commit a53def3b08248af132fb32429a2909082571aa7b Author: ModelHub XC Date: Tue Jun 16 08:06:17 2026 +0800 初始化项目，由ModelHub XC社区提供模型 Model: pvs333/supergames-grpo Source: Original Platform diff --git a/.gitattributes b/.gitattributes new file mode 100644 index 0000000..a017861 --- /dev/null +++ b/.gitattributes @@ -0,0 +1,46 @@ +*.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 +checkpoint-100/tokenizer.json filter=lfs diff=lfs merge=lfs -text +checkpoint-120/tokenizer.json filter=lfs diff=lfs merge=lfs -text +checkpoint-140/tokenizer.json filter=lfs diff=lfs merge=lfs -text +checkpoint-160/tokenizer.json filter=lfs diff=lfs merge=lfs -text +checkpoint-180/tokenizer.json filter=lfs diff=lfs merge=lfs -text +checkpoint-20/tokenizer.json filter=lfs diff=lfs merge=lfs -text +checkpoint-200/tokenizer.json filter=lfs diff=lfs merge=lfs -text +checkpoint-40/tokenizer.json filter=lfs diff=lfs merge=lfs -text +checkpoint-60/tokenizer.json filter=lfs diff=lfs merge=lfs -text +checkpoint-80/tokenizer.json filter=lfs diff=lfs merge=lfs -text +tokenizer.json filter=lfs diff=lfs merge=lfs -text diff --git a/README.md b/README.md new file mode 100644 index 0000000..670dfc2 --- /dev/null +++ b/README.md @@ -0,0 +1,67 @@ +--- +base_model: Qwen/Qwen2.5-1.5B-Instruct +library_name: transformers +model_name: supergames-grpo200-1.5B +tags: +- generated_from_trainer +- grpo +- trl +licence: license +--- + +# Model Card for supergames-grpo200-1.5B + +This model is a fine-tuned version of [Qwen/Qwen2.5-1.5B-Instruct](https://huggingface.co/Qwen/Qwen2.5-1.5B-Instruct). +It has been trained using [TRL](https://github.com/huggingface/trl). + +## Quick start + +```python +from transformers import pipeline + +question = "If you had a time machine, but could only go to the past or the future once and never return, which would you choose and why?" +generator = pipeline("text-generation", model="None", device="cuda") +output = generator([{"role": "user", "content": question}], max_new_tokens=128, return_full_text=False)[0] +print(output["generated_text"]) +``` + +## Training procedure + + + + + +This model was trained with GRPO, a method introduced in [DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Models](https://huggingface.co/papers/2402.03300). + +### Framework versions + +- TRL: 1.2.0 +- Transformers: 5.6.2 +- Pytorch: 2.11.0 +- Datasets: 4.8.4 +- Tokenizers: 0.22.2 + +## Citations + +Cite GRPO as: + +```bibtex +@article{shao2024deepseekmath, + title = {{DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Models}}, + author = {Zhihong Shao and Peiyi Wang and Qihao Zhu and Runxin Xu and Junxiao Song and Mingchuan Zhang and Y. K. Li and Y. Wu and Daya Guo}, + year = 2024, + eprint = {arXiv:2402.03300}, +} +``` + +Cite TRL as: + +```bibtex +@software{vonwerra2020trl, + title = {{TRL: Transformers Reinforcement Learning}}, + author = {von Werra, Leandro and Belkada, Younes and Tunstall, Lewis and Beeching, Edward and Thrush, Tristan and Lambert, Nathan and Huang, Shengyi and Rasul, Kashif and Gallouédec, Quentin}, + license = {Apache-2.0}, + url = {https://github.com/huggingface/trl}, + year = {2020} +} +``` \ No newline at end of file diff --git a/chat_template.jinja b/chat_template.jinja new file mode 100644 index 0000000..bdf7919 --- /dev/null +++ b/chat_template.jinja @@ -0,0 +1,54 @@ +{%- if tools %} + {{- '<|im_start|>system\n' }} + {%- if messages[0]['role'] == 'system' %} + {{- messages[0]['content'] }} + {%- else %} + {{- 'You are Qwen, created by Alibaba Cloud. You are a helpful assistant.' }} + {%- endif %} + {{- "\n\n# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }} + {%- for tool in tools %} + {{- "\n" }} + {{- tool | tojson }} + {%- endfor %} + {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }} +{%- else %} + {%- if messages[0]['role'] == 'system' %} + {{- '<|im_start|>system\n' + messages[0]['content'] + '<|im_end|>\n' }} + {%- else %} + {{- '<|im_start|>system\nYou are Qwen, created by Alibaba Cloud. You are a helpful assistant.<|im_end|>\n' }} + {%- endif %} +{%- endif %} +{%- for message in messages %} + {%- if (message.role == "user") or (message.role == "system" and not loop.first) or (message.role == "assistant" and not message.tool_calls) %} + {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }} + {%- elif message.role == "assistant" %} + {{- '<|im_start|>' + message.role }} + {%- if message.content %} + {{- '\n' + message.content }} + {%- endif %} + {%- for tool_call in message.tool_calls %} + {%- if tool_call.function is defined %} + {%- set tool_call = tool_call.function %} + {%- endif %} + {{- '\n\n{"name": "' }} + {{- tool_call.name }} + {{- '", "arguments": ' }} + {{- tool_call.arguments | tojson }} + {{- '}\n' }} + {%- endfor %} + {{- '<|im_end|>\n' }} + {%- elif message.role == "tool" %} + {%- if (loop.index0 == 0) or (messages[loop.index0 - 1].role != "tool") %} + {{- '<|im_start|>user' }} + {%- endif %} + {{- '\n\n' }} + {{- message.content }} + {{- '\n' }} + {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %} + {{- '<|im_end|>\n' }} + {%- endif %} + {%- endif %} +{%- endfor %} +{%- if add_generation_prompt %} + {{- '<|im_start|>assistant\n' }} +{%- endif %} diff --git a/checkpoint-100/chat_template.jinja b/checkpoint-100/chat_template.jinja new file mode 100644 index 0000000..bdf7919 --- /dev/null +++ b/checkpoint-100/chat_template.jinja @@ -0,0 +1,54 @@ +{%- if tools %} + {{- '<|im_start|>system\n' }} + {%- if messages[0]['role'] == 'system' %} + {{- messages[0]['content'] }} + {%- else %} + {{- 'You are Qwen, created by Alibaba Cloud. You are a helpful assistant.' }} + {%- endif %} + {{- "\n\n# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }} + {%- for tool in tools %} + {{- "\n" }} + {{- tool | tojson }} + {%- endfor %} + {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }} +{%- else %} + {%- if messages[0]['role'] == 'system' %} + {{- '<|im_start|>system\n' + messages[0]['content'] + '<|im_end|>\n' }} + {%- else %} + {{- '<|im_start|>system\nYou are Qwen, created by Alibaba Cloud. You are a helpful assistant.<|im_end|>\n' }} + {%- endif %} +{%- endif %} +{%- for message in messages %} + {%- if (message.role == "user") or (message.role == "system" and not loop.first) or (message.role == "assistant" and not message.tool_calls) %} + {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }} + {%- elif message.role == "assistant" %} + {{- '<|im_start|>' + message.role }} + {%- if message.content %} + {{- '\n' + message.content }} + {%- endif %} + {%- for tool_call in message.tool_calls %} + {%- if tool_call.function is defined %} + {%- set tool_call = tool_call.function %} + {%- endif %} + {{- '\n\n{"name": "' }} + {{- tool_call.name }} + {{- '", "arguments": ' }} + {{- tool_call.arguments | tojson }} + {{- '}\n' }} + {%- endfor %} + {{- '<|im_end|>\n' }} + {%- elif message.role == "tool" %} + {%- if (loop.index0 == 0) or (messages[loop.index0 - 1].role != "tool") %} + {{- '<|im_start|>user' }} + {%- endif %} + {{- '\n\n' }} + {{- message.content }} + {{- '\n' }} + {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %} + {{- '<|im_end|>\n' }} + {%- endif %} + {%- endif %} +{%- endfor %} +{%- if add_generation_prompt %} + {{- '<|im_start|>assistant\n' }} +{%- endif %} diff --git a/checkpoint-100/config.json b/checkpoint-100/config.json new file mode 100644 index 0000000..f57b09b --- /dev/null +++ b/checkpoint-100/config.json @@ -0,0 +1,61 @@ +{ + "architectures": [ + "Qwen2ForCausalLM" + ], + "attention_dropout": 0.0, + "bos_token_id": null, + "dtype": "float32", + "eos_token_id": 151645, + "hidden_act": "silu", + "hidden_size": 1536, + "initializer_range": 0.02, + "intermediate_size": 8960, + "layer_types": [ + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention" + ], + "max_position_embeddings": 32768, + "max_window_layers": 21, + "model_type": "qwen2", + "num_attention_heads": 12, + "num_hidden_layers": 28, + "num_key_value_heads": 2, + "pad_token_id": 151643, + "rms_norm_eps": 1e-06, + "rope_parameters": { + "rope_theta": 1000000.0, + "rope_type": "default" + }, + "sliding_window": null, + "tie_word_embeddings": true, + "transformers_version": "5.6.2", + "use_cache": false, + "use_sliding_window": false, + "vocab_size": 151936 +} diff --git a/checkpoint-100/generation_config.json b/checkpoint-100/generation_config.json new file mode 100644 index 0000000..1b2bba9 --- /dev/null +++ b/checkpoint-100/generation_config.json @@ -0,0 +1,13 @@ +{ + "do_sample": true, + "eos_token_id": [ + 151645, + 151643 + ], + "pad_token_id": 151643, + "repetition_penalty": 1.1, + "temperature": 0.7, + "top_k": 20, + "top_p": 0.8, + "transformers_version": "5.6.2" +} diff --git a/checkpoint-100/model.safetensors b/checkpoint-100/model.safetensors new file mode 100644 index 0000000..bb9332f --- /dev/null +++ b/checkpoint-100/model.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:016ca4a7af59ae5340813c9f0b26ea5ebd58afe4cb3ed06f59d7fa1fe9828095 +size 6174895536 diff --git a/checkpoint-100/optimizer.pt b/checkpoint-100/optimizer.pt new file mode 100644 index 0000000..37c939b --- /dev/null +++ b/checkpoint-100/optimizer.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:b8fa58ba705beb0a7bbc366e9d67c072abecfdd9f655213dfa382d9e628925a6 +size 12350013801 diff --git a/checkpoint-100/rng_state.pth b/checkpoint-100/rng_state.pth new file mode 100644 index 0000000..406bb59 --- /dev/null +++ b/checkpoint-100/rng_state.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:9869cc2d5f67dae6ae945d3ceb0eb17366022213b0956cc572cab18268befe71 +size 14645 diff --git a/checkpoint-100/scheduler.pt b/checkpoint-100/scheduler.pt new file mode 100644 index 0000000..8a88dc3 --- /dev/null +++ b/checkpoint-100/scheduler.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:c6151631075f7a98a9131e237ab6732efc7b20c8af25b6f984599c180e47967c +size 1465 diff --git a/checkpoint-100/tokenizer.json b/checkpoint-100/tokenizer.json new file mode 100644 index 0000000..34510ff --- /dev/null +++ b/checkpoint-100/tokenizer.json @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:3fd169731d2cbde95e10bf356d66d5997fd885dd8dbb6fb4684da3f23b2585d8 +size 11421892 diff --git a/checkpoint-100/tokenizer_config.json b/checkpoint-100/tokenizer_config.json new file mode 100644 index 0000000..770e41d --- /dev/null +++ b/checkpoint-100/tokenizer_config.json @@ -0,0 +1,30 @@ +{ + "add_prefix_space": false, + "backend": "tokenizers", + "bos_token": null, + "clean_up_tokenization_spaces": false, + "eos_token": "<|im_end|>", + "errors": "replace", + "extra_special_tokens": [ + "<|im_start|>", + "<|im_end|>", + "<|object_ref_start|>", + "<|object_ref_end|>", + "<|box_start|>", + "<|box_end|>", + "<|quad_start|>", + "<|quad_end|>", + "<|vision_start|>", + "<|vision_end|>", + "<|vision_pad|>", + "<|image_pad|>", + "<|video_pad|>" + ], + "is_local": false, + "local_files_only": false, + "model_max_length": 131072, + "pad_token": "<|endoftext|>", + "split_special_tokens": false, + "tokenizer_class": "Qwen2Tokenizer", + "unk_token": null +} diff --git a/checkpoint-100/trainer_state.json b/checkpoint-100/trainer_state.json new file mode 100644 index 0000000..83c67d3 --- /dev/null +++ b/checkpoint-100/trainer_state.json @@ -0,0 +1,2934 @@ +{ + "best_global_step": null, + "best_metric": null, + "best_model_checkpoint": null, + "epoch": 0.5, + "eval_steps": 500, + "global_step": 100, + "is_hyper_param_search": false, + "is_local_process_zero": true, + "is_world_process_zero": true, + "log_history": [ + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 478.0, + "completions/max_terminated_length": 478.0, + "completions/mean_length": 191.0, + "completions/mean_terminated_length": 191.0, + "completions/min_length": 66.0, + "completions/min_terminated_length": 66.0, + "entropy": 0.5566893219947815, + "epoch": 0.005, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.370361328125, + "learning_rate": 1e-06, + "loss": -0.24658073484897614, + "num_tokens": 10832.0, + "reward": -3.3000030517578125, + "reward_std": 85.62333679199219, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -3.4000015258789062, + "rewards/supergames_reward/std": 85.62333679199219, + "step": 1, + "step_time": 22.779711072013015 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 216.0, + "completions/max_terminated_length": 216.0, + "completions/mean_length": 131.5, + "completions/mean_terminated_length": 131.5, + "completions/min_length": 60.0, + "completions/min_terminated_length": 60.0, + "entropy": 0.6395841240882874, + "epoch": 0.01, + "frac_reward_zero_std": 0.0, + "grad_norm": 8.09774398803711, + "learning_rate": 9.95e-07, + "loss": 0.20567649602890015, + "num_tokens": 16404.0, + "reward": -12.422499656677246, + "reward_std": 7.134707450866699, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -12.522500038146973, + "rewards/supergames_reward/std": 7.134707927703857, + "step": 2, + "step_time": 9.960156448010821 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 287.0, + "completions/max_terminated_length": 287.0, + "completions/mean_length": 156.625, + "completions/mean_terminated_length": 156.625, + "completions/min_length": 77.0, + "completions/min_terminated_length": 77.0, + "entropy": 0.562222421169281, + "epoch": 0.015, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.9053616523742676, + "learning_rate": 9.9e-07, + "loss": 0.09602774679660797, + "num_tokens": 26953.0, + "reward": -74.94999694824219, + "reward_std": 70.73094177246094, + "rewards/json_format_reward/mean": 0.05000000074505806, + "rewards/json_format_reward/std": 0.09258200973272324, + "rewards/supergames_reward/mean": -75.0, + "rewards/supergames_reward/std": 70.71067810058594, + "step": 3, + "step_time": 13.686320498993155 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 268.0, + "completions/max_terminated_length": 268.0, + "completions/mean_length": 154.125, + "completions/mean_terminated_length": 154.125, + "completions/min_length": 53.0, + "completions/min_terminated_length": 53.0, + "entropy": 0.7323317527770996, + "epoch": 0.02, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.933310031890869, + "learning_rate": 9.849999999999999e-07, + "loss": 0.28249427676200867, + "num_tokens": 36514.0, + "reward": -40.7599983215332, + "reward_std": 81.73140716552734, + "rewards/json_format_reward/mean": 0.05000000074505806, + "rewards/json_format_reward/std": 0.09258200973272324, + "rewards/supergames_reward/mean": -40.80999755859375, + "rewards/supergames_reward/std": 81.68995666503906, + "step": 4, + "step_time": 12.805880262021674 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 401.0, + "completions/max_terminated_length": 401.0, + "completions/mean_length": 205.25, + "completions/mean_terminated_length": 205.25, + "completions/min_length": 115.0, + "completions/min_terminated_length": 115.0, + "entropy": 0.6167430281639099, + "epoch": 0.025, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.344135046005249, + "learning_rate": 9.8e-07, + "loss": 0.1433994024991989, + "num_tokens": 47476.0, + "reward": -97.6612548828125, + "reward_std": 6.3321428298950195, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -97.76124572753906, + "rewards/supergames_reward/std": 6.3321428298950195, + "step": 5, + "step_time": 18.789364666008623 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 276.0, + "completions/max_terminated_length": 276.0, + "completions/mean_length": 167.25, + "completions/mean_terminated_length": 167.25, + "completions/min_length": 73.0, + "completions/min_terminated_length": 73.0, + "entropy": 0.7583790421485901, + "epoch": 0.03, + "frac_reward_zero_std": 0.0, + "grad_norm": 5.172371864318848, + "learning_rate": 9.75e-07, + "loss": -0.12127404659986496, + "num_tokens": 55246.0, + "reward": -58.687503814697266, + "reward_std": 58.64059829711914, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": -58.76250457763672, + "rewards/supergames_reward/std": 58.620460510253906, + "step": 6, + "step_time": 12.948570273991209 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 261.0, + "completions/max_terminated_length": 261.0, + "completions/mean_length": 133.875, + "completions/mean_terminated_length": 133.875, + "completions/min_length": 57.0, + "completions/min_terminated_length": 57.0, + "entropy": 0.7447654604911804, + "epoch": 0.035, + "frac_reward_zero_std": 0.0, + "grad_norm": 6.428595066070557, + "learning_rate": 9.7e-07, + "loss": 0.10135584324598312, + "num_tokens": 60837.0, + "reward": -32.45000076293945, + "reward_std": 41.754486083984375, + "rewards/json_format_reward/mean": 0.05000000074505806, + "rewards/json_format_reward/std": 0.09258200973272324, + "rewards/supergames_reward/mean": -32.5, + "rewards/supergames_reward/std": 41.66190719604492, + "step": 7, + "step_time": 12.058315072004916 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 414.0, + "completions/max_terminated_length": 414.0, + "completions/mean_length": 209.125, + "completions/mean_terminated_length": 209.125, + "completions/min_length": 104.0, + "completions/min_terminated_length": 104.0, + "entropy": 0.3860666751861572, + "epoch": 0.04, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.806255578994751, + "learning_rate": 9.649999999999999e-07, + "loss": -0.33238139748573303, + "num_tokens": 70838.0, + "reward": -80.50375366210938, + "reward_std": 38.82380294799805, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -80.60375213623047, + "rewards/supergames_reward/std": 38.82379913330078, + "step": 8, + "step_time": 19.323370319994865 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.25, + "completions/max_length": 512.0, + "completions/max_terminated_length": 199.0, + "completions/mean_length": 247.125, + "completions/mean_terminated_length": 158.83334350585938, + "completions/min_length": 106.0, + "completions/min_terminated_length": 106.0, + "entropy": 0.5816237330436707, + "epoch": 0.045, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.3209943771362305, + "learning_rate": 9.6e-07, + "loss": 0.36752766370773315, + "num_tokens": 82135.0, + "reward": -27.688751220703125, + "reward_std": 82.8358154296875, + "rewards/json_format_reward/mean": 0.05000000074505806, + "rewards/json_format_reward/std": 0.09258200973272324, + "rewards/supergames_reward/mean": -27.738750457763672, + "rewards/supergames_reward/std": 82.78590393066406, + "step": 9, + "step_time": 21.948575104994234 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 240.0, + "completions/max_terminated_length": 240.0, + "completions/mean_length": 153.0, + "completions/mean_terminated_length": 153.0, + "completions/min_length": 51.0, + "completions/min_terminated_length": 51.0, + "entropy": 0.6301521062850952, + "epoch": 0.05, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.125359058380127, + "learning_rate": 9.55e-07, + "loss": 0.06129350885748863, + "num_tokens": 89807.0, + "reward": -78.39624786376953, + "reward_std": 40.17042922973633, + "rewards/json_format_reward/mean": 0.05000000074505806, + "rewards/json_format_reward/std": 0.09258200973272324, + "rewards/supergames_reward/mean": -78.44625091552734, + "rewards/supergames_reward/std": 40.1396484375, + "step": 10, + "step_time": 9.098202993016457 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 133.0, + "completions/max_terminated_length": 133.0, + "completions/mean_length": 112.0, + "completions/mean_terminated_length": 112.0, + "completions/min_length": 68.0, + "completions/min_terminated_length": 68.0, + "entropy": 0.5521990656852722, + "epoch": 0.055, + "frac_reward_zero_std": 0.0, + "grad_norm": 6.616060256958008, + "learning_rate": 9.499999999999999e-07, + "loss": -0.006313305348157883, + "num_tokens": 95231.0, + "reward": -12.161249160766602, + "reward_std": 6.39578104019165, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -12.261249542236328, + "rewards/supergames_reward/std": 6.39578104019165, + "step": 11, + "step_time": 5.136311663984088 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 396.0, + "completions/max_terminated_length": 396.0, + "completions/mean_length": 214.875, + "completions/mean_terminated_length": 214.875, + "completions/min_length": 55.0, + "completions/min_terminated_length": 55.0, + "entropy": 0.5505905151367188, + "epoch": 0.06, + "frac_reward_zero_std": 0.0, + "grad_norm": 5.289583206176758, + "learning_rate": 9.45e-07, + "loss": -0.29802003502845764, + "num_tokens": 106238.0, + "reward": -91.66999816894531, + "reward_std": 23.2779541015625, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -91.77000427246094, + "rewards/supergames_reward/std": 23.277956008911133, + "step": 12, + "step_time": 14.945365622988902 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 269.0, + "completions/max_terminated_length": 269.0, + "completions/mean_length": 167.375, + "completions/mean_terminated_length": 167.375, + "completions/min_length": 113.0, + "completions/min_terminated_length": 113.0, + "entropy": 0.540415346622467, + "epoch": 0.065, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.858291149139404, + "learning_rate": 9.399999999999999e-07, + "loss": 0.24004128575325012, + "num_tokens": 116937.0, + "reward": -20.38249969482422, + "reward_std": 87.55204010009766, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": -20.457500457763672, + "rewards/supergames_reward/std": 87.52605438232422, + "step": 13, + "step_time": 10.482285185018554 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 254.0, + "completions/max_terminated_length": 254.0, + "completions/mean_length": 156.875, + "completions/mean_terminated_length": 156.875, + "completions/min_length": 114.0, + "completions/min_terminated_length": 114.0, + "entropy": 0.7237679958343506, + "epoch": 0.07, + "frac_reward_zero_std": 0.0, + "grad_norm": 5.083980560302734, + "learning_rate": 9.35e-07, + "loss": -0.19025824964046478, + "num_tokens": 124640.0, + "reward": -51.397499084472656, + "reward_std": 54.486454010009766, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -51.49749755859375, + "rewards/supergames_reward/std": 54.486454010009766, + "step": 14, + "step_time": 9.597231683001155 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 236.0, + "completions/max_terminated_length": 236.0, + "completions/mean_length": 141.375, + "completions/mean_terminated_length": 141.375, + "completions/min_length": 91.0, + "completions/min_terminated_length": 91.0, + "entropy": 0.754334568977356, + "epoch": 0.075, + "frac_reward_zero_std": 0.0, + "grad_norm": 5.720428943634033, + "learning_rate": 9.3e-07, + "loss": -0.16830675303936005, + "num_tokens": 132219.0, + "reward": -56.226253509521484, + "reward_std": 60.115657806396484, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -56.32625198364258, + "rewards/supergames_reward/std": 60.11566162109375, + "step": 15, + "step_time": 8.998362872982398 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 209.0, + "completions/max_terminated_length": 209.0, + "completions/mean_length": 136.375, + "completions/mean_terminated_length": 136.375, + "completions/min_length": 97.0, + "completions/min_terminated_length": 97.0, + "entropy": 0.6280568838119507, + "epoch": 0.08, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.9681806564331055, + "learning_rate": 9.25e-07, + "loss": -0.01910916529595852, + "num_tokens": 139726.0, + "reward": -68.24000549316406, + "reward_std": 52.238914489746094, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -68.33999633789062, + "rewards/supergames_reward/std": 52.238914489746094, + "step": 16, + "step_time": 8.034480294008972 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 365.0, + "completions/max_terminated_length": 365.0, + "completions/mean_length": 211.875, + "completions/mean_terminated_length": 211.875, + "completions/min_length": 100.0, + "completions/min_terminated_length": 100.0, + "entropy": 0.899176836013794, + "epoch": 0.085, + "frac_reward_zero_std": 0.0, + "grad_norm": 6.691746711730957, + "learning_rate": 9.2e-07, + "loss": 0.25513291358947754, + "num_tokens": 147869.0, + "reward": -99.92500305175781, + "reward_std": 0.07070883363485336, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": -100.0, + "rewards/supergames_reward/std": 0.0, + "step": 17, + "step_time": 13.655792869016295 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 387.0, + "completions/max_terminated_length": 387.0, + "completions/mean_length": 238.0, + "completions/mean_terminated_length": 238.0, + "completions/min_length": 87.0, + "completions/min_terminated_length": 87.0, + "entropy": 0.5596873760223389, + "epoch": 0.09, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.6711347103118896, + "learning_rate": 9.15e-07, + "loss": -0.2628335654735565, + "num_tokens": 159101.0, + "reward": -24.900001525878906, + "reward_std": 103.50983428955078, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -25.0, + "rewards/supergames_reward/std": 103.50983428955078, + "step": 18, + "step_time": 14.784329420013819 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.125, + "completions/max_length": 512.0, + "completions/max_terminated_length": 277.0, + "completions/mean_length": 219.75, + "completions/mean_terminated_length": 178.00001525878906, + "completions/min_length": 90.0, + "completions/min_terminated_length": 90.0, + "entropy": 0.9245517253875732, + "epoch": 0.095, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.55502986907959, + "learning_rate": 9.1e-07, + "loss": 0.13060179352760315, + "num_tokens": 167299.0, + "reward": -86.89125061035156, + "reward_std": 36.95603561401367, + "rewards/json_format_reward/mean": 0.05000000074505806, + "rewards/json_format_reward/std": 0.09258200973272324, + "rewards/supergames_reward/mean": -86.94125366210938, + "rewards/supergames_reward/std": 36.93572235107422, + "step": 19, + "step_time": 18.832684700988466 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 267.0, + "completions/max_terminated_length": 267.0, + "completions/mean_length": 177.25, + "completions/mean_terminated_length": 177.25, + "completions/min_length": 93.0, + "completions/min_terminated_length": 93.0, + "entropy": 0.7331254482269287, + "epoch": 0.1, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.538497447967529, + "learning_rate": 9.05e-07, + "loss": -0.16418980062007904, + "num_tokens": 175141.0, + "reward": -68.52874755859375, + "reward_std": 58.48719787597656, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -68.62875366210938, + "rewards/supergames_reward/std": 58.48720169067383, + "step": 20, + "step_time": 10.090975169994636 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 300.0, + "completions/max_terminated_length": 300.0, + "completions/mean_length": 190.625, + "completions/mean_terminated_length": 190.625, + "completions/min_length": 103.0, + "completions/min_terminated_length": 103.0, + "entropy": 0.6659425497055054, + "epoch": 0.105, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.485373497009277, + "learning_rate": 9e-07, + "loss": -0.10106432437896729, + "num_tokens": 183098.0, + "reward": -71.48875427246094, + "reward_std": 55.33565139770508, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -71.5887451171875, + "rewards/supergames_reward/std": 55.33565139770508, + "step": 21, + "step_time": 11.255001295008697 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 279.0, + "completions/max_terminated_length": 279.0, + "completions/mean_length": 179.0, + "completions/mean_terminated_length": 179.0, + "completions/min_length": 103.0, + "completions/min_terminated_length": 103.0, + "entropy": 0.49365949630737305, + "epoch": 0.11, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.829419136047363, + "learning_rate": 8.95e-07, + "loss": 0.10587231814861298, + "num_tokens": 192826.0, + "reward": -75.62000274658203, + "reward_std": 45.01959991455078, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": -75.69499969482422, + "rewards/supergames_reward/std": 45.00411605834961, + "step": 22, + "step_time": 10.611246599000879 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 315.0, + "completions/max_terminated_length": 315.0, + "completions/mean_length": 212.625, + "completions/mean_terminated_length": 212.625, + "completions/min_length": 100.0, + "completions/min_terminated_length": 100.0, + "entropy": 0.43159469962120056, + "epoch": 0.115, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.541125774383545, + "learning_rate": 8.9e-07, + "loss": -0.1681259125471115, + "num_tokens": 202815.0, + "reward": 23.67624855041504, + "reward_std": 67.74340057373047, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 23.576250076293945, + "rewards/supergames_reward/std": 67.743408203125, + "step": 23, + "step_time": 11.731771531980485 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 168.0, + "completions/max_terminated_length": 168.0, + "completions/mean_length": 101.625, + "completions/mean_terminated_length": 101.625, + "completions/min_length": 65.0, + "completions/min_terminated_length": 65.0, + "entropy": 0.3972298502922058, + "epoch": 0.12, + "frac_reward_zero_std": 0.0, + "grad_norm": 5.8285417556762695, + "learning_rate": 8.85e-07, + "loss": -0.011585958302021027, + "num_tokens": 208188.0, + "reward": 58.849998474121094, + "reward_std": 56.9304084777832, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 58.75, + "rewards/supergames_reward/std": 56.9304084777832, + "step": 24, + "step_time": 6.187504950998118 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 303.0, + "completions/max_terminated_length": 303.0, + "completions/mean_length": 165.375, + "completions/mean_terminated_length": 165.375, + "completions/min_length": 69.0, + "completions/min_terminated_length": 69.0, + "entropy": 0.6339899301528931, + "epoch": 0.125, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.823799133300781, + "learning_rate": 8.799999999999999e-07, + "loss": 0.05159185826778412, + "num_tokens": 215935.0, + "reward": -23.521251678466797, + "reward_std": 74.23067474365234, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -23.621248245239258, + "rewards/supergames_reward/std": 74.23067474365234, + "step": 25, + "step_time": 11.059416300005978 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 349.0, + "completions/max_terminated_length": 349.0, + "completions/mean_length": 204.0, + "completions/mean_terminated_length": 204.0, + "completions/min_length": 143.0, + "completions/min_terminated_length": 143.0, + "entropy": 0.6462895274162292, + "epoch": 0.13, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.969292640686035, + "learning_rate": 8.75e-07, + "loss": 0.0115677984431386, + "num_tokens": 223999.0, + "reward": -74.8175048828125, + "reward_std": 40.94221115112305, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": -74.89250183105469, + "rewards/supergames_reward/std": 40.924625396728516, + "step": 26, + "step_time": 12.666237785975682 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 299.0, + "completions/max_terminated_length": 299.0, + "completions/mean_length": 166.5, + "completions/mean_terminated_length": 166.5, + "completions/min_length": 87.0, + "completions/min_terminated_length": 87.0, + "entropy": 0.4756850600242615, + "epoch": 0.135, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.915069103240967, + "learning_rate": 8.699999999999999e-07, + "loss": -0.13321346044540405, + "num_tokens": 233643.0, + "reward": -52.125, + "reward_std": 19.405282974243164, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": -52.20000076293945, + "rewards/supergames_reward/std": 19.33464813232422, + "step": 27, + "step_time": 11.172539752995363 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 184.0, + "completions/max_terminated_length": 184.0, + "completions/mean_length": 137.125, + "completions/mean_terminated_length": 137.125, + "completions/min_length": 100.0, + "completions/min_terminated_length": 100.0, + "entropy": 0.5834778547286987, + "epoch": 0.14, + "frac_reward_zero_std": 0.0, + "grad_norm": 5.146666526794434, + "learning_rate": 8.65e-07, + "loss": 0.060454584658145905, + "num_tokens": 241204.0, + "reward": -68.99250793457031, + "reward_std": 51.010581970214844, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -69.09249877929688, + "rewards/supergames_reward/std": 51.010581970214844, + "step": 28, + "step_time": 7.025046669004951 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 331.0, + "completions/max_terminated_length": 331.0, + "completions/mean_length": 182.375, + "completions/mean_terminated_length": 182.375, + "completions/min_length": 106.0, + "completions/min_terminated_length": 106.0, + "entropy": 0.6368551254272461, + "epoch": 0.145, + "frac_reward_zero_std": 0.0, + "grad_norm": 5.15535306930542, + "learning_rate": 8.599999999999999e-07, + "loss": -0.3323673903942108, + "num_tokens": 249087.0, + "reward": -44.17000198364258, + "reward_std": 80.5373306274414, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -44.27000045776367, + "rewards/supergames_reward/std": 80.5373306274414, + "step": 29, + "step_time": 12.079259724996518 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 190.0, + "completions/max_terminated_length": 190.0, + "completions/mean_length": 127.75, + "completions/mean_terminated_length": 127.75, + "completions/min_length": 80.0, + "completions/min_terminated_length": 80.0, + "entropy": 0.4491196870803833, + "epoch": 0.15, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.896186828613281, + "learning_rate": 8.55e-07, + "loss": -0.12887656688690186, + "num_tokens": 259445.0, + "reward": 57.446250915527344, + "reward_std": 38.78837585449219, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 57.34625244140625, + "rewards/supergames_reward/std": 38.78837966918945, + "step": 30, + "step_time": 7.483972909016302 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 166.0, + "completions/max_terminated_length": 166.0, + "completions/mean_length": 112.875, + "completions/mean_terminated_length": 112.875, + "completions/min_length": 85.0, + "completions/min_terminated_length": 85.0, + "entropy": 0.4858454465866089, + "epoch": 0.155, + "frac_reward_zero_std": 0.0, + "grad_norm": 6.742906093597412, + "learning_rate": 8.499999999999999e-07, + "loss": 0.06930096447467804, + "num_tokens": 264924.0, + "reward": -21.174999237060547, + "reward_std": 31.89051628112793, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": -21.25, + "rewards/supergames_reward/std": 31.819807052612305, + "step": 31, + "step_time": 6.1866529019898735 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 148.0, + "completions/max_terminated_length": 148.0, + "completions/mean_length": 118.125, + "completions/mean_terminated_length": 118.125, + "completions/min_length": 98.0, + "completions/min_terminated_length": 98.0, + "entropy": 0.45428749918937683, + "epoch": 0.16, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.663414478302002, + "learning_rate": 8.45e-07, + "loss": 0.05735419690608978, + "num_tokens": 270381.0, + "reward": 0.4687504768371582, + "reward_std": 38.66371154785156, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 0.36875057220458984, + "rewards/supergames_reward/std": 38.6637077331543, + "step": 32, + "step_time": 5.577042757999152 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 345.0, + "completions/max_terminated_length": 345.0, + "completions/mean_length": 166.875, + "completions/mean_terminated_length": 166.875, + "completions/min_length": 86.0, + "completions/min_terminated_length": 86.0, + "entropy": 0.6107293367385864, + "epoch": 0.165, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.75006628036499, + "learning_rate": 8.399999999999999e-07, + "loss": -0.14760127663612366, + "num_tokens": 278156.0, + "reward": -87.84750366210938, + "reward_std": 26.84708023071289, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -87.94749450683594, + "rewards/supergames_reward/std": 26.847076416015625, + "step": 33, + "step_time": 12.552876825997373 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.125, + "completions/max_length": 512.0, + "completions/max_terminated_length": 360.0, + "completions/mean_length": 231.0, + "completions/mean_terminated_length": 190.85714721679688, + "completions/min_length": 74.0, + "completions/min_terminated_length": 74.0, + "entropy": 0.6089653372764587, + "epoch": 0.17, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.781267166137695, + "learning_rate": 8.349999999999999e-07, + "loss": 0.4294504225254059, + "num_tokens": 289364.0, + "reward": -99.92500305175781, + "reward_std": 0.07070883363485336, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": -100.0, + "rewards/supergames_reward/std": 0.0, + "step": 34, + "step_time": 18.961819477990502 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 218.0, + "completions/max_terminated_length": 218.0, + "completions/mean_length": 134.125, + "completions/mean_terminated_length": 134.125, + "completions/min_length": 104.0, + "completions/min_terminated_length": 104.0, + "entropy": 0.46158918738365173, + "epoch": 0.175, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.016353607177734, + "learning_rate": 8.299999999999999e-07, + "loss": 0.09251818805932999, + "num_tokens": 298733.0, + "reward": -35.558753967285156, + "reward_std": 15.881204605102539, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -35.65875244140625, + "rewards/supergames_reward/std": 15.881203651428223, + "step": 35, + "step_time": 8.38491243700264 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 207.0, + "completions/max_terminated_length": 207.0, + "completions/mean_length": 156.875, + "completions/mean_terminated_length": 156.875, + "completions/min_length": 111.0, + "completions/min_terminated_length": 111.0, + "entropy": 0.5104788541793823, + "epoch": 0.18, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.592893123626709, + "learning_rate": 8.249999999999999e-07, + "loss": -0.0410832017660141, + "num_tokens": 306420.0, + "reward": 5.4212493896484375, + "reward_std": 93.50404357910156, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": 5.346250534057617, + "rewards/supergames_reward/std": 93.4718246459961, + "step": 36, + "step_time": 7.811868985998444 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 241.0, + "completions/max_terminated_length": 241.0, + "completions/mean_length": 178.0, + "completions/mean_terminated_length": 178.0, + "completions/min_length": 106.0, + "completions/min_terminated_length": 106.0, + "entropy": 0.7527878880500793, + "epoch": 0.185, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.8812546730041504, + "learning_rate": 8.199999999999999e-07, + "loss": 0.12539231777191162, + "num_tokens": 314252.0, + "reward": -89.58500671386719, + "reward_std": 29.256120681762695, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": -89.66000366210938, + "rewards/supergames_reward/std": 29.24593734741211, + "step": 37, + "step_time": 8.959742914012168 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 199.0, + "completions/max_terminated_length": 199.0, + "completions/mean_length": 133.375, + "completions/mean_terminated_length": 133.375, + "completions/min_length": 87.0, + "completions/min_terminated_length": 87.0, + "entropy": 0.3347901403903961, + "epoch": 0.19, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.205414295196533, + "learning_rate": 8.149999999999999e-07, + "loss": -0.06991486996412277, + "num_tokens": 319839.0, + "reward": -14.498749732971191, + "reward_std": 13.007229804992676, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -14.598750114440918, + "rewards/supergames_reward/std": 13.007229804992676, + "step": 38, + "step_time": 7.377183554985095 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 181.0, + "completions/max_terminated_length": 181.0, + "completions/mean_length": 114.375, + "completions/mean_terminated_length": 114.375, + "completions/min_length": 58.0, + "completions/min_terminated_length": 58.0, + "entropy": 0.4445965886116028, + "epoch": 0.195, + "frac_reward_zero_std": 0.0, + "grad_norm": 5.561298847198486, + "learning_rate": 8.1e-07, + "loss": -0.06235164776444435, + "num_tokens": 325290.0, + "reward": -7.425000190734863, + "reward_std": 53.701602935791016, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": -7.5, + "rewards/supergames_reward/std": 53.652320861816406, + "step": 39, + "step_time": 6.753862089011818 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 370.0, + "completions/max_terminated_length": 370.0, + "completions/mean_length": 187.625, + "completions/mean_terminated_length": 187.625, + "completions/min_length": 60.0, + "completions/min_terminated_length": 60.0, + "entropy": 0.3950234353542328, + "epoch": 0.2, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.5988712310791016, + "learning_rate": 8.05e-07, + "loss": -0.1381440907716751, + "num_tokens": 335119.0, + "reward": -33.68375015258789, + "reward_std": 42.069435119628906, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -33.783748626708984, + "rewards/supergames_reward/std": 42.069435119628906, + "step": 40, + "step_time": 13.718958162004128 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.125, + "completions/max_length": 512.0, + "completions/max_terminated_length": 371.0, + "completions/mean_length": 251.125, + "completions/mean_terminated_length": 213.85714721679688, + "completions/min_length": 121.0, + "completions/min_terminated_length": 121.0, + "entropy": 0.507067084312439, + "epoch": 0.205, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.3898658752441406, + "learning_rate": 8e-07, + "loss": 0.0078964838758111, + "num_tokens": 346464.0, + "reward": -86.48625183105469, + "reward_std": 38.020694732666016, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": -86.56124877929688, + "rewards/supergames_reward/std": 38.010528564453125, + "step": 41, + "step_time": 18.90960379401804 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 152.0, + "completions/max_terminated_length": 152.0, + "completions/mean_length": 116.375, + "completions/mean_terminated_length": 116.375, + "completions/min_length": 93.0, + "completions/min_terminated_length": 93.0, + "entropy": 0.3801707923412323, + "epoch": 0.21, + "frac_reward_zero_std": 1.0, + "grad_norm": 0.0, + "learning_rate": 7.95e-07, + "loss": 0.0, + "num_tokens": 351955.0, + "reward": -9.899999618530273, + "reward_std": 0.0, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -10.0, + "rewards/supergames_reward/std": 0.0, + "step": 42, + "step_time": 5.711630532023264 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 308.0, + "completions/max_terminated_length": 308.0, + "completions/mean_length": 247.375, + "completions/mean_terminated_length": 247.375, + "completions/min_length": 116.0, + "completions/min_terminated_length": 116.0, + "entropy": 0.3185043931007385, + "epoch": 0.215, + "frac_reward_zero_std": 0.0, + "grad_norm": 2.168762445449829, + "learning_rate": 7.9e-07, + "loss": 0.1970764398574829, + "num_tokens": 363278.0, + "reward": -53.89249801635742, + "reward_std": 63.02619934082031, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -53.99250030517578, + "rewards/supergames_reward/std": 63.02619934082031, + "step": 43, + "step_time": 11.661934906995157 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 330.0, + "completions/max_terminated_length": 330.0, + "completions/mean_length": 212.5, + "completions/mean_terminated_length": 212.5, + "completions/min_length": 106.0, + "completions/min_terminated_length": 106.0, + "entropy": 0.4446268379688263, + "epoch": 0.22, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.305844306945801, + "learning_rate": 7.85e-07, + "loss": -0.06735197454690933, + "num_tokens": 374306.0, + "reward": 18.059999465942383, + "reward_std": 98.9417953491211, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 17.959999084472656, + "rewards/supergames_reward/std": 98.9417953491211, + "step": 44, + "step_time": 12.492453911982011 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 316.0, + "completions/max_terminated_length": 316.0, + "completions/mean_length": 183.0, + "completions/mean_terminated_length": 183.0, + "completions/min_length": 103.0, + "completions/min_terminated_length": 103.0, + "entropy": 0.5269614458084106, + "epoch": 0.225, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.8006370067596436, + "learning_rate": 7.799999999999999e-07, + "loss": -0.08870815485715866, + "num_tokens": 382210.0, + "reward": -54.928749084472656, + "reward_std": 72.5127944946289, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -55.02874755859375, + "rewards/supergames_reward/std": 72.5127944946289, + "step": 45, + "step_time": 11.7636467939883 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 244.0, + "completions/max_terminated_length": 244.0, + "completions/mean_length": 136.625, + "completions/mean_terminated_length": 136.625, + "completions/min_length": 85.0, + "completions/min_terminated_length": 85.0, + "entropy": 0.34820204973220825, + "epoch": 0.23, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.911736011505127, + "learning_rate": 7.75e-07, + "loss": 0.04397330805659294, + "num_tokens": 391583.0, + "reward": -19.10375213623047, + "reward_std": 23.967702865600586, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -19.203750610351562, + "rewards/supergames_reward/std": 23.967702865600586, + "step": 46, + "step_time": 10.119426151999505 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 218.0, + "completions/max_terminated_length": 218.0, + "completions/mean_length": 160.375, + "completions/mean_terminated_length": 160.375, + "completions/min_length": 93.0, + "completions/min_terminated_length": 93.0, + "entropy": 0.611914336681366, + "epoch": 0.235, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.571913242340088, + "learning_rate": 7.699999999999999e-07, + "loss": -0.044293053448200226, + "num_tokens": 399290.0, + "reward": -25.186250686645508, + "reward_std": 71.46305084228516, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -25.286251068115234, + "rewards/supergames_reward/std": 71.46305084228516, + "step": 47, + "step_time": 8.563868903991533 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 167.0, + "completions/max_terminated_length": 167.0, + "completions/mean_length": 111.125, + "completions/mean_terminated_length": 111.125, + "completions/min_length": 56.0, + "completions/min_terminated_length": 56.0, + "entropy": 0.46585777401924133, + "epoch": 0.24, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.654109954833984, + "learning_rate": 7.65e-07, + "loss": -0.035151124000549316, + "num_tokens": 404707.0, + "reward": -0.408750057220459, + "reward_std": 31.939748764038086, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -0.5087499618530273, + "rewards/supergames_reward/std": 31.93975257873535, + "step": 48, + "step_time": 6.438482580007985 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 306.0, + "completions/max_terminated_length": 306.0, + "completions/mean_length": 216.625, + "completions/mean_terminated_length": 216.625, + "completions/min_length": 132.0, + "completions/min_terminated_length": 132.0, + "entropy": 0.33054471015930176, + "epoch": 0.245, + "frac_reward_zero_std": 0.0, + "grad_norm": 2.700303792953491, + "learning_rate": 7.599999999999999e-07, + "loss": -0.04393656551837921, + "num_tokens": 414736.0, + "reward": -97.26000213623047, + "reward_std": 1.6294406652450562, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -97.36000061035156, + "rewards/supergames_reward/std": 1.629441738128662, + "step": 49, + "step_time": 11.636040106008295 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 378.0, + "completions/max_terminated_length": 378.0, + "completions/mean_length": 218.375, + "completions/mean_terminated_length": 218.375, + "completions/min_length": 104.0, + "completions/min_terminated_length": 104.0, + "entropy": 0.48711177706718445, + "epoch": 0.25, + "frac_reward_zero_std": 0.0, + "grad_norm": 2.9650564193725586, + "learning_rate": 7.55e-07, + "loss": -0.3529520332813263, + "num_tokens": 424795.0, + "reward": 47.408748626708984, + "reward_std": 72.72083282470703, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 47.30875015258789, + "rewards/supergames_reward/std": 72.72083282470703, + "step": 50, + "step_time": 14.258096842997475 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.125, + "completions/max_length": 512.0, + "completions/max_terminated_length": 492.0, + "completions/mean_length": 267.625, + "completions/mean_terminated_length": 232.71429443359375, + "completions/min_length": 90.0, + "completions/min_terminated_length": 90.0, + "entropy": 0.4482860863208771, + "epoch": 0.255, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.2998197078704834, + "learning_rate": 7.5e-07, + "loss": 0.2493654191493988, + "num_tokens": 436288.0, + "reward": 50.61125183105469, + "reward_std": 75.3349838256836, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": 50.53624725341797, + "rewards/supergames_reward/std": 75.27783966064453, + "step": 51, + "step_time": 19.20827590499539 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 176.0, + "completions/max_terminated_length": 176.0, + "completions/mean_length": 144.5, + "completions/mean_terminated_length": 144.5, + "completions/min_length": 99.0, + "completions/min_terminated_length": 99.0, + "entropy": 0.5378735661506653, + "epoch": 0.26, + "frac_reward_zero_std": 0.0, + "grad_norm": 5.114461898803711, + "learning_rate": 7.45e-07, + "loss": 0.0036597400903701782, + "num_tokens": 441980.0, + "reward": -15.20625114440918, + "reward_std": 45.35526657104492, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": -15.28125, + "rewards/supergames_reward/std": 45.301815032958984, + "step": 52, + "step_time": 6.660627231001854 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 136.0, + "completions/max_terminated_length": 136.0, + "completions/mean_length": 119.0, + "completions/mean_terminated_length": 119.0, + "completions/min_length": 105.0, + "completions/min_terminated_length": 105.0, + "entropy": 0.41796669363975525, + "epoch": 0.265, + "frac_reward_zero_std": 0.0, + "grad_norm": 5.059505462646484, + "learning_rate": 7.4e-07, + "loss": -0.014973883517086506, + "num_tokens": 447468.0, + "reward": 15.170000076293945, + "reward_std": 46.658329010009766, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 15.070000648498535, + "rewards/supergames_reward/std": 46.658329010009766, + "step": 53, + "step_time": 5.271571868011961 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 259.0, + "completions/max_terminated_length": 259.0, + "completions/mean_length": 191.125, + "completions/mean_terminated_length": 191.125, + "completions/min_length": 91.0, + "completions/min_terminated_length": 91.0, + "entropy": 0.45421987771987915, + "epoch": 0.27, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.4454867839813232, + "learning_rate": 7.35e-07, + "loss": -0.1128973588347435, + "num_tokens": 457301.0, + "reward": 38.89875030517578, + "reward_std": 62.1772575378418, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 38.79874801635742, + "rewards/supergames_reward/std": 62.1772575378418, + "step": 54, + "step_time": 10.524528659996577 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 319.0, + "completions/max_terminated_length": 319.0, + "completions/mean_length": 233.375, + "completions/mean_terminated_length": 233.375, + "completions/min_length": 136.0, + "completions/min_terminated_length": 136.0, + "entropy": 0.4449213445186615, + "epoch": 0.275, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.2065398693084717, + "learning_rate": 7.3e-07, + "loss": 0.0016277075046673417, + "num_tokens": 467520.0, + "reward": 27.372499465942383, + "reward_std": 81.23126983642578, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 27.27250099182129, + "rewards/supergames_reward/std": 81.23126983642578, + "step": 55, + "step_time": 12.592280682991259 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.125, + "completions/max_length": 512.0, + "completions/max_terminated_length": 357.0, + "completions/mean_length": 261.0, + "completions/mean_terminated_length": 225.1428680419922, + "completions/min_length": 95.0, + "completions/min_terminated_length": 95.0, + "entropy": 0.506920337677002, + "epoch": 0.28, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.058388710021973, + "learning_rate": 7.249999999999999e-07, + "loss": 0.3289242386817932, + "num_tokens": 476056.0, + "reward": 47.05249786376953, + "reward_std": 80.97753143310547, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": 46.977500915527344, + "rewards/supergames_reward/std": 80.9256362915039, + "step": 56, + "step_time": 18.845177220006008 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 250.0, + "completions/max_terminated_length": 250.0, + "completions/mean_length": 192.75, + "completions/mean_terminated_length": 192.75, + "completions/min_length": 122.0, + "completions/min_terminated_length": 122.0, + "entropy": 0.5197336077690125, + "epoch": 0.285, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.230815410614014, + "learning_rate": 7.2e-07, + "loss": -0.15186546742916107, + "num_tokens": 484014.0, + "reward": -64.5050048828125, + "reward_std": 50.35145568847656, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -64.60499572753906, + "rewards/supergames_reward/std": 50.3514518737793, + "step": 57, + "step_time": 9.576232638006331 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 324.0, + "completions/max_terminated_length": 324.0, + "completions/mean_length": 249.0, + "completions/mean_terminated_length": 249.0, + "completions/min_length": 129.0, + "completions/min_terminated_length": 129.0, + "entropy": 0.42913514375686646, + "epoch": 0.29, + "frac_reward_zero_std": 0.0, + "grad_norm": 2.8544528484344482, + "learning_rate": 7.149999999999999e-07, + "loss": -0.056405920535326004, + "num_tokens": 494390.0, + "reward": 40.123748779296875, + "reward_std": 84.164794921875, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 40.02375030517578, + "rewards/supergames_reward/std": 84.164794921875, + "step": 58, + "step_time": 12.385353341000155 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 138.0, + "completions/max_terminated_length": 138.0, + "completions/mean_length": 107.25, + "completions/mean_terminated_length": 107.25, + "completions/min_length": 81.0, + "completions/min_terminated_length": 81.0, + "entropy": 0.34386488795280457, + "epoch": 0.295, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.6915154457092285, + "learning_rate": 7.1e-07, + "loss": 0.0865321159362793, + "num_tokens": 499816.0, + "reward": -8.244999885559082, + "reward_std": 4.681046962738037, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -8.345000267028809, + "rewards/supergames_reward/std": 4.681046962738037, + "step": 59, + "step_time": 5.42325339600211 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 187.0, + "completions/max_terminated_length": 187.0, + "completions/mean_length": 127.25, + "completions/mean_terminated_length": 127.25, + "completions/min_length": 97.0, + "completions/min_terminated_length": 97.0, + "entropy": 0.5430045127868652, + "epoch": 0.3, + "frac_reward_zero_std": 0.0, + "grad_norm": 5.635311126708984, + "learning_rate": 7.049999999999999e-07, + "loss": -0.05412375554442406, + "num_tokens": 505410.0, + "reward": 20.005001068115234, + "reward_std": 55.156005859375, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 19.904998779296875, + "rewards/supergames_reward/std": 55.156005859375, + "step": 60, + "step_time": 7.068010902003152 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 286.0, + "completions/max_terminated_length": 286.0, + "completions/mean_length": 192.125, + "completions/mean_terminated_length": 192.125, + "completions/min_length": 95.0, + "completions/min_terminated_length": 95.0, + "entropy": 0.4213869869709015, + "epoch": 0.305, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.4664618968963623, + "learning_rate": 7e-07, + "loss": -0.0495159812271595, + "num_tokens": 516283.0, + "reward": -47.212501525878906, + "reward_std": 91.2253646850586, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -47.3125, + "rewards/supergames_reward/std": 91.2253646850586, + "step": 61, + "step_time": 11.176304235996213 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 324.0, + "completions/max_terminated_length": 324.0, + "completions/mean_length": 171.0, + "completions/mean_terminated_length": 171.0, + "completions/min_length": 92.0, + "completions/min_terminated_length": 92.0, + "entropy": 0.4626811146736145, + "epoch": 0.31, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.949252128601074, + "learning_rate": 6.949999999999999e-07, + "loss": 0.192047581076622, + "num_tokens": 522251.0, + "reward": -15.092500686645508, + "reward_std": 49.82962417602539, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": -15.167499542236328, + "rewards/supergames_reward/std": 49.78091049194336, + "step": 62, + "step_time": 12.075224861997413 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.125, + "completions/max_length": 512.0, + "completions/max_terminated_length": 429.0, + "completions/mean_length": 283.5, + "completions/mean_terminated_length": 250.85714721679688, + "completions/min_length": 85.0, + "completions/min_terminated_length": 85.0, + "entropy": 0.4814216196537018, + "epoch": 0.315, + "frac_reward_zero_std": 0.0, + "grad_norm": 2.7918753623962402, + "learning_rate": 6.9e-07, + "loss": 0.15706156194210052, + "num_tokens": 533815.0, + "reward": -20.86625099182129, + "reward_std": 85.99250030517578, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": -20.94124984741211, + "rewards/supergames_reward/std": 85.96620178222656, + "step": 63, + "step_time": 19.448832260008203 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 189.0, + "completions/max_terminated_length": 189.0, + "completions/mean_length": 131.25, + "completions/mean_terminated_length": 131.25, + "completions/min_length": 106.0, + "completions/min_terminated_length": 106.0, + "entropy": 0.4929414987564087, + "epoch": 0.32, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.774101734161377, + "learning_rate": 6.85e-07, + "loss": -0.060691747814416885, + "num_tokens": 539465.0, + "reward": -21.0674991607666, + "reward_std": 6.892730236053467, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -21.167499542236328, + "rewards/supergames_reward/std": 6.892730236053467, + "step": 64, + "step_time": 7.141569407976931 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 454.0, + "completions/max_terminated_length": 454.0, + "completions/mean_length": 281.375, + "completions/mean_terminated_length": 281.375, + "completions/min_length": 107.0, + "completions/min_terminated_length": 107.0, + "entropy": 0.45866572856903076, + "epoch": 0.325, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.092717409133911, + "learning_rate": 6.800000000000001e-07, + "loss": 0.11694959551095963, + "num_tokens": 551012.0, + "reward": 49.938751220703125, + "reward_std": 79.1865005493164, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 49.8387451171875, + "rewards/supergames_reward/std": 79.1865005493164, + "step": 65, + "step_time": 17.06170882002334 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 203.0, + "completions/max_terminated_length": 203.0, + "completions/mean_length": 128.625, + "completions/mean_terminated_length": 128.625, + "completions/min_length": 88.0, + "completions/min_terminated_length": 88.0, + "entropy": 0.5005782246589661, + "epoch": 0.33, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.77616024017334, + "learning_rate": 6.75e-07, + "loss": 0.11242837458848953, + "num_tokens": 556577.0, + "reward": 48.875, + "reward_std": 49.13539123535156, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 48.775001525878906, + "rewards/supergames_reward/std": 49.13539123535156, + "step": 66, + "step_time": 7.695410129002994 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 436.0, + "completions/max_terminated_length": 436.0, + "completions/mean_length": 281.875, + "completions/mean_terminated_length": 281.875, + "completions/min_length": 133.0, + "completions/min_terminated_length": 133.0, + "entropy": 0.507097065448761, + "epoch": 0.335, + "frac_reward_zero_std": 0.0, + "grad_norm": 2.911184787750244, + "learning_rate": 6.7e-07, + "loss": -0.08151137083768845, + "num_tokens": 567120.0, + "reward": -12.886249542236328, + "reward_std": 84.04269409179688, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -12.986251831054688, + "rewards/supergames_reward/std": 84.04269409179688, + "step": 67, + "step_time": 16.409127942984924 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.125, + "completions/max_length": 512.0, + "completions/max_terminated_length": 501.0, + "completions/mean_length": 318.375, + "completions/mean_terminated_length": 290.71429443359375, + "completions/min_length": 65.0, + "completions/min_terminated_length": 65.0, + "entropy": 0.486402690410614, + "epoch": 0.34, + "frac_reward_zero_std": 0.0, + "grad_norm": 2.9461607933044434, + "learning_rate": 6.65e-07, + "loss": 0.2442178875207901, + "num_tokens": 578995.0, + "reward": 2.8387489318847656, + "reward_std": 98.47579956054688, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": 2.7637500762939453, + "rewards/supergames_reward/std": 98.44595336914062, + "step": 68, + "step_time": 19.32364999302081 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 291.0, + "completions/max_terminated_length": 291.0, + "completions/mean_length": 219.625, + "completions/mean_terminated_length": 219.625, + "completions/min_length": 117.0, + "completions/min_terminated_length": 117.0, + "entropy": 0.37856727838516235, + "epoch": 0.345, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.0204732418060303, + "learning_rate": 6.6e-07, + "loss": -0.08764916658401489, + "num_tokens": 589072.0, + "reward": -54.18375015258789, + "reward_std": 71.45012664794922, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -54.283748626708984, + "rewards/supergames_reward/std": 71.45013427734375, + "step": 69, + "step_time": 11.126611230982235 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.125, + "completions/max_length": 512.0, + "completions/max_terminated_length": 305.0, + "completions/mean_length": 254.75, + "completions/mean_terminated_length": 218.00001525878906, + "completions/min_length": 104.0, + "completions/min_terminated_length": 104.0, + "entropy": 0.46423614025115967, + "epoch": 0.35, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.0140154361724854, + "learning_rate": 6.55e-07, + "loss": 0.26797160506248474, + "num_tokens": 600438.0, + "reward": -32.04875183105469, + "reward_std": 69.86957550048828, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": -32.12375259399414, + "rewards/supergames_reward/std": 69.84178161621094, + "step": 70, + "step_time": 19.12617294798838 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 372.0, + "completions/max_terminated_length": 372.0, + "completions/mean_length": 194.125, + "completions/mean_terminated_length": 194.125, + "completions/min_length": 94.0, + "completions/min_terminated_length": 94.0, + "entropy": 0.3427293300628662, + "epoch": 0.355, + "frac_reward_zero_std": 0.0, + "grad_norm": 2.781383991241455, + "learning_rate": 6.5e-07, + "loss": 0.354524701833725, + "num_tokens": 610327.0, + "reward": -6.913749694824219, + "reward_std": 87.56071472167969, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": -6.988750457763672, + "rewards/supergames_reward/std": 87.53033447265625, + "step": 71, + "step_time": 14.041668199002743 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 197.0, + "completions/max_terminated_length": 197.0, + "completions/mean_length": 124.75, + "completions/mean_terminated_length": 124.75, + "completions/min_length": 93.0, + "completions/min_terminated_length": 93.0, + "entropy": 0.4455118477344513, + "epoch": 0.36, + "frac_reward_zero_std": 0.0, + "grad_norm": 5.726032733917236, + "learning_rate": 6.45e-07, + "loss": 0.18206289410591125, + "num_tokens": 615853.0, + "reward": 41.95124816894531, + "reward_std": 60.640811920166016, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 41.85124969482422, + "rewards/supergames_reward/std": 60.640811920166016, + "step": 72, + "step_time": 7.416493425989756 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 323.0, + "completions/max_terminated_length": 323.0, + "completions/mean_length": 147.375, + "completions/mean_terminated_length": 147.375, + "completions/min_length": 94.0, + "completions/min_terminated_length": 94.0, + "entropy": 0.480733186006546, + "epoch": 0.365, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.853621482849121, + "learning_rate": 6.4e-07, + "loss": 0.04558124020695686, + "num_tokens": 621584.0, + "reward": 14.884998321533203, + "reward_std": 46.25857162475586, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 14.78499984741211, + "rewards/supergames_reward/std": 46.25857162475586, + "step": 73, + "step_time": 11.95379540900467 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 283.0, + "completions/max_terminated_length": 283.0, + "completions/mean_length": 188.125, + "completions/mean_terminated_length": 188.125, + "completions/min_length": 116.0, + "completions/min_terminated_length": 116.0, + "entropy": 0.5526013374328613, + "epoch": 0.37, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.729883909225464, + "learning_rate": 6.35e-07, + "loss": 0.06268303096294403, + "num_tokens": 629545.0, + "reward": -68.05750274658203, + "reward_std": 59.44808578491211, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -68.15750122070312, + "rewards/supergames_reward/std": 59.44808578491211, + "step": 74, + "step_time": 10.64124580900534 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 141.0, + "completions/max_terminated_length": 141.0, + "completions/mean_length": 117.375, + "completions/mean_terminated_length": 117.375, + "completions/min_length": 100.0, + "completions/min_terminated_length": 100.0, + "entropy": 0.3744319677352905, + "epoch": 0.375, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.775154113769531, + "learning_rate": 6.3e-07, + "loss": -0.023284845054149628, + "num_tokens": 635052.0, + "reward": 3.9374990463256836, + "reward_std": 61.85771179199219, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": 3.862499237060547, + "rewards/supergames_reward/std": 61.80967712402344, + "step": 75, + "step_time": 5.432648951013107 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 268.0, + "completions/max_terminated_length": 268.0, + "completions/mean_length": 157.0, + "completions/mean_terminated_length": 157.0, + "completions/min_length": 104.0, + "completions/min_terminated_length": 104.0, + "entropy": 0.4822140336036682, + "epoch": 0.38, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.715704917907715, + "learning_rate": 6.249999999999999e-07, + "loss": -0.04918142035603523, + "num_tokens": 644644.0, + "reward": 48.64125061035156, + "reward_std": 78.0205078125, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": 48.56624984741211, + "rewards/supergames_reward/std": 77.96605682373047, + "step": 76, + "step_time": 10.348935816989979 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 380.0, + "completions/max_terminated_length": 380.0, + "completions/mean_length": 210.875, + "completions/mean_terminated_length": 210.875, + "completions/min_length": 126.0, + "completions/min_terminated_length": 126.0, + "entropy": 0.45722317695617676, + "epoch": 0.385, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.6057567596435547, + "learning_rate": 6.2e-07, + "loss": 0.058352649211883545, + "num_tokens": 652771.0, + "reward": -8.283750534057617, + "reward_std": 77.68436431884766, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -8.383749008178711, + "rewards/supergames_reward/std": 77.68437194824219, + "step": 77, + "step_time": 14.142948755004909 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 324.0, + "completions/max_terminated_length": 324.0, + "completions/mean_length": 188.25, + "completions/mean_terminated_length": 188.25, + "completions/min_length": 68.0, + "completions/min_terminated_length": 68.0, + "entropy": 0.33380764722824097, + "epoch": 0.39, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.2239863872528076, + "learning_rate": 6.149999999999999e-07, + "loss": -0.04592633992433548, + "num_tokens": 662653.0, + "reward": 11.153749465942383, + "reward_std": 89.88329315185547, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 11.053749084472656, + "rewards/supergames_reward/std": 89.88329315185547, + "step": 78, + "step_time": 12.277474621019792 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 143.0, + "completions/max_terminated_length": 143.0, + "completions/mean_length": 108.875, + "completions/mean_terminated_length": 108.875, + "completions/min_length": 85.0, + "completions/min_terminated_length": 85.0, + "entropy": 0.3518812656402588, + "epoch": 0.395, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.967776298522949, + "learning_rate": 6.1e-07, + "loss": -0.10800496488809586, + "num_tokens": 668076.0, + "reward": 70.33250427246094, + "reward_std": 49.9145393371582, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 70.23249816894531, + "rewards/supergames_reward/std": 49.91453552246094, + "step": 79, + "step_time": 5.52714040101273 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 184.0, + "completions/max_terminated_length": 184.0, + "completions/mean_length": 152.125, + "completions/mean_terminated_length": 152.125, + "completions/min_length": 104.0, + "completions/min_terminated_length": 104.0, + "entropy": 0.4699896574020386, + "epoch": 0.4, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.561161994934082, + "learning_rate": 6.049999999999999e-07, + "loss": 0.09505834430456161, + "num_tokens": 675701.0, + "reward": -20.25625228881836, + "reward_std": 53.78542709350586, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -20.35624885559082, + "rewards/supergames_reward/std": 53.78542709350586, + "step": 80, + "step_time": 7.155081019998761 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.125, + "completions/max_length": 512.0, + "completions/max_terminated_length": 506.0, + "completions/mean_length": 288.875, + "completions/mean_terminated_length": 257.0, + "completions/min_length": 115.0, + "completions/min_terminated_length": 115.0, + "entropy": 0.4086494445800781, + "epoch": 0.405, + "frac_reward_zero_std": 0.0, + "grad_norm": 2.3719611167907715, + "learning_rate": 6e-07, + "loss": 0.04727627709507942, + "num_tokens": 687332.0, + "reward": -37.20750045776367, + "reward_std": 91.7676010131836, + "rewards/json_format_reward/mean": 0.05000000074505806, + "rewards/json_format_reward/std": 0.09258200973272324, + "rewards/supergames_reward/mean": -37.25749969482422, + "rewards/supergames_reward/std": 91.72847747802734, + "step": 81, + "step_time": 19.231399144016905 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 376.0, + "completions/max_terminated_length": 376.0, + "completions/mean_length": 227.75, + "completions/mean_terminated_length": 227.75, + "completions/min_length": 39.0, + "completions/min_terminated_length": 39.0, + "entropy": 0.43233707547187805, + "epoch": 0.41, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.3391714096069336, + "learning_rate": 5.949999999999999e-07, + "loss": -0.25524550676345825, + "num_tokens": 698474.0, + "reward": -3.7287511825561523, + "reward_std": 91.25379943847656, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -3.828749656677246, + "rewards/supergames_reward/std": 91.25379943847656, + "step": 82, + "step_time": 14.329176368017215 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 299.0, + "completions/max_terminated_length": 299.0, + "completions/mean_length": 178.125, + "completions/mean_terminated_length": 178.125, + "completions/min_length": 91.0, + "completions/min_terminated_length": 91.0, + "entropy": 0.4293636977672577, + "epoch": 0.415, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.7271742820739746, + "learning_rate": 5.9e-07, + "loss": 0.09273672848939896, + "num_tokens": 708243.0, + "reward": 12.802498817443848, + "reward_std": 86.23526000976562, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 12.702500343322754, + "rewards/supergames_reward/std": 86.23526000976562, + "step": 83, + "step_time": 11.419686011999147 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 425.0, + "completions/max_terminated_length": 425.0, + "completions/mean_length": 255.375, + "completions/mean_terminated_length": 255.375, + "completions/min_length": 172.0, + "completions/min_terminated_length": 172.0, + "entropy": 0.38068246841430664, + "epoch": 0.42, + "frac_reward_zero_std": 0.0, + "grad_norm": 2.6244184970855713, + "learning_rate": 5.849999999999999e-07, + "loss": 0.02006213366985321, + "num_tokens": 718606.0, + "reward": 37.803749084472656, + "reward_std": 57.58624267578125, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 37.70375061035156, + "rewards/supergames_reward/std": 57.58624267578125, + "step": 84, + "step_time": 15.975198492989875 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 116.0, + "completions/max_terminated_length": 116.0, + "completions/mean_length": 108.125, + "completions/mean_terminated_length": 108.125, + "completions/min_length": 87.0, + "completions/min_terminated_length": 87.0, + "entropy": 0.339572012424469, + "epoch": 0.425, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.754759311676025, + "learning_rate": 5.8e-07, + "loss": -0.011015941388905048, + "num_tokens": 723983.0, + "reward": -3.8212504386901855, + "reward_std": 61.624786376953125, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -3.921250820159912, + "rewards/supergames_reward/std": 61.624786376953125, + "step": 85, + "step_time": 4.542777584982105 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 447.0, + "completions/max_terminated_length": 447.0, + "completions/mean_length": 291.375, + "completions/mean_terminated_length": 291.375, + "completions/min_length": 163.0, + "completions/min_terminated_length": 163.0, + "entropy": 0.5846173763275146, + "epoch": 0.43, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.55724835395813, + "learning_rate": 5.749999999999999e-07, + "loss": -0.04981120675802231, + "num_tokens": 735642.0, + "reward": 50.07499694824219, + "reward_std": 92.6283187866211, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": 50.0, + "rewards/supergames_reward/std": 92.5820083618164, + "step": 86, + "step_time": 16.777178087009815 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 232.0, + "completions/max_terminated_length": 232.0, + "completions/mean_length": 192.0, + "completions/mean_terminated_length": 192.0, + "completions/min_length": 139.0, + "completions/min_terminated_length": 139.0, + "entropy": 0.5139474272727966, + "epoch": 0.435, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.8860411643981934, + "learning_rate": 5.699999999999999e-07, + "loss": -0.02194221317768097, + "num_tokens": 743602.0, + "reward": -52.522499084472656, + "reward_std": 73.8686294555664, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -52.62249755859375, + "rewards/supergames_reward/std": 73.86863708496094, + "step": 87, + "step_time": 8.772893911984283 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 266.0, + "completions/max_terminated_length": 266.0, + "completions/mean_length": 146.875, + "completions/mean_terminated_length": 146.875, + "completions/min_length": 89.0, + "completions/min_terminated_length": 89.0, + "entropy": 0.3428497910499573, + "epoch": 0.44, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.509483337402344, + "learning_rate": 5.649999999999999e-07, + "loss": 0.19857533276081085, + "num_tokens": 753065.0, + "reward": 63.63874816894531, + "reward_std": 67.54963684082031, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 63.53874969482422, + "rewards/supergames_reward/std": 67.54963684082031, + "step": 88, + "step_time": 10.230529835011112 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 446.0, + "completions/max_terminated_length": 446.0, + "completions/mean_length": 270.625, + "completions/mean_terminated_length": 270.625, + "completions/min_length": 103.0, + "completions/min_terminated_length": 103.0, + "entropy": 0.4417419731616974, + "epoch": 0.445, + "frac_reward_zero_std": 0.0, + "grad_norm": 2.973951816558838, + "learning_rate": 5.6e-07, + "loss": -0.00031440958264283836, + "num_tokens": 764534.0, + "reward": -9.736251831054688, + "reward_std": 92.25879669189453, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -9.836250305175781, + "rewards/supergames_reward/std": 92.25880432128906, + "step": 89, + "step_time": 17.40752330099349 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 398.0, + "completions/max_terminated_length": 398.0, + "completions/mean_length": 227.375, + "completions/mean_terminated_length": 227.375, + "completions/min_length": 98.0, + "completions/min_terminated_length": 98.0, + "entropy": 0.4732024073600769, + "epoch": 0.45, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.2560815811157227, + "learning_rate": 5.55e-07, + "loss": 0.03820464387536049, + "num_tokens": 772809.0, + "reward": 5.982499599456787, + "reward_std": 58.833797454833984, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 5.8824992179870605, + "rewards/supergames_reward/std": 58.83380126953125, + "step": 90, + "step_time": 15.007269965979503 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 269.0, + "completions/max_terminated_length": 269.0, + "completions/mean_length": 207.5, + "completions/mean_terminated_length": 207.5, + "completions/min_length": 92.0, + "completions/min_terminated_length": 92.0, + "entropy": 0.42038559913635254, + "epoch": 0.455, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.3649117946624756, + "learning_rate": 5.5e-07, + "loss": 0.1510048508644104, + "num_tokens": 782805.0, + "reward": -5.61500358581543, + "reward_std": 82.67152404785156, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -5.714998245239258, + "rewards/supergames_reward/std": 82.67152404785156, + "step": 91, + "step_time": 10.386137172987219 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 416.0, + "completions/max_terminated_length": 416.0, + "completions/mean_length": 218.375, + "completions/mean_terminated_length": 218.375, + "completions/min_length": 95.0, + "completions/min_terminated_length": 95.0, + "entropy": 0.3971734642982483, + "epoch": 0.46, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.433772563934326, + "learning_rate": 5.45e-07, + "loss": -0.3027213513851166, + "num_tokens": 793840.0, + "reward": -29.06500244140625, + "reward_std": 91.68325805664062, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -29.16499900817871, + "rewards/supergames_reward/std": 91.68325805664062, + "step": 92, + "step_time": 15.790611553995404 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 345.0, + "completions/max_terminated_length": 345.0, + "completions/mean_length": 206.625, + "completions/mean_terminated_length": 206.625, + "completions/min_length": 90.0, + "completions/min_terminated_length": 90.0, + "entropy": 0.5906907320022583, + "epoch": 0.465, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.293088912963867, + "learning_rate": 5.4e-07, + "loss": 0.09760032594203949, + "num_tokens": 803821.0, + "reward": -72.73999786376953, + "reward_std": 50.72484588623047, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": -72.81500244140625, + "rewards/supergames_reward/std": 50.7094841003418, + "step": 93, + "step_time": 13.043119941983605 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 143.0, + "completions/max_terminated_length": 143.0, + "completions/mean_length": 115.75, + "completions/mean_terminated_length": 115.75, + "completions/min_length": 80.0, + "completions/min_terminated_length": 80.0, + "entropy": 0.4233115315437317, + "epoch": 0.47, + "frac_reward_zero_std": 0.0, + "grad_norm": 5.58886194229126, + "learning_rate": 5.35e-07, + "loss": -0.04040650278329849, + "num_tokens": 809299.0, + "reward": -3.3050003051757812, + "reward_std": 24.443714141845703, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -3.405000686645508, + "rewards/supergames_reward/std": 24.443714141845703, + "step": 94, + "step_time": 5.512874762003776 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 122.0, + "completions/max_terminated_length": 122.0, + "completions/mean_length": 103.125, + "completions/mean_terminated_length": 103.125, + "completions/min_length": 86.0, + "completions/min_terminated_length": 86.0, + "entropy": 0.4055764377117157, + "epoch": 0.475, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.981549263000488, + "learning_rate": 5.3e-07, + "loss": -0.07588938623666763, + "num_tokens": 814716.0, + "reward": 60.06999969482422, + "reward_std": 42.708473205566406, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 59.970001220703125, + "rewards/supergames_reward/std": 42.708473205566406, + "step": 95, + "step_time": 4.766389057011111 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 475.0, + "completions/max_terminated_length": 475.0, + "completions/mean_length": 218.375, + "completions/mean_terminated_length": 218.375, + "completions/min_length": 81.0, + "completions/min_terminated_length": 81.0, + "entropy": 0.408913254737854, + "epoch": 0.48, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.79402756690979, + "learning_rate": 5.25e-07, + "loss": -0.3087541460990906, + "num_tokens": 824791.0, + "reward": -15.042500495910645, + "reward_std": 84.37246704101562, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -15.142499923706055, + "rewards/supergames_reward/std": 84.37246704101562, + "step": 96, + "step_time": 17.779843941010768 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 154.0, + "completions/max_terminated_length": 154.0, + "completions/mean_length": 121.625, + "completions/mean_terminated_length": 121.625, + "completions/min_length": 80.0, + "completions/min_terminated_length": 80.0, + "entropy": 0.529015839099884, + "epoch": 0.485, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.142816066741943, + "learning_rate": 5.2e-07, + "loss": 0.07216037809848785, + "num_tokens": 830300.0, + "reward": -32.743751525878906, + "reward_std": 54.61775588989258, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": -32.81875228881836, + "rewards/supergames_reward/std": 54.58256149291992, + "step": 97, + "step_time": 5.913989890017547 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 126.0, + "completions/max_terminated_length": 126.0, + "completions/mean_length": 95.625, + "completions/mean_terminated_length": 95.625, + "completions/min_length": 74.0, + "completions/min_terminated_length": 74.0, + "entropy": 0.33743274211883545, + "epoch": 0.49, + "frac_reward_zero_std": 0.0, + "grad_norm": 5.197273254394531, + "learning_rate": 5.149999999999999e-07, + "loss": 0.07500407844781876, + "num_tokens": 835633.0, + "reward": 11.00999927520752, + "reward_std": 62.73848342895508, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 10.910000801086426, + "rewards/supergames_reward/std": 62.73848342895508, + "step": 98, + "step_time": 4.9155233050114475 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 149.0, + "completions/max_terminated_length": 149.0, + "completions/mean_length": 108.5, + "completions/mean_terminated_length": 108.5, + "completions/min_length": 86.0, + "completions/min_terminated_length": 86.0, + "entropy": 0.3106803297996521, + "epoch": 0.495, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.864002227783203, + "learning_rate": 5.1e-07, + "loss": -0.1206185445189476, + "num_tokens": 841045.0, + "reward": 14.238749504089355, + "reward_std": 51.77016830444336, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 14.138750076293945, + "rewards/supergames_reward/std": 51.77016830444336, + "step": 99, + "step_time": 5.7507751359953545 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 149.0, + "completions/max_terminated_length": 149.0, + "completions/mean_length": 97.125, + "completions/mean_terminated_length": 97.125, + "completions/min_length": 73.0, + "completions/min_terminated_length": 73.0, + "entropy": 0.35160186886787415, + "epoch": 0.5, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.915731906890869, + "learning_rate": 5.049999999999999e-07, + "loss": -0.07951541244983673, + "num_tokens": 846398.0, + "reward": 67.78125, + "reward_std": 38.15436553955078, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 67.68124389648438, + "rewards/supergames_reward/std": 38.15436553955078, + "step": 100, + "step_time": 5.994720440998208 + } + ], + "logging_steps": 1, + "max_steps": 200, + "num_input_tokens_seen": 846398, + "num_train_epochs": 1, + "save_steps": 20, + "stateful_callbacks": { + "TrainerControl": { + "args": { + "should_epoch_stop": false, + "should_evaluate": false, + "should_log": false, + "should_save": true, + "should_training_stop": false + }, + "attributes": {} + } + }, + "total_flos": 0.0, + "train_batch_size": 8, + "trial_name": null, + "trial_params": null +} diff --git a/checkpoint-100/training_args.bin b/checkpoint-100/training_args.bin new file mode 100644 index 0000000..730dec0 --- /dev/null +++ b/checkpoint-100/training_args.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:691fe5be2d7d1d56160dae63866d136877a72efb30e15ff7b3249192f998a788 +size 7185 diff --git a/checkpoint-120/chat_template.jinja b/checkpoint-120/chat_template.jinja new file mode 100644 index 0000000..bdf7919 --- /dev/null +++ b/checkpoint-120/chat_template.jinja @@ -0,0 +1,54 @@ +{%- if tools %} + {{- '<|im_start|>system\n' }} + {%- if messages[0]['role'] == 'system' %} + {{- messages[0]['content'] }} + {%- else %} + {{- 'You are Qwen, created by Alibaba Cloud. You are a helpful assistant.' }} + {%- endif %} + {{- "\n\n# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }} + {%- for tool in tools %} + {{- "\n" }} + {{- tool | tojson }} + {%- endfor %} + {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }} +{%- else %} + {%- if messages[0]['role'] == 'system' %} + {{- '<|im_start|>system\n' + messages[0]['content'] + '<|im_end|>\n' }} + {%- else %} + {{- '<|im_start|>system\nYou are Qwen, created by Alibaba Cloud. You are a helpful assistant.<|im_end|>\n' }} + {%- endif %} +{%- endif %} +{%- for message in messages %} + {%- if (message.role == "user") or (message.role == "system" and not loop.first) or (message.role == "assistant" and not message.tool_calls) %} + {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }} + {%- elif message.role == "assistant" %} + {{- '<|im_start|>' + message.role }} + {%- if message.content %} + {{- '\n' + message.content }} + {%- endif %} + {%- for tool_call in message.tool_calls %} + {%- if tool_call.function is defined %} + {%- set tool_call = tool_call.function %} + {%- endif %} + {{- '\n\n{"name": "' }} + {{- tool_call.name }} + {{- '", "arguments": ' }} + {{- tool_call.arguments | tojson }} + {{- '}\n' }} + {%- endfor %} + {{- '<|im_end|>\n' }} + {%- elif message.role == "tool" %} + {%- if (loop.index0 == 0) or (messages[loop.index0 - 1].role != "tool") %} + {{- '<|im_start|>user' }} + {%- endif %} + {{- '\n\n' }} + {{- message.content }} + {{- '\n' }} + {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %} + {{- '<|im_end|>\n' }} + {%- endif %} + {%- endif %} +{%- endfor %} +{%- if add_generation_prompt %} + {{- '<|im_start|>assistant\n' }} +{%- endif %} diff --git a/checkpoint-120/config.json b/checkpoint-120/config.json new file mode 100644 index 0000000..f57b09b --- /dev/null +++ b/checkpoint-120/config.json @@ -0,0 +1,61 @@ +{ + "architectures": [ + "Qwen2ForCausalLM" + ], + "attention_dropout": 0.0, + "bos_token_id": null, + "dtype": "float32", + "eos_token_id": 151645, + "hidden_act": "silu", + "hidden_size": 1536, + "initializer_range": 0.02, + "intermediate_size": 8960, + "layer_types": [ + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention" + ], + "max_position_embeddings": 32768, + "max_window_layers": 21, + "model_type": "qwen2", + "num_attention_heads": 12, + "num_hidden_layers": 28, + "num_key_value_heads": 2, + "pad_token_id": 151643, + "rms_norm_eps": 1e-06, + "rope_parameters": { + "rope_theta": 1000000.0, + "rope_type": "default" + }, + "sliding_window": null, + "tie_word_embeddings": true, + "transformers_version": "5.6.2", + "use_cache": false, + "use_sliding_window": false, + "vocab_size": 151936 +} diff --git a/checkpoint-120/generation_config.json b/checkpoint-120/generation_config.json new file mode 100644 index 0000000..1b2bba9 --- /dev/null +++ b/checkpoint-120/generation_config.json @@ -0,0 +1,13 @@ +{ + "do_sample": true, + "eos_token_id": [ + 151645, + 151643 + ], + "pad_token_id": 151643, + "repetition_penalty": 1.1, + "temperature": 0.7, + "top_k": 20, + "top_p": 0.8, + "transformers_version": "5.6.2" +} diff --git a/checkpoint-120/model.safetensors b/checkpoint-120/model.safetensors new file mode 100644 index 0000000..6a9d42b --- /dev/null +++ b/checkpoint-120/model.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:9c4da3a283becce8e8206f10d185aee8fe0be0f1db7ddd6e86f8bb6c27c8c355 +size 6174895536 diff --git a/checkpoint-120/optimizer.pt b/checkpoint-120/optimizer.pt new file mode 100644 index 0000000..370be41 --- /dev/null +++ b/checkpoint-120/optimizer.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:2d819fb438d02d0584096eedc109a92dcee131a733136d155f10f0608cd23c99 +size 12350013801 diff --git a/checkpoint-120/rng_state.pth b/checkpoint-120/rng_state.pth new file mode 100644 index 0000000..fab1d60 --- /dev/null +++ b/checkpoint-120/rng_state.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:23f4512500fe387abc57142c0ca62eb2d6ae3b2f6fa7cfce743b4111042fa89b +size 14645 diff --git a/checkpoint-120/scheduler.pt b/checkpoint-120/scheduler.pt new file mode 100644 index 0000000..7ffc8c1 --- /dev/null +++ b/checkpoint-120/scheduler.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:440de387e2beaf1948cdb5c020c66cf05b31275efda10896ac069f091dc596e7 +size 1465 diff --git a/checkpoint-120/tokenizer.json b/checkpoint-120/tokenizer.json new file mode 100644 index 0000000..34510ff --- /dev/null +++ b/checkpoint-120/tokenizer.json @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:3fd169731d2cbde95e10bf356d66d5997fd885dd8dbb6fb4684da3f23b2585d8 +size 11421892 diff --git a/checkpoint-120/tokenizer_config.json b/checkpoint-120/tokenizer_config.json new file mode 100644 index 0000000..770e41d --- /dev/null +++ b/checkpoint-120/tokenizer_config.json @@ -0,0 +1,30 @@ +{ + "add_prefix_space": false, + "backend": "tokenizers", + "bos_token": null, + "clean_up_tokenization_spaces": false, + "eos_token": "<|im_end|>", + "errors": "replace", + "extra_special_tokens": [ + "<|im_start|>", + "<|im_end|>", + "<|object_ref_start|>", + "<|object_ref_end|>", + "<|box_start|>", + "<|box_end|>", + "<|quad_start|>", + "<|quad_end|>", + "<|vision_start|>", + "<|vision_end|>", + "<|vision_pad|>", + "<|image_pad|>", + "<|video_pad|>" + ], + "is_local": false, + "local_files_only": false, + "model_max_length": 131072, + "pad_token": "<|endoftext|>", + "split_special_tokens": false, + "tokenizer_class": "Qwen2Tokenizer", + "unk_token": null +} diff --git a/checkpoint-120/trainer_state.json b/checkpoint-120/trainer_state.json new file mode 100644 index 0000000..3960e1f --- /dev/null +++ b/checkpoint-120/trainer_state.json @@ -0,0 +1,3514 @@ +{ + "best_global_step": null, + "best_metric": null, + "best_model_checkpoint": null, + "epoch": 0.6, + "eval_steps": 500, + "global_step": 120, + "is_hyper_param_search": false, + "is_local_process_zero": true, + "is_world_process_zero": true, + "log_history": [ + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 478.0, + "completions/max_terminated_length": 478.0, + "completions/mean_length": 191.0, + "completions/mean_terminated_length": 191.0, + "completions/min_length": 66.0, + "completions/min_terminated_length": 66.0, + "entropy": 0.5566893219947815, + "epoch": 0.005, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.370361328125, + "learning_rate": 1e-06, + "loss": -0.24658073484897614, + "num_tokens": 10832.0, + "reward": -3.3000030517578125, + "reward_std": 85.62333679199219, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -3.4000015258789062, + "rewards/supergames_reward/std": 85.62333679199219, + "step": 1, + "step_time": 22.779711072013015 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 216.0, + "completions/max_terminated_length": 216.0, + "completions/mean_length": 131.5, + "completions/mean_terminated_length": 131.5, + "completions/min_length": 60.0, + "completions/min_terminated_length": 60.0, + "entropy": 0.6395841240882874, + "epoch": 0.01, + "frac_reward_zero_std": 0.0, + "grad_norm": 8.09774398803711, + "learning_rate": 9.95e-07, + "loss": 0.20567649602890015, + "num_tokens": 16404.0, + "reward": -12.422499656677246, + "reward_std": 7.134707450866699, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -12.522500038146973, + "rewards/supergames_reward/std": 7.134707927703857, + "step": 2, + "step_time": 9.960156448010821 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 287.0, + "completions/max_terminated_length": 287.0, + "completions/mean_length": 156.625, + "completions/mean_terminated_length": 156.625, + "completions/min_length": 77.0, + "completions/min_terminated_length": 77.0, + "entropy": 0.562222421169281, + "epoch": 0.015, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.9053616523742676, + "learning_rate": 9.9e-07, + "loss": 0.09602774679660797, + "num_tokens": 26953.0, + "reward": -74.94999694824219, + "reward_std": 70.73094177246094, + "rewards/json_format_reward/mean": 0.05000000074505806, + "rewards/json_format_reward/std": 0.09258200973272324, + "rewards/supergames_reward/mean": -75.0, + "rewards/supergames_reward/std": 70.71067810058594, + "step": 3, + "step_time": 13.686320498993155 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 268.0, + "completions/max_terminated_length": 268.0, + "completions/mean_length": 154.125, + "completions/mean_terminated_length": 154.125, + "completions/min_length": 53.0, + "completions/min_terminated_length": 53.0, + "entropy": 0.7323317527770996, + "epoch": 0.02, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.933310031890869, + "learning_rate": 9.849999999999999e-07, + "loss": 0.28249427676200867, + "num_tokens": 36514.0, + "reward": -40.7599983215332, + "reward_std": 81.73140716552734, + "rewards/json_format_reward/mean": 0.05000000074505806, + "rewards/json_format_reward/std": 0.09258200973272324, + "rewards/supergames_reward/mean": -40.80999755859375, + "rewards/supergames_reward/std": 81.68995666503906, + "step": 4, + "step_time": 12.805880262021674 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 401.0, + "completions/max_terminated_length": 401.0, + "completions/mean_length": 205.25, + "completions/mean_terminated_length": 205.25, + "completions/min_length": 115.0, + "completions/min_terminated_length": 115.0, + "entropy": 0.6167430281639099, + "epoch": 0.025, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.344135046005249, + "learning_rate": 9.8e-07, + "loss": 0.1433994024991989, + "num_tokens": 47476.0, + "reward": -97.6612548828125, + "reward_std": 6.3321428298950195, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -97.76124572753906, + "rewards/supergames_reward/std": 6.3321428298950195, + "step": 5, + "step_time": 18.789364666008623 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 276.0, + "completions/max_terminated_length": 276.0, + "completions/mean_length": 167.25, + "completions/mean_terminated_length": 167.25, + "completions/min_length": 73.0, + "completions/min_terminated_length": 73.0, + "entropy": 0.7583790421485901, + "epoch": 0.03, + "frac_reward_zero_std": 0.0, + "grad_norm": 5.172371864318848, + "learning_rate": 9.75e-07, + "loss": -0.12127404659986496, + "num_tokens": 55246.0, + "reward": -58.687503814697266, + "reward_std": 58.64059829711914, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": -58.76250457763672, + "rewards/supergames_reward/std": 58.620460510253906, + "step": 6, + "step_time": 12.948570273991209 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 261.0, + "completions/max_terminated_length": 261.0, + "completions/mean_length": 133.875, + "completions/mean_terminated_length": 133.875, + "completions/min_length": 57.0, + "completions/min_terminated_length": 57.0, + "entropy": 0.7447654604911804, + "epoch": 0.035, + "frac_reward_zero_std": 0.0, + "grad_norm": 6.428595066070557, + "learning_rate": 9.7e-07, + "loss": 0.10135584324598312, + "num_tokens": 60837.0, + "reward": -32.45000076293945, + "reward_std": 41.754486083984375, + "rewards/json_format_reward/mean": 0.05000000074505806, + "rewards/json_format_reward/std": 0.09258200973272324, + "rewards/supergames_reward/mean": -32.5, + "rewards/supergames_reward/std": 41.66190719604492, + "step": 7, + "step_time": 12.058315072004916 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 414.0, + "completions/max_terminated_length": 414.0, + "completions/mean_length": 209.125, + "completions/mean_terminated_length": 209.125, + "completions/min_length": 104.0, + "completions/min_terminated_length": 104.0, + "entropy": 0.3860666751861572, + "epoch": 0.04, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.806255578994751, + "learning_rate": 9.649999999999999e-07, + "loss": -0.33238139748573303, + "num_tokens": 70838.0, + "reward": -80.50375366210938, + "reward_std": 38.82380294799805, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -80.60375213623047, + "rewards/supergames_reward/std": 38.82379913330078, + "step": 8, + "step_time": 19.323370319994865 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.25, + "completions/max_length": 512.0, + "completions/max_terminated_length": 199.0, + "completions/mean_length": 247.125, + "completions/mean_terminated_length": 158.83334350585938, + "completions/min_length": 106.0, + "completions/min_terminated_length": 106.0, + "entropy": 0.5816237330436707, + "epoch": 0.045, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.3209943771362305, + "learning_rate": 9.6e-07, + "loss": 0.36752766370773315, + "num_tokens": 82135.0, + "reward": -27.688751220703125, + "reward_std": 82.8358154296875, + "rewards/json_format_reward/mean": 0.05000000074505806, + "rewards/json_format_reward/std": 0.09258200973272324, + "rewards/supergames_reward/mean": -27.738750457763672, + "rewards/supergames_reward/std": 82.78590393066406, + "step": 9, + "step_time": 21.948575104994234 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 240.0, + "completions/max_terminated_length": 240.0, + "completions/mean_length": 153.0, + "completions/mean_terminated_length": 153.0, + "completions/min_length": 51.0, + "completions/min_terminated_length": 51.0, + "entropy": 0.6301521062850952, + "epoch": 0.05, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.125359058380127, + "learning_rate": 9.55e-07, + "loss": 0.06129350885748863, + "num_tokens": 89807.0, + "reward": -78.39624786376953, + "reward_std": 40.17042922973633, + "rewards/json_format_reward/mean": 0.05000000074505806, + "rewards/json_format_reward/std": 0.09258200973272324, + "rewards/supergames_reward/mean": -78.44625091552734, + "rewards/supergames_reward/std": 40.1396484375, + "step": 10, + "step_time": 9.098202993016457 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 133.0, + "completions/max_terminated_length": 133.0, + "completions/mean_length": 112.0, + "completions/mean_terminated_length": 112.0, + "completions/min_length": 68.0, + "completions/min_terminated_length": 68.0, + "entropy": 0.5521990656852722, + "epoch": 0.055, + "frac_reward_zero_std": 0.0, + "grad_norm": 6.616060256958008, + "learning_rate": 9.499999999999999e-07, + "loss": -0.006313305348157883, + "num_tokens": 95231.0, + "reward": -12.161249160766602, + "reward_std": 6.39578104019165, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -12.261249542236328, + "rewards/supergames_reward/std": 6.39578104019165, + "step": 11, + "step_time": 5.136311663984088 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 396.0, + "completions/max_terminated_length": 396.0, + "completions/mean_length": 214.875, + "completions/mean_terminated_length": 214.875, + "completions/min_length": 55.0, + "completions/min_terminated_length": 55.0, + "entropy": 0.5505905151367188, + "epoch": 0.06, + "frac_reward_zero_std": 0.0, + "grad_norm": 5.289583206176758, + "learning_rate": 9.45e-07, + "loss": -0.29802003502845764, + "num_tokens": 106238.0, + "reward": -91.66999816894531, + "reward_std": 23.2779541015625, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -91.77000427246094, + "rewards/supergames_reward/std": 23.277956008911133, + "step": 12, + "step_time": 14.945365622988902 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 269.0, + "completions/max_terminated_length": 269.0, + "completions/mean_length": 167.375, + "completions/mean_terminated_length": 167.375, + "completions/min_length": 113.0, + "completions/min_terminated_length": 113.0, + "entropy": 0.540415346622467, + "epoch": 0.065, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.858291149139404, + "learning_rate": 9.399999999999999e-07, + "loss": 0.24004128575325012, + "num_tokens": 116937.0, + "reward": -20.38249969482422, + "reward_std": 87.55204010009766, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": -20.457500457763672, + "rewards/supergames_reward/std": 87.52605438232422, + "step": 13, + "step_time": 10.482285185018554 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 254.0, + "completions/max_terminated_length": 254.0, + "completions/mean_length": 156.875, + "completions/mean_terminated_length": 156.875, + "completions/min_length": 114.0, + "completions/min_terminated_length": 114.0, + "entropy": 0.7237679958343506, + "epoch": 0.07, + "frac_reward_zero_std": 0.0, + "grad_norm": 5.083980560302734, + "learning_rate": 9.35e-07, + "loss": -0.19025824964046478, + "num_tokens": 124640.0, + "reward": -51.397499084472656, + "reward_std": 54.486454010009766, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -51.49749755859375, + "rewards/supergames_reward/std": 54.486454010009766, + "step": 14, + "step_time": 9.597231683001155 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 236.0, + "completions/max_terminated_length": 236.0, + "completions/mean_length": 141.375, + "completions/mean_terminated_length": 141.375, + "completions/min_length": 91.0, + "completions/min_terminated_length": 91.0, + "entropy": 0.754334568977356, + "epoch": 0.075, + "frac_reward_zero_std": 0.0, + "grad_norm": 5.720428943634033, + "learning_rate": 9.3e-07, + "loss": -0.16830675303936005, + "num_tokens": 132219.0, + "reward": -56.226253509521484, + "reward_std": 60.115657806396484, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -56.32625198364258, + "rewards/supergames_reward/std": 60.11566162109375, + "step": 15, + "step_time": 8.998362872982398 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 209.0, + "completions/max_terminated_length": 209.0, + "completions/mean_length": 136.375, + "completions/mean_terminated_length": 136.375, + "completions/min_length": 97.0, + "completions/min_terminated_length": 97.0, + "entropy": 0.6280568838119507, + "epoch": 0.08, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.9681806564331055, + "learning_rate": 9.25e-07, + "loss": -0.01910916529595852, + "num_tokens": 139726.0, + "reward": -68.24000549316406, + "reward_std": 52.238914489746094, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -68.33999633789062, + "rewards/supergames_reward/std": 52.238914489746094, + "step": 16, + "step_time": 8.034480294008972 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 365.0, + "completions/max_terminated_length": 365.0, + "completions/mean_length": 211.875, + "completions/mean_terminated_length": 211.875, + "completions/min_length": 100.0, + "completions/min_terminated_length": 100.0, + "entropy": 0.899176836013794, + "epoch": 0.085, + "frac_reward_zero_std": 0.0, + "grad_norm": 6.691746711730957, + "learning_rate": 9.2e-07, + "loss": 0.25513291358947754, + "num_tokens": 147869.0, + "reward": -99.92500305175781, + "reward_std": 0.07070883363485336, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": -100.0, + "rewards/supergames_reward/std": 0.0, + "step": 17, + "step_time": 13.655792869016295 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 387.0, + "completions/max_terminated_length": 387.0, + "completions/mean_length": 238.0, + "completions/mean_terminated_length": 238.0, + "completions/min_length": 87.0, + "completions/min_terminated_length": 87.0, + "entropy": 0.5596873760223389, + "epoch": 0.09, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.6711347103118896, + "learning_rate": 9.15e-07, + "loss": -0.2628335654735565, + "num_tokens": 159101.0, + "reward": -24.900001525878906, + "reward_std": 103.50983428955078, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -25.0, + "rewards/supergames_reward/std": 103.50983428955078, + "step": 18, + "step_time": 14.784329420013819 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.125, + "completions/max_length": 512.0, + "completions/max_terminated_length": 277.0, + "completions/mean_length": 219.75, + "completions/mean_terminated_length": 178.00001525878906, + "completions/min_length": 90.0, + "completions/min_terminated_length": 90.0, + "entropy": 0.9245517253875732, + "epoch": 0.095, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.55502986907959, + "learning_rate": 9.1e-07, + "loss": 0.13060179352760315, + "num_tokens": 167299.0, + "reward": -86.89125061035156, + "reward_std": 36.95603561401367, + "rewards/json_format_reward/mean": 0.05000000074505806, + "rewards/json_format_reward/std": 0.09258200973272324, + "rewards/supergames_reward/mean": -86.94125366210938, + "rewards/supergames_reward/std": 36.93572235107422, + "step": 19, + "step_time": 18.832684700988466 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 267.0, + "completions/max_terminated_length": 267.0, + "completions/mean_length": 177.25, + "completions/mean_terminated_length": 177.25, + "completions/min_length": 93.0, + "completions/min_terminated_length": 93.0, + "entropy": 0.7331254482269287, + "epoch": 0.1, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.538497447967529, + "learning_rate": 9.05e-07, + "loss": -0.16418980062007904, + "num_tokens": 175141.0, + "reward": -68.52874755859375, + "reward_std": 58.48719787597656, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -68.62875366210938, + "rewards/supergames_reward/std": 58.48720169067383, + "step": 20, + "step_time": 10.090975169994636 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 300.0, + "completions/max_terminated_length": 300.0, + "completions/mean_length": 190.625, + "completions/mean_terminated_length": 190.625, + "completions/min_length": 103.0, + "completions/min_terminated_length": 103.0, + "entropy": 0.6659425497055054, + "epoch": 0.105, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.485373497009277, + "learning_rate": 9e-07, + "loss": -0.10106432437896729, + "num_tokens": 183098.0, + "reward": -71.48875427246094, + "reward_std": 55.33565139770508, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -71.5887451171875, + "rewards/supergames_reward/std": 55.33565139770508, + "step": 21, + "step_time": 11.255001295008697 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 279.0, + "completions/max_terminated_length": 279.0, + "completions/mean_length": 179.0, + "completions/mean_terminated_length": 179.0, + "completions/min_length": 103.0, + "completions/min_terminated_length": 103.0, + "entropy": 0.49365949630737305, + "epoch": 0.11, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.829419136047363, + "learning_rate": 8.95e-07, + "loss": 0.10587231814861298, + "num_tokens": 192826.0, + "reward": -75.62000274658203, + "reward_std": 45.01959991455078, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": -75.69499969482422, + "rewards/supergames_reward/std": 45.00411605834961, + "step": 22, + "step_time": 10.611246599000879 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 315.0, + "completions/max_terminated_length": 315.0, + "completions/mean_length": 212.625, + "completions/mean_terminated_length": 212.625, + "completions/min_length": 100.0, + "completions/min_terminated_length": 100.0, + "entropy": 0.43159469962120056, + "epoch": 0.115, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.541125774383545, + "learning_rate": 8.9e-07, + "loss": -0.1681259125471115, + "num_tokens": 202815.0, + "reward": 23.67624855041504, + "reward_std": 67.74340057373047, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 23.576250076293945, + "rewards/supergames_reward/std": 67.743408203125, + "step": 23, + "step_time": 11.731771531980485 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 168.0, + "completions/max_terminated_length": 168.0, + "completions/mean_length": 101.625, + "completions/mean_terminated_length": 101.625, + "completions/min_length": 65.0, + "completions/min_terminated_length": 65.0, + "entropy": 0.3972298502922058, + "epoch": 0.12, + "frac_reward_zero_std": 0.0, + "grad_norm": 5.8285417556762695, + "learning_rate": 8.85e-07, + "loss": -0.011585958302021027, + "num_tokens": 208188.0, + "reward": 58.849998474121094, + "reward_std": 56.9304084777832, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 58.75, + "rewards/supergames_reward/std": 56.9304084777832, + "step": 24, + "step_time": 6.187504950998118 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 303.0, + "completions/max_terminated_length": 303.0, + "completions/mean_length": 165.375, + "completions/mean_terminated_length": 165.375, + "completions/min_length": 69.0, + "completions/min_terminated_length": 69.0, + "entropy": 0.6339899301528931, + "epoch": 0.125, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.823799133300781, + "learning_rate": 8.799999999999999e-07, + "loss": 0.05159185826778412, + "num_tokens": 215935.0, + "reward": -23.521251678466797, + "reward_std": 74.23067474365234, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -23.621248245239258, + "rewards/supergames_reward/std": 74.23067474365234, + "step": 25, + "step_time": 11.059416300005978 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 349.0, + "completions/max_terminated_length": 349.0, + "completions/mean_length": 204.0, + "completions/mean_terminated_length": 204.0, + "completions/min_length": 143.0, + "completions/min_terminated_length": 143.0, + "entropy": 0.6462895274162292, + "epoch": 0.13, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.969292640686035, + "learning_rate": 8.75e-07, + "loss": 0.0115677984431386, + "num_tokens": 223999.0, + "reward": -74.8175048828125, + "reward_std": 40.94221115112305, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": -74.89250183105469, + "rewards/supergames_reward/std": 40.924625396728516, + "step": 26, + "step_time": 12.666237785975682 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 299.0, + "completions/max_terminated_length": 299.0, + "completions/mean_length": 166.5, + "completions/mean_terminated_length": 166.5, + "completions/min_length": 87.0, + "completions/min_terminated_length": 87.0, + "entropy": 0.4756850600242615, + "epoch": 0.135, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.915069103240967, + "learning_rate": 8.699999999999999e-07, + "loss": -0.13321346044540405, + "num_tokens": 233643.0, + "reward": -52.125, + "reward_std": 19.405282974243164, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": -52.20000076293945, + "rewards/supergames_reward/std": 19.33464813232422, + "step": 27, + "step_time": 11.172539752995363 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 184.0, + "completions/max_terminated_length": 184.0, + "completions/mean_length": 137.125, + "completions/mean_terminated_length": 137.125, + "completions/min_length": 100.0, + "completions/min_terminated_length": 100.0, + "entropy": 0.5834778547286987, + "epoch": 0.14, + "frac_reward_zero_std": 0.0, + "grad_norm": 5.146666526794434, + "learning_rate": 8.65e-07, + "loss": 0.060454584658145905, + "num_tokens": 241204.0, + "reward": -68.99250793457031, + "reward_std": 51.010581970214844, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -69.09249877929688, + "rewards/supergames_reward/std": 51.010581970214844, + "step": 28, + "step_time": 7.025046669004951 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 331.0, + "completions/max_terminated_length": 331.0, + "completions/mean_length": 182.375, + "completions/mean_terminated_length": 182.375, + "completions/min_length": 106.0, + "completions/min_terminated_length": 106.0, + "entropy": 0.6368551254272461, + "epoch": 0.145, + "frac_reward_zero_std": 0.0, + "grad_norm": 5.15535306930542, + "learning_rate": 8.599999999999999e-07, + "loss": -0.3323673903942108, + "num_tokens": 249087.0, + "reward": -44.17000198364258, + "reward_std": 80.5373306274414, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -44.27000045776367, + "rewards/supergames_reward/std": 80.5373306274414, + "step": 29, + "step_time": 12.079259724996518 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 190.0, + "completions/max_terminated_length": 190.0, + "completions/mean_length": 127.75, + "completions/mean_terminated_length": 127.75, + "completions/min_length": 80.0, + "completions/min_terminated_length": 80.0, + "entropy": 0.4491196870803833, + "epoch": 0.15, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.896186828613281, + "learning_rate": 8.55e-07, + "loss": -0.12887656688690186, + "num_tokens": 259445.0, + "reward": 57.446250915527344, + "reward_std": 38.78837585449219, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 57.34625244140625, + "rewards/supergames_reward/std": 38.78837966918945, + "step": 30, + "step_time": 7.483972909016302 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 166.0, + "completions/max_terminated_length": 166.0, + "completions/mean_length": 112.875, + "completions/mean_terminated_length": 112.875, + "completions/min_length": 85.0, + "completions/min_terminated_length": 85.0, + "entropy": 0.4858454465866089, + "epoch": 0.155, + "frac_reward_zero_std": 0.0, + "grad_norm": 6.742906093597412, + "learning_rate": 8.499999999999999e-07, + "loss": 0.06930096447467804, + "num_tokens": 264924.0, + "reward": -21.174999237060547, + "reward_std": 31.89051628112793, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": -21.25, + "rewards/supergames_reward/std": 31.819807052612305, + "step": 31, + "step_time": 6.1866529019898735 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 148.0, + "completions/max_terminated_length": 148.0, + "completions/mean_length": 118.125, + "completions/mean_terminated_length": 118.125, + "completions/min_length": 98.0, + "completions/min_terminated_length": 98.0, + "entropy": 0.45428749918937683, + "epoch": 0.16, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.663414478302002, + "learning_rate": 8.45e-07, + "loss": 0.05735419690608978, + "num_tokens": 270381.0, + "reward": 0.4687504768371582, + "reward_std": 38.66371154785156, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 0.36875057220458984, + "rewards/supergames_reward/std": 38.6637077331543, + "step": 32, + "step_time": 5.577042757999152 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 345.0, + "completions/max_terminated_length": 345.0, + "completions/mean_length": 166.875, + "completions/mean_terminated_length": 166.875, + "completions/min_length": 86.0, + "completions/min_terminated_length": 86.0, + "entropy": 0.6107293367385864, + "epoch": 0.165, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.75006628036499, + "learning_rate": 8.399999999999999e-07, + "loss": -0.14760127663612366, + "num_tokens": 278156.0, + "reward": -87.84750366210938, + "reward_std": 26.84708023071289, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -87.94749450683594, + "rewards/supergames_reward/std": 26.847076416015625, + "step": 33, + "step_time": 12.552876825997373 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.125, + "completions/max_length": 512.0, + "completions/max_terminated_length": 360.0, + "completions/mean_length": 231.0, + "completions/mean_terminated_length": 190.85714721679688, + "completions/min_length": 74.0, + "completions/min_terminated_length": 74.0, + "entropy": 0.6089653372764587, + "epoch": 0.17, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.781267166137695, + "learning_rate": 8.349999999999999e-07, + "loss": 0.4294504225254059, + "num_tokens": 289364.0, + "reward": -99.92500305175781, + "reward_std": 0.07070883363485336, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": -100.0, + "rewards/supergames_reward/std": 0.0, + "step": 34, + "step_time": 18.961819477990502 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 218.0, + "completions/max_terminated_length": 218.0, + "completions/mean_length": 134.125, + "completions/mean_terminated_length": 134.125, + "completions/min_length": 104.0, + "completions/min_terminated_length": 104.0, + "entropy": 0.46158918738365173, + "epoch": 0.175, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.016353607177734, + "learning_rate": 8.299999999999999e-07, + "loss": 0.09251818805932999, + "num_tokens": 298733.0, + "reward": -35.558753967285156, + "reward_std": 15.881204605102539, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -35.65875244140625, + "rewards/supergames_reward/std": 15.881203651428223, + "step": 35, + "step_time": 8.38491243700264 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 207.0, + "completions/max_terminated_length": 207.0, + "completions/mean_length": 156.875, + "completions/mean_terminated_length": 156.875, + "completions/min_length": 111.0, + "completions/min_terminated_length": 111.0, + "entropy": 0.5104788541793823, + "epoch": 0.18, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.592893123626709, + "learning_rate": 8.249999999999999e-07, + "loss": -0.0410832017660141, + "num_tokens": 306420.0, + "reward": 5.4212493896484375, + "reward_std": 93.50404357910156, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": 5.346250534057617, + "rewards/supergames_reward/std": 93.4718246459961, + "step": 36, + "step_time": 7.811868985998444 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 241.0, + "completions/max_terminated_length": 241.0, + "completions/mean_length": 178.0, + "completions/mean_terminated_length": 178.0, + "completions/min_length": 106.0, + "completions/min_terminated_length": 106.0, + "entropy": 0.7527878880500793, + "epoch": 0.185, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.8812546730041504, + "learning_rate": 8.199999999999999e-07, + "loss": 0.12539231777191162, + "num_tokens": 314252.0, + "reward": -89.58500671386719, + "reward_std": 29.256120681762695, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": -89.66000366210938, + "rewards/supergames_reward/std": 29.24593734741211, + "step": 37, + "step_time": 8.959742914012168 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 199.0, + "completions/max_terminated_length": 199.0, + "completions/mean_length": 133.375, + "completions/mean_terminated_length": 133.375, + "completions/min_length": 87.0, + "completions/min_terminated_length": 87.0, + "entropy": 0.3347901403903961, + "epoch": 0.19, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.205414295196533, + "learning_rate": 8.149999999999999e-07, + "loss": -0.06991486996412277, + "num_tokens": 319839.0, + "reward": -14.498749732971191, + "reward_std": 13.007229804992676, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -14.598750114440918, + "rewards/supergames_reward/std": 13.007229804992676, + "step": 38, + "step_time": 7.377183554985095 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 181.0, + "completions/max_terminated_length": 181.0, + "completions/mean_length": 114.375, + "completions/mean_terminated_length": 114.375, + "completions/min_length": 58.0, + "completions/min_terminated_length": 58.0, + "entropy": 0.4445965886116028, + "epoch": 0.195, + "frac_reward_zero_std": 0.0, + "grad_norm": 5.561298847198486, + "learning_rate": 8.1e-07, + "loss": -0.06235164776444435, + "num_tokens": 325290.0, + "reward": -7.425000190734863, + "reward_std": 53.701602935791016, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": -7.5, + "rewards/supergames_reward/std": 53.652320861816406, + "step": 39, + "step_time": 6.753862089011818 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 370.0, + "completions/max_terminated_length": 370.0, + "completions/mean_length": 187.625, + "completions/mean_terminated_length": 187.625, + "completions/min_length": 60.0, + "completions/min_terminated_length": 60.0, + "entropy": 0.3950234353542328, + "epoch": 0.2, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.5988712310791016, + "learning_rate": 8.05e-07, + "loss": -0.1381440907716751, + "num_tokens": 335119.0, + "reward": -33.68375015258789, + "reward_std": 42.069435119628906, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -33.783748626708984, + "rewards/supergames_reward/std": 42.069435119628906, + "step": 40, + "step_time": 13.718958162004128 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.125, + "completions/max_length": 512.0, + "completions/max_terminated_length": 371.0, + "completions/mean_length": 251.125, + "completions/mean_terminated_length": 213.85714721679688, + "completions/min_length": 121.0, + "completions/min_terminated_length": 121.0, + "entropy": 0.507067084312439, + "epoch": 0.205, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.3898658752441406, + "learning_rate": 8e-07, + "loss": 0.0078964838758111, + "num_tokens": 346464.0, + "reward": -86.48625183105469, + "reward_std": 38.020694732666016, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": -86.56124877929688, + "rewards/supergames_reward/std": 38.010528564453125, + "step": 41, + "step_time": 18.90960379401804 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 152.0, + "completions/max_terminated_length": 152.0, + "completions/mean_length": 116.375, + "completions/mean_terminated_length": 116.375, + "completions/min_length": 93.0, + "completions/min_terminated_length": 93.0, + "entropy": 0.3801707923412323, + "epoch": 0.21, + "frac_reward_zero_std": 1.0, + "grad_norm": 0.0, + "learning_rate": 7.95e-07, + "loss": 0.0, + "num_tokens": 351955.0, + "reward": -9.899999618530273, + "reward_std": 0.0, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -10.0, + "rewards/supergames_reward/std": 0.0, + "step": 42, + "step_time": 5.711630532023264 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 308.0, + "completions/max_terminated_length": 308.0, + "completions/mean_length": 247.375, + "completions/mean_terminated_length": 247.375, + "completions/min_length": 116.0, + "completions/min_terminated_length": 116.0, + "entropy": 0.3185043931007385, + "epoch": 0.215, + "frac_reward_zero_std": 0.0, + "grad_norm": 2.168762445449829, + "learning_rate": 7.9e-07, + "loss": 0.1970764398574829, + "num_tokens": 363278.0, + "reward": -53.89249801635742, + "reward_std": 63.02619934082031, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -53.99250030517578, + "rewards/supergames_reward/std": 63.02619934082031, + "step": 43, + "step_time": 11.661934906995157 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 330.0, + "completions/max_terminated_length": 330.0, + "completions/mean_length": 212.5, + "completions/mean_terminated_length": 212.5, + "completions/min_length": 106.0, + "completions/min_terminated_length": 106.0, + "entropy": 0.4446268379688263, + "epoch": 0.22, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.305844306945801, + "learning_rate": 7.85e-07, + "loss": -0.06735197454690933, + "num_tokens": 374306.0, + "reward": 18.059999465942383, + "reward_std": 98.9417953491211, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 17.959999084472656, + "rewards/supergames_reward/std": 98.9417953491211, + "step": 44, + "step_time": 12.492453911982011 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 316.0, + "completions/max_terminated_length": 316.0, + "completions/mean_length": 183.0, + "completions/mean_terminated_length": 183.0, + "completions/min_length": 103.0, + "completions/min_terminated_length": 103.0, + "entropy": 0.5269614458084106, + "epoch": 0.225, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.8006370067596436, + "learning_rate": 7.799999999999999e-07, + "loss": -0.08870815485715866, + "num_tokens": 382210.0, + "reward": -54.928749084472656, + "reward_std": 72.5127944946289, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -55.02874755859375, + "rewards/supergames_reward/std": 72.5127944946289, + "step": 45, + "step_time": 11.7636467939883 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 244.0, + "completions/max_terminated_length": 244.0, + "completions/mean_length": 136.625, + "completions/mean_terminated_length": 136.625, + "completions/min_length": 85.0, + "completions/min_terminated_length": 85.0, + "entropy": 0.34820204973220825, + "epoch": 0.23, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.911736011505127, + "learning_rate": 7.75e-07, + "loss": 0.04397330805659294, + "num_tokens": 391583.0, + "reward": -19.10375213623047, + "reward_std": 23.967702865600586, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -19.203750610351562, + "rewards/supergames_reward/std": 23.967702865600586, + "step": 46, + "step_time": 10.119426151999505 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 218.0, + "completions/max_terminated_length": 218.0, + "completions/mean_length": 160.375, + "completions/mean_terminated_length": 160.375, + "completions/min_length": 93.0, + "completions/min_terminated_length": 93.0, + "entropy": 0.611914336681366, + "epoch": 0.235, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.571913242340088, + "learning_rate": 7.699999999999999e-07, + "loss": -0.044293053448200226, + "num_tokens": 399290.0, + "reward": -25.186250686645508, + "reward_std": 71.46305084228516, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -25.286251068115234, + "rewards/supergames_reward/std": 71.46305084228516, + "step": 47, + "step_time": 8.563868903991533 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 167.0, + "completions/max_terminated_length": 167.0, + "completions/mean_length": 111.125, + "completions/mean_terminated_length": 111.125, + "completions/min_length": 56.0, + "completions/min_terminated_length": 56.0, + "entropy": 0.46585777401924133, + "epoch": 0.24, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.654109954833984, + "learning_rate": 7.65e-07, + "loss": -0.035151124000549316, + "num_tokens": 404707.0, + "reward": -0.408750057220459, + "reward_std": 31.939748764038086, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -0.5087499618530273, + "rewards/supergames_reward/std": 31.93975257873535, + "step": 48, + "step_time": 6.438482580007985 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 306.0, + "completions/max_terminated_length": 306.0, + "completions/mean_length": 216.625, + "completions/mean_terminated_length": 216.625, + "completions/min_length": 132.0, + "completions/min_terminated_length": 132.0, + "entropy": 0.33054471015930176, + "epoch": 0.245, + "frac_reward_zero_std": 0.0, + "grad_norm": 2.700303792953491, + "learning_rate": 7.599999999999999e-07, + "loss": -0.04393656551837921, + "num_tokens": 414736.0, + "reward": -97.26000213623047, + "reward_std": 1.6294406652450562, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -97.36000061035156, + "rewards/supergames_reward/std": 1.629441738128662, + "step": 49, + "step_time": 11.636040106008295 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 378.0, + "completions/max_terminated_length": 378.0, + "completions/mean_length": 218.375, + "completions/mean_terminated_length": 218.375, + "completions/min_length": 104.0, + "completions/min_terminated_length": 104.0, + "entropy": 0.48711177706718445, + "epoch": 0.25, + "frac_reward_zero_std": 0.0, + "grad_norm": 2.9650564193725586, + "learning_rate": 7.55e-07, + "loss": -0.3529520332813263, + "num_tokens": 424795.0, + "reward": 47.408748626708984, + "reward_std": 72.72083282470703, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 47.30875015258789, + "rewards/supergames_reward/std": 72.72083282470703, + "step": 50, + "step_time": 14.258096842997475 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.125, + "completions/max_length": 512.0, + "completions/max_terminated_length": 492.0, + "completions/mean_length": 267.625, + "completions/mean_terminated_length": 232.71429443359375, + "completions/min_length": 90.0, + "completions/min_terminated_length": 90.0, + "entropy": 0.4482860863208771, + "epoch": 0.255, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.2998197078704834, + "learning_rate": 7.5e-07, + "loss": 0.2493654191493988, + "num_tokens": 436288.0, + "reward": 50.61125183105469, + "reward_std": 75.3349838256836, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": 50.53624725341797, + "rewards/supergames_reward/std": 75.27783966064453, + "step": 51, + "step_time": 19.20827590499539 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 176.0, + "completions/max_terminated_length": 176.0, + "completions/mean_length": 144.5, + "completions/mean_terminated_length": 144.5, + "completions/min_length": 99.0, + "completions/min_terminated_length": 99.0, + "entropy": 0.5378735661506653, + "epoch": 0.26, + "frac_reward_zero_std": 0.0, + "grad_norm": 5.114461898803711, + "learning_rate": 7.45e-07, + "loss": 0.0036597400903701782, + "num_tokens": 441980.0, + "reward": -15.20625114440918, + "reward_std": 45.35526657104492, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": -15.28125, + "rewards/supergames_reward/std": 45.301815032958984, + "step": 52, + "step_time": 6.660627231001854 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 136.0, + "completions/max_terminated_length": 136.0, + "completions/mean_length": 119.0, + "completions/mean_terminated_length": 119.0, + "completions/min_length": 105.0, + "completions/min_terminated_length": 105.0, + "entropy": 0.41796669363975525, + "epoch": 0.265, + "frac_reward_zero_std": 0.0, + "grad_norm": 5.059505462646484, + "learning_rate": 7.4e-07, + "loss": -0.014973883517086506, + "num_tokens": 447468.0, + "reward": 15.170000076293945, + "reward_std": 46.658329010009766, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 15.070000648498535, + "rewards/supergames_reward/std": 46.658329010009766, + "step": 53, + "step_time": 5.271571868011961 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 259.0, + "completions/max_terminated_length": 259.0, + "completions/mean_length": 191.125, + "completions/mean_terminated_length": 191.125, + "completions/min_length": 91.0, + "completions/min_terminated_length": 91.0, + "entropy": 0.45421987771987915, + "epoch": 0.27, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.4454867839813232, + "learning_rate": 7.35e-07, + "loss": -0.1128973588347435, + "num_tokens": 457301.0, + "reward": 38.89875030517578, + "reward_std": 62.1772575378418, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 38.79874801635742, + "rewards/supergames_reward/std": 62.1772575378418, + "step": 54, + "step_time": 10.524528659996577 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 319.0, + "completions/max_terminated_length": 319.0, + "completions/mean_length": 233.375, + "completions/mean_terminated_length": 233.375, + "completions/min_length": 136.0, + "completions/min_terminated_length": 136.0, + "entropy": 0.4449213445186615, + "epoch": 0.275, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.2065398693084717, + "learning_rate": 7.3e-07, + "loss": 0.0016277075046673417, + "num_tokens": 467520.0, + "reward": 27.372499465942383, + "reward_std": 81.23126983642578, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 27.27250099182129, + "rewards/supergames_reward/std": 81.23126983642578, + "step": 55, + "step_time": 12.592280682991259 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.125, + "completions/max_length": 512.0, + "completions/max_terminated_length": 357.0, + "completions/mean_length": 261.0, + "completions/mean_terminated_length": 225.1428680419922, + "completions/min_length": 95.0, + "completions/min_terminated_length": 95.0, + "entropy": 0.506920337677002, + "epoch": 0.28, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.058388710021973, + "learning_rate": 7.249999999999999e-07, + "loss": 0.3289242386817932, + "num_tokens": 476056.0, + "reward": 47.05249786376953, + "reward_std": 80.97753143310547, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": 46.977500915527344, + "rewards/supergames_reward/std": 80.9256362915039, + "step": 56, + "step_time": 18.845177220006008 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 250.0, + "completions/max_terminated_length": 250.0, + "completions/mean_length": 192.75, + "completions/mean_terminated_length": 192.75, + "completions/min_length": 122.0, + "completions/min_terminated_length": 122.0, + "entropy": 0.5197336077690125, + "epoch": 0.285, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.230815410614014, + "learning_rate": 7.2e-07, + "loss": -0.15186546742916107, + "num_tokens": 484014.0, + "reward": -64.5050048828125, + "reward_std": 50.35145568847656, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -64.60499572753906, + "rewards/supergames_reward/std": 50.3514518737793, + "step": 57, + "step_time": 9.576232638006331 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 324.0, + "completions/max_terminated_length": 324.0, + "completions/mean_length": 249.0, + "completions/mean_terminated_length": 249.0, + "completions/min_length": 129.0, + "completions/min_terminated_length": 129.0, + "entropy": 0.42913514375686646, + "epoch": 0.29, + "frac_reward_zero_std": 0.0, + "grad_norm": 2.8544528484344482, + "learning_rate": 7.149999999999999e-07, + "loss": -0.056405920535326004, + "num_tokens": 494390.0, + "reward": 40.123748779296875, + "reward_std": 84.164794921875, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 40.02375030517578, + "rewards/supergames_reward/std": 84.164794921875, + "step": 58, + "step_time": 12.385353341000155 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 138.0, + "completions/max_terminated_length": 138.0, + "completions/mean_length": 107.25, + "completions/mean_terminated_length": 107.25, + "completions/min_length": 81.0, + "completions/min_terminated_length": 81.0, + "entropy": 0.34386488795280457, + "epoch": 0.295, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.6915154457092285, + "learning_rate": 7.1e-07, + "loss": 0.0865321159362793, + "num_tokens": 499816.0, + "reward": -8.244999885559082, + "reward_std": 4.681046962738037, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -8.345000267028809, + "rewards/supergames_reward/std": 4.681046962738037, + "step": 59, + "step_time": 5.42325339600211 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 187.0, + "completions/max_terminated_length": 187.0, + "completions/mean_length": 127.25, + "completions/mean_terminated_length": 127.25, + "completions/min_length": 97.0, + "completions/min_terminated_length": 97.0, + "entropy": 0.5430045127868652, + "epoch": 0.3, + "frac_reward_zero_std": 0.0, + "grad_norm": 5.635311126708984, + "learning_rate": 7.049999999999999e-07, + "loss": -0.05412375554442406, + "num_tokens": 505410.0, + "reward": 20.005001068115234, + "reward_std": 55.156005859375, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 19.904998779296875, + "rewards/supergames_reward/std": 55.156005859375, + "step": 60, + "step_time": 7.068010902003152 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 286.0, + "completions/max_terminated_length": 286.0, + "completions/mean_length": 192.125, + "completions/mean_terminated_length": 192.125, + "completions/min_length": 95.0, + "completions/min_terminated_length": 95.0, + "entropy": 0.4213869869709015, + "epoch": 0.305, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.4664618968963623, + "learning_rate": 7e-07, + "loss": -0.0495159812271595, + "num_tokens": 516283.0, + "reward": -47.212501525878906, + "reward_std": 91.2253646850586, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -47.3125, + "rewards/supergames_reward/std": 91.2253646850586, + "step": 61, + "step_time": 11.176304235996213 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 324.0, + "completions/max_terminated_length": 324.0, + "completions/mean_length": 171.0, + "completions/mean_terminated_length": 171.0, + "completions/min_length": 92.0, + "completions/min_terminated_length": 92.0, + "entropy": 0.4626811146736145, + "epoch": 0.31, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.949252128601074, + "learning_rate": 6.949999999999999e-07, + "loss": 0.192047581076622, + "num_tokens": 522251.0, + "reward": -15.092500686645508, + "reward_std": 49.82962417602539, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": -15.167499542236328, + "rewards/supergames_reward/std": 49.78091049194336, + "step": 62, + "step_time": 12.075224861997413 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.125, + "completions/max_length": 512.0, + "completions/max_terminated_length": 429.0, + "completions/mean_length": 283.5, + "completions/mean_terminated_length": 250.85714721679688, + "completions/min_length": 85.0, + "completions/min_terminated_length": 85.0, + "entropy": 0.4814216196537018, + "epoch": 0.315, + "frac_reward_zero_std": 0.0, + "grad_norm": 2.7918753623962402, + "learning_rate": 6.9e-07, + "loss": 0.15706156194210052, + "num_tokens": 533815.0, + "reward": -20.86625099182129, + "reward_std": 85.99250030517578, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": -20.94124984741211, + "rewards/supergames_reward/std": 85.96620178222656, + "step": 63, + "step_time": 19.448832260008203 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 189.0, + "completions/max_terminated_length": 189.0, + "completions/mean_length": 131.25, + "completions/mean_terminated_length": 131.25, + "completions/min_length": 106.0, + "completions/min_terminated_length": 106.0, + "entropy": 0.4929414987564087, + "epoch": 0.32, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.774101734161377, + "learning_rate": 6.85e-07, + "loss": -0.060691747814416885, + "num_tokens": 539465.0, + "reward": -21.0674991607666, + "reward_std": 6.892730236053467, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -21.167499542236328, + "rewards/supergames_reward/std": 6.892730236053467, + "step": 64, + "step_time": 7.141569407976931 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 454.0, + "completions/max_terminated_length": 454.0, + "completions/mean_length": 281.375, + "completions/mean_terminated_length": 281.375, + "completions/min_length": 107.0, + "completions/min_terminated_length": 107.0, + "entropy": 0.45866572856903076, + "epoch": 0.325, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.092717409133911, + "learning_rate": 6.800000000000001e-07, + "loss": 0.11694959551095963, + "num_tokens": 551012.0, + "reward": 49.938751220703125, + "reward_std": 79.1865005493164, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 49.8387451171875, + "rewards/supergames_reward/std": 79.1865005493164, + "step": 65, + "step_time": 17.06170882002334 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 203.0, + "completions/max_terminated_length": 203.0, + "completions/mean_length": 128.625, + "completions/mean_terminated_length": 128.625, + "completions/min_length": 88.0, + "completions/min_terminated_length": 88.0, + "entropy": 0.5005782246589661, + "epoch": 0.33, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.77616024017334, + "learning_rate": 6.75e-07, + "loss": 0.11242837458848953, + "num_tokens": 556577.0, + "reward": 48.875, + "reward_std": 49.13539123535156, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 48.775001525878906, + "rewards/supergames_reward/std": 49.13539123535156, + "step": 66, + "step_time": 7.695410129002994 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 436.0, + "completions/max_terminated_length": 436.0, + "completions/mean_length": 281.875, + "completions/mean_terminated_length": 281.875, + "completions/min_length": 133.0, + "completions/min_terminated_length": 133.0, + "entropy": 0.507097065448761, + "epoch": 0.335, + "frac_reward_zero_std": 0.0, + "grad_norm": 2.911184787750244, + "learning_rate": 6.7e-07, + "loss": -0.08151137083768845, + "num_tokens": 567120.0, + "reward": -12.886249542236328, + "reward_std": 84.04269409179688, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -12.986251831054688, + "rewards/supergames_reward/std": 84.04269409179688, + "step": 67, + "step_time": 16.409127942984924 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.125, + "completions/max_length": 512.0, + "completions/max_terminated_length": 501.0, + "completions/mean_length": 318.375, + "completions/mean_terminated_length": 290.71429443359375, + "completions/min_length": 65.0, + "completions/min_terminated_length": 65.0, + "entropy": 0.486402690410614, + "epoch": 0.34, + "frac_reward_zero_std": 0.0, + "grad_norm": 2.9461607933044434, + "learning_rate": 6.65e-07, + "loss": 0.2442178875207901, + "num_tokens": 578995.0, + "reward": 2.8387489318847656, + "reward_std": 98.47579956054688, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": 2.7637500762939453, + "rewards/supergames_reward/std": 98.44595336914062, + "step": 68, + "step_time": 19.32364999302081 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 291.0, + "completions/max_terminated_length": 291.0, + "completions/mean_length": 219.625, + "completions/mean_terminated_length": 219.625, + "completions/min_length": 117.0, + "completions/min_terminated_length": 117.0, + "entropy": 0.37856727838516235, + "epoch": 0.345, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.0204732418060303, + "learning_rate": 6.6e-07, + "loss": -0.08764916658401489, + "num_tokens": 589072.0, + "reward": -54.18375015258789, + "reward_std": 71.45012664794922, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -54.283748626708984, + "rewards/supergames_reward/std": 71.45013427734375, + "step": 69, + "step_time": 11.126611230982235 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.125, + "completions/max_length": 512.0, + "completions/max_terminated_length": 305.0, + "completions/mean_length": 254.75, + "completions/mean_terminated_length": 218.00001525878906, + "completions/min_length": 104.0, + "completions/min_terminated_length": 104.0, + "entropy": 0.46423614025115967, + "epoch": 0.35, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.0140154361724854, + "learning_rate": 6.55e-07, + "loss": 0.26797160506248474, + "num_tokens": 600438.0, + "reward": -32.04875183105469, + "reward_std": 69.86957550048828, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": -32.12375259399414, + "rewards/supergames_reward/std": 69.84178161621094, + "step": 70, + "step_time": 19.12617294798838 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 372.0, + "completions/max_terminated_length": 372.0, + "completions/mean_length": 194.125, + "completions/mean_terminated_length": 194.125, + "completions/min_length": 94.0, + "completions/min_terminated_length": 94.0, + "entropy": 0.3427293300628662, + "epoch": 0.355, + "frac_reward_zero_std": 0.0, + "grad_norm": 2.781383991241455, + "learning_rate": 6.5e-07, + "loss": 0.354524701833725, + "num_tokens": 610327.0, + "reward": -6.913749694824219, + "reward_std": 87.56071472167969, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": -6.988750457763672, + "rewards/supergames_reward/std": 87.53033447265625, + "step": 71, + "step_time": 14.041668199002743 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 197.0, + "completions/max_terminated_length": 197.0, + "completions/mean_length": 124.75, + "completions/mean_terminated_length": 124.75, + "completions/min_length": 93.0, + "completions/min_terminated_length": 93.0, + "entropy": 0.4455118477344513, + "epoch": 0.36, + "frac_reward_zero_std": 0.0, + "grad_norm": 5.726032733917236, + "learning_rate": 6.45e-07, + "loss": 0.18206289410591125, + "num_tokens": 615853.0, + "reward": 41.95124816894531, + "reward_std": 60.640811920166016, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 41.85124969482422, + "rewards/supergames_reward/std": 60.640811920166016, + "step": 72, + "step_time": 7.416493425989756 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 323.0, + "completions/max_terminated_length": 323.0, + "completions/mean_length": 147.375, + "completions/mean_terminated_length": 147.375, + "completions/min_length": 94.0, + "completions/min_terminated_length": 94.0, + "entropy": 0.480733186006546, + "epoch": 0.365, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.853621482849121, + "learning_rate": 6.4e-07, + "loss": 0.04558124020695686, + "num_tokens": 621584.0, + "reward": 14.884998321533203, + "reward_std": 46.25857162475586, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 14.78499984741211, + "rewards/supergames_reward/std": 46.25857162475586, + "step": 73, + "step_time": 11.95379540900467 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 283.0, + "completions/max_terminated_length": 283.0, + "completions/mean_length": 188.125, + "completions/mean_terminated_length": 188.125, + "completions/min_length": 116.0, + "completions/min_terminated_length": 116.0, + "entropy": 0.5526013374328613, + "epoch": 0.37, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.729883909225464, + "learning_rate": 6.35e-07, + "loss": 0.06268303096294403, + "num_tokens": 629545.0, + "reward": -68.05750274658203, + "reward_std": 59.44808578491211, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -68.15750122070312, + "rewards/supergames_reward/std": 59.44808578491211, + "step": 74, + "step_time": 10.64124580900534 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 141.0, + "completions/max_terminated_length": 141.0, + "completions/mean_length": 117.375, + "completions/mean_terminated_length": 117.375, + "completions/min_length": 100.0, + "completions/min_terminated_length": 100.0, + "entropy": 0.3744319677352905, + "epoch": 0.375, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.775154113769531, + "learning_rate": 6.3e-07, + "loss": -0.023284845054149628, + "num_tokens": 635052.0, + "reward": 3.9374990463256836, + "reward_std": 61.85771179199219, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": 3.862499237060547, + "rewards/supergames_reward/std": 61.80967712402344, + "step": 75, + "step_time": 5.432648951013107 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 268.0, + "completions/max_terminated_length": 268.0, + "completions/mean_length": 157.0, + "completions/mean_terminated_length": 157.0, + "completions/min_length": 104.0, + "completions/min_terminated_length": 104.0, + "entropy": 0.4822140336036682, + "epoch": 0.38, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.715704917907715, + "learning_rate": 6.249999999999999e-07, + "loss": -0.04918142035603523, + "num_tokens": 644644.0, + "reward": 48.64125061035156, + "reward_std": 78.0205078125, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": 48.56624984741211, + "rewards/supergames_reward/std": 77.96605682373047, + "step": 76, + "step_time": 10.348935816989979 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 380.0, + "completions/max_terminated_length": 380.0, + "completions/mean_length": 210.875, + "completions/mean_terminated_length": 210.875, + "completions/min_length": 126.0, + "completions/min_terminated_length": 126.0, + "entropy": 0.45722317695617676, + "epoch": 0.385, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.6057567596435547, + "learning_rate": 6.2e-07, + "loss": 0.058352649211883545, + "num_tokens": 652771.0, + "reward": -8.283750534057617, + "reward_std": 77.68436431884766, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -8.383749008178711, + "rewards/supergames_reward/std": 77.68437194824219, + "step": 77, + "step_time": 14.142948755004909 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 324.0, + "completions/max_terminated_length": 324.0, + "completions/mean_length": 188.25, + "completions/mean_terminated_length": 188.25, + "completions/min_length": 68.0, + "completions/min_terminated_length": 68.0, + "entropy": 0.33380764722824097, + "epoch": 0.39, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.2239863872528076, + "learning_rate": 6.149999999999999e-07, + "loss": -0.04592633992433548, + "num_tokens": 662653.0, + "reward": 11.153749465942383, + "reward_std": 89.88329315185547, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 11.053749084472656, + "rewards/supergames_reward/std": 89.88329315185547, + "step": 78, + "step_time": 12.277474621019792 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 143.0, + "completions/max_terminated_length": 143.0, + "completions/mean_length": 108.875, + "completions/mean_terminated_length": 108.875, + "completions/min_length": 85.0, + "completions/min_terminated_length": 85.0, + "entropy": 0.3518812656402588, + "epoch": 0.395, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.967776298522949, + "learning_rate": 6.1e-07, + "loss": -0.10800496488809586, + "num_tokens": 668076.0, + "reward": 70.33250427246094, + "reward_std": 49.9145393371582, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 70.23249816894531, + "rewards/supergames_reward/std": 49.91453552246094, + "step": 79, + "step_time": 5.52714040101273 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 184.0, + "completions/max_terminated_length": 184.0, + "completions/mean_length": 152.125, + "completions/mean_terminated_length": 152.125, + "completions/min_length": 104.0, + "completions/min_terminated_length": 104.0, + "entropy": 0.4699896574020386, + "epoch": 0.4, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.561161994934082, + "learning_rate": 6.049999999999999e-07, + "loss": 0.09505834430456161, + "num_tokens": 675701.0, + "reward": -20.25625228881836, + "reward_std": 53.78542709350586, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -20.35624885559082, + "rewards/supergames_reward/std": 53.78542709350586, + "step": 80, + "step_time": 7.155081019998761 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.125, + "completions/max_length": 512.0, + "completions/max_terminated_length": 506.0, + "completions/mean_length": 288.875, + "completions/mean_terminated_length": 257.0, + "completions/min_length": 115.0, + "completions/min_terminated_length": 115.0, + "entropy": 0.4086494445800781, + "epoch": 0.405, + "frac_reward_zero_std": 0.0, + "grad_norm": 2.3719611167907715, + "learning_rate": 6e-07, + "loss": 0.04727627709507942, + "num_tokens": 687332.0, + "reward": -37.20750045776367, + "reward_std": 91.7676010131836, + "rewards/json_format_reward/mean": 0.05000000074505806, + "rewards/json_format_reward/std": 0.09258200973272324, + "rewards/supergames_reward/mean": -37.25749969482422, + "rewards/supergames_reward/std": 91.72847747802734, + "step": 81, + "step_time": 19.231399144016905 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 376.0, + "completions/max_terminated_length": 376.0, + "completions/mean_length": 227.75, + "completions/mean_terminated_length": 227.75, + "completions/min_length": 39.0, + "completions/min_terminated_length": 39.0, + "entropy": 0.43233707547187805, + "epoch": 0.41, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.3391714096069336, + "learning_rate": 5.949999999999999e-07, + "loss": -0.25524550676345825, + "num_tokens": 698474.0, + "reward": -3.7287511825561523, + "reward_std": 91.25379943847656, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -3.828749656677246, + "rewards/supergames_reward/std": 91.25379943847656, + "step": 82, + "step_time": 14.329176368017215 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 299.0, + "completions/max_terminated_length": 299.0, + "completions/mean_length": 178.125, + "completions/mean_terminated_length": 178.125, + "completions/min_length": 91.0, + "completions/min_terminated_length": 91.0, + "entropy": 0.4293636977672577, + "epoch": 0.415, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.7271742820739746, + "learning_rate": 5.9e-07, + "loss": 0.09273672848939896, + "num_tokens": 708243.0, + "reward": 12.802498817443848, + "reward_std": 86.23526000976562, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 12.702500343322754, + "rewards/supergames_reward/std": 86.23526000976562, + "step": 83, + "step_time": 11.419686011999147 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 425.0, + "completions/max_terminated_length": 425.0, + "completions/mean_length": 255.375, + "completions/mean_terminated_length": 255.375, + "completions/min_length": 172.0, + "completions/min_terminated_length": 172.0, + "entropy": 0.38068246841430664, + "epoch": 0.42, + "frac_reward_zero_std": 0.0, + "grad_norm": 2.6244184970855713, + "learning_rate": 5.849999999999999e-07, + "loss": 0.02006213366985321, + "num_tokens": 718606.0, + "reward": 37.803749084472656, + "reward_std": 57.58624267578125, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 37.70375061035156, + "rewards/supergames_reward/std": 57.58624267578125, + "step": 84, + "step_time": 15.975198492989875 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 116.0, + "completions/max_terminated_length": 116.0, + "completions/mean_length": 108.125, + "completions/mean_terminated_length": 108.125, + "completions/min_length": 87.0, + "completions/min_terminated_length": 87.0, + "entropy": 0.339572012424469, + "epoch": 0.425, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.754759311676025, + "learning_rate": 5.8e-07, + "loss": -0.011015941388905048, + "num_tokens": 723983.0, + "reward": -3.8212504386901855, + "reward_std": 61.624786376953125, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -3.921250820159912, + "rewards/supergames_reward/std": 61.624786376953125, + "step": 85, + "step_time": 4.542777584982105 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 447.0, + "completions/max_terminated_length": 447.0, + "completions/mean_length": 291.375, + "completions/mean_terminated_length": 291.375, + "completions/min_length": 163.0, + "completions/min_terminated_length": 163.0, + "entropy": 0.5846173763275146, + "epoch": 0.43, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.55724835395813, + "learning_rate": 5.749999999999999e-07, + "loss": -0.04981120675802231, + "num_tokens": 735642.0, + "reward": 50.07499694824219, + "reward_std": 92.6283187866211, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": 50.0, + "rewards/supergames_reward/std": 92.5820083618164, + "step": 86, + "step_time": 16.777178087009815 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 232.0, + "completions/max_terminated_length": 232.0, + "completions/mean_length": 192.0, + "completions/mean_terminated_length": 192.0, + "completions/min_length": 139.0, + "completions/min_terminated_length": 139.0, + "entropy": 0.5139474272727966, + "epoch": 0.435, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.8860411643981934, + "learning_rate": 5.699999999999999e-07, + "loss": -0.02194221317768097, + "num_tokens": 743602.0, + "reward": -52.522499084472656, + "reward_std": 73.8686294555664, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -52.62249755859375, + "rewards/supergames_reward/std": 73.86863708496094, + "step": 87, + "step_time": 8.772893911984283 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 266.0, + "completions/max_terminated_length": 266.0, + "completions/mean_length": 146.875, + "completions/mean_terminated_length": 146.875, + "completions/min_length": 89.0, + "completions/min_terminated_length": 89.0, + "entropy": 0.3428497910499573, + "epoch": 0.44, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.509483337402344, + "learning_rate": 5.649999999999999e-07, + "loss": 0.19857533276081085, + "num_tokens": 753065.0, + "reward": 63.63874816894531, + "reward_std": 67.54963684082031, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 63.53874969482422, + "rewards/supergames_reward/std": 67.54963684082031, + "step": 88, + "step_time": 10.230529835011112 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 446.0, + "completions/max_terminated_length": 446.0, + "completions/mean_length": 270.625, + "completions/mean_terminated_length": 270.625, + "completions/min_length": 103.0, + "completions/min_terminated_length": 103.0, + "entropy": 0.4417419731616974, + "epoch": 0.445, + "frac_reward_zero_std": 0.0, + "grad_norm": 2.973951816558838, + "learning_rate": 5.6e-07, + "loss": -0.00031440958264283836, + "num_tokens": 764534.0, + "reward": -9.736251831054688, + "reward_std": 92.25879669189453, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -9.836250305175781, + "rewards/supergames_reward/std": 92.25880432128906, + "step": 89, + "step_time": 17.40752330099349 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 398.0, + "completions/max_terminated_length": 398.0, + "completions/mean_length": 227.375, + "completions/mean_terminated_length": 227.375, + "completions/min_length": 98.0, + "completions/min_terminated_length": 98.0, + "entropy": 0.4732024073600769, + "epoch": 0.45, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.2560815811157227, + "learning_rate": 5.55e-07, + "loss": 0.03820464387536049, + "num_tokens": 772809.0, + "reward": 5.982499599456787, + "reward_std": 58.833797454833984, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 5.8824992179870605, + "rewards/supergames_reward/std": 58.83380126953125, + "step": 90, + "step_time": 15.007269965979503 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 269.0, + "completions/max_terminated_length": 269.0, + "completions/mean_length": 207.5, + "completions/mean_terminated_length": 207.5, + "completions/min_length": 92.0, + "completions/min_terminated_length": 92.0, + "entropy": 0.42038559913635254, + "epoch": 0.455, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.3649117946624756, + "learning_rate": 5.5e-07, + "loss": 0.1510048508644104, + "num_tokens": 782805.0, + "reward": -5.61500358581543, + "reward_std": 82.67152404785156, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -5.714998245239258, + "rewards/supergames_reward/std": 82.67152404785156, + "step": 91, + "step_time": 10.386137172987219 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 416.0, + "completions/max_terminated_length": 416.0, + "completions/mean_length": 218.375, + "completions/mean_terminated_length": 218.375, + "completions/min_length": 95.0, + "completions/min_terminated_length": 95.0, + "entropy": 0.3971734642982483, + "epoch": 0.46, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.433772563934326, + "learning_rate": 5.45e-07, + "loss": -0.3027213513851166, + "num_tokens": 793840.0, + "reward": -29.06500244140625, + "reward_std": 91.68325805664062, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -29.16499900817871, + "rewards/supergames_reward/std": 91.68325805664062, + "step": 92, + "step_time": 15.790611553995404 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 345.0, + "completions/max_terminated_length": 345.0, + "completions/mean_length": 206.625, + "completions/mean_terminated_length": 206.625, + "completions/min_length": 90.0, + "completions/min_terminated_length": 90.0, + "entropy": 0.5906907320022583, + "epoch": 0.465, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.293088912963867, + "learning_rate": 5.4e-07, + "loss": 0.09760032594203949, + "num_tokens": 803821.0, + "reward": -72.73999786376953, + "reward_std": 50.72484588623047, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": -72.81500244140625, + "rewards/supergames_reward/std": 50.7094841003418, + "step": 93, + "step_time": 13.043119941983605 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 143.0, + "completions/max_terminated_length": 143.0, + "completions/mean_length": 115.75, + "completions/mean_terminated_length": 115.75, + "completions/min_length": 80.0, + "completions/min_terminated_length": 80.0, + "entropy": 0.4233115315437317, + "epoch": 0.47, + "frac_reward_zero_std": 0.0, + "grad_norm": 5.58886194229126, + "learning_rate": 5.35e-07, + "loss": -0.04040650278329849, + "num_tokens": 809299.0, + "reward": -3.3050003051757812, + "reward_std": 24.443714141845703, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -3.405000686645508, + "rewards/supergames_reward/std": 24.443714141845703, + "step": 94, + "step_time": 5.512874762003776 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 122.0, + "completions/max_terminated_length": 122.0, + "completions/mean_length": 103.125, + "completions/mean_terminated_length": 103.125, + "completions/min_length": 86.0, + "completions/min_terminated_length": 86.0, + "entropy": 0.4055764377117157, + "epoch": 0.475, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.981549263000488, + "learning_rate": 5.3e-07, + "loss": -0.07588938623666763, + "num_tokens": 814716.0, + "reward": 60.06999969482422, + "reward_std": 42.708473205566406, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 59.970001220703125, + "rewards/supergames_reward/std": 42.708473205566406, + "step": 95, + "step_time": 4.766389057011111 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 475.0, + "completions/max_terminated_length": 475.0, + "completions/mean_length": 218.375, + "completions/mean_terminated_length": 218.375, + "completions/min_length": 81.0, + "completions/min_terminated_length": 81.0, + "entropy": 0.408913254737854, + "epoch": 0.48, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.79402756690979, + "learning_rate": 5.25e-07, + "loss": -0.3087541460990906, + "num_tokens": 824791.0, + "reward": -15.042500495910645, + "reward_std": 84.37246704101562, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -15.142499923706055, + "rewards/supergames_reward/std": 84.37246704101562, + "step": 96, + "step_time": 17.779843941010768 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 154.0, + "completions/max_terminated_length": 154.0, + "completions/mean_length": 121.625, + "completions/mean_terminated_length": 121.625, + "completions/min_length": 80.0, + "completions/min_terminated_length": 80.0, + "entropy": 0.529015839099884, + "epoch": 0.485, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.142816066741943, + "learning_rate": 5.2e-07, + "loss": 0.07216037809848785, + "num_tokens": 830300.0, + "reward": -32.743751525878906, + "reward_std": 54.61775588989258, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": -32.81875228881836, + "rewards/supergames_reward/std": 54.58256149291992, + "step": 97, + "step_time": 5.913989890017547 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 126.0, + "completions/max_terminated_length": 126.0, + "completions/mean_length": 95.625, + "completions/mean_terminated_length": 95.625, + "completions/min_length": 74.0, + "completions/min_terminated_length": 74.0, + "entropy": 0.33743274211883545, + "epoch": 0.49, + "frac_reward_zero_std": 0.0, + "grad_norm": 5.197273254394531, + "learning_rate": 5.149999999999999e-07, + "loss": 0.07500407844781876, + "num_tokens": 835633.0, + "reward": 11.00999927520752, + "reward_std": 62.73848342895508, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 10.910000801086426, + "rewards/supergames_reward/std": 62.73848342895508, + "step": 98, + "step_time": 4.9155233050114475 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 149.0, + "completions/max_terminated_length": 149.0, + "completions/mean_length": 108.5, + "completions/mean_terminated_length": 108.5, + "completions/min_length": 86.0, + "completions/min_terminated_length": 86.0, + "entropy": 0.3106803297996521, + "epoch": 0.495, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.864002227783203, + "learning_rate": 5.1e-07, + "loss": -0.1206185445189476, + "num_tokens": 841045.0, + "reward": 14.238749504089355, + "reward_std": 51.77016830444336, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 14.138750076293945, + "rewards/supergames_reward/std": 51.77016830444336, + "step": 99, + "step_time": 5.7507751359953545 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 149.0, + "completions/max_terminated_length": 149.0, + "completions/mean_length": 97.125, + "completions/mean_terminated_length": 97.125, + "completions/min_length": 73.0, + "completions/min_terminated_length": 73.0, + "entropy": 0.35160186886787415, + "epoch": 0.5, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.915731906890869, + "learning_rate": 5.049999999999999e-07, + "loss": -0.07951541244983673, + "num_tokens": 846398.0, + "reward": 67.78125, + "reward_std": 38.15436553955078, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 67.68124389648438, + "rewards/supergames_reward/std": 38.15436553955078, + "step": 100, + "step_time": 5.994720440998208 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 320.0, + "completions/max_terminated_length": 320.0, + "completions/mean_length": 208.75, + "completions/mean_terminated_length": 208.75, + "completions/min_length": 95.0, + "completions/min_terminated_length": 95.0, + "entropy": 0.452095091342926, + "epoch": 0.505, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.065176963806152, + "learning_rate": 5e-07, + "loss": -0.1397111564874649, + "num_tokens": 857420.0, + "reward": -5.801251411437988, + "reward_std": 73.56117248535156, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -5.90125036239624, + "rewards/supergames_reward/std": 73.5611801147461, + "step": 101, + "step_time": 12.28178753197426 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 230.0, + "completions/max_terminated_length": 230.0, + "completions/mean_length": 171.375, + "completions/mean_terminated_length": 171.375, + "completions/min_length": 95.0, + "completions/min_terminated_length": 95.0, + "entropy": 0.4138447642326355, + "epoch": 0.51, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.998823404312134, + "learning_rate": 4.95e-07, + "loss": -0.08941879868507385, + "num_tokens": 867103.0, + "reward": 92.24749755859375, + "reward_std": 14.540005683898926, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 92.14749908447266, + "rewards/supergames_reward/std": 14.540006637573242, + "step": 102, + "step_time": 8.992682139010867 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 311.0, + "completions/max_terminated_length": 311.0, + "completions/mean_length": 163.0, + "completions/mean_terminated_length": 163.0, + "completions/min_length": 74.0, + "completions/min_terminated_length": 74.0, + "entropy": 0.4182547330856323, + "epoch": 0.515, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.419595241546631, + "learning_rate": 4.9e-07, + "loss": -0.19304415583610535, + "num_tokens": 876679.0, + "reward": 79.13249969482422, + "reward_std": 59.3050422668457, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 79.03250122070312, + "rewards/supergames_reward/std": 59.30504608154297, + "step": 103, + "step_time": 11.836970244999975 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 253.0, + "completions/max_terminated_length": 253.0, + "completions/mean_length": 155.625, + "completions/mean_terminated_length": 155.625, + "completions/min_length": 94.0, + "completions/min_terminated_length": 94.0, + "entropy": 0.49889081716537476, + "epoch": 0.52, + "frac_reward_zero_std": 0.0, + "grad_norm": 5.463719367980957, + "learning_rate": 4.85e-07, + "loss": -0.04786720871925354, + "num_tokens": 884348.0, + "reward": 48.616249084472656, + "reward_std": 73.30741119384766, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 48.51625061035156, + "rewards/supergames_reward/std": 73.30741119384766, + "step": 104, + "step_time": 9.579507841990562 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 380.0, + "completions/max_terminated_length": 380.0, + "completions/mean_length": 269.5, + "completions/mean_terminated_length": 269.5, + "completions/min_length": 101.0, + "completions/min_terminated_length": 101.0, + "entropy": 0.5286482572555542, + "epoch": 0.525, + "frac_reward_zero_std": 0.0, + "grad_norm": 2.993295431137085, + "learning_rate": 4.8e-07, + "loss": -0.08071611076593399, + "num_tokens": 894856.0, + "reward": 34.970001220703125, + "reward_std": 85.324462890625, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 34.869998931884766, + "rewards/supergames_reward/std": 85.324462890625, + "step": 105, + "step_time": 14.295730627985904 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 261.0, + "completions/max_terminated_length": 261.0, + "completions/mean_length": 203.25, + "completions/mean_terminated_length": 203.25, + "completions/min_length": 91.0, + "completions/min_terminated_length": 91.0, + "entropy": 0.6049264073371887, + "epoch": 0.53, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.787306785583496, + "learning_rate": 4.7499999999999995e-07, + "loss": 0.19928883016109467, + "num_tokens": 902922.0, + "reward": 32.44499969482422, + "reward_std": 70.4663314819336, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": 32.369998931884766, + "rewards/supergames_reward/std": 70.41261291503906, + "step": 106, + "step_time": 9.84894504098338 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 279.0, + "completions/max_terminated_length": 279.0, + "completions/mean_length": 194.0, + "completions/mean_terminated_length": 194.0, + "completions/min_length": 109.0, + "completions/min_terminated_length": 109.0, + "entropy": 0.5115755200386047, + "epoch": 0.535, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.277207374572754, + "learning_rate": 4.6999999999999995e-07, + "loss": 0.09118300676345825, + "num_tokens": 910906.0, + "reward": 40.432498931884766, + "reward_std": 48.54548263549805, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 40.33250045776367, + "rewards/supergames_reward/std": 48.54548263549805, + "step": 107, + "step_time": 10.466603949986165 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 332.0, + "completions/max_terminated_length": 332.0, + "completions/mean_length": 195.875, + "completions/mean_terminated_length": 195.875, + "completions/min_length": 87.0, + "completions/min_terminated_length": 87.0, + "entropy": 0.4192962944507599, + "epoch": 0.54, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.308061122894287, + "learning_rate": 4.65e-07, + "loss": 0.28064602613449097, + "num_tokens": 920745.0, + "reward": -21.982500076293945, + "reward_std": 66.31195068359375, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -22.08249855041504, + "rewards/supergames_reward/std": 66.31195068359375, + "step": 108, + "step_time": 12.595848233992001 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 269.0, + "completions/max_terminated_length": 269.0, + "completions/mean_length": 195.125, + "completions/mean_terminated_length": 195.125, + "completions/min_length": 111.0, + "completions/min_terminated_length": 111.0, + "entropy": 0.37523961067199707, + "epoch": 0.545, + "frac_reward_zero_std": 0.0, + "grad_norm": 2.802706718444824, + "learning_rate": 4.6e-07, + "loss": -0.06614465266466141, + "num_tokens": 930618.0, + "reward": 71.75, + "reward_std": 51.14524459838867, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 71.6500015258789, + "rewards/supergames_reward/std": 51.1452522277832, + "step": 109, + "step_time": 10.33884758799104 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 331.0, + "completions/max_terminated_length": 331.0, + "completions/mean_length": 167.75, + "completions/mean_terminated_length": 167.75, + "completions/min_length": 89.0, + "completions/min_terminated_length": 89.0, + "entropy": 0.4339042007923126, + "epoch": 0.55, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.254003047943115, + "learning_rate": 4.55e-07, + "loss": 0.34405598044395447, + "num_tokens": 941256.0, + "reward": 99.21875, + "reward_std": 2.492553234100342, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 99.1187515258789, + "rewards/supergames_reward/std": 2.4925525188446045, + "step": 110, + "step_time": 12.657525141985388 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 285.0, + "completions/max_terminated_length": 285.0, + "completions/mean_length": 211.0, + "completions/mean_terminated_length": 211.0, + "completions/min_length": 112.0, + "completions/min_terminated_length": 112.0, + "entropy": 0.46809861063957214, + "epoch": 0.555, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.811166286468506, + "learning_rate": 4.5e-07, + "loss": 0.045113153755664825, + "num_tokens": 951296.0, + "reward": 55.27750015258789, + "reward_std": 47.90687561035156, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 55.1775016784668, + "rewards/supergames_reward/std": 47.9068717956543, + "step": 111, + "step_time": 10.91118389699841 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 145.0, + "completions/max_terminated_length": 145.0, + "completions/mean_length": 108.375, + "completions/mean_terminated_length": 108.375, + "completions/min_length": 86.0, + "completions/min_terminated_length": 86.0, + "entropy": 0.34181955456733704, + "epoch": 0.56, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.537296772003174, + "learning_rate": 4.45e-07, + "loss": -0.052413541823625565, + "num_tokens": 956675.0, + "reward": 47.56624984741211, + "reward_std": 49.185157775878906, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 47.466251373291016, + "rewards/supergames_reward/std": 49.18516159057617, + "step": 112, + "step_time": 5.552288047998445 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 245.0, + "completions/max_terminated_length": 245.0, + "completions/mean_length": 189.125, + "completions/mean_terminated_length": 189.125, + "completions/min_length": 102.0, + "completions/min_terminated_length": 102.0, + "entropy": 0.5034605860710144, + "epoch": 0.565, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.349876642227173, + "learning_rate": 4.3999999999999997e-07, + "loss": -0.08186715841293335, + "num_tokens": 964636.0, + "reward": 81.32374572753906, + "reward_std": 36.85074996948242, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 81.2237548828125, + "rewards/supergames_reward/std": 36.85074996948242, + "step": 113, + "step_time": 9.267611294984818 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 178.0, + "completions/max_terminated_length": 178.0, + "completions/mean_length": 111.25, + "completions/mean_terminated_length": 111.25, + "completions/min_length": 67.0, + "completions/min_terminated_length": 67.0, + "entropy": 0.3229002058506012, + "epoch": 0.57, + "frac_reward_zero_std": 0.0, + "grad_norm": 5.245575428009033, + "learning_rate": 4.3499999999999996e-07, + "loss": 0.06544198095798492, + "num_tokens": 970070.0, + "reward": 63.1775016784668, + "reward_std": 52.30624008178711, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 63.07749938964844, + "rewards/supergames_reward/std": 52.30624008178711, + "step": 114, + "step_time": 6.731139309995342 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 224.0, + "completions/max_terminated_length": 224.0, + "completions/mean_length": 170.0, + "completions/mean_terminated_length": 170.0, + "completions/min_length": 94.0, + "completions/min_terminated_length": 94.0, + "entropy": 0.3274807929992676, + "epoch": 0.575, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.5544140338897705, + "learning_rate": 4.2999999999999996e-07, + "loss": -0.06953569501638412, + "num_tokens": 979702.0, + "reward": 25.62125015258789, + "reward_std": 64.52003479003906, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 25.521244049072266, + "rewards/supergames_reward/std": 64.52003479003906, + "step": 115, + "step_time": 8.702618776995223 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 427.0, + "completions/max_terminated_length": 427.0, + "completions/mean_length": 209.5, + "completions/mean_terminated_length": 209.5, + "completions/min_length": 100.0, + "completions/min_terminated_length": 100.0, + "entropy": 0.4970880150794983, + "epoch": 0.58, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.4932165145874023, + "learning_rate": 4.2499999999999995e-07, + "loss": -0.2281273603439331, + "num_tokens": 990690.0, + "reward": 50.54750061035156, + "reward_std": 70.5126953125, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 50.4474983215332, + "rewards/supergames_reward/std": 70.51270294189453, + "step": 116, + "step_time": 16.204386608995264 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 228.0, + "completions/max_terminated_length": 228.0, + "completions/mean_length": 132.375, + "completions/mean_terminated_length": 132.375, + "completions/min_length": 85.0, + "completions/min_terminated_length": 85.0, + "entropy": 0.4380227029323578, + "epoch": 0.585, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.089909076690674, + "learning_rate": 4.1999999999999995e-07, + "loss": 0.12121254950761795, + "num_tokens": 996333.0, + "reward": -15.333749771118164, + "reward_std": 47.200164794921875, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -15.433748245239258, + "rewards/supergames_reward/std": 47.200164794921875, + "step": 117, + "step_time": 8.546233230998041 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 323.0, + "completions/max_terminated_length": 323.0, + "completions/mean_length": 137.625, + "completions/mean_terminated_length": 137.625, + "completions/min_length": 90.0, + "completions/min_terminated_length": 90.0, + "entropy": 0.4177803695201874, + "epoch": 0.59, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.8712472915649414, + "learning_rate": 4.1499999999999994e-07, + "loss": 0.14192955195903778, + "num_tokens": 1002010.0, + "reward": -30.700000762939453, + "reward_std": 17.224035263061523, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -30.799999237060547, + "rewards/supergames_reward/std": 17.224035263061523, + "step": 118, + "step_time": 11.878434990998358 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 453.0, + "completions/max_terminated_length": 453.0, + "completions/mean_length": 238.125, + "completions/mean_terminated_length": 238.125, + "completions/min_length": 109.0, + "completions/min_terminated_length": 109.0, + "entropy": 0.5640217065811157, + "epoch": 0.595, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.064737319946289, + "learning_rate": 4.0999999999999994e-07, + "loss": -0.023621462285518646, + "num_tokens": 1012259.0, + "reward": 43.683746337890625, + "reward_std": 70.4210205078125, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 43.5837516784668, + "rewards/supergames_reward/std": 70.4210205078125, + "step": 119, + "step_time": 16.829514264973113 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 236.0, + "completions/max_terminated_length": 236.0, + "completions/mean_length": 193.75, + "completions/mean_terminated_length": 193.75, + "completions/min_length": 151.0, + "completions/min_terminated_length": 151.0, + "entropy": 0.5795091390609741, + "epoch": 0.6, + "frac_reward_zero_std": 0.0, + "grad_norm": 5.423007965087891, + "learning_rate": 4.05e-07, + "loss": -0.0666111558675766, + "num_tokens": 1020241.0, + "reward": -85.86000061035156, + "reward_std": 27.4442081451416, + "rewards/json_format_reward/mean": 0.05000000074505806, + "rewards/json_format_reward/std": 0.09258200973272324, + "rewards/supergames_reward/mean": -85.90999603271484, + "rewards/supergames_reward/std": 27.41469955444336, + "step": 120, + "step_time": 8.95496519000153 + } + ], + "logging_steps": 1, + "max_steps": 200, + "num_input_tokens_seen": 1020241, + "num_train_epochs": 1, + "save_steps": 20, + "stateful_callbacks": { + "TrainerControl": { + "args": { + "should_epoch_stop": false, + "should_evaluate": false, + "should_log": false, + "should_save": true, + "should_training_stop": false + }, + "attributes": {} + } + }, + "total_flos": 0.0, + "train_batch_size": 8, + "trial_name": null, + "trial_params": null +} diff --git a/checkpoint-120/training_args.bin b/checkpoint-120/training_args.bin new file mode 100644 index 0000000..730dec0 --- /dev/null +++ b/checkpoint-120/training_args.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:691fe5be2d7d1d56160dae63866d136877a72efb30e15ff7b3249192f998a788 +size 7185 diff --git a/checkpoint-140/chat_template.jinja b/checkpoint-140/chat_template.jinja new file mode 100644 index 0000000..bdf7919 --- /dev/null +++ b/checkpoint-140/chat_template.jinja @@ -0,0 +1,54 @@ +{%- if tools %} + {{- '<|im_start|>system\n' }} + {%- if messages[0]['role'] == 'system' %} + {{- messages[0]['content'] }} + {%- else %} + {{- 'You are Qwen, created by Alibaba Cloud. You are a helpful assistant.' }} + {%- endif %} + {{- "\n\n# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }} + {%- for tool in tools %} + {{- "\n" }} + {{- tool | tojson }} + {%- endfor %} + {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }} +{%- else %} + {%- if messages[0]['role'] == 'system' %} + {{- '<|im_start|>system\n' + messages[0]['content'] + '<|im_end|>\n' }} + {%- else %} + {{- '<|im_start|>system\nYou are Qwen, created by Alibaba Cloud. You are a helpful assistant.<|im_end|>\n' }} + {%- endif %} +{%- endif %} +{%- for message in messages %} + {%- if (message.role == "user") or (message.role == "system" and not loop.first) or (message.role == "assistant" and not message.tool_calls) %} + {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }} + {%- elif message.role == "assistant" %} + {{- '<|im_start|>' + message.role }} + {%- if message.content %} + {{- '\n' + message.content }} + {%- endif %} + {%- for tool_call in message.tool_calls %} + {%- if tool_call.function is defined %} + {%- set tool_call = tool_call.function %} + {%- endif %} + {{- '\n\n{"name": "' }} + {{- tool_call.name }} + {{- '", "arguments": ' }} + {{- tool_call.arguments | tojson }} + {{- '}\n' }} + {%- endfor %} + {{- '<|im_end|>\n' }} + {%- elif message.role == "tool" %} + {%- if (loop.index0 == 0) or (messages[loop.index0 - 1].role != "tool") %} + {{- '<|im_start|>user' }} + {%- endif %} + {{- '\n\n' }} + {{- message.content }} + {{- '\n' }} + {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %} + {{- '<|im_end|>\n' }} + {%- endif %} + {%- endif %} +{%- endfor %} +{%- if add_generation_prompt %} + {{- '<|im_start|>assistant\n' }} +{%- endif %} diff --git a/checkpoint-140/config.json b/checkpoint-140/config.json new file mode 100644 index 0000000..f57b09b --- /dev/null +++ b/checkpoint-140/config.json @@ -0,0 +1,61 @@ +{ + "architectures": [ + "Qwen2ForCausalLM" + ], + "attention_dropout": 0.0, + "bos_token_id": null, + "dtype": "float32", + "eos_token_id": 151645, + "hidden_act": "silu", + "hidden_size": 1536, + "initializer_range": 0.02, + "intermediate_size": 8960, + "layer_types": [ + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention" + ], + "max_position_embeddings": 32768, + "max_window_layers": 21, + "model_type": "qwen2", + "num_attention_heads": 12, + "num_hidden_layers": 28, + "num_key_value_heads": 2, + "pad_token_id": 151643, + "rms_norm_eps": 1e-06, + "rope_parameters": { + "rope_theta": 1000000.0, + "rope_type": "default" + }, + "sliding_window": null, + "tie_word_embeddings": true, + "transformers_version": "5.6.2", + "use_cache": false, + "use_sliding_window": false, + "vocab_size": 151936 +} diff --git a/checkpoint-140/generation_config.json b/checkpoint-140/generation_config.json new file mode 100644 index 0000000..1b2bba9 --- /dev/null +++ b/checkpoint-140/generation_config.json @@ -0,0 +1,13 @@ +{ + "do_sample": true, + "eos_token_id": [ + 151645, + 151643 + ], + "pad_token_id": 151643, + "repetition_penalty": 1.1, + "temperature": 0.7, + "top_k": 20, + "top_p": 0.8, + "transformers_version": "5.6.2" +} diff --git a/checkpoint-140/model.safetensors b/checkpoint-140/model.safetensors new file mode 100644 index 0000000..387af85 --- /dev/null +++ b/checkpoint-140/model.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:7fdd60d1fa7be0cb10b8b346de5fc6cae3622c2276f19c10936a3d951a7a51af +size 6174895536 diff --git a/checkpoint-140/optimizer.pt b/checkpoint-140/optimizer.pt new file mode 100644 index 0000000..0bf27c4 --- /dev/null +++ b/checkpoint-140/optimizer.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:8249d1c911756c4ae68efa3d86cd5d15c283e6a678dc1914f57d936d46e7e7b7 +size 12350013801 diff --git a/checkpoint-140/rng_state.pth b/checkpoint-140/rng_state.pth new file mode 100644 index 0000000..1cbff49 --- /dev/null +++ b/checkpoint-140/rng_state.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:f30aa6f3347e5a84497bbadb0dd2e02b31bd49b38300a56ce489f81df37084c6 +size 14645 diff --git a/checkpoint-140/scheduler.pt b/checkpoint-140/scheduler.pt new file mode 100644 index 0000000..4bcce52 --- /dev/null +++ b/checkpoint-140/scheduler.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:06e07d7192e2daaedeea2396d67721d207d9a0175b75327949bf0087d7d181eb +size 1465 diff --git a/checkpoint-140/tokenizer.json b/checkpoint-140/tokenizer.json new file mode 100644 index 0000000..34510ff --- /dev/null +++ b/checkpoint-140/tokenizer.json @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:3fd169731d2cbde95e10bf356d66d5997fd885dd8dbb6fb4684da3f23b2585d8 +size 11421892 diff --git a/checkpoint-140/tokenizer_config.json b/checkpoint-140/tokenizer_config.json new file mode 100644 index 0000000..770e41d --- /dev/null +++ b/checkpoint-140/tokenizer_config.json @@ -0,0 +1,30 @@ +{ + "add_prefix_space": false, + "backend": "tokenizers", + "bos_token": null, + "clean_up_tokenization_spaces": false, + "eos_token": "<|im_end|>", + "errors": "replace", + "extra_special_tokens": [ + "<|im_start|>", + "<|im_end|>", + "<|object_ref_start|>", + "<|object_ref_end|>", + "<|box_start|>", + "<|box_end|>", + "<|quad_start|>", + "<|quad_end|>", + "<|vision_start|>", + "<|vision_end|>", + "<|vision_pad|>", + "<|image_pad|>", + "<|video_pad|>" + ], + "is_local": false, + "local_files_only": false, + "model_max_length": 131072, + "pad_token": "<|endoftext|>", + "split_special_tokens": false, + "tokenizer_class": "Qwen2Tokenizer", + "unk_token": null +} diff --git a/checkpoint-140/trainer_state.json b/checkpoint-140/trainer_state.json new file mode 100644 index 0000000..cfef8e8 --- /dev/null +++ b/checkpoint-140/trainer_state.json @@ -0,0 +1,4094 @@ +{ + "best_global_step": null, + "best_metric": null, + "best_model_checkpoint": null, + "epoch": 0.7, + "eval_steps": 500, + "global_step": 140, + "is_hyper_param_search": false, + "is_local_process_zero": true, + "is_world_process_zero": true, + "log_history": [ + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 478.0, + "completions/max_terminated_length": 478.0, + "completions/mean_length": 191.0, + "completions/mean_terminated_length": 191.0, + "completions/min_length": 66.0, + "completions/min_terminated_length": 66.0, + "entropy": 0.5566893219947815, + "epoch": 0.005, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.370361328125, + "learning_rate": 1e-06, + "loss": -0.24658073484897614, + "num_tokens": 10832.0, + "reward": -3.3000030517578125, + "reward_std": 85.62333679199219, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -3.4000015258789062, + "rewards/supergames_reward/std": 85.62333679199219, + "step": 1, + "step_time": 22.779711072013015 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 216.0, + "completions/max_terminated_length": 216.0, + "completions/mean_length": 131.5, + "completions/mean_terminated_length": 131.5, + "completions/min_length": 60.0, + "completions/min_terminated_length": 60.0, + "entropy": 0.6395841240882874, + "epoch": 0.01, + "frac_reward_zero_std": 0.0, + "grad_norm": 8.09774398803711, + "learning_rate": 9.95e-07, + "loss": 0.20567649602890015, + "num_tokens": 16404.0, + "reward": -12.422499656677246, + "reward_std": 7.134707450866699, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -12.522500038146973, + "rewards/supergames_reward/std": 7.134707927703857, + "step": 2, + "step_time": 9.960156448010821 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 287.0, + "completions/max_terminated_length": 287.0, + "completions/mean_length": 156.625, + "completions/mean_terminated_length": 156.625, + "completions/min_length": 77.0, + "completions/min_terminated_length": 77.0, + "entropy": 0.562222421169281, + "epoch": 0.015, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.9053616523742676, + "learning_rate": 9.9e-07, + "loss": 0.09602774679660797, + "num_tokens": 26953.0, + "reward": -74.94999694824219, + "reward_std": 70.73094177246094, + "rewards/json_format_reward/mean": 0.05000000074505806, + "rewards/json_format_reward/std": 0.09258200973272324, + "rewards/supergames_reward/mean": -75.0, + "rewards/supergames_reward/std": 70.71067810058594, + "step": 3, + "step_time": 13.686320498993155 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 268.0, + "completions/max_terminated_length": 268.0, + "completions/mean_length": 154.125, + "completions/mean_terminated_length": 154.125, + "completions/min_length": 53.0, + "completions/min_terminated_length": 53.0, + "entropy": 0.7323317527770996, + "epoch": 0.02, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.933310031890869, + "learning_rate": 9.849999999999999e-07, + "loss": 0.28249427676200867, + "num_tokens": 36514.0, + "reward": -40.7599983215332, + "reward_std": 81.73140716552734, + "rewards/json_format_reward/mean": 0.05000000074505806, + "rewards/json_format_reward/std": 0.09258200973272324, + "rewards/supergames_reward/mean": -40.80999755859375, + "rewards/supergames_reward/std": 81.68995666503906, + "step": 4, + "step_time": 12.805880262021674 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 401.0, + "completions/max_terminated_length": 401.0, + "completions/mean_length": 205.25, + "completions/mean_terminated_length": 205.25, + "completions/min_length": 115.0, + "completions/min_terminated_length": 115.0, + "entropy": 0.6167430281639099, + "epoch": 0.025, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.344135046005249, + "learning_rate": 9.8e-07, + "loss": 0.1433994024991989, + "num_tokens": 47476.0, + "reward": -97.6612548828125, + "reward_std": 6.3321428298950195, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -97.76124572753906, + "rewards/supergames_reward/std": 6.3321428298950195, + "step": 5, + "step_time": 18.789364666008623 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 276.0, + "completions/max_terminated_length": 276.0, + "completions/mean_length": 167.25, + "completions/mean_terminated_length": 167.25, + "completions/min_length": 73.0, + "completions/min_terminated_length": 73.0, + "entropy": 0.7583790421485901, + "epoch": 0.03, + "frac_reward_zero_std": 0.0, + "grad_norm": 5.172371864318848, + "learning_rate": 9.75e-07, + "loss": -0.12127404659986496, + "num_tokens": 55246.0, + "reward": -58.687503814697266, + "reward_std": 58.64059829711914, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": -58.76250457763672, + "rewards/supergames_reward/std": 58.620460510253906, + "step": 6, + "step_time": 12.948570273991209 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 261.0, + "completions/max_terminated_length": 261.0, + "completions/mean_length": 133.875, + "completions/mean_terminated_length": 133.875, + "completions/min_length": 57.0, + "completions/min_terminated_length": 57.0, + "entropy": 0.7447654604911804, + "epoch": 0.035, + "frac_reward_zero_std": 0.0, + "grad_norm": 6.428595066070557, + "learning_rate": 9.7e-07, + "loss": 0.10135584324598312, + "num_tokens": 60837.0, + "reward": -32.45000076293945, + "reward_std": 41.754486083984375, + "rewards/json_format_reward/mean": 0.05000000074505806, + "rewards/json_format_reward/std": 0.09258200973272324, + "rewards/supergames_reward/mean": -32.5, + "rewards/supergames_reward/std": 41.66190719604492, + "step": 7, + "step_time": 12.058315072004916 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 414.0, + "completions/max_terminated_length": 414.0, + "completions/mean_length": 209.125, + "completions/mean_terminated_length": 209.125, + "completions/min_length": 104.0, + "completions/min_terminated_length": 104.0, + "entropy": 0.3860666751861572, + "epoch": 0.04, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.806255578994751, + "learning_rate": 9.649999999999999e-07, + "loss": -0.33238139748573303, + "num_tokens": 70838.0, + "reward": -80.50375366210938, + "reward_std": 38.82380294799805, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -80.60375213623047, + "rewards/supergames_reward/std": 38.82379913330078, + "step": 8, + "step_time": 19.323370319994865 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.25, + "completions/max_length": 512.0, + "completions/max_terminated_length": 199.0, + "completions/mean_length": 247.125, + "completions/mean_terminated_length": 158.83334350585938, + "completions/min_length": 106.0, + "completions/min_terminated_length": 106.0, + "entropy": 0.5816237330436707, + "epoch": 0.045, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.3209943771362305, + "learning_rate": 9.6e-07, + "loss": 0.36752766370773315, + "num_tokens": 82135.0, + "reward": -27.688751220703125, + "reward_std": 82.8358154296875, + "rewards/json_format_reward/mean": 0.05000000074505806, + "rewards/json_format_reward/std": 0.09258200973272324, + "rewards/supergames_reward/mean": -27.738750457763672, + "rewards/supergames_reward/std": 82.78590393066406, + "step": 9, + "step_time": 21.948575104994234 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 240.0, + "completions/max_terminated_length": 240.0, + "completions/mean_length": 153.0, + "completions/mean_terminated_length": 153.0, + "completions/min_length": 51.0, + "completions/min_terminated_length": 51.0, + "entropy": 0.6301521062850952, + "epoch": 0.05, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.125359058380127, + "learning_rate": 9.55e-07, + "loss": 0.06129350885748863, + "num_tokens": 89807.0, + "reward": -78.39624786376953, + "reward_std": 40.17042922973633, + "rewards/json_format_reward/mean": 0.05000000074505806, + "rewards/json_format_reward/std": 0.09258200973272324, + "rewards/supergames_reward/mean": -78.44625091552734, + "rewards/supergames_reward/std": 40.1396484375, + "step": 10, + "step_time": 9.098202993016457 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 133.0, + "completions/max_terminated_length": 133.0, + "completions/mean_length": 112.0, + "completions/mean_terminated_length": 112.0, + "completions/min_length": 68.0, + "completions/min_terminated_length": 68.0, + "entropy": 0.5521990656852722, + "epoch": 0.055, + "frac_reward_zero_std": 0.0, + "grad_norm": 6.616060256958008, + "learning_rate": 9.499999999999999e-07, + "loss": -0.006313305348157883, + "num_tokens": 95231.0, + "reward": -12.161249160766602, + "reward_std": 6.39578104019165, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -12.261249542236328, + "rewards/supergames_reward/std": 6.39578104019165, + "step": 11, + "step_time": 5.136311663984088 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 396.0, + "completions/max_terminated_length": 396.0, + "completions/mean_length": 214.875, + "completions/mean_terminated_length": 214.875, + "completions/min_length": 55.0, + "completions/min_terminated_length": 55.0, + "entropy": 0.5505905151367188, + "epoch": 0.06, + "frac_reward_zero_std": 0.0, + "grad_norm": 5.289583206176758, + "learning_rate": 9.45e-07, + "loss": -0.29802003502845764, + "num_tokens": 106238.0, + "reward": -91.66999816894531, + "reward_std": 23.2779541015625, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -91.77000427246094, + "rewards/supergames_reward/std": 23.277956008911133, + "step": 12, + "step_time": 14.945365622988902 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 269.0, + "completions/max_terminated_length": 269.0, + "completions/mean_length": 167.375, + "completions/mean_terminated_length": 167.375, + "completions/min_length": 113.0, + "completions/min_terminated_length": 113.0, + "entropy": 0.540415346622467, + "epoch": 0.065, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.858291149139404, + "learning_rate": 9.399999999999999e-07, + "loss": 0.24004128575325012, + "num_tokens": 116937.0, + "reward": -20.38249969482422, + "reward_std": 87.55204010009766, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": -20.457500457763672, + "rewards/supergames_reward/std": 87.52605438232422, + "step": 13, + "step_time": 10.482285185018554 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 254.0, + "completions/max_terminated_length": 254.0, + "completions/mean_length": 156.875, + "completions/mean_terminated_length": 156.875, + "completions/min_length": 114.0, + "completions/min_terminated_length": 114.0, + "entropy": 0.7237679958343506, + "epoch": 0.07, + "frac_reward_zero_std": 0.0, + "grad_norm": 5.083980560302734, + "learning_rate": 9.35e-07, + "loss": -0.19025824964046478, + "num_tokens": 124640.0, + "reward": -51.397499084472656, + "reward_std": 54.486454010009766, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -51.49749755859375, + "rewards/supergames_reward/std": 54.486454010009766, + "step": 14, + "step_time": 9.597231683001155 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 236.0, + "completions/max_terminated_length": 236.0, + "completions/mean_length": 141.375, + "completions/mean_terminated_length": 141.375, + "completions/min_length": 91.0, + "completions/min_terminated_length": 91.0, + "entropy": 0.754334568977356, + "epoch": 0.075, + "frac_reward_zero_std": 0.0, + "grad_norm": 5.720428943634033, + "learning_rate": 9.3e-07, + "loss": -0.16830675303936005, + "num_tokens": 132219.0, + "reward": -56.226253509521484, + "reward_std": 60.115657806396484, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -56.32625198364258, + "rewards/supergames_reward/std": 60.11566162109375, + "step": 15, + "step_time": 8.998362872982398 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 209.0, + "completions/max_terminated_length": 209.0, + "completions/mean_length": 136.375, + "completions/mean_terminated_length": 136.375, + "completions/min_length": 97.0, + "completions/min_terminated_length": 97.0, + "entropy": 0.6280568838119507, + "epoch": 0.08, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.9681806564331055, + "learning_rate": 9.25e-07, + "loss": -0.01910916529595852, + "num_tokens": 139726.0, + "reward": -68.24000549316406, + "reward_std": 52.238914489746094, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -68.33999633789062, + "rewards/supergames_reward/std": 52.238914489746094, + "step": 16, + "step_time": 8.034480294008972 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 365.0, + "completions/max_terminated_length": 365.0, + "completions/mean_length": 211.875, + "completions/mean_terminated_length": 211.875, + "completions/min_length": 100.0, + "completions/min_terminated_length": 100.0, + "entropy": 0.899176836013794, + "epoch": 0.085, + "frac_reward_zero_std": 0.0, + "grad_norm": 6.691746711730957, + "learning_rate": 9.2e-07, + "loss": 0.25513291358947754, + "num_tokens": 147869.0, + "reward": -99.92500305175781, + "reward_std": 0.07070883363485336, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": -100.0, + "rewards/supergames_reward/std": 0.0, + "step": 17, + "step_time": 13.655792869016295 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 387.0, + "completions/max_terminated_length": 387.0, + "completions/mean_length": 238.0, + "completions/mean_terminated_length": 238.0, + "completions/min_length": 87.0, + "completions/min_terminated_length": 87.0, + "entropy": 0.5596873760223389, + "epoch": 0.09, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.6711347103118896, + "learning_rate": 9.15e-07, + "loss": -0.2628335654735565, + "num_tokens": 159101.0, + "reward": -24.900001525878906, + "reward_std": 103.50983428955078, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -25.0, + "rewards/supergames_reward/std": 103.50983428955078, + "step": 18, + "step_time": 14.784329420013819 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.125, + "completions/max_length": 512.0, + "completions/max_terminated_length": 277.0, + "completions/mean_length": 219.75, + "completions/mean_terminated_length": 178.00001525878906, + "completions/min_length": 90.0, + "completions/min_terminated_length": 90.0, + "entropy": 0.9245517253875732, + "epoch": 0.095, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.55502986907959, + "learning_rate": 9.1e-07, + "loss": 0.13060179352760315, + "num_tokens": 167299.0, + "reward": -86.89125061035156, + "reward_std": 36.95603561401367, + "rewards/json_format_reward/mean": 0.05000000074505806, + "rewards/json_format_reward/std": 0.09258200973272324, + "rewards/supergames_reward/mean": -86.94125366210938, + "rewards/supergames_reward/std": 36.93572235107422, + "step": 19, + "step_time": 18.832684700988466 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 267.0, + "completions/max_terminated_length": 267.0, + "completions/mean_length": 177.25, + "completions/mean_terminated_length": 177.25, + "completions/min_length": 93.0, + "completions/min_terminated_length": 93.0, + "entropy": 0.7331254482269287, + "epoch": 0.1, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.538497447967529, + "learning_rate": 9.05e-07, + "loss": -0.16418980062007904, + "num_tokens": 175141.0, + "reward": -68.52874755859375, + "reward_std": 58.48719787597656, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -68.62875366210938, + "rewards/supergames_reward/std": 58.48720169067383, + "step": 20, + "step_time": 10.090975169994636 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 300.0, + "completions/max_terminated_length": 300.0, + "completions/mean_length": 190.625, + "completions/mean_terminated_length": 190.625, + "completions/min_length": 103.0, + "completions/min_terminated_length": 103.0, + "entropy": 0.6659425497055054, + "epoch": 0.105, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.485373497009277, + "learning_rate": 9e-07, + "loss": -0.10106432437896729, + "num_tokens": 183098.0, + "reward": -71.48875427246094, + "reward_std": 55.33565139770508, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -71.5887451171875, + "rewards/supergames_reward/std": 55.33565139770508, + "step": 21, + "step_time": 11.255001295008697 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 279.0, + "completions/max_terminated_length": 279.0, + "completions/mean_length": 179.0, + "completions/mean_terminated_length": 179.0, + "completions/min_length": 103.0, + "completions/min_terminated_length": 103.0, + "entropy": 0.49365949630737305, + "epoch": 0.11, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.829419136047363, + "learning_rate": 8.95e-07, + "loss": 0.10587231814861298, + "num_tokens": 192826.0, + "reward": -75.62000274658203, + "reward_std": 45.01959991455078, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": -75.69499969482422, + "rewards/supergames_reward/std": 45.00411605834961, + "step": 22, + "step_time": 10.611246599000879 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 315.0, + "completions/max_terminated_length": 315.0, + "completions/mean_length": 212.625, + "completions/mean_terminated_length": 212.625, + "completions/min_length": 100.0, + "completions/min_terminated_length": 100.0, + "entropy": 0.43159469962120056, + "epoch": 0.115, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.541125774383545, + "learning_rate": 8.9e-07, + "loss": -0.1681259125471115, + "num_tokens": 202815.0, + "reward": 23.67624855041504, + "reward_std": 67.74340057373047, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 23.576250076293945, + "rewards/supergames_reward/std": 67.743408203125, + "step": 23, + "step_time": 11.731771531980485 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 168.0, + "completions/max_terminated_length": 168.0, + "completions/mean_length": 101.625, + "completions/mean_terminated_length": 101.625, + "completions/min_length": 65.0, + "completions/min_terminated_length": 65.0, + "entropy": 0.3972298502922058, + "epoch": 0.12, + "frac_reward_zero_std": 0.0, + "grad_norm": 5.8285417556762695, + "learning_rate": 8.85e-07, + "loss": -0.011585958302021027, + "num_tokens": 208188.0, + "reward": 58.849998474121094, + "reward_std": 56.9304084777832, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 58.75, + "rewards/supergames_reward/std": 56.9304084777832, + "step": 24, + "step_time": 6.187504950998118 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 303.0, + "completions/max_terminated_length": 303.0, + "completions/mean_length": 165.375, + "completions/mean_terminated_length": 165.375, + "completions/min_length": 69.0, + "completions/min_terminated_length": 69.0, + "entropy": 0.6339899301528931, + "epoch": 0.125, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.823799133300781, + "learning_rate": 8.799999999999999e-07, + "loss": 0.05159185826778412, + "num_tokens": 215935.0, + "reward": -23.521251678466797, + "reward_std": 74.23067474365234, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -23.621248245239258, + "rewards/supergames_reward/std": 74.23067474365234, + "step": 25, + "step_time": 11.059416300005978 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 349.0, + "completions/max_terminated_length": 349.0, + "completions/mean_length": 204.0, + "completions/mean_terminated_length": 204.0, + "completions/min_length": 143.0, + "completions/min_terminated_length": 143.0, + "entropy": 0.6462895274162292, + "epoch": 0.13, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.969292640686035, + "learning_rate": 8.75e-07, + "loss": 0.0115677984431386, + "num_tokens": 223999.0, + "reward": -74.8175048828125, + "reward_std": 40.94221115112305, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": -74.89250183105469, + "rewards/supergames_reward/std": 40.924625396728516, + "step": 26, + "step_time": 12.666237785975682 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 299.0, + "completions/max_terminated_length": 299.0, + "completions/mean_length": 166.5, + "completions/mean_terminated_length": 166.5, + "completions/min_length": 87.0, + "completions/min_terminated_length": 87.0, + "entropy": 0.4756850600242615, + "epoch": 0.135, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.915069103240967, + "learning_rate": 8.699999999999999e-07, + "loss": -0.13321346044540405, + "num_tokens": 233643.0, + "reward": -52.125, + "reward_std": 19.405282974243164, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": -52.20000076293945, + "rewards/supergames_reward/std": 19.33464813232422, + "step": 27, + "step_time": 11.172539752995363 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 184.0, + "completions/max_terminated_length": 184.0, + "completions/mean_length": 137.125, + "completions/mean_terminated_length": 137.125, + "completions/min_length": 100.0, + "completions/min_terminated_length": 100.0, + "entropy": 0.5834778547286987, + "epoch": 0.14, + "frac_reward_zero_std": 0.0, + "grad_norm": 5.146666526794434, + "learning_rate": 8.65e-07, + "loss": 0.060454584658145905, + "num_tokens": 241204.0, + "reward": -68.99250793457031, + "reward_std": 51.010581970214844, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -69.09249877929688, + "rewards/supergames_reward/std": 51.010581970214844, + "step": 28, + "step_time": 7.025046669004951 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 331.0, + "completions/max_terminated_length": 331.0, + "completions/mean_length": 182.375, + "completions/mean_terminated_length": 182.375, + "completions/min_length": 106.0, + "completions/min_terminated_length": 106.0, + "entropy": 0.6368551254272461, + "epoch": 0.145, + "frac_reward_zero_std": 0.0, + "grad_norm": 5.15535306930542, + "learning_rate": 8.599999999999999e-07, + "loss": -0.3323673903942108, + "num_tokens": 249087.0, + "reward": -44.17000198364258, + "reward_std": 80.5373306274414, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -44.27000045776367, + "rewards/supergames_reward/std": 80.5373306274414, + "step": 29, + "step_time": 12.079259724996518 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 190.0, + "completions/max_terminated_length": 190.0, + "completions/mean_length": 127.75, + "completions/mean_terminated_length": 127.75, + "completions/min_length": 80.0, + "completions/min_terminated_length": 80.0, + "entropy": 0.4491196870803833, + "epoch": 0.15, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.896186828613281, + "learning_rate": 8.55e-07, + "loss": -0.12887656688690186, + "num_tokens": 259445.0, + "reward": 57.446250915527344, + "reward_std": 38.78837585449219, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 57.34625244140625, + "rewards/supergames_reward/std": 38.78837966918945, + "step": 30, + "step_time": 7.483972909016302 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 166.0, + "completions/max_terminated_length": 166.0, + "completions/mean_length": 112.875, + "completions/mean_terminated_length": 112.875, + "completions/min_length": 85.0, + "completions/min_terminated_length": 85.0, + "entropy": 0.4858454465866089, + "epoch": 0.155, + "frac_reward_zero_std": 0.0, + "grad_norm": 6.742906093597412, + "learning_rate": 8.499999999999999e-07, + "loss": 0.06930096447467804, + "num_tokens": 264924.0, + "reward": -21.174999237060547, + "reward_std": 31.89051628112793, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": -21.25, + "rewards/supergames_reward/std": 31.819807052612305, + "step": 31, + "step_time": 6.1866529019898735 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 148.0, + "completions/max_terminated_length": 148.0, + "completions/mean_length": 118.125, + "completions/mean_terminated_length": 118.125, + "completions/min_length": 98.0, + "completions/min_terminated_length": 98.0, + "entropy": 0.45428749918937683, + "epoch": 0.16, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.663414478302002, + "learning_rate": 8.45e-07, + "loss": 0.05735419690608978, + "num_tokens": 270381.0, + "reward": 0.4687504768371582, + "reward_std": 38.66371154785156, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 0.36875057220458984, + "rewards/supergames_reward/std": 38.6637077331543, + "step": 32, + "step_time": 5.577042757999152 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 345.0, + "completions/max_terminated_length": 345.0, + "completions/mean_length": 166.875, + "completions/mean_terminated_length": 166.875, + "completions/min_length": 86.0, + "completions/min_terminated_length": 86.0, + "entropy": 0.6107293367385864, + "epoch": 0.165, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.75006628036499, + "learning_rate": 8.399999999999999e-07, + "loss": -0.14760127663612366, + "num_tokens": 278156.0, + "reward": -87.84750366210938, + "reward_std": 26.84708023071289, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -87.94749450683594, + "rewards/supergames_reward/std": 26.847076416015625, + "step": 33, + "step_time": 12.552876825997373 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.125, + "completions/max_length": 512.0, + "completions/max_terminated_length": 360.0, + "completions/mean_length": 231.0, + "completions/mean_terminated_length": 190.85714721679688, + "completions/min_length": 74.0, + "completions/min_terminated_length": 74.0, + "entropy": 0.6089653372764587, + "epoch": 0.17, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.781267166137695, + "learning_rate": 8.349999999999999e-07, + "loss": 0.4294504225254059, + "num_tokens": 289364.0, + "reward": -99.92500305175781, + "reward_std": 0.07070883363485336, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": -100.0, + "rewards/supergames_reward/std": 0.0, + "step": 34, + "step_time": 18.961819477990502 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 218.0, + "completions/max_terminated_length": 218.0, + "completions/mean_length": 134.125, + "completions/mean_terminated_length": 134.125, + "completions/min_length": 104.0, + "completions/min_terminated_length": 104.0, + "entropy": 0.46158918738365173, + "epoch": 0.175, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.016353607177734, + "learning_rate": 8.299999999999999e-07, + "loss": 0.09251818805932999, + "num_tokens": 298733.0, + "reward": -35.558753967285156, + "reward_std": 15.881204605102539, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -35.65875244140625, + "rewards/supergames_reward/std": 15.881203651428223, + "step": 35, + "step_time": 8.38491243700264 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 207.0, + "completions/max_terminated_length": 207.0, + "completions/mean_length": 156.875, + "completions/mean_terminated_length": 156.875, + "completions/min_length": 111.0, + "completions/min_terminated_length": 111.0, + "entropy": 0.5104788541793823, + "epoch": 0.18, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.592893123626709, + "learning_rate": 8.249999999999999e-07, + "loss": -0.0410832017660141, + "num_tokens": 306420.0, + "reward": 5.4212493896484375, + "reward_std": 93.50404357910156, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": 5.346250534057617, + "rewards/supergames_reward/std": 93.4718246459961, + "step": 36, + "step_time": 7.811868985998444 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 241.0, + "completions/max_terminated_length": 241.0, + "completions/mean_length": 178.0, + "completions/mean_terminated_length": 178.0, + "completions/min_length": 106.0, + "completions/min_terminated_length": 106.0, + "entropy": 0.7527878880500793, + "epoch": 0.185, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.8812546730041504, + "learning_rate": 8.199999999999999e-07, + "loss": 0.12539231777191162, + "num_tokens": 314252.0, + "reward": -89.58500671386719, + "reward_std": 29.256120681762695, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": -89.66000366210938, + "rewards/supergames_reward/std": 29.24593734741211, + "step": 37, + "step_time": 8.959742914012168 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 199.0, + "completions/max_terminated_length": 199.0, + "completions/mean_length": 133.375, + "completions/mean_terminated_length": 133.375, + "completions/min_length": 87.0, + "completions/min_terminated_length": 87.0, + "entropy": 0.3347901403903961, + "epoch": 0.19, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.205414295196533, + "learning_rate": 8.149999999999999e-07, + "loss": -0.06991486996412277, + "num_tokens": 319839.0, + "reward": -14.498749732971191, + "reward_std": 13.007229804992676, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -14.598750114440918, + "rewards/supergames_reward/std": 13.007229804992676, + "step": 38, + "step_time": 7.377183554985095 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 181.0, + "completions/max_terminated_length": 181.0, + "completions/mean_length": 114.375, + "completions/mean_terminated_length": 114.375, + "completions/min_length": 58.0, + "completions/min_terminated_length": 58.0, + "entropy": 0.4445965886116028, + "epoch": 0.195, + "frac_reward_zero_std": 0.0, + "grad_norm": 5.561298847198486, + "learning_rate": 8.1e-07, + "loss": -0.06235164776444435, + "num_tokens": 325290.0, + "reward": -7.425000190734863, + "reward_std": 53.701602935791016, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": -7.5, + "rewards/supergames_reward/std": 53.652320861816406, + "step": 39, + "step_time": 6.753862089011818 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 370.0, + "completions/max_terminated_length": 370.0, + "completions/mean_length": 187.625, + "completions/mean_terminated_length": 187.625, + "completions/min_length": 60.0, + "completions/min_terminated_length": 60.0, + "entropy": 0.3950234353542328, + "epoch": 0.2, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.5988712310791016, + "learning_rate": 8.05e-07, + "loss": -0.1381440907716751, + "num_tokens": 335119.0, + "reward": -33.68375015258789, + "reward_std": 42.069435119628906, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -33.783748626708984, + "rewards/supergames_reward/std": 42.069435119628906, + "step": 40, + "step_time": 13.718958162004128 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.125, + "completions/max_length": 512.0, + "completions/max_terminated_length": 371.0, + "completions/mean_length": 251.125, + "completions/mean_terminated_length": 213.85714721679688, + "completions/min_length": 121.0, + "completions/min_terminated_length": 121.0, + "entropy": 0.507067084312439, + "epoch": 0.205, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.3898658752441406, + "learning_rate": 8e-07, + "loss": 0.0078964838758111, + "num_tokens": 346464.0, + "reward": -86.48625183105469, + "reward_std": 38.020694732666016, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": -86.56124877929688, + "rewards/supergames_reward/std": 38.010528564453125, + "step": 41, + "step_time": 18.90960379401804 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 152.0, + "completions/max_terminated_length": 152.0, + "completions/mean_length": 116.375, + "completions/mean_terminated_length": 116.375, + "completions/min_length": 93.0, + "completions/min_terminated_length": 93.0, + "entropy": 0.3801707923412323, + "epoch": 0.21, + "frac_reward_zero_std": 1.0, + "grad_norm": 0.0, + "learning_rate": 7.95e-07, + "loss": 0.0, + "num_tokens": 351955.0, + "reward": -9.899999618530273, + "reward_std": 0.0, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -10.0, + "rewards/supergames_reward/std": 0.0, + "step": 42, + "step_time": 5.711630532023264 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 308.0, + "completions/max_terminated_length": 308.0, + "completions/mean_length": 247.375, + "completions/mean_terminated_length": 247.375, + "completions/min_length": 116.0, + "completions/min_terminated_length": 116.0, + "entropy": 0.3185043931007385, + "epoch": 0.215, + "frac_reward_zero_std": 0.0, + "grad_norm": 2.168762445449829, + "learning_rate": 7.9e-07, + "loss": 0.1970764398574829, + "num_tokens": 363278.0, + "reward": -53.89249801635742, + "reward_std": 63.02619934082031, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -53.99250030517578, + "rewards/supergames_reward/std": 63.02619934082031, + "step": 43, + "step_time": 11.661934906995157 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 330.0, + "completions/max_terminated_length": 330.0, + "completions/mean_length": 212.5, + "completions/mean_terminated_length": 212.5, + "completions/min_length": 106.0, + "completions/min_terminated_length": 106.0, + "entropy": 0.4446268379688263, + "epoch": 0.22, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.305844306945801, + "learning_rate": 7.85e-07, + "loss": -0.06735197454690933, + "num_tokens": 374306.0, + "reward": 18.059999465942383, + "reward_std": 98.9417953491211, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 17.959999084472656, + "rewards/supergames_reward/std": 98.9417953491211, + "step": 44, + "step_time": 12.492453911982011 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 316.0, + "completions/max_terminated_length": 316.0, + "completions/mean_length": 183.0, + "completions/mean_terminated_length": 183.0, + "completions/min_length": 103.0, + "completions/min_terminated_length": 103.0, + "entropy": 0.5269614458084106, + "epoch": 0.225, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.8006370067596436, + "learning_rate": 7.799999999999999e-07, + "loss": -0.08870815485715866, + "num_tokens": 382210.0, + "reward": -54.928749084472656, + "reward_std": 72.5127944946289, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -55.02874755859375, + "rewards/supergames_reward/std": 72.5127944946289, + "step": 45, + "step_time": 11.7636467939883 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 244.0, + "completions/max_terminated_length": 244.0, + "completions/mean_length": 136.625, + "completions/mean_terminated_length": 136.625, + "completions/min_length": 85.0, + "completions/min_terminated_length": 85.0, + "entropy": 0.34820204973220825, + "epoch": 0.23, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.911736011505127, + "learning_rate": 7.75e-07, + "loss": 0.04397330805659294, + "num_tokens": 391583.0, + "reward": -19.10375213623047, + "reward_std": 23.967702865600586, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -19.203750610351562, + "rewards/supergames_reward/std": 23.967702865600586, + "step": 46, + "step_time": 10.119426151999505 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 218.0, + "completions/max_terminated_length": 218.0, + "completions/mean_length": 160.375, + "completions/mean_terminated_length": 160.375, + "completions/min_length": 93.0, + "completions/min_terminated_length": 93.0, + "entropy": 0.611914336681366, + "epoch": 0.235, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.571913242340088, + "learning_rate": 7.699999999999999e-07, + "loss": -0.044293053448200226, + "num_tokens": 399290.0, + "reward": -25.186250686645508, + "reward_std": 71.46305084228516, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -25.286251068115234, + "rewards/supergames_reward/std": 71.46305084228516, + "step": 47, + "step_time": 8.563868903991533 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 167.0, + "completions/max_terminated_length": 167.0, + "completions/mean_length": 111.125, + "completions/mean_terminated_length": 111.125, + "completions/min_length": 56.0, + "completions/min_terminated_length": 56.0, + "entropy": 0.46585777401924133, + "epoch": 0.24, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.654109954833984, + "learning_rate": 7.65e-07, + "loss": -0.035151124000549316, + "num_tokens": 404707.0, + "reward": -0.408750057220459, + "reward_std": 31.939748764038086, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -0.5087499618530273, + "rewards/supergames_reward/std": 31.93975257873535, + "step": 48, + "step_time": 6.438482580007985 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 306.0, + "completions/max_terminated_length": 306.0, + "completions/mean_length": 216.625, + "completions/mean_terminated_length": 216.625, + "completions/min_length": 132.0, + "completions/min_terminated_length": 132.0, + "entropy": 0.33054471015930176, + "epoch": 0.245, + "frac_reward_zero_std": 0.0, + "grad_norm": 2.700303792953491, + "learning_rate": 7.599999999999999e-07, + "loss": -0.04393656551837921, + "num_tokens": 414736.0, + "reward": -97.26000213623047, + "reward_std": 1.6294406652450562, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -97.36000061035156, + "rewards/supergames_reward/std": 1.629441738128662, + "step": 49, + "step_time": 11.636040106008295 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 378.0, + "completions/max_terminated_length": 378.0, + "completions/mean_length": 218.375, + "completions/mean_terminated_length": 218.375, + "completions/min_length": 104.0, + "completions/min_terminated_length": 104.0, + "entropy": 0.48711177706718445, + "epoch": 0.25, + "frac_reward_zero_std": 0.0, + "grad_norm": 2.9650564193725586, + "learning_rate": 7.55e-07, + "loss": -0.3529520332813263, + "num_tokens": 424795.0, + "reward": 47.408748626708984, + "reward_std": 72.72083282470703, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 47.30875015258789, + "rewards/supergames_reward/std": 72.72083282470703, + "step": 50, + "step_time": 14.258096842997475 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.125, + "completions/max_length": 512.0, + "completions/max_terminated_length": 492.0, + "completions/mean_length": 267.625, + "completions/mean_terminated_length": 232.71429443359375, + "completions/min_length": 90.0, + "completions/min_terminated_length": 90.0, + "entropy": 0.4482860863208771, + "epoch": 0.255, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.2998197078704834, + "learning_rate": 7.5e-07, + "loss": 0.2493654191493988, + "num_tokens": 436288.0, + "reward": 50.61125183105469, + "reward_std": 75.3349838256836, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": 50.53624725341797, + "rewards/supergames_reward/std": 75.27783966064453, + "step": 51, + "step_time": 19.20827590499539 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 176.0, + "completions/max_terminated_length": 176.0, + "completions/mean_length": 144.5, + "completions/mean_terminated_length": 144.5, + "completions/min_length": 99.0, + "completions/min_terminated_length": 99.0, + "entropy": 0.5378735661506653, + "epoch": 0.26, + "frac_reward_zero_std": 0.0, + "grad_norm": 5.114461898803711, + "learning_rate": 7.45e-07, + "loss": 0.0036597400903701782, + "num_tokens": 441980.0, + "reward": -15.20625114440918, + "reward_std": 45.35526657104492, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": -15.28125, + "rewards/supergames_reward/std": 45.301815032958984, + "step": 52, + "step_time": 6.660627231001854 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 136.0, + "completions/max_terminated_length": 136.0, + "completions/mean_length": 119.0, + "completions/mean_terminated_length": 119.0, + "completions/min_length": 105.0, + "completions/min_terminated_length": 105.0, + "entropy": 0.41796669363975525, + "epoch": 0.265, + "frac_reward_zero_std": 0.0, + "grad_norm": 5.059505462646484, + "learning_rate": 7.4e-07, + "loss": -0.014973883517086506, + "num_tokens": 447468.0, + "reward": 15.170000076293945, + "reward_std": 46.658329010009766, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 15.070000648498535, + "rewards/supergames_reward/std": 46.658329010009766, + "step": 53, + "step_time": 5.271571868011961 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 259.0, + "completions/max_terminated_length": 259.0, + "completions/mean_length": 191.125, + "completions/mean_terminated_length": 191.125, + "completions/min_length": 91.0, + "completions/min_terminated_length": 91.0, + "entropy": 0.45421987771987915, + "epoch": 0.27, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.4454867839813232, + "learning_rate": 7.35e-07, + "loss": -0.1128973588347435, + "num_tokens": 457301.0, + "reward": 38.89875030517578, + "reward_std": 62.1772575378418, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 38.79874801635742, + "rewards/supergames_reward/std": 62.1772575378418, + "step": 54, + "step_time": 10.524528659996577 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 319.0, + "completions/max_terminated_length": 319.0, + "completions/mean_length": 233.375, + "completions/mean_terminated_length": 233.375, + "completions/min_length": 136.0, + "completions/min_terminated_length": 136.0, + "entropy": 0.4449213445186615, + "epoch": 0.275, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.2065398693084717, + "learning_rate": 7.3e-07, + "loss": 0.0016277075046673417, + "num_tokens": 467520.0, + "reward": 27.372499465942383, + "reward_std": 81.23126983642578, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 27.27250099182129, + "rewards/supergames_reward/std": 81.23126983642578, + "step": 55, + "step_time": 12.592280682991259 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.125, + "completions/max_length": 512.0, + "completions/max_terminated_length": 357.0, + "completions/mean_length": 261.0, + "completions/mean_terminated_length": 225.1428680419922, + "completions/min_length": 95.0, + "completions/min_terminated_length": 95.0, + "entropy": 0.506920337677002, + "epoch": 0.28, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.058388710021973, + "learning_rate": 7.249999999999999e-07, + "loss": 0.3289242386817932, + "num_tokens": 476056.0, + "reward": 47.05249786376953, + "reward_std": 80.97753143310547, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": 46.977500915527344, + "rewards/supergames_reward/std": 80.9256362915039, + "step": 56, + "step_time": 18.845177220006008 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 250.0, + "completions/max_terminated_length": 250.0, + "completions/mean_length": 192.75, + "completions/mean_terminated_length": 192.75, + "completions/min_length": 122.0, + "completions/min_terminated_length": 122.0, + "entropy": 0.5197336077690125, + "epoch": 0.285, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.230815410614014, + "learning_rate": 7.2e-07, + "loss": -0.15186546742916107, + "num_tokens": 484014.0, + "reward": -64.5050048828125, + "reward_std": 50.35145568847656, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -64.60499572753906, + "rewards/supergames_reward/std": 50.3514518737793, + "step": 57, + "step_time": 9.576232638006331 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 324.0, + "completions/max_terminated_length": 324.0, + "completions/mean_length": 249.0, + "completions/mean_terminated_length": 249.0, + "completions/min_length": 129.0, + "completions/min_terminated_length": 129.0, + "entropy": 0.42913514375686646, + "epoch": 0.29, + "frac_reward_zero_std": 0.0, + "grad_norm": 2.8544528484344482, + "learning_rate": 7.149999999999999e-07, + "loss": -0.056405920535326004, + "num_tokens": 494390.0, + "reward": 40.123748779296875, + "reward_std": 84.164794921875, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 40.02375030517578, + "rewards/supergames_reward/std": 84.164794921875, + "step": 58, + "step_time": 12.385353341000155 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 138.0, + "completions/max_terminated_length": 138.0, + "completions/mean_length": 107.25, + "completions/mean_terminated_length": 107.25, + "completions/min_length": 81.0, + "completions/min_terminated_length": 81.0, + "entropy": 0.34386488795280457, + "epoch": 0.295, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.6915154457092285, + "learning_rate": 7.1e-07, + "loss": 0.0865321159362793, + "num_tokens": 499816.0, + "reward": -8.244999885559082, + "reward_std": 4.681046962738037, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -8.345000267028809, + "rewards/supergames_reward/std": 4.681046962738037, + "step": 59, + "step_time": 5.42325339600211 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 187.0, + "completions/max_terminated_length": 187.0, + "completions/mean_length": 127.25, + "completions/mean_terminated_length": 127.25, + "completions/min_length": 97.0, + "completions/min_terminated_length": 97.0, + "entropy": 0.5430045127868652, + "epoch": 0.3, + "frac_reward_zero_std": 0.0, + "grad_norm": 5.635311126708984, + "learning_rate": 7.049999999999999e-07, + "loss": -0.05412375554442406, + "num_tokens": 505410.0, + "reward": 20.005001068115234, + "reward_std": 55.156005859375, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 19.904998779296875, + "rewards/supergames_reward/std": 55.156005859375, + "step": 60, + "step_time": 7.068010902003152 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 286.0, + "completions/max_terminated_length": 286.0, + "completions/mean_length": 192.125, + "completions/mean_terminated_length": 192.125, + "completions/min_length": 95.0, + "completions/min_terminated_length": 95.0, + "entropy": 0.4213869869709015, + "epoch": 0.305, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.4664618968963623, + "learning_rate": 7e-07, + "loss": -0.0495159812271595, + "num_tokens": 516283.0, + "reward": -47.212501525878906, + "reward_std": 91.2253646850586, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -47.3125, + "rewards/supergames_reward/std": 91.2253646850586, + "step": 61, + "step_time": 11.176304235996213 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 324.0, + "completions/max_terminated_length": 324.0, + "completions/mean_length": 171.0, + "completions/mean_terminated_length": 171.0, + "completions/min_length": 92.0, + "completions/min_terminated_length": 92.0, + "entropy": 0.4626811146736145, + "epoch": 0.31, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.949252128601074, + "learning_rate": 6.949999999999999e-07, + "loss": 0.192047581076622, + "num_tokens": 522251.0, + "reward": -15.092500686645508, + "reward_std": 49.82962417602539, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": -15.167499542236328, + "rewards/supergames_reward/std": 49.78091049194336, + "step": 62, + "step_time": 12.075224861997413 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.125, + "completions/max_length": 512.0, + "completions/max_terminated_length": 429.0, + "completions/mean_length": 283.5, + "completions/mean_terminated_length": 250.85714721679688, + "completions/min_length": 85.0, + "completions/min_terminated_length": 85.0, + "entropy": 0.4814216196537018, + "epoch": 0.315, + "frac_reward_zero_std": 0.0, + "grad_norm": 2.7918753623962402, + "learning_rate": 6.9e-07, + "loss": 0.15706156194210052, + "num_tokens": 533815.0, + "reward": -20.86625099182129, + "reward_std": 85.99250030517578, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": -20.94124984741211, + "rewards/supergames_reward/std": 85.96620178222656, + "step": 63, + "step_time": 19.448832260008203 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 189.0, + "completions/max_terminated_length": 189.0, + "completions/mean_length": 131.25, + "completions/mean_terminated_length": 131.25, + "completions/min_length": 106.0, + "completions/min_terminated_length": 106.0, + "entropy": 0.4929414987564087, + "epoch": 0.32, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.774101734161377, + "learning_rate": 6.85e-07, + "loss": -0.060691747814416885, + "num_tokens": 539465.0, + "reward": -21.0674991607666, + "reward_std": 6.892730236053467, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -21.167499542236328, + "rewards/supergames_reward/std": 6.892730236053467, + "step": 64, + "step_time": 7.141569407976931 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 454.0, + "completions/max_terminated_length": 454.0, + "completions/mean_length": 281.375, + "completions/mean_terminated_length": 281.375, + "completions/min_length": 107.0, + "completions/min_terminated_length": 107.0, + "entropy": 0.45866572856903076, + "epoch": 0.325, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.092717409133911, + "learning_rate": 6.800000000000001e-07, + "loss": 0.11694959551095963, + "num_tokens": 551012.0, + "reward": 49.938751220703125, + "reward_std": 79.1865005493164, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 49.8387451171875, + "rewards/supergames_reward/std": 79.1865005493164, + "step": 65, + "step_time": 17.06170882002334 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 203.0, + "completions/max_terminated_length": 203.0, + "completions/mean_length": 128.625, + "completions/mean_terminated_length": 128.625, + "completions/min_length": 88.0, + "completions/min_terminated_length": 88.0, + "entropy": 0.5005782246589661, + "epoch": 0.33, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.77616024017334, + "learning_rate": 6.75e-07, + "loss": 0.11242837458848953, + "num_tokens": 556577.0, + "reward": 48.875, + "reward_std": 49.13539123535156, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 48.775001525878906, + "rewards/supergames_reward/std": 49.13539123535156, + "step": 66, + "step_time": 7.695410129002994 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 436.0, + "completions/max_terminated_length": 436.0, + "completions/mean_length": 281.875, + "completions/mean_terminated_length": 281.875, + "completions/min_length": 133.0, + "completions/min_terminated_length": 133.0, + "entropy": 0.507097065448761, + "epoch": 0.335, + "frac_reward_zero_std": 0.0, + "grad_norm": 2.911184787750244, + "learning_rate": 6.7e-07, + "loss": -0.08151137083768845, + "num_tokens": 567120.0, + "reward": -12.886249542236328, + "reward_std": 84.04269409179688, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -12.986251831054688, + "rewards/supergames_reward/std": 84.04269409179688, + "step": 67, + "step_time": 16.409127942984924 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.125, + "completions/max_length": 512.0, + "completions/max_terminated_length": 501.0, + "completions/mean_length": 318.375, + "completions/mean_terminated_length": 290.71429443359375, + "completions/min_length": 65.0, + "completions/min_terminated_length": 65.0, + "entropy": 0.486402690410614, + "epoch": 0.34, + "frac_reward_zero_std": 0.0, + "grad_norm": 2.9461607933044434, + "learning_rate": 6.65e-07, + "loss": 0.2442178875207901, + "num_tokens": 578995.0, + "reward": 2.8387489318847656, + "reward_std": 98.47579956054688, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": 2.7637500762939453, + "rewards/supergames_reward/std": 98.44595336914062, + "step": 68, + "step_time": 19.32364999302081 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 291.0, + "completions/max_terminated_length": 291.0, + "completions/mean_length": 219.625, + "completions/mean_terminated_length": 219.625, + "completions/min_length": 117.0, + "completions/min_terminated_length": 117.0, + "entropy": 0.37856727838516235, + "epoch": 0.345, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.0204732418060303, + "learning_rate": 6.6e-07, + "loss": -0.08764916658401489, + "num_tokens": 589072.0, + "reward": -54.18375015258789, + "reward_std": 71.45012664794922, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -54.283748626708984, + "rewards/supergames_reward/std": 71.45013427734375, + "step": 69, + "step_time": 11.126611230982235 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.125, + "completions/max_length": 512.0, + "completions/max_terminated_length": 305.0, + "completions/mean_length": 254.75, + "completions/mean_terminated_length": 218.00001525878906, + "completions/min_length": 104.0, + "completions/min_terminated_length": 104.0, + "entropy": 0.46423614025115967, + "epoch": 0.35, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.0140154361724854, + "learning_rate": 6.55e-07, + "loss": 0.26797160506248474, + "num_tokens": 600438.0, + "reward": -32.04875183105469, + "reward_std": 69.86957550048828, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": -32.12375259399414, + "rewards/supergames_reward/std": 69.84178161621094, + "step": 70, + "step_time": 19.12617294798838 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 372.0, + "completions/max_terminated_length": 372.0, + "completions/mean_length": 194.125, + "completions/mean_terminated_length": 194.125, + "completions/min_length": 94.0, + "completions/min_terminated_length": 94.0, + "entropy": 0.3427293300628662, + "epoch": 0.355, + "frac_reward_zero_std": 0.0, + "grad_norm": 2.781383991241455, + "learning_rate": 6.5e-07, + "loss": 0.354524701833725, + "num_tokens": 610327.0, + "reward": -6.913749694824219, + "reward_std": 87.56071472167969, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": -6.988750457763672, + "rewards/supergames_reward/std": 87.53033447265625, + "step": 71, + "step_time": 14.041668199002743 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 197.0, + "completions/max_terminated_length": 197.0, + "completions/mean_length": 124.75, + "completions/mean_terminated_length": 124.75, + "completions/min_length": 93.0, + "completions/min_terminated_length": 93.0, + "entropy": 0.4455118477344513, + "epoch": 0.36, + "frac_reward_zero_std": 0.0, + "grad_norm": 5.726032733917236, + "learning_rate": 6.45e-07, + "loss": 0.18206289410591125, + "num_tokens": 615853.0, + "reward": 41.95124816894531, + "reward_std": 60.640811920166016, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 41.85124969482422, + "rewards/supergames_reward/std": 60.640811920166016, + "step": 72, + "step_time": 7.416493425989756 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 323.0, + "completions/max_terminated_length": 323.0, + "completions/mean_length": 147.375, + "completions/mean_terminated_length": 147.375, + "completions/min_length": 94.0, + "completions/min_terminated_length": 94.0, + "entropy": 0.480733186006546, + "epoch": 0.365, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.853621482849121, + "learning_rate": 6.4e-07, + "loss": 0.04558124020695686, + "num_tokens": 621584.0, + "reward": 14.884998321533203, + "reward_std": 46.25857162475586, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 14.78499984741211, + "rewards/supergames_reward/std": 46.25857162475586, + "step": 73, + "step_time": 11.95379540900467 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 283.0, + "completions/max_terminated_length": 283.0, + "completions/mean_length": 188.125, + "completions/mean_terminated_length": 188.125, + "completions/min_length": 116.0, + "completions/min_terminated_length": 116.0, + "entropy": 0.5526013374328613, + "epoch": 0.37, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.729883909225464, + "learning_rate": 6.35e-07, + "loss": 0.06268303096294403, + "num_tokens": 629545.0, + "reward": -68.05750274658203, + "reward_std": 59.44808578491211, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -68.15750122070312, + "rewards/supergames_reward/std": 59.44808578491211, + "step": 74, + "step_time": 10.64124580900534 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 141.0, + "completions/max_terminated_length": 141.0, + "completions/mean_length": 117.375, + "completions/mean_terminated_length": 117.375, + "completions/min_length": 100.0, + "completions/min_terminated_length": 100.0, + "entropy": 0.3744319677352905, + "epoch": 0.375, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.775154113769531, + "learning_rate": 6.3e-07, + "loss": -0.023284845054149628, + "num_tokens": 635052.0, + "reward": 3.9374990463256836, + "reward_std": 61.85771179199219, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": 3.862499237060547, + "rewards/supergames_reward/std": 61.80967712402344, + "step": 75, + "step_time": 5.432648951013107 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 268.0, + "completions/max_terminated_length": 268.0, + "completions/mean_length": 157.0, + "completions/mean_terminated_length": 157.0, + "completions/min_length": 104.0, + "completions/min_terminated_length": 104.0, + "entropy": 0.4822140336036682, + "epoch": 0.38, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.715704917907715, + "learning_rate": 6.249999999999999e-07, + "loss": -0.04918142035603523, + "num_tokens": 644644.0, + "reward": 48.64125061035156, + "reward_std": 78.0205078125, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": 48.56624984741211, + "rewards/supergames_reward/std": 77.96605682373047, + "step": 76, + "step_time": 10.348935816989979 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 380.0, + "completions/max_terminated_length": 380.0, + "completions/mean_length": 210.875, + "completions/mean_terminated_length": 210.875, + "completions/min_length": 126.0, + "completions/min_terminated_length": 126.0, + "entropy": 0.45722317695617676, + "epoch": 0.385, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.6057567596435547, + "learning_rate": 6.2e-07, + "loss": 0.058352649211883545, + "num_tokens": 652771.0, + "reward": -8.283750534057617, + "reward_std": 77.68436431884766, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -8.383749008178711, + "rewards/supergames_reward/std": 77.68437194824219, + "step": 77, + "step_time": 14.142948755004909 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 324.0, + "completions/max_terminated_length": 324.0, + "completions/mean_length": 188.25, + "completions/mean_terminated_length": 188.25, + "completions/min_length": 68.0, + "completions/min_terminated_length": 68.0, + "entropy": 0.33380764722824097, + "epoch": 0.39, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.2239863872528076, + "learning_rate": 6.149999999999999e-07, + "loss": -0.04592633992433548, + "num_tokens": 662653.0, + "reward": 11.153749465942383, + "reward_std": 89.88329315185547, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 11.053749084472656, + "rewards/supergames_reward/std": 89.88329315185547, + "step": 78, + "step_time": 12.277474621019792 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 143.0, + "completions/max_terminated_length": 143.0, + "completions/mean_length": 108.875, + "completions/mean_terminated_length": 108.875, + "completions/min_length": 85.0, + "completions/min_terminated_length": 85.0, + "entropy": 0.3518812656402588, + "epoch": 0.395, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.967776298522949, + "learning_rate": 6.1e-07, + "loss": -0.10800496488809586, + "num_tokens": 668076.0, + "reward": 70.33250427246094, + "reward_std": 49.9145393371582, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 70.23249816894531, + "rewards/supergames_reward/std": 49.91453552246094, + "step": 79, + "step_time": 5.52714040101273 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 184.0, + "completions/max_terminated_length": 184.0, + "completions/mean_length": 152.125, + "completions/mean_terminated_length": 152.125, + "completions/min_length": 104.0, + "completions/min_terminated_length": 104.0, + "entropy": 0.4699896574020386, + "epoch": 0.4, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.561161994934082, + "learning_rate": 6.049999999999999e-07, + "loss": 0.09505834430456161, + "num_tokens": 675701.0, + "reward": -20.25625228881836, + "reward_std": 53.78542709350586, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -20.35624885559082, + "rewards/supergames_reward/std": 53.78542709350586, + "step": 80, + "step_time": 7.155081019998761 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.125, + "completions/max_length": 512.0, + "completions/max_terminated_length": 506.0, + "completions/mean_length": 288.875, + "completions/mean_terminated_length": 257.0, + "completions/min_length": 115.0, + "completions/min_terminated_length": 115.0, + "entropy": 0.4086494445800781, + "epoch": 0.405, + "frac_reward_zero_std": 0.0, + "grad_norm": 2.3719611167907715, + "learning_rate": 6e-07, + "loss": 0.04727627709507942, + "num_tokens": 687332.0, + "reward": -37.20750045776367, + "reward_std": 91.7676010131836, + "rewards/json_format_reward/mean": 0.05000000074505806, + "rewards/json_format_reward/std": 0.09258200973272324, + "rewards/supergames_reward/mean": -37.25749969482422, + "rewards/supergames_reward/std": 91.72847747802734, + "step": 81, + "step_time": 19.231399144016905 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 376.0, + "completions/max_terminated_length": 376.0, + "completions/mean_length": 227.75, + "completions/mean_terminated_length": 227.75, + "completions/min_length": 39.0, + "completions/min_terminated_length": 39.0, + "entropy": 0.43233707547187805, + "epoch": 0.41, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.3391714096069336, + "learning_rate": 5.949999999999999e-07, + "loss": -0.25524550676345825, + "num_tokens": 698474.0, + "reward": -3.7287511825561523, + "reward_std": 91.25379943847656, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -3.828749656677246, + "rewards/supergames_reward/std": 91.25379943847656, + "step": 82, + "step_time": 14.329176368017215 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 299.0, + "completions/max_terminated_length": 299.0, + "completions/mean_length": 178.125, + "completions/mean_terminated_length": 178.125, + "completions/min_length": 91.0, + "completions/min_terminated_length": 91.0, + "entropy": 0.4293636977672577, + "epoch": 0.415, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.7271742820739746, + "learning_rate": 5.9e-07, + "loss": 0.09273672848939896, + "num_tokens": 708243.0, + "reward": 12.802498817443848, + "reward_std": 86.23526000976562, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 12.702500343322754, + "rewards/supergames_reward/std": 86.23526000976562, + "step": 83, + "step_time": 11.419686011999147 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 425.0, + "completions/max_terminated_length": 425.0, + "completions/mean_length": 255.375, + "completions/mean_terminated_length": 255.375, + "completions/min_length": 172.0, + "completions/min_terminated_length": 172.0, + "entropy": 0.38068246841430664, + "epoch": 0.42, + "frac_reward_zero_std": 0.0, + "grad_norm": 2.6244184970855713, + "learning_rate": 5.849999999999999e-07, + "loss": 0.02006213366985321, + "num_tokens": 718606.0, + "reward": 37.803749084472656, + "reward_std": 57.58624267578125, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 37.70375061035156, + "rewards/supergames_reward/std": 57.58624267578125, + "step": 84, + "step_time": 15.975198492989875 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 116.0, + "completions/max_terminated_length": 116.0, + "completions/mean_length": 108.125, + "completions/mean_terminated_length": 108.125, + "completions/min_length": 87.0, + "completions/min_terminated_length": 87.0, + "entropy": 0.339572012424469, + "epoch": 0.425, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.754759311676025, + "learning_rate": 5.8e-07, + "loss": -0.011015941388905048, + "num_tokens": 723983.0, + "reward": -3.8212504386901855, + "reward_std": 61.624786376953125, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -3.921250820159912, + "rewards/supergames_reward/std": 61.624786376953125, + "step": 85, + "step_time": 4.542777584982105 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 447.0, + "completions/max_terminated_length": 447.0, + "completions/mean_length": 291.375, + "completions/mean_terminated_length": 291.375, + "completions/min_length": 163.0, + "completions/min_terminated_length": 163.0, + "entropy": 0.5846173763275146, + "epoch": 0.43, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.55724835395813, + "learning_rate": 5.749999999999999e-07, + "loss": -0.04981120675802231, + "num_tokens": 735642.0, + "reward": 50.07499694824219, + "reward_std": 92.6283187866211, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": 50.0, + "rewards/supergames_reward/std": 92.5820083618164, + "step": 86, + "step_time": 16.777178087009815 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 232.0, + "completions/max_terminated_length": 232.0, + "completions/mean_length": 192.0, + "completions/mean_terminated_length": 192.0, + "completions/min_length": 139.0, + "completions/min_terminated_length": 139.0, + "entropy": 0.5139474272727966, + "epoch": 0.435, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.8860411643981934, + "learning_rate": 5.699999999999999e-07, + "loss": -0.02194221317768097, + "num_tokens": 743602.0, + "reward": -52.522499084472656, + "reward_std": 73.8686294555664, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -52.62249755859375, + "rewards/supergames_reward/std": 73.86863708496094, + "step": 87, + "step_time": 8.772893911984283 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 266.0, + "completions/max_terminated_length": 266.0, + "completions/mean_length": 146.875, + "completions/mean_terminated_length": 146.875, + "completions/min_length": 89.0, + "completions/min_terminated_length": 89.0, + "entropy": 0.3428497910499573, + "epoch": 0.44, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.509483337402344, + "learning_rate": 5.649999999999999e-07, + "loss": 0.19857533276081085, + "num_tokens": 753065.0, + "reward": 63.63874816894531, + "reward_std": 67.54963684082031, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 63.53874969482422, + "rewards/supergames_reward/std": 67.54963684082031, + "step": 88, + "step_time": 10.230529835011112 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 446.0, + "completions/max_terminated_length": 446.0, + "completions/mean_length": 270.625, + "completions/mean_terminated_length": 270.625, + "completions/min_length": 103.0, + "completions/min_terminated_length": 103.0, + "entropy": 0.4417419731616974, + "epoch": 0.445, + "frac_reward_zero_std": 0.0, + "grad_norm": 2.973951816558838, + "learning_rate": 5.6e-07, + "loss": -0.00031440958264283836, + "num_tokens": 764534.0, + "reward": -9.736251831054688, + "reward_std": 92.25879669189453, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -9.836250305175781, + "rewards/supergames_reward/std": 92.25880432128906, + "step": 89, + "step_time": 17.40752330099349 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 398.0, + "completions/max_terminated_length": 398.0, + "completions/mean_length": 227.375, + "completions/mean_terminated_length": 227.375, + "completions/min_length": 98.0, + "completions/min_terminated_length": 98.0, + "entropy": 0.4732024073600769, + "epoch": 0.45, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.2560815811157227, + "learning_rate": 5.55e-07, + "loss": 0.03820464387536049, + "num_tokens": 772809.0, + "reward": 5.982499599456787, + "reward_std": 58.833797454833984, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 5.8824992179870605, + "rewards/supergames_reward/std": 58.83380126953125, + "step": 90, + "step_time": 15.007269965979503 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 269.0, + "completions/max_terminated_length": 269.0, + "completions/mean_length": 207.5, + "completions/mean_terminated_length": 207.5, + "completions/min_length": 92.0, + "completions/min_terminated_length": 92.0, + "entropy": 0.42038559913635254, + "epoch": 0.455, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.3649117946624756, + "learning_rate": 5.5e-07, + "loss": 0.1510048508644104, + "num_tokens": 782805.0, + "reward": -5.61500358581543, + "reward_std": 82.67152404785156, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -5.714998245239258, + "rewards/supergames_reward/std": 82.67152404785156, + "step": 91, + "step_time": 10.386137172987219 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 416.0, + "completions/max_terminated_length": 416.0, + "completions/mean_length": 218.375, + "completions/mean_terminated_length": 218.375, + "completions/min_length": 95.0, + "completions/min_terminated_length": 95.0, + "entropy": 0.3971734642982483, + "epoch": 0.46, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.433772563934326, + "learning_rate": 5.45e-07, + "loss": -0.3027213513851166, + "num_tokens": 793840.0, + "reward": -29.06500244140625, + "reward_std": 91.68325805664062, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -29.16499900817871, + "rewards/supergames_reward/std": 91.68325805664062, + "step": 92, + "step_time": 15.790611553995404 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 345.0, + "completions/max_terminated_length": 345.0, + "completions/mean_length": 206.625, + "completions/mean_terminated_length": 206.625, + "completions/min_length": 90.0, + "completions/min_terminated_length": 90.0, + "entropy": 0.5906907320022583, + "epoch": 0.465, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.293088912963867, + "learning_rate": 5.4e-07, + "loss": 0.09760032594203949, + "num_tokens": 803821.0, + "reward": -72.73999786376953, + "reward_std": 50.72484588623047, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": -72.81500244140625, + "rewards/supergames_reward/std": 50.7094841003418, + "step": 93, + "step_time": 13.043119941983605 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 143.0, + "completions/max_terminated_length": 143.0, + "completions/mean_length": 115.75, + "completions/mean_terminated_length": 115.75, + "completions/min_length": 80.0, + "completions/min_terminated_length": 80.0, + "entropy": 0.4233115315437317, + "epoch": 0.47, + "frac_reward_zero_std": 0.0, + "grad_norm": 5.58886194229126, + "learning_rate": 5.35e-07, + "loss": -0.04040650278329849, + "num_tokens": 809299.0, + "reward": -3.3050003051757812, + "reward_std": 24.443714141845703, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -3.405000686645508, + "rewards/supergames_reward/std": 24.443714141845703, + "step": 94, + "step_time": 5.512874762003776 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 122.0, + "completions/max_terminated_length": 122.0, + "completions/mean_length": 103.125, + "completions/mean_terminated_length": 103.125, + "completions/min_length": 86.0, + "completions/min_terminated_length": 86.0, + "entropy": 0.4055764377117157, + "epoch": 0.475, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.981549263000488, + "learning_rate": 5.3e-07, + "loss": -0.07588938623666763, + "num_tokens": 814716.0, + "reward": 60.06999969482422, + "reward_std": 42.708473205566406, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 59.970001220703125, + "rewards/supergames_reward/std": 42.708473205566406, + "step": 95, + "step_time": 4.766389057011111 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 475.0, + "completions/max_terminated_length": 475.0, + "completions/mean_length": 218.375, + "completions/mean_terminated_length": 218.375, + "completions/min_length": 81.0, + "completions/min_terminated_length": 81.0, + "entropy": 0.408913254737854, + "epoch": 0.48, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.79402756690979, + "learning_rate": 5.25e-07, + "loss": -0.3087541460990906, + "num_tokens": 824791.0, + "reward": -15.042500495910645, + "reward_std": 84.37246704101562, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -15.142499923706055, + "rewards/supergames_reward/std": 84.37246704101562, + "step": 96, + "step_time": 17.779843941010768 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 154.0, + "completions/max_terminated_length": 154.0, + "completions/mean_length": 121.625, + "completions/mean_terminated_length": 121.625, + "completions/min_length": 80.0, + "completions/min_terminated_length": 80.0, + "entropy": 0.529015839099884, + "epoch": 0.485, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.142816066741943, + "learning_rate": 5.2e-07, + "loss": 0.07216037809848785, + "num_tokens": 830300.0, + "reward": -32.743751525878906, + "reward_std": 54.61775588989258, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": -32.81875228881836, + "rewards/supergames_reward/std": 54.58256149291992, + "step": 97, + "step_time": 5.913989890017547 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 126.0, + "completions/max_terminated_length": 126.0, + "completions/mean_length": 95.625, + "completions/mean_terminated_length": 95.625, + "completions/min_length": 74.0, + "completions/min_terminated_length": 74.0, + "entropy": 0.33743274211883545, + "epoch": 0.49, + "frac_reward_zero_std": 0.0, + "grad_norm": 5.197273254394531, + "learning_rate": 5.149999999999999e-07, + "loss": 0.07500407844781876, + "num_tokens": 835633.0, + "reward": 11.00999927520752, + "reward_std": 62.73848342895508, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 10.910000801086426, + "rewards/supergames_reward/std": 62.73848342895508, + "step": 98, + "step_time": 4.9155233050114475 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 149.0, + "completions/max_terminated_length": 149.0, + "completions/mean_length": 108.5, + "completions/mean_terminated_length": 108.5, + "completions/min_length": 86.0, + "completions/min_terminated_length": 86.0, + "entropy": 0.3106803297996521, + "epoch": 0.495, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.864002227783203, + "learning_rate": 5.1e-07, + "loss": -0.1206185445189476, + "num_tokens": 841045.0, + "reward": 14.238749504089355, + "reward_std": 51.77016830444336, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 14.138750076293945, + "rewards/supergames_reward/std": 51.77016830444336, + "step": 99, + "step_time": 5.7507751359953545 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 149.0, + "completions/max_terminated_length": 149.0, + "completions/mean_length": 97.125, + "completions/mean_terminated_length": 97.125, + "completions/min_length": 73.0, + "completions/min_terminated_length": 73.0, + "entropy": 0.35160186886787415, + "epoch": 0.5, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.915731906890869, + "learning_rate": 5.049999999999999e-07, + "loss": -0.07951541244983673, + "num_tokens": 846398.0, + "reward": 67.78125, + "reward_std": 38.15436553955078, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 67.68124389648438, + "rewards/supergames_reward/std": 38.15436553955078, + "step": 100, + "step_time": 5.994720440998208 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 320.0, + "completions/max_terminated_length": 320.0, + "completions/mean_length": 208.75, + "completions/mean_terminated_length": 208.75, + "completions/min_length": 95.0, + "completions/min_terminated_length": 95.0, + "entropy": 0.452095091342926, + "epoch": 0.505, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.065176963806152, + "learning_rate": 5e-07, + "loss": -0.1397111564874649, + "num_tokens": 857420.0, + "reward": -5.801251411437988, + "reward_std": 73.56117248535156, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -5.90125036239624, + "rewards/supergames_reward/std": 73.5611801147461, + "step": 101, + "step_time": 12.28178753197426 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 230.0, + "completions/max_terminated_length": 230.0, + "completions/mean_length": 171.375, + "completions/mean_terminated_length": 171.375, + "completions/min_length": 95.0, + "completions/min_terminated_length": 95.0, + "entropy": 0.4138447642326355, + "epoch": 0.51, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.998823404312134, + "learning_rate": 4.95e-07, + "loss": -0.08941879868507385, + "num_tokens": 867103.0, + "reward": 92.24749755859375, + "reward_std": 14.540005683898926, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 92.14749908447266, + "rewards/supergames_reward/std": 14.540006637573242, + "step": 102, + "step_time": 8.992682139010867 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 311.0, + "completions/max_terminated_length": 311.0, + "completions/mean_length": 163.0, + "completions/mean_terminated_length": 163.0, + "completions/min_length": 74.0, + "completions/min_terminated_length": 74.0, + "entropy": 0.4182547330856323, + "epoch": 0.515, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.419595241546631, + "learning_rate": 4.9e-07, + "loss": -0.19304415583610535, + "num_tokens": 876679.0, + "reward": 79.13249969482422, + "reward_std": 59.3050422668457, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 79.03250122070312, + "rewards/supergames_reward/std": 59.30504608154297, + "step": 103, + "step_time": 11.836970244999975 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 253.0, + "completions/max_terminated_length": 253.0, + "completions/mean_length": 155.625, + "completions/mean_terminated_length": 155.625, + "completions/min_length": 94.0, + "completions/min_terminated_length": 94.0, + "entropy": 0.49889081716537476, + "epoch": 0.52, + "frac_reward_zero_std": 0.0, + "grad_norm": 5.463719367980957, + "learning_rate": 4.85e-07, + "loss": -0.04786720871925354, + "num_tokens": 884348.0, + "reward": 48.616249084472656, + "reward_std": 73.30741119384766, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 48.51625061035156, + "rewards/supergames_reward/std": 73.30741119384766, + "step": 104, + "step_time": 9.579507841990562 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 380.0, + "completions/max_terminated_length": 380.0, + "completions/mean_length": 269.5, + "completions/mean_terminated_length": 269.5, + "completions/min_length": 101.0, + "completions/min_terminated_length": 101.0, + "entropy": 0.5286482572555542, + "epoch": 0.525, + "frac_reward_zero_std": 0.0, + "grad_norm": 2.993295431137085, + "learning_rate": 4.8e-07, + "loss": -0.08071611076593399, + "num_tokens": 894856.0, + "reward": 34.970001220703125, + "reward_std": 85.324462890625, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 34.869998931884766, + "rewards/supergames_reward/std": 85.324462890625, + "step": 105, + "step_time": 14.295730627985904 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 261.0, + "completions/max_terminated_length": 261.0, + "completions/mean_length": 203.25, + "completions/mean_terminated_length": 203.25, + "completions/min_length": 91.0, + "completions/min_terminated_length": 91.0, + "entropy": 0.6049264073371887, + "epoch": 0.53, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.787306785583496, + "learning_rate": 4.7499999999999995e-07, + "loss": 0.19928883016109467, + "num_tokens": 902922.0, + "reward": 32.44499969482422, + "reward_std": 70.4663314819336, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": 32.369998931884766, + "rewards/supergames_reward/std": 70.41261291503906, + "step": 106, + "step_time": 9.84894504098338 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 279.0, + "completions/max_terminated_length": 279.0, + "completions/mean_length": 194.0, + "completions/mean_terminated_length": 194.0, + "completions/min_length": 109.0, + "completions/min_terminated_length": 109.0, + "entropy": 0.5115755200386047, + "epoch": 0.535, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.277207374572754, + "learning_rate": 4.6999999999999995e-07, + "loss": 0.09118300676345825, + "num_tokens": 910906.0, + "reward": 40.432498931884766, + "reward_std": 48.54548263549805, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 40.33250045776367, + "rewards/supergames_reward/std": 48.54548263549805, + "step": 107, + "step_time": 10.466603949986165 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 332.0, + "completions/max_terminated_length": 332.0, + "completions/mean_length": 195.875, + "completions/mean_terminated_length": 195.875, + "completions/min_length": 87.0, + "completions/min_terminated_length": 87.0, + "entropy": 0.4192962944507599, + "epoch": 0.54, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.308061122894287, + "learning_rate": 4.65e-07, + "loss": 0.28064602613449097, + "num_tokens": 920745.0, + "reward": -21.982500076293945, + "reward_std": 66.31195068359375, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -22.08249855041504, + "rewards/supergames_reward/std": 66.31195068359375, + "step": 108, + "step_time": 12.595848233992001 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 269.0, + "completions/max_terminated_length": 269.0, + "completions/mean_length": 195.125, + "completions/mean_terminated_length": 195.125, + "completions/min_length": 111.0, + "completions/min_terminated_length": 111.0, + "entropy": 0.37523961067199707, + "epoch": 0.545, + "frac_reward_zero_std": 0.0, + "grad_norm": 2.802706718444824, + "learning_rate": 4.6e-07, + "loss": -0.06614465266466141, + "num_tokens": 930618.0, + "reward": 71.75, + "reward_std": 51.14524459838867, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 71.6500015258789, + "rewards/supergames_reward/std": 51.1452522277832, + "step": 109, + "step_time": 10.33884758799104 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 331.0, + "completions/max_terminated_length": 331.0, + "completions/mean_length": 167.75, + "completions/mean_terminated_length": 167.75, + "completions/min_length": 89.0, + "completions/min_terminated_length": 89.0, + "entropy": 0.4339042007923126, + "epoch": 0.55, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.254003047943115, + "learning_rate": 4.55e-07, + "loss": 0.34405598044395447, + "num_tokens": 941256.0, + "reward": 99.21875, + "reward_std": 2.492553234100342, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 99.1187515258789, + "rewards/supergames_reward/std": 2.4925525188446045, + "step": 110, + "step_time": 12.657525141985388 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 285.0, + "completions/max_terminated_length": 285.0, + "completions/mean_length": 211.0, + "completions/mean_terminated_length": 211.0, + "completions/min_length": 112.0, + "completions/min_terminated_length": 112.0, + "entropy": 0.46809861063957214, + "epoch": 0.555, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.811166286468506, + "learning_rate": 4.5e-07, + "loss": 0.045113153755664825, + "num_tokens": 951296.0, + "reward": 55.27750015258789, + "reward_std": 47.90687561035156, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 55.1775016784668, + "rewards/supergames_reward/std": 47.9068717956543, + "step": 111, + "step_time": 10.91118389699841 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 145.0, + "completions/max_terminated_length": 145.0, + "completions/mean_length": 108.375, + "completions/mean_terminated_length": 108.375, + "completions/min_length": 86.0, + "completions/min_terminated_length": 86.0, + "entropy": 0.34181955456733704, + "epoch": 0.56, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.537296772003174, + "learning_rate": 4.45e-07, + "loss": -0.052413541823625565, + "num_tokens": 956675.0, + "reward": 47.56624984741211, + "reward_std": 49.185157775878906, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 47.466251373291016, + "rewards/supergames_reward/std": 49.18516159057617, + "step": 112, + "step_time": 5.552288047998445 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 245.0, + "completions/max_terminated_length": 245.0, + "completions/mean_length": 189.125, + "completions/mean_terminated_length": 189.125, + "completions/min_length": 102.0, + "completions/min_terminated_length": 102.0, + "entropy": 0.5034605860710144, + "epoch": 0.565, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.349876642227173, + "learning_rate": 4.3999999999999997e-07, + "loss": -0.08186715841293335, + "num_tokens": 964636.0, + "reward": 81.32374572753906, + "reward_std": 36.85074996948242, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 81.2237548828125, + "rewards/supergames_reward/std": 36.85074996948242, + "step": 113, + "step_time": 9.267611294984818 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 178.0, + "completions/max_terminated_length": 178.0, + "completions/mean_length": 111.25, + "completions/mean_terminated_length": 111.25, + "completions/min_length": 67.0, + "completions/min_terminated_length": 67.0, + "entropy": 0.3229002058506012, + "epoch": 0.57, + "frac_reward_zero_std": 0.0, + "grad_norm": 5.245575428009033, + "learning_rate": 4.3499999999999996e-07, + "loss": 0.06544198095798492, + "num_tokens": 970070.0, + "reward": 63.1775016784668, + "reward_std": 52.30624008178711, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 63.07749938964844, + "rewards/supergames_reward/std": 52.30624008178711, + "step": 114, + "step_time": 6.731139309995342 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 224.0, + "completions/max_terminated_length": 224.0, + "completions/mean_length": 170.0, + "completions/mean_terminated_length": 170.0, + "completions/min_length": 94.0, + "completions/min_terminated_length": 94.0, + "entropy": 0.3274807929992676, + "epoch": 0.575, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.5544140338897705, + "learning_rate": 4.2999999999999996e-07, + "loss": -0.06953569501638412, + "num_tokens": 979702.0, + "reward": 25.62125015258789, + "reward_std": 64.52003479003906, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 25.521244049072266, + "rewards/supergames_reward/std": 64.52003479003906, + "step": 115, + "step_time": 8.702618776995223 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 427.0, + "completions/max_terminated_length": 427.0, + "completions/mean_length": 209.5, + "completions/mean_terminated_length": 209.5, + "completions/min_length": 100.0, + "completions/min_terminated_length": 100.0, + "entropy": 0.4970880150794983, + "epoch": 0.58, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.4932165145874023, + "learning_rate": 4.2499999999999995e-07, + "loss": -0.2281273603439331, + "num_tokens": 990690.0, + "reward": 50.54750061035156, + "reward_std": 70.5126953125, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 50.4474983215332, + "rewards/supergames_reward/std": 70.51270294189453, + "step": 116, + "step_time": 16.204386608995264 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 228.0, + "completions/max_terminated_length": 228.0, + "completions/mean_length": 132.375, + "completions/mean_terminated_length": 132.375, + "completions/min_length": 85.0, + "completions/min_terminated_length": 85.0, + "entropy": 0.4380227029323578, + "epoch": 0.585, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.089909076690674, + "learning_rate": 4.1999999999999995e-07, + "loss": 0.12121254950761795, + "num_tokens": 996333.0, + "reward": -15.333749771118164, + "reward_std": 47.200164794921875, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -15.433748245239258, + "rewards/supergames_reward/std": 47.200164794921875, + "step": 117, + "step_time": 8.546233230998041 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 323.0, + "completions/max_terminated_length": 323.0, + "completions/mean_length": 137.625, + "completions/mean_terminated_length": 137.625, + "completions/min_length": 90.0, + "completions/min_terminated_length": 90.0, + "entropy": 0.4177803695201874, + "epoch": 0.59, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.8712472915649414, + "learning_rate": 4.1499999999999994e-07, + "loss": 0.14192955195903778, + "num_tokens": 1002010.0, + "reward": -30.700000762939453, + "reward_std": 17.224035263061523, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -30.799999237060547, + "rewards/supergames_reward/std": 17.224035263061523, + "step": 118, + "step_time": 11.878434990998358 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 453.0, + "completions/max_terminated_length": 453.0, + "completions/mean_length": 238.125, + "completions/mean_terminated_length": 238.125, + "completions/min_length": 109.0, + "completions/min_terminated_length": 109.0, + "entropy": 0.5640217065811157, + "epoch": 0.595, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.064737319946289, + "learning_rate": 4.0999999999999994e-07, + "loss": -0.023621462285518646, + "num_tokens": 1012259.0, + "reward": 43.683746337890625, + "reward_std": 70.4210205078125, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 43.5837516784668, + "rewards/supergames_reward/std": 70.4210205078125, + "step": 119, + "step_time": 16.829514264973113 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 236.0, + "completions/max_terminated_length": 236.0, + "completions/mean_length": 193.75, + "completions/mean_terminated_length": 193.75, + "completions/min_length": 151.0, + "completions/min_terminated_length": 151.0, + "entropy": 0.5795091390609741, + "epoch": 0.6, + "frac_reward_zero_std": 0.0, + "grad_norm": 5.423007965087891, + "learning_rate": 4.05e-07, + "loss": -0.0666111558675766, + "num_tokens": 1020241.0, + "reward": -85.86000061035156, + "reward_std": 27.4442081451416, + "rewards/json_format_reward/mean": 0.05000000074505806, + "rewards/json_format_reward/std": 0.09258200973272324, + "rewards/supergames_reward/mean": -85.90999603271484, + "rewards/supergames_reward/std": 27.41469955444336, + "step": 120, + "step_time": 8.95496519000153 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 331.0, + "completions/max_terminated_length": 331.0, + "completions/mean_length": 231.0, + "completions/mean_terminated_length": 231.0, + "completions/min_length": 165.0, + "completions/min_terminated_length": 165.0, + "entropy": 0.3767699897289276, + "epoch": 0.605, + "frac_reward_zero_std": 0.0, + "grad_norm": 2.7434473037719727, + "learning_rate": 4e-07, + "loss": -0.1165132150053978, + "num_tokens": 1031465.0, + "reward": 61.45624923706055, + "reward_std": 72.21080780029297, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 61.35625076293945, + "rewards/supergames_reward/std": 72.2108154296875, + "step": 121, + "step_time": 12.588820598000893 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 297.0, + "completions/max_terminated_length": 297.0, + "completions/mean_length": 223.375, + "completions/mean_terminated_length": 223.375, + "completions/min_length": 112.0, + "completions/min_terminated_length": 112.0, + "entropy": 0.4537277817726135, + "epoch": 0.61, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.4382436275482178, + "learning_rate": 3.95e-07, + "loss": -0.1038316935300827, + "num_tokens": 1042548.0, + "reward": 5.658749580383301, + "reward_std": 83.77420043945312, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 5.558750152587891, + "rewards/supergames_reward/std": 83.77420043945312, + "step": 122, + "step_time": 11.39525846898323 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 456.0, + "completions/max_terminated_length": 456.0, + "completions/mean_length": 314.25, + "completions/mean_terminated_length": 314.25, + "completions/min_length": 95.0, + "completions/min_terminated_length": 95.0, + "entropy": 0.4113953411579132, + "epoch": 0.615, + "frac_reward_zero_std": 0.0, + "grad_norm": 2.449777841567993, + "learning_rate": 3.8999999999999997e-07, + "loss": -0.1303664743900299, + "num_tokens": 1054422.0, + "reward": 35.98249816894531, + "reward_std": 88.6131362915039, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 35.88249969482422, + "rewards/supergames_reward/std": 88.6131362915039, + "step": 123, + "step_time": 17.001118954998674 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 242.0, + "completions/max_terminated_length": 242.0, + "completions/mean_length": 172.5, + "completions/mean_terminated_length": 172.5, + "completions/min_length": 137.0, + "completions/min_terminated_length": 137.0, + "entropy": 0.4670729339122772, + "epoch": 0.62, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.951946973800659, + "learning_rate": 3.8499999999999997e-07, + "loss": 0.05888795852661133, + "num_tokens": 1062210.0, + "reward": 43.368751525878906, + "reward_std": 40.30733871459961, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 43.26874542236328, + "rewards/supergames_reward/std": 40.307342529296875, + "step": 124, + "step_time": 9.142582255997695 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 148.0, + "completions/max_terminated_length": 148.0, + "completions/mean_length": 106.625, + "completions/mean_terminated_length": 106.625, + "completions/min_length": 79.0, + "completions/min_terminated_length": 79.0, + "entropy": 0.30219921469688416, + "epoch": 0.625, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.391322135925293, + "learning_rate": 3.7999999999999996e-07, + "loss": 0.14367049932479858, + "num_tokens": 1067599.0, + "reward": 42.63374710083008, + "reward_std": 58.20897674560547, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 42.533748626708984, + "rewards/supergames_reward/std": 58.208984375, + "step": 125, + "step_time": 5.636063549987739 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 243.0, + "completions/max_terminated_length": 243.0, + "completions/mean_length": 179.25, + "completions/mean_terminated_length": 179.25, + "completions/min_length": 134.0, + "completions/min_terminated_length": 134.0, + "entropy": 0.4834887683391571, + "epoch": 0.63, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.9391286373138428, + "learning_rate": 3.75e-07, + "loss": -0.06614185124635696, + "num_tokens": 1075457.0, + "reward": 8.302498817443848, + "reward_std": 75.10364532470703, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 8.202500343322754, + "rewards/supergames_reward/std": 75.10364532470703, + "step": 126, + "step_time": 9.130600645992672 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 196.0, + "completions/max_terminated_length": 196.0, + "completions/mean_length": 157.0, + "completions/mean_terminated_length": 157.0, + "completions/min_length": 117.0, + "completions/min_terminated_length": 117.0, + "entropy": 0.4801807701587677, + "epoch": 0.635, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.425615310668945, + "learning_rate": 3.7e-07, + "loss": -0.08830522745847702, + "num_tokens": 1083129.0, + "reward": -10.563751220703125, + "reward_std": 64.00443267822266, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -10.663749694824219, + "rewards/supergames_reward/std": 64.00444030761719, + "step": 127, + "step_time": 7.4972667430120055 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 412.0, + "completions/max_terminated_length": 412.0, + "completions/mean_length": 239.25, + "completions/mean_terminated_length": 239.25, + "completions/min_length": 89.0, + "completions/min_terminated_length": 89.0, + "entropy": 0.4667063355445862, + "epoch": 0.64, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.199113130569458, + "learning_rate": 3.65e-07, + "loss": -0.043856702744960785, + "num_tokens": 1094355.0, + "reward": -55.89250183105469, + "reward_std": 53.05980682373047, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -55.99250030517578, + "rewards/supergames_reward/std": 53.05980682373047, + "step": 128, + "step_time": 15.429580625001108 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 370.0, + "completions/max_terminated_length": 370.0, + "completions/mean_length": 244.75, + "completions/mean_terminated_length": 244.75, + "completions/min_length": 127.0, + "completions/min_terminated_length": 127.0, + "entropy": 0.57573401927948, + "epoch": 0.645, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.6800079345703125, + "learning_rate": 3.6e-07, + "loss": 0.09521033614873886, + "num_tokens": 1105633.0, + "reward": -23.35250473022461, + "reward_std": 102.31401062011719, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -23.452499389648438, + "rewards/supergames_reward/std": 102.31402587890625, + "step": 129, + "step_time": 13.920327747007832 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 258.0, + "completions/max_terminated_length": 258.0, + "completions/mean_length": 188.25, + "completions/mean_terminated_length": 188.25, + "completions/min_length": 131.0, + "completions/min_terminated_length": 131.0, + "entropy": 0.5341438055038452, + "epoch": 0.65, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.97955060005188, + "learning_rate": 3.55e-07, + "loss": -0.09512650966644287, + "num_tokens": 1113571.0, + "reward": 29.181249618530273, + "reward_std": 80.10585021972656, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 29.081249237060547, + "rewards/supergames_reward/std": 80.10585021972656, + "step": 130, + "step_time": 9.69286120700417 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 257.0, + "completions/max_terminated_length": 257.0, + "completions/mean_length": 180.75, + "completions/mean_terminated_length": 180.75, + "completions/min_length": 136.0, + "completions/min_terminated_length": 136.0, + "entropy": 0.46591487526893616, + "epoch": 0.655, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.3518130779266357, + "learning_rate": 3.5e-07, + "loss": 0.04714229330420494, + "num_tokens": 1121457.0, + "reward": -8.287501335144043, + "reward_std": 76.35875701904297, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -8.387499809265137, + "rewards/supergames_reward/std": 76.35875701904297, + "step": 131, + "step_time": 9.636637121002423 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 190.0, + "completions/max_terminated_length": 190.0, + "completions/mean_length": 147.75, + "completions/mean_terminated_length": 147.75, + "completions/min_length": 90.0, + "completions/min_terminated_length": 90.0, + "entropy": 0.4407735764980316, + "epoch": 0.66, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.498741626739502, + "learning_rate": 3.45e-07, + "loss": -0.03996007889509201, + "num_tokens": 1129095.0, + "reward": 0.1999988555908203, + "reward_std": 84.63554382324219, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": 0.125, + "rewards/supergames_reward/std": 84.6017074584961, + "step": 132, + "step_time": 7.510833573003765 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 358.0, + "completions/max_terminated_length": 358.0, + "completions/mean_length": 213.875, + "completions/mean_terminated_length": 213.875, + "completions/min_length": 137.0, + "completions/min_terminated_length": 137.0, + "entropy": 0.5186149477958679, + "epoch": 0.665, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.3987107276916504, + "learning_rate": 3.4000000000000003e-07, + "loss": 0.1863594949245453, + "num_tokens": 1139110.0, + "reward": -21.328752517700195, + "reward_std": 88.9102783203125, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -21.428749084472656, + "rewards/supergames_reward/std": 88.9102783203125, + "step": 133, + "step_time": 13.393839450000087 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 230.0, + "completions/max_terminated_length": 230.0, + "completions/mean_length": 169.875, + "completions/mean_terminated_length": 169.875, + "completions/min_length": 121.0, + "completions/min_terminated_length": 121.0, + "entropy": 0.49445784091949463, + "epoch": 0.67, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.479673862457275, + "learning_rate": 3.35e-07, + "loss": 0.03618276119232178, + "num_tokens": 1146885.0, + "reward": -14.042501449584961, + "reward_std": 83.22671508789062, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -14.142499923706055, + "rewards/supergames_reward/std": 83.22671508789062, + "step": 134, + "step_time": 8.703399487014394 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 168.0, + "completions/max_terminated_length": 168.0, + "completions/mean_length": 124.75, + "completions/mean_terminated_length": 124.75, + "completions/min_length": 83.0, + "completions/min_terminated_length": 83.0, + "entropy": 0.2997814118862152, + "epoch": 0.675, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.440976619720459, + "learning_rate": 3.3e-07, + "loss": -0.029299044981598854, + "num_tokens": 1152427.0, + "reward": 45.274993896484375, + "reward_std": 44.401519775390625, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 45.17499923706055, + "rewards/supergames_reward/std": 44.40152359008789, + "step": 135, + "step_time": 6.334954546997324 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 161.0, + "completions/max_terminated_length": 161.0, + "completions/mean_length": 123.625, + "completions/mean_terminated_length": 123.625, + "completions/min_length": 93.0, + "completions/min_terminated_length": 93.0, + "entropy": 0.3462243974208832, + "epoch": 0.68, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.1878814697265625, + "learning_rate": 3.25e-07, + "loss": -0.10772529244422913, + "num_tokens": 1157952.0, + "reward": 44.186248779296875, + "reward_std": 51.78197479248047, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 44.08625030517578, + "rewards/supergames_reward/std": 51.7819709777832, + "step": 136, + "step_time": 6.067531232984038 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 155.0, + "completions/max_terminated_length": 155.0, + "completions/mean_length": 106.75, + "completions/mean_terminated_length": 106.75, + "completions/min_length": 74.0, + "completions/min_terminated_length": 74.0, + "entropy": 0.3662136197090149, + "epoch": 0.685, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.508971214294434, + "learning_rate": 3.2e-07, + "loss": -0.11172202974557877, + "num_tokens": 1163334.0, + "reward": 53.0574951171875, + "reward_std": 58.36653518676758, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 52.957496643066406, + "rewards/supergames_reward/std": 58.36653518676758, + "step": 137, + "step_time": 5.8906258280039765 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 175.0, + "completions/max_terminated_length": 175.0, + "completions/mean_length": 125.375, + "completions/mean_terminated_length": 125.375, + "completions/min_length": 90.0, + "completions/min_terminated_length": 90.0, + "entropy": 0.3352622091770172, + "epoch": 0.69, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.1937079429626465, + "learning_rate": 3.15e-07, + "loss": 0.08307601511478424, + "num_tokens": 1168929.0, + "reward": 11.796250343322754, + "reward_std": 53.48870849609375, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 11.696250915527344, + "rewards/supergames_reward/std": 53.48870849609375, + "step": 138, + "step_time": 6.595962718012743 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 254.0, + "completions/max_terminated_length": 254.0, + "completions/mean_length": 193.0, + "completions/mean_terminated_length": 193.0, + "completions/min_length": 111.0, + "completions/min_terminated_length": 111.0, + "entropy": 0.5486535429954529, + "epoch": 0.695, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.861871242523193, + "learning_rate": 3.1e-07, + "loss": 0.0321456640958786, + "num_tokens": 1176937.0, + "reward": 50.463748931884766, + "reward_std": 70.37553405761719, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": 50.38875198364258, + "rewards/supergames_reward/std": 70.31440734863281, + "step": 139, + "step_time": 9.544256388006033 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 261.0, + "completions/max_terminated_length": 261.0, + "completions/mean_length": 189.5, + "completions/mean_terminated_length": 189.5, + "completions/min_length": 95.0, + "completions/min_terminated_length": 95.0, + "entropy": 0.42230603098869324, + "epoch": 0.7, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.830677032470703, + "learning_rate": 3.05e-07, + "loss": -0.07222295552492142, + "num_tokens": 1186789.0, + "reward": 33.44999694824219, + "reward_std": 67.1298599243164, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 33.349998474121094, + "rewards/supergames_reward/std": 67.1298599243164, + "step": 140, + "step_time": 10.002322309010196 + } + ], + "logging_steps": 1, + "max_steps": 200, + "num_input_tokens_seen": 1186789, + "num_train_epochs": 1, + "save_steps": 20, + "stateful_callbacks": { + "TrainerControl": { + "args": { + "should_epoch_stop": false, + "should_evaluate": false, + "should_log": false, + "should_save": true, + "should_training_stop": false + }, + "attributes": {} + } + }, + "total_flos": 0.0, + "train_batch_size": 8, + "trial_name": null, + "trial_params": null +} diff --git a/checkpoint-140/training_args.bin b/checkpoint-140/training_args.bin new file mode 100644 index 0000000..730dec0 --- /dev/null +++ b/checkpoint-140/training_args.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:691fe5be2d7d1d56160dae63866d136877a72efb30e15ff7b3249192f998a788 +size 7185 diff --git a/checkpoint-160/chat_template.jinja b/checkpoint-160/chat_template.jinja new file mode 100644 index 0000000..bdf7919 --- /dev/null +++ b/checkpoint-160/chat_template.jinja @@ -0,0 +1,54 @@ +{%- if tools %} + {{- '<|im_start|>system\n' }} + {%- if messages[0]['role'] == 'system' %} + {{- messages[0]['content'] }} + {%- else %} + {{- 'You are Qwen, created by Alibaba Cloud. You are a helpful assistant.' }} + {%- endif %} + {{- "\n\n# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }} + {%- for tool in tools %} + {{- "\n" }} + {{- tool | tojson }} + {%- endfor %} + {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }} +{%- else %} + {%- if messages[0]['role'] == 'system' %} + {{- '<|im_start|>system\n' + messages[0]['content'] + '<|im_end|>\n' }} + {%- else %} + {{- '<|im_start|>system\nYou are Qwen, created by Alibaba Cloud. You are a helpful assistant.<|im_end|>\n' }} + {%- endif %} +{%- endif %} +{%- for message in messages %} + {%- if (message.role == "user") or (message.role == "system" and not loop.first) or (message.role == "assistant" and not message.tool_calls) %} + {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }} + {%- elif message.role == "assistant" %} + {{- '<|im_start|>' + message.role }} + {%- if message.content %} + {{- '\n' + message.content }} + {%- endif %} + {%- for tool_call in message.tool_calls %} + {%- if tool_call.function is defined %} + {%- set tool_call = tool_call.function %} + {%- endif %} + {{- '\n\n{"name": "' }} + {{- tool_call.name }} + {{- '", "arguments": ' }} + {{- tool_call.arguments | tojson }} + {{- '}\n' }} + {%- endfor %} + {{- '<|im_end|>\n' }} + {%- elif message.role == "tool" %} + {%- if (loop.index0 == 0) or (messages[loop.index0 - 1].role != "tool") %} + {{- '<|im_start|>user' }} + {%- endif %} + {{- '\n\n' }} + {{- message.content }} + {{- '\n' }} + {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %} + {{- '<|im_end|>\n' }} + {%- endif %} + {%- endif %} +{%- endfor %} +{%- if add_generation_prompt %} + {{- '<|im_start|>assistant\n' }} +{%- endif %} diff --git a/checkpoint-160/config.json b/checkpoint-160/config.json new file mode 100644 index 0000000..f57b09b --- /dev/null +++ b/checkpoint-160/config.json @@ -0,0 +1,61 @@ +{ + "architectures": [ + "Qwen2ForCausalLM" + ], + "attention_dropout": 0.0, + "bos_token_id": null, + "dtype": "float32", + "eos_token_id": 151645, + "hidden_act": "silu", + "hidden_size": 1536, + "initializer_range": 0.02, + "intermediate_size": 8960, + "layer_types": [ + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention" + ], + "max_position_embeddings": 32768, + "max_window_layers": 21, + "model_type": "qwen2", + "num_attention_heads": 12, + "num_hidden_layers": 28, + "num_key_value_heads": 2, + "pad_token_id": 151643, + "rms_norm_eps": 1e-06, + "rope_parameters": { + "rope_theta": 1000000.0, + "rope_type": "default" + }, + "sliding_window": null, + "tie_word_embeddings": true, + "transformers_version": "5.6.2", + "use_cache": false, + "use_sliding_window": false, + "vocab_size": 151936 +} diff --git a/checkpoint-160/generation_config.json b/checkpoint-160/generation_config.json new file mode 100644 index 0000000..1b2bba9 --- /dev/null +++ b/checkpoint-160/generation_config.json @@ -0,0 +1,13 @@ +{ + "do_sample": true, + "eos_token_id": [ + 151645, + 151643 + ], + "pad_token_id": 151643, + "repetition_penalty": 1.1, + "temperature": 0.7, + "top_k": 20, + "top_p": 0.8, + "transformers_version": "5.6.2" +} diff --git a/checkpoint-160/model.safetensors b/checkpoint-160/model.safetensors new file mode 100644 index 0000000..1f2a9b9 --- /dev/null +++ b/checkpoint-160/model.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:1e0d5c0c17e78ede34e38c84f732b9754796cabe771cda20dfba617e134780ff +size 6174895536 diff --git a/checkpoint-160/optimizer.pt b/checkpoint-160/optimizer.pt new file mode 100644 index 0000000..b8f2764 --- /dev/null +++ b/checkpoint-160/optimizer.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:50807e72f23133d36d3657a582732486baf0740b84182caf65002921eddd9a47 +size 12350013801 diff --git a/checkpoint-160/rng_state.pth b/checkpoint-160/rng_state.pth new file mode 100644 index 0000000..9d5e465 --- /dev/null +++ b/checkpoint-160/rng_state.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:573f5fe5c1703fc9f10a995c424d9deec8188be21ffead3f210839eff346240e +size 14645 diff --git a/checkpoint-160/scheduler.pt b/checkpoint-160/scheduler.pt new file mode 100644 index 0000000..66de7f6 --- /dev/null +++ b/checkpoint-160/scheduler.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:a7254de8fa1e1e54523ae878ad08e005e688587df236e1a0c0daa9706e54ac55 +size 1465 diff --git a/checkpoint-160/tokenizer.json b/checkpoint-160/tokenizer.json new file mode 100644 index 0000000..34510ff --- /dev/null +++ b/checkpoint-160/tokenizer.json @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:3fd169731d2cbde95e10bf356d66d5997fd885dd8dbb6fb4684da3f23b2585d8 +size 11421892 diff --git a/checkpoint-160/tokenizer_config.json b/checkpoint-160/tokenizer_config.json new file mode 100644 index 0000000..770e41d --- /dev/null +++ b/checkpoint-160/tokenizer_config.json @@ -0,0 +1,30 @@ +{ + "add_prefix_space": false, + "backend": "tokenizers", + "bos_token": null, + "clean_up_tokenization_spaces": false, + "eos_token": "<|im_end|>", + "errors": "replace", + "extra_special_tokens": [ + "<|im_start|>", + "<|im_end|>", + "<|object_ref_start|>", + "<|object_ref_end|>", + "<|box_start|>", + "<|box_end|>", + "<|quad_start|>", + "<|quad_end|>", + "<|vision_start|>", + "<|vision_end|>", + "<|vision_pad|>", + "<|image_pad|>", + "<|video_pad|>" + ], + "is_local": false, + "local_files_only": false, + "model_max_length": 131072, + "pad_token": "<|endoftext|>", + "split_special_tokens": false, + "tokenizer_class": "Qwen2Tokenizer", + "unk_token": null +} diff --git a/checkpoint-160/trainer_state.json b/checkpoint-160/trainer_state.json new file mode 100644 index 0000000..acc872a --- /dev/null +++ b/checkpoint-160/trainer_state.json @@ -0,0 +1,4674 @@ +{ + "best_global_step": null, + "best_metric": null, + "best_model_checkpoint": null, + "epoch": 0.8, + "eval_steps": 500, + "global_step": 160, + "is_hyper_param_search": false, + "is_local_process_zero": true, + "is_world_process_zero": true, + "log_history": [ + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 478.0, + "completions/max_terminated_length": 478.0, + "completions/mean_length": 191.0, + "completions/mean_terminated_length": 191.0, + "completions/min_length": 66.0, + "completions/min_terminated_length": 66.0, + "entropy": 0.5566893219947815, + "epoch": 0.005, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.370361328125, + "learning_rate": 1e-06, + "loss": -0.24658073484897614, + "num_tokens": 10832.0, + "reward": -3.3000030517578125, + "reward_std": 85.62333679199219, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -3.4000015258789062, + "rewards/supergames_reward/std": 85.62333679199219, + "step": 1, + "step_time": 22.779711072013015 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 216.0, + "completions/max_terminated_length": 216.0, + "completions/mean_length": 131.5, + "completions/mean_terminated_length": 131.5, + "completions/min_length": 60.0, + "completions/min_terminated_length": 60.0, + "entropy": 0.6395841240882874, + "epoch": 0.01, + "frac_reward_zero_std": 0.0, + "grad_norm": 8.09774398803711, + "learning_rate": 9.95e-07, + "loss": 0.20567649602890015, + "num_tokens": 16404.0, + "reward": -12.422499656677246, + "reward_std": 7.134707450866699, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -12.522500038146973, + "rewards/supergames_reward/std": 7.134707927703857, + "step": 2, + "step_time": 9.960156448010821 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 287.0, + "completions/max_terminated_length": 287.0, + "completions/mean_length": 156.625, + "completions/mean_terminated_length": 156.625, + "completions/min_length": 77.0, + "completions/min_terminated_length": 77.0, + "entropy": 0.562222421169281, + "epoch": 0.015, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.9053616523742676, + "learning_rate": 9.9e-07, + "loss": 0.09602774679660797, + "num_tokens": 26953.0, + "reward": -74.94999694824219, + "reward_std": 70.73094177246094, + "rewards/json_format_reward/mean": 0.05000000074505806, + "rewards/json_format_reward/std": 0.09258200973272324, + "rewards/supergames_reward/mean": -75.0, + "rewards/supergames_reward/std": 70.71067810058594, + "step": 3, + "step_time": 13.686320498993155 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 268.0, + "completions/max_terminated_length": 268.0, + "completions/mean_length": 154.125, + "completions/mean_terminated_length": 154.125, + "completions/min_length": 53.0, + "completions/min_terminated_length": 53.0, + "entropy": 0.7323317527770996, + "epoch": 0.02, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.933310031890869, + "learning_rate": 9.849999999999999e-07, + "loss": 0.28249427676200867, + "num_tokens": 36514.0, + "reward": -40.7599983215332, + "reward_std": 81.73140716552734, + "rewards/json_format_reward/mean": 0.05000000074505806, + "rewards/json_format_reward/std": 0.09258200973272324, + "rewards/supergames_reward/mean": -40.80999755859375, + "rewards/supergames_reward/std": 81.68995666503906, + "step": 4, + "step_time": 12.805880262021674 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 401.0, + "completions/max_terminated_length": 401.0, + "completions/mean_length": 205.25, + "completions/mean_terminated_length": 205.25, + "completions/min_length": 115.0, + "completions/min_terminated_length": 115.0, + "entropy": 0.6167430281639099, + "epoch": 0.025, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.344135046005249, + "learning_rate": 9.8e-07, + "loss": 0.1433994024991989, + "num_tokens": 47476.0, + "reward": -97.6612548828125, + "reward_std": 6.3321428298950195, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -97.76124572753906, + "rewards/supergames_reward/std": 6.3321428298950195, + "step": 5, + "step_time": 18.789364666008623 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 276.0, + "completions/max_terminated_length": 276.0, + "completions/mean_length": 167.25, + "completions/mean_terminated_length": 167.25, + "completions/min_length": 73.0, + "completions/min_terminated_length": 73.0, + "entropy": 0.7583790421485901, + "epoch": 0.03, + "frac_reward_zero_std": 0.0, + "grad_norm": 5.172371864318848, + "learning_rate": 9.75e-07, + "loss": -0.12127404659986496, + "num_tokens": 55246.0, + "reward": -58.687503814697266, + "reward_std": 58.64059829711914, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": -58.76250457763672, + "rewards/supergames_reward/std": 58.620460510253906, + "step": 6, + "step_time": 12.948570273991209 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 261.0, + "completions/max_terminated_length": 261.0, + "completions/mean_length": 133.875, + "completions/mean_terminated_length": 133.875, + "completions/min_length": 57.0, + "completions/min_terminated_length": 57.0, + "entropy": 0.7447654604911804, + "epoch": 0.035, + "frac_reward_zero_std": 0.0, + "grad_norm": 6.428595066070557, + "learning_rate": 9.7e-07, + "loss": 0.10135584324598312, + "num_tokens": 60837.0, + "reward": -32.45000076293945, + "reward_std": 41.754486083984375, + "rewards/json_format_reward/mean": 0.05000000074505806, + "rewards/json_format_reward/std": 0.09258200973272324, + "rewards/supergames_reward/mean": -32.5, + "rewards/supergames_reward/std": 41.66190719604492, + "step": 7, + "step_time": 12.058315072004916 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 414.0, + "completions/max_terminated_length": 414.0, + "completions/mean_length": 209.125, + "completions/mean_terminated_length": 209.125, + "completions/min_length": 104.0, + "completions/min_terminated_length": 104.0, + "entropy": 0.3860666751861572, + "epoch": 0.04, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.806255578994751, + "learning_rate": 9.649999999999999e-07, + "loss": -0.33238139748573303, + "num_tokens": 70838.0, + "reward": -80.50375366210938, + "reward_std": 38.82380294799805, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -80.60375213623047, + "rewards/supergames_reward/std": 38.82379913330078, + "step": 8, + "step_time": 19.323370319994865 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.25, + "completions/max_length": 512.0, + "completions/max_terminated_length": 199.0, + "completions/mean_length": 247.125, + "completions/mean_terminated_length": 158.83334350585938, + "completions/min_length": 106.0, + "completions/min_terminated_length": 106.0, + "entropy": 0.5816237330436707, + "epoch": 0.045, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.3209943771362305, + "learning_rate": 9.6e-07, + "loss": 0.36752766370773315, + "num_tokens": 82135.0, + "reward": -27.688751220703125, + "reward_std": 82.8358154296875, + "rewards/json_format_reward/mean": 0.05000000074505806, + "rewards/json_format_reward/std": 0.09258200973272324, + "rewards/supergames_reward/mean": -27.738750457763672, + "rewards/supergames_reward/std": 82.78590393066406, + "step": 9, + "step_time": 21.948575104994234 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 240.0, + "completions/max_terminated_length": 240.0, + "completions/mean_length": 153.0, + "completions/mean_terminated_length": 153.0, + "completions/min_length": 51.0, + "completions/min_terminated_length": 51.0, + "entropy": 0.6301521062850952, + "epoch": 0.05, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.125359058380127, + "learning_rate": 9.55e-07, + "loss": 0.06129350885748863, + "num_tokens": 89807.0, + "reward": -78.39624786376953, + "reward_std": 40.17042922973633, + "rewards/json_format_reward/mean": 0.05000000074505806, + "rewards/json_format_reward/std": 0.09258200973272324, + "rewards/supergames_reward/mean": -78.44625091552734, + "rewards/supergames_reward/std": 40.1396484375, + "step": 10, + "step_time": 9.098202993016457 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 133.0, + "completions/max_terminated_length": 133.0, + "completions/mean_length": 112.0, + "completions/mean_terminated_length": 112.0, + "completions/min_length": 68.0, + "completions/min_terminated_length": 68.0, + "entropy": 0.5521990656852722, + "epoch": 0.055, + "frac_reward_zero_std": 0.0, + "grad_norm": 6.616060256958008, + "learning_rate": 9.499999999999999e-07, + "loss": -0.006313305348157883, + "num_tokens": 95231.0, + "reward": -12.161249160766602, + "reward_std": 6.39578104019165, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -12.261249542236328, + "rewards/supergames_reward/std": 6.39578104019165, + "step": 11, + "step_time": 5.136311663984088 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 396.0, + "completions/max_terminated_length": 396.0, + "completions/mean_length": 214.875, + "completions/mean_terminated_length": 214.875, + "completions/min_length": 55.0, + "completions/min_terminated_length": 55.0, + "entropy": 0.5505905151367188, + "epoch": 0.06, + "frac_reward_zero_std": 0.0, + "grad_norm": 5.289583206176758, + "learning_rate": 9.45e-07, + "loss": -0.29802003502845764, + "num_tokens": 106238.0, + "reward": -91.66999816894531, + "reward_std": 23.2779541015625, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -91.77000427246094, + "rewards/supergames_reward/std": 23.277956008911133, + "step": 12, + "step_time": 14.945365622988902 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 269.0, + "completions/max_terminated_length": 269.0, + "completions/mean_length": 167.375, + "completions/mean_terminated_length": 167.375, + "completions/min_length": 113.0, + "completions/min_terminated_length": 113.0, + "entropy": 0.540415346622467, + "epoch": 0.065, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.858291149139404, + "learning_rate": 9.399999999999999e-07, + "loss": 0.24004128575325012, + "num_tokens": 116937.0, + "reward": -20.38249969482422, + "reward_std": 87.55204010009766, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": -20.457500457763672, + "rewards/supergames_reward/std": 87.52605438232422, + "step": 13, + "step_time": 10.482285185018554 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 254.0, + "completions/max_terminated_length": 254.0, + "completions/mean_length": 156.875, + "completions/mean_terminated_length": 156.875, + "completions/min_length": 114.0, + "completions/min_terminated_length": 114.0, + "entropy": 0.7237679958343506, + "epoch": 0.07, + "frac_reward_zero_std": 0.0, + "grad_norm": 5.083980560302734, + "learning_rate": 9.35e-07, + "loss": -0.19025824964046478, + "num_tokens": 124640.0, + "reward": -51.397499084472656, + "reward_std": 54.486454010009766, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -51.49749755859375, + "rewards/supergames_reward/std": 54.486454010009766, + "step": 14, + "step_time": 9.597231683001155 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 236.0, + "completions/max_terminated_length": 236.0, + "completions/mean_length": 141.375, + "completions/mean_terminated_length": 141.375, + "completions/min_length": 91.0, + "completions/min_terminated_length": 91.0, + "entropy": 0.754334568977356, + "epoch": 0.075, + "frac_reward_zero_std": 0.0, + "grad_norm": 5.720428943634033, + "learning_rate": 9.3e-07, + "loss": -0.16830675303936005, + "num_tokens": 132219.0, + "reward": -56.226253509521484, + "reward_std": 60.115657806396484, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -56.32625198364258, + "rewards/supergames_reward/std": 60.11566162109375, + "step": 15, + "step_time": 8.998362872982398 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 209.0, + "completions/max_terminated_length": 209.0, + "completions/mean_length": 136.375, + "completions/mean_terminated_length": 136.375, + "completions/min_length": 97.0, + "completions/min_terminated_length": 97.0, + "entropy": 0.6280568838119507, + "epoch": 0.08, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.9681806564331055, + "learning_rate": 9.25e-07, + "loss": -0.01910916529595852, + "num_tokens": 139726.0, + "reward": -68.24000549316406, + "reward_std": 52.238914489746094, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -68.33999633789062, + "rewards/supergames_reward/std": 52.238914489746094, + "step": 16, + "step_time": 8.034480294008972 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 365.0, + "completions/max_terminated_length": 365.0, + "completions/mean_length": 211.875, + "completions/mean_terminated_length": 211.875, + "completions/min_length": 100.0, + "completions/min_terminated_length": 100.0, + "entropy": 0.899176836013794, + "epoch": 0.085, + "frac_reward_zero_std": 0.0, + "grad_norm": 6.691746711730957, + "learning_rate": 9.2e-07, + "loss": 0.25513291358947754, + "num_tokens": 147869.0, + "reward": -99.92500305175781, + "reward_std": 0.07070883363485336, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": -100.0, + "rewards/supergames_reward/std": 0.0, + "step": 17, + "step_time": 13.655792869016295 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 387.0, + "completions/max_terminated_length": 387.0, + "completions/mean_length": 238.0, + "completions/mean_terminated_length": 238.0, + "completions/min_length": 87.0, + "completions/min_terminated_length": 87.0, + "entropy": 0.5596873760223389, + "epoch": 0.09, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.6711347103118896, + "learning_rate": 9.15e-07, + "loss": -0.2628335654735565, + "num_tokens": 159101.0, + "reward": -24.900001525878906, + "reward_std": 103.50983428955078, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -25.0, + "rewards/supergames_reward/std": 103.50983428955078, + "step": 18, + "step_time": 14.784329420013819 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.125, + "completions/max_length": 512.0, + "completions/max_terminated_length": 277.0, + "completions/mean_length": 219.75, + "completions/mean_terminated_length": 178.00001525878906, + "completions/min_length": 90.0, + "completions/min_terminated_length": 90.0, + "entropy": 0.9245517253875732, + "epoch": 0.095, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.55502986907959, + "learning_rate": 9.1e-07, + "loss": 0.13060179352760315, + "num_tokens": 167299.0, + "reward": -86.89125061035156, + "reward_std": 36.95603561401367, + "rewards/json_format_reward/mean": 0.05000000074505806, + "rewards/json_format_reward/std": 0.09258200973272324, + "rewards/supergames_reward/mean": -86.94125366210938, + "rewards/supergames_reward/std": 36.93572235107422, + "step": 19, + "step_time": 18.832684700988466 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 267.0, + "completions/max_terminated_length": 267.0, + "completions/mean_length": 177.25, + "completions/mean_terminated_length": 177.25, + "completions/min_length": 93.0, + "completions/min_terminated_length": 93.0, + "entropy": 0.7331254482269287, + "epoch": 0.1, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.538497447967529, + "learning_rate": 9.05e-07, + "loss": -0.16418980062007904, + "num_tokens": 175141.0, + "reward": -68.52874755859375, + "reward_std": 58.48719787597656, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -68.62875366210938, + "rewards/supergames_reward/std": 58.48720169067383, + "step": 20, + "step_time": 10.090975169994636 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 300.0, + "completions/max_terminated_length": 300.0, + "completions/mean_length": 190.625, + "completions/mean_terminated_length": 190.625, + "completions/min_length": 103.0, + "completions/min_terminated_length": 103.0, + "entropy": 0.6659425497055054, + "epoch": 0.105, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.485373497009277, + "learning_rate": 9e-07, + "loss": -0.10106432437896729, + "num_tokens": 183098.0, + "reward": -71.48875427246094, + "reward_std": 55.33565139770508, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -71.5887451171875, + "rewards/supergames_reward/std": 55.33565139770508, + "step": 21, + "step_time": 11.255001295008697 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 279.0, + "completions/max_terminated_length": 279.0, + "completions/mean_length": 179.0, + "completions/mean_terminated_length": 179.0, + "completions/min_length": 103.0, + "completions/min_terminated_length": 103.0, + "entropy": 0.49365949630737305, + "epoch": 0.11, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.829419136047363, + "learning_rate": 8.95e-07, + "loss": 0.10587231814861298, + "num_tokens": 192826.0, + "reward": -75.62000274658203, + "reward_std": 45.01959991455078, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": -75.69499969482422, + "rewards/supergames_reward/std": 45.00411605834961, + "step": 22, + "step_time": 10.611246599000879 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 315.0, + "completions/max_terminated_length": 315.0, + "completions/mean_length": 212.625, + "completions/mean_terminated_length": 212.625, + "completions/min_length": 100.0, + "completions/min_terminated_length": 100.0, + "entropy": 0.43159469962120056, + "epoch": 0.115, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.541125774383545, + "learning_rate": 8.9e-07, + "loss": -0.1681259125471115, + "num_tokens": 202815.0, + "reward": 23.67624855041504, + "reward_std": 67.74340057373047, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 23.576250076293945, + "rewards/supergames_reward/std": 67.743408203125, + "step": 23, + "step_time": 11.731771531980485 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 168.0, + "completions/max_terminated_length": 168.0, + "completions/mean_length": 101.625, + "completions/mean_terminated_length": 101.625, + "completions/min_length": 65.0, + "completions/min_terminated_length": 65.0, + "entropy": 0.3972298502922058, + "epoch": 0.12, + "frac_reward_zero_std": 0.0, + "grad_norm": 5.8285417556762695, + "learning_rate": 8.85e-07, + "loss": -0.011585958302021027, + "num_tokens": 208188.0, + "reward": 58.849998474121094, + "reward_std": 56.9304084777832, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 58.75, + "rewards/supergames_reward/std": 56.9304084777832, + "step": 24, + "step_time": 6.187504950998118 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 303.0, + "completions/max_terminated_length": 303.0, + "completions/mean_length": 165.375, + "completions/mean_terminated_length": 165.375, + "completions/min_length": 69.0, + "completions/min_terminated_length": 69.0, + "entropy": 0.6339899301528931, + "epoch": 0.125, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.823799133300781, + "learning_rate": 8.799999999999999e-07, + "loss": 0.05159185826778412, + "num_tokens": 215935.0, + "reward": -23.521251678466797, + "reward_std": 74.23067474365234, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -23.621248245239258, + "rewards/supergames_reward/std": 74.23067474365234, + "step": 25, + "step_time": 11.059416300005978 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 349.0, + "completions/max_terminated_length": 349.0, + "completions/mean_length": 204.0, + "completions/mean_terminated_length": 204.0, + "completions/min_length": 143.0, + "completions/min_terminated_length": 143.0, + "entropy": 0.6462895274162292, + "epoch": 0.13, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.969292640686035, + "learning_rate": 8.75e-07, + "loss": 0.0115677984431386, + "num_tokens": 223999.0, + "reward": -74.8175048828125, + "reward_std": 40.94221115112305, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": -74.89250183105469, + "rewards/supergames_reward/std": 40.924625396728516, + "step": 26, + "step_time": 12.666237785975682 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 299.0, + "completions/max_terminated_length": 299.0, + "completions/mean_length": 166.5, + "completions/mean_terminated_length": 166.5, + "completions/min_length": 87.0, + "completions/min_terminated_length": 87.0, + "entropy": 0.4756850600242615, + "epoch": 0.135, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.915069103240967, + "learning_rate": 8.699999999999999e-07, + "loss": -0.13321346044540405, + "num_tokens": 233643.0, + "reward": -52.125, + "reward_std": 19.405282974243164, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": -52.20000076293945, + "rewards/supergames_reward/std": 19.33464813232422, + "step": 27, + "step_time": 11.172539752995363 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 184.0, + "completions/max_terminated_length": 184.0, + "completions/mean_length": 137.125, + "completions/mean_terminated_length": 137.125, + "completions/min_length": 100.0, + "completions/min_terminated_length": 100.0, + "entropy": 0.5834778547286987, + "epoch": 0.14, + "frac_reward_zero_std": 0.0, + "grad_norm": 5.146666526794434, + "learning_rate": 8.65e-07, + "loss": 0.060454584658145905, + "num_tokens": 241204.0, + "reward": -68.99250793457031, + "reward_std": 51.010581970214844, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -69.09249877929688, + "rewards/supergames_reward/std": 51.010581970214844, + "step": 28, + "step_time": 7.025046669004951 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 331.0, + "completions/max_terminated_length": 331.0, + "completions/mean_length": 182.375, + "completions/mean_terminated_length": 182.375, + "completions/min_length": 106.0, + "completions/min_terminated_length": 106.0, + "entropy": 0.6368551254272461, + "epoch": 0.145, + "frac_reward_zero_std": 0.0, + "grad_norm": 5.15535306930542, + "learning_rate": 8.599999999999999e-07, + "loss": -0.3323673903942108, + "num_tokens": 249087.0, + "reward": -44.17000198364258, + "reward_std": 80.5373306274414, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -44.27000045776367, + "rewards/supergames_reward/std": 80.5373306274414, + "step": 29, + "step_time": 12.079259724996518 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 190.0, + "completions/max_terminated_length": 190.0, + "completions/mean_length": 127.75, + "completions/mean_terminated_length": 127.75, + "completions/min_length": 80.0, + "completions/min_terminated_length": 80.0, + "entropy": 0.4491196870803833, + "epoch": 0.15, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.896186828613281, + "learning_rate": 8.55e-07, + "loss": -0.12887656688690186, + "num_tokens": 259445.0, + "reward": 57.446250915527344, + "reward_std": 38.78837585449219, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 57.34625244140625, + "rewards/supergames_reward/std": 38.78837966918945, + "step": 30, + "step_time": 7.483972909016302 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 166.0, + "completions/max_terminated_length": 166.0, + "completions/mean_length": 112.875, + "completions/mean_terminated_length": 112.875, + "completions/min_length": 85.0, + "completions/min_terminated_length": 85.0, + "entropy": 0.4858454465866089, + "epoch": 0.155, + "frac_reward_zero_std": 0.0, + "grad_norm": 6.742906093597412, + "learning_rate": 8.499999999999999e-07, + "loss": 0.06930096447467804, + "num_tokens": 264924.0, + "reward": -21.174999237060547, + "reward_std": 31.89051628112793, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": -21.25, + "rewards/supergames_reward/std": 31.819807052612305, + "step": 31, + "step_time": 6.1866529019898735 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 148.0, + "completions/max_terminated_length": 148.0, + "completions/mean_length": 118.125, + "completions/mean_terminated_length": 118.125, + "completions/min_length": 98.0, + "completions/min_terminated_length": 98.0, + "entropy": 0.45428749918937683, + "epoch": 0.16, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.663414478302002, + "learning_rate": 8.45e-07, + "loss": 0.05735419690608978, + "num_tokens": 270381.0, + "reward": 0.4687504768371582, + "reward_std": 38.66371154785156, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 0.36875057220458984, + "rewards/supergames_reward/std": 38.6637077331543, + "step": 32, + "step_time": 5.577042757999152 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 345.0, + "completions/max_terminated_length": 345.0, + "completions/mean_length": 166.875, + "completions/mean_terminated_length": 166.875, + "completions/min_length": 86.0, + "completions/min_terminated_length": 86.0, + "entropy": 0.6107293367385864, + "epoch": 0.165, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.75006628036499, + "learning_rate": 8.399999999999999e-07, + "loss": -0.14760127663612366, + "num_tokens": 278156.0, + "reward": -87.84750366210938, + "reward_std": 26.84708023071289, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -87.94749450683594, + "rewards/supergames_reward/std": 26.847076416015625, + "step": 33, + "step_time": 12.552876825997373 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.125, + "completions/max_length": 512.0, + "completions/max_terminated_length": 360.0, + "completions/mean_length": 231.0, + "completions/mean_terminated_length": 190.85714721679688, + "completions/min_length": 74.0, + "completions/min_terminated_length": 74.0, + "entropy": 0.6089653372764587, + "epoch": 0.17, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.781267166137695, + "learning_rate": 8.349999999999999e-07, + "loss": 0.4294504225254059, + "num_tokens": 289364.0, + "reward": -99.92500305175781, + "reward_std": 0.07070883363485336, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": -100.0, + "rewards/supergames_reward/std": 0.0, + "step": 34, + "step_time": 18.961819477990502 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 218.0, + "completions/max_terminated_length": 218.0, + "completions/mean_length": 134.125, + "completions/mean_terminated_length": 134.125, + "completions/min_length": 104.0, + "completions/min_terminated_length": 104.0, + "entropy": 0.46158918738365173, + "epoch": 0.175, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.016353607177734, + "learning_rate": 8.299999999999999e-07, + "loss": 0.09251818805932999, + "num_tokens": 298733.0, + "reward": -35.558753967285156, + "reward_std": 15.881204605102539, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -35.65875244140625, + "rewards/supergames_reward/std": 15.881203651428223, + "step": 35, + "step_time": 8.38491243700264 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 207.0, + "completions/max_terminated_length": 207.0, + "completions/mean_length": 156.875, + "completions/mean_terminated_length": 156.875, + "completions/min_length": 111.0, + "completions/min_terminated_length": 111.0, + "entropy": 0.5104788541793823, + "epoch": 0.18, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.592893123626709, + "learning_rate": 8.249999999999999e-07, + "loss": -0.0410832017660141, + "num_tokens": 306420.0, + "reward": 5.4212493896484375, + "reward_std": 93.50404357910156, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": 5.346250534057617, + "rewards/supergames_reward/std": 93.4718246459961, + "step": 36, + "step_time": 7.811868985998444 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 241.0, + "completions/max_terminated_length": 241.0, + "completions/mean_length": 178.0, + "completions/mean_terminated_length": 178.0, + "completions/min_length": 106.0, + "completions/min_terminated_length": 106.0, + "entropy": 0.7527878880500793, + "epoch": 0.185, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.8812546730041504, + "learning_rate": 8.199999999999999e-07, + "loss": 0.12539231777191162, + "num_tokens": 314252.0, + "reward": -89.58500671386719, + "reward_std": 29.256120681762695, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": -89.66000366210938, + "rewards/supergames_reward/std": 29.24593734741211, + "step": 37, + "step_time": 8.959742914012168 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 199.0, + "completions/max_terminated_length": 199.0, + "completions/mean_length": 133.375, + "completions/mean_terminated_length": 133.375, + "completions/min_length": 87.0, + "completions/min_terminated_length": 87.0, + "entropy": 0.3347901403903961, + "epoch": 0.19, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.205414295196533, + "learning_rate": 8.149999999999999e-07, + "loss": -0.06991486996412277, + "num_tokens": 319839.0, + "reward": -14.498749732971191, + "reward_std": 13.007229804992676, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -14.598750114440918, + "rewards/supergames_reward/std": 13.007229804992676, + "step": 38, + "step_time": 7.377183554985095 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 181.0, + "completions/max_terminated_length": 181.0, + "completions/mean_length": 114.375, + "completions/mean_terminated_length": 114.375, + "completions/min_length": 58.0, + "completions/min_terminated_length": 58.0, + "entropy": 0.4445965886116028, + "epoch": 0.195, + "frac_reward_zero_std": 0.0, + "grad_norm": 5.561298847198486, + "learning_rate": 8.1e-07, + "loss": -0.06235164776444435, + "num_tokens": 325290.0, + "reward": -7.425000190734863, + "reward_std": 53.701602935791016, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": -7.5, + "rewards/supergames_reward/std": 53.652320861816406, + "step": 39, + "step_time": 6.753862089011818 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 370.0, + "completions/max_terminated_length": 370.0, + "completions/mean_length": 187.625, + "completions/mean_terminated_length": 187.625, + "completions/min_length": 60.0, + "completions/min_terminated_length": 60.0, + "entropy": 0.3950234353542328, + "epoch": 0.2, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.5988712310791016, + "learning_rate": 8.05e-07, + "loss": -0.1381440907716751, + "num_tokens": 335119.0, + "reward": -33.68375015258789, + "reward_std": 42.069435119628906, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -33.783748626708984, + "rewards/supergames_reward/std": 42.069435119628906, + "step": 40, + "step_time": 13.718958162004128 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.125, + "completions/max_length": 512.0, + "completions/max_terminated_length": 371.0, + "completions/mean_length": 251.125, + "completions/mean_terminated_length": 213.85714721679688, + "completions/min_length": 121.0, + "completions/min_terminated_length": 121.0, + "entropy": 0.507067084312439, + "epoch": 0.205, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.3898658752441406, + "learning_rate": 8e-07, + "loss": 0.0078964838758111, + "num_tokens": 346464.0, + "reward": -86.48625183105469, + "reward_std": 38.020694732666016, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": -86.56124877929688, + "rewards/supergames_reward/std": 38.010528564453125, + "step": 41, + "step_time": 18.90960379401804 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 152.0, + "completions/max_terminated_length": 152.0, + "completions/mean_length": 116.375, + "completions/mean_terminated_length": 116.375, + "completions/min_length": 93.0, + "completions/min_terminated_length": 93.0, + "entropy": 0.3801707923412323, + "epoch": 0.21, + "frac_reward_zero_std": 1.0, + "grad_norm": 0.0, + "learning_rate": 7.95e-07, + "loss": 0.0, + "num_tokens": 351955.0, + "reward": -9.899999618530273, + "reward_std": 0.0, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -10.0, + "rewards/supergames_reward/std": 0.0, + "step": 42, + "step_time": 5.711630532023264 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 308.0, + "completions/max_terminated_length": 308.0, + "completions/mean_length": 247.375, + "completions/mean_terminated_length": 247.375, + "completions/min_length": 116.0, + "completions/min_terminated_length": 116.0, + "entropy": 0.3185043931007385, + "epoch": 0.215, + "frac_reward_zero_std": 0.0, + "grad_norm": 2.168762445449829, + "learning_rate": 7.9e-07, + "loss": 0.1970764398574829, + "num_tokens": 363278.0, + "reward": -53.89249801635742, + "reward_std": 63.02619934082031, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -53.99250030517578, + "rewards/supergames_reward/std": 63.02619934082031, + "step": 43, + "step_time": 11.661934906995157 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 330.0, + "completions/max_terminated_length": 330.0, + "completions/mean_length": 212.5, + "completions/mean_terminated_length": 212.5, + "completions/min_length": 106.0, + "completions/min_terminated_length": 106.0, + "entropy": 0.4446268379688263, + "epoch": 0.22, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.305844306945801, + "learning_rate": 7.85e-07, + "loss": -0.06735197454690933, + "num_tokens": 374306.0, + "reward": 18.059999465942383, + "reward_std": 98.9417953491211, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 17.959999084472656, + "rewards/supergames_reward/std": 98.9417953491211, + "step": 44, + "step_time": 12.492453911982011 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 316.0, + "completions/max_terminated_length": 316.0, + "completions/mean_length": 183.0, + "completions/mean_terminated_length": 183.0, + "completions/min_length": 103.0, + "completions/min_terminated_length": 103.0, + "entropy": 0.5269614458084106, + "epoch": 0.225, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.8006370067596436, + "learning_rate": 7.799999999999999e-07, + "loss": -0.08870815485715866, + "num_tokens": 382210.0, + "reward": -54.928749084472656, + "reward_std": 72.5127944946289, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -55.02874755859375, + "rewards/supergames_reward/std": 72.5127944946289, + "step": 45, + "step_time": 11.7636467939883 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 244.0, + "completions/max_terminated_length": 244.0, + "completions/mean_length": 136.625, + "completions/mean_terminated_length": 136.625, + "completions/min_length": 85.0, + "completions/min_terminated_length": 85.0, + "entropy": 0.34820204973220825, + "epoch": 0.23, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.911736011505127, + "learning_rate": 7.75e-07, + "loss": 0.04397330805659294, + "num_tokens": 391583.0, + "reward": -19.10375213623047, + "reward_std": 23.967702865600586, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -19.203750610351562, + "rewards/supergames_reward/std": 23.967702865600586, + "step": 46, + "step_time": 10.119426151999505 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 218.0, + "completions/max_terminated_length": 218.0, + "completions/mean_length": 160.375, + "completions/mean_terminated_length": 160.375, + "completions/min_length": 93.0, + "completions/min_terminated_length": 93.0, + "entropy": 0.611914336681366, + "epoch": 0.235, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.571913242340088, + "learning_rate": 7.699999999999999e-07, + "loss": -0.044293053448200226, + "num_tokens": 399290.0, + "reward": -25.186250686645508, + "reward_std": 71.46305084228516, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -25.286251068115234, + "rewards/supergames_reward/std": 71.46305084228516, + "step": 47, + "step_time": 8.563868903991533 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 167.0, + "completions/max_terminated_length": 167.0, + "completions/mean_length": 111.125, + "completions/mean_terminated_length": 111.125, + "completions/min_length": 56.0, + "completions/min_terminated_length": 56.0, + "entropy": 0.46585777401924133, + "epoch": 0.24, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.654109954833984, + "learning_rate": 7.65e-07, + "loss": -0.035151124000549316, + "num_tokens": 404707.0, + "reward": -0.408750057220459, + "reward_std": 31.939748764038086, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -0.5087499618530273, + "rewards/supergames_reward/std": 31.93975257873535, + "step": 48, + "step_time": 6.438482580007985 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 306.0, + "completions/max_terminated_length": 306.0, + "completions/mean_length": 216.625, + "completions/mean_terminated_length": 216.625, + "completions/min_length": 132.0, + "completions/min_terminated_length": 132.0, + "entropy": 0.33054471015930176, + "epoch": 0.245, + "frac_reward_zero_std": 0.0, + "grad_norm": 2.700303792953491, + "learning_rate": 7.599999999999999e-07, + "loss": -0.04393656551837921, + "num_tokens": 414736.0, + "reward": -97.26000213623047, + "reward_std": 1.6294406652450562, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -97.36000061035156, + "rewards/supergames_reward/std": 1.629441738128662, + "step": 49, + "step_time": 11.636040106008295 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 378.0, + "completions/max_terminated_length": 378.0, + "completions/mean_length": 218.375, + "completions/mean_terminated_length": 218.375, + "completions/min_length": 104.0, + "completions/min_terminated_length": 104.0, + "entropy": 0.48711177706718445, + "epoch": 0.25, + "frac_reward_zero_std": 0.0, + "grad_norm": 2.9650564193725586, + "learning_rate": 7.55e-07, + "loss": -0.3529520332813263, + "num_tokens": 424795.0, + "reward": 47.408748626708984, + "reward_std": 72.72083282470703, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 47.30875015258789, + "rewards/supergames_reward/std": 72.72083282470703, + "step": 50, + "step_time": 14.258096842997475 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.125, + "completions/max_length": 512.0, + "completions/max_terminated_length": 492.0, + "completions/mean_length": 267.625, + "completions/mean_terminated_length": 232.71429443359375, + "completions/min_length": 90.0, + "completions/min_terminated_length": 90.0, + "entropy": 0.4482860863208771, + "epoch": 0.255, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.2998197078704834, + "learning_rate": 7.5e-07, + "loss": 0.2493654191493988, + "num_tokens": 436288.0, + "reward": 50.61125183105469, + "reward_std": 75.3349838256836, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": 50.53624725341797, + "rewards/supergames_reward/std": 75.27783966064453, + "step": 51, + "step_time": 19.20827590499539 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 176.0, + "completions/max_terminated_length": 176.0, + "completions/mean_length": 144.5, + "completions/mean_terminated_length": 144.5, + "completions/min_length": 99.0, + "completions/min_terminated_length": 99.0, + "entropy": 0.5378735661506653, + "epoch": 0.26, + "frac_reward_zero_std": 0.0, + "grad_norm": 5.114461898803711, + "learning_rate": 7.45e-07, + "loss": 0.0036597400903701782, + "num_tokens": 441980.0, + "reward": -15.20625114440918, + "reward_std": 45.35526657104492, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": -15.28125, + "rewards/supergames_reward/std": 45.301815032958984, + "step": 52, + "step_time": 6.660627231001854 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 136.0, + "completions/max_terminated_length": 136.0, + "completions/mean_length": 119.0, + "completions/mean_terminated_length": 119.0, + "completions/min_length": 105.0, + "completions/min_terminated_length": 105.0, + "entropy": 0.41796669363975525, + "epoch": 0.265, + "frac_reward_zero_std": 0.0, + "grad_norm": 5.059505462646484, + "learning_rate": 7.4e-07, + "loss": -0.014973883517086506, + "num_tokens": 447468.0, + "reward": 15.170000076293945, + "reward_std": 46.658329010009766, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 15.070000648498535, + "rewards/supergames_reward/std": 46.658329010009766, + "step": 53, + "step_time": 5.271571868011961 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 259.0, + "completions/max_terminated_length": 259.0, + "completions/mean_length": 191.125, + "completions/mean_terminated_length": 191.125, + "completions/min_length": 91.0, + "completions/min_terminated_length": 91.0, + "entropy": 0.45421987771987915, + "epoch": 0.27, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.4454867839813232, + "learning_rate": 7.35e-07, + "loss": -0.1128973588347435, + "num_tokens": 457301.0, + "reward": 38.89875030517578, + "reward_std": 62.1772575378418, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 38.79874801635742, + "rewards/supergames_reward/std": 62.1772575378418, + "step": 54, + "step_time": 10.524528659996577 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 319.0, + "completions/max_terminated_length": 319.0, + "completions/mean_length": 233.375, + "completions/mean_terminated_length": 233.375, + "completions/min_length": 136.0, + "completions/min_terminated_length": 136.0, + "entropy": 0.4449213445186615, + "epoch": 0.275, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.2065398693084717, + "learning_rate": 7.3e-07, + "loss": 0.0016277075046673417, + "num_tokens": 467520.0, + "reward": 27.372499465942383, + "reward_std": 81.23126983642578, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 27.27250099182129, + "rewards/supergames_reward/std": 81.23126983642578, + "step": 55, + "step_time": 12.592280682991259 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.125, + "completions/max_length": 512.0, + "completions/max_terminated_length": 357.0, + "completions/mean_length": 261.0, + "completions/mean_terminated_length": 225.1428680419922, + "completions/min_length": 95.0, + "completions/min_terminated_length": 95.0, + "entropy": 0.506920337677002, + "epoch": 0.28, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.058388710021973, + "learning_rate": 7.249999999999999e-07, + "loss": 0.3289242386817932, + "num_tokens": 476056.0, + "reward": 47.05249786376953, + "reward_std": 80.97753143310547, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": 46.977500915527344, + "rewards/supergames_reward/std": 80.9256362915039, + "step": 56, + "step_time": 18.845177220006008 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 250.0, + "completions/max_terminated_length": 250.0, + "completions/mean_length": 192.75, + "completions/mean_terminated_length": 192.75, + "completions/min_length": 122.0, + "completions/min_terminated_length": 122.0, + "entropy": 0.5197336077690125, + "epoch": 0.285, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.230815410614014, + "learning_rate": 7.2e-07, + "loss": -0.15186546742916107, + "num_tokens": 484014.0, + "reward": -64.5050048828125, + "reward_std": 50.35145568847656, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -64.60499572753906, + "rewards/supergames_reward/std": 50.3514518737793, + "step": 57, + "step_time": 9.576232638006331 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 324.0, + "completions/max_terminated_length": 324.0, + "completions/mean_length": 249.0, + "completions/mean_terminated_length": 249.0, + "completions/min_length": 129.0, + "completions/min_terminated_length": 129.0, + "entropy": 0.42913514375686646, + "epoch": 0.29, + "frac_reward_zero_std": 0.0, + "grad_norm": 2.8544528484344482, + "learning_rate": 7.149999999999999e-07, + "loss": -0.056405920535326004, + "num_tokens": 494390.0, + "reward": 40.123748779296875, + "reward_std": 84.164794921875, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 40.02375030517578, + "rewards/supergames_reward/std": 84.164794921875, + "step": 58, + "step_time": 12.385353341000155 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 138.0, + "completions/max_terminated_length": 138.0, + "completions/mean_length": 107.25, + "completions/mean_terminated_length": 107.25, + "completions/min_length": 81.0, + "completions/min_terminated_length": 81.0, + "entropy": 0.34386488795280457, + "epoch": 0.295, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.6915154457092285, + "learning_rate": 7.1e-07, + "loss": 0.0865321159362793, + "num_tokens": 499816.0, + "reward": -8.244999885559082, + "reward_std": 4.681046962738037, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -8.345000267028809, + "rewards/supergames_reward/std": 4.681046962738037, + "step": 59, + "step_time": 5.42325339600211 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 187.0, + "completions/max_terminated_length": 187.0, + "completions/mean_length": 127.25, + "completions/mean_terminated_length": 127.25, + "completions/min_length": 97.0, + "completions/min_terminated_length": 97.0, + "entropy": 0.5430045127868652, + "epoch": 0.3, + "frac_reward_zero_std": 0.0, + "grad_norm": 5.635311126708984, + "learning_rate": 7.049999999999999e-07, + "loss": -0.05412375554442406, + "num_tokens": 505410.0, + "reward": 20.005001068115234, + "reward_std": 55.156005859375, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 19.904998779296875, + "rewards/supergames_reward/std": 55.156005859375, + "step": 60, + "step_time": 7.068010902003152 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 286.0, + "completions/max_terminated_length": 286.0, + "completions/mean_length": 192.125, + "completions/mean_terminated_length": 192.125, + "completions/min_length": 95.0, + "completions/min_terminated_length": 95.0, + "entropy": 0.4213869869709015, + "epoch": 0.305, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.4664618968963623, + "learning_rate": 7e-07, + "loss": -0.0495159812271595, + "num_tokens": 516283.0, + "reward": -47.212501525878906, + "reward_std": 91.2253646850586, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -47.3125, + "rewards/supergames_reward/std": 91.2253646850586, + "step": 61, + "step_time": 11.176304235996213 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 324.0, + "completions/max_terminated_length": 324.0, + "completions/mean_length": 171.0, + "completions/mean_terminated_length": 171.0, + "completions/min_length": 92.0, + "completions/min_terminated_length": 92.0, + "entropy": 0.4626811146736145, + "epoch": 0.31, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.949252128601074, + "learning_rate": 6.949999999999999e-07, + "loss": 0.192047581076622, + "num_tokens": 522251.0, + "reward": -15.092500686645508, + "reward_std": 49.82962417602539, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": -15.167499542236328, + "rewards/supergames_reward/std": 49.78091049194336, + "step": 62, + "step_time": 12.075224861997413 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.125, + "completions/max_length": 512.0, + "completions/max_terminated_length": 429.0, + "completions/mean_length": 283.5, + "completions/mean_terminated_length": 250.85714721679688, + "completions/min_length": 85.0, + "completions/min_terminated_length": 85.0, + "entropy": 0.4814216196537018, + "epoch": 0.315, + "frac_reward_zero_std": 0.0, + "grad_norm": 2.7918753623962402, + "learning_rate": 6.9e-07, + "loss": 0.15706156194210052, + "num_tokens": 533815.0, + "reward": -20.86625099182129, + "reward_std": 85.99250030517578, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": -20.94124984741211, + "rewards/supergames_reward/std": 85.96620178222656, + "step": 63, + "step_time": 19.448832260008203 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 189.0, + "completions/max_terminated_length": 189.0, + "completions/mean_length": 131.25, + "completions/mean_terminated_length": 131.25, + "completions/min_length": 106.0, + "completions/min_terminated_length": 106.0, + "entropy": 0.4929414987564087, + "epoch": 0.32, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.774101734161377, + "learning_rate": 6.85e-07, + "loss": -0.060691747814416885, + "num_tokens": 539465.0, + "reward": -21.0674991607666, + "reward_std": 6.892730236053467, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -21.167499542236328, + "rewards/supergames_reward/std": 6.892730236053467, + "step": 64, + "step_time": 7.141569407976931 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 454.0, + "completions/max_terminated_length": 454.0, + "completions/mean_length": 281.375, + "completions/mean_terminated_length": 281.375, + "completions/min_length": 107.0, + "completions/min_terminated_length": 107.0, + "entropy": 0.45866572856903076, + "epoch": 0.325, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.092717409133911, + "learning_rate": 6.800000000000001e-07, + "loss": 0.11694959551095963, + "num_tokens": 551012.0, + "reward": 49.938751220703125, + "reward_std": 79.1865005493164, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 49.8387451171875, + "rewards/supergames_reward/std": 79.1865005493164, + "step": 65, + "step_time": 17.06170882002334 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 203.0, + "completions/max_terminated_length": 203.0, + "completions/mean_length": 128.625, + "completions/mean_terminated_length": 128.625, + "completions/min_length": 88.0, + "completions/min_terminated_length": 88.0, + "entropy": 0.5005782246589661, + "epoch": 0.33, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.77616024017334, + "learning_rate": 6.75e-07, + "loss": 0.11242837458848953, + "num_tokens": 556577.0, + "reward": 48.875, + "reward_std": 49.13539123535156, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 48.775001525878906, + "rewards/supergames_reward/std": 49.13539123535156, + "step": 66, + "step_time": 7.695410129002994 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 436.0, + "completions/max_terminated_length": 436.0, + "completions/mean_length": 281.875, + "completions/mean_terminated_length": 281.875, + "completions/min_length": 133.0, + "completions/min_terminated_length": 133.0, + "entropy": 0.507097065448761, + "epoch": 0.335, + "frac_reward_zero_std": 0.0, + "grad_norm": 2.911184787750244, + "learning_rate": 6.7e-07, + "loss": -0.08151137083768845, + "num_tokens": 567120.0, + "reward": -12.886249542236328, + "reward_std": 84.04269409179688, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -12.986251831054688, + "rewards/supergames_reward/std": 84.04269409179688, + "step": 67, + "step_time": 16.409127942984924 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.125, + "completions/max_length": 512.0, + "completions/max_terminated_length": 501.0, + "completions/mean_length": 318.375, + "completions/mean_terminated_length": 290.71429443359375, + "completions/min_length": 65.0, + "completions/min_terminated_length": 65.0, + "entropy": 0.486402690410614, + "epoch": 0.34, + "frac_reward_zero_std": 0.0, + "grad_norm": 2.9461607933044434, + "learning_rate": 6.65e-07, + "loss": 0.2442178875207901, + "num_tokens": 578995.0, + "reward": 2.8387489318847656, + "reward_std": 98.47579956054688, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": 2.7637500762939453, + "rewards/supergames_reward/std": 98.44595336914062, + "step": 68, + "step_time": 19.32364999302081 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 291.0, + "completions/max_terminated_length": 291.0, + "completions/mean_length": 219.625, + "completions/mean_terminated_length": 219.625, + "completions/min_length": 117.0, + "completions/min_terminated_length": 117.0, + "entropy": 0.37856727838516235, + "epoch": 0.345, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.0204732418060303, + "learning_rate": 6.6e-07, + "loss": -0.08764916658401489, + "num_tokens": 589072.0, + "reward": -54.18375015258789, + "reward_std": 71.45012664794922, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -54.283748626708984, + "rewards/supergames_reward/std": 71.45013427734375, + "step": 69, + "step_time": 11.126611230982235 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.125, + "completions/max_length": 512.0, + "completions/max_terminated_length": 305.0, + "completions/mean_length": 254.75, + "completions/mean_terminated_length": 218.00001525878906, + "completions/min_length": 104.0, + "completions/min_terminated_length": 104.0, + "entropy": 0.46423614025115967, + "epoch": 0.35, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.0140154361724854, + "learning_rate": 6.55e-07, + "loss": 0.26797160506248474, + "num_tokens": 600438.0, + "reward": -32.04875183105469, + "reward_std": 69.86957550048828, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": -32.12375259399414, + "rewards/supergames_reward/std": 69.84178161621094, + "step": 70, + "step_time": 19.12617294798838 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 372.0, + "completions/max_terminated_length": 372.0, + "completions/mean_length": 194.125, + "completions/mean_terminated_length": 194.125, + "completions/min_length": 94.0, + "completions/min_terminated_length": 94.0, + "entropy": 0.3427293300628662, + "epoch": 0.355, + "frac_reward_zero_std": 0.0, + "grad_norm": 2.781383991241455, + "learning_rate": 6.5e-07, + "loss": 0.354524701833725, + "num_tokens": 610327.0, + "reward": -6.913749694824219, + "reward_std": 87.56071472167969, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": -6.988750457763672, + "rewards/supergames_reward/std": 87.53033447265625, + "step": 71, + "step_time": 14.041668199002743 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 197.0, + "completions/max_terminated_length": 197.0, + "completions/mean_length": 124.75, + "completions/mean_terminated_length": 124.75, + "completions/min_length": 93.0, + "completions/min_terminated_length": 93.0, + "entropy": 0.4455118477344513, + "epoch": 0.36, + "frac_reward_zero_std": 0.0, + "grad_norm": 5.726032733917236, + "learning_rate": 6.45e-07, + "loss": 0.18206289410591125, + "num_tokens": 615853.0, + "reward": 41.95124816894531, + "reward_std": 60.640811920166016, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 41.85124969482422, + "rewards/supergames_reward/std": 60.640811920166016, + "step": 72, + "step_time": 7.416493425989756 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 323.0, + "completions/max_terminated_length": 323.0, + "completions/mean_length": 147.375, + "completions/mean_terminated_length": 147.375, + "completions/min_length": 94.0, + "completions/min_terminated_length": 94.0, + "entropy": 0.480733186006546, + "epoch": 0.365, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.853621482849121, + "learning_rate": 6.4e-07, + "loss": 0.04558124020695686, + "num_tokens": 621584.0, + "reward": 14.884998321533203, + "reward_std": 46.25857162475586, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 14.78499984741211, + "rewards/supergames_reward/std": 46.25857162475586, + "step": 73, + "step_time": 11.95379540900467 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 283.0, + "completions/max_terminated_length": 283.0, + "completions/mean_length": 188.125, + "completions/mean_terminated_length": 188.125, + "completions/min_length": 116.0, + "completions/min_terminated_length": 116.0, + "entropy": 0.5526013374328613, + "epoch": 0.37, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.729883909225464, + "learning_rate": 6.35e-07, + "loss": 0.06268303096294403, + "num_tokens": 629545.0, + "reward": -68.05750274658203, + "reward_std": 59.44808578491211, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -68.15750122070312, + "rewards/supergames_reward/std": 59.44808578491211, + "step": 74, + "step_time": 10.64124580900534 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 141.0, + "completions/max_terminated_length": 141.0, + "completions/mean_length": 117.375, + "completions/mean_terminated_length": 117.375, + "completions/min_length": 100.0, + "completions/min_terminated_length": 100.0, + "entropy": 0.3744319677352905, + "epoch": 0.375, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.775154113769531, + "learning_rate": 6.3e-07, + "loss": -0.023284845054149628, + "num_tokens": 635052.0, + "reward": 3.9374990463256836, + "reward_std": 61.85771179199219, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": 3.862499237060547, + "rewards/supergames_reward/std": 61.80967712402344, + "step": 75, + "step_time": 5.432648951013107 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 268.0, + "completions/max_terminated_length": 268.0, + "completions/mean_length": 157.0, + "completions/mean_terminated_length": 157.0, + "completions/min_length": 104.0, + "completions/min_terminated_length": 104.0, + "entropy": 0.4822140336036682, + "epoch": 0.38, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.715704917907715, + "learning_rate": 6.249999999999999e-07, + "loss": -0.04918142035603523, + "num_tokens": 644644.0, + "reward": 48.64125061035156, + "reward_std": 78.0205078125, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": 48.56624984741211, + "rewards/supergames_reward/std": 77.96605682373047, + "step": 76, + "step_time": 10.348935816989979 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 380.0, + "completions/max_terminated_length": 380.0, + "completions/mean_length": 210.875, + "completions/mean_terminated_length": 210.875, + "completions/min_length": 126.0, + "completions/min_terminated_length": 126.0, + "entropy": 0.45722317695617676, + "epoch": 0.385, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.6057567596435547, + "learning_rate": 6.2e-07, + "loss": 0.058352649211883545, + "num_tokens": 652771.0, + "reward": -8.283750534057617, + "reward_std": 77.68436431884766, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -8.383749008178711, + "rewards/supergames_reward/std": 77.68437194824219, + "step": 77, + "step_time": 14.142948755004909 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 324.0, + "completions/max_terminated_length": 324.0, + "completions/mean_length": 188.25, + "completions/mean_terminated_length": 188.25, + "completions/min_length": 68.0, + "completions/min_terminated_length": 68.0, + "entropy": 0.33380764722824097, + "epoch": 0.39, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.2239863872528076, + "learning_rate": 6.149999999999999e-07, + "loss": -0.04592633992433548, + "num_tokens": 662653.0, + "reward": 11.153749465942383, + "reward_std": 89.88329315185547, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 11.053749084472656, + "rewards/supergames_reward/std": 89.88329315185547, + "step": 78, + "step_time": 12.277474621019792 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 143.0, + "completions/max_terminated_length": 143.0, + "completions/mean_length": 108.875, + "completions/mean_terminated_length": 108.875, + "completions/min_length": 85.0, + "completions/min_terminated_length": 85.0, + "entropy": 0.3518812656402588, + "epoch": 0.395, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.967776298522949, + "learning_rate": 6.1e-07, + "loss": -0.10800496488809586, + "num_tokens": 668076.0, + "reward": 70.33250427246094, + "reward_std": 49.9145393371582, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 70.23249816894531, + "rewards/supergames_reward/std": 49.91453552246094, + "step": 79, + "step_time": 5.52714040101273 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 184.0, + "completions/max_terminated_length": 184.0, + "completions/mean_length": 152.125, + "completions/mean_terminated_length": 152.125, + "completions/min_length": 104.0, + "completions/min_terminated_length": 104.0, + "entropy": 0.4699896574020386, + "epoch": 0.4, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.561161994934082, + "learning_rate": 6.049999999999999e-07, + "loss": 0.09505834430456161, + "num_tokens": 675701.0, + "reward": -20.25625228881836, + "reward_std": 53.78542709350586, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -20.35624885559082, + "rewards/supergames_reward/std": 53.78542709350586, + "step": 80, + "step_time": 7.155081019998761 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.125, + "completions/max_length": 512.0, + "completions/max_terminated_length": 506.0, + "completions/mean_length": 288.875, + "completions/mean_terminated_length": 257.0, + "completions/min_length": 115.0, + "completions/min_terminated_length": 115.0, + "entropy": 0.4086494445800781, + "epoch": 0.405, + "frac_reward_zero_std": 0.0, + "grad_norm": 2.3719611167907715, + "learning_rate": 6e-07, + "loss": 0.04727627709507942, + "num_tokens": 687332.0, + "reward": -37.20750045776367, + "reward_std": 91.7676010131836, + "rewards/json_format_reward/mean": 0.05000000074505806, + "rewards/json_format_reward/std": 0.09258200973272324, + "rewards/supergames_reward/mean": -37.25749969482422, + "rewards/supergames_reward/std": 91.72847747802734, + "step": 81, + "step_time": 19.231399144016905 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 376.0, + "completions/max_terminated_length": 376.0, + "completions/mean_length": 227.75, + "completions/mean_terminated_length": 227.75, + "completions/min_length": 39.0, + "completions/min_terminated_length": 39.0, + "entropy": 0.43233707547187805, + "epoch": 0.41, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.3391714096069336, + "learning_rate": 5.949999999999999e-07, + "loss": -0.25524550676345825, + "num_tokens": 698474.0, + "reward": -3.7287511825561523, + "reward_std": 91.25379943847656, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -3.828749656677246, + "rewards/supergames_reward/std": 91.25379943847656, + "step": 82, + "step_time": 14.329176368017215 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 299.0, + "completions/max_terminated_length": 299.0, + "completions/mean_length": 178.125, + "completions/mean_terminated_length": 178.125, + "completions/min_length": 91.0, + "completions/min_terminated_length": 91.0, + "entropy": 0.4293636977672577, + "epoch": 0.415, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.7271742820739746, + "learning_rate": 5.9e-07, + "loss": 0.09273672848939896, + "num_tokens": 708243.0, + "reward": 12.802498817443848, + "reward_std": 86.23526000976562, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 12.702500343322754, + "rewards/supergames_reward/std": 86.23526000976562, + "step": 83, + "step_time": 11.419686011999147 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 425.0, + "completions/max_terminated_length": 425.0, + "completions/mean_length": 255.375, + "completions/mean_terminated_length": 255.375, + "completions/min_length": 172.0, + "completions/min_terminated_length": 172.0, + "entropy": 0.38068246841430664, + "epoch": 0.42, + "frac_reward_zero_std": 0.0, + "grad_norm": 2.6244184970855713, + "learning_rate": 5.849999999999999e-07, + "loss": 0.02006213366985321, + "num_tokens": 718606.0, + "reward": 37.803749084472656, + "reward_std": 57.58624267578125, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 37.70375061035156, + "rewards/supergames_reward/std": 57.58624267578125, + "step": 84, + "step_time": 15.975198492989875 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 116.0, + "completions/max_terminated_length": 116.0, + "completions/mean_length": 108.125, + "completions/mean_terminated_length": 108.125, + "completions/min_length": 87.0, + "completions/min_terminated_length": 87.0, + "entropy": 0.339572012424469, + "epoch": 0.425, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.754759311676025, + "learning_rate": 5.8e-07, + "loss": -0.011015941388905048, + "num_tokens": 723983.0, + "reward": -3.8212504386901855, + "reward_std": 61.624786376953125, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -3.921250820159912, + "rewards/supergames_reward/std": 61.624786376953125, + "step": 85, + "step_time": 4.542777584982105 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 447.0, + "completions/max_terminated_length": 447.0, + "completions/mean_length": 291.375, + "completions/mean_terminated_length": 291.375, + "completions/min_length": 163.0, + "completions/min_terminated_length": 163.0, + "entropy": 0.5846173763275146, + "epoch": 0.43, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.55724835395813, + "learning_rate": 5.749999999999999e-07, + "loss": -0.04981120675802231, + "num_tokens": 735642.0, + "reward": 50.07499694824219, + "reward_std": 92.6283187866211, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": 50.0, + "rewards/supergames_reward/std": 92.5820083618164, + "step": 86, + "step_time": 16.777178087009815 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 232.0, + "completions/max_terminated_length": 232.0, + "completions/mean_length": 192.0, + "completions/mean_terminated_length": 192.0, + "completions/min_length": 139.0, + "completions/min_terminated_length": 139.0, + "entropy": 0.5139474272727966, + "epoch": 0.435, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.8860411643981934, + "learning_rate": 5.699999999999999e-07, + "loss": -0.02194221317768097, + "num_tokens": 743602.0, + "reward": -52.522499084472656, + "reward_std": 73.8686294555664, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -52.62249755859375, + "rewards/supergames_reward/std": 73.86863708496094, + "step": 87, + "step_time": 8.772893911984283 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 266.0, + "completions/max_terminated_length": 266.0, + "completions/mean_length": 146.875, + "completions/mean_terminated_length": 146.875, + "completions/min_length": 89.0, + "completions/min_terminated_length": 89.0, + "entropy": 0.3428497910499573, + "epoch": 0.44, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.509483337402344, + "learning_rate": 5.649999999999999e-07, + "loss": 0.19857533276081085, + "num_tokens": 753065.0, + "reward": 63.63874816894531, + "reward_std": 67.54963684082031, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 63.53874969482422, + "rewards/supergames_reward/std": 67.54963684082031, + "step": 88, + "step_time": 10.230529835011112 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 446.0, + "completions/max_terminated_length": 446.0, + "completions/mean_length": 270.625, + "completions/mean_terminated_length": 270.625, + "completions/min_length": 103.0, + "completions/min_terminated_length": 103.0, + "entropy": 0.4417419731616974, + "epoch": 0.445, + "frac_reward_zero_std": 0.0, + "grad_norm": 2.973951816558838, + "learning_rate": 5.6e-07, + "loss": -0.00031440958264283836, + "num_tokens": 764534.0, + "reward": -9.736251831054688, + "reward_std": 92.25879669189453, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -9.836250305175781, + "rewards/supergames_reward/std": 92.25880432128906, + "step": 89, + "step_time": 17.40752330099349 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 398.0, + "completions/max_terminated_length": 398.0, + "completions/mean_length": 227.375, + "completions/mean_terminated_length": 227.375, + "completions/min_length": 98.0, + "completions/min_terminated_length": 98.0, + "entropy": 0.4732024073600769, + "epoch": 0.45, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.2560815811157227, + "learning_rate": 5.55e-07, + "loss": 0.03820464387536049, + "num_tokens": 772809.0, + "reward": 5.982499599456787, + "reward_std": 58.833797454833984, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 5.8824992179870605, + "rewards/supergames_reward/std": 58.83380126953125, + "step": 90, + "step_time": 15.007269965979503 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 269.0, + "completions/max_terminated_length": 269.0, + "completions/mean_length": 207.5, + "completions/mean_terminated_length": 207.5, + "completions/min_length": 92.0, + "completions/min_terminated_length": 92.0, + "entropy": 0.42038559913635254, + "epoch": 0.455, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.3649117946624756, + "learning_rate": 5.5e-07, + "loss": 0.1510048508644104, + "num_tokens": 782805.0, + "reward": -5.61500358581543, + "reward_std": 82.67152404785156, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -5.714998245239258, + "rewards/supergames_reward/std": 82.67152404785156, + "step": 91, + "step_time": 10.386137172987219 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 416.0, + "completions/max_terminated_length": 416.0, + "completions/mean_length": 218.375, + "completions/mean_terminated_length": 218.375, + "completions/min_length": 95.0, + "completions/min_terminated_length": 95.0, + "entropy": 0.3971734642982483, + "epoch": 0.46, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.433772563934326, + "learning_rate": 5.45e-07, + "loss": -0.3027213513851166, + "num_tokens": 793840.0, + "reward": -29.06500244140625, + "reward_std": 91.68325805664062, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -29.16499900817871, + "rewards/supergames_reward/std": 91.68325805664062, + "step": 92, + "step_time": 15.790611553995404 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 345.0, + "completions/max_terminated_length": 345.0, + "completions/mean_length": 206.625, + "completions/mean_terminated_length": 206.625, + "completions/min_length": 90.0, + "completions/min_terminated_length": 90.0, + "entropy": 0.5906907320022583, + "epoch": 0.465, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.293088912963867, + "learning_rate": 5.4e-07, + "loss": 0.09760032594203949, + "num_tokens": 803821.0, + "reward": -72.73999786376953, + "reward_std": 50.72484588623047, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": -72.81500244140625, + "rewards/supergames_reward/std": 50.7094841003418, + "step": 93, + "step_time": 13.043119941983605 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 143.0, + "completions/max_terminated_length": 143.0, + "completions/mean_length": 115.75, + "completions/mean_terminated_length": 115.75, + "completions/min_length": 80.0, + "completions/min_terminated_length": 80.0, + "entropy": 0.4233115315437317, + "epoch": 0.47, + "frac_reward_zero_std": 0.0, + "grad_norm": 5.58886194229126, + "learning_rate": 5.35e-07, + "loss": -0.04040650278329849, + "num_tokens": 809299.0, + "reward": -3.3050003051757812, + "reward_std": 24.443714141845703, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -3.405000686645508, + "rewards/supergames_reward/std": 24.443714141845703, + "step": 94, + "step_time": 5.512874762003776 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 122.0, + "completions/max_terminated_length": 122.0, + "completions/mean_length": 103.125, + "completions/mean_terminated_length": 103.125, + "completions/min_length": 86.0, + "completions/min_terminated_length": 86.0, + "entropy": 0.4055764377117157, + "epoch": 0.475, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.981549263000488, + "learning_rate": 5.3e-07, + "loss": -0.07588938623666763, + "num_tokens": 814716.0, + "reward": 60.06999969482422, + "reward_std": 42.708473205566406, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 59.970001220703125, + "rewards/supergames_reward/std": 42.708473205566406, + "step": 95, + "step_time": 4.766389057011111 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 475.0, + "completions/max_terminated_length": 475.0, + "completions/mean_length": 218.375, + "completions/mean_terminated_length": 218.375, + "completions/min_length": 81.0, + "completions/min_terminated_length": 81.0, + "entropy": 0.408913254737854, + "epoch": 0.48, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.79402756690979, + "learning_rate": 5.25e-07, + "loss": -0.3087541460990906, + "num_tokens": 824791.0, + "reward": -15.042500495910645, + "reward_std": 84.37246704101562, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -15.142499923706055, + "rewards/supergames_reward/std": 84.37246704101562, + "step": 96, + "step_time": 17.779843941010768 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 154.0, + "completions/max_terminated_length": 154.0, + "completions/mean_length": 121.625, + "completions/mean_terminated_length": 121.625, + "completions/min_length": 80.0, + "completions/min_terminated_length": 80.0, + "entropy": 0.529015839099884, + "epoch": 0.485, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.142816066741943, + "learning_rate": 5.2e-07, + "loss": 0.07216037809848785, + "num_tokens": 830300.0, + "reward": -32.743751525878906, + "reward_std": 54.61775588989258, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": -32.81875228881836, + "rewards/supergames_reward/std": 54.58256149291992, + "step": 97, + "step_time": 5.913989890017547 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 126.0, + "completions/max_terminated_length": 126.0, + "completions/mean_length": 95.625, + "completions/mean_terminated_length": 95.625, + "completions/min_length": 74.0, + "completions/min_terminated_length": 74.0, + "entropy": 0.33743274211883545, + "epoch": 0.49, + "frac_reward_zero_std": 0.0, + "grad_norm": 5.197273254394531, + "learning_rate": 5.149999999999999e-07, + "loss": 0.07500407844781876, + "num_tokens": 835633.0, + "reward": 11.00999927520752, + "reward_std": 62.73848342895508, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 10.910000801086426, + "rewards/supergames_reward/std": 62.73848342895508, + "step": 98, + "step_time": 4.9155233050114475 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 149.0, + "completions/max_terminated_length": 149.0, + "completions/mean_length": 108.5, + "completions/mean_terminated_length": 108.5, + "completions/min_length": 86.0, + "completions/min_terminated_length": 86.0, + "entropy": 0.3106803297996521, + "epoch": 0.495, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.864002227783203, + "learning_rate": 5.1e-07, + "loss": -0.1206185445189476, + "num_tokens": 841045.0, + "reward": 14.238749504089355, + "reward_std": 51.77016830444336, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 14.138750076293945, + "rewards/supergames_reward/std": 51.77016830444336, + "step": 99, + "step_time": 5.7507751359953545 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 149.0, + "completions/max_terminated_length": 149.0, + "completions/mean_length": 97.125, + "completions/mean_terminated_length": 97.125, + "completions/min_length": 73.0, + "completions/min_terminated_length": 73.0, + "entropy": 0.35160186886787415, + "epoch": 0.5, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.915731906890869, + "learning_rate": 5.049999999999999e-07, + "loss": -0.07951541244983673, + "num_tokens": 846398.0, + "reward": 67.78125, + "reward_std": 38.15436553955078, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 67.68124389648438, + "rewards/supergames_reward/std": 38.15436553955078, + "step": 100, + "step_time": 5.994720440998208 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 320.0, + "completions/max_terminated_length": 320.0, + "completions/mean_length": 208.75, + "completions/mean_terminated_length": 208.75, + "completions/min_length": 95.0, + "completions/min_terminated_length": 95.0, + "entropy": 0.452095091342926, + "epoch": 0.505, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.065176963806152, + "learning_rate": 5e-07, + "loss": -0.1397111564874649, + "num_tokens": 857420.0, + "reward": -5.801251411437988, + "reward_std": 73.56117248535156, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -5.90125036239624, + "rewards/supergames_reward/std": 73.5611801147461, + "step": 101, + "step_time": 12.28178753197426 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 230.0, + "completions/max_terminated_length": 230.0, + "completions/mean_length": 171.375, + "completions/mean_terminated_length": 171.375, + "completions/min_length": 95.0, + "completions/min_terminated_length": 95.0, + "entropy": 0.4138447642326355, + "epoch": 0.51, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.998823404312134, + "learning_rate": 4.95e-07, + "loss": -0.08941879868507385, + "num_tokens": 867103.0, + "reward": 92.24749755859375, + "reward_std": 14.540005683898926, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 92.14749908447266, + "rewards/supergames_reward/std": 14.540006637573242, + "step": 102, + "step_time": 8.992682139010867 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 311.0, + "completions/max_terminated_length": 311.0, + "completions/mean_length": 163.0, + "completions/mean_terminated_length": 163.0, + "completions/min_length": 74.0, + "completions/min_terminated_length": 74.0, + "entropy": 0.4182547330856323, + "epoch": 0.515, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.419595241546631, + "learning_rate": 4.9e-07, + "loss": -0.19304415583610535, + "num_tokens": 876679.0, + "reward": 79.13249969482422, + "reward_std": 59.3050422668457, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 79.03250122070312, + "rewards/supergames_reward/std": 59.30504608154297, + "step": 103, + "step_time": 11.836970244999975 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 253.0, + "completions/max_terminated_length": 253.0, + "completions/mean_length": 155.625, + "completions/mean_terminated_length": 155.625, + "completions/min_length": 94.0, + "completions/min_terminated_length": 94.0, + "entropy": 0.49889081716537476, + "epoch": 0.52, + "frac_reward_zero_std": 0.0, + "grad_norm": 5.463719367980957, + "learning_rate": 4.85e-07, + "loss": -0.04786720871925354, + "num_tokens": 884348.0, + "reward": 48.616249084472656, + "reward_std": 73.30741119384766, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 48.51625061035156, + "rewards/supergames_reward/std": 73.30741119384766, + "step": 104, + "step_time": 9.579507841990562 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 380.0, + "completions/max_terminated_length": 380.0, + "completions/mean_length": 269.5, + "completions/mean_terminated_length": 269.5, + "completions/min_length": 101.0, + "completions/min_terminated_length": 101.0, + "entropy": 0.5286482572555542, + "epoch": 0.525, + "frac_reward_zero_std": 0.0, + "grad_norm": 2.993295431137085, + "learning_rate": 4.8e-07, + "loss": -0.08071611076593399, + "num_tokens": 894856.0, + "reward": 34.970001220703125, + "reward_std": 85.324462890625, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 34.869998931884766, + "rewards/supergames_reward/std": 85.324462890625, + "step": 105, + "step_time": 14.295730627985904 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 261.0, + "completions/max_terminated_length": 261.0, + "completions/mean_length": 203.25, + "completions/mean_terminated_length": 203.25, + "completions/min_length": 91.0, + "completions/min_terminated_length": 91.0, + "entropy": 0.6049264073371887, + "epoch": 0.53, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.787306785583496, + "learning_rate": 4.7499999999999995e-07, + "loss": 0.19928883016109467, + "num_tokens": 902922.0, + "reward": 32.44499969482422, + "reward_std": 70.4663314819336, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": 32.369998931884766, + "rewards/supergames_reward/std": 70.41261291503906, + "step": 106, + "step_time": 9.84894504098338 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 279.0, + "completions/max_terminated_length": 279.0, + "completions/mean_length": 194.0, + "completions/mean_terminated_length": 194.0, + "completions/min_length": 109.0, + "completions/min_terminated_length": 109.0, + "entropy": 0.5115755200386047, + "epoch": 0.535, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.277207374572754, + "learning_rate": 4.6999999999999995e-07, + "loss": 0.09118300676345825, + "num_tokens": 910906.0, + "reward": 40.432498931884766, + "reward_std": 48.54548263549805, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 40.33250045776367, + "rewards/supergames_reward/std": 48.54548263549805, + "step": 107, + "step_time": 10.466603949986165 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 332.0, + "completions/max_terminated_length": 332.0, + "completions/mean_length": 195.875, + "completions/mean_terminated_length": 195.875, + "completions/min_length": 87.0, + "completions/min_terminated_length": 87.0, + "entropy": 0.4192962944507599, + "epoch": 0.54, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.308061122894287, + "learning_rate": 4.65e-07, + "loss": 0.28064602613449097, + "num_tokens": 920745.0, + "reward": -21.982500076293945, + "reward_std": 66.31195068359375, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -22.08249855041504, + "rewards/supergames_reward/std": 66.31195068359375, + "step": 108, + "step_time": 12.595848233992001 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 269.0, + "completions/max_terminated_length": 269.0, + "completions/mean_length": 195.125, + "completions/mean_terminated_length": 195.125, + "completions/min_length": 111.0, + "completions/min_terminated_length": 111.0, + "entropy": 0.37523961067199707, + "epoch": 0.545, + "frac_reward_zero_std": 0.0, + "grad_norm": 2.802706718444824, + "learning_rate": 4.6e-07, + "loss": -0.06614465266466141, + "num_tokens": 930618.0, + "reward": 71.75, + "reward_std": 51.14524459838867, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 71.6500015258789, + "rewards/supergames_reward/std": 51.1452522277832, + "step": 109, + "step_time": 10.33884758799104 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 331.0, + "completions/max_terminated_length": 331.0, + "completions/mean_length": 167.75, + "completions/mean_terminated_length": 167.75, + "completions/min_length": 89.0, + "completions/min_terminated_length": 89.0, + "entropy": 0.4339042007923126, + "epoch": 0.55, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.254003047943115, + "learning_rate": 4.55e-07, + "loss": 0.34405598044395447, + "num_tokens": 941256.0, + "reward": 99.21875, + "reward_std": 2.492553234100342, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 99.1187515258789, + "rewards/supergames_reward/std": 2.4925525188446045, + "step": 110, + "step_time": 12.657525141985388 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 285.0, + "completions/max_terminated_length": 285.0, + "completions/mean_length": 211.0, + "completions/mean_terminated_length": 211.0, + "completions/min_length": 112.0, + "completions/min_terminated_length": 112.0, + "entropy": 0.46809861063957214, + "epoch": 0.555, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.811166286468506, + "learning_rate": 4.5e-07, + "loss": 0.045113153755664825, + "num_tokens": 951296.0, + "reward": 55.27750015258789, + "reward_std": 47.90687561035156, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 55.1775016784668, + "rewards/supergames_reward/std": 47.9068717956543, + "step": 111, + "step_time": 10.91118389699841 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 145.0, + "completions/max_terminated_length": 145.0, + "completions/mean_length": 108.375, + "completions/mean_terminated_length": 108.375, + "completions/min_length": 86.0, + "completions/min_terminated_length": 86.0, + "entropy": 0.34181955456733704, + "epoch": 0.56, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.537296772003174, + "learning_rate": 4.45e-07, + "loss": -0.052413541823625565, + "num_tokens": 956675.0, + "reward": 47.56624984741211, + "reward_std": 49.185157775878906, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 47.466251373291016, + "rewards/supergames_reward/std": 49.18516159057617, + "step": 112, + "step_time": 5.552288047998445 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 245.0, + "completions/max_terminated_length": 245.0, + "completions/mean_length": 189.125, + "completions/mean_terminated_length": 189.125, + "completions/min_length": 102.0, + "completions/min_terminated_length": 102.0, + "entropy": 0.5034605860710144, + "epoch": 0.565, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.349876642227173, + "learning_rate": 4.3999999999999997e-07, + "loss": -0.08186715841293335, + "num_tokens": 964636.0, + "reward": 81.32374572753906, + "reward_std": 36.85074996948242, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 81.2237548828125, + "rewards/supergames_reward/std": 36.85074996948242, + "step": 113, + "step_time": 9.267611294984818 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 178.0, + "completions/max_terminated_length": 178.0, + "completions/mean_length": 111.25, + "completions/mean_terminated_length": 111.25, + "completions/min_length": 67.0, + "completions/min_terminated_length": 67.0, + "entropy": 0.3229002058506012, + "epoch": 0.57, + "frac_reward_zero_std": 0.0, + "grad_norm": 5.245575428009033, + "learning_rate": 4.3499999999999996e-07, + "loss": 0.06544198095798492, + "num_tokens": 970070.0, + "reward": 63.1775016784668, + "reward_std": 52.30624008178711, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 63.07749938964844, + "rewards/supergames_reward/std": 52.30624008178711, + "step": 114, + "step_time": 6.731139309995342 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 224.0, + "completions/max_terminated_length": 224.0, + "completions/mean_length": 170.0, + "completions/mean_terminated_length": 170.0, + "completions/min_length": 94.0, + "completions/min_terminated_length": 94.0, + "entropy": 0.3274807929992676, + "epoch": 0.575, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.5544140338897705, + "learning_rate": 4.2999999999999996e-07, + "loss": -0.06953569501638412, + "num_tokens": 979702.0, + "reward": 25.62125015258789, + "reward_std": 64.52003479003906, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 25.521244049072266, + "rewards/supergames_reward/std": 64.52003479003906, + "step": 115, + "step_time": 8.702618776995223 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 427.0, + "completions/max_terminated_length": 427.0, + "completions/mean_length": 209.5, + "completions/mean_terminated_length": 209.5, + "completions/min_length": 100.0, + "completions/min_terminated_length": 100.0, + "entropy": 0.4970880150794983, + "epoch": 0.58, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.4932165145874023, + "learning_rate": 4.2499999999999995e-07, + "loss": -0.2281273603439331, + "num_tokens": 990690.0, + "reward": 50.54750061035156, + "reward_std": 70.5126953125, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 50.4474983215332, + "rewards/supergames_reward/std": 70.51270294189453, + "step": 116, + "step_time": 16.204386608995264 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 228.0, + "completions/max_terminated_length": 228.0, + "completions/mean_length": 132.375, + "completions/mean_terminated_length": 132.375, + "completions/min_length": 85.0, + "completions/min_terminated_length": 85.0, + "entropy": 0.4380227029323578, + "epoch": 0.585, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.089909076690674, + "learning_rate": 4.1999999999999995e-07, + "loss": 0.12121254950761795, + "num_tokens": 996333.0, + "reward": -15.333749771118164, + "reward_std": 47.200164794921875, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -15.433748245239258, + "rewards/supergames_reward/std": 47.200164794921875, + "step": 117, + "step_time": 8.546233230998041 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 323.0, + "completions/max_terminated_length": 323.0, + "completions/mean_length": 137.625, + "completions/mean_terminated_length": 137.625, + "completions/min_length": 90.0, + "completions/min_terminated_length": 90.0, + "entropy": 0.4177803695201874, + "epoch": 0.59, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.8712472915649414, + "learning_rate": 4.1499999999999994e-07, + "loss": 0.14192955195903778, + "num_tokens": 1002010.0, + "reward": -30.700000762939453, + "reward_std": 17.224035263061523, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -30.799999237060547, + "rewards/supergames_reward/std": 17.224035263061523, + "step": 118, + "step_time": 11.878434990998358 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 453.0, + "completions/max_terminated_length": 453.0, + "completions/mean_length": 238.125, + "completions/mean_terminated_length": 238.125, + "completions/min_length": 109.0, + "completions/min_terminated_length": 109.0, + "entropy": 0.5640217065811157, + "epoch": 0.595, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.064737319946289, + "learning_rate": 4.0999999999999994e-07, + "loss": -0.023621462285518646, + "num_tokens": 1012259.0, + "reward": 43.683746337890625, + "reward_std": 70.4210205078125, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 43.5837516784668, + "rewards/supergames_reward/std": 70.4210205078125, + "step": 119, + "step_time": 16.829514264973113 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 236.0, + "completions/max_terminated_length": 236.0, + "completions/mean_length": 193.75, + "completions/mean_terminated_length": 193.75, + "completions/min_length": 151.0, + "completions/min_terminated_length": 151.0, + "entropy": 0.5795091390609741, + "epoch": 0.6, + "frac_reward_zero_std": 0.0, + "grad_norm": 5.423007965087891, + "learning_rate": 4.05e-07, + "loss": -0.0666111558675766, + "num_tokens": 1020241.0, + "reward": -85.86000061035156, + "reward_std": 27.4442081451416, + "rewards/json_format_reward/mean": 0.05000000074505806, + "rewards/json_format_reward/std": 0.09258200973272324, + "rewards/supergames_reward/mean": -85.90999603271484, + "rewards/supergames_reward/std": 27.41469955444336, + "step": 120, + "step_time": 8.95496519000153 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 331.0, + "completions/max_terminated_length": 331.0, + "completions/mean_length": 231.0, + "completions/mean_terminated_length": 231.0, + "completions/min_length": 165.0, + "completions/min_terminated_length": 165.0, + "entropy": 0.3767699897289276, + "epoch": 0.605, + "frac_reward_zero_std": 0.0, + "grad_norm": 2.7434473037719727, + "learning_rate": 4e-07, + "loss": -0.1165132150053978, + "num_tokens": 1031465.0, + "reward": 61.45624923706055, + "reward_std": 72.21080780029297, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 61.35625076293945, + "rewards/supergames_reward/std": 72.2108154296875, + "step": 121, + "step_time": 12.588820598000893 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 297.0, + "completions/max_terminated_length": 297.0, + "completions/mean_length": 223.375, + "completions/mean_terminated_length": 223.375, + "completions/min_length": 112.0, + "completions/min_terminated_length": 112.0, + "entropy": 0.4537277817726135, + "epoch": 0.61, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.4382436275482178, + "learning_rate": 3.95e-07, + "loss": -0.1038316935300827, + "num_tokens": 1042548.0, + "reward": 5.658749580383301, + "reward_std": 83.77420043945312, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 5.558750152587891, + "rewards/supergames_reward/std": 83.77420043945312, + "step": 122, + "step_time": 11.39525846898323 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 456.0, + "completions/max_terminated_length": 456.0, + "completions/mean_length": 314.25, + "completions/mean_terminated_length": 314.25, + "completions/min_length": 95.0, + "completions/min_terminated_length": 95.0, + "entropy": 0.4113953411579132, + "epoch": 0.615, + "frac_reward_zero_std": 0.0, + "grad_norm": 2.449777841567993, + "learning_rate": 3.8999999999999997e-07, + "loss": -0.1303664743900299, + "num_tokens": 1054422.0, + "reward": 35.98249816894531, + "reward_std": 88.6131362915039, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 35.88249969482422, + "rewards/supergames_reward/std": 88.6131362915039, + "step": 123, + "step_time": 17.001118954998674 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 242.0, + "completions/max_terminated_length": 242.0, + "completions/mean_length": 172.5, + "completions/mean_terminated_length": 172.5, + "completions/min_length": 137.0, + "completions/min_terminated_length": 137.0, + "entropy": 0.4670729339122772, + "epoch": 0.62, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.951946973800659, + "learning_rate": 3.8499999999999997e-07, + "loss": 0.05888795852661133, + "num_tokens": 1062210.0, + "reward": 43.368751525878906, + "reward_std": 40.30733871459961, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 43.26874542236328, + "rewards/supergames_reward/std": 40.307342529296875, + "step": 124, + "step_time": 9.142582255997695 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 148.0, + "completions/max_terminated_length": 148.0, + "completions/mean_length": 106.625, + "completions/mean_terminated_length": 106.625, + "completions/min_length": 79.0, + "completions/min_terminated_length": 79.0, + "entropy": 0.30219921469688416, + "epoch": 0.625, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.391322135925293, + "learning_rate": 3.7999999999999996e-07, + "loss": 0.14367049932479858, + "num_tokens": 1067599.0, + "reward": 42.63374710083008, + "reward_std": 58.20897674560547, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 42.533748626708984, + "rewards/supergames_reward/std": 58.208984375, + "step": 125, + "step_time": 5.636063549987739 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 243.0, + "completions/max_terminated_length": 243.0, + "completions/mean_length": 179.25, + "completions/mean_terminated_length": 179.25, + "completions/min_length": 134.0, + "completions/min_terminated_length": 134.0, + "entropy": 0.4834887683391571, + "epoch": 0.63, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.9391286373138428, + "learning_rate": 3.75e-07, + "loss": -0.06614185124635696, + "num_tokens": 1075457.0, + "reward": 8.302498817443848, + "reward_std": 75.10364532470703, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 8.202500343322754, + "rewards/supergames_reward/std": 75.10364532470703, + "step": 126, + "step_time": 9.130600645992672 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 196.0, + "completions/max_terminated_length": 196.0, + "completions/mean_length": 157.0, + "completions/mean_terminated_length": 157.0, + "completions/min_length": 117.0, + "completions/min_terminated_length": 117.0, + "entropy": 0.4801807701587677, + "epoch": 0.635, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.425615310668945, + "learning_rate": 3.7e-07, + "loss": -0.08830522745847702, + "num_tokens": 1083129.0, + "reward": -10.563751220703125, + "reward_std": 64.00443267822266, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -10.663749694824219, + "rewards/supergames_reward/std": 64.00444030761719, + "step": 127, + "step_time": 7.4972667430120055 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 412.0, + "completions/max_terminated_length": 412.0, + "completions/mean_length": 239.25, + "completions/mean_terminated_length": 239.25, + "completions/min_length": 89.0, + "completions/min_terminated_length": 89.0, + "entropy": 0.4667063355445862, + "epoch": 0.64, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.199113130569458, + "learning_rate": 3.65e-07, + "loss": -0.043856702744960785, + "num_tokens": 1094355.0, + "reward": -55.89250183105469, + "reward_std": 53.05980682373047, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -55.99250030517578, + "rewards/supergames_reward/std": 53.05980682373047, + "step": 128, + "step_time": 15.429580625001108 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 370.0, + "completions/max_terminated_length": 370.0, + "completions/mean_length": 244.75, + "completions/mean_terminated_length": 244.75, + "completions/min_length": 127.0, + "completions/min_terminated_length": 127.0, + "entropy": 0.57573401927948, + "epoch": 0.645, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.6800079345703125, + "learning_rate": 3.6e-07, + "loss": 0.09521033614873886, + "num_tokens": 1105633.0, + "reward": -23.35250473022461, + "reward_std": 102.31401062011719, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -23.452499389648438, + "rewards/supergames_reward/std": 102.31402587890625, + "step": 129, + "step_time": 13.920327747007832 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 258.0, + "completions/max_terminated_length": 258.0, + "completions/mean_length": 188.25, + "completions/mean_terminated_length": 188.25, + "completions/min_length": 131.0, + "completions/min_terminated_length": 131.0, + "entropy": 0.5341438055038452, + "epoch": 0.65, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.97955060005188, + "learning_rate": 3.55e-07, + "loss": -0.09512650966644287, + "num_tokens": 1113571.0, + "reward": 29.181249618530273, + "reward_std": 80.10585021972656, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 29.081249237060547, + "rewards/supergames_reward/std": 80.10585021972656, + "step": 130, + "step_time": 9.69286120700417 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 257.0, + "completions/max_terminated_length": 257.0, + "completions/mean_length": 180.75, + "completions/mean_terminated_length": 180.75, + "completions/min_length": 136.0, + "completions/min_terminated_length": 136.0, + "entropy": 0.46591487526893616, + "epoch": 0.655, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.3518130779266357, + "learning_rate": 3.5e-07, + "loss": 0.04714229330420494, + "num_tokens": 1121457.0, + "reward": -8.287501335144043, + "reward_std": 76.35875701904297, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -8.387499809265137, + "rewards/supergames_reward/std": 76.35875701904297, + "step": 131, + "step_time": 9.636637121002423 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 190.0, + "completions/max_terminated_length": 190.0, + "completions/mean_length": 147.75, + "completions/mean_terminated_length": 147.75, + "completions/min_length": 90.0, + "completions/min_terminated_length": 90.0, + "entropy": 0.4407735764980316, + "epoch": 0.66, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.498741626739502, + "learning_rate": 3.45e-07, + "loss": -0.03996007889509201, + "num_tokens": 1129095.0, + "reward": 0.1999988555908203, + "reward_std": 84.63554382324219, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": 0.125, + "rewards/supergames_reward/std": 84.6017074584961, + "step": 132, + "step_time": 7.510833573003765 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 358.0, + "completions/max_terminated_length": 358.0, + "completions/mean_length": 213.875, + "completions/mean_terminated_length": 213.875, + "completions/min_length": 137.0, + "completions/min_terminated_length": 137.0, + "entropy": 0.5186149477958679, + "epoch": 0.665, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.3987107276916504, + "learning_rate": 3.4000000000000003e-07, + "loss": 0.1863594949245453, + "num_tokens": 1139110.0, + "reward": -21.328752517700195, + "reward_std": 88.9102783203125, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -21.428749084472656, + "rewards/supergames_reward/std": 88.9102783203125, + "step": 133, + "step_time": 13.393839450000087 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 230.0, + "completions/max_terminated_length": 230.0, + "completions/mean_length": 169.875, + "completions/mean_terminated_length": 169.875, + "completions/min_length": 121.0, + "completions/min_terminated_length": 121.0, + "entropy": 0.49445784091949463, + "epoch": 0.67, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.479673862457275, + "learning_rate": 3.35e-07, + "loss": 0.03618276119232178, + "num_tokens": 1146885.0, + "reward": -14.042501449584961, + "reward_std": 83.22671508789062, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -14.142499923706055, + "rewards/supergames_reward/std": 83.22671508789062, + "step": 134, + "step_time": 8.703399487014394 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 168.0, + "completions/max_terminated_length": 168.0, + "completions/mean_length": 124.75, + "completions/mean_terminated_length": 124.75, + "completions/min_length": 83.0, + "completions/min_terminated_length": 83.0, + "entropy": 0.2997814118862152, + "epoch": 0.675, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.440976619720459, + "learning_rate": 3.3e-07, + "loss": -0.029299044981598854, + "num_tokens": 1152427.0, + "reward": 45.274993896484375, + "reward_std": 44.401519775390625, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 45.17499923706055, + "rewards/supergames_reward/std": 44.40152359008789, + "step": 135, + "step_time": 6.334954546997324 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 161.0, + "completions/max_terminated_length": 161.0, + "completions/mean_length": 123.625, + "completions/mean_terminated_length": 123.625, + "completions/min_length": 93.0, + "completions/min_terminated_length": 93.0, + "entropy": 0.3462243974208832, + "epoch": 0.68, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.1878814697265625, + "learning_rate": 3.25e-07, + "loss": -0.10772529244422913, + "num_tokens": 1157952.0, + "reward": 44.186248779296875, + "reward_std": 51.78197479248047, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 44.08625030517578, + "rewards/supergames_reward/std": 51.7819709777832, + "step": 136, + "step_time": 6.067531232984038 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 155.0, + "completions/max_terminated_length": 155.0, + "completions/mean_length": 106.75, + "completions/mean_terminated_length": 106.75, + "completions/min_length": 74.0, + "completions/min_terminated_length": 74.0, + "entropy": 0.3662136197090149, + "epoch": 0.685, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.508971214294434, + "learning_rate": 3.2e-07, + "loss": -0.11172202974557877, + "num_tokens": 1163334.0, + "reward": 53.0574951171875, + "reward_std": 58.36653518676758, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 52.957496643066406, + "rewards/supergames_reward/std": 58.36653518676758, + "step": 137, + "step_time": 5.8906258280039765 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 175.0, + "completions/max_terminated_length": 175.0, + "completions/mean_length": 125.375, + "completions/mean_terminated_length": 125.375, + "completions/min_length": 90.0, + "completions/min_terminated_length": 90.0, + "entropy": 0.3352622091770172, + "epoch": 0.69, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.1937079429626465, + "learning_rate": 3.15e-07, + "loss": 0.08307601511478424, + "num_tokens": 1168929.0, + "reward": 11.796250343322754, + "reward_std": 53.48870849609375, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 11.696250915527344, + "rewards/supergames_reward/std": 53.48870849609375, + "step": 138, + "step_time": 6.595962718012743 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 254.0, + "completions/max_terminated_length": 254.0, + "completions/mean_length": 193.0, + "completions/mean_terminated_length": 193.0, + "completions/min_length": 111.0, + "completions/min_terminated_length": 111.0, + "entropy": 0.5486535429954529, + "epoch": 0.695, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.861871242523193, + "learning_rate": 3.1e-07, + "loss": 0.0321456640958786, + "num_tokens": 1176937.0, + "reward": 50.463748931884766, + "reward_std": 70.37553405761719, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": 50.38875198364258, + "rewards/supergames_reward/std": 70.31440734863281, + "step": 139, + "step_time": 9.544256388006033 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 261.0, + "completions/max_terminated_length": 261.0, + "completions/mean_length": 189.5, + "completions/mean_terminated_length": 189.5, + "completions/min_length": 95.0, + "completions/min_terminated_length": 95.0, + "entropy": 0.42230603098869324, + "epoch": 0.7, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.830677032470703, + "learning_rate": 3.05e-07, + "loss": -0.07222295552492142, + "num_tokens": 1186789.0, + "reward": 33.44999694824219, + "reward_std": 67.1298599243164, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 33.349998474121094, + "rewards/supergames_reward/std": 67.1298599243164, + "step": 140, + "step_time": 10.002322309010196 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 282.0, + "completions/max_terminated_length": 282.0, + "completions/mean_length": 197.375, + "completions/mean_terminated_length": 197.375, + "completions/min_length": 92.0, + "completions/min_terminated_length": 92.0, + "entropy": 0.33588480949401855, + "epoch": 0.705, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.028353452682495, + "learning_rate": 3e-07, + "loss": -0.07543312013149261, + "num_tokens": 1196672.0, + "reward": 17.10249900817871, + "reward_std": 73.94812774658203, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 17.002500534057617, + "rewards/supergames_reward/std": 73.94813537597656, + "step": 141, + "step_time": 10.684560916008195 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 235.0, + "completions/max_terminated_length": 235.0, + "completions/mean_length": 209.25, + "completions/mean_terminated_length": 209.25, + "completions/min_length": 167.0, + "completions/min_terminated_length": 167.0, + "entropy": 0.3622831702232361, + "epoch": 0.71, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.047222137451172, + "learning_rate": 2.95e-07, + "loss": -0.04569581151008606, + "num_tokens": 1206658.0, + "reward": 49.57624816894531, + "reward_std": 78.21060180664062, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 49.476253509521484, + "rewards/supergames_reward/std": 78.21060943603516, + "step": 142, + "step_time": 8.97679864402744 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 161.0, + "completions/max_terminated_length": 161.0, + "completions/mean_length": 116.25, + "completions/mean_terminated_length": 116.25, + "completions/min_length": 50.0, + "completions/min_terminated_length": 50.0, + "entropy": 0.43277010321617126, + "epoch": 0.715, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.345203876495361, + "learning_rate": 2.9e-07, + "loss": -0.16899466514587402, + "num_tokens": 1212140.0, + "reward": 27.42875099182129, + "reward_std": 48.59138870239258, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 27.328750610351562, + "rewards/supergames_reward/std": 48.591392517089844, + "step": 143, + "step_time": 6.1606669370085 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 205.0, + "completions/max_terminated_length": 205.0, + "completions/mean_length": 173.75, + "completions/mean_terminated_length": 173.75, + "completions/min_length": 134.0, + "completions/min_terminated_length": 134.0, + "entropy": 0.43806371092796326, + "epoch": 0.72, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.858840227127075, + "learning_rate": 2.8499999999999997e-07, + "loss": -0.06502802670001984, + "num_tokens": 1219954.0, + "reward": 38.69249725341797, + "reward_std": 68.78203582763672, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 38.592498779296875, + "rewards/supergames_reward/std": 68.78202819824219, + "step": 144, + "step_time": 7.88639413099736 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 169.0, + "completions/max_terminated_length": 169.0, + "completions/mean_length": 110.625, + "completions/mean_terminated_length": 110.625, + "completions/min_length": 85.0, + "completions/min_terminated_length": 85.0, + "entropy": 0.3383479118347168, + "epoch": 0.725, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.225733757019043, + "learning_rate": 2.8e-07, + "loss": -0.042058318853378296, + "num_tokens": 1225399.0, + "reward": 64.85874938964844, + "reward_std": 25.170387268066406, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 64.75875091552734, + "rewards/supergames_reward/std": 25.170391082763672, + "step": 145, + "step_time": 6.432729878026294 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 147.0, + "completions/max_terminated_length": 147.0, + "completions/mean_length": 129.5, + "completions/mean_terminated_length": 129.5, + "completions/min_length": 93.0, + "completions/min_terminated_length": 93.0, + "entropy": 0.3088987171649933, + "epoch": 0.73, + "frac_reward_zero_std": 1.0, + "grad_norm": 0.0, + "learning_rate": 2.75e-07, + "loss": 0.0, + "num_tokens": 1231043.0, + "reward": 100.0999984741211, + "reward_std": 0.0, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 100.0, + "rewards/supergames_reward/std": 0.0, + "step": 146, + "step_time": 5.634907905012369 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 160.0, + "completions/max_terminated_length": 160.0, + "completions/mean_length": 120.5, + "completions/mean_terminated_length": 120.5, + "completions/min_length": 80.0, + "completions/min_terminated_length": 80.0, + "entropy": 0.37292322516441345, + "epoch": 0.735, + "frac_reward_zero_std": 0.0, + "grad_norm": 5.174577713012695, + "learning_rate": 2.7e-07, + "loss": 0.11778222769498825, + "num_tokens": 1236607.0, + "reward": 34.82374572753906, + "reward_std": 58.513397216796875, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 34.72374725341797, + "rewards/supergames_reward/std": 58.513397216796875, + "step": 147, + "step_time": 6.078690663998714 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 172.0, + "completions/max_terminated_length": 172.0, + "completions/mean_length": 125.625, + "completions/mean_terminated_length": 125.625, + "completions/min_length": 88.0, + "completions/min_terminated_length": 88.0, + "entropy": 0.308654248714447, + "epoch": 0.74, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.8483362197875977, + "learning_rate": 2.65e-07, + "loss": -0.09181805700063705, + "num_tokens": 1242140.0, + "reward": 85.06375122070312, + "reward_std": 42.52893829345703, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 84.9637451171875, + "rewards/supergames_reward/std": 42.52893829345703, + "step": 148, + "step_time": 6.501119122985983 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 275.0, + "completions/max_terminated_length": 275.0, + "completions/mean_length": 211.125, + "completions/mean_terminated_length": 211.125, + "completions/min_length": 157.0, + "completions/min_terminated_length": 157.0, + "entropy": 0.5405827164649963, + "epoch": 0.745, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.774259567260742, + "learning_rate": 2.6e-07, + "loss": -0.03056861087679863, + "num_tokens": 1250205.0, + "reward": 84.63125610351562, + "reward_std": 34.3358154296875, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 84.53125, + "rewards/supergames_reward/std": 34.3358154296875, + "step": 149, + "step_time": 10.352729418984381 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 257.0, + "completions/max_terminated_length": 257.0, + "completions/mean_length": 218.5, + "completions/mean_terminated_length": 218.5, + "completions/min_length": 96.0, + "completions/min_terminated_length": 96.0, + "entropy": 0.3852894604206085, + "epoch": 0.75, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.6103010177612305, + "learning_rate": 2.55e-07, + "loss": 0.034789010882377625, + "num_tokens": 1260249.0, + "reward": 81.81375122070312, + "reward_std": 51.72132873535156, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 81.7137451171875, + "rewards/supergames_reward/std": 51.72132873535156, + "step": 150, + "step_time": 9.937249842012534 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 351.0, + "completions/max_terminated_length": 351.0, + "completions/mean_length": 232.625, + "completions/mean_terminated_length": 232.625, + "completions/min_length": 89.0, + "completions/min_terminated_length": 89.0, + "entropy": 0.4323120415210724, + "epoch": 0.755, + "frac_reward_zero_std": 0.0, + "grad_norm": 2.640308380126953, + "learning_rate": 2.5e-07, + "loss": -0.11181996762752533, + "num_tokens": 1271430.0, + "reward": 56.33625030517578, + "reward_std": 78.97452545166016, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 56.23625183105469, + "rewards/supergames_reward/std": 78.97452545166016, + "step": 151, + "step_time": 13.323547195002902 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 359.0, + "completions/max_terminated_length": 359.0, + "completions/mean_length": 247.5, + "completions/mean_terminated_length": 247.5, + "completions/min_length": 161.0, + "completions/min_terminated_length": 161.0, + "entropy": 0.4038960039615631, + "epoch": 0.76, + "frac_reward_zero_std": 0.0, + "grad_norm": 2.5704824924468994, + "learning_rate": 2.45e-07, + "loss": 0.060525521636009216, + "num_tokens": 1282722.0, + "reward": -68.53875732421875, + "reward_std": 50.590599060058594, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -68.63874816894531, + "rewards/supergames_reward/std": 50.590599060058594, + "step": 152, + "step_time": 13.593127576023107 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 383.0, + "completions/max_terminated_length": 383.0, + "completions/mean_length": 246.375, + "completions/mean_terminated_length": 246.375, + "completions/min_length": 150.0, + "completions/min_terminated_length": 150.0, + "entropy": 0.4110933542251587, + "epoch": 0.765, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.233025312423706, + "learning_rate": 2.4e-07, + "loss": 0.06547188758850098, + "num_tokens": 1294061.0, + "reward": 95.91499328613281, + "reward_std": 11.836968421936035, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 95.81500244140625, + "rewards/supergames_reward/std": 11.836968421936035, + "step": 153, + "step_time": 14.499402174988063 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 338.0, + "completions/max_terminated_length": 338.0, + "completions/mean_length": 233.125, + "completions/mean_terminated_length": 233.125, + "completions/min_length": 88.0, + "completions/min_terminated_length": 88.0, + "entropy": 0.34559860825538635, + "epoch": 0.77, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.6887102127075195, + "learning_rate": 2.3499999999999997e-07, + "loss": 0.15905122458934784, + "num_tokens": 1304262.0, + "reward": 88.92375183105469, + "reward_std": 31.611207962036133, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 88.82374572753906, + "rewards/supergames_reward/std": 31.6112117767334, + "step": 154, + "step_time": 12.704337224975461 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 208.0, + "completions/max_terminated_length": 208.0, + "completions/mean_length": 178.5, + "completions/mean_terminated_length": 178.5, + "completions/min_length": 125.0, + "completions/min_terminated_length": 125.0, + "entropy": 0.4646885097026825, + "epoch": 0.775, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.161937236785889, + "learning_rate": 2.3e-07, + "loss": -0.04354584217071533, + "num_tokens": 1312098.0, + "reward": 27.80875015258789, + "reward_std": 60.81097412109375, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 27.708751678466797, + "rewards/supergames_reward/std": 60.81097412109375, + "step": 155, + "step_time": 7.911955468007363 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 391.0, + "completions/max_terminated_length": 391.0, + "completions/mean_length": 235.375, + "completions/mean_terminated_length": 235.375, + "completions/min_length": 118.0, + "completions/min_terminated_length": 118.0, + "entropy": 0.40985700488090515, + "epoch": 0.78, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.12153959274292, + "learning_rate": 2.25e-07, + "loss": -0.024320494383573532, + "num_tokens": 1322301.0, + "reward": 64.14624786376953, + "reward_std": 49.90530014038086, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 64.04624938964844, + "rewards/supergames_reward/std": 49.90530014038086, + "step": 156, + "step_time": 14.631677505996777 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 330.0, + "completions/max_terminated_length": 330.0, + "completions/mean_length": 220.625, + "completions/mean_terminated_length": 220.625, + "completions/min_length": 93.0, + "completions/min_terminated_length": 93.0, + "entropy": 0.3543764054775238, + "epoch": 0.785, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.0111091136932373, + "learning_rate": 2.1999999999999998e-07, + "loss": -0.13996481895446777, + "num_tokens": 1333386.0, + "reward": 3.6812496185302734, + "reward_std": 103.5096664428711, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": 3.606250762939453, + "rewards/supergames_reward/std": 103.48104858398438, + "step": 157, + "step_time": 12.613216657977318 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 225.0, + "completions/max_terminated_length": 225.0, + "completions/mean_length": 193.125, + "completions/mean_terminated_length": 193.125, + "completions/min_length": 157.0, + "completions/min_terminated_length": 157.0, + "entropy": 0.5682947635650635, + "epoch": 0.79, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.67927885055542, + "learning_rate": 2.1499999999999998e-07, + "loss": 0.0535324327647686, + "num_tokens": 1341355.0, + "reward": -19.78999900817871, + "reward_std": 69.72611236572266, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -19.889997482299805, + "rewards/supergames_reward/std": 69.72611236572266, + "step": 158, + "step_time": 8.585773958999198 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 316.0, + "completions/max_terminated_length": 316.0, + "completions/mean_length": 237.625, + "completions/mean_terminated_length": 237.625, + "completions/min_length": 156.0, + "completions/min_terminated_length": 156.0, + "entropy": 0.3637359142303467, + "epoch": 0.795, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.033409357070923, + "learning_rate": 2.0999999999999997e-07, + "loss": 0.018884683027863503, + "num_tokens": 1352632.0, + "reward": 52.063751220703125, + "reward_std": 65.35551452636719, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 51.96375274658203, + "rewards/supergames_reward/std": 65.35551452636719, + "step": 159, + "step_time": 12.306456914986484 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 225.0, + "completions/max_terminated_length": 225.0, + "completions/mean_length": 184.0, + "completions/mean_terminated_length": 184.0, + "completions/min_length": 143.0, + "completions/min_terminated_length": 143.0, + "entropy": 0.47005388140678406, + "epoch": 0.8, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.585808515548706, + "learning_rate": 2.0499999999999997e-07, + "loss": -0.021751077845692635, + "num_tokens": 1360528.0, + "reward": 33.73124694824219, + "reward_std": 78.67134857177734, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 33.631248474121094, + "rewards/supergames_reward/std": 78.67134857177734, + "step": 160, + "step_time": 8.517202848976012 + } + ], + "logging_steps": 1, + "max_steps": 200, + "num_input_tokens_seen": 1360528, + "num_train_epochs": 1, + "save_steps": 20, + "stateful_callbacks": { + "TrainerControl": { + "args": { + "should_epoch_stop": false, + "should_evaluate": false, + "should_log": false, + "should_save": true, + "should_training_stop": false + }, + "attributes": {} + } + }, + "total_flos": 0.0, + "train_batch_size": 8, + "trial_name": null, + "trial_params": null +} diff --git a/checkpoint-160/training_args.bin b/checkpoint-160/training_args.bin new file mode 100644 index 0000000..730dec0 --- /dev/null +++ b/checkpoint-160/training_args.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:691fe5be2d7d1d56160dae63866d136877a72efb30e15ff7b3249192f998a788 +size 7185 diff --git a/checkpoint-180/chat_template.jinja b/checkpoint-180/chat_template.jinja new file mode 100644 index 0000000..bdf7919 --- /dev/null +++ b/checkpoint-180/chat_template.jinja @@ -0,0 +1,54 @@ +{%- if tools %} + {{- '<|im_start|>system\n' }} + {%- if messages[0]['role'] == 'system' %} + {{- messages[0]['content'] }} + {%- else %} + {{- 'You are Qwen, created by Alibaba Cloud. You are a helpful assistant.' }} + {%- endif %} + {{- "\n\n# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }} + {%- for tool in tools %} + {{- "\n" }} + {{- tool | tojson }} + {%- endfor %} + {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }} +{%- else %} + {%- if messages[0]['role'] == 'system' %} + {{- '<|im_start|>system\n' + messages[0]['content'] + '<|im_end|>\n' }} + {%- else %} + {{- '<|im_start|>system\nYou are Qwen, created by Alibaba Cloud. You are a helpful assistant.<|im_end|>\n' }} + {%- endif %} +{%- endif %} +{%- for message in messages %} + {%- if (message.role == "user") or (message.role == "system" and not loop.first) or (message.role == "assistant" and not message.tool_calls) %} + {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }} + {%- elif message.role == "assistant" %} + {{- '<|im_start|>' + message.role }} + {%- if message.content %} + {{- '\n' + message.content }} + {%- endif %} + {%- for tool_call in message.tool_calls %} + {%- if tool_call.function is defined %} + {%- set tool_call = tool_call.function %} + {%- endif %} + {{- '\n\n{"name": "' }} + {{- tool_call.name }} + {{- '", "arguments": ' }} + {{- tool_call.arguments | tojson }} + {{- '}\n' }} + {%- endfor %} + {{- '<|im_end|>\n' }} + {%- elif message.role == "tool" %} + {%- if (loop.index0 == 0) or (messages[loop.index0 - 1].role != "tool") %} + {{- '<|im_start|>user' }} + {%- endif %} + {{- '\n\n' }} + {{- message.content }} + {{- '\n' }} + {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %} + {{- '<|im_end|>\n' }} + {%- endif %} + {%- endif %} +{%- endfor %} +{%- if add_generation_prompt %} + {{- '<|im_start|>assistant\n' }} +{%- endif %} diff --git a/checkpoint-180/config.json b/checkpoint-180/config.json new file mode 100644 index 0000000..f57b09b --- /dev/null +++ b/checkpoint-180/config.json @@ -0,0 +1,61 @@ +{ + "architectures": [ + "Qwen2ForCausalLM" + ], + "attention_dropout": 0.0, + "bos_token_id": null, + "dtype": "float32", + "eos_token_id": 151645, + "hidden_act": "silu", + "hidden_size": 1536, + "initializer_range": 0.02, + "intermediate_size": 8960, + "layer_types": [ + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention" + ], + "max_position_embeddings": 32768, + "max_window_layers": 21, + "model_type": "qwen2", + "num_attention_heads": 12, + "num_hidden_layers": 28, + "num_key_value_heads": 2, + "pad_token_id": 151643, + "rms_norm_eps": 1e-06, + "rope_parameters": { + "rope_theta": 1000000.0, + "rope_type": "default" + }, + "sliding_window": null, + "tie_word_embeddings": true, + "transformers_version": "5.6.2", + "use_cache": false, + "use_sliding_window": false, + "vocab_size": 151936 +} diff --git a/checkpoint-180/generation_config.json b/checkpoint-180/generation_config.json new file mode 100644 index 0000000..1b2bba9 --- /dev/null +++ b/checkpoint-180/generation_config.json @@ -0,0 +1,13 @@ +{ + "do_sample": true, + "eos_token_id": [ + 151645, + 151643 + ], + "pad_token_id": 151643, + "repetition_penalty": 1.1, + "temperature": 0.7, + "top_k": 20, + "top_p": 0.8, + "transformers_version": "5.6.2" +} diff --git a/checkpoint-180/model.safetensors b/checkpoint-180/model.safetensors new file mode 100644 index 0000000..19d6536 --- /dev/null +++ b/checkpoint-180/model.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:aa72ec17dea3b938582c9d0c66180015e24ba492541ada4b1ff62122b46a81f6 +size 6174895536 diff --git a/checkpoint-180/optimizer.pt b/checkpoint-180/optimizer.pt new file mode 100644 index 0000000..60dda89 --- /dev/null +++ b/checkpoint-180/optimizer.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:32366444f69b40b1c07efaa61259dd4d894f310967b0b29be3249b78176ab9cb +size 12350013801 diff --git a/checkpoint-180/rng_state.pth b/checkpoint-180/rng_state.pth new file mode 100644 index 0000000..cdda4a8 --- /dev/null +++ b/checkpoint-180/rng_state.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:09b3b8e2168fb1e067c78a7650e6b875052ff8cd483bbf8e2cf27a7d0b6ed6dc +size 14645 diff --git a/checkpoint-180/scheduler.pt b/checkpoint-180/scheduler.pt new file mode 100644 index 0000000..d8c6f76 --- /dev/null +++ b/checkpoint-180/scheduler.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:3906df9ece476453bfcc305277c81bbdd9072fad7ee821bcdaff0ae528f6fa6b +size 1465 diff --git a/checkpoint-180/tokenizer.json b/checkpoint-180/tokenizer.json new file mode 100644 index 0000000..34510ff --- /dev/null +++ b/checkpoint-180/tokenizer.json @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:3fd169731d2cbde95e10bf356d66d5997fd885dd8dbb6fb4684da3f23b2585d8 +size 11421892 diff --git a/checkpoint-180/tokenizer_config.json b/checkpoint-180/tokenizer_config.json new file mode 100644 index 0000000..770e41d --- /dev/null +++ b/checkpoint-180/tokenizer_config.json @@ -0,0 +1,30 @@ +{ + "add_prefix_space": false, + "backend": "tokenizers", + "bos_token": null, + "clean_up_tokenization_spaces": false, + "eos_token": "<|im_end|>", + "errors": "replace", + "extra_special_tokens": [ + "<|im_start|>", + "<|im_end|>", + "<|object_ref_start|>", + "<|object_ref_end|>", + "<|box_start|>", + "<|box_end|>", + "<|quad_start|>", + "<|quad_end|>", + "<|vision_start|>", + "<|vision_end|>", + "<|vision_pad|>", + "<|image_pad|>", + "<|video_pad|>" + ], + "is_local": false, + "local_files_only": false, + "model_max_length": 131072, + "pad_token": "<|endoftext|>", + "split_special_tokens": false, + "tokenizer_class": "Qwen2Tokenizer", + "unk_token": null +} diff --git a/checkpoint-180/trainer_state.json b/checkpoint-180/trainer_state.json new file mode 100644 index 0000000..b5fb2c5 --- /dev/null +++ b/checkpoint-180/trainer_state.json @@ -0,0 +1,5254 @@ +{ + "best_global_step": null, + "best_metric": null, + "best_model_checkpoint": null, + "epoch": 0.9, + "eval_steps": 500, + "global_step": 180, + "is_hyper_param_search": false, + "is_local_process_zero": true, + "is_world_process_zero": true, + "log_history": [ + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 478.0, + "completions/max_terminated_length": 478.0, + "completions/mean_length": 191.0, + "completions/mean_terminated_length": 191.0, + "completions/min_length": 66.0, + "completions/min_terminated_length": 66.0, + "entropy": 0.5566893219947815, + "epoch": 0.005, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.370361328125, + "learning_rate": 1e-06, + "loss": -0.24658073484897614, + "num_tokens": 10832.0, + "reward": -3.3000030517578125, + "reward_std": 85.62333679199219, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -3.4000015258789062, + "rewards/supergames_reward/std": 85.62333679199219, + "step": 1, + "step_time": 22.779711072013015 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 216.0, + "completions/max_terminated_length": 216.0, + "completions/mean_length": 131.5, + "completions/mean_terminated_length": 131.5, + "completions/min_length": 60.0, + "completions/min_terminated_length": 60.0, + "entropy": 0.6395841240882874, + "epoch": 0.01, + "frac_reward_zero_std": 0.0, + "grad_norm": 8.09774398803711, + "learning_rate": 9.95e-07, + "loss": 0.20567649602890015, + "num_tokens": 16404.0, + "reward": -12.422499656677246, + "reward_std": 7.134707450866699, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -12.522500038146973, + "rewards/supergames_reward/std": 7.134707927703857, + "step": 2, + "step_time": 9.960156448010821 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 287.0, + "completions/max_terminated_length": 287.0, + "completions/mean_length": 156.625, + "completions/mean_terminated_length": 156.625, + "completions/min_length": 77.0, + "completions/min_terminated_length": 77.0, + "entropy": 0.562222421169281, + "epoch": 0.015, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.9053616523742676, + "learning_rate": 9.9e-07, + "loss": 0.09602774679660797, + "num_tokens": 26953.0, + "reward": -74.94999694824219, + "reward_std": 70.73094177246094, + "rewards/json_format_reward/mean": 0.05000000074505806, + "rewards/json_format_reward/std": 0.09258200973272324, + "rewards/supergames_reward/mean": -75.0, + "rewards/supergames_reward/std": 70.71067810058594, + "step": 3, + "step_time": 13.686320498993155 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 268.0, + "completions/max_terminated_length": 268.0, + "completions/mean_length": 154.125, + "completions/mean_terminated_length": 154.125, + "completions/min_length": 53.0, + "completions/min_terminated_length": 53.0, + "entropy": 0.7323317527770996, + "epoch": 0.02, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.933310031890869, + "learning_rate": 9.849999999999999e-07, + "loss": 0.28249427676200867, + "num_tokens": 36514.0, + "reward": -40.7599983215332, + "reward_std": 81.73140716552734, + "rewards/json_format_reward/mean": 0.05000000074505806, + "rewards/json_format_reward/std": 0.09258200973272324, + "rewards/supergames_reward/mean": -40.80999755859375, + "rewards/supergames_reward/std": 81.68995666503906, + "step": 4, + "step_time": 12.805880262021674 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 401.0, + "completions/max_terminated_length": 401.0, + "completions/mean_length": 205.25, + "completions/mean_terminated_length": 205.25, + "completions/min_length": 115.0, + "completions/min_terminated_length": 115.0, + "entropy": 0.6167430281639099, + "epoch": 0.025, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.344135046005249, + "learning_rate": 9.8e-07, + "loss": 0.1433994024991989, + "num_tokens": 47476.0, + "reward": -97.6612548828125, + "reward_std": 6.3321428298950195, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -97.76124572753906, + "rewards/supergames_reward/std": 6.3321428298950195, + "step": 5, + "step_time": 18.789364666008623 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 276.0, + "completions/max_terminated_length": 276.0, + "completions/mean_length": 167.25, + "completions/mean_terminated_length": 167.25, + "completions/min_length": 73.0, + "completions/min_terminated_length": 73.0, + "entropy": 0.7583790421485901, + "epoch": 0.03, + "frac_reward_zero_std": 0.0, + "grad_norm": 5.172371864318848, + "learning_rate": 9.75e-07, + "loss": -0.12127404659986496, + "num_tokens": 55246.0, + "reward": -58.687503814697266, + "reward_std": 58.64059829711914, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": -58.76250457763672, + "rewards/supergames_reward/std": 58.620460510253906, + "step": 6, + "step_time": 12.948570273991209 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 261.0, + "completions/max_terminated_length": 261.0, + "completions/mean_length": 133.875, + "completions/mean_terminated_length": 133.875, + "completions/min_length": 57.0, + "completions/min_terminated_length": 57.0, + "entropy": 0.7447654604911804, + "epoch": 0.035, + "frac_reward_zero_std": 0.0, + "grad_norm": 6.428595066070557, + "learning_rate": 9.7e-07, + "loss": 0.10135584324598312, + "num_tokens": 60837.0, + "reward": -32.45000076293945, + "reward_std": 41.754486083984375, + "rewards/json_format_reward/mean": 0.05000000074505806, + "rewards/json_format_reward/std": 0.09258200973272324, + "rewards/supergames_reward/mean": -32.5, + "rewards/supergames_reward/std": 41.66190719604492, + "step": 7, + "step_time": 12.058315072004916 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 414.0, + "completions/max_terminated_length": 414.0, + "completions/mean_length": 209.125, + "completions/mean_terminated_length": 209.125, + "completions/min_length": 104.0, + "completions/min_terminated_length": 104.0, + "entropy": 0.3860666751861572, + "epoch": 0.04, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.806255578994751, + "learning_rate": 9.649999999999999e-07, + "loss": -0.33238139748573303, + "num_tokens": 70838.0, + "reward": -80.50375366210938, + "reward_std": 38.82380294799805, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -80.60375213623047, + "rewards/supergames_reward/std": 38.82379913330078, + "step": 8, + "step_time": 19.323370319994865 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.25, + "completions/max_length": 512.0, + "completions/max_terminated_length": 199.0, + "completions/mean_length": 247.125, + "completions/mean_terminated_length": 158.83334350585938, + "completions/min_length": 106.0, + "completions/min_terminated_length": 106.0, + "entropy": 0.5816237330436707, + "epoch": 0.045, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.3209943771362305, + "learning_rate": 9.6e-07, + "loss": 0.36752766370773315, + "num_tokens": 82135.0, + "reward": -27.688751220703125, + "reward_std": 82.8358154296875, + "rewards/json_format_reward/mean": 0.05000000074505806, + "rewards/json_format_reward/std": 0.09258200973272324, + "rewards/supergames_reward/mean": -27.738750457763672, + "rewards/supergames_reward/std": 82.78590393066406, + "step": 9, + "step_time": 21.948575104994234 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 240.0, + "completions/max_terminated_length": 240.0, + "completions/mean_length": 153.0, + "completions/mean_terminated_length": 153.0, + "completions/min_length": 51.0, + "completions/min_terminated_length": 51.0, + "entropy": 0.6301521062850952, + "epoch": 0.05, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.125359058380127, + "learning_rate": 9.55e-07, + "loss": 0.06129350885748863, + "num_tokens": 89807.0, + "reward": -78.39624786376953, + "reward_std": 40.17042922973633, + "rewards/json_format_reward/mean": 0.05000000074505806, + "rewards/json_format_reward/std": 0.09258200973272324, + "rewards/supergames_reward/mean": -78.44625091552734, + "rewards/supergames_reward/std": 40.1396484375, + "step": 10, + "step_time": 9.098202993016457 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 133.0, + "completions/max_terminated_length": 133.0, + "completions/mean_length": 112.0, + "completions/mean_terminated_length": 112.0, + "completions/min_length": 68.0, + "completions/min_terminated_length": 68.0, + "entropy": 0.5521990656852722, + "epoch": 0.055, + "frac_reward_zero_std": 0.0, + "grad_norm": 6.616060256958008, + "learning_rate": 9.499999999999999e-07, + "loss": -0.006313305348157883, + "num_tokens": 95231.0, + "reward": -12.161249160766602, + "reward_std": 6.39578104019165, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -12.261249542236328, + "rewards/supergames_reward/std": 6.39578104019165, + "step": 11, + "step_time": 5.136311663984088 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 396.0, + "completions/max_terminated_length": 396.0, + "completions/mean_length": 214.875, + "completions/mean_terminated_length": 214.875, + "completions/min_length": 55.0, + "completions/min_terminated_length": 55.0, + "entropy": 0.5505905151367188, + "epoch": 0.06, + "frac_reward_zero_std": 0.0, + "grad_norm": 5.289583206176758, + "learning_rate": 9.45e-07, + "loss": -0.29802003502845764, + "num_tokens": 106238.0, + "reward": -91.66999816894531, + "reward_std": 23.2779541015625, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -91.77000427246094, + "rewards/supergames_reward/std": 23.277956008911133, + "step": 12, + "step_time": 14.945365622988902 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 269.0, + "completions/max_terminated_length": 269.0, + "completions/mean_length": 167.375, + "completions/mean_terminated_length": 167.375, + "completions/min_length": 113.0, + "completions/min_terminated_length": 113.0, + "entropy": 0.540415346622467, + "epoch": 0.065, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.858291149139404, + "learning_rate": 9.399999999999999e-07, + "loss": 0.24004128575325012, + "num_tokens": 116937.0, + "reward": -20.38249969482422, + "reward_std": 87.55204010009766, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": -20.457500457763672, + "rewards/supergames_reward/std": 87.52605438232422, + "step": 13, + "step_time": 10.482285185018554 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 254.0, + "completions/max_terminated_length": 254.0, + "completions/mean_length": 156.875, + "completions/mean_terminated_length": 156.875, + "completions/min_length": 114.0, + "completions/min_terminated_length": 114.0, + "entropy": 0.7237679958343506, + "epoch": 0.07, + "frac_reward_zero_std": 0.0, + "grad_norm": 5.083980560302734, + "learning_rate": 9.35e-07, + "loss": -0.19025824964046478, + "num_tokens": 124640.0, + "reward": -51.397499084472656, + "reward_std": 54.486454010009766, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -51.49749755859375, + "rewards/supergames_reward/std": 54.486454010009766, + "step": 14, + "step_time": 9.597231683001155 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 236.0, + "completions/max_terminated_length": 236.0, + "completions/mean_length": 141.375, + "completions/mean_terminated_length": 141.375, + "completions/min_length": 91.0, + "completions/min_terminated_length": 91.0, + "entropy": 0.754334568977356, + "epoch": 0.075, + "frac_reward_zero_std": 0.0, + "grad_norm": 5.720428943634033, + "learning_rate": 9.3e-07, + "loss": -0.16830675303936005, + "num_tokens": 132219.0, + "reward": -56.226253509521484, + "reward_std": 60.115657806396484, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -56.32625198364258, + "rewards/supergames_reward/std": 60.11566162109375, + "step": 15, + "step_time": 8.998362872982398 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 209.0, + "completions/max_terminated_length": 209.0, + "completions/mean_length": 136.375, + "completions/mean_terminated_length": 136.375, + "completions/min_length": 97.0, + "completions/min_terminated_length": 97.0, + "entropy": 0.6280568838119507, + "epoch": 0.08, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.9681806564331055, + "learning_rate": 9.25e-07, + "loss": -0.01910916529595852, + "num_tokens": 139726.0, + "reward": -68.24000549316406, + "reward_std": 52.238914489746094, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -68.33999633789062, + "rewards/supergames_reward/std": 52.238914489746094, + "step": 16, + "step_time": 8.034480294008972 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 365.0, + "completions/max_terminated_length": 365.0, + "completions/mean_length": 211.875, + "completions/mean_terminated_length": 211.875, + "completions/min_length": 100.0, + "completions/min_terminated_length": 100.0, + "entropy": 0.899176836013794, + "epoch": 0.085, + "frac_reward_zero_std": 0.0, + "grad_norm": 6.691746711730957, + "learning_rate": 9.2e-07, + "loss": 0.25513291358947754, + "num_tokens": 147869.0, + "reward": -99.92500305175781, + "reward_std": 0.07070883363485336, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": -100.0, + "rewards/supergames_reward/std": 0.0, + "step": 17, + "step_time": 13.655792869016295 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 387.0, + "completions/max_terminated_length": 387.0, + "completions/mean_length": 238.0, + "completions/mean_terminated_length": 238.0, + "completions/min_length": 87.0, + "completions/min_terminated_length": 87.0, + "entropy": 0.5596873760223389, + "epoch": 0.09, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.6711347103118896, + "learning_rate": 9.15e-07, + "loss": -0.2628335654735565, + "num_tokens": 159101.0, + "reward": -24.900001525878906, + "reward_std": 103.50983428955078, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -25.0, + "rewards/supergames_reward/std": 103.50983428955078, + "step": 18, + "step_time": 14.784329420013819 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.125, + "completions/max_length": 512.0, + "completions/max_terminated_length": 277.0, + "completions/mean_length": 219.75, + "completions/mean_terminated_length": 178.00001525878906, + "completions/min_length": 90.0, + "completions/min_terminated_length": 90.0, + "entropy": 0.9245517253875732, + "epoch": 0.095, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.55502986907959, + "learning_rate": 9.1e-07, + "loss": 0.13060179352760315, + "num_tokens": 167299.0, + "reward": -86.89125061035156, + "reward_std": 36.95603561401367, + "rewards/json_format_reward/mean": 0.05000000074505806, + "rewards/json_format_reward/std": 0.09258200973272324, + "rewards/supergames_reward/mean": -86.94125366210938, + "rewards/supergames_reward/std": 36.93572235107422, + "step": 19, + "step_time": 18.832684700988466 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 267.0, + "completions/max_terminated_length": 267.0, + "completions/mean_length": 177.25, + "completions/mean_terminated_length": 177.25, + "completions/min_length": 93.0, + "completions/min_terminated_length": 93.0, + "entropy": 0.7331254482269287, + "epoch": 0.1, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.538497447967529, + "learning_rate": 9.05e-07, + "loss": -0.16418980062007904, + "num_tokens": 175141.0, + "reward": -68.52874755859375, + "reward_std": 58.48719787597656, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -68.62875366210938, + "rewards/supergames_reward/std": 58.48720169067383, + "step": 20, + "step_time": 10.090975169994636 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 300.0, + "completions/max_terminated_length": 300.0, + "completions/mean_length": 190.625, + "completions/mean_terminated_length": 190.625, + "completions/min_length": 103.0, + "completions/min_terminated_length": 103.0, + "entropy": 0.6659425497055054, + "epoch": 0.105, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.485373497009277, + "learning_rate": 9e-07, + "loss": -0.10106432437896729, + "num_tokens": 183098.0, + "reward": -71.48875427246094, + "reward_std": 55.33565139770508, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -71.5887451171875, + "rewards/supergames_reward/std": 55.33565139770508, + "step": 21, + "step_time": 11.255001295008697 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 279.0, + "completions/max_terminated_length": 279.0, + "completions/mean_length": 179.0, + "completions/mean_terminated_length": 179.0, + "completions/min_length": 103.0, + "completions/min_terminated_length": 103.0, + "entropy": 0.49365949630737305, + "epoch": 0.11, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.829419136047363, + "learning_rate": 8.95e-07, + "loss": 0.10587231814861298, + "num_tokens": 192826.0, + "reward": -75.62000274658203, + "reward_std": 45.01959991455078, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": -75.69499969482422, + "rewards/supergames_reward/std": 45.00411605834961, + "step": 22, + "step_time": 10.611246599000879 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 315.0, + "completions/max_terminated_length": 315.0, + "completions/mean_length": 212.625, + "completions/mean_terminated_length": 212.625, + "completions/min_length": 100.0, + "completions/min_terminated_length": 100.0, + "entropy": 0.43159469962120056, + "epoch": 0.115, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.541125774383545, + "learning_rate": 8.9e-07, + "loss": -0.1681259125471115, + "num_tokens": 202815.0, + "reward": 23.67624855041504, + "reward_std": 67.74340057373047, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 23.576250076293945, + "rewards/supergames_reward/std": 67.743408203125, + "step": 23, + "step_time": 11.731771531980485 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 168.0, + "completions/max_terminated_length": 168.0, + "completions/mean_length": 101.625, + "completions/mean_terminated_length": 101.625, + "completions/min_length": 65.0, + "completions/min_terminated_length": 65.0, + "entropy": 0.3972298502922058, + "epoch": 0.12, + "frac_reward_zero_std": 0.0, + "grad_norm": 5.8285417556762695, + "learning_rate": 8.85e-07, + "loss": -0.011585958302021027, + "num_tokens": 208188.0, + "reward": 58.849998474121094, + "reward_std": 56.9304084777832, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 58.75, + "rewards/supergames_reward/std": 56.9304084777832, + "step": 24, + "step_time": 6.187504950998118 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 303.0, + "completions/max_terminated_length": 303.0, + "completions/mean_length": 165.375, + "completions/mean_terminated_length": 165.375, + "completions/min_length": 69.0, + "completions/min_terminated_length": 69.0, + "entropy": 0.6339899301528931, + "epoch": 0.125, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.823799133300781, + "learning_rate": 8.799999999999999e-07, + "loss": 0.05159185826778412, + "num_tokens": 215935.0, + "reward": -23.521251678466797, + "reward_std": 74.23067474365234, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -23.621248245239258, + "rewards/supergames_reward/std": 74.23067474365234, + "step": 25, + "step_time": 11.059416300005978 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 349.0, + "completions/max_terminated_length": 349.0, + "completions/mean_length": 204.0, + "completions/mean_terminated_length": 204.0, + "completions/min_length": 143.0, + "completions/min_terminated_length": 143.0, + "entropy": 0.6462895274162292, + "epoch": 0.13, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.969292640686035, + "learning_rate": 8.75e-07, + "loss": 0.0115677984431386, + "num_tokens": 223999.0, + "reward": -74.8175048828125, + "reward_std": 40.94221115112305, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": -74.89250183105469, + "rewards/supergames_reward/std": 40.924625396728516, + "step": 26, + "step_time": 12.666237785975682 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 299.0, + "completions/max_terminated_length": 299.0, + "completions/mean_length": 166.5, + "completions/mean_terminated_length": 166.5, + "completions/min_length": 87.0, + "completions/min_terminated_length": 87.0, + "entropy": 0.4756850600242615, + "epoch": 0.135, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.915069103240967, + "learning_rate": 8.699999999999999e-07, + "loss": -0.13321346044540405, + "num_tokens": 233643.0, + "reward": -52.125, + "reward_std": 19.405282974243164, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": -52.20000076293945, + "rewards/supergames_reward/std": 19.33464813232422, + "step": 27, + "step_time": 11.172539752995363 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 184.0, + "completions/max_terminated_length": 184.0, + "completions/mean_length": 137.125, + "completions/mean_terminated_length": 137.125, + "completions/min_length": 100.0, + "completions/min_terminated_length": 100.0, + "entropy": 0.5834778547286987, + "epoch": 0.14, + "frac_reward_zero_std": 0.0, + "grad_norm": 5.146666526794434, + "learning_rate": 8.65e-07, + "loss": 0.060454584658145905, + "num_tokens": 241204.0, + "reward": -68.99250793457031, + "reward_std": 51.010581970214844, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -69.09249877929688, + "rewards/supergames_reward/std": 51.010581970214844, + "step": 28, + "step_time": 7.025046669004951 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 331.0, + "completions/max_terminated_length": 331.0, + "completions/mean_length": 182.375, + "completions/mean_terminated_length": 182.375, + "completions/min_length": 106.0, + "completions/min_terminated_length": 106.0, + "entropy": 0.6368551254272461, + "epoch": 0.145, + "frac_reward_zero_std": 0.0, + "grad_norm": 5.15535306930542, + "learning_rate": 8.599999999999999e-07, + "loss": -0.3323673903942108, + "num_tokens": 249087.0, + "reward": -44.17000198364258, + "reward_std": 80.5373306274414, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -44.27000045776367, + "rewards/supergames_reward/std": 80.5373306274414, + "step": 29, + "step_time": 12.079259724996518 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 190.0, + "completions/max_terminated_length": 190.0, + "completions/mean_length": 127.75, + "completions/mean_terminated_length": 127.75, + "completions/min_length": 80.0, + "completions/min_terminated_length": 80.0, + "entropy": 0.4491196870803833, + "epoch": 0.15, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.896186828613281, + "learning_rate": 8.55e-07, + "loss": -0.12887656688690186, + "num_tokens": 259445.0, + "reward": 57.446250915527344, + "reward_std": 38.78837585449219, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 57.34625244140625, + "rewards/supergames_reward/std": 38.78837966918945, + "step": 30, + "step_time": 7.483972909016302 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 166.0, + "completions/max_terminated_length": 166.0, + "completions/mean_length": 112.875, + "completions/mean_terminated_length": 112.875, + "completions/min_length": 85.0, + "completions/min_terminated_length": 85.0, + "entropy": 0.4858454465866089, + "epoch": 0.155, + "frac_reward_zero_std": 0.0, + "grad_norm": 6.742906093597412, + "learning_rate": 8.499999999999999e-07, + "loss": 0.06930096447467804, + "num_tokens": 264924.0, + "reward": -21.174999237060547, + "reward_std": 31.89051628112793, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": -21.25, + "rewards/supergames_reward/std": 31.819807052612305, + "step": 31, + "step_time": 6.1866529019898735 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 148.0, + "completions/max_terminated_length": 148.0, + "completions/mean_length": 118.125, + "completions/mean_terminated_length": 118.125, + "completions/min_length": 98.0, + "completions/min_terminated_length": 98.0, + "entropy": 0.45428749918937683, + "epoch": 0.16, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.663414478302002, + "learning_rate": 8.45e-07, + "loss": 0.05735419690608978, + "num_tokens": 270381.0, + "reward": 0.4687504768371582, + "reward_std": 38.66371154785156, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 0.36875057220458984, + "rewards/supergames_reward/std": 38.6637077331543, + "step": 32, + "step_time": 5.577042757999152 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 345.0, + "completions/max_terminated_length": 345.0, + "completions/mean_length": 166.875, + "completions/mean_terminated_length": 166.875, + "completions/min_length": 86.0, + "completions/min_terminated_length": 86.0, + "entropy": 0.6107293367385864, + "epoch": 0.165, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.75006628036499, + "learning_rate": 8.399999999999999e-07, + "loss": -0.14760127663612366, + "num_tokens": 278156.0, + "reward": -87.84750366210938, + "reward_std": 26.84708023071289, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -87.94749450683594, + "rewards/supergames_reward/std": 26.847076416015625, + "step": 33, + "step_time": 12.552876825997373 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.125, + "completions/max_length": 512.0, + "completions/max_terminated_length": 360.0, + "completions/mean_length": 231.0, + "completions/mean_terminated_length": 190.85714721679688, + "completions/min_length": 74.0, + "completions/min_terminated_length": 74.0, + "entropy": 0.6089653372764587, + "epoch": 0.17, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.781267166137695, + "learning_rate": 8.349999999999999e-07, + "loss": 0.4294504225254059, + "num_tokens": 289364.0, + "reward": -99.92500305175781, + "reward_std": 0.07070883363485336, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": -100.0, + "rewards/supergames_reward/std": 0.0, + "step": 34, + "step_time": 18.961819477990502 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 218.0, + "completions/max_terminated_length": 218.0, + "completions/mean_length": 134.125, + "completions/mean_terminated_length": 134.125, + "completions/min_length": 104.0, + "completions/min_terminated_length": 104.0, + "entropy": 0.46158918738365173, + "epoch": 0.175, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.016353607177734, + "learning_rate": 8.299999999999999e-07, + "loss": 0.09251818805932999, + "num_tokens": 298733.0, + "reward": -35.558753967285156, + "reward_std": 15.881204605102539, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -35.65875244140625, + "rewards/supergames_reward/std": 15.881203651428223, + "step": 35, + "step_time": 8.38491243700264 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 207.0, + "completions/max_terminated_length": 207.0, + "completions/mean_length": 156.875, + "completions/mean_terminated_length": 156.875, + "completions/min_length": 111.0, + "completions/min_terminated_length": 111.0, + "entropy": 0.5104788541793823, + "epoch": 0.18, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.592893123626709, + "learning_rate": 8.249999999999999e-07, + "loss": -0.0410832017660141, + "num_tokens": 306420.0, + "reward": 5.4212493896484375, + "reward_std": 93.50404357910156, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": 5.346250534057617, + "rewards/supergames_reward/std": 93.4718246459961, + "step": 36, + "step_time": 7.811868985998444 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 241.0, + "completions/max_terminated_length": 241.0, + "completions/mean_length": 178.0, + "completions/mean_terminated_length": 178.0, + "completions/min_length": 106.0, + "completions/min_terminated_length": 106.0, + "entropy": 0.7527878880500793, + "epoch": 0.185, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.8812546730041504, + "learning_rate": 8.199999999999999e-07, + "loss": 0.12539231777191162, + "num_tokens": 314252.0, + "reward": -89.58500671386719, + "reward_std": 29.256120681762695, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": -89.66000366210938, + "rewards/supergames_reward/std": 29.24593734741211, + "step": 37, + "step_time": 8.959742914012168 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 199.0, + "completions/max_terminated_length": 199.0, + "completions/mean_length": 133.375, + "completions/mean_terminated_length": 133.375, + "completions/min_length": 87.0, + "completions/min_terminated_length": 87.0, + "entropy": 0.3347901403903961, + "epoch": 0.19, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.205414295196533, + "learning_rate": 8.149999999999999e-07, + "loss": -0.06991486996412277, + "num_tokens": 319839.0, + "reward": -14.498749732971191, + "reward_std": 13.007229804992676, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -14.598750114440918, + "rewards/supergames_reward/std": 13.007229804992676, + "step": 38, + "step_time": 7.377183554985095 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 181.0, + "completions/max_terminated_length": 181.0, + "completions/mean_length": 114.375, + "completions/mean_terminated_length": 114.375, + "completions/min_length": 58.0, + "completions/min_terminated_length": 58.0, + "entropy": 0.4445965886116028, + "epoch": 0.195, + "frac_reward_zero_std": 0.0, + "grad_norm": 5.561298847198486, + "learning_rate": 8.1e-07, + "loss": -0.06235164776444435, + "num_tokens": 325290.0, + "reward": -7.425000190734863, + "reward_std": 53.701602935791016, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": -7.5, + "rewards/supergames_reward/std": 53.652320861816406, + "step": 39, + "step_time": 6.753862089011818 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 370.0, + "completions/max_terminated_length": 370.0, + "completions/mean_length": 187.625, + "completions/mean_terminated_length": 187.625, + "completions/min_length": 60.0, + "completions/min_terminated_length": 60.0, + "entropy": 0.3950234353542328, + "epoch": 0.2, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.5988712310791016, + "learning_rate": 8.05e-07, + "loss": -0.1381440907716751, + "num_tokens": 335119.0, + "reward": -33.68375015258789, + "reward_std": 42.069435119628906, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -33.783748626708984, + "rewards/supergames_reward/std": 42.069435119628906, + "step": 40, + "step_time": 13.718958162004128 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.125, + "completions/max_length": 512.0, + "completions/max_terminated_length": 371.0, + "completions/mean_length": 251.125, + "completions/mean_terminated_length": 213.85714721679688, + "completions/min_length": 121.0, + "completions/min_terminated_length": 121.0, + "entropy": 0.507067084312439, + "epoch": 0.205, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.3898658752441406, + "learning_rate": 8e-07, + "loss": 0.0078964838758111, + "num_tokens": 346464.0, + "reward": -86.48625183105469, + "reward_std": 38.020694732666016, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": -86.56124877929688, + "rewards/supergames_reward/std": 38.010528564453125, + "step": 41, + "step_time": 18.90960379401804 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 152.0, + "completions/max_terminated_length": 152.0, + "completions/mean_length": 116.375, + "completions/mean_terminated_length": 116.375, + "completions/min_length": 93.0, + "completions/min_terminated_length": 93.0, + "entropy": 0.3801707923412323, + "epoch": 0.21, + "frac_reward_zero_std": 1.0, + "grad_norm": 0.0, + "learning_rate": 7.95e-07, + "loss": 0.0, + "num_tokens": 351955.0, + "reward": -9.899999618530273, + "reward_std": 0.0, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -10.0, + "rewards/supergames_reward/std": 0.0, + "step": 42, + "step_time": 5.711630532023264 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 308.0, + "completions/max_terminated_length": 308.0, + "completions/mean_length": 247.375, + "completions/mean_terminated_length": 247.375, + "completions/min_length": 116.0, + "completions/min_terminated_length": 116.0, + "entropy": 0.3185043931007385, + "epoch": 0.215, + "frac_reward_zero_std": 0.0, + "grad_norm": 2.168762445449829, + "learning_rate": 7.9e-07, + "loss": 0.1970764398574829, + "num_tokens": 363278.0, + "reward": -53.89249801635742, + "reward_std": 63.02619934082031, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -53.99250030517578, + "rewards/supergames_reward/std": 63.02619934082031, + "step": 43, + "step_time": 11.661934906995157 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 330.0, + "completions/max_terminated_length": 330.0, + "completions/mean_length": 212.5, + "completions/mean_terminated_length": 212.5, + "completions/min_length": 106.0, + "completions/min_terminated_length": 106.0, + "entropy": 0.4446268379688263, + "epoch": 0.22, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.305844306945801, + "learning_rate": 7.85e-07, + "loss": -0.06735197454690933, + "num_tokens": 374306.0, + "reward": 18.059999465942383, + "reward_std": 98.9417953491211, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 17.959999084472656, + "rewards/supergames_reward/std": 98.9417953491211, + "step": 44, + "step_time": 12.492453911982011 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 316.0, + "completions/max_terminated_length": 316.0, + "completions/mean_length": 183.0, + "completions/mean_terminated_length": 183.0, + "completions/min_length": 103.0, + "completions/min_terminated_length": 103.0, + "entropy": 0.5269614458084106, + "epoch": 0.225, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.8006370067596436, + "learning_rate": 7.799999999999999e-07, + "loss": -0.08870815485715866, + "num_tokens": 382210.0, + "reward": -54.928749084472656, + "reward_std": 72.5127944946289, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -55.02874755859375, + "rewards/supergames_reward/std": 72.5127944946289, + "step": 45, + "step_time": 11.7636467939883 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 244.0, + "completions/max_terminated_length": 244.0, + "completions/mean_length": 136.625, + "completions/mean_terminated_length": 136.625, + "completions/min_length": 85.0, + "completions/min_terminated_length": 85.0, + "entropy": 0.34820204973220825, + "epoch": 0.23, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.911736011505127, + "learning_rate": 7.75e-07, + "loss": 0.04397330805659294, + "num_tokens": 391583.0, + "reward": -19.10375213623047, + "reward_std": 23.967702865600586, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -19.203750610351562, + "rewards/supergames_reward/std": 23.967702865600586, + "step": 46, + "step_time": 10.119426151999505 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 218.0, + "completions/max_terminated_length": 218.0, + "completions/mean_length": 160.375, + "completions/mean_terminated_length": 160.375, + "completions/min_length": 93.0, + "completions/min_terminated_length": 93.0, + "entropy": 0.611914336681366, + "epoch": 0.235, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.571913242340088, + "learning_rate": 7.699999999999999e-07, + "loss": -0.044293053448200226, + "num_tokens": 399290.0, + "reward": -25.186250686645508, + "reward_std": 71.46305084228516, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -25.286251068115234, + "rewards/supergames_reward/std": 71.46305084228516, + "step": 47, + "step_time": 8.563868903991533 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 167.0, + "completions/max_terminated_length": 167.0, + "completions/mean_length": 111.125, + "completions/mean_terminated_length": 111.125, + "completions/min_length": 56.0, + "completions/min_terminated_length": 56.0, + "entropy": 0.46585777401924133, + "epoch": 0.24, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.654109954833984, + "learning_rate": 7.65e-07, + "loss": -0.035151124000549316, + "num_tokens": 404707.0, + "reward": -0.408750057220459, + "reward_std": 31.939748764038086, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -0.5087499618530273, + "rewards/supergames_reward/std": 31.93975257873535, + "step": 48, + "step_time": 6.438482580007985 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 306.0, + "completions/max_terminated_length": 306.0, + "completions/mean_length": 216.625, + "completions/mean_terminated_length": 216.625, + "completions/min_length": 132.0, + "completions/min_terminated_length": 132.0, + "entropy": 0.33054471015930176, + "epoch": 0.245, + "frac_reward_zero_std": 0.0, + "grad_norm": 2.700303792953491, + "learning_rate": 7.599999999999999e-07, + "loss": -0.04393656551837921, + "num_tokens": 414736.0, + "reward": -97.26000213623047, + "reward_std": 1.6294406652450562, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -97.36000061035156, + "rewards/supergames_reward/std": 1.629441738128662, + "step": 49, + "step_time": 11.636040106008295 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 378.0, + "completions/max_terminated_length": 378.0, + "completions/mean_length": 218.375, + "completions/mean_terminated_length": 218.375, + "completions/min_length": 104.0, + "completions/min_terminated_length": 104.0, + "entropy": 0.48711177706718445, + "epoch": 0.25, + "frac_reward_zero_std": 0.0, + "grad_norm": 2.9650564193725586, + "learning_rate": 7.55e-07, + "loss": -0.3529520332813263, + "num_tokens": 424795.0, + "reward": 47.408748626708984, + "reward_std": 72.72083282470703, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 47.30875015258789, + "rewards/supergames_reward/std": 72.72083282470703, + "step": 50, + "step_time": 14.258096842997475 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.125, + "completions/max_length": 512.0, + "completions/max_terminated_length": 492.0, + "completions/mean_length": 267.625, + "completions/mean_terminated_length": 232.71429443359375, + "completions/min_length": 90.0, + "completions/min_terminated_length": 90.0, + "entropy": 0.4482860863208771, + "epoch": 0.255, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.2998197078704834, + "learning_rate": 7.5e-07, + "loss": 0.2493654191493988, + "num_tokens": 436288.0, + "reward": 50.61125183105469, + "reward_std": 75.3349838256836, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": 50.53624725341797, + "rewards/supergames_reward/std": 75.27783966064453, + "step": 51, + "step_time": 19.20827590499539 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 176.0, + "completions/max_terminated_length": 176.0, + "completions/mean_length": 144.5, + "completions/mean_terminated_length": 144.5, + "completions/min_length": 99.0, + "completions/min_terminated_length": 99.0, + "entropy": 0.5378735661506653, + "epoch": 0.26, + "frac_reward_zero_std": 0.0, + "grad_norm": 5.114461898803711, + "learning_rate": 7.45e-07, + "loss": 0.0036597400903701782, + "num_tokens": 441980.0, + "reward": -15.20625114440918, + "reward_std": 45.35526657104492, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": -15.28125, + "rewards/supergames_reward/std": 45.301815032958984, + "step": 52, + "step_time": 6.660627231001854 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 136.0, + "completions/max_terminated_length": 136.0, + "completions/mean_length": 119.0, + "completions/mean_terminated_length": 119.0, + "completions/min_length": 105.0, + "completions/min_terminated_length": 105.0, + "entropy": 0.41796669363975525, + "epoch": 0.265, + "frac_reward_zero_std": 0.0, + "grad_norm": 5.059505462646484, + "learning_rate": 7.4e-07, + "loss": -0.014973883517086506, + "num_tokens": 447468.0, + "reward": 15.170000076293945, + "reward_std": 46.658329010009766, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 15.070000648498535, + "rewards/supergames_reward/std": 46.658329010009766, + "step": 53, + "step_time": 5.271571868011961 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 259.0, + "completions/max_terminated_length": 259.0, + "completions/mean_length": 191.125, + "completions/mean_terminated_length": 191.125, + "completions/min_length": 91.0, + "completions/min_terminated_length": 91.0, + "entropy": 0.45421987771987915, + "epoch": 0.27, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.4454867839813232, + "learning_rate": 7.35e-07, + "loss": -0.1128973588347435, + "num_tokens": 457301.0, + "reward": 38.89875030517578, + "reward_std": 62.1772575378418, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 38.79874801635742, + "rewards/supergames_reward/std": 62.1772575378418, + "step": 54, + "step_time": 10.524528659996577 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 319.0, + "completions/max_terminated_length": 319.0, + "completions/mean_length": 233.375, + "completions/mean_terminated_length": 233.375, + "completions/min_length": 136.0, + "completions/min_terminated_length": 136.0, + "entropy": 0.4449213445186615, + "epoch": 0.275, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.2065398693084717, + "learning_rate": 7.3e-07, + "loss": 0.0016277075046673417, + "num_tokens": 467520.0, + "reward": 27.372499465942383, + "reward_std": 81.23126983642578, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 27.27250099182129, + "rewards/supergames_reward/std": 81.23126983642578, + "step": 55, + "step_time": 12.592280682991259 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.125, + "completions/max_length": 512.0, + "completions/max_terminated_length": 357.0, + "completions/mean_length": 261.0, + "completions/mean_terminated_length": 225.1428680419922, + "completions/min_length": 95.0, + "completions/min_terminated_length": 95.0, + "entropy": 0.506920337677002, + "epoch": 0.28, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.058388710021973, + "learning_rate": 7.249999999999999e-07, + "loss": 0.3289242386817932, + "num_tokens": 476056.0, + "reward": 47.05249786376953, + "reward_std": 80.97753143310547, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": 46.977500915527344, + "rewards/supergames_reward/std": 80.9256362915039, + "step": 56, + "step_time": 18.845177220006008 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 250.0, + "completions/max_terminated_length": 250.0, + "completions/mean_length": 192.75, + "completions/mean_terminated_length": 192.75, + "completions/min_length": 122.0, + "completions/min_terminated_length": 122.0, + "entropy": 0.5197336077690125, + "epoch": 0.285, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.230815410614014, + "learning_rate": 7.2e-07, + "loss": -0.15186546742916107, + "num_tokens": 484014.0, + "reward": -64.5050048828125, + "reward_std": 50.35145568847656, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -64.60499572753906, + "rewards/supergames_reward/std": 50.3514518737793, + "step": 57, + "step_time": 9.576232638006331 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 324.0, + "completions/max_terminated_length": 324.0, + "completions/mean_length": 249.0, + "completions/mean_terminated_length": 249.0, + "completions/min_length": 129.0, + "completions/min_terminated_length": 129.0, + "entropy": 0.42913514375686646, + "epoch": 0.29, + "frac_reward_zero_std": 0.0, + "grad_norm": 2.8544528484344482, + "learning_rate": 7.149999999999999e-07, + "loss": -0.056405920535326004, + "num_tokens": 494390.0, + "reward": 40.123748779296875, + "reward_std": 84.164794921875, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 40.02375030517578, + "rewards/supergames_reward/std": 84.164794921875, + "step": 58, + "step_time": 12.385353341000155 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 138.0, + "completions/max_terminated_length": 138.0, + "completions/mean_length": 107.25, + "completions/mean_terminated_length": 107.25, + "completions/min_length": 81.0, + "completions/min_terminated_length": 81.0, + "entropy": 0.34386488795280457, + "epoch": 0.295, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.6915154457092285, + "learning_rate": 7.1e-07, + "loss": 0.0865321159362793, + "num_tokens": 499816.0, + "reward": -8.244999885559082, + "reward_std": 4.681046962738037, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -8.345000267028809, + "rewards/supergames_reward/std": 4.681046962738037, + "step": 59, + "step_time": 5.42325339600211 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 187.0, + "completions/max_terminated_length": 187.0, + "completions/mean_length": 127.25, + "completions/mean_terminated_length": 127.25, + "completions/min_length": 97.0, + "completions/min_terminated_length": 97.0, + "entropy": 0.5430045127868652, + "epoch": 0.3, + "frac_reward_zero_std": 0.0, + "grad_norm": 5.635311126708984, + "learning_rate": 7.049999999999999e-07, + "loss": -0.05412375554442406, + "num_tokens": 505410.0, + "reward": 20.005001068115234, + "reward_std": 55.156005859375, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 19.904998779296875, + "rewards/supergames_reward/std": 55.156005859375, + "step": 60, + "step_time": 7.068010902003152 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 286.0, + "completions/max_terminated_length": 286.0, + "completions/mean_length": 192.125, + "completions/mean_terminated_length": 192.125, + "completions/min_length": 95.0, + "completions/min_terminated_length": 95.0, + "entropy": 0.4213869869709015, + "epoch": 0.305, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.4664618968963623, + "learning_rate": 7e-07, + "loss": -0.0495159812271595, + "num_tokens": 516283.0, + "reward": -47.212501525878906, + "reward_std": 91.2253646850586, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -47.3125, + "rewards/supergames_reward/std": 91.2253646850586, + "step": 61, + "step_time": 11.176304235996213 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 324.0, + "completions/max_terminated_length": 324.0, + "completions/mean_length": 171.0, + "completions/mean_terminated_length": 171.0, + "completions/min_length": 92.0, + "completions/min_terminated_length": 92.0, + "entropy": 0.4626811146736145, + "epoch": 0.31, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.949252128601074, + "learning_rate": 6.949999999999999e-07, + "loss": 0.192047581076622, + "num_tokens": 522251.0, + "reward": -15.092500686645508, + "reward_std": 49.82962417602539, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": -15.167499542236328, + "rewards/supergames_reward/std": 49.78091049194336, + "step": 62, + "step_time": 12.075224861997413 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.125, + "completions/max_length": 512.0, + "completions/max_terminated_length": 429.0, + "completions/mean_length": 283.5, + "completions/mean_terminated_length": 250.85714721679688, + "completions/min_length": 85.0, + "completions/min_terminated_length": 85.0, + "entropy": 0.4814216196537018, + "epoch": 0.315, + "frac_reward_zero_std": 0.0, + "grad_norm": 2.7918753623962402, + "learning_rate": 6.9e-07, + "loss": 0.15706156194210052, + "num_tokens": 533815.0, + "reward": -20.86625099182129, + "reward_std": 85.99250030517578, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": -20.94124984741211, + "rewards/supergames_reward/std": 85.96620178222656, + "step": 63, + "step_time": 19.448832260008203 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 189.0, + "completions/max_terminated_length": 189.0, + "completions/mean_length": 131.25, + "completions/mean_terminated_length": 131.25, + "completions/min_length": 106.0, + "completions/min_terminated_length": 106.0, + "entropy": 0.4929414987564087, + "epoch": 0.32, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.774101734161377, + "learning_rate": 6.85e-07, + "loss": -0.060691747814416885, + "num_tokens": 539465.0, + "reward": -21.0674991607666, + "reward_std": 6.892730236053467, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -21.167499542236328, + "rewards/supergames_reward/std": 6.892730236053467, + "step": 64, + "step_time": 7.141569407976931 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 454.0, + "completions/max_terminated_length": 454.0, + "completions/mean_length": 281.375, + "completions/mean_terminated_length": 281.375, + "completions/min_length": 107.0, + "completions/min_terminated_length": 107.0, + "entropy": 0.45866572856903076, + "epoch": 0.325, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.092717409133911, + "learning_rate": 6.800000000000001e-07, + "loss": 0.11694959551095963, + "num_tokens": 551012.0, + "reward": 49.938751220703125, + "reward_std": 79.1865005493164, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 49.8387451171875, + "rewards/supergames_reward/std": 79.1865005493164, + "step": 65, + "step_time": 17.06170882002334 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 203.0, + "completions/max_terminated_length": 203.0, + "completions/mean_length": 128.625, + "completions/mean_terminated_length": 128.625, + "completions/min_length": 88.0, + "completions/min_terminated_length": 88.0, + "entropy": 0.5005782246589661, + "epoch": 0.33, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.77616024017334, + "learning_rate": 6.75e-07, + "loss": 0.11242837458848953, + "num_tokens": 556577.0, + "reward": 48.875, + "reward_std": 49.13539123535156, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 48.775001525878906, + "rewards/supergames_reward/std": 49.13539123535156, + "step": 66, + "step_time": 7.695410129002994 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 436.0, + "completions/max_terminated_length": 436.0, + "completions/mean_length": 281.875, + "completions/mean_terminated_length": 281.875, + "completions/min_length": 133.0, + "completions/min_terminated_length": 133.0, + "entropy": 0.507097065448761, + "epoch": 0.335, + "frac_reward_zero_std": 0.0, + "grad_norm": 2.911184787750244, + "learning_rate": 6.7e-07, + "loss": -0.08151137083768845, + "num_tokens": 567120.0, + "reward": -12.886249542236328, + "reward_std": 84.04269409179688, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -12.986251831054688, + "rewards/supergames_reward/std": 84.04269409179688, + "step": 67, + "step_time": 16.409127942984924 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.125, + "completions/max_length": 512.0, + "completions/max_terminated_length": 501.0, + "completions/mean_length": 318.375, + "completions/mean_terminated_length": 290.71429443359375, + "completions/min_length": 65.0, + "completions/min_terminated_length": 65.0, + "entropy": 0.486402690410614, + "epoch": 0.34, + "frac_reward_zero_std": 0.0, + "grad_norm": 2.9461607933044434, + "learning_rate": 6.65e-07, + "loss": 0.2442178875207901, + "num_tokens": 578995.0, + "reward": 2.8387489318847656, + "reward_std": 98.47579956054688, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": 2.7637500762939453, + "rewards/supergames_reward/std": 98.44595336914062, + "step": 68, + "step_time": 19.32364999302081 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 291.0, + "completions/max_terminated_length": 291.0, + "completions/mean_length": 219.625, + "completions/mean_terminated_length": 219.625, + "completions/min_length": 117.0, + "completions/min_terminated_length": 117.0, + "entropy": 0.37856727838516235, + "epoch": 0.345, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.0204732418060303, + "learning_rate": 6.6e-07, + "loss": -0.08764916658401489, + "num_tokens": 589072.0, + "reward": -54.18375015258789, + "reward_std": 71.45012664794922, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -54.283748626708984, + "rewards/supergames_reward/std": 71.45013427734375, + "step": 69, + "step_time": 11.126611230982235 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.125, + "completions/max_length": 512.0, + "completions/max_terminated_length": 305.0, + "completions/mean_length": 254.75, + "completions/mean_terminated_length": 218.00001525878906, + "completions/min_length": 104.0, + "completions/min_terminated_length": 104.0, + "entropy": 0.46423614025115967, + "epoch": 0.35, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.0140154361724854, + "learning_rate": 6.55e-07, + "loss": 0.26797160506248474, + "num_tokens": 600438.0, + "reward": -32.04875183105469, + "reward_std": 69.86957550048828, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": -32.12375259399414, + "rewards/supergames_reward/std": 69.84178161621094, + "step": 70, + "step_time": 19.12617294798838 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 372.0, + "completions/max_terminated_length": 372.0, + "completions/mean_length": 194.125, + "completions/mean_terminated_length": 194.125, + "completions/min_length": 94.0, + "completions/min_terminated_length": 94.0, + "entropy": 0.3427293300628662, + "epoch": 0.355, + "frac_reward_zero_std": 0.0, + "grad_norm": 2.781383991241455, + "learning_rate": 6.5e-07, + "loss": 0.354524701833725, + "num_tokens": 610327.0, + "reward": -6.913749694824219, + "reward_std": 87.56071472167969, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": -6.988750457763672, + "rewards/supergames_reward/std": 87.53033447265625, + "step": 71, + "step_time": 14.041668199002743 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 197.0, + "completions/max_terminated_length": 197.0, + "completions/mean_length": 124.75, + "completions/mean_terminated_length": 124.75, + "completions/min_length": 93.0, + "completions/min_terminated_length": 93.0, + "entropy": 0.4455118477344513, + "epoch": 0.36, + "frac_reward_zero_std": 0.0, + "grad_norm": 5.726032733917236, + "learning_rate": 6.45e-07, + "loss": 0.18206289410591125, + "num_tokens": 615853.0, + "reward": 41.95124816894531, + "reward_std": 60.640811920166016, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 41.85124969482422, + "rewards/supergames_reward/std": 60.640811920166016, + "step": 72, + "step_time": 7.416493425989756 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 323.0, + "completions/max_terminated_length": 323.0, + "completions/mean_length": 147.375, + "completions/mean_terminated_length": 147.375, + "completions/min_length": 94.0, + "completions/min_terminated_length": 94.0, + "entropy": 0.480733186006546, + "epoch": 0.365, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.853621482849121, + "learning_rate": 6.4e-07, + "loss": 0.04558124020695686, + "num_tokens": 621584.0, + "reward": 14.884998321533203, + "reward_std": 46.25857162475586, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 14.78499984741211, + "rewards/supergames_reward/std": 46.25857162475586, + "step": 73, + "step_time": 11.95379540900467 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 283.0, + "completions/max_terminated_length": 283.0, + "completions/mean_length": 188.125, + "completions/mean_terminated_length": 188.125, + "completions/min_length": 116.0, + "completions/min_terminated_length": 116.0, + "entropy": 0.5526013374328613, + "epoch": 0.37, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.729883909225464, + "learning_rate": 6.35e-07, + "loss": 0.06268303096294403, + "num_tokens": 629545.0, + "reward": -68.05750274658203, + "reward_std": 59.44808578491211, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -68.15750122070312, + "rewards/supergames_reward/std": 59.44808578491211, + "step": 74, + "step_time": 10.64124580900534 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 141.0, + "completions/max_terminated_length": 141.0, + "completions/mean_length": 117.375, + "completions/mean_terminated_length": 117.375, + "completions/min_length": 100.0, + "completions/min_terminated_length": 100.0, + "entropy": 0.3744319677352905, + "epoch": 0.375, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.775154113769531, + "learning_rate": 6.3e-07, + "loss": -0.023284845054149628, + "num_tokens": 635052.0, + "reward": 3.9374990463256836, + "reward_std": 61.85771179199219, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": 3.862499237060547, + "rewards/supergames_reward/std": 61.80967712402344, + "step": 75, + "step_time": 5.432648951013107 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 268.0, + "completions/max_terminated_length": 268.0, + "completions/mean_length": 157.0, + "completions/mean_terminated_length": 157.0, + "completions/min_length": 104.0, + "completions/min_terminated_length": 104.0, + "entropy": 0.4822140336036682, + "epoch": 0.38, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.715704917907715, + "learning_rate": 6.249999999999999e-07, + "loss": -0.04918142035603523, + "num_tokens": 644644.0, + "reward": 48.64125061035156, + "reward_std": 78.0205078125, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": 48.56624984741211, + "rewards/supergames_reward/std": 77.96605682373047, + "step": 76, + "step_time": 10.348935816989979 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 380.0, + "completions/max_terminated_length": 380.0, + "completions/mean_length": 210.875, + "completions/mean_terminated_length": 210.875, + "completions/min_length": 126.0, + "completions/min_terminated_length": 126.0, + "entropy": 0.45722317695617676, + "epoch": 0.385, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.6057567596435547, + "learning_rate": 6.2e-07, + "loss": 0.058352649211883545, + "num_tokens": 652771.0, + "reward": -8.283750534057617, + "reward_std": 77.68436431884766, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -8.383749008178711, + "rewards/supergames_reward/std": 77.68437194824219, + "step": 77, + "step_time": 14.142948755004909 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 324.0, + "completions/max_terminated_length": 324.0, + "completions/mean_length": 188.25, + "completions/mean_terminated_length": 188.25, + "completions/min_length": 68.0, + "completions/min_terminated_length": 68.0, + "entropy": 0.33380764722824097, + "epoch": 0.39, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.2239863872528076, + "learning_rate": 6.149999999999999e-07, + "loss": -0.04592633992433548, + "num_tokens": 662653.0, + "reward": 11.153749465942383, + "reward_std": 89.88329315185547, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 11.053749084472656, + "rewards/supergames_reward/std": 89.88329315185547, + "step": 78, + "step_time": 12.277474621019792 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 143.0, + "completions/max_terminated_length": 143.0, + "completions/mean_length": 108.875, + "completions/mean_terminated_length": 108.875, + "completions/min_length": 85.0, + "completions/min_terminated_length": 85.0, + "entropy": 0.3518812656402588, + "epoch": 0.395, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.967776298522949, + "learning_rate": 6.1e-07, + "loss": -0.10800496488809586, + "num_tokens": 668076.0, + "reward": 70.33250427246094, + "reward_std": 49.9145393371582, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 70.23249816894531, + "rewards/supergames_reward/std": 49.91453552246094, + "step": 79, + "step_time": 5.52714040101273 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 184.0, + "completions/max_terminated_length": 184.0, + "completions/mean_length": 152.125, + "completions/mean_terminated_length": 152.125, + "completions/min_length": 104.0, + "completions/min_terminated_length": 104.0, + "entropy": 0.4699896574020386, + "epoch": 0.4, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.561161994934082, + "learning_rate": 6.049999999999999e-07, + "loss": 0.09505834430456161, + "num_tokens": 675701.0, + "reward": -20.25625228881836, + "reward_std": 53.78542709350586, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -20.35624885559082, + "rewards/supergames_reward/std": 53.78542709350586, + "step": 80, + "step_time": 7.155081019998761 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.125, + "completions/max_length": 512.0, + "completions/max_terminated_length": 506.0, + "completions/mean_length": 288.875, + "completions/mean_terminated_length": 257.0, + "completions/min_length": 115.0, + "completions/min_terminated_length": 115.0, + "entropy": 0.4086494445800781, + "epoch": 0.405, + "frac_reward_zero_std": 0.0, + "grad_norm": 2.3719611167907715, + "learning_rate": 6e-07, + "loss": 0.04727627709507942, + "num_tokens": 687332.0, + "reward": -37.20750045776367, + "reward_std": 91.7676010131836, + "rewards/json_format_reward/mean": 0.05000000074505806, + "rewards/json_format_reward/std": 0.09258200973272324, + "rewards/supergames_reward/mean": -37.25749969482422, + "rewards/supergames_reward/std": 91.72847747802734, + "step": 81, + "step_time": 19.231399144016905 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 376.0, + "completions/max_terminated_length": 376.0, + "completions/mean_length": 227.75, + "completions/mean_terminated_length": 227.75, + "completions/min_length": 39.0, + "completions/min_terminated_length": 39.0, + "entropy": 0.43233707547187805, + "epoch": 0.41, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.3391714096069336, + "learning_rate": 5.949999999999999e-07, + "loss": -0.25524550676345825, + "num_tokens": 698474.0, + "reward": -3.7287511825561523, + "reward_std": 91.25379943847656, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -3.828749656677246, + "rewards/supergames_reward/std": 91.25379943847656, + "step": 82, + "step_time": 14.329176368017215 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 299.0, + "completions/max_terminated_length": 299.0, + "completions/mean_length": 178.125, + "completions/mean_terminated_length": 178.125, + "completions/min_length": 91.0, + "completions/min_terminated_length": 91.0, + "entropy": 0.4293636977672577, + "epoch": 0.415, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.7271742820739746, + "learning_rate": 5.9e-07, + "loss": 0.09273672848939896, + "num_tokens": 708243.0, + "reward": 12.802498817443848, + "reward_std": 86.23526000976562, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 12.702500343322754, + "rewards/supergames_reward/std": 86.23526000976562, + "step": 83, + "step_time": 11.419686011999147 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 425.0, + "completions/max_terminated_length": 425.0, + "completions/mean_length": 255.375, + "completions/mean_terminated_length": 255.375, + "completions/min_length": 172.0, + "completions/min_terminated_length": 172.0, + "entropy": 0.38068246841430664, + "epoch": 0.42, + "frac_reward_zero_std": 0.0, + "grad_norm": 2.6244184970855713, + "learning_rate": 5.849999999999999e-07, + "loss": 0.02006213366985321, + "num_tokens": 718606.0, + "reward": 37.803749084472656, + "reward_std": 57.58624267578125, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 37.70375061035156, + "rewards/supergames_reward/std": 57.58624267578125, + "step": 84, + "step_time": 15.975198492989875 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 116.0, + "completions/max_terminated_length": 116.0, + "completions/mean_length": 108.125, + "completions/mean_terminated_length": 108.125, + "completions/min_length": 87.0, + "completions/min_terminated_length": 87.0, + "entropy": 0.339572012424469, + "epoch": 0.425, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.754759311676025, + "learning_rate": 5.8e-07, + "loss": -0.011015941388905048, + "num_tokens": 723983.0, + "reward": -3.8212504386901855, + "reward_std": 61.624786376953125, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -3.921250820159912, + "rewards/supergames_reward/std": 61.624786376953125, + "step": 85, + "step_time": 4.542777584982105 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 447.0, + "completions/max_terminated_length": 447.0, + "completions/mean_length": 291.375, + "completions/mean_terminated_length": 291.375, + "completions/min_length": 163.0, + "completions/min_terminated_length": 163.0, + "entropy": 0.5846173763275146, + "epoch": 0.43, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.55724835395813, + "learning_rate": 5.749999999999999e-07, + "loss": -0.04981120675802231, + "num_tokens": 735642.0, + "reward": 50.07499694824219, + "reward_std": 92.6283187866211, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": 50.0, + "rewards/supergames_reward/std": 92.5820083618164, + "step": 86, + "step_time": 16.777178087009815 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 232.0, + "completions/max_terminated_length": 232.0, + "completions/mean_length": 192.0, + "completions/mean_terminated_length": 192.0, + "completions/min_length": 139.0, + "completions/min_terminated_length": 139.0, + "entropy": 0.5139474272727966, + "epoch": 0.435, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.8860411643981934, + "learning_rate": 5.699999999999999e-07, + "loss": -0.02194221317768097, + "num_tokens": 743602.0, + "reward": -52.522499084472656, + "reward_std": 73.8686294555664, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -52.62249755859375, + "rewards/supergames_reward/std": 73.86863708496094, + "step": 87, + "step_time": 8.772893911984283 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 266.0, + "completions/max_terminated_length": 266.0, + "completions/mean_length": 146.875, + "completions/mean_terminated_length": 146.875, + "completions/min_length": 89.0, + "completions/min_terminated_length": 89.0, + "entropy": 0.3428497910499573, + "epoch": 0.44, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.509483337402344, + "learning_rate": 5.649999999999999e-07, + "loss": 0.19857533276081085, + "num_tokens": 753065.0, + "reward": 63.63874816894531, + "reward_std": 67.54963684082031, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 63.53874969482422, + "rewards/supergames_reward/std": 67.54963684082031, + "step": 88, + "step_time": 10.230529835011112 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 446.0, + "completions/max_terminated_length": 446.0, + "completions/mean_length": 270.625, + "completions/mean_terminated_length": 270.625, + "completions/min_length": 103.0, + "completions/min_terminated_length": 103.0, + "entropy": 0.4417419731616974, + "epoch": 0.445, + "frac_reward_zero_std": 0.0, + "grad_norm": 2.973951816558838, + "learning_rate": 5.6e-07, + "loss": -0.00031440958264283836, + "num_tokens": 764534.0, + "reward": -9.736251831054688, + "reward_std": 92.25879669189453, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -9.836250305175781, + "rewards/supergames_reward/std": 92.25880432128906, + "step": 89, + "step_time": 17.40752330099349 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 398.0, + "completions/max_terminated_length": 398.0, + "completions/mean_length": 227.375, + "completions/mean_terminated_length": 227.375, + "completions/min_length": 98.0, + "completions/min_terminated_length": 98.0, + "entropy": 0.4732024073600769, + "epoch": 0.45, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.2560815811157227, + "learning_rate": 5.55e-07, + "loss": 0.03820464387536049, + "num_tokens": 772809.0, + "reward": 5.982499599456787, + "reward_std": 58.833797454833984, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 5.8824992179870605, + "rewards/supergames_reward/std": 58.83380126953125, + "step": 90, + "step_time": 15.007269965979503 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 269.0, + "completions/max_terminated_length": 269.0, + "completions/mean_length": 207.5, + "completions/mean_terminated_length": 207.5, + "completions/min_length": 92.0, + "completions/min_terminated_length": 92.0, + "entropy": 0.42038559913635254, + "epoch": 0.455, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.3649117946624756, + "learning_rate": 5.5e-07, + "loss": 0.1510048508644104, + "num_tokens": 782805.0, + "reward": -5.61500358581543, + "reward_std": 82.67152404785156, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -5.714998245239258, + "rewards/supergames_reward/std": 82.67152404785156, + "step": 91, + "step_time": 10.386137172987219 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 416.0, + "completions/max_terminated_length": 416.0, + "completions/mean_length": 218.375, + "completions/mean_terminated_length": 218.375, + "completions/min_length": 95.0, + "completions/min_terminated_length": 95.0, + "entropy": 0.3971734642982483, + "epoch": 0.46, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.433772563934326, + "learning_rate": 5.45e-07, + "loss": -0.3027213513851166, + "num_tokens": 793840.0, + "reward": -29.06500244140625, + "reward_std": 91.68325805664062, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -29.16499900817871, + "rewards/supergames_reward/std": 91.68325805664062, + "step": 92, + "step_time": 15.790611553995404 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 345.0, + "completions/max_terminated_length": 345.0, + "completions/mean_length": 206.625, + "completions/mean_terminated_length": 206.625, + "completions/min_length": 90.0, + "completions/min_terminated_length": 90.0, + "entropy": 0.5906907320022583, + "epoch": 0.465, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.293088912963867, + "learning_rate": 5.4e-07, + "loss": 0.09760032594203949, + "num_tokens": 803821.0, + "reward": -72.73999786376953, + "reward_std": 50.72484588623047, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": -72.81500244140625, + "rewards/supergames_reward/std": 50.7094841003418, + "step": 93, + "step_time": 13.043119941983605 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 143.0, + "completions/max_terminated_length": 143.0, + "completions/mean_length": 115.75, + "completions/mean_terminated_length": 115.75, + "completions/min_length": 80.0, + "completions/min_terminated_length": 80.0, + "entropy": 0.4233115315437317, + "epoch": 0.47, + "frac_reward_zero_std": 0.0, + "grad_norm": 5.58886194229126, + "learning_rate": 5.35e-07, + "loss": -0.04040650278329849, + "num_tokens": 809299.0, + "reward": -3.3050003051757812, + "reward_std": 24.443714141845703, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -3.405000686645508, + "rewards/supergames_reward/std": 24.443714141845703, + "step": 94, + "step_time": 5.512874762003776 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 122.0, + "completions/max_terminated_length": 122.0, + "completions/mean_length": 103.125, + "completions/mean_terminated_length": 103.125, + "completions/min_length": 86.0, + "completions/min_terminated_length": 86.0, + "entropy": 0.4055764377117157, + "epoch": 0.475, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.981549263000488, + "learning_rate": 5.3e-07, + "loss": -0.07588938623666763, + "num_tokens": 814716.0, + "reward": 60.06999969482422, + "reward_std": 42.708473205566406, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 59.970001220703125, + "rewards/supergames_reward/std": 42.708473205566406, + "step": 95, + "step_time": 4.766389057011111 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 475.0, + "completions/max_terminated_length": 475.0, + "completions/mean_length": 218.375, + "completions/mean_terminated_length": 218.375, + "completions/min_length": 81.0, + "completions/min_terminated_length": 81.0, + "entropy": 0.408913254737854, + "epoch": 0.48, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.79402756690979, + "learning_rate": 5.25e-07, + "loss": -0.3087541460990906, + "num_tokens": 824791.0, + "reward": -15.042500495910645, + "reward_std": 84.37246704101562, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -15.142499923706055, + "rewards/supergames_reward/std": 84.37246704101562, + "step": 96, + "step_time": 17.779843941010768 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 154.0, + "completions/max_terminated_length": 154.0, + "completions/mean_length": 121.625, + "completions/mean_terminated_length": 121.625, + "completions/min_length": 80.0, + "completions/min_terminated_length": 80.0, + "entropy": 0.529015839099884, + "epoch": 0.485, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.142816066741943, + "learning_rate": 5.2e-07, + "loss": 0.07216037809848785, + "num_tokens": 830300.0, + "reward": -32.743751525878906, + "reward_std": 54.61775588989258, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": -32.81875228881836, + "rewards/supergames_reward/std": 54.58256149291992, + "step": 97, + "step_time": 5.913989890017547 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 126.0, + "completions/max_terminated_length": 126.0, + "completions/mean_length": 95.625, + "completions/mean_terminated_length": 95.625, + "completions/min_length": 74.0, + "completions/min_terminated_length": 74.0, + "entropy": 0.33743274211883545, + "epoch": 0.49, + "frac_reward_zero_std": 0.0, + "grad_norm": 5.197273254394531, + "learning_rate": 5.149999999999999e-07, + "loss": 0.07500407844781876, + "num_tokens": 835633.0, + "reward": 11.00999927520752, + "reward_std": 62.73848342895508, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 10.910000801086426, + "rewards/supergames_reward/std": 62.73848342895508, + "step": 98, + "step_time": 4.9155233050114475 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 149.0, + "completions/max_terminated_length": 149.0, + "completions/mean_length": 108.5, + "completions/mean_terminated_length": 108.5, + "completions/min_length": 86.0, + "completions/min_terminated_length": 86.0, + "entropy": 0.3106803297996521, + "epoch": 0.495, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.864002227783203, + "learning_rate": 5.1e-07, + "loss": -0.1206185445189476, + "num_tokens": 841045.0, + "reward": 14.238749504089355, + "reward_std": 51.77016830444336, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 14.138750076293945, + "rewards/supergames_reward/std": 51.77016830444336, + "step": 99, + "step_time": 5.7507751359953545 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 149.0, + "completions/max_terminated_length": 149.0, + "completions/mean_length": 97.125, + "completions/mean_terminated_length": 97.125, + "completions/min_length": 73.0, + "completions/min_terminated_length": 73.0, + "entropy": 0.35160186886787415, + "epoch": 0.5, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.915731906890869, + "learning_rate": 5.049999999999999e-07, + "loss": -0.07951541244983673, + "num_tokens": 846398.0, + "reward": 67.78125, + "reward_std": 38.15436553955078, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 67.68124389648438, + "rewards/supergames_reward/std": 38.15436553955078, + "step": 100, + "step_time": 5.994720440998208 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 320.0, + "completions/max_terminated_length": 320.0, + "completions/mean_length": 208.75, + "completions/mean_terminated_length": 208.75, + "completions/min_length": 95.0, + "completions/min_terminated_length": 95.0, + "entropy": 0.452095091342926, + "epoch": 0.505, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.065176963806152, + "learning_rate": 5e-07, + "loss": -0.1397111564874649, + "num_tokens": 857420.0, + "reward": -5.801251411437988, + "reward_std": 73.56117248535156, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -5.90125036239624, + "rewards/supergames_reward/std": 73.5611801147461, + "step": 101, + "step_time": 12.28178753197426 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 230.0, + "completions/max_terminated_length": 230.0, + "completions/mean_length": 171.375, + "completions/mean_terminated_length": 171.375, + "completions/min_length": 95.0, + "completions/min_terminated_length": 95.0, + "entropy": 0.4138447642326355, + "epoch": 0.51, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.998823404312134, + "learning_rate": 4.95e-07, + "loss": -0.08941879868507385, + "num_tokens": 867103.0, + "reward": 92.24749755859375, + "reward_std": 14.540005683898926, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 92.14749908447266, + "rewards/supergames_reward/std": 14.540006637573242, + "step": 102, + "step_time": 8.992682139010867 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 311.0, + "completions/max_terminated_length": 311.0, + "completions/mean_length": 163.0, + "completions/mean_terminated_length": 163.0, + "completions/min_length": 74.0, + "completions/min_terminated_length": 74.0, + "entropy": 0.4182547330856323, + "epoch": 0.515, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.419595241546631, + "learning_rate": 4.9e-07, + "loss": -0.19304415583610535, + "num_tokens": 876679.0, + "reward": 79.13249969482422, + "reward_std": 59.3050422668457, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 79.03250122070312, + "rewards/supergames_reward/std": 59.30504608154297, + "step": 103, + "step_time": 11.836970244999975 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 253.0, + "completions/max_terminated_length": 253.0, + "completions/mean_length": 155.625, + "completions/mean_terminated_length": 155.625, + "completions/min_length": 94.0, + "completions/min_terminated_length": 94.0, + "entropy": 0.49889081716537476, + "epoch": 0.52, + "frac_reward_zero_std": 0.0, + "grad_norm": 5.463719367980957, + "learning_rate": 4.85e-07, + "loss": -0.04786720871925354, + "num_tokens": 884348.0, + "reward": 48.616249084472656, + "reward_std": 73.30741119384766, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 48.51625061035156, + "rewards/supergames_reward/std": 73.30741119384766, + "step": 104, + "step_time": 9.579507841990562 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 380.0, + "completions/max_terminated_length": 380.0, + "completions/mean_length": 269.5, + "completions/mean_terminated_length": 269.5, + "completions/min_length": 101.0, + "completions/min_terminated_length": 101.0, + "entropy": 0.5286482572555542, + "epoch": 0.525, + "frac_reward_zero_std": 0.0, + "grad_norm": 2.993295431137085, + "learning_rate": 4.8e-07, + "loss": -0.08071611076593399, + "num_tokens": 894856.0, + "reward": 34.970001220703125, + "reward_std": 85.324462890625, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 34.869998931884766, + "rewards/supergames_reward/std": 85.324462890625, + "step": 105, + "step_time": 14.295730627985904 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 261.0, + "completions/max_terminated_length": 261.0, + "completions/mean_length": 203.25, + "completions/mean_terminated_length": 203.25, + "completions/min_length": 91.0, + "completions/min_terminated_length": 91.0, + "entropy": 0.6049264073371887, + "epoch": 0.53, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.787306785583496, + "learning_rate": 4.7499999999999995e-07, + "loss": 0.19928883016109467, + "num_tokens": 902922.0, + "reward": 32.44499969482422, + "reward_std": 70.4663314819336, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": 32.369998931884766, + "rewards/supergames_reward/std": 70.41261291503906, + "step": 106, + "step_time": 9.84894504098338 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 279.0, + "completions/max_terminated_length": 279.0, + "completions/mean_length": 194.0, + "completions/mean_terminated_length": 194.0, + "completions/min_length": 109.0, + "completions/min_terminated_length": 109.0, + "entropy": 0.5115755200386047, + "epoch": 0.535, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.277207374572754, + "learning_rate": 4.6999999999999995e-07, + "loss": 0.09118300676345825, + "num_tokens": 910906.0, + "reward": 40.432498931884766, + "reward_std": 48.54548263549805, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 40.33250045776367, + "rewards/supergames_reward/std": 48.54548263549805, + "step": 107, + "step_time": 10.466603949986165 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 332.0, + "completions/max_terminated_length": 332.0, + "completions/mean_length": 195.875, + "completions/mean_terminated_length": 195.875, + "completions/min_length": 87.0, + "completions/min_terminated_length": 87.0, + "entropy": 0.4192962944507599, + "epoch": 0.54, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.308061122894287, + "learning_rate": 4.65e-07, + "loss": 0.28064602613449097, + "num_tokens": 920745.0, + "reward": -21.982500076293945, + "reward_std": 66.31195068359375, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -22.08249855041504, + "rewards/supergames_reward/std": 66.31195068359375, + "step": 108, + "step_time": 12.595848233992001 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 269.0, + "completions/max_terminated_length": 269.0, + "completions/mean_length": 195.125, + "completions/mean_terminated_length": 195.125, + "completions/min_length": 111.0, + "completions/min_terminated_length": 111.0, + "entropy": 0.37523961067199707, + "epoch": 0.545, + "frac_reward_zero_std": 0.0, + "grad_norm": 2.802706718444824, + "learning_rate": 4.6e-07, + "loss": -0.06614465266466141, + "num_tokens": 930618.0, + "reward": 71.75, + "reward_std": 51.14524459838867, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 71.6500015258789, + "rewards/supergames_reward/std": 51.1452522277832, + "step": 109, + "step_time": 10.33884758799104 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 331.0, + "completions/max_terminated_length": 331.0, + "completions/mean_length": 167.75, + "completions/mean_terminated_length": 167.75, + "completions/min_length": 89.0, + "completions/min_terminated_length": 89.0, + "entropy": 0.4339042007923126, + "epoch": 0.55, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.254003047943115, + "learning_rate": 4.55e-07, + "loss": 0.34405598044395447, + "num_tokens": 941256.0, + "reward": 99.21875, + "reward_std": 2.492553234100342, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 99.1187515258789, + "rewards/supergames_reward/std": 2.4925525188446045, + "step": 110, + "step_time": 12.657525141985388 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 285.0, + "completions/max_terminated_length": 285.0, + "completions/mean_length": 211.0, + "completions/mean_terminated_length": 211.0, + "completions/min_length": 112.0, + "completions/min_terminated_length": 112.0, + "entropy": 0.46809861063957214, + "epoch": 0.555, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.811166286468506, + "learning_rate": 4.5e-07, + "loss": 0.045113153755664825, + "num_tokens": 951296.0, + "reward": 55.27750015258789, + "reward_std": 47.90687561035156, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 55.1775016784668, + "rewards/supergames_reward/std": 47.9068717956543, + "step": 111, + "step_time": 10.91118389699841 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 145.0, + "completions/max_terminated_length": 145.0, + "completions/mean_length": 108.375, + "completions/mean_terminated_length": 108.375, + "completions/min_length": 86.0, + "completions/min_terminated_length": 86.0, + "entropy": 0.34181955456733704, + "epoch": 0.56, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.537296772003174, + "learning_rate": 4.45e-07, + "loss": -0.052413541823625565, + "num_tokens": 956675.0, + "reward": 47.56624984741211, + "reward_std": 49.185157775878906, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 47.466251373291016, + "rewards/supergames_reward/std": 49.18516159057617, + "step": 112, + "step_time": 5.552288047998445 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 245.0, + "completions/max_terminated_length": 245.0, + "completions/mean_length": 189.125, + "completions/mean_terminated_length": 189.125, + "completions/min_length": 102.0, + "completions/min_terminated_length": 102.0, + "entropy": 0.5034605860710144, + "epoch": 0.565, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.349876642227173, + "learning_rate": 4.3999999999999997e-07, + "loss": -0.08186715841293335, + "num_tokens": 964636.0, + "reward": 81.32374572753906, + "reward_std": 36.85074996948242, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 81.2237548828125, + "rewards/supergames_reward/std": 36.85074996948242, + "step": 113, + "step_time": 9.267611294984818 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 178.0, + "completions/max_terminated_length": 178.0, + "completions/mean_length": 111.25, + "completions/mean_terminated_length": 111.25, + "completions/min_length": 67.0, + "completions/min_terminated_length": 67.0, + "entropy": 0.3229002058506012, + "epoch": 0.57, + "frac_reward_zero_std": 0.0, + "grad_norm": 5.245575428009033, + "learning_rate": 4.3499999999999996e-07, + "loss": 0.06544198095798492, + "num_tokens": 970070.0, + "reward": 63.1775016784668, + "reward_std": 52.30624008178711, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 63.07749938964844, + "rewards/supergames_reward/std": 52.30624008178711, + "step": 114, + "step_time": 6.731139309995342 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 224.0, + "completions/max_terminated_length": 224.0, + "completions/mean_length": 170.0, + "completions/mean_terminated_length": 170.0, + "completions/min_length": 94.0, + "completions/min_terminated_length": 94.0, + "entropy": 0.3274807929992676, + "epoch": 0.575, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.5544140338897705, + "learning_rate": 4.2999999999999996e-07, + "loss": -0.06953569501638412, + "num_tokens": 979702.0, + "reward": 25.62125015258789, + "reward_std": 64.52003479003906, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 25.521244049072266, + "rewards/supergames_reward/std": 64.52003479003906, + "step": 115, + "step_time": 8.702618776995223 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 427.0, + "completions/max_terminated_length": 427.0, + "completions/mean_length": 209.5, + "completions/mean_terminated_length": 209.5, + "completions/min_length": 100.0, + "completions/min_terminated_length": 100.0, + "entropy": 0.4970880150794983, + "epoch": 0.58, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.4932165145874023, + "learning_rate": 4.2499999999999995e-07, + "loss": -0.2281273603439331, + "num_tokens": 990690.0, + "reward": 50.54750061035156, + "reward_std": 70.5126953125, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 50.4474983215332, + "rewards/supergames_reward/std": 70.51270294189453, + "step": 116, + "step_time": 16.204386608995264 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 228.0, + "completions/max_terminated_length": 228.0, + "completions/mean_length": 132.375, + "completions/mean_terminated_length": 132.375, + "completions/min_length": 85.0, + "completions/min_terminated_length": 85.0, + "entropy": 0.4380227029323578, + "epoch": 0.585, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.089909076690674, + "learning_rate": 4.1999999999999995e-07, + "loss": 0.12121254950761795, + "num_tokens": 996333.0, + "reward": -15.333749771118164, + "reward_std": 47.200164794921875, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -15.433748245239258, + "rewards/supergames_reward/std": 47.200164794921875, + "step": 117, + "step_time": 8.546233230998041 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 323.0, + "completions/max_terminated_length": 323.0, + "completions/mean_length": 137.625, + "completions/mean_terminated_length": 137.625, + "completions/min_length": 90.0, + "completions/min_terminated_length": 90.0, + "entropy": 0.4177803695201874, + "epoch": 0.59, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.8712472915649414, + "learning_rate": 4.1499999999999994e-07, + "loss": 0.14192955195903778, + "num_tokens": 1002010.0, + "reward": -30.700000762939453, + "reward_std": 17.224035263061523, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -30.799999237060547, + "rewards/supergames_reward/std": 17.224035263061523, + "step": 118, + "step_time": 11.878434990998358 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 453.0, + "completions/max_terminated_length": 453.0, + "completions/mean_length": 238.125, + "completions/mean_terminated_length": 238.125, + "completions/min_length": 109.0, + "completions/min_terminated_length": 109.0, + "entropy": 0.5640217065811157, + "epoch": 0.595, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.064737319946289, + "learning_rate": 4.0999999999999994e-07, + "loss": -0.023621462285518646, + "num_tokens": 1012259.0, + "reward": 43.683746337890625, + "reward_std": 70.4210205078125, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 43.5837516784668, + "rewards/supergames_reward/std": 70.4210205078125, + "step": 119, + "step_time": 16.829514264973113 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 236.0, + "completions/max_terminated_length": 236.0, + "completions/mean_length": 193.75, + "completions/mean_terminated_length": 193.75, + "completions/min_length": 151.0, + "completions/min_terminated_length": 151.0, + "entropy": 0.5795091390609741, + "epoch": 0.6, + "frac_reward_zero_std": 0.0, + "grad_norm": 5.423007965087891, + "learning_rate": 4.05e-07, + "loss": -0.0666111558675766, + "num_tokens": 1020241.0, + "reward": -85.86000061035156, + "reward_std": 27.4442081451416, + "rewards/json_format_reward/mean": 0.05000000074505806, + "rewards/json_format_reward/std": 0.09258200973272324, + "rewards/supergames_reward/mean": -85.90999603271484, + "rewards/supergames_reward/std": 27.41469955444336, + "step": 120, + "step_time": 8.95496519000153 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 331.0, + "completions/max_terminated_length": 331.0, + "completions/mean_length": 231.0, + "completions/mean_terminated_length": 231.0, + "completions/min_length": 165.0, + "completions/min_terminated_length": 165.0, + "entropy": 0.3767699897289276, + "epoch": 0.605, + "frac_reward_zero_std": 0.0, + "grad_norm": 2.7434473037719727, + "learning_rate": 4e-07, + "loss": -0.1165132150053978, + "num_tokens": 1031465.0, + "reward": 61.45624923706055, + "reward_std": 72.21080780029297, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 61.35625076293945, + "rewards/supergames_reward/std": 72.2108154296875, + "step": 121, + "step_time": 12.588820598000893 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 297.0, + "completions/max_terminated_length": 297.0, + "completions/mean_length": 223.375, + "completions/mean_terminated_length": 223.375, + "completions/min_length": 112.0, + "completions/min_terminated_length": 112.0, + "entropy": 0.4537277817726135, + "epoch": 0.61, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.4382436275482178, + "learning_rate": 3.95e-07, + "loss": -0.1038316935300827, + "num_tokens": 1042548.0, + "reward": 5.658749580383301, + "reward_std": 83.77420043945312, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 5.558750152587891, + "rewards/supergames_reward/std": 83.77420043945312, + "step": 122, + "step_time": 11.39525846898323 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 456.0, + "completions/max_terminated_length": 456.0, + "completions/mean_length": 314.25, + "completions/mean_terminated_length": 314.25, + "completions/min_length": 95.0, + "completions/min_terminated_length": 95.0, + "entropy": 0.4113953411579132, + "epoch": 0.615, + "frac_reward_zero_std": 0.0, + "grad_norm": 2.449777841567993, + "learning_rate": 3.8999999999999997e-07, + "loss": -0.1303664743900299, + "num_tokens": 1054422.0, + "reward": 35.98249816894531, + "reward_std": 88.6131362915039, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 35.88249969482422, + "rewards/supergames_reward/std": 88.6131362915039, + "step": 123, + "step_time": 17.001118954998674 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 242.0, + "completions/max_terminated_length": 242.0, + "completions/mean_length": 172.5, + "completions/mean_terminated_length": 172.5, + "completions/min_length": 137.0, + "completions/min_terminated_length": 137.0, + "entropy": 0.4670729339122772, + "epoch": 0.62, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.951946973800659, + "learning_rate": 3.8499999999999997e-07, + "loss": 0.05888795852661133, + "num_tokens": 1062210.0, + "reward": 43.368751525878906, + "reward_std": 40.30733871459961, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 43.26874542236328, + "rewards/supergames_reward/std": 40.307342529296875, + "step": 124, + "step_time": 9.142582255997695 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 148.0, + "completions/max_terminated_length": 148.0, + "completions/mean_length": 106.625, + "completions/mean_terminated_length": 106.625, + "completions/min_length": 79.0, + "completions/min_terminated_length": 79.0, + "entropy": 0.30219921469688416, + "epoch": 0.625, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.391322135925293, + "learning_rate": 3.7999999999999996e-07, + "loss": 0.14367049932479858, + "num_tokens": 1067599.0, + "reward": 42.63374710083008, + "reward_std": 58.20897674560547, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 42.533748626708984, + "rewards/supergames_reward/std": 58.208984375, + "step": 125, + "step_time": 5.636063549987739 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 243.0, + "completions/max_terminated_length": 243.0, + "completions/mean_length": 179.25, + "completions/mean_terminated_length": 179.25, + "completions/min_length": 134.0, + "completions/min_terminated_length": 134.0, + "entropy": 0.4834887683391571, + "epoch": 0.63, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.9391286373138428, + "learning_rate": 3.75e-07, + "loss": -0.06614185124635696, + "num_tokens": 1075457.0, + "reward": 8.302498817443848, + "reward_std": 75.10364532470703, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 8.202500343322754, + "rewards/supergames_reward/std": 75.10364532470703, + "step": 126, + "step_time": 9.130600645992672 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 196.0, + "completions/max_terminated_length": 196.0, + "completions/mean_length": 157.0, + "completions/mean_terminated_length": 157.0, + "completions/min_length": 117.0, + "completions/min_terminated_length": 117.0, + "entropy": 0.4801807701587677, + "epoch": 0.635, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.425615310668945, + "learning_rate": 3.7e-07, + "loss": -0.08830522745847702, + "num_tokens": 1083129.0, + "reward": -10.563751220703125, + "reward_std": 64.00443267822266, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -10.663749694824219, + "rewards/supergames_reward/std": 64.00444030761719, + "step": 127, + "step_time": 7.4972667430120055 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 412.0, + "completions/max_terminated_length": 412.0, + "completions/mean_length": 239.25, + "completions/mean_terminated_length": 239.25, + "completions/min_length": 89.0, + "completions/min_terminated_length": 89.0, + "entropy": 0.4667063355445862, + "epoch": 0.64, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.199113130569458, + "learning_rate": 3.65e-07, + "loss": -0.043856702744960785, + "num_tokens": 1094355.0, + "reward": -55.89250183105469, + "reward_std": 53.05980682373047, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -55.99250030517578, + "rewards/supergames_reward/std": 53.05980682373047, + "step": 128, + "step_time": 15.429580625001108 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 370.0, + "completions/max_terminated_length": 370.0, + "completions/mean_length": 244.75, + "completions/mean_terminated_length": 244.75, + "completions/min_length": 127.0, + "completions/min_terminated_length": 127.0, + "entropy": 0.57573401927948, + "epoch": 0.645, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.6800079345703125, + "learning_rate": 3.6e-07, + "loss": 0.09521033614873886, + "num_tokens": 1105633.0, + "reward": -23.35250473022461, + "reward_std": 102.31401062011719, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -23.452499389648438, + "rewards/supergames_reward/std": 102.31402587890625, + "step": 129, + "step_time": 13.920327747007832 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 258.0, + "completions/max_terminated_length": 258.0, + "completions/mean_length": 188.25, + "completions/mean_terminated_length": 188.25, + "completions/min_length": 131.0, + "completions/min_terminated_length": 131.0, + "entropy": 0.5341438055038452, + "epoch": 0.65, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.97955060005188, + "learning_rate": 3.55e-07, + "loss": -0.09512650966644287, + "num_tokens": 1113571.0, + "reward": 29.181249618530273, + "reward_std": 80.10585021972656, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 29.081249237060547, + "rewards/supergames_reward/std": 80.10585021972656, + "step": 130, + "step_time": 9.69286120700417 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 257.0, + "completions/max_terminated_length": 257.0, + "completions/mean_length": 180.75, + "completions/mean_terminated_length": 180.75, + "completions/min_length": 136.0, + "completions/min_terminated_length": 136.0, + "entropy": 0.46591487526893616, + "epoch": 0.655, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.3518130779266357, + "learning_rate": 3.5e-07, + "loss": 0.04714229330420494, + "num_tokens": 1121457.0, + "reward": -8.287501335144043, + "reward_std": 76.35875701904297, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -8.387499809265137, + "rewards/supergames_reward/std": 76.35875701904297, + "step": 131, + "step_time": 9.636637121002423 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 190.0, + "completions/max_terminated_length": 190.0, + "completions/mean_length": 147.75, + "completions/mean_terminated_length": 147.75, + "completions/min_length": 90.0, + "completions/min_terminated_length": 90.0, + "entropy": 0.4407735764980316, + "epoch": 0.66, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.498741626739502, + "learning_rate": 3.45e-07, + "loss": -0.03996007889509201, + "num_tokens": 1129095.0, + "reward": 0.1999988555908203, + "reward_std": 84.63554382324219, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": 0.125, + "rewards/supergames_reward/std": 84.6017074584961, + "step": 132, + "step_time": 7.510833573003765 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 358.0, + "completions/max_terminated_length": 358.0, + "completions/mean_length": 213.875, + "completions/mean_terminated_length": 213.875, + "completions/min_length": 137.0, + "completions/min_terminated_length": 137.0, + "entropy": 0.5186149477958679, + "epoch": 0.665, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.3987107276916504, + "learning_rate": 3.4000000000000003e-07, + "loss": 0.1863594949245453, + "num_tokens": 1139110.0, + "reward": -21.328752517700195, + "reward_std": 88.9102783203125, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -21.428749084472656, + "rewards/supergames_reward/std": 88.9102783203125, + "step": 133, + "step_time": 13.393839450000087 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 230.0, + "completions/max_terminated_length": 230.0, + "completions/mean_length": 169.875, + "completions/mean_terminated_length": 169.875, + "completions/min_length": 121.0, + "completions/min_terminated_length": 121.0, + "entropy": 0.49445784091949463, + "epoch": 0.67, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.479673862457275, + "learning_rate": 3.35e-07, + "loss": 0.03618276119232178, + "num_tokens": 1146885.0, + "reward": -14.042501449584961, + "reward_std": 83.22671508789062, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -14.142499923706055, + "rewards/supergames_reward/std": 83.22671508789062, + "step": 134, + "step_time": 8.703399487014394 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 168.0, + "completions/max_terminated_length": 168.0, + "completions/mean_length": 124.75, + "completions/mean_terminated_length": 124.75, + "completions/min_length": 83.0, + "completions/min_terminated_length": 83.0, + "entropy": 0.2997814118862152, + "epoch": 0.675, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.440976619720459, + "learning_rate": 3.3e-07, + "loss": -0.029299044981598854, + "num_tokens": 1152427.0, + "reward": 45.274993896484375, + "reward_std": 44.401519775390625, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 45.17499923706055, + "rewards/supergames_reward/std": 44.40152359008789, + "step": 135, + "step_time": 6.334954546997324 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 161.0, + "completions/max_terminated_length": 161.0, + "completions/mean_length": 123.625, + "completions/mean_terminated_length": 123.625, + "completions/min_length": 93.0, + "completions/min_terminated_length": 93.0, + "entropy": 0.3462243974208832, + "epoch": 0.68, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.1878814697265625, + "learning_rate": 3.25e-07, + "loss": -0.10772529244422913, + "num_tokens": 1157952.0, + "reward": 44.186248779296875, + "reward_std": 51.78197479248047, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 44.08625030517578, + "rewards/supergames_reward/std": 51.7819709777832, + "step": 136, + "step_time": 6.067531232984038 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 155.0, + "completions/max_terminated_length": 155.0, + "completions/mean_length": 106.75, + "completions/mean_terminated_length": 106.75, + "completions/min_length": 74.0, + "completions/min_terminated_length": 74.0, + "entropy": 0.3662136197090149, + "epoch": 0.685, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.508971214294434, + "learning_rate": 3.2e-07, + "loss": -0.11172202974557877, + "num_tokens": 1163334.0, + "reward": 53.0574951171875, + "reward_std": 58.36653518676758, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 52.957496643066406, + "rewards/supergames_reward/std": 58.36653518676758, + "step": 137, + "step_time": 5.8906258280039765 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 175.0, + "completions/max_terminated_length": 175.0, + "completions/mean_length": 125.375, + "completions/mean_terminated_length": 125.375, + "completions/min_length": 90.0, + "completions/min_terminated_length": 90.0, + "entropy": 0.3352622091770172, + "epoch": 0.69, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.1937079429626465, + "learning_rate": 3.15e-07, + "loss": 0.08307601511478424, + "num_tokens": 1168929.0, + "reward": 11.796250343322754, + "reward_std": 53.48870849609375, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 11.696250915527344, + "rewards/supergames_reward/std": 53.48870849609375, + "step": 138, + "step_time": 6.595962718012743 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 254.0, + "completions/max_terminated_length": 254.0, + "completions/mean_length": 193.0, + "completions/mean_terminated_length": 193.0, + "completions/min_length": 111.0, + "completions/min_terminated_length": 111.0, + "entropy": 0.5486535429954529, + "epoch": 0.695, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.861871242523193, + "learning_rate": 3.1e-07, + "loss": 0.0321456640958786, + "num_tokens": 1176937.0, + "reward": 50.463748931884766, + "reward_std": 70.37553405761719, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": 50.38875198364258, + "rewards/supergames_reward/std": 70.31440734863281, + "step": 139, + "step_time": 9.544256388006033 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 261.0, + "completions/max_terminated_length": 261.0, + "completions/mean_length": 189.5, + "completions/mean_terminated_length": 189.5, + "completions/min_length": 95.0, + "completions/min_terminated_length": 95.0, + "entropy": 0.42230603098869324, + "epoch": 0.7, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.830677032470703, + "learning_rate": 3.05e-07, + "loss": -0.07222295552492142, + "num_tokens": 1186789.0, + "reward": 33.44999694824219, + "reward_std": 67.1298599243164, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 33.349998474121094, + "rewards/supergames_reward/std": 67.1298599243164, + "step": 140, + "step_time": 10.002322309010196 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 282.0, + "completions/max_terminated_length": 282.0, + "completions/mean_length": 197.375, + "completions/mean_terminated_length": 197.375, + "completions/min_length": 92.0, + "completions/min_terminated_length": 92.0, + "entropy": 0.33588480949401855, + "epoch": 0.705, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.028353452682495, + "learning_rate": 3e-07, + "loss": -0.07543312013149261, + "num_tokens": 1196672.0, + "reward": 17.10249900817871, + "reward_std": 73.94812774658203, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 17.002500534057617, + "rewards/supergames_reward/std": 73.94813537597656, + "step": 141, + "step_time": 10.684560916008195 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 235.0, + "completions/max_terminated_length": 235.0, + "completions/mean_length": 209.25, + "completions/mean_terminated_length": 209.25, + "completions/min_length": 167.0, + "completions/min_terminated_length": 167.0, + "entropy": 0.3622831702232361, + "epoch": 0.71, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.047222137451172, + "learning_rate": 2.95e-07, + "loss": -0.04569581151008606, + "num_tokens": 1206658.0, + "reward": 49.57624816894531, + "reward_std": 78.21060180664062, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 49.476253509521484, + "rewards/supergames_reward/std": 78.21060943603516, + "step": 142, + "step_time": 8.97679864402744 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 161.0, + "completions/max_terminated_length": 161.0, + "completions/mean_length": 116.25, + "completions/mean_terminated_length": 116.25, + "completions/min_length": 50.0, + "completions/min_terminated_length": 50.0, + "entropy": 0.43277010321617126, + "epoch": 0.715, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.345203876495361, + "learning_rate": 2.9e-07, + "loss": -0.16899466514587402, + "num_tokens": 1212140.0, + "reward": 27.42875099182129, + "reward_std": 48.59138870239258, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 27.328750610351562, + "rewards/supergames_reward/std": 48.591392517089844, + "step": 143, + "step_time": 6.1606669370085 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 205.0, + "completions/max_terminated_length": 205.0, + "completions/mean_length": 173.75, + "completions/mean_terminated_length": 173.75, + "completions/min_length": 134.0, + "completions/min_terminated_length": 134.0, + "entropy": 0.43806371092796326, + "epoch": 0.72, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.858840227127075, + "learning_rate": 2.8499999999999997e-07, + "loss": -0.06502802670001984, + "num_tokens": 1219954.0, + "reward": 38.69249725341797, + "reward_std": 68.78203582763672, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 38.592498779296875, + "rewards/supergames_reward/std": 68.78202819824219, + "step": 144, + "step_time": 7.88639413099736 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 169.0, + "completions/max_terminated_length": 169.0, + "completions/mean_length": 110.625, + "completions/mean_terminated_length": 110.625, + "completions/min_length": 85.0, + "completions/min_terminated_length": 85.0, + "entropy": 0.3383479118347168, + "epoch": 0.725, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.225733757019043, + "learning_rate": 2.8e-07, + "loss": -0.042058318853378296, + "num_tokens": 1225399.0, + "reward": 64.85874938964844, + "reward_std": 25.170387268066406, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 64.75875091552734, + "rewards/supergames_reward/std": 25.170391082763672, + "step": 145, + "step_time": 6.432729878026294 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 147.0, + "completions/max_terminated_length": 147.0, + "completions/mean_length": 129.5, + "completions/mean_terminated_length": 129.5, + "completions/min_length": 93.0, + "completions/min_terminated_length": 93.0, + "entropy": 0.3088987171649933, + "epoch": 0.73, + "frac_reward_zero_std": 1.0, + "grad_norm": 0.0, + "learning_rate": 2.75e-07, + "loss": 0.0, + "num_tokens": 1231043.0, + "reward": 100.0999984741211, + "reward_std": 0.0, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 100.0, + "rewards/supergames_reward/std": 0.0, + "step": 146, + "step_time": 5.634907905012369 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 160.0, + "completions/max_terminated_length": 160.0, + "completions/mean_length": 120.5, + "completions/mean_terminated_length": 120.5, + "completions/min_length": 80.0, + "completions/min_terminated_length": 80.0, + "entropy": 0.37292322516441345, + "epoch": 0.735, + "frac_reward_zero_std": 0.0, + "grad_norm": 5.174577713012695, + "learning_rate": 2.7e-07, + "loss": 0.11778222769498825, + "num_tokens": 1236607.0, + "reward": 34.82374572753906, + "reward_std": 58.513397216796875, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 34.72374725341797, + "rewards/supergames_reward/std": 58.513397216796875, + "step": 147, + "step_time": 6.078690663998714 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 172.0, + "completions/max_terminated_length": 172.0, + "completions/mean_length": 125.625, + "completions/mean_terminated_length": 125.625, + "completions/min_length": 88.0, + "completions/min_terminated_length": 88.0, + "entropy": 0.308654248714447, + "epoch": 0.74, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.8483362197875977, + "learning_rate": 2.65e-07, + "loss": -0.09181805700063705, + "num_tokens": 1242140.0, + "reward": 85.06375122070312, + "reward_std": 42.52893829345703, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 84.9637451171875, + "rewards/supergames_reward/std": 42.52893829345703, + "step": 148, + "step_time": 6.501119122985983 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 275.0, + "completions/max_terminated_length": 275.0, + "completions/mean_length": 211.125, + "completions/mean_terminated_length": 211.125, + "completions/min_length": 157.0, + "completions/min_terminated_length": 157.0, + "entropy": 0.5405827164649963, + "epoch": 0.745, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.774259567260742, + "learning_rate": 2.6e-07, + "loss": -0.03056861087679863, + "num_tokens": 1250205.0, + "reward": 84.63125610351562, + "reward_std": 34.3358154296875, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 84.53125, + "rewards/supergames_reward/std": 34.3358154296875, + "step": 149, + "step_time": 10.352729418984381 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 257.0, + "completions/max_terminated_length": 257.0, + "completions/mean_length": 218.5, + "completions/mean_terminated_length": 218.5, + "completions/min_length": 96.0, + "completions/min_terminated_length": 96.0, + "entropy": 0.3852894604206085, + "epoch": 0.75, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.6103010177612305, + "learning_rate": 2.55e-07, + "loss": 0.034789010882377625, + "num_tokens": 1260249.0, + "reward": 81.81375122070312, + "reward_std": 51.72132873535156, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 81.7137451171875, + "rewards/supergames_reward/std": 51.72132873535156, + "step": 150, + "step_time": 9.937249842012534 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 351.0, + "completions/max_terminated_length": 351.0, + "completions/mean_length": 232.625, + "completions/mean_terminated_length": 232.625, + "completions/min_length": 89.0, + "completions/min_terminated_length": 89.0, + "entropy": 0.4323120415210724, + "epoch": 0.755, + "frac_reward_zero_std": 0.0, + "grad_norm": 2.640308380126953, + "learning_rate": 2.5e-07, + "loss": -0.11181996762752533, + "num_tokens": 1271430.0, + "reward": 56.33625030517578, + "reward_std": 78.97452545166016, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 56.23625183105469, + "rewards/supergames_reward/std": 78.97452545166016, + "step": 151, + "step_time": 13.323547195002902 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 359.0, + "completions/max_terminated_length": 359.0, + "completions/mean_length": 247.5, + "completions/mean_terminated_length": 247.5, + "completions/min_length": 161.0, + "completions/min_terminated_length": 161.0, + "entropy": 0.4038960039615631, + "epoch": 0.76, + "frac_reward_zero_std": 0.0, + "grad_norm": 2.5704824924468994, + "learning_rate": 2.45e-07, + "loss": 0.060525521636009216, + "num_tokens": 1282722.0, + "reward": -68.53875732421875, + "reward_std": 50.590599060058594, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -68.63874816894531, + "rewards/supergames_reward/std": 50.590599060058594, + "step": 152, + "step_time": 13.593127576023107 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 383.0, + "completions/max_terminated_length": 383.0, + "completions/mean_length": 246.375, + "completions/mean_terminated_length": 246.375, + "completions/min_length": 150.0, + "completions/min_terminated_length": 150.0, + "entropy": 0.4110933542251587, + "epoch": 0.765, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.233025312423706, + "learning_rate": 2.4e-07, + "loss": 0.06547188758850098, + "num_tokens": 1294061.0, + "reward": 95.91499328613281, + "reward_std": 11.836968421936035, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 95.81500244140625, + "rewards/supergames_reward/std": 11.836968421936035, + "step": 153, + "step_time": 14.499402174988063 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 338.0, + "completions/max_terminated_length": 338.0, + "completions/mean_length": 233.125, + "completions/mean_terminated_length": 233.125, + "completions/min_length": 88.0, + "completions/min_terminated_length": 88.0, + "entropy": 0.34559860825538635, + "epoch": 0.77, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.6887102127075195, + "learning_rate": 2.3499999999999997e-07, + "loss": 0.15905122458934784, + "num_tokens": 1304262.0, + "reward": 88.92375183105469, + "reward_std": 31.611207962036133, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 88.82374572753906, + "rewards/supergames_reward/std": 31.6112117767334, + "step": 154, + "step_time": 12.704337224975461 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 208.0, + "completions/max_terminated_length": 208.0, + "completions/mean_length": 178.5, + "completions/mean_terminated_length": 178.5, + "completions/min_length": 125.0, + "completions/min_terminated_length": 125.0, + "entropy": 0.4646885097026825, + "epoch": 0.775, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.161937236785889, + "learning_rate": 2.3e-07, + "loss": -0.04354584217071533, + "num_tokens": 1312098.0, + "reward": 27.80875015258789, + "reward_std": 60.81097412109375, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 27.708751678466797, + "rewards/supergames_reward/std": 60.81097412109375, + "step": 155, + "step_time": 7.911955468007363 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 391.0, + "completions/max_terminated_length": 391.0, + "completions/mean_length": 235.375, + "completions/mean_terminated_length": 235.375, + "completions/min_length": 118.0, + "completions/min_terminated_length": 118.0, + "entropy": 0.40985700488090515, + "epoch": 0.78, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.12153959274292, + "learning_rate": 2.25e-07, + "loss": -0.024320494383573532, + "num_tokens": 1322301.0, + "reward": 64.14624786376953, + "reward_std": 49.90530014038086, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 64.04624938964844, + "rewards/supergames_reward/std": 49.90530014038086, + "step": 156, + "step_time": 14.631677505996777 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 330.0, + "completions/max_terminated_length": 330.0, + "completions/mean_length": 220.625, + "completions/mean_terminated_length": 220.625, + "completions/min_length": 93.0, + "completions/min_terminated_length": 93.0, + "entropy": 0.3543764054775238, + "epoch": 0.785, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.0111091136932373, + "learning_rate": 2.1999999999999998e-07, + "loss": -0.13996481895446777, + "num_tokens": 1333386.0, + "reward": 3.6812496185302734, + "reward_std": 103.5096664428711, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": 3.606250762939453, + "rewards/supergames_reward/std": 103.48104858398438, + "step": 157, + "step_time": 12.613216657977318 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 225.0, + "completions/max_terminated_length": 225.0, + "completions/mean_length": 193.125, + "completions/mean_terminated_length": 193.125, + "completions/min_length": 157.0, + "completions/min_terminated_length": 157.0, + "entropy": 0.5682947635650635, + "epoch": 0.79, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.67927885055542, + "learning_rate": 2.1499999999999998e-07, + "loss": 0.0535324327647686, + "num_tokens": 1341355.0, + "reward": -19.78999900817871, + "reward_std": 69.72611236572266, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -19.889997482299805, + "rewards/supergames_reward/std": 69.72611236572266, + "step": 158, + "step_time": 8.585773958999198 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 316.0, + "completions/max_terminated_length": 316.0, + "completions/mean_length": 237.625, + "completions/mean_terminated_length": 237.625, + "completions/min_length": 156.0, + "completions/min_terminated_length": 156.0, + "entropy": 0.3637359142303467, + "epoch": 0.795, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.033409357070923, + "learning_rate": 2.0999999999999997e-07, + "loss": 0.018884683027863503, + "num_tokens": 1352632.0, + "reward": 52.063751220703125, + "reward_std": 65.35551452636719, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 51.96375274658203, + "rewards/supergames_reward/std": 65.35551452636719, + "step": 159, + "step_time": 12.306456914986484 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 225.0, + "completions/max_terminated_length": 225.0, + "completions/mean_length": 184.0, + "completions/mean_terminated_length": 184.0, + "completions/min_length": 143.0, + "completions/min_terminated_length": 143.0, + "entropy": 0.47005388140678406, + "epoch": 0.8, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.585808515548706, + "learning_rate": 2.0499999999999997e-07, + "loss": -0.021751077845692635, + "num_tokens": 1360528.0, + "reward": 33.73124694824219, + "reward_std": 78.67134857177734, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 33.631248474121094, + "rewards/supergames_reward/std": 78.67134857177734, + "step": 160, + "step_time": 8.517202848976012 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 456.0, + "completions/max_terminated_length": 456.0, + "completions/mean_length": 248.75, + "completions/mean_terminated_length": 248.75, + "completions/min_length": 85.0, + "completions/min_terminated_length": 85.0, + "entropy": 0.39807969331741333, + "epoch": 0.805, + "frac_reward_zero_std": 0.0, + "grad_norm": 2.8931398391723633, + "learning_rate": 2e-07, + "loss": -0.05018138885498047, + "num_tokens": 1371854.0, + "reward": 4.002498626708984, + "reward_std": 103.03063201904297, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": 3.9274978637695312, + "rewards/supergames_reward/std": 103.00178527832031, + "step": 161, + "step_time": 17.225145012984285 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 400.0, + "completions/max_terminated_length": 400.0, + "completions/mean_length": 231.125, + "completions/mean_terminated_length": 231.125, + "completions/min_length": 95.0, + "completions/min_terminated_length": 95.0, + "entropy": 0.49391740560531616, + "epoch": 0.81, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.5471365451812744, + "learning_rate": 1.9499999999999999e-07, + "loss": -0.18011468648910522, + "num_tokens": 1383015.0, + "reward": 67.2449951171875, + "reward_std": 69.01223754882812, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 67.14500427246094, + "rewards/supergames_reward/std": 69.01224517822266, + "step": 162, + "step_time": 15.188760302000446 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 262.0, + "completions/max_terminated_length": 262.0, + "completions/mean_length": 221.875, + "completions/mean_terminated_length": 221.875, + "completions/min_length": 89.0, + "completions/min_terminated_length": 89.0, + "entropy": 0.388366162776947, + "epoch": 0.815, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.662609815597534, + "learning_rate": 1.8999999999999998e-07, + "loss": 1.8455075405654497e-05, + "num_tokens": 1393150.0, + "reward": 59.23624801635742, + "reward_std": 76.00599670410156, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 59.13624954223633, + "rewards/supergames_reward/std": 76.00599670410156, + "step": 163, + "step_time": 10.180389262997778 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 396.0, + "completions/max_terminated_length": 396.0, + "completions/mean_length": 323.625, + "completions/mean_terminated_length": 323.625, + "completions/min_length": 256.0, + "completions/min_terminated_length": 256.0, + "entropy": 0.45717528462409973, + "epoch": 0.82, + "frac_reward_zero_std": 0.0, + "grad_norm": 2.580068826675415, + "learning_rate": 1.85e-07, + "loss": 0.024594159796833992, + "num_tokens": 1405067.0, + "reward": 6.318748474121094, + "reward_std": 88.21497344970703, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 6.21875, + "rewards/supergames_reward/std": 88.21497344970703, + "step": 164, + "step_time": 15.074436638999032 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 163.0, + "completions/max_terminated_length": 163.0, + "completions/mean_length": 133.875, + "completions/mean_terminated_length": 133.875, + "completions/min_length": 90.0, + "completions/min_terminated_length": 90.0, + "entropy": 0.3449660539627075, + "epoch": 0.825, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.045867443084717, + "learning_rate": 1.8e-07, + "loss": -0.02517889253795147, + "num_tokens": 1410730.0, + "reward": 84.35499572753906, + "reward_std": 38.135921478271484, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 84.2550048828125, + "rewards/supergames_reward/std": 38.13592529296875, + "step": 165, + "step_time": 6.275273430015659 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 144.0, + "completions/max_terminated_length": 144.0, + "completions/mean_length": 105.875, + "completions/mean_terminated_length": 105.875, + "completions/min_length": 83.0, + "completions/min_terminated_length": 83.0, + "entropy": 0.286411851644516, + "epoch": 0.83, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.855896472930908, + "learning_rate": 1.75e-07, + "loss": 0.014497723430395126, + "num_tokens": 1416081.0, + "reward": 48.40374755859375, + "reward_std": 47.61183547973633, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 48.303749084472656, + "rewards/supergames_reward/std": 47.611839294433594, + "step": 166, + "step_time": 5.5562305669882335 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 269.0, + "completions/max_terminated_length": 269.0, + "completions/mean_length": 226.125, + "completions/mean_terminated_length": 226.125, + "completions/min_length": 146.0, + "completions/min_terminated_length": 146.0, + "entropy": 0.39676305651664734, + "epoch": 0.835, + "frac_reward_zero_std": 0.0, + "grad_norm": 2.633286237716675, + "learning_rate": 1.7000000000000001e-07, + "loss": -0.008210637606680393, + "num_tokens": 1426194.0, + "reward": 67.51000213623047, + "reward_std": 58.9198112487793, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 67.41000366210938, + "rewards/supergames_reward/std": 58.9198112487793, + "step": 167, + "step_time": 10.435288048989605 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 210.0, + "completions/max_terminated_length": 210.0, + "completions/mean_length": 168.5, + "completions/mean_terminated_length": 168.5, + "completions/min_length": 137.0, + "completions/min_terminated_length": 137.0, + "entropy": 0.4233987629413605, + "epoch": 0.84, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.654274940490723, + "learning_rate": 1.65e-07, + "loss": -0.0319184847176075, + "num_tokens": 1433990.0, + "reward": 50.342498779296875, + "reward_std": 67.66273498535156, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": 50.26750183105469, + "rewards/supergames_reward/std": 67.5992202758789, + "step": 168, + "step_time": 8.111304188991198 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 397.0, + "completions/max_terminated_length": 397.0, + "completions/mean_length": 217.75, + "completions/mean_terminated_length": 217.75, + "completions/min_length": 90.0, + "completions/min_terminated_length": 90.0, + "entropy": 0.41511332988739014, + "epoch": 0.845, + "frac_reward_zero_std": 0.0, + "grad_norm": 5.156782150268555, + "learning_rate": 1.6e-07, + "loss": 0.30166593194007874, + "num_tokens": 1444068.0, + "reward": 48.64500045776367, + "reward_std": 51.50307846069336, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 48.54499816894531, + "rewards/supergames_reward/std": 51.50307846069336, + "step": 169, + "step_time": 14.915170635009417 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 245.0, + "completions/max_terminated_length": 245.0, + "completions/mean_length": 218.625, + "completions/mean_terminated_length": 218.625, + "completions/min_length": 140.0, + "completions/min_terminated_length": 140.0, + "entropy": 0.3981763422489166, + "epoch": 0.85, + "frac_reward_zero_std": 0.0, + "grad_norm": 2.818580389022827, + "learning_rate": 1.55e-07, + "loss": -0.08915036171674728, + "num_tokens": 1454153.0, + "reward": 82.69249725341797, + "reward_std": 24.65846824645996, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 82.59249877929688, + "rewards/supergames_reward/std": 24.65846824645996, + "step": 170, + "step_time": 9.446723325003404 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 386.0, + "completions/max_terminated_length": 386.0, + "completions/mean_length": 280.375, + "completions/mean_terminated_length": 280.375, + "completions/min_length": 98.0, + "completions/min_terminated_length": 98.0, + "entropy": 0.39121532440185547, + "epoch": 0.855, + "frac_reward_zero_std": 0.0, + "grad_norm": 2.067988634109497, + "learning_rate": 1.5e-07, + "loss": -0.28749391436576843, + "num_tokens": 1465692.0, + "reward": 66.02375030517578, + "reward_std": 70.814697265625, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 65.92375183105469, + "rewards/supergames_reward/std": 70.814697265625, + "step": 171, + "step_time": 14.603269515006104 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 188.0, + "completions/max_terminated_length": 188.0, + "completions/mean_length": 131.125, + "completions/mean_terminated_length": 131.125, + "completions/min_length": 87.0, + "completions/min_terminated_length": 87.0, + "entropy": 0.435172438621521, + "epoch": 0.86, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.832774639129639, + "learning_rate": 1.45e-07, + "loss": -0.1204412579536438, + "num_tokens": 1471253.0, + "reward": 39.80875015258789, + "reward_std": 51.58554458618164, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 39.7087516784668, + "rewards/supergames_reward/std": 51.585548400878906, + "step": 172, + "step_time": 7.057335219986271 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 391.0, + "completions/max_terminated_length": 391.0, + "completions/mean_length": 271.375, + "completions/mean_terminated_length": 271.375, + "completions/min_length": 176.0, + "completions/min_terminated_length": 176.0, + "entropy": 0.4081321060657501, + "epoch": 0.865, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.2880308628082275, + "learning_rate": 1.4e-07, + "loss": 0.03748488798737526, + "num_tokens": 1482744.0, + "reward": 53.62249755859375, + "reward_std": 75.1306381225586, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 53.522499084472656, + "rewards/supergames_reward/std": 75.1306381225586, + "step": 173, + "step_time": 14.96632127999328 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 304.0, + "completions/max_terminated_length": 304.0, + "completions/mean_length": 203.875, + "completions/mean_terminated_length": 203.875, + "completions/min_length": 140.0, + "completions/min_terminated_length": 140.0, + "entropy": 0.4152996242046356, + "epoch": 0.87, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.1565890312194824, + "learning_rate": 1.35e-07, + "loss": -0.047037553042173386, + "num_tokens": 1492679.0, + "reward": 37.48374938964844, + "reward_std": 72.71965789794922, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 37.383750915527344, + "rewards/supergames_reward/std": 72.71966552734375, + "step": 174, + "step_time": 11.566132662002929 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 187.0, + "completions/max_terminated_length": 187.0, + "completions/mean_length": 133.375, + "completions/mean_terminated_length": 133.375, + "completions/min_length": 95.0, + "completions/min_terminated_length": 95.0, + "entropy": 0.3746980130672455, + "epoch": 0.875, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.724977970123291, + "learning_rate": 1.3e-07, + "loss": 0.05197479575872421, + "num_tokens": 1498330.0, + "reward": 37.86625289916992, + "reward_std": 52.43158721923828, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 37.76625061035156, + "rewards/supergames_reward/std": 52.43158721923828, + "step": 175, + "step_time": 7.080777855007909 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 448.0, + "completions/max_terminated_length": 448.0, + "completions/mean_length": 277.25, + "completions/mean_terminated_length": 277.25, + "completions/min_length": 96.0, + "completions/min_terminated_length": 96.0, + "entropy": 0.397225022315979, + "epoch": 0.88, + "frac_reward_zero_std": 0.0, + "grad_norm": 2.438502788543701, + "learning_rate": 1.25e-07, + "loss": 0.1022070124745369, + "num_tokens": 1509860.0, + "reward": -41.323753356933594, + "reward_std": 79.704345703125, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -41.42374801635742, + "rewards/supergames_reward/std": 79.704345703125, + "step": 176, + "step_time": 16.852883081999607 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 187.0, + "completions/max_terminated_length": 187.0, + "completions/mean_length": 145.875, + "completions/mean_terminated_length": 145.875, + "completions/min_length": 92.0, + "completions/min_terminated_length": 92.0, + "entropy": 0.34882763028144836, + "epoch": 0.885, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.683617115020752, + "learning_rate": 1.2e-07, + "loss": 0.0609549917280674, + "num_tokens": 1515587.0, + "reward": 84.27999877929688, + "reward_std": 38.06671905517578, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 84.17999267578125, + "rewards/supergames_reward/std": 38.06672286987305, + "step": 177, + "step_time": 7.09286071601673 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 217.0, + "completions/max_terminated_length": 217.0, + "completions/mean_length": 181.875, + "completions/mean_terminated_length": 181.875, + "completions/min_length": 146.0, + "completions/min_terminated_length": 146.0, + "entropy": 0.423494815826416, + "epoch": 0.89, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.3993117809295654, + "learning_rate": 1.15e-07, + "loss": 0.04823008552193642, + "num_tokens": 1523490.0, + "reward": 22.46124839782715, + "reward_std": 89.67146301269531, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 22.361249923706055, + "rewards/supergames_reward/std": 89.67146301269531, + "step": 178, + "step_time": 8.281584940006724 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 315.0, + "completions/max_terminated_length": 315.0, + "completions/mean_length": 273.25, + "completions/mean_terminated_length": 273.25, + "completions/min_length": 234.0, + "completions/min_terminated_length": 234.0, + "entropy": 0.3831624984741211, + "epoch": 0.895, + "frac_reward_zero_std": 0.0, + "grad_norm": 2.7383997440338135, + "learning_rate": 1.0999999999999999e-07, + "loss": -0.010000350885093212, + "num_tokens": 1535012.0, + "reward": 2.237499237060547, + "reward_std": 93.32353210449219, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 2.137500762939453, + "rewards/supergames_reward/std": 93.32353973388672, + "step": 179, + "step_time": 12.098909514024854 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 395.0, + "completions/max_terminated_length": 395.0, + "completions/mean_length": 197.375, + "completions/mean_terminated_length": 197.375, + "completions/min_length": 132.0, + "completions/min_terminated_length": 132.0, + "entropy": 0.5201653242111206, + "epoch": 0.9, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.4596073627471924, + "learning_rate": 1.0499999999999999e-07, + "loss": -0.11710439622402191, + "num_tokens": 1543023.0, + "reward": 88.1875, + "reward_std": 33.69363784790039, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 88.0875015258789, + "rewards/supergames_reward/std": 33.693641662597656, + "step": 180, + "step_time": 14.662479719001567 + } + ], + "logging_steps": 1, + "max_steps": 200, + "num_input_tokens_seen": 1543023, + "num_train_epochs": 1, + "save_steps": 20, + "stateful_callbacks": { + "TrainerControl": { + "args": { + "should_epoch_stop": false, + "should_evaluate": false, + "should_log": false, + "should_save": true, + "should_training_stop": false + }, + "attributes": {} + } + }, + "total_flos": 0.0, + "train_batch_size": 8, + "trial_name": null, + "trial_params": null +} diff --git a/checkpoint-180/training_args.bin b/checkpoint-180/training_args.bin new file mode 100644 index 0000000..730dec0 --- /dev/null +++ b/checkpoint-180/training_args.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:691fe5be2d7d1d56160dae63866d136877a72efb30e15ff7b3249192f998a788 +size 7185 diff --git a/checkpoint-20/chat_template.jinja b/checkpoint-20/chat_template.jinja new file mode 100644 index 0000000..bdf7919 --- /dev/null +++ b/checkpoint-20/chat_template.jinja @@ -0,0 +1,54 @@ +{%- if tools %} + {{- '<|im_start|>system\n' }} + {%- if messages[0]['role'] == 'system' %} + {{- messages[0]['content'] }} + {%- else %} + {{- 'You are Qwen, created by Alibaba Cloud. You are a helpful assistant.' }} + {%- endif %} + {{- "\n\n# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }} + {%- for tool in tools %} + {{- "\n" }} + {{- tool | tojson }} + {%- endfor %} + {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }} +{%- else %} + {%- if messages[0]['role'] == 'system' %} + {{- '<|im_start|>system\n' + messages[0]['content'] + '<|im_end|>\n' }} + {%- else %} + {{- '<|im_start|>system\nYou are Qwen, created by Alibaba Cloud. You are a helpful assistant.<|im_end|>\n' }} + {%- endif %} +{%- endif %} +{%- for message in messages %} + {%- if (message.role == "user") or (message.role == "system" and not loop.first) or (message.role == "assistant" and not message.tool_calls) %} + {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }} + {%- elif message.role == "assistant" %} + {{- '<|im_start|>' + message.role }} + {%- if message.content %} + {{- '\n' + message.content }} + {%- endif %} + {%- for tool_call in message.tool_calls %} + {%- if tool_call.function is defined %} + {%- set tool_call = tool_call.function %} + {%- endif %} + {{- '\n\n{"name": "' }} + {{- tool_call.name }} + {{- '", "arguments": ' }} + {{- tool_call.arguments | tojson }} + {{- '}\n' }} + {%- endfor %} + {{- '<|im_end|>\n' }} + {%- elif message.role == "tool" %} + {%- if (loop.index0 == 0) or (messages[loop.index0 - 1].role != "tool") %} + {{- '<|im_start|>user' }} + {%- endif %} + {{- '\n\n' }} + {{- message.content }} + {{- '\n' }} + {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %} + {{- '<|im_end|>\n' }} + {%- endif %} + {%- endif %} +{%- endfor %} +{%- if add_generation_prompt %} + {{- '<|im_start|>assistant\n' }} +{%- endif %} diff --git a/checkpoint-20/config.json b/checkpoint-20/config.json new file mode 100644 index 0000000..f57b09b --- /dev/null +++ b/checkpoint-20/config.json @@ -0,0 +1,61 @@ +{ + "architectures": [ + "Qwen2ForCausalLM" + ], + "attention_dropout": 0.0, + "bos_token_id": null, + "dtype": "float32", + "eos_token_id": 151645, + "hidden_act": "silu", + "hidden_size": 1536, + "initializer_range": 0.02, + "intermediate_size": 8960, + "layer_types": [ + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention" + ], + "max_position_embeddings": 32768, + "max_window_layers": 21, + "model_type": "qwen2", + "num_attention_heads": 12, + "num_hidden_layers": 28, + "num_key_value_heads": 2, + "pad_token_id": 151643, + "rms_norm_eps": 1e-06, + "rope_parameters": { + "rope_theta": 1000000.0, + "rope_type": "default" + }, + "sliding_window": null, + "tie_word_embeddings": true, + "transformers_version": "5.6.2", + "use_cache": false, + "use_sliding_window": false, + "vocab_size": 151936 +} diff --git a/checkpoint-20/generation_config.json b/checkpoint-20/generation_config.json new file mode 100644 index 0000000..1b2bba9 --- /dev/null +++ b/checkpoint-20/generation_config.json @@ -0,0 +1,13 @@ +{ + "do_sample": true, + "eos_token_id": [ + 151645, + 151643 + ], + "pad_token_id": 151643, + "repetition_penalty": 1.1, + "temperature": 0.7, + "top_k": 20, + "top_p": 0.8, + "transformers_version": "5.6.2" +} diff --git a/checkpoint-20/model.safetensors b/checkpoint-20/model.safetensors new file mode 100644 index 0000000..db32b2f --- /dev/null +++ b/checkpoint-20/model.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:a39aec9f313a668fc041aff46341277464455362e1e1a310b26d9e9d14adf24d +size 6174895536 diff --git a/checkpoint-20/optimizer.pt b/checkpoint-20/optimizer.pt new file mode 100644 index 0000000..1dcbec2 --- /dev/null +++ b/checkpoint-20/optimizer.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:7700c459a5327b2ab9aee3a5bb3d23e73c8bb01230b33ba79014e18313adfef8 +size 12350013801 diff --git a/checkpoint-20/rng_state.pth b/checkpoint-20/rng_state.pth new file mode 100644 index 0000000..f9f8073 --- /dev/null +++ b/checkpoint-20/rng_state.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:93bf4adc8891698c782639b2c71621fcef18a7f4aa56009307e85e227fac0b8f +size 14645 diff --git a/checkpoint-20/scheduler.pt b/checkpoint-20/scheduler.pt new file mode 100644 index 0000000..1d7ac3e --- /dev/null +++ b/checkpoint-20/scheduler.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:6532ac530958ee6e596569b01cb0102818e20b71baecdf6e49fd861fbe7fab75 +size 1465 diff --git a/checkpoint-20/tokenizer.json b/checkpoint-20/tokenizer.json new file mode 100644 index 0000000..34510ff --- /dev/null +++ b/checkpoint-20/tokenizer.json @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:3fd169731d2cbde95e10bf356d66d5997fd885dd8dbb6fb4684da3f23b2585d8 +size 11421892 diff --git a/checkpoint-20/tokenizer_config.json b/checkpoint-20/tokenizer_config.json new file mode 100644 index 0000000..770e41d --- /dev/null +++ b/checkpoint-20/tokenizer_config.json @@ -0,0 +1,30 @@ +{ + "add_prefix_space": false, + "backend": "tokenizers", + "bos_token": null, + "clean_up_tokenization_spaces": false, + "eos_token": "<|im_end|>", + "errors": "replace", + "extra_special_tokens": [ + "<|im_start|>", + "<|im_end|>", + "<|object_ref_start|>", + "<|object_ref_end|>", + "<|box_start|>", + "<|box_end|>", + "<|quad_start|>", + "<|quad_end|>", + "<|vision_start|>", + "<|vision_end|>", + "<|vision_pad|>", + "<|image_pad|>", + "<|video_pad|>" + ], + "is_local": false, + "local_files_only": false, + "model_max_length": 131072, + "pad_token": "<|endoftext|>", + "split_special_tokens": false, + "tokenizer_class": "Qwen2Tokenizer", + "unk_token": null +} diff --git a/checkpoint-20/trainer_state.json b/checkpoint-20/trainer_state.json new file mode 100644 index 0000000..a4963a7 --- /dev/null +++ b/checkpoint-20/trainer_state.json @@ -0,0 +1,614 @@ +{ + "best_global_step": null, + "best_metric": null, + "best_model_checkpoint": null, + "epoch": 0.1, + "eval_steps": 500, + "global_step": 20, + "is_hyper_param_search": false, + "is_local_process_zero": true, + "is_world_process_zero": true, + "log_history": [ + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 478.0, + "completions/max_terminated_length": 478.0, + "completions/mean_length": 191.0, + "completions/mean_terminated_length": 191.0, + "completions/min_length": 66.0, + "completions/min_terminated_length": 66.0, + "entropy": 0.5566893219947815, + "epoch": 0.005, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.370361328125, + "learning_rate": 1e-06, + "loss": -0.24658073484897614, + "num_tokens": 10832.0, + "reward": -3.3000030517578125, + "reward_std": 85.62333679199219, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -3.4000015258789062, + "rewards/supergames_reward/std": 85.62333679199219, + "step": 1, + "step_time": 22.779711072013015 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 216.0, + "completions/max_terminated_length": 216.0, + "completions/mean_length": 131.5, + "completions/mean_terminated_length": 131.5, + "completions/min_length": 60.0, + "completions/min_terminated_length": 60.0, + "entropy": 0.6395841240882874, + "epoch": 0.01, + "frac_reward_zero_std": 0.0, + "grad_norm": 8.09774398803711, + "learning_rate": 9.95e-07, + "loss": 0.20567649602890015, + "num_tokens": 16404.0, + "reward": -12.422499656677246, + "reward_std": 7.134707450866699, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -12.522500038146973, + "rewards/supergames_reward/std": 7.134707927703857, + "step": 2, + "step_time": 9.960156448010821 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 287.0, + "completions/max_terminated_length": 287.0, + "completions/mean_length": 156.625, + "completions/mean_terminated_length": 156.625, + "completions/min_length": 77.0, + "completions/min_terminated_length": 77.0, + "entropy": 0.562222421169281, + "epoch": 0.015, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.9053616523742676, + "learning_rate": 9.9e-07, + "loss": 0.09602774679660797, + "num_tokens": 26953.0, + "reward": -74.94999694824219, + "reward_std": 70.73094177246094, + "rewards/json_format_reward/mean": 0.05000000074505806, + "rewards/json_format_reward/std": 0.09258200973272324, + "rewards/supergames_reward/mean": -75.0, + "rewards/supergames_reward/std": 70.71067810058594, + "step": 3, + "step_time": 13.686320498993155 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 268.0, + "completions/max_terminated_length": 268.0, + "completions/mean_length": 154.125, + "completions/mean_terminated_length": 154.125, + "completions/min_length": 53.0, + "completions/min_terminated_length": 53.0, + "entropy": 0.7323317527770996, + "epoch": 0.02, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.933310031890869, + "learning_rate": 9.849999999999999e-07, + "loss": 0.28249427676200867, + "num_tokens": 36514.0, + "reward": -40.7599983215332, + "reward_std": 81.73140716552734, + "rewards/json_format_reward/mean": 0.05000000074505806, + "rewards/json_format_reward/std": 0.09258200973272324, + "rewards/supergames_reward/mean": -40.80999755859375, + "rewards/supergames_reward/std": 81.68995666503906, + "step": 4, + "step_time": 12.805880262021674 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 401.0, + "completions/max_terminated_length": 401.0, + "completions/mean_length": 205.25, + "completions/mean_terminated_length": 205.25, + "completions/min_length": 115.0, + "completions/min_terminated_length": 115.0, + "entropy": 0.6167430281639099, + "epoch": 0.025, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.344135046005249, + "learning_rate": 9.8e-07, + "loss": 0.1433994024991989, + "num_tokens": 47476.0, + "reward": -97.6612548828125, + "reward_std": 6.3321428298950195, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -97.76124572753906, + "rewards/supergames_reward/std": 6.3321428298950195, + "step": 5, + "step_time": 18.789364666008623 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 276.0, + "completions/max_terminated_length": 276.0, + "completions/mean_length": 167.25, + "completions/mean_terminated_length": 167.25, + "completions/min_length": 73.0, + "completions/min_terminated_length": 73.0, + "entropy": 0.7583790421485901, + "epoch": 0.03, + "frac_reward_zero_std": 0.0, + "grad_norm": 5.172371864318848, + "learning_rate": 9.75e-07, + "loss": -0.12127404659986496, + "num_tokens": 55246.0, + "reward": -58.687503814697266, + "reward_std": 58.64059829711914, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": -58.76250457763672, + "rewards/supergames_reward/std": 58.620460510253906, + "step": 6, + "step_time": 12.948570273991209 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 261.0, + "completions/max_terminated_length": 261.0, + "completions/mean_length": 133.875, + "completions/mean_terminated_length": 133.875, + "completions/min_length": 57.0, + "completions/min_terminated_length": 57.0, + "entropy": 0.7447654604911804, + "epoch": 0.035, + "frac_reward_zero_std": 0.0, + "grad_norm": 6.428595066070557, + "learning_rate": 9.7e-07, + "loss": 0.10135584324598312, + "num_tokens": 60837.0, + "reward": -32.45000076293945, + "reward_std": 41.754486083984375, + "rewards/json_format_reward/mean": 0.05000000074505806, + "rewards/json_format_reward/std": 0.09258200973272324, + "rewards/supergames_reward/mean": -32.5, + "rewards/supergames_reward/std": 41.66190719604492, + "step": 7, + "step_time": 12.058315072004916 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 414.0, + "completions/max_terminated_length": 414.0, + "completions/mean_length": 209.125, + "completions/mean_terminated_length": 209.125, + "completions/min_length": 104.0, + "completions/min_terminated_length": 104.0, + "entropy": 0.3860666751861572, + "epoch": 0.04, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.806255578994751, + "learning_rate": 9.649999999999999e-07, + "loss": -0.33238139748573303, + "num_tokens": 70838.0, + "reward": -80.50375366210938, + "reward_std": 38.82380294799805, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -80.60375213623047, + "rewards/supergames_reward/std": 38.82379913330078, + "step": 8, + "step_time": 19.323370319994865 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.25, + "completions/max_length": 512.0, + "completions/max_terminated_length": 199.0, + "completions/mean_length": 247.125, + "completions/mean_terminated_length": 158.83334350585938, + "completions/min_length": 106.0, + "completions/min_terminated_length": 106.0, + "entropy": 0.5816237330436707, + "epoch": 0.045, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.3209943771362305, + "learning_rate": 9.6e-07, + "loss": 0.36752766370773315, + "num_tokens": 82135.0, + "reward": -27.688751220703125, + "reward_std": 82.8358154296875, + "rewards/json_format_reward/mean": 0.05000000074505806, + "rewards/json_format_reward/std": 0.09258200973272324, + "rewards/supergames_reward/mean": -27.738750457763672, + "rewards/supergames_reward/std": 82.78590393066406, + "step": 9, + "step_time": 21.948575104994234 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 240.0, + "completions/max_terminated_length": 240.0, + "completions/mean_length": 153.0, + "completions/mean_terminated_length": 153.0, + "completions/min_length": 51.0, + "completions/min_terminated_length": 51.0, + "entropy": 0.6301521062850952, + "epoch": 0.05, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.125359058380127, + "learning_rate": 9.55e-07, + "loss": 0.06129350885748863, + "num_tokens": 89807.0, + "reward": -78.39624786376953, + "reward_std": 40.17042922973633, + "rewards/json_format_reward/mean": 0.05000000074505806, + "rewards/json_format_reward/std": 0.09258200973272324, + "rewards/supergames_reward/mean": -78.44625091552734, + "rewards/supergames_reward/std": 40.1396484375, + "step": 10, + "step_time": 9.098202993016457 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 133.0, + "completions/max_terminated_length": 133.0, + "completions/mean_length": 112.0, + "completions/mean_terminated_length": 112.0, + "completions/min_length": 68.0, + "completions/min_terminated_length": 68.0, + "entropy": 0.5521990656852722, + "epoch": 0.055, + "frac_reward_zero_std": 0.0, + "grad_norm": 6.616060256958008, + "learning_rate": 9.499999999999999e-07, + "loss": -0.006313305348157883, + "num_tokens": 95231.0, + "reward": -12.161249160766602, + "reward_std": 6.39578104019165, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -12.261249542236328, + "rewards/supergames_reward/std": 6.39578104019165, + "step": 11, + "step_time": 5.136311663984088 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 396.0, + "completions/max_terminated_length": 396.0, + "completions/mean_length": 214.875, + "completions/mean_terminated_length": 214.875, + "completions/min_length": 55.0, + "completions/min_terminated_length": 55.0, + "entropy": 0.5505905151367188, + "epoch": 0.06, + "frac_reward_zero_std": 0.0, + "grad_norm": 5.289583206176758, + "learning_rate": 9.45e-07, + "loss": -0.29802003502845764, + "num_tokens": 106238.0, + "reward": -91.66999816894531, + "reward_std": 23.2779541015625, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -91.77000427246094, + "rewards/supergames_reward/std": 23.277956008911133, + "step": 12, + "step_time": 14.945365622988902 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 269.0, + "completions/max_terminated_length": 269.0, + "completions/mean_length": 167.375, + "completions/mean_terminated_length": 167.375, + "completions/min_length": 113.0, + "completions/min_terminated_length": 113.0, + "entropy": 0.540415346622467, + "epoch": 0.065, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.858291149139404, + "learning_rate": 9.399999999999999e-07, + "loss": 0.24004128575325012, + "num_tokens": 116937.0, + "reward": -20.38249969482422, + "reward_std": 87.55204010009766, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": -20.457500457763672, + "rewards/supergames_reward/std": 87.52605438232422, + "step": 13, + "step_time": 10.482285185018554 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 254.0, + "completions/max_terminated_length": 254.0, + "completions/mean_length": 156.875, + "completions/mean_terminated_length": 156.875, + "completions/min_length": 114.0, + "completions/min_terminated_length": 114.0, + "entropy": 0.7237679958343506, + "epoch": 0.07, + "frac_reward_zero_std": 0.0, + "grad_norm": 5.083980560302734, + "learning_rate": 9.35e-07, + "loss": -0.19025824964046478, + "num_tokens": 124640.0, + "reward": -51.397499084472656, + "reward_std": 54.486454010009766, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -51.49749755859375, + "rewards/supergames_reward/std": 54.486454010009766, + "step": 14, + "step_time": 9.597231683001155 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 236.0, + "completions/max_terminated_length": 236.0, + "completions/mean_length": 141.375, + "completions/mean_terminated_length": 141.375, + "completions/min_length": 91.0, + "completions/min_terminated_length": 91.0, + "entropy": 0.754334568977356, + "epoch": 0.075, + "frac_reward_zero_std": 0.0, + "grad_norm": 5.720428943634033, + "learning_rate": 9.3e-07, + "loss": -0.16830675303936005, + "num_tokens": 132219.0, + "reward": -56.226253509521484, + "reward_std": 60.115657806396484, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -56.32625198364258, + "rewards/supergames_reward/std": 60.11566162109375, + "step": 15, + "step_time": 8.998362872982398 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 209.0, + "completions/max_terminated_length": 209.0, + "completions/mean_length": 136.375, + "completions/mean_terminated_length": 136.375, + "completions/min_length": 97.0, + "completions/min_terminated_length": 97.0, + "entropy": 0.6280568838119507, + "epoch": 0.08, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.9681806564331055, + "learning_rate": 9.25e-07, + "loss": -0.01910916529595852, + "num_tokens": 139726.0, + "reward": -68.24000549316406, + "reward_std": 52.238914489746094, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -68.33999633789062, + "rewards/supergames_reward/std": 52.238914489746094, + "step": 16, + "step_time": 8.034480294008972 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 365.0, + "completions/max_terminated_length": 365.0, + "completions/mean_length": 211.875, + "completions/mean_terminated_length": 211.875, + "completions/min_length": 100.0, + "completions/min_terminated_length": 100.0, + "entropy": 0.899176836013794, + "epoch": 0.085, + "frac_reward_zero_std": 0.0, + "grad_norm": 6.691746711730957, + "learning_rate": 9.2e-07, + "loss": 0.25513291358947754, + "num_tokens": 147869.0, + "reward": -99.92500305175781, + "reward_std": 0.07070883363485336, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": -100.0, + "rewards/supergames_reward/std": 0.0, + "step": 17, + "step_time": 13.655792869016295 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 387.0, + "completions/max_terminated_length": 387.0, + "completions/mean_length": 238.0, + "completions/mean_terminated_length": 238.0, + "completions/min_length": 87.0, + "completions/min_terminated_length": 87.0, + "entropy": 0.5596873760223389, + "epoch": 0.09, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.6711347103118896, + "learning_rate": 9.15e-07, + "loss": -0.2628335654735565, + "num_tokens": 159101.0, + "reward": -24.900001525878906, + "reward_std": 103.50983428955078, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -25.0, + "rewards/supergames_reward/std": 103.50983428955078, + "step": 18, + "step_time": 14.784329420013819 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.125, + "completions/max_length": 512.0, + "completions/max_terminated_length": 277.0, + "completions/mean_length": 219.75, + "completions/mean_terminated_length": 178.00001525878906, + "completions/min_length": 90.0, + "completions/min_terminated_length": 90.0, + "entropy": 0.9245517253875732, + "epoch": 0.095, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.55502986907959, + "learning_rate": 9.1e-07, + "loss": 0.13060179352760315, + "num_tokens": 167299.0, + "reward": -86.89125061035156, + "reward_std": 36.95603561401367, + "rewards/json_format_reward/mean": 0.05000000074505806, + "rewards/json_format_reward/std": 0.09258200973272324, + "rewards/supergames_reward/mean": -86.94125366210938, + "rewards/supergames_reward/std": 36.93572235107422, + "step": 19, + "step_time": 18.832684700988466 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 267.0, + "completions/max_terminated_length": 267.0, + "completions/mean_length": 177.25, + "completions/mean_terminated_length": 177.25, + "completions/min_length": 93.0, + "completions/min_terminated_length": 93.0, + "entropy": 0.7331254482269287, + "epoch": 0.1, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.538497447967529, + "learning_rate": 9.05e-07, + "loss": -0.16418980062007904, + "num_tokens": 175141.0, + "reward": -68.52874755859375, + "reward_std": 58.48719787597656, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -68.62875366210938, + "rewards/supergames_reward/std": 58.48720169067383, + "step": 20, + "step_time": 10.090975169994636 + } + ], + "logging_steps": 1, + "max_steps": 200, + "num_input_tokens_seen": 175141, + "num_train_epochs": 1, + "save_steps": 20, + "stateful_callbacks": { + "TrainerControl": { + "args": { + "should_epoch_stop": false, + "should_evaluate": false, + "should_log": false, + "should_save": true, + "should_training_stop": false + }, + "attributes": {} + } + }, + "total_flos": 0.0, + "train_batch_size": 8, + "trial_name": null, + "trial_params": null +} diff --git a/checkpoint-20/training_args.bin b/checkpoint-20/training_args.bin new file mode 100644 index 0000000..730dec0 --- /dev/null +++ b/checkpoint-20/training_args.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:691fe5be2d7d1d56160dae63866d136877a72efb30e15ff7b3249192f998a788 +size 7185 diff --git a/checkpoint-200/chat_template.jinja b/checkpoint-200/chat_template.jinja new file mode 100644 index 0000000..bdf7919 --- /dev/null +++ b/checkpoint-200/chat_template.jinja @@ -0,0 +1,54 @@ +{%- if tools %} + {{- '<|im_start|>system\n' }} + {%- if messages[0]['role'] == 'system' %} + {{- messages[0]['content'] }} + {%- else %} + {{- 'You are Qwen, created by Alibaba Cloud. You are a helpful assistant.' }} + {%- endif %} + {{- "\n\n# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }} + {%- for tool in tools %} + {{- "\n" }} + {{- tool | tojson }} + {%- endfor %} + {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }} +{%- else %} + {%- if messages[0]['role'] == 'system' %} + {{- '<|im_start|>system\n' + messages[0]['content'] + '<|im_end|>\n' }} + {%- else %} + {{- '<|im_start|>system\nYou are Qwen, created by Alibaba Cloud. You are a helpful assistant.<|im_end|>\n' }} + {%- endif %} +{%- endif %} +{%- for message in messages %} + {%- if (message.role == "user") or (message.role == "system" and not loop.first) or (message.role == "assistant" and not message.tool_calls) %} + {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }} + {%- elif message.role == "assistant" %} + {{- '<|im_start|>' + message.role }} + {%- if message.content %} + {{- '\n' + message.content }} + {%- endif %} + {%- for tool_call in message.tool_calls %} + {%- if tool_call.function is defined %} + {%- set tool_call = tool_call.function %} + {%- endif %} + {{- '\n\n{"name": "' }} + {{- tool_call.name }} + {{- '", "arguments": ' }} + {{- tool_call.arguments | tojson }} + {{- '}\n' }} + {%- endfor %} + {{- '<|im_end|>\n' }} + {%- elif message.role == "tool" %} + {%- if (loop.index0 == 0) or (messages[loop.index0 - 1].role != "tool") %} + {{- '<|im_start|>user' }} + {%- endif %} + {{- '\n\n' }} + {{- message.content }} + {{- '\n' }} + {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %} + {{- '<|im_end|>\n' }} + {%- endif %} + {%- endif %} +{%- endfor %} +{%- if add_generation_prompt %} + {{- '<|im_start|>assistant\n' }} +{%- endif %} diff --git a/checkpoint-200/config.json b/checkpoint-200/config.json new file mode 100644 index 0000000..f57b09b --- /dev/null +++ b/checkpoint-200/config.json @@ -0,0 +1,61 @@ +{ + "architectures": [ + "Qwen2ForCausalLM" + ], + "attention_dropout": 0.0, + "bos_token_id": null, + "dtype": "float32", + "eos_token_id": 151645, + "hidden_act": "silu", + "hidden_size": 1536, + "initializer_range": 0.02, + "intermediate_size": 8960, + "layer_types": [ + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention" + ], + "max_position_embeddings": 32768, + "max_window_layers": 21, + "model_type": "qwen2", + "num_attention_heads": 12, + "num_hidden_layers": 28, + "num_key_value_heads": 2, + "pad_token_id": 151643, + "rms_norm_eps": 1e-06, + "rope_parameters": { + "rope_theta": 1000000.0, + "rope_type": "default" + }, + "sliding_window": null, + "tie_word_embeddings": true, + "transformers_version": "5.6.2", + "use_cache": false, + "use_sliding_window": false, + "vocab_size": 151936 +} diff --git a/checkpoint-200/generation_config.json b/checkpoint-200/generation_config.json new file mode 100644 index 0000000..1b2bba9 --- /dev/null +++ b/checkpoint-200/generation_config.json @@ -0,0 +1,13 @@ +{ + "do_sample": true, + "eos_token_id": [ + 151645, + 151643 + ], + "pad_token_id": 151643, + "repetition_penalty": 1.1, + "temperature": 0.7, + "top_k": 20, + "top_p": 0.8, + "transformers_version": "5.6.2" +} diff --git a/checkpoint-200/model.safetensors b/checkpoint-200/model.safetensors new file mode 100644 index 0000000..73c5aaa --- /dev/null +++ b/checkpoint-200/model.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:d768f5bced7c176c364bd8caa06f355a13ba8e84bfa6afc5000b0943d729a1b0 +size 6174895536 diff --git a/checkpoint-200/optimizer.pt b/checkpoint-200/optimizer.pt new file mode 100644 index 0000000..a9aa0ca --- /dev/null +++ b/checkpoint-200/optimizer.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:5ecab9b19baa7f145e68200d06e20d1eabf0b812395b056a997f9ad356d423e9 +size 12350013801 diff --git a/checkpoint-200/rng_state.pth b/checkpoint-200/rng_state.pth new file mode 100644 index 0000000..bc226e1 --- /dev/null +++ b/checkpoint-200/rng_state.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:bb4f2fcdc9d655fb6008b359f9ac492f3657acb42eb854239f64bcbd2e7ef3c8 +size 14645 diff --git a/checkpoint-200/scheduler.pt b/checkpoint-200/scheduler.pt new file mode 100644 index 0000000..f468c57 --- /dev/null +++ b/checkpoint-200/scheduler.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:62627619e905cc6929ad64a59b955d88cc2414468a59bffeba792e0b2b5af7d8 +size 1465 diff --git a/checkpoint-200/tokenizer.json b/checkpoint-200/tokenizer.json new file mode 100644 index 0000000..34510ff --- /dev/null +++ b/checkpoint-200/tokenizer.json @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:3fd169731d2cbde95e10bf356d66d5997fd885dd8dbb6fb4684da3f23b2585d8 +size 11421892 diff --git a/checkpoint-200/tokenizer_config.json b/checkpoint-200/tokenizer_config.json new file mode 100644 index 0000000..770e41d --- /dev/null +++ b/checkpoint-200/tokenizer_config.json @@ -0,0 +1,30 @@ +{ + "add_prefix_space": false, + "backend": "tokenizers", + "bos_token": null, + "clean_up_tokenization_spaces": false, + "eos_token": "<|im_end|>", + "errors": "replace", + "extra_special_tokens": [ + "<|im_start|>", + "<|im_end|>", + "<|object_ref_start|>", + "<|object_ref_end|>", + "<|box_start|>", + "<|box_end|>", + "<|quad_start|>", + "<|quad_end|>", + "<|vision_start|>", + "<|vision_end|>", + "<|vision_pad|>", + "<|image_pad|>", + "<|video_pad|>" + ], + "is_local": false, + "local_files_only": false, + "model_max_length": 131072, + "pad_token": "<|endoftext|>", + "split_special_tokens": false, + "tokenizer_class": "Qwen2Tokenizer", + "unk_token": null +} diff --git a/checkpoint-200/trainer_state.json b/checkpoint-200/trainer_state.json new file mode 100644 index 0000000..ef8c921 --- /dev/null +++ b/checkpoint-200/trainer_state.json @@ -0,0 +1,5834 @@ +{ + "best_global_step": null, + "best_metric": null, + "best_model_checkpoint": null, + "epoch": 1.0, + "eval_steps": 500, + "global_step": 200, + "is_hyper_param_search": false, + "is_local_process_zero": true, + "is_world_process_zero": true, + "log_history": [ + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 478.0, + "completions/max_terminated_length": 478.0, + "completions/mean_length": 191.0, + "completions/mean_terminated_length": 191.0, + "completions/min_length": 66.0, + "completions/min_terminated_length": 66.0, + "entropy": 0.5566893219947815, + "epoch": 0.005, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.370361328125, + "learning_rate": 1e-06, + "loss": -0.24658073484897614, + "num_tokens": 10832.0, + "reward": -3.3000030517578125, + "reward_std": 85.62333679199219, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -3.4000015258789062, + "rewards/supergames_reward/std": 85.62333679199219, + "step": 1, + "step_time": 22.779711072013015 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 216.0, + "completions/max_terminated_length": 216.0, + "completions/mean_length": 131.5, + "completions/mean_terminated_length": 131.5, + "completions/min_length": 60.0, + "completions/min_terminated_length": 60.0, + "entropy": 0.6395841240882874, + "epoch": 0.01, + "frac_reward_zero_std": 0.0, + "grad_norm": 8.09774398803711, + "learning_rate": 9.95e-07, + "loss": 0.20567649602890015, + "num_tokens": 16404.0, + "reward": -12.422499656677246, + "reward_std": 7.134707450866699, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -12.522500038146973, + "rewards/supergames_reward/std": 7.134707927703857, + "step": 2, + "step_time": 9.960156448010821 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 287.0, + "completions/max_terminated_length": 287.0, + "completions/mean_length": 156.625, + "completions/mean_terminated_length": 156.625, + "completions/min_length": 77.0, + "completions/min_terminated_length": 77.0, + "entropy": 0.562222421169281, + "epoch": 0.015, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.9053616523742676, + "learning_rate": 9.9e-07, + "loss": 0.09602774679660797, + "num_tokens": 26953.0, + "reward": -74.94999694824219, + "reward_std": 70.73094177246094, + "rewards/json_format_reward/mean": 0.05000000074505806, + "rewards/json_format_reward/std": 0.09258200973272324, + "rewards/supergames_reward/mean": -75.0, + "rewards/supergames_reward/std": 70.71067810058594, + "step": 3, + "step_time": 13.686320498993155 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 268.0, + "completions/max_terminated_length": 268.0, + "completions/mean_length": 154.125, + "completions/mean_terminated_length": 154.125, + "completions/min_length": 53.0, + "completions/min_terminated_length": 53.0, + "entropy": 0.7323317527770996, + "epoch": 0.02, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.933310031890869, + "learning_rate": 9.849999999999999e-07, + "loss": 0.28249427676200867, + "num_tokens": 36514.0, + "reward": -40.7599983215332, + "reward_std": 81.73140716552734, + "rewards/json_format_reward/mean": 0.05000000074505806, + "rewards/json_format_reward/std": 0.09258200973272324, + "rewards/supergames_reward/mean": -40.80999755859375, + "rewards/supergames_reward/std": 81.68995666503906, + "step": 4, + "step_time": 12.805880262021674 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 401.0, + "completions/max_terminated_length": 401.0, + "completions/mean_length": 205.25, + "completions/mean_terminated_length": 205.25, + "completions/min_length": 115.0, + "completions/min_terminated_length": 115.0, + "entropy": 0.6167430281639099, + "epoch": 0.025, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.344135046005249, + "learning_rate": 9.8e-07, + "loss": 0.1433994024991989, + "num_tokens": 47476.0, + "reward": -97.6612548828125, + "reward_std": 6.3321428298950195, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -97.76124572753906, + "rewards/supergames_reward/std": 6.3321428298950195, + "step": 5, + "step_time": 18.789364666008623 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 276.0, + "completions/max_terminated_length": 276.0, + "completions/mean_length": 167.25, + "completions/mean_terminated_length": 167.25, + "completions/min_length": 73.0, + "completions/min_terminated_length": 73.0, + "entropy": 0.7583790421485901, + "epoch": 0.03, + "frac_reward_zero_std": 0.0, + "grad_norm": 5.172371864318848, + "learning_rate": 9.75e-07, + "loss": -0.12127404659986496, + "num_tokens": 55246.0, + "reward": -58.687503814697266, + "reward_std": 58.64059829711914, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": -58.76250457763672, + "rewards/supergames_reward/std": 58.620460510253906, + "step": 6, + "step_time": 12.948570273991209 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 261.0, + "completions/max_terminated_length": 261.0, + "completions/mean_length": 133.875, + "completions/mean_terminated_length": 133.875, + "completions/min_length": 57.0, + "completions/min_terminated_length": 57.0, + "entropy": 0.7447654604911804, + "epoch": 0.035, + "frac_reward_zero_std": 0.0, + "grad_norm": 6.428595066070557, + "learning_rate": 9.7e-07, + "loss": 0.10135584324598312, + "num_tokens": 60837.0, + "reward": -32.45000076293945, + "reward_std": 41.754486083984375, + "rewards/json_format_reward/mean": 0.05000000074505806, + "rewards/json_format_reward/std": 0.09258200973272324, + "rewards/supergames_reward/mean": -32.5, + "rewards/supergames_reward/std": 41.66190719604492, + "step": 7, + "step_time": 12.058315072004916 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 414.0, + "completions/max_terminated_length": 414.0, + "completions/mean_length": 209.125, + "completions/mean_terminated_length": 209.125, + "completions/min_length": 104.0, + "completions/min_terminated_length": 104.0, + "entropy": 0.3860666751861572, + "epoch": 0.04, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.806255578994751, + "learning_rate": 9.649999999999999e-07, + "loss": -0.33238139748573303, + "num_tokens": 70838.0, + "reward": -80.50375366210938, + "reward_std": 38.82380294799805, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -80.60375213623047, + "rewards/supergames_reward/std": 38.82379913330078, + "step": 8, + "step_time": 19.323370319994865 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.25, + "completions/max_length": 512.0, + "completions/max_terminated_length": 199.0, + "completions/mean_length": 247.125, + "completions/mean_terminated_length": 158.83334350585938, + "completions/min_length": 106.0, + "completions/min_terminated_length": 106.0, + "entropy": 0.5816237330436707, + "epoch": 0.045, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.3209943771362305, + "learning_rate": 9.6e-07, + "loss": 0.36752766370773315, + "num_tokens": 82135.0, + "reward": -27.688751220703125, + "reward_std": 82.8358154296875, + "rewards/json_format_reward/mean": 0.05000000074505806, + "rewards/json_format_reward/std": 0.09258200973272324, + "rewards/supergames_reward/mean": -27.738750457763672, + "rewards/supergames_reward/std": 82.78590393066406, + "step": 9, + "step_time": 21.948575104994234 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 240.0, + "completions/max_terminated_length": 240.0, + "completions/mean_length": 153.0, + "completions/mean_terminated_length": 153.0, + "completions/min_length": 51.0, + "completions/min_terminated_length": 51.0, + "entropy": 0.6301521062850952, + "epoch": 0.05, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.125359058380127, + "learning_rate": 9.55e-07, + "loss": 0.06129350885748863, + "num_tokens": 89807.0, + "reward": -78.39624786376953, + "reward_std": 40.17042922973633, + "rewards/json_format_reward/mean": 0.05000000074505806, + "rewards/json_format_reward/std": 0.09258200973272324, + "rewards/supergames_reward/mean": -78.44625091552734, + "rewards/supergames_reward/std": 40.1396484375, + "step": 10, + "step_time": 9.098202993016457 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 133.0, + "completions/max_terminated_length": 133.0, + "completions/mean_length": 112.0, + "completions/mean_terminated_length": 112.0, + "completions/min_length": 68.0, + "completions/min_terminated_length": 68.0, + "entropy": 0.5521990656852722, + "epoch": 0.055, + "frac_reward_zero_std": 0.0, + "grad_norm": 6.616060256958008, + "learning_rate": 9.499999999999999e-07, + "loss": -0.006313305348157883, + "num_tokens": 95231.0, + "reward": -12.161249160766602, + "reward_std": 6.39578104019165, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -12.261249542236328, + "rewards/supergames_reward/std": 6.39578104019165, + "step": 11, + "step_time": 5.136311663984088 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 396.0, + "completions/max_terminated_length": 396.0, + "completions/mean_length": 214.875, + "completions/mean_terminated_length": 214.875, + "completions/min_length": 55.0, + "completions/min_terminated_length": 55.0, + "entropy": 0.5505905151367188, + "epoch": 0.06, + "frac_reward_zero_std": 0.0, + "grad_norm": 5.289583206176758, + "learning_rate": 9.45e-07, + "loss": -0.29802003502845764, + "num_tokens": 106238.0, + "reward": -91.66999816894531, + "reward_std": 23.2779541015625, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -91.77000427246094, + "rewards/supergames_reward/std": 23.277956008911133, + "step": 12, + "step_time": 14.945365622988902 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 269.0, + "completions/max_terminated_length": 269.0, + "completions/mean_length": 167.375, + "completions/mean_terminated_length": 167.375, + "completions/min_length": 113.0, + "completions/min_terminated_length": 113.0, + "entropy": 0.540415346622467, + "epoch": 0.065, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.858291149139404, + "learning_rate": 9.399999999999999e-07, + "loss": 0.24004128575325012, + "num_tokens": 116937.0, + "reward": -20.38249969482422, + "reward_std": 87.55204010009766, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": -20.457500457763672, + "rewards/supergames_reward/std": 87.52605438232422, + "step": 13, + "step_time": 10.482285185018554 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 254.0, + "completions/max_terminated_length": 254.0, + "completions/mean_length": 156.875, + "completions/mean_terminated_length": 156.875, + "completions/min_length": 114.0, + "completions/min_terminated_length": 114.0, + "entropy": 0.7237679958343506, + "epoch": 0.07, + "frac_reward_zero_std": 0.0, + "grad_norm": 5.083980560302734, + "learning_rate": 9.35e-07, + "loss": -0.19025824964046478, + "num_tokens": 124640.0, + "reward": -51.397499084472656, + "reward_std": 54.486454010009766, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -51.49749755859375, + "rewards/supergames_reward/std": 54.486454010009766, + "step": 14, + "step_time": 9.597231683001155 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 236.0, + "completions/max_terminated_length": 236.0, + "completions/mean_length": 141.375, + "completions/mean_terminated_length": 141.375, + "completions/min_length": 91.0, + "completions/min_terminated_length": 91.0, + "entropy": 0.754334568977356, + "epoch": 0.075, + "frac_reward_zero_std": 0.0, + "grad_norm": 5.720428943634033, + "learning_rate": 9.3e-07, + "loss": -0.16830675303936005, + "num_tokens": 132219.0, + "reward": -56.226253509521484, + "reward_std": 60.115657806396484, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -56.32625198364258, + "rewards/supergames_reward/std": 60.11566162109375, + "step": 15, + "step_time": 8.998362872982398 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 209.0, + "completions/max_terminated_length": 209.0, + "completions/mean_length": 136.375, + "completions/mean_terminated_length": 136.375, + "completions/min_length": 97.0, + "completions/min_terminated_length": 97.0, + "entropy": 0.6280568838119507, + "epoch": 0.08, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.9681806564331055, + "learning_rate": 9.25e-07, + "loss": -0.01910916529595852, + "num_tokens": 139726.0, + "reward": -68.24000549316406, + "reward_std": 52.238914489746094, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -68.33999633789062, + "rewards/supergames_reward/std": 52.238914489746094, + "step": 16, + "step_time": 8.034480294008972 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 365.0, + "completions/max_terminated_length": 365.0, + "completions/mean_length": 211.875, + "completions/mean_terminated_length": 211.875, + "completions/min_length": 100.0, + "completions/min_terminated_length": 100.0, + "entropy": 0.899176836013794, + "epoch": 0.085, + "frac_reward_zero_std": 0.0, + "grad_norm": 6.691746711730957, + "learning_rate": 9.2e-07, + "loss": 0.25513291358947754, + "num_tokens": 147869.0, + "reward": -99.92500305175781, + "reward_std": 0.07070883363485336, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": -100.0, + "rewards/supergames_reward/std": 0.0, + "step": 17, + "step_time": 13.655792869016295 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 387.0, + "completions/max_terminated_length": 387.0, + "completions/mean_length": 238.0, + "completions/mean_terminated_length": 238.0, + "completions/min_length": 87.0, + "completions/min_terminated_length": 87.0, + "entropy": 0.5596873760223389, + "epoch": 0.09, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.6711347103118896, + "learning_rate": 9.15e-07, + "loss": -0.2628335654735565, + "num_tokens": 159101.0, + "reward": -24.900001525878906, + "reward_std": 103.50983428955078, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -25.0, + "rewards/supergames_reward/std": 103.50983428955078, + "step": 18, + "step_time": 14.784329420013819 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.125, + "completions/max_length": 512.0, + "completions/max_terminated_length": 277.0, + "completions/mean_length": 219.75, + "completions/mean_terminated_length": 178.00001525878906, + "completions/min_length": 90.0, + "completions/min_terminated_length": 90.0, + "entropy": 0.9245517253875732, + "epoch": 0.095, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.55502986907959, + "learning_rate": 9.1e-07, + "loss": 0.13060179352760315, + "num_tokens": 167299.0, + "reward": -86.89125061035156, + "reward_std": 36.95603561401367, + "rewards/json_format_reward/mean": 0.05000000074505806, + "rewards/json_format_reward/std": 0.09258200973272324, + "rewards/supergames_reward/mean": -86.94125366210938, + "rewards/supergames_reward/std": 36.93572235107422, + "step": 19, + "step_time": 18.832684700988466 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 267.0, + "completions/max_terminated_length": 267.0, + "completions/mean_length": 177.25, + "completions/mean_terminated_length": 177.25, + "completions/min_length": 93.0, + "completions/min_terminated_length": 93.0, + "entropy": 0.7331254482269287, + "epoch": 0.1, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.538497447967529, + "learning_rate": 9.05e-07, + "loss": -0.16418980062007904, + "num_tokens": 175141.0, + "reward": -68.52874755859375, + "reward_std": 58.48719787597656, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -68.62875366210938, + "rewards/supergames_reward/std": 58.48720169067383, + "step": 20, + "step_time": 10.090975169994636 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 300.0, + "completions/max_terminated_length": 300.0, + "completions/mean_length": 190.625, + "completions/mean_terminated_length": 190.625, + "completions/min_length": 103.0, + "completions/min_terminated_length": 103.0, + "entropy": 0.6659425497055054, + "epoch": 0.105, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.485373497009277, + "learning_rate": 9e-07, + "loss": -0.10106432437896729, + "num_tokens": 183098.0, + "reward": -71.48875427246094, + "reward_std": 55.33565139770508, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -71.5887451171875, + "rewards/supergames_reward/std": 55.33565139770508, + "step": 21, + "step_time": 11.255001295008697 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 279.0, + "completions/max_terminated_length": 279.0, + "completions/mean_length": 179.0, + "completions/mean_terminated_length": 179.0, + "completions/min_length": 103.0, + "completions/min_terminated_length": 103.0, + "entropy": 0.49365949630737305, + "epoch": 0.11, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.829419136047363, + "learning_rate": 8.95e-07, + "loss": 0.10587231814861298, + "num_tokens": 192826.0, + "reward": -75.62000274658203, + "reward_std": 45.01959991455078, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": -75.69499969482422, + "rewards/supergames_reward/std": 45.00411605834961, + "step": 22, + "step_time": 10.611246599000879 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 315.0, + "completions/max_terminated_length": 315.0, + "completions/mean_length": 212.625, + "completions/mean_terminated_length": 212.625, + "completions/min_length": 100.0, + "completions/min_terminated_length": 100.0, + "entropy": 0.43159469962120056, + "epoch": 0.115, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.541125774383545, + "learning_rate": 8.9e-07, + "loss": -0.1681259125471115, + "num_tokens": 202815.0, + "reward": 23.67624855041504, + "reward_std": 67.74340057373047, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 23.576250076293945, + "rewards/supergames_reward/std": 67.743408203125, + "step": 23, + "step_time": 11.731771531980485 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 168.0, + "completions/max_terminated_length": 168.0, + "completions/mean_length": 101.625, + "completions/mean_terminated_length": 101.625, + "completions/min_length": 65.0, + "completions/min_terminated_length": 65.0, + "entropy": 0.3972298502922058, + "epoch": 0.12, + "frac_reward_zero_std": 0.0, + "grad_norm": 5.8285417556762695, + "learning_rate": 8.85e-07, + "loss": -0.011585958302021027, + "num_tokens": 208188.0, + "reward": 58.849998474121094, + "reward_std": 56.9304084777832, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 58.75, + "rewards/supergames_reward/std": 56.9304084777832, + "step": 24, + "step_time": 6.187504950998118 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 303.0, + "completions/max_terminated_length": 303.0, + "completions/mean_length": 165.375, + "completions/mean_terminated_length": 165.375, + "completions/min_length": 69.0, + "completions/min_terminated_length": 69.0, + "entropy": 0.6339899301528931, + "epoch": 0.125, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.823799133300781, + "learning_rate": 8.799999999999999e-07, + "loss": 0.05159185826778412, + "num_tokens": 215935.0, + "reward": -23.521251678466797, + "reward_std": 74.23067474365234, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -23.621248245239258, + "rewards/supergames_reward/std": 74.23067474365234, + "step": 25, + "step_time": 11.059416300005978 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 349.0, + "completions/max_terminated_length": 349.0, + "completions/mean_length": 204.0, + "completions/mean_terminated_length": 204.0, + "completions/min_length": 143.0, + "completions/min_terminated_length": 143.0, + "entropy": 0.6462895274162292, + "epoch": 0.13, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.969292640686035, + "learning_rate": 8.75e-07, + "loss": 0.0115677984431386, + "num_tokens": 223999.0, + "reward": -74.8175048828125, + "reward_std": 40.94221115112305, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": -74.89250183105469, + "rewards/supergames_reward/std": 40.924625396728516, + "step": 26, + "step_time": 12.666237785975682 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 299.0, + "completions/max_terminated_length": 299.0, + "completions/mean_length": 166.5, + "completions/mean_terminated_length": 166.5, + "completions/min_length": 87.0, + "completions/min_terminated_length": 87.0, + "entropy": 0.4756850600242615, + "epoch": 0.135, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.915069103240967, + "learning_rate": 8.699999999999999e-07, + "loss": -0.13321346044540405, + "num_tokens": 233643.0, + "reward": -52.125, + "reward_std": 19.405282974243164, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": -52.20000076293945, + "rewards/supergames_reward/std": 19.33464813232422, + "step": 27, + "step_time": 11.172539752995363 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 184.0, + "completions/max_terminated_length": 184.0, + "completions/mean_length": 137.125, + "completions/mean_terminated_length": 137.125, + "completions/min_length": 100.0, + "completions/min_terminated_length": 100.0, + "entropy": 0.5834778547286987, + "epoch": 0.14, + "frac_reward_zero_std": 0.0, + "grad_norm": 5.146666526794434, + "learning_rate": 8.65e-07, + "loss": 0.060454584658145905, + "num_tokens": 241204.0, + "reward": -68.99250793457031, + "reward_std": 51.010581970214844, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -69.09249877929688, + "rewards/supergames_reward/std": 51.010581970214844, + "step": 28, + "step_time": 7.025046669004951 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 331.0, + "completions/max_terminated_length": 331.0, + "completions/mean_length": 182.375, + "completions/mean_terminated_length": 182.375, + "completions/min_length": 106.0, + "completions/min_terminated_length": 106.0, + "entropy": 0.6368551254272461, + "epoch": 0.145, + "frac_reward_zero_std": 0.0, + "grad_norm": 5.15535306930542, + "learning_rate": 8.599999999999999e-07, + "loss": -0.3323673903942108, + "num_tokens": 249087.0, + "reward": -44.17000198364258, + "reward_std": 80.5373306274414, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -44.27000045776367, + "rewards/supergames_reward/std": 80.5373306274414, + "step": 29, + "step_time": 12.079259724996518 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 190.0, + "completions/max_terminated_length": 190.0, + "completions/mean_length": 127.75, + "completions/mean_terminated_length": 127.75, + "completions/min_length": 80.0, + "completions/min_terminated_length": 80.0, + "entropy": 0.4491196870803833, + "epoch": 0.15, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.896186828613281, + "learning_rate": 8.55e-07, + "loss": -0.12887656688690186, + "num_tokens": 259445.0, + "reward": 57.446250915527344, + "reward_std": 38.78837585449219, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 57.34625244140625, + "rewards/supergames_reward/std": 38.78837966918945, + "step": 30, + "step_time": 7.483972909016302 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 166.0, + "completions/max_terminated_length": 166.0, + "completions/mean_length": 112.875, + "completions/mean_terminated_length": 112.875, + "completions/min_length": 85.0, + "completions/min_terminated_length": 85.0, + "entropy": 0.4858454465866089, + "epoch": 0.155, + "frac_reward_zero_std": 0.0, + "grad_norm": 6.742906093597412, + "learning_rate": 8.499999999999999e-07, + "loss": 0.06930096447467804, + "num_tokens": 264924.0, + "reward": -21.174999237060547, + "reward_std": 31.89051628112793, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": -21.25, + "rewards/supergames_reward/std": 31.819807052612305, + "step": 31, + "step_time": 6.1866529019898735 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 148.0, + "completions/max_terminated_length": 148.0, + "completions/mean_length": 118.125, + "completions/mean_terminated_length": 118.125, + "completions/min_length": 98.0, + "completions/min_terminated_length": 98.0, + "entropy": 0.45428749918937683, + "epoch": 0.16, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.663414478302002, + "learning_rate": 8.45e-07, + "loss": 0.05735419690608978, + "num_tokens": 270381.0, + "reward": 0.4687504768371582, + "reward_std": 38.66371154785156, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 0.36875057220458984, + "rewards/supergames_reward/std": 38.6637077331543, + "step": 32, + "step_time": 5.577042757999152 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 345.0, + "completions/max_terminated_length": 345.0, + "completions/mean_length": 166.875, + "completions/mean_terminated_length": 166.875, + "completions/min_length": 86.0, + "completions/min_terminated_length": 86.0, + "entropy": 0.6107293367385864, + "epoch": 0.165, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.75006628036499, + "learning_rate": 8.399999999999999e-07, + "loss": -0.14760127663612366, + "num_tokens": 278156.0, + "reward": -87.84750366210938, + "reward_std": 26.84708023071289, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -87.94749450683594, + "rewards/supergames_reward/std": 26.847076416015625, + "step": 33, + "step_time": 12.552876825997373 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.125, + "completions/max_length": 512.0, + "completions/max_terminated_length": 360.0, + "completions/mean_length": 231.0, + "completions/mean_terminated_length": 190.85714721679688, + "completions/min_length": 74.0, + "completions/min_terminated_length": 74.0, + "entropy": 0.6089653372764587, + "epoch": 0.17, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.781267166137695, + "learning_rate": 8.349999999999999e-07, + "loss": 0.4294504225254059, + "num_tokens": 289364.0, + "reward": -99.92500305175781, + "reward_std": 0.07070883363485336, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": -100.0, + "rewards/supergames_reward/std": 0.0, + "step": 34, + "step_time": 18.961819477990502 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 218.0, + "completions/max_terminated_length": 218.0, + "completions/mean_length": 134.125, + "completions/mean_terminated_length": 134.125, + "completions/min_length": 104.0, + "completions/min_terminated_length": 104.0, + "entropy": 0.46158918738365173, + "epoch": 0.175, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.016353607177734, + "learning_rate": 8.299999999999999e-07, + "loss": 0.09251818805932999, + "num_tokens": 298733.0, + "reward": -35.558753967285156, + "reward_std": 15.881204605102539, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -35.65875244140625, + "rewards/supergames_reward/std": 15.881203651428223, + "step": 35, + "step_time": 8.38491243700264 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 207.0, + "completions/max_terminated_length": 207.0, + "completions/mean_length": 156.875, + "completions/mean_terminated_length": 156.875, + "completions/min_length": 111.0, + "completions/min_terminated_length": 111.0, + "entropy": 0.5104788541793823, + "epoch": 0.18, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.592893123626709, + "learning_rate": 8.249999999999999e-07, + "loss": -0.0410832017660141, + "num_tokens": 306420.0, + "reward": 5.4212493896484375, + "reward_std": 93.50404357910156, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": 5.346250534057617, + "rewards/supergames_reward/std": 93.4718246459961, + "step": 36, + "step_time": 7.811868985998444 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 241.0, + "completions/max_terminated_length": 241.0, + "completions/mean_length": 178.0, + "completions/mean_terminated_length": 178.0, + "completions/min_length": 106.0, + "completions/min_terminated_length": 106.0, + "entropy": 0.7527878880500793, + "epoch": 0.185, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.8812546730041504, + "learning_rate": 8.199999999999999e-07, + "loss": 0.12539231777191162, + "num_tokens": 314252.0, + "reward": -89.58500671386719, + "reward_std": 29.256120681762695, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": -89.66000366210938, + "rewards/supergames_reward/std": 29.24593734741211, + "step": 37, + "step_time": 8.959742914012168 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 199.0, + "completions/max_terminated_length": 199.0, + "completions/mean_length": 133.375, + "completions/mean_terminated_length": 133.375, + "completions/min_length": 87.0, + "completions/min_terminated_length": 87.0, + "entropy": 0.3347901403903961, + "epoch": 0.19, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.205414295196533, + "learning_rate": 8.149999999999999e-07, + "loss": -0.06991486996412277, + "num_tokens": 319839.0, + "reward": -14.498749732971191, + "reward_std": 13.007229804992676, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -14.598750114440918, + "rewards/supergames_reward/std": 13.007229804992676, + "step": 38, + "step_time": 7.377183554985095 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 181.0, + "completions/max_terminated_length": 181.0, + "completions/mean_length": 114.375, + "completions/mean_terminated_length": 114.375, + "completions/min_length": 58.0, + "completions/min_terminated_length": 58.0, + "entropy": 0.4445965886116028, + "epoch": 0.195, + "frac_reward_zero_std": 0.0, + "grad_norm": 5.561298847198486, + "learning_rate": 8.1e-07, + "loss": -0.06235164776444435, + "num_tokens": 325290.0, + "reward": -7.425000190734863, + "reward_std": 53.701602935791016, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": -7.5, + "rewards/supergames_reward/std": 53.652320861816406, + "step": 39, + "step_time": 6.753862089011818 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 370.0, + "completions/max_terminated_length": 370.0, + "completions/mean_length": 187.625, + "completions/mean_terminated_length": 187.625, + "completions/min_length": 60.0, + "completions/min_terminated_length": 60.0, + "entropy": 0.3950234353542328, + "epoch": 0.2, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.5988712310791016, + "learning_rate": 8.05e-07, + "loss": -0.1381440907716751, + "num_tokens": 335119.0, + "reward": -33.68375015258789, + "reward_std": 42.069435119628906, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -33.783748626708984, + "rewards/supergames_reward/std": 42.069435119628906, + "step": 40, + "step_time": 13.718958162004128 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.125, + "completions/max_length": 512.0, + "completions/max_terminated_length": 371.0, + "completions/mean_length": 251.125, + "completions/mean_terminated_length": 213.85714721679688, + "completions/min_length": 121.0, + "completions/min_terminated_length": 121.0, + "entropy": 0.507067084312439, + "epoch": 0.205, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.3898658752441406, + "learning_rate": 8e-07, + "loss": 0.0078964838758111, + "num_tokens": 346464.0, + "reward": -86.48625183105469, + "reward_std": 38.020694732666016, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": -86.56124877929688, + "rewards/supergames_reward/std": 38.010528564453125, + "step": 41, + "step_time": 18.90960379401804 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 152.0, + "completions/max_terminated_length": 152.0, + "completions/mean_length": 116.375, + "completions/mean_terminated_length": 116.375, + "completions/min_length": 93.0, + "completions/min_terminated_length": 93.0, + "entropy": 0.3801707923412323, + "epoch": 0.21, + "frac_reward_zero_std": 1.0, + "grad_norm": 0.0, + "learning_rate": 7.95e-07, + "loss": 0.0, + "num_tokens": 351955.0, + "reward": -9.899999618530273, + "reward_std": 0.0, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -10.0, + "rewards/supergames_reward/std": 0.0, + "step": 42, + "step_time": 5.711630532023264 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 308.0, + "completions/max_terminated_length": 308.0, + "completions/mean_length": 247.375, + "completions/mean_terminated_length": 247.375, + "completions/min_length": 116.0, + "completions/min_terminated_length": 116.0, + "entropy": 0.3185043931007385, + "epoch": 0.215, + "frac_reward_zero_std": 0.0, + "grad_norm": 2.168762445449829, + "learning_rate": 7.9e-07, + "loss": 0.1970764398574829, + "num_tokens": 363278.0, + "reward": -53.89249801635742, + "reward_std": 63.02619934082031, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -53.99250030517578, + "rewards/supergames_reward/std": 63.02619934082031, + "step": 43, + "step_time": 11.661934906995157 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 330.0, + "completions/max_terminated_length": 330.0, + "completions/mean_length": 212.5, + "completions/mean_terminated_length": 212.5, + "completions/min_length": 106.0, + "completions/min_terminated_length": 106.0, + "entropy": 0.4446268379688263, + "epoch": 0.22, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.305844306945801, + "learning_rate": 7.85e-07, + "loss": -0.06735197454690933, + "num_tokens": 374306.0, + "reward": 18.059999465942383, + "reward_std": 98.9417953491211, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 17.959999084472656, + "rewards/supergames_reward/std": 98.9417953491211, + "step": 44, + "step_time": 12.492453911982011 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 316.0, + "completions/max_terminated_length": 316.0, + "completions/mean_length": 183.0, + "completions/mean_terminated_length": 183.0, + "completions/min_length": 103.0, + "completions/min_terminated_length": 103.0, + "entropy": 0.5269614458084106, + "epoch": 0.225, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.8006370067596436, + "learning_rate": 7.799999999999999e-07, + "loss": -0.08870815485715866, + "num_tokens": 382210.0, + "reward": -54.928749084472656, + "reward_std": 72.5127944946289, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -55.02874755859375, + "rewards/supergames_reward/std": 72.5127944946289, + "step": 45, + "step_time": 11.7636467939883 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 244.0, + "completions/max_terminated_length": 244.0, + "completions/mean_length": 136.625, + "completions/mean_terminated_length": 136.625, + "completions/min_length": 85.0, + "completions/min_terminated_length": 85.0, + "entropy": 0.34820204973220825, + "epoch": 0.23, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.911736011505127, + "learning_rate": 7.75e-07, + "loss": 0.04397330805659294, + "num_tokens": 391583.0, + "reward": -19.10375213623047, + "reward_std": 23.967702865600586, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -19.203750610351562, + "rewards/supergames_reward/std": 23.967702865600586, + "step": 46, + "step_time": 10.119426151999505 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 218.0, + "completions/max_terminated_length": 218.0, + "completions/mean_length": 160.375, + "completions/mean_terminated_length": 160.375, + "completions/min_length": 93.0, + "completions/min_terminated_length": 93.0, + "entropy": 0.611914336681366, + "epoch": 0.235, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.571913242340088, + "learning_rate": 7.699999999999999e-07, + "loss": -0.044293053448200226, + "num_tokens": 399290.0, + "reward": -25.186250686645508, + "reward_std": 71.46305084228516, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -25.286251068115234, + "rewards/supergames_reward/std": 71.46305084228516, + "step": 47, + "step_time": 8.563868903991533 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 167.0, + "completions/max_terminated_length": 167.0, + "completions/mean_length": 111.125, + "completions/mean_terminated_length": 111.125, + "completions/min_length": 56.0, + "completions/min_terminated_length": 56.0, + "entropy": 0.46585777401924133, + "epoch": 0.24, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.654109954833984, + "learning_rate": 7.65e-07, + "loss": -0.035151124000549316, + "num_tokens": 404707.0, + "reward": -0.408750057220459, + "reward_std": 31.939748764038086, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -0.5087499618530273, + "rewards/supergames_reward/std": 31.93975257873535, + "step": 48, + "step_time": 6.438482580007985 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 306.0, + "completions/max_terminated_length": 306.0, + "completions/mean_length": 216.625, + "completions/mean_terminated_length": 216.625, + "completions/min_length": 132.0, + "completions/min_terminated_length": 132.0, + "entropy": 0.33054471015930176, + "epoch": 0.245, + "frac_reward_zero_std": 0.0, + "grad_norm": 2.700303792953491, + "learning_rate": 7.599999999999999e-07, + "loss": -0.04393656551837921, + "num_tokens": 414736.0, + "reward": -97.26000213623047, + "reward_std": 1.6294406652450562, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -97.36000061035156, + "rewards/supergames_reward/std": 1.629441738128662, + "step": 49, + "step_time": 11.636040106008295 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 378.0, + "completions/max_terminated_length": 378.0, + "completions/mean_length": 218.375, + "completions/mean_terminated_length": 218.375, + "completions/min_length": 104.0, + "completions/min_terminated_length": 104.0, + "entropy": 0.48711177706718445, + "epoch": 0.25, + "frac_reward_zero_std": 0.0, + "grad_norm": 2.9650564193725586, + "learning_rate": 7.55e-07, + "loss": -0.3529520332813263, + "num_tokens": 424795.0, + "reward": 47.408748626708984, + "reward_std": 72.72083282470703, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 47.30875015258789, + "rewards/supergames_reward/std": 72.72083282470703, + "step": 50, + "step_time": 14.258096842997475 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.125, + "completions/max_length": 512.0, + "completions/max_terminated_length": 492.0, + "completions/mean_length": 267.625, + "completions/mean_terminated_length": 232.71429443359375, + "completions/min_length": 90.0, + "completions/min_terminated_length": 90.0, + "entropy": 0.4482860863208771, + "epoch": 0.255, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.2998197078704834, + "learning_rate": 7.5e-07, + "loss": 0.2493654191493988, + "num_tokens": 436288.0, + "reward": 50.61125183105469, + "reward_std": 75.3349838256836, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": 50.53624725341797, + "rewards/supergames_reward/std": 75.27783966064453, + "step": 51, + "step_time": 19.20827590499539 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 176.0, + "completions/max_terminated_length": 176.0, + "completions/mean_length": 144.5, + "completions/mean_terminated_length": 144.5, + "completions/min_length": 99.0, + "completions/min_terminated_length": 99.0, + "entropy": 0.5378735661506653, + "epoch": 0.26, + "frac_reward_zero_std": 0.0, + "grad_norm": 5.114461898803711, + "learning_rate": 7.45e-07, + "loss": 0.0036597400903701782, + "num_tokens": 441980.0, + "reward": -15.20625114440918, + "reward_std": 45.35526657104492, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": -15.28125, + "rewards/supergames_reward/std": 45.301815032958984, + "step": 52, + "step_time": 6.660627231001854 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 136.0, + "completions/max_terminated_length": 136.0, + "completions/mean_length": 119.0, + "completions/mean_terminated_length": 119.0, + "completions/min_length": 105.0, + "completions/min_terminated_length": 105.0, + "entropy": 0.41796669363975525, + "epoch": 0.265, + "frac_reward_zero_std": 0.0, + "grad_norm": 5.059505462646484, + "learning_rate": 7.4e-07, + "loss": -0.014973883517086506, + "num_tokens": 447468.0, + "reward": 15.170000076293945, + "reward_std": 46.658329010009766, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 15.070000648498535, + "rewards/supergames_reward/std": 46.658329010009766, + "step": 53, + "step_time": 5.271571868011961 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 259.0, + "completions/max_terminated_length": 259.0, + "completions/mean_length": 191.125, + "completions/mean_terminated_length": 191.125, + "completions/min_length": 91.0, + "completions/min_terminated_length": 91.0, + "entropy": 0.45421987771987915, + "epoch": 0.27, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.4454867839813232, + "learning_rate": 7.35e-07, + "loss": -0.1128973588347435, + "num_tokens": 457301.0, + "reward": 38.89875030517578, + "reward_std": 62.1772575378418, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 38.79874801635742, + "rewards/supergames_reward/std": 62.1772575378418, + "step": 54, + "step_time": 10.524528659996577 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 319.0, + "completions/max_terminated_length": 319.0, + "completions/mean_length": 233.375, + "completions/mean_terminated_length": 233.375, + "completions/min_length": 136.0, + "completions/min_terminated_length": 136.0, + "entropy": 0.4449213445186615, + "epoch": 0.275, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.2065398693084717, + "learning_rate": 7.3e-07, + "loss": 0.0016277075046673417, + "num_tokens": 467520.0, + "reward": 27.372499465942383, + "reward_std": 81.23126983642578, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 27.27250099182129, + "rewards/supergames_reward/std": 81.23126983642578, + "step": 55, + "step_time": 12.592280682991259 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.125, + "completions/max_length": 512.0, + "completions/max_terminated_length": 357.0, + "completions/mean_length": 261.0, + "completions/mean_terminated_length": 225.1428680419922, + "completions/min_length": 95.0, + "completions/min_terminated_length": 95.0, + "entropy": 0.506920337677002, + "epoch": 0.28, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.058388710021973, + "learning_rate": 7.249999999999999e-07, + "loss": 0.3289242386817932, + "num_tokens": 476056.0, + "reward": 47.05249786376953, + "reward_std": 80.97753143310547, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": 46.977500915527344, + "rewards/supergames_reward/std": 80.9256362915039, + "step": 56, + "step_time": 18.845177220006008 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 250.0, + "completions/max_terminated_length": 250.0, + "completions/mean_length": 192.75, + "completions/mean_terminated_length": 192.75, + "completions/min_length": 122.0, + "completions/min_terminated_length": 122.0, + "entropy": 0.5197336077690125, + "epoch": 0.285, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.230815410614014, + "learning_rate": 7.2e-07, + "loss": -0.15186546742916107, + "num_tokens": 484014.0, + "reward": -64.5050048828125, + "reward_std": 50.35145568847656, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -64.60499572753906, + "rewards/supergames_reward/std": 50.3514518737793, + "step": 57, + "step_time": 9.576232638006331 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 324.0, + "completions/max_terminated_length": 324.0, + "completions/mean_length": 249.0, + "completions/mean_terminated_length": 249.0, + "completions/min_length": 129.0, + "completions/min_terminated_length": 129.0, + "entropy": 0.42913514375686646, + "epoch": 0.29, + "frac_reward_zero_std": 0.0, + "grad_norm": 2.8544528484344482, + "learning_rate": 7.149999999999999e-07, + "loss": -0.056405920535326004, + "num_tokens": 494390.0, + "reward": 40.123748779296875, + "reward_std": 84.164794921875, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 40.02375030517578, + "rewards/supergames_reward/std": 84.164794921875, + "step": 58, + "step_time": 12.385353341000155 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 138.0, + "completions/max_terminated_length": 138.0, + "completions/mean_length": 107.25, + "completions/mean_terminated_length": 107.25, + "completions/min_length": 81.0, + "completions/min_terminated_length": 81.0, + "entropy": 0.34386488795280457, + "epoch": 0.295, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.6915154457092285, + "learning_rate": 7.1e-07, + "loss": 0.0865321159362793, + "num_tokens": 499816.0, + "reward": -8.244999885559082, + "reward_std": 4.681046962738037, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -8.345000267028809, + "rewards/supergames_reward/std": 4.681046962738037, + "step": 59, + "step_time": 5.42325339600211 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 187.0, + "completions/max_terminated_length": 187.0, + "completions/mean_length": 127.25, + "completions/mean_terminated_length": 127.25, + "completions/min_length": 97.0, + "completions/min_terminated_length": 97.0, + "entropy": 0.5430045127868652, + "epoch": 0.3, + "frac_reward_zero_std": 0.0, + "grad_norm": 5.635311126708984, + "learning_rate": 7.049999999999999e-07, + "loss": -0.05412375554442406, + "num_tokens": 505410.0, + "reward": 20.005001068115234, + "reward_std": 55.156005859375, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 19.904998779296875, + "rewards/supergames_reward/std": 55.156005859375, + "step": 60, + "step_time": 7.068010902003152 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 286.0, + "completions/max_terminated_length": 286.0, + "completions/mean_length": 192.125, + "completions/mean_terminated_length": 192.125, + "completions/min_length": 95.0, + "completions/min_terminated_length": 95.0, + "entropy": 0.4213869869709015, + "epoch": 0.305, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.4664618968963623, + "learning_rate": 7e-07, + "loss": -0.0495159812271595, + "num_tokens": 516283.0, + "reward": -47.212501525878906, + "reward_std": 91.2253646850586, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -47.3125, + "rewards/supergames_reward/std": 91.2253646850586, + "step": 61, + "step_time": 11.176304235996213 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 324.0, + "completions/max_terminated_length": 324.0, + "completions/mean_length": 171.0, + "completions/mean_terminated_length": 171.0, + "completions/min_length": 92.0, + "completions/min_terminated_length": 92.0, + "entropy": 0.4626811146736145, + "epoch": 0.31, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.949252128601074, + "learning_rate": 6.949999999999999e-07, + "loss": 0.192047581076622, + "num_tokens": 522251.0, + "reward": -15.092500686645508, + "reward_std": 49.82962417602539, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": -15.167499542236328, + "rewards/supergames_reward/std": 49.78091049194336, + "step": 62, + "step_time": 12.075224861997413 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.125, + "completions/max_length": 512.0, + "completions/max_terminated_length": 429.0, + "completions/mean_length": 283.5, + "completions/mean_terminated_length": 250.85714721679688, + "completions/min_length": 85.0, + "completions/min_terminated_length": 85.0, + "entropy": 0.4814216196537018, + "epoch": 0.315, + "frac_reward_zero_std": 0.0, + "grad_norm": 2.7918753623962402, + "learning_rate": 6.9e-07, + "loss": 0.15706156194210052, + "num_tokens": 533815.0, + "reward": -20.86625099182129, + "reward_std": 85.99250030517578, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": -20.94124984741211, + "rewards/supergames_reward/std": 85.96620178222656, + "step": 63, + "step_time": 19.448832260008203 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 189.0, + "completions/max_terminated_length": 189.0, + "completions/mean_length": 131.25, + "completions/mean_terminated_length": 131.25, + "completions/min_length": 106.0, + "completions/min_terminated_length": 106.0, + "entropy": 0.4929414987564087, + "epoch": 0.32, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.774101734161377, + "learning_rate": 6.85e-07, + "loss": -0.060691747814416885, + "num_tokens": 539465.0, + "reward": -21.0674991607666, + "reward_std": 6.892730236053467, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -21.167499542236328, + "rewards/supergames_reward/std": 6.892730236053467, + "step": 64, + "step_time": 7.141569407976931 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 454.0, + "completions/max_terminated_length": 454.0, + "completions/mean_length": 281.375, + "completions/mean_terminated_length": 281.375, + "completions/min_length": 107.0, + "completions/min_terminated_length": 107.0, + "entropy": 0.45866572856903076, + "epoch": 0.325, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.092717409133911, + "learning_rate": 6.800000000000001e-07, + "loss": 0.11694959551095963, + "num_tokens": 551012.0, + "reward": 49.938751220703125, + "reward_std": 79.1865005493164, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 49.8387451171875, + "rewards/supergames_reward/std": 79.1865005493164, + "step": 65, + "step_time": 17.06170882002334 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 203.0, + "completions/max_terminated_length": 203.0, + "completions/mean_length": 128.625, + "completions/mean_terminated_length": 128.625, + "completions/min_length": 88.0, + "completions/min_terminated_length": 88.0, + "entropy": 0.5005782246589661, + "epoch": 0.33, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.77616024017334, + "learning_rate": 6.75e-07, + "loss": 0.11242837458848953, + "num_tokens": 556577.0, + "reward": 48.875, + "reward_std": 49.13539123535156, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 48.775001525878906, + "rewards/supergames_reward/std": 49.13539123535156, + "step": 66, + "step_time": 7.695410129002994 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 436.0, + "completions/max_terminated_length": 436.0, + "completions/mean_length": 281.875, + "completions/mean_terminated_length": 281.875, + "completions/min_length": 133.0, + "completions/min_terminated_length": 133.0, + "entropy": 0.507097065448761, + "epoch": 0.335, + "frac_reward_zero_std": 0.0, + "grad_norm": 2.911184787750244, + "learning_rate": 6.7e-07, + "loss": -0.08151137083768845, + "num_tokens": 567120.0, + "reward": -12.886249542236328, + "reward_std": 84.04269409179688, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -12.986251831054688, + "rewards/supergames_reward/std": 84.04269409179688, + "step": 67, + "step_time": 16.409127942984924 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.125, + "completions/max_length": 512.0, + "completions/max_terminated_length": 501.0, + "completions/mean_length": 318.375, + "completions/mean_terminated_length": 290.71429443359375, + "completions/min_length": 65.0, + "completions/min_terminated_length": 65.0, + "entropy": 0.486402690410614, + "epoch": 0.34, + "frac_reward_zero_std": 0.0, + "grad_norm": 2.9461607933044434, + "learning_rate": 6.65e-07, + "loss": 0.2442178875207901, + "num_tokens": 578995.0, + "reward": 2.8387489318847656, + "reward_std": 98.47579956054688, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": 2.7637500762939453, + "rewards/supergames_reward/std": 98.44595336914062, + "step": 68, + "step_time": 19.32364999302081 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 291.0, + "completions/max_terminated_length": 291.0, + "completions/mean_length": 219.625, + "completions/mean_terminated_length": 219.625, + "completions/min_length": 117.0, + "completions/min_terminated_length": 117.0, + "entropy": 0.37856727838516235, + "epoch": 0.345, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.0204732418060303, + "learning_rate": 6.6e-07, + "loss": -0.08764916658401489, + "num_tokens": 589072.0, + "reward": -54.18375015258789, + "reward_std": 71.45012664794922, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -54.283748626708984, + "rewards/supergames_reward/std": 71.45013427734375, + "step": 69, + "step_time": 11.126611230982235 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.125, + "completions/max_length": 512.0, + "completions/max_terminated_length": 305.0, + "completions/mean_length": 254.75, + "completions/mean_terminated_length": 218.00001525878906, + "completions/min_length": 104.0, + "completions/min_terminated_length": 104.0, + "entropy": 0.46423614025115967, + "epoch": 0.35, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.0140154361724854, + "learning_rate": 6.55e-07, + "loss": 0.26797160506248474, + "num_tokens": 600438.0, + "reward": -32.04875183105469, + "reward_std": 69.86957550048828, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": -32.12375259399414, + "rewards/supergames_reward/std": 69.84178161621094, + "step": 70, + "step_time": 19.12617294798838 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 372.0, + "completions/max_terminated_length": 372.0, + "completions/mean_length": 194.125, + "completions/mean_terminated_length": 194.125, + "completions/min_length": 94.0, + "completions/min_terminated_length": 94.0, + "entropy": 0.3427293300628662, + "epoch": 0.355, + "frac_reward_zero_std": 0.0, + "grad_norm": 2.781383991241455, + "learning_rate": 6.5e-07, + "loss": 0.354524701833725, + "num_tokens": 610327.0, + "reward": -6.913749694824219, + "reward_std": 87.56071472167969, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": -6.988750457763672, + "rewards/supergames_reward/std": 87.53033447265625, + "step": 71, + "step_time": 14.041668199002743 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 197.0, + "completions/max_terminated_length": 197.0, + "completions/mean_length": 124.75, + "completions/mean_terminated_length": 124.75, + "completions/min_length": 93.0, + "completions/min_terminated_length": 93.0, + "entropy": 0.4455118477344513, + "epoch": 0.36, + "frac_reward_zero_std": 0.0, + "grad_norm": 5.726032733917236, + "learning_rate": 6.45e-07, + "loss": 0.18206289410591125, + "num_tokens": 615853.0, + "reward": 41.95124816894531, + "reward_std": 60.640811920166016, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 41.85124969482422, + "rewards/supergames_reward/std": 60.640811920166016, + "step": 72, + "step_time": 7.416493425989756 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 323.0, + "completions/max_terminated_length": 323.0, + "completions/mean_length": 147.375, + "completions/mean_terminated_length": 147.375, + "completions/min_length": 94.0, + "completions/min_terminated_length": 94.0, + "entropy": 0.480733186006546, + "epoch": 0.365, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.853621482849121, + "learning_rate": 6.4e-07, + "loss": 0.04558124020695686, + "num_tokens": 621584.0, + "reward": 14.884998321533203, + "reward_std": 46.25857162475586, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 14.78499984741211, + "rewards/supergames_reward/std": 46.25857162475586, + "step": 73, + "step_time": 11.95379540900467 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 283.0, + "completions/max_terminated_length": 283.0, + "completions/mean_length": 188.125, + "completions/mean_terminated_length": 188.125, + "completions/min_length": 116.0, + "completions/min_terminated_length": 116.0, + "entropy": 0.5526013374328613, + "epoch": 0.37, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.729883909225464, + "learning_rate": 6.35e-07, + "loss": 0.06268303096294403, + "num_tokens": 629545.0, + "reward": -68.05750274658203, + "reward_std": 59.44808578491211, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -68.15750122070312, + "rewards/supergames_reward/std": 59.44808578491211, + "step": 74, + "step_time": 10.64124580900534 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 141.0, + "completions/max_terminated_length": 141.0, + "completions/mean_length": 117.375, + "completions/mean_terminated_length": 117.375, + "completions/min_length": 100.0, + "completions/min_terminated_length": 100.0, + "entropy": 0.3744319677352905, + "epoch": 0.375, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.775154113769531, + "learning_rate": 6.3e-07, + "loss": -0.023284845054149628, + "num_tokens": 635052.0, + "reward": 3.9374990463256836, + "reward_std": 61.85771179199219, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": 3.862499237060547, + "rewards/supergames_reward/std": 61.80967712402344, + "step": 75, + "step_time": 5.432648951013107 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 268.0, + "completions/max_terminated_length": 268.0, + "completions/mean_length": 157.0, + "completions/mean_terminated_length": 157.0, + "completions/min_length": 104.0, + "completions/min_terminated_length": 104.0, + "entropy": 0.4822140336036682, + "epoch": 0.38, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.715704917907715, + "learning_rate": 6.249999999999999e-07, + "loss": -0.04918142035603523, + "num_tokens": 644644.0, + "reward": 48.64125061035156, + "reward_std": 78.0205078125, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": 48.56624984741211, + "rewards/supergames_reward/std": 77.96605682373047, + "step": 76, + "step_time": 10.348935816989979 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 380.0, + "completions/max_terminated_length": 380.0, + "completions/mean_length": 210.875, + "completions/mean_terminated_length": 210.875, + "completions/min_length": 126.0, + "completions/min_terminated_length": 126.0, + "entropy": 0.45722317695617676, + "epoch": 0.385, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.6057567596435547, + "learning_rate": 6.2e-07, + "loss": 0.058352649211883545, + "num_tokens": 652771.0, + "reward": -8.283750534057617, + "reward_std": 77.68436431884766, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -8.383749008178711, + "rewards/supergames_reward/std": 77.68437194824219, + "step": 77, + "step_time": 14.142948755004909 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 324.0, + "completions/max_terminated_length": 324.0, + "completions/mean_length": 188.25, + "completions/mean_terminated_length": 188.25, + "completions/min_length": 68.0, + "completions/min_terminated_length": 68.0, + "entropy": 0.33380764722824097, + "epoch": 0.39, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.2239863872528076, + "learning_rate": 6.149999999999999e-07, + "loss": -0.04592633992433548, + "num_tokens": 662653.0, + "reward": 11.153749465942383, + "reward_std": 89.88329315185547, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 11.053749084472656, + "rewards/supergames_reward/std": 89.88329315185547, + "step": 78, + "step_time": 12.277474621019792 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 143.0, + "completions/max_terminated_length": 143.0, + "completions/mean_length": 108.875, + "completions/mean_terminated_length": 108.875, + "completions/min_length": 85.0, + "completions/min_terminated_length": 85.0, + "entropy": 0.3518812656402588, + "epoch": 0.395, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.967776298522949, + "learning_rate": 6.1e-07, + "loss": -0.10800496488809586, + "num_tokens": 668076.0, + "reward": 70.33250427246094, + "reward_std": 49.9145393371582, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 70.23249816894531, + "rewards/supergames_reward/std": 49.91453552246094, + "step": 79, + "step_time": 5.52714040101273 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 184.0, + "completions/max_terminated_length": 184.0, + "completions/mean_length": 152.125, + "completions/mean_terminated_length": 152.125, + "completions/min_length": 104.0, + "completions/min_terminated_length": 104.0, + "entropy": 0.4699896574020386, + "epoch": 0.4, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.561161994934082, + "learning_rate": 6.049999999999999e-07, + "loss": 0.09505834430456161, + "num_tokens": 675701.0, + "reward": -20.25625228881836, + "reward_std": 53.78542709350586, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -20.35624885559082, + "rewards/supergames_reward/std": 53.78542709350586, + "step": 80, + "step_time": 7.155081019998761 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.125, + "completions/max_length": 512.0, + "completions/max_terminated_length": 506.0, + "completions/mean_length": 288.875, + "completions/mean_terminated_length": 257.0, + "completions/min_length": 115.0, + "completions/min_terminated_length": 115.0, + "entropy": 0.4086494445800781, + "epoch": 0.405, + "frac_reward_zero_std": 0.0, + "grad_norm": 2.3719611167907715, + "learning_rate": 6e-07, + "loss": 0.04727627709507942, + "num_tokens": 687332.0, + "reward": -37.20750045776367, + "reward_std": 91.7676010131836, + "rewards/json_format_reward/mean": 0.05000000074505806, + "rewards/json_format_reward/std": 0.09258200973272324, + "rewards/supergames_reward/mean": -37.25749969482422, + "rewards/supergames_reward/std": 91.72847747802734, + "step": 81, + "step_time": 19.231399144016905 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 376.0, + "completions/max_terminated_length": 376.0, + "completions/mean_length": 227.75, + "completions/mean_terminated_length": 227.75, + "completions/min_length": 39.0, + "completions/min_terminated_length": 39.0, + "entropy": 0.43233707547187805, + "epoch": 0.41, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.3391714096069336, + "learning_rate": 5.949999999999999e-07, + "loss": -0.25524550676345825, + "num_tokens": 698474.0, + "reward": -3.7287511825561523, + "reward_std": 91.25379943847656, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -3.828749656677246, + "rewards/supergames_reward/std": 91.25379943847656, + "step": 82, + "step_time": 14.329176368017215 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 299.0, + "completions/max_terminated_length": 299.0, + "completions/mean_length": 178.125, + "completions/mean_terminated_length": 178.125, + "completions/min_length": 91.0, + "completions/min_terminated_length": 91.0, + "entropy": 0.4293636977672577, + "epoch": 0.415, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.7271742820739746, + "learning_rate": 5.9e-07, + "loss": 0.09273672848939896, + "num_tokens": 708243.0, + "reward": 12.802498817443848, + "reward_std": 86.23526000976562, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 12.702500343322754, + "rewards/supergames_reward/std": 86.23526000976562, + "step": 83, + "step_time": 11.419686011999147 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 425.0, + "completions/max_terminated_length": 425.0, + "completions/mean_length": 255.375, + "completions/mean_terminated_length": 255.375, + "completions/min_length": 172.0, + "completions/min_terminated_length": 172.0, + "entropy": 0.38068246841430664, + "epoch": 0.42, + "frac_reward_zero_std": 0.0, + "grad_norm": 2.6244184970855713, + "learning_rate": 5.849999999999999e-07, + "loss": 0.02006213366985321, + "num_tokens": 718606.0, + "reward": 37.803749084472656, + "reward_std": 57.58624267578125, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 37.70375061035156, + "rewards/supergames_reward/std": 57.58624267578125, + "step": 84, + "step_time": 15.975198492989875 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 116.0, + "completions/max_terminated_length": 116.0, + "completions/mean_length": 108.125, + "completions/mean_terminated_length": 108.125, + "completions/min_length": 87.0, + "completions/min_terminated_length": 87.0, + "entropy": 0.339572012424469, + "epoch": 0.425, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.754759311676025, + "learning_rate": 5.8e-07, + "loss": -0.011015941388905048, + "num_tokens": 723983.0, + "reward": -3.8212504386901855, + "reward_std": 61.624786376953125, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -3.921250820159912, + "rewards/supergames_reward/std": 61.624786376953125, + "step": 85, + "step_time": 4.542777584982105 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 447.0, + "completions/max_terminated_length": 447.0, + "completions/mean_length": 291.375, + "completions/mean_terminated_length": 291.375, + "completions/min_length": 163.0, + "completions/min_terminated_length": 163.0, + "entropy": 0.5846173763275146, + "epoch": 0.43, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.55724835395813, + "learning_rate": 5.749999999999999e-07, + "loss": -0.04981120675802231, + "num_tokens": 735642.0, + "reward": 50.07499694824219, + "reward_std": 92.6283187866211, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": 50.0, + "rewards/supergames_reward/std": 92.5820083618164, + "step": 86, + "step_time": 16.777178087009815 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 232.0, + "completions/max_terminated_length": 232.0, + "completions/mean_length": 192.0, + "completions/mean_terminated_length": 192.0, + "completions/min_length": 139.0, + "completions/min_terminated_length": 139.0, + "entropy": 0.5139474272727966, + "epoch": 0.435, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.8860411643981934, + "learning_rate": 5.699999999999999e-07, + "loss": -0.02194221317768097, + "num_tokens": 743602.0, + "reward": -52.522499084472656, + "reward_std": 73.8686294555664, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -52.62249755859375, + "rewards/supergames_reward/std": 73.86863708496094, + "step": 87, + "step_time": 8.772893911984283 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 266.0, + "completions/max_terminated_length": 266.0, + "completions/mean_length": 146.875, + "completions/mean_terminated_length": 146.875, + "completions/min_length": 89.0, + "completions/min_terminated_length": 89.0, + "entropy": 0.3428497910499573, + "epoch": 0.44, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.509483337402344, + "learning_rate": 5.649999999999999e-07, + "loss": 0.19857533276081085, + "num_tokens": 753065.0, + "reward": 63.63874816894531, + "reward_std": 67.54963684082031, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 63.53874969482422, + "rewards/supergames_reward/std": 67.54963684082031, + "step": 88, + "step_time": 10.230529835011112 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 446.0, + "completions/max_terminated_length": 446.0, + "completions/mean_length": 270.625, + "completions/mean_terminated_length": 270.625, + "completions/min_length": 103.0, + "completions/min_terminated_length": 103.0, + "entropy": 0.4417419731616974, + "epoch": 0.445, + "frac_reward_zero_std": 0.0, + "grad_norm": 2.973951816558838, + "learning_rate": 5.6e-07, + "loss": -0.00031440958264283836, + "num_tokens": 764534.0, + "reward": -9.736251831054688, + "reward_std": 92.25879669189453, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -9.836250305175781, + "rewards/supergames_reward/std": 92.25880432128906, + "step": 89, + "step_time": 17.40752330099349 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 398.0, + "completions/max_terminated_length": 398.0, + "completions/mean_length": 227.375, + "completions/mean_terminated_length": 227.375, + "completions/min_length": 98.0, + "completions/min_terminated_length": 98.0, + "entropy": 0.4732024073600769, + "epoch": 0.45, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.2560815811157227, + "learning_rate": 5.55e-07, + "loss": 0.03820464387536049, + "num_tokens": 772809.0, + "reward": 5.982499599456787, + "reward_std": 58.833797454833984, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 5.8824992179870605, + "rewards/supergames_reward/std": 58.83380126953125, + "step": 90, + "step_time": 15.007269965979503 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 269.0, + "completions/max_terminated_length": 269.0, + "completions/mean_length": 207.5, + "completions/mean_terminated_length": 207.5, + "completions/min_length": 92.0, + "completions/min_terminated_length": 92.0, + "entropy": 0.42038559913635254, + "epoch": 0.455, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.3649117946624756, + "learning_rate": 5.5e-07, + "loss": 0.1510048508644104, + "num_tokens": 782805.0, + "reward": -5.61500358581543, + "reward_std": 82.67152404785156, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -5.714998245239258, + "rewards/supergames_reward/std": 82.67152404785156, + "step": 91, + "step_time": 10.386137172987219 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 416.0, + "completions/max_terminated_length": 416.0, + "completions/mean_length": 218.375, + "completions/mean_terminated_length": 218.375, + "completions/min_length": 95.0, + "completions/min_terminated_length": 95.0, + "entropy": 0.3971734642982483, + "epoch": 0.46, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.433772563934326, + "learning_rate": 5.45e-07, + "loss": -0.3027213513851166, + "num_tokens": 793840.0, + "reward": -29.06500244140625, + "reward_std": 91.68325805664062, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -29.16499900817871, + "rewards/supergames_reward/std": 91.68325805664062, + "step": 92, + "step_time": 15.790611553995404 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 345.0, + "completions/max_terminated_length": 345.0, + "completions/mean_length": 206.625, + "completions/mean_terminated_length": 206.625, + "completions/min_length": 90.0, + "completions/min_terminated_length": 90.0, + "entropy": 0.5906907320022583, + "epoch": 0.465, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.293088912963867, + "learning_rate": 5.4e-07, + "loss": 0.09760032594203949, + "num_tokens": 803821.0, + "reward": -72.73999786376953, + "reward_std": 50.72484588623047, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": -72.81500244140625, + "rewards/supergames_reward/std": 50.7094841003418, + "step": 93, + "step_time": 13.043119941983605 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 143.0, + "completions/max_terminated_length": 143.0, + "completions/mean_length": 115.75, + "completions/mean_terminated_length": 115.75, + "completions/min_length": 80.0, + "completions/min_terminated_length": 80.0, + "entropy": 0.4233115315437317, + "epoch": 0.47, + "frac_reward_zero_std": 0.0, + "grad_norm": 5.58886194229126, + "learning_rate": 5.35e-07, + "loss": -0.04040650278329849, + "num_tokens": 809299.0, + "reward": -3.3050003051757812, + "reward_std": 24.443714141845703, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -3.405000686645508, + "rewards/supergames_reward/std": 24.443714141845703, + "step": 94, + "step_time": 5.512874762003776 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 122.0, + "completions/max_terminated_length": 122.0, + "completions/mean_length": 103.125, + "completions/mean_terminated_length": 103.125, + "completions/min_length": 86.0, + "completions/min_terminated_length": 86.0, + "entropy": 0.4055764377117157, + "epoch": 0.475, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.981549263000488, + "learning_rate": 5.3e-07, + "loss": -0.07588938623666763, + "num_tokens": 814716.0, + "reward": 60.06999969482422, + "reward_std": 42.708473205566406, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 59.970001220703125, + "rewards/supergames_reward/std": 42.708473205566406, + "step": 95, + "step_time": 4.766389057011111 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 475.0, + "completions/max_terminated_length": 475.0, + "completions/mean_length": 218.375, + "completions/mean_terminated_length": 218.375, + "completions/min_length": 81.0, + "completions/min_terminated_length": 81.0, + "entropy": 0.408913254737854, + "epoch": 0.48, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.79402756690979, + "learning_rate": 5.25e-07, + "loss": -0.3087541460990906, + "num_tokens": 824791.0, + "reward": -15.042500495910645, + "reward_std": 84.37246704101562, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -15.142499923706055, + "rewards/supergames_reward/std": 84.37246704101562, + "step": 96, + "step_time": 17.779843941010768 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 154.0, + "completions/max_terminated_length": 154.0, + "completions/mean_length": 121.625, + "completions/mean_terminated_length": 121.625, + "completions/min_length": 80.0, + "completions/min_terminated_length": 80.0, + "entropy": 0.529015839099884, + "epoch": 0.485, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.142816066741943, + "learning_rate": 5.2e-07, + "loss": 0.07216037809848785, + "num_tokens": 830300.0, + "reward": -32.743751525878906, + "reward_std": 54.61775588989258, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": -32.81875228881836, + "rewards/supergames_reward/std": 54.58256149291992, + "step": 97, + "step_time": 5.913989890017547 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 126.0, + "completions/max_terminated_length": 126.0, + "completions/mean_length": 95.625, + "completions/mean_terminated_length": 95.625, + "completions/min_length": 74.0, + "completions/min_terminated_length": 74.0, + "entropy": 0.33743274211883545, + "epoch": 0.49, + "frac_reward_zero_std": 0.0, + "grad_norm": 5.197273254394531, + "learning_rate": 5.149999999999999e-07, + "loss": 0.07500407844781876, + "num_tokens": 835633.0, + "reward": 11.00999927520752, + "reward_std": 62.73848342895508, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 10.910000801086426, + "rewards/supergames_reward/std": 62.73848342895508, + "step": 98, + "step_time": 4.9155233050114475 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 149.0, + "completions/max_terminated_length": 149.0, + "completions/mean_length": 108.5, + "completions/mean_terminated_length": 108.5, + "completions/min_length": 86.0, + "completions/min_terminated_length": 86.0, + "entropy": 0.3106803297996521, + "epoch": 0.495, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.864002227783203, + "learning_rate": 5.1e-07, + "loss": -0.1206185445189476, + "num_tokens": 841045.0, + "reward": 14.238749504089355, + "reward_std": 51.77016830444336, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 14.138750076293945, + "rewards/supergames_reward/std": 51.77016830444336, + "step": 99, + "step_time": 5.7507751359953545 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 149.0, + "completions/max_terminated_length": 149.0, + "completions/mean_length": 97.125, + "completions/mean_terminated_length": 97.125, + "completions/min_length": 73.0, + "completions/min_terminated_length": 73.0, + "entropy": 0.35160186886787415, + "epoch": 0.5, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.915731906890869, + "learning_rate": 5.049999999999999e-07, + "loss": -0.07951541244983673, + "num_tokens": 846398.0, + "reward": 67.78125, + "reward_std": 38.15436553955078, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 67.68124389648438, + "rewards/supergames_reward/std": 38.15436553955078, + "step": 100, + "step_time": 5.994720440998208 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 320.0, + "completions/max_terminated_length": 320.0, + "completions/mean_length": 208.75, + "completions/mean_terminated_length": 208.75, + "completions/min_length": 95.0, + "completions/min_terminated_length": 95.0, + "entropy": 0.452095091342926, + "epoch": 0.505, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.065176963806152, + "learning_rate": 5e-07, + "loss": -0.1397111564874649, + "num_tokens": 857420.0, + "reward": -5.801251411437988, + "reward_std": 73.56117248535156, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -5.90125036239624, + "rewards/supergames_reward/std": 73.5611801147461, + "step": 101, + "step_time": 12.28178753197426 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 230.0, + "completions/max_terminated_length": 230.0, + "completions/mean_length": 171.375, + "completions/mean_terminated_length": 171.375, + "completions/min_length": 95.0, + "completions/min_terminated_length": 95.0, + "entropy": 0.4138447642326355, + "epoch": 0.51, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.998823404312134, + "learning_rate": 4.95e-07, + "loss": -0.08941879868507385, + "num_tokens": 867103.0, + "reward": 92.24749755859375, + "reward_std": 14.540005683898926, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 92.14749908447266, + "rewards/supergames_reward/std": 14.540006637573242, + "step": 102, + "step_time": 8.992682139010867 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 311.0, + "completions/max_terminated_length": 311.0, + "completions/mean_length": 163.0, + "completions/mean_terminated_length": 163.0, + "completions/min_length": 74.0, + "completions/min_terminated_length": 74.0, + "entropy": 0.4182547330856323, + "epoch": 0.515, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.419595241546631, + "learning_rate": 4.9e-07, + "loss": -0.19304415583610535, + "num_tokens": 876679.0, + "reward": 79.13249969482422, + "reward_std": 59.3050422668457, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 79.03250122070312, + "rewards/supergames_reward/std": 59.30504608154297, + "step": 103, + "step_time": 11.836970244999975 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 253.0, + "completions/max_terminated_length": 253.0, + "completions/mean_length": 155.625, + "completions/mean_terminated_length": 155.625, + "completions/min_length": 94.0, + "completions/min_terminated_length": 94.0, + "entropy": 0.49889081716537476, + "epoch": 0.52, + "frac_reward_zero_std": 0.0, + "grad_norm": 5.463719367980957, + "learning_rate": 4.85e-07, + "loss": -0.04786720871925354, + "num_tokens": 884348.0, + "reward": 48.616249084472656, + "reward_std": 73.30741119384766, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 48.51625061035156, + "rewards/supergames_reward/std": 73.30741119384766, + "step": 104, + "step_time": 9.579507841990562 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 380.0, + "completions/max_terminated_length": 380.0, + "completions/mean_length": 269.5, + "completions/mean_terminated_length": 269.5, + "completions/min_length": 101.0, + "completions/min_terminated_length": 101.0, + "entropy": 0.5286482572555542, + "epoch": 0.525, + "frac_reward_zero_std": 0.0, + "grad_norm": 2.993295431137085, + "learning_rate": 4.8e-07, + "loss": -0.08071611076593399, + "num_tokens": 894856.0, + "reward": 34.970001220703125, + "reward_std": 85.324462890625, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 34.869998931884766, + "rewards/supergames_reward/std": 85.324462890625, + "step": 105, + "step_time": 14.295730627985904 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 261.0, + "completions/max_terminated_length": 261.0, + "completions/mean_length": 203.25, + "completions/mean_terminated_length": 203.25, + "completions/min_length": 91.0, + "completions/min_terminated_length": 91.0, + "entropy": 0.6049264073371887, + "epoch": 0.53, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.787306785583496, + "learning_rate": 4.7499999999999995e-07, + "loss": 0.19928883016109467, + "num_tokens": 902922.0, + "reward": 32.44499969482422, + "reward_std": 70.4663314819336, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": 32.369998931884766, + "rewards/supergames_reward/std": 70.41261291503906, + "step": 106, + "step_time": 9.84894504098338 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 279.0, + "completions/max_terminated_length": 279.0, + "completions/mean_length": 194.0, + "completions/mean_terminated_length": 194.0, + "completions/min_length": 109.0, + "completions/min_terminated_length": 109.0, + "entropy": 0.5115755200386047, + "epoch": 0.535, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.277207374572754, + "learning_rate": 4.6999999999999995e-07, + "loss": 0.09118300676345825, + "num_tokens": 910906.0, + "reward": 40.432498931884766, + "reward_std": 48.54548263549805, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 40.33250045776367, + "rewards/supergames_reward/std": 48.54548263549805, + "step": 107, + "step_time": 10.466603949986165 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 332.0, + "completions/max_terminated_length": 332.0, + "completions/mean_length": 195.875, + "completions/mean_terminated_length": 195.875, + "completions/min_length": 87.0, + "completions/min_terminated_length": 87.0, + "entropy": 0.4192962944507599, + "epoch": 0.54, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.308061122894287, + "learning_rate": 4.65e-07, + "loss": 0.28064602613449097, + "num_tokens": 920745.0, + "reward": -21.982500076293945, + "reward_std": 66.31195068359375, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -22.08249855041504, + "rewards/supergames_reward/std": 66.31195068359375, + "step": 108, + "step_time": 12.595848233992001 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 269.0, + "completions/max_terminated_length": 269.0, + "completions/mean_length": 195.125, + "completions/mean_terminated_length": 195.125, + "completions/min_length": 111.0, + "completions/min_terminated_length": 111.0, + "entropy": 0.37523961067199707, + "epoch": 0.545, + "frac_reward_zero_std": 0.0, + "grad_norm": 2.802706718444824, + "learning_rate": 4.6e-07, + "loss": -0.06614465266466141, + "num_tokens": 930618.0, + "reward": 71.75, + "reward_std": 51.14524459838867, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 71.6500015258789, + "rewards/supergames_reward/std": 51.1452522277832, + "step": 109, + "step_time": 10.33884758799104 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 331.0, + "completions/max_terminated_length": 331.0, + "completions/mean_length": 167.75, + "completions/mean_terminated_length": 167.75, + "completions/min_length": 89.0, + "completions/min_terminated_length": 89.0, + "entropy": 0.4339042007923126, + "epoch": 0.55, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.254003047943115, + "learning_rate": 4.55e-07, + "loss": 0.34405598044395447, + "num_tokens": 941256.0, + "reward": 99.21875, + "reward_std": 2.492553234100342, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 99.1187515258789, + "rewards/supergames_reward/std": 2.4925525188446045, + "step": 110, + "step_time": 12.657525141985388 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 285.0, + "completions/max_terminated_length": 285.0, + "completions/mean_length": 211.0, + "completions/mean_terminated_length": 211.0, + "completions/min_length": 112.0, + "completions/min_terminated_length": 112.0, + "entropy": 0.46809861063957214, + "epoch": 0.555, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.811166286468506, + "learning_rate": 4.5e-07, + "loss": 0.045113153755664825, + "num_tokens": 951296.0, + "reward": 55.27750015258789, + "reward_std": 47.90687561035156, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 55.1775016784668, + "rewards/supergames_reward/std": 47.9068717956543, + "step": 111, + "step_time": 10.91118389699841 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 145.0, + "completions/max_terminated_length": 145.0, + "completions/mean_length": 108.375, + "completions/mean_terminated_length": 108.375, + "completions/min_length": 86.0, + "completions/min_terminated_length": 86.0, + "entropy": 0.34181955456733704, + "epoch": 0.56, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.537296772003174, + "learning_rate": 4.45e-07, + "loss": -0.052413541823625565, + "num_tokens": 956675.0, + "reward": 47.56624984741211, + "reward_std": 49.185157775878906, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 47.466251373291016, + "rewards/supergames_reward/std": 49.18516159057617, + "step": 112, + "step_time": 5.552288047998445 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 245.0, + "completions/max_terminated_length": 245.0, + "completions/mean_length": 189.125, + "completions/mean_terminated_length": 189.125, + "completions/min_length": 102.0, + "completions/min_terminated_length": 102.0, + "entropy": 0.5034605860710144, + "epoch": 0.565, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.349876642227173, + "learning_rate": 4.3999999999999997e-07, + "loss": -0.08186715841293335, + "num_tokens": 964636.0, + "reward": 81.32374572753906, + "reward_std": 36.85074996948242, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 81.2237548828125, + "rewards/supergames_reward/std": 36.85074996948242, + "step": 113, + "step_time": 9.267611294984818 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 178.0, + "completions/max_terminated_length": 178.0, + "completions/mean_length": 111.25, + "completions/mean_terminated_length": 111.25, + "completions/min_length": 67.0, + "completions/min_terminated_length": 67.0, + "entropy": 0.3229002058506012, + "epoch": 0.57, + "frac_reward_zero_std": 0.0, + "grad_norm": 5.245575428009033, + "learning_rate": 4.3499999999999996e-07, + "loss": 0.06544198095798492, + "num_tokens": 970070.0, + "reward": 63.1775016784668, + "reward_std": 52.30624008178711, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 63.07749938964844, + "rewards/supergames_reward/std": 52.30624008178711, + "step": 114, + "step_time": 6.731139309995342 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 224.0, + "completions/max_terminated_length": 224.0, + "completions/mean_length": 170.0, + "completions/mean_terminated_length": 170.0, + "completions/min_length": 94.0, + "completions/min_terminated_length": 94.0, + "entropy": 0.3274807929992676, + "epoch": 0.575, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.5544140338897705, + "learning_rate": 4.2999999999999996e-07, + "loss": -0.06953569501638412, + "num_tokens": 979702.0, + "reward": 25.62125015258789, + "reward_std": 64.52003479003906, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 25.521244049072266, + "rewards/supergames_reward/std": 64.52003479003906, + "step": 115, + "step_time": 8.702618776995223 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 427.0, + "completions/max_terminated_length": 427.0, + "completions/mean_length": 209.5, + "completions/mean_terminated_length": 209.5, + "completions/min_length": 100.0, + "completions/min_terminated_length": 100.0, + "entropy": 0.4970880150794983, + "epoch": 0.58, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.4932165145874023, + "learning_rate": 4.2499999999999995e-07, + "loss": -0.2281273603439331, + "num_tokens": 990690.0, + "reward": 50.54750061035156, + "reward_std": 70.5126953125, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 50.4474983215332, + "rewards/supergames_reward/std": 70.51270294189453, + "step": 116, + "step_time": 16.204386608995264 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 228.0, + "completions/max_terminated_length": 228.0, + "completions/mean_length": 132.375, + "completions/mean_terminated_length": 132.375, + "completions/min_length": 85.0, + "completions/min_terminated_length": 85.0, + "entropy": 0.4380227029323578, + "epoch": 0.585, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.089909076690674, + "learning_rate": 4.1999999999999995e-07, + "loss": 0.12121254950761795, + "num_tokens": 996333.0, + "reward": -15.333749771118164, + "reward_std": 47.200164794921875, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -15.433748245239258, + "rewards/supergames_reward/std": 47.200164794921875, + "step": 117, + "step_time": 8.546233230998041 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 323.0, + "completions/max_terminated_length": 323.0, + "completions/mean_length": 137.625, + "completions/mean_terminated_length": 137.625, + "completions/min_length": 90.0, + "completions/min_terminated_length": 90.0, + "entropy": 0.4177803695201874, + "epoch": 0.59, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.8712472915649414, + "learning_rate": 4.1499999999999994e-07, + "loss": 0.14192955195903778, + "num_tokens": 1002010.0, + "reward": -30.700000762939453, + "reward_std": 17.224035263061523, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -30.799999237060547, + "rewards/supergames_reward/std": 17.224035263061523, + "step": 118, + "step_time": 11.878434990998358 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 453.0, + "completions/max_terminated_length": 453.0, + "completions/mean_length": 238.125, + "completions/mean_terminated_length": 238.125, + "completions/min_length": 109.0, + "completions/min_terminated_length": 109.0, + "entropy": 0.5640217065811157, + "epoch": 0.595, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.064737319946289, + "learning_rate": 4.0999999999999994e-07, + "loss": -0.023621462285518646, + "num_tokens": 1012259.0, + "reward": 43.683746337890625, + "reward_std": 70.4210205078125, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 43.5837516784668, + "rewards/supergames_reward/std": 70.4210205078125, + "step": 119, + "step_time": 16.829514264973113 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 236.0, + "completions/max_terminated_length": 236.0, + "completions/mean_length": 193.75, + "completions/mean_terminated_length": 193.75, + "completions/min_length": 151.0, + "completions/min_terminated_length": 151.0, + "entropy": 0.5795091390609741, + "epoch": 0.6, + "frac_reward_zero_std": 0.0, + "grad_norm": 5.423007965087891, + "learning_rate": 4.05e-07, + "loss": -0.0666111558675766, + "num_tokens": 1020241.0, + "reward": -85.86000061035156, + "reward_std": 27.4442081451416, + "rewards/json_format_reward/mean": 0.05000000074505806, + "rewards/json_format_reward/std": 0.09258200973272324, + "rewards/supergames_reward/mean": -85.90999603271484, + "rewards/supergames_reward/std": 27.41469955444336, + "step": 120, + "step_time": 8.95496519000153 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 331.0, + "completions/max_terminated_length": 331.0, + "completions/mean_length": 231.0, + "completions/mean_terminated_length": 231.0, + "completions/min_length": 165.0, + "completions/min_terminated_length": 165.0, + "entropy": 0.3767699897289276, + "epoch": 0.605, + "frac_reward_zero_std": 0.0, + "grad_norm": 2.7434473037719727, + "learning_rate": 4e-07, + "loss": -0.1165132150053978, + "num_tokens": 1031465.0, + "reward": 61.45624923706055, + "reward_std": 72.21080780029297, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 61.35625076293945, + "rewards/supergames_reward/std": 72.2108154296875, + "step": 121, + "step_time": 12.588820598000893 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 297.0, + "completions/max_terminated_length": 297.0, + "completions/mean_length": 223.375, + "completions/mean_terminated_length": 223.375, + "completions/min_length": 112.0, + "completions/min_terminated_length": 112.0, + "entropy": 0.4537277817726135, + "epoch": 0.61, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.4382436275482178, + "learning_rate": 3.95e-07, + "loss": -0.1038316935300827, + "num_tokens": 1042548.0, + "reward": 5.658749580383301, + "reward_std": 83.77420043945312, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 5.558750152587891, + "rewards/supergames_reward/std": 83.77420043945312, + "step": 122, + "step_time": 11.39525846898323 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 456.0, + "completions/max_terminated_length": 456.0, + "completions/mean_length": 314.25, + "completions/mean_terminated_length": 314.25, + "completions/min_length": 95.0, + "completions/min_terminated_length": 95.0, + "entropy": 0.4113953411579132, + "epoch": 0.615, + "frac_reward_zero_std": 0.0, + "grad_norm": 2.449777841567993, + "learning_rate": 3.8999999999999997e-07, + "loss": -0.1303664743900299, + "num_tokens": 1054422.0, + "reward": 35.98249816894531, + "reward_std": 88.6131362915039, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 35.88249969482422, + "rewards/supergames_reward/std": 88.6131362915039, + "step": 123, + "step_time": 17.001118954998674 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 242.0, + "completions/max_terminated_length": 242.0, + "completions/mean_length": 172.5, + "completions/mean_terminated_length": 172.5, + "completions/min_length": 137.0, + "completions/min_terminated_length": 137.0, + "entropy": 0.4670729339122772, + "epoch": 0.62, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.951946973800659, + "learning_rate": 3.8499999999999997e-07, + "loss": 0.05888795852661133, + "num_tokens": 1062210.0, + "reward": 43.368751525878906, + "reward_std": 40.30733871459961, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 43.26874542236328, + "rewards/supergames_reward/std": 40.307342529296875, + "step": 124, + "step_time": 9.142582255997695 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 148.0, + "completions/max_terminated_length": 148.0, + "completions/mean_length": 106.625, + "completions/mean_terminated_length": 106.625, + "completions/min_length": 79.0, + "completions/min_terminated_length": 79.0, + "entropy": 0.30219921469688416, + "epoch": 0.625, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.391322135925293, + "learning_rate": 3.7999999999999996e-07, + "loss": 0.14367049932479858, + "num_tokens": 1067599.0, + "reward": 42.63374710083008, + "reward_std": 58.20897674560547, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 42.533748626708984, + "rewards/supergames_reward/std": 58.208984375, + "step": 125, + "step_time": 5.636063549987739 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 243.0, + "completions/max_terminated_length": 243.0, + "completions/mean_length": 179.25, + "completions/mean_terminated_length": 179.25, + "completions/min_length": 134.0, + "completions/min_terminated_length": 134.0, + "entropy": 0.4834887683391571, + "epoch": 0.63, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.9391286373138428, + "learning_rate": 3.75e-07, + "loss": -0.06614185124635696, + "num_tokens": 1075457.0, + "reward": 8.302498817443848, + "reward_std": 75.10364532470703, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 8.202500343322754, + "rewards/supergames_reward/std": 75.10364532470703, + "step": 126, + "step_time": 9.130600645992672 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 196.0, + "completions/max_terminated_length": 196.0, + "completions/mean_length": 157.0, + "completions/mean_terminated_length": 157.0, + "completions/min_length": 117.0, + "completions/min_terminated_length": 117.0, + "entropy": 0.4801807701587677, + "epoch": 0.635, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.425615310668945, + "learning_rate": 3.7e-07, + "loss": -0.08830522745847702, + "num_tokens": 1083129.0, + "reward": -10.563751220703125, + "reward_std": 64.00443267822266, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -10.663749694824219, + "rewards/supergames_reward/std": 64.00444030761719, + "step": 127, + "step_time": 7.4972667430120055 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 412.0, + "completions/max_terminated_length": 412.0, + "completions/mean_length": 239.25, + "completions/mean_terminated_length": 239.25, + "completions/min_length": 89.0, + "completions/min_terminated_length": 89.0, + "entropy": 0.4667063355445862, + "epoch": 0.64, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.199113130569458, + "learning_rate": 3.65e-07, + "loss": -0.043856702744960785, + "num_tokens": 1094355.0, + "reward": -55.89250183105469, + "reward_std": 53.05980682373047, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -55.99250030517578, + "rewards/supergames_reward/std": 53.05980682373047, + "step": 128, + "step_time": 15.429580625001108 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 370.0, + "completions/max_terminated_length": 370.0, + "completions/mean_length": 244.75, + "completions/mean_terminated_length": 244.75, + "completions/min_length": 127.0, + "completions/min_terminated_length": 127.0, + "entropy": 0.57573401927948, + "epoch": 0.645, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.6800079345703125, + "learning_rate": 3.6e-07, + "loss": 0.09521033614873886, + "num_tokens": 1105633.0, + "reward": -23.35250473022461, + "reward_std": 102.31401062011719, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -23.452499389648438, + "rewards/supergames_reward/std": 102.31402587890625, + "step": 129, + "step_time": 13.920327747007832 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 258.0, + "completions/max_terminated_length": 258.0, + "completions/mean_length": 188.25, + "completions/mean_terminated_length": 188.25, + "completions/min_length": 131.0, + "completions/min_terminated_length": 131.0, + "entropy": 0.5341438055038452, + "epoch": 0.65, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.97955060005188, + "learning_rate": 3.55e-07, + "loss": -0.09512650966644287, + "num_tokens": 1113571.0, + "reward": 29.181249618530273, + "reward_std": 80.10585021972656, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 29.081249237060547, + "rewards/supergames_reward/std": 80.10585021972656, + "step": 130, + "step_time": 9.69286120700417 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 257.0, + "completions/max_terminated_length": 257.0, + "completions/mean_length": 180.75, + "completions/mean_terminated_length": 180.75, + "completions/min_length": 136.0, + "completions/min_terminated_length": 136.0, + "entropy": 0.46591487526893616, + "epoch": 0.655, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.3518130779266357, + "learning_rate": 3.5e-07, + "loss": 0.04714229330420494, + "num_tokens": 1121457.0, + "reward": -8.287501335144043, + "reward_std": 76.35875701904297, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -8.387499809265137, + "rewards/supergames_reward/std": 76.35875701904297, + "step": 131, + "step_time": 9.636637121002423 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 190.0, + "completions/max_terminated_length": 190.0, + "completions/mean_length": 147.75, + "completions/mean_terminated_length": 147.75, + "completions/min_length": 90.0, + "completions/min_terminated_length": 90.0, + "entropy": 0.4407735764980316, + "epoch": 0.66, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.498741626739502, + "learning_rate": 3.45e-07, + "loss": -0.03996007889509201, + "num_tokens": 1129095.0, + "reward": 0.1999988555908203, + "reward_std": 84.63554382324219, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": 0.125, + "rewards/supergames_reward/std": 84.6017074584961, + "step": 132, + "step_time": 7.510833573003765 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 358.0, + "completions/max_terminated_length": 358.0, + "completions/mean_length": 213.875, + "completions/mean_terminated_length": 213.875, + "completions/min_length": 137.0, + "completions/min_terminated_length": 137.0, + "entropy": 0.5186149477958679, + "epoch": 0.665, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.3987107276916504, + "learning_rate": 3.4000000000000003e-07, + "loss": 0.1863594949245453, + "num_tokens": 1139110.0, + "reward": -21.328752517700195, + "reward_std": 88.9102783203125, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -21.428749084472656, + "rewards/supergames_reward/std": 88.9102783203125, + "step": 133, + "step_time": 13.393839450000087 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 230.0, + "completions/max_terminated_length": 230.0, + "completions/mean_length": 169.875, + "completions/mean_terminated_length": 169.875, + "completions/min_length": 121.0, + "completions/min_terminated_length": 121.0, + "entropy": 0.49445784091949463, + "epoch": 0.67, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.479673862457275, + "learning_rate": 3.35e-07, + "loss": 0.03618276119232178, + "num_tokens": 1146885.0, + "reward": -14.042501449584961, + "reward_std": 83.22671508789062, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -14.142499923706055, + "rewards/supergames_reward/std": 83.22671508789062, + "step": 134, + "step_time": 8.703399487014394 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 168.0, + "completions/max_terminated_length": 168.0, + "completions/mean_length": 124.75, + "completions/mean_terminated_length": 124.75, + "completions/min_length": 83.0, + "completions/min_terminated_length": 83.0, + "entropy": 0.2997814118862152, + "epoch": 0.675, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.440976619720459, + "learning_rate": 3.3e-07, + "loss": -0.029299044981598854, + "num_tokens": 1152427.0, + "reward": 45.274993896484375, + "reward_std": 44.401519775390625, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 45.17499923706055, + "rewards/supergames_reward/std": 44.40152359008789, + "step": 135, + "step_time": 6.334954546997324 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 161.0, + "completions/max_terminated_length": 161.0, + "completions/mean_length": 123.625, + "completions/mean_terminated_length": 123.625, + "completions/min_length": 93.0, + "completions/min_terminated_length": 93.0, + "entropy": 0.3462243974208832, + "epoch": 0.68, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.1878814697265625, + "learning_rate": 3.25e-07, + "loss": -0.10772529244422913, + "num_tokens": 1157952.0, + "reward": 44.186248779296875, + "reward_std": 51.78197479248047, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 44.08625030517578, + "rewards/supergames_reward/std": 51.7819709777832, + "step": 136, + "step_time": 6.067531232984038 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 155.0, + "completions/max_terminated_length": 155.0, + "completions/mean_length": 106.75, + "completions/mean_terminated_length": 106.75, + "completions/min_length": 74.0, + "completions/min_terminated_length": 74.0, + "entropy": 0.3662136197090149, + "epoch": 0.685, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.508971214294434, + "learning_rate": 3.2e-07, + "loss": -0.11172202974557877, + "num_tokens": 1163334.0, + "reward": 53.0574951171875, + "reward_std": 58.36653518676758, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 52.957496643066406, + "rewards/supergames_reward/std": 58.36653518676758, + "step": 137, + "step_time": 5.8906258280039765 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 175.0, + "completions/max_terminated_length": 175.0, + "completions/mean_length": 125.375, + "completions/mean_terminated_length": 125.375, + "completions/min_length": 90.0, + "completions/min_terminated_length": 90.0, + "entropy": 0.3352622091770172, + "epoch": 0.69, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.1937079429626465, + "learning_rate": 3.15e-07, + "loss": 0.08307601511478424, + "num_tokens": 1168929.0, + "reward": 11.796250343322754, + "reward_std": 53.48870849609375, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 11.696250915527344, + "rewards/supergames_reward/std": 53.48870849609375, + "step": 138, + "step_time": 6.595962718012743 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 254.0, + "completions/max_terminated_length": 254.0, + "completions/mean_length": 193.0, + "completions/mean_terminated_length": 193.0, + "completions/min_length": 111.0, + "completions/min_terminated_length": 111.0, + "entropy": 0.5486535429954529, + "epoch": 0.695, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.861871242523193, + "learning_rate": 3.1e-07, + "loss": 0.0321456640958786, + "num_tokens": 1176937.0, + "reward": 50.463748931884766, + "reward_std": 70.37553405761719, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": 50.38875198364258, + "rewards/supergames_reward/std": 70.31440734863281, + "step": 139, + "step_time": 9.544256388006033 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 261.0, + "completions/max_terminated_length": 261.0, + "completions/mean_length": 189.5, + "completions/mean_terminated_length": 189.5, + "completions/min_length": 95.0, + "completions/min_terminated_length": 95.0, + "entropy": 0.42230603098869324, + "epoch": 0.7, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.830677032470703, + "learning_rate": 3.05e-07, + "loss": -0.07222295552492142, + "num_tokens": 1186789.0, + "reward": 33.44999694824219, + "reward_std": 67.1298599243164, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 33.349998474121094, + "rewards/supergames_reward/std": 67.1298599243164, + "step": 140, + "step_time": 10.002322309010196 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 282.0, + "completions/max_terminated_length": 282.0, + "completions/mean_length": 197.375, + "completions/mean_terminated_length": 197.375, + "completions/min_length": 92.0, + "completions/min_terminated_length": 92.0, + "entropy": 0.33588480949401855, + "epoch": 0.705, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.028353452682495, + "learning_rate": 3e-07, + "loss": -0.07543312013149261, + "num_tokens": 1196672.0, + "reward": 17.10249900817871, + "reward_std": 73.94812774658203, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 17.002500534057617, + "rewards/supergames_reward/std": 73.94813537597656, + "step": 141, + "step_time": 10.684560916008195 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 235.0, + "completions/max_terminated_length": 235.0, + "completions/mean_length": 209.25, + "completions/mean_terminated_length": 209.25, + "completions/min_length": 167.0, + "completions/min_terminated_length": 167.0, + "entropy": 0.3622831702232361, + "epoch": 0.71, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.047222137451172, + "learning_rate": 2.95e-07, + "loss": -0.04569581151008606, + "num_tokens": 1206658.0, + "reward": 49.57624816894531, + "reward_std": 78.21060180664062, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 49.476253509521484, + "rewards/supergames_reward/std": 78.21060943603516, + "step": 142, + "step_time": 8.97679864402744 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 161.0, + "completions/max_terminated_length": 161.0, + "completions/mean_length": 116.25, + "completions/mean_terminated_length": 116.25, + "completions/min_length": 50.0, + "completions/min_terminated_length": 50.0, + "entropy": 0.43277010321617126, + "epoch": 0.715, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.345203876495361, + "learning_rate": 2.9e-07, + "loss": -0.16899466514587402, + "num_tokens": 1212140.0, + "reward": 27.42875099182129, + "reward_std": 48.59138870239258, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 27.328750610351562, + "rewards/supergames_reward/std": 48.591392517089844, + "step": 143, + "step_time": 6.1606669370085 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 205.0, + "completions/max_terminated_length": 205.0, + "completions/mean_length": 173.75, + "completions/mean_terminated_length": 173.75, + "completions/min_length": 134.0, + "completions/min_terminated_length": 134.0, + "entropy": 0.43806371092796326, + "epoch": 0.72, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.858840227127075, + "learning_rate": 2.8499999999999997e-07, + "loss": -0.06502802670001984, + "num_tokens": 1219954.0, + "reward": 38.69249725341797, + "reward_std": 68.78203582763672, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 38.592498779296875, + "rewards/supergames_reward/std": 68.78202819824219, + "step": 144, + "step_time": 7.88639413099736 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 169.0, + "completions/max_terminated_length": 169.0, + "completions/mean_length": 110.625, + "completions/mean_terminated_length": 110.625, + "completions/min_length": 85.0, + "completions/min_terminated_length": 85.0, + "entropy": 0.3383479118347168, + "epoch": 0.725, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.225733757019043, + "learning_rate": 2.8e-07, + "loss": -0.042058318853378296, + "num_tokens": 1225399.0, + "reward": 64.85874938964844, + "reward_std": 25.170387268066406, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 64.75875091552734, + "rewards/supergames_reward/std": 25.170391082763672, + "step": 145, + "step_time": 6.432729878026294 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 147.0, + "completions/max_terminated_length": 147.0, + "completions/mean_length": 129.5, + "completions/mean_terminated_length": 129.5, + "completions/min_length": 93.0, + "completions/min_terminated_length": 93.0, + "entropy": 0.3088987171649933, + "epoch": 0.73, + "frac_reward_zero_std": 1.0, + "grad_norm": 0.0, + "learning_rate": 2.75e-07, + "loss": 0.0, + "num_tokens": 1231043.0, + "reward": 100.0999984741211, + "reward_std": 0.0, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 100.0, + "rewards/supergames_reward/std": 0.0, + "step": 146, + "step_time": 5.634907905012369 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 160.0, + "completions/max_terminated_length": 160.0, + "completions/mean_length": 120.5, + "completions/mean_terminated_length": 120.5, + "completions/min_length": 80.0, + "completions/min_terminated_length": 80.0, + "entropy": 0.37292322516441345, + "epoch": 0.735, + "frac_reward_zero_std": 0.0, + "grad_norm": 5.174577713012695, + "learning_rate": 2.7e-07, + "loss": 0.11778222769498825, + "num_tokens": 1236607.0, + "reward": 34.82374572753906, + "reward_std": 58.513397216796875, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 34.72374725341797, + "rewards/supergames_reward/std": 58.513397216796875, + "step": 147, + "step_time": 6.078690663998714 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 172.0, + "completions/max_terminated_length": 172.0, + "completions/mean_length": 125.625, + "completions/mean_terminated_length": 125.625, + "completions/min_length": 88.0, + "completions/min_terminated_length": 88.0, + "entropy": 0.308654248714447, + "epoch": 0.74, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.8483362197875977, + "learning_rate": 2.65e-07, + "loss": -0.09181805700063705, + "num_tokens": 1242140.0, + "reward": 85.06375122070312, + "reward_std": 42.52893829345703, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 84.9637451171875, + "rewards/supergames_reward/std": 42.52893829345703, + "step": 148, + "step_time": 6.501119122985983 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 275.0, + "completions/max_terminated_length": 275.0, + "completions/mean_length": 211.125, + "completions/mean_terminated_length": 211.125, + "completions/min_length": 157.0, + "completions/min_terminated_length": 157.0, + "entropy": 0.5405827164649963, + "epoch": 0.745, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.774259567260742, + "learning_rate": 2.6e-07, + "loss": -0.03056861087679863, + "num_tokens": 1250205.0, + "reward": 84.63125610351562, + "reward_std": 34.3358154296875, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 84.53125, + "rewards/supergames_reward/std": 34.3358154296875, + "step": 149, + "step_time": 10.352729418984381 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 257.0, + "completions/max_terminated_length": 257.0, + "completions/mean_length": 218.5, + "completions/mean_terminated_length": 218.5, + "completions/min_length": 96.0, + "completions/min_terminated_length": 96.0, + "entropy": 0.3852894604206085, + "epoch": 0.75, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.6103010177612305, + "learning_rate": 2.55e-07, + "loss": 0.034789010882377625, + "num_tokens": 1260249.0, + "reward": 81.81375122070312, + "reward_std": 51.72132873535156, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 81.7137451171875, + "rewards/supergames_reward/std": 51.72132873535156, + "step": 150, + "step_time": 9.937249842012534 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 351.0, + "completions/max_terminated_length": 351.0, + "completions/mean_length": 232.625, + "completions/mean_terminated_length": 232.625, + "completions/min_length": 89.0, + "completions/min_terminated_length": 89.0, + "entropy": 0.4323120415210724, + "epoch": 0.755, + "frac_reward_zero_std": 0.0, + "grad_norm": 2.640308380126953, + "learning_rate": 2.5e-07, + "loss": -0.11181996762752533, + "num_tokens": 1271430.0, + "reward": 56.33625030517578, + "reward_std": 78.97452545166016, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 56.23625183105469, + "rewards/supergames_reward/std": 78.97452545166016, + "step": 151, + "step_time": 13.323547195002902 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 359.0, + "completions/max_terminated_length": 359.0, + "completions/mean_length": 247.5, + "completions/mean_terminated_length": 247.5, + "completions/min_length": 161.0, + "completions/min_terminated_length": 161.0, + "entropy": 0.4038960039615631, + "epoch": 0.76, + "frac_reward_zero_std": 0.0, + "grad_norm": 2.5704824924468994, + "learning_rate": 2.45e-07, + "loss": 0.060525521636009216, + "num_tokens": 1282722.0, + "reward": -68.53875732421875, + "reward_std": 50.590599060058594, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -68.63874816894531, + "rewards/supergames_reward/std": 50.590599060058594, + "step": 152, + "step_time": 13.593127576023107 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 383.0, + "completions/max_terminated_length": 383.0, + "completions/mean_length": 246.375, + "completions/mean_terminated_length": 246.375, + "completions/min_length": 150.0, + "completions/min_terminated_length": 150.0, + "entropy": 0.4110933542251587, + "epoch": 0.765, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.233025312423706, + "learning_rate": 2.4e-07, + "loss": 0.06547188758850098, + "num_tokens": 1294061.0, + "reward": 95.91499328613281, + "reward_std": 11.836968421936035, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 95.81500244140625, + "rewards/supergames_reward/std": 11.836968421936035, + "step": 153, + "step_time": 14.499402174988063 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 338.0, + "completions/max_terminated_length": 338.0, + "completions/mean_length": 233.125, + "completions/mean_terminated_length": 233.125, + "completions/min_length": 88.0, + "completions/min_terminated_length": 88.0, + "entropy": 0.34559860825538635, + "epoch": 0.77, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.6887102127075195, + "learning_rate": 2.3499999999999997e-07, + "loss": 0.15905122458934784, + "num_tokens": 1304262.0, + "reward": 88.92375183105469, + "reward_std": 31.611207962036133, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 88.82374572753906, + "rewards/supergames_reward/std": 31.6112117767334, + "step": 154, + "step_time": 12.704337224975461 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 208.0, + "completions/max_terminated_length": 208.0, + "completions/mean_length": 178.5, + "completions/mean_terminated_length": 178.5, + "completions/min_length": 125.0, + "completions/min_terminated_length": 125.0, + "entropy": 0.4646885097026825, + "epoch": 0.775, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.161937236785889, + "learning_rate": 2.3e-07, + "loss": -0.04354584217071533, + "num_tokens": 1312098.0, + "reward": 27.80875015258789, + "reward_std": 60.81097412109375, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 27.708751678466797, + "rewards/supergames_reward/std": 60.81097412109375, + "step": 155, + "step_time": 7.911955468007363 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 391.0, + "completions/max_terminated_length": 391.0, + "completions/mean_length": 235.375, + "completions/mean_terminated_length": 235.375, + "completions/min_length": 118.0, + "completions/min_terminated_length": 118.0, + "entropy": 0.40985700488090515, + "epoch": 0.78, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.12153959274292, + "learning_rate": 2.25e-07, + "loss": -0.024320494383573532, + "num_tokens": 1322301.0, + "reward": 64.14624786376953, + "reward_std": 49.90530014038086, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 64.04624938964844, + "rewards/supergames_reward/std": 49.90530014038086, + "step": 156, + "step_time": 14.631677505996777 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 330.0, + "completions/max_terminated_length": 330.0, + "completions/mean_length": 220.625, + "completions/mean_terminated_length": 220.625, + "completions/min_length": 93.0, + "completions/min_terminated_length": 93.0, + "entropy": 0.3543764054775238, + "epoch": 0.785, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.0111091136932373, + "learning_rate": 2.1999999999999998e-07, + "loss": -0.13996481895446777, + "num_tokens": 1333386.0, + "reward": 3.6812496185302734, + "reward_std": 103.5096664428711, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": 3.606250762939453, + "rewards/supergames_reward/std": 103.48104858398438, + "step": 157, + "step_time": 12.613216657977318 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 225.0, + "completions/max_terminated_length": 225.0, + "completions/mean_length": 193.125, + "completions/mean_terminated_length": 193.125, + "completions/min_length": 157.0, + "completions/min_terminated_length": 157.0, + "entropy": 0.5682947635650635, + "epoch": 0.79, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.67927885055542, + "learning_rate": 2.1499999999999998e-07, + "loss": 0.0535324327647686, + "num_tokens": 1341355.0, + "reward": -19.78999900817871, + "reward_std": 69.72611236572266, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -19.889997482299805, + "rewards/supergames_reward/std": 69.72611236572266, + "step": 158, + "step_time": 8.585773958999198 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 316.0, + "completions/max_terminated_length": 316.0, + "completions/mean_length": 237.625, + "completions/mean_terminated_length": 237.625, + "completions/min_length": 156.0, + "completions/min_terminated_length": 156.0, + "entropy": 0.3637359142303467, + "epoch": 0.795, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.033409357070923, + "learning_rate": 2.0999999999999997e-07, + "loss": 0.018884683027863503, + "num_tokens": 1352632.0, + "reward": 52.063751220703125, + "reward_std": 65.35551452636719, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 51.96375274658203, + "rewards/supergames_reward/std": 65.35551452636719, + "step": 159, + "step_time": 12.306456914986484 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 225.0, + "completions/max_terminated_length": 225.0, + "completions/mean_length": 184.0, + "completions/mean_terminated_length": 184.0, + "completions/min_length": 143.0, + "completions/min_terminated_length": 143.0, + "entropy": 0.47005388140678406, + "epoch": 0.8, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.585808515548706, + "learning_rate": 2.0499999999999997e-07, + "loss": -0.021751077845692635, + "num_tokens": 1360528.0, + "reward": 33.73124694824219, + "reward_std": 78.67134857177734, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 33.631248474121094, + "rewards/supergames_reward/std": 78.67134857177734, + "step": 160, + "step_time": 8.517202848976012 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 456.0, + "completions/max_terminated_length": 456.0, + "completions/mean_length": 248.75, + "completions/mean_terminated_length": 248.75, + "completions/min_length": 85.0, + "completions/min_terminated_length": 85.0, + "entropy": 0.39807969331741333, + "epoch": 0.805, + "frac_reward_zero_std": 0.0, + "grad_norm": 2.8931398391723633, + "learning_rate": 2e-07, + "loss": -0.05018138885498047, + "num_tokens": 1371854.0, + "reward": 4.002498626708984, + "reward_std": 103.03063201904297, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": 3.9274978637695312, + "rewards/supergames_reward/std": 103.00178527832031, + "step": 161, + "step_time": 17.225145012984285 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 400.0, + "completions/max_terminated_length": 400.0, + "completions/mean_length": 231.125, + "completions/mean_terminated_length": 231.125, + "completions/min_length": 95.0, + "completions/min_terminated_length": 95.0, + "entropy": 0.49391740560531616, + "epoch": 0.81, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.5471365451812744, + "learning_rate": 1.9499999999999999e-07, + "loss": -0.18011468648910522, + "num_tokens": 1383015.0, + "reward": 67.2449951171875, + "reward_std": 69.01223754882812, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 67.14500427246094, + "rewards/supergames_reward/std": 69.01224517822266, + "step": 162, + "step_time": 15.188760302000446 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 262.0, + "completions/max_terminated_length": 262.0, + "completions/mean_length": 221.875, + "completions/mean_terminated_length": 221.875, + "completions/min_length": 89.0, + "completions/min_terminated_length": 89.0, + "entropy": 0.388366162776947, + "epoch": 0.815, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.662609815597534, + "learning_rate": 1.8999999999999998e-07, + "loss": 1.8455075405654497e-05, + "num_tokens": 1393150.0, + "reward": 59.23624801635742, + "reward_std": 76.00599670410156, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 59.13624954223633, + "rewards/supergames_reward/std": 76.00599670410156, + "step": 163, + "step_time": 10.180389262997778 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 396.0, + "completions/max_terminated_length": 396.0, + "completions/mean_length": 323.625, + "completions/mean_terminated_length": 323.625, + "completions/min_length": 256.0, + "completions/min_terminated_length": 256.0, + "entropy": 0.45717528462409973, + "epoch": 0.82, + "frac_reward_zero_std": 0.0, + "grad_norm": 2.580068826675415, + "learning_rate": 1.85e-07, + "loss": 0.024594159796833992, + "num_tokens": 1405067.0, + "reward": 6.318748474121094, + "reward_std": 88.21497344970703, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 6.21875, + "rewards/supergames_reward/std": 88.21497344970703, + "step": 164, + "step_time": 15.074436638999032 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 163.0, + "completions/max_terminated_length": 163.0, + "completions/mean_length": 133.875, + "completions/mean_terminated_length": 133.875, + "completions/min_length": 90.0, + "completions/min_terminated_length": 90.0, + "entropy": 0.3449660539627075, + "epoch": 0.825, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.045867443084717, + "learning_rate": 1.8e-07, + "loss": -0.02517889253795147, + "num_tokens": 1410730.0, + "reward": 84.35499572753906, + "reward_std": 38.135921478271484, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 84.2550048828125, + "rewards/supergames_reward/std": 38.13592529296875, + "step": 165, + "step_time": 6.275273430015659 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 144.0, + "completions/max_terminated_length": 144.0, + "completions/mean_length": 105.875, + "completions/mean_terminated_length": 105.875, + "completions/min_length": 83.0, + "completions/min_terminated_length": 83.0, + "entropy": 0.286411851644516, + "epoch": 0.83, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.855896472930908, + "learning_rate": 1.75e-07, + "loss": 0.014497723430395126, + "num_tokens": 1416081.0, + "reward": 48.40374755859375, + "reward_std": 47.61183547973633, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 48.303749084472656, + "rewards/supergames_reward/std": 47.611839294433594, + "step": 166, + "step_time": 5.5562305669882335 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 269.0, + "completions/max_terminated_length": 269.0, + "completions/mean_length": 226.125, + "completions/mean_terminated_length": 226.125, + "completions/min_length": 146.0, + "completions/min_terminated_length": 146.0, + "entropy": 0.39676305651664734, + "epoch": 0.835, + "frac_reward_zero_std": 0.0, + "grad_norm": 2.633286237716675, + "learning_rate": 1.7000000000000001e-07, + "loss": -0.008210637606680393, + "num_tokens": 1426194.0, + "reward": 67.51000213623047, + "reward_std": 58.9198112487793, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 67.41000366210938, + "rewards/supergames_reward/std": 58.9198112487793, + "step": 167, + "step_time": 10.435288048989605 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 210.0, + "completions/max_terminated_length": 210.0, + "completions/mean_length": 168.5, + "completions/mean_terminated_length": 168.5, + "completions/min_length": 137.0, + "completions/min_terminated_length": 137.0, + "entropy": 0.4233987629413605, + "epoch": 0.84, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.654274940490723, + "learning_rate": 1.65e-07, + "loss": -0.0319184847176075, + "num_tokens": 1433990.0, + "reward": 50.342498779296875, + "reward_std": 67.66273498535156, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": 50.26750183105469, + "rewards/supergames_reward/std": 67.5992202758789, + "step": 168, + "step_time": 8.111304188991198 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 397.0, + "completions/max_terminated_length": 397.0, + "completions/mean_length": 217.75, + "completions/mean_terminated_length": 217.75, + "completions/min_length": 90.0, + "completions/min_terminated_length": 90.0, + "entropy": 0.41511332988739014, + "epoch": 0.845, + "frac_reward_zero_std": 0.0, + "grad_norm": 5.156782150268555, + "learning_rate": 1.6e-07, + "loss": 0.30166593194007874, + "num_tokens": 1444068.0, + "reward": 48.64500045776367, + "reward_std": 51.50307846069336, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 48.54499816894531, + "rewards/supergames_reward/std": 51.50307846069336, + "step": 169, + "step_time": 14.915170635009417 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 245.0, + "completions/max_terminated_length": 245.0, + "completions/mean_length": 218.625, + "completions/mean_terminated_length": 218.625, + "completions/min_length": 140.0, + "completions/min_terminated_length": 140.0, + "entropy": 0.3981763422489166, + "epoch": 0.85, + "frac_reward_zero_std": 0.0, + "grad_norm": 2.818580389022827, + "learning_rate": 1.55e-07, + "loss": -0.08915036171674728, + "num_tokens": 1454153.0, + "reward": 82.69249725341797, + "reward_std": 24.65846824645996, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 82.59249877929688, + "rewards/supergames_reward/std": 24.65846824645996, + "step": 170, + "step_time": 9.446723325003404 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 386.0, + "completions/max_terminated_length": 386.0, + "completions/mean_length": 280.375, + "completions/mean_terminated_length": 280.375, + "completions/min_length": 98.0, + "completions/min_terminated_length": 98.0, + "entropy": 0.39121532440185547, + "epoch": 0.855, + "frac_reward_zero_std": 0.0, + "grad_norm": 2.067988634109497, + "learning_rate": 1.5e-07, + "loss": -0.28749391436576843, + "num_tokens": 1465692.0, + "reward": 66.02375030517578, + "reward_std": 70.814697265625, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 65.92375183105469, + "rewards/supergames_reward/std": 70.814697265625, + "step": 171, + "step_time": 14.603269515006104 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 188.0, + "completions/max_terminated_length": 188.0, + "completions/mean_length": 131.125, + "completions/mean_terminated_length": 131.125, + "completions/min_length": 87.0, + "completions/min_terminated_length": 87.0, + "entropy": 0.435172438621521, + "epoch": 0.86, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.832774639129639, + "learning_rate": 1.45e-07, + "loss": -0.1204412579536438, + "num_tokens": 1471253.0, + "reward": 39.80875015258789, + "reward_std": 51.58554458618164, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 39.7087516784668, + "rewards/supergames_reward/std": 51.585548400878906, + "step": 172, + "step_time": 7.057335219986271 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 391.0, + "completions/max_terminated_length": 391.0, + "completions/mean_length": 271.375, + "completions/mean_terminated_length": 271.375, + "completions/min_length": 176.0, + "completions/min_terminated_length": 176.0, + "entropy": 0.4081321060657501, + "epoch": 0.865, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.2880308628082275, + "learning_rate": 1.4e-07, + "loss": 0.03748488798737526, + "num_tokens": 1482744.0, + "reward": 53.62249755859375, + "reward_std": 75.1306381225586, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 53.522499084472656, + "rewards/supergames_reward/std": 75.1306381225586, + "step": 173, + "step_time": 14.96632127999328 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 304.0, + "completions/max_terminated_length": 304.0, + "completions/mean_length": 203.875, + "completions/mean_terminated_length": 203.875, + "completions/min_length": 140.0, + "completions/min_terminated_length": 140.0, + "entropy": 0.4152996242046356, + "epoch": 0.87, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.1565890312194824, + "learning_rate": 1.35e-07, + "loss": -0.047037553042173386, + "num_tokens": 1492679.0, + "reward": 37.48374938964844, + "reward_std": 72.71965789794922, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 37.383750915527344, + "rewards/supergames_reward/std": 72.71966552734375, + "step": 174, + "step_time": 11.566132662002929 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 187.0, + "completions/max_terminated_length": 187.0, + "completions/mean_length": 133.375, + "completions/mean_terminated_length": 133.375, + "completions/min_length": 95.0, + "completions/min_terminated_length": 95.0, + "entropy": 0.3746980130672455, + "epoch": 0.875, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.724977970123291, + "learning_rate": 1.3e-07, + "loss": 0.05197479575872421, + "num_tokens": 1498330.0, + "reward": 37.86625289916992, + "reward_std": 52.43158721923828, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 37.76625061035156, + "rewards/supergames_reward/std": 52.43158721923828, + "step": 175, + "step_time": 7.080777855007909 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 448.0, + "completions/max_terminated_length": 448.0, + "completions/mean_length": 277.25, + "completions/mean_terminated_length": 277.25, + "completions/min_length": 96.0, + "completions/min_terminated_length": 96.0, + "entropy": 0.397225022315979, + "epoch": 0.88, + "frac_reward_zero_std": 0.0, + "grad_norm": 2.438502788543701, + "learning_rate": 1.25e-07, + "loss": 0.1022070124745369, + "num_tokens": 1509860.0, + "reward": -41.323753356933594, + "reward_std": 79.704345703125, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -41.42374801635742, + "rewards/supergames_reward/std": 79.704345703125, + "step": 176, + "step_time": 16.852883081999607 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 187.0, + "completions/max_terminated_length": 187.0, + "completions/mean_length": 145.875, + "completions/mean_terminated_length": 145.875, + "completions/min_length": 92.0, + "completions/min_terminated_length": 92.0, + "entropy": 0.34882763028144836, + "epoch": 0.885, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.683617115020752, + "learning_rate": 1.2e-07, + "loss": 0.0609549917280674, + "num_tokens": 1515587.0, + "reward": 84.27999877929688, + "reward_std": 38.06671905517578, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 84.17999267578125, + "rewards/supergames_reward/std": 38.06672286987305, + "step": 177, + "step_time": 7.09286071601673 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 217.0, + "completions/max_terminated_length": 217.0, + "completions/mean_length": 181.875, + "completions/mean_terminated_length": 181.875, + "completions/min_length": 146.0, + "completions/min_terminated_length": 146.0, + "entropy": 0.423494815826416, + "epoch": 0.89, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.3993117809295654, + "learning_rate": 1.15e-07, + "loss": 0.04823008552193642, + "num_tokens": 1523490.0, + "reward": 22.46124839782715, + "reward_std": 89.67146301269531, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 22.361249923706055, + "rewards/supergames_reward/std": 89.67146301269531, + "step": 178, + "step_time": 8.281584940006724 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 315.0, + "completions/max_terminated_length": 315.0, + "completions/mean_length": 273.25, + "completions/mean_terminated_length": 273.25, + "completions/min_length": 234.0, + "completions/min_terminated_length": 234.0, + "entropy": 0.3831624984741211, + "epoch": 0.895, + "frac_reward_zero_std": 0.0, + "grad_norm": 2.7383997440338135, + "learning_rate": 1.0999999999999999e-07, + "loss": -0.010000350885093212, + "num_tokens": 1535012.0, + "reward": 2.237499237060547, + "reward_std": 93.32353210449219, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 2.137500762939453, + "rewards/supergames_reward/std": 93.32353973388672, + "step": 179, + "step_time": 12.098909514024854 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 395.0, + "completions/max_terminated_length": 395.0, + "completions/mean_length": 197.375, + "completions/mean_terminated_length": 197.375, + "completions/min_length": 132.0, + "completions/min_terminated_length": 132.0, + "entropy": 0.5201653242111206, + "epoch": 0.9, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.4596073627471924, + "learning_rate": 1.0499999999999999e-07, + "loss": -0.11710439622402191, + "num_tokens": 1543023.0, + "reward": 88.1875, + "reward_std": 33.69363784790039, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 88.0875015258789, + "rewards/supergames_reward/std": 33.693641662597656, + "step": 180, + "step_time": 14.662479719001567 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 351.0, + "completions/max_terminated_length": 351.0, + "completions/mean_length": 218.75, + "completions/mean_terminated_length": 218.75, + "completions/min_length": 159.0, + "completions/min_terminated_length": 159.0, + "entropy": 0.4402031898498535, + "epoch": 0.905, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.2664496898651123, + "learning_rate": 1e-07, + "loss": -0.038904689252376556, + "num_tokens": 1554085.0, + "reward": 71.10874938964844, + "reward_std": 69.99549102783203, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 71.00875091552734, + "rewards/supergames_reward/std": 69.99549865722656, + "step": 181, + "step_time": 13.245673589000944 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 328.0, + "completions/max_terminated_length": 328.0, + "completions/mean_length": 263.375, + "completions/mean_terminated_length": 263.375, + "completions/min_length": 182.0, + "completions/min_terminated_length": 182.0, + "entropy": 0.37136757373809814, + "epoch": 0.91, + "frac_reward_zero_std": 0.0, + "grad_norm": 2.946763277053833, + "learning_rate": 9.499999999999999e-08, + "loss": 0.010829905979335308, + "num_tokens": 1565512.0, + "reward": 52.63500213623047, + "reward_std": 73.81818389892578, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 52.534996032714844, + "rewards/supergames_reward/std": 73.81818389892578, + "step": 182, + "step_time": 12.547108427999774 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 333.0, + "completions/max_terminated_length": 333.0, + "completions/mean_length": 263.0, + "completions/mean_terminated_length": 263.0, + "completions/min_length": 177.0, + "completions/min_terminated_length": 177.0, + "entropy": 0.3846859633922577, + "epoch": 0.915, + "frac_reward_zero_std": 0.0, + "grad_norm": 2.7999610900878906, + "learning_rate": 9e-08, + "loss": 0.004038939252495766, + "num_tokens": 1576960.0, + "reward": 71.89500427246094, + "reward_std": 45.18279266357422, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 71.79499816894531, + "rewards/supergames_reward/std": 45.182796478271484, + "step": 183, + "step_time": 12.604182851006044 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 268.0, + "completions/max_terminated_length": 268.0, + "completions/mean_length": 216.375, + "completions/mean_terminated_length": 216.375, + "completions/min_length": 138.0, + "completions/min_terminated_length": 138.0, + "entropy": 0.36285799741744995, + "epoch": 0.92, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.070614814758301, + "learning_rate": 8.500000000000001e-08, + "loss": 0.08384927362203598, + "num_tokens": 1587019.0, + "reward": -29.672500610351562, + "reward_std": 94.68096923828125, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -29.772499084472656, + "rewards/supergames_reward/std": 94.68096923828125, + "step": 184, + "step_time": 10.202588220010512 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 314.0, + "completions/max_terminated_length": 314.0, + "completions/mean_length": 239.375, + "completions/mean_terminated_length": 239.375, + "completions/min_length": 173.0, + "completions/min_terminated_length": 173.0, + "entropy": 0.43123888969421387, + "epoch": 0.925, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.0129008293151855, + "learning_rate": 8e-08, + "loss": 0.03122909925878048, + "num_tokens": 1597230.0, + "reward": 10.326248168945312, + "reward_std": 66.74771881103516, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 10.226249694824219, + "rewards/supergames_reward/std": 66.74771881103516, + "step": 185, + "step_time": 11.8433545169828 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 266.0, + "completions/max_terminated_length": 266.0, + "completions/mean_length": 219.75, + "completions/mean_terminated_length": 219.75, + "completions/min_length": 147.0, + "completions/min_terminated_length": 147.0, + "entropy": 0.39456790685653687, + "epoch": 0.93, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.786937952041626, + "learning_rate": 7.5e-08, + "loss": 0.01745336316525936, + "num_tokens": 1607300.0, + "reward": 61.45499801635742, + "reward_std": 59.6201171875, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 61.35499954223633, + "rewards/supergames_reward/std": 59.620121002197266, + "step": 186, + "step_time": 10.153366808983264 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 221.0, + "completions/max_terminated_length": 221.0, + "completions/mean_length": 158.5, + "completions/mean_terminated_length": 158.5, + "completions/min_length": 121.0, + "completions/min_terminated_length": 121.0, + "entropy": 0.399124413728714, + "epoch": 0.935, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.657733917236328, + "learning_rate": 7e-08, + "loss": -0.05617351084947586, + "num_tokens": 1615000.0, + "reward": -12.976250648498535, + "reward_std": 83.94307708740234, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": -13.051249504089355, + "rewards/supergames_reward/std": 83.9134521484375, + "step": 187, + "step_time": 8.401661271986086 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 190.0, + "completions/max_terminated_length": 190.0, + "completions/mean_length": 131.25, + "completions/mean_terminated_length": 131.25, + "completions/min_length": 92.0, + "completions/min_terminated_length": 92.0, + "entropy": 0.38221123814582825, + "epoch": 0.94, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.81663703918457, + "learning_rate": 6.5e-08, + "loss": 0.1486133486032486, + "num_tokens": 1620586.0, + "reward": 32.564998626708984, + "reward_std": 68.18844604492188, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 32.46500015258789, + "rewards/supergames_reward/std": 68.18844604492188, + "step": 188, + "step_time": 7.095077903999481 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 369.0, + "completions/max_terminated_length": 369.0, + "completions/mean_length": 273.75, + "completions/mean_terminated_length": 273.75, + "completions/min_length": 189.0, + "completions/min_terminated_length": 189.0, + "entropy": 0.4164618253707886, + "epoch": 0.945, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.0384137630462646, + "learning_rate": 6e-08, + "loss": -0.013533106073737144, + "num_tokens": 1632056.0, + "reward": 28.412498474121094, + "reward_std": 99.22362518310547, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 28.3125, + "rewards/supergames_reward/std": 99.22362518310547, + "step": 189, + "step_time": 13.92420879600104 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 146.0, + "completions/max_terminated_length": 146.0, + "completions/mean_length": 116.25, + "completions/mean_terminated_length": 116.25, + "completions/min_length": 88.0, + "completions/min_terminated_length": 88.0, + "entropy": 0.2739017605781555, + "epoch": 0.95, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.924455165863037, + "learning_rate": 5.4999999999999996e-08, + "loss": -0.10890161991119385, + "num_tokens": 1637538.0, + "reward": 57.61750030517578, + "reward_std": 71.02250671386719, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 57.51750183105469, + "rewards/supergames_reward/std": 71.02251434326172, + "step": 190, + "step_time": 5.542224484990584 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 351.0, + "completions/max_terminated_length": 351.0, + "completions/mean_length": 265.375, + "completions/mean_terminated_length": 265.375, + "completions/min_length": 133.0, + "completions/min_terminated_length": 133.0, + "entropy": 0.38974297046661377, + "epoch": 0.955, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.5310401916503906, + "learning_rate": 5e-08, + "loss": 0.10060633718967438, + "num_tokens": 1649029.0, + "reward": 68.22624206542969, + "reward_std": 69.13253021240234, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 68.12625122070312, + "rewards/supergames_reward/std": 69.13253021240234, + "step": 191, + "step_time": 13.248786654992728 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 309.0, + "completions/max_terminated_length": 309.0, + "completions/mean_length": 228.25, + "completions/mean_terminated_length": 228.25, + "completions/min_length": 153.0, + "completions/min_terminated_length": 153.0, + "entropy": 0.4095747470855713, + "epoch": 0.96, + "frac_reward_zero_std": 0.0, + "grad_norm": 2.988938331604004, + "learning_rate": 4.5e-08, + "loss": 0.012724814936518669, + "num_tokens": 1659151.0, + "reward": 22.46999740600586, + "reward_std": 71.24070739746094, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 22.369998931884766, + "rewards/supergames_reward/std": 71.2406997680664, + "step": 192, + "step_time": 11.694355098996311 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 332.0, + "completions/max_terminated_length": 332.0, + "completions/mean_length": 183.375, + "completions/mean_terminated_length": 183.375, + "completions/min_length": 129.0, + "completions/min_terminated_length": 129.0, + "entropy": 0.3741522431373596, + "epoch": 0.965, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.120519161224365, + "learning_rate": 4e-08, + "loss": 0.18378837406635284, + "num_tokens": 1667042.0, + "reward": 18.967496871948242, + "reward_std": 91.04219818115234, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 18.86750030517578, + "rewards/supergames_reward/std": 91.04219818115234, + "step": 193, + "step_time": 12.186350238014711 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.125, + "completions/max_length": 512.0, + "completions/max_terminated_length": 287.0, + "completions/mean_length": 275.125, + "completions/mean_terminated_length": 241.2857208251953, + "completions/min_length": 193.0, + "completions/min_terminated_length": 193.0, + "entropy": 0.31828516721725464, + "epoch": 0.97, + "frac_reward_zero_std": 0.0, + "grad_norm": 2.3239247798919678, + "learning_rate": 3.5e-08, + "loss": 0.2439030557870865, + "num_tokens": 1677579.0, + "reward": 22.55124855041504, + "reward_std": 74.4148941040039, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": 22.47624969482422, + "rewards/supergames_reward/std": 74.3678207397461, + "step": 194, + "step_time": 18.785672573983902 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 183.0, + "completions/max_terminated_length": 183.0, + "completions/mean_length": 162.625, + "completions/mean_terminated_length": 162.625, + "completions/min_length": 140.0, + "completions/min_terminated_length": 140.0, + "entropy": 0.37672579288482666, + "epoch": 0.975, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.524571657180786, + "learning_rate": 3e-08, + "loss": -0.002952136332169175, + "num_tokens": 1685304.0, + "reward": 17.362499237060547, + "reward_std": 67.76500701904297, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 17.262500762939453, + "rewards/supergames_reward/std": 67.76500701904297, + "step": 195, + "step_time": 7.0316317139950115 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 391.0, + "completions/max_terminated_length": 391.0, + "completions/mean_length": 290.0, + "completions/mean_terminated_length": 290.0, + "completions/min_length": 144.0, + "completions/min_terminated_length": 144.0, + "entropy": 0.3798663914203644, + "epoch": 0.98, + "frac_reward_zero_std": 1.0, + "grad_norm": 0.0, + "learning_rate": 2.5e-08, + "loss": 0.0, + "num_tokens": 1696944.0, + "reward": 100.0999984741211, + "reward_std": 0.0, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 100.0, + "rewards/supergames_reward/std": 0.0, + "step": 196, + "step_time": 14.659201105998363 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 307.0, + "completions/max_terminated_length": 307.0, + "completions/mean_length": 230.375, + "completions/mean_terminated_length": 230.375, + "completions/min_length": 166.0, + "completions/min_terminated_length": 166.0, + "entropy": 0.3792188763618469, + "epoch": 0.985, + "frac_reward_zero_std": 0.0, + "grad_norm": 2.9155354499816895, + "learning_rate": 2e-08, + "loss": 0.08655659109354019, + "num_tokens": 1707123.0, + "reward": 80.99250030517578, + "reward_std": 47.69900131225586, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 80.89250183105469, + "rewards/supergames_reward/std": 47.69900131225586, + "step": 197, + "step_time": 11.552438066981267 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 420.0, + "completions/max_terminated_length": 420.0, + "completions/mean_length": 268.0, + "completions/mean_terminated_length": 268.0, + "completions/min_length": 171.0, + "completions/min_terminated_length": 171.0, + "entropy": 0.5007117986679077, + "epoch": 0.99, + "frac_reward_zero_std": 0.0, + "grad_norm": 2.9498138427734375, + "learning_rate": 1.5e-08, + "loss": 0.0345655158162117, + "num_tokens": 1718579.0, + "reward": 51.65749740600586, + "reward_std": 83.32228088378906, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 51.557498931884766, + "rewards/supergames_reward/std": 83.3222885131836, + "step": 198, + "step_time": 15.654064546979498 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 157.0, + "completions/max_terminated_length": 157.0, + "completions/mean_length": 120.375, + "completions/mean_terminated_length": 120.375, + "completions/min_length": 82.0, + "completions/min_terminated_length": 82.0, + "entropy": 0.3440852761268616, + "epoch": 0.995, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.353044509887695, + "learning_rate": 1e-08, + "loss": -0.08407377451658249, + "num_tokens": 1724094.0, + "reward": 77.20124816894531, + "reward_std": 9.252492904663086, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 77.10124969482422, + "rewards/supergames_reward/std": 9.25249195098877, + "step": 199, + "step_time": 5.929525154992007 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 278.0, + "completions/max_terminated_length": 278.0, + "completions/mean_length": 182.5, + "completions/mean_terminated_length": 182.5, + "completions/min_length": 118.0, + "completions/min_terminated_length": 118.0, + "entropy": 0.46504852175712585, + "epoch": 1.0, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.5138256549835205, + "learning_rate": 5e-09, + "loss": -0.025602273643016815, + "num_tokens": 1731978.0, + "reward": 24.10624885559082, + "reward_std": 56.38955307006836, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 24.006248474121094, + "rewards/supergames_reward/std": 56.38955307006836, + "step": 200, + "step_time": 10.38404856598936 + } + ], + "logging_steps": 1, + "max_steps": 200, + "num_input_tokens_seen": 1731978, + "num_train_epochs": 1, + "save_steps": 20, + "stateful_callbacks": { + "TrainerControl": { + "args": { + "should_epoch_stop": false, + "should_evaluate": false, + "should_log": false, + "should_save": true, + "should_training_stop": true + }, + "attributes": {} + } + }, + "total_flos": 0.0, + "train_batch_size": 8, + "trial_name": null, + "trial_params": null +} diff --git a/checkpoint-200/training_args.bin b/checkpoint-200/training_args.bin new file mode 100644 index 0000000..730dec0 --- /dev/null +++ b/checkpoint-200/training_args.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:691fe5be2d7d1d56160dae63866d136877a72efb30e15ff7b3249192f998a788 +size 7185 diff --git a/checkpoint-40/chat_template.jinja b/checkpoint-40/chat_template.jinja new file mode 100644 index 0000000..bdf7919 --- /dev/null +++ b/checkpoint-40/chat_template.jinja @@ -0,0 +1,54 @@ +{%- if tools %} + {{- '<|im_start|>system\n' }} + {%- if messages[0]['role'] == 'system' %} + {{- messages[0]['content'] }} + {%- else %} + {{- 'You are Qwen, created by Alibaba Cloud. You are a helpful assistant.' }} + {%- endif %} + {{- "\n\n# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }} + {%- for tool in tools %} + {{- "\n" }} + {{- tool | tojson }} + {%- endfor %} + {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }} +{%- else %} + {%- if messages[0]['role'] == 'system' %} + {{- '<|im_start|>system\n' + messages[0]['content'] + '<|im_end|>\n' }} + {%- else %} + {{- '<|im_start|>system\nYou are Qwen, created by Alibaba Cloud. You are a helpful assistant.<|im_end|>\n' }} + {%- endif %} +{%- endif %} +{%- for message in messages %} + {%- if (message.role == "user") or (message.role == "system" and not loop.first) or (message.role == "assistant" and not message.tool_calls) %} + {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }} + {%- elif message.role == "assistant" %} + {{- '<|im_start|>' + message.role }} + {%- if message.content %} + {{- '\n' + message.content }} + {%- endif %} + {%- for tool_call in message.tool_calls %} + {%- if tool_call.function is defined %} + {%- set tool_call = tool_call.function %} + {%- endif %} + {{- '\n\n{"name": "' }} + {{- tool_call.name }} + {{- '", "arguments": ' }} + {{- tool_call.arguments | tojson }} + {{- '}\n' }} + {%- endfor %} + {{- '<|im_end|>\n' }} + {%- elif message.role == "tool" %} + {%- if (loop.index0 == 0) or (messages[loop.index0 - 1].role != "tool") %} + {{- '<|im_start|>user' }} + {%- endif %} + {{- '\n\n' }} + {{- message.content }} + {{- '\n' }} + {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %} + {{- '<|im_end|>\n' }} + {%- endif %} + {%- endif %} +{%- endfor %} +{%- if add_generation_prompt %} + {{- '<|im_start|>assistant\n' }} +{%- endif %} diff --git a/checkpoint-40/config.json b/checkpoint-40/config.json new file mode 100644 index 0000000..f57b09b --- /dev/null +++ b/checkpoint-40/config.json @@ -0,0 +1,61 @@ +{ + "architectures": [ + "Qwen2ForCausalLM" + ], + "attention_dropout": 0.0, + "bos_token_id": null, + "dtype": "float32", + "eos_token_id": 151645, + "hidden_act": "silu", + "hidden_size": 1536, + "initializer_range": 0.02, + "intermediate_size": 8960, + "layer_types": [ + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention" + ], + "max_position_embeddings": 32768, + "max_window_layers": 21, + "model_type": "qwen2", + "num_attention_heads": 12, + "num_hidden_layers": 28, + "num_key_value_heads": 2, + "pad_token_id": 151643, + "rms_norm_eps": 1e-06, + "rope_parameters": { + "rope_theta": 1000000.0, + "rope_type": "default" + }, + "sliding_window": null, + "tie_word_embeddings": true, + "transformers_version": "5.6.2", + "use_cache": false, + "use_sliding_window": false, + "vocab_size": 151936 +} diff --git a/checkpoint-40/generation_config.json b/checkpoint-40/generation_config.json new file mode 100644 index 0000000..1b2bba9 --- /dev/null +++ b/checkpoint-40/generation_config.json @@ -0,0 +1,13 @@ +{ + "do_sample": true, + "eos_token_id": [ + 151645, + 151643 + ], + "pad_token_id": 151643, + "repetition_penalty": 1.1, + "temperature": 0.7, + "top_k": 20, + "top_p": 0.8, + "transformers_version": "5.6.2" +} diff --git a/checkpoint-40/model.safetensors b/checkpoint-40/model.safetensors new file mode 100644 index 0000000..e8583b3 --- /dev/null +++ b/checkpoint-40/model.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:ff9808a347f7c1f6723676d9b64c93e8fb373f523e804c982e79f9232ad6dff9 +size 6174895536 diff --git a/checkpoint-40/optimizer.pt b/checkpoint-40/optimizer.pt new file mode 100644 index 0000000..d87ec02 --- /dev/null +++ b/checkpoint-40/optimizer.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:e3d9bf1d973ca75b6888d27fd8f016d457eb9428e6ae62d0e481aca478fb2502 +size 12350013801 diff --git a/checkpoint-40/rng_state.pth b/checkpoint-40/rng_state.pth new file mode 100644 index 0000000..b1f5e0b --- /dev/null +++ b/checkpoint-40/rng_state.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:0cd4224f54462732698d27ac3a85f81b4b84f7f8f6f508de186eac62c4e74852 +size 14645 diff --git a/checkpoint-40/scheduler.pt b/checkpoint-40/scheduler.pt new file mode 100644 index 0000000..a3822b0 --- /dev/null +++ b/checkpoint-40/scheduler.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:d12172e6d50575e5620355d22f9f62e7d87de349d6a965fe2f8a67e390b9f8cf +size 1465 diff --git a/checkpoint-40/tokenizer.json b/checkpoint-40/tokenizer.json new file mode 100644 index 0000000..34510ff --- /dev/null +++ b/checkpoint-40/tokenizer.json @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:3fd169731d2cbde95e10bf356d66d5997fd885dd8dbb6fb4684da3f23b2585d8 +size 11421892 diff --git a/checkpoint-40/tokenizer_config.json b/checkpoint-40/tokenizer_config.json new file mode 100644 index 0000000..770e41d --- /dev/null +++ b/checkpoint-40/tokenizer_config.json @@ -0,0 +1,30 @@ +{ + "add_prefix_space": false, + "backend": "tokenizers", + "bos_token": null, + "clean_up_tokenization_spaces": false, + "eos_token": "<|im_end|>", + "errors": "replace", + "extra_special_tokens": [ + "<|im_start|>", + "<|im_end|>", + "<|object_ref_start|>", + "<|object_ref_end|>", + "<|box_start|>", + "<|box_end|>", + "<|quad_start|>", + "<|quad_end|>", + "<|vision_start|>", + "<|vision_end|>", + "<|vision_pad|>", + "<|image_pad|>", + "<|video_pad|>" + ], + "is_local": false, + "local_files_only": false, + "model_max_length": 131072, + "pad_token": "<|endoftext|>", + "split_special_tokens": false, + "tokenizer_class": "Qwen2Tokenizer", + "unk_token": null +} diff --git a/checkpoint-40/trainer_state.json b/checkpoint-40/trainer_state.json new file mode 100644 index 0000000..b1f562a --- /dev/null +++ b/checkpoint-40/trainer_state.json @@ -0,0 +1,1194 @@ +{ + "best_global_step": null, + "best_metric": null, + "best_model_checkpoint": null, + "epoch": 0.2, + "eval_steps": 500, + "global_step": 40, + "is_hyper_param_search": false, + "is_local_process_zero": true, + "is_world_process_zero": true, + "log_history": [ + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 478.0, + "completions/max_terminated_length": 478.0, + "completions/mean_length": 191.0, + "completions/mean_terminated_length": 191.0, + "completions/min_length": 66.0, + "completions/min_terminated_length": 66.0, + "entropy": 0.5566893219947815, + "epoch": 0.005, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.370361328125, + "learning_rate": 1e-06, + "loss": -0.24658073484897614, + "num_tokens": 10832.0, + "reward": -3.3000030517578125, + "reward_std": 85.62333679199219, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -3.4000015258789062, + "rewards/supergames_reward/std": 85.62333679199219, + "step": 1, + "step_time": 22.779711072013015 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 216.0, + "completions/max_terminated_length": 216.0, + "completions/mean_length": 131.5, + "completions/mean_terminated_length": 131.5, + "completions/min_length": 60.0, + "completions/min_terminated_length": 60.0, + "entropy": 0.6395841240882874, + "epoch": 0.01, + "frac_reward_zero_std": 0.0, + "grad_norm": 8.09774398803711, + "learning_rate": 9.95e-07, + "loss": 0.20567649602890015, + "num_tokens": 16404.0, + "reward": -12.422499656677246, + "reward_std": 7.134707450866699, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -12.522500038146973, + "rewards/supergames_reward/std": 7.134707927703857, + "step": 2, + "step_time": 9.960156448010821 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 287.0, + "completions/max_terminated_length": 287.0, + "completions/mean_length": 156.625, + "completions/mean_terminated_length": 156.625, + "completions/min_length": 77.0, + "completions/min_terminated_length": 77.0, + "entropy": 0.562222421169281, + "epoch": 0.015, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.9053616523742676, + "learning_rate": 9.9e-07, + "loss": 0.09602774679660797, + "num_tokens": 26953.0, + "reward": -74.94999694824219, + "reward_std": 70.73094177246094, + "rewards/json_format_reward/mean": 0.05000000074505806, + "rewards/json_format_reward/std": 0.09258200973272324, + "rewards/supergames_reward/mean": -75.0, + "rewards/supergames_reward/std": 70.71067810058594, + "step": 3, + "step_time": 13.686320498993155 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 268.0, + "completions/max_terminated_length": 268.0, + "completions/mean_length": 154.125, + "completions/mean_terminated_length": 154.125, + "completions/min_length": 53.0, + "completions/min_terminated_length": 53.0, + "entropy": 0.7323317527770996, + "epoch": 0.02, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.933310031890869, + "learning_rate": 9.849999999999999e-07, + "loss": 0.28249427676200867, + "num_tokens": 36514.0, + "reward": -40.7599983215332, + "reward_std": 81.73140716552734, + "rewards/json_format_reward/mean": 0.05000000074505806, + "rewards/json_format_reward/std": 0.09258200973272324, + "rewards/supergames_reward/mean": -40.80999755859375, + "rewards/supergames_reward/std": 81.68995666503906, + "step": 4, + "step_time": 12.805880262021674 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 401.0, + "completions/max_terminated_length": 401.0, + "completions/mean_length": 205.25, + "completions/mean_terminated_length": 205.25, + "completions/min_length": 115.0, + "completions/min_terminated_length": 115.0, + "entropy": 0.6167430281639099, + "epoch": 0.025, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.344135046005249, + "learning_rate": 9.8e-07, + "loss": 0.1433994024991989, + "num_tokens": 47476.0, + "reward": -97.6612548828125, + "reward_std": 6.3321428298950195, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -97.76124572753906, + "rewards/supergames_reward/std": 6.3321428298950195, + "step": 5, + "step_time": 18.789364666008623 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 276.0, + "completions/max_terminated_length": 276.0, + "completions/mean_length": 167.25, + "completions/mean_terminated_length": 167.25, + "completions/min_length": 73.0, + "completions/min_terminated_length": 73.0, + "entropy": 0.7583790421485901, + "epoch": 0.03, + "frac_reward_zero_std": 0.0, + "grad_norm": 5.172371864318848, + "learning_rate": 9.75e-07, + "loss": -0.12127404659986496, + "num_tokens": 55246.0, + "reward": -58.687503814697266, + "reward_std": 58.64059829711914, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": -58.76250457763672, + "rewards/supergames_reward/std": 58.620460510253906, + "step": 6, + "step_time": 12.948570273991209 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 261.0, + "completions/max_terminated_length": 261.0, + "completions/mean_length": 133.875, + "completions/mean_terminated_length": 133.875, + "completions/min_length": 57.0, + "completions/min_terminated_length": 57.0, + "entropy": 0.7447654604911804, + "epoch": 0.035, + "frac_reward_zero_std": 0.0, + "grad_norm": 6.428595066070557, + "learning_rate": 9.7e-07, + "loss": 0.10135584324598312, + "num_tokens": 60837.0, + "reward": -32.45000076293945, + "reward_std": 41.754486083984375, + "rewards/json_format_reward/mean": 0.05000000074505806, + "rewards/json_format_reward/std": 0.09258200973272324, + "rewards/supergames_reward/mean": -32.5, + "rewards/supergames_reward/std": 41.66190719604492, + "step": 7, + "step_time": 12.058315072004916 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 414.0, + "completions/max_terminated_length": 414.0, + "completions/mean_length": 209.125, + "completions/mean_terminated_length": 209.125, + "completions/min_length": 104.0, + "completions/min_terminated_length": 104.0, + "entropy": 0.3860666751861572, + "epoch": 0.04, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.806255578994751, + "learning_rate": 9.649999999999999e-07, + "loss": -0.33238139748573303, + "num_tokens": 70838.0, + "reward": -80.50375366210938, + "reward_std": 38.82380294799805, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -80.60375213623047, + "rewards/supergames_reward/std": 38.82379913330078, + "step": 8, + "step_time": 19.323370319994865 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.25, + "completions/max_length": 512.0, + "completions/max_terminated_length": 199.0, + "completions/mean_length": 247.125, + "completions/mean_terminated_length": 158.83334350585938, + "completions/min_length": 106.0, + "completions/min_terminated_length": 106.0, + "entropy": 0.5816237330436707, + "epoch": 0.045, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.3209943771362305, + "learning_rate": 9.6e-07, + "loss": 0.36752766370773315, + "num_tokens": 82135.0, + "reward": -27.688751220703125, + "reward_std": 82.8358154296875, + "rewards/json_format_reward/mean": 0.05000000074505806, + "rewards/json_format_reward/std": 0.09258200973272324, + "rewards/supergames_reward/mean": -27.738750457763672, + "rewards/supergames_reward/std": 82.78590393066406, + "step": 9, + "step_time": 21.948575104994234 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 240.0, + "completions/max_terminated_length": 240.0, + "completions/mean_length": 153.0, + "completions/mean_terminated_length": 153.0, + "completions/min_length": 51.0, + "completions/min_terminated_length": 51.0, + "entropy": 0.6301521062850952, + "epoch": 0.05, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.125359058380127, + "learning_rate": 9.55e-07, + "loss": 0.06129350885748863, + "num_tokens": 89807.0, + "reward": -78.39624786376953, + "reward_std": 40.17042922973633, + "rewards/json_format_reward/mean": 0.05000000074505806, + "rewards/json_format_reward/std": 0.09258200973272324, + "rewards/supergames_reward/mean": -78.44625091552734, + "rewards/supergames_reward/std": 40.1396484375, + "step": 10, + "step_time": 9.098202993016457 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 133.0, + "completions/max_terminated_length": 133.0, + "completions/mean_length": 112.0, + "completions/mean_terminated_length": 112.0, + "completions/min_length": 68.0, + "completions/min_terminated_length": 68.0, + "entropy": 0.5521990656852722, + "epoch": 0.055, + "frac_reward_zero_std": 0.0, + "grad_norm": 6.616060256958008, + "learning_rate": 9.499999999999999e-07, + "loss": -0.006313305348157883, + "num_tokens": 95231.0, + "reward": -12.161249160766602, + "reward_std": 6.39578104019165, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -12.261249542236328, + "rewards/supergames_reward/std": 6.39578104019165, + "step": 11, + "step_time": 5.136311663984088 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 396.0, + "completions/max_terminated_length": 396.0, + "completions/mean_length": 214.875, + "completions/mean_terminated_length": 214.875, + "completions/min_length": 55.0, + "completions/min_terminated_length": 55.0, + "entropy": 0.5505905151367188, + "epoch": 0.06, + "frac_reward_zero_std": 0.0, + "grad_norm": 5.289583206176758, + "learning_rate": 9.45e-07, + "loss": -0.29802003502845764, + "num_tokens": 106238.0, + "reward": -91.66999816894531, + "reward_std": 23.2779541015625, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -91.77000427246094, + "rewards/supergames_reward/std": 23.277956008911133, + "step": 12, + "step_time": 14.945365622988902 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 269.0, + "completions/max_terminated_length": 269.0, + "completions/mean_length": 167.375, + "completions/mean_terminated_length": 167.375, + "completions/min_length": 113.0, + "completions/min_terminated_length": 113.0, + "entropy": 0.540415346622467, + "epoch": 0.065, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.858291149139404, + "learning_rate": 9.399999999999999e-07, + "loss": 0.24004128575325012, + "num_tokens": 116937.0, + "reward": -20.38249969482422, + "reward_std": 87.55204010009766, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": -20.457500457763672, + "rewards/supergames_reward/std": 87.52605438232422, + "step": 13, + "step_time": 10.482285185018554 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 254.0, + "completions/max_terminated_length": 254.0, + "completions/mean_length": 156.875, + "completions/mean_terminated_length": 156.875, + "completions/min_length": 114.0, + "completions/min_terminated_length": 114.0, + "entropy": 0.7237679958343506, + "epoch": 0.07, + "frac_reward_zero_std": 0.0, + "grad_norm": 5.083980560302734, + "learning_rate": 9.35e-07, + "loss": -0.19025824964046478, + "num_tokens": 124640.0, + "reward": -51.397499084472656, + "reward_std": 54.486454010009766, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -51.49749755859375, + "rewards/supergames_reward/std": 54.486454010009766, + "step": 14, + "step_time": 9.597231683001155 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 236.0, + "completions/max_terminated_length": 236.0, + "completions/mean_length": 141.375, + "completions/mean_terminated_length": 141.375, + "completions/min_length": 91.0, + "completions/min_terminated_length": 91.0, + "entropy": 0.754334568977356, + "epoch": 0.075, + "frac_reward_zero_std": 0.0, + "grad_norm": 5.720428943634033, + "learning_rate": 9.3e-07, + "loss": -0.16830675303936005, + "num_tokens": 132219.0, + "reward": -56.226253509521484, + "reward_std": 60.115657806396484, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -56.32625198364258, + "rewards/supergames_reward/std": 60.11566162109375, + "step": 15, + "step_time": 8.998362872982398 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 209.0, + "completions/max_terminated_length": 209.0, + "completions/mean_length": 136.375, + "completions/mean_terminated_length": 136.375, + "completions/min_length": 97.0, + "completions/min_terminated_length": 97.0, + "entropy": 0.6280568838119507, + "epoch": 0.08, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.9681806564331055, + "learning_rate": 9.25e-07, + "loss": -0.01910916529595852, + "num_tokens": 139726.0, + "reward": -68.24000549316406, + "reward_std": 52.238914489746094, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -68.33999633789062, + "rewards/supergames_reward/std": 52.238914489746094, + "step": 16, + "step_time": 8.034480294008972 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 365.0, + "completions/max_terminated_length": 365.0, + "completions/mean_length": 211.875, + "completions/mean_terminated_length": 211.875, + "completions/min_length": 100.0, + "completions/min_terminated_length": 100.0, + "entropy": 0.899176836013794, + "epoch": 0.085, + "frac_reward_zero_std": 0.0, + "grad_norm": 6.691746711730957, + "learning_rate": 9.2e-07, + "loss": 0.25513291358947754, + "num_tokens": 147869.0, + "reward": -99.92500305175781, + "reward_std": 0.07070883363485336, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": -100.0, + "rewards/supergames_reward/std": 0.0, + "step": 17, + "step_time": 13.655792869016295 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 387.0, + "completions/max_terminated_length": 387.0, + "completions/mean_length": 238.0, + "completions/mean_terminated_length": 238.0, + "completions/min_length": 87.0, + "completions/min_terminated_length": 87.0, + "entropy": 0.5596873760223389, + "epoch": 0.09, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.6711347103118896, + "learning_rate": 9.15e-07, + "loss": -0.2628335654735565, + "num_tokens": 159101.0, + "reward": -24.900001525878906, + "reward_std": 103.50983428955078, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -25.0, + "rewards/supergames_reward/std": 103.50983428955078, + "step": 18, + "step_time": 14.784329420013819 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.125, + "completions/max_length": 512.0, + "completions/max_terminated_length": 277.0, + "completions/mean_length": 219.75, + "completions/mean_terminated_length": 178.00001525878906, + "completions/min_length": 90.0, + "completions/min_terminated_length": 90.0, + "entropy": 0.9245517253875732, + "epoch": 0.095, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.55502986907959, + "learning_rate": 9.1e-07, + "loss": 0.13060179352760315, + "num_tokens": 167299.0, + "reward": -86.89125061035156, + "reward_std": 36.95603561401367, + "rewards/json_format_reward/mean": 0.05000000074505806, + "rewards/json_format_reward/std": 0.09258200973272324, + "rewards/supergames_reward/mean": -86.94125366210938, + "rewards/supergames_reward/std": 36.93572235107422, + "step": 19, + "step_time": 18.832684700988466 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 267.0, + "completions/max_terminated_length": 267.0, + "completions/mean_length": 177.25, + "completions/mean_terminated_length": 177.25, + "completions/min_length": 93.0, + "completions/min_terminated_length": 93.0, + "entropy": 0.7331254482269287, + "epoch": 0.1, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.538497447967529, + "learning_rate": 9.05e-07, + "loss": -0.16418980062007904, + "num_tokens": 175141.0, + "reward": -68.52874755859375, + "reward_std": 58.48719787597656, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -68.62875366210938, + "rewards/supergames_reward/std": 58.48720169067383, + "step": 20, + "step_time": 10.090975169994636 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 300.0, + "completions/max_terminated_length": 300.0, + "completions/mean_length": 190.625, + "completions/mean_terminated_length": 190.625, + "completions/min_length": 103.0, + "completions/min_terminated_length": 103.0, + "entropy": 0.6659425497055054, + "epoch": 0.105, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.485373497009277, + "learning_rate": 9e-07, + "loss": -0.10106432437896729, + "num_tokens": 183098.0, + "reward": -71.48875427246094, + "reward_std": 55.33565139770508, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -71.5887451171875, + "rewards/supergames_reward/std": 55.33565139770508, + "step": 21, + "step_time": 11.255001295008697 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 279.0, + "completions/max_terminated_length": 279.0, + "completions/mean_length": 179.0, + "completions/mean_terminated_length": 179.0, + "completions/min_length": 103.0, + "completions/min_terminated_length": 103.0, + "entropy": 0.49365949630737305, + "epoch": 0.11, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.829419136047363, + "learning_rate": 8.95e-07, + "loss": 0.10587231814861298, + "num_tokens": 192826.0, + "reward": -75.62000274658203, + "reward_std": 45.01959991455078, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": -75.69499969482422, + "rewards/supergames_reward/std": 45.00411605834961, + "step": 22, + "step_time": 10.611246599000879 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 315.0, + "completions/max_terminated_length": 315.0, + "completions/mean_length": 212.625, + "completions/mean_terminated_length": 212.625, + "completions/min_length": 100.0, + "completions/min_terminated_length": 100.0, + "entropy": 0.43159469962120056, + "epoch": 0.115, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.541125774383545, + "learning_rate": 8.9e-07, + "loss": -0.1681259125471115, + "num_tokens": 202815.0, + "reward": 23.67624855041504, + "reward_std": 67.74340057373047, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 23.576250076293945, + "rewards/supergames_reward/std": 67.743408203125, + "step": 23, + "step_time": 11.731771531980485 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 168.0, + "completions/max_terminated_length": 168.0, + "completions/mean_length": 101.625, + "completions/mean_terminated_length": 101.625, + "completions/min_length": 65.0, + "completions/min_terminated_length": 65.0, + "entropy": 0.3972298502922058, + "epoch": 0.12, + "frac_reward_zero_std": 0.0, + "grad_norm": 5.8285417556762695, + "learning_rate": 8.85e-07, + "loss": -0.011585958302021027, + "num_tokens": 208188.0, + "reward": 58.849998474121094, + "reward_std": 56.9304084777832, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 58.75, + "rewards/supergames_reward/std": 56.9304084777832, + "step": 24, + "step_time": 6.187504950998118 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 303.0, + "completions/max_terminated_length": 303.0, + "completions/mean_length": 165.375, + "completions/mean_terminated_length": 165.375, + "completions/min_length": 69.0, + "completions/min_terminated_length": 69.0, + "entropy": 0.6339899301528931, + "epoch": 0.125, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.823799133300781, + "learning_rate": 8.799999999999999e-07, + "loss": 0.05159185826778412, + "num_tokens": 215935.0, + "reward": -23.521251678466797, + "reward_std": 74.23067474365234, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -23.621248245239258, + "rewards/supergames_reward/std": 74.23067474365234, + "step": 25, + "step_time": 11.059416300005978 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 349.0, + "completions/max_terminated_length": 349.0, + "completions/mean_length": 204.0, + "completions/mean_terminated_length": 204.0, + "completions/min_length": 143.0, + "completions/min_terminated_length": 143.0, + "entropy": 0.6462895274162292, + "epoch": 0.13, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.969292640686035, + "learning_rate": 8.75e-07, + "loss": 0.0115677984431386, + "num_tokens": 223999.0, + "reward": -74.8175048828125, + "reward_std": 40.94221115112305, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": -74.89250183105469, + "rewards/supergames_reward/std": 40.924625396728516, + "step": 26, + "step_time": 12.666237785975682 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 299.0, + "completions/max_terminated_length": 299.0, + "completions/mean_length": 166.5, + "completions/mean_terminated_length": 166.5, + "completions/min_length": 87.0, + "completions/min_terminated_length": 87.0, + "entropy": 0.4756850600242615, + "epoch": 0.135, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.915069103240967, + "learning_rate": 8.699999999999999e-07, + "loss": -0.13321346044540405, + "num_tokens": 233643.0, + "reward": -52.125, + "reward_std": 19.405282974243164, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": -52.20000076293945, + "rewards/supergames_reward/std": 19.33464813232422, + "step": 27, + "step_time": 11.172539752995363 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 184.0, + "completions/max_terminated_length": 184.0, + "completions/mean_length": 137.125, + "completions/mean_terminated_length": 137.125, + "completions/min_length": 100.0, + "completions/min_terminated_length": 100.0, + "entropy": 0.5834778547286987, + "epoch": 0.14, + "frac_reward_zero_std": 0.0, + "grad_norm": 5.146666526794434, + "learning_rate": 8.65e-07, + "loss": 0.060454584658145905, + "num_tokens": 241204.0, + "reward": -68.99250793457031, + "reward_std": 51.010581970214844, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -69.09249877929688, + "rewards/supergames_reward/std": 51.010581970214844, + "step": 28, + "step_time": 7.025046669004951 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 331.0, + "completions/max_terminated_length": 331.0, + "completions/mean_length": 182.375, + "completions/mean_terminated_length": 182.375, + "completions/min_length": 106.0, + "completions/min_terminated_length": 106.0, + "entropy": 0.6368551254272461, + "epoch": 0.145, + "frac_reward_zero_std": 0.0, + "grad_norm": 5.15535306930542, + "learning_rate": 8.599999999999999e-07, + "loss": -0.3323673903942108, + "num_tokens": 249087.0, + "reward": -44.17000198364258, + "reward_std": 80.5373306274414, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -44.27000045776367, + "rewards/supergames_reward/std": 80.5373306274414, + "step": 29, + "step_time": 12.079259724996518 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 190.0, + "completions/max_terminated_length": 190.0, + "completions/mean_length": 127.75, + "completions/mean_terminated_length": 127.75, + "completions/min_length": 80.0, + "completions/min_terminated_length": 80.0, + "entropy": 0.4491196870803833, + "epoch": 0.15, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.896186828613281, + "learning_rate": 8.55e-07, + "loss": -0.12887656688690186, + "num_tokens": 259445.0, + "reward": 57.446250915527344, + "reward_std": 38.78837585449219, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 57.34625244140625, + "rewards/supergames_reward/std": 38.78837966918945, + "step": 30, + "step_time": 7.483972909016302 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 166.0, + "completions/max_terminated_length": 166.0, + "completions/mean_length": 112.875, + "completions/mean_terminated_length": 112.875, + "completions/min_length": 85.0, + "completions/min_terminated_length": 85.0, + "entropy": 0.4858454465866089, + "epoch": 0.155, + "frac_reward_zero_std": 0.0, + "grad_norm": 6.742906093597412, + "learning_rate": 8.499999999999999e-07, + "loss": 0.06930096447467804, + "num_tokens": 264924.0, + "reward": -21.174999237060547, + "reward_std": 31.89051628112793, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": -21.25, + "rewards/supergames_reward/std": 31.819807052612305, + "step": 31, + "step_time": 6.1866529019898735 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 148.0, + "completions/max_terminated_length": 148.0, + "completions/mean_length": 118.125, + "completions/mean_terminated_length": 118.125, + "completions/min_length": 98.0, + "completions/min_terminated_length": 98.0, + "entropy": 0.45428749918937683, + "epoch": 0.16, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.663414478302002, + "learning_rate": 8.45e-07, + "loss": 0.05735419690608978, + "num_tokens": 270381.0, + "reward": 0.4687504768371582, + "reward_std": 38.66371154785156, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 0.36875057220458984, + "rewards/supergames_reward/std": 38.6637077331543, + "step": 32, + "step_time": 5.577042757999152 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 345.0, + "completions/max_terminated_length": 345.0, + "completions/mean_length": 166.875, + "completions/mean_terminated_length": 166.875, + "completions/min_length": 86.0, + "completions/min_terminated_length": 86.0, + "entropy": 0.6107293367385864, + "epoch": 0.165, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.75006628036499, + "learning_rate": 8.399999999999999e-07, + "loss": -0.14760127663612366, + "num_tokens": 278156.0, + "reward": -87.84750366210938, + "reward_std": 26.84708023071289, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -87.94749450683594, + "rewards/supergames_reward/std": 26.847076416015625, + "step": 33, + "step_time": 12.552876825997373 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.125, + "completions/max_length": 512.0, + "completions/max_terminated_length": 360.0, + "completions/mean_length": 231.0, + "completions/mean_terminated_length": 190.85714721679688, + "completions/min_length": 74.0, + "completions/min_terminated_length": 74.0, + "entropy": 0.6089653372764587, + "epoch": 0.17, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.781267166137695, + "learning_rate": 8.349999999999999e-07, + "loss": 0.4294504225254059, + "num_tokens": 289364.0, + "reward": -99.92500305175781, + "reward_std": 0.07070883363485336, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": -100.0, + "rewards/supergames_reward/std": 0.0, + "step": 34, + "step_time": 18.961819477990502 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 218.0, + "completions/max_terminated_length": 218.0, + "completions/mean_length": 134.125, + "completions/mean_terminated_length": 134.125, + "completions/min_length": 104.0, + "completions/min_terminated_length": 104.0, + "entropy": 0.46158918738365173, + "epoch": 0.175, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.016353607177734, + "learning_rate": 8.299999999999999e-07, + "loss": 0.09251818805932999, + "num_tokens": 298733.0, + "reward": -35.558753967285156, + "reward_std": 15.881204605102539, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -35.65875244140625, + "rewards/supergames_reward/std": 15.881203651428223, + "step": 35, + "step_time": 8.38491243700264 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 207.0, + "completions/max_terminated_length": 207.0, + "completions/mean_length": 156.875, + "completions/mean_terminated_length": 156.875, + "completions/min_length": 111.0, + "completions/min_terminated_length": 111.0, + "entropy": 0.5104788541793823, + "epoch": 0.18, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.592893123626709, + "learning_rate": 8.249999999999999e-07, + "loss": -0.0410832017660141, + "num_tokens": 306420.0, + "reward": 5.4212493896484375, + "reward_std": 93.50404357910156, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": 5.346250534057617, + "rewards/supergames_reward/std": 93.4718246459961, + "step": 36, + "step_time": 7.811868985998444 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 241.0, + "completions/max_terminated_length": 241.0, + "completions/mean_length": 178.0, + "completions/mean_terminated_length": 178.0, + "completions/min_length": 106.0, + "completions/min_terminated_length": 106.0, + "entropy": 0.7527878880500793, + "epoch": 0.185, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.8812546730041504, + "learning_rate": 8.199999999999999e-07, + "loss": 0.12539231777191162, + "num_tokens": 314252.0, + "reward": -89.58500671386719, + "reward_std": 29.256120681762695, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": -89.66000366210938, + "rewards/supergames_reward/std": 29.24593734741211, + "step": 37, + "step_time": 8.959742914012168 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 199.0, + "completions/max_terminated_length": 199.0, + "completions/mean_length": 133.375, + "completions/mean_terminated_length": 133.375, + "completions/min_length": 87.0, + "completions/min_terminated_length": 87.0, + "entropy": 0.3347901403903961, + "epoch": 0.19, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.205414295196533, + "learning_rate": 8.149999999999999e-07, + "loss": -0.06991486996412277, + "num_tokens": 319839.0, + "reward": -14.498749732971191, + "reward_std": 13.007229804992676, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -14.598750114440918, + "rewards/supergames_reward/std": 13.007229804992676, + "step": 38, + "step_time": 7.377183554985095 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 181.0, + "completions/max_terminated_length": 181.0, + "completions/mean_length": 114.375, + "completions/mean_terminated_length": 114.375, + "completions/min_length": 58.0, + "completions/min_terminated_length": 58.0, + "entropy": 0.4445965886116028, + "epoch": 0.195, + "frac_reward_zero_std": 0.0, + "grad_norm": 5.561298847198486, + "learning_rate": 8.1e-07, + "loss": -0.06235164776444435, + "num_tokens": 325290.0, + "reward": -7.425000190734863, + "reward_std": 53.701602935791016, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": -7.5, + "rewards/supergames_reward/std": 53.652320861816406, + "step": 39, + "step_time": 6.753862089011818 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 370.0, + "completions/max_terminated_length": 370.0, + "completions/mean_length": 187.625, + "completions/mean_terminated_length": 187.625, + "completions/min_length": 60.0, + "completions/min_terminated_length": 60.0, + "entropy": 0.3950234353542328, + "epoch": 0.2, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.5988712310791016, + "learning_rate": 8.05e-07, + "loss": -0.1381440907716751, + "num_tokens": 335119.0, + "reward": -33.68375015258789, + "reward_std": 42.069435119628906, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -33.783748626708984, + "rewards/supergames_reward/std": 42.069435119628906, + "step": 40, + "step_time": 13.718958162004128 + } + ], + "logging_steps": 1, + "max_steps": 200, + "num_input_tokens_seen": 335119, + "num_train_epochs": 1, + "save_steps": 20, + "stateful_callbacks": { + "TrainerControl": { + "args": { + "should_epoch_stop": false, + "should_evaluate": false, + "should_log": false, + "should_save": true, + "should_training_stop": false + }, + "attributes": {} + } + }, + "total_flos": 0.0, + "train_batch_size": 8, + "trial_name": null, + "trial_params": null +} diff --git a/checkpoint-40/training_args.bin b/checkpoint-40/training_args.bin new file mode 100644 index 0000000..730dec0 --- /dev/null +++ b/checkpoint-40/training_args.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:691fe5be2d7d1d56160dae63866d136877a72efb30e15ff7b3249192f998a788 +size 7185 diff --git a/checkpoint-60/chat_template.jinja b/checkpoint-60/chat_template.jinja new file mode 100644 index 0000000..bdf7919 --- /dev/null +++ b/checkpoint-60/chat_template.jinja @@ -0,0 +1,54 @@ +{%- if tools %} + {{- '<|im_start|>system\n' }} + {%- if messages[0]['role'] == 'system' %} + {{- messages[0]['content'] }} + {%- else %} + {{- 'You are Qwen, created by Alibaba Cloud. You are a helpful assistant.' }} + {%- endif %} + {{- "\n\n# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }} + {%- for tool in tools %} + {{- "\n" }} + {{- tool | tojson }} + {%- endfor %} + {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }} +{%- else %} + {%- if messages[0]['role'] == 'system' %} + {{- '<|im_start|>system\n' + messages[0]['content'] + '<|im_end|>\n' }} + {%- else %} + {{- '<|im_start|>system\nYou are Qwen, created by Alibaba Cloud. You are a helpful assistant.<|im_end|>\n' }} + {%- endif %} +{%- endif %} +{%- for message in messages %} + {%- if (message.role == "user") or (message.role == "system" and not loop.first) or (message.role == "assistant" and not message.tool_calls) %} + {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }} + {%- elif message.role == "assistant" %} + {{- '<|im_start|>' + message.role }} + {%- if message.content %} + {{- '\n' + message.content }} + {%- endif %} + {%- for tool_call in message.tool_calls %} + {%- if tool_call.function is defined %} + {%- set tool_call = tool_call.function %} + {%- endif %} + {{- '\n\n{"name": "' }} + {{- tool_call.name }} + {{- '", "arguments": ' }} + {{- tool_call.arguments | tojson }} + {{- '}\n' }} + {%- endfor %} + {{- '<|im_end|>\n' }} + {%- elif message.role == "tool" %} + {%- if (loop.index0 == 0) or (messages[loop.index0 - 1].role != "tool") %} + {{- '<|im_start|>user' }} + {%- endif %} + {{- '\n\n' }} + {{- message.content }} + {{- '\n' }} + {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %} + {{- '<|im_end|>\n' }} + {%- endif %} + {%- endif %} +{%- endfor %} +{%- if add_generation_prompt %} + {{- '<|im_start|>assistant\n' }} +{%- endif %} diff --git a/checkpoint-60/config.json b/checkpoint-60/config.json new file mode 100644 index 0000000..f57b09b --- /dev/null +++ b/checkpoint-60/config.json @@ -0,0 +1,61 @@ +{ + "architectures": [ + "Qwen2ForCausalLM" + ], + "attention_dropout": 0.0, + "bos_token_id": null, + "dtype": "float32", + "eos_token_id": 151645, + "hidden_act": "silu", + "hidden_size": 1536, + "initializer_range": 0.02, + "intermediate_size": 8960, + "layer_types": [ + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention" + ], + "max_position_embeddings": 32768, + "max_window_layers": 21, + "model_type": "qwen2", + "num_attention_heads": 12, + "num_hidden_layers": 28, + "num_key_value_heads": 2, + "pad_token_id": 151643, + "rms_norm_eps": 1e-06, + "rope_parameters": { + "rope_theta": 1000000.0, + "rope_type": "default" + }, + "sliding_window": null, + "tie_word_embeddings": true, + "transformers_version": "5.6.2", + "use_cache": false, + "use_sliding_window": false, + "vocab_size": 151936 +} diff --git a/checkpoint-60/generation_config.json b/checkpoint-60/generation_config.json new file mode 100644 index 0000000..1b2bba9 --- /dev/null +++ b/checkpoint-60/generation_config.json @@ -0,0 +1,13 @@ +{ + "do_sample": true, + "eos_token_id": [ + 151645, + 151643 + ], + "pad_token_id": 151643, + "repetition_penalty": 1.1, + "temperature": 0.7, + "top_k": 20, + "top_p": 0.8, + "transformers_version": "5.6.2" +} diff --git a/checkpoint-60/model.safetensors b/checkpoint-60/model.safetensors new file mode 100644 index 0000000..19e80b5 --- /dev/null +++ b/checkpoint-60/model.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:e62af89c25fad92b5ed3fcff2d61135e77afdbbac6bf1de213b138bc5c886202 +size 6174895536 diff --git a/checkpoint-60/optimizer.pt b/checkpoint-60/optimizer.pt new file mode 100644 index 0000000..b432bf9 --- /dev/null +++ b/checkpoint-60/optimizer.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:ff7d04ae88053c662cef57e6d5cc192af79abe8d30a164edb4b312e98b6908dd +size 12350013801 diff --git a/checkpoint-60/rng_state.pth b/checkpoint-60/rng_state.pth new file mode 100644 index 0000000..8b8159f --- /dev/null +++ b/checkpoint-60/rng_state.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:fae16bd6d032942a248f17a0f77c550f505e9c87781ba9a4b79a3b2d897cd14d +size 14645 diff --git a/checkpoint-60/scheduler.pt b/checkpoint-60/scheduler.pt new file mode 100644 index 0000000..ff57925 --- /dev/null +++ b/checkpoint-60/scheduler.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:de04036787404fb7ee6853d8c3f6a8fa92951d7f795a7b73b185729920015ad6 +size 1465 diff --git a/checkpoint-60/tokenizer.json b/checkpoint-60/tokenizer.json new file mode 100644 index 0000000..34510ff --- /dev/null +++ b/checkpoint-60/tokenizer.json @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:3fd169731d2cbde95e10bf356d66d5997fd885dd8dbb6fb4684da3f23b2585d8 +size 11421892 diff --git a/checkpoint-60/tokenizer_config.json b/checkpoint-60/tokenizer_config.json new file mode 100644 index 0000000..770e41d --- /dev/null +++ b/checkpoint-60/tokenizer_config.json @@ -0,0 +1,30 @@ +{ + "add_prefix_space": false, + "backend": "tokenizers", + "bos_token": null, + "clean_up_tokenization_spaces": false, + "eos_token": "<|im_end|>", + "errors": "replace", + "extra_special_tokens": [ + "<|im_start|>", + "<|im_end|>", + "<|object_ref_start|>", + "<|object_ref_end|>", + "<|box_start|>", + "<|box_end|>", + "<|quad_start|>", + "<|quad_end|>", + "<|vision_start|>", + "<|vision_end|>", + "<|vision_pad|>", + "<|image_pad|>", + "<|video_pad|>" + ], + "is_local": false, + "local_files_only": false, + "model_max_length": 131072, + "pad_token": "<|endoftext|>", + "split_special_tokens": false, + "tokenizer_class": "Qwen2Tokenizer", + "unk_token": null +} diff --git a/checkpoint-60/trainer_state.json b/checkpoint-60/trainer_state.json new file mode 100644 index 0000000..c15b75a --- /dev/null +++ b/checkpoint-60/trainer_state.json @@ -0,0 +1,1774 @@ +{ + "best_global_step": null, + "best_metric": null, + "best_model_checkpoint": null, + "epoch": 0.3, + "eval_steps": 500, + "global_step": 60, + "is_hyper_param_search": false, + "is_local_process_zero": true, + "is_world_process_zero": true, + "log_history": [ + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 478.0, + "completions/max_terminated_length": 478.0, + "completions/mean_length": 191.0, + "completions/mean_terminated_length": 191.0, + "completions/min_length": 66.0, + "completions/min_terminated_length": 66.0, + "entropy": 0.5566893219947815, + "epoch": 0.005, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.370361328125, + "learning_rate": 1e-06, + "loss": -0.24658073484897614, + "num_tokens": 10832.0, + "reward": -3.3000030517578125, + "reward_std": 85.62333679199219, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -3.4000015258789062, + "rewards/supergames_reward/std": 85.62333679199219, + "step": 1, + "step_time": 22.779711072013015 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 216.0, + "completions/max_terminated_length": 216.0, + "completions/mean_length": 131.5, + "completions/mean_terminated_length": 131.5, + "completions/min_length": 60.0, + "completions/min_terminated_length": 60.0, + "entropy": 0.6395841240882874, + "epoch": 0.01, + "frac_reward_zero_std": 0.0, + "grad_norm": 8.09774398803711, + "learning_rate": 9.95e-07, + "loss": 0.20567649602890015, + "num_tokens": 16404.0, + "reward": -12.422499656677246, + "reward_std": 7.134707450866699, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -12.522500038146973, + "rewards/supergames_reward/std": 7.134707927703857, + "step": 2, + "step_time": 9.960156448010821 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 287.0, + "completions/max_terminated_length": 287.0, + "completions/mean_length": 156.625, + "completions/mean_terminated_length": 156.625, + "completions/min_length": 77.0, + "completions/min_terminated_length": 77.0, + "entropy": 0.562222421169281, + "epoch": 0.015, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.9053616523742676, + "learning_rate": 9.9e-07, + "loss": 0.09602774679660797, + "num_tokens": 26953.0, + "reward": -74.94999694824219, + "reward_std": 70.73094177246094, + "rewards/json_format_reward/mean": 0.05000000074505806, + "rewards/json_format_reward/std": 0.09258200973272324, + "rewards/supergames_reward/mean": -75.0, + "rewards/supergames_reward/std": 70.71067810058594, + "step": 3, + "step_time": 13.686320498993155 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 268.0, + "completions/max_terminated_length": 268.0, + "completions/mean_length": 154.125, + "completions/mean_terminated_length": 154.125, + "completions/min_length": 53.0, + "completions/min_terminated_length": 53.0, + "entropy": 0.7323317527770996, + "epoch": 0.02, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.933310031890869, + "learning_rate": 9.849999999999999e-07, + "loss": 0.28249427676200867, + "num_tokens": 36514.0, + "reward": -40.7599983215332, + "reward_std": 81.73140716552734, + "rewards/json_format_reward/mean": 0.05000000074505806, + "rewards/json_format_reward/std": 0.09258200973272324, + "rewards/supergames_reward/mean": -40.80999755859375, + "rewards/supergames_reward/std": 81.68995666503906, + "step": 4, + "step_time": 12.805880262021674 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 401.0, + "completions/max_terminated_length": 401.0, + "completions/mean_length": 205.25, + "completions/mean_terminated_length": 205.25, + "completions/min_length": 115.0, + "completions/min_terminated_length": 115.0, + "entropy": 0.6167430281639099, + "epoch": 0.025, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.344135046005249, + "learning_rate": 9.8e-07, + "loss": 0.1433994024991989, + "num_tokens": 47476.0, + "reward": -97.6612548828125, + "reward_std": 6.3321428298950195, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -97.76124572753906, + "rewards/supergames_reward/std": 6.3321428298950195, + "step": 5, + "step_time": 18.789364666008623 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 276.0, + "completions/max_terminated_length": 276.0, + "completions/mean_length": 167.25, + "completions/mean_terminated_length": 167.25, + "completions/min_length": 73.0, + "completions/min_terminated_length": 73.0, + "entropy": 0.7583790421485901, + "epoch": 0.03, + "frac_reward_zero_std": 0.0, + "grad_norm": 5.172371864318848, + "learning_rate": 9.75e-07, + "loss": -0.12127404659986496, + "num_tokens": 55246.0, + "reward": -58.687503814697266, + "reward_std": 58.64059829711914, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": -58.76250457763672, + "rewards/supergames_reward/std": 58.620460510253906, + "step": 6, + "step_time": 12.948570273991209 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 261.0, + "completions/max_terminated_length": 261.0, + "completions/mean_length": 133.875, + "completions/mean_terminated_length": 133.875, + "completions/min_length": 57.0, + "completions/min_terminated_length": 57.0, + "entropy": 0.7447654604911804, + "epoch": 0.035, + "frac_reward_zero_std": 0.0, + "grad_norm": 6.428595066070557, + "learning_rate": 9.7e-07, + "loss": 0.10135584324598312, + "num_tokens": 60837.0, + "reward": -32.45000076293945, + "reward_std": 41.754486083984375, + "rewards/json_format_reward/mean": 0.05000000074505806, + "rewards/json_format_reward/std": 0.09258200973272324, + "rewards/supergames_reward/mean": -32.5, + "rewards/supergames_reward/std": 41.66190719604492, + "step": 7, + "step_time": 12.058315072004916 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 414.0, + "completions/max_terminated_length": 414.0, + "completions/mean_length": 209.125, + "completions/mean_terminated_length": 209.125, + "completions/min_length": 104.0, + "completions/min_terminated_length": 104.0, + "entropy": 0.3860666751861572, + "epoch": 0.04, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.806255578994751, + "learning_rate": 9.649999999999999e-07, + "loss": -0.33238139748573303, + "num_tokens": 70838.0, + "reward": -80.50375366210938, + "reward_std": 38.82380294799805, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -80.60375213623047, + "rewards/supergames_reward/std": 38.82379913330078, + "step": 8, + "step_time": 19.323370319994865 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.25, + "completions/max_length": 512.0, + "completions/max_terminated_length": 199.0, + "completions/mean_length": 247.125, + "completions/mean_terminated_length": 158.83334350585938, + "completions/min_length": 106.0, + "completions/min_terminated_length": 106.0, + "entropy": 0.5816237330436707, + "epoch": 0.045, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.3209943771362305, + "learning_rate": 9.6e-07, + "loss": 0.36752766370773315, + "num_tokens": 82135.0, + "reward": -27.688751220703125, + "reward_std": 82.8358154296875, + "rewards/json_format_reward/mean": 0.05000000074505806, + "rewards/json_format_reward/std": 0.09258200973272324, + "rewards/supergames_reward/mean": -27.738750457763672, + "rewards/supergames_reward/std": 82.78590393066406, + "step": 9, + "step_time": 21.948575104994234 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 240.0, + "completions/max_terminated_length": 240.0, + "completions/mean_length": 153.0, + "completions/mean_terminated_length": 153.0, + "completions/min_length": 51.0, + "completions/min_terminated_length": 51.0, + "entropy": 0.6301521062850952, + "epoch": 0.05, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.125359058380127, + "learning_rate": 9.55e-07, + "loss": 0.06129350885748863, + "num_tokens": 89807.0, + "reward": -78.39624786376953, + "reward_std": 40.17042922973633, + "rewards/json_format_reward/mean": 0.05000000074505806, + "rewards/json_format_reward/std": 0.09258200973272324, + "rewards/supergames_reward/mean": -78.44625091552734, + "rewards/supergames_reward/std": 40.1396484375, + "step": 10, + "step_time": 9.098202993016457 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 133.0, + "completions/max_terminated_length": 133.0, + "completions/mean_length": 112.0, + "completions/mean_terminated_length": 112.0, + "completions/min_length": 68.0, + "completions/min_terminated_length": 68.0, + "entropy": 0.5521990656852722, + "epoch": 0.055, + "frac_reward_zero_std": 0.0, + "grad_norm": 6.616060256958008, + "learning_rate": 9.499999999999999e-07, + "loss": -0.006313305348157883, + "num_tokens": 95231.0, + "reward": -12.161249160766602, + "reward_std": 6.39578104019165, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -12.261249542236328, + "rewards/supergames_reward/std": 6.39578104019165, + "step": 11, + "step_time": 5.136311663984088 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 396.0, + "completions/max_terminated_length": 396.0, + "completions/mean_length": 214.875, + "completions/mean_terminated_length": 214.875, + "completions/min_length": 55.0, + "completions/min_terminated_length": 55.0, + "entropy": 0.5505905151367188, + "epoch": 0.06, + "frac_reward_zero_std": 0.0, + "grad_norm": 5.289583206176758, + "learning_rate": 9.45e-07, + "loss": -0.29802003502845764, + "num_tokens": 106238.0, + "reward": -91.66999816894531, + "reward_std": 23.2779541015625, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -91.77000427246094, + "rewards/supergames_reward/std": 23.277956008911133, + "step": 12, + "step_time": 14.945365622988902 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 269.0, + "completions/max_terminated_length": 269.0, + "completions/mean_length": 167.375, + "completions/mean_terminated_length": 167.375, + "completions/min_length": 113.0, + "completions/min_terminated_length": 113.0, + "entropy": 0.540415346622467, + "epoch": 0.065, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.858291149139404, + "learning_rate": 9.399999999999999e-07, + "loss": 0.24004128575325012, + "num_tokens": 116937.0, + "reward": -20.38249969482422, + "reward_std": 87.55204010009766, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": -20.457500457763672, + "rewards/supergames_reward/std": 87.52605438232422, + "step": 13, + "step_time": 10.482285185018554 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 254.0, + "completions/max_terminated_length": 254.0, + "completions/mean_length": 156.875, + "completions/mean_terminated_length": 156.875, + "completions/min_length": 114.0, + "completions/min_terminated_length": 114.0, + "entropy": 0.7237679958343506, + "epoch": 0.07, + "frac_reward_zero_std": 0.0, + "grad_norm": 5.083980560302734, + "learning_rate": 9.35e-07, + "loss": -0.19025824964046478, + "num_tokens": 124640.0, + "reward": -51.397499084472656, + "reward_std": 54.486454010009766, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -51.49749755859375, + "rewards/supergames_reward/std": 54.486454010009766, + "step": 14, + "step_time": 9.597231683001155 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 236.0, + "completions/max_terminated_length": 236.0, + "completions/mean_length": 141.375, + "completions/mean_terminated_length": 141.375, + "completions/min_length": 91.0, + "completions/min_terminated_length": 91.0, + "entropy": 0.754334568977356, + "epoch": 0.075, + "frac_reward_zero_std": 0.0, + "grad_norm": 5.720428943634033, + "learning_rate": 9.3e-07, + "loss": -0.16830675303936005, + "num_tokens": 132219.0, + "reward": -56.226253509521484, + "reward_std": 60.115657806396484, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -56.32625198364258, + "rewards/supergames_reward/std": 60.11566162109375, + "step": 15, + "step_time": 8.998362872982398 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 209.0, + "completions/max_terminated_length": 209.0, + "completions/mean_length": 136.375, + "completions/mean_terminated_length": 136.375, + "completions/min_length": 97.0, + "completions/min_terminated_length": 97.0, + "entropy": 0.6280568838119507, + "epoch": 0.08, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.9681806564331055, + "learning_rate": 9.25e-07, + "loss": -0.01910916529595852, + "num_tokens": 139726.0, + "reward": -68.24000549316406, + "reward_std": 52.238914489746094, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -68.33999633789062, + "rewards/supergames_reward/std": 52.238914489746094, + "step": 16, + "step_time": 8.034480294008972 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 365.0, + "completions/max_terminated_length": 365.0, + "completions/mean_length": 211.875, + "completions/mean_terminated_length": 211.875, + "completions/min_length": 100.0, + "completions/min_terminated_length": 100.0, + "entropy": 0.899176836013794, + "epoch": 0.085, + "frac_reward_zero_std": 0.0, + "grad_norm": 6.691746711730957, + "learning_rate": 9.2e-07, + "loss": 0.25513291358947754, + "num_tokens": 147869.0, + "reward": -99.92500305175781, + "reward_std": 0.07070883363485336, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": -100.0, + "rewards/supergames_reward/std": 0.0, + "step": 17, + "step_time": 13.655792869016295 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 387.0, + "completions/max_terminated_length": 387.0, + "completions/mean_length": 238.0, + "completions/mean_terminated_length": 238.0, + "completions/min_length": 87.0, + "completions/min_terminated_length": 87.0, + "entropy": 0.5596873760223389, + "epoch": 0.09, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.6711347103118896, + "learning_rate": 9.15e-07, + "loss": -0.2628335654735565, + "num_tokens": 159101.0, + "reward": -24.900001525878906, + "reward_std": 103.50983428955078, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -25.0, + "rewards/supergames_reward/std": 103.50983428955078, + "step": 18, + "step_time": 14.784329420013819 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.125, + "completions/max_length": 512.0, + "completions/max_terminated_length": 277.0, + "completions/mean_length": 219.75, + "completions/mean_terminated_length": 178.00001525878906, + "completions/min_length": 90.0, + "completions/min_terminated_length": 90.0, + "entropy": 0.9245517253875732, + "epoch": 0.095, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.55502986907959, + "learning_rate": 9.1e-07, + "loss": 0.13060179352760315, + "num_tokens": 167299.0, + "reward": -86.89125061035156, + "reward_std": 36.95603561401367, + "rewards/json_format_reward/mean": 0.05000000074505806, + "rewards/json_format_reward/std": 0.09258200973272324, + "rewards/supergames_reward/mean": -86.94125366210938, + "rewards/supergames_reward/std": 36.93572235107422, + "step": 19, + "step_time": 18.832684700988466 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 267.0, + "completions/max_terminated_length": 267.0, + "completions/mean_length": 177.25, + "completions/mean_terminated_length": 177.25, + "completions/min_length": 93.0, + "completions/min_terminated_length": 93.0, + "entropy": 0.7331254482269287, + "epoch": 0.1, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.538497447967529, + "learning_rate": 9.05e-07, + "loss": -0.16418980062007904, + "num_tokens": 175141.0, + "reward": -68.52874755859375, + "reward_std": 58.48719787597656, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -68.62875366210938, + "rewards/supergames_reward/std": 58.48720169067383, + "step": 20, + "step_time": 10.090975169994636 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 300.0, + "completions/max_terminated_length": 300.0, + "completions/mean_length": 190.625, + "completions/mean_terminated_length": 190.625, + "completions/min_length": 103.0, + "completions/min_terminated_length": 103.0, + "entropy": 0.6659425497055054, + "epoch": 0.105, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.485373497009277, + "learning_rate": 9e-07, + "loss": -0.10106432437896729, + "num_tokens": 183098.0, + "reward": -71.48875427246094, + "reward_std": 55.33565139770508, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -71.5887451171875, + "rewards/supergames_reward/std": 55.33565139770508, + "step": 21, + "step_time": 11.255001295008697 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 279.0, + "completions/max_terminated_length": 279.0, + "completions/mean_length": 179.0, + "completions/mean_terminated_length": 179.0, + "completions/min_length": 103.0, + "completions/min_terminated_length": 103.0, + "entropy": 0.49365949630737305, + "epoch": 0.11, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.829419136047363, + "learning_rate": 8.95e-07, + "loss": 0.10587231814861298, + "num_tokens": 192826.0, + "reward": -75.62000274658203, + "reward_std": 45.01959991455078, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": -75.69499969482422, + "rewards/supergames_reward/std": 45.00411605834961, + "step": 22, + "step_time": 10.611246599000879 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 315.0, + "completions/max_terminated_length": 315.0, + "completions/mean_length": 212.625, + "completions/mean_terminated_length": 212.625, + "completions/min_length": 100.0, + "completions/min_terminated_length": 100.0, + "entropy": 0.43159469962120056, + "epoch": 0.115, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.541125774383545, + "learning_rate": 8.9e-07, + "loss": -0.1681259125471115, + "num_tokens": 202815.0, + "reward": 23.67624855041504, + "reward_std": 67.74340057373047, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 23.576250076293945, + "rewards/supergames_reward/std": 67.743408203125, + "step": 23, + "step_time": 11.731771531980485 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 168.0, + "completions/max_terminated_length": 168.0, + "completions/mean_length": 101.625, + "completions/mean_terminated_length": 101.625, + "completions/min_length": 65.0, + "completions/min_terminated_length": 65.0, + "entropy": 0.3972298502922058, + "epoch": 0.12, + "frac_reward_zero_std": 0.0, + "grad_norm": 5.8285417556762695, + "learning_rate": 8.85e-07, + "loss": -0.011585958302021027, + "num_tokens": 208188.0, + "reward": 58.849998474121094, + "reward_std": 56.9304084777832, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 58.75, + "rewards/supergames_reward/std": 56.9304084777832, + "step": 24, + "step_time": 6.187504950998118 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 303.0, + "completions/max_terminated_length": 303.0, + "completions/mean_length": 165.375, + "completions/mean_terminated_length": 165.375, + "completions/min_length": 69.0, + "completions/min_terminated_length": 69.0, + "entropy": 0.6339899301528931, + "epoch": 0.125, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.823799133300781, + "learning_rate": 8.799999999999999e-07, + "loss": 0.05159185826778412, + "num_tokens": 215935.0, + "reward": -23.521251678466797, + "reward_std": 74.23067474365234, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -23.621248245239258, + "rewards/supergames_reward/std": 74.23067474365234, + "step": 25, + "step_time": 11.059416300005978 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 349.0, + "completions/max_terminated_length": 349.0, + "completions/mean_length": 204.0, + "completions/mean_terminated_length": 204.0, + "completions/min_length": 143.0, + "completions/min_terminated_length": 143.0, + "entropy": 0.6462895274162292, + "epoch": 0.13, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.969292640686035, + "learning_rate": 8.75e-07, + "loss": 0.0115677984431386, + "num_tokens": 223999.0, + "reward": -74.8175048828125, + "reward_std": 40.94221115112305, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": -74.89250183105469, + "rewards/supergames_reward/std": 40.924625396728516, + "step": 26, + "step_time": 12.666237785975682 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 299.0, + "completions/max_terminated_length": 299.0, + "completions/mean_length": 166.5, + "completions/mean_terminated_length": 166.5, + "completions/min_length": 87.0, + "completions/min_terminated_length": 87.0, + "entropy": 0.4756850600242615, + "epoch": 0.135, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.915069103240967, + "learning_rate": 8.699999999999999e-07, + "loss": -0.13321346044540405, + "num_tokens": 233643.0, + "reward": -52.125, + "reward_std": 19.405282974243164, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": -52.20000076293945, + "rewards/supergames_reward/std": 19.33464813232422, + "step": 27, + "step_time": 11.172539752995363 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 184.0, + "completions/max_terminated_length": 184.0, + "completions/mean_length": 137.125, + "completions/mean_terminated_length": 137.125, + "completions/min_length": 100.0, + "completions/min_terminated_length": 100.0, + "entropy": 0.5834778547286987, + "epoch": 0.14, + "frac_reward_zero_std": 0.0, + "grad_norm": 5.146666526794434, + "learning_rate": 8.65e-07, + "loss": 0.060454584658145905, + "num_tokens": 241204.0, + "reward": -68.99250793457031, + "reward_std": 51.010581970214844, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -69.09249877929688, + "rewards/supergames_reward/std": 51.010581970214844, + "step": 28, + "step_time": 7.025046669004951 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 331.0, + "completions/max_terminated_length": 331.0, + "completions/mean_length": 182.375, + "completions/mean_terminated_length": 182.375, + "completions/min_length": 106.0, + "completions/min_terminated_length": 106.0, + "entropy": 0.6368551254272461, + "epoch": 0.145, + "frac_reward_zero_std": 0.0, + "grad_norm": 5.15535306930542, + "learning_rate": 8.599999999999999e-07, + "loss": -0.3323673903942108, + "num_tokens": 249087.0, + "reward": -44.17000198364258, + "reward_std": 80.5373306274414, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -44.27000045776367, + "rewards/supergames_reward/std": 80.5373306274414, + "step": 29, + "step_time": 12.079259724996518 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 190.0, + "completions/max_terminated_length": 190.0, + "completions/mean_length": 127.75, + "completions/mean_terminated_length": 127.75, + "completions/min_length": 80.0, + "completions/min_terminated_length": 80.0, + "entropy": 0.4491196870803833, + "epoch": 0.15, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.896186828613281, + "learning_rate": 8.55e-07, + "loss": -0.12887656688690186, + "num_tokens": 259445.0, + "reward": 57.446250915527344, + "reward_std": 38.78837585449219, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 57.34625244140625, + "rewards/supergames_reward/std": 38.78837966918945, + "step": 30, + "step_time": 7.483972909016302 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 166.0, + "completions/max_terminated_length": 166.0, + "completions/mean_length": 112.875, + "completions/mean_terminated_length": 112.875, + "completions/min_length": 85.0, + "completions/min_terminated_length": 85.0, + "entropy": 0.4858454465866089, + "epoch": 0.155, + "frac_reward_zero_std": 0.0, + "grad_norm": 6.742906093597412, + "learning_rate": 8.499999999999999e-07, + "loss": 0.06930096447467804, + "num_tokens": 264924.0, + "reward": -21.174999237060547, + "reward_std": 31.89051628112793, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": -21.25, + "rewards/supergames_reward/std": 31.819807052612305, + "step": 31, + "step_time": 6.1866529019898735 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 148.0, + "completions/max_terminated_length": 148.0, + "completions/mean_length": 118.125, + "completions/mean_terminated_length": 118.125, + "completions/min_length": 98.0, + "completions/min_terminated_length": 98.0, + "entropy": 0.45428749918937683, + "epoch": 0.16, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.663414478302002, + "learning_rate": 8.45e-07, + "loss": 0.05735419690608978, + "num_tokens": 270381.0, + "reward": 0.4687504768371582, + "reward_std": 38.66371154785156, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 0.36875057220458984, + "rewards/supergames_reward/std": 38.6637077331543, + "step": 32, + "step_time": 5.577042757999152 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 345.0, + "completions/max_terminated_length": 345.0, + "completions/mean_length": 166.875, + "completions/mean_terminated_length": 166.875, + "completions/min_length": 86.0, + "completions/min_terminated_length": 86.0, + "entropy": 0.6107293367385864, + "epoch": 0.165, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.75006628036499, + "learning_rate": 8.399999999999999e-07, + "loss": -0.14760127663612366, + "num_tokens": 278156.0, + "reward": -87.84750366210938, + "reward_std": 26.84708023071289, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -87.94749450683594, + "rewards/supergames_reward/std": 26.847076416015625, + "step": 33, + "step_time": 12.552876825997373 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.125, + "completions/max_length": 512.0, + "completions/max_terminated_length": 360.0, + "completions/mean_length": 231.0, + "completions/mean_terminated_length": 190.85714721679688, + "completions/min_length": 74.0, + "completions/min_terminated_length": 74.0, + "entropy": 0.6089653372764587, + "epoch": 0.17, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.781267166137695, + "learning_rate": 8.349999999999999e-07, + "loss": 0.4294504225254059, + "num_tokens": 289364.0, + "reward": -99.92500305175781, + "reward_std": 0.07070883363485336, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": -100.0, + "rewards/supergames_reward/std": 0.0, + "step": 34, + "step_time": 18.961819477990502 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 218.0, + "completions/max_terminated_length": 218.0, + "completions/mean_length": 134.125, + "completions/mean_terminated_length": 134.125, + "completions/min_length": 104.0, + "completions/min_terminated_length": 104.0, + "entropy": 0.46158918738365173, + "epoch": 0.175, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.016353607177734, + "learning_rate": 8.299999999999999e-07, + "loss": 0.09251818805932999, + "num_tokens": 298733.0, + "reward": -35.558753967285156, + "reward_std": 15.881204605102539, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -35.65875244140625, + "rewards/supergames_reward/std": 15.881203651428223, + "step": 35, + "step_time": 8.38491243700264 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 207.0, + "completions/max_terminated_length": 207.0, + "completions/mean_length": 156.875, + "completions/mean_terminated_length": 156.875, + "completions/min_length": 111.0, + "completions/min_terminated_length": 111.0, + "entropy": 0.5104788541793823, + "epoch": 0.18, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.592893123626709, + "learning_rate": 8.249999999999999e-07, + "loss": -0.0410832017660141, + "num_tokens": 306420.0, + "reward": 5.4212493896484375, + "reward_std": 93.50404357910156, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": 5.346250534057617, + "rewards/supergames_reward/std": 93.4718246459961, + "step": 36, + "step_time": 7.811868985998444 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 241.0, + "completions/max_terminated_length": 241.0, + "completions/mean_length": 178.0, + "completions/mean_terminated_length": 178.0, + "completions/min_length": 106.0, + "completions/min_terminated_length": 106.0, + "entropy": 0.7527878880500793, + "epoch": 0.185, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.8812546730041504, + "learning_rate": 8.199999999999999e-07, + "loss": 0.12539231777191162, + "num_tokens": 314252.0, + "reward": -89.58500671386719, + "reward_std": 29.256120681762695, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": -89.66000366210938, + "rewards/supergames_reward/std": 29.24593734741211, + "step": 37, + "step_time": 8.959742914012168 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 199.0, + "completions/max_terminated_length": 199.0, + "completions/mean_length": 133.375, + "completions/mean_terminated_length": 133.375, + "completions/min_length": 87.0, + "completions/min_terminated_length": 87.0, + "entropy": 0.3347901403903961, + "epoch": 0.19, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.205414295196533, + "learning_rate": 8.149999999999999e-07, + "loss": -0.06991486996412277, + "num_tokens": 319839.0, + "reward": -14.498749732971191, + "reward_std": 13.007229804992676, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -14.598750114440918, + "rewards/supergames_reward/std": 13.007229804992676, + "step": 38, + "step_time": 7.377183554985095 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 181.0, + "completions/max_terminated_length": 181.0, + "completions/mean_length": 114.375, + "completions/mean_terminated_length": 114.375, + "completions/min_length": 58.0, + "completions/min_terminated_length": 58.0, + "entropy": 0.4445965886116028, + "epoch": 0.195, + "frac_reward_zero_std": 0.0, + "grad_norm": 5.561298847198486, + "learning_rate": 8.1e-07, + "loss": -0.06235164776444435, + "num_tokens": 325290.0, + "reward": -7.425000190734863, + "reward_std": 53.701602935791016, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": -7.5, + "rewards/supergames_reward/std": 53.652320861816406, + "step": 39, + "step_time": 6.753862089011818 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 370.0, + "completions/max_terminated_length": 370.0, + "completions/mean_length": 187.625, + "completions/mean_terminated_length": 187.625, + "completions/min_length": 60.0, + "completions/min_terminated_length": 60.0, + "entropy": 0.3950234353542328, + "epoch": 0.2, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.5988712310791016, + "learning_rate": 8.05e-07, + "loss": -0.1381440907716751, + "num_tokens": 335119.0, + "reward": -33.68375015258789, + "reward_std": 42.069435119628906, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -33.783748626708984, + "rewards/supergames_reward/std": 42.069435119628906, + "step": 40, + "step_time": 13.718958162004128 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.125, + "completions/max_length": 512.0, + "completions/max_terminated_length": 371.0, + "completions/mean_length": 251.125, + "completions/mean_terminated_length": 213.85714721679688, + "completions/min_length": 121.0, + "completions/min_terminated_length": 121.0, + "entropy": 0.507067084312439, + "epoch": 0.205, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.3898658752441406, + "learning_rate": 8e-07, + "loss": 0.0078964838758111, + "num_tokens": 346464.0, + "reward": -86.48625183105469, + "reward_std": 38.020694732666016, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": -86.56124877929688, + "rewards/supergames_reward/std": 38.010528564453125, + "step": 41, + "step_time": 18.90960379401804 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 152.0, + "completions/max_terminated_length": 152.0, + "completions/mean_length": 116.375, + "completions/mean_terminated_length": 116.375, + "completions/min_length": 93.0, + "completions/min_terminated_length": 93.0, + "entropy": 0.3801707923412323, + "epoch": 0.21, + "frac_reward_zero_std": 1.0, + "grad_norm": 0.0, + "learning_rate": 7.95e-07, + "loss": 0.0, + "num_tokens": 351955.0, + "reward": -9.899999618530273, + "reward_std": 0.0, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -10.0, + "rewards/supergames_reward/std": 0.0, + "step": 42, + "step_time": 5.711630532023264 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 308.0, + "completions/max_terminated_length": 308.0, + "completions/mean_length": 247.375, + "completions/mean_terminated_length": 247.375, + "completions/min_length": 116.0, + "completions/min_terminated_length": 116.0, + "entropy": 0.3185043931007385, + "epoch": 0.215, + "frac_reward_zero_std": 0.0, + "grad_norm": 2.168762445449829, + "learning_rate": 7.9e-07, + "loss": 0.1970764398574829, + "num_tokens": 363278.0, + "reward": -53.89249801635742, + "reward_std": 63.02619934082031, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -53.99250030517578, + "rewards/supergames_reward/std": 63.02619934082031, + "step": 43, + "step_time": 11.661934906995157 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 330.0, + "completions/max_terminated_length": 330.0, + "completions/mean_length": 212.5, + "completions/mean_terminated_length": 212.5, + "completions/min_length": 106.0, + "completions/min_terminated_length": 106.0, + "entropy": 0.4446268379688263, + "epoch": 0.22, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.305844306945801, + "learning_rate": 7.85e-07, + "loss": -0.06735197454690933, + "num_tokens": 374306.0, + "reward": 18.059999465942383, + "reward_std": 98.9417953491211, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 17.959999084472656, + "rewards/supergames_reward/std": 98.9417953491211, + "step": 44, + "step_time": 12.492453911982011 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 316.0, + "completions/max_terminated_length": 316.0, + "completions/mean_length": 183.0, + "completions/mean_terminated_length": 183.0, + "completions/min_length": 103.0, + "completions/min_terminated_length": 103.0, + "entropy": 0.5269614458084106, + "epoch": 0.225, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.8006370067596436, + "learning_rate": 7.799999999999999e-07, + "loss": -0.08870815485715866, + "num_tokens": 382210.0, + "reward": -54.928749084472656, + "reward_std": 72.5127944946289, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -55.02874755859375, + "rewards/supergames_reward/std": 72.5127944946289, + "step": 45, + "step_time": 11.7636467939883 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 244.0, + "completions/max_terminated_length": 244.0, + "completions/mean_length": 136.625, + "completions/mean_terminated_length": 136.625, + "completions/min_length": 85.0, + "completions/min_terminated_length": 85.0, + "entropy": 0.34820204973220825, + "epoch": 0.23, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.911736011505127, + "learning_rate": 7.75e-07, + "loss": 0.04397330805659294, + "num_tokens": 391583.0, + "reward": -19.10375213623047, + "reward_std": 23.967702865600586, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -19.203750610351562, + "rewards/supergames_reward/std": 23.967702865600586, + "step": 46, + "step_time": 10.119426151999505 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 218.0, + "completions/max_terminated_length": 218.0, + "completions/mean_length": 160.375, + "completions/mean_terminated_length": 160.375, + "completions/min_length": 93.0, + "completions/min_terminated_length": 93.0, + "entropy": 0.611914336681366, + "epoch": 0.235, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.571913242340088, + "learning_rate": 7.699999999999999e-07, + "loss": -0.044293053448200226, + "num_tokens": 399290.0, + "reward": -25.186250686645508, + "reward_std": 71.46305084228516, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -25.286251068115234, + "rewards/supergames_reward/std": 71.46305084228516, + "step": 47, + "step_time": 8.563868903991533 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 167.0, + "completions/max_terminated_length": 167.0, + "completions/mean_length": 111.125, + "completions/mean_terminated_length": 111.125, + "completions/min_length": 56.0, + "completions/min_terminated_length": 56.0, + "entropy": 0.46585777401924133, + "epoch": 0.24, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.654109954833984, + "learning_rate": 7.65e-07, + "loss": -0.035151124000549316, + "num_tokens": 404707.0, + "reward": -0.408750057220459, + "reward_std": 31.939748764038086, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -0.5087499618530273, + "rewards/supergames_reward/std": 31.93975257873535, + "step": 48, + "step_time": 6.438482580007985 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 306.0, + "completions/max_terminated_length": 306.0, + "completions/mean_length": 216.625, + "completions/mean_terminated_length": 216.625, + "completions/min_length": 132.0, + "completions/min_terminated_length": 132.0, + "entropy": 0.33054471015930176, + "epoch": 0.245, + "frac_reward_zero_std": 0.0, + "grad_norm": 2.700303792953491, + "learning_rate": 7.599999999999999e-07, + "loss": -0.04393656551837921, + "num_tokens": 414736.0, + "reward": -97.26000213623047, + "reward_std": 1.6294406652450562, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -97.36000061035156, + "rewards/supergames_reward/std": 1.629441738128662, + "step": 49, + "step_time": 11.636040106008295 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 378.0, + "completions/max_terminated_length": 378.0, + "completions/mean_length": 218.375, + "completions/mean_terminated_length": 218.375, + "completions/min_length": 104.0, + "completions/min_terminated_length": 104.0, + "entropy": 0.48711177706718445, + "epoch": 0.25, + "frac_reward_zero_std": 0.0, + "grad_norm": 2.9650564193725586, + "learning_rate": 7.55e-07, + "loss": -0.3529520332813263, + "num_tokens": 424795.0, + "reward": 47.408748626708984, + "reward_std": 72.72083282470703, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 47.30875015258789, + "rewards/supergames_reward/std": 72.72083282470703, + "step": 50, + "step_time": 14.258096842997475 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.125, + "completions/max_length": 512.0, + "completions/max_terminated_length": 492.0, + "completions/mean_length": 267.625, + "completions/mean_terminated_length": 232.71429443359375, + "completions/min_length": 90.0, + "completions/min_terminated_length": 90.0, + "entropy": 0.4482860863208771, + "epoch": 0.255, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.2998197078704834, + "learning_rate": 7.5e-07, + "loss": 0.2493654191493988, + "num_tokens": 436288.0, + "reward": 50.61125183105469, + "reward_std": 75.3349838256836, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": 50.53624725341797, + "rewards/supergames_reward/std": 75.27783966064453, + "step": 51, + "step_time": 19.20827590499539 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 176.0, + "completions/max_terminated_length": 176.0, + "completions/mean_length": 144.5, + "completions/mean_terminated_length": 144.5, + "completions/min_length": 99.0, + "completions/min_terminated_length": 99.0, + "entropy": 0.5378735661506653, + "epoch": 0.26, + "frac_reward_zero_std": 0.0, + "grad_norm": 5.114461898803711, + "learning_rate": 7.45e-07, + "loss": 0.0036597400903701782, + "num_tokens": 441980.0, + "reward": -15.20625114440918, + "reward_std": 45.35526657104492, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": -15.28125, + "rewards/supergames_reward/std": 45.301815032958984, + "step": 52, + "step_time": 6.660627231001854 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 136.0, + "completions/max_terminated_length": 136.0, + "completions/mean_length": 119.0, + "completions/mean_terminated_length": 119.0, + "completions/min_length": 105.0, + "completions/min_terminated_length": 105.0, + "entropy": 0.41796669363975525, + "epoch": 0.265, + "frac_reward_zero_std": 0.0, + "grad_norm": 5.059505462646484, + "learning_rate": 7.4e-07, + "loss": -0.014973883517086506, + "num_tokens": 447468.0, + "reward": 15.170000076293945, + "reward_std": 46.658329010009766, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 15.070000648498535, + "rewards/supergames_reward/std": 46.658329010009766, + "step": 53, + "step_time": 5.271571868011961 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 259.0, + "completions/max_terminated_length": 259.0, + "completions/mean_length": 191.125, + "completions/mean_terminated_length": 191.125, + "completions/min_length": 91.0, + "completions/min_terminated_length": 91.0, + "entropy": 0.45421987771987915, + "epoch": 0.27, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.4454867839813232, + "learning_rate": 7.35e-07, + "loss": -0.1128973588347435, + "num_tokens": 457301.0, + "reward": 38.89875030517578, + "reward_std": 62.1772575378418, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 38.79874801635742, + "rewards/supergames_reward/std": 62.1772575378418, + "step": 54, + "step_time": 10.524528659996577 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 319.0, + "completions/max_terminated_length": 319.0, + "completions/mean_length": 233.375, + "completions/mean_terminated_length": 233.375, + "completions/min_length": 136.0, + "completions/min_terminated_length": 136.0, + "entropy": 0.4449213445186615, + "epoch": 0.275, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.2065398693084717, + "learning_rate": 7.3e-07, + "loss": 0.0016277075046673417, + "num_tokens": 467520.0, + "reward": 27.372499465942383, + "reward_std": 81.23126983642578, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 27.27250099182129, + "rewards/supergames_reward/std": 81.23126983642578, + "step": 55, + "step_time": 12.592280682991259 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.125, + "completions/max_length": 512.0, + "completions/max_terminated_length": 357.0, + "completions/mean_length": 261.0, + "completions/mean_terminated_length": 225.1428680419922, + "completions/min_length": 95.0, + "completions/min_terminated_length": 95.0, + "entropy": 0.506920337677002, + "epoch": 0.28, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.058388710021973, + "learning_rate": 7.249999999999999e-07, + "loss": 0.3289242386817932, + "num_tokens": 476056.0, + "reward": 47.05249786376953, + "reward_std": 80.97753143310547, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": 46.977500915527344, + "rewards/supergames_reward/std": 80.9256362915039, + "step": 56, + "step_time": 18.845177220006008 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 250.0, + "completions/max_terminated_length": 250.0, + "completions/mean_length": 192.75, + "completions/mean_terminated_length": 192.75, + "completions/min_length": 122.0, + "completions/min_terminated_length": 122.0, + "entropy": 0.5197336077690125, + "epoch": 0.285, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.230815410614014, + "learning_rate": 7.2e-07, + "loss": -0.15186546742916107, + "num_tokens": 484014.0, + "reward": -64.5050048828125, + "reward_std": 50.35145568847656, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -64.60499572753906, + "rewards/supergames_reward/std": 50.3514518737793, + "step": 57, + "step_time": 9.576232638006331 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 324.0, + "completions/max_terminated_length": 324.0, + "completions/mean_length": 249.0, + "completions/mean_terminated_length": 249.0, + "completions/min_length": 129.0, + "completions/min_terminated_length": 129.0, + "entropy": 0.42913514375686646, + "epoch": 0.29, + "frac_reward_zero_std": 0.0, + "grad_norm": 2.8544528484344482, + "learning_rate": 7.149999999999999e-07, + "loss": -0.056405920535326004, + "num_tokens": 494390.0, + "reward": 40.123748779296875, + "reward_std": 84.164794921875, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 40.02375030517578, + "rewards/supergames_reward/std": 84.164794921875, + "step": 58, + "step_time": 12.385353341000155 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 138.0, + "completions/max_terminated_length": 138.0, + "completions/mean_length": 107.25, + "completions/mean_terminated_length": 107.25, + "completions/min_length": 81.0, + "completions/min_terminated_length": 81.0, + "entropy": 0.34386488795280457, + "epoch": 0.295, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.6915154457092285, + "learning_rate": 7.1e-07, + "loss": 0.0865321159362793, + "num_tokens": 499816.0, + "reward": -8.244999885559082, + "reward_std": 4.681046962738037, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -8.345000267028809, + "rewards/supergames_reward/std": 4.681046962738037, + "step": 59, + "step_time": 5.42325339600211 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 187.0, + "completions/max_terminated_length": 187.0, + "completions/mean_length": 127.25, + "completions/mean_terminated_length": 127.25, + "completions/min_length": 97.0, + "completions/min_terminated_length": 97.0, + "entropy": 0.5430045127868652, + "epoch": 0.3, + "frac_reward_zero_std": 0.0, + "grad_norm": 5.635311126708984, + "learning_rate": 7.049999999999999e-07, + "loss": -0.05412375554442406, + "num_tokens": 505410.0, + "reward": 20.005001068115234, + "reward_std": 55.156005859375, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 19.904998779296875, + "rewards/supergames_reward/std": 55.156005859375, + "step": 60, + "step_time": 7.068010902003152 + } + ], + "logging_steps": 1, + "max_steps": 200, + "num_input_tokens_seen": 505410, + "num_train_epochs": 1, + "save_steps": 20, + "stateful_callbacks": { + "TrainerControl": { + "args": { + "should_epoch_stop": false, + "should_evaluate": false, + "should_log": false, + "should_save": true, + "should_training_stop": false + }, + "attributes": {} + } + }, + "total_flos": 0.0, + "train_batch_size": 8, + "trial_name": null, + "trial_params": null +} diff --git a/checkpoint-60/training_args.bin b/checkpoint-60/training_args.bin new file mode 100644 index 0000000..730dec0 --- /dev/null +++ b/checkpoint-60/training_args.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:691fe5be2d7d1d56160dae63866d136877a72efb30e15ff7b3249192f998a788 +size 7185 diff --git a/checkpoint-80/chat_template.jinja b/checkpoint-80/chat_template.jinja new file mode 100644 index 0000000..bdf7919 --- /dev/null +++ b/checkpoint-80/chat_template.jinja @@ -0,0 +1,54 @@ +{%- if tools %} + {{- '<|im_start|>system\n' }} + {%- if messages[0]['role'] == 'system' %} + {{- messages[0]['content'] }} + {%- else %} + {{- 'You are Qwen, created by Alibaba Cloud. You are a helpful assistant.' }} + {%- endif %} + {{- "\n\n# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }} + {%- for tool in tools %} + {{- "\n" }} + {{- tool | tojson }} + {%- endfor %} + {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }} +{%- else %} + {%- if messages[0]['role'] == 'system' %} + {{- '<|im_start|>system\n' + messages[0]['content'] + '<|im_end|>\n' }} + {%- else %} + {{- '<|im_start|>system\nYou are Qwen, created by Alibaba Cloud. You are a helpful assistant.<|im_end|>\n' }} + {%- endif %} +{%- endif %} +{%- for message in messages %} + {%- if (message.role == "user") or (message.role == "system" and not loop.first) or (message.role == "assistant" and not message.tool_calls) %} + {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }} + {%- elif message.role == "assistant" %} + {{- '<|im_start|>' + message.role }} + {%- if message.content %} + {{- '\n' + message.content }} + {%- endif %} + {%- for tool_call in message.tool_calls %} + {%- if tool_call.function is defined %} + {%- set tool_call = tool_call.function %} + {%- endif %} + {{- '\n\n{"name": "' }} + {{- tool_call.name }} + {{- '", "arguments": ' }} + {{- tool_call.arguments | tojson }} + {{- '}\n' }} + {%- endfor %} + {{- '<|im_end|>\n' }} + {%- elif message.role == "tool" %} + {%- if (loop.index0 == 0) or (messages[loop.index0 - 1].role != "tool") %} + {{- '<|im_start|>user' }} + {%- endif %} + {{- '\n\n' }} + {{- message.content }} + {{- '\n' }} + {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %} + {{- '<|im_end|>\n' }} + {%- endif %} + {%- endif %} +{%- endfor %} +{%- if add_generation_prompt %} + {{- '<|im_start|>assistant\n' }} +{%- endif %} diff --git a/checkpoint-80/config.json b/checkpoint-80/config.json new file mode 100644 index 0000000..f57b09b --- /dev/null +++ b/checkpoint-80/config.json @@ -0,0 +1,61 @@ +{ + "architectures": [ + "Qwen2ForCausalLM" + ], + "attention_dropout": 0.0, + "bos_token_id": null, + "dtype": "float32", + "eos_token_id": 151645, + "hidden_act": "silu", + "hidden_size": 1536, + "initializer_range": 0.02, + "intermediate_size": 8960, + "layer_types": [ + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention" + ], + "max_position_embeddings": 32768, + "max_window_layers": 21, + "model_type": "qwen2", + "num_attention_heads": 12, + "num_hidden_layers": 28, + "num_key_value_heads": 2, + "pad_token_id": 151643, + "rms_norm_eps": 1e-06, + "rope_parameters": { + "rope_theta": 1000000.0, + "rope_type": "default" + }, + "sliding_window": null, + "tie_word_embeddings": true, + "transformers_version": "5.6.2", + "use_cache": false, + "use_sliding_window": false, + "vocab_size": 151936 +} diff --git a/checkpoint-80/generation_config.json b/checkpoint-80/generation_config.json new file mode 100644 index 0000000..1b2bba9 --- /dev/null +++ b/checkpoint-80/generation_config.json @@ -0,0 +1,13 @@ +{ + "do_sample": true, + "eos_token_id": [ + 151645, + 151643 + ], + "pad_token_id": 151643, + "repetition_penalty": 1.1, + "temperature": 0.7, + "top_k": 20, + "top_p": 0.8, + "transformers_version": "5.6.2" +} diff --git a/checkpoint-80/model.safetensors b/checkpoint-80/model.safetensors new file mode 100644 index 0000000..895468f --- /dev/null +++ b/checkpoint-80/model.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:36c49868659b2f8b24ac9a86bf6e80c90079b3530402b89634f2af283a1ea452 +size 6174895536 diff --git a/checkpoint-80/optimizer.pt b/checkpoint-80/optimizer.pt new file mode 100644 index 0000000..81d9d05 --- /dev/null +++ b/checkpoint-80/optimizer.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:e39d224c64f43b5aacbde8ab914fec80a9f819231b78731c937ba3f159056817 +size 12350013801 diff --git a/checkpoint-80/rng_state.pth b/checkpoint-80/rng_state.pth new file mode 100644 index 0000000..a1fe3d4 --- /dev/null +++ b/checkpoint-80/rng_state.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:772106dc1aba21ee6aaf31ea19134157618b434ed364c663ce63e5ff7344641c +size 14645 diff --git a/checkpoint-80/scheduler.pt b/checkpoint-80/scheduler.pt new file mode 100644 index 0000000..30fa8ce --- /dev/null +++ b/checkpoint-80/scheduler.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:a0f1c734dfd575182ad7d0aca3bf6db0d25d9f1d262a3050bff25c7518303a48 +size 1465 diff --git a/checkpoint-80/tokenizer.json b/checkpoint-80/tokenizer.json new file mode 100644 index 0000000..34510ff --- /dev/null +++ b/checkpoint-80/tokenizer.json @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:3fd169731d2cbde95e10bf356d66d5997fd885dd8dbb6fb4684da3f23b2585d8 +size 11421892 diff --git a/checkpoint-80/tokenizer_config.json b/checkpoint-80/tokenizer_config.json new file mode 100644 index 0000000..770e41d --- /dev/null +++ b/checkpoint-80/tokenizer_config.json @@ -0,0 +1,30 @@ +{ + "add_prefix_space": false, + "backend": "tokenizers", + "bos_token": null, + "clean_up_tokenization_spaces": false, + "eos_token": "<|im_end|>", + "errors": "replace", + "extra_special_tokens": [ + "<|im_start|>", + "<|im_end|>", + "<|object_ref_start|>", + "<|object_ref_end|>", + "<|box_start|>", + "<|box_end|>", + "<|quad_start|>", + "<|quad_end|>", + "<|vision_start|>", + "<|vision_end|>", + "<|vision_pad|>", + "<|image_pad|>", + "<|video_pad|>" + ], + "is_local": false, + "local_files_only": false, + "model_max_length": 131072, + "pad_token": "<|endoftext|>", + "split_special_tokens": false, + "tokenizer_class": "Qwen2Tokenizer", + "unk_token": null +} diff --git a/checkpoint-80/trainer_state.json b/checkpoint-80/trainer_state.json new file mode 100644 index 0000000..b696646 --- /dev/null +++ b/checkpoint-80/trainer_state.json @@ -0,0 +1,2354 @@ +{ + "best_global_step": null, + "best_metric": null, + "best_model_checkpoint": null, + "epoch": 0.4, + "eval_steps": 500, + "global_step": 80, + "is_hyper_param_search": false, + "is_local_process_zero": true, + "is_world_process_zero": true, + "log_history": [ + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 478.0, + "completions/max_terminated_length": 478.0, + "completions/mean_length": 191.0, + "completions/mean_terminated_length": 191.0, + "completions/min_length": 66.0, + "completions/min_terminated_length": 66.0, + "entropy": 0.5566893219947815, + "epoch": 0.005, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.370361328125, + "learning_rate": 1e-06, + "loss": -0.24658073484897614, + "num_tokens": 10832.0, + "reward": -3.3000030517578125, + "reward_std": 85.62333679199219, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -3.4000015258789062, + "rewards/supergames_reward/std": 85.62333679199219, + "step": 1, + "step_time": 22.779711072013015 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 216.0, + "completions/max_terminated_length": 216.0, + "completions/mean_length": 131.5, + "completions/mean_terminated_length": 131.5, + "completions/min_length": 60.0, + "completions/min_terminated_length": 60.0, + "entropy": 0.6395841240882874, + "epoch": 0.01, + "frac_reward_zero_std": 0.0, + "grad_norm": 8.09774398803711, + "learning_rate": 9.95e-07, + "loss": 0.20567649602890015, + "num_tokens": 16404.0, + "reward": -12.422499656677246, + "reward_std": 7.134707450866699, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -12.522500038146973, + "rewards/supergames_reward/std": 7.134707927703857, + "step": 2, + "step_time": 9.960156448010821 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 287.0, + "completions/max_terminated_length": 287.0, + "completions/mean_length": 156.625, + "completions/mean_terminated_length": 156.625, + "completions/min_length": 77.0, + "completions/min_terminated_length": 77.0, + "entropy": 0.562222421169281, + "epoch": 0.015, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.9053616523742676, + "learning_rate": 9.9e-07, + "loss": 0.09602774679660797, + "num_tokens": 26953.0, + "reward": -74.94999694824219, + "reward_std": 70.73094177246094, + "rewards/json_format_reward/mean": 0.05000000074505806, + "rewards/json_format_reward/std": 0.09258200973272324, + "rewards/supergames_reward/mean": -75.0, + "rewards/supergames_reward/std": 70.71067810058594, + "step": 3, + "step_time": 13.686320498993155 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 268.0, + "completions/max_terminated_length": 268.0, + "completions/mean_length": 154.125, + "completions/mean_terminated_length": 154.125, + "completions/min_length": 53.0, + "completions/min_terminated_length": 53.0, + "entropy": 0.7323317527770996, + "epoch": 0.02, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.933310031890869, + "learning_rate": 9.849999999999999e-07, + "loss": 0.28249427676200867, + "num_tokens": 36514.0, + "reward": -40.7599983215332, + "reward_std": 81.73140716552734, + "rewards/json_format_reward/mean": 0.05000000074505806, + "rewards/json_format_reward/std": 0.09258200973272324, + "rewards/supergames_reward/mean": -40.80999755859375, + "rewards/supergames_reward/std": 81.68995666503906, + "step": 4, + "step_time": 12.805880262021674 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 401.0, + "completions/max_terminated_length": 401.0, + "completions/mean_length": 205.25, + "completions/mean_terminated_length": 205.25, + "completions/min_length": 115.0, + "completions/min_terminated_length": 115.0, + "entropy": 0.6167430281639099, + "epoch": 0.025, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.344135046005249, + "learning_rate": 9.8e-07, + "loss": 0.1433994024991989, + "num_tokens": 47476.0, + "reward": -97.6612548828125, + "reward_std": 6.3321428298950195, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -97.76124572753906, + "rewards/supergames_reward/std": 6.3321428298950195, + "step": 5, + "step_time": 18.789364666008623 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 276.0, + "completions/max_terminated_length": 276.0, + "completions/mean_length": 167.25, + "completions/mean_terminated_length": 167.25, + "completions/min_length": 73.0, + "completions/min_terminated_length": 73.0, + "entropy": 0.7583790421485901, + "epoch": 0.03, + "frac_reward_zero_std": 0.0, + "grad_norm": 5.172371864318848, + "learning_rate": 9.75e-07, + "loss": -0.12127404659986496, + "num_tokens": 55246.0, + "reward": -58.687503814697266, + "reward_std": 58.64059829711914, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": -58.76250457763672, + "rewards/supergames_reward/std": 58.620460510253906, + "step": 6, + "step_time": 12.948570273991209 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 261.0, + "completions/max_terminated_length": 261.0, + "completions/mean_length": 133.875, + "completions/mean_terminated_length": 133.875, + "completions/min_length": 57.0, + "completions/min_terminated_length": 57.0, + "entropy": 0.7447654604911804, + "epoch": 0.035, + "frac_reward_zero_std": 0.0, + "grad_norm": 6.428595066070557, + "learning_rate": 9.7e-07, + "loss": 0.10135584324598312, + "num_tokens": 60837.0, + "reward": -32.45000076293945, + "reward_std": 41.754486083984375, + "rewards/json_format_reward/mean": 0.05000000074505806, + "rewards/json_format_reward/std": 0.09258200973272324, + "rewards/supergames_reward/mean": -32.5, + "rewards/supergames_reward/std": 41.66190719604492, + "step": 7, + "step_time": 12.058315072004916 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 414.0, + "completions/max_terminated_length": 414.0, + "completions/mean_length": 209.125, + "completions/mean_terminated_length": 209.125, + "completions/min_length": 104.0, + "completions/min_terminated_length": 104.0, + "entropy": 0.3860666751861572, + "epoch": 0.04, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.806255578994751, + "learning_rate": 9.649999999999999e-07, + "loss": -0.33238139748573303, + "num_tokens": 70838.0, + "reward": -80.50375366210938, + "reward_std": 38.82380294799805, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -80.60375213623047, + "rewards/supergames_reward/std": 38.82379913330078, + "step": 8, + "step_time": 19.323370319994865 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.25, + "completions/max_length": 512.0, + "completions/max_terminated_length": 199.0, + "completions/mean_length": 247.125, + "completions/mean_terminated_length": 158.83334350585938, + "completions/min_length": 106.0, + "completions/min_terminated_length": 106.0, + "entropy": 0.5816237330436707, + "epoch": 0.045, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.3209943771362305, + "learning_rate": 9.6e-07, + "loss": 0.36752766370773315, + "num_tokens": 82135.0, + "reward": -27.688751220703125, + "reward_std": 82.8358154296875, + "rewards/json_format_reward/mean": 0.05000000074505806, + "rewards/json_format_reward/std": 0.09258200973272324, + "rewards/supergames_reward/mean": -27.738750457763672, + "rewards/supergames_reward/std": 82.78590393066406, + "step": 9, + "step_time": 21.948575104994234 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 240.0, + "completions/max_terminated_length": 240.0, + "completions/mean_length": 153.0, + "completions/mean_terminated_length": 153.0, + "completions/min_length": 51.0, + "completions/min_terminated_length": 51.0, + "entropy": 0.6301521062850952, + "epoch": 0.05, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.125359058380127, + "learning_rate": 9.55e-07, + "loss": 0.06129350885748863, + "num_tokens": 89807.0, + "reward": -78.39624786376953, + "reward_std": 40.17042922973633, + "rewards/json_format_reward/mean": 0.05000000074505806, + "rewards/json_format_reward/std": 0.09258200973272324, + "rewards/supergames_reward/mean": -78.44625091552734, + "rewards/supergames_reward/std": 40.1396484375, + "step": 10, + "step_time": 9.098202993016457 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 133.0, + "completions/max_terminated_length": 133.0, + "completions/mean_length": 112.0, + "completions/mean_terminated_length": 112.0, + "completions/min_length": 68.0, + "completions/min_terminated_length": 68.0, + "entropy": 0.5521990656852722, + "epoch": 0.055, + "frac_reward_zero_std": 0.0, + "grad_norm": 6.616060256958008, + "learning_rate": 9.499999999999999e-07, + "loss": -0.006313305348157883, + "num_tokens": 95231.0, + "reward": -12.161249160766602, + "reward_std": 6.39578104019165, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -12.261249542236328, + "rewards/supergames_reward/std": 6.39578104019165, + "step": 11, + "step_time": 5.136311663984088 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 396.0, + "completions/max_terminated_length": 396.0, + "completions/mean_length": 214.875, + "completions/mean_terminated_length": 214.875, + "completions/min_length": 55.0, + "completions/min_terminated_length": 55.0, + "entropy": 0.5505905151367188, + "epoch": 0.06, + "frac_reward_zero_std": 0.0, + "grad_norm": 5.289583206176758, + "learning_rate": 9.45e-07, + "loss": -0.29802003502845764, + "num_tokens": 106238.0, + "reward": -91.66999816894531, + "reward_std": 23.2779541015625, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -91.77000427246094, + "rewards/supergames_reward/std": 23.277956008911133, + "step": 12, + "step_time": 14.945365622988902 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 269.0, + "completions/max_terminated_length": 269.0, + "completions/mean_length": 167.375, + "completions/mean_terminated_length": 167.375, + "completions/min_length": 113.0, + "completions/min_terminated_length": 113.0, + "entropy": 0.540415346622467, + "epoch": 0.065, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.858291149139404, + "learning_rate": 9.399999999999999e-07, + "loss": 0.24004128575325012, + "num_tokens": 116937.0, + "reward": -20.38249969482422, + "reward_std": 87.55204010009766, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": -20.457500457763672, + "rewards/supergames_reward/std": 87.52605438232422, + "step": 13, + "step_time": 10.482285185018554 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 254.0, + "completions/max_terminated_length": 254.0, + "completions/mean_length": 156.875, + "completions/mean_terminated_length": 156.875, + "completions/min_length": 114.0, + "completions/min_terminated_length": 114.0, + "entropy": 0.7237679958343506, + "epoch": 0.07, + "frac_reward_zero_std": 0.0, + "grad_norm": 5.083980560302734, + "learning_rate": 9.35e-07, + "loss": -0.19025824964046478, + "num_tokens": 124640.0, + "reward": -51.397499084472656, + "reward_std": 54.486454010009766, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -51.49749755859375, + "rewards/supergames_reward/std": 54.486454010009766, + "step": 14, + "step_time": 9.597231683001155 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 236.0, + "completions/max_terminated_length": 236.0, + "completions/mean_length": 141.375, + "completions/mean_terminated_length": 141.375, + "completions/min_length": 91.0, + "completions/min_terminated_length": 91.0, + "entropy": 0.754334568977356, + "epoch": 0.075, + "frac_reward_zero_std": 0.0, + "grad_norm": 5.720428943634033, + "learning_rate": 9.3e-07, + "loss": -0.16830675303936005, + "num_tokens": 132219.0, + "reward": -56.226253509521484, + "reward_std": 60.115657806396484, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -56.32625198364258, + "rewards/supergames_reward/std": 60.11566162109375, + "step": 15, + "step_time": 8.998362872982398 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 209.0, + "completions/max_terminated_length": 209.0, + "completions/mean_length": 136.375, + "completions/mean_terminated_length": 136.375, + "completions/min_length": 97.0, + "completions/min_terminated_length": 97.0, + "entropy": 0.6280568838119507, + "epoch": 0.08, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.9681806564331055, + "learning_rate": 9.25e-07, + "loss": -0.01910916529595852, + "num_tokens": 139726.0, + "reward": -68.24000549316406, + "reward_std": 52.238914489746094, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -68.33999633789062, + "rewards/supergames_reward/std": 52.238914489746094, + "step": 16, + "step_time": 8.034480294008972 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 365.0, + "completions/max_terminated_length": 365.0, + "completions/mean_length": 211.875, + "completions/mean_terminated_length": 211.875, + "completions/min_length": 100.0, + "completions/min_terminated_length": 100.0, + "entropy": 0.899176836013794, + "epoch": 0.085, + "frac_reward_zero_std": 0.0, + "grad_norm": 6.691746711730957, + "learning_rate": 9.2e-07, + "loss": 0.25513291358947754, + "num_tokens": 147869.0, + "reward": -99.92500305175781, + "reward_std": 0.07070883363485336, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": -100.0, + "rewards/supergames_reward/std": 0.0, + "step": 17, + "step_time": 13.655792869016295 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 387.0, + "completions/max_terminated_length": 387.0, + "completions/mean_length": 238.0, + "completions/mean_terminated_length": 238.0, + "completions/min_length": 87.0, + "completions/min_terminated_length": 87.0, + "entropy": 0.5596873760223389, + "epoch": 0.09, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.6711347103118896, + "learning_rate": 9.15e-07, + "loss": -0.2628335654735565, + "num_tokens": 159101.0, + "reward": -24.900001525878906, + "reward_std": 103.50983428955078, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -25.0, + "rewards/supergames_reward/std": 103.50983428955078, + "step": 18, + "step_time": 14.784329420013819 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.125, + "completions/max_length": 512.0, + "completions/max_terminated_length": 277.0, + "completions/mean_length": 219.75, + "completions/mean_terminated_length": 178.00001525878906, + "completions/min_length": 90.0, + "completions/min_terminated_length": 90.0, + "entropy": 0.9245517253875732, + "epoch": 0.095, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.55502986907959, + "learning_rate": 9.1e-07, + "loss": 0.13060179352760315, + "num_tokens": 167299.0, + "reward": -86.89125061035156, + "reward_std": 36.95603561401367, + "rewards/json_format_reward/mean": 0.05000000074505806, + "rewards/json_format_reward/std": 0.09258200973272324, + "rewards/supergames_reward/mean": -86.94125366210938, + "rewards/supergames_reward/std": 36.93572235107422, + "step": 19, + "step_time": 18.832684700988466 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 267.0, + "completions/max_terminated_length": 267.0, + "completions/mean_length": 177.25, + "completions/mean_terminated_length": 177.25, + "completions/min_length": 93.0, + "completions/min_terminated_length": 93.0, + "entropy": 0.7331254482269287, + "epoch": 0.1, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.538497447967529, + "learning_rate": 9.05e-07, + "loss": -0.16418980062007904, + "num_tokens": 175141.0, + "reward": -68.52874755859375, + "reward_std": 58.48719787597656, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -68.62875366210938, + "rewards/supergames_reward/std": 58.48720169067383, + "step": 20, + "step_time": 10.090975169994636 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 300.0, + "completions/max_terminated_length": 300.0, + "completions/mean_length": 190.625, + "completions/mean_terminated_length": 190.625, + "completions/min_length": 103.0, + "completions/min_terminated_length": 103.0, + "entropy": 0.6659425497055054, + "epoch": 0.105, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.485373497009277, + "learning_rate": 9e-07, + "loss": -0.10106432437896729, + "num_tokens": 183098.0, + "reward": -71.48875427246094, + "reward_std": 55.33565139770508, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -71.5887451171875, + "rewards/supergames_reward/std": 55.33565139770508, + "step": 21, + "step_time": 11.255001295008697 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 279.0, + "completions/max_terminated_length": 279.0, + "completions/mean_length": 179.0, + "completions/mean_terminated_length": 179.0, + "completions/min_length": 103.0, + "completions/min_terminated_length": 103.0, + "entropy": 0.49365949630737305, + "epoch": 0.11, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.829419136047363, + "learning_rate": 8.95e-07, + "loss": 0.10587231814861298, + "num_tokens": 192826.0, + "reward": -75.62000274658203, + "reward_std": 45.01959991455078, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": -75.69499969482422, + "rewards/supergames_reward/std": 45.00411605834961, + "step": 22, + "step_time": 10.611246599000879 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 315.0, + "completions/max_terminated_length": 315.0, + "completions/mean_length": 212.625, + "completions/mean_terminated_length": 212.625, + "completions/min_length": 100.0, + "completions/min_terminated_length": 100.0, + "entropy": 0.43159469962120056, + "epoch": 0.115, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.541125774383545, + "learning_rate": 8.9e-07, + "loss": -0.1681259125471115, + "num_tokens": 202815.0, + "reward": 23.67624855041504, + "reward_std": 67.74340057373047, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 23.576250076293945, + "rewards/supergames_reward/std": 67.743408203125, + "step": 23, + "step_time": 11.731771531980485 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 168.0, + "completions/max_terminated_length": 168.0, + "completions/mean_length": 101.625, + "completions/mean_terminated_length": 101.625, + "completions/min_length": 65.0, + "completions/min_terminated_length": 65.0, + "entropy": 0.3972298502922058, + "epoch": 0.12, + "frac_reward_zero_std": 0.0, + "grad_norm": 5.8285417556762695, + "learning_rate": 8.85e-07, + "loss": -0.011585958302021027, + "num_tokens": 208188.0, + "reward": 58.849998474121094, + "reward_std": 56.9304084777832, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 58.75, + "rewards/supergames_reward/std": 56.9304084777832, + "step": 24, + "step_time": 6.187504950998118 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 303.0, + "completions/max_terminated_length": 303.0, + "completions/mean_length": 165.375, + "completions/mean_terminated_length": 165.375, + "completions/min_length": 69.0, + "completions/min_terminated_length": 69.0, + "entropy": 0.6339899301528931, + "epoch": 0.125, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.823799133300781, + "learning_rate": 8.799999999999999e-07, + "loss": 0.05159185826778412, + "num_tokens": 215935.0, + "reward": -23.521251678466797, + "reward_std": 74.23067474365234, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -23.621248245239258, + "rewards/supergames_reward/std": 74.23067474365234, + "step": 25, + "step_time": 11.059416300005978 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 349.0, + "completions/max_terminated_length": 349.0, + "completions/mean_length": 204.0, + "completions/mean_terminated_length": 204.0, + "completions/min_length": 143.0, + "completions/min_terminated_length": 143.0, + "entropy": 0.6462895274162292, + "epoch": 0.13, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.969292640686035, + "learning_rate": 8.75e-07, + "loss": 0.0115677984431386, + "num_tokens": 223999.0, + "reward": -74.8175048828125, + "reward_std": 40.94221115112305, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": -74.89250183105469, + "rewards/supergames_reward/std": 40.924625396728516, + "step": 26, + "step_time": 12.666237785975682 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 299.0, + "completions/max_terminated_length": 299.0, + "completions/mean_length": 166.5, + "completions/mean_terminated_length": 166.5, + "completions/min_length": 87.0, + "completions/min_terminated_length": 87.0, + "entropy": 0.4756850600242615, + "epoch": 0.135, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.915069103240967, + "learning_rate": 8.699999999999999e-07, + "loss": -0.13321346044540405, + "num_tokens": 233643.0, + "reward": -52.125, + "reward_std": 19.405282974243164, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": -52.20000076293945, + "rewards/supergames_reward/std": 19.33464813232422, + "step": 27, + "step_time": 11.172539752995363 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 184.0, + "completions/max_terminated_length": 184.0, + "completions/mean_length": 137.125, + "completions/mean_terminated_length": 137.125, + "completions/min_length": 100.0, + "completions/min_terminated_length": 100.0, + "entropy": 0.5834778547286987, + "epoch": 0.14, + "frac_reward_zero_std": 0.0, + "grad_norm": 5.146666526794434, + "learning_rate": 8.65e-07, + "loss": 0.060454584658145905, + "num_tokens": 241204.0, + "reward": -68.99250793457031, + "reward_std": 51.010581970214844, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -69.09249877929688, + "rewards/supergames_reward/std": 51.010581970214844, + "step": 28, + "step_time": 7.025046669004951 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 331.0, + "completions/max_terminated_length": 331.0, + "completions/mean_length": 182.375, + "completions/mean_terminated_length": 182.375, + "completions/min_length": 106.0, + "completions/min_terminated_length": 106.0, + "entropy": 0.6368551254272461, + "epoch": 0.145, + "frac_reward_zero_std": 0.0, + "grad_norm": 5.15535306930542, + "learning_rate": 8.599999999999999e-07, + "loss": -0.3323673903942108, + "num_tokens": 249087.0, + "reward": -44.17000198364258, + "reward_std": 80.5373306274414, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -44.27000045776367, + "rewards/supergames_reward/std": 80.5373306274414, + "step": 29, + "step_time": 12.079259724996518 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 190.0, + "completions/max_terminated_length": 190.0, + "completions/mean_length": 127.75, + "completions/mean_terminated_length": 127.75, + "completions/min_length": 80.0, + "completions/min_terminated_length": 80.0, + "entropy": 0.4491196870803833, + "epoch": 0.15, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.896186828613281, + "learning_rate": 8.55e-07, + "loss": -0.12887656688690186, + "num_tokens": 259445.0, + "reward": 57.446250915527344, + "reward_std": 38.78837585449219, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 57.34625244140625, + "rewards/supergames_reward/std": 38.78837966918945, + "step": 30, + "step_time": 7.483972909016302 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 166.0, + "completions/max_terminated_length": 166.0, + "completions/mean_length": 112.875, + "completions/mean_terminated_length": 112.875, + "completions/min_length": 85.0, + "completions/min_terminated_length": 85.0, + "entropy": 0.4858454465866089, + "epoch": 0.155, + "frac_reward_zero_std": 0.0, + "grad_norm": 6.742906093597412, + "learning_rate": 8.499999999999999e-07, + "loss": 0.06930096447467804, + "num_tokens": 264924.0, + "reward": -21.174999237060547, + "reward_std": 31.89051628112793, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": -21.25, + "rewards/supergames_reward/std": 31.819807052612305, + "step": 31, + "step_time": 6.1866529019898735 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 148.0, + "completions/max_terminated_length": 148.0, + "completions/mean_length": 118.125, + "completions/mean_terminated_length": 118.125, + "completions/min_length": 98.0, + "completions/min_terminated_length": 98.0, + "entropy": 0.45428749918937683, + "epoch": 0.16, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.663414478302002, + "learning_rate": 8.45e-07, + "loss": 0.05735419690608978, + "num_tokens": 270381.0, + "reward": 0.4687504768371582, + "reward_std": 38.66371154785156, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 0.36875057220458984, + "rewards/supergames_reward/std": 38.6637077331543, + "step": 32, + "step_time": 5.577042757999152 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 345.0, + "completions/max_terminated_length": 345.0, + "completions/mean_length": 166.875, + "completions/mean_terminated_length": 166.875, + "completions/min_length": 86.0, + "completions/min_terminated_length": 86.0, + "entropy": 0.6107293367385864, + "epoch": 0.165, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.75006628036499, + "learning_rate": 8.399999999999999e-07, + "loss": -0.14760127663612366, + "num_tokens": 278156.0, + "reward": -87.84750366210938, + "reward_std": 26.84708023071289, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -87.94749450683594, + "rewards/supergames_reward/std": 26.847076416015625, + "step": 33, + "step_time": 12.552876825997373 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.125, + "completions/max_length": 512.0, + "completions/max_terminated_length": 360.0, + "completions/mean_length": 231.0, + "completions/mean_terminated_length": 190.85714721679688, + "completions/min_length": 74.0, + "completions/min_terminated_length": 74.0, + "entropy": 0.6089653372764587, + "epoch": 0.17, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.781267166137695, + "learning_rate": 8.349999999999999e-07, + "loss": 0.4294504225254059, + "num_tokens": 289364.0, + "reward": -99.92500305175781, + "reward_std": 0.07070883363485336, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": -100.0, + "rewards/supergames_reward/std": 0.0, + "step": 34, + "step_time": 18.961819477990502 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 218.0, + "completions/max_terminated_length": 218.0, + "completions/mean_length": 134.125, + "completions/mean_terminated_length": 134.125, + "completions/min_length": 104.0, + "completions/min_terminated_length": 104.0, + "entropy": 0.46158918738365173, + "epoch": 0.175, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.016353607177734, + "learning_rate": 8.299999999999999e-07, + "loss": 0.09251818805932999, + "num_tokens": 298733.0, + "reward": -35.558753967285156, + "reward_std": 15.881204605102539, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -35.65875244140625, + "rewards/supergames_reward/std": 15.881203651428223, + "step": 35, + "step_time": 8.38491243700264 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 207.0, + "completions/max_terminated_length": 207.0, + "completions/mean_length": 156.875, + "completions/mean_terminated_length": 156.875, + "completions/min_length": 111.0, + "completions/min_terminated_length": 111.0, + "entropy": 0.5104788541793823, + "epoch": 0.18, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.592893123626709, + "learning_rate": 8.249999999999999e-07, + "loss": -0.0410832017660141, + "num_tokens": 306420.0, + "reward": 5.4212493896484375, + "reward_std": 93.50404357910156, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": 5.346250534057617, + "rewards/supergames_reward/std": 93.4718246459961, + "step": 36, + "step_time": 7.811868985998444 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 241.0, + "completions/max_terminated_length": 241.0, + "completions/mean_length": 178.0, + "completions/mean_terminated_length": 178.0, + "completions/min_length": 106.0, + "completions/min_terminated_length": 106.0, + "entropy": 0.7527878880500793, + "epoch": 0.185, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.8812546730041504, + "learning_rate": 8.199999999999999e-07, + "loss": 0.12539231777191162, + "num_tokens": 314252.0, + "reward": -89.58500671386719, + "reward_std": 29.256120681762695, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": -89.66000366210938, + "rewards/supergames_reward/std": 29.24593734741211, + "step": 37, + "step_time": 8.959742914012168 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 199.0, + "completions/max_terminated_length": 199.0, + "completions/mean_length": 133.375, + "completions/mean_terminated_length": 133.375, + "completions/min_length": 87.0, + "completions/min_terminated_length": 87.0, + "entropy": 0.3347901403903961, + "epoch": 0.19, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.205414295196533, + "learning_rate": 8.149999999999999e-07, + "loss": -0.06991486996412277, + "num_tokens": 319839.0, + "reward": -14.498749732971191, + "reward_std": 13.007229804992676, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -14.598750114440918, + "rewards/supergames_reward/std": 13.007229804992676, + "step": 38, + "step_time": 7.377183554985095 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 181.0, + "completions/max_terminated_length": 181.0, + "completions/mean_length": 114.375, + "completions/mean_terminated_length": 114.375, + "completions/min_length": 58.0, + "completions/min_terminated_length": 58.0, + "entropy": 0.4445965886116028, + "epoch": 0.195, + "frac_reward_zero_std": 0.0, + "grad_norm": 5.561298847198486, + "learning_rate": 8.1e-07, + "loss": -0.06235164776444435, + "num_tokens": 325290.0, + "reward": -7.425000190734863, + "reward_std": 53.701602935791016, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": -7.5, + "rewards/supergames_reward/std": 53.652320861816406, + "step": 39, + "step_time": 6.753862089011818 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 370.0, + "completions/max_terminated_length": 370.0, + "completions/mean_length": 187.625, + "completions/mean_terminated_length": 187.625, + "completions/min_length": 60.0, + "completions/min_terminated_length": 60.0, + "entropy": 0.3950234353542328, + "epoch": 0.2, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.5988712310791016, + "learning_rate": 8.05e-07, + "loss": -0.1381440907716751, + "num_tokens": 335119.0, + "reward": -33.68375015258789, + "reward_std": 42.069435119628906, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -33.783748626708984, + "rewards/supergames_reward/std": 42.069435119628906, + "step": 40, + "step_time": 13.718958162004128 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.125, + "completions/max_length": 512.0, + "completions/max_terminated_length": 371.0, + "completions/mean_length": 251.125, + "completions/mean_terminated_length": 213.85714721679688, + "completions/min_length": 121.0, + "completions/min_terminated_length": 121.0, + "entropy": 0.507067084312439, + "epoch": 0.205, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.3898658752441406, + "learning_rate": 8e-07, + "loss": 0.0078964838758111, + "num_tokens": 346464.0, + "reward": -86.48625183105469, + "reward_std": 38.020694732666016, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": -86.56124877929688, + "rewards/supergames_reward/std": 38.010528564453125, + "step": 41, + "step_time": 18.90960379401804 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 152.0, + "completions/max_terminated_length": 152.0, + "completions/mean_length": 116.375, + "completions/mean_terminated_length": 116.375, + "completions/min_length": 93.0, + "completions/min_terminated_length": 93.0, + "entropy": 0.3801707923412323, + "epoch": 0.21, + "frac_reward_zero_std": 1.0, + "grad_norm": 0.0, + "learning_rate": 7.95e-07, + "loss": 0.0, + "num_tokens": 351955.0, + "reward": -9.899999618530273, + "reward_std": 0.0, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -10.0, + "rewards/supergames_reward/std": 0.0, + "step": 42, + "step_time": 5.711630532023264 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 308.0, + "completions/max_terminated_length": 308.0, + "completions/mean_length": 247.375, + "completions/mean_terminated_length": 247.375, + "completions/min_length": 116.0, + "completions/min_terminated_length": 116.0, + "entropy": 0.3185043931007385, + "epoch": 0.215, + "frac_reward_zero_std": 0.0, + "grad_norm": 2.168762445449829, + "learning_rate": 7.9e-07, + "loss": 0.1970764398574829, + "num_tokens": 363278.0, + "reward": -53.89249801635742, + "reward_std": 63.02619934082031, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -53.99250030517578, + "rewards/supergames_reward/std": 63.02619934082031, + "step": 43, + "step_time": 11.661934906995157 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 330.0, + "completions/max_terminated_length": 330.0, + "completions/mean_length": 212.5, + "completions/mean_terminated_length": 212.5, + "completions/min_length": 106.0, + "completions/min_terminated_length": 106.0, + "entropy": 0.4446268379688263, + "epoch": 0.22, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.305844306945801, + "learning_rate": 7.85e-07, + "loss": -0.06735197454690933, + "num_tokens": 374306.0, + "reward": 18.059999465942383, + "reward_std": 98.9417953491211, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 17.959999084472656, + "rewards/supergames_reward/std": 98.9417953491211, + "step": 44, + "step_time": 12.492453911982011 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 316.0, + "completions/max_terminated_length": 316.0, + "completions/mean_length": 183.0, + "completions/mean_terminated_length": 183.0, + "completions/min_length": 103.0, + "completions/min_terminated_length": 103.0, + "entropy": 0.5269614458084106, + "epoch": 0.225, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.8006370067596436, + "learning_rate": 7.799999999999999e-07, + "loss": -0.08870815485715866, + "num_tokens": 382210.0, + "reward": -54.928749084472656, + "reward_std": 72.5127944946289, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -55.02874755859375, + "rewards/supergames_reward/std": 72.5127944946289, + "step": 45, + "step_time": 11.7636467939883 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 244.0, + "completions/max_terminated_length": 244.0, + "completions/mean_length": 136.625, + "completions/mean_terminated_length": 136.625, + "completions/min_length": 85.0, + "completions/min_terminated_length": 85.0, + "entropy": 0.34820204973220825, + "epoch": 0.23, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.911736011505127, + "learning_rate": 7.75e-07, + "loss": 0.04397330805659294, + "num_tokens": 391583.0, + "reward": -19.10375213623047, + "reward_std": 23.967702865600586, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -19.203750610351562, + "rewards/supergames_reward/std": 23.967702865600586, + "step": 46, + "step_time": 10.119426151999505 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 218.0, + "completions/max_terminated_length": 218.0, + "completions/mean_length": 160.375, + "completions/mean_terminated_length": 160.375, + "completions/min_length": 93.0, + "completions/min_terminated_length": 93.0, + "entropy": 0.611914336681366, + "epoch": 0.235, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.571913242340088, + "learning_rate": 7.699999999999999e-07, + "loss": -0.044293053448200226, + "num_tokens": 399290.0, + "reward": -25.186250686645508, + "reward_std": 71.46305084228516, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -25.286251068115234, + "rewards/supergames_reward/std": 71.46305084228516, + "step": 47, + "step_time": 8.563868903991533 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 167.0, + "completions/max_terminated_length": 167.0, + "completions/mean_length": 111.125, + "completions/mean_terminated_length": 111.125, + "completions/min_length": 56.0, + "completions/min_terminated_length": 56.0, + "entropy": 0.46585777401924133, + "epoch": 0.24, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.654109954833984, + "learning_rate": 7.65e-07, + "loss": -0.035151124000549316, + "num_tokens": 404707.0, + "reward": -0.408750057220459, + "reward_std": 31.939748764038086, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -0.5087499618530273, + "rewards/supergames_reward/std": 31.93975257873535, + "step": 48, + "step_time": 6.438482580007985 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 306.0, + "completions/max_terminated_length": 306.0, + "completions/mean_length": 216.625, + "completions/mean_terminated_length": 216.625, + "completions/min_length": 132.0, + "completions/min_terminated_length": 132.0, + "entropy": 0.33054471015930176, + "epoch": 0.245, + "frac_reward_zero_std": 0.0, + "grad_norm": 2.700303792953491, + "learning_rate": 7.599999999999999e-07, + "loss": -0.04393656551837921, + "num_tokens": 414736.0, + "reward": -97.26000213623047, + "reward_std": 1.6294406652450562, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -97.36000061035156, + "rewards/supergames_reward/std": 1.629441738128662, + "step": 49, + "step_time": 11.636040106008295 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 378.0, + "completions/max_terminated_length": 378.0, + "completions/mean_length": 218.375, + "completions/mean_terminated_length": 218.375, + "completions/min_length": 104.0, + "completions/min_terminated_length": 104.0, + "entropy": 0.48711177706718445, + "epoch": 0.25, + "frac_reward_zero_std": 0.0, + "grad_norm": 2.9650564193725586, + "learning_rate": 7.55e-07, + "loss": -0.3529520332813263, + "num_tokens": 424795.0, + "reward": 47.408748626708984, + "reward_std": 72.72083282470703, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 47.30875015258789, + "rewards/supergames_reward/std": 72.72083282470703, + "step": 50, + "step_time": 14.258096842997475 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.125, + "completions/max_length": 512.0, + "completions/max_terminated_length": 492.0, + "completions/mean_length": 267.625, + "completions/mean_terminated_length": 232.71429443359375, + "completions/min_length": 90.0, + "completions/min_terminated_length": 90.0, + "entropy": 0.4482860863208771, + "epoch": 0.255, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.2998197078704834, + "learning_rate": 7.5e-07, + "loss": 0.2493654191493988, + "num_tokens": 436288.0, + "reward": 50.61125183105469, + "reward_std": 75.3349838256836, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": 50.53624725341797, + "rewards/supergames_reward/std": 75.27783966064453, + "step": 51, + "step_time": 19.20827590499539 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 176.0, + "completions/max_terminated_length": 176.0, + "completions/mean_length": 144.5, + "completions/mean_terminated_length": 144.5, + "completions/min_length": 99.0, + "completions/min_terminated_length": 99.0, + "entropy": 0.5378735661506653, + "epoch": 0.26, + "frac_reward_zero_std": 0.0, + "grad_norm": 5.114461898803711, + "learning_rate": 7.45e-07, + "loss": 0.0036597400903701782, + "num_tokens": 441980.0, + "reward": -15.20625114440918, + "reward_std": 45.35526657104492, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": -15.28125, + "rewards/supergames_reward/std": 45.301815032958984, + "step": 52, + "step_time": 6.660627231001854 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 136.0, + "completions/max_terminated_length": 136.0, + "completions/mean_length": 119.0, + "completions/mean_terminated_length": 119.0, + "completions/min_length": 105.0, + "completions/min_terminated_length": 105.0, + "entropy": 0.41796669363975525, + "epoch": 0.265, + "frac_reward_zero_std": 0.0, + "grad_norm": 5.059505462646484, + "learning_rate": 7.4e-07, + "loss": -0.014973883517086506, + "num_tokens": 447468.0, + "reward": 15.170000076293945, + "reward_std": 46.658329010009766, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 15.070000648498535, + "rewards/supergames_reward/std": 46.658329010009766, + "step": 53, + "step_time": 5.271571868011961 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 259.0, + "completions/max_terminated_length": 259.0, + "completions/mean_length": 191.125, + "completions/mean_terminated_length": 191.125, + "completions/min_length": 91.0, + "completions/min_terminated_length": 91.0, + "entropy": 0.45421987771987915, + "epoch": 0.27, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.4454867839813232, + "learning_rate": 7.35e-07, + "loss": -0.1128973588347435, + "num_tokens": 457301.0, + "reward": 38.89875030517578, + "reward_std": 62.1772575378418, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 38.79874801635742, + "rewards/supergames_reward/std": 62.1772575378418, + "step": 54, + "step_time": 10.524528659996577 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 319.0, + "completions/max_terminated_length": 319.0, + "completions/mean_length": 233.375, + "completions/mean_terminated_length": 233.375, + "completions/min_length": 136.0, + "completions/min_terminated_length": 136.0, + "entropy": 0.4449213445186615, + "epoch": 0.275, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.2065398693084717, + "learning_rate": 7.3e-07, + "loss": 0.0016277075046673417, + "num_tokens": 467520.0, + "reward": 27.372499465942383, + "reward_std": 81.23126983642578, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 27.27250099182129, + "rewards/supergames_reward/std": 81.23126983642578, + "step": 55, + "step_time": 12.592280682991259 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.125, + "completions/max_length": 512.0, + "completions/max_terminated_length": 357.0, + "completions/mean_length": 261.0, + "completions/mean_terminated_length": 225.1428680419922, + "completions/min_length": 95.0, + "completions/min_terminated_length": 95.0, + "entropy": 0.506920337677002, + "epoch": 0.28, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.058388710021973, + "learning_rate": 7.249999999999999e-07, + "loss": 0.3289242386817932, + "num_tokens": 476056.0, + "reward": 47.05249786376953, + "reward_std": 80.97753143310547, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": 46.977500915527344, + "rewards/supergames_reward/std": 80.9256362915039, + "step": 56, + "step_time": 18.845177220006008 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 250.0, + "completions/max_terminated_length": 250.0, + "completions/mean_length": 192.75, + "completions/mean_terminated_length": 192.75, + "completions/min_length": 122.0, + "completions/min_terminated_length": 122.0, + "entropy": 0.5197336077690125, + "epoch": 0.285, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.230815410614014, + "learning_rate": 7.2e-07, + "loss": -0.15186546742916107, + "num_tokens": 484014.0, + "reward": -64.5050048828125, + "reward_std": 50.35145568847656, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -64.60499572753906, + "rewards/supergames_reward/std": 50.3514518737793, + "step": 57, + "step_time": 9.576232638006331 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 324.0, + "completions/max_terminated_length": 324.0, + "completions/mean_length": 249.0, + "completions/mean_terminated_length": 249.0, + "completions/min_length": 129.0, + "completions/min_terminated_length": 129.0, + "entropy": 0.42913514375686646, + "epoch": 0.29, + "frac_reward_zero_std": 0.0, + "grad_norm": 2.8544528484344482, + "learning_rate": 7.149999999999999e-07, + "loss": -0.056405920535326004, + "num_tokens": 494390.0, + "reward": 40.123748779296875, + "reward_std": 84.164794921875, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 40.02375030517578, + "rewards/supergames_reward/std": 84.164794921875, + "step": 58, + "step_time": 12.385353341000155 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 138.0, + "completions/max_terminated_length": 138.0, + "completions/mean_length": 107.25, + "completions/mean_terminated_length": 107.25, + "completions/min_length": 81.0, + "completions/min_terminated_length": 81.0, + "entropy": 0.34386488795280457, + "epoch": 0.295, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.6915154457092285, + "learning_rate": 7.1e-07, + "loss": 0.0865321159362793, + "num_tokens": 499816.0, + "reward": -8.244999885559082, + "reward_std": 4.681046962738037, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -8.345000267028809, + "rewards/supergames_reward/std": 4.681046962738037, + "step": 59, + "step_time": 5.42325339600211 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 187.0, + "completions/max_terminated_length": 187.0, + "completions/mean_length": 127.25, + "completions/mean_terminated_length": 127.25, + "completions/min_length": 97.0, + "completions/min_terminated_length": 97.0, + "entropy": 0.5430045127868652, + "epoch": 0.3, + "frac_reward_zero_std": 0.0, + "grad_norm": 5.635311126708984, + "learning_rate": 7.049999999999999e-07, + "loss": -0.05412375554442406, + "num_tokens": 505410.0, + "reward": 20.005001068115234, + "reward_std": 55.156005859375, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 19.904998779296875, + "rewards/supergames_reward/std": 55.156005859375, + "step": 60, + "step_time": 7.068010902003152 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 286.0, + "completions/max_terminated_length": 286.0, + "completions/mean_length": 192.125, + "completions/mean_terminated_length": 192.125, + "completions/min_length": 95.0, + "completions/min_terminated_length": 95.0, + "entropy": 0.4213869869709015, + "epoch": 0.305, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.4664618968963623, + "learning_rate": 7e-07, + "loss": -0.0495159812271595, + "num_tokens": 516283.0, + "reward": -47.212501525878906, + "reward_std": 91.2253646850586, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -47.3125, + "rewards/supergames_reward/std": 91.2253646850586, + "step": 61, + "step_time": 11.176304235996213 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 324.0, + "completions/max_terminated_length": 324.0, + "completions/mean_length": 171.0, + "completions/mean_terminated_length": 171.0, + "completions/min_length": 92.0, + "completions/min_terminated_length": 92.0, + "entropy": 0.4626811146736145, + "epoch": 0.31, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.949252128601074, + "learning_rate": 6.949999999999999e-07, + "loss": 0.192047581076622, + "num_tokens": 522251.0, + "reward": -15.092500686645508, + "reward_std": 49.82962417602539, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": -15.167499542236328, + "rewards/supergames_reward/std": 49.78091049194336, + "step": 62, + "step_time": 12.075224861997413 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.125, + "completions/max_length": 512.0, + "completions/max_terminated_length": 429.0, + "completions/mean_length": 283.5, + "completions/mean_terminated_length": 250.85714721679688, + "completions/min_length": 85.0, + "completions/min_terminated_length": 85.0, + "entropy": 0.4814216196537018, + "epoch": 0.315, + "frac_reward_zero_std": 0.0, + "grad_norm": 2.7918753623962402, + "learning_rate": 6.9e-07, + "loss": 0.15706156194210052, + "num_tokens": 533815.0, + "reward": -20.86625099182129, + "reward_std": 85.99250030517578, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": -20.94124984741211, + "rewards/supergames_reward/std": 85.96620178222656, + "step": 63, + "step_time": 19.448832260008203 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 189.0, + "completions/max_terminated_length": 189.0, + "completions/mean_length": 131.25, + "completions/mean_terminated_length": 131.25, + "completions/min_length": 106.0, + "completions/min_terminated_length": 106.0, + "entropy": 0.4929414987564087, + "epoch": 0.32, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.774101734161377, + "learning_rate": 6.85e-07, + "loss": -0.060691747814416885, + "num_tokens": 539465.0, + "reward": -21.0674991607666, + "reward_std": 6.892730236053467, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -21.167499542236328, + "rewards/supergames_reward/std": 6.892730236053467, + "step": 64, + "step_time": 7.141569407976931 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 454.0, + "completions/max_terminated_length": 454.0, + "completions/mean_length": 281.375, + "completions/mean_terminated_length": 281.375, + "completions/min_length": 107.0, + "completions/min_terminated_length": 107.0, + "entropy": 0.45866572856903076, + "epoch": 0.325, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.092717409133911, + "learning_rate": 6.800000000000001e-07, + "loss": 0.11694959551095963, + "num_tokens": 551012.0, + "reward": 49.938751220703125, + "reward_std": 79.1865005493164, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 49.8387451171875, + "rewards/supergames_reward/std": 79.1865005493164, + "step": 65, + "step_time": 17.06170882002334 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 203.0, + "completions/max_terminated_length": 203.0, + "completions/mean_length": 128.625, + "completions/mean_terminated_length": 128.625, + "completions/min_length": 88.0, + "completions/min_terminated_length": 88.0, + "entropy": 0.5005782246589661, + "epoch": 0.33, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.77616024017334, + "learning_rate": 6.75e-07, + "loss": 0.11242837458848953, + "num_tokens": 556577.0, + "reward": 48.875, + "reward_std": 49.13539123535156, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 48.775001525878906, + "rewards/supergames_reward/std": 49.13539123535156, + "step": 66, + "step_time": 7.695410129002994 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 436.0, + "completions/max_terminated_length": 436.0, + "completions/mean_length": 281.875, + "completions/mean_terminated_length": 281.875, + "completions/min_length": 133.0, + "completions/min_terminated_length": 133.0, + "entropy": 0.507097065448761, + "epoch": 0.335, + "frac_reward_zero_std": 0.0, + "grad_norm": 2.911184787750244, + "learning_rate": 6.7e-07, + "loss": -0.08151137083768845, + "num_tokens": 567120.0, + "reward": -12.886249542236328, + "reward_std": 84.04269409179688, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -12.986251831054688, + "rewards/supergames_reward/std": 84.04269409179688, + "step": 67, + "step_time": 16.409127942984924 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.125, + "completions/max_length": 512.0, + "completions/max_terminated_length": 501.0, + "completions/mean_length": 318.375, + "completions/mean_terminated_length": 290.71429443359375, + "completions/min_length": 65.0, + "completions/min_terminated_length": 65.0, + "entropy": 0.486402690410614, + "epoch": 0.34, + "frac_reward_zero_std": 0.0, + "grad_norm": 2.9461607933044434, + "learning_rate": 6.65e-07, + "loss": 0.2442178875207901, + "num_tokens": 578995.0, + "reward": 2.8387489318847656, + "reward_std": 98.47579956054688, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": 2.7637500762939453, + "rewards/supergames_reward/std": 98.44595336914062, + "step": 68, + "step_time": 19.32364999302081 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 291.0, + "completions/max_terminated_length": 291.0, + "completions/mean_length": 219.625, + "completions/mean_terminated_length": 219.625, + "completions/min_length": 117.0, + "completions/min_terminated_length": 117.0, + "entropy": 0.37856727838516235, + "epoch": 0.345, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.0204732418060303, + "learning_rate": 6.6e-07, + "loss": -0.08764916658401489, + "num_tokens": 589072.0, + "reward": -54.18375015258789, + "reward_std": 71.45012664794922, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -54.283748626708984, + "rewards/supergames_reward/std": 71.45013427734375, + "step": 69, + "step_time": 11.126611230982235 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.125, + "completions/max_length": 512.0, + "completions/max_terminated_length": 305.0, + "completions/mean_length": 254.75, + "completions/mean_terminated_length": 218.00001525878906, + "completions/min_length": 104.0, + "completions/min_terminated_length": 104.0, + "entropy": 0.46423614025115967, + "epoch": 0.35, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.0140154361724854, + "learning_rate": 6.55e-07, + "loss": 0.26797160506248474, + "num_tokens": 600438.0, + "reward": -32.04875183105469, + "reward_std": 69.86957550048828, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": -32.12375259399414, + "rewards/supergames_reward/std": 69.84178161621094, + "step": 70, + "step_time": 19.12617294798838 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 372.0, + "completions/max_terminated_length": 372.0, + "completions/mean_length": 194.125, + "completions/mean_terminated_length": 194.125, + "completions/min_length": 94.0, + "completions/min_terminated_length": 94.0, + "entropy": 0.3427293300628662, + "epoch": 0.355, + "frac_reward_zero_std": 0.0, + "grad_norm": 2.781383991241455, + "learning_rate": 6.5e-07, + "loss": 0.354524701833725, + "num_tokens": 610327.0, + "reward": -6.913749694824219, + "reward_std": 87.56071472167969, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": -6.988750457763672, + "rewards/supergames_reward/std": 87.53033447265625, + "step": 71, + "step_time": 14.041668199002743 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 197.0, + "completions/max_terminated_length": 197.0, + "completions/mean_length": 124.75, + "completions/mean_terminated_length": 124.75, + "completions/min_length": 93.0, + "completions/min_terminated_length": 93.0, + "entropy": 0.4455118477344513, + "epoch": 0.36, + "frac_reward_zero_std": 0.0, + "grad_norm": 5.726032733917236, + "learning_rate": 6.45e-07, + "loss": 0.18206289410591125, + "num_tokens": 615853.0, + "reward": 41.95124816894531, + "reward_std": 60.640811920166016, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 41.85124969482422, + "rewards/supergames_reward/std": 60.640811920166016, + "step": 72, + "step_time": 7.416493425989756 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 323.0, + "completions/max_terminated_length": 323.0, + "completions/mean_length": 147.375, + "completions/mean_terminated_length": 147.375, + "completions/min_length": 94.0, + "completions/min_terminated_length": 94.0, + "entropy": 0.480733186006546, + "epoch": 0.365, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.853621482849121, + "learning_rate": 6.4e-07, + "loss": 0.04558124020695686, + "num_tokens": 621584.0, + "reward": 14.884998321533203, + "reward_std": 46.25857162475586, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 14.78499984741211, + "rewards/supergames_reward/std": 46.25857162475586, + "step": 73, + "step_time": 11.95379540900467 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 283.0, + "completions/max_terminated_length": 283.0, + "completions/mean_length": 188.125, + "completions/mean_terminated_length": 188.125, + "completions/min_length": 116.0, + "completions/min_terminated_length": 116.0, + "entropy": 0.5526013374328613, + "epoch": 0.37, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.729883909225464, + "learning_rate": 6.35e-07, + "loss": 0.06268303096294403, + "num_tokens": 629545.0, + "reward": -68.05750274658203, + "reward_std": 59.44808578491211, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -68.15750122070312, + "rewards/supergames_reward/std": 59.44808578491211, + "step": 74, + "step_time": 10.64124580900534 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 141.0, + "completions/max_terminated_length": 141.0, + "completions/mean_length": 117.375, + "completions/mean_terminated_length": 117.375, + "completions/min_length": 100.0, + "completions/min_terminated_length": 100.0, + "entropy": 0.3744319677352905, + "epoch": 0.375, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.775154113769531, + "learning_rate": 6.3e-07, + "loss": -0.023284845054149628, + "num_tokens": 635052.0, + "reward": 3.9374990463256836, + "reward_std": 61.85771179199219, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": 3.862499237060547, + "rewards/supergames_reward/std": 61.80967712402344, + "step": 75, + "step_time": 5.432648951013107 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 268.0, + "completions/max_terminated_length": 268.0, + "completions/mean_length": 157.0, + "completions/mean_terminated_length": 157.0, + "completions/min_length": 104.0, + "completions/min_terminated_length": 104.0, + "entropy": 0.4822140336036682, + "epoch": 0.38, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.715704917907715, + "learning_rate": 6.249999999999999e-07, + "loss": -0.04918142035603523, + "num_tokens": 644644.0, + "reward": 48.64125061035156, + "reward_std": 78.0205078125, + "rewards/json_format_reward/mean": 0.07500000298023224, + "rewards/json_format_reward/std": 0.0707106813788414, + "rewards/supergames_reward/mean": 48.56624984741211, + "rewards/supergames_reward/std": 77.96605682373047, + "step": 76, + "step_time": 10.348935816989979 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 380.0, + "completions/max_terminated_length": 380.0, + "completions/mean_length": 210.875, + "completions/mean_terminated_length": 210.875, + "completions/min_length": 126.0, + "completions/min_terminated_length": 126.0, + "entropy": 0.45722317695617676, + "epoch": 0.385, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.6057567596435547, + "learning_rate": 6.2e-07, + "loss": 0.058352649211883545, + "num_tokens": 652771.0, + "reward": -8.283750534057617, + "reward_std": 77.68436431884766, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -8.383749008178711, + "rewards/supergames_reward/std": 77.68437194824219, + "step": 77, + "step_time": 14.142948755004909 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 324.0, + "completions/max_terminated_length": 324.0, + "completions/mean_length": 188.25, + "completions/mean_terminated_length": 188.25, + "completions/min_length": 68.0, + "completions/min_terminated_length": 68.0, + "entropy": 0.33380764722824097, + "epoch": 0.39, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.2239863872528076, + "learning_rate": 6.149999999999999e-07, + "loss": -0.04592633992433548, + "num_tokens": 662653.0, + "reward": 11.153749465942383, + "reward_std": 89.88329315185547, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 11.053749084472656, + "rewards/supergames_reward/std": 89.88329315185547, + "step": 78, + "step_time": 12.277474621019792 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 143.0, + "completions/max_terminated_length": 143.0, + "completions/mean_length": 108.875, + "completions/mean_terminated_length": 108.875, + "completions/min_length": 85.0, + "completions/min_terminated_length": 85.0, + "entropy": 0.3518812656402588, + "epoch": 0.395, + "frac_reward_zero_std": 0.0, + "grad_norm": 3.967776298522949, + "learning_rate": 6.1e-07, + "loss": -0.10800496488809586, + "num_tokens": 668076.0, + "reward": 70.33250427246094, + "reward_std": 49.9145393371582, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": 70.23249816894531, + "rewards/supergames_reward/std": 49.91453552246094, + "step": 79, + "step_time": 5.52714040101273 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0, + "completions/max_length": 184.0, + "completions/max_terminated_length": 184.0, + "completions/mean_length": 152.125, + "completions/mean_terminated_length": 152.125, + "completions/min_length": 104.0, + "completions/min_terminated_length": 104.0, + "entropy": 0.4699896574020386, + "epoch": 0.4, + "frac_reward_zero_std": 0.0, + "grad_norm": 4.561161994934082, + "learning_rate": 6.049999999999999e-07, + "loss": 0.09505834430456161, + "num_tokens": 675701.0, + "reward": -20.25625228881836, + "reward_std": 53.78542709350586, + "rewards/json_format_reward/mean": 0.10000000149011612, + "rewards/json_format_reward/std": 0.0, + "rewards/supergames_reward/mean": -20.35624885559082, + "rewards/supergames_reward/std": 53.78542709350586, + "step": 80, + "step_time": 7.155081019998761 + } + ], + "logging_steps": 1, + "max_steps": 200, + "num_input_tokens_seen": 675701, + "num_train_epochs": 1, + "save_steps": 20, + "stateful_callbacks": { + "TrainerControl": { + "args": { + "should_epoch_stop": false, + "should_evaluate": false, + "should_log": false, + "should_save": true, + "should_training_stop": false + }, + "attributes": {} + } + }, + "total_flos": 0.0, + "train_batch_size": 8, + "trial_name": null, + "trial_params": null +} diff --git a/checkpoint-80/training_args.bin b/checkpoint-80/training_args.bin new file mode 100644 index 0000000..730dec0 --- /dev/null +++ b/checkpoint-80/training_args.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:691fe5be2d7d1d56160dae63866d136877a72efb30e15ff7b3249192f998a788 +size 7185 diff --git a/config.json b/config.json new file mode 100644 index 0000000..f57b09b --- /dev/null +++ b/config.json @@ -0,0 +1,61 @@ +{ + "architectures": [ + "Qwen2ForCausalLM" + ], + "attention_dropout": 0.0, + "bos_token_id": null, + "dtype": "float32", + "eos_token_id": 151645, + "hidden_act": "silu", + "hidden_size": 1536, + "initializer_range": 0.02, + "intermediate_size": 8960, + "layer_types": [ + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention", + "full_attention" + ], + "max_position_embeddings": 32768, + "max_window_layers": 21, + "model_type": "qwen2", + "num_attention_heads": 12, + "num_hidden_layers": 28, + "num_key_value_heads": 2, + "pad_token_id": 151643, + "rms_norm_eps": 1e-06, + "rope_parameters": { + "rope_theta": 1000000.0, + "rope_type": "default" + }, + "sliding_window": null, + "tie_word_embeddings": true, + "transformers_version": "5.6.2", + "use_cache": false, + "use_sliding_window": false, + "vocab_size": 151936 +} diff --git a/generation_config.json b/generation_config.json new file mode 100644 index 0000000..1b2bba9 --- /dev/null +++ b/generation_config.json @@ -0,0 +1,13 @@ +{ + "do_sample": true, + "eos_token_id": [ + 151645, + 151643 + ], + "pad_token_id": 151643, + "repetition_penalty": 1.1, + "temperature": 0.7, + "top_k": 20, + "top_p": 0.8, + "transformers_version": "5.6.2" +} diff --git a/model.safetensors b/model.safetensors new file mode 100644 index 0000000..73c5aaa --- /dev/null +++ b/model.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:d768f5bced7c176c364bd8caa06f355a13ba8e84bfa6afc5000b0943d729a1b0 +size 6174895536 diff --git a/tokenizer.json b/tokenizer.json new file mode 100644 index 0000000..34510ff --- /dev/null +++ b/tokenizer.json @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:3fd169731d2cbde95e10bf356d66d5997fd885dd8dbb6fb4684da3f23b2585d8 +size 11421892 diff --git a/tokenizer_config.json b/tokenizer_config.json new file mode 100644 index 0000000..770e41d --- /dev/null +++ b/tokenizer_config.json @@ -0,0 +1,30 @@ +{ + "add_prefix_space": false, + "backend": "tokenizers", + "bos_token": null, + "clean_up_tokenization_spaces": false, + "eos_token": "<|im_end|>", + "errors": "replace", + "extra_special_tokens": [ + "<|im_start|>", + "<|im_end|>", + "<|object_ref_start|>", + "<|object_ref_end|>", + "<|box_start|>", + "<|box_end|>", + "<|quad_start|>", + "<|quad_end|>", + "<|vision_start|>", + "<|vision_end|>", + "<|vision_pad|>", + "<|image_pad|>", + "<|video_pad|>" + ], + "is_local": false, + "local_files_only": false, + "model_max_length": 131072, + "pad_token": "<|endoftext|>", + "split_special_tokens": false, + "tokenizer_class": "Qwen2Tokenizer", + "unk_token": null +} diff --git a/training_args.bin b/training_args.bin new file mode 100644 index 0000000..730dec0 --- /dev/null +++ b/training_args.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:691fe5be2d7d1d56160dae63866d136877a72efb30e15ff7b3249192f998a788 +size 7185 diff --git a/untrained_vs_trained.png b/untrained_vs_trained.png new file mode 100644 index 0000000..345dcda Binary files /dev/null and b/untrained_vs_trained.png differ