, + "loss": 0.8309, + "step": 1105 + }, + { + "epoch": 1.46, + "grad_norm": 2.6875, + "learning_rate": 3.521153400194659e-06, + "loss": 0.8321, + "step": 1106 + }, + { + "epoch": 1.46, + "grad_norm": 2.703125, + "learning_rate": 3.5187584086546455e-06, + "loss": 0.8507, + "step": 1107 + }, + { + "epoch": 1.46, + "grad_norm": 2.6875, + "learning_rate": 3.516362295438911e-06, + "loss": 0.8564, + "step": 1108 + }, + { + "epoch": 1.46, + "grad_norm": 2.84375, + "learning_rate": 3.5139650631856283e-06, + "loss": 0.8545, + "step": 1109 + }, + { + "epoch": 1.46, + "grad_norm": 2.515625, + "learning_rate": 3.511566714534203e-06, + "loss": 0.8378, + "step": 1110 + }, + { + "epoch": 1.46, + "grad_norm": 2.609375, + "learning_rate": 3.5091672521252724e-06, + "loss": 0.8484, + "step": 1111 + }, + { + "epoch": 1.46, + "grad_norm": 2.75, + "learning_rate": 3.5067666786006964e-06, + "loss": 0.8638, + "step": 1112 + }, + { + "epoch": 1.47, + "grad_norm": 2.65625, + "learning_rate": 3.504364996603561e-06, + "loss": 0.7978, + "step": 1113 + }, + { + "epoch": 1.47, + "grad_norm": 2.5625, + "learning_rate": 3.501962208778172e-06, + "loss": 0.8169, + "step": 1114 + }, + { + "epoch": 1.47, + "grad_norm": 2.609375, + "learning_rate": 3.4995583177700506e-06, + "loss": 0.7856, + "step": 1115 + }, + { + "epoch": 1.47, + "grad_norm": 2.546875, + "learning_rate": 3.497153326225936e-06, + "loss": 0.8123, + "step": 1116 + }, + { + "epoch": 1.47, + "grad_norm": 2.734375, + "learning_rate": 3.4947472367937775e-06, + "loss": 0.7838, + "step": 1117 + }, + { + "epoch": 1.47, + "grad_norm": 2.546875, + "learning_rate": 3.4923400521227317e-06, + "loss": 0.7954, + "step": 1118 + }, + { + "epoch": 1.47, + "grad_norm": 2.625, + "learning_rate": 3.489931774863164e-06, + "loss": 0.8407, + "step": 1119 + }, + { + "epoch": 1.48, + "grad_norm": 2.6875, + "learning_rate": 3.487522407666641e-06, + "loss": 0.8349, + "step": 1120 + }, + { + "epoch": 1.48, + "grad_norm": 2.625, + "learning_rate": 3.485111953185929e-06, + "loss": 0.8175, + "step": 1121 + }, + { + "epoch": 1.48, + "grad_norm": 2.65625, + "learning_rate": 3.482700414074994e-06, + "loss": 0.8165, + "step": 1122 + }, + { + "epoch": 1.48, + "grad_norm": 2.546875, + "learning_rate": 3.4802877929889924e-06, + "loss": 0.7937, + "step": 1123 + }, + { + "epoch": 1.48, + "grad_norm": 2.625, + "learning_rate": 3.4778740925842742e-06, + "loss": 0.8399, + "step": 1124 + }, + { + "epoch": 1.48, + "grad_norm": 2.65625, + "learning_rate": 3.475459315518378e-06, + "loss": 0.8227, + "step": 1125 + }, + { + "epoch": 1.48, + "grad_norm": 2.515625, + "learning_rate": 3.473043464450027e-06, + "loss": 0.7987, + "step": 1126 + }, + { + "epoch": 1.48, + "grad_norm": 2.53125, + "learning_rate": 3.470626542039127e-06, + "loss": 0.8154, + "step": 1127 + }, + { + "epoch": 1.49, + "grad_norm": 2.53125, + "learning_rate": 3.4682085509467634e-06, + "loss": 0.8399, + "step": 1128 + }, + { + "epoch": 1.49, + "grad_norm": 2.609375, + "learning_rate": 3.4657894938351976e-06, + "loss": 0.8184, + "step": 1129 + }, + { + "epoch": 1.49, + "grad_norm": 2.578125, + "learning_rate": 3.4633693733678668e-06, + "loss": 0.8365, + "step": 1130 + }, + { + "epoch": 1.49, + "grad_norm": 2.546875, + "learning_rate": 3.460948192209377e-06, + "loss": 0.8183, + "step": 1131 + }, + { + "epoch": 1.49, + "grad_norm": 2.71875, + "learning_rate": 3.458525953025503e-06, + "loss": 0.8131, + "step": 1132 + }, + { + "epoch": 1.49, + "grad_norm": 2.6875, + "learning_rate": 3.4561026584831827e-06, + "loss": 0.8251, + "step": 1133 + }, + { + "epoch": 1.49, + "grad_norm": 2.734375, + "learning_rate": 3.4536783112505183e-06, + "loss": 0.794, + "step": 1134 + }, + { + "epoch": 1.5, + "grad_norm": 2.59375, + "learning_rate": 3.451252913996771e-06, + "loss": 0.8197, + "step": 1135 + }, + { + "epoch": 1.5, + "grad_norm": 2.703125, + "learning_rate": 3.4488264693923568e-06, + "loss": 0.8027, + "step": 1136 + }, + { + "epoch": 1.5, + "grad_norm": 2.625, + "learning_rate": 3.4463989801088444e-06, + "loss": 0.8359, + "step": 1137 + }, + { + "epoch": 1.5, + "grad_norm": 2.515625, + "learning_rate": 3.443970448818954e-06, + "loss": 0.7968, + "step": 1138 + }, + { + "epoch": 1.5, + "grad_norm": 2.65625, + "learning_rate": 3.441540878196553e-06, + "loss": 0.8219, + "step": 1139 + }, + { + "epoch": 1.5, + "grad_norm": 2.5625, + "learning_rate": 3.4391102709166536e-06, + "loss": 0.8199, + "step": 1140 + }, + { + "epoch": 1.5, + "grad_norm": 2.484375, + "learning_rate": 3.4366786296554078e-06, + "loss": 0.827, + "step": 1141 + }, + { + "epoch": 1.5, + "grad_norm": 2.5625, + "learning_rate": 3.4342459570901055e-06, + "loss": 0.8125, + "step": 1142 + }, + { + "epoch": 1.51, + "grad_norm": 2.609375, + "learning_rate": 3.431812255899176e-06, + "loss": 0.8207, + "step": 1143 + }, + { + "epoch": 1.51, + "grad_norm": 2.53125, + "learning_rate": 3.429377528762177e-06, + "loss": 0.8138, + "step": 1144 + }, + { + "epoch": 1.51, + "grad_norm": 2.546875, + "learning_rate": 3.426941778359798e-06, + "loss": 0.8369, + "step": 1145 + }, + { + "epoch": 1.51, + "grad_norm": 2.546875, + "learning_rate": 3.424505007373854e-06, + "loss": 0.8256, + "step": 1146 + }, + { + "epoch": 1.51, + "grad_norm": 2.515625, + "learning_rate": 3.4220672184872838e-06, + "loss": 0.804, + "step": 1147 + }, + { + "epoch": 1.51, + "grad_norm": 2.59375, + "learning_rate": 3.4196284143841487e-06, + "loss": 0.8195, + "step": 1148 + }, + { + "epoch": 1.51, + "grad_norm": 2.53125, + "learning_rate": 3.4171885977496255e-06, + "loss": 0.8143, + "step": 1149 + }, + { + "epoch": 1.52, + "grad_norm": 2.578125, + "learning_rate": 3.414747771270007e-06, + "loss": 0.8435, + "step": 1150 + }, + { + "epoch": 1.52, + "grad_norm": 2.46875, + "learning_rate": 3.4123059376326965e-06, + "loss": 0.8251, + "step": 1151 + }, + { + "epoch": 1.52, + "grad_norm": 2.546875, + "learning_rate": 3.4098630995262085e-06, + "loss": 0.8159, + "step": 1152 + }, + { + "epoch": 1.52, + "grad_norm": 2.578125, + "learning_rate": 3.4074192596401624e-06, + "loss": 0.8085, + "step": 1153 + }, + { + "epoch": 1.52, + "grad_norm": 2.640625, + "learning_rate": 3.404974420665279e-06, + "loss": 0.8136, + "step": 1154 + }, + { + "epoch": 1.52, + "grad_norm": 2.734375, + "learning_rate": 3.402528585293382e-06, + "loss": 0.8209, + "step": 1155 + }, + { + "epoch": 1.52, + "grad_norm": 2.5, + "learning_rate": 3.40008175621739e-06, + "loss": 0.8152, + "step": 1156 + }, + { + "epoch": 1.52, + "grad_norm": 2.5, + "learning_rate": 3.397633936131316e-06, + "loss": 0.8225, + "step": 1157 + }, + { + "epoch": 1.53, + "grad_norm": 2.796875, + "learning_rate": 3.395185127730266e-06, + "loss": 0.8237, + "step": 1158 + }, + { + "epoch": 1.53, + "grad_norm": 2.625, + "learning_rate": 3.3927353337104306e-06, + "loss": 0.8226, + "step": 1159 + }, + { + "epoch": 1.53, + "grad_norm": 2.53125, + "learning_rate": 3.390284556769089e-06, + "loss": 0.7966, + "step": 1160 + }, + { + "epoch": 1.53, + "grad_norm": 2.515625, + "learning_rate": 3.387832799604601e-06, + "loss": 0.8097, + "step": 1161 + }, + { + "epoch": 1.53, + "grad_norm": 2.65625, + "learning_rate": 3.3853800649164053e-06, + "loss": 0.7866, + "step": 1162 + }, + { + "epoch": 1.53, + "grad_norm": 2.609375, + "learning_rate": 3.3829263554050184e-06, + "loss": 0.8148, + "step": 1163 + }, + { + "epoch": 1.53, + "grad_norm": 2.46875, + "learning_rate": 3.3804716737720287e-06, + "loss": 0.8204, + "step": 1164 + }, + { + "epoch": 1.54, + "grad_norm": 2.609375, + "learning_rate": 3.3780160227200953e-06, + "loss": 0.7955, + "step": 1165 + }, + { + "epoch": 1.54, + "grad_norm": 2.71875, + "learning_rate": 3.3755594049529443e-06, + "loss": 0.7997, + "step": 1166 + }, + { + "epoch": 1.54, + "grad_norm": 2.625, + "learning_rate": 3.3731018231753687e-06, + "loss": 0.8245, + "step": 1167 + }, + { + "epoch": 1.54, + "grad_norm": 2.640625, + "learning_rate": 3.3706432800932184e-06, + "loss": 0.8573, + "step": 1168 + }, + { + "epoch": 1.54, + "grad_norm": 2.765625, + "learning_rate": 3.3681837784134055e-06, + "loss": 0.8205, + "step": 1169 + }, + { + "epoch": 1.54, + "grad_norm": 2.609375, + "learning_rate": 3.365723320843895e-06, + "loss": 0.8411, + "step": 1170 + }, + { + "epoch": 1.54, + "grad_norm": 2.59375, + "learning_rate": 3.363261910093707e-06, + "loss": 0.8359, + "step": 1171 + }, + { + "epoch": 1.54, + "grad_norm": 2.625, + "learning_rate": 3.3607995488729082e-06, + "loss": 0.8088, + "step": 1172 + }, + { + "epoch": 1.55, + "grad_norm": 2.71875, + "learning_rate": 3.3583362398926156e-06, + "loss": 0.8081, + "step": 1173 + }, + { + "epoch": 1.55, + "grad_norm": 2.53125, + "learning_rate": 3.3558719858649835e-06, + "loss": 0.8248, + "step": 1174 + }, + { + "epoch": 1.55, + "grad_norm": 2.5625, + "learning_rate": 3.353406789503214e-06, + "loss": 0.8296, + "step": 1175 + }, + { + "epoch": 1.55, + "grad_norm": 2.609375, + "learning_rate": 3.3509406535215423e-06, + "loss": 0.8206, + "step": 1176 + }, + { + "epoch": 1.55, + "grad_norm": 2.59375, + "learning_rate": 3.348473580635238e-06, + "loss": 0.8297, + "step": 1177 + }, + { + "epoch": 1.55, + "grad_norm": 2.53125, + "learning_rate": 3.3460055735606045e-06, + "loss": 0.807, + "step": 1178 + }, + { + "epoch": 1.55, + "grad_norm": 2.625, + "learning_rate": 3.3435366350149716e-06, + "loss": 0.8304, + "step": 1179 + }, + { + "epoch": 1.56, + "grad_norm": 2.734375, + "learning_rate": 3.341066767716697e-06, + "loss": 0.8024, + "step": 1180 + }, + { + "epoch": 1.56, + "grad_norm": 2.578125, + "learning_rate": 3.3385959743851583e-06, + "loss": 0.8097, + "step": 1181 + }, + { + "epoch": 1.56, + "grad_norm": 2.59375, + "learning_rate": 3.3361242577407556e-06, + "loss": 0.8112, + "step": 1182 + }, + { + "epoch": 1.56, + "grad_norm": 2.59375, + "learning_rate": 3.3336516205049017e-06, + "loss": 0.8025, + "step": 1183 + }, + { + "epoch": 1.56, + "grad_norm": 2.625, + "learning_rate": 3.331178065400028e-06, + "loss": 0.8227, + "step": 1184 + }, + { + "epoch": 1.56, + "grad_norm": 2.71875, + "learning_rate": 3.328703595149573e-06, + "loss": 0.8239, + "step": 1185 + }, + { + "epoch": 1.56, + "grad_norm": 2.625, + "learning_rate": 3.3262282124779823e-06, + "loss": 0.8142, + "step": 1186 + }, + { + "epoch": 1.56, + "grad_norm": 2.640625, + "learning_rate": 3.3237519201107084e-06, + "loss": 0.8219, + "step": 1187 + }, + { + "epoch": 1.57, + "grad_norm": 2.625, + "learning_rate": 3.3212747207742047e-06, + "loss": 0.8029, + "step": 1188 + }, + { + "epoch": 1.57, + "grad_norm": 2.59375, + "learning_rate": 3.318796617195922e-06, + "loss": 0.8152, + "step": 1189 + }, + { + "epoch": 1.57, + "grad_norm": 2.765625, + "learning_rate": 3.3163176121043086e-06, + "loss": 0.8256, + "step": 1190 + }, + { + "epoch": 1.57, + "grad_norm": 2.71875, + "learning_rate": 3.313837708228804e-06, + "loss": 0.8236, + "step": 1191 + }, + { + "epoch": 1.57, + "grad_norm": 2.59375, + "learning_rate": 3.3113569082998367e-06, + "loss": 0.8423, + "step": 1192 + }, + { + "epoch": 1.57, + "grad_norm": 2.515625, + "learning_rate": 3.3088752150488236e-06, + "loss": 0.7812, + "step": 1193 + }, + { + "epoch": 1.57, + "grad_norm": 2.734375, + "learning_rate": 3.306392631208165e-06, + "loss": 0.8388, + "step": 1194 + }, + { + "epoch": 1.58, + "grad_norm": 2.515625, + "learning_rate": 3.3039091595112404e-06, + "loss": 0.8182, + "step": 1195 + }, + { + "epoch": 1.58, + "grad_norm": 2.546875, + "learning_rate": 3.3014248026924073e-06, + "loss": 0.8358, + "step": 1196 + }, + { + "epoch": 1.58, + "grad_norm": 2.59375, + "learning_rate": 3.298939563486998e-06, + "loss": 0.7901, + "step": 1197 + }, + { + "epoch": 1.58, + "grad_norm": 2.453125, + "learning_rate": 3.2964534446313163e-06, + "loss": 0.7996, + "step": 1198 + }, + { + "epoch": 1.58, + "grad_norm": 2.5625, + "learning_rate": 3.2939664488626344e-06, + "loss": 0.8174, + "step": 1199 + }, + { + "epoch": 1.58, + "grad_norm": 2.59375, + "learning_rate": 3.291478578919191e-06, + "loss": 0.8265, + "step": 1200 + }, + { + "epoch": 1.58, + "grad_norm": 2.515625, + "learning_rate": 3.2889898375401847e-06, + "loss": 0.8335, + "step": 1201 + }, + { + "epoch": 1.58, + "grad_norm": 2.546875, + "learning_rate": 3.2865002274657765e-06, + "loss": 0.8129, + "step": 1202 + }, + { + "epoch": 1.59, + "grad_norm": 2.65625, + "learning_rate": 3.2840097514370827e-06, + "loss": 0.8271, + "step": 1203 + }, + { + "epoch": 1.59, + "grad_norm": 2.546875, + "learning_rate": 3.2815184121961725e-06, + "loss": 0.849, + "step": 1204 + }, + { + "epoch": 1.59, + "grad_norm": 2.546875, + "learning_rate": 3.2790262124860657e-06, + "loss": 0.7985, + "step": 1205 + }, + { + "epoch": 1.59, + "grad_norm": 2.59375, + "learning_rate": 3.2765331550507296e-06, + "loss": 0.8146, + "step": 1206 + }, + { + "epoch": 1.59, + "grad_norm": 2.625, + "learning_rate": 3.2740392426350763e-06, + "loss": 0.8387, + "step": 1207 + }, + { + "epoch": 1.59, + "grad_norm": 2.546875, + "learning_rate": 3.27154447798496e-06, + "loss": 0.8311, + "step": 1208 + }, + { + "epoch": 1.59, + "grad_norm": 2.546875, + "learning_rate": 3.2690488638471713e-06, + "loss": 0.8238, + "step": 1209 + }, + { + "epoch": 1.6, + "grad_norm": 2.5625, + "learning_rate": 3.266552402969437e-06, + "loss": 0.7965, + "step": 1210 + }, + { + "epoch": 1.6, + "grad_norm": 2.609375, + "learning_rate": 3.2640550981004155e-06, + "loss": 0.8509, + "step": 1211 + }, + { + "epoch": 1.6, + "grad_norm": 2.71875, + "learning_rate": 3.2615569519896962e-06, + "loss": 0.8154, + "step": 1212 + }, + { + "epoch": 1.6, + "grad_norm": 2.65625, + "learning_rate": 3.2590579673877932e-06, + "loss": 0.8234, + "step": 1213 + }, + { + "epoch": 1.6, + "grad_norm": 2.546875, + "learning_rate": 3.2565581470461444e-06, + "loss": 0.7961, + "step": 1214 + }, + { + "epoch": 1.6, + "grad_norm": 2.71875, + "learning_rate": 3.2540574937171076e-06, + "loss": 0.8078, + "step": 1215 + }, + { + "epoch": 1.6, + "grad_norm": 2.5, + "learning_rate": 3.251556010153958e-06, + "loss": 0.7844, + "step": 1216 + }, + { + "epoch": 1.6, + "grad_norm": 2.671875, + "learning_rate": 3.2490536991108857e-06, + "loss": 0.8181, + "step": 1217 + }, + { + "epoch": 1.61, + "grad_norm": 2.546875, + "learning_rate": 3.246550563342989e-06, + "loss": 0.8263, + "step": 1218 + }, + { + "epoch": 1.61, + "grad_norm": 2.703125, + "learning_rate": 3.2440466056062776e-06, + "loss": 0.8043, + "step": 1219 + }, + { + "epoch": 1.61, + "grad_norm": 2.578125, + "learning_rate": 3.241541828657664e-06, + "loss": 0.8111, + "step": 1220 + }, + { + "epoch": 1.61, + "grad_norm": 2.578125, + "learning_rate": 3.2390362352549638e-06, + "loss": 0.7945, + "step": 1221 + }, + { + "epoch": 1.61, + "grad_norm": 2.609375, + "learning_rate": 3.2365298281568913e-06, + "loss": 0.7872, + "step": 1222 + }, + { + "epoch": 1.61, + "grad_norm": 2.625, + "learning_rate": 3.234022610123057e-06, + "loss": 0.8479, + "step": 1223 + }, + { + "epoch": 1.61, + "grad_norm": 2.640625, + "learning_rate": 3.231514583913963e-06, + "loss": 0.823, + "step": 1224 + }, + { + "epoch": 1.62, + "grad_norm": 2.515625, + "learning_rate": 3.2290057522910024e-06, + "loss": 0.8021, + "step": 1225 + }, + { + "epoch": 1.62, + "grad_norm": 2.609375, + "learning_rate": 3.2264961180164558e-06, + "loss": 0.8254, + "step": 1226 + }, + { + "epoch": 1.62, + "grad_norm": 2.5625, + "learning_rate": 3.2239856838534846e-06, + "loss": 0.8329, + "step": 1227 + }, + { + "epoch": 1.62, + "grad_norm": 2.671875, + "learning_rate": 3.2214744525661336e-06, + "loss": 0.8378, + "step": 1228 + }, + { + "epoch": 1.62, + "grad_norm": 2.5625, + "learning_rate": 3.218962426919324e-06, + "loss": 0.8514, + "step": 1229 + }, + { + "epoch": 1.62, + "grad_norm": 2.5, + "learning_rate": 3.2164496096788534e-06, + "loss": 0.7885, + "step": 1230 + }, + { + "epoch": 1.62, + "grad_norm": 2.515625, + "learning_rate": 3.2139360036113885e-06, + "loss": 0.8149, + "step": 1231 + }, + { + "epoch": 1.62, + "grad_norm": 2.609375, + "learning_rate": 3.211421611484467e-06, + "loss": 0.8419, + "step": 1232 + }, + { + "epoch": 1.63, + "grad_norm": 2.65625, + "learning_rate": 3.208906436066488e-06, + "loss": 0.8511, + "step": 1233 + }, + { + "epoch": 1.63, + "grad_norm": 2.671875, + "learning_rate": 3.2063904801267184e-06, + "loss": 0.8266, + "step": 1234 + }, + { + "epoch": 1.63, + "grad_norm": 2.546875, + "learning_rate": 3.203873746435281e-06, + "loss": 0.7924, + "step": 1235 + }, + { + "epoch": 1.63, + "grad_norm": 2.6875, + "learning_rate": 3.2013562377631546e-06, + "loss": 0.8432, + "step": 1236 + }, + { + "epoch": 1.63, + "grad_norm": 2.5625, + "learning_rate": 3.198837956882173e-06, + "loss": 0.821, + "step": 1237 + }, + { + "epoch": 1.63, + "grad_norm": 2.53125, + "learning_rate": 3.19631890656502e-06, + "loss": 0.7895, + "step": 1238 + }, + { + "epoch": 1.63, + "grad_norm": 2.53125, + "learning_rate": 3.1937990895852263e-06, + "loss": 0.8083, + "step": 1239 + }, + { + "epoch": 1.64, + "grad_norm": 2.53125, + "learning_rate": 3.191278508717166e-06, + "loss": 0.8354, + "step": 1240 + }, + { + "epoch": 1.64, + "grad_norm": 2.53125, + "learning_rate": 3.1887571667360557e-06, + "loss": 0.8007, + "step": 1241 + }, + { + "epoch": 1.64, + "grad_norm": 2.625, + "learning_rate": 3.186235066417947e-06, + "loss": 0.8017, + "step": 1242 + }, + { + "epoch": 1.64, + "grad_norm": 2.609375, + "learning_rate": 3.1837122105397312e-06, + "loss": 0.8297, + "step": 1243 + }, + { + "epoch": 1.64, + "grad_norm": 2.671875, + "learning_rate": 3.1811886018791276e-06, + "loss": 0.8202, + "step": 1244 + }, + { + "epoch": 1.64, + "grad_norm": 2.5625, + "learning_rate": 3.1786642432146857e-06, + "loss": 0.8341, + "step": 1245 + }, + { + "epoch": 1.64, + "grad_norm": 2.59375, + "learning_rate": 3.176139137325781e-06, + "loss": 0.8524, + "step": 1246 + }, + { + "epoch": 1.64, + "grad_norm": 2.546875, + "learning_rate": 3.1736132869926105e-06, + "loss": 0.8099, + "step": 1247 + }, + { + "epoch": 1.65, + "grad_norm": 2.546875, + "learning_rate": 3.1710866949961927e-06, + "loss": 0.8119, + "step": 1248 + }, + { + "epoch": 1.65, + "grad_norm": 2.578125, + "learning_rate": 3.168559364118361e-06, + "loss": 0.8109, + "step": 1249 + }, + { + "epoch": 1.65, + "grad_norm": 2.640625, + "learning_rate": 3.166031297141764e-06, + "loss": 0.8546, + "step": 1250 + }, + { + "epoch": 1.65, + "grad_norm": 2.65625, + "learning_rate": 3.1635024968498583e-06, + "loss": 0.8182, + "step": 1251 + }, + { + "epoch": 1.65, + "grad_norm": 2.640625, + "learning_rate": 3.1609729660269114e-06, + "loss": 0.8266, + "step": 1252 + }, + { + "epoch": 1.65, + "grad_norm": 2.671875, + "learning_rate": 3.1584427074579917e-06, + "loss": 0.8208, + "step": 1253 + }, + { + "epoch": 1.65, + "grad_norm": 2.671875, + "learning_rate": 3.1559117239289707e-06, + "loss": 0.8143, + "step": 1254 + }, + { + "epoch": 1.66, + "grad_norm": 2.65625, + "learning_rate": 3.1533800182265173e-06, + "loss": 0.8284, + "step": 1255 + }, + { + "epoch": 1.66, + "grad_norm": 2.703125, + "learning_rate": 3.150847593138096e-06, + "loss": 0.8091, + "step": 1256 + }, + { + "epoch": 1.66, + "grad_norm": 2.5625, + "learning_rate": 3.1483144514519637e-06, + "loss": 0.8322, + "step": 1257 + }, + { + "epoch": 1.66, + "grad_norm": 2.59375, + "learning_rate": 3.1457805959571663e-06, + "loss": 0.8297, + "step": 1258 + }, + { + "epoch": 1.66, + "grad_norm": 2.625, + "learning_rate": 3.1432460294435346e-06, + "loss": 0.7879, + "step": 1259 + }, + { + "epoch": 1.66, + "grad_norm": 2.5, + "learning_rate": 3.1407107547016818e-06, + "loss": 0.8403, + "step": 1260 + }, + { + "epoch": 1.66, + "grad_norm": 2.46875, + "learning_rate": 3.138174774523004e-06, + "loss": 0.7887, + "step": 1261 + }, + { + "epoch": 1.66, + "grad_norm": 2.53125, + "learning_rate": 3.135638091699671e-06, + "loss": 0.7863, + "step": 1262 + }, + { + "epoch": 1.67, + "grad_norm": 2.71875, + "learning_rate": 3.1331007090246263e-06, + "loss": 0.8039, + "step": 1263 + }, + { + "epoch": 1.67, + "grad_norm": 2.640625, + "learning_rate": 3.130562629291586e-06, + "loss": 0.8041, + "step": 1264 + }, + { + "epoch": 1.67, + "grad_norm": 2.671875, + "learning_rate": 3.1280238552950315e-06, + "loss": 0.8182, + "step": 1265 + }, + { + "epoch": 1.67, + "grad_norm": 2.84375, + "learning_rate": 3.1254843898302104e-06, + "loss": 0.8406, + "step": 1266 + }, + { + "epoch": 1.67, + "grad_norm": 2.65625, + "learning_rate": 3.122944235693131e-06, + "loss": 0.8179, + "step": 1267 + }, + { + "epoch": 1.67, + "grad_norm": 2.5, + "learning_rate": 3.1204033956805583e-06, + "loss": 0.8321, + "step": 1268 + }, + { + "epoch": 1.67, + "grad_norm": 2.65625, + "learning_rate": 3.117861872590015e-06, + "loss": 0.8054, + "step": 1269 + }, + { + "epoch": 1.67, + "grad_norm": 2.515625, + "learning_rate": 3.1153196692197747e-06, + "loss": 0.8033, + "step": 1270 + }, + { + "epoch": 1.68, + "grad_norm": 2.65625, + "learning_rate": 3.112776788368859e-06, + "loss": 0.8301, + "step": 1271 + }, + { + "epoch": 1.68, + "grad_norm": 2.53125, + "learning_rate": 3.1102332328370373e-06, + "loss": 0.8151, + "step": 1272 + }, + { + "epoch": 1.68, + "grad_norm": 2.546875, + "learning_rate": 3.1076890054248206e-06, + "loss": 0.8303, + "step": 1273 + }, + { + "epoch": 1.68, + "grad_norm": 2.640625, + "learning_rate": 3.1051441089334596e-06, + "loss": 0.8148, + "step": 1274 + }, + { + "epoch": 1.68, + "grad_norm": 2.796875, + "learning_rate": 3.102598546164943e-06, + "loss": 0.8369, + "step": 1275 + }, + { + "epoch": 1.68, + "grad_norm": 2.828125, + "learning_rate": 3.100052319921992e-06, + "loss": 0.8843, + "step": 1276 + }, + { + "epoch": 1.68, + "grad_norm": 2.578125, + "learning_rate": 3.097505433008058e-06, + "loss": 0.7975, + "step": 1277 + }, + { + "epoch": 1.69, + "grad_norm": 2.625, + "learning_rate": 3.09495788822732e-06, + "loss": 0.8324, + "step": 1278 + }, + { + "epoch": 1.69, + "grad_norm": 2.6875, + "learning_rate": 3.092409688384682e-06, + "loss": 0.835, + "step": 1279 + }, + { + "epoch": 1.69, + "grad_norm": 2.5, + "learning_rate": 3.0898608362857695e-06, + "loss": 0.7959, + "step": 1280 + }, + { + "epoch": 1.69, + "grad_norm": 2.5625, + "learning_rate": 3.087311334736925e-06, + "loss": 0.8066, + "step": 1281 + }, + { + "epoch": 1.69, + "grad_norm": 2.609375, + "learning_rate": 3.0847611865452064e-06, + "loss": 0.8064, + "step": 1282 + }, + { + "epoch": 1.69, + "grad_norm": 2.6875, + "learning_rate": 3.082210394518383e-06, + "loss": 0.8514, + "step": 1283 + }, + { + "epoch": 1.69, + "grad_norm": 2.5625, + "learning_rate": 3.0796589614649347e-06, + "loss": 0.8229, + "step": 1284 + }, + { + "epoch": 1.69, + "grad_norm": 2.59375, + "learning_rate": 3.0771068901940467e-06, + "loss": 0.8198, + "step": 1285 + }, + { + "epoch": 1.7, + "grad_norm": 2.640625, + "learning_rate": 3.0745541835156047e-06, + "loss": 0.8296, + "step": 1286 + }, + { + "epoch": 1.7, + "grad_norm": 2.59375, + "learning_rate": 3.0720008442401966e-06, + "loss": 0.7983, + "step": 1287 + }, + { + "epoch": 1.7, + "grad_norm": 2.640625, + "learning_rate": 3.069446875179106e-06, + "loss": 0.8149, + "step": 1288 + }, + { + "epoch": 1.7, + "grad_norm": 2.5, + "learning_rate": 3.0668922791443097e-06, + "loss": 0.8253, + "step": 1289 + }, + { + "epoch": 1.7, + "grad_norm": 2.5, + "learning_rate": 3.0643370589484745e-06, + "loss": 0.8231, + "step": 1290 + }, + { + "epoch": 1.7, + "grad_norm": 2.46875, + "learning_rate": 3.061781217404956e-06, + "loss": 0.7827, + "step": 1291 + }, + { + "epoch": 1.7, + "grad_norm": 2.65625, + "learning_rate": 3.0592247573277905e-06, + "loss": 0.8325, + "step": 1292 + }, + { + "epoch": 1.71, + "grad_norm": 2.59375, + "learning_rate": 3.0566676815317e-06, + "loss": 0.7855, + "step": 1293 + }, + { + "epoch": 1.71, + "grad_norm": 2.5, + "learning_rate": 3.0541099928320806e-06, + "loss": 0.7898, + "step": 1294 + }, + { + "epoch": 1.71, + "grad_norm": 2.546875, + "learning_rate": 3.0515516940450046e-06, + "loss": 0.8148, + "step": 1295 + }, + { + "epoch": 1.71, + "grad_norm": 2.65625, + "learning_rate": 3.048992787987216e-06, + "loss": 0.8184, + "step": 1296 + }, + { + "epoch": 1.71, + "grad_norm": 2.734375, + "learning_rate": 3.046433277476128e-06, + "loss": 0.8459, + "step": 1297 + }, + { + "epoch": 1.71, + "grad_norm": 2.546875, + "learning_rate": 3.0438731653298183e-06, + "loss": 0.8259, + "step": 1298 + }, + { + "epoch": 1.71, + "grad_norm": 2.5, + "learning_rate": 3.041312454367027e-06, + "loss": 0.8183, + "step": 1299 + }, + { + "epoch": 1.71, + "grad_norm": 2.6875, + "learning_rate": 3.0387511474071556e-06, + "loss": 0.8266, + "step": 1300 + }, + { + "epoch": 1.72, + "grad_norm": 2.5, + "learning_rate": 3.036189247270258e-06, + "loss": 0.8084, + "step": 1301 + }, + { + "epoch": 1.72, + "grad_norm": 2.578125, + "learning_rate": 3.033626756777045e-06, + "loss": 0.8341, + "step": 1302 + }, + { + "epoch": 1.72, + "grad_norm": 2.59375, + "learning_rate": 3.031063678748875e-06, + "loss": 0.8315, + "step": 1303 + }, + { + "epoch": 1.72, + "grad_norm": 2.578125, + "learning_rate": 3.0285000160077545e-06, + "loss": 0.8339, + "step": 1304 + }, + { + "epoch": 1.72, + "grad_norm": 2.609375, + "learning_rate": 3.0259357713763325e-06, + "loss": 0.8275, + "step": 1305 + }, + { + "epoch": 1.72, + "grad_norm": 2.625, + "learning_rate": 3.023370947677901e-06, + "loss": 0.8073, + "step": 1306 + }, + { + "epoch": 1.72, + "grad_norm": 2.65625, + "learning_rate": 3.020805547736387e-06, + "loss": 0.8124, + "step": 1307 + }, + { + "epoch": 1.73, + "grad_norm": 2.515625, + "learning_rate": 3.0182395743763528e-06, + "loss": 0.8072, + "step": 1308 + }, + { + "epoch": 1.73, + "grad_norm": 2.578125, + "learning_rate": 3.015673030422993e-06, + "loss": 0.8062, + "step": 1309 + }, + { + "epoch": 1.73, + "grad_norm": 2.53125, + "learning_rate": 3.013105918702128e-06, + "loss": 0.8039, + "step": 1310 + }, + { + "epoch": 1.73, + "grad_norm": 2.53125, + "learning_rate": 3.0105382420402085e-06, + "loss": 0.8028, + "step": 1311 + }, + { + "epoch": 1.73, + "grad_norm": 2.484375, + "learning_rate": 3.007970003264301e-06, + "loss": 0.8047, + "step": 1312 + }, + { + "epoch": 1.73, + "grad_norm": 2.625, + "learning_rate": 3.005401205202094e-06, + "loss": 0.8184, + "step": 1313 + }, + { + "epoch": 1.73, + "grad_norm": 2.625, + "learning_rate": 3.002831850681892e-06, + "loss": 0.817, + "step": 1314 + }, + { + "epoch": 1.73, + "grad_norm": 2.578125, + "learning_rate": 3.000261942532612e-06, + "loss": 0.7928, + "step": 1315 + }, + { + "epoch": 1.74, + "grad_norm": 2.640625, + "learning_rate": 2.9976914835837787e-06, + "loss": 0.8264, + "step": 1316 + }, + { + "epoch": 1.74, + "grad_norm": 2.59375, + "learning_rate": 2.995120476665526e-06, + "loss": 0.8076, + "step": 1317 + }, + { + "epoch": 1.74, + "grad_norm": 2.5, + "learning_rate": 2.99254892460859e-06, + "loss": 0.824, + "step": 1318 + }, + { + "epoch": 1.74, + "grad_norm": 2.453125, + "learning_rate": 2.9899768302443044e-06, + "loss": 0.8053, + "step": 1319 + }, + { + "epoch": 1.74, + "grad_norm": 2.734375, + "learning_rate": 2.9874041964046064e-06, + "loss": 0.819, + "step": 1320 + }, + { + "epoch": 1.74, + "grad_norm": 2.671875, + "learning_rate": 2.9848310259220203e-06, + "loss": 0.8158, + "step": 1321 + }, + { + "epoch": 1.74, + "grad_norm": 2.59375, + "learning_rate": 2.9822573216296647e-06, + "loss": 0.8007, + "step": 1322 + }, + { + "epoch": 1.75, + "grad_norm": 2.484375, + "learning_rate": 2.979683086361246e-06, + "loss": 0.8101, + "step": 1323 + }, + { + "epoch": 1.75, + "grad_norm": 2.5625, + "learning_rate": 2.9771083229510543e-06, + "loss": 0.8146, + "step": 1324 + }, + { + "epoch": 1.75, + "grad_norm": 2.703125, + "learning_rate": 2.9745330342339618e-06, + "loss": 0.8116, + "step": 1325 + }, + { + "epoch": 1.75, + "grad_norm": 2.671875, + "learning_rate": 2.9719572230454184e-06, + "loss": 0.7909, + "step": 1326 + }, + { + "epoch": 1.75, + "grad_norm": 2.484375, + "learning_rate": 2.96938089222145e-06, + "loss": 0.7935, + "step": 1327 + }, + { + "epoch": 1.75, + "grad_norm": 2.71875, + "learning_rate": 2.9668040445986536e-06, + "loss": 0.8465, + "step": 1328 + }, + { + "epoch": 1.75, + "grad_norm": 2.46875, + "learning_rate": 2.9642266830141965e-06, + "loss": 0.8147, + "step": 1329 + }, + { + "epoch": 1.75, + "grad_norm": 2.546875, + "learning_rate": 2.9616488103058115e-06, + "loss": 0.8281, + "step": 1330 + }, + { + "epoch": 1.76, + "grad_norm": 2.515625, + "learning_rate": 2.959070429311793e-06, + "loss": 0.8391, + "step": 1331 + }, + { + "epoch": 1.76, + "grad_norm": 2.609375, + "learning_rate": 2.9564915428709968e-06, + "loss": 0.8322, + "step": 1332 + }, + { + "epoch": 1.76, + "grad_norm": 2.625, + "learning_rate": 2.953912153822834e-06, + "loss": 0.8182, + "step": 1333 + }, + { + "epoch": 1.76, + "grad_norm": 2.609375, + "learning_rate": 2.95133226500727e-06, + "loss": 0.8099, + "step": 1334 + }, + { + "epoch": 1.76, + "grad_norm": 2.46875, + "learning_rate": 2.9487518792648194e-06, + "loss": 0.8105, + "step": 1335 + }, + { + "epoch": 1.76, + "grad_norm": 2.5625, + "learning_rate": 2.9461709994365445e-06, + "loss": 0.8055, + "step": 1336 + }, + { + "epoch": 1.76, + "grad_norm": 2.65625, + "learning_rate": 2.9435896283640507e-06, + "loss": 0.8236, + "step": 1337 + }, + { + "epoch": 1.77, + "grad_norm": 2.71875, + "learning_rate": 2.9410077688894868e-06, + "loss": 0.8359, + "step": 1338 + }, + { + "epoch": 1.77, + "grad_norm": 2.53125, + "learning_rate": 2.9384254238555364e-06, + "loss": 0.8079, + "step": 1339 + }, + { + "epoch": 1.77, + "grad_norm": 2.4375, + "learning_rate": 2.9358425961054194e-06, + "loss": 0.8302, + "step": 1340 + }, + { + "epoch": 1.77, + "grad_norm": 2.609375, + "learning_rate": 2.933259288482887e-06, + "loss": 0.8205, + "step": 1341 + }, + { + "epoch": 1.77, + "grad_norm": 2.5, + "learning_rate": 2.930675503832217e-06, + "loss": 0.8231, + "step": 1342 + }, + { + "epoch": 1.77, + "grad_norm": 2.609375, + "learning_rate": 2.9280912449982164e-06, + "loss": 0.8343, + "step": 1343 + }, + { + "epoch": 1.77, + "grad_norm": 2.4375, + "learning_rate": 2.9255065148262102e-06, + "loss": 0.7995, + "step": 1344 + }, + { + "epoch": 1.77, + "grad_norm": 2.5625, + "learning_rate": 2.922921316162044e-06, + "loss": 0.7969, + "step": 1345 + }, + { + "epoch": 1.78, + "grad_norm": 2.515625, + "learning_rate": 2.920335651852079e-06, + "loss": 0.8198, + "step": 1346 + }, + { + "epoch": 1.78, + "grad_norm": 2.53125, + "learning_rate": 2.9177495247431898e-06, + "loss": 0.8297, + "step": 1347 + }, + { + "epoch": 1.78, + "grad_norm": 2.5625, + "learning_rate": 2.91516293768276e-06, + "loss": 0.7851, + "step": 1348 + }, + { + "epoch": 1.78, + "grad_norm": 2.65625, + "learning_rate": 2.9125758935186793e-06, + "loss": 0.8066, + "step": 1349 + }, + { + "epoch": 1.78, + "grad_norm": 2.578125, + "learning_rate": 2.9099883950993417e-06, + "loss": 0.7993, + "step": 1350 + }, + { + "epoch": 1.78, + "grad_norm": 2.734375, + "learning_rate": 2.907400445273639e-06, + "loss": 0.8126, + "step": 1351 + }, + { + "epoch": 1.78, + "grad_norm": 2.609375, + "learning_rate": 2.904812046890964e-06, + "loss": 0.8313, + "step": 1352 + }, + { + "epoch": 1.79, + "grad_norm": 2.53125, + "learning_rate": 2.9022232028012e-06, + "loss": 0.8223, + "step": 1353 + }, + { + "epoch": 1.79, + "grad_norm": 2.578125, + "learning_rate": 2.899633915854721e-06, + "loss": 0.8414, + "step": 1354 + }, + { + "epoch": 1.79, + "grad_norm": 2.578125, + "learning_rate": 2.8970441889023905e-06, + "loss": 0.8014, + "step": 1355 + }, + { + "epoch": 1.79, + "grad_norm": 2.65625, + "learning_rate": 2.8944540247955554e-06, + "loss": 0.8216, + "step": 1356 + }, + { + "epoch": 1.79, + "grad_norm": 2.71875, + "learning_rate": 2.8918634263860453e-06, + "loss": 0.8292, + "step": 1357 + }, + { + "epoch": 1.79, + "grad_norm": 2.609375, + "learning_rate": 2.889272396526165e-06, + "loss": 0.8159, + "step": 1358 + }, + { + "epoch": 1.79, + "grad_norm": 2.671875, + "learning_rate": 2.8866809380686977e-06, + "loss": 0.8279, + "step": 1359 + }, + { + "epoch": 1.79, + "grad_norm": 2.65625, + "learning_rate": 2.8840890538668955e-06, + "loss": 0.8002, + "step": 1360 + }, + { + "epoch": 1.8, + "grad_norm": 2.609375, + "learning_rate": 2.881496746774482e-06, + "loss": 0.8291, + "step": 1361 + }, + { + "epoch": 1.8, + "grad_norm": 2.578125, + "learning_rate": 2.878904019645645e-06, + "loss": 0.7963, + "step": 1362 + }, + { + "epoch": 1.8, + "grad_norm": 2.5, + "learning_rate": 2.8763108753350338e-06, + "loss": 0.824, + "step": 1363 + }, + { + "epoch": 1.8, + "grad_norm": 2.671875, + "learning_rate": 2.873717316697759e-06, + "loss": 0.7929, + "step": 1364 + }, + { + "epoch": 1.8, + "grad_norm": 2.671875, + "learning_rate": 2.8711233465893855e-06, + "loss": 0.8015, + "step": 1365 + }, + { + "epoch": 1.8, + "grad_norm": 2.5625, + "learning_rate": 2.868528967865934e-06, + "loss": 0.8074, + "step": 1366 + }, + { + "epoch": 1.8, + "grad_norm": 2.625, + "learning_rate": 2.8659341833838715e-06, + "loss": 0.8242, + "step": 1367 + }, + { + "epoch": 1.81, + "grad_norm": 2.609375, + "learning_rate": 2.863338996000115e-06, + "loss": 0.8033, + "step": 1368 + }, + { + "epoch": 1.81, + "grad_norm": 2.703125, + "learning_rate": 2.860743408572022e-06, + "loss": 0.8203, + "step": 1369 + }, + { + "epoch": 1.81, + "grad_norm": 2.625, + "learning_rate": 2.858147423957394e-06, + "loss": 0.8312, + "step": 1370 + }, + { + "epoch": 1.81, + "grad_norm": 2.546875, + "learning_rate": 2.855551045014467e-06, + "loss": 0.8427, + "step": 1371 + }, + { + "epoch": 1.81, + "grad_norm": 2.5625, + "learning_rate": 2.8529542746019118e-06, + "loss": 0.8146, + "step": 1372 + }, + { + "epoch": 1.81, + "grad_norm": 2.5, + "learning_rate": 2.8503571155788312e-06, + "loss": 0.7866, + "step": 1373 + }, + { + "epoch": 1.81, + "grad_norm": 2.5, + "learning_rate": 2.8477595708047543e-06, + "loss": 0.8114, + "step": 1374 + }, + { + "epoch": 1.81, + "grad_norm": 2.5, + "learning_rate": 2.8451616431396374e-06, + "loss": 0.7991, + "step": 1375 + }, + { + "epoch": 1.82, + "grad_norm": 2.5625, + "learning_rate": 2.842563335443856e-06, + "loss": 0.8275, + "step": 1376 + }, + { + "epoch": 1.82, + "grad_norm": 2.453125, + "learning_rate": 2.8399646505782035e-06, + "loss": 0.7847, + "step": 1377 + }, + { + "epoch": 1.82, + "grad_norm": 2.5625, + "learning_rate": 2.8373655914038907e-06, + "loss": 0.8341, + "step": 1378 + }, + { + "epoch": 1.82, + "grad_norm": 2.609375, + "learning_rate": 2.834766160782541e-06, + "loss": 0.7867, + "step": 1379 + }, + { + "epoch": 1.82, + "grad_norm": 2.59375, + "learning_rate": 2.832166361576184e-06, + "loss": 0.804, + "step": 1380 + }, + { + "epoch": 1.82, + "grad_norm": 2.625, + "learning_rate": 2.829566196647257e-06, + "loss": 0.8323, + "step": 1381 + }, + { + "epoch": 1.82, + "grad_norm": 2.59375, + "learning_rate": 2.8269656688585994e-06, + "loss": 0.8342, + "step": 1382 + }, + { + "epoch": 1.83, + "grad_norm": 2.625, + "learning_rate": 2.8243647810734505e-06, + "loss": 0.8321, + "step": 1383 + }, + { + "epoch": 1.83, + "grad_norm": 2.546875, + "learning_rate": 2.821763536155446e-06, + "loss": 0.8248, + "step": 1384 + }, + { + "epoch": 1.83, + "grad_norm": 2.46875, + "learning_rate": 2.819161936968615e-06, + "loss": 0.8087, + "step": 1385 + }, + { + "epoch": 1.83, + "grad_norm": 2.609375, + "learning_rate": 2.816559986377374e-06, + "loss": 0.8085, + "step": 1386 + }, + { + "epoch": 1.83, + "grad_norm": 2.5, + "learning_rate": 2.8139576872465306e-06, + "loss": 0.7812, + "step": 1387 + }, + { + "epoch": 1.83, + "grad_norm": 2.59375, + "learning_rate": 2.8113550424412733e-06, + "loss": 0.828, + "step": 1388 + }, + { + "epoch": 1.83, + "grad_norm": 2.515625, + "learning_rate": 2.808752054827173e-06, + "loss": 0.7832, + "step": 1389 + }, + { + "epoch": 1.83, + "grad_norm": 2.59375, + "learning_rate": 2.806148727270176e-06, + "loss": 0.8053, + "step": 1390 + }, + { + "epoch": 1.84, + "grad_norm": 2.53125, + "learning_rate": 2.8035450626366047e-06, + "loss": 0.8171, + "step": 1391 + }, + { + "epoch": 1.84, + "grad_norm": 2.53125, + "learning_rate": 2.8009410637931516e-06, + "loss": 0.8164, + "step": 1392 + }, + { + "epoch": 1.84, + "grad_norm": 2.5625, + "learning_rate": 2.798336733606878e-06, + "loss": 0.8025, + "step": 1393 + }, + { + "epoch": 1.84, + "grad_norm": 2.5625, + "learning_rate": 2.795732074945209e-06, + "loss": 0.8056, + "step": 1394 + }, + { + "epoch": 1.84, + "grad_norm": 2.640625, + "learning_rate": 2.7931270906759317e-06, + "loss": 0.8168, + "step": 1395 + }, + { + "epoch": 1.84, + "grad_norm": 2.59375, + "learning_rate": 2.7905217836671915e-06, + "loss": 0.8085, + "step": 1396 + }, + { + "epoch": 1.84, + "grad_norm": 2.5625, + "learning_rate": 2.7879161567874905e-06, + "loss": 0.8186, + "step": 1397 + }, + { + "epoch": 1.85, + "grad_norm": 2.640625, + "learning_rate": 2.785310212905682e-06, + "loss": 0.8307, + "step": 1398 + }, + { + "epoch": 1.85, + "grad_norm": 2.53125, + "learning_rate": 2.7827039548909675e-06, + "loss": 0.8165, + "step": 1399 + }, + { + "epoch": 1.85, + "grad_norm": 2.515625, + "learning_rate": 2.780097385612896e-06, + "loss": 0.7957, + "step": 1400 + }, + { + "epoch": 1.85, + "grad_norm": 2.703125, + "learning_rate": 2.7774905079413573e-06, + "loss": 0.8171, + "step": 1401 + }, + { + "epoch": 1.85, + "grad_norm": 2.859375, + "learning_rate": 2.774883324746583e-06, + "loss": 0.8398, + "step": 1402 + }, + { + "epoch": 1.85, + "grad_norm": 2.625, + "learning_rate": 2.7722758388991394e-06, + "loss": 0.8108, + "step": 1403 + }, + { + "epoch": 1.85, + "grad_norm": 2.578125, + "learning_rate": 2.7696680532699265e-06, + "loss": 0.7974, + "step": 1404 + }, + { + "epoch": 1.85, + "grad_norm": 2.5625, + "learning_rate": 2.767059970730175e-06, + "loss": 0.8036, + "step": 1405 + }, + { + "epoch": 1.86, + "grad_norm": 2.6875, + "learning_rate": 2.7644515941514405e-06, + "loss": 0.8148, + "step": 1406 + }, + { + "epoch": 1.86, + "grad_norm": 2.65625, + "learning_rate": 2.761842926405605e-06, + "loss": 0.8053, + "step": 1407 + }, + { + "epoch": 1.86, + "grad_norm": 2.59375, + "learning_rate": 2.7592339703648696e-06, + "loss": 0.8307, + "step": 1408 + }, + { + "epoch": 1.86, + "grad_norm": 2.5625, + "learning_rate": 2.7566247289017537e-06, + "loss": 0.8193, + "step": 1409 + }, + { + "epoch": 1.86, + "grad_norm": 2.5625, + "learning_rate": 2.7540152048890885e-06, + "loss": 0.8107, + "step": 1410 + }, + { + "epoch": 1.86, + "grad_norm": 2.609375, + "learning_rate": 2.7514054012000197e-06, + "loss": 0.8431, + "step": 1411 + }, + { + "epoch": 1.86, + "grad_norm": 2.5625, + "learning_rate": 2.748795320708e-06, + "loss": 0.8243, + "step": 1412 + }, + { + "epoch": 1.87, + "grad_norm": 2.5625, + "learning_rate": 2.746184966286784e-06, + "loss": 0.8236, + "step": 1413 + }, + { + "epoch": 1.87, + "grad_norm": 2.578125, + "learning_rate": 2.743574340810431e-06, + "loss": 0.8046, + "step": 1414 + }, + { + "epoch": 1.87, + "grad_norm": 2.5625, + "learning_rate": 2.7409634471532987e-06, + "loss": 0.8243, + "step": 1415 + }, + { + "epoch": 1.87, + "grad_norm": 2.5, + "learning_rate": 2.7383522881900382e-06, + "loss": 0.8437, + "step": 1416 + }, + { + "epoch": 1.87, + "grad_norm": 2.59375, + "learning_rate": 2.735740866795595e-06, + "loss": 0.8108, + "step": 1417 + }, + { + "epoch": 1.87, + "grad_norm": 2.671875, + "learning_rate": 2.733129185845201e-06, + "loss": 0.8425, + "step": 1418 + }, + { + "epoch": 1.87, + "grad_norm": 2.609375, + "learning_rate": 2.730517248214375e-06, + "loss": 0.8367, + "step": 1419 + }, + { + "epoch": 1.87, + "grad_norm": 2.65625, + "learning_rate": 2.7279050567789195e-06, + "loss": 0.8231, + "step": 1420 + }, + { + "epoch": 1.88, + "grad_norm": 2.515625, + "learning_rate": 2.725292614414916e-06, + "loss": 0.8273, + "step": 1421 + }, + { + "epoch": 1.88, + "grad_norm": 2.84375, + "learning_rate": 2.72267992399872e-06, + "loss": 0.855, + "step": 1422 + }, + { + "epoch": 1.88, + "grad_norm": 2.671875, + "learning_rate": 2.720066988406963e-06, + "loss": 0.7913, + "step": 1423 + }, + { + "epoch": 1.88, + "grad_norm": 2.578125, + "learning_rate": 2.7174538105165454e-06, + "loss": 0.8044, + "step": 1424 + }, + { + "epoch": 1.88, + "grad_norm": 2.546875, + "learning_rate": 2.714840393204634e-06, + "loss": 0.8004, + "step": 1425 + }, + { + "epoch": 1.88, + "grad_norm": 2.53125, + "learning_rate": 2.7122267393486605e-06, + "loss": 0.7999, + "step": 1426 + }, + { + "epoch": 1.88, + "grad_norm": 2.5625, + "learning_rate": 2.709612851826315e-06, + "loss": 0.8258, + "step": 1427 + }, + { + "epoch": 1.89, + "grad_norm": 2.546875, + "learning_rate": 2.7069987335155445e-06, + "loss": 0.8334, + "step": 1428 + }, + { + "epoch": 1.89, + "grad_norm": 2.53125, + "learning_rate": 2.7043843872945554e-06, + "loss": 0.8485, + "step": 1429 + }, + { + "epoch": 1.89, + "grad_norm": 2.625, + "learning_rate": 2.7017698160417976e-06, + "loss": 0.8396, + "step": 1430 + }, + { + "epoch": 1.89, + "grad_norm": 2.640625, + "learning_rate": 2.699155022635974e-06, + "loss": 0.8221, + "step": 1431 + }, + { + "epoch": 1.89, + "grad_norm": 2.59375, + "learning_rate": 2.6965400099560305e-06, + "loss": 0.8276, + "step": 1432 + }, + { + "epoch": 1.89, + "grad_norm": 2.625, + "learning_rate": 2.693924780881154e-06, + "loss": 0.8274, + "step": 1433 + }, + { + "epoch": 1.89, + "grad_norm": 2.546875, + "learning_rate": 2.6913093382907696e-06, + "loss": 0.8359, + "step": 1434 + }, + { + "epoch": 1.89, + "grad_norm": 2.5625, + "learning_rate": 2.6886936850645397e-06, + "loss": 0.8093, + "step": 1435 + }, + { + "epoch": 1.9, + "grad_norm": 2.609375, + "learning_rate": 2.6860778240823536e-06, + "loss": 0.803, + "step": 1436 + }, + { + "epoch": 1.9, + "grad_norm": 2.578125, + "learning_rate": 2.6834617582243348e-06, + "loss": 0.8266, + "step": 1437 + }, + { + "epoch": 1.9, + "grad_norm": 2.609375, + "learning_rate": 2.6808454903708313e-06, + "loss": 0.8398, + "step": 1438 + }, + { + "epoch": 1.9, + "grad_norm": 2.578125, + "learning_rate": 2.6782290234024092e-06, + "loss": 0.8067, + "step": 1439 + }, + { + "epoch": 1.9, + "grad_norm": 2.53125, + "learning_rate": 2.6756123601998598e-06, + "loss": 0.8159, + "step": 1440 + }, + { + "epoch": 1.9, + "grad_norm": 2.625, + "learning_rate": 2.672995503644186e-06, + "loss": 0.8172, + "step": 1441 + }, + { + "epoch": 1.9, + "grad_norm": 2.546875, + "learning_rate": 2.6703784566166053e-06, + "loss": 0.7978, + "step": 1442 + }, + { + "epoch": 1.91, + "grad_norm": 2.59375, + "learning_rate": 2.667761221998545e-06, + "loss": 0.8112, + "step": 1443 + }, + { + "epoch": 1.91, + "grad_norm": 2.5, + "learning_rate": 2.66514380267164e-06, + "loss": 0.7821, + "step": 1444 + }, + { + "epoch": 1.91, + "grad_norm": 2.59375, + "learning_rate": 2.662526201517725e-06, + "loss": 0.821, + "step": 1445 + }, + { + "epoch": 1.91, + "grad_norm": 2.546875, + "learning_rate": 2.6599084214188393e-06, + "loss": 0.8074, + "step": 1446 + }, + { + "epoch": 1.91, + "grad_norm": 2.546875, + "learning_rate": 2.6572904652572155e-06, + "loss": 0.856, + "step": 1447 + }, + { + "epoch": 1.91, + "grad_norm": 2.46875, + "learning_rate": 2.6546723359152834e-06, + "loss": 0.7976, + "step": 1448 + }, + { + "epoch": 1.91, + "grad_norm": 2.65625, + "learning_rate": 2.6520540362756606e-06, + "loss": 0.7961, + "step": 1449 + }, + { + "epoch": 1.91, + "grad_norm": 2.59375, + "learning_rate": 2.6494355692211537e-06, + "loss": 0.8359, + "step": 1450 + }, + { + "epoch": 1.92, + "grad_norm": 2.65625, + "learning_rate": 2.6468169376347535e-06, + "loss": 0.8104, + "step": 1451 + }, + { + "epoch": 1.92, + "grad_norm": 2.828125, + "learning_rate": 2.644198144399631e-06, + "loss": 0.7916, + "step": 1452 + }, + { + "epoch": 1.92, + "grad_norm": 2.625, + "learning_rate": 2.6415791923991373e-06, + "loss": 0.81, + "step": 1453 + }, + { + "epoch": 1.92, + "grad_norm": 2.5625, + "learning_rate": 2.638960084516796e-06, + "loss": 0.822, + "step": 1454 + }, + { + "epoch": 1.92, + "grad_norm": 2.546875, + "learning_rate": 2.636340823636303e-06, + "loss": 0.8044, + "step": 1455 + }, + { + "epoch": 1.92, + "grad_norm": 2.609375, + "learning_rate": 2.6337214126415237e-06, + "loss": 0.7878, + "step": 1456 + }, + { + "epoch": 1.92, + "grad_norm": 2.6875, + "learning_rate": 2.631101854416488e-06, + "loss": 0.8245, + "step": 1457 + }, + { + "epoch": 1.93, + "grad_norm": 2.46875, + "learning_rate": 2.628482151845387e-06, + "loss": 0.7784, + "step": 1458 + }, + { + "epoch": 1.93, + "grad_norm": 2.640625, + "learning_rate": 2.625862307812573e-06, + "loss": 0.791, + "step": 1459 + }, + { + "epoch": 1.93, + "grad_norm": 2.5, + "learning_rate": 2.62324232520255e-06, + "loss": 0.807, + "step": 1460 + }, + { + "epoch": 1.93, + "grad_norm": 2.640625, + "learning_rate": 2.620622206899981e-06, + "loss": 0.8206, + "step": 1461 + }, + { + "epoch": 1.93, + "grad_norm": 2.5, + "learning_rate": 2.6180019557896725e-06, + "loss": 0.7767, + "step": 1462 + }, + { + "epoch": 1.93, + "grad_norm": 2.53125, + "learning_rate": 2.61538157475658e-06, + "loss": 0.8287, + "step": 1463 + }, + { + "epoch": 1.93, + "grad_norm": 2.546875, + "learning_rate": 2.6127610666858e-06, + "loss": 0.8129, + "step": 1464 + }, + { + "epoch": 1.93, + "grad_norm": 2.546875, + "learning_rate": 2.6101404344625725e-06, + "loss": 0.8164, + "step": 1465 + }, + { + "epoch": 1.94, + "grad_norm": 2.59375, + "learning_rate": 2.6075196809722704e-06, + "loss": 0.8131, + "step": 1466 + }, + { + "epoch": 1.94, + "grad_norm": 2.609375, + "learning_rate": 2.6048988091004023e-06, + "loss": 0.8195, + "step": 1467 + }, + { + "epoch": 1.94, + "grad_norm": 2.609375, + "learning_rate": 2.6022778217326077e-06, + "loss": 0.7988, + "step": 1468 + }, + { + "epoch": 1.94, + "grad_norm": 2.625, + "learning_rate": 2.59965672175465e-06, + "loss": 0.8218, + "step": 1469 + }, + { + "epoch": 1.94, + "grad_norm": 2.484375, + "learning_rate": 2.5970355120524194e-06, + "loss": 0.7993, + "step": 1470 + }, + { + "epoch": 1.94, + "grad_norm": 2.5625, + "learning_rate": 2.5944141955119277e-06, + "loss": 0.829, + "step": 1471 + }, + { + "epoch": 1.94, + "grad_norm": 2.546875, + "learning_rate": 2.591792775019301e-06, + "loss": 0.8012, + "step": 1472 + }, + { + "epoch": 1.95, + "grad_norm": 2.484375, + "learning_rate": 2.589171253460783e-06, + "loss": 0.793, + "step": 1473 + }, + { + "epoch": 1.95, + "grad_norm": 2.640625, + "learning_rate": 2.586549633722726e-06, + "loss": 0.858, + "step": 1474 + }, + { + "epoch": 1.95, + "grad_norm": 2.71875, + "learning_rate": 2.583927918691593e-06, + "loss": 0.8305, + "step": 1475 + }, + { + "epoch": 1.95, + "grad_norm": 2.59375, + "learning_rate": 2.5813061112539502e-06, + "loss": 0.8243, + "step": 1476 + }, + { + "epoch": 1.95, + "grad_norm": 2.578125, + "learning_rate": 2.5786842142964663e-06, + "loss": 0.8179, + "step": 1477 + }, + { + "epoch": 1.95, + "grad_norm": 2.53125, + "learning_rate": 2.5760622307059073e-06, + "loss": 0.8258, + "step": 1478 + }, + { + "epoch": 1.95, + "grad_norm": 2.5, + "learning_rate": 2.5734401633691373e-06, + "loss": 0.7936, + "step": 1479 + }, + { + "epoch": 1.95, + "grad_norm": 2.546875, + "learning_rate": 2.5708180151731105e-06, + "loss": 0.8248, + "step": 1480 + }, + { + "epoch": 1.96, + "grad_norm": 2.671875, + "learning_rate": 2.5681957890048693e-06, + "loss": 0.8664, + "step": 1481 + }, + { + "epoch": 1.96, + "grad_norm": 2.609375, + "learning_rate": 2.5655734877515437e-06, + "loss": 0.8316, + "step": 1482 + }, + { + "epoch": 1.96, + "grad_norm": 2.671875, + "learning_rate": 2.5629511143003466e-06, + "loss": 0.8261, + "step": 1483 + }, + { + "epoch": 1.96, + "grad_norm": 2.578125, + "learning_rate": 2.5603286715385687e-06, + "loss": 0.8197, + "step": 1484 + }, + { + "epoch": 1.96, + "grad_norm": 2.546875, + "learning_rate": 2.5577061623535793e-06, + "loss": 0.8028, + "step": 1485 + }, + { + "epoch": 1.96, + "grad_norm": 2.578125, + "learning_rate": 2.555083589632818e-06, + "loss": 0.8401, + "step": 1486 + }, + { + "epoch": 1.96, + "grad_norm": 2.515625, + "learning_rate": 2.5524609562637952e-06, + "loss": 0.822, + "step": 1487 + }, + { + "epoch": 1.97, + "grad_norm": 2.53125, + "learning_rate": 2.5498382651340916e-06, + "loss": 0.7908, + "step": 1488 + }, + { + "epoch": 1.97, + "grad_norm": 2.5, + "learning_rate": 2.5472155191313464e-06, + "loss": 0.7987, + "step": 1489 + }, + { + "epoch": 1.97, + "grad_norm": 2.609375, + "learning_rate": 2.5445927211432616e-06, + "loss": 0.8136, + "step": 1490 + }, + { + "epoch": 1.97, + "grad_norm": 2.640625, + "learning_rate": 2.5419698740575967e-06, + "loss": 0.789, + "step": 1491 + }, + { + "epoch": 1.97, + "grad_norm": 2.59375, + "learning_rate": 2.5393469807621646e-06, + "loss": 0.8183, + "step": 1492 + }, + { + "epoch": 1.97, + "grad_norm": 2.578125, + "learning_rate": 2.5367240441448297e-06, + "loss": 0.8176, + "step": 1493 + }, + { + "epoch": 1.97, + "grad_norm": 2.59375, + "learning_rate": 2.534101067093504e-06, + "loss": 0.8545, + "step": 1494 + }, + { + "epoch": 1.97, + "grad_norm": 2.5625, + "learning_rate": 2.5314780524961426e-06, + "loss": 0.8306, + "step": 1495 + }, + { + "epoch": 1.98, + "grad_norm": 2.625, + "learning_rate": 2.528855003240742e-06, + "loss": 0.8058, + "step": 1496 + }, + { + "epoch": 1.98, + "grad_norm": 2.484375, + "learning_rate": 2.526231922215342e-06, + "loss": 0.8133, + "step": 1497 + }, + { + "epoch": 1.98, + "grad_norm": 2.59375, + "learning_rate": 2.523608812308009e-06, + "loss": 0.8205, + "step": 1498 + }, + { + "epoch": 1.98, + "grad_norm": 2.53125, + "learning_rate": 2.5209856764068484e-06, + "loss": 0.798, + "step": 1499 + }, + { + "epoch": 1.98, + "grad_norm": 2.484375, + "learning_rate": 2.51836251739999e-06, + "loss": 0.7946, + "step": 1500 + }, + { + "epoch": 1.98, + "grad_norm": 2.515625, + "learning_rate": 2.5157393381755896e-06, + "loss": 0.7726, + "step": 1501 + }, + { + "epoch": 1.98, + "grad_norm": 2.5625, + "learning_rate": 2.513116141621828e-06, + "loss": 0.7807, + "step": 1502 + }, + { + "epoch": 1.99, + "grad_norm": 2.625, + "learning_rate": 2.510492930626902e-06, + "loss": 0.8439, + "step": 1503 + }, + { + "epoch": 1.99, + "grad_norm": 2.671875, + "learning_rate": 2.5078697080790248e-06, + "loss": 0.7976, + "step": 1504 + }, + { + "epoch": 1.99, + "grad_norm": 2.65625, + "learning_rate": 2.5052464768664243e-06, + "loss": 0.8466, + "step": 1505 + }, + { + "epoch": 1.99, + "grad_norm": 2.4375, + "learning_rate": 2.502623239877335e-06, + "loss": 0.7916, + "step": 1506 + }, + { + "epoch": 1.99, + "grad_norm": 2.578125, + "learning_rate": 2.5e-06, + "loss": 0.8214, + "step": 1507 + }, + { + "epoch": 1.99, + "grad_norm": 2.546875, + "learning_rate": 2.497376760122666e-06, + "loss": 0.8393, + "step": 1508 + }, + { + "epoch": 1.99, + "grad_norm": 2.609375, + "learning_rate": 2.494753523133577e-06, + "loss": 0.7873, + "step": 1509 + }, + { + "epoch": 1.99, + "grad_norm": 2.578125, + "learning_rate": 2.4921302919209765e-06, + "loss": 0.8238, + "step": 1510 + }, + { + "epoch": 2.0, + "grad_norm": 2.609375, + "learning_rate": 2.4895070693730996e-06, + "loss": 0.821, + "step": 1511 + }, + { + "epoch": 2.0, + "grad_norm": 2.609375, + "learning_rate": 2.486883858378173e-06, + "loss": 0.8023, + "step": 1512 + }, + { + "epoch": 2.0, + "grad_norm": 2.6875, + "learning_rate": 2.484260661824411e-06, + "loss": 0.7973, + "step": 1513 + }, + { + "epoch": 2.0, + "grad_norm": 2.609375, + "learning_rate": 2.481637482600011e-06, + "loss": 0.8436, + "step": 1514 + }, + { + "epoch": 2.0, + "grad_norm": 2.59375, + "learning_rate": 2.479014323593152e-06, + "loss": 0.8069, + "step": 1515 + }, + { + "epoch": 2.0, + "grad_norm": 2.703125, + "learning_rate": 2.476391187691992e-06, + "loss": 0.7881, + "step": 1516 + }, + { + "epoch": 2.0, + "grad_norm": 2.703125, + "learning_rate": 2.4737680777846596e-06, + "loss": 0.8321, + "step": 1517 + }, + { + "epoch": 2.0, + "grad_norm": 2.59375, + "learning_rate": 2.4711449967592584e-06, + "loss": 0.8096, + "step": 1518 + }, + { + "epoch": 2.01, + "grad_norm": 2.5, + "learning_rate": 2.4685219475038587e-06, + "loss": 0.7918, + "step": 1519 + }, + { + "epoch": 2.01, + "grad_norm": 2.46875, + "learning_rate": 2.465898932906497e-06, + "loss": 0.7983, + "step": 1520 + }, + { + "epoch": 2.01, + "grad_norm": 2.453125, + "learning_rate": 2.4632759558551707e-06, + "loss": 0.7984, + "step": 1521 + }, + { + "epoch": 2.01, + "grad_norm": 2.578125, + "learning_rate": 2.4606530192378358e-06, + "loss": 0.8128, + "step": 1522 + }, + { + "epoch": 2.01, + "grad_norm": 2.453125, + "learning_rate": 2.458030125942404e-06, + "loss": 0.7922, + "step": 1523 + }, + { + "epoch": 2.01, + "grad_norm": 2.59375, + "learning_rate": 2.455407278856739e-06, + "loss": 0.8341, + "step": 1524 + }, + { + "epoch": 2.01, + "grad_norm": 2.484375, + "learning_rate": 2.4527844808686544e-06, + "loss": 0.8039, + "step": 1525 + }, + { + "epoch": 2.0, + "grad_norm": 2.75, + "learning_rate": 2.450161734865909e-06, + "loss": 0.8008, + "step": 1526 + }, + { + "epoch": 2.0, + "grad_norm": 3.328125, + "learning_rate": 2.4475390437362048e-06, + "loss": 0.7544, + "step": 1527 + }, + { + "epoch": 2.0, + "grad_norm": 3.0625, + "learning_rate": 2.4449164103671834e-06, + "loss": 0.7761, + "step": 1528 + }, + { + "epoch": 2.0, + "grad_norm": 2.84375, + "learning_rate": 2.442293837646422e-06, + "loss": 0.7529, + "step": 1529 + }, + { + "epoch": 2.01, + "grad_norm": 2.8125, + "learning_rate": 2.4396713284614317e-06, + "loss": 0.7564, + "step": 1530 + }, + { + "epoch": 2.01, + "grad_norm": 2.96875, + "learning_rate": 2.437048885699654e-06, + "loss": 0.7614, + "step": 1531 + }, + { + "epoch": 2.01, + "grad_norm": 3.03125, + "learning_rate": 2.4344265122484568e-06, + "loss": 0.7568, + "step": 1532 + }, + { + "epoch": 2.01, + "grad_norm": 3.421875, + "learning_rate": 2.431804210995131e-06, + "loss": 0.7829, + "step": 1533 + }, + { + "epoch": 2.01, + "grad_norm": 3.265625, + "learning_rate": 2.4291819848268908e-06, + "loss": 0.786, + "step": 1534 + }, + { + "epoch": 2.01, + "grad_norm": 3.015625, + "learning_rate": 2.4265598366308636e-06, + "loss": 0.7774, + "step": 1535 + }, + { + "epoch": 2.01, + "grad_norm": 2.796875, + "learning_rate": 2.423937769294093e-06, + "loss": 0.7518, + "step": 1536 + }, + { + "epoch": 2.02, + "grad_norm": 2.71875, + "learning_rate": 2.4213157857035345e-06, + "loss": 0.7451, + "step": 1537 + }, + { + "epoch": 2.02, + "grad_norm": 2.765625, + "learning_rate": 2.41869388874605e-06, + "loss": 0.7497, + "step": 1538 + }, + { + "epoch": 2.02, + "grad_norm": 3.078125, + "learning_rate": 2.4160720813084075e-06, + "loss": 0.7748, + "step": 1539 + }, + { + "epoch": 2.02, + "grad_norm": 2.921875, + "learning_rate": 2.4134503662772754e-06, + "loss": 0.7539, + "step": 1540 + }, + { + "epoch": 2.02, + "grad_norm": 2.78125, + "learning_rate": 2.410828746539218e-06, + "loss": 0.744, + "step": 1541 + }, + { + "epoch": 2.02, + "grad_norm": 2.84375, + "learning_rate": 2.4082072249806994e-06, + "loss": 0.757, + "step": 1542 + }, + { + "epoch": 2.02, + "grad_norm": 2.734375, + "learning_rate": 2.4055858044880727e-06, + "loss": 0.7831, + "step": 1543 + }, + { + "epoch": 2.02, + "grad_norm": 2.6875, + "learning_rate": 2.4029644879475806e-06, + "loss": 0.7734, + "step": 1544 + }, + { + "epoch": 2.03, + "grad_norm": 2.609375, + "learning_rate": 2.4003432782453505e-06, + "loss": 0.7307, + "step": 1545 + }, + { + "epoch": 2.03, + "grad_norm": 2.796875, + "learning_rate": 2.3977221782673936e-06, + "loss": 0.7791, + "step": 1546 + }, + { + "epoch": 2.03, + "grad_norm": 2.875, + "learning_rate": 2.395101190899598e-06, + "loss": 0.7583, + "step": 1547 + }, + { + "epoch": 2.03, + "grad_norm": 2.796875, + "learning_rate": 2.39248031902773e-06, + "loss": 0.7502, + "step": 1548 + }, + { + "epoch": 2.03, + "grad_norm": 2.765625, + "learning_rate": 2.3898595655374284e-06, + "loss": 0.7454, + "step": 1549 + }, + { + "epoch": 2.03, + "grad_norm": 2.75, + "learning_rate": 2.387238933314201e-06, + "loss": 0.7403, + "step": 1550 + }, + { + "epoch": 2.03, + "grad_norm": 2.6875, + "learning_rate": 2.3846184252434214e-06, + "loss": 0.7238, + "step": 1551 + }, + { + "epoch": 2.04, + "grad_norm": 2.75, + "learning_rate": 2.3819980442103288e-06, + "loss": 0.7521, + "step": 1552 + }, + { + "epoch": 2.04, + "grad_norm": 2.75, + "learning_rate": 2.37937779310002e-06, + "loss": 0.7936, + "step": 1553 + }, + { + "epoch": 2.04, + "grad_norm": 2.796875, + "learning_rate": 2.3767576747974504e-06, + "loss": 0.7511, + "step": 1554 + }, + { + "epoch": 2.04, + "grad_norm": 2.796875, + "learning_rate": 2.3741376921874283e-06, + "loss": 0.7682, + "step": 1555 + }, + { + "epoch": 2.04, + "grad_norm": 2.859375, + "learning_rate": 2.371517848154613e-06, + "loss": 0.7722, + "step": 1556 + }, + { + "epoch": 2.04, + "grad_norm": 2.6875, + "learning_rate": 2.3688981455835124e-06, + "loss": 0.7853, + "step": 1557 + }, + { + "epoch": 2.04, + "grad_norm": 2.703125, + "learning_rate": 2.3662785873584775e-06, + "loss": 0.7421, + "step": 1558 + }, + { + "epoch": 2.04, + "grad_norm": 2.734375, + "learning_rate": 2.3636591763636974e-06, + "loss": 0.7611, + "step": 1559 + }, + { + "epoch": 2.05, + "grad_norm": 2.671875, + "learning_rate": 2.3610399154832047e-06, + "loss": 0.7649, + "step": 1560 + }, + { + "epoch": 2.05, + "grad_norm": 2.6875, + "learning_rate": 2.358420807600863e-06, + "loss": 0.7626, + "step": 1561 + }, + { + "epoch": 2.05, + "grad_norm": 2.71875, + "learning_rate": 2.355801855600369e-06, + "loss": 0.7748, + "step": 1562 + }, + { + "epoch": 2.05, + "grad_norm": 2.71875, + "learning_rate": 2.3531830623652473e-06, + "loss": 0.7254, + "step": 1563 + }, + { + "epoch": 2.05, + "grad_norm": 2.828125, + "learning_rate": 2.350564430778847e-06, + "loss": 0.7543, + "step": 1564 + }, + { + "epoch": 2.05, + "grad_norm": 2.671875, + "learning_rate": 2.34794596372434e-06, + "loss": 0.7762, + "step": 1565 + }, + { + "epoch": 2.05, + "grad_norm": 2.734375, + "learning_rate": 2.345327664084717e-06, + "loss": 0.7352, + "step": 1566 + }, + { + "epoch": 2.06, + "grad_norm": 2.75, + "learning_rate": 2.342709534742785e-06, + "loss": 0.7659, + "step": 1567 + }, + { + "epoch": 2.06, + "grad_norm": 2.8125, + "learning_rate": 2.3400915785811616e-06, + "loss": 0.7804, + "step": 1568 + }, + { + "epoch": 2.06, + "grad_norm": 2.6875, + "learning_rate": 2.337473798482276e-06, + "loss": 0.7535, + "step": 1569 + }, + { + "epoch": 2.06, + "grad_norm": 2.640625, + "learning_rate": 2.3348561973283613e-06, + "loss": 0.7304, + "step": 1570 + }, + { + "epoch": 2.06, + "grad_norm": 2.71875, + "learning_rate": 2.3322387780014558e-06, + "loss": 0.7828, + "step": 1571 + }, + { + "epoch": 2.06, + "grad_norm": 2.75, + "learning_rate": 2.329621543383396e-06, + "loss": 0.765, + "step": 1572 + }, + { + "epoch": 2.06, + "grad_norm": 2.6875, + "learning_rate": 2.3270044963558154e-06, + "loss": 0.7431, + "step": 1573 + }, + { + "epoch": 2.06, + "grad_norm": 2.71875, + "learning_rate": 2.3243876398001407e-06, + "loss": 0.7541, + "step": 1574 + }, + { + "epoch": 2.07, + "grad_norm": 2.90625, + "learning_rate": 2.3217709765975916e-06, + "loss": 0.7435, + "step": 1575 + }, + { + "epoch": 2.07, + "grad_norm": 2.8125, + "learning_rate": 2.31915450962917e-06, + "loss": 0.764, + "step": 1576 + }, + { + "epoch": 2.07, + "grad_norm": 2.78125, + "learning_rate": 2.3165382417756656e-06, + "loss": 0.7557, + "step": 1577 + }, + { + "epoch": 2.07, + "grad_norm": 2.96875, + "learning_rate": 2.3139221759176468e-06, + "loss": 0.7801, + "step": 1578 + }, + { + "epoch": 2.07, + "grad_norm": 2.828125, + "learning_rate": 2.3113063149354615e-06, + "loss": 0.7661, + "step": 1579 + }, + { + "epoch": 2.07, + "grad_norm": 2.828125, + "learning_rate": 2.3086906617092308e-06, + "loss": 0.7439, + "step": 1580 + }, + { + "epoch": 2.07, + "grad_norm": 2.71875, + "learning_rate": 2.3060752191188466e-06, + "loss": 0.7443, + "step": 1581 + }, + { + "epoch": 2.08, + "grad_norm": 2.71875, + "learning_rate": 2.3034599900439703e-06, + "loss": 0.7865, + "step": 1582 + }, + { + "epoch": 2.08, + "grad_norm": 2.765625, + "learning_rate": 2.3008449773640264e-06, + "loss": 0.7489, + "step": 1583 + }, + { + "epoch": 2.08, + "grad_norm": 2.71875, + "learning_rate": 2.298230183958203e-06, + "loss": 0.7463, + "step": 1584 + }, + { + "epoch": 2.08, + "grad_norm": 2.71875, + "learning_rate": 2.2956156127054454e-06, + "loss": 0.7719, + "step": 1585 + }, + { + "epoch": 2.08, + "grad_norm": 2.84375, + "learning_rate": 2.2930012664844555e-06, + "loss": 0.7655, + "step": 1586 + }, + { + "epoch": 2.08, + "grad_norm": 2.75, + "learning_rate": 2.2903871481736867e-06, + "loss": 0.7458, + "step": 1587 + }, + { + "epoch": 2.08, + "grad_norm": 2.734375, + "learning_rate": 2.2877732606513407e-06, + "loss": 0.7751, + "step": 1588 + }, + { + "epoch": 2.08, + "grad_norm": 2.609375, + "learning_rate": 2.2851596067953664e-06, + "loss": 0.756, + "step": 1589 + }, + { + "epoch": 2.09, + "grad_norm": 2.734375, + "learning_rate": 2.282546189483455e-06, + "loss": 0.7639, + "step": 1590 + }, + { + "epoch": 2.09, + "grad_norm": 2.75, + "learning_rate": 2.2799330115930374e-06, + "loss": 0.7352, + "step": 1591 + }, + { + "epoch": 2.09, + "grad_norm": 2.8125, + "learning_rate": 2.27732007600128e-06, + "loss": 0.7706, + "step": 1592 + }, + { + "epoch": 2.09, + "grad_norm": 2.828125, + "learning_rate": 2.2747073855850855e-06, + "loss": 0.7496, + "step": 1593 + }, + { + "epoch": 2.09, + "grad_norm": 2.75, + "learning_rate": 2.2720949432210813e-06, + "loss": 0.7577, + "step": 1594 + }, + { + "epoch": 2.09, + "grad_norm": 2.796875, + "learning_rate": 2.269482751785626e-06, + "loss": 0.7652, + "step": 1595 + }, + { + "epoch": 2.09, + "grad_norm": 2.671875, + "learning_rate": 2.2668708141547996e-06, + "loss": 0.7472, + "step": 1596 + }, + { + "epoch": 2.1, + "grad_norm": 2.828125, + "learning_rate": 2.2642591332044055e-06, + "loss": 0.7474, + "step": 1597 + }, + { + "epoch": 2.1, + "grad_norm": 2.75, + "learning_rate": 2.2616477118099617e-06, + "loss": 0.7609, + "step": 1598 + }, + { + "epoch": 2.1, + "grad_norm": 2.796875, + "learning_rate": 2.2590365528467025e-06, + "loss": 0.7512, + "step": 1599 + }, + { + "epoch": 2.1, + "grad_norm": 2.859375, + "learning_rate": 2.2564256591895695e-06, + "loss": 0.7365, + "step": 1600 + }, + { + "epoch": 2.1, + "grad_norm": 2.625, + "learning_rate": 2.253815033713217e-06, + "loss": 0.7538, + "step": 1601 + }, + { + "epoch": 2.1, + "grad_norm": 2.671875, + "learning_rate": 2.2512046792920013e-06, + "loss": 0.7421, + "step": 1602 + }, + { + "epoch": 2.1, + "grad_norm": 2.75, + "learning_rate": 2.2485945987999803e-06, + "loss": 0.7184, + "step": 1603 + }, + { + "epoch": 2.1, + "grad_norm": 2.78125, + "learning_rate": 2.2459847951109115e-06, + "loss": 0.7389, + "step": 1604 + }, + { + "epoch": 2.11, + "grad_norm": 2.75, + "learning_rate": 2.2433752710982475e-06, + "loss": 0.7541, + "step": 1605 + }, + { + "epoch": 2.11, + "grad_norm": 2.875, + "learning_rate": 2.2407660296351313e-06, + "loss": 0.7693, + "step": 1606 + }, + { + "epoch": 2.11, + "grad_norm": 2.75, + "learning_rate": 2.2381570735943957e-06, + "loss": 0.732, + "step": 1607 + }, + { + "epoch": 2.11, + "grad_norm": 2.75, + "learning_rate": 2.2355484058485603e-06, + "loss": 0.735, + "step": 1608 + }, + { + "epoch": 2.11, + "grad_norm": 2.75, + "learning_rate": 2.2329400292698262e-06, + "loss": 0.7695, + "step": 1609 + }, + { + "epoch": 2.11, + "grad_norm": 2.75, + "learning_rate": 2.2303319467300735e-06, + "loss": 0.7368, + "step": 1610 + }, + { + "epoch": 2.11, + "grad_norm": 2.796875, + "learning_rate": 2.2277241611008615e-06, + "loss": 0.7635, + "step": 1611 + }, + { + "epoch": 2.12, + "grad_norm": 2.765625, + "learning_rate": 2.225116675253418e-06, + "loss": 0.7715, + "step": 1612 + }, + { + "epoch": 2.12, + "grad_norm": 2.671875, + "learning_rate": 2.2225094920586435e-06, + "loss": 0.7499, + "step": 1613 + }, + { + "epoch": 2.12, + "grad_norm": 2.71875, + "learning_rate": 2.2199026143871043e-06, + "loss": 0.7539, + "step": 1614 + }, + { + "epoch": 2.12, + "grad_norm": 2.65625, + "learning_rate": 2.2172960451090325e-06, + "loss": 0.76, + "step": 1615 + }, + { + "epoch": 2.12, + "grad_norm": 2.75, + "learning_rate": 2.214689787094318e-06, + "loss": 0.7284, + "step": 1616 + }, + { + "epoch": 2.12, + "grad_norm": 2.84375, + "learning_rate": 2.2120838432125103e-06, + "loss": 0.7618, + "step": 1617 + }, + { + "epoch": 2.12, + "grad_norm": 2.75, + "learning_rate": 2.209478216332809e-06, + "loss": 0.7684, + "step": 1618 + }, + { + "epoch": 2.12, + "grad_norm": 2.71875, + "learning_rate": 2.206872909324069e-06, + "loss": 0.7566, + "step": 1619 + }, + { + "epoch": 2.13, + "grad_norm": 2.78125, + "learning_rate": 2.204267925054792e-06, + "loss": 0.7509, + "step": 1620 + }, + { + "epoch": 2.13, + "grad_norm": 2.6875, + "learning_rate": 2.2016632663931224e-06, + "loss": 0.7443, + "step": 1621 + }, + { + "epoch": 2.13, + "grad_norm": 2.703125, + "learning_rate": 2.199058936206849e-06, + "loss": 0.7503, + "step": 1622 + }, + { + "epoch": 2.13, + "grad_norm": 2.6875, + "learning_rate": 2.1964549373633957e-06, + "loss": 0.7323, + "step": 1623 + }, + { + "epoch": 2.13, + "grad_norm": 2.828125, + "learning_rate": 2.193851272729825e-06, + "loss": 0.7674, + "step": 1624 + }, + { + "epoch": 2.13, + "grad_norm": 2.765625, + "learning_rate": 2.191247945172828e-06, + "loss": 0.7694, + "step": 1625 + }, + { + "epoch": 2.13, + "grad_norm": 2.71875, + "learning_rate": 2.188644957558727e-06, + "loss": 0.7329, + "step": 1626 + }, + { + "epoch": 2.14, + "grad_norm": 2.75, + "learning_rate": 2.18604231275347e-06, + "loss": 0.7653, + "step": 1627 + }, + { + "epoch": 2.14, + "grad_norm": 2.796875, + "learning_rate": 2.1834400136226276e-06, + "loss": 0.7648, + "step": 1628 + }, + { + "epoch": 2.14, + "grad_norm": 2.78125, + "learning_rate": 2.180838063031387e-06, + "loss": 0.759, + "step": 1629 + }, + { + "epoch": 2.14, + "grad_norm": 2.8125, + "learning_rate": 2.1782364638445545e-06, + "loss": 0.7648, + "step": 1630 + }, + { + "epoch": 2.14, + "grad_norm": 2.734375, + "learning_rate": 2.17563521892655e-06, + "loss": 0.7652, + "step": 1631 + }, + { + "epoch": 2.14, + "grad_norm": 2.71875, + "learning_rate": 2.173034331141401e-06, + "loss": 0.7612, + "step": 1632 + }, + { + "epoch": 2.14, + "grad_norm": 2.703125, + "learning_rate": 2.1704338033527434e-06, + "loss": 0.7619, + "step": 1633 + }, + { + "epoch": 2.14, + "grad_norm": 2.734375, + "learning_rate": 2.1678336384238172e-06, + "loss": 0.7516, + "step": 1634 + }, + { + "epoch": 2.15, + "grad_norm": 2.84375, + "learning_rate": 2.1652338392174605e-06, + "loss": 0.7708, + "step": 1635 + }, + { + "epoch": 2.15, + "grad_norm": 2.8125, + "learning_rate": 2.16263440859611e-06, + "loss": 0.7625, + "step": 1636 + }, + { + "epoch": 2.15, + "grad_norm": 2.734375, + "learning_rate": 2.1600353494217973e-06, + "loss": 0.7605, + "step": 1637 + }, + { + "epoch": 2.15, + "grad_norm": 2.828125, + "learning_rate": 2.157436664556145e-06, + "loss": 0.7799, + "step": 1638 + }, + { + "epoch": 2.15, + "grad_norm": 2.65625, + "learning_rate": 2.154838356860363e-06, + "loss": 0.7291, + "step": 1639 + }, + { + "epoch": 2.15, + "grad_norm": 2.703125, + "learning_rate": 2.152240429195246e-06, + "loss": 0.7575, + "step": 1640 + }, + { + "epoch": 2.15, + "grad_norm": 2.859375, + "learning_rate": 2.1496428844211696e-06, + "loss": 0.7495, + "step": 1641 + }, + { + "epoch": 2.16, + "grad_norm": 2.78125, + "learning_rate": 2.1470457253980887e-06, + "loss": 0.751, + "step": 1642 + }, + { + "epoch": 2.16, + "grad_norm": 2.71875, + "learning_rate": 2.1444489549855335e-06, + "loss": 0.7237, + "step": 1643 + }, + { + "epoch": 2.16, + "grad_norm": 2.734375, + "learning_rate": 2.1418525760426063e-06, + "loss": 0.755, + "step": 1644 + }, + { + "epoch": 2.16, + "grad_norm": 2.6875, + "learning_rate": 2.139256591427978e-06, + "loss": 0.7471, + "step": 1645 + }, + { + "epoch": 2.16, + "grad_norm": 2.84375, + "learning_rate": 2.1366610039998863e-06, + "loss": 0.7686, + "step": 1646 + }, + { + "epoch": 2.16, + "grad_norm": 2.75, + "learning_rate": 2.1340658166161293e-06, + "loss": 0.7633, + "step": 1647 + }, + { + "epoch": 2.16, + "grad_norm": 2.875, + "learning_rate": 2.131471032134067e-06, + "loss": 0.7891, + "step": 1648 + }, + { + "epoch": 2.16, + "grad_norm": 2.71875, + "learning_rate": 2.128876653410615e-06, + "loss": 0.7714, + "step": 1649 + }, + { + "epoch": 2.17, + "grad_norm": 2.734375, + "learning_rate": 2.126282683302242e-06, + "loss": 0.7766, + "step": 1650 + }, + { + "epoch": 2.17, + "grad_norm": 2.6875, + "learning_rate": 2.123689124664966e-06, + "loss": 0.7433, + "step": 1651 + }, + { + "epoch": 2.17, + "grad_norm": 2.71875, + "learning_rate": 2.1210959803543564e-06, + "loss": 0.7322, + "step": 1652 + }, + { + "epoch": 2.17, + "grad_norm": 2.796875, + "learning_rate": 2.1185032532255187e-06, + "loss": 0.7632, + "step": 1653 + }, + { + "epoch": 2.17, + "grad_norm": 2.796875, + "learning_rate": 2.115910946133105e-06, + "loss": 0.7882, + "step": 1654 + }, + { + "epoch": 2.17, + "grad_norm": 2.8125, + "learning_rate": 2.1133190619313027e-06, + "loss": 0.7501, + "step": 1655 + }, + { + "epoch": 2.17, + "grad_norm": 2.671875, + "learning_rate": 2.110727603473835e-06, + "loss": 0.7349, + "step": 1656 + }, + { + "epoch": 2.17, + "grad_norm": 2.6875, + "learning_rate": 2.108136573613955e-06, + "loss": 0.7906, + "step": 1657 + }, + { + "epoch": 2.18, + "grad_norm": 2.796875, + "learning_rate": 2.1055459752044454e-06, + "loss": 0.7523, + "step": 1658 + }, + { + "epoch": 2.18, + "grad_norm": 2.796875, + "learning_rate": 2.1029558110976103e-06, + "loss": 0.7449, + "step": 1659 + }, + { + "epoch": 2.18, + "grad_norm": 2.765625, + "learning_rate": 2.10036608414528e-06, + "loss": 0.7735, + "step": 1660 + }, + { + "epoch": 2.18, + "grad_norm": 2.71875, + "learning_rate": 2.0977767971988012e-06, + "loss": 0.7306, + "step": 1661 + }, + { + "epoch": 2.18, + "grad_norm": 2.796875, + "learning_rate": 2.0951879531090365e-06, + "loss": 0.7488, + "step": 1662 + }, + { + "epoch": 2.18, + "grad_norm": 2.84375, + "learning_rate": 2.0925995547263604e-06, + "loss": 0.7563, + "step": 1663 + }, + { + "epoch": 2.18, + "grad_norm": 2.765625, + "learning_rate": 2.090011604900659e-06, + "loss": 0.764, + "step": 1664 + }, + { + "epoch": 2.19, + "grad_norm": 2.765625, + "learning_rate": 2.087424106481321e-06, + "loss": 0.7741, + "step": 1665 + }, + { + "epoch": 2.19, + "grad_norm": 2.765625, + "learning_rate": 2.084837062317241e-06, + "loss": 0.7565, + "step": 1666 + }, + { + "epoch": 2.19, + "grad_norm": 2.703125, + "learning_rate": 2.0822504752568106e-06, + "loss": 0.7521, + "step": 1667 + }, + { + "epoch": 2.19, + "grad_norm": 2.703125, + "learning_rate": 2.0796643481479216e-06, + "loss": 0.7521, + "step": 1668 + }, + { + "epoch": 2.19, + "grad_norm": 2.859375, + "learning_rate": 2.0770786838379564e-06, + "loss": 0.7979, + "step": 1669 + }, + { + "epoch": 2.19, + "grad_norm": 2.828125, + "learning_rate": 2.074493485173791e-06, + "loss": 0.7764, + "step": 1670 + }, + { + "epoch": 2.19, + "grad_norm": 2.84375, + "learning_rate": 2.071908755001785e-06, + "loss": 0.7833, + "step": 1671 + }, + { + "epoch": 2.19, + "grad_norm": 2.8125, + "learning_rate": 2.0693244961677834e-06, + "loss": 0.8034, + "step": 1672 + }, + { + "epoch": 2.2, + "grad_norm": 2.703125, + "learning_rate": 2.0667407115171136e-06, + "loss": 0.7293, + "step": 1673 + }, + { + "epoch": 2.2, + "grad_norm": 2.828125, + "learning_rate": 2.064157403894581e-06, + "loss": 0.7494, + "step": 1674 + }, + { + "epoch": 2.2, + "grad_norm": 2.75, + "learning_rate": 2.061574576144465e-06, + "loss": 0.7451, + "step": 1675 + }, + { + "epoch": 2.2, + "grad_norm": 2.78125, + "learning_rate": 2.0589922311105145e-06, + "loss": 0.7802, + "step": 1676 + }, + { + "epoch": 2.2, + "grad_norm": 2.734375, + "learning_rate": 2.0564103716359497e-06, + "loss": 0.7701, + "step": 1677 + }, + { + "epoch": 2.2, + "grad_norm": 2.703125, + "learning_rate": 2.0538290005634564e-06, + "loss": 0.7564, + "step": 1678 + }, + { + "epoch": 2.2, + "grad_norm": 2.65625, + "learning_rate": 2.051248120735181e-06, + "loss": 0.7423, + "step": 1679 + }, + { + "epoch": 2.21, + "grad_norm": 2.8125, + "learning_rate": 2.0486677349927303e-06, + "loss": 0.7648, + "step": 1680 + }, + { + "epoch": 2.21, + "grad_norm": 2.734375, + "learning_rate": 2.0460878461771666e-06, + "loss": 0.7377, + "step": 1681 + }, + { + "epoch": 2.21, + "grad_norm": 2.65625, + "learning_rate": 2.0435084571290036e-06, + "loss": 0.7573, + "step": 1682 + }, + { + "epoch": 2.21, + "grad_norm": 2.71875, + "learning_rate": 2.0409295706882076e-06, + "loss": 0.7436, + "step": 1683 + }, + { + "epoch": 2.21, + "grad_norm": 2.75, + "learning_rate": 2.0383511896941894e-06, + "loss": 0.7759, + "step": 1684 + }, + { + "epoch": 2.21, + "grad_norm": 2.734375, + "learning_rate": 2.035773316985804e-06, + "loss": 0.7332, + "step": 1685 + }, + { + "epoch": 2.21, + "grad_norm": 2.703125, + "learning_rate": 2.0331959554013468e-06, + "loss": 0.7453, + "step": 1686 + }, + { + "epoch": 2.21, + "grad_norm": 2.765625, + "learning_rate": 2.0306191077785512e-06, + "loss": 0.7703, + "step": 1687 + }, + { + "epoch": 2.22, + "grad_norm": 2.71875, + "learning_rate": 2.0280427769545825e-06, + "loss": 0.7806, + "step": 1688 + }, + { + "epoch": 2.22, + "grad_norm": 2.765625, + "learning_rate": 2.025466965766039e-06, + "loss": 0.7538, + "step": 1689 + }, + { + "epoch": 2.22, + "grad_norm": 2.65625, + "learning_rate": 2.0228916770489466e-06, + "loss": 0.7476, + "step": 1690 + }, + { + "epoch": 2.22, + "grad_norm": 2.796875, + "learning_rate": 2.0203169136387545e-06, + "loss": 0.7489, + "step": 1691 + }, + { + "epoch": 2.22, + "grad_norm": 2.765625, + "learning_rate": 2.0177426783703353e-06, + "loss": 0.7628, + "step": 1692 + }, + { + "epoch": 2.22, + "grad_norm": 2.796875, + "learning_rate": 2.0151689740779813e-06, + "loss": 0.7436, + "step": 1693 + }, + { + "epoch": 2.22, + "grad_norm": 2.765625, + "learning_rate": 2.012595803595395e-06, + "loss": 0.7739, + "step": 1694 + }, + { + "epoch": 2.23, + "grad_norm": 2.703125, + "learning_rate": 2.010023169755696e-06, + "loss": 0.7613, + "step": 1695 + }, + { + "epoch": 2.23, + "grad_norm": 2.6875, + "learning_rate": 2.0074510753914113e-06, + "loss": 0.7553, + "step": 1696 + }, + { + "epoch": 2.23, + "grad_norm": 2.75, + "learning_rate": 2.0048795233344743e-06, + "loss": 0.7456, + "step": 1697 + }, + { + "epoch": 2.23, + "grad_norm": 2.765625, + "learning_rate": 2.0023085164162217e-06, + "loss": 0.773, + "step": 1698 + }, + { + "epoch": 2.23, + "grad_norm": 2.734375, + "learning_rate": 1.999738057467389e-06, + "loss": 0.7611, + "step": 1699 + }, + { + "epoch": 2.23, + "grad_norm": 2.71875, + "learning_rate": 1.9971681493181084e-06, + "loss": 0.7545, + "step": 1700 + }, + { + "epoch": 2.23, + "grad_norm": 2.703125, + "learning_rate": 1.9945987947979063e-06, + "loss": 0.7676, + "step": 1701 + }, + { + "epoch": 2.23, + "grad_norm": 2.671875, + "learning_rate": 1.9920299967356995e-06, + "loss": 0.7356, + "step": 1702 + }, + { + "epoch": 2.24, + "grad_norm": 2.75, + "learning_rate": 1.989461757959792e-06, + "loss": 0.7551, + "step": 1703 + }, + { + "epoch": 2.24, + "grad_norm": 2.75, + "learning_rate": 1.9868940812978714e-06, + "loss": 0.7792, + "step": 1704 + }, + { + "epoch": 2.24, + "grad_norm": 2.78125, + "learning_rate": 1.9843269695770083e-06, + "loss": 0.7379, + "step": 1705 + }, + { + "epoch": 2.24, + "grad_norm": 2.734375, + "learning_rate": 1.9817604256236485e-06, + "loss": 0.7595, + "step": 1706 + }, + { + "epoch": 2.24, + "grad_norm": 2.765625, + "learning_rate": 1.9791944522636143e-06, + "loss": 0.7529, + "step": 1707 + }, + { + "epoch": 2.24, + "grad_norm": 2.765625, + "learning_rate": 1.9766290523220997e-06, + "loss": 0.7391, + "step": 1708 + }, + { + "epoch": 2.24, + "grad_norm": 2.8125, + "learning_rate": 1.974064228623668e-06, + "loss": 0.7397, + "step": 1709 + }, + { + "epoch": 2.25, + "grad_norm": 2.78125, + "learning_rate": 1.9714999839922455e-06, + "loss": 0.7609, + "step": 1710 + }, + { + "epoch": 2.25, + "grad_norm": 2.890625, + "learning_rate": 1.968936321251126e-06, + "loss": 0.7562, + "step": 1711 + }, + { + "epoch": 2.25, + "grad_norm": 2.78125, + "learning_rate": 1.966373243222956e-06, + "loss": 0.7674, + "step": 1712 + }, + { + "epoch": 2.25, + "grad_norm": 2.640625, + "learning_rate": 1.963810752729743e-06, + "loss": 0.7548, + "step": 1713 + }, + { + "epoch": 2.25, + "grad_norm": 2.71875, + "learning_rate": 1.9612488525928453e-06, + "loss": 0.7846, + "step": 1714 + }, + { + "epoch": 2.25, + "grad_norm": 2.84375, + "learning_rate": 1.958687545632973e-06, + "loss": 0.754, + "step": 1715 + }, + { + "epoch": 2.25, + "grad_norm": 2.71875, + "learning_rate": 1.956126834670182e-06, + "loss": 0.7446, + "step": 1716 + }, + { + "epoch": 2.25, + "grad_norm": 2.84375, + "learning_rate": 1.9535667225238734e-06, + "loss": 0.7855, + "step": 1717 + }, + { + "epoch": 2.26, + "grad_norm": 2.75, + "learning_rate": 1.951007212012785e-06, + "loss": 0.783, + "step": 1718 + }, + { + "epoch": 2.26, + "grad_norm": 2.765625, + "learning_rate": 1.9484483059549962e-06, + "loss": 0.7805, + "step": 1719 + }, + { + "epoch": 2.26, + "grad_norm": 2.734375, + "learning_rate": 1.94589000716792e-06, + "loss": 0.7468, + "step": 1720 + }, + { + "epoch": 2.26, + "grad_norm": 2.71875, + "learning_rate": 1.9433323184683006e-06, + "loss": 0.7926, + "step": 1721 + }, + { + "epoch": 2.26, + "grad_norm": 2.71875, + "learning_rate": 1.9407752426722095e-06, + "loss": 0.7673, + "step": 1722 + }, + { + "epoch": 2.26, + "grad_norm": 2.796875, + "learning_rate": 1.9382187825950454e-06, + "loss": 0.7785, + "step": 1723 + }, + { + "epoch": 2.26, + "grad_norm": 2.625, + "learning_rate": 1.935662941051526e-06, + "loss": 0.7351, + "step": 1724 + }, + { + "epoch": 2.27, + "grad_norm": 2.765625, + "learning_rate": 1.933107720855691e-06, + "loss": 0.7488, + "step": 1725 + }, + { + "epoch": 2.27, + "grad_norm": 2.71875, + "learning_rate": 1.9305531248208948e-06, + "loss": 0.7559, + "step": 1726 + }, + { + "epoch": 2.27, + "grad_norm": 2.8125, + "learning_rate": 1.9279991557598042e-06, + "loss": 0.7567, + "step": 1727 + }, + { + "epoch": 2.27, + "grad_norm": 2.71875, + "learning_rate": 1.9254458164843957e-06, + "loss": 0.747, + "step": 1728 + }, + { + "epoch": 2.27, + "grad_norm": 2.8125, + "learning_rate": 1.922893109805955e-06, + "loss": 0.7483, + "step": 1729 + }, + { + "epoch": 2.27, + "grad_norm": 2.703125, + "learning_rate": 1.920341038535066e-06, + "loss": 0.763, + "step": 1730 + }, + { + "epoch": 2.27, + "grad_norm": 2.78125, + "learning_rate": 1.9177896054816182e-06, + "loss": 0.732, + "step": 1731 + }, + { + "epoch": 2.27, + "grad_norm": 2.71875, + "learning_rate": 1.9152388134547944e-06, + "loss": 0.7591, + "step": 1732 + }, + { + "epoch": 2.28, + "grad_norm": 2.78125, + "learning_rate": 1.9126886652630755e-06, + "loss": 0.7606, + "step": 1733 + }, + { + "epoch": 2.28, + "grad_norm": 2.6875, + "learning_rate": 1.9101391637142317e-06, + "loss": 0.7557, + "step": 1734 + }, + { + "epoch": 2.28, + "grad_norm": 2.78125, + "learning_rate": 1.907590311615319e-06, + "loss": 0.7671, + "step": 1735 + }, + { + "epoch": 2.28, + "grad_norm": 2.703125, + "learning_rate": 1.9050421117726806e-06, + "loss": 0.7616, + "step": 1736 + }, + { + "epoch": 2.28, + "grad_norm": 2.703125, + "learning_rate": 1.902494566991943e-06, + "loss": 0.7643, + "step": 1737 + }, + { + "epoch": 2.28, + "grad_norm": 2.734375, + "learning_rate": 1.8999476800780087e-06, + "loss": 0.7692, + "step": 1738 + }, + { + "epoch": 2.28, + "grad_norm": 2.890625, + "learning_rate": 1.897401453835057e-06, + "loss": 0.759, + "step": 1739 + }, + { + "epoch": 2.29, + "grad_norm": 2.828125, + "learning_rate": 1.894855891066541e-06, + "loss": 0.7704, + "step": 1740 + }, + { + "epoch": 2.29, + "grad_norm": 2.765625, + "learning_rate": 1.8923109945751806e-06, + "loss": 0.7645, + "step": 1741 + }, + { + "epoch": 2.29, + "grad_norm": 2.75, + "learning_rate": 1.8897667671629638e-06, + "loss": 0.7635, + "step": 1742 + }, + { + "epoch": 2.29, + "grad_norm": 2.71875, + "learning_rate": 1.8872232116311418e-06, + "loss": 0.7677, + "step": 1743 + }, + { + "epoch": 2.29, + "grad_norm": 2.765625, + "learning_rate": 1.8846803307802263e-06, + "loss": 0.7537, + "step": 1744 + }, + { + "epoch": 2.29, + "grad_norm": 2.890625, + "learning_rate": 1.8821381274099854e-06, + "loss": 0.7639, + "step": 1745 + }, + { + "epoch": 2.29, + "grad_norm": 2.734375, + "learning_rate": 1.8795966043194425e-06, + "loss": 0.7486, + "step": 1746 + }, + { + "epoch": 2.29, + "grad_norm": 2.875, + "learning_rate": 1.8770557643068701e-06, + "loss": 0.7792, + "step": 1747 + }, + { + "epoch": 2.3, + "grad_norm": 2.75, + "learning_rate": 1.8745156101697904e-06, + "loss": 0.7651, + "step": 1748 + }, + { + "epoch": 2.3, + "grad_norm": 2.84375, + "learning_rate": 1.8719761447049691e-06, + "loss": 0.7823, + "step": 1749 + }, + { + "epoch": 2.3, + "grad_norm": 2.875, + "learning_rate": 1.8694373707084148e-06, + "loss": 0.7592, + "step": 1750 + }, + { + "epoch": 2.3, + "grad_norm": 2.75, + "learning_rate": 1.866899290975374e-06, + "loss": 0.7799, + "step": 1751 + }, + { + "epoch": 2.3, + "grad_norm": 2.703125, + "learning_rate": 1.8643619083003306e-06, + "loss": 0.727, + "step": 1752 + }, + { + "epoch": 2.3, + "grad_norm": 2.765625, + "learning_rate": 1.861825225476997e-06, + "loss": 0.7345, + "step": 1753 + }, + { + "epoch": 2.3, + "grad_norm": 2.703125, + "learning_rate": 1.8592892452983188e-06, + "loss": 0.751, + "step": 1754 + }, + { + "epoch": 2.31, + "grad_norm": 2.78125, + "learning_rate": 1.8567539705564664e-06, + "loss": 0.7902, + "step": 1755 + }, + { + "epoch": 2.31, + "grad_norm": 2.703125, + "learning_rate": 1.854219404042834e-06, + "loss": 0.7477, + "step": 1756 + }, + { + "epoch": 2.31, + "grad_norm": 2.78125, + "learning_rate": 1.8516855485480361e-06, + "loss": 0.754, + "step": 1757 + }, + { + "epoch": 2.31, + "grad_norm": 2.75, + "learning_rate": 1.8491524068619049e-06, + "loss": 0.7617, + "step": 1758 + }, + { + "epoch": 2.31, + "grad_norm": 2.796875, + "learning_rate": 1.8466199817734836e-06, + "loss": 0.7522, + "step": 1759 + }, + { + "epoch": 2.31, + "grad_norm": 2.90625, + "learning_rate": 1.8440882760710304e-06, + "loss": 0.7819, + "step": 1760 + }, + { + "epoch": 2.31, + "grad_norm": 2.734375, + "learning_rate": 1.841557292542009e-06, + "loss": 0.7423, + "step": 1761 + }, + { + "epoch": 2.31, + "grad_norm": 2.6875, + "learning_rate": 1.8390270339730892e-06, + "loss": 0.748, + "step": 1762 + }, + { + "epoch": 2.32, + "grad_norm": 2.75, + "learning_rate": 1.8364975031501417e-06, + "loss": 0.757, + "step": 1763 + }, + { + "epoch": 2.32, + "grad_norm": 2.734375, + "learning_rate": 1.833968702858237e-06, + "loss": 0.7684, + "step": 1764 + }, + { + "epoch": 2.32, + "grad_norm": 2.671875, + "learning_rate": 1.8314406358816398e-06, + "loss": 0.7664, + "step": 1765 + }, + { + "epoch": 2.32, + "grad_norm": 2.640625, + "learning_rate": 1.8289133050038083e-06, + "loss": 0.7277, + "step": 1766 + }, + { + "epoch": 2.32, + "grad_norm": 2.765625, + "learning_rate": 1.8263867130073903e-06, + "loss": 0.7548, + "step": 1767 + }, + { + "epoch": 2.32, + "grad_norm": 2.78125, + "learning_rate": 1.82386086267422e-06, + "loss": 0.7769, + "step": 1768 + }, + { + "epoch": 2.32, + "grad_norm": 2.75, + "learning_rate": 1.8213357567853142e-06, + "loss": 0.7829, + "step": 1769 + }, + { + "epoch": 2.33, + "grad_norm": 2.734375, + "learning_rate": 1.8188113981208732e-06, + "loss": 0.7579, + "step": 1770 + }, + { + "epoch": 2.33, + "grad_norm": 2.65625, + "learning_rate": 1.8162877894602698e-06, + "loss": 0.7662, + "step": 1771 + }, + { + "epoch": 2.33, + "grad_norm": 2.828125, + "learning_rate": 1.8137649335820534e-06, + "loss": 0.7526, + "step": 1772 + }, + { + "epoch": 2.33, + "grad_norm": 2.75, + "learning_rate": 1.8112428332639454e-06, + "loss": 0.78, + "step": 1773 + }, + { + "epoch": 2.33, + "grad_norm": 2.734375, + "learning_rate": 1.8087214912828343e-06, + "loss": 0.7517, + "step": 1774 + }, + { + "epoch": 2.33, + "grad_norm": 2.734375, + "learning_rate": 1.8062009104147737e-06, + "loss": 0.7465, + "step": 1775 + }, + { + "epoch": 2.33, + "grad_norm": 2.71875, + "learning_rate": 1.803681093434981e-06, + "loss": 0.7402, + "step": 1776 + }, + { + "epoch": 2.33, + "grad_norm": 2.703125, + "learning_rate": 1.8011620431178276e-06, + "loss": 0.7535, + "step": 1777 + }, + { + "epoch": 2.34, + "grad_norm": 2.703125, + "learning_rate": 1.7986437622368462e-06, + "loss": 0.761, + "step": 1778 + }, + { + "epoch": 2.34, + "grad_norm": 2.6875, + "learning_rate": 1.7961262535647201e-06, + "loss": 0.7536, + "step": 1779 + }, + { + "epoch": 2.34, + "grad_norm": 2.765625, + "learning_rate": 1.793609519873282e-06, + "loss": 0.7551, + "step": 1780 + }, + { + "epoch": 2.34, + "grad_norm": 2.828125, + "learning_rate": 1.791093563933512e-06, + "loss": 0.7624, + "step": 1781 + }, + { + "epoch": 2.34, + "grad_norm": 2.828125, + "learning_rate": 1.7885783885155342e-06, + "loss": 0.759, + "step": 1782 + }, + { + "epoch": 2.34, + "grad_norm": 2.734375, + "learning_rate": 1.7860639963886117e-06, + "loss": 0.7781, + "step": 1783 + }, + { + "epoch": 2.34, + "grad_norm": 2.71875, + "learning_rate": 1.783550390321147e-06, + "loss": 0.7386, + "step": 1784 + }, + { + "epoch": 2.35, + "grad_norm": 2.71875, + "learning_rate": 1.781037573080676e-06, + "loss": 0.7479, + "step": 1785 + }, + { + "epoch": 2.35, + "grad_norm": 2.875, + "learning_rate": 1.7785255474338675e-06, + "loss": 0.7596, + "step": 1786 + }, + { + "epoch": 2.35, + "grad_norm": 2.75, + "learning_rate": 1.776014316146517e-06, + "loss": 0.7572, + "step": 1787 + }, + { + "epoch": 2.35, + "grad_norm": 2.75, + "learning_rate": 1.773503881983546e-06, + "loss": 0.7483, + "step": 1788 + }, + { + "epoch": 2.35, + "grad_norm": 2.703125, + "learning_rate": 1.7709942477089986e-06, + "loss": 0.7398, + "step": 1789 + }, + { + "epoch": 2.35, + "grad_norm": 2.890625, + "learning_rate": 1.7684854160860381e-06, + "loss": 0.7468, + "step": 1790 + }, + { + "epoch": 2.35, + "grad_norm": 2.703125, + "learning_rate": 1.7659773898769433e-06, + "loss": 0.7465, + "step": 1791 + }, + { + "epoch": 2.35, + "grad_norm": 2.765625, + "learning_rate": 1.763470171843109e-06, + "loss": 0.77, + "step": 1792 + }, + { + "epoch": 2.36, + "grad_norm": 2.890625, + "learning_rate": 1.7609637647450375e-06, + "loss": 0.7866, + "step": 1793 + }, + { + "epoch": 2.36, + "grad_norm": 2.703125, + "learning_rate": 1.7584581713423374e-06, + "loss": 0.7589, + "step": 1794 + }, + { + "epoch": 2.36, + "grad_norm": 2.71875, + "learning_rate": 1.7559533943937235e-06, + "loss": 0.753, + "step": 1795 + }, + { + "epoch": 2.36, + "grad_norm": 2.828125, + "learning_rate": 1.7534494366570115e-06, + "loss": 0.7453, + "step": 1796 + }, + { + "epoch": 2.36, + "grad_norm": 2.78125, + "learning_rate": 1.7509463008891153e-06, + "loss": 0.7718, + "step": 1797 + }, + { + "epoch": 2.36, + "grad_norm": 2.703125, + "learning_rate": 1.748443989846042e-06, + "loss": 0.7504, + "step": 1798 + }, + { + "epoch": 2.36, + "grad_norm": 3.203125, + "learning_rate": 1.745942506282893e-06, + "loss": 0.7782, + "step": 1799 + }, + { + "epoch": 2.37, + "grad_norm": 2.765625, + "learning_rate": 1.7434418529538564e-06, + "loss": 0.741, + "step": 1800 + }, + { + "epoch": 2.37, + "grad_norm": 2.8125, + "learning_rate": 1.7409420326122076e-06, + "loss": 0.7288, + "step": 1801 + }, + { + "epoch": 2.37, + "grad_norm": 2.734375, + "learning_rate": 1.7384430480103046e-06, + "loss": 0.7302, + "step": 1802 + }, + { + "epoch": 2.37, + "grad_norm": 2.734375, + "learning_rate": 1.7359449018995849e-06, + "loss": 0.7618, + "step": 1803 + }, + { + "epoch": 2.37, + "grad_norm": 2.671875, + "learning_rate": 1.7334475970305639e-06, + "loss": 0.7374, + "step": 1804 + }, + { + "epoch": 2.37, + "grad_norm": 2.8125, + "learning_rate": 1.7309511361528297e-06, + "loss": 0.7677, + "step": 1805 + }, + { + "epoch": 2.37, + "grad_norm": 2.703125, + "learning_rate": 1.7284555220150406e-06, + "loss": 0.7553, + "step": 1806 + }, + { + "epoch": 2.37, + "grad_norm": 2.734375, + "learning_rate": 1.725960757364924e-06, + "loss": 0.7615, + "step": 1807 + }, + { + "epoch": 2.38, + "grad_norm": 2.671875, + "learning_rate": 1.7234668449492712e-06, + "loss": 0.7622, + "step": 1808 + }, + { + "epoch": 2.38, + "grad_norm": 2.6875, + "learning_rate": 1.7209737875139353e-06, + "loss": 0.7432, + "step": 1809 + }, + { + "epoch": 2.38, + "grad_norm": 2.734375, + "learning_rate": 1.718481587803828e-06, + "loss": 0.7563, + "step": 1810 + }, + { + "epoch": 2.38, + "grad_norm": 2.828125, + "learning_rate": 1.7159902485629186e-06, + "loss": 0.7829, + "step": 1811 + }, + { + "epoch": 2.38, + "grad_norm": 2.796875, + "learning_rate": 1.7134997725342243e-06, + "loss": 0.7713, + "step": 1812 + }, + { + "epoch": 2.38, + "grad_norm": 2.796875, + "learning_rate": 1.711010162459816e-06, + "loss": 0.7514, + "step": 1813 + }, + { + "epoch": 2.38, + "grad_norm": 2.640625, + "learning_rate": 1.7085214210808096e-06, + "loss": 0.7487, + "step": 1814 + }, + { + "epoch": 2.39, + "grad_norm": 2.78125, + "learning_rate": 1.7060335511373658e-06, + "loss": 0.7552, + "step": 1815 + }, + { + "epoch": 2.39, + "grad_norm": 2.6875, + "learning_rate": 1.7035465553686839e-06, + "loss": 0.7404, + "step": 1816 + }, + { + "epoch": 2.39, + "grad_norm": 2.859375, + "learning_rate": 1.701060436513003e-06, + "loss": 0.7738, + "step": 1817 + }, + { + "epoch": 2.39, + "grad_norm": 2.6875, + "learning_rate": 1.6985751973075935e-06, + "loss": 0.7224, + "step": 1818 + }, + { + "epoch": 2.39, + "grad_norm": 2.71875, + "learning_rate": 1.6960908404887602e-06, + "loss": 0.7771, + "step": 1819 + }, + { + "epoch": 2.39, + "grad_norm": 2.84375, + "learning_rate": 1.6936073687918352e-06, + "loss": 0.767, + "step": 1820 + }, + { + "epoch": 2.39, + "grad_norm": 2.6875, + "learning_rate": 1.6911247849511764e-06, + "loss": 0.7398, + "step": 1821 + }, + { + "epoch": 2.39, + "grad_norm": 2.78125, + "learning_rate": 1.688643091700164e-06, + "loss": 0.7455, + "step": 1822 + }, + { + "epoch": 2.4, + "grad_norm": 2.765625, + "learning_rate": 1.6861622917711972e-06, + "loss": 0.7423, + "step": 1823 + }, + { + "epoch": 2.4, + "grad_norm": 2.859375, + "learning_rate": 1.6836823878956927e-06, + "loss": 0.7429, + "step": 1824 + }, + { + "epoch": 2.4, + "grad_norm": 2.78125, + "learning_rate": 1.681203382804079e-06, + "loss": 0.7719, + "step": 1825 + }, + { + "epoch": 2.4, + "grad_norm": 2.78125, + "learning_rate": 1.6787252792257964e-06, + "loss": 0.7488, + "step": 1826 + }, + { + "epoch": 2.4, + "grad_norm": 2.859375, + "learning_rate": 1.6762480798892922e-06, + "loss": 0.7653, + "step": 1827 + }, + { + "epoch": 2.4, + "grad_norm": 2.875, + "learning_rate": 1.6737717875220177e-06, + "loss": 0.7687, + "step": 1828 + }, + { + "epoch": 2.4, + "grad_norm": 2.75, + "learning_rate": 1.6712964048504286e-06, + "loss": 0.7601, + "step": 1829 + }, + { + "epoch": 2.41, + "grad_norm": 2.84375, + "learning_rate": 1.6688219345999727e-06, + "loss": 0.7937, + "step": 1830 + }, + { + "epoch": 2.41, + "grad_norm": 2.75, + "learning_rate": 1.6663483794950985e-06, + "loss": 0.7454, + "step": 1831 + }, + { + "epoch": 2.41, + "grad_norm": 2.75, + "learning_rate": 1.663875742259245e-06, + "loss": 0.7588, + "step": 1832 + }, + { + "epoch": 2.41, + "grad_norm": 2.765625, + "learning_rate": 1.6614040256148417e-06, + "loss": 0.7859, + "step": 1833 + }, + { + "epoch": 2.41, + "grad_norm": 2.6875, + "learning_rate": 1.6589332322833035e-06, + "loss": 0.7584, + "step": 1834 + }, + { + "epoch": 2.41, + "grad_norm": 2.828125, + "learning_rate": 1.6564633649850295e-06, + "loss": 0.7866, + "step": 1835 + }, + { + "epoch": 2.41, + "grad_norm": 2.765625, + "learning_rate": 1.6539944264393965e-06, + "loss": 0.7737, + "step": 1836 + }, + { + "epoch": 2.41, + "grad_norm": 2.71875, + "learning_rate": 1.6515264193647627e-06, + "loss": 0.7288, + "step": 1837 + }, + { + "epoch": 2.42, + "grad_norm": 2.734375, + "learning_rate": 1.6490593464784586e-06, + "loss": 0.7597, + "step": 1838 + }, + { + "epoch": 2.42, + "grad_norm": 2.765625, + "learning_rate": 1.6465932104967863e-06, + "loss": 0.7353, + "step": 1839 + }, + { + "epoch": 2.42, + "grad_norm": 2.71875, + "learning_rate": 1.6441280141350169e-06, + "loss": 0.7568, + "step": 1840 + }, + { + "epoch": 2.42, + "grad_norm": 2.71875, + "learning_rate": 1.641663760107386e-06, + "loss": 0.7427, + "step": 1841 + }, + { + "epoch": 2.42, + "grad_norm": 2.71875, + "learning_rate": 1.639200451127092e-06, + "loss": 0.758, + "step": 1842 + }, + { + "epoch": 2.42, + "grad_norm": 2.75, + "learning_rate": 1.636738089906294e-06, + "loss": 0.7799, + "step": 1843 + }, + { + "epoch": 2.42, + "grad_norm": 2.640625, + "learning_rate": 1.6342766791561058e-06, + "loss": 0.7302, + "step": 1844 + }, + { + "epoch": 2.43, + "grad_norm": 2.6875, + "learning_rate": 1.6318162215865958e-06, + "loss": 0.7362, + "step": 1845 + }, + { + "epoch": 2.43, + "grad_norm": 2.71875, + "learning_rate": 1.6293567199067827e-06, + "loss": 0.7406, + "step": 1846 + }, + { + "epoch": 2.43, + "grad_norm": 2.90625, + "learning_rate": 1.6268981768246327e-06, + "loss": 0.7561, + "step": 1847 + }, + { + "epoch": 2.43, + "grad_norm": 2.78125, + "learning_rate": 1.624440595047056e-06, + "loss": 0.7794, + "step": 1848 + }, + { + "epoch": 2.43, + "grad_norm": 2.78125, + "learning_rate": 1.6219839772799057e-06, + "loss": 0.7504, + "step": 1849 + }, + { + "epoch": 2.43, + "grad_norm": 2.75, + "learning_rate": 1.6195283262279715e-06, + "loss": 0.7448, + "step": 1850 + }, + { + "epoch": 2.43, + "grad_norm": 2.78125, + "learning_rate": 1.617073644594982e-06, + "loss": 0.7844, + "step": 1851 + }, + { + "epoch": 2.43, + "grad_norm": 2.9375, + "learning_rate": 1.6146199350835956e-06, + "loss": 0.7575, + "step": 1852 + }, + { + "epoch": 2.44, + "grad_norm": 2.734375, + "learning_rate": 1.6121672003954003e-06, + "loss": 0.7345, + "step": 1853 + }, + { + "epoch": 2.44, + "grad_norm": 2.8125, + "learning_rate": 1.6097154432309114e-06, + "loss": 0.7701, + "step": 1854 + }, + { + "epoch": 2.44, + "grad_norm": 2.6875, + "learning_rate": 1.6072646662895696e-06, + "loss": 0.7505, + "step": 1855 + }, + { + "epoch": 2.44, + "grad_norm": 2.671875, + "learning_rate": 1.604814872269735e-06, + "loss": 0.7485, + "step": 1856 + }, + { + "epoch": 2.44, + "grad_norm": 2.703125, + "learning_rate": 1.6023660638686839e-06, + "loss": 0.7359, + "step": 1857 + }, + { + "epoch": 2.44, + "grad_norm": 2.65625, + "learning_rate": 1.5999182437826111e-06, + "loss": 0.7316, + "step": 1858 + }, + { + "epoch": 2.44, + "grad_norm": 2.875, + "learning_rate": 1.5974714147066188e-06, + "loss": 0.7864, + "step": 1859 + }, + { + "epoch": 2.45, + "grad_norm": 2.828125, + "learning_rate": 1.5950255793347217e-06, + "loss": 0.7477, + "step": 1860 + }, + { + "epoch": 2.45, + "grad_norm": 2.640625, + "learning_rate": 1.5925807403598387e-06, + "loss": 0.7377, + "step": 1861 + }, + { + "epoch": 2.45, + "grad_norm": 2.75, + "learning_rate": 1.5901369004737922e-06, + "loss": 0.743, + "step": 1862 + }, + { + "epoch": 2.45, + "grad_norm": 2.84375, + "learning_rate": 1.5876940623673043e-06, + "loss": 0.7496, + "step": 1863 + }, + { + "epoch": 2.45, + "grad_norm": 2.75, + "learning_rate": 1.5852522287299946e-06, + "loss": 0.7413, + "step": 1864 + }, + { + "epoch": 2.45, + "grad_norm": 2.734375, + "learning_rate": 1.5828114022503755e-06, + "loss": 0.7474, + "step": 1865 + }, + { + "epoch": 2.45, + "grad_norm": 2.875, + "learning_rate": 1.5803715856158521e-06, + "loss": 0.7492, + "step": 1866 + }, + { + "epoch": 2.45, + "grad_norm": 3.90625, + "learning_rate": 1.5779327815127166e-06, + "loss": 0.775, + "step": 1867 + }, + { + "epoch": 2.46, + "grad_norm": 2.6875, + "learning_rate": 1.5754949926261471e-06, + "loss": 0.7528, + "step": 1868 + }, + { + "epoch": 2.46, + "grad_norm": 2.625, + "learning_rate": 1.5730582216402023e-06, + "loss": 0.7498, + "step": 1869 + }, + { + "epoch": 2.46, + "grad_norm": 2.703125, + "learning_rate": 1.5706224712378237e-06, + "loss": 0.7661, + "step": 1870 + }, + { + "epoch": 2.46, + "grad_norm": 2.65625, + "learning_rate": 1.5681877441008248e-06, + "loss": 0.7422, + "step": 1871 + }, + { + "epoch": 2.46, + "grad_norm": 2.671875, + "learning_rate": 1.565754042909895e-06, + "loss": 0.7388, + "step": 1872 + }, + { + "epoch": 2.46, + "grad_norm": 2.71875, + "learning_rate": 1.5633213703445928e-06, + "loss": 0.7457, + "step": 1873 + }, + { + "epoch": 2.46, + "grad_norm": 2.8125, + "learning_rate": 1.5608897290833466e-06, + "loss": 0.7625, + "step": 1874 + }, + { + "epoch": 2.47, + "grad_norm": 2.71875, + "learning_rate": 1.5584591218034468e-06, + "loss": 0.7661, + "step": 1875 + }, + { + "epoch": 2.47, + "grad_norm": 2.75, + "learning_rate": 1.5560295511810465e-06, + "loss": 0.7422, + "step": 1876 + }, + { + "epoch": 2.47, + "grad_norm": 2.875, + "learning_rate": 1.5536010198911566e-06, + "loss": 0.7557, + "step": 1877 + }, + { + "epoch": 2.47, + "grad_norm": 2.859375, + "learning_rate": 1.5511735306076443e-06, + "loss": 0.7571, + "step": 1878 + }, + { + "epoch": 2.47, + "grad_norm": 2.75, + "learning_rate": 1.5487470860032293e-06, + "loss": 0.7262, + "step": 1879 + }, + { + "epoch": 2.47, + "grad_norm": 2.84375, + "learning_rate": 1.5463216887494819e-06, + "loss": 0.7954, + "step": 1880 + }, + { + "epoch": 2.47, + "grad_norm": 2.65625, + "learning_rate": 1.543897341516818e-06, + "loss": 0.7426, + "step": 1881 + }, + { + "epoch": 2.47, + "grad_norm": 2.90625, + "learning_rate": 1.5414740469744986e-06, + "loss": 0.7654, + "step": 1882 + }, + { + "epoch": 2.48, + "grad_norm": 2.734375, + "learning_rate": 1.5390518077906235e-06, + "loss": 0.7477, + "step": 1883 + }, + { + "epoch": 2.48, + "grad_norm": 2.796875, + "learning_rate": 1.5366306266321337e-06, + "loss": 0.7503, + "step": 1884 + }, + { + "epoch": 2.48, + "grad_norm": 2.6875, + "learning_rate": 1.5342105061648028e-06, + "loss": 0.7304, + "step": 1885 + }, + { + "epoch": 2.48, + "grad_norm": 2.734375, + "learning_rate": 1.5317914490532376e-06, + "loss": 0.7892, + "step": 1886 + }, + { + "epoch": 2.48, + "grad_norm": 2.75, + "learning_rate": 1.5293734579608732e-06, + "loss": 0.7754, + "step": 1887 + }, + { + "epoch": 2.48, + "grad_norm": 2.765625, + "learning_rate": 1.5269565355499738e-06, + "loss": 0.7555, + "step": 1888 + }, + { + "epoch": 2.48, + "grad_norm": 2.8125, + "learning_rate": 1.5245406844816229e-06, + "loss": 0.7702, + "step": 1889 + }, + { + "epoch": 2.49, + "grad_norm": 2.71875, + "learning_rate": 1.5221259074157266e-06, + "loss": 0.7335, + "step": 1890 + }, + { + "epoch": 2.49, + "grad_norm": 2.6875, + "learning_rate": 1.5197122070110082e-06, + "loss": 0.7516, + "step": 1891 + }, + { + "epoch": 2.49, + "grad_norm": 2.703125, + "learning_rate": 1.5172995859250067e-06, + "loss": 0.757, + "step": 1892 + }, + { + "epoch": 2.49, + "grad_norm": 2.671875, + "learning_rate": 1.5148880468140708e-06, + "loss": 0.7382, + "step": 1893 + }, + { + "epoch": 2.49, + "grad_norm": 2.765625, + "learning_rate": 1.5124775923333604e-06, + "loss": 0.7828, + "step": 1894 + }, + { + "epoch": 2.49, + "grad_norm": 2.734375, + "learning_rate": 1.5100682251368367e-06, + "loss": 0.7608, + "step": 1895 + }, + { + "epoch": 2.49, + "grad_norm": 2.6875, + "learning_rate": 1.5076599478772691e-06, + "loss": 0.767, + "step": 1896 + }, + { + "epoch": 2.49, + "grad_norm": 2.796875, + "learning_rate": 1.5052527632062236e-06, + "loss": 0.7687, + "step": 1897 + }, + { + "epoch": 2.5, + "grad_norm": 2.703125, + "learning_rate": 1.5028466737740643e-06, + "loss": 0.7354, + "step": 1898 + }, + { + "epoch": 2.5, + "grad_norm": 2.875, + "learning_rate": 1.5004416822299502e-06, + "loss": 0.7817, + "step": 1899 + }, + { + "epoch": 2.5, + "grad_norm": 2.6875, + "learning_rate": 1.4980377912218291e-06, + "loss": 0.7343, + "step": 1900 + }, + { + "epoch": 2.5, + "grad_norm": 2.703125, + "learning_rate": 1.4956350033964395e-06, + "loss": 0.7407, + "step": 1901 + }, + { + "epoch": 2.5, + "grad_norm": 2.75, + "learning_rate": 1.4932333213993043e-06, + "loss": 0.7499, + "step": 1902 + }, + { + "epoch": 2.5, + "grad_norm": 2.75, + "learning_rate": 1.4908327478747287e-06, + "loss": 0.7599, + "step": 1903 + }, + { + "epoch": 2.5, + "grad_norm": 2.71875, + "learning_rate": 1.4884332854657973e-06, + "loss": 0.7294, + "step": 1904 + }, + { + "epoch": 2.5, + "grad_norm": 2.71875, + "learning_rate": 1.4860349368143734e-06, + "loss": 0.7463, + "step": 1905 + }, + { + "epoch": 2.51, + "grad_norm": 2.78125, + "learning_rate": 1.4836377045610903e-06, + "loss": 0.756, + "step": 1906 + }, + { + "epoch": 2.51, + "grad_norm": 2.671875, + "learning_rate": 1.481241591345355e-06, + "loss": 0.7494, + "step": 1907 + }, + { + "epoch": 2.51, + "grad_norm": 2.84375, + "learning_rate": 1.4788465998053417e-06, + "loss": 0.7749, + "step": 1908 + }, + { + "epoch": 2.51, + "grad_norm": 2.8125, + "learning_rate": 1.4764527325779886e-06, + "loss": 0.7566, + "step": 1909 + }, + { + "epoch": 2.51, + "grad_norm": 2.765625, + "learning_rate": 1.4740599922989985e-06, + "loss": 0.7418, + "step": 1910 + }, + { + "epoch": 2.51, + "grad_norm": 2.765625, + "learning_rate": 1.4716683816028314e-06, + "loss": 0.7508, + "step": 1911 + }, + { + "epoch": 2.51, + "grad_norm": 2.703125, + "learning_rate": 1.4692779031227032e-06, + "loss": 0.7634, + "step": 1912 + }, + { + "epoch": 2.52, + "grad_norm": 2.734375, + "learning_rate": 1.4668885594905835e-06, + "loss": 0.785, + "step": 1913 + }, + { + "epoch": 2.52, + "grad_norm": 2.78125, + "learning_rate": 1.4645003533371943e-06, + "loss": 0.7655, + "step": 1914 + }, + { + "epoch": 2.52, + "grad_norm": 2.75, + "learning_rate": 1.4621132872920034e-06, + "loss": 0.7492, + "step": 1915 + }, + { + "epoch": 2.52, + "grad_norm": 2.703125, + "learning_rate": 1.4597273639832232e-06, + "loss": 0.7356, + "step": 1916 + }, + { + "epoch": 2.52, + "grad_norm": 2.765625, + "learning_rate": 1.4573425860378102e-06, + "loss": 0.7588, + "step": 1917 + }, + { + "epoch": 2.52, + "grad_norm": 2.703125, + "learning_rate": 1.4549589560814558e-06, + "loss": 0.7601, + "step": 1918 + }, + { + "epoch": 2.52, + "grad_norm": 2.703125, + "learning_rate": 1.4525764767385906e-06, + "loss": 0.7489, + "step": 1919 + }, + { + "epoch": 2.52, + "grad_norm": 2.765625, + "learning_rate": 1.4501951506323765e-06, + "loss": 0.7494, + "step": 1920 + }, + { + "epoch": 2.53, + "grad_norm": 2.796875, + "learning_rate": 1.4478149803847087e-06, + "loss": 0.7521, + "step": 1921 + }, + { + "epoch": 2.53, + "grad_norm": 2.71875, + "learning_rate": 1.4454359686162067e-06, + "loss": 0.782, + "step": 1922 + }, + { + "epoch": 2.53, + "grad_norm": 2.6875, + "learning_rate": 1.4430581179462167e-06, + "loss": 0.7726, + "step": 1923 + }, + { + "epoch": 2.53, + "grad_norm": 2.875, + "learning_rate": 1.440681430992803e-06, + "loss": 0.7734, + "step": 1924 + }, + { + "epoch": 2.53, + "grad_norm": 2.734375, + "learning_rate": 1.4383059103727518e-06, + "loss": 0.757, + "step": 1925 + }, + { + "epoch": 2.53, + "grad_norm": 2.78125, + "learning_rate": 1.4359315587015648e-06, + "loss": 0.7652, + "step": 1926 + }, + { + "epoch": 2.53, + "grad_norm": 2.78125, + "learning_rate": 1.4335583785934556e-06, + "loss": 0.7448, + "step": 1927 + }, + { + "epoch": 2.54, + "grad_norm": 2.84375, + "learning_rate": 1.431186372661349e-06, + "loss": 0.7565, + "step": 1928 + }, + { + "epoch": 2.54, + "grad_norm": 2.734375, + "learning_rate": 1.4288155435168755e-06, + "loss": 0.765, + "step": 1929 + }, + { + "epoch": 2.54, + "grad_norm": 2.734375, + "learning_rate": 1.4264458937703717e-06, + "loss": 0.7733, + "step": 1930 + }, + { + "epoch": 2.54, + "grad_norm": 2.84375, + "learning_rate": 1.4240774260308749e-06, + "loss": 0.7811, + "step": 1931 + }, + { + "epoch": 2.54, + "grad_norm": 2.78125, + "learning_rate": 1.4217101429061197e-06, + "loss": 0.7747, + "step": 1932 + }, + { + "epoch": 2.54, + "grad_norm": 2.75, + "learning_rate": 1.4193440470025391e-06, + "loss": 0.7529, + "step": 1933 + }, + { + "epoch": 2.54, + "grad_norm": 2.65625, + "learning_rate": 1.4169791409252565e-06, + "loss": 0.7592, + "step": 1934 + }, + { + "epoch": 2.54, + "grad_norm": 2.625, + "learning_rate": 1.4146154272780874e-06, + "loss": 0.7299, + "step": 1935 + }, + { + "epoch": 2.55, + "grad_norm": 2.734375, + "learning_rate": 1.4122529086635312e-06, + "loss": 0.7578, + "step": 1936 + }, + { + "epoch": 2.55, + "grad_norm": 2.796875, + "learning_rate": 1.4098915876827748e-06, + "loss": 0.782, + "step": 1937 + }, + { + "epoch": 2.55, + "grad_norm": 2.703125, + "learning_rate": 1.4075314669356852e-06, + "loss": 0.7855, + "step": 1938 + }, + { + "epoch": 2.55, + "grad_norm": 2.78125, + "learning_rate": 1.405172549020807e-06, + "loss": 0.76, + "step": 1939 + }, + { + "epoch": 2.55, + "grad_norm": 2.8125, + "learning_rate": 1.4028148365353627e-06, + "loss": 0.7782, + "step": 1940 + }, + { + "epoch": 2.55, + "grad_norm": 2.671875, + "learning_rate": 1.400458332075247e-06, + "loss": 0.7459, + "step": 1941 + }, + { + "epoch": 2.55, + "grad_norm": 2.703125, + "learning_rate": 1.398103038235022e-06, + "loss": 0.7724, + "step": 1942 + }, + { + "epoch": 2.56, + "grad_norm": 2.796875, + "learning_rate": 1.3957489576079196e-06, + "loss": 0.7647, + "step": 1943 + }, + { + "epoch": 2.56, + "grad_norm": 2.8125, + "learning_rate": 1.3933960927858347e-06, + "loss": 0.7584, + "step": 1944 + }, + { + "epoch": 2.56, + "grad_norm": 2.78125, + "learning_rate": 1.3910444463593243e-06, + "loss": 0.7537, + "step": 1945 + }, + { + "epoch": 2.56, + "grad_norm": 3.09375, + "learning_rate": 1.3886940209176031e-06, + "loss": 0.7376, + "step": 1946 + }, + { + "epoch": 2.56, + "grad_norm": 2.78125, + "learning_rate": 1.3863448190485426e-06, + "loss": 0.7632, + "step": 1947 + }, + { + "epoch": 2.56, + "grad_norm": 2.6875, + "learning_rate": 1.3839968433386659e-06, + "loss": 0.7594, + "step": 1948 + }, + { + "epoch": 2.56, + "grad_norm": 2.828125, + "learning_rate": 1.3816500963731467e-06, + "loss": 0.772, + "step": 1949 + }, + { + "epoch": 2.56, + "grad_norm": 2.859375, + "learning_rate": 1.379304580735806e-06, + "loss": 0.7833, + "step": 1950 + }, + { + "epoch": 2.57, + "grad_norm": 2.6875, + "learning_rate": 1.3769602990091086e-06, + "loss": 0.7702, + "step": 1951 + }, + { + "epoch": 2.57, + "grad_norm": 2.765625, + "learning_rate": 1.3746172537741618e-06, + "loss": 0.7831, + "step": 1952 + }, + { + "epoch": 2.57, + "grad_norm": 2.671875, + "learning_rate": 1.3722754476107086e-06, + "loss": 0.7345, + "step": 1953 + }, + { + "epoch": 2.57, + "grad_norm": 2.828125, + "learning_rate": 1.3699348830971316e-06, + "loss": 0.7745, + "step": 1954 + }, + { + "epoch": 2.57, + "grad_norm": 2.625, + "learning_rate": 1.3675955628104435e-06, + "loss": 0.7231, + "step": 1955 + }, + { + "epoch": 2.57, + "grad_norm": 2.703125, + "learning_rate": 1.3652574893262885e-06, + "loss": 0.731, + "step": 1956 + }, + { + "epoch": 2.57, + "grad_norm": 2.75, + "learning_rate": 1.3629206652189366e-06, + "loss": 0.7157, + "step": 1957 + }, + { + "epoch": 2.58, + "grad_norm": 2.6875, + "learning_rate": 1.3605850930612862e-06, + "loss": 0.7594, + "step": 1958 + }, + { + "epoch": 2.58, + "grad_norm": 2.765625, + "learning_rate": 1.3582507754248519e-06, + "loss": 0.75, + "step": 1959 + }, + { + "epoch": 2.58, + "grad_norm": 2.78125, + "learning_rate": 1.3559177148797698e-06, + "loss": 0.7552, + "step": 1960 + }, + { + "epoch": 2.58, + "grad_norm": 2.75, + "learning_rate": 1.3535859139947916e-06, + "loss": 0.7446, + "step": 1961 + }, + { + "epoch": 2.58, + "grad_norm": 2.78125, + "learning_rate": 1.351255375337282e-06, + "loss": 0.7801, + "step": 1962 + }, + { + "epoch": 2.58, + "grad_norm": 2.75, + "learning_rate": 1.3489261014732164e-06, + "loss": 0.7507, + "step": 1963 + }, + { + "epoch": 2.58, + "grad_norm": 2.78125, + "learning_rate": 1.346598094967178e-06, + "loss": 0.7308, + "step": 1964 + }, + { + "epoch": 2.58, + "grad_norm": 2.703125, + "learning_rate": 1.3442713583823514e-06, + "loss": 0.7528, + "step": 1965 + }, + { + "epoch": 2.59, + "grad_norm": 2.78125, + "learning_rate": 1.3419458942805274e-06, + "loss": 0.7576, + "step": 1966 + }, + { + "epoch": 2.59, + "grad_norm": 2.765625, + "learning_rate": 1.3396217052220939e-06, + "loss": 0.7399, + "step": 1967 + }, + { + "epoch": 2.59, + "grad_norm": 2.6875, + "learning_rate": 1.3372987937660342e-06, + "loss": 0.7796, + "step": 1968 + }, + { + "epoch": 2.59, + "grad_norm": 2.796875, + "learning_rate": 1.334977162469926e-06, + "loss": 0.7599, + "step": 1969 + }, + { + "epoch": 2.59, + "grad_norm": 2.765625, + "learning_rate": 1.3326568138899376e-06, + "loss": 0.7622, + "step": 1970 + }, + { + "epoch": 2.59, + "grad_norm": 2.671875, + "learning_rate": 1.330337750580823e-06, + "loss": 0.7363, + "step": 1971 + }, + { + "epoch": 2.59, + "grad_norm": 2.6875, + "learning_rate": 1.3280199750959233e-06, + "loss": 0.7409, + "step": 1972 + }, + { + "epoch": 2.6, + "grad_norm": 2.78125, + "learning_rate": 1.3257034899871613e-06, + "loss": 0.7654, + "step": 1973 + }, + { + "epoch": 2.6, + "grad_norm": 2.65625, + "learning_rate": 1.3233882978050389e-06, + "loss": 0.7545, + "step": 1974 + }, + { + "epoch": 2.6, + "grad_norm": 2.796875, + "learning_rate": 1.3210744010986337e-06, + "loss": 0.7623, + "step": 1975 + }, + { + "epoch": 2.6, + "grad_norm": 2.625, + "learning_rate": 1.3187618024155979e-06, + "loss": 0.7535, + "step": 1976 + }, + { + "epoch": 2.6, + "grad_norm": 2.90625, + "learning_rate": 1.316450504302154e-06, + "loss": 0.7667, + "step": 1977 + }, + { + "epoch": 2.6, + "grad_norm": 2.765625, + "learning_rate": 1.314140509303093e-06, + "loss": 0.7595, + "step": 1978 + }, + { + "epoch": 2.6, + "grad_norm": 2.75, + "learning_rate": 1.311831819961771e-06, + "loss": 0.7699, + "step": 1979 + }, + { + "epoch": 2.6, + "grad_norm": 2.796875, + "learning_rate": 1.3095244388201062e-06, + "loss": 0.7518, + "step": 1980 + }, + { + "epoch": 2.61, + "grad_norm": 2.734375, + "learning_rate": 1.3072183684185769e-06, + "loss": 0.7655, + "step": 1981 + }, + { + "epoch": 2.61, + "grad_norm": 2.78125, + "learning_rate": 1.3049136112962188e-06, + "loss": 0.7507, + "step": 1982 + }, + { + "epoch": 2.61, + "grad_norm": 2.8125, + "learning_rate": 1.302610169990618e-06, + "loss": 0.7712, + "step": 1983 + }, + { + "epoch": 2.61, + "grad_norm": 2.71875, + "learning_rate": 1.3003080470379176e-06, + "loss": 0.7452, + "step": 1984 + }, + { + "epoch": 2.61, + "grad_norm": 2.703125, + "learning_rate": 1.2980072449728054e-06, + "loss": 0.7791, + "step": 1985 + }, + { + "epoch": 2.61, + "grad_norm": 2.796875, + "learning_rate": 1.2957077663285155e-06, + "loss": 0.7586, + "step": 1986 + }, + { + "epoch": 2.61, + "grad_norm": 2.78125, + "learning_rate": 1.2934096136368256e-06, + "loss": 0.7557, + "step": 1987 + }, + { + "epoch": 2.62, + "grad_norm": 2.78125, + "learning_rate": 1.2911127894280542e-06, + "loss": 0.7872, + "step": 1988 + }, + { + "epoch": 2.62, + "grad_norm": 2.765625, + "learning_rate": 1.2888172962310535e-06, + "loss": 0.7798, + "step": 1989 + }, + { + "epoch": 2.62, + "grad_norm": 2.859375, + "learning_rate": 1.2865231365732136e-06, + "loss": 0.7474, + "step": 1990 + }, + { + "epoch": 2.62, + "grad_norm": 2.796875, + "learning_rate": 1.2842303129804559e-06, + "loss": 0.741, + "step": 1991 + }, + { + "epoch": 2.62, + "grad_norm": 2.796875, + "learning_rate": 1.2819388279772298e-06, + "loss": 0.7628, + "step": 1992 + }, + { + "epoch": 2.62, + "grad_norm": 2.765625, + "learning_rate": 1.279648684086512e-06, + "loss": 0.7504, + "step": 1993 + }, + { + "epoch": 2.62, + "grad_norm": 2.671875, + "learning_rate": 1.2773598838298012e-06, + "loss": 0.7745, + "step": 1994 + }, + { + "epoch": 2.62, + "grad_norm": 2.75, + "learning_rate": 1.2750724297271184e-06, + "loss": 0.7362, + "step": 1995 + }, + { + "epoch": 2.63, + "grad_norm": 2.75, + "learning_rate": 1.2727863242970007e-06, + "loss": 0.748, + "step": 1996 + }, + { + "epoch": 2.63, + "grad_norm": 2.796875, + "learning_rate": 1.2705015700565015e-06, + "loss": 0.7439, + "step": 1997 + }, + { + "epoch": 2.63, + "grad_norm": 2.78125, + "learning_rate": 1.2682181695211865e-06, + "loss": 0.7709, + "step": 1998 + }, + { + "epoch": 2.63, + "grad_norm": 2.78125, + "learning_rate": 1.2659361252051303e-06, + "loss": 0.744, + "step": 1999 + }, + { + "epoch": 2.63, + "grad_norm": 2.734375, + "learning_rate": 1.2636554396209155e-06, + "loss": 0.7478, + "step": 2000 + }, + { + "epoch": 2.63, + "grad_norm": 2.75, + "learning_rate": 1.2613761152796255e-06, + "loss": 0.7561, + "step": 2001 + }, + { + "epoch": 2.63, + "grad_norm": 2.671875, + "learning_rate": 1.2590981546908481e-06, + "loss": 0.7515, + "step": 2002 + }, + { + "epoch": 2.64, + "grad_norm": 2.75, + "learning_rate": 1.25682156036267e-06, + "loss": 0.7656, + "step": 2003 + }, + { + "epoch": 2.64, + "grad_norm": 2.671875, + "learning_rate": 1.2545463348016715e-06, + "loss": 0.7389, + "step": 2004 + }, + { + "epoch": 2.64, + "grad_norm": 2.765625, + "learning_rate": 1.2522724805129267e-06, + "loss": 0.7932, + "step": 2005 + }, + { + "epoch": 2.64, + "grad_norm": 2.78125, + "learning_rate": 1.2500000000000007e-06, + "loss": 0.755, + "step": 2006 + }, + { + "epoch": 2.64, + "grad_norm": 2.75, + "learning_rate": 1.2477288957649434e-06, + "loss": 0.7707, + "step": 2007 + }, + { + "epoch": 2.64, + "grad_norm": 2.828125, + "learning_rate": 1.245459170308292e-06, + "loss": 0.7851, + "step": 2008 + }, + { + "epoch": 2.64, + "grad_norm": 2.734375, + "learning_rate": 1.2431908261290651e-06, + "loss": 0.7658, + "step": 2009 + }, + { + "epoch": 2.64, + "grad_norm": 2.765625, + "learning_rate": 1.2409238657247598e-06, + "loss": 0.763, + "step": 2010 + }, + { + "epoch": 2.65, + "grad_norm": 2.8125, + "learning_rate": 1.2386582915913509e-06, + "loss": 0.7939, + "step": 2011 + }, + { + "epoch": 2.65, + "grad_norm": 2.65625, + "learning_rate": 1.236394106223285e-06, + "loss": 0.7503, + "step": 2012 + }, + { + "epoch": 2.65, + "grad_norm": 2.765625, + "learning_rate": 1.2341313121134813e-06, + "loss": 0.7518, + "step": 2013 + }, + { + "epoch": 2.65, + "grad_norm": 2.65625, + "learning_rate": 1.2318699117533264e-06, + "loss": 0.7408, + "step": 2014 + }, + { + "epoch": 2.65, + "grad_norm": 2.78125, + "learning_rate": 1.2296099076326725e-06, + "loss": 0.7606, + "step": 2015 + }, + { + "epoch": 2.65, + "grad_norm": 2.78125, + "learning_rate": 1.2273513022398345e-06, + "loss": 0.7497, + "step": 2016 + }, + { + "epoch": 2.65, + "grad_norm": 2.765625, + "learning_rate": 1.2250940980615885e-06, + "loss": 0.7305, + "step": 2017 + }, + { + "epoch": 2.66, + "grad_norm": 2.71875, + "learning_rate": 1.2228382975831643e-06, + "loss": 0.754, + "step": 2018 + }, + { + "epoch": 2.66, + "grad_norm": 2.71875, + "learning_rate": 1.2205839032882497e-06, + "loss": 0.7555, + "step": 2019 + }, + { + "epoch": 2.66, + "grad_norm": 2.875, + "learning_rate": 1.2183309176589822e-06, + "loss": 0.7863, + "step": 2020 + }, + { + "epoch": 2.66, + "grad_norm": 2.796875, + "learning_rate": 1.216079343175951e-06, + "loss": 0.772, + "step": 2021 + }, + { + "epoch": 2.66, + "grad_norm": 2.71875, + "learning_rate": 1.2138291823181889e-06, + "loss": 0.7623, + "step": 2022 + }, + { + "epoch": 2.66, + "grad_norm": 2.734375, + "learning_rate": 1.211580437563174e-06, + "loss": 0.7612, + "step": 2023 + }, + { + "epoch": 2.66, + "grad_norm": 2.6875, + "learning_rate": 1.209333111386823e-06, + "loss": 0.7407, + "step": 2024 + }, + { + "epoch": 2.66, + "grad_norm": 2.765625, + "learning_rate": 1.2070872062634923e-06, + "loss": 0.7768, + "step": 2025 + }, + { + "epoch": 2.67, + "grad_norm": 2.703125, + "learning_rate": 1.2048427246659738e-06, + "loss": 0.7493, + "step": 2026 + }, + { + "epoch": 2.67, + "grad_norm": 2.71875, + "learning_rate": 1.2025996690654923e-06, + "loss": 0.7443, + "step": 2027 + }, + { + "epoch": 2.67, + "grad_norm": 2.671875, + "learning_rate": 1.2003580419317012e-06, + "loss": 0.7756, + "step": 2028 + }, + { + "epoch": 2.67, + "grad_norm": 2.671875, + "learning_rate": 1.1981178457326827e-06, + "loss": 0.7269, + "step": 2029 + }, + { + "epoch": 2.67, + "grad_norm": 2.796875, + "learning_rate": 1.1958790829349423e-06, + "loss": 0.7283, + "step": 2030 + }, + { + "epoch": 2.67, + "grad_norm": 2.734375, + "learning_rate": 1.193641756003408e-06, + "loss": 0.7455, + "step": 2031 + }, + { + "epoch": 2.67, + "grad_norm": 2.71875, + "learning_rate": 1.1914058674014264e-06, + "loss": 0.7539, + "step": 2032 + }, + { + "epoch": 2.68, + "grad_norm": 2.84375, + "learning_rate": 1.1891714195907608e-06, + "loss": 0.7683, + "step": 2033 + }, + { + "epoch": 2.68, + "grad_norm": 2.75, + "learning_rate": 1.1869384150315883e-06, + "loss": 0.7567, + "step": 2034 + }, + { + "epoch": 2.68, + "grad_norm": 2.71875, + "learning_rate": 1.1847068561824972e-06, + "loss": 0.7352, + "step": 2035 + }, + { + "epoch": 2.68, + "grad_norm": 2.703125, + "learning_rate": 1.182476745500482e-06, + "loss": 0.7756, + "step": 2036 + }, + { + "epoch": 2.68, + "grad_norm": 2.765625, + "learning_rate": 1.180248085440945e-06, + "loss": 0.7366, + "step": 2037 + }, + { + "epoch": 2.68, + "grad_norm": 2.765625, + "learning_rate": 1.17802087845769e-06, + "loss": 0.7701, + "step": 2038 + }, + { + "epoch": 2.68, + "grad_norm": 2.78125, + "learning_rate": 1.175795127002923e-06, + "loss": 0.7726, + "step": 2039 + }, + { + "epoch": 2.68, + "grad_norm": 2.640625, + "learning_rate": 1.1735708335272452e-06, + "loss": 0.7539, + "step": 2040 + }, + { + "epoch": 2.69, + "grad_norm": 2.71875, + "learning_rate": 1.1713480004796546e-06, + "loss": 0.7493, + "step": 2041 + }, + { + "epoch": 2.69, + "grad_norm": 2.65625, + "learning_rate": 1.1691266303075377e-06, + "loss": 0.7476, + "step": 2042 + }, + { + "epoch": 2.69, + "grad_norm": 2.796875, + "learning_rate": 1.1669067254566735e-06, + "loss": 0.7847, + "step": 2043 + }, + { + "epoch": 2.69, + "grad_norm": 2.6875, + "learning_rate": 1.164688288371227e-06, + "loss": 0.7266, + "step": 2044 + }, + { + "epoch": 2.69, + "grad_norm": 2.8125, + "learning_rate": 1.1624713214937464e-06, + "loss": 0.7495, + "step": 2045 + }, + { + "epoch": 2.69, + "grad_norm": 2.765625, + "learning_rate": 1.1602558272651623e-06, + "loss": 0.7904, + "step": 2046 + }, + { + "epoch": 2.69, + "grad_norm": 2.875, + "learning_rate": 1.1580418081247826e-06, + "loss": 0.783, + "step": 2047 + }, + { + "epoch": 2.7, + "grad_norm": 2.78125, + "learning_rate": 1.1558292665102923e-06, + "loss": 0.773, + "step": 2048 + }, + { + "epoch": 2.7, + "grad_norm": 2.734375, + "learning_rate": 1.1536182048577488e-06, + "loss": 0.7444, + "step": 2049 + }, + { + "epoch": 2.7, + "grad_norm": 2.71875, + "learning_rate": 1.1514086256015803e-06, + "loss": 0.7702, + "step": 2050 + }, + { + "epoch": 2.7, + "grad_norm": 2.625, + "learning_rate": 1.1492005311745827e-06, + "loss": 0.7549, + "step": 2051 + }, + { + "epoch": 2.7, + "grad_norm": 2.640625, + "learning_rate": 1.1469939240079176e-06, + "loss": 0.7494, + "step": 2052 + }, + { + "epoch": 2.7, + "grad_norm": 2.71875, + "learning_rate": 1.1447888065311092e-06, + "loss": 0.7195, + "step": 2053 + }, + { + "epoch": 2.7, + "grad_norm": 2.734375, + "learning_rate": 1.142585181172039e-06, + "loss": 0.7614, + "step": 2054 + }, + { + "epoch": 2.7, + "grad_norm": 2.765625, + "learning_rate": 1.140383050356949e-06, + "loss": 0.7734, + "step": 2055 + }, + { + "epoch": 2.71, + "grad_norm": 2.796875, + "learning_rate": 1.138182416510434e-06, + "loss": 0.7666, + "step": 2056 + }, + { + "epoch": 2.71, + "grad_norm": 2.828125, + "learning_rate": 1.1359832820554406e-06, + "loss": 0.7949, + "step": 2057 + }, + { + "epoch": 2.71, + "grad_norm": 2.6875, + "learning_rate": 1.1337856494132663e-06, + "loss": 0.7712, + "step": 2058 + }, + { + "epoch": 2.71, + "grad_norm": 2.890625, + "learning_rate": 1.1315895210035532e-06, + "loss": 0.7696, + "step": 2059 + }, + { + "epoch": 2.71, + "grad_norm": 2.78125, + "learning_rate": 1.1293948992442866e-06, + "loss": 0.7466, + "step": 2060 + }, + { + "epoch": 2.71, + "grad_norm": 2.75, + "learning_rate": 1.1272017865517942e-06, + "loss": 0.7584, + "step": 2061 + }, + { + "epoch": 2.71, + "grad_norm": 2.703125, + "learning_rate": 1.1250101853407427e-06, + "loss": 0.7616, + "step": 2062 + }, + { + "epoch": 2.72, + "grad_norm": 2.75, + "learning_rate": 1.1228200980241334e-06, + "loss": 0.7566, + "step": 2063 + }, + { + "epoch": 2.72, + "grad_norm": 2.640625, + "learning_rate": 1.1206315270133023e-06, + "loss": 0.7701, + "step": 2064 + }, + { + "epoch": 2.72, + "grad_norm": 2.71875, + "learning_rate": 1.1184444747179118e-06, + "loss": 0.7535, + "step": 2065 + }, + { + "epoch": 2.72, + "grad_norm": 2.890625, + "learning_rate": 1.1162589435459587e-06, + "loss": 0.7602, + "step": 2066 + }, + { + "epoch": 2.72, + "grad_norm": 2.8125, + "learning_rate": 1.11407493590376e-06, + "loss": 0.7559, + "step": 2067 + }, + { + "epoch": 2.72, + "grad_norm": 2.75, + "learning_rate": 1.1118924541959573e-06, + "loss": 0.7498, + "step": 2068 + }, + { + "epoch": 2.72, + "grad_norm": 2.765625, + "learning_rate": 1.1097115008255113e-06, + "loss": 0.7616, + "step": 2069 + }, + { + "epoch": 2.72, + "grad_norm": 2.78125, + "learning_rate": 1.107532078193701e-06, + "loss": 0.7547, + "step": 2070 + }, + { + "epoch": 2.73, + "grad_norm": 2.921875, + "learning_rate": 1.1053541887001181e-06, + "loss": 0.7561, + "step": 2071 + }, + { + "epoch": 2.73, + "grad_norm": 2.734375, + "learning_rate": 1.103177834742668e-06, + "loss": 0.7742, + "step": 2072 + }, + { + "epoch": 2.73, + "grad_norm": 2.828125, + "learning_rate": 1.1010030187175649e-06, + "loss": 0.7814, + "step": 2073 + }, + { + "epoch": 2.73, + "grad_norm": 2.765625, + "learning_rate": 1.09882974301933e-06, + "loss": 0.7533, + "step": 2074 + }, + { + "epoch": 2.73, + "grad_norm": 2.6875, + "learning_rate": 1.0966580100407867e-06, + "loss": 0.7555, + "step": 2075 + }, + { + "epoch": 2.73, + "grad_norm": 2.84375, + "learning_rate": 1.094487822173065e-06, + "loss": 0.7399, + "step": 2076 + }, + { + "epoch": 2.73, + "grad_norm": 2.75, + "learning_rate": 1.0923191818055873e-06, + "loss": 0.746, + "step": 2077 + }, + { + "epoch": 2.74, + "grad_norm": 2.75, + "learning_rate": 1.0901520913260754e-06, + "loss": 0.7698, + "step": 2078 + }, + { + "epoch": 2.74, + "grad_norm": 2.8125, + "learning_rate": 1.0879865531205445e-06, + "loss": 0.7472, + "step": 2079 + }, + { + "epoch": 2.74, + "grad_norm": 2.734375, + "learning_rate": 1.0858225695733006e-06, + "loss": 0.7547, + "step": 2080 + }, + { + "epoch": 2.74, + "grad_norm": 2.734375, + "learning_rate": 1.0836601430669375e-06, + "loss": 0.7849, + "step": 2081 + }, + { + "epoch": 2.74, + "grad_norm": 2.703125, + "learning_rate": 1.0814992759823362e-06, + "loss": 0.752, + "step": 2082 + }, + { + "epoch": 2.74, + "grad_norm": 2.734375, + "learning_rate": 1.0793399706986568e-06, + "loss": 0.7723, + "step": 2083 + }, + { + "epoch": 2.74, + "grad_norm": 2.625, + "learning_rate": 1.0771822295933452e-06, + "loss": 0.7661, + "step": 2084 + }, + { + "epoch": 2.74, + "grad_norm": 2.71875, + "learning_rate": 1.0750260550421213e-06, + "loss": 0.7349, + "step": 2085 + }, + { + "epoch": 2.75, + "grad_norm": 2.828125, + "learning_rate": 1.072871449418982e-06, + "loss": 0.7516, + "step": 2086 + }, + { + "epoch": 2.75, + "grad_norm": 2.828125, + "learning_rate": 1.0707184150961952e-06, + "loss": 0.7643, + "step": 2087 + }, + { + "epoch": 2.75, + "grad_norm": 2.703125, + "learning_rate": 1.0685669544443014e-06, + "loss": 0.706, + "step": 2088 + }, + { + "epoch": 2.75, + "grad_norm": 2.859375, + "learning_rate": 1.0664170698321047e-06, + "loss": 0.7712, + "step": 2089 + }, + { + "epoch": 2.75, + "grad_norm": 2.71875, + "learning_rate": 1.064268763626677e-06, + "loss": 0.7324, + "step": 2090 + }, + { + "epoch": 2.75, + "grad_norm": 2.703125, + "learning_rate": 1.0621220381933508e-06, + "loss": 0.7532, + "step": 2091 + }, + { + "epoch": 2.75, + "grad_norm": 2.640625, + "learning_rate": 1.0599768958957193e-06, + "loss": 0.7601, + "step": 2092 + }, + { + "epoch": 2.76, + "grad_norm": 2.71875, + "learning_rate": 1.0578333390956317e-06, + "loss": 0.7398, + "step": 2093 + }, + { + "epoch": 2.76, + "grad_norm": 2.6875, + "learning_rate": 1.055691370153192e-06, + "loss": 0.7552, + "step": 2094 + }, + { + "epoch": 2.76, + "grad_norm": 2.734375, + "learning_rate": 1.0535509914267555e-06, + "loss": 0.7654, + "step": 2095 + }, + { + "epoch": 2.76, + "grad_norm": 2.796875, + "learning_rate": 1.051412205272927e-06, + "loss": 0.7987, + "step": 2096 + }, + { + "epoch": 2.76, + "grad_norm": 2.734375, + "learning_rate": 1.0492750140465579e-06, + "loss": 0.7698, + "step": 2097 + }, + { + "epoch": 2.76, + "grad_norm": 2.671875, + "learning_rate": 1.0471394201007435e-06, + "loss": 0.7388, + "step": 2098 + }, + { + "epoch": 2.76, + "grad_norm": 2.640625, + "learning_rate": 1.0450054257868197e-06, + "loss": 0.7583, + "step": 2099 + }, + { + "epoch": 2.76, + "grad_norm": 2.796875, + "learning_rate": 1.0428730334543636e-06, + "loss": 0.7343, + "step": 2100 + }, + { + "epoch": 2.77, + "grad_norm": 2.75, + "learning_rate": 1.0407422454511832e-06, + "loss": 0.7661, + "step": 2101 + }, + { + "epoch": 2.77, + "grad_norm": 2.65625, + "learning_rate": 1.0386130641233266e-06, + "loss": 0.7402, + "step": 2102 + }, + { + "epoch": 2.77, + "grad_norm": 2.84375, + "learning_rate": 1.0364854918150695e-06, + "loss": 0.7614, + "step": 2103 + }, + { + "epoch": 2.77, + "grad_norm": 2.75, + "learning_rate": 1.0343595308689156e-06, + "loss": 0.7466, + "step": 2104 + }, + { + "epoch": 2.77, + "grad_norm": 2.6875, + "learning_rate": 1.0322351836255955e-06, + "loss": 0.7368, + "step": 2105 + }, + { + "epoch": 2.77, + "grad_norm": 2.6875, + "learning_rate": 1.0301124524240645e-06, + "loss": 0.7474, + "step": 2106 + }, + { + "epoch": 2.77, + "grad_norm": 2.671875, + "learning_rate": 1.0279913396014944e-06, + "loss": 0.7587, + "step": 2107 + }, + { + "epoch": 2.78, + "grad_norm": 2.6875, + "learning_rate": 1.0258718474932788e-06, + "loss": 0.7471, + "step": 2108 + }, + { + "epoch": 2.78, + "grad_norm": 2.6875, + "learning_rate": 1.023753978433026e-06, + "loss": 0.7526, + "step": 2109 + }, + { + "epoch": 2.78, + "grad_norm": 2.671875, + "learning_rate": 1.021637734752557e-06, + "loss": 0.7497, + "step": 2110 + }, + { + "epoch": 2.78, + "grad_norm": 2.8125, + "learning_rate": 1.019523118781903e-06, + "loss": 0.7699, + "step": 2111 + }, + { + "epoch": 2.78, + "grad_norm": 2.796875, + "learning_rate": 1.0174101328493035e-06, + "loss": 0.7815, + "step": 2112 + }, + { + "epoch": 2.78, + "grad_norm": 2.78125, + "learning_rate": 1.015298779281203e-06, + "loss": 0.7576, + "step": 2113 + }, + { + "epoch": 2.78, + "grad_norm": 2.75, + "learning_rate": 1.0131890604022494e-06, + "loss": 0.7473, + "step": 2114 + }, + { + "epoch": 2.78, + "grad_norm": 2.765625, + "learning_rate": 1.0110809785352892e-06, + "loss": 0.773, + "step": 2115 + }, + { + "epoch": 2.79, + "grad_norm": 2.734375, + "learning_rate": 1.0089745360013685e-06, + "loss": 0.7607, + "step": 2116 + }, + { + "epoch": 2.79, + "grad_norm": 2.703125, + "learning_rate": 1.0068697351197282e-06, + "loss": 0.7515, + "step": 2117 + }, + { + "epoch": 2.79, + "grad_norm": 2.8125, + "learning_rate": 1.004766578207799e-06, + "loss": 0.7855, + "step": 2118 + }, + { + "epoch": 2.79, + "grad_norm": 2.703125, + "learning_rate": 1.0026650675812041e-06, + "loss": 0.7424, + "step": 2119 + }, + { + "epoch": 2.79, + "grad_norm": 2.796875, + "learning_rate": 1.0005652055537532e-06, + "loss": 0.7477, + "step": 2120 + }, + { + "epoch": 2.79, + "grad_norm": 2.640625, + "learning_rate": 9.984669944374431e-07, + "loss": 0.7713, + "step": 2121 + }, + { + "epoch": 2.79, + "grad_norm": 2.78125, + "learning_rate": 9.963704365424494e-07, + "loss": 0.7391, + "step": 2122 + }, + { + "epoch": 2.8, + "grad_norm": 2.6875, + "learning_rate": 9.94275534177131e-07, + "loss": 0.7412, + "step": 2123 + }, + { + "epoch": 2.8, + "grad_norm": 2.75, + "learning_rate": 9.921822896480194e-07, + "loss": 0.7501, + "step": 2124 + }, + { + "epoch": 2.8, + "grad_norm": 2.703125, + "learning_rate": 9.90090705259825e-07, + "loss": 0.7745, + "step": 2125 + }, + { + "epoch": 2.8, + "grad_norm": 2.671875, + "learning_rate": 9.880007833154284e-07, + "loss": 0.7363, + "step": 2126 + }, + { + "epoch": 2.8, + "grad_norm": 2.71875, + "learning_rate": 9.859125261158803e-07, + "loss": 0.7714, + "step": 2127 + }, + { + "epoch": 2.8, + "grad_norm": 2.765625, + "learning_rate": 9.838259359603987e-07, + "loss": 0.7714, + "step": 2128 + }, + { + "epoch": 2.8, + "grad_norm": 2.765625, + "learning_rate": 9.817410151463661e-07, + "loss": 0.7713, + "step": 2129 + }, + { + "epoch": 2.8, + "grad_norm": 2.71875, + "learning_rate": 9.796577659693262e-07, + "loss": 0.7567, + "step": 2130 + }, + { + "epoch": 2.81, + "grad_norm": 2.796875, + "learning_rate": 9.775761907229833e-07, + "loss": 0.7657, + "step": 2131 + }, + { + "epoch": 2.81, + "grad_norm": 2.75, + "learning_rate": 9.754962916991978e-07, + "loss": 0.7543, + "step": 2132 + }, + { + "epoch": 2.81, + "grad_norm": 2.671875, + "learning_rate": 9.73418071187985e-07, + "loss": 0.768, + "step": 2133 + }, + { + "epoch": 2.81, + "grad_norm": 2.65625, + "learning_rate": 9.713415314775122e-07, + "loss": 0.7543, + "step": 2134 + }, + { + "epoch": 2.81, + "grad_norm": 2.734375, + "learning_rate": 9.692666748540964e-07, + "loss": 0.7673, + "step": 2135 + }, + { + "epoch": 2.81, + "grad_norm": 2.75, + "learning_rate": 9.671935036021996e-07, + "loss": 0.7736, + "step": 2136 + }, + { + "epoch": 2.81, + "grad_norm": 2.671875, + "learning_rate": 9.651220200044303e-07, + "loss": 0.7508, + "step": 2137 + }, + { + "epoch": 2.82, + "grad_norm": 2.796875, + "learning_rate": 9.630522263415373e-07, + "loss": 0.7729, + "step": 2138 + }, + { + "epoch": 2.82, + "grad_norm": 2.671875, + "learning_rate": 9.609841248924115e-07, + "loss": 0.7631, + "step": 2139 + }, + { + "epoch": 2.82, + "grad_norm": 2.765625, + "learning_rate": 9.589177179340775e-07, + "loss": 0.7809, + "step": 2140 + }, + { + "epoch": 2.82, + "grad_norm": 2.859375, + "learning_rate": 9.568530077416966e-07, + "loss": 0.7715, + "step": 2141 + }, + { + "epoch": 2.82, + "grad_norm": 2.828125, + "learning_rate": 9.547899965885584e-07, + "loss": 0.7741, + "step": 2142 + }, + { + "epoch": 2.82, + "grad_norm": 2.9375, + "learning_rate": 9.527286867460859e-07, + "loss": 0.7761, + "step": 2143 + }, + { + "epoch": 2.82, + "grad_norm": 2.71875, + "learning_rate": 9.506690804838267e-07, + "loss": 0.7391, + "step": 2144 + }, + { + "epoch": 2.82, + "grad_norm": 2.75, + "learning_rate": 9.486111800694533e-07, + "loss": 0.778, + "step": 2145 + }, + { + "epoch": 2.83, + "grad_norm": 2.875, + "learning_rate": 9.465549877687602e-07, + "loss": 0.7719, + "step": 2146 + }, + { + "epoch": 2.83, + "grad_norm": 2.8125, + "learning_rate": 9.445005058456605e-07, + "loss": 0.7745, + "step": 2147 + }, + { + "epoch": 2.83, + "grad_norm": 2.640625, + "learning_rate": 9.424477365621848e-07, + "loss": 0.7478, + "step": 2148 + }, + { + "epoch": 2.83, + "grad_norm": 2.765625, + "learning_rate": 9.40396682178478e-07, + "loss": 0.7939, + "step": 2149 + }, + { + "epoch": 2.83, + "grad_norm": 2.828125, + "learning_rate": 9.383473449527963e-07, + "loss": 0.7486, + "step": 2150 + }, + { + "epoch": 2.83, + "grad_norm": 2.8125, + "learning_rate": 9.362997271415062e-07, + "loss": 0.7732, + "step": 2151 + }, + { + "epoch": 2.83, + "grad_norm": 2.703125, + "learning_rate": 9.342538309990804e-07, + "loss": 0.7198, + "step": 2152 + }, + { + "epoch": 2.83, + "grad_norm": 2.796875, + "learning_rate": 9.322096587780968e-07, + "loss": 0.7527, + "step": 2153 + }, + { + "epoch": 2.84, + "grad_norm": 2.734375, + "learning_rate": 9.30167212729233e-07, + "loss": 0.7459, + "step": 2154 + }, + { + "epoch": 2.84, + "grad_norm": 2.71875, + "learning_rate": 9.281264951012686e-07, + "loss": 0.7619, + "step": 2155 + }, + { + "epoch": 2.84, + "grad_norm": 2.734375, + "learning_rate": 9.260875081410783e-07, + "loss": 0.7705, + "step": 2156 + }, + { + "epoch": 2.84, + "grad_norm": 2.890625, + "learning_rate": 9.240502540936338e-07, + "loss": 0.7803, + "step": 2157 + }, + { + "epoch": 2.84, + "grad_norm": 2.765625, + "learning_rate": 9.220147352019965e-07, + "loss": 0.7465, + "step": 2158 + }, + { + "epoch": 2.84, + "grad_norm": 2.703125, + "learning_rate": 9.19980953707319e-07, + "loss": 0.7489, + "step": 2159 + }, + { + "epoch": 2.84, + "grad_norm": 2.6875, + "learning_rate": 9.17948911848838e-07, + "loss": 0.7689, + "step": 2160 + }, + { + "epoch": 2.85, + "grad_norm": 2.75, + "learning_rate": 9.159186118638783e-07, + "loss": 0.7427, + "step": 2161 + }, + { + "epoch": 2.85, + "grad_norm": 2.765625, + "learning_rate": 9.13890055987845e-07, + "loss": 0.7654, + "step": 2162 + }, + { + "epoch": 2.85, + "grad_norm": 2.8125, + "learning_rate": 9.118632464542235e-07, + "loss": 0.7351, + "step": 2163 + }, + { + "epoch": 2.85, + "grad_norm": 2.65625, + "learning_rate": 9.098381854945762e-07, + "loss": 0.7563, + "step": 2164 + }, + { + "epoch": 2.85, + "grad_norm": 2.609375, + "learning_rate": 9.078148753385404e-07, + "loss": 0.7409, + "step": 2165 + }, + { + "epoch": 2.85, + "grad_norm": 2.71875, + "learning_rate": 9.057933182138259e-07, + "loss": 0.749, + "step": 2166 + }, + { + "epoch": 2.85, + "grad_norm": 2.71875, + "learning_rate": 9.037735163462122e-07, + "loss": 0.7476, + "step": 2167 + }, + { + "epoch": 2.85, + "grad_norm": 2.765625, + "learning_rate": 9.017554719595458e-07, + "loss": 0.7527, + "step": 2168 + }, + { + "epoch": 2.86, + "grad_norm": 2.78125, + "learning_rate": 8.997391872757389e-07, + "loss": 0.7775, + "step": 2169 + }, + { + "epoch": 2.86, + "grad_norm": 2.796875, + "learning_rate": 8.977246645147655e-07, + "loss": 0.7566, + "step": 2170 + }, + { + "epoch": 2.86, + "grad_norm": 2.75, + "learning_rate": 8.957119058946614e-07, + "loss": 0.7725, + "step": 2171 + }, + { + "epoch": 2.86, + "grad_norm": 2.84375, + "learning_rate": 8.937009136315167e-07, + "loss": 0.7577, + "step": 2172 + }, + { + "epoch": 2.86, + "grad_norm": 2.796875, + "learning_rate": 8.916916899394798e-07, + "loss": 0.7659, + "step": 2173 + }, + { + "epoch": 2.86, + "grad_norm": 2.75, + "learning_rate": 8.896842370307504e-07, + "loss": 0.7715, + "step": 2174 + }, + { + "epoch": 2.86, + "grad_norm": 2.734375, + "learning_rate": 8.876785571155785e-07, + "loss": 0.7576, + "step": 2175 + }, + { + "epoch": 2.87, + "grad_norm": 2.765625, + "learning_rate": 8.856746524022647e-07, + "loss": 0.7742, + "step": 2176 + }, + { + "epoch": 2.87, + "grad_norm": 2.75, + "learning_rate": 8.836725250971503e-07, + "loss": 0.7574, + "step": 2177 + }, + { + "epoch": 2.87, + "grad_norm": 2.75, + "learning_rate": 8.81672177404623e-07, + "loss": 0.7683, + "step": 2178 + }, + { + "epoch": 2.87, + "grad_norm": 2.703125, + "learning_rate": 8.796736115271101e-07, + "loss": 0.7744, + "step": 2179 + }, + { + "epoch": 2.87, + "grad_norm": 2.828125, + "learning_rate": 8.776768296650772e-07, + "loss": 0.7512, + "step": 2180 + }, + { + "epoch": 2.87, + "grad_norm": 2.78125, + "learning_rate": 8.756818340170256e-07, + "loss": 0.7661, + "step": 2181 + }, + { + "epoch": 2.87, + "grad_norm": 2.625, + "learning_rate": 8.736886267794911e-07, + "loss": 0.7414, + "step": 2182 + }, + { + "epoch": 2.87, + "grad_norm": 2.6875, + "learning_rate": 8.716972101470364e-07, + "loss": 0.7602, + "step": 2183 + }, + { + "epoch": 2.88, + "grad_norm": 2.734375, + "learning_rate": 8.697075863122587e-07, + "loss": 0.7387, + "step": 2184 + }, + { + "epoch": 2.88, + "grad_norm": 2.640625, + "learning_rate": 8.677197574657767e-07, + "loss": 0.759, + "step": 2185 + }, + { + "epoch": 2.88, + "grad_norm": 2.71875, + "learning_rate": 8.657337257962344e-07, + "loss": 0.7669, + "step": 2186 + }, + { + "epoch": 2.88, + "grad_norm": 2.84375, + "learning_rate": 8.637494934902973e-07, + "loss": 0.7895, + "step": 2187 + }, + { + "epoch": 2.88, + "grad_norm": 2.703125, + "learning_rate": 8.617670627326503e-07, + "loss": 0.76, + "step": 2188 + }, + { + "epoch": 2.88, + "grad_norm": 2.765625, + "learning_rate": 8.597864357059918e-07, + "loss": 0.7618, + "step": 2189 + }, + { + "epoch": 2.88, + "grad_norm": 2.71875, + "learning_rate": 8.578076145910372e-07, + "loss": 0.7697, + "step": 2190 + }, + { + "epoch": 2.89, + "grad_norm": 2.875, + "learning_rate": 8.558306015665128e-07, + "loss": 0.7941, + "step": 2191 + }, + { + "epoch": 2.89, + "grad_norm": 2.609375, + "learning_rate": 8.538553988091541e-07, + "loss": 0.7495, + "step": 2192 + }, + { + "epoch": 2.89, + "grad_norm": 2.734375, + "learning_rate": 8.518820084937024e-07, + "loss": 0.7756, + "step": 2193 + }, + { + "epoch": 2.89, + "grad_norm": 2.703125, + "learning_rate": 8.49910432792907e-07, + "loss": 0.7257, + "step": 2194 + }, + { + "epoch": 2.89, + "grad_norm": 2.75, + "learning_rate": 8.479406738775142e-07, + "loss": 0.7665, + "step": 2195 + }, + { + "epoch": 2.89, + "grad_norm": 2.71875, + "learning_rate": 8.459727339162732e-07, + "loss": 0.7481, + "step": 2196 + }, + { + "epoch": 2.89, + "grad_norm": 2.75, + "learning_rate": 8.440066150759299e-07, + "loss": 0.7593, + "step": 2197 + }, + { + "epoch": 2.89, + "grad_norm": 2.65625, + "learning_rate": 8.420423195212249e-07, + "loss": 0.7572, + "step": 2198 + }, + { + "epoch": 2.9, + "grad_norm": 2.75, + "learning_rate": 8.400798494148915e-07, + "loss": 0.7587, + "step": 2199 + }, + { + "epoch": 2.9, + "grad_norm": 2.84375, + "learning_rate": 8.381192069176539e-07, + "loss": 0.7664, + "step": 2200 + }, + { + "epoch": 2.9, + "grad_norm": 2.75, + "learning_rate": 8.361603941882204e-07, + "loss": 0.7429, + "step": 2201 + }, + { + "epoch": 2.9, + "grad_norm": 2.8125, + "learning_rate": 8.342034133832904e-07, + "loss": 0.7337, + "step": 2202 + }, + { + "epoch": 2.9, + "grad_norm": 2.65625, + "learning_rate": 8.322482666575424e-07, + "loss": 0.7329, + "step": 2203 + }, + { + "epoch": 2.9, + "grad_norm": 2.71875, + "learning_rate": 8.302949561636364e-07, + "loss": 0.7462, + "step": 2204 + }, + { + "epoch": 2.9, + "grad_norm": 2.765625, + "learning_rate": 8.283434840522112e-07, + "loss": 0.741, + "step": 2205 + }, + { + "epoch": 2.91, + "grad_norm": 2.75, + "learning_rate": 8.263938524718812e-07, + "loss": 0.7775, + "step": 2206 + }, + { + "epoch": 2.91, + "grad_norm": 2.78125, + "learning_rate": 8.244460635692334e-07, + "loss": 0.7577, + "step": 2207 + }, + { + "epoch": 2.91, + "grad_norm": 2.75, + "learning_rate": 8.225001194888271e-07, + "loss": 0.7676, + "step": 2208 + }, + { + "epoch": 2.91, + "grad_norm": 2.75, + "learning_rate": 8.205560223731904e-07, + "loss": 0.7607, + "step": 2209 + }, + { + "epoch": 2.91, + "grad_norm": 2.734375, + "learning_rate": 8.186137743628172e-07, + "loss": 0.7507, + "step": 2210 + }, + { + "epoch": 2.91, + "grad_norm": 2.6875, + "learning_rate": 8.166733775961663e-07, + "loss": 0.747, + "step": 2211 + }, + { + "epoch": 2.91, + "grad_norm": 2.8125, + "learning_rate": 8.147348342096579e-07, + "loss": 0.7689, + "step": 2212 + }, + { + "epoch": 2.91, + "grad_norm": 2.859375, + "learning_rate": 8.127981463376708e-07, + "loss": 0.7526, + "step": 2213 + }, + { + "epoch": 2.92, + "grad_norm": 2.75, + "learning_rate": 8.10863316112542e-07, + "loss": 0.7672, + "step": 2214 + }, + { + "epoch": 2.92, + "grad_norm": 2.828125, + "learning_rate": 8.089303456645631e-07, + "loss": 0.7434, + "step": 2215 + }, + { + "epoch": 2.92, + "grad_norm": 2.6875, + "learning_rate": 8.069992371219767e-07, + "loss": 0.7328, + "step": 2216 + }, + { + "epoch": 2.92, + "grad_norm": 2.75, + "learning_rate": 8.050699926109771e-07, + "loss": 0.7488, + "step": 2217 + }, + { + "epoch": 2.92, + "grad_norm": 2.96875, + "learning_rate": 8.031426142557061e-07, + "loss": 0.7636, + "step": 2218 + }, + { + "epoch": 2.92, + "grad_norm": 2.75, + "learning_rate": 8.012171041782477e-07, + "loss": 0.7454, + "step": 2219 + }, + { + "epoch": 2.92, + "grad_norm": 2.6875, + "learning_rate": 7.992934644986341e-07, + "loss": 0.7532, + "step": 2220 + }, + { + "epoch": 2.93, + "grad_norm": 2.90625, + "learning_rate": 7.973716973348347e-07, + "loss": 0.7431, + "step": 2221 + }, + { + "epoch": 2.93, + "grad_norm": 2.65625, + "learning_rate": 7.954518048027576e-07, + "loss": 0.7641, + "step": 2222 + }, + { + "epoch": 2.93, + "grad_norm": 2.75, + "learning_rate": 7.935337890162475e-07, + "loss": 0.7955, + "step": 2223 + }, + { + "epoch": 2.93, + "grad_norm": 2.703125, + "learning_rate": 7.916176520870836e-07, + "loss": 0.7607, + "step": 2224 + }, + { + "epoch": 2.93, + "grad_norm": 2.765625, + "learning_rate": 7.897033961249733e-07, + "loss": 0.7664, + "step": 2225 + }, + { + "epoch": 2.93, + "grad_norm": 2.796875, + "learning_rate": 7.877910232375563e-07, + "loss": 0.7587, + "step": 2226 + }, + { + "epoch": 2.93, + "grad_norm": 2.703125, + "learning_rate": 7.858805355303975e-07, + "loss": 0.7407, + "step": 2227 + }, + { + "epoch": 2.93, + "grad_norm": 2.734375, + "learning_rate": 7.839719351069866e-07, + "loss": 0.749, + "step": 2228 + }, + { + "epoch": 2.94, + "grad_norm": 2.734375, + "learning_rate": 7.820652240687354e-07, + "loss": 0.743, + "step": 2229 + }, + { + "epoch": 2.94, + "grad_norm": 2.765625, + "learning_rate": 7.80160404514975e-07, + "loss": 0.7608, + "step": 2230 + }, + { + "epoch": 2.94, + "grad_norm": 2.75, + "learning_rate": 7.782574785429542e-07, + "loss": 0.7673, + "step": 2231 + }, + { + "epoch": 2.94, + "grad_norm": 2.796875, + "learning_rate": 7.76356448247837e-07, + "loss": 0.7541, + "step": 2232 + }, + { + "epoch": 2.94, + "grad_norm": 2.625, + "learning_rate": 7.744573157227001e-07, + "loss": 0.7161, + "step": 2233 + }, + { + "epoch": 2.94, + "grad_norm": 3.125, + "learning_rate": 7.725600830585309e-07, + "loss": 0.7508, + "step": 2234 + }, + { + "epoch": 2.94, + "grad_norm": 2.65625, + "learning_rate": 7.706647523442254e-07, + "loss": 0.7446, + "step": 2235 + }, + { + "epoch": 2.95, + "grad_norm": 2.6875, + "learning_rate": 7.687713256665835e-07, + "loss": 0.7442, + "step": 2236 + }, + { + "epoch": 2.95, + "grad_norm": 2.671875, + "learning_rate": 7.66879805110311e-07, + "loss": 0.753, + "step": 2237 + }, + { + "epoch": 2.95, + "grad_norm": 2.796875, + "learning_rate": 7.649901927580136e-07, + "loss": 0.7654, + "step": 2238 + }, + { + "epoch": 2.95, + "grad_norm": 2.75, + "learning_rate": 7.631024906901979e-07, + "loss": 0.7868, + "step": 2239 + }, + { + "epoch": 2.95, + "grad_norm": 2.671875, + "learning_rate": 7.612167009852656e-07, + "loss": 0.767, + "step": 2240 + }, + { + "epoch": 2.95, + "grad_norm": 2.75, + "learning_rate": 7.593328257195134e-07, + "loss": 0.7534, + "step": 2241 + }, + { + "epoch": 2.95, + "grad_norm": 2.78125, + "learning_rate": 7.574508669671288e-07, + "loss": 0.7723, + "step": 2242 + }, + { + "epoch": 2.95, + "grad_norm": 2.796875, + "learning_rate": 7.55570826800191e-07, + "loss": 0.7757, + "step": 2243 + }, + { + "epoch": 2.96, + "grad_norm": 2.734375, + "learning_rate": 7.536927072886663e-07, + "loss": 0.757, + "step": 2244 + }, + { + "epoch": 2.96, + "grad_norm": 2.71875, + "learning_rate": 7.518165105004058e-07, + "loss": 0.772, + "step": 2245 + }, + { + "epoch": 2.96, + "grad_norm": 2.6875, + "learning_rate": 7.499422385011445e-07, + "loss": 0.7527, + "step": 2246 + }, + { + "epoch": 2.96, + "grad_norm": 2.625, + "learning_rate": 7.480698933544972e-07, + "loss": 0.7562, + "step": 2247 + }, + { + "epoch": 2.96, + "grad_norm": 2.859375, + "learning_rate": 7.46199477121958e-07, + "loss": 0.7848, + "step": 2248 + }, + { + "epoch": 2.96, + "grad_norm": 2.671875, + "learning_rate": 7.443309918628966e-07, + "loss": 0.7499, + "step": 2249 + }, + { + "epoch": 2.96, + "grad_norm": 2.703125, + "learning_rate": 7.42464439634557e-07, + "loss": 0.7547, + "step": 2250 + }, + { + "epoch": 2.97, + "grad_norm": 2.671875, + "learning_rate": 7.405998224920549e-07, + "loss": 0.7585, + "step": 2251 + }, + { + "epoch": 2.97, + "grad_norm": 2.703125, + "learning_rate": 7.387371424883754e-07, + "loss": 0.7589, + "step": 2252 + }, + { + "epoch": 2.97, + "grad_norm": 2.71875, + "learning_rate": 7.368764016743712e-07, + "loss": 0.7304, + "step": 2253 + }, + { + "epoch": 2.97, + "grad_norm": 2.734375, + "learning_rate": 7.350176020987585e-07, + "loss": 0.7597, + "step": 2254 + }, + { + "epoch": 2.97, + "grad_norm": 2.84375, + "learning_rate": 7.331607458081169e-07, + "loss": 0.7793, + "step": 2255 + }, + { + "epoch": 2.97, + "grad_norm": 2.6875, + "learning_rate": 7.313058348468866e-07, + "loss": 0.7716, + "step": 2256 + }, + { + "epoch": 2.97, + "grad_norm": 2.65625, + "learning_rate": 7.29452871257367e-07, + "loss": 0.7296, + "step": 2257 + }, + { + "epoch": 2.97, + "grad_norm": 2.75, + "learning_rate": 7.276018570797119e-07, + "loss": 0.766, + "step": 2258 + }, + { + "epoch": 2.98, + "grad_norm": 2.671875, + "learning_rate": 7.257527943519299e-07, + "loss": 0.7454, + "step": 2259 + }, + { + "epoch": 2.98, + "grad_norm": 2.65625, + "learning_rate": 7.239056851098785e-07, + "loss": 0.7622, + "step": 2260 + }, + { + "epoch": 2.98, + "grad_norm": 2.75, + "learning_rate": 7.220605313872672e-07, + "loss": 0.7468, + "step": 2261 + }, + { + "epoch": 2.98, + "grad_norm": 2.65625, + "learning_rate": 7.202173352156511e-07, + "loss": 0.7373, + "step": 2262 + }, + { + "epoch": 2.98, + "grad_norm": 2.625, + "learning_rate": 7.183760986244307e-07, + "loss": 0.7259, + "step": 2263 + }, + { + "epoch": 2.98, + "grad_norm": 2.703125, + "learning_rate": 7.16536823640848e-07, + "loss": 0.755, + "step": 2264 + }, + { + "epoch": 2.98, + "grad_norm": 2.6875, + "learning_rate": 7.146995122899861e-07, + "loss": 0.7958, + "step": 2265 + }, + { + "epoch": 2.99, + "grad_norm": 2.703125, + "learning_rate": 7.128641665947658e-07, + "loss": 0.7552, + "step": 2266 + }, + { + "epoch": 2.99, + "grad_norm": 2.78125, + "learning_rate": 7.110307885759435e-07, + "loss": 0.7518, + "step": 2267 + }, + { + "epoch": 2.99, + "grad_norm": 2.703125, + "learning_rate": 7.091993802521094e-07, + "loss": 0.7552, + "step": 2268 + }, + { + "epoch": 2.99, + "grad_norm": 2.703125, + "learning_rate": 7.073699436396847e-07, + "loss": 0.7494, + "step": 2269 + }, + { + "epoch": 2.99, + "grad_norm": 2.734375, + "learning_rate": 7.055424807529199e-07, + "loss": 0.7877, + "step": 2270 + }, + { + "epoch": 2.99, + "grad_norm": 2.6875, + "learning_rate": 7.037169936038935e-07, + "loss": 0.7652, + "step": 2271 + }, + { + "epoch": 2.99, + "grad_norm": 2.734375, + "learning_rate": 7.018934842025058e-07, + "loss": 0.75, + "step": 2272 + }, + { + "epoch": 2.99, + "grad_norm": 2.6875, + "learning_rate": 7.000719545564819e-07, + "loss": 0.738, + "step": 2273 + }, + { + "epoch": 3.0, + "grad_norm": 2.703125, + "learning_rate": 6.982524066713658e-07, + "loss": 0.7532, + "step": 2274 + }, + { + "epoch": 3.0, + "grad_norm": 2.75, + "learning_rate": 6.964348425505222e-07, + "loss": 0.7637, + "step": 2275 + }, + { + "epoch": 3.0, + "grad_norm": 2.6875, + "learning_rate": 6.946192641951285e-07, + "loss": 0.7558, + "step": 2276 + }, + { + "epoch": 3.0, + "grad_norm": 2.703125, + "learning_rate": 6.928056736041783e-07, + "loss": 0.7313, + "step": 2277 + }, + { + "epoch": 3.0, + "grad_norm": 2.703125, + "learning_rate": 6.90994072774473e-07, + "loss": 0.7377, + "step": 2278 + }, + { + "epoch": 3.0, + "grad_norm": 2.75, + "learning_rate": 6.891844637006268e-07, + "loss": 0.733, + "step": 2279 + }, + { + "epoch": 3.0, + "grad_norm": 2.609375, + "learning_rate": 6.873768483750595e-07, + "loss": 0.7361, + "step": 2280 + }, + { + "epoch": 3.01, + "grad_norm": 2.6875, + "learning_rate": 6.855712287879957e-07, + "loss": 0.7741, + "step": 2281 + }, + { + "epoch": 3.01, + "grad_norm": 2.75, + "learning_rate": 6.837676069274632e-07, + "loss": 0.7451, + "step": 2282 + }, + { + "epoch": 3.01, + "grad_norm": 2.734375, + "learning_rate": 6.819659847792895e-07, + "loss": 0.7632, + "step": 2283 + }, + { + "epoch": 3.01, + "grad_norm": 2.75, + "learning_rate": 6.801663643271012e-07, + "loss": 0.765, + "step": 2284 + }, + { + "epoch": 3.01, + "grad_norm": 2.71875, + "learning_rate": 6.783687475523199e-07, + "loss": 0.7495, + "step": 2285 + }, + { + "epoch": 3.01, + "grad_norm": 2.765625, + "learning_rate": 6.765731364341622e-07, + "loss": 0.7506, + "step": 2286 + }, + { + "epoch": 3.01, + "grad_norm": 2.703125, + "learning_rate": 6.747795329496359e-07, + "loss": 0.7544, + "step": 2287 + }, + { + "epoch": 3.01, + "grad_norm": 2.703125, + "learning_rate": 6.729879390735391e-07, + "loss": 0.7334, + "step": 2288 + }, + { + "epoch": 3.0, + "grad_norm": 2.859375, + "learning_rate": 6.711983567784552e-07, + "loss": 0.7458, + "step": 2289 + }, + { + "epoch": 3.0, + "grad_norm": 2.734375, + "learning_rate": 6.69410788034755e-07, + "loss": 0.736, + "step": 2290 + }, + { + "epoch": 3.0, + "grad_norm": 2.8125, + "learning_rate": 6.676252348105916e-07, + "loss": 0.7523, + "step": 2291 + }, + { + "epoch": 3.0, + "grad_norm": 2.9375, + "learning_rate": 6.658416990718986e-07, + "loss": 0.743, + "step": 2292 + }, + { + "epoch": 3.01, + "grad_norm": 2.671875, + "learning_rate": 6.640601827823878e-07, + "loss": 0.7135, + "step": 2293 + }, + { + "epoch": 3.01, + "grad_norm": 2.859375, + "learning_rate": 6.622806879035512e-07, + "loss": 0.7385, + "step": 2294 + }, + { + "epoch": 3.01, + "grad_norm": 2.828125, + "learning_rate": 6.605032163946496e-07, + "loss": 0.7395, + "step": 2295 + }, + { + "epoch": 3.01, + "grad_norm": 2.875, + "learning_rate": 6.587277702127196e-07, + "loss": 0.7255, + "step": 2296 + }, + { + "epoch": 3.01, + "grad_norm": 2.84375, + "learning_rate": 6.56954351312567e-07, + "loss": 0.754, + "step": 2297 + }, + { + "epoch": 3.01, + "grad_norm": 2.6875, + "learning_rate": 6.551829616467656e-07, + "loss": 0.7245, + "step": 2298 + }, + { + "epoch": 3.01, + "grad_norm": 2.75, + "learning_rate": 6.534136031656552e-07, + "loss": 0.7448, + "step": 2299 + }, + { + "epoch": 3.02, + "grad_norm": 2.6875, + "learning_rate": 6.516462778173383e-07, + "loss": 0.7134, + "step": 2300 + }, + { + "epoch": 3.02, + "grad_norm": 2.75, + "learning_rate": 6.498809875476799e-07, + "loss": 0.7394, + "step": 2301 + }, + { + "epoch": 3.02, + "grad_norm": 2.8125, + "learning_rate": 6.481177343003043e-07, + "loss": 0.7496, + "step": 2302 + }, + { + "epoch": 3.02, + "grad_norm": 2.828125, + "learning_rate": 6.463565200165919e-07, + "loss": 0.75, + "step": 2303 + }, + { + "epoch": 3.02, + "grad_norm": 2.78125, + "learning_rate": 6.445973466356794e-07, + "loss": 0.7282, + "step": 2304 + }, + { + "epoch": 3.02, + "grad_norm": 2.6875, + "learning_rate": 6.428402160944558e-07, + "loss": 0.7266, + "step": 2305 + }, + { + "epoch": 3.02, + "grad_norm": 2.8125, + "learning_rate": 6.410851303275617e-07, + "loss": 0.7538, + "step": 2306 + }, + { + "epoch": 3.02, + "grad_norm": 2.703125, + "learning_rate": 6.393320912673842e-07, + "loss": 0.7137, + "step": 2307 + }, + { + "epoch": 3.03, + "grad_norm": 2.671875, + "learning_rate": 6.375811008440591e-07, + "loss": 0.7237, + "step": 2308 + }, + { + "epoch": 3.03, + "grad_norm": 2.90625, + "learning_rate": 6.358321609854659e-07, + "loss": 0.7391, + "step": 2309 + }, + { + "epoch": 3.03, + "grad_norm": 2.84375, + "learning_rate": 6.340852736172265e-07, + "loss": 0.7325, + "step": 2310 + }, + { + "epoch": 3.03, + "grad_norm": 2.78125, + "learning_rate": 6.323404406627018e-07, + "loss": 0.7259, + "step": 2311 + }, + { + "epoch": 3.03, + "grad_norm": 2.75, + "learning_rate": 6.305976640429945e-07, + "loss": 0.7299, + "step": 2312 + }, + { + "epoch": 3.03, + "grad_norm": 2.796875, + "learning_rate": 6.288569456769377e-07, + "loss": 0.7378, + "step": 2313 + }, + { + "epoch": 3.03, + "grad_norm": 2.75, + "learning_rate": 6.271182874811024e-07, + "loss": 0.7302, + "step": 2314 + }, + { + "epoch": 3.04, + "grad_norm": 2.796875, + "learning_rate": 6.253816913697896e-07, + "loss": 0.7195, + "step": 2315 + }, + { + "epoch": 3.04, + "grad_norm": 2.65625, + "learning_rate": 6.236471592550305e-07, + "loss": 0.7367, + "step": 2316 + }, + { + "epoch": 3.04, + "grad_norm": 2.8125, + "learning_rate": 6.219146930465838e-07, + "loss": 0.7477, + "step": 2317 + }, + { + "epoch": 3.04, + "grad_norm": 2.71875, + "learning_rate": 6.201842946519343e-07, + "loss": 0.7284, + "step": 2318 + }, + { + "epoch": 3.04, + "grad_norm": 2.84375, + "learning_rate": 6.184559659762865e-07, + "loss": 0.741, + "step": 2319 + }, + { + "epoch": 3.04, + "grad_norm": 2.71875, + "learning_rate": 6.167297089225713e-07, + "loss": 0.7472, + "step": 2320 + }, + { + "epoch": 3.04, + "grad_norm": 2.734375, + "learning_rate": 6.150055253914353e-07, + "loss": 0.724, + "step": 2321 + }, + { + "epoch": 3.04, + "grad_norm": 2.75, + "learning_rate": 6.132834172812432e-07, + "loss": 0.7345, + "step": 2322 + }, + { + "epoch": 3.05, + "grad_norm": 2.765625, + "learning_rate": 6.115633864880743e-07, + "loss": 0.7521, + "step": 2323 + }, + { + "epoch": 3.05, + "grad_norm": 2.796875, + "learning_rate": 6.098454349057218e-07, + "loss": 0.7718, + "step": 2324 + }, + { + "epoch": 3.05, + "grad_norm": 2.84375, + "learning_rate": 6.081295644256868e-07, + "loss": 0.7365, + "step": 2325 + }, + { + "epoch": 3.05, + "grad_norm": 2.75, + "learning_rate": 6.064157769371823e-07, + "loss": 0.7374, + "step": 2326 + }, + { + "epoch": 3.05, + "grad_norm": 2.65625, + "learning_rate": 6.04704074327126e-07, + "loss": 0.7485, + "step": 2327 + }, + { + "epoch": 3.05, + "grad_norm": 2.75, + "learning_rate": 6.029944584801408e-07, + "loss": 0.7654, + "step": 2328 + }, + { + "epoch": 3.05, + "grad_norm": 2.78125, + "learning_rate": 6.012869312785513e-07, + "loss": 0.7542, + "step": 2329 + }, + { + "epoch": 3.06, + "grad_norm": 2.75, + "learning_rate": 5.995814946023848e-07, + "loss": 0.7427, + "step": 2330 + }, + { + "epoch": 3.06, + "grad_norm": 2.78125, + "learning_rate": 5.978781503293634e-07, + "loss": 0.7387, + "step": 2331 + }, + { + "epoch": 3.06, + "grad_norm": 2.78125, + "learning_rate": 5.961769003349077e-07, + "loss": 0.7572, + "step": 2332 + }, + { + "epoch": 3.06, + "grad_norm": 2.8125, + "learning_rate": 5.944777464921317e-07, + "loss": 0.7625, + "step": 2333 + }, + { + "epoch": 3.06, + "grad_norm": 2.671875, + "learning_rate": 5.92780690671842e-07, + "loss": 0.7181, + "step": 2334 + }, + { + "epoch": 3.06, + "grad_norm": 2.65625, + "learning_rate": 5.910857347425345e-07, + "loss": 0.7116, + "step": 2335 + }, + { + "epoch": 3.06, + "grad_norm": 2.765625, + "learning_rate": 5.893928805703943e-07, + "loss": 0.7554, + "step": 2336 + }, + { + "epoch": 3.06, + "grad_norm": 2.828125, + "learning_rate": 5.877021300192895e-07, + "loss": 0.7872, + "step": 2337 + }, + { + "epoch": 3.07, + "grad_norm": 2.734375, + "learning_rate": 5.860134849507765e-07, + "loss": 0.7269, + "step": 2338 + }, + { + "epoch": 3.07, + "grad_norm": 2.78125, + "learning_rate": 5.843269472240901e-07, + "loss": 0.7573, + "step": 2339 + }, + { + "epoch": 3.07, + "grad_norm": 2.75, + "learning_rate": 5.82642518696146e-07, + "loss": 0.735, + "step": 2340 + }, + { + "epoch": 3.07, + "grad_norm": 2.78125, + "learning_rate": 5.809602012215385e-07, + "loss": 0.75, + "step": 2341 + }, + { + "epoch": 3.07, + "grad_norm": 2.71875, + "learning_rate": 5.79279996652535e-07, + "loss": 0.6935, + "step": 2342 + }, + { + "epoch": 3.07, + "grad_norm": 2.796875, + "learning_rate": 5.776019068390792e-07, + "loss": 0.7396, + "step": 2343 + }, + { + "epoch": 3.07, + "grad_norm": 2.71875, + "learning_rate": 5.759259336287851e-07, + "loss": 0.7488, + "step": 2344 + }, + { + "epoch": 3.08, + "grad_norm": 2.78125, + "learning_rate": 5.742520788669367e-07, + "loss": 0.7453, + "step": 2345 + }, + { + "epoch": 3.08, + "grad_norm": 2.78125, + "learning_rate": 5.725803443964853e-07, + "loss": 0.7617, + "step": 2346 + }, + { + "epoch": 3.08, + "grad_norm": 2.78125, + "learning_rate": 5.709107320580479e-07, + "loss": 0.7488, + "step": 2347 + }, + { + "epoch": 3.08, + "grad_norm": 2.6875, + "learning_rate": 5.692432436899051e-07, + "loss": 0.7188, + "step": 2348 + }, + { + "epoch": 3.08, + "grad_norm": 2.734375, + "learning_rate": 5.675778811279983e-07, + "loss": 0.7247, + "step": 2349 + }, + { + "epoch": 3.08, + "grad_norm": 2.765625, + "learning_rate": 5.659146462059292e-07, + "loss": 0.7439, + "step": 2350 + }, + { + "epoch": 3.08, + "grad_norm": 2.671875, + "learning_rate": 5.642535407549565e-07, + "loss": 0.7298, + "step": 2351 + }, + { + "epoch": 3.08, + "grad_norm": 2.71875, + "learning_rate": 5.625945666039942e-07, + "loss": 0.7119, + "step": 2352 + }, + { + "epoch": 3.09, + "grad_norm": 2.640625, + "learning_rate": 5.609377255796106e-07, + "loss": 0.7445, + "step": 2353 + }, + { + "epoch": 3.09, + "grad_norm": 2.734375, + "learning_rate": 5.592830195060234e-07, + "loss": 0.7476, + "step": 2354 + }, + { + "epoch": 3.09, + "grad_norm": 2.65625, + "learning_rate": 5.576304502051011e-07, + "loss": 0.7004, + "step": 2355 + }, + { + "epoch": 3.09, + "grad_norm": 2.734375, + "learning_rate": 5.559800194963591e-07, + "loss": 0.7178, + "step": 2356 + }, + { + "epoch": 3.09, + "grad_norm": 2.71875, + "learning_rate": 5.543317291969597e-07, + "loss": 0.7292, + "step": 2357 + }, + { + "epoch": 3.09, + "grad_norm": 2.8125, + "learning_rate": 5.526855811217064e-07, + "loss": 0.7513, + "step": 2358 + }, + { + "epoch": 3.09, + "grad_norm": 2.734375, + "learning_rate": 5.510415770830457e-07, + "loss": 0.7337, + "step": 2359 + }, + { + "epoch": 3.1, + "grad_norm": 2.78125, + "learning_rate": 5.493997188910613e-07, + "loss": 0.7513, + "step": 2360 + }, + { + "epoch": 3.1, + "grad_norm": 2.6875, + "learning_rate": 5.477600083534764e-07, + "loss": 0.7437, + "step": 2361 + }, + { + "epoch": 3.1, + "grad_norm": 2.75, + "learning_rate": 5.46122447275649e-07, + "loss": 0.741, + "step": 2362 + }, + { + "epoch": 3.1, + "grad_norm": 2.78125, + "learning_rate": 5.444870374605704e-07, + "loss": 0.7225, + "step": 2363 + }, + { + "epoch": 3.1, + "grad_norm": 2.8125, + "learning_rate": 5.428537807088627e-07, + "loss": 0.6944, + "step": 2364 + }, + { + "epoch": 3.1, + "grad_norm": 2.75, + "learning_rate": 5.412226788187788e-07, + "loss": 0.7508, + "step": 2365 + }, + { + "epoch": 3.1, + "grad_norm": 2.75, + "learning_rate": 5.395937335861975e-07, + "loss": 0.725, + "step": 2366 + }, + { + "epoch": 3.1, + "grad_norm": 2.75, + "learning_rate": 5.379669468046245e-07, + "loss": 0.7428, + "step": 2367 + }, + { + "epoch": 3.11, + "grad_norm": 2.734375, + "learning_rate": 5.363423202651876e-07, + "loss": 0.7218, + "step": 2368 + }, + { + "epoch": 3.11, + "grad_norm": 2.765625, + "learning_rate": 5.347198557566374e-07, + "loss": 0.7236, + "step": 2369 + }, + { + "epoch": 3.11, + "grad_norm": 2.71875, + "learning_rate": 5.330995550653434e-07, + "loss": 0.7174, + "step": 2370 + }, + { + "epoch": 3.11, + "grad_norm": 2.703125, + "learning_rate": 5.314814199752935e-07, + "loss": 0.7339, + "step": 2371 + }, + { + "epoch": 3.11, + "grad_norm": 2.78125, + "learning_rate": 5.298654522680888e-07, + "loss": 0.7511, + "step": 2372 + }, + { + "epoch": 3.11, + "grad_norm": 2.890625, + "learning_rate": 5.282516537229468e-07, + "loss": 0.7366, + "step": 2373 + }, + { + "epoch": 3.11, + "grad_norm": 2.796875, + "learning_rate": 5.266400261166951e-07, + "loss": 0.755, + "step": 2374 + }, + { + "epoch": 3.12, + "grad_norm": 2.765625, + "learning_rate": 5.250305712237724e-07, + "loss": 0.7411, + "step": 2375 + }, + { + "epoch": 3.12, + "grad_norm": 2.796875, + "learning_rate": 5.234232908162243e-07, + "loss": 0.7681, + "step": 2376 + }, + { + "epoch": 3.12, + "grad_norm": 2.765625, + "learning_rate": 5.218181866637029e-07, + "loss": 0.7097, + "step": 2377 + }, + { + "epoch": 3.12, + "grad_norm": 2.78125, + "learning_rate": 5.202152605334623e-07, + "loss": 0.7398, + "step": 2378 + }, + { + "epoch": 3.12, + "grad_norm": 2.625, + "learning_rate": 5.186145141903606e-07, + "loss": 0.7076, + "step": 2379 + }, + { + "epoch": 3.12, + "grad_norm": 2.828125, + "learning_rate": 5.170159493968549e-07, + "loss": 0.7502, + "step": 2380 + }, + { + "epoch": 3.12, + "grad_norm": 2.78125, + "learning_rate": 5.154195679130012e-07, + "loss": 0.7569, + "step": 2381 + }, + { + "epoch": 3.12, + "grad_norm": 2.921875, + "learning_rate": 5.138253714964509e-07, + "loss": 0.7543, + "step": 2382 + }, + { + "epoch": 3.13, + "grad_norm": 2.90625, + "learning_rate": 5.122333619024494e-07, + "loss": 0.7527, + "step": 2383 + }, + { + "epoch": 3.13, + "grad_norm": 2.8125, + "learning_rate": 5.106435408838354e-07, + "loss": 0.7354, + "step": 2384 + }, + { + "epoch": 3.13, + "grad_norm": 2.703125, + "learning_rate": 5.090559101910369e-07, + "loss": 0.7404, + "step": 2385 + }, + { + "epoch": 3.13, + "grad_norm": 2.8125, + "learning_rate": 5.074704715720711e-07, + "loss": 0.7498, + "step": 2386 + }, + { + "epoch": 3.13, + "grad_norm": 2.828125, + "learning_rate": 5.058872267725407e-07, + "loss": 0.7304, + "step": 2387 + }, + { + "epoch": 3.13, + "grad_norm": 2.8125, + "learning_rate": 5.043061775356336e-07, + "loss": 0.7505, + "step": 2388 + }, + { + "epoch": 3.13, + "grad_norm": 2.765625, + "learning_rate": 5.027273256021212e-07, + "loss": 0.7263, + "step": 2389 + }, + { + "epoch": 3.14, + "grad_norm": 2.8125, + "learning_rate": 5.011506727103536e-07, + "loss": 0.7544, + "step": 2390 + }, + { + "epoch": 3.14, + "grad_norm": 2.75, + "learning_rate": 4.995762205962607e-07, + "loss": 0.756, + "step": 2391 + }, + { + "epoch": 3.14, + "grad_norm": 2.71875, + "learning_rate": 4.980039709933492e-07, + "loss": 0.7044, + "step": 2392 + }, + { + "epoch": 3.14, + "grad_norm": 2.6875, + "learning_rate": 4.964339256327022e-07, + "loss": 0.7123, + "step": 2393 + }, + { + "epoch": 3.14, + "grad_norm": 2.71875, + "learning_rate": 4.948660862429741e-07, + "loss": 0.7342, + "step": 2394 + }, + { + "epoch": 3.14, + "grad_norm": 2.765625, + "learning_rate": 4.933004545503917e-07, + "loss": 0.7532, + "step": 2395 + }, + { + "epoch": 3.14, + "grad_norm": 2.734375, + "learning_rate": 4.917370322787488e-07, + "loss": 0.7435, + "step": 2396 + }, + { + "epoch": 3.14, + "grad_norm": 2.8125, + "learning_rate": 4.901758211494089e-07, + "loss": 0.7251, + "step": 2397 + }, + { + "epoch": 3.15, + "grad_norm": 2.859375, + "learning_rate": 4.886168228813007e-07, + "loss": 0.7306, + "step": 2398 + }, + { + "epoch": 3.15, + "grad_norm": 2.6875, + "learning_rate": 4.870600391909156e-07, + "loss": 0.7276, + "step": 2399 + }, + { + "epoch": 3.15, + "grad_norm": 2.6875, + "learning_rate": 4.855054717923072e-07, + "loss": 0.7128, + "step": 2400 + }, + { + "epoch": 3.15, + "grad_norm": 2.78125, + "learning_rate": 4.839531223970889e-07, + "loss": 0.7291, + "step": 2401 + }, + { + "epoch": 3.15, + "grad_norm": 2.71875, + "learning_rate": 4.824029927144321e-07, + "loss": 0.7587, + "step": 2402 + }, + { + "epoch": 3.15, + "grad_norm": 2.84375, + "learning_rate": 4.80855084451064e-07, + "loss": 0.7431, + "step": 2403 + }, + { + "epoch": 3.15, + "grad_norm": 2.6875, + "learning_rate": 4.793093993112663e-07, + "loss": 0.7269, + "step": 2404 + }, + { + "epoch": 3.16, + "grad_norm": 2.796875, + "learning_rate": 4.777659389968728e-07, + "loss": 0.7415, + "step": 2405 + }, + { + "epoch": 3.16, + "grad_norm": 2.859375, + "learning_rate": 4.76224705207268e-07, + "loss": 0.7523, + "step": 2406 + }, + { + "epoch": 3.16, + "grad_norm": 2.71875, + "learning_rate": 4.7468569963938365e-07, + "loss": 0.7234, + "step": 2407 + }, + { + "epoch": 3.16, + "grad_norm": 2.734375, + "learning_rate": 4.731489239876999e-07, + "loss": 0.7303, + "step": 2408 + }, + { + "epoch": 3.16, + "grad_norm": 2.734375, + "learning_rate": 4.71614379944241e-07, + "loss": 0.7389, + "step": 2409 + }, + { + "epoch": 3.16, + "grad_norm": 2.8125, + "learning_rate": 4.700820691985739e-07, + "loss": 0.7597, + "step": 2410 + }, + { + "epoch": 3.16, + "grad_norm": 2.703125, + "learning_rate": 4.685519934378066e-07, + "loss": 0.7248, + "step": 2411 + }, + { + "epoch": 3.16, + "grad_norm": 2.734375, + "learning_rate": 4.670241543465881e-07, + "loss": 0.7269, + "step": 2412 + }, + { + "epoch": 3.17, + "grad_norm": 2.75, + "learning_rate": 4.654985536071019e-07, + "loss": 0.7445, + "step": 2413 + }, + { + "epoch": 3.17, + "grad_norm": 2.734375, + "learning_rate": 4.639751928990688e-07, + "loss": 0.7453, + "step": 2414 + }, + { + "epoch": 3.17, + "grad_norm": 2.734375, + "learning_rate": 4.6245407389974293e-07, + "loss": 0.7385, + "step": 2415 + }, + { + "epoch": 3.17, + "grad_norm": 2.734375, + "learning_rate": 4.6093519828391025e-07, + "loss": 0.7419, + "step": 2416 + }, + { + "epoch": 3.17, + "grad_norm": 2.765625, + "learning_rate": 4.594185677238866e-07, + "loss": 0.7306, + "step": 2417 + }, + { + "epoch": 3.17, + "grad_norm": 2.78125, + "learning_rate": 4.5790418388951596e-07, + "loss": 0.7604, + "step": 2418 + }, + { + "epoch": 3.17, + "grad_norm": 2.6875, + "learning_rate": 4.5639204844816904e-07, + "loss": 0.7203, + "step": 2419 + }, + { + "epoch": 3.18, + "grad_norm": 2.796875, + "learning_rate": 4.5488216306474044e-07, + "loss": 0.7542, + "step": 2420 + }, + { + "epoch": 3.18, + "grad_norm": 2.796875, + "learning_rate": 4.5337452940164757e-07, + "loss": 0.7379, + "step": 2421 + }, + { + "epoch": 3.18, + "grad_norm": 2.703125, + "learning_rate": 4.51869149118829e-07, + "loss": 0.7314, + "step": 2422 + }, + { + "epoch": 3.18, + "grad_norm": 2.703125, + "learning_rate": 4.5036602387374143e-07, + "loss": 0.7354, + "step": 2423 + }, + { + "epoch": 3.18, + "grad_norm": 2.71875, + "learning_rate": 4.488651553213605e-07, + "loss": 0.7207, + "step": 2424 + }, + { + "epoch": 3.18, + "grad_norm": 2.78125, + "learning_rate": 4.4736654511417456e-07, + "loss": 0.7549, + "step": 2425 + }, + { + "epoch": 3.18, + "grad_norm": 2.78125, + "learning_rate": 4.458701949021871e-07, + "loss": 0.733, + "step": 2426 + }, + { + "epoch": 3.18, + "grad_norm": 2.75, + "learning_rate": 4.443761063329133e-07, + "loss": 0.7572, + "step": 2427 + }, + { + "epoch": 3.19, + "grad_norm": 2.640625, + "learning_rate": 4.428842810513784e-07, + "loss": 0.7242, + "step": 2428 + }, + { + "epoch": 3.19, + "grad_norm": 2.6875, + "learning_rate": 4.4139472070011374e-07, + "loss": 0.7304, + "step": 2429 + }, + { + "epoch": 3.19, + "grad_norm": 2.796875, + "learning_rate": 4.399074269191614e-07, + "loss": 0.7434, + "step": 2430 + }, + { + "epoch": 3.19, + "grad_norm": 2.796875, + "learning_rate": 4.3842240134606225e-07, + "loss": 0.7429, + "step": 2431 + }, + { + "epoch": 3.19, + "grad_norm": 2.78125, + "learning_rate": 4.369396456158639e-07, + "loss": 0.7203, + "step": 2432 + }, + { + "epoch": 3.19, + "grad_norm": 2.671875, + "learning_rate": 4.3545916136111284e-07, + "loss": 0.7554, + "step": 2433 + }, + { + "epoch": 3.19, + "grad_norm": 2.78125, + "learning_rate": 4.3398095021185557e-07, + "loss": 0.7229, + "step": 2434 + }, + { + "epoch": 3.2, + "grad_norm": 2.78125, + "learning_rate": 4.325050137956355e-07, + "loss": 0.731, + "step": 2435 + }, + { + "epoch": 3.2, + "grad_norm": 2.828125, + "learning_rate": 4.3103135373749216e-07, + "loss": 0.7836, + "step": 2436 + }, + { + "epoch": 3.2, + "grad_norm": 2.8125, + "learning_rate": 4.2955997165995634e-07, + "loss": 0.7388, + "step": 2437 + }, + { + "epoch": 3.2, + "grad_norm": 2.859375, + "learning_rate": 4.2809086918305403e-07, + "loss": 0.7573, + "step": 2438 + }, + { + "epoch": 3.2, + "grad_norm": 2.84375, + "learning_rate": 4.266240479242997e-07, + "loss": 0.7542, + "step": 2439 + }, + { + "epoch": 3.2, + "grad_norm": 2.703125, + "learning_rate": 4.251595094986957e-07, + "loss": 0.7314, + "step": 2440 + }, + { + "epoch": 3.2, + "grad_norm": 2.796875, + "learning_rate": 4.236972555187319e-07, + "loss": 0.7696, + "step": 2441 + }, + { + "epoch": 3.2, + "grad_norm": 2.75, + "learning_rate": 4.222372875943831e-07, + "loss": 0.7625, + "step": 2442 + }, + { + "epoch": 3.21, + "grad_norm": 2.796875, + "learning_rate": 4.207796073331052e-07, + "loss": 0.754, + "step": 2443 + }, + { + "epoch": 3.21, + "grad_norm": 2.703125, + "learning_rate": 4.1932421633983726e-07, + "loss": 0.7264, + "step": 2444 + }, + { + "epoch": 3.21, + "grad_norm": 2.8125, + "learning_rate": 4.178711162169974e-07, + "loss": 0.7295, + "step": 2445 + }, + { + "epoch": 3.21, + "grad_norm": 2.8125, + "learning_rate": 4.1642030856448104e-07, + "loss": 0.7379, + "step": 2446 + }, + { + "epoch": 3.21, + "grad_norm": 2.75, + "learning_rate": 4.1497179497965915e-07, + "loss": 0.7712, + "step": 2447 + }, + { + "epoch": 3.21, + "grad_norm": 2.734375, + "learning_rate": 4.1352557705737957e-07, + "loss": 0.7512, + "step": 2448 + }, + { + "epoch": 3.21, + "grad_norm": 2.6875, + "learning_rate": 4.120816563899585e-07, + "loss": 0.7678, + "step": 2449 + }, + { + "epoch": 3.22, + "grad_norm": 2.734375, + "learning_rate": 4.1064003456718517e-07, + "loss": 0.7376, + "step": 2450 + }, + { + "epoch": 3.22, + "grad_norm": 2.75, + "learning_rate": 4.092007131763176e-07, + "loss": 0.7255, + "step": 2451 + }, + { + "epoch": 3.22, + "grad_norm": 2.671875, + "learning_rate": 4.077636938020807e-07, + "loss": 0.7056, + "step": 2452 + }, + { + "epoch": 3.22, + "grad_norm": 2.671875, + "learning_rate": 4.063289780266652e-07, + "loss": 0.7429, + "step": 2453 + }, + { + "epoch": 3.22, + "grad_norm": 2.859375, + "learning_rate": 4.048965674297245e-07, + "loss": 0.7441, + "step": 2454 + }, + { + "epoch": 3.22, + "grad_norm": 2.640625, + "learning_rate": 4.034664635883742e-07, + "loss": 0.6863, + "step": 2455 + }, + { + "epoch": 3.22, + "grad_norm": 2.6875, + "learning_rate": 4.0203866807719176e-07, + "loss": 0.7189, + "step": 2456 + }, + { + "epoch": 3.22, + "grad_norm": 2.8125, + "learning_rate": 4.0061318246821105e-07, + "loss": 0.7478, + "step": 2457 + }, + { + "epoch": 3.23, + "grad_norm": 2.78125, + "learning_rate": 3.991900083309241e-07, + "loss": 0.7417, + "step": 2458 + }, + { + "epoch": 3.23, + "grad_norm": 2.71875, + "learning_rate": 3.9776914723227753e-07, + "loss": 0.7149, + "step": 2459 + }, + { + "epoch": 3.23, + "grad_norm": 2.765625, + "learning_rate": 3.9635060073666993e-07, + "loss": 0.7243, + "step": 2460 + }, + { + "epoch": 3.23, + "grad_norm": 2.75, + "learning_rate": 3.9493437040595327e-07, + "loss": 0.7261, + "step": 2461 + }, + { + "epoch": 3.23, + "grad_norm": 2.890625, + "learning_rate": 3.935204577994289e-07, + "loss": 0.7668, + "step": 2462 + }, + { + "epoch": 3.23, + "grad_norm": 2.71875, + "learning_rate": 3.921088644738458e-07, + "loss": 0.7382, + "step": 2463 + }, + { + "epoch": 3.23, + "grad_norm": 2.75, + "learning_rate": 3.906995919833997e-07, + "loss": 0.743, + "step": 2464 + }, + { + "epoch": 3.24, + "grad_norm": 2.78125, + "learning_rate": 3.892926418797313e-07, + "loss": 0.7242, + "step": 2465 + }, + { + "epoch": 3.24, + "grad_norm": 2.765625, + "learning_rate": 3.878880157119241e-07, + "loss": 0.7465, + "step": 2466 + }, + { + "epoch": 3.24, + "grad_norm": 2.8125, + "learning_rate": 3.8648571502650253e-07, + "loss": 0.741, + "step": 2467 + }, + { + "epoch": 3.24, + "grad_norm": 2.75, + "learning_rate": 3.8508574136743085e-07, + "loss": 0.7638, + "step": 2468 + }, + { + "epoch": 3.24, + "grad_norm": 2.84375, + "learning_rate": 3.8368809627611135e-07, + "loss": 0.7375, + "step": 2469 + }, + { + "epoch": 3.24, + "grad_norm": 2.828125, + "learning_rate": 3.8229278129138293e-07, + "loss": 0.7364, + "step": 2470 + }, + { + "epoch": 3.24, + "grad_norm": 2.90625, + "learning_rate": 3.8089979794951825e-07, + "loss": 0.7208, + "step": 2471 + }, + { + "epoch": 3.24, + "grad_norm": 2.84375, + "learning_rate": 3.795091477842225e-07, + "loss": 0.7375, + "step": 2472 + }, + { + "epoch": 3.25, + "grad_norm": 2.765625, + "learning_rate": 3.7812083232663306e-07, + "loss": 0.7208, + "step": 2473 + }, + { + "epoch": 3.25, + "grad_norm": 2.859375, + "learning_rate": 3.7673485310531525e-07, + "loss": 0.7518, + "step": 2474 + }, + { + "epoch": 3.25, + "grad_norm": 2.796875, + "learning_rate": 3.753512116462643e-07, + "loss": 0.7345, + "step": 2475 + }, + { + "epoch": 3.25, + "grad_norm": 2.859375, + "learning_rate": 3.739699094729002e-07, + "loss": 0.7367, + "step": 2476 + }, + { + "epoch": 3.25, + "grad_norm": 2.75, + "learning_rate": 3.7259094810606745e-07, + "loss": 0.7352, + "step": 2477 + }, + { + "epoch": 3.25, + "grad_norm": 2.734375, + "learning_rate": 3.7121432906403244e-07, + "loss": 0.7244, + "step": 2478 + }, + { + "epoch": 3.25, + "grad_norm": 2.640625, + "learning_rate": 3.6984005386248385e-07, + "loss": 0.7126, + "step": 2479 + }, + { + "epoch": 3.26, + "grad_norm": 2.796875, + "learning_rate": 3.684681240145296e-07, + "loss": 0.7566, + "step": 2480 + }, + { + "epoch": 3.26, + "grad_norm": 2.6875, + "learning_rate": 3.6709854103069457e-07, + "loss": 0.7272, + "step": 2481 + }, + { + "epoch": 3.26, + "grad_norm": 2.796875, + "learning_rate": 3.6573130641892053e-07, + "loss": 0.7387, + "step": 2482 + }, + { + "epoch": 3.26, + "grad_norm": 2.8125, + "learning_rate": 3.6436642168456325e-07, + "loss": 0.7194, + "step": 2483 + }, + { + "epoch": 3.26, + "grad_norm": 2.859375, + "learning_rate": 3.63003888330391e-07, + "loss": 0.7669, + "step": 2484 + }, + { + "epoch": 3.26, + "grad_norm": 2.71875, + "learning_rate": 3.6164370785658346e-07, + "loss": 0.7403, + "step": 2485 + }, + { + "epoch": 3.26, + "grad_norm": 2.734375, + "learning_rate": 3.602858817607299e-07, + "loss": 0.739, + "step": 2486 + }, + { + "epoch": 3.26, + "grad_norm": 2.75, + "learning_rate": 3.5893041153782697e-07, + "loss": 0.7241, + "step": 2487 + }, + { + "epoch": 3.27, + "grad_norm": 2.75, + "learning_rate": 3.575772986802775e-07, + "loss": 0.7532, + "step": 2488 + }, + { + "epoch": 3.27, + "grad_norm": 2.75, + "learning_rate": 3.562265446778898e-07, + "loss": 0.7488, + "step": 2489 + }, + { + "epoch": 3.27, + "grad_norm": 2.75, + "learning_rate": 3.5487815101787253e-07, + "loss": 0.7359, + "step": 2490 + }, + { + "epoch": 3.27, + "grad_norm": 2.78125, + "learning_rate": 3.5353211918483813e-07, + "loss": 0.7444, + "step": 2491 + }, + { + "epoch": 3.27, + "grad_norm": 2.8125, + "learning_rate": 3.52188450660797e-07, + "loss": 0.7389, + "step": 2492 + }, + { + "epoch": 3.27, + "grad_norm": 2.71875, + "learning_rate": 3.50847146925159e-07, + "loss": 0.7365, + "step": 2493 + }, + { + "epoch": 3.27, + "grad_norm": 2.828125, + "learning_rate": 3.4950820945472945e-07, + "loss": 0.714, + "step": 2494 + }, + { + "epoch": 3.28, + "grad_norm": 2.78125, + "learning_rate": 3.481716397237084e-07, + "loss": 0.7675, + "step": 2495 + }, + { + "epoch": 3.28, + "grad_norm": 2.75, + "learning_rate": 3.46837439203688e-07, + "loss": 0.7285, + "step": 2496 + }, + { + "epoch": 3.28, + "grad_norm": 2.75, + "learning_rate": 3.455056093636533e-07, + "loss": 0.7343, + "step": 2497 + }, + { + "epoch": 3.28, + "grad_norm": 2.734375, + "learning_rate": 3.441761516699788e-07, + "loss": 0.7157, + "step": 2498 + }, + { + "epoch": 3.28, + "grad_norm": 2.703125, + "learning_rate": 3.4284906758642683e-07, + "loss": 0.726, + "step": 2499 + }, + { + "epoch": 3.28, + "grad_norm": 2.84375, + "learning_rate": 3.4152435857414676e-07, + "loss": 0.7455, + "step": 2500 + }, + { + "epoch": 3.28, + "grad_norm": 2.828125, + "learning_rate": 3.4020202609167245e-07, + "loss": 0.7319, + "step": 2501 + }, + { + "epoch": 3.28, + "grad_norm": 2.765625, + "learning_rate": 3.3888207159492177e-07, + "loss": 0.7075, + "step": 2502 + }, + { + "epoch": 3.29, + "grad_norm": 2.78125, + "learning_rate": 3.375644965371941e-07, + "loss": 0.7237, + "step": 2503 + }, + { + "epoch": 3.29, + "grad_norm": 2.984375, + "learning_rate": 3.362493023691685e-07, + "loss": 0.738, + "step": 2504 + }, + { + "epoch": 3.29, + "grad_norm": 2.734375, + "learning_rate": 3.3493649053890325e-07, + "loss": 0.7032, + "step": 2505 + }, + { + "epoch": 3.29, + "grad_norm": 2.8125, + "learning_rate": 3.3362606249183446e-07, + "loss": 0.744, + "step": 2506 + }, + { + "epoch": 3.29, + "grad_norm": 2.8125, + "learning_rate": 3.323180196707709e-07, + "loss": 0.7664, + "step": 2507 + }, + { + "epoch": 3.29, + "grad_norm": 2.765625, + "learning_rate": 3.3101236351589764e-07, + "loss": 0.7668, + "step": 2508 + }, + { + "epoch": 3.29, + "grad_norm": 2.78125, + "learning_rate": 3.297090954647714e-07, + "loss": 0.7308, + "step": 2509 + }, + { + "epoch": 3.3, + "grad_norm": 2.765625, + "learning_rate": 3.2840821695231867e-07, + "loss": 0.7502, + "step": 2510 + }, + { + "epoch": 3.3, + "grad_norm": 2.75, + "learning_rate": 3.271097294108369e-07, + "loss": 0.7618, + "step": 2511 + }, + { + "epoch": 3.3, + "grad_norm": 2.734375, + "learning_rate": 3.2581363426998966e-07, + "loss": 0.7359, + "step": 2512 + }, + { + "epoch": 3.3, + "grad_norm": 2.859375, + "learning_rate": 3.245199329568055e-07, + "loss": 0.7454, + "step": 2513 + }, + { + "epoch": 3.3, + "grad_norm": 2.796875, + "learning_rate": 3.2322862689567885e-07, + "loss": 0.7373, + "step": 2514 + }, + { + "epoch": 3.3, + "grad_norm": 2.8125, + "learning_rate": 3.219397175083669e-07, + "loss": 0.7672, + "step": 2515 + }, + { + "epoch": 3.3, + "grad_norm": 2.703125, + "learning_rate": 3.206532062139875e-07, + "loss": 0.7409, + "step": 2516 + }, + { + "epoch": 3.3, + "grad_norm": 2.71875, + "learning_rate": 3.193690944290179e-07, + "loss": 0.7164, + "step": 2517 + }, + { + "epoch": 3.31, + "grad_norm": 2.78125, + "learning_rate": 3.18087383567294e-07, + "loss": 0.7414, + "step": 2518 + }, + { + "epoch": 3.31, + "grad_norm": 2.6875, + "learning_rate": 3.168080750400082e-07, + "loss": 0.7373, + "step": 2519 + }, + { + "epoch": 3.31, + "grad_norm": 2.75, + "learning_rate": 3.1553117025570763e-07, + "loss": 0.7407, + "step": 2520 + }, + { + "epoch": 3.31, + "grad_norm": 2.765625, + "learning_rate": 3.142566706202929e-07, + "loss": 0.7237, + "step": 2521 + }, + { + "epoch": 3.31, + "grad_norm": 2.71875, + "learning_rate": 3.129845775370163e-07, + "loss": 0.7437, + "step": 2522 + }, + { + "epoch": 3.31, + "grad_norm": 2.765625, + "learning_rate": 3.117148924064811e-07, + "loss": 0.7347, + "step": 2523 + }, + { + "epoch": 3.31, + "grad_norm": 2.6875, + "learning_rate": 3.1044761662663933e-07, + "loss": 0.7156, + "step": 2524 + }, + { + "epoch": 3.32, + "grad_norm": 2.796875, + "learning_rate": 3.091827515927884e-07, + "loss": 0.7528, + "step": 2525 + }, + { + "epoch": 3.32, + "grad_norm": 2.75, + "learning_rate": 3.079202986975741e-07, + "loss": 0.7298, + "step": 2526 + }, + { + "epoch": 3.32, + "grad_norm": 2.828125, + "learning_rate": 3.0666025933098475e-07, + "loss": 0.7426, + "step": 2527 + }, + { + "epoch": 3.32, + "grad_norm": 2.765625, + "learning_rate": 3.054026348803521e-07, + "loss": 0.7318, + "step": 2528 + }, + { + "epoch": 3.32, + "grad_norm": 2.71875, + "learning_rate": 3.041474267303479e-07, + "loss": 0.7163, + "step": 2529 + }, + { + "epoch": 3.32, + "grad_norm": 2.828125, + "learning_rate": 3.0289463626298585e-07, + "loss": 0.7589, + "step": 2530 + }, + { + "epoch": 3.32, + "grad_norm": 2.765625, + "learning_rate": 3.016442648576151e-07, + "loss": 0.7362, + "step": 2531 + }, + { + "epoch": 3.32, + "grad_norm": 2.84375, + "learning_rate": 3.003963138909224e-07, + "loss": 0.7456, + "step": 2532 + }, + { + "epoch": 3.33, + "grad_norm": 2.71875, + "learning_rate": 2.991507847369299e-07, + "loss": 0.7276, + "step": 2533 + }, + { + "epoch": 3.33, + "grad_norm": 2.734375, + "learning_rate": 2.9790767876699293e-07, + "loss": 0.755, + "step": 2534 + }, + { + "epoch": 3.33, + "grad_norm": 2.84375, + "learning_rate": 2.9666699734979875e-07, + "loss": 0.7558, + "step": 2535 + }, + { + "epoch": 3.33, + "grad_norm": 2.75, + "learning_rate": 2.9542874185136545e-07, + "loss": 0.7497, + "step": 2536 + }, + { + "epoch": 3.33, + "grad_norm": 2.796875, + "learning_rate": 2.941929136350399e-07, + "loss": 0.7451, + "step": 2537 + }, + { + "epoch": 3.33, + "grad_norm": 2.796875, + "learning_rate": 2.929595140614963e-07, + "loss": 0.7497, + "step": 2538 + }, + { + "epoch": 3.33, + "grad_norm": 2.875, + "learning_rate": 2.917285444887355e-07, + "loss": 0.7591, + "step": 2539 + }, + { + "epoch": 3.33, + "grad_norm": 2.703125, + "learning_rate": 2.9050000627208195e-07, + "loss": 0.7396, + "step": 2540 + }, + { + "epoch": 3.34, + "grad_norm": 2.6875, + "learning_rate": 2.89273900764184e-07, + "loss": 0.7328, + "step": 2541 + }, + { + "epoch": 3.34, + "grad_norm": 2.78125, + "learning_rate": 2.880502293150117e-07, + "loss": 0.7459, + "step": 2542 + }, + { + "epoch": 3.34, + "grad_norm": 2.703125, + "learning_rate": 2.8682899327185293e-07, + "loss": 0.7186, + "step": 2543 + }, + { + "epoch": 3.34, + "grad_norm": 2.765625, + "learning_rate": 2.856101939793171e-07, + "loss": 0.729, + "step": 2544 + }, + { + "epoch": 3.34, + "grad_norm": 2.8125, + "learning_rate": 2.843938327793291e-07, + "loss": 0.7485, + "step": 2545 + }, + { + "epoch": 3.34, + "grad_norm": 2.6875, + "learning_rate": 2.8317991101112966e-07, + "loss": 0.7478, + "step": 2546 + }, + { + "epoch": 3.34, + "grad_norm": 2.828125, + "learning_rate": 2.819684300112732e-07, + "loss": 0.7725, + "step": 2547 + }, + { + "epoch": 3.35, + "grad_norm": 2.734375, + "learning_rate": 2.8075939111362915e-07, + "loss": 0.7284, + "step": 2548 + }, + { + "epoch": 3.35, + "grad_norm": 2.75, + "learning_rate": 2.7955279564937433e-07, + "loss": 0.7398, + "step": 2549 + }, + { + "epoch": 3.35, + "grad_norm": 2.875, + "learning_rate": 2.783486449469977e-07, + "loss": 0.7423, + "step": 2550 + }, + { + "epoch": 3.35, + "grad_norm": 2.75, + "learning_rate": 2.7714694033229656e-07, + "loss": 0.7381, + "step": 2551 + }, + { + "epoch": 3.35, + "grad_norm": 2.703125, + "learning_rate": 2.759476831283742e-07, + "loss": 0.7243, + "step": 2552 + }, + { + "epoch": 3.35, + "grad_norm": 2.765625, + "learning_rate": 2.747508746556396e-07, + "loss": 0.765, + "step": 2553 + }, + { + "epoch": 3.35, + "grad_norm": 2.9375, + "learning_rate": 2.7355651623180574e-07, + "loss": 0.7317, + "step": 2554 + }, + { + "epoch": 3.35, + "grad_norm": 2.796875, + "learning_rate": 2.723646091718868e-07, + "loss": 0.7361, + "step": 2555 + }, + { + "epoch": 3.36, + "grad_norm": 2.71875, + "learning_rate": 2.711751547882005e-07, + "loss": 0.7174, + "step": 2556 + }, + { + "epoch": 3.36, + "grad_norm": 2.765625, + "learning_rate": 2.699881543903618e-07, + "loss": 0.7314, + "step": 2557 + }, + { + "epoch": 3.36, + "grad_norm": 2.703125, + "learning_rate": 2.6880360928528497e-07, + "loss": 0.7426, + "step": 2558 + }, + { + "epoch": 3.36, + "grad_norm": 2.953125, + "learning_rate": 2.676215207771804e-07, + "loss": 0.7566, + "step": 2559 + }, + { + "epoch": 3.36, + "grad_norm": 2.84375, + "learning_rate": 2.6644189016755415e-07, + "loss": 0.7516, + "step": 2560 + }, + { + "epoch": 3.36, + "grad_norm": 2.84375, + "learning_rate": 2.652647187552054e-07, + "loss": 0.7347, + "step": 2561 + }, + { + "epoch": 3.36, + "grad_norm": 2.734375, + "learning_rate": 2.640900078362263e-07, + "loss": 0.7097, + "step": 2562 + }, + { + "epoch": 3.37, + "grad_norm": 2.734375, + "learning_rate": 2.629177587039997e-07, + "loss": 0.7516, + "step": 2563 + }, + { + "epoch": 3.37, + "grad_norm": 2.796875, + "learning_rate": 2.617479726491981e-07, + "loss": 0.7262, + "step": 2564 + }, + { + "epoch": 3.37, + "grad_norm": 2.796875, + "learning_rate": 2.605806509597819e-07, + "loss": 0.7511, + "step": 2565 + }, + { + "epoch": 3.37, + "grad_norm": 2.8125, + "learning_rate": 2.5941579492099853e-07, + "loss": 0.742, + "step": 2566 + }, + { + "epoch": 3.37, + "grad_norm": 2.71875, + "learning_rate": 2.5825340581538043e-07, + "loss": 0.7357, + "step": 2567 + }, + { + "epoch": 3.37, + "grad_norm": 2.8125, + "learning_rate": 2.5709348492274384e-07, + "loss": 0.7173, + "step": 2568 + }, + { + "epoch": 3.37, + "grad_norm": 2.796875, + "learning_rate": 2.559360335201874e-07, + "loss": 0.7532, + "step": 2569 + }, + { + "epoch": 3.37, + "grad_norm": 2.8125, + "learning_rate": 2.547810528820913e-07, + "loss": 0.7346, + "step": 2570 + }, + { + "epoch": 3.38, + "grad_norm": 2.765625, + "learning_rate": 2.536285442801148e-07, + "loss": 0.7426, + "step": 2571 + }, + { + "epoch": 3.38, + "grad_norm": 2.84375, + "learning_rate": 2.524785089831955e-07, + "loss": 0.7343, + "step": 2572 + }, + { + "epoch": 3.38, + "grad_norm": 2.75, + "learning_rate": 2.513309482575471e-07, + "loss": 0.7296, + "step": 2573 + }, + { + "epoch": 3.38, + "grad_norm": 2.8125, + "learning_rate": 2.5018586336666094e-07, + "loss": 0.7495, + "step": 2574 + }, + { + "epoch": 3.38, + "grad_norm": 2.75, + "learning_rate": 2.490432555713002e-07, + "loss": 0.7513, + "step": 2575 + }, + { + "epoch": 3.38, + "grad_norm": 2.796875, + "learning_rate": 2.47903126129502e-07, + "loss": 0.7371, + "step": 2576 + }, + { + "epoch": 3.38, + "grad_norm": 2.859375, + "learning_rate": 2.467654762965743e-07, + "loss": 0.7392, + "step": 2577 + }, + { + "epoch": 3.39, + "grad_norm": 2.84375, + "learning_rate": 2.456303073250943e-07, + "loss": 0.7331, + "step": 2578 + }, + { + "epoch": 3.39, + "grad_norm": 2.75, + "learning_rate": 2.4449762046490885e-07, + "loss": 0.7307, + "step": 2579 + }, + { + "epoch": 3.39, + "grad_norm": 2.8125, + "learning_rate": 2.4336741696313113e-07, + "loss": 0.7557, + "step": 2580 + }, + { + "epoch": 3.39, + "grad_norm": 2.75, + "learning_rate": 2.422396980641406e-07, + "loss": 0.722, + "step": 2581 + }, + { + "epoch": 3.39, + "grad_norm": 2.8125, + "learning_rate": 2.411144650095809e-07, + "loss": 0.7376, + "step": 2582 + }, + { + "epoch": 3.39, + "grad_norm": 2.734375, + "learning_rate": 2.3999171903835853e-07, + "loss": 0.7393, + "step": 2583 + }, + { + "epoch": 3.39, + "grad_norm": 2.734375, + "learning_rate": 2.388714613866422e-07, + "loss": 0.7223, + "step": 2584 + }, + { + "epoch": 3.39, + "grad_norm": 2.765625, + "learning_rate": 2.377536932878602e-07, + "loss": 0.7306, + "step": 2585 + }, + { + "epoch": 3.4, + "grad_norm": 2.75, + "learning_rate": 2.3663841597270032e-07, + "loss": 0.7416, + "step": 2586 + }, + { + "epoch": 3.4, + "grad_norm": 2.65625, + "learning_rate": 2.3552563066910756e-07, + "loss": 0.7443, + "step": 2587 + }, + { + "epoch": 3.4, + "grad_norm": 2.890625, + "learning_rate": 2.3441533860228376e-07, + "loss": 0.7527, + "step": 2588 + }, + { + "epoch": 3.4, + "grad_norm": 2.765625, + "learning_rate": 2.3330754099468528e-07, + "loss": 0.7381, + "step": 2589 + }, + { + "epoch": 3.4, + "grad_norm": 2.703125, + "learning_rate": 2.3220223906602113e-07, + "loss": 0.7393, + "step": 2590 + }, + { + "epoch": 3.4, + "grad_norm": 2.75, + "learning_rate": 2.3109943403325397e-07, + "loss": 0.7519, + "step": 2591 + }, + { + "epoch": 3.4, + "grad_norm": 2.78125, + "learning_rate": 2.299991271105964e-07, + "loss": 0.717, + "step": 2592 + }, + { + "epoch": 3.41, + "grad_norm": 2.703125, + "learning_rate": 2.289013195095116e-07, + "loss": 0.7463, + "step": 2593 + }, + { + "epoch": 3.41, + "grad_norm": 2.78125, + "learning_rate": 2.2780601243870964e-07, + "loss": 0.7392, + "step": 2594 + }, + { + "epoch": 3.41, + "grad_norm": 2.78125, + "learning_rate": 2.2671320710414861e-07, + "loss": 0.7651, + "step": 2595 + }, + { + "epoch": 3.41, + "grad_norm": 2.859375, + "learning_rate": 2.2562290470903082e-07, + "loss": 0.73, + "step": 2596 + }, + { + "epoch": 3.41, + "grad_norm": 2.75, + "learning_rate": 2.2453510645380344e-07, + "loss": 0.7186, + "step": 2597 + }, + { + "epoch": 3.41, + "grad_norm": 2.734375, + "learning_rate": 2.2344981353615737e-07, + "loss": 0.704, + "step": 2598 + }, + { + "epoch": 3.41, + "grad_norm": 2.796875, + "learning_rate": 2.2236702715102366e-07, + "loss": 0.7348, + "step": 2599 + }, + { + "epoch": 3.41, + "grad_norm": 2.671875, + "learning_rate": 2.212867484905745e-07, + "loss": 0.7304, + "step": 2600 + }, + { + "epoch": 3.42, + "grad_norm": 2.828125, + "learning_rate": 2.2020897874422076e-07, + "loss": 0.7573, + "step": 2601 + }, + { + "epoch": 3.42, + "grad_norm": 2.8125, + "learning_rate": 2.191337190986112e-07, + "loss": 0.7553, + "step": 2602 + }, + { + "epoch": 3.42, + "grad_norm": 2.734375, + "learning_rate": 2.180609707376302e-07, + "loss": 0.7278, + "step": 2603 + }, + { + "epoch": 3.42, + "grad_norm": 2.921875, + "learning_rate": 2.16990734842398e-07, + "loss": 0.7378, + "step": 2604 + }, + { + "epoch": 3.42, + "grad_norm": 2.703125, + "learning_rate": 2.1592301259126787e-07, + "loss": 0.718, + "step": 2605 + }, + { + "epoch": 3.42, + "grad_norm": 2.734375, + "learning_rate": 2.148578051598263e-07, + "loss": 0.7515, + "step": 2606 + }, + { + "epoch": 3.42, + "grad_norm": 2.765625, + "learning_rate": 2.1379511372089018e-07, + "loss": 0.7162, + "step": 2607 + }, + { + "epoch": 3.43, + "grad_norm": 2.78125, + "learning_rate": 2.1273493944450634e-07, + "loss": 0.7593, + "step": 2608 + }, + { + "epoch": 3.43, + "grad_norm": 2.78125, + "learning_rate": 2.1167728349795025e-07, + "loss": 0.7483, + "step": 2609 + }, + { + "epoch": 3.43, + "grad_norm": 2.75, + "learning_rate": 2.1062214704572426e-07, + "loss": 0.7329, + "step": 2610 + }, + { + "epoch": 3.43, + "grad_norm": 2.765625, + "learning_rate": 2.09569531249558e-07, + "loss": 0.73, + "step": 2611 + }, + { + "epoch": 3.43, + "grad_norm": 2.78125, + "learning_rate": 2.085194372684049e-07, + "loss": 0.7422, + "step": 2612 + }, + { + "epoch": 3.43, + "grad_norm": 2.65625, + "learning_rate": 2.0747186625844183e-07, + "loss": 0.7077, + "step": 2613 + }, + { + "epoch": 3.43, + "grad_norm": 2.734375, + "learning_rate": 2.06426819373067e-07, + "loss": 0.7269, + "step": 2614 + }, + { + "epoch": 3.43, + "grad_norm": 2.78125, + "learning_rate": 2.0538429776290114e-07, + "loss": 0.7389, + "step": 2615 + }, + { + "epoch": 3.44, + "grad_norm": 2.71875, + "learning_rate": 2.0434430257578325e-07, + "loss": 0.7308, + "step": 2616 + }, + { + "epoch": 3.44, + "grad_norm": 2.84375, + "learning_rate": 2.0330683495677182e-07, + "loss": 0.7502, + "step": 2617 + }, + { + "epoch": 3.44, + "grad_norm": 2.734375, + "learning_rate": 2.0227189604814124e-07, + "loss": 0.7286, + "step": 2618 + }, + { + "epoch": 3.44, + "grad_norm": 2.796875, + "learning_rate": 2.0123948698938272e-07, + "loss": 0.7527, + "step": 2619 + }, + { + "epoch": 3.44, + "grad_norm": 2.75, + "learning_rate": 2.0020960891720147e-07, + "loss": 0.7329, + "step": 2620 + }, + { + "epoch": 3.44, + "grad_norm": 2.734375, + "learning_rate": 1.991822629655163e-07, + "loss": 0.7064, + "step": 2621 + }, + { + "epoch": 3.44, + "grad_norm": 2.734375, + "learning_rate": 1.981574502654579e-07, + "loss": 0.7558, + "step": 2622 + }, + { + "epoch": 3.45, + "grad_norm": 2.78125, + "learning_rate": 1.971351719453679e-07, + "loss": 0.7569, + "step": 2623 + }, + { + "epoch": 3.45, + "grad_norm": 2.71875, + "learning_rate": 1.9611542913079817e-07, + "loss": 0.7268, + "step": 2624 + }, + { + "epoch": 3.45, + "grad_norm": 2.78125, + "learning_rate": 1.9509822294450725e-07, + "loss": 0.7576, + "step": 2625 + }, + { + "epoch": 3.45, + "grad_norm": 2.828125, + "learning_rate": 1.9408355450646234e-07, + "loss": 0.7543, + "step": 2626 + }, + { + "epoch": 3.45, + "grad_norm": 2.75, + "learning_rate": 1.9307142493383585e-07, + "loss": 0.7103, + "step": 2627 + }, + { + "epoch": 3.45, + "grad_norm": 2.734375, + "learning_rate": 1.9206183534100474e-07, + "loss": 0.7038, + "step": 2628 + }, + { + "epoch": 3.45, + "grad_norm": 2.796875, + "learning_rate": 1.9105478683955042e-07, + "loss": 0.7318, + "step": 2629 + }, + { + "epoch": 3.45, + "grad_norm": 2.703125, + "learning_rate": 1.9005028053825592e-07, + "loss": 0.7094, + "step": 2630 + }, + { + "epoch": 3.46, + "grad_norm": 2.75, + "learning_rate": 1.8904831754310376e-07, + "loss": 0.7539, + "step": 2631 + }, + { + "epoch": 3.46, + "grad_norm": 2.78125, + "learning_rate": 1.8804889895727872e-07, + "loss": 0.7186, + "step": 2632 + }, + { + "epoch": 3.46, + "grad_norm": 2.75, + "learning_rate": 1.8705202588116227e-07, + "loss": 0.744, + "step": 2633 + }, + { + "epoch": 3.46, + "grad_norm": 2.9375, + "learning_rate": 1.8605769941233421e-07, + "loss": 0.7349, + "step": 2634 + }, + { + "epoch": 3.46, + "grad_norm": 2.71875, + "learning_rate": 1.850659206455699e-07, + "loss": 0.7295, + "step": 2635 + }, + { + "epoch": 3.46, + "grad_norm": 2.703125, + "learning_rate": 1.8407669067284028e-07, + "loss": 0.7094, + "step": 2636 + }, + { + "epoch": 3.46, + "grad_norm": 2.953125, + "learning_rate": 1.8309001058330944e-07, + "loss": 0.7257, + "step": 2637 + }, + { + "epoch": 3.47, + "grad_norm": 2.921875, + "learning_rate": 1.821058814633339e-07, + "loss": 0.7217, + "step": 2638 + }, + { + "epoch": 3.47, + "grad_norm": 2.765625, + "learning_rate": 1.8112430439646194e-07, + "loss": 0.7139, + "step": 2639 + }, + { + "epoch": 3.47, + "grad_norm": 2.828125, + "learning_rate": 1.8014528046343183e-07, + "loss": 0.7362, + "step": 2640 + }, + { + "epoch": 3.47, + "grad_norm": 2.765625, + "learning_rate": 1.7916881074217102e-07, + "loss": 0.7666, + "step": 2641 + }, + { + "epoch": 3.47, + "grad_norm": 2.96875, + "learning_rate": 1.7819489630779423e-07, + "loss": 0.7341, + "step": 2642 + }, + { + "epoch": 3.47, + "grad_norm": 2.71875, + "learning_rate": 1.7722353823260262e-07, + "loss": 0.7503, + "step": 2643 + }, + { + "epoch": 3.47, + "grad_norm": 2.75, + "learning_rate": 1.762547375860832e-07, + "loss": 0.7234, + "step": 2644 + }, + { + "epoch": 3.47, + "grad_norm": 2.734375, + "learning_rate": 1.752884954349074e-07, + "loss": 0.759, + "step": 2645 + }, + { + "epoch": 3.48, + "grad_norm": 2.75, + "learning_rate": 1.7432481284292923e-07, + "loss": 0.7275, + "step": 2646 + }, + { + "epoch": 3.48, + "grad_norm": 2.75, + "learning_rate": 1.7336369087118443e-07, + "loss": 0.7429, + "step": 2647 + }, + { + "epoch": 3.48, + "grad_norm": 2.78125, + "learning_rate": 1.7240513057789065e-07, + "loss": 0.7283, + "step": 2648 + }, + { + "epoch": 3.48, + "grad_norm": 2.75, + "learning_rate": 1.7144913301844313e-07, + "loss": 0.7485, + "step": 2649 + }, + { + "epoch": 3.48, + "grad_norm": 2.78125, + "learning_rate": 1.7049569924541653e-07, + "loss": 0.7214, + "step": 2650 + }, + { + "epoch": 3.48, + "grad_norm": 2.71875, + "learning_rate": 1.695448303085634e-07, + "loss": 0.7184, + "step": 2651 + }, + { + "epoch": 3.48, + "grad_norm": 2.75, + "learning_rate": 1.6859652725481095e-07, + "loss": 0.7651, + "step": 2652 + }, + { + "epoch": 3.49, + "grad_norm": 2.734375, + "learning_rate": 1.6765079112826266e-07, + "loss": 0.7198, + "step": 2653 + }, + { + "epoch": 3.49, + "grad_norm": 2.78125, + "learning_rate": 1.6670762297019444e-07, + "loss": 0.7473, + "step": 2654 + }, + { + "epoch": 3.49, + "grad_norm": 2.71875, + "learning_rate": 1.657670238190559e-07, + "loss": 0.7111, + "step": 2655 + }, + { + "epoch": 3.49, + "grad_norm": 2.765625, + "learning_rate": 1.6482899471046726e-07, + "loss": 0.7149, + "step": 2656 + }, + { + "epoch": 3.49, + "grad_norm": 2.890625, + "learning_rate": 1.6389353667721984e-07, + "loss": 0.7794, + "step": 2657 + }, + { + "epoch": 3.49, + "grad_norm": 2.78125, + "learning_rate": 1.629606507492737e-07, + "loss": 0.751, + "step": 2658 + }, + { + "epoch": 3.49, + "grad_norm": 2.75, + "learning_rate": 1.6203033795375717e-07, + "loss": 0.739, + "step": 2659 + }, + { + "epoch": 3.49, + "grad_norm": 2.75, + "learning_rate": 1.6110259931496564e-07, + "loss": 0.7302, + "step": 2660 + }, + { + "epoch": 3.5, + "grad_norm": 2.75, + "learning_rate": 1.6017743585435897e-07, + "loss": 0.7629, + "step": 2661 + }, + { + "epoch": 3.5, + "grad_norm": 2.796875, + "learning_rate": 1.5925484859056372e-07, + "loss": 0.7414, + "step": 2662 + }, + { + "epoch": 3.5, + "grad_norm": 2.890625, + "learning_rate": 1.583348385393685e-07, + "loss": 0.7227, + "step": 2663 + }, + { + "epoch": 3.5, + "grad_norm": 2.703125, + "learning_rate": 1.5741740671372548e-07, + "loss": 0.7169, + "step": 2664 + }, + { + "epoch": 3.5, + "grad_norm": 2.765625, + "learning_rate": 1.5650255412374638e-07, + "loss": 0.7348, + "step": 2665 + }, + { + "epoch": 3.5, + "grad_norm": 2.703125, + "learning_rate": 1.555902817767063e-07, + "loss": 0.71, + "step": 2666 + }, + { + "epoch": 3.5, + "grad_norm": 2.75, + "learning_rate": 1.5468059067703571e-07, + "loss": 0.7191, + "step": 2667 + }, + { + "epoch": 3.51, + "grad_norm": 2.78125, + "learning_rate": 1.5377348182632536e-07, + "loss": 0.7435, + "step": 2668 + }, + { + "epoch": 3.51, + "grad_norm": 2.75, + "learning_rate": 1.528689562233221e-07, + "loss": 0.7107, + "step": 2669 + }, + { + "epoch": 3.51, + "grad_norm": 2.796875, + "learning_rate": 1.5196701486392906e-07, + "loss": 0.7216, + "step": 2670 + }, + { + "epoch": 3.51, + "grad_norm": 2.75, + "learning_rate": 1.5106765874120382e-07, + "loss": 0.7617, + "step": 2671 + }, + { + "epoch": 3.51, + "grad_norm": 2.75, + "learning_rate": 1.501708888453579e-07, + "loss": 0.7454, + "step": 2672 + }, + { + "epoch": 3.51, + "grad_norm": 2.859375, + "learning_rate": 1.4927670616375345e-07, + "loss": 0.7424, + "step": 2673 + }, + { + "epoch": 3.51, + "grad_norm": 3.09375, + "learning_rate": 1.4838511168090707e-07, + "loss": 0.7219, + "step": 2674 + }, + { + "epoch": 3.51, + "grad_norm": 2.828125, + "learning_rate": 1.4749610637848383e-07, + "loss": 0.7569, + "step": 2675 + }, + { + "epoch": 3.52, + "grad_norm": 2.828125, + "learning_rate": 1.4660969123529795e-07, + "loss": 0.7219, + "step": 2676 + }, + { + "epoch": 3.52, + "grad_norm": 2.796875, + "learning_rate": 1.4572586722731292e-07, + "loss": 0.7286, + "step": 2677 + }, + { + "epoch": 3.52, + "grad_norm": 2.75, + "learning_rate": 1.4484463532763809e-07, + "loss": 0.729, + "step": 2678 + }, + { + "epoch": 3.52, + "grad_norm": 2.65625, + "learning_rate": 1.4396599650652959e-07, + "loss": 0.7401, + "step": 2679 + }, + { + "epoch": 3.52, + "grad_norm": 2.71875, + "learning_rate": 1.4308995173138828e-07, + "loss": 0.7234, + "step": 2680 + }, + { + "epoch": 3.52, + "grad_norm": 2.828125, + "learning_rate": 1.4221650196675901e-07, + "loss": 0.7358, + "step": 2681 + }, + { + "epoch": 3.52, + "grad_norm": 2.796875, + "learning_rate": 1.4134564817432972e-07, + "loss": 0.7345, + "step": 2682 + }, + { + "epoch": 3.53, + "grad_norm": 2.890625, + "learning_rate": 1.4047739131292953e-07, + "loss": 0.758, + "step": 2683 + }, + { + "epoch": 3.53, + "grad_norm": 2.796875, + "learning_rate": 1.3961173233852876e-07, + "loss": 0.7417, + "step": 2684 + }, + { + "epoch": 3.53, + "grad_norm": 2.78125, + "learning_rate": 1.3874867220423721e-07, + "loss": 0.72, + "step": 2685 + }, + { + "epoch": 3.53, + "grad_norm": 2.84375, + "learning_rate": 1.3788821186030338e-07, + "loss": 0.7787, + "step": 2686 + }, + { + "epoch": 3.53, + "grad_norm": 2.703125, + "learning_rate": 1.370303522541133e-07, + "loss": 0.7291, + "step": 2687 + }, + { + "epoch": 3.53, + "grad_norm": 2.71875, + "learning_rate": 1.3617509433018928e-07, + "loss": 0.7344, + "step": 2688 + }, + { + "epoch": 3.53, + "grad_norm": 2.71875, + "learning_rate": 1.3532243903018993e-07, + "loss": 0.7444, + "step": 2689 + }, + { + "epoch": 3.53, + "grad_norm": 2.703125, + "learning_rate": 1.3447238729290713e-07, + "loss": 0.7318, + "step": 2690 + }, + { + "epoch": 3.54, + "grad_norm": 2.703125, + "learning_rate": 1.3362494005426662e-07, + "loss": 0.7339, + "step": 2691 + }, + { + "epoch": 3.54, + "grad_norm": 2.78125, + "learning_rate": 1.3278009824732763e-07, + "loss": 0.748, + "step": 2692 + }, + { + "epoch": 3.54, + "grad_norm": 2.8125, + "learning_rate": 1.3193786280227912e-07, + "loss": 0.7342, + "step": 2693 + }, + { + "epoch": 3.54, + "grad_norm": 2.8125, + "learning_rate": 1.310982346464415e-07, + "loss": 0.7375, + "step": 2694 + }, + { + "epoch": 3.54, + "grad_norm": 2.796875, + "learning_rate": 1.3026121470426394e-07, + "loss": 0.7616, + "step": 2695 + }, + { + "epoch": 3.54, + "grad_norm": 2.6875, + "learning_rate": 1.2942680389732398e-07, + "loss": 0.7177, + "step": 2696 + }, + { + "epoch": 3.54, + "grad_norm": 2.65625, + "learning_rate": 1.2859500314432616e-07, + "loss": 0.7198, + "step": 2697 + }, + { + "epoch": 3.55, + "grad_norm": 2.84375, + "learning_rate": 1.2776581336110234e-07, + "loss": 0.7457, + "step": 2698 + }, + { + "epoch": 3.55, + "grad_norm": 2.796875, + "learning_rate": 1.269392354606086e-07, + "loss": 0.7488, + "step": 2699 + }, + { + "epoch": 3.55, + "grad_norm": 2.828125, + "learning_rate": 1.2611527035292558e-07, + "loss": 0.7396, + "step": 2700 + }, + { + "epoch": 3.55, + "grad_norm": 2.84375, + "learning_rate": 1.2529391894525733e-07, + "loss": 0.7652, + "step": 2701 + }, + { + "epoch": 3.55, + "grad_norm": 2.84375, + "learning_rate": 1.2447518214193017e-07, + "loss": 0.7386, + "step": 2702 + }, + { + "epoch": 3.55, + "grad_norm": 2.859375, + "learning_rate": 1.2365906084439133e-07, + "loss": 0.7597, + "step": 2703 + }, + { + "epoch": 3.55, + "grad_norm": 2.78125, + "learning_rate": 1.2284555595120901e-07, + "loss": 0.7328, + "step": 2704 + }, + { + "epoch": 3.55, + "grad_norm": 2.75, + "learning_rate": 1.2203466835806977e-07, + "loss": 0.7354, + "step": 2705 + }, + { + "epoch": 3.56, + "grad_norm": 2.734375, + "learning_rate": 1.2122639895777887e-07, + "loss": 0.7442, + "step": 2706 + }, + { + "epoch": 3.56, + "grad_norm": 2.71875, + "learning_rate": 1.2042074864025977e-07, + "loss": 0.7365, + "step": 2707 + }, + { + "epoch": 3.56, + "grad_norm": 2.84375, + "learning_rate": 1.1961771829255037e-07, + "loss": 0.7599, + "step": 2708 + }, + { + "epoch": 3.56, + "grad_norm": 2.75, + "learning_rate": 1.1881730879880565e-07, + "loss": 0.751, + "step": 2709 + }, + { + "epoch": 3.56, + "grad_norm": 2.671875, + "learning_rate": 1.1801952104029347e-07, + "loss": 0.713, + "step": 2710 + }, + { + "epoch": 3.56, + "grad_norm": 2.78125, + "learning_rate": 1.1722435589539699e-07, + "loss": 0.7472, + "step": 2711 + }, + { + "epoch": 3.56, + "grad_norm": 2.734375, + "learning_rate": 1.1643181423961037e-07, + "loss": 0.7608, + "step": 2712 + }, + { + "epoch": 3.57, + "grad_norm": 2.796875, + "learning_rate": 1.1564189694554001e-07, + "loss": 0.7604, + "step": 2713 + }, + { + "epoch": 3.57, + "grad_norm": 2.796875, + "learning_rate": 1.1485460488290163e-07, + "loss": 0.7267, + "step": 2714 + }, + { + "epoch": 3.57, + "grad_norm": 2.734375, + "learning_rate": 1.140699389185218e-07, + "loss": 0.7258, + "step": 2715 + }, + { + "epoch": 3.57, + "grad_norm": 2.734375, + "learning_rate": 1.1328789991633532e-07, + "loss": 0.7486, + "step": 2716 + }, + { + "epoch": 3.57, + "grad_norm": 2.890625, + "learning_rate": 1.1250848873738452e-07, + "loss": 0.7492, + "step": 2717 + }, + { + "epoch": 3.57, + "grad_norm": 2.765625, + "learning_rate": 1.1173170623981822e-07, + "loss": 0.7401, + "step": 2718 + }, + { + "epoch": 3.57, + "grad_norm": 2.765625, + "learning_rate": 1.1095755327889146e-07, + "loss": 0.7704, + "step": 2719 + }, + { + "epoch": 3.57, + "grad_norm": 2.703125, + "learning_rate": 1.1018603070696409e-07, + "loss": 0.7402, + "step": 2720 + }, + { + "epoch": 3.58, + "grad_norm": 2.71875, + "learning_rate": 1.094171393734994e-07, + "loss": 0.7158, + "step": 2721 + }, + { + "epoch": 3.58, + "grad_norm": 2.8125, + "learning_rate": 1.0865088012506408e-07, + "loss": 0.7411, + "step": 2722 + }, + { + "epoch": 3.58, + "grad_norm": 2.734375, + "learning_rate": 1.078872538053266e-07, + "loss": 0.7364, + "step": 2723 + }, + { + "epoch": 3.58, + "grad_norm": 2.84375, + "learning_rate": 1.0712626125505665e-07, + "loss": 0.7643, + "step": 2724 + }, + { + "epoch": 3.58, + "grad_norm": 2.78125, + "learning_rate": 1.0636790331212399e-07, + "loss": 0.727, + "step": 2725 + }, + { + "epoch": 3.58, + "grad_norm": 2.75, + "learning_rate": 1.0561218081149711e-07, + "loss": 0.7175, + "step": 2726 + }, + { + "epoch": 3.58, + "grad_norm": 2.875, + "learning_rate": 1.0485909458524402e-07, + "loss": 0.7593, + "step": 2727 + }, + { + "epoch": 3.59, + "grad_norm": 2.796875, + "learning_rate": 1.0410864546252841e-07, + "loss": 0.7395, + "step": 2728 + }, + { + "epoch": 3.59, + "grad_norm": 2.59375, + "learning_rate": 1.0336083426961296e-07, + "loss": 0.6924, + "step": 2729 + }, + { + "epoch": 3.59, + "grad_norm": 2.859375, + "learning_rate": 1.0261566182985378e-07, + "loss": 0.7466, + "step": 2730 + }, + { + "epoch": 3.59, + "grad_norm": 2.75, + "learning_rate": 1.0187312896370233e-07, + "loss": 0.727, + "step": 2731 + }, + { + "epoch": 3.59, + "grad_norm": 2.796875, + "learning_rate": 1.0113323648870354e-07, + "loss": 0.7462, + "step": 2732 + }, + { + "epoch": 3.59, + "grad_norm": 2.859375, + "learning_rate": 1.0039598521949578e-07, + "loss": 0.7598, + "step": 2733 + }, + { + "epoch": 3.59, + "grad_norm": 2.796875, + "learning_rate": 9.966137596780945e-08, + "loss": 0.7507, + "step": 2734 + }, + { + "epoch": 3.59, + "grad_norm": 2.75, + "learning_rate": 9.892940954246532e-08, + "loss": 0.7347, + "step": 2735 + }, + { + "epoch": 3.6, + "grad_norm": 2.8125, + "learning_rate": 9.820008674937459e-08, + "loss": 0.7634, + "step": 2736 + }, + { + "epoch": 3.6, + "grad_norm": 2.78125, + "learning_rate": 9.747340839153852e-08, + "loss": 0.7517, + "step": 2737 + }, + { + "epoch": 3.6, + "grad_norm": 2.75, + "learning_rate": 9.674937526904571e-08, + "loss": 0.7389, + "step": 2738 + }, + { + "epoch": 3.6, + "grad_norm": 2.828125, + "learning_rate": 9.602798817907294e-08, + "loss": 0.7393, + "step": 2739 + }, + { + "epoch": 3.6, + "grad_norm": 2.796875, + "learning_rate": 9.530924791588319e-08, + "loss": 0.7612, + "step": 2740 + }, + { + "epoch": 3.6, + "grad_norm": 2.734375, + "learning_rate": 9.459315527082624e-08, + "loss": 0.7261, + "step": 2741 + }, + { + "epoch": 3.6, + "grad_norm": 2.78125, + "learning_rate": 9.387971103233556e-08, + "loss": 0.7272, + "step": 2742 + }, + { + "epoch": 3.61, + "grad_norm": 2.703125, + "learning_rate": 9.316891598592892e-08, + "loss": 0.7141, + "step": 2743 + }, + { + "epoch": 3.61, + "grad_norm": 2.765625, + "learning_rate": 9.24607709142078e-08, + "loss": 0.7343, + "step": 2744 + }, + { + "epoch": 3.61, + "grad_norm": 2.78125, + "learning_rate": 9.175527659685573e-08, + "loss": 0.7291, + "step": 2745 + }, + { + "epoch": 3.61, + "grad_norm": 2.703125, + "learning_rate": 9.10524338106375e-08, + "loss": 0.6963, + "step": 2746 + }, + { + "epoch": 3.61, + "grad_norm": 2.734375, + "learning_rate": 9.035224332939879e-08, + "loss": 0.7264, + "step": 2747 + }, + { + "epoch": 3.61, + "grad_norm": 2.828125, + "learning_rate": 8.965470592406544e-08, + "loss": 0.758, + "step": 2748 + }, + { + "epoch": 3.61, + "grad_norm": 2.734375, + "learning_rate": 8.895982236264117e-08, + "loss": 0.7443, + "step": 2749 + }, + { + "epoch": 3.61, + "grad_norm": 2.78125, + "learning_rate": 8.826759341020813e-08, + "loss": 0.7427, + "step": 2750 + }, + { + "epoch": 3.62, + "grad_norm": 2.84375, + "learning_rate": 8.757801982892583e-08, + "loss": 0.7344, + "step": 2751 + }, + { + "epoch": 3.62, + "grad_norm": 2.75, + "learning_rate": 8.689110237803056e-08, + "loss": 0.7442, + "step": 2752 + }, + { + "epoch": 3.62, + "grad_norm": 2.765625, + "learning_rate": 8.620684181383343e-08, + "loss": 0.7205, + "step": 2753 + }, + { + "epoch": 3.62, + "grad_norm": 2.765625, + "learning_rate": 8.552523888972097e-08, + "loss": 0.7302, + "step": 2754 + }, + { + "epoch": 3.62, + "grad_norm": 2.75, + "learning_rate": 8.48462943561526e-08, + "loss": 0.745, + "step": 2755 + }, + { + "epoch": 3.62, + "grad_norm": 2.75, + "learning_rate": 8.417000896066201e-08, + "loss": 0.7418, + "step": 2756 + }, + { + "epoch": 3.62, + "grad_norm": 2.734375, + "learning_rate": 8.349638344785388e-08, + "loss": 0.7604, + "step": 2757 + }, + { + "epoch": 3.63, + "grad_norm": 2.84375, + "learning_rate": 8.282541855940546e-08, + "loss": 0.7529, + "step": 2758 + }, + { + "epoch": 3.63, + "grad_norm": 2.796875, + "learning_rate": 8.21571150340636e-08, + "loss": 0.7441, + "step": 2759 + }, + { + "epoch": 3.63, + "grad_norm": 2.796875, + "learning_rate": 8.149147360764609e-08, + "loss": 0.7247, + "step": 2760 + }, + { + "epoch": 3.63, + "grad_norm": 2.796875, + "learning_rate": 8.082849501303836e-08, + "loss": 0.7345, + "step": 2761 + }, + { + "epoch": 3.63, + "grad_norm": 2.71875, + "learning_rate": 8.016817998019455e-08, + "loss": 0.7331, + "step": 2762 + }, + { + "epoch": 3.63, + "grad_norm": 2.75, + "learning_rate": 7.951052923613644e-08, + "loss": 0.726, + "step": 2763 + }, + { + "epoch": 3.63, + "grad_norm": 2.78125, + "learning_rate": 7.885554350495206e-08, + "loss": 0.7535, + "step": 2764 + }, + { + "epoch": 3.63, + "grad_norm": 2.78125, + "learning_rate": 7.820322350779536e-08, + "loss": 0.7317, + "step": 2765 + }, + { + "epoch": 3.64, + "grad_norm": 2.6875, + "learning_rate": 7.755356996288571e-08, + "loss": 0.7013, + "step": 2766 + }, + { + "epoch": 3.64, + "grad_norm": 2.78125, + "learning_rate": 7.69065835855054e-08, + "loss": 0.7384, + "step": 2767 + }, + { + "epoch": 3.64, + "grad_norm": 2.71875, + "learning_rate": 7.626226508800128e-08, + "loss": 0.7478, + "step": 2768 + }, + { + "epoch": 3.64, + "grad_norm": 2.828125, + "learning_rate": 7.562061517978198e-08, + "loss": 0.7591, + "step": 2769 + }, + { + "epoch": 3.64, + "grad_norm": 2.796875, + "learning_rate": 7.498163456731878e-08, + "loss": 0.7609, + "step": 2770 + }, + { + "epoch": 3.64, + "grad_norm": 2.78125, + "learning_rate": 7.434532395414335e-08, + "loss": 0.7136, + "step": 2771 + }, + { + "epoch": 3.64, + "grad_norm": 2.765625, + "learning_rate": 7.371168404084806e-08, + "loss": 0.7357, + "step": 2772 + }, + { + "epoch": 3.65, + "grad_norm": 2.78125, + "learning_rate": 7.308071552508455e-08, + "loss": 0.7532, + "step": 2773 + }, + { + "epoch": 3.65, + "grad_norm": 2.765625, + "learning_rate": 7.245241910156297e-08, + "loss": 0.7328, + "step": 2774 + }, + { + "epoch": 3.65, + "grad_norm": 2.765625, + "learning_rate": 7.18267954620519e-08, + "loss": 0.7677, + "step": 2775 + }, + { + "epoch": 3.65, + "grad_norm": 2.78125, + "learning_rate": 7.120384529537672e-08, + "loss": 0.7534, + "step": 2776 + }, + { + "epoch": 3.65, + "grad_norm": 2.859375, + "learning_rate": 7.058356928741988e-08, + "loss": 0.7279, + "step": 2777 + }, + { + "epoch": 3.65, + "grad_norm": 2.703125, + "learning_rate": 6.996596812111872e-08, + "loss": 0.7258, + "step": 2778 + }, + { + "epoch": 3.65, + "grad_norm": 2.765625, + "learning_rate": 6.935104247646596e-08, + "loss": 0.7477, + "step": 2779 + }, + { + "epoch": 3.65, + "grad_norm": 2.703125, + "learning_rate": 6.873879303050863e-08, + "loss": 0.7258, + "step": 2780 + }, + { + "epoch": 3.66, + "grad_norm": 2.859375, + "learning_rate": 6.812922045734666e-08, + "loss": 0.7537, + "step": 2781 + }, + { + "epoch": 3.66, + "grad_norm": 2.796875, + "learning_rate": 6.752232542813319e-08, + "loss": 0.7369, + "step": 2782 + }, + { + "epoch": 3.66, + "grad_norm": 2.78125, + "learning_rate": 6.691810861107318e-08, + "loss": 0.7418, + "step": 2783 + }, + { + "epoch": 3.66, + "grad_norm": 2.78125, + "learning_rate": 6.631657067142306e-08, + "loss": 0.7394, + "step": 2784 + }, + { + "epoch": 3.66, + "grad_norm": 2.75, + "learning_rate": 6.571771227148943e-08, + "loss": 0.7529, + "step": 2785 + }, + { + "epoch": 3.66, + "grad_norm": 2.765625, + "learning_rate": 6.512153407062849e-08, + "loss": 0.7551, + "step": 2786 + }, + { + "epoch": 3.66, + "grad_norm": 2.828125, + "learning_rate": 6.452803672524571e-08, + "loss": 0.746, + "step": 2787 + }, + { + "epoch": 3.67, + "grad_norm": 2.78125, + "learning_rate": 6.393722088879534e-08, + "loss": 0.7208, + "step": 2788 + }, + { + "epoch": 3.67, + "grad_norm": 2.8125, + "learning_rate": 6.334908721177868e-08, + "loss": 0.7459, + "step": 2789 + }, + { + "epoch": 3.67, + "grad_norm": 2.734375, + "learning_rate": 6.276363634174359e-08, + "loss": 0.7333, + "step": 2790 + }, + { + "epoch": 3.67, + "grad_norm": 2.8125, + "learning_rate": 6.218086892328446e-08, + "loss": 0.7524, + "step": 2791 + }, + { + "epoch": 3.67, + "grad_norm": 2.75, + "learning_rate": 6.160078559804189e-08, + "loss": 0.7351, + "step": 2792 + }, + { + "epoch": 3.67, + "grad_norm": 2.75, + "learning_rate": 6.102338700470028e-08, + "loss": 0.7482, + "step": 2793 + }, + { + "epoch": 3.67, + "grad_norm": 2.78125, + "learning_rate": 6.044867377898806e-08, + "loss": 0.7403, + "step": 2794 + }, + { + "epoch": 3.67, + "grad_norm": 2.765625, + "learning_rate": 5.987664655367792e-08, + "loss": 0.7417, + "step": 2795 + }, + { + "epoch": 3.68, + "grad_norm": 2.671875, + "learning_rate": 5.930730595858414e-08, + "loss": 0.7322, + "step": 2796 + }, + { + "epoch": 3.68, + "grad_norm": 2.828125, + "learning_rate": 5.8740652620563334e-08, + "loss": 0.7684, + "step": 2797 + }, + { + "epoch": 3.68, + "grad_norm": 2.734375, + "learning_rate": 5.817668716351338e-08, + "loss": 0.7392, + "step": 2798 + }, + { + "epoch": 3.68, + "grad_norm": 2.796875, + "learning_rate": 5.7615410208373415e-08, + "loss": 0.7334, + "step": 2799 + }, + { + "epoch": 3.68, + "grad_norm": 2.75, + "learning_rate": 5.7056822373121324e-08, + "loss": 0.7448, + "step": 2800 + }, + { + "epoch": 3.68, + "grad_norm": 2.734375, + "learning_rate": 5.650092427277487e-08, + "loss": 0.7345, + "step": 2801 + }, + { + "epoch": 3.68, + "grad_norm": 2.796875, + "learning_rate": 5.594771651939057e-08, + "loss": 0.7373, + "step": 2802 + }, + { + "epoch": 3.68, + "grad_norm": 2.71875, + "learning_rate": 5.5397199722062034e-08, + "loss": 0.7372, + "step": 2803 + }, + { + "epoch": 3.69, + "grad_norm": 2.796875, + "learning_rate": 5.484937448692079e-08, + "loss": 0.7445, + "step": 2804 + }, + { + "epoch": 3.69, + "grad_norm": 2.71875, + "learning_rate": 5.430424141713464e-08, + "loss": 0.7263, + "step": 2805 + }, + { + "epoch": 3.69, + "grad_norm": 2.78125, + "learning_rate": 5.3761801112907356e-08, + "loss": 0.7268, + "step": 2806 + }, + { + "epoch": 3.69, + "grad_norm": 2.84375, + "learning_rate": 5.3222054171477574e-08, + "loss": 0.7446, + "step": 2807 + }, + { + "epoch": 3.69, + "grad_norm": 2.828125, + "learning_rate": 5.268500118711883e-08, + "loss": 0.7415, + "step": 2808 + }, + { + "epoch": 3.69, + "grad_norm": 2.71875, + "learning_rate": 5.2150642751138124e-08, + "loss": 0.7326, + "step": 2809 + }, + { + "epoch": 3.69, + "grad_norm": 2.765625, + "learning_rate": 5.16189794518765e-08, + "loss": 0.7352, + "step": 2810 + }, + { + "epoch": 3.7, + "grad_norm": 2.765625, + "learning_rate": 5.10900118747068e-08, + "loss": 0.7468, + "step": 2811 + }, + { + "epoch": 3.7, + "grad_norm": 2.8125, + "learning_rate": 5.0563740602034284e-08, + "loss": 0.7509, + "step": 2812 + }, + { + "epoch": 3.7, + "grad_norm": 2.734375, + "learning_rate": 5.004016621329544e-08, + "loss": 0.7191, + "step": 2813 + }, + { + "epoch": 3.7, + "grad_norm": 2.671875, + "learning_rate": 4.9519289284956916e-08, + "loss": 0.7324, + "step": 2814 + }, + { + "epoch": 3.7, + "grad_norm": 2.796875, + "learning_rate": 4.900111039051636e-08, + "loss": 0.7422, + "step": 2815 + }, + { + "epoch": 3.7, + "grad_norm": 2.796875, + "learning_rate": 4.848563010049989e-08, + "loss": 0.7299, + "step": 2816 + }, + { + "epoch": 3.7, + "grad_norm": 2.78125, + "learning_rate": 4.797284898246296e-08, + "loss": 0.7403, + "step": 2817 + }, + { + "epoch": 3.7, + "grad_norm": 2.84375, + "learning_rate": 4.746276760098867e-08, + "loss": 0.7352, + "step": 2818 + }, + { + "epoch": 3.71, + "grad_norm": 2.765625, + "learning_rate": 4.695538651768861e-08, + "loss": 0.7482, + "step": 2819 + }, + { + "epoch": 3.71, + "grad_norm": 2.734375, + "learning_rate": 4.6450706291200354e-08, + "loss": 0.7571, + "step": 2820 + }, + { + "epoch": 3.71, + "grad_norm": 2.8125, + "learning_rate": 4.594872747718804e-08, + "loss": 0.752, + "step": 2821 + }, + { + "epoch": 3.71, + "grad_norm": 2.890625, + "learning_rate": 4.54494506283415e-08, + "loss": 0.7472, + "step": 2822 + }, + { + "epoch": 3.71, + "grad_norm": 2.734375, + "learning_rate": 4.4952876294376e-08, + "loss": 0.7175, + "step": 2823 + }, + { + "epoch": 3.71, + "grad_norm": 2.75, + "learning_rate": 4.44590050220306e-08, + "loss": 0.7253, + "step": 2824 + }, + { + "epoch": 3.71, + "grad_norm": 2.796875, + "learning_rate": 4.396783735506893e-08, + "loss": 0.7376, + "step": 2825 + }, + { + "epoch": 3.72, + "grad_norm": 2.734375, + "learning_rate": 4.3479373834277026e-08, + "loss": 0.7321, + "step": 2826 + }, + { + "epoch": 3.72, + "grad_norm": 2.71875, + "learning_rate": 4.299361499746441e-08, + "loss": 0.739, + "step": 2827 + }, + { + "epoch": 3.72, + "grad_norm": 2.75, + "learning_rate": 4.2510561379462136e-08, + "loss": 0.7638, + "step": 2828 + }, + { + "epoch": 3.72, + "grad_norm": 2.765625, + "learning_rate": 4.2030213512123384e-08, + "loss": 0.7394, + "step": 2829 + }, + { + "epoch": 3.72, + "grad_norm": 2.796875, + "learning_rate": 4.155257192432205e-08, + "loss": 0.7281, + "step": 2830 + }, + { + "epoch": 3.72, + "grad_norm": 2.765625, + "learning_rate": 4.1077637141951624e-08, + "loss": 0.7392, + "step": 2831 + }, + { + "epoch": 3.72, + "grad_norm": 2.875, + "learning_rate": 4.060540968792631e-08, + "loss": 0.7199, + "step": 2832 + }, + { + "epoch": 3.72, + "grad_norm": 2.71875, + "learning_rate": 4.013589008217883e-08, + "loss": 0.7199, + "step": 2833 + }, + { + "epoch": 3.73, + "grad_norm": 2.765625, + "learning_rate": 3.9669078841660655e-08, + "loss": 0.7228, + "step": 2834 + }, + { + "epoch": 3.73, + "grad_norm": 2.71875, + "learning_rate": 3.92049764803415e-08, + "loss": 0.7264, + "step": 2835 + }, + { + "epoch": 3.73, + "grad_norm": 2.765625, + "learning_rate": 3.874358350920843e-08, + "loss": 0.7295, + "step": 2836 + }, + { + "epoch": 3.73, + "grad_norm": 2.75, + "learning_rate": 3.828490043626537e-08, + "loss": 0.7348, + "step": 2837 + }, + { + "epoch": 3.73, + "grad_norm": 2.71875, + "learning_rate": 3.782892776653252e-08, + "loss": 0.7246, + "step": 2838 + }, + { + "epoch": 3.73, + "grad_norm": 2.78125, + "learning_rate": 3.7375666002046064e-08, + "loss": 0.7531, + "step": 2839 + }, + { + "epoch": 3.73, + "grad_norm": 2.84375, + "learning_rate": 3.6925115641857356e-08, + "loss": 0.7359, + "step": 2840 + }, + { + "epoch": 3.74, + "grad_norm": 2.765625, + "learning_rate": 3.6477277182032365e-08, + "loss": 0.7443, + "step": 2841 + }, + { + "epoch": 3.74, + "grad_norm": 2.78125, + "learning_rate": 3.603215111565139e-08, + "loss": 0.7426, + "step": 2842 + }, + { + "epoch": 3.74, + "grad_norm": 2.75, + "learning_rate": 3.5589737932807674e-08, + "loss": 0.7321, + "step": 2843 + }, + { + "epoch": 3.74, + "grad_norm": 2.734375, + "learning_rate": 3.515003812060824e-08, + "loss": 0.7242, + "step": 2844 + }, + { + "epoch": 3.74, + "grad_norm": 2.8125, + "learning_rate": 3.47130521631725e-08, + "loss": 0.7508, + "step": 2845 + }, + { + "epoch": 3.74, + "grad_norm": 2.703125, + "learning_rate": 3.427878054163142e-08, + "loss": 0.7402, + "step": 2846 + }, + { + "epoch": 3.74, + "grad_norm": 2.703125, + "learning_rate": 3.384722373412835e-08, + "loss": 0.7309, + "step": 2847 + }, + { + "epoch": 3.74, + "grad_norm": 2.734375, + "learning_rate": 3.341838221581656e-08, + "loss": 0.7722, + "step": 2848 + }, + { + "epoch": 3.75, + "grad_norm": 2.78125, + "learning_rate": 3.2992256458860006e-08, + "loss": 0.74, + "step": 2849 + }, + { + "epoch": 3.75, + "grad_norm": 2.765625, + "learning_rate": 3.2568846932432565e-08, + "loss": 0.7351, + "step": 2850 + }, + { + "epoch": 3.75, + "grad_norm": 2.671875, + "learning_rate": 3.2148154102717985e-08, + "loss": 0.7087, + "step": 2851 + }, + { + "epoch": 3.75, + "grad_norm": 2.78125, + "learning_rate": 3.1730178432908266e-08, + "loss": 0.7378, + "step": 2852 + }, + { + "epoch": 3.75, + "grad_norm": 2.71875, + "learning_rate": 3.1314920383203886e-08, + "loss": 0.7254, + "step": 2853 + }, + { + "epoch": 3.75, + "grad_norm": 3.0625, + "learning_rate": 3.090238041081328e-08, + "loss": 0.7478, + "step": 2854 + }, + { + "epoch": 3.75, + "grad_norm": 2.765625, + "learning_rate": 3.049255896995201e-08, + "loss": 0.7333, + "step": 2855 + }, + { + "epoch": 3.76, + "grad_norm": 2.890625, + "learning_rate": 3.008545651184275e-08, + "loss": 0.7478, + "step": 2856 + }, + { + "epoch": 3.76, + "grad_norm": 2.71875, + "learning_rate": 2.9681073484714438e-08, + "loss": 0.7359, + "step": 2857 + }, + { + "epoch": 3.76, + "grad_norm": 2.859375, + "learning_rate": 2.9279410333801773e-08, + "loss": 0.7365, + "step": 2858 + }, + { + "epoch": 3.76, + "grad_norm": 2.75, + "learning_rate": 2.88804675013446e-08, + "loss": 0.7288, + "step": 2859 + }, + { + "epoch": 3.76, + "grad_norm": 2.765625, + "learning_rate": 2.848424542658823e-08, + "loss": 0.7377, + "step": 2860 + }, + { + "epoch": 3.76, + "grad_norm": 2.71875, + "learning_rate": 2.8090744545782034e-08, + "loss": 0.7067, + "step": 2861 + }, + { + "epoch": 3.76, + "grad_norm": 2.78125, + "learning_rate": 2.769996529217861e-08, + "loss": 0.7117, + "step": 2862 + }, + { + "epoch": 3.76, + "grad_norm": 2.78125, + "learning_rate": 2.7311908096035454e-08, + "loss": 0.7122, + "step": 2863 + }, + { + "epoch": 3.77, + "grad_norm": 2.6875, + "learning_rate": 2.6926573384611353e-08, + "loss": 0.7102, + "step": 2864 + }, + { + "epoch": 3.77, + "grad_norm": 2.75, + "learning_rate": 2.6543961582169153e-08, + "loss": 0.7175, + "step": 2865 + }, + { + "epoch": 3.77, + "grad_norm": 2.734375, + "learning_rate": 2.6164073109972986e-08, + "loss": 0.7483, + "step": 2866 + }, + { + "epoch": 3.77, + "grad_norm": 2.71875, + "learning_rate": 2.5786908386288e-08, + "loss": 0.7433, + "step": 2867 + }, + { + "epoch": 3.77, + "grad_norm": 2.796875, + "learning_rate": 2.5412467826381182e-08, + "loss": 0.7559, + "step": 2868 + }, + { + "epoch": 3.77, + "grad_norm": 2.765625, + "learning_rate": 2.504075184251997e-08, + "loss": 0.7435, + "step": 2869 + }, + { + "epoch": 3.77, + "grad_norm": 2.78125, + "learning_rate": 2.4671760843971992e-08, + "loss": 0.7434, + "step": 2870 + }, + { + "epoch": 3.78, + "grad_norm": 2.734375, + "learning_rate": 2.4305495237004762e-08, + "loss": 0.7279, + "step": 2871 + }, + { + "epoch": 3.78, + "grad_norm": 2.703125, + "learning_rate": 2.3941955424884312e-08, + "loss": 0.7393, + "step": 2872 + }, + { + "epoch": 3.78, + "grad_norm": 2.765625, + "learning_rate": 2.3581141807876573e-08, + "loss": 0.7192, + "step": 2873 + }, + { + "epoch": 3.78, + "grad_norm": 2.8125, + "learning_rate": 2.322305478324516e-08, + "loss": 0.731, + "step": 2874 + }, + { + "epoch": 3.78, + "grad_norm": 2.828125, + "learning_rate": 2.2867694745251634e-08, + "loss": 0.7385, + "step": 2875 + }, + { + "epoch": 3.78, + "grad_norm": 2.765625, + "learning_rate": 2.251506208515608e-08, + "loss": 0.7166, + "step": 2876 + }, + { + "epoch": 3.78, + "grad_norm": 2.8125, + "learning_rate": 2.2165157191214314e-08, + "loss": 0.7436, + "step": 2877 + }, + { + "epoch": 3.78, + "grad_norm": 2.796875, + "learning_rate": 2.1817980448679553e-08, + "loss": 0.728, + "step": 2878 + }, + { + "epoch": 3.79, + "grad_norm": 2.78125, + "learning_rate": 2.147353223980103e-08, + "loss": 0.7318, + "step": 2879 + }, + { + "epoch": 3.79, + "grad_norm": 2.71875, + "learning_rate": 2.1131812943823994e-08, + "loss": 0.7152, + "step": 2880 + }, + { + "epoch": 3.79, + "grad_norm": 2.796875, + "learning_rate": 2.0792822936989142e-08, + "loss": 0.7667, + "step": 2881 + }, + { + "epoch": 3.79, + "grad_norm": 2.78125, + "learning_rate": 2.0456562592532093e-08, + "loss": 0.7233, + "step": 2882 + }, + { + "epoch": 3.79, + "grad_norm": 2.796875, + "learning_rate": 2.0123032280682242e-08, + "loss": 0.7389, + "step": 2883 + }, + { + "epoch": 3.79, + "grad_norm": 2.765625, + "learning_rate": 1.979223236866501e-08, + "loss": 0.7327, + "step": 2884 + }, + { + "epoch": 3.79, + "grad_norm": 2.65625, + "learning_rate": 1.946416322069794e-08, + "loss": 0.733, + "step": 2885 + }, + { + "epoch": 3.8, + "grad_norm": 2.78125, + "learning_rate": 1.9138825197992096e-08, + "loss": 0.7327, + "step": 2886 + }, + { + "epoch": 3.8, + "grad_norm": 2.734375, + "learning_rate": 1.8816218658752605e-08, + "loss": 0.7271, + "step": 2887 + }, + { + "epoch": 3.8, + "grad_norm": 2.765625, + "learning_rate": 1.8496343958175898e-08, + "loss": 0.7392, + "step": 2888 + }, + { + "epoch": 3.8, + "grad_norm": 2.796875, + "learning_rate": 1.8179201448451368e-08, + "loss": 0.769, + "step": 2889 + }, + { + "epoch": 3.8, + "grad_norm": 2.765625, + "learning_rate": 1.7864791478760245e-08, + "loss": 0.7105, + "step": 2890 + }, + { + "epoch": 3.8, + "grad_norm": 2.78125, + "learning_rate": 1.755311439527424e-08, + "loss": 0.737, + "step": 2891 + }, + { + "epoch": 3.8, + "grad_norm": 2.8125, + "learning_rate": 1.7244170541157455e-08, + "loss": 0.749, + "step": 2892 + }, + { + "epoch": 3.8, + "grad_norm": 2.703125, + "learning_rate": 1.6937960256563914e-08, + "loss": 0.7221, + "step": 2893 + }, + { + "epoch": 3.81, + "grad_norm": 2.828125, + "learning_rate": 1.663448387863753e-08, + "loss": 0.7467, + "step": 2894 + }, + { + "epoch": 3.81, + "grad_norm": 2.796875, + "learning_rate": 1.6333741741512977e-08, + "loss": 0.7574, + "step": 2895 + }, + { + "epoch": 3.81, + "grad_norm": 2.765625, + "learning_rate": 1.603573417631371e-08, + "loss": 0.7659, + "step": 2896 + }, + { + "epoch": 3.81, + "grad_norm": 2.765625, + "learning_rate": 1.5740461511152827e-08, + "loss": 0.7434, + "step": 2897 + }, + { + "epoch": 3.81, + "grad_norm": 2.875, + "learning_rate": 1.5447924071131937e-08, + "loss": 0.7636, + "step": 2898 + }, + { + "epoch": 3.81, + "grad_norm": 2.8125, + "learning_rate": 1.5158122178341727e-08, + "loss": 0.7282, + "step": 2899 + }, + { + "epoch": 3.81, + "grad_norm": 2.765625, + "learning_rate": 1.4871056151860296e-08, + "loss": 0.7234, + "step": 2900 + }, + { + "epoch": 3.82, + "grad_norm": 2.796875, + "learning_rate": 1.4586726307753985e-08, + "loss": 0.7264, + "step": 2901 + }, + { + "epoch": 3.82, + "grad_norm": 2.859375, + "learning_rate": 1.4305132959075706e-08, + "loss": 0.7561, + "step": 2902 + }, + { + "epoch": 3.82, + "grad_norm": 2.796875, + "learning_rate": 1.4026276415866624e-08, + "loss": 0.7512, + "step": 2903 + }, + { + "epoch": 3.82, + "grad_norm": 2.78125, + "learning_rate": 1.3750156985153361e-08, + "loss": 0.7299, + "step": 2904 + }, + { + "epoch": 3.82, + "grad_norm": 2.78125, + "learning_rate": 1.3476774970950236e-08, + "loss": 0.7348, + "step": 2905 + }, + { + "epoch": 3.82, + "grad_norm": 2.734375, + "learning_rate": 1.3206130674256746e-08, + "loss": 0.7432, + "step": 2906 + }, + { + "epoch": 3.82, + "grad_norm": 2.84375, + "learning_rate": 1.293822439305814e-08, + "loss": 0.7636, + "step": 2907 + }, + { + "epoch": 3.82, + "grad_norm": 2.765625, + "learning_rate": 1.2673056422325413e-08, + "loss": 0.7549, + "step": 2908 + }, + { + "epoch": 3.83, + "grad_norm": 2.828125, + "learning_rate": 1.2410627054013913e-08, + "loss": 0.772, + "step": 2909 + }, + { + "epoch": 3.83, + "grad_norm": 2.796875, + "learning_rate": 1.2150936577065008e-08, + "loss": 0.766, + "step": 2910 + }, + { + "epoch": 3.83, + "grad_norm": 2.703125, + "learning_rate": 1.1893985277403041e-08, + "loss": 0.704, + "step": 2911 + }, + { + "epoch": 3.83, + "grad_norm": 2.84375, + "learning_rate": 1.1639773437937263e-08, + "loss": 0.7266, + "step": 2912 + }, + { + "epoch": 3.83, + "grad_norm": 2.796875, + "learning_rate": 1.1388301338561004e-08, + "loss": 0.7554, + "step": 2913 + }, + { + "epoch": 3.83, + "grad_norm": 2.703125, + "learning_rate": 1.1139569256150285e-08, + "loss": 0.7383, + "step": 2914 + }, + { + "epoch": 3.83, + "grad_norm": 2.796875, + "learning_rate": 1.0893577464564654e-08, + "loss": 0.7462, + "step": 2915 + }, + { + "epoch": 3.84, + "grad_norm": 2.796875, + "learning_rate": 1.0650326234646347e-08, + "loss": 0.7399, + "step": 2916 + }, + { + "epoch": 3.84, + "grad_norm": 2.84375, + "learning_rate": 1.0409815834221126e-08, + "loss": 0.7402, + "step": 2917 + }, + { + "epoch": 3.84, + "grad_norm": 2.75, + "learning_rate": 1.0172046528095502e-08, + "loss": 0.7301, + "step": 2918 + }, + { + "epoch": 3.84, + "grad_norm": 2.671875, + "learning_rate": 9.937018578059232e-09, + "loss": 0.6891, + "step": 2919 + }, + { + "epoch": 3.84, + "grad_norm": 2.71875, + "learning_rate": 9.704732242883374e-09, + "loss": 0.7348, + "step": 2920 + }, + { + "epoch": 3.84, + "grad_norm": 2.75, + "learning_rate": 9.475187778320017e-09, + "loss": 0.7366, + "step": 2921 + }, + { + "epoch": 3.84, + "grad_norm": 2.796875, + "learning_rate": 9.248385437103102e-09, + "loss": 0.7531, + "step": 2922 + }, + { + "epoch": 3.84, + "grad_norm": 2.8125, + "learning_rate": 9.024325468946771e-09, + "loss": 0.7097, + "step": 2923 + }, + { + "epoch": 3.85, + "grad_norm": 2.8125, + "learning_rate": 8.803008120546186e-09, + "loss": 0.7508, + "step": 2924 + }, + { + "epoch": 3.85, + "grad_norm": 2.921875, + "learning_rate": 8.584433635576706e-09, + "loss": 0.7541, + "step": 2925 + }, + { + "epoch": 3.85, + "grad_norm": 2.75, + "learning_rate": 8.368602254693603e-09, + "loss": 0.7324, + "step": 2926 + }, + { + "epoch": 3.85, + "grad_norm": 2.6875, + "learning_rate": 8.155514215532067e-09, + "loss": 0.7129, + "step": 2927 + }, + { + "epoch": 3.85, + "grad_norm": 2.734375, + "learning_rate": 7.945169752706927e-09, + "loss": 0.7325, + "step": 2928 + }, + { + "epoch": 3.85, + "grad_norm": 2.765625, + "learning_rate": 7.737569097811814e-09, + "loss": 0.7265, + "step": 2929 + }, + { + "epoch": 3.85, + "grad_norm": 2.703125, + "learning_rate": 7.532712479420279e-09, + "loss": 0.7415, + "step": 2930 + }, + { + "epoch": 3.86, + "grad_norm": 2.78125, + "learning_rate": 7.330600123083287e-09, + "loss": 0.7386, + "step": 2931 + }, + { + "epoch": 3.86, + "grad_norm": 2.875, + "learning_rate": 7.131232251331721e-09, + "loss": 0.7384, + "step": 2932 + }, + { + "epoch": 3.86, + "grad_norm": 2.75, + "learning_rate": 6.9346090836733275e-09, + "loss": 0.7664, + "step": 2933 + }, + { + "epoch": 3.86, + "grad_norm": 2.796875, + "learning_rate": 6.740730836595211e-09, + "loss": 0.7617, + "step": 2934 + }, + { + "epoch": 3.86, + "grad_norm": 2.828125, + "learning_rate": 6.5495977235613405e-09, + "loss": 0.7337, + "step": 2935 + }, + { + "epoch": 3.86, + "grad_norm": 2.765625, + "learning_rate": 6.361209955013659e-09, + "loss": 0.7311, + "step": 2936 + }, + { + "epoch": 3.86, + "grad_norm": 2.734375, + "learning_rate": 6.1755677383712465e-09, + "loss": 0.735, + "step": 2937 + }, + { + "epoch": 3.86, + "grad_norm": 2.78125, + "learning_rate": 5.992671278030327e-09, + "loss": 0.7274, + "step": 2938 + }, + { + "epoch": 3.87, + "grad_norm": 2.765625, + "learning_rate": 5.812520775363984e-09, + "loss": 0.7217, + "step": 2939 + }, + { + "epoch": 3.87, + "grad_norm": 2.828125, + "learning_rate": 5.635116428722165e-09, + "loss": 0.7365, + "step": 2940 + }, + { + "epoch": 3.87, + "grad_norm": 2.859375, + "learning_rate": 5.460458433430571e-09, + "loss": 0.7187, + "step": 2941 + }, + { + "epoch": 3.87, + "grad_norm": 2.78125, + "learning_rate": 5.288546981791764e-09, + "loss": 0.7244, + "step": 2942 + }, + { + "epoch": 3.87, + "grad_norm": 2.78125, + "learning_rate": 5.1193822630840605e-09, + "loss": 0.7478, + "step": 2943 + }, + { + "epoch": 3.87, + "grad_norm": 2.78125, + "learning_rate": 4.952964463561805e-09, + "loss": 0.763, + "step": 2944 + }, + { + "epoch": 3.87, + "grad_norm": 2.671875, + "learning_rate": 4.789293766454262e-09, + "loss": 0.7294, + "step": 2945 + }, + { + "epoch": 3.88, + "grad_norm": 2.78125, + "learning_rate": 4.628370351967004e-09, + "loss": 0.7239, + "step": 2946 + }, + { + "epoch": 3.88, + "grad_norm": 2.703125, + "learning_rate": 4.47019439727997e-09, + "loss": 0.7102, + "step": 2947 + }, + { + "epoch": 3.88, + "grad_norm": 2.796875, + "learning_rate": 4.314766076548571e-09, + "loss": 0.719, + "step": 2948 + }, + { + "epoch": 3.88, + "grad_norm": 2.875, + "learning_rate": 4.162085560902585e-09, + "loss": 0.7664, + "step": 2949 + }, + { + "epoch": 3.88, + "grad_norm": 2.75, + "learning_rate": 4.012153018446984e-09, + "loss": 0.7484, + "step": 2950 + }, + { + "epoch": 3.88, + "grad_norm": 2.921875, + "learning_rate": 3.864968614260834e-09, + "loss": 0.7535, + "step": 2951 + }, + { + "epoch": 3.88, + "grad_norm": 2.75, + "learning_rate": 3.720532510397279e-09, + "loss": 0.7329, + "step": 2952 + }, + { + "epoch": 3.88, + "grad_norm": 2.765625, + "learning_rate": 3.5788448658838347e-09, + "loss": 0.7396, + "step": 2953 + }, + { + "epoch": 3.89, + "grad_norm": 2.75, + "learning_rate": 3.4399058367215466e-09, + "loss": 0.7397, + "step": 2954 + }, + { + "epoch": 3.89, + "grad_norm": 2.796875, + "learning_rate": 3.3037155758855466e-09, + "loss": 0.7373, + "step": 2955 + }, + { + "epoch": 3.89, + "grad_norm": 2.765625, + "learning_rate": 3.170274233324222e-09, + "loss": 0.7425, + "step": 2956 + }, + { + "epoch": 3.89, + "grad_norm": 2.703125, + "learning_rate": 3.039581955959492e-09, + "loss": 0.7203, + "step": 2957 + }, + { + "epoch": 3.89, + "grad_norm": 2.734375, + "learning_rate": 2.9116388876868075e-09, + "loss": 0.7311, + "step": 2958 + }, + { + "epoch": 3.89, + "grad_norm": 2.828125, + "learning_rate": 2.7864451693740412e-09, + "loss": 0.7643, + "step": 2959 + }, + { + "epoch": 3.89, + "grad_norm": 2.75, + "learning_rate": 2.6640009388620434e-09, + "loss": 0.753, + "step": 2960 + }, + { + "epoch": 3.9, + "grad_norm": 2.78125, + "learning_rate": 2.544306330965196e-09, + "loss": 0.7482, + "step": 2961 + }, + { + "epoch": 3.9, + "grad_norm": 2.75, + "learning_rate": 2.4273614774691923e-09, + "loss": 0.7422, + "step": 2962 + }, + { + "epoch": 3.9, + "grad_norm": 2.734375, + "learning_rate": 2.3131665071338127e-09, + "loss": 0.7334, + "step": 2963 + }, + { + "epoch": 3.9, + "grad_norm": 2.71875, + "learning_rate": 2.2017215456895947e-09, + "loss": 0.7389, + "step": 2964 + }, + { + "epoch": 3.9, + "grad_norm": 2.734375, + "learning_rate": 2.093026715840052e-09, + "loss": 0.7329, + "step": 2965 + }, + { + "epoch": 3.9, + "grad_norm": 2.71875, + "learning_rate": 1.987082137260843e-09, + "loss": 0.7374, + "step": 2966 + }, + { + "epoch": 3.9, + "grad_norm": 2.78125, + "learning_rate": 1.883887926598937e-09, + "loss": 0.7191, + "step": 2967 + }, + { + "epoch": 3.9, + "grad_norm": 2.859375, + "learning_rate": 1.7834441974740047e-09, + "loss": 0.7467, + "step": 2968 + }, + { + "epoch": 3.91, + "grad_norm": 2.703125, + "learning_rate": 1.6857510604764704e-09, + "loss": 0.7067, + "step": 2969 + }, + { + "epoch": 3.91, + "grad_norm": 2.828125, + "learning_rate": 1.590808623168627e-09, + "loss": 0.7633, + "step": 2970 + }, + { + "epoch": 3.91, + "grad_norm": 2.8125, + "learning_rate": 1.4986169900840785e-09, + "loss": 0.7312, + "step": 2971 + }, + { + "epoch": 3.91, + "grad_norm": 2.6875, + "learning_rate": 1.4091762627282956e-09, + "loss": 0.7373, + "step": 2972 + }, + { + "epoch": 3.91, + "grad_norm": 2.828125, + "learning_rate": 1.3224865395775054e-09, + "loss": 0.7671, + "step": 2973 + }, + { + "epoch": 3.91, + "grad_norm": 2.765625, + "learning_rate": 1.2385479160784141e-09, + "loss": 0.7335, + "step": 2974 + }, + { + "epoch": 3.91, + "grad_norm": 2.765625, + "learning_rate": 1.1573604846501496e-09, + "loss": 0.7544, + "step": 2975 + }, + { + "epoch": 3.92, + "grad_norm": 2.828125, + "learning_rate": 1.0789243346812083e-09, + "loss": 0.7404, + "step": 2976 + }, + { + "epoch": 3.92, + "grad_norm": 2.78125, + "learning_rate": 1.0032395525316762e-09, + "loss": 0.7736, + "step": 2977 + }, + { + "epoch": 3.92, + "grad_norm": 2.796875, + "learning_rate": 9.303062215323955e-10, + "loss": 0.7172, + "step": 2978 + }, + { + "epoch": 3.92, + "grad_norm": 2.921875, + "learning_rate": 8.601244219846872e-10, + "loss": 0.746, + "step": 2979 + }, + { + "epoch": 3.92, + "grad_norm": 2.859375, + "learning_rate": 7.926942311597962e-10, + "loss": 0.7433, + "step": 2980 + }, + { + "epoch": 3.92, + "grad_norm": 2.796875, + "learning_rate": 7.280157233002794e-10, + "loss": 0.7349, + "step": 2981 + }, + { + "epoch": 3.92, + "grad_norm": 2.703125, + "learning_rate": 6.660889696186168e-10, + "loss": 0.742, + "step": 2982 + }, + { + "epoch": 3.92, + "grad_norm": 2.8125, + "learning_rate": 6.069140382972128e-10, + "loss": 0.7588, + "step": 2983 + }, + { + "epoch": 3.93, + "grad_norm": 2.875, + "learning_rate": 5.504909944895054e-10, + "loss": 0.7558, + "step": 2984 + }, + { + "epoch": 3.93, + "grad_norm": 2.796875, + "learning_rate": 4.968199003177466e-10, + "loss": 0.7512, + "step": 2985 + }, + { + "epoch": 3.93, + "grad_norm": 2.796875, + "learning_rate": 4.4590081487577706e-10, + "loss": 0.7457, + "step": 2986 + }, + { + "epoch": 3.93, + "grad_norm": 2.890625, + "learning_rate": 3.9773379422625156e-10, + "loss": 0.7603, + "step": 2987 + }, + { + "epoch": 3.93, + "grad_norm": 2.828125, + "learning_rate": 3.5231889140230347e-10, + "loss": 0.7217, + "step": 2988 + }, + { + "epoch": 3.93, + "grad_norm": 2.734375, + "learning_rate": 3.0965615640671243e-10, + "loss": 0.7448, + "step": 2989 + }, + { + "epoch": 3.93, + "grad_norm": 2.71875, + "learning_rate": 2.697456362119044e-10, + "loss": 0.7266, + "step": 2990 + }, + { + "epoch": 3.94, + "grad_norm": 2.75, + "learning_rate": 2.3258737476050674e-10, + "loss": 0.7257, + "step": 2991 + }, + { + "epoch": 3.94, + "grad_norm": 2.78125, + "learning_rate": 1.9818141296451544e-10, + "loss": 0.7411, + "step": 2992 + }, + { + "epoch": 3.94, + "grad_norm": 2.84375, + "learning_rate": 1.6652778870557274e-10, + "loss": 0.7428, + "step": 2993 + }, + { + "epoch": 3.94, + "grad_norm": 2.796875, + "learning_rate": 1.3762653683496718e-10, + "loss": 0.7325, + "step": 2994 + }, + { + "epoch": 3.94, + "grad_norm": 2.796875, + "learning_rate": 1.1147768917391111e-10, + "loss": 0.7328, + "step": 2995 + }, + { + "epoch": 3.94, + "grad_norm": 2.765625, + "learning_rate": 8.808127451270798e-11, + "loss": 0.7472, + "step": 2996 + }, + { + "epoch": 3.94, + "grad_norm": 2.765625, + "learning_rate": 6.743731861130753e-11, + "loss": 0.7252, + "step": 2997 + }, + { + "epoch": 3.94, + "grad_norm": 2.6875, + "learning_rate": 4.954584419930575e-11, + "loss": 0.7301, + "step": 2998 + }, + { + "epoch": 3.95, + "grad_norm": 2.71875, + "learning_rate": 3.440687097538975e-11, + "loss": 0.7206, + "step": 2999 + }, + { + "epoch": 3.95, + "grad_norm": 2.734375, + "learning_rate": 2.202041560789292e-11, + "loss": 0.7362, + "step": 3000 + }, + { + "epoch": 3.95, + "grad_norm": 2.828125, + "learning_rate": 1.2386491734794891e-11, + "loss": 0.7393, + "step": 3001 + }, + { + "epoch": 3.95, + "grad_norm": 2.796875, + "learning_rate": 5.505109963166444e-12, + "loss": 0.7456, + "step": 3002 + }, + { + "epoch": 3.95, + "grad_norm": 2.875, + "learning_rate": 1.376277869724607e-12, + "loss": 0.7549, + "step": 3003 + }, + { + "epoch": 3.95, + "grad_norm": 2.65625, + "learning_rate": 0.0, + "loss": 0.7152, + "step": 3004 + } + ], + "logging_steps": 1, + "max_steps": 3004, + "num_input_tokens_seen": 0, + "num_train_epochs": 4, + "save_steps": 751, + "total_flos": 3.359704024882977e+19, + "train_batch_size": 2, + "trial_name": null, + "trial_params": null +} diff --git a/last-checkpoint/training_args.bin b/last-checkpoint/training_args.bin new file mode 100644 index 0000000..68b1431 --- /dev/null +++ b/last-checkpoint/training_args.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:00d5675931773fafeff74351dc5a9cbbf81577167988949e58fbb52a3bd05d94 +size 5880 diff --git a/model-00001-of-00003.safetensors b/model-00001-of-00003.safetensors new file mode 100644 index 0000000..7a7a35a --- /dev/null +++ b/model-00001-of-00003.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:1918667d6c8ea19c91894aed25da7c6c3a707bbca0e11dcfba9225a78109ed5c +size 4943162336 diff --git a/model-00002-of-00003.safetensors b/model-00002-of-00003.safetensors new file mode 100644 index 0000000..9178ff6 --- /dev/null +++ b/model-00002-of-00003.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:7abfc057a799175c2b8cb179b3f28455cb1f2ded40c9ffab62d44b51b4b54db7 +size 4999819336 diff --git a/model-00003-of-00003.safetensors b/model-00003-of-00003.safetensors new file mode 100644 index 0000000..0c7ecde --- /dev/null +++ b/model-00003-of-00003.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:82c09b650cdaaf22e170b97b68130c35b65aa74418387c100a25980f525e1835 +size 4540516344 diff --git a/model.safetensors.index.json b/model.safetensors.index.json new file mode 100644 index 0000000..b349bc0 --- /dev/null +++ b/model.safetensors.index.json @@ -0,0 +1,298 @@ +{ + "metadata": { + "total_size": 14483464192 + }, + "weight_map": { + "lm_head.weight": "model-00003-of-00003.safetensors", + "model.embed_tokens.weight": "model-00001-of-00003.safetensors", + "model.layers.0.input_layernorm.weight": "model-00001-of-00003.safetensors", + "model.layers.0.mlp.down_proj.weight": "model-00001-of-00003.safetensors", + "model.layers.0.mlp.gate_proj.weight": "model-00001-of-00003.safetensors", + "model.layers.0.mlp.up_proj.weight": "model-00001-of-00003.safetensors", + "model.layers.0.post_attention_layernorm.weight": "model-00001-of-00003.safetensors", + "model.layers.0.self_attn.k_proj.weight": "model-00001-of-00003.safetensors", + "model.layers.0.self_attn.o_proj.weight": "model-00001-of-00003.safetensors", + "model.layers.0.self_attn.q_proj.weight": "model-00001-of-00003.safetensors", + "model.layers.0.self_attn.v_proj.weight": "model-00001-of-00003.safetensors", + "model.layers.1.input_layernorm.weight": "model-00001-of-00003.safetensors", + "model.layers.1.mlp.down_proj.weight": "model-00001-of-00003.safetensors", + "model.layers.1.mlp.gate_proj.weight": "model-00001-of-00003.safetensors", + "model.layers.1.mlp.up_proj.weight": "model-00001-of-00003.safetensors", + "model.layers.1.post_attention_layernorm.weight": "model-00001-of-00003.safetensors", + "model.layers.1.self_attn.k_proj.weight": "model-00001-of-00003.safetensors", + "model.layers.1.self_attn.o_proj.weight": "model-00001-of-00003.safetensors", + "model.layers.1.self_attn.q_proj.weight": "model-00001-of-00003.safetensors", + "model.layers.1.self_attn.v_proj.weight": "model-00001-of-00003.safetensors", + "model.layers.10.input_layernorm.weight": "model-00002-of-00003.safetensors", + "model.layers.10.mlp.down_proj.weight": "model-00002-of-00003.safetensors", + "model.layers.10.mlp.gate_proj.weight": "model-00001-of-00003.safetensors", + "model.layers.10.mlp.up_proj.weight": "model-00001-of-00003.safetensors", + "model.layers.10.post_attention_layernorm.weight": "model-00002-of-00003.safetensors", + "model.layers.10.self_attn.k_proj.weight": "model-00001-of-00003.safetensors", + "model.layers.10.self_attn.o_proj.weight": "model-00001-of-00003.safetensors", + "model.layers.10.self_attn.q_proj.weight": "model-00001-of-00003.safetensors", + "model.layers.10.self_attn.v_proj.weight": "model-00001-of-00003.safetensors", + "model.layers.11.input_layernorm.weight": "model-00002-of-00003.safetensors", + "model.layers.11.mlp.down_proj.weight": "model-00002-of-00003.safetensors", + "model.layers.11.mlp.gate_proj.weight": "model-00002-of-00003.safetensors", + "model.layers.11.mlp.up_proj.weight": "model-00002-of-00003.safetensors", + "model.layers.11.post_attention_layernorm.weight": "model-00002-of-00003.safetensors", + "model.layers.11.self_attn.k_proj.weight": "model-00002-of-00003.safetensors", + "model.layers.11.self_attn.o_proj.weight": "model-00002-of-00003.safetensors", + "model.layers.11.self_attn.q_proj.weight": "model-00002-of-00003.safetensors", + "model.layers.11.self_attn.v_proj.weight": "model-00002-of-00003.safetensors", + "model.layers.12.input_layernorm.weight": "model-00002-of-00003.safetensors", + "model.layers.12.mlp.down_proj.weight": "model-00002-of-00003.safetensors", + "model.layers.12.mlp.gate_proj.weight": "model-00002-of-00003.safetensors", + "model.layers.12.mlp.up_proj.weight": "model-00002-of-00003.safetensors", + "model.layers.12.post_attention_layernorm.weight": "model-00002-of-00003.safetensors", + "model.layers.12.self_attn.k_proj.weight": "model-00002-of-00003.safetensors", + "model.layers.12.self_attn.o_proj.weight": "model-00002-of-00003.safetensors", + "model.layers.12.self_attn.q_proj.weight": "model-00002-of-00003.safetensors", + "model.layers.12.self_attn.v_proj.weight": "model-00002-of-00003.safetensors", + "model.layers.13.input_layernorm.weight": "model-00002-of-00003.safetensors", + "model.layers.13.mlp.down_proj.weight": "model-00002-of-00003.safetensors", + "model.layers.13.mlp.gate_proj.weight": "model-00002-of-00003.safetensors", + "model.layers.13.mlp.up_proj.weight": "model-00002-of-00003.safetensors", + "model.layers.13.post_attention_layernorm.weight": "model-00002-of-00003.safetensors", + "model.layers.13.self_attn.k_proj.weight": "model-00002-of-00003.safetensors", + "model.layers.13.self_attn.o_proj.weight": "model-00002-of-00003.safetensors", + "model.layers.13.self_attn.q_proj.weight": "model-00002-of-00003.safetensors", + "model.layers.13.self_attn.v_proj.weight": "model-00002-of-00003.safetensors", + "model.layers.14.input_layernorm.weight": "model-00002-of-00003.safetensors", + "model.layers.14.mlp.down_proj.weight": "model-00002-of-00003.safetensors", + "model.layers.14.mlp.gate_proj.weight": "model-00002-of-00003.safetensors", + "model.layers.14.mlp.up_proj.weight": "model-00002-of-00003.safetensors", + "model.layers.14.post_attention_layernorm.weight": "model-00002-of-00003.safetensors", + "model.layers.14.self_attn.k_proj.weight": "model-00002-of-00003.safetensors", + "model.layers.14.self_attn.o_proj.weight": "model-00002-of-00003.safetensors", + "model.layers.14.self_attn.q_proj.weight": "model-00002-of-00003.safetensors", + "model.layers.14.self_attn.v_proj.weight": "model-00002-of-00003.safetensors", + "model.layers.15.input_layernorm.weight": "model-00002-of-00003.safetensors", + "model.layers.15.mlp.down_proj.weight": "model-00002-of-00003.safetensors", + "model.layers.15.mlp.gate_proj.weight": "model-00002-of-00003.safetensors", + "model.layers.15.mlp.up_proj.weight": "model-00002-of-00003.safetensors", + "model.layers.15.post_attention_layernorm.weight": "model-00002-of-00003.safetensors", + "model.layers.15.self_attn.k_proj.weight": "model-00002-of-00003.safetensors", + "model.layers.15.self_attn.o_proj.weight": "model-00002-of-00003.safetensors", + "model.layers.15.self_attn.q_proj.weight": "model-00002-of-00003.safetensors", + "model.layers.15.self_attn.v_proj.weight": "model-00002-of-00003.safetensors", + "model.layers.16.input_layernorm.weight": "model-00002-of-00003.safetensors", + "model.layers.16.mlp.down_proj.weight": "model-00002-of-00003.safetensors", + "model.layers.16.mlp.gate_proj.weight": "model-00002-of-00003.safetensors", + "model.layers.16.mlp.up_proj.weight": "model-00002-of-00003.safetensors", + "model.layers.16.post_attention_layernorm.weight": "model-00002-of-00003.safetensors", + "model.layers.16.self_attn.k_proj.weight": "model-00002-of-00003.safetensors", + "model.layers.16.self_attn.o_proj.weight": "model-00002-of-00003.safetensors", + "model.layers.16.self_attn.q_proj.weight": "model-00002-of-00003.safetensors", + "model.layers.16.self_attn.v_proj.weight": "model-00002-of-00003.safetensors", + "model.layers.17.input_layernorm.weight": "model-00002-of-00003.safetensors", + "model.layers.17.mlp.down_proj.weight": "model-00002-of-00003.safetensors", + "model.layers.17.mlp.gate_proj.weight": "model-00002-of-00003.safetensors", + "model.layers.17.mlp.up_proj.weight": "model-00002-of-00003.safetensors", + "model.layers.17.post_attention_layernorm.weight": "model-00002-of-00003.safetensors", + "model.layers.17.self_attn.k_proj.weight": "model-00002-of-00003.safetensors", + "model.layers.17.self_attn.o_proj.weight": "model-00002-of-00003.safetensors", + "model.layers.17.self_attn.q_proj.weight": "model-00002-of-00003.safetensors", + "model.layers.17.self_attn.v_proj.weight": "model-00002-of-00003.safetensors", + "model.layers.18.input_layernorm.weight": "model-00002-of-00003.safetensors", + "model.layers.18.mlp.down_proj.weight": "model-00002-of-00003.safetensors", + "model.layers.18.mlp.gate_proj.weight": "model-00002-of-00003.safetensors", + "model.layers.18.mlp.up_proj.weight": "model-00002-of-00003.safetensors", + "model.layers.18.post_attention_layernorm.weight": "model-00002-of-00003.safetensors", + "model.layers.18.self_attn.k_proj.weight": "model-00002-of-00003.safetensors", + "model.layers.18.self_attn.o_proj.weight": "model-00002-of-00003.safetensors", + "model.layers.18.self_attn.q_proj.weight": "model-00002-of-00003.safetensors", + "model.layers.18.self_attn.v_proj.weight": "model-00002-of-00003.safetensors", + "model.layers.19.input_layernorm.weight": "model-00002-of-00003.safetensors", + "model.layers.19.mlp.down_proj.weight": "model-00002-of-00003.safetensors", + "model.layers.19.mlp.gate_proj.weight": "model-00002-of-00003.safetensors", + "model.layers.19.mlp.up_proj.weight": "model-00002-of-00003.safetensors", + "model.layers.19.post_attention_layernorm.weight": "model-00002-of-00003.safetensors", + "model.layers.19.self_attn.k_proj.weight": "model-00002-of-00003.safetensors", + "model.layers.19.self_attn.o_proj.weight": "model-00002-of-00003.safetensors", + "model.layers.19.self_attn.q_proj.weight": "model-00002-of-00003.safetensors", + "model.layers.19.self_attn.v_proj.weight": "model-00002-of-00003.safetensors", + "model.layers.2.input_layernorm.weight": "model-00001-of-00003.safetensors", + "model.layers.2.mlp.down_proj.weight": "model-00001-of-00003.safetensors", + "model.layers.2.mlp.gate_proj.weight": "model-00001-of-00003.safetensors", + "model.layers.2.mlp.up_proj.weight": "model-00001-of-00003.safetensors", + "model.layers.2.post_attention_layernorm.weight": "model-00001-of-00003.safetensors", + "model.layers.2.self_attn.k_proj.weight": "model-00001-of-00003.safetensors", + "model.layers.2.self_attn.o_proj.weight": "model-00001-of-00003.safetensors", + "model.layers.2.self_attn.q_proj.weight": "model-00001-of-00003.safetensors", + "model.layers.2.self_attn.v_proj.weight": "model-00001-of-00003.safetensors", + "model.layers.20.input_layernorm.weight": "model-00002-of-00003.safetensors", + "model.layers.20.mlp.down_proj.weight": "model-00002-of-00003.safetensors", + "model.layers.20.mlp.gate_proj.weight": "model-00002-of-00003.safetensors", + "model.layers.20.mlp.up_proj.weight": "model-00002-of-00003.safetensors", + "model.layers.20.post_attention_layernorm.weight": "model-00002-of-00003.safetensors", + "model.layers.20.self_attn.k_proj.weight": "model-00002-of-00003.safetensors", + "model.layers.20.self_attn.o_proj.weight": "model-00002-of-00003.safetensors", + "model.layers.20.self_attn.q_proj.weight": "model-00002-of-00003.safetensors", + "model.layers.20.self_attn.v_proj.weight": "model-00002-of-00003.safetensors", + "model.layers.21.input_layernorm.weight": "model-00002-of-00003.safetensors", + "model.layers.21.mlp.down_proj.weight": "model-00002-of-00003.safetensors", + "model.layers.21.mlp.gate_proj.weight": "model-00002-of-00003.safetensors", + "model.layers.21.mlp.up_proj.weight": "model-00002-of-00003.safetensors", + "model.layers.21.post_attention_layernorm.weight": "model-00002-of-00003.safetensors", + "model.layers.21.self_attn.k_proj.weight": "model-00002-of-00003.safetensors", + "model.layers.21.self_attn.o_proj.weight": "model-00002-of-00003.safetensors", + "model.layers.21.self_attn.q_proj.weight": "model-00002-of-00003.safetensors", + "model.layers.21.self_attn.v_proj.weight": "model-00002-of-00003.safetensors", + "model.layers.22.input_layernorm.weight": "model-00003-of-00003.safetensors", + "model.layers.22.mlp.down_proj.weight": "model-00003-of-00003.safetensors", + "model.layers.22.mlp.gate_proj.weight": "model-00003-of-00003.safetensors", + "model.layers.22.mlp.up_proj.weight": "model-00003-of-00003.safetensors", + "model.layers.22.post_attention_layernorm.weight": "model-00003-of-00003.safetensors", + "model.layers.22.self_attn.k_proj.weight": "model-00002-of-00003.safetensors", + "model.layers.22.self_attn.o_proj.weight": "model-00002-of-00003.safetensors", + "model.layers.22.self_attn.q_proj.weight": "model-00002-of-00003.safetensors", + "model.layers.22.self_attn.v_proj.weight": "model-00002-of-00003.safetensors", + "model.layers.23.input_layernorm.weight": "model-00003-of-00003.safetensors", + "model.layers.23.mlp.down_proj.weight": "model-00003-of-00003.safetensors", + "model.layers.23.mlp.gate_proj.weight": "model-00003-of-00003.safetensors", + "model.layers.23.mlp.up_proj.weight": "model-00003-of-00003.safetensors", + "model.layers.23.post_attention_layernorm.weight": "model-00003-of-00003.safetensors", + "model.layers.23.self_attn.k_proj.weight": "model-00003-of-00003.safetensors", + "model.layers.23.self_attn.o_proj.weight": "model-00003-of-00003.safetensors", + "model.layers.23.self_attn.q_proj.weight": "model-00003-of-00003.safetensors", + "model.layers.23.self_attn.v_proj.weight": "model-00003-of-00003.safetensors", + "model.layers.24.input_layernorm.weight": "model-00003-of-00003.safetensors", + "model.layers.24.mlp.down_proj.weight": "model-00003-of-00003.safetensors", + "model.layers.24.mlp.gate_proj.weight": "model-00003-of-00003.safetensors", + "model.layers.24.mlp.up_proj.weight": "model-00003-of-00003.safetensors", + "model.layers.24.post_attention_layernorm.weight": "model-00003-of-00003.safetensors", + "model.layers.24.self_attn.k_proj.weight": "model-00003-of-00003.safetensors", + "model.layers.24.self_attn.o_proj.weight": "model-00003-of-00003.safetensors", + "model.layers.24.self_attn.q_proj.weight": "model-00003-of-00003.safetensors", + "model.layers.24.self_attn.v_proj.weight": "model-00003-of-00003.safetensors", + "model.layers.25.input_layernorm.weight": "model-00003-of-00003.safetensors", + "model.layers.25.mlp.down_proj.weight": "model-00003-of-00003.safetensors", + "model.layers.25.mlp.gate_proj.weight": "model-00003-of-00003.safetensors", + "model.layers.25.mlp.up_proj.weight": "model-00003-of-00003.safetensors", + "model.layers.25.post_attention_layernorm.weight": "model-00003-of-00003.safetensors", + "model.layers.25.self_attn.k_proj.weight": "model-00003-of-00003.safetensors", + "model.layers.25.self_attn.o_proj.weight": "model-00003-of-00003.safetensors", + "model.layers.25.self_attn.q_proj.weight": "model-00003-of-00003.safetensors", + "model.layers.25.self_attn.v_proj.weight": "model-00003-of-00003.safetensors", + "model.layers.26.input_layernorm.weight": "model-00003-of-00003.safetensors", + "model.layers.26.mlp.down_proj.weight": "model-00003-of-00003.safetensors", + "model.layers.26.mlp.gate_proj.weight": "model-00003-of-00003.safetensors", + "model.layers.26.mlp.up_proj.weight": "model-00003-of-00003.safetensors", + "model.layers.26.post_attention_layernorm.weight": "model-00003-of-00003.safetensors", + "model.layers.26.self_attn.k_proj.weight": "model-00003-of-00003.safetensors", + "model.layers.26.self_attn.o_proj.weight": "model-00003-of-00003.safetensors", + "model.layers.26.self_attn.q_proj.weight": "model-00003-of-00003.safetensors", + "model.layers.26.self_attn.v_proj.weight": "model-00003-of-00003.safetensors", + "model.layers.27.input_layernorm.weight": "model-00003-of-00003.safetensors", + "model.layers.27.mlp.down_proj.weight": "model-00003-of-00003.safetensors", + "model.layers.27.mlp.gate_proj.weight": "model-00003-of-00003.safetensors", + "model.layers.27.mlp.up_proj.weight": "model-00003-of-00003.safetensors", + "model.layers.27.post_attention_layernorm.weight": "model-00003-of-00003.safetensors", + "model.layers.27.self_attn.k_proj.weight": "model-00003-of-00003.safetensors", + "model.layers.27.self_attn.o_proj.weight": "model-00003-of-00003.safetensors", + "model.layers.27.self_attn.q_proj.weight": "model-00003-of-00003.safetensors", + "model.layers.27.self_attn.v_proj.weight": "model-00003-of-00003.safetensors", + "model.layers.28.input_layernorm.weight": "model-00003-of-00003.safetensors", + "model.layers.28.mlp.down_proj.weight": "model-00003-of-00003.safetensors", + "model.layers.28.mlp.gate_proj.weight": "model-00003-of-00003.safetensors", + "model.layers.28.mlp.up_proj.weight": "model-00003-of-00003.safetensors", + "model.layers.28.post_attention_layernorm.weight": "model-00003-of-00003.safetensors", + "model.layers.28.self_attn.k_proj.weight": "model-00003-of-00003.safetensors", + "model.layers.28.self_attn.o_proj.weight": "model-00003-of-00003.safetensors", + "model.layers.28.self_attn.q_proj.weight": "model-00003-of-00003.safetensors", + "model.layers.28.self_attn.v_proj.weight": "model-00003-of-00003.safetensors", + "model.layers.29.input_layernorm.weight": "model-00003-of-00003.safetensors", + "model.layers.29.mlp.down_proj.weight": "model-00003-of-00003.safetensors", + "model.layers.29.mlp.gate_proj.weight": "model-00003-of-00003.safetensors", + "model.layers.29.mlp.up_proj.weight": "model-00003-of-00003.safetensors", + "model.layers.29.post_attention_layernorm.weight": "model-00003-of-00003.safetensors", + "model.layers.29.self_attn.k_proj.weight": "model-00003-of-00003.safetensors", + "model.layers.29.self_attn.o_proj.weight": "model-00003-of-00003.safetensors", + "model.layers.29.self_attn.q_proj.weight": "model-00003-of-00003.safetensors", + "model.layers.29.self_attn.v_proj.weight": "model-00003-of-00003.safetensors", + "model.layers.3.input_layernorm.weight": "model-00001-of-00003.safetensors", + "model.layers.3.mlp.down_proj.weight": "model-00001-of-00003.safetensors", + "model.layers.3.mlp.gate_proj.weight": "model-00001-of-00003.safetensors", + "model.layers.3.mlp.up_proj.weight": "model-00001-of-00003.safetensors", + "model.layers.3.post_attention_layernorm.weight": "model-00001-of-00003.safetensors", + "model.layers.3.self_attn.k_proj.weight": "model-00001-of-00003.safetensors", + "model.layers.3.self_attn.o_proj.weight": "model-00001-of-00003.safetensors", + "model.layers.3.self_attn.q_proj.weight": "model-00001-of-00003.safetensors", + "model.layers.3.self_attn.v_proj.weight": "model-00001-of-00003.safetensors", + "model.layers.30.input_layernorm.weight": "model-00003-of-00003.safetensors", + "model.layers.30.mlp.down_proj.weight": "model-00003-of-00003.safetensors", + "model.layers.30.mlp.gate_proj.weight": "model-00003-of-00003.safetensors", + "model.layers.30.mlp.up_proj.weight": "model-00003-of-00003.safetensors", + "model.layers.30.post_attention_layernorm.weight": "model-00003-of-00003.safetensors", + "model.layers.30.self_attn.k_proj.weight": "model-00003-of-00003.safetensors", + "model.layers.30.self_attn.o_proj.weight": "model-00003-of-00003.safetensors", + "model.layers.30.self_attn.q_proj.weight": "model-00003-of-00003.safetensors", + "model.layers.30.self_attn.v_proj.weight": "model-00003-of-00003.safetensors", + "model.layers.31.input_layernorm.weight": "model-00003-of-00003.safetensors", + "model.layers.31.mlp.down_proj.weight": "model-00003-of-00003.safetensors", + "model.layers.31.mlp.gate_proj.weight": "model-00003-of-00003.safetensors", + "model.layers.31.mlp.up_proj.weight": "model-00003-of-00003.safetensors", + "model.layers.31.post_attention_layernorm.weight": "model-00003-of-00003.safetensors", + "model.layers.31.self_attn.k_proj.weight": "model-00003-of-00003.safetensors", + "model.layers.31.self_attn.o_proj.weight": "model-00003-of-00003.safetensors", + "model.layers.31.self_attn.q_proj.weight": "model-00003-of-00003.safetensors", + "model.layers.31.self_attn.v_proj.weight": "model-00003-of-00003.safetensors", + "model.layers.4.input_layernorm.weight": "model-00001-of-00003.safetensors", + "model.layers.4.mlp.down_proj.weight": "model-00001-of-00003.safetensors", + "model.layers.4.mlp.gate_proj.weight": "model-00001-of-00003.safetensors", + "model.layers.4.mlp.up_proj.weight": "model-00001-of-00003.safetensors", + "model.layers.4.post_attention_layernorm.weight": "model-00001-of-00003.safetensors", + "model.layers.4.self_attn.k_proj.weight": "model-00001-of-00003.safetensors", + "model.layers.4.self_attn.o_proj.weight": "model-00001-of-00003.safetensors", + "model.layers.4.self_attn.q_proj.weight": "model-00001-of-00003.safetensors", + "model.layers.4.self_attn.v_proj.weight": "model-00001-of-00003.safetensors", + "model.layers.5.input_layernorm.weight": "model-00001-of-00003.safetensors", + "model.layers.5.mlp.down_proj.weight": "model-00001-of-00003.safetensors", + "model.layers.5.mlp.gate_proj.weight": "model-00001-of-00003.safetensors", + "model.layers.5.mlp.up_proj.weight": "model-00001-of-00003.safetensors", + "model.layers.5.post_attention_layernorm.weight": "model-00001-of-00003.safetensors", + "model.layers.5.self_attn.k_proj.weight": "model-00001-of-00003.safetensors", + "model.layers.5.self_attn.o_proj.weight": "model-00001-of-00003.safetensors", + "model.layers.5.self_attn.q_proj.weight": "model-00001-of-00003.safetensors", + "model.layers.5.self_attn.v_proj.weight": "model-00001-of-00003.safetensors", + "model.layers.6.input_layernorm.weight": "model-00001-of-00003.safetensors", + "model.layers.6.mlp.down_proj.weight": "model-00001-of-00003.safetensors", + "model.layers.6.mlp.gate_proj.weight": "model-00001-of-00003.safetensors", + "model.layers.6.mlp.up_proj.weight": "model-00001-of-00003.safetensors", + "model.layers.6.post_attention_layernorm.weight": "model-00001-of-00003.safetensors", + "model.layers.6.self_attn.k_proj.weight": "model-00001-of-00003.safetensors", + "model.layers.6.self_attn.o_proj.weight": "model-00001-of-00003.safetensors", + "model.layers.6.self_attn.q_proj.weight": "model-00001-of-00003.safetensors", + "model.layers.6.self_attn.v_proj.weight": "model-00001-of-00003.safetensors", + "model.layers.7.input_layernorm.weight": "model-00001-of-00003.safetensors", + "model.layers.7.mlp.down_proj.weight": "model-00001-of-00003.safetensors", + "model.layers.7.mlp.gate_proj.weight": "model-00001-of-00003.safetensors", + "model.layers.7.mlp.up_proj.weight": "model-00001-of-00003.safetensors", + "model.layers.7.post_attention_layernorm.weight": "model-00001-of-00003.safetensors", + "model.layers.7.self_attn.k_proj.weight": "model-00001-of-00003.safetensors", + "model.layers.7.self_attn.o_proj.weight": "model-00001-of-00003.safetensors", + "model.layers.7.self_attn.q_proj.weight": "model-00001-of-00003.safetensors", + "model.layers.7.self_attn.v_proj.weight": "model-00001-of-00003.safetensors", + "model.layers.8.input_layernorm.weight": "model-00001-of-00003.safetensors", + "model.layers.8.mlp.down_proj.weight": "model-00001-of-00003.safetensors", + "model.layers.8.mlp.gate_proj.weight": "model-00001-of-00003.safetensors", + "model.layers.8.mlp.up_proj.weight": "model-00001-of-00003.safetensors", + "model.layers.8.post_attention_layernorm.weight": "model-00001-of-00003.safetensors", + "model.layers.8.self_attn.k_proj.weight": "model-00001-of-00003.safetensors", + "model.layers.8.self_attn.o_proj.weight": "model-00001-of-00003.safetensors", + "model.layers.8.self_attn.q_proj.weight": "model-00001-of-00003.safetensors", + "model.layers.8.self_attn.v_proj.weight": "model-00001-of-00003.safetensors", + "model.layers.9.input_layernorm.weight": "model-00001-of-00003.safetensors", + "model.layers.9.mlp.down_proj.weight": "model-00001-of-00003.safetensors", + "model.layers.9.mlp.gate_proj.weight": "model-00001-of-00003.safetensors", + "model.layers.9.mlp.up_proj.weight": "model-00001-of-00003.safetensors", + "model.layers.9.post_attention_layernorm.weight": "model-00001-of-00003.safetensors", + "model.layers.9.self_attn.k_proj.weight": "model-00001-of-00003.safetensors", + "model.layers.9.self_attn.o_proj.weight": "model-00001-of-00003.safetensors", + "model.layers.9.self_attn.q_proj.weight": "model-00001-of-00003.safetensors", + "model.layers.9.self_attn.v_proj.weight": "model-00001-of-00003.safetensors", + "model.norm.weight": "model-00003-of-00003.safetensors" + } +} diff --git a/pytorch_model-00001-of-00003.bin b/pytorch_model-00001-of-00003.bin new file mode 100644 index 0000000..3983916 --- /dev/null +++ b/pytorch_model-00001-of-00003.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:e3d2743f123d1cb1f2d4747beeb99e3ab21912a57e7dbd32bea0cef345d7a8c6 +size 4943185632 diff --git a/pytorch_model-00002-of-00003.bin b/pytorch_model-00002-of-00003.bin new file mode 100644 index 0000000..4f45940 --- /dev/null +++ b/pytorch_model-00002-of-00003.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:3b6433ed3e6f3d3b8025c954fbdd51566b43feb3c232c515be90148c630af5db +size 4999844744 diff --git a/pytorch_model-00003-of-00003.bin b/pytorch_model-00003-of-00003.bin new file mode 100644 index 0000000..84f5e17 --- /dev/null +++ b/pytorch_model-00003-of-00003.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:e93e4f4d775814a5708aa647ec80493ff6bdd79a9a5e144667daa624422ffee5 +size 4540537414 diff --git a/pytorch_model.bin.index.json b/pytorch_model.bin.index.json new file mode 100644 index 0000000..76fccc2 --- /dev/null +++ b/pytorch_model.bin.index.json @@ -0,0 +1,298 @@ +{ + "metadata": { + "total_size": 14483464192 + }, + "weight_map": { + "lm_head.weight": "pytorch_model-00003-of-00003.bin", + "model.embed_tokens.weight": "pytorch_model-00001-of-00003.bin", + "model.layers.0.input_layernorm.weight": "pytorch_model-00001-of-00003.bin", + "model.layers.0.mlp.down_proj.weight": "pytorch_model-00001-of-00003.bin", + "model.layers.0.mlp.gate_proj.weight": "pytorch_model-00001-of-00003.bin", + "model.layers.0.mlp.up_proj.weight": "pytorch_model-00001-of-00003.bin", + "model.layers.0.post_attention_layernorm.weight": "pytorch_model-00001-of-00003.bin", + "model.layers.0.self_attn.k_proj.weight": "pytorch_model-00001-of-00003.bin", + "model.layers.0.self_attn.o_proj.weight": "pytorch_model-00001-of-00003.bin", + "model.layers.0.self_attn.q_proj.weight": "pytorch_model-00001-of-00003.bin", + "model.layers.0.self_attn.v_proj.weight": "pytorch_model-00001-of-00003.bin", + "model.layers.1.input_layernorm.weight": "pytorch_model-00001-of-00003.bin", + "model.layers.1.mlp.down_proj.weight": "pytorch_model-00001-of-00003.bin", + "model.layers.1.mlp.gate_proj.weight": "pytorch_model-00001-of-00003.bin", + "model.layers.1.mlp.up_proj.weight": "pytorch_model-00001-of-00003.bin", + "model.layers.1.post_attention_layernorm.weight": "pytorch_model-00001-of-00003.bin", + "model.layers.1.self_attn.k_proj.weight": "pytorch_model-00001-of-00003.bin", + "model.layers.1.self_attn.o_proj.weight": "pytorch_model-00001-of-00003.bin", + "model.layers.1.self_attn.q_proj.weight": "pytorch_model-00001-of-00003.bin", + "model.layers.1.self_attn.v_proj.weight": "pytorch_model-00001-of-00003.bin", + "model.layers.10.input_layernorm.weight": "pytorch_model-00002-of-00003.bin", + "model.layers.10.mlp.down_proj.weight": "pytorch_model-00002-of-00003.bin", + "model.layers.10.mlp.gate_proj.weight": "pytorch_model-00001-of-00003.bin", + "model.layers.10.mlp.up_proj.weight": "pytorch_model-00001-of-00003.bin", + "model.layers.10.post_attention_layernorm.weight": "pytorch_model-00002-of-00003.bin", + "model.layers.10.self_attn.k_proj.weight": "pytorch_model-00001-of-00003.bin", + "model.layers.10.self_attn.o_proj.weight": "pytorch_model-00001-of-00003.bin", + "model.layers.10.self_attn.q_proj.weight": "pytorch_model-00001-of-00003.bin", + "model.layers.10.self_attn.v_proj.weight": "pytorch_model-00001-of-00003.bin", + "model.layers.11.input_layernorm.weight": "pytorch_model-00002-of-00003.bin", + "model.layers.11.mlp.down_proj.weight": "pytorch_model-00002-of-00003.bin", + "model.layers.11.mlp.gate_proj.weight": "pytorch_model-00002-of-00003.bin", + "model.layers.11.mlp.up_proj.weight": "pytorch_model-00002-of-00003.bin", + "model.layers.11.post_attention_layernorm.weight": "pytorch_model-00002-of-00003.bin", + "model.layers.11.self_attn.k_proj.weight": "pytorch_model-00002-of-00003.bin", + "model.layers.11.self_attn.o_proj.weight": "pytorch_model-00002-of-00003.bin", + "model.layers.11.self_attn.q_proj.weight": "pytorch_model-00002-of-00003.bin", + "model.layers.11.self_attn.v_proj.weight": "pytorch_model-00002-of-00003.bin", + "model.layers.12.input_layernorm.weight": "pytorch_model-00002-of-00003.bin", + "model.layers.12.mlp.down_proj.weight": "pytorch_model-00002-of-00003.bin", + "model.layers.12.mlp.gate_proj.weight": "pytorch_model-00002-of-00003.bin", + "model.layers.12.mlp.up_proj.weight": "pytorch_model-00002-of-00003.bin", + "model.layers.12.post_attention_layernorm.weight": "pytorch_model-00002-of-00003.bin", + "model.layers.12.self_attn.k_proj.weight": "pytorch_model-00002-of-00003.bin", + "model.layers.12.self_attn.o_proj.weight": "pytorch_model-00002-of-00003.bin", + "model.layers.12.self_attn.q_proj.weight": "pytorch_model-00002-of-00003.bin", + "model.layers.12.self_attn.v_proj.weight": "pytorch_model-00002-of-00003.bin", + "model.layers.13.input_layernorm.weight": "pytorch_model-00002-of-00003.bin", + "model.layers.13.mlp.down_proj.weight": "pytorch_model-00002-of-00003.bin", + "model.layers.13.mlp.gate_proj.weight": "pytorch_model-00002-of-00003.bin", + "model.layers.13.mlp.up_proj.weight": "pytorch_model-00002-of-00003.bin", + "model.layers.13.post_attention_layernorm.weight": "pytorch_model-00002-of-00003.bin", + "model.layers.13.self_attn.k_proj.weight": "pytorch_model-00002-of-00003.bin", + "model.layers.13.self_attn.o_proj.weight": "pytorch_model-00002-of-00003.bin", + "model.layers.13.self_attn.q_proj.weight": "pytorch_model-00002-of-00003.bin", + "model.layers.13.self_attn.v_proj.weight": "pytorch_model-00002-of-00003.bin", + "model.layers.14.input_layernorm.weight": "pytorch_model-00002-of-00003.bin", + "model.layers.14.mlp.down_proj.weight": "pytorch_model-00002-of-00003.bin", + "model.layers.14.mlp.gate_proj.weight": "pytorch_model-00002-of-00003.bin", + "model.layers.14.mlp.up_proj.weight": "pytorch_model-00002-of-00003.bin", + "model.layers.14.post_attention_layernorm.weight": "pytorch_model-00002-of-00003.bin", + "model.layers.14.self_attn.k_proj.weight": "pytorch_model-00002-of-00003.bin", + "model.layers.14.self_attn.o_proj.weight": "pytorch_model-00002-of-00003.bin", + "model.layers.14.self_attn.q_proj.weight": "pytorch_model-00002-of-00003.bin", + "model.layers.14.self_attn.v_proj.weight": "pytorch_model-00002-of-00003.bin", + "model.layers.15.input_layernorm.weight": "pytorch_model-00002-of-00003.bin", + "model.layers.15.mlp.down_proj.weight": "pytorch_model-00002-of-00003.bin", + "model.layers.15.mlp.gate_proj.weight": "pytorch_model-00002-of-00003.bin", + "model.layers.15.mlp.up_proj.weight": "pytorch_model-00002-of-00003.bin", + "model.layers.15.post_attention_layernorm.weight": "pytorch_model-00002-of-00003.bin", + "model.layers.15.self_attn.k_proj.weight": "pytorch_model-00002-of-00003.bin", + "model.layers.15.self_attn.o_proj.weight": "pytorch_model-00002-of-00003.bin", + "model.layers.15.self_attn.q_proj.weight": "pytorch_model-00002-of-00003.bin", + "model.layers.15.self_attn.v_proj.weight": "pytorch_model-00002-of-00003.bin", + "model.layers.16.input_layernorm.weight": "pytorch_model-00002-of-00003.bin", + "model.layers.16.mlp.down_proj.weight": "pytorch_model-00002-of-00003.bin", + "model.layers.16.mlp.gate_proj.weight": "pytorch_model-00002-of-00003.bin", + "model.layers.16.mlp.up_proj.weight": "pytorch_model-00002-of-00003.bin", + "model.layers.16.post_attention_layernorm.weight": "pytorch_model-00002-of-00003.bin", + "model.layers.16.self_attn.k_proj.weight": "pytorch_model-00002-of-00003.bin", + "model.layers.16.self_attn.o_proj.weight": "pytorch_model-00002-of-00003.bin", + "model.layers.16.self_attn.q_proj.weight": "pytorch_model-00002-of-00003.bin", + "model.layers.16.self_attn.v_proj.weight": "pytorch_model-00002-of-00003.bin", + "model.layers.17.input_layernorm.weight": "pytorch_model-00002-of-00003.bin", + "model.layers.17.mlp.down_proj.weight": "pytorch_model-00002-of-00003.bin", + "model.layers.17.mlp.gate_proj.weight": "pytorch_model-00002-of-00003.bin", + "model.layers.17.mlp.up_proj.weight": "pytorch_model-00002-of-00003.bin", + "model.layers.17.post_attention_layernorm.weight": "pytorch_model-00002-of-00003.bin", + "model.layers.17.self_attn.k_proj.weight": "pytorch_model-00002-of-00003.bin", + "model.layers.17.self_attn.o_proj.weight": "pytorch_model-00002-of-00003.bin", + "model.layers.17.self_attn.q_proj.weight": "pytorch_model-00002-of-00003.bin", + "model.layers.17.self_attn.v_proj.weight": "pytorch_model-00002-of-00003.bin", + "model.layers.18.input_layernorm.weight": "pytorch_model-00002-of-00003.bin", + "model.layers.18.mlp.down_proj.weight": "pytorch_model-00002-of-00003.bin", + "model.layers.18.mlp.gate_proj.weight": "pytorch_model-00002-of-00003.bin", + "model.layers.18.mlp.up_proj.weight": "pytorch_model-00002-of-00003.bin", + "model.layers.18.post_attention_layernorm.weight": "pytorch_model-00002-of-00003.bin", + "model.layers.18.self_attn.k_proj.weight": "pytorch_model-00002-of-00003.bin", + "model.layers.18.self_attn.o_proj.weight": "pytorch_model-00002-of-00003.bin", + "model.layers.18.self_attn.q_proj.weight": "pytorch_model-00002-of-00003.bin", + "model.layers.18.self_attn.v_proj.weight": "pytorch_model-00002-of-00003.bin", + "model.layers.19.input_layernorm.weight": "pytorch_model-00002-of-00003.bin", + "model.layers.19.mlp.down_proj.weight": "pytorch_model-00002-of-00003.bin", + "model.layers.19.mlp.gate_proj.weight": "pytorch_model-00002-of-00003.bin", + "model.layers.19.mlp.up_proj.weight": "pytorch_model-00002-of-00003.bin", + "model.layers.19.post_attention_layernorm.weight": "pytorch_model-00002-of-00003.bin", + "model.layers.19.self_attn.k_proj.weight": "pytorch_model-00002-of-00003.bin", + "model.layers.19.self_attn.o_proj.weight": "pytorch_model-00002-of-00003.bin", + "model.layers.19.self_attn.q_proj.weight": "pytorch_model-00002-of-00003.bin", + "model.layers.19.self_attn.v_proj.weight": "pytorch_model-00002-of-00003.bin", + "model.layers.2.input_layernorm.weight": "pytorch_model-00001-of-00003.bin", + "model.layers.2.mlp.down_proj.weight": "pytorch_model-00001-of-00003.bin", + "model.layers.2.mlp.gate_proj.weight": "pytorch_model-00001-of-00003.bin", + "model.layers.2.mlp.up_proj.weight": "pytorch_model-00001-of-00003.bin", + "model.layers.2.post_attention_layernorm.weight": "pytorch_model-00001-of-00003.bin", + "model.layers.2.self_attn.k_proj.weight": "pytorch_model-00001-of-00003.bin", + "model.layers.2.self_attn.o_proj.weight": "pytorch_model-00001-of-00003.bin", + "model.layers.2.self_attn.q_proj.weight": "pytorch_model-00001-of-00003.bin", + "model.layers.2.self_attn.v_proj.weight": "pytorch_model-00001-of-00003.bin", + "model.layers.20.input_layernorm.weight": "pytorch_model-00002-of-00003.bin", + "model.layers.20.mlp.down_proj.weight": "pytorch_model-00002-of-00003.bin", + "model.layers.20.mlp.gate_proj.weight": "pytorch_model-00002-of-00003.bin", + "model.layers.20.mlp.up_proj.weight": "pytorch_model-00002-of-00003.bin", + "model.layers.20.post_attention_layernorm.weight": "pytorch_model-00002-of-00003.bin", + "model.layers.20.self_attn.k_proj.weight": "pytorch_model-00002-of-00003.bin", + "model.layers.20.self_attn.o_proj.weight": "pytorch_model-00002-of-00003.bin", + "model.layers.20.self_attn.q_proj.weight": "pytorch_model-00002-of-00003.bin", + "model.layers.20.self_attn.v_proj.weight": "pytorch_model-00002-of-00003.bin", + "model.layers.21.input_layernorm.weight": "pytorch_model-00002-of-00003.bin", + "model.layers.21.mlp.down_proj.weight": "pytorch_model-00002-of-00003.bin", + "model.layers.21.mlp.gate_proj.weight": "pytorch_model-00002-of-00003.bin", + "model.layers.21.mlp.up_proj.weight": "pytorch_model-00002-of-00003.bin", + "model.layers.21.post_attention_layernorm.weight": "pytorch_model-00002-of-00003.bin", + "model.layers.21.self_attn.k_proj.weight": "pytorch_model-00002-of-00003.bin", + "model.layers.21.self_attn.o_proj.weight": "pytorch_model-00002-of-00003.bin", + "model.layers.21.self_attn.q_proj.weight": "pytorch_model-00002-of-00003.bin", + "model.layers.21.self_attn.v_proj.weight": "pytorch_model-00002-of-00003.bin", + "model.layers.22.input_layernorm.weight": "pytorch_model-00003-of-00003.bin", + "model.layers.22.mlp.down_proj.weight": "pytorch_model-00003-of-00003.bin", + "model.layers.22.mlp.gate_proj.weight": "pytorch_model-00003-of-00003.bin", + "model.layers.22.mlp.up_proj.weight": "pytorch_model-00003-of-00003.bin", + "model.layers.22.post_attention_layernorm.weight": "pytorch_model-00003-of-00003.bin", + "model.layers.22.self_attn.k_proj.weight": "pytorch_model-00002-of-00003.bin", + "model.layers.22.self_attn.o_proj.weight": "pytorch_model-00002-of-00003.bin", + "model.layers.22.self_attn.q_proj.weight": "pytorch_model-00002-of-00003.bin", + "model.layers.22.self_attn.v_proj.weight": "pytorch_model-00002-of-00003.bin", + "model.layers.23.input_layernorm.weight": "pytorch_model-00003-of-00003.bin", + "model.layers.23.mlp.down_proj.weight": "pytorch_model-00003-of-00003.bin", + "model.layers.23.mlp.gate_proj.weight": "pytorch_model-00003-of-00003.bin", + "model.layers.23.mlp.up_proj.weight": "pytorch_model-00003-of-00003.bin", + "model.layers.23.post_attention_layernorm.weight": "pytorch_model-00003-of-00003.bin", + "model.layers.23.self_attn.k_proj.weight": "pytorch_model-00003-of-00003.bin", + "model.layers.23.self_attn.o_proj.weight": "pytorch_model-00003-of-00003.bin", + "model.layers.23.self_attn.q_proj.weight": "pytorch_model-00003-of-00003.bin", + "model.layers.23.self_attn.v_proj.weight": "pytorch_model-00003-of-00003.bin", + "model.layers.24.input_layernorm.weight": "pytorch_model-00003-of-00003.bin", + "model.layers.24.mlp.down_proj.weight": "pytorch_model-00003-of-00003.bin", + "model.layers.24.mlp.gate_proj.weight": "pytorch_model-00003-of-00003.bin", + "model.layers.24.mlp.up_proj.weight": "pytorch_model-00003-of-00003.bin", + "model.layers.24.post_attention_layernorm.weight": "pytorch_model-00003-of-00003.bin", + "model.layers.24.self_attn.k_proj.weight": "pytorch_model-00003-of-00003.bin", + "model.layers.24.self_attn.o_proj.weight": "pytorch_model-00003-of-00003.bin", + "model.layers.24.self_attn.q_proj.weight": "pytorch_model-00003-of-00003.bin", + "model.layers.24.self_attn.v_proj.weight": "pytorch_model-00003-of-00003.bin", + "model.layers.25.input_layernorm.weight": "pytorch_model-00003-of-00003.bin", + "model.layers.25.mlp.down_proj.weight": "pytorch_model-00003-of-00003.bin", + "model.layers.25.mlp.gate_proj.weight": "pytorch_model-00003-of-00003.bin", + "model.layers.25.mlp.up_proj.weight": "pytorch_model-00003-of-00003.bin", + "model.layers.25.post_attention_layernorm.weight": "pytorch_model-00003-of-00003.bin", + "model.layers.25.self_attn.k_proj.weight": "pytorch_model-00003-of-00003.bin", + "model.layers.25.self_attn.o_proj.weight": "pytorch_model-00003-of-00003.bin", + "model.layers.25.self_attn.q_proj.weight": "pytorch_model-00003-of-00003.bin", + "model.layers.25.self_attn.v_proj.weight": "pytorch_model-00003-of-00003.bin", + "model.layers.26.input_layernorm.weight": "pytorch_model-00003-of-00003.bin", + "model.layers.26.mlp.down_proj.weight": "pytorch_model-00003-of-00003.bin", + "model.layers.26.mlp.gate_proj.weight": "pytorch_model-00003-of-00003.bin", + "model.layers.26.mlp.up_proj.weight": "pytorch_model-00003-of-00003.bin", + "model.layers.26.post_attention_layernorm.weight": "pytorch_model-00003-of-00003.bin", + "model.layers.26.self_attn.k_proj.weight": "pytorch_model-00003-of-00003.bin", + "model.layers.26.self_attn.o_proj.weight": "pytorch_model-00003-of-00003.bin", + "model.layers.26.self_attn.q_proj.weight": "pytorch_model-00003-of-00003.bin", + "model.layers.26.self_attn.v_proj.weight": "pytorch_model-00003-of-00003.bin", + "model.layers.27.input_layernorm.weight": "pytorch_model-00003-of-00003.bin", + "model.layers.27.mlp.down_proj.weight": "pytorch_model-00003-of-00003.bin", + "model.layers.27.mlp.gate_proj.weight": "pytorch_model-00003-of-00003.bin", + "model.layers.27.mlp.up_proj.weight": "pytorch_model-00003-of-00003.bin", + "model.layers.27.post_attention_layernorm.weight": "pytorch_model-00003-of-00003.bin", + "model.layers.27.self_attn.k_proj.weight": "pytorch_model-00003-of-00003.bin", + "model.layers.27.self_attn.o_proj.weight": "pytorch_model-00003-of-00003.bin", + "model.layers.27.self_attn.q_proj.weight": "pytorch_model-00003-of-00003.bin", + "model.layers.27.self_attn.v_proj.weight": "pytorch_model-00003-of-00003.bin", + "model.layers.28.input_layernorm.weight": "pytorch_model-00003-of-00003.bin", + "model.layers.28.mlp.down_proj.weight": "pytorch_model-00003-of-00003.bin", + "model.layers.28.mlp.gate_proj.weight": "pytorch_model-00003-of-00003.bin", + "model.layers.28.mlp.up_proj.weight": "pytorch_model-00003-of-00003.bin", + "model.layers.28.post_attention_layernorm.weight": "pytorch_model-00003-of-00003.bin", + "model.layers.28.self_attn.k_proj.weight": "pytorch_model-00003-of-00003.bin", + "model.layers.28.self_attn.o_proj.weight": "pytorch_model-00003-of-00003.bin", + "model.layers.28.self_attn.q_proj.weight": "pytorch_model-00003-of-00003.bin", + "model.layers.28.self_attn.v_proj.weight": "pytorch_model-00003-of-00003.bin", + "model.layers.29.input_layernorm.weight": "pytorch_model-00003-of-00003.bin", + "model.layers.29.mlp.down_proj.weight": "pytorch_model-00003-of-00003.bin", + "model.layers.29.mlp.gate_proj.weight": "pytorch_model-00003-of-00003.bin", + "model.layers.29.mlp.up_proj.weight": "pytorch_model-00003-of-00003.bin", + "model.layers.29.post_attention_layernorm.weight": "pytorch_model-00003-of-00003.bin", + "model.layers.29.self_attn.k_proj.weight": "pytorch_model-00003-of-00003.bin", + "model.layers.29.self_attn.o_proj.weight": "pytorch_model-00003-of-00003.bin", + "model.layers.29.self_attn.q_proj.weight": "pytorch_model-00003-of-00003.bin", + "model.layers.29.self_attn.v_proj.weight": "pytorch_model-00003-of-00003.bin", + "model.layers.3.input_layernorm.weight": "pytorch_model-00001-of-00003.bin", + "model.layers.3.mlp.down_proj.weight": "pytorch_model-00001-of-00003.bin", + "model.layers.3.mlp.gate_proj.weight": "pytorch_model-00001-of-00003.bin", + "model.layers.3.mlp.up_proj.weight": "pytorch_model-00001-of-00003.bin", + "model.layers.3.post_attention_layernorm.weight": "pytorch_model-00001-of-00003.bin", + "model.layers.3.self_attn.k_proj.weight": "pytorch_model-00001-of-00003.bin", + "model.layers.3.self_attn.o_proj.weight": "pytorch_model-00001-of-00003.bin", + "model.layers.3.self_attn.q_proj.weight": "pytorch_model-00001-of-00003.bin", + "model.layers.3.self_attn.v_proj.weight": "pytorch_model-00001-of-00003.bin", + "model.layers.30.input_layernorm.weight": "pytorch_model-00003-of-00003.bin", + "model.layers.30.mlp.down_proj.weight": "pytorch_model-00003-of-00003.bin", + "model.layers.30.mlp.gate_proj.weight": "pytorch_model-00003-of-00003.bin", + "model.layers.30.mlp.up_proj.weight": "pytorch_model-00003-of-00003.bin", + "model.layers.30.post_attention_layernorm.weight": "pytorch_model-00003-of-00003.bin", + "model.layers.30.self_attn.k_proj.weight": "pytorch_model-00003-of-00003.bin", + "model.layers.30.self_attn.o_proj.weight": "pytorch_model-00003-of-00003.bin", + "model.layers.30.self_attn.q_proj.weight": "pytorch_model-00003-of-00003.bin", + "model.layers.30.self_attn.v_proj.weight": "pytorch_model-00003-of-00003.bin", + "model.layers.31.input_layernorm.weight": "pytorch_model-00003-of-00003.bin", + "model.layers.31.mlp.down_proj.weight": "pytorch_model-00003-of-00003.bin", + "model.layers.31.mlp.gate_proj.weight": "pytorch_model-00003-of-00003.bin", + "model.layers.31.mlp.up_proj.weight": "pytorch_model-00003-of-00003.bin", + "model.layers.31.post_attention_layernorm.weight": "pytorch_model-00003-of-00003.bin", + "model.layers.31.self_attn.k_proj.weight": "pytorch_model-00003-of-00003.bin", + "model.layers.31.self_attn.o_proj.weight": "pytorch_model-00003-of-00003.bin", + "model.layers.31.self_attn.q_proj.weight": "pytorch_model-00003-of-00003.bin", + "model.layers.31.self_attn.v_proj.weight": "pytorch_model-00003-of-00003.bin", + "model.layers.4.input_layernorm.weight": "pytorch_model-00001-of-00003.bin", + "model.layers.4.mlp.down_proj.weight": "pytorch_model-00001-of-00003.bin", + "model.layers.4.mlp.gate_proj.weight": "pytorch_model-00001-of-00003.bin", + "model.layers.4.mlp.up_proj.weight": "pytorch_model-00001-of-00003.bin", + "model.layers.4.post_attention_layernorm.weight": "pytorch_model-00001-of-00003.bin", + "model.layers.4.self_attn.k_proj.weight": "pytorch_model-00001-of-00003.bin", + "model.layers.4.self_attn.o_proj.weight": "pytorch_model-00001-of-00003.bin", + "model.layers.4.self_attn.q_proj.weight": "pytorch_model-00001-of-00003.bin", + "model.layers.4.self_attn.v_proj.weight": "pytorch_model-00001-of-00003.bin", + "model.layers.5.input_layernorm.weight": "pytorch_model-00001-of-00003.bin", + "model.layers.5.mlp.down_proj.weight": "pytorch_model-00001-of-00003.bin", + "model.layers.5.mlp.gate_proj.weight": "pytorch_model-00001-of-00003.bin", + "model.layers.5.mlp.up_proj.weight": "pytorch_model-00001-of-00003.bin", + "model.layers.5.post_attention_layernorm.weight": "pytorch_model-00001-of-00003.bin", + "model.layers.5.self_attn.k_proj.weight": "pytorch_model-00001-of-00003.bin", + "model.layers.5.self_attn.o_proj.weight": "pytorch_model-00001-of-00003.bin", + "model.layers.5.self_attn.q_proj.weight": "pytorch_model-00001-of-00003.bin", + "model.layers.5.self_attn.v_proj.weight": "pytorch_model-00001-of-00003.bin", + "model.layers.6.input_layernorm.weight": "pytorch_model-00001-of-00003.bin", + "model.layers.6.mlp.down_proj.weight": "pytorch_model-00001-of-00003.bin", + "model.layers.6.mlp.gate_proj.weight": "pytorch_model-00001-of-00003.bin", + "model.layers.6.mlp.up_proj.weight": "pytorch_model-00001-of-00003.bin", + "model.layers.6.post_attention_layernorm.weight": "pytorch_model-00001-of-00003.bin", + "model.layers.6.self_attn.k_proj.weight": "pytorch_model-00001-of-00003.bin", + "model.layers.6.self_attn.o_proj.weight": "pytorch_model-00001-of-00003.bin", + "model.layers.6.self_attn.q_proj.weight": "pytorch_model-00001-of-00003.bin", + "model.layers.6.self_attn.v_proj.weight": "pytorch_model-00001-of-00003.bin", + "model.layers.7.input_layernorm.weight": "pytorch_model-00001-of-00003.bin", + "model.layers.7.mlp.down_proj.weight": "pytorch_model-00001-of-00003.bin", + "model.layers.7.mlp.gate_proj.weight": "pytorch_model-00001-of-00003.bin", + "model.layers.7.mlp.up_proj.weight": "pytorch_model-00001-of-00003.bin", + "model.layers.7.post_attention_layernorm.weight": "pytorch_model-00001-of-00003.bin", + "model.layers.7.self_attn.k_proj.weight": "pytorch_model-00001-of-00003.bin", + "model.layers.7.self_attn.o_proj.weight": "pytorch_model-00001-of-00003.bin", + "model.layers.7.self_attn.q_proj.weight": "pytorch_model-00001-of-00003.bin", + "model.layers.7.self_attn.v_proj.weight": "pytorch_model-00001-of-00003.bin", + "model.layers.8.input_layernorm.weight": "pytorch_model-00001-of-00003.bin", + "model.layers.8.mlp.down_proj.weight": "pytorch_model-00001-of-00003.bin", + "model.layers.8.mlp.gate_proj.weight": "pytorch_model-00001-of-00003.bin", + "model.layers.8.mlp.up_proj.weight": "pytorch_model-00001-of-00003.bin", + "model.layers.8.post_attention_layernorm.weight": "pytorch_model-00001-of-00003.bin", + "model.layers.8.self_attn.k_proj.weight": "pytorch_model-00001-of-00003.bin", + "model.layers.8.self_attn.o_proj.weight": "pytorch_model-00001-of-00003.bin", + "model.layers.8.self_attn.q_proj.weight": "pytorch_model-00001-of-00003.bin", + "model.layers.8.self_attn.v_proj.weight": "pytorch_model-00001-of-00003.bin", + "model.layers.9.input_layernorm.weight": "pytorch_model-00001-of-00003.bin", + "model.layers.9.mlp.down_proj.weight": "pytorch_model-00001-of-00003.bin", + "model.layers.9.mlp.gate_proj.weight": "pytorch_model-00001-of-00003.bin", + "model.layers.9.mlp.up_proj.weight": "pytorch_model-00001-of-00003.bin", + "model.layers.9.post_attention_layernorm.weight": "pytorch_model-00001-of-00003.bin", + "model.layers.9.self_attn.k_proj.weight": "pytorch_model-00001-of-00003.bin", + "model.layers.9.self_attn.o_proj.weight": "pytorch_model-00001-of-00003.bin", + "model.layers.9.self_attn.q_proj.weight": "pytorch_model-00001-of-00003.bin", + "model.layers.9.self_attn.v_proj.weight": "pytorch_model-00001-of-00003.bin", + "model.norm.weight": "pytorch_model-00003-of-00003.bin" + } +} diff --git a/special_tokens_map.json b/special_tokens_map.json new file mode 100644 index 0000000..72ecfee --- /dev/null +++ b/special_tokens_map.json @@ -0,0 +1,24 @@ +{ + "bos_token": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false + }, + "eos_token": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false + }, + "pad_token": "", + "unk_token": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false + } +} diff --git a/tokenizer.model b/tokenizer.model new file mode 100644 index 0000000..8b443ef --- /dev/null +++ b/tokenizer.model @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:dadfd56d766715c61d2ef780a525ab43b8e6da4de6865bda3d95fdef5e134055 +size 493443 diff --git a/tokenizer_config.json b/tokenizer_config.json new file mode 100644 index 0000000..c25b6d0 --- /dev/null +++ b/tokenizer_config.json @@ -0,0 +1,43 @@ +{ + "add_bos_token": true, + "add_eos_token": false, + "add_prefix_space": true, + "added_tokens_decoder": { + "0": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "1": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "2": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + } + }, + "bos_token": "", + "clean_up_tokenization_spaces": false, + "eos_token": "", + "legacy": true, + "model_max_length": 1000000000000000019884624838656, + "pad_token": "", + "sp_model_kwargs": {}, + "spaces_between_special_tokens": false, + "tokenizer_class": "LlamaTokenizer", + "unk_token": "", + "use_default_system_prompt": false, + "use_fast": true +} diff --git a/training_args.bin b/training_args.bin new file mode 100644 index 0000000..68b1431 --- /dev/null +++ b/training_args.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:00d5675931773fafeff74351dc5a9cbbf81577167988949e58fbb52a3bd05d94 +size 5880