Nirbhayhero07/deepsentinel-overseer-small
1
0
Fork 0
Code Issues Pull Requests Actions Projects Releases Wiki Activity

初始化项目，由ModelHub XC社区提供模型

Browse Source 
Model: Nirbhayhero07/deepsentinel-overseer-small
Source: Original Platform
This commit is contained in:
ModelHub XC
2026-09-06 18:29:15 +08:00
commit 42a8f75134
45 changed files with 2445 additions and 0 deletions
Download Patch File Download Diff File Expand all files Collapse all files

54
checkpoint-250/chat_template.jinja Normal file
View File

@@ -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 <tools></tools> XML tags:\n<tools>" }}
{%- for tool in tools %}
{{- "\n" }}
{{- tool | tojson }}
{%- endfor %}
{{- "\n</tools>\n\nFor each function call, return a json object with function name and arguments within <tool_call></tool_call> XML tags:\n<tool_call>\n{\"name\": <function-name>, \"arguments\": <args-json-object>}\n</tool_call><|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<tool_call>\n{"name": "' }}
{{- tool_call.name }}
{{- '", "arguments": ' }}
{{- tool_call.arguments | tojson }}
{{- '}\n</tool_call>' }}
{%- endfor %}
{{- '<|im_end|>\n' }}
{%- elif message.role == "tool" %}
{%- if (loop.index0 == 0) or (messages[loop.index0 - 1].role != "tool") %}
{{- '<|im_start|>user' }}
{%- endif %}
{{- '\n<tool_response>\n' }}
{{- message.content }}
{{- '\n</tool_response>' }}
{%- 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 %}

57
checkpoint-250/config.json Normal file
View File

@@ -0,0 +1,57 @@
{
"architectures": [
"Qwen2ForCausalLM"
],
"attention_dropout": 0.0,
"bos_token_id": null,
"dtype": "bfloat16",
"eos_token_id": 151645,
"hidden_act": "silu",
"hidden_size": 896,
"initializer_range": 0.02,
"intermediate_size": 4864,
"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"
],
"max_position_embeddings": 32768,
"max_window_layers": 21,
"model_type": "qwen2",
"num_attention_heads": 14,
"num_hidden_layers": 24,
"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": "5.0.0",
"use_cache": false,
"use_sliding_window": false,
"vocab_size": 151936
}

13
checkpoint-250/generation_config.json Normal file
View File

@@ -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": "5.0.0"
}

3
checkpoint-250/model.safetensors Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:2f6ecf368a5199e23b82403ab68284fec38fb1f66520d261e9e8cb1f6b4ede31
size 988097824

3
checkpoint-250/optimizer.pt Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:2d8ddb2a1eae3170886a27db3b5ac9bf3aedea6f613bc1f3564d112947682900
size 1976378699

3
checkpoint-250/rng_state.pth Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:979e24915498a0c3479f76d4e80c72ae05419a1971ce8f05dcfdb1d82b4c524f
size 14645

3
checkpoint-250/scheduler.pt Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:5b04611bea80597dee677a377375d057cf3fa3437fb4ed8809d4662285b8811b
size 1465

3
checkpoint-250/tokenizer.json Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:3fd169731d2cbde95e10bf356d66d5997fd885dd8dbb6fb4684da3f23b2585d8
size 11421892

29
checkpoint-250/tokenizer_config.json Normal file
View File

@@ -0,0 +1,29 @@
{
"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,
"model_max_length": 131072,
"pad_token": "<|endoftext|>",
"split_special_tokens": false,
"tokenizer_class": "Qwen2Tokenizer",
"unk_token": null
}

709
checkpoint-250/trainer_state.json Normal file
View File

@@ -0,0 +1,709 @@
{
"best_global_step": null,
"best_metric": null,
"best_model_checkpoint": null,
"epoch": 8.333333333333334,
"eval_steps": 500,
"global_step": 250,
"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": 0.8625,
"completions/max_length": 379.3,
"completions/max_terminated_length": 57.8,
"completions/mean_length": 362.7625,
"completions/mean_terminated_length": 46.4125,
"completions/min_length": 318.1,
"completions/min_terminated_length": 38.1,
"entropy": 0.5937434114515782,
"epoch": 0.3333333333333333,
"frac_reward_zero_std": 0.8,
"grad_norm": 0.0,
"learning_rate": 1.9280000000000002e-05,
"loss": -0.017474760115146638,
"num_tokens": 61847.0,
"reward": -0.0987912505865097,
"reward_std": 0.9303328216075897,
"rewards/reward_fn/mean": -0.0987912505865097,
"rewards/reward_fn/std": 0.9303328156471252,
"step": 10,
"step_time": 31.724731107300066
},
{
"clip_ratio/high_max": 0.0,
"clip_ratio/high_mean": 0.0,
"clip_ratio/low_mean": 0.0,
"clip_ratio/low_min": 0.0,
"clip_ratio/region_mean": 0.0,
"completions/clipped_ratio": 1.0,
"completions/max_length": 400.0,
"completions/max_terminated_length": 0.0,
"completions/mean_length": 400.0,
"completions/mean_terminated_length": 0.0,
"completions/min_length": 400.0,
"completions/min_terminated_length": 0.0,
"entropy": 0.6593769532628357,
"epoch": 0.6666666666666666,
"frac_reward_zero_std": 1.0,
"grad_norm": 0.0,
"learning_rate": 1.8480000000000003e-05,
"loss": 0.0,
"num_tokens": 126843.0,
"reward": 0.048249998688697816,
"reward_std": 0.8191121600568294,
"rewards/reward_fn/mean": 0.048249998688697816,
"rewards/reward_fn/std": 0.8191121838986873,
"step": 20,
"step_time": 33.554111711299946
},
{
"clip_ratio/high_max": 0.0,
"clip_ratio/high_mean": 0.0,
"clip_ratio/low_mean": 0.0,
"clip_ratio/low_min": 0.0,
"clip_ratio/region_mean": 0.0,
"completions/clipped_ratio": 0.975,
"completions/max_length": 400.0,
"completions/max_terminated_length": 49.7,
"completions/mean_length": 396.2125,
"completions/mean_terminated_length": 49.7,
"completions/min_length": 369.7,
"completions/min_terminated_length": 49.7,
"entropy": 0.5578590292367153,
"epoch": 1.0,
"frac_reward_zero_std": 0.9,
"grad_norm": 1.8046875,
"learning_rate": 1.768e-05,
"loss": -0.005067326501011849,
"num_tokens": 191538.0,
"reward": 0.11049998998641967,
"reward_std": 0.711869715154171,
"rewards/reward_fn/mean": 0.11049998998641967,
"rewards/reward_fn/std": 0.7118697345256806,
"step": 30,
"step_time": 33.49379607419994
},
{
"clip_ratio/high_max": 0.0,
"clip_ratio/high_mean": 0.0,
"clip_ratio/low_mean": 0.0,
"clip_ratio/low_min": 0.0,
"clip_ratio/region_mean": 0.0,
"completions/clipped_ratio": 1.0,
"completions/max_length": 400.0,
"completions/max_terminated_length": 0.0,
"completions/mean_length": 400.0,
"completions/mean_terminated_length": 0.0,
"completions/min_length": 400.0,
"completions/min_terminated_length": 0.0,
"entropy": 0.004339599603554234,
"epoch": 1.3333333333333333,
"frac_reward_zero_std": 0.95,
"grad_norm": 0.0,
"learning_rate": 1.688e-05,
"loss": 0.0,
"num_tokens": 256536.0,
"reward": 0.07887499779462814,
"reward_std": 0.8938172444701195,
"rewards/reward_fn/mean": 0.07887499779462814,
"rewards/reward_fn/std": 0.8938172534108162,
"step": 40,
"step_time": 33.4393088197
},
{
"clip_ratio/high_max": 0.0,
"clip_ratio/high_mean": 0.0,
"clip_ratio/low_mean": 0.0,
"clip_ratio/low_min": 0.0,
"clip_ratio/region_mean": 0.0,
"completions/clipped_ratio": 0.9875,
"completions/max_length": 400.0,
"completions/max_terminated_length": 0.9,
"completions/mean_length": 395.1125,
"completions/mean_terminated_length": 0.9,
"completions/min_length": 360.9,
"completions/min_terminated_length": 0.9,
"entropy": 0.004430754791246727,
"epoch": 1.6666666666666665,
"frac_reward_zero_std": 0.975,
"grad_norm": 0.384765625,
"learning_rate": 1.6080000000000002e-05,
"loss": 0.0,
"num_tokens": 321071.0,
"reward": -0.16437500417232515,
"reward_std": 1.063568216562271,
"rewards/reward_fn/mean": -0.16437500417232515,
"rewards/reward_fn/std": 1.0635682106018067,
"step": 50,
"step_time": 33.5646041849001
},
{
"clip_ratio/high_max": 0.0,
"clip_ratio/high_mean": 0.0,
"clip_ratio/low_mean": 0.0,
"clip_ratio/low_min": 0.0,
"clip_ratio/region_mean": 0.0,
"completions/clipped_ratio": 0.9625,
"completions/max_length": 400.0,
"completions/max_terminated_length": 3.7,
"completions/mean_length": 385.4625,
"completions/mean_terminated_length": 3.7,
"completions/min_length": 283.7,
"completions/min_terminated_length": 3.7,
"entropy": 0.09348616125644185,
"epoch": 2.0,
"frac_reward_zero_std": 0.925,
"grad_norm": 0.0,
"learning_rate": 1.5280000000000003e-05,
"loss": 0.009755914658308029,
"num_tokens": 384804.0,
"reward": -0.16562500596046448,
"reward_std": 0.9360396310687065,
"rewards/reward_fn/mean": -0.16562500596046448,
"rewards/reward_fn/std": 0.9360396608710289,
"step": 60,
"step_time": 33.48749011520003
},
{
"clip_ratio/high_max": 0.0,
"clip_ratio/high_mean": 0.0,
"clip_ratio/low_mean": 0.0,
"clip_ratio/low_min": 0.0,
"clip_ratio/region_mean": 0.0,
"completions/clipped_ratio": 0.8625,
"completions/max_length": 400.0,
"completions/max_terminated_length": 44.3,
"completions/mean_length": 351.7625,
"completions/mean_terminated_length": 42.26666679382324,
"completions/min_length": 159.8,
"completions/min_terminated_length": 39.8,
"entropy": 0.6013902719132602,
"epoch": 2.3333333333333335,
"frac_reward_zero_std": 0.925,
"grad_norm": 2.40625,
"learning_rate": 1.448e-05,
"loss": -0.011668374389410019,
"num_tokens": 445825.0,
"reward": 0.08912500143051147,
"reward_std": 0.9901719689369202,
"rewards/reward_fn/mean": 0.08912500143051147,
"rewards/reward_fn/std": 0.9901719927787781,
"step": 70,
"step_time": 33.48351585430014
},
{
"clip_ratio/high_max": 0.0,
"clip_ratio/high_mean": 0.0,
"clip_ratio/low_mean": 0.0,
"clip_ratio/low_min": 0.0,
"clip_ratio/region_mean": 0.0,
"completions/clipped_ratio": 0.9,
"completions/max_length": 400.0,
"completions/max_terminated_length": 36.8,
"completions/mean_length": 365.4875,
"completions/mean_terminated_length": 34.75,
"completions/min_length": 192.7,
"completions/min_terminated_length": 32.7,
"entropy": 0.7647990029305219,
"epoch": 2.6666666666666665,
"frac_reward_zero_std": 0.975,
"grad_norm": 0.0,
"learning_rate": 1.3680000000000003e-05,
"loss": 0.0,
"num_tokens": 507944.0,
"reward": -0.14375001043081284,
"reward_std": 0.6796593680977822,
"rewards/reward_fn/mean": -0.14375001043081284,
"rewards/reward_fn/std": 0.6796593815088272,
"step": 80,
"step_time": 33.443731949500034
},
{
"clip_ratio/high_max": 0.0,
"clip_ratio/high_mean": 0.0,
"clip_ratio/low_mean": 0.0,
"clip_ratio/low_min": 0.0,
"clip_ratio/region_mean": 0.0,
"completions/clipped_ratio": 0.95,
"completions/max_length": 400.0,
"completions/max_terminated_length": 38.7,
"completions/mean_length": 386.45,
"completions/mean_terminated_length": 35.4,
"completions/min_length": 312.1,
"completions/min_terminated_length": 32.1,
"entropy": 0.817728553712368,
"epoch": 3.0,
"frac_reward_zero_std": 1.0,
"grad_norm": 0.0,
"learning_rate": 1.2880000000000002e-05,
"loss": 0.0,
"num_tokens": 571920.0,
"reward": -0.09100000560283661,
"reward_std": 0.7491292104125022,
"rewards/reward_fn/mean": -0.09100000560283661,
"rewards/reward_fn/std": 0.7491292074322701,
"step": 90,
"step_time": 33.4239154713001
},
{
"clip_ratio/high_max": 0.0,
"clip_ratio/high_mean": 0.0,
"clip_ratio/low_mean": 0.0,
"clip_ratio/low_min": 0.0,
"clip_ratio/region_mean": 0.0,
"completions/clipped_ratio": 0.95,
"completions/max_length": 400.0,
"completions/max_terminated_length": 52.1,
"completions/mean_length": 386.5125,
"completions/mean_terminated_length": 52.1,
"completions/min_length": 292.1,
"completions/min_terminated_length": 52.1,
"entropy": 0.7871452756226063,
"epoch": 3.3333333333333335,
"frac_reward_zero_std": 1.0,
"grad_norm": 0.0,
"learning_rate": 1.2080000000000001e-05,
"loss": 0.0,
"num_tokens": 635831.0,
"reward": -0.11849999725818634,
"reward_std": 0.9780218183994294,
"rewards/reward_fn/mean": -0.11849999725818634,
"rewards/reward_fn/std": 0.9780218482017518,
"step": 100,
"step_time": 33.44708331829988
},
{
"clip_ratio/high_max": 0.0,
"clip_ratio/high_mean": 0.0,
"clip_ratio/low_mean": 0.0,
"clip_ratio/low_min": 0.0,
"clip_ratio/region_mean": 0.0,
"completions/clipped_ratio": 0.95,
"completions/max_length": 400.0,
"completions/max_terminated_length": 75.3,
"completions/mean_length": 389.775,
"completions/mean_terminated_length": 65.95,
"completions/min_length": 336.6,
"completions/min_terminated_length": 56.6,
"entropy": 0.8245016686618328,
"epoch": 3.6666666666666665,
"frac_reward_zero_std": 1.0,
"grad_norm": 0.0,
"learning_rate": 1.128e-05,
"loss": 0.0,
"num_tokens": 699917.0,
"reward": 0.06824999898672104,
"reward_std": 0.8597677208483219,
"rewards/reward_fn/mean": 0.06824999898672104,
"rewards/reward_fn/std": 0.8597677327692509,
"step": 110,
"step_time": 33.43369797569976
},
{
"clip_ratio/high_max": 0.0,
"clip_ratio/high_mean": 0.0,
"clip_ratio/low_mean": 0.0,
"clip_ratio/low_min": 0.0,
"clip_ratio/region_mean": 0.0,
"completions/clipped_ratio": 0.9875,
"completions/max_length": 400.0,
"completions/max_terminated_length": 29.0,
"completions/mean_length": 398.625,
"completions/mean_terminated_length": 29.0,
"completions/min_length": 389.0,
"completions/min_terminated_length": 29.0,
"entropy": 0.8096485503017903,
"epoch": 4.0,
"frac_reward_zero_std": 1.0,
"grad_norm": 0.0,
"learning_rate": 1.0480000000000001e-05,
"loss": 0.0,
"num_tokens": 764733.0,
"reward": -0.05875001698732376,
"reward_std": 0.7073742881417274,
"rewards/reward_fn/mean": -0.05875001698732376,
"rewards/reward_fn/std": 0.7073742985725403,
"step": 120,
"step_time": 33.47654410239984
},
{
"clip_ratio/high_max": 0.0,
"clip_ratio/high_mean": 0.0,
"clip_ratio/low_mean": 0.0,
"clip_ratio/low_min": 0.0,
"clip_ratio/region_mean": 0.0,
"completions/clipped_ratio": 0.9625,
"completions/max_length": 400.0,
"completions/max_terminated_length": 51.8,
"completions/mean_length": 391.475,
"completions/mean_terminated_length": 51.8,
"completions/min_length": 331.8,
"completions/min_terminated_length": 51.8,
"entropy": 0.8178432904183864,
"epoch": 4.333333333333333,
"frac_reward_zero_std": 0.975,
"grad_norm": 0.0,
"learning_rate": 9.68e-06,
"loss": 0.0,
"num_tokens": 828855.0,
"reward": -0.017000006139278413,
"reward_std": 0.7560826383531094,
"rewards/reward_fn/mean": -0.017000006139278413,
"rewards/reward_fn/std": 0.7560826435685157,
"step": 130,
"step_time": 33.514995820599964
},
{
"clip_ratio/high_max": 0.0,
"clip_ratio/high_mean": 0.0,
"clip_ratio/low_mean": 0.0,
"clip_ratio/low_min": 0.0,
"clip_ratio/region_mean": 0.0,
"completions/clipped_ratio": 0.9375,
"completions/max_length": 400.0,
"completions/max_terminated_length": 46.2,
"completions/mean_length": 380.775,
"completions/mean_terminated_length": 46.2,
"completions/min_length": 246.2,
"completions/min_terminated_length": 46.2,
"entropy": 0.8854692216962576,
"epoch": 4.666666666666667,
"frac_reward_zero_std": 0.95,
"grad_norm": 0.0,
"learning_rate": 8.880000000000001e-06,
"loss": -0.012354153394699096,
"num_tokens": 892373.0,
"reward": -0.13475000411272048,
"reward_std": 0.9627669990062714,
"rewards/reward_fn/mean": -0.13475000411272048,
"rewards/reward_fn/std": 0.9627670347690582,
"step": 140,
"step_time": 33.50577180429955
},
{
"clip_ratio/high_max": 0.0,
"clip_ratio/high_mean": 0.0,
"clip_ratio/low_mean": 0.0,
"clip_ratio/low_min": 0.0,
"clip_ratio/region_mean": 0.0,
"completions/clipped_ratio": 0.975,
"completions/max_length": 400.0,
"completions/max_terminated_length": 32.3,
"completions/mean_length": 394.275,
"completions/mean_terminated_length": 17.1,
"completions/min_length": 361.9,
"completions/min_terminated_length": 1.9,
"entropy": 0.8237141810357571,
"epoch": 5.0,
"frac_reward_zero_std": 1.0,
"grad_norm": 0.0,
"learning_rate": 8.08e-06,
"loss": 0.0,
"num_tokens": 956875.0,
"reward": -0.030000001192092896,
"reward_std": 0.7297740295529366,
"rewards/reward_fn/mean": -0.030000001192092896,
"rewards/reward_fn/std": 0.7297740370035172,
"step": 150,
"step_time": 33.46883516300004
},
{
"clip_ratio/high_max": 0.0,
"clip_ratio/high_mean": 0.0,
"clip_ratio/low_mean": 0.0,
"clip_ratio/low_min": 0.0,
"clip_ratio/region_mean": 0.0,
"completions/clipped_ratio": 0.975,
"completions/max_length": 400.0,
"completions/max_terminated_length": 38.5,
"completions/mean_length": 394.8125,
"completions/mean_terminated_length": 38.5,
"completions/min_length": 358.5,
"completions/min_terminated_length": 38.5,
"entropy": 0.9815139673650265,
"epoch": 5.333333333333333,
"frac_reward_zero_std": 1.0,
"grad_norm": 0.0,
"learning_rate": 7.280000000000001e-06,
"loss": 0.0,
"num_tokens": 1021402.0,
"reward": -0.06250000596046448,
"reward_std": 0.9604951858520507,
"rewards/reward_fn/mean": -0.06250000596046448,
"rewards/reward_fn/std": 0.9604952037334442,
"step": 160,
"step_time": 33.41488298210034
},
{
"clip_ratio/high_max": 0.0,
"clip_ratio/high_mean": 0.0,
"clip_ratio/low_mean": 0.0,
"clip_ratio/low_min": 0.0,
"clip_ratio/region_mean": 0.0,
"completions/clipped_ratio": 0.975,
"completions/max_length": 400.0,
"completions/max_terminated_length": 21.2,
"completions/mean_length": 392.65,
"completions/mean_terminated_length": 21.2,
"completions/min_length": 341.2,
"completions/min_terminated_length": 21.2,
"entropy": 0.9070465706288815,
"epoch": 5.666666666666667,
"frac_reward_zero_std": 0.975,
"grad_norm": 0.0,
"learning_rate": 6.480000000000001e-06,
"loss": 0.0,
"num_tokens": 1085632.0,
"reward": -0.1513125091791153,
"reward_std": 0.855255764722824,
"rewards/reward_fn/mean": -0.1513125091791153,
"rewards/reward_fn/std": 0.8552557826042175,
"step": 170,
"step_time": 33.4777201513999
},
{
"clip_ratio/high_max": 0.0,
"clip_ratio/high_mean": 0.0,
"clip_ratio/low_mean": 0.0,
"clip_ratio/low_min": 0.0,
"clip_ratio/region_mean": 0.0,
"completions/clipped_ratio": 1.0,
"completions/max_length": 400.0,
"completions/max_terminated_length": 0.0,
"completions/mean_length": 400.0,
"completions/mean_terminated_length": 0.0,
"completions/min_length": 400.0,
"completions/min_terminated_length": 0.0,
"entropy": 0.9226562466472388,
"epoch": 6.0,
"frac_reward_zero_std": 1.0,
"grad_norm": 0.0,
"learning_rate": 5.68e-06,
"loss": 0.0,
"num_tokens": 1150692.0,
"reward": 0.08200000673532486,
"reward_std": 0.9346169531345367,
"rewards/reward_fn/mean": 0.08200000673532486,
"rewards/reward_fn/std": 0.9346169650554657,
"step": 180,
"step_time": 33.47372673979989
},
{
"clip_ratio/high_max": 0.0,
"clip_ratio/high_mean": 0.0,
"clip_ratio/low_mean": 0.0,
"clip_ratio/low_min": 0.0,
"clip_ratio/region_mean": 0.0,
"completions/clipped_ratio": 0.975,
"completions/max_length": 400.0,
"completions/max_terminated_length": 35.5,
"completions/mean_length": 394.4375,
"completions/mean_terminated_length": 35.5,
"completions/min_length": 355.5,
"completions/min_terminated_length": 35.5,
"entropy": 0.9981254413723946,
"epoch": 6.333333333333333,
"frac_reward_zero_std": 0.975,
"grad_norm": 0.0,
"learning_rate": 4.880000000000001e-06,
"loss": 0.0,
"num_tokens": 1215133.0,
"reward": -0.08775000870227814,
"reward_std": 0.8883017361164093,
"rewards/reward_fn/mean": -0.08775000870227814,
"rewards/reward_fn/std": 0.8883017897605896,
"step": 190,
"step_time": 33.41868621859985
},
{
"clip_ratio/high_max": 0.0,
"clip_ratio/high_mean": 0.0,
"clip_ratio/low_mean": 0.0,
"clip_ratio/low_min": 0.0,
"clip_ratio/region_mean": 0.0,
"completions/clipped_ratio": 0.9875,
"completions/max_length": 400.0,
"completions/max_terminated_length": 30.4,
"completions/mean_length": 398.8,
"completions/mean_terminated_length": 30.4,
"completions/min_length": 390.4,
"completions/min_terminated_length": 30.4,
"entropy": 0.9789559677243233,
"epoch": 6.666666666666667,
"frac_reward_zero_std": 1.0,
"grad_norm": 0.0,
"learning_rate": 4.08e-06,
"loss": 0.0,
"num_tokens": 1279879.0,
"reward": -0.03525000065565109,
"reward_std": 0.8935358546674251,
"rewards/reward_fn/mean": -0.03525000065565109,
"rewards/reward_fn/std": 0.8935358554124833,
"step": 200,
"step_time": 33.424166100800086
},
{
"clip_ratio/high_max": 0.0,
"clip_ratio/high_mean": 0.0,
"clip_ratio/low_mean": 0.0,
"clip_ratio/low_min": 0.0,
"clip_ratio/region_mean": 0.0,
"completions/clipped_ratio": 0.9625,
"completions/max_length": 400.0,
"completions/max_terminated_length": 57.2,
"completions/mean_length": 392.15,
"completions/mean_terminated_length": 57.2,
"completions/min_length": 337.2,
"completions/min_terminated_length": 57.2,
"entropy": 0.9129896691069007,
"epoch": 7.0,
"frac_reward_zero_std": 1.0,
"grad_norm": 0.0,
"learning_rate": 3.2800000000000004e-06,
"loss": 0.0,
"num_tokens": 1344343.0,
"reward": -0.011500009894371032,
"reward_std": 0.8285948574543,
"rewards/reward_fn/mean": -0.011500009894371032,
"rewards/reward_fn/std": 0.8285948634147644,
"step": 210,
"step_time": 33.59373969100034
},
{
"clip_ratio/high_max": 0.0,
"clip_ratio/high_mean": 0.0,
"clip_ratio/low_mean": 0.0,
"clip_ratio/low_min": 0.0,
"clip_ratio/region_mean": 0.0,
"completions/clipped_ratio": 1.0,
"completions/max_length": 400.0,
"completions/max_terminated_length": 0.0,
"completions/mean_length": 400.0,
"completions/mean_terminated_length": 0.0,
"completions/min_length": 400.0,
"completions/min_terminated_length": 0.0,
"entropy": 0.960437498241663,
"epoch": 7.333333333333333,
"frac_reward_zero_std": 1.0,
"grad_norm": 0.0,
"learning_rate": 2.4800000000000004e-06,
"loss": 0.0,
"num_tokens": 1409195.0,
"reward": 0.018250004947185518,
"reward_std": 0.8383742928504944,
"rewards/reward_fn/mean": 0.018250004947185518,
"rewards/reward_fn/std": 0.8383742868900299,
"step": 220,
"step_time": 33.32045125410023
},
{
"clip_ratio/high_max": 0.0,
"clip_ratio/high_mean": 0.0,
"clip_ratio/low_mean": 0.0,
"clip_ratio/low_min": 0.0,
"clip_ratio/region_mean": 0.0,
"completions/clipped_ratio": 0.9875,
"completions/max_length": 400.0,
"completions/max_terminated_length": 19.6,
"completions/mean_length": 397.45,
"completions/mean_terminated_length": 19.6,
"completions/min_length": 379.6,
"completions/min_terminated_length": 19.6,
"entropy": 0.9672328025102616,
"epoch": 7.666666666666667,
"frac_reward_zero_std": 1.0,
"grad_norm": 0.0,
"learning_rate": 1.6800000000000002e-06,
"loss": 0.0,
"num_tokens": 1473919.0,
"reward": -0.11225000917911529,
"reward_std": 0.8593939155340194,
"rewards/reward_fn/mean": -0.11225000917911529,
"rewards/reward_fn/std": 0.8593939125537873,
"step": 230,
"step_time": 33.41978766690026
},
{
"clip_ratio/high_max": 0.0,
"clip_ratio/high_mean": 0.0,
"clip_ratio/low_mean": 0.0,
"clip_ratio/low_min": 0.0,
"clip_ratio/region_mean": 0.0,
"completions/clipped_ratio": 0.9625,
"completions/max_length": 400.0,
"completions/max_terminated_length": 36.1,
"completions/mean_length": 389.5125,
"completions/mean_terminated_length": 36.1,
"completions/min_length": 316.1,
"completions/min_terminated_length": 36.1,
"entropy": 1.013204599916935,
"epoch": 8.0,
"frac_reward_zero_std": 1.0,
"grad_norm": 0.0,
"learning_rate": 8.8e-07,
"loss": 0.0,
"num_tokens": 1538120.0,
"reward": -0.015000000596046448,
"reward_std": 0.9198581218719483,
"rewards/reward_fn/mean": -0.015000000596046448,
"rewards/reward_fn/std": 0.9198581159114838,
"step": 240,
"step_time": 33.588851060100204
},
{
"clip_ratio/high_max": 0.0,
"clip_ratio/high_mean": 0.0,
"clip_ratio/low_mean": 0.0,
"clip_ratio/low_min": 0.0,
"clip_ratio/region_mean": 0.0,
"completions/clipped_ratio": 0.9625,
"completions/max_length": 400.0,
"completions/max_terminated_length": 54.6,
"completions/mean_length": 391.825,
"completions/mean_terminated_length": 54.6,
"completions/min_length": 334.6,
"completions/min_terminated_length": 54.6,
"entropy": 0.9223409309983254,
"epoch": 8.333333333333334,
"frac_reward_zero_std": 1.0,
"grad_norm": 0.0,
"learning_rate": 8e-08,
"loss": 0.0,
"num_tokens": 1602540.0,
"reward": 0.044499991834163664,
"reward_std": 0.8610172867774963,
"rewards/reward_fn/mean": 0.044499991834163664,
"rewards/reward_fn/std": 0.8610173106193543,
"step": 250,
"step_time": 33.55163867890042
}
],
"logging_steps": 10,
"max_steps": 250,
"num_input_tokens_seen": 1602540,
"num_train_epochs": 9,
"save_steps": 100,
"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": 2,
"trial_name": null,
"trial_params": null
}

3
checkpoint-250/training_args.bin Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:0f0dd240a760a35edcf19bfedd721f31bea53099058c16d322d918d5f830b64a
size 7057