Hugging Face's logo

Siddeshwar1625
/
osint-checkpoints

Safetensors
Model card Files
xet
 Community
Siddeshwar1625 commited on
Commit
353ce8d
·
verified ·
1 Parent(s): 2c9954a

Intermediate checkpoint upload step=90 (generator_train)

Browse files
Files changed (12) hide show
  1. .gitattributes +1 -0
  2. self_play_hf_l40s_full/round_002/generator_train/checkpoint-90/chat_template.jinja +54 -0
  3. self_play_hf_l40s_full/round_002/generator_train/checkpoint-90/config.json +57 -0
  4. self_play_hf_l40s_full/round_002/generator_train/checkpoint-90/generation_config.json +13 -0
  5. self_play_hf_l40s_full/round_002/generator_train/checkpoint-90/model.safetensors +3 -0
  6. self_play_hf_l40s_full/round_002/generator_train/checkpoint-90/optimizer.pt +3 -0
  7. self_play_hf_l40s_full/round_002/generator_train/checkpoint-90/rng_state.pth +3 -0
  8. self_play_hf_l40s_full/round_002/generator_train/checkpoint-90/scheduler.pt +3 -0
  9. self_play_hf_l40s_full/round_002/generator_train/checkpoint-90/tokenizer.json +3 -0
  10. self_play_hf_l40s_full/round_002/generator_train/checkpoint-90/tokenizer_config.json +32 -0
  11. self_play_hf_l40s_full/round_002/generator_train/checkpoint-90/trainer_state.json +1683 -0
  12. self_play_hf_l40s_full/round_002/generator_train/checkpoint-90/training_args.bin +3 -0
.gitattributes CHANGED
@@ -54,3 +54,4 @@ self_play_hf_l40s_full/round_001/answerer_train/checkpoint-120/tokenizer.json fi
54
  self_play_hf_l40s_full/round_001/answerer_train/final_model/tokenizer.json filter=lfs diff=lfs merge=lfs -text
55
  self_play_hf_l40s_full/round_002/generator_train/checkpoint-30/tokenizer.json filter=lfs diff=lfs merge=lfs -text
56
  self_play_hf_l40s_full/round_002/generator_train/checkpoint-60/tokenizer.json filter=lfs diff=lfs merge=lfs -text
 
 
54
  self_play_hf_l40s_full/round_001/answerer_train/final_model/tokenizer.json filter=lfs diff=lfs merge=lfs -text
55
  self_play_hf_l40s_full/round_002/generator_train/checkpoint-30/tokenizer.json filter=lfs diff=lfs merge=lfs -text
56
  self_play_hf_l40s_full/round_002/generator_train/checkpoint-60/tokenizer.json filter=lfs diff=lfs merge=lfs -text
57
+ self_play_hf_l40s_full/round_002/generator_train/checkpoint-90/tokenizer.json filter=lfs diff=lfs merge=lfs -text
self_play_hf_l40s_full/round_002/generator_train/checkpoint-90/chat_template.jinja ADDED
@@ -0,0 +1,54 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {%- if tools %}
2
+ {{- '<|im_start|>system\n' }}
3
+ {%- if messages[0]['role'] == 'system' %}
4
+ {{- messages[0]['content'] }}
5
+ {%- else %}
6
+ {{- 'You are Qwen, created by Alibaba Cloud. You are a helpful assistant.' }}
7
+ {%- endif %}
8
+ {{- "\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>" }}
9
+ {%- for tool in tools %}
10
+ {{- "\n" }}
11
+ {{- tool | tojson }}
12
+ {%- endfor %}
13
+ {{- "\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" }}
14
+ {%- else %}
15
+ {%- if messages[0]['role'] == 'system' %}
16
+ {{- '<|im_start|>system\n' + messages[0]['content'] + '<|im_end|>\n' }}
17
+ {%- else %}
18
+ {{- '<|im_start|>system\nYou are Qwen, created by Alibaba Cloud. You are a helpful assistant.<|im_end|>\n' }}
19
+ {%- endif %}
20
+ {%- endif %}
21
+ {%- for message in messages %}
22
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) or (message.role == "assistant" and not message.tool_calls) %}
23
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
24
+ {%- elif message.role == "assistant" %}
25
+ {{- '<|im_start|>' + message.role }}
26
+ {%- if message.content %}
27
+ {{- '\n' + message.content }}
28
+ {%- endif %}
29
+ {%- for tool_call in message.tool_calls %}
30
+ {%- if tool_call.function is defined %}
31
+ {%- set tool_call = tool_call.function %}
32
+ {%- endif %}
33
+ {{- '\n<tool_call>\n{"name": "' }}
34
+ {{- tool_call.name }}
35
+ {{- '", "arguments": ' }}
36
+ {{- tool_call.arguments | tojson }}
37
+ {{- '}\n</tool_call>' }}
38
+ {%- endfor %}
39
+ {{- '<|im_end|>\n' }}
40
+ {%- elif message.role == "tool" %}
41
+ {%- if (loop.index0 == 0) or (messages[loop.index0 - 1].role != "tool") %}
42
+ {{- '<|im_start|>user' }}
43
+ {%- endif %}
44
+ {{- '\n<tool_response>\n' }}
45
+ {{- message.content }}
46
+ {{- '\n</tool_response>' }}
47
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
48
+ {{- '<|im_end|>\n' }}
49
+ {%- endif %}
50
+ {%- endif %}
51
+ {%- endfor %}
52
+ {%- if add_generation_prompt %}
53
+ {{- '<|im_start|>assistant\n' }}
54
+ {%- endif %}
self_play_hf_l40s_full/round_002/generator_train/checkpoint-90/config.json ADDED
@@ -0,0 +1,57 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "architectures": [
3
+ "Qwen2ForCausalLM"
4
+ ],
5
+ "attention_dropout": 0.0,
6
+ "bos_token_id": null,
7
+ "dtype": "float32",
8
+ "eos_token_id": 151645,
9
+ "hidden_act": "silu",
10
+ "hidden_size": 896,
11
+ "initializer_range": 0.02,
12
+ "intermediate_size": 4864,
13
+ "layer_types": [
14
+ "full_attention",
15
+ "full_attention",
16
+ "full_attention",
17
+ "full_attention",
18
+ "full_attention",
19
+ "full_attention",
20
+ "full_attention",
21
+ "full_attention",
22
+ "full_attention",
23
+ "full_attention",
24
+ "full_attention",
25
+ "full_attention",
26
+ "full_attention",
27
+ "full_attention",
28
+ "full_attention",
29
+ "full_attention",
30
+ "full_attention",
31
+ "full_attention",
32
+ "full_attention",
33
+ "full_attention",
34
+ "full_attention",
35
+ "full_attention",
36
+ "full_attention",
37
+ "full_attention"
38
+ ],
39
+ "max_position_embeddings": 32768,
40
+ "max_window_layers": 21,
41
+ "model_type": "qwen2",
42
+ "num_attention_heads": 14,
43
+ "num_hidden_layers": 24,
44
+ "num_key_value_heads": 2,
45
+ "pad_token_id": 151643,
46
+ "rms_norm_eps": 1e-06,
47
+ "rope_parameters": {
48
+ "rope_theta": 1000000.0,
49
+ "rope_type": "default"
50
+ },
51
+ "sliding_window": null,
52
+ "tie_word_embeddings": true,
53
+ "transformers_version": "5.6.2",
54
+ "use_cache": false,
55
+ "use_sliding_window": false,
56
+ "vocab_size": 151936
57
+ }
self_play_hf_l40s_full/round_002/generator_train/checkpoint-90/generation_config.json ADDED
@@ -0,0 +1,13 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "do_sample": true,
3
+ "eos_token_id": [
4
+ 151645,
5
+ 151643
6
+ ],
7
+ "pad_token_id": 151643,
8
+ "repetition_penalty": 1.1,
9
+ "temperature": 0.7,
10
+ "top_k": 20,
11
+ "top_p": 0.8,
12
+ "transformers_version": "5.6.2"
13
+ }
self_play_hf_l40s_full/round_002/generator_train/checkpoint-90/model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:795d4902fb523a56e1fdadec29b397f6d85823a6db98628452bb6706cf986a88
3
+ size 1976163472
self_play_hf_l40s_full/round_002/generator_train/checkpoint-90/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:646da4572e89e1f3601f400342c6ebefb01ba9e33a8d08181d56c6956a9d7e93
3
+ size 3952509771
self_play_hf_l40s_full/round_002/generator_train/checkpoint-90/rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b5001a1413e890182ec459c1edbf745017684c89f471fbb0cafcaadbc1070a9f
3
+ size 14645
self_play_hf_l40s_full/round_002/generator_train/checkpoint-90/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9466ecaa07ea9f7aa3b70a452615802059d7a87e6220a892a7f18c7703f928c0
3
+ size 1465
self_play_hf_l40s_full/round_002/generator_train/checkpoint-90/tokenizer.json ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:3fd169731d2cbde95e10bf356d66d5997fd885dd8dbb6fb4684da3f23b2585d8
3
+ size 11421892
self_play_hf_l40s_full/round_002/generator_train/checkpoint-90/tokenizer_config.json ADDED
@@ -0,0 +1,32 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "add_prefix_space": false,
3
+ "backend": "tokenizers",
4
+ "bos_token": null,
5
+ "clean_up_tokenization_spaces": false,
6
+ "eos_token": "<|im_end|>",
7
+ "errors": "replace",
8
+ "extra_special_tokens": [
9
+ "<|im_start|>",
10
+ "<|im_end|>",
11
+ "<|object_ref_start|>",
12
+ "<|object_ref_end|>",
13
+ "<|box_start|>",
14
+ "<|box_end|>",
15
+ "<|quad_start|>",
16
+ "<|quad_end|>",
17
+ "<|vision_start|>",
18
+ "<|vision_end|>",
19
+ "<|vision_pad|>",
20
+ "<|image_pad|>",
21
+ "<|video_pad|>"
22
+ ],
23
+ "is_local": true,
24
+ "local_files_only": false,
25
+ "model_max_length": 131072,
26
+ "pad_token": "<|endoftext|>",
27
+ "padding_side": "left",
28
+ "split_special_tokens": false,
29
+ "tokenizer_class": "Qwen2Tokenizer",
30
+ "truncation_side": "left",
31
+ "unk_token": null
32
+ }
self_play_hf_l40s_full/round_002/generator_train/checkpoint-90/trainer_state.json ADDED
@@ -0,0 +1,1683 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_global_step": null,
3
+ "best_metric": null,
4
+ "best_model_checkpoint": null,
5
+ "epoch": 3.75,
6
+ "eval_steps": 500,
7
+ "global_step": 90,
8
+ "is_hyper_param_search": false,
9
+ "is_local_process_zero": true,
10
+ "is_world_process_zero": true,
11
+ "log_history": [
12
+ {
13
+ "clip_ratio/high_max": 0.0006510416860692203,
14
+ "clip_ratio/high_mean": 0.0006510416860692203,
15
+ "clip_ratio/low_mean": 0.0,
16
+ "clip_ratio/low_min": 0.0,
17
+ "clip_ratio/region_mean": 0.0006510416860692203,
18
+ "completions/clipped_ratio": 1.0,
19
+ "completions/max_length": 384.0,
20
+ "completions/max_terminated_length": 0.0,
21
+ "completions/mean_length": 384.0,
22
+ "completions/mean_terminated_length": 0.0,
23
+ "completions/min_length": 384.0,
24
+ "completions/min_terminated_length": 0.0,
25
+ "entropy": 2.2909066677093506,
26
+ "epoch": 0.041666666666666664,
27
+ "frac_reward_zero_std": 0.0,
28
+ "grad_norm": 3.419240951538086,
29
+ "kl": 0.0005526020540855825,
30
+ "learning_rate": 5e-06,
31
+ "loss": 0.12301453948020935,
32
+ "num_tokens": 25288.0,
33
+ "reward": -0.13312500715255737,
34
+ "reward_std": 0.1580703854560852,
35
+ "rewards/GeneratorRewardFunction/mean": -0.13312500715255737,
36
+ "rewards/GeneratorRewardFunction/std": 0.1580704003572464,
37
+ "step": 1,
38
+ "step_time": 12.38097839000011
39
+ },
40
+ {
41
+ "clip_ratio/high_max": 0.025390625,
42
+ "clip_ratio/high_mean": 0.025390625,
43
+ "clip_ratio/low_mean": 0.0481770820915699,
44
+ "clip_ratio/low_min": 0.0481770820915699,
45
+ "clip_ratio/region_mean": 0.0735677108168602,
46
+ "entropy": 2.2825264930725098,
47
+ "epoch": 0.08333333333333333,
48
+ "grad_norm": 3.043118476867676,
49
+ "kl": 0.017308469861745834,
50
+ "learning_rate": 4.958333333333334e-06,
51
+ "loss": 0.12409868091344833,
52
+ "step": 2,
53
+ "step_time": 0.22687196500010032
54
+ },
55
+ {
56
+ "clip_ratio/high_max": 0.068359375,
57
+ "clip_ratio/high_mean": 0.068359375,
58
+ "clip_ratio/low_mean": 0.0,
59
+ "clip_ratio/low_min": 0.0,
60
+ "clip_ratio/region_mean": 0.068359375,
61
+ "entropy": 1.7901748418807983,
62
+ "epoch": 0.125,
63
+ "grad_norm": 1.792228102684021,
64
+ "kl": 0.042044539004564285,
65
+ "learning_rate": 4.9166666666666665e-06,
66
+ "loss": -0.11261526495218277,
67
+ "step": 3,
68
+ "step_time": 0.22754332899967267
69
+ },
70
+ {
71
+ "clip_ratio/high_max": 0.1139322891831398,
72
+ "clip_ratio/high_mean": 0.1139322891831398,
73
+ "clip_ratio/low_mean": 0.0,
74
+ "clip_ratio/low_min": 0.0,
75
+ "clip_ratio/region_mean": 0.1139322891831398,
76
+ "entropy": 2.698014974594116,
77
+ "epoch": 0.16666666666666666,
78
+ "grad_norm": 2.728670835494995,
79
+ "kl": 0.11385751515626907,
80
+ "learning_rate": 4.875e-06,
81
+ "loss": -0.11856894940137863,
82
+ "step": 4,
83
+ "step_time": 0.22612155700016956
84
+ },
85
+ {
86
+ "clip_ratio/high_max": 0.0006510416860692203,
87
+ "clip_ratio/high_mean": 0.0006510416860692203,
88
+ "clip_ratio/low_mean": 0.0006510416860692203,
89
+ "clip_ratio/low_min": 0.0006510416860692203,
90
+ "clip_ratio/region_mean": 0.0013020833721384406,
91
+ "completions/clipped_ratio": 1.0,
92
+ "completions/max_length": 384.0,
93
+ "completions/max_terminated_length": 0.0,
94
+ "completions/mean_length": 384.0,
95
+ "completions/mean_terminated_length": 0.0,
96
+ "completions/min_length": 384.0,
97
+ "completions/min_terminated_length": 0.0,
98
+ "entropy": 1.2396337985992432,
99
+ "epoch": 0.20833333333333334,
100
+ "frac_reward_zero_std": 0.25,
101
+ "grad_norm": 3.7660365104675293,
102
+ "kl": 0.027865365147590637,
103
+ "learning_rate": 4.833333333333333e-06,
104
+ "loss": 0.07652663439512253,
105
+ "num_tokens": 50484.0,
106
+ "reward": -0.1990624964237213,
107
+ "reward_std": 0.22889748215675354,
108
+ "rewards/GeneratorRewardFunction/mean": -0.1990624964237213,
109
+ "rewards/GeneratorRewardFunction/std": 0.22889748215675354,
110
+ "step": 5,
111
+ "step_time": 12.267690716000288
112
+ },
113
+ {
114
+ "clip_ratio/high_max": 0.0065104165114462376,
115
+ "clip_ratio/high_mean": 0.0065104165114462376,
116
+ "clip_ratio/low_mean": 0.025390625,
117
+ "clip_ratio/low_min": 0.025390625,
118
+ "clip_ratio/region_mean": 0.0319010429084301,
119
+ "entropy": 1.3415206670761108,
120
+ "epoch": 0.25,
121
+ "grad_norm": 1.7361595630645752,
122
+ "kl": 0.03342495858669281,
123
+ "learning_rate": 4.791666666666668e-06,
124
+ "loss": 0.06931179016828537,
125
+ "step": 6,
126
+ "step_time": 0.22098127500021292
127
+ },
128
+ {
129
+ "clip_ratio/high_max": 0.01888020895421505,
130
+ "clip_ratio/high_mean": 0.01888020895421505,
131
+ "clip_ratio/low_mean": 0.0,
132
+ "clip_ratio/low_min": 0.0,
133
+ "clip_ratio/region_mean": 0.01888020895421505,
134
+ "entropy": 1.5484552383422852,
135
+ "epoch": 0.2916666666666667,
136
+ "grad_norm": 0.7583158612251282,
137
+ "kl": 0.0418914258480072,
138
+ "learning_rate": 4.75e-06,
139
+ "loss": -0.051571328192949295,
140
+ "step": 7,
141
+ "step_time": 0.22065312600034304
142
+ },
143
+ {
144
+ "clip_ratio/high_max": 0.1178385391831398,
145
+ "clip_ratio/high_mean": 0.1178385391831398,
146
+ "clip_ratio/low_mean": 0.0,
147
+ "clip_ratio/low_min": 0.0,
148
+ "clip_ratio/region_mean": 0.1178385391831398,
149
+ "entropy": 2.2375190258026123,
150
+ "epoch": 0.3333333333333333,
151
+ "grad_norm": 1.100609302520752,
152
+ "kl": 0.05581093952059746,
153
+ "learning_rate": 4.708333333333334e-06,
154
+ "loss": -0.09243377298116684,
155
+ "step": 8,
156
+ "step_time": 0.2210044349999407
157
+ },
158
+ {
159
+ "clip_ratio/high_max": 0.0,
160
+ "clip_ratio/high_mean": 0.0,
161
+ "clip_ratio/low_mean": 0.0,
162
+ "clip_ratio/low_min": 0.0,
163
+ "clip_ratio/region_mean": 0.0,
164
+ "completions/clipped_ratio": 1.0,
165
+ "completions/max_length": 384.0,
166
+ "completions/max_terminated_length": 0.0,
167
+ "completions/mean_length": 384.0,
168
+ "completions/mean_terminated_length": 0.0,
169
+ "completions/min_length": 384.0,
170
+ "completions/min_terminated_length": 0.0,
171
+ "entropy": 1.8528798818588257,
172
+ "epoch": 0.375,
173
+ "frac_reward_zero_std": 0.0,
174
+ "grad_norm": 2.863710880279541,
175
+ "kl": 0.13026608526706696,
176
+ "learning_rate": 4.666666666666667e-06,
177
+ "loss": -0.01838863641023636,
178
+ "num_tokens": 75740.0,
179
+ "reward": -0.1399218738079071,
180
+ "reward_std": 0.2057529091835022,
181
+ "rewards/GeneratorRewardFunction/mean": -0.1399218738079071,
182
+ "rewards/GeneratorRewardFunction/std": 0.20575293898582458,
183
+ "step": 9,
184
+ "step_time": 12.445934573000159
185
+ },
186
+ {
187
+ "clip_ratio/high_max": 0.008463541977107525,
188
+ "clip_ratio/high_mean": 0.008463541977107525,
189
+ "clip_ratio/low_mean": 0.0013020833721384406,
190
+ "clip_ratio/low_min": 0.0013020833721384406,
191
+ "clip_ratio/region_mean": 0.009765625,
192
+ "entropy": 1.8978748321533203,
193
+ "epoch": 0.4166666666666667,
194
+ "grad_norm": 2.6196653842926025,
195
+ "kl": 0.0548124723136425,
196
+ "learning_rate": 4.625000000000001e-06,
197
+ "loss": -0.02614506147801876,
198
+ "step": 10,
199
+ "step_time": 0.2266816760002257
200
+ },
201
+ {
202
+ "clip_ratio/high_max": 0.013671875,
203
+ "clip_ratio/high_mean": 0.013671875,
204
+ "clip_ratio/low_mean": 0.013020833022892475,
205
+ "clip_ratio/low_min": 0.013020833022892475,
206
+ "clip_ratio/region_mean": 0.02669270895421505,
207
+ "entropy": 1.9386297464370728,
208
+ "epoch": 0.4583333333333333,
209
+ "grad_norm": 3.514759063720703,
210
+ "kl": 0.0901331901550293,
211
+ "learning_rate": 4.583333333333333e-06,
212
+ "loss": 0.12482144683599472,
213
+ "step": 11,
214
+ "step_time": 0.22661321499981568
215
+ },
216
+ {
217
+ "clip_ratio/high_max": 0.04296875,
218
+ "clip_ratio/high_mean": 0.04296875,
219
+ "clip_ratio/low_mean": 0.0,
220
+ "clip_ratio/low_min": 0.0,
221
+ "clip_ratio/region_mean": 0.04296875,
222
+ "entropy": 2.489102602005005,
223
+ "epoch": 0.5,
224
+ "grad_norm": 1.1873637437820435,
225
+ "kl": 0.06463827937841415,
226
+ "learning_rate": 4.541666666666667e-06,
227
+ "loss": -0.07826828956604004,
228
+ "step": 12,
229
+ "step_time": 0.22605949299986605
230
+ },
231
+ {
232
+ "clip_ratio/high_max": 0.0,
233
+ "clip_ratio/high_mean": 0.0,
234
+ "clip_ratio/low_mean": 0.0,
235
+ "clip_ratio/low_min": 0.0,
236
+ "clip_ratio/region_mean": 0.0,
237
+ "completions/clipped_ratio": 1.0,
238
+ "completions/max_length": 384.0,
239
+ "completions/max_terminated_length": 0.0,
240
+ "completions/mean_length": 384.0,
241
+ "completions/mean_terminated_length": 0.0,
242
+ "completions/min_length": 384.0,
243
+ "completions/min_terminated_length": 0.0,
244
+ "entropy": 2.2705869674682617,
245
+ "epoch": 0.5416666666666666,
246
+ "frac_reward_zero_std": 0.0,
247
+ "grad_norm": 1.5841100215911865,
248
+ "kl": 0.07928314059972763,
249
+ "learning_rate": 4.5e-06,
250
+ "loss": -0.05041670799255371,
251
+ "num_tokens": 100992.0,
252
+ "reward": -0.06781250238418579,
253
+ "reward_std": 0.02886570803821087,
254
+ "rewards/GeneratorRewardFunction/mean": -0.06781250238418579,
255
+ "rewards/GeneratorRewardFunction/std": 0.028865709900856018,
256
+ "step": 13,
257
+ "step_time": 12.360951734000082
258
+ },
259
+ {
260
+ "clip_ratio/high_max": 0.015625,
261
+ "clip_ratio/high_mean": 0.015625,
262
+ "clip_ratio/low_mean": 0.0006510416860692203,
263
+ "clip_ratio/low_min": 0.0006510416860692203,
264
+ "clip_ratio/region_mean": 0.01627604104578495,
265
+ "entropy": 1.7322059869766235,
266
+ "epoch": 0.5833333333333334,
267
+ "grad_norm": 2.0741848945617676,
268
+ "kl": 0.060624316334724426,
269
+ "learning_rate": 4.4583333333333336e-06,
270
+ "loss": -0.04468907043337822,
271
+ "step": 14,
272
+ "step_time": 0.21993768200036357
273
+ },
274
+ {
275
+ "clip_ratio/high_max": 0.01627604104578495,
276
+ "clip_ratio/high_mean": 0.01627604104578495,
277
+ "clip_ratio/low_mean": 0.0026041667442768812,
278
+ "clip_ratio/low_min": 0.0026041667442768812,
279
+ "clip_ratio/region_mean": 0.01888020895421505,
280
+ "entropy": 2.18737530708313,
281
+ "epoch": 0.625,
282
+ "grad_norm": 4.066506385803223,
283
+ "kl": 0.07642111927270889,
284
+ "learning_rate": 4.416666666666667e-06,
285
+ "loss": 0.1044829785823822,
286
+ "step": 15,
287
+ "step_time": 0.22013131699986843
288
+ },
289
+ {
290
+ "clip_ratio/high_max": 0.0442708320915699,
291
+ "clip_ratio/high_mean": 0.0442708320915699,
292
+ "clip_ratio/low_mean": 0.009765625,
293
+ "clip_ratio/low_min": 0.009765625,
294
+ "clip_ratio/region_mean": 0.0540364570915699,
295
+ "entropy": 2.622234582901001,
296
+ "epoch": 0.6666666666666666,
297
+ "grad_norm": 3.0252487659454346,
298
+ "kl": 0.10498789697885513,
299
+ "learning_rate": 4.3750000000000005e-06,
300
+ "loss": -0.003130403347313404,
301
+ "step": 16,
302
+ "step_time": 0.21912707499996031
303
+ },
304
+ {
305
+ "clip_ratio/high_max": 0.0,
306
+ "clip_ratio/high_mean": 0.0,
307
+ "clip_ratio/low_mean": 0.0006510416860692203,
308
+ "clip_ratio/low_min": 0.0006510416860692203,
309
+ "clip_ratio/region_mean": 0.0006510416860692203,
310
+ "completions/clipped_ratio": 1.0,
311
+ "completions/max_length": 384.0,
312
+ "completions/max_terminated_length": 0.0,
313
+ "completions/mean_length": 384.0,
314
+ "completions/mean_terminated_length": 0.0,
315
+ "completions/min_length": 384.0,
316
+ "completions/min_terminated_length": 0.0,
317
+ "entropy": 0.9028136134147644,
318
+ "epoch": 0.7083333333333334,
319
+ "frac_reward_zero_std": 0.25,
320
+ "grad_norm": 1.3516546487808228,
321
+ "kl": 0.08489418029785156,
322
+ "learning_rate": 4.333333333333334e-06,
323
+ "loss": 0.03129524365067482,
324
+ "num_tokens": 126380.0,
325
+ "reward": -0.18468749523162842,
326
+ "reward_std": 0.27848079800605774,
327
+ "rewards/GeneratorRewardFunction/mean": -0.18468749523162842,
328
+ "rewards/GeneratorRewardFunction/std": 0.27848079800605774,
329
+ "step": 17,
330
+ "step_time": 12.429075290000128
331
+ },
332
+ {
333
+ "clip_ratio/high_max": 0.001953125,
334
+ "clip_ratio/high_mean": 0.001953125,
335
+ "clip_ratio/low_mean": 0.0,
336
+ "clip_ratio/low_min": 0.0,
337
+ "clip_ratio/region_mean": 0.001953125,
338
+ "entropy": 2.785872220993042,
339
+ "epoch": 0.75,
340
+ "grad_norm": 1.737163782119751,
341
+ "kl": 0.10512515157461166,
342
+ "learning_rate": 4.2916666666666665e-06,
343
+ "loss": -0.0846581980586052,
344
+ "step": 18,
345
+ "step_time": 0.2297946270000466
346
+ },
347
+ {
348
+ "clip_ratio/high_max": 0.01171875,
349
+ "clip_ratio/high_mean": 0.01171875,
350
+ "clip_ratio/low_mean": 0.0006510416860692203,
351
+ "clip_ratio/low_min": 0.0006510416860692203,
352
+ "clip_ratio/region_mean": 0.012369791977107525,
353
+ "entropy": 2.066610097885132,
354
+ "epoch": 0.7916666666666666,
355
+ "grad_norm": 2.454716920852661,
356
+ "kl": 0.10243833065032959,
357
+ "learning_rate": 4.25e-06,
358
+ "loss": 0.031895026564598083,
359
+ "step": 19,
360
+ "step_time": 0.23007028499978333
361
+ },
362
+ {
363
+ "clip_ratio/high_max": 0.013671875,
364
+ "clip_ratio/high_mean": 0.013671875,
365
+ "clip_ratio/low_mean": 0.0032552082557231188,
366
+ "clip_ratio/low_min": 0.0032552082557231188,
367
+ "clip_ratio/region_mean": 0.01692708395421505,
368
+ "entropy": 1.9649995565414429,
369
+ "epoch": 0.8333333333333334,
370
+ "grad_norm": 2.563176155090332,
371
+ "kl": 0.07682739198207855,
372
+ "learning_rate": 4.208333333333333e-06,
373
+ "loss": 0.022962143644690514,
374
+ "step": 20,
375
+ "step_time": 0.23112243600007787
376
+ },
377
+ {
378
+ "clip_ratio/high_max": 0.0,
379
+ "clip_ratio/high_mean": 0.0,
380
+ "clip_ratio/low_mean": 0.0,
381
+ "clip_ratio/low_min": 0.0,
382
+ "clip_ratio/region_mean": 0.0,
383
+ "completions/clipped_ratio": 1.0,
384
+ "completions/max_length": 384.0,
385
+ "completions/max_terminated_length": 0.0,
386
+ "completions/mean_length": 384.0,
387
+ "completions/mean_terminated_length": 0.0,
388
+ "completions/min_length": 384.0,
389
+ "completions/min_terminated_length": 0.0,
390
+ "entropy": 2.3113908767700195,
391
+ "epoch": 0.875,
392
+ "frac_reward_zero_std": 0.0,
393
+ "grad_norm": 2.672222375869751,
394
+ "kl": 0.09574653953313828,
395
+ "learning_rate": 4.166666666666667e-06,
396
+ "loss": -0.04630495235323906,
397
+ "num_tokens": 151916.0,
398
+ "reward": -0.18390625715255737,
399
+ "reward_std": 0.24478043615818024,
400
+ "rewards/GeneratorRewardFunction/mean": -0.18390625715255737,
401
+ "rewards/GeneratorRewardFunction/std": 0.24478045105934143,
402
+ "step": 21,
403
+ "step_time": 12.423363883000093
404
+ },
405
+ {
406
+ "clip_ratio/high_max": 0.0013020833721384406,
407
+ "clip_ratio/high_mean": 0.0013020833721384406,
408
+ "clip_ratio/low_mean": 0.00390625,
409
+ "clip_ratio/low_min": 0.00390625,
410
+ "clip_ratio/region_mean": 0.0052083334885537624,
411
+ "entropy": 1.7172327041625977,
412
+ "epoch": 0.9166666666666666,
413
+ "grad_norm": 2.3833675384521484,
414
+ "kl": 0.10857907682657242,
415
+ "learning_rate": 4.125e-06,
416
+ "loss": 0.08606918901205063,
417
+ "step": 22,
418
+ "step_time": 0.23097956299989164
419
+ },
420
+ {
421
+ "clip_ratio/high_max": 0.0071614584885537624,
422
+ "clip_ratio/high_mean": 0.0071614584885537624,
423
+ "clip_ratio/low_mean": 0.001953125,
424
+ "clip_ratio/low_min": 0.001953125,
425
+ "clip_ratio/region_mean": 0.009114583022892475,
426
+ "entropy": 2.305208444595337,
427
+ "epoch": 0.9583333333333334,
428
+ "grad_norm": 1.6843225955963135,
429
+ "kl": 0.14590007066726685,
430
+ "learning_rate": 4.083333333333334e-06,
431
+ "loss": -0.037479907274246216,
432
+ "step": 23,
433
+ "step_time": 0.2308707210004286
434
+ },
435
+ {
436
+ "clip_ratio/high_max": 0.014973958022892475,
437
+ "clip_ratio/high_mean": 0.014973958022892475,
438
+ "clip_ratio/low_mean": 0.0052083334885537624,
439
+ "clip_ratio/low_min": 0.0052083334885537624,
440
+ "clip_ratio/region_mean": 0.02018229104578495,
441
+ "entropy": 1.4579397439956665,
442
+ "epoch": 1.0,
443
+ "grad_norm": 2.526630163192749,
444
+ "kl": 0.10583696514368057,
445
+ "learning_rate": 4.041666666666667e-06,
446
+ "loss": -0.0003783749125432223,
447
+ "step": 24,
448
+ "step_time": 0.23123665799994342
449
+ },
450
+ {
451
+ "clip_ratio/high_max": 0.0,
452
+ "clip_ratio/high_mean": 0.0,
453
+ "clip_ratio/low_mean": 0.0,
454
+ "clip_ratio/low_min": 0.0,
455
+ "clip_ratio/region_mean": 0.0,
456
+ "completions/clipped_ratio": 1.0,
457
+ "completions/max_length": 384.0,
458
+ "completions/max_terminated_length": 0.0,
459
+ "completions/mean_length": 384.0,
460
+ "completions/mean_terminated_length": 0.0,
461
+ "completions/min_length": 384.0,
462
+ "completions/min_terminated_length": 0.0,
463
+ "entropy": 2.00585675239563,
464
+ "epoch": 1.0416666666666667,
465
+ "frac_reward_zero_std": 0.5,
466
+ "grad_norm": 0.705661952495575,
467
+ "kl": 0.10460541397333145,
468
+ "learning_rate": 4.000000000000001e-06,
469
+ "loss": -0.030740460380911827,
470
+ "num_tokens": 177032.0,
471
+ "reward": -0.054375000298023224,
472
+ "reward_std": 0.012499997392296791,
473
+ "rewards/GeneratorRewardFunction/mean": -0.054375000298023224,
474
+ "rewards/GeneratorRewardFunction/std": 0.012500000186264515,
475
+ "step": 25,
476
+ "step_time": 12.342589669000063
477
+ },
478
+ {
479
+ "clip_ratio/high_max": 0.001953125,
480
+ "clip_ratio/high_mean": 0.001953125,
481
+ "clip_ratio/low_mean": 0.0,
482
+ "clip_ratio/low_min": 0.0,
483
+ "clip_ratio/region_mean": 0.001953125,
484
+ "entropy": 1.3793340921401978,
485
+ "epoch": 1.0833333333333333,
486
+ "grad_norm": 1.3432904481887817,
487
+ "kl": 0.10281199216842651,
488
+ "learning_rate": 3.958333333333333e-06,
489
+ "loss": -0.0930887758731842,
490
+ "step": 26,
491
+ "step_time": 0.22512248300017745
492
+ },
493
+ {
494
+ "clip_ratio/high_max": 0.005859375,
495
+ "clip_ratio/high_mean": 0.005859375,
496
+ "clip_ratio/low_mean": 0.0,
497
+ "clip_ratio/low_min": 0.0,
498
+ "clip_ratio/region_mean": 0.005859375,
499
+ "entropy": 1.6825193166732788,
500
+ "epoch": 1.125,
501
+ "grad_norm": 2.242509603500366,
502
+ "kl": 0.10501086711883545,
503
+ "learning_rate": 3.916666666666667e-06,
504
+ "loss": 0.031402457505464554,
505
+ "step": 27,
506
+ "step_time": 0.22399900899972636
507
+ },
508
+ {
509
+ "clip_ratio/high_max": 0.0,
510
+ "clip_ratio/high_mean": 0.0,
511
+ "clip_ratio/low_mean": 0.0026041667442768812,
512
+ "clip_ratio/low_min": 0.0026041667442768812,
513
+ "clip_ratio/region_mean": 0.0026041667442768812,
514
+ "entropy": 1.8964576721191406,
515
+ "epoch": 1.1666666666666667,
516
+ "grad_norm": 2.502171039581299,
517
+ "kl": 0.12420836836099625,
518
+ "learning_rate": 3.875e-06,
519
+ "loss": 0.09433440119028091,
520
+ "step": 28,
521
+ "step_time": 0.22452458999987357
522
+ },
523
+ {
524
+ "clip_ratio/high_max": 0.0,
525
+ "clip_ratio/high_mean": 0.0,
526
+ "clip_ratio/low_mean": 0.0,
527
+ "clip_ratio/low_min": 0.0,
528
+ "clip_ratio/region_mean": 0.0,
529
+ "completions/clipped_ratio": 1.0,
530
+ "completions/max_length": 384.0,
531
+ "completions/max_terminated_length": 0.0,
532
+ "completions/mean_length": 384.0,
533
+ "completions/mean_terminated_length": 0.0,
534
+ "completions/min_length": 384.0,
535
+ "completions/min_terminated_length": 0.0,
536
+ "entropy": 1.6724086999893188,
537
+ "epoch": 1.2083333333333333,
538
+ "frac_reward_zero_std": 0.5,
539
+ "grad_norm": 1.0264368057250977,
540
+ "kl": 0.15805059671401978,
541
+ "learning_rate": 3.833333333333334e-06,
542
+ "loss": -0.06202007457613945,
543
+ "num_tokens": 202536.0,
544
+ "reward": -0.125,
545
+ "reward_std": 0.20493900775909424,
546
+ "rewards/GeneratorRewardFunction/mean": -0.125,
547
+ "rewards/GeneratorRewardFunction/std": 0.20493900775909424,
548
+ "step": 29,
549
+ "step_time": 12.317649789999905
550
+ },
551
+ {
552
+ "clip_ratio/high_max": 0.0032552082557231188,
553
+ "clip_ratio/high_mean": 0.0032552082557231188,
554
+ "clip_ratio/low_mean": 0.0,
555
+ "clip_ratio/low_min": 0.0,
556
+ "clip_ratio/region_mean": 0.0032552082557231188,
557
+ "entropy": 1.8093080520629883,
558
+ "epoch": 1.25,
559
+ "grad_norm": 1.1415506601333618,
560
+ "kl": 0.12885110080242157,
561
+ "learning_rate": 3.7916666666666666e-06,
562
+ "loss": -0.06245793402194977,
563
+ "step": 30,
564
+ "step_time": 0.22891983299996355
565
+ },
566
+ {
567
+ "clip_ratio/high_max": 0.0065104165114462376,
568
+ "clip_ratio/high_mean": 0.0065104165114462376,
569
+ "clip_ratio/low_mean": 0.0013020833721384406,
570
+ "clip_ratio/low_min": 0.0013020833721384406,
571
+ "clip_ratio/region_mean": 0.0078125,
572
+ "entropy": 2.2301156520843506,
573
+ "epoch": 1.2916666666666667,
574
+ "grad_norm": 2.2105777263641357,
575
+ "kl": 0.19130028784275055,
576
+ "learning_rate": 3.7500000000000005e-06,
577
+ "loss": 0.06395368278026581,
578
+ "step": 31,
579
+ "step_time": 0.22761336599978677
580
+ },
581
+ {
582
+ "clip_ratio/high_max": 0.0071614584885537624,
583
+ "clip_ratio/high_mean": 0.0071614584885537624,
584
+ "clip_ratio/low_mean": 0.005859375,
585
+ "clip_ratio/low_min": 0.005859375,
586
+ "clip_ratio/region_mean": 0.013020833022892475,
587
+ "entropy": 1.473127007484436,
588
+ "epoch": 1.3333333333333333,
589
+ "grad_norm": 2.4992551803588867,
590
+ "kl": 0.17668001353740692,
591
+ "learning_rate": 3.708333333333334e-06,
592
+ "loss": 0.06458758562803268,
593
+ "step": 32,
594
+ "step_time": 0.22745083499967222
595
+ },
596
+ {
597
+ "clip_ratio/high_max": 0.0,
598
+ "clip_ratio/high_mean": 0.0,
599
+ "clip_ratio/low_mean": 0.0,
600
+ "clip_ratio/low_min": 0.0,
601
+ "clip_ratio/region_mean": 0.0,
602
+ "completions/clipped_ratio": 1.0,
603
+ "completions/max_length": 384.0,
604
+ "completions/max_terminated_length": 0.0,
605
+ "completions/mean_length": 384.0,
606
+ "completions/mean_terminated_length": 0.0,
607
+ "completions/min_length": 384.0,
608
+ "completions/min_terminated_length": 0.0,
609
+ "entropy": 2.6607534885406494,
610
+ "epoch": 1.375,
611
+ "frac_reward_zero_std": 0.25,
612
+ "grad_norm": 2.440457582473755,
613
+ "kl": 0.16098029911518097,
614
+ "learning_rate": 3.6666666666666666e-06,
615
+ "loss": 0.08189795166254044,
616
+ "num_tokens": 227804.0,
617
+ "reward": -0.06499999761581421,
618
+ "reward_std": 0.03568379953503609,
619
+ "rewards/GeneratorRewardFunction/mean": -0.06499999761581421,
620
+ "rewards/GeneratorRewardFunction/std": 0.03568379580974579,
621
+ "step": 33,
622
+ "step_time": 12.333147503999953
623
+ },
624
+ {
625
+ "clip_ratio/high_max": 0.0026041667442768812,
626
+ "clip_ratio/high_mean": 0.0026041667442768812,
627
+ "clip_ratio/low_mean": 0.0,
628
+ "clip_ratio/low_min": 0.0,
629
+ "clip_ratio/region_mean": 0.0026041667442768812,
630
+ "entropy": 2.67105770111084,
631
+ "epoch": 1.4166666666666667,
632
+ "grad_norm": 1.7682558298110962,
633
+ "kl": 0.15667910873889923,
634
+ "learning_rate": 3.625e-06,
635
+ "loss": -0.11175209283828735,
636
+ "step": 34,
637
+ "step_time": 0.2226674130001811
638
+ },
639
+ {
640
+ "clip_ratio/high_max": 0.011067708022892475,
641
+ "clip_ratio/high_mean": 0.011067708022892475,
642
+ "clip_ratio/low_mean": 0.0,
643
+ "clip_ratio/low_min": 0.0,
644
+ "clip_ratio/region_mean": 0.011067708022892475,
645
+ "entropy": 2.243622064590454,
646
+ "epoch": 1.4583333333333333,
647
+ "grad_norm": 1.3864362239837646,
648
+ "kl": 0.1985555738210678,
649
+ "learning_rate": 3.5833333333333335e-06,
650
+ "loss": -0.08945892006158829,
651
+ "step": 35,
652
+ "step_time": 0.2224721780003165
653
+ },
654
+ {
655
+ "clip_ratio/high_max": 0.01888020895421505,
656
+ "clip_ratio/high_mean": 0.01888020895421505,
657
+ "clip_ratio/low_mean": 0.00390625,
658
+ "clip_ratio/low_min": 0.00390625,
659
+ "clip_ratio/region_mean": 0.02278645895421505,
660
+ "entropy": 2.7333145141601562,
661
+ "epoch": 1.5,
662
+ "grad_norm": 3.525688886642456,
663
+ "kl": 0.1788162738084793,
664
+ "learning_rate": 3.5416666666666673e-06,
665
+ "loss": 0.12119659781455994,
666
+ "step": 36,
667
+ "step_time": 0.22364275199970507
668
+ },
669
+ {
670
+ "clip_ratio/high_max": 0.0,
671
+ "clip_ratio/high_mean": 0.0,
672
+ "clip_ratio/low_mean": 0.0,
673
+ "clip_ratio/low_min": 0.0,
674
+ "clip_ratio/region_mean": 0.0,
675
+ "completions/clipped_ratio": 1.0,
676
+ "completions/max_length": 384.0,
677
+ "completions/max_terminated_length": 0.0,
678
+ "completions/mean_length": 384.0,
679
+ "completions/mean_terminated_length": 0.0,
680
+ "completions/min_length": 384.0,
681
+ "completions/min_terminated_length": 0.0,
682
+ "entropy": 2.323140859603882,
683
+ "epoch": 1.5416666666666665,
684
+ "frac_reward_zero_std": 0.25,
685
+ "grad_norm": 1.3344190120697021,
686
+ "kl": 0.19980604946613312,
687
+ "learning_rate": 3.5e-06,
688
+ "loss": -0.08966616541147232,
689
+ "num_tokens": 253140.0,
690
+ "reward": -0.05218750238418579,
691
+ "reward_std": 0.006316314451396465,
692
+ "rewards/GeneratorRewardFunction/mean": -0.05218750238418579,
693
+ "rewards/GeneratorRewardFunction/std": 0.00631631538271904,
694
+ "step": 37,
695
+ "step_time": 12.417013897999823
696
+ },
697
+ {
698
+ "clip_ratio/high_max": 0.0013020833721384406,
699
+ "clip_ratio/high_mean": 0.0013020833721384406,
700
+ "clip_ratio/low_mean": 0.0013020833721384406,
701
+ "clip_ratio/low_min": 0.0013020833721384406,
702
+ "clip_ratio/region_mean": 0.0026041667442768812,
703
+ "entropy": 2.036484956741333,
704
+ "epoch": 1.5833333333333335,
705
+ "grad_norm": 2.302572250366211,
706
+ "kl": 0.17958642542362213,
707
+ "learning_rate": 3.4583333333333334e-06,
708
+ "loss": 0.05969787761569023,
709
+ "step": 38,
710
+ "step_time": 0.23123699799998576
711
+ },
712
+ {
713
+ "clip_ratio/high_max": 0.010416666977107525,
714
+ "clip_ratio/high_mean": 0.010416666977107525,
715
+ "clip_ratio/low_mean": 0.0,
716
+ "clip_ratio/low_min": 0.0,
717
+ "clip_ratio/region_mean": 0.010416666977107525,
718
+ "entropy": 2.5753612518310547,
719
+ "epoch": 1.625,
720
+ "grad_norm": 2.368976354598999,
721
+ "kl": 0.1637336164712906,
722
+ "learning_rate": 3.416666666666667e-06,
723
+ "loss": -0.0014251094544306397,
724
+ "step": 39,
725
+ "step_time": 0.23036246999981813
726
+ },
727
+ {
728
+ "clip_ratio/high_max": 0.02213541604578495,
729
+ "clip_ratio/high_mean": 0.02213541604578495,
730
+ "clip_ratio/low_mean": 0.0006510416860692203,
731
+ "clip_ratio/low_min": 0.0006510416860692203,
732
+ "clip_ratio/region_mean": 0.02278645895421505,
733
+ "entropy": 2.702845573425293,
734
+ "epoch": 1.6666666666666665,
735
+ "grad_norm": 2.3713412284851074,
736
+ "kl": 0.1807432919740677,
737
+ "learning_rate": 3.3750000000000003e-06,
738
+ "loss": 0.034005194902420044,
739
+ "step": 40,
740
+ "step_time": 0.23214888699976655
741
+ },
742
+ {
743
+ "clip_ratio/high_max": 0.0,
744
+ "clip_ratio/high_mean": 0.0,
745
+ "clip_ratio/low_mean": 0.0,
746
+ "clip_ratio/low_min": 0.0,
747
+ "clip_ratio/region_mean": 0.0,
748
+ "completions/clipped_ratio": 1.0,
749
+ "completions/max_length": 384.0,
750
+ "completions/max_terminated_length": 0.0,
751
+ "completions/mean_length": 384.0,
752
+ "completions/mean_terminated_length": 0.0,
753
+ "completions/min_length": 384.0,
754
+ "completions/min_terminated_length": 0.0,
755
+ "entropy": 1.8684877157211304,
756
+ "epoch": 1.7083333333333335,
757
+ "frac_reward_zero_std": 0.0,
758
+ "grad_norm": 4.169064521789551,
759
+ "kl": 0.15921029448509216,
760
+ "learning_rate": 3.3333333333333333e-06,
761
+ "loss": 0.24258767068386078,
762
+ "num_tokens": 278460.0,
763
+ "reward": -0.07437500357627869,
764
+ "reward_std": 0.07449552416801453,
765
+ "rewards/GeneratorRewardFunction/mean": -0.07437500357627869,
766
+ "rewards/GeneratorRewardFunction/std": 0.07449552416801453,
767
+ "step": 41,
768
+ "step_time": 12.523336845000358
769
+ },
770
+ {
771
+ "clip_ratio/high_max": 0.001953125,
772
+ "clip_ratio/high_mean": 0.001953125,
773
+ "clip_ratio/low_mean": 0.0,
774
+ "clip_ratio/low_min": 0.0,
775
+ "clip_ratio/region_mean": 0.001953125,
776
+ "entropy": 1.8914880752563477,
777
+ "epoch": 1.75,
778
+ "grad_norm": 2.1851789951324463,
779
+ "kl": 0.16675592958927155,
780
+ "learning_rate": 3.2916666666666668e-06,
781
+ "loss": 0.002008717739954591,
782
+ "step": 42,
783
+ "step_time": 0.22089864800000214
784
+ },
785
+ {
786
+ "clip_ratio/high_max": 0.0052083334885537624,
787
+ "clip_ratio/high_mean": 0.0052083334885537624,
788
+ "clip_ratio/low_mean": 0.0,
789
+ "clip_ratio/low_min": 0.0,
790
+ "clip_ratio/region_mean": 0.0052083334885537624,
791
+ "entropy": 2.0609495639801025,
792
+ "epoch": 1.7916666666666665,
793
+ "grad_norm": 1.6196449995040894,
794
+ "kl": 0.1755555123090744,
795
+ "learning_rate": 3.2500000000000002e-06,
796
+ "loss": -0.11282453685998917,
797
+ "step": 43,
798
+ "step_time": 0.22081435700010843
799
+ },
800
+ {
801
+ "clip_ratio/high_max": 0.021484375,
802
+ "clip_ratio/high_mean": 0.021484375,
803
+ "clip_ratio/low_mean": 0.0,
804
+ "clip_ratio/low_min": 0.0,
805
+ "clip_ratio/region_mean": 0.021484375,
806
+ "entropy": 1.6751976013183594,
807
+ "epoch": 1.8333333333333335,
808
+ "grad_norm": 1.4664008617401123,
809
+ "kl": 0.1903287172317505,
810
+ "learning_rate": 3.2083333333333337e-06,
811
+ "loss": -0.12895803153514862,
812
+ "step": 44,
813
+ "step_time": 0.22197585999992953
814
+ },
815
+ {
816
+ "clip_ratio/high_max": 0.0,
817
+ "clip_ratio/high_mean": 0.0,
818
+ "clip_ratio/low_mean": 0.0,
819
+ "clip_ratio/low_min": 0.0,
820
+ "clip_ratio/region_mean": 0.0,
821
+ "completions/clipped_ratio": 1.0,
822
+ "completions/max_length": 384.0,
823
+ "completions/max_terminated_length": 0.0,
824
+ "completions/mean_length": 384.0,
825
+ "completions/mean_terminated_length": 0.0,
826
+ "completions/min_length": 384.0,
827
+ "completions/min_terminated_length": 0.0,
828
+ "entropy": 1.793825626373291,
829
+ "epoch": 1.875,
830
+ "frac_reward_zero_std": 0.25,
831
+ "grad_norm": 2.486213445663452,
832
+ "kl": 0.1978270262479782,
833
+ "learning_rate": 3.1666666666666667e-06,
834
+ "loss": 0.02904059924185276,
835
+ "num_tokens": 303832.0,
836
+ "reward": -0.06187500059604645,
837
+ "reward_std": 0.018427787348628044,
838
+ "rewards/GeneratorRewardFunction/mean": -0.06187500059604645,
839
+ "rewards/GeneratorRewardFunction/std": 0.018427787348628044,
840
+ "step": 45,
841
+ "step_time": 12.442938505999791
842
+ },
843
+ {
844
+ "clip_ratio/high_max": 0.0032552082557231188,
845
+ "clip_ratio/high_mean": 0.0032552082557231188,
846
+ "clip_ratio/low_mean": 0.0013020833721384406,
847
+ "clip_ratio/low_min": 0.0013020833721384406,
848
+ "clip_ratio/region_mean": 0.0045572915114462376,
849
+ "entropy": 1.8483558893203735,
850
+ "epoch": 1.9166666666666665,
851
+ "grad_norm": 1.7908422946929932,
852
+ "kl": 0.18597811460494995,
853
+ "learning_rate": 3.125e-06,
854
+ "loss": -0.035078153014183044,
855
+ "step": 46,
856
+ "step_time": 0.23107473800018852
857
+ },
858
+ {
859
+ "clip_ratio/high_max": 0.0026041667442768812,
860
+ "clip_ratio/high_mean": 0.0026041667442768812,
861
+ "clip_ratio/low_mean": 0.001953125,
862
+ "clip_ratio/low_min": 0.001953125,
863
+ "clip_ratio/region_mean": 0.0045572915114462376,
864
+ "entropy": 2.0933265686035156,
865
+ "epoch": 1.9583333333333335,
866
+ "grad_norm": 2.3887951374053955,
867
+ "kl": 0.24478764832019806,
868
+ "learning_rate": 3.0833333333333336e-06,
869
+ "loss": 0.03933021053671837,
870
+ "step": 47,
871
+ "step_time": 0.2313891750000039
872
+ },
873
+ {
874
+ "clip_ratio/high_max": 0.010416666977107525,
875
+ "clip_ratio/high_mean": 0.010416666977107525,
876
+ "clip_ratio/low_mean": 0.0006510416860692203,
877
+ "clip_ratio/low_min": 0.0006510416860692203,
878
+ "clip_ratio/region_mean": 0.011067708022892475,
879
+ "entropy": 1.6857815980911255,
880
+ "epoch": 2.0,
881
+ "grad_norm": 1.1888960599899292,
882
+ "kl": 0.2191638946533203,
883
+ "learning_rate": 3.0416666666666666e-06,
884
+ "loss": -0.03249615058302879,
885
+ "step": 48,
886
+ "step_time": 0.23116914000001998
887
+ },
888
+ {
889
+ "clip_ratio/high_max": 0.0,
890
+ "clip_ratio/high_mean": 0.0,
891
+ "clip_ratio/low_mean": 0.0,
892
+ "clip_ratio/low_min": 0.0,
893
+ "clip_ratio/region_mean": 0.0,
894
+ "completions/clipped_ratio": 1.0,
895
+ "completions/max_length": 384.0,
896
+ "completions/max_terminated_length": 0.0,
897
+ "completions/mean_length": 384.0,
898
+ "completions/mean_terminated_length": 0.0,
899
+ "completions/min_length": 384.0,
900
+ "completions/min_terminated_length": 0.0,
901
+ "entropy": 2.1281673908233643,
902
+ "epoch": 2.0416666666666665,
903
+ "frac_reward_zero_std": 0.25,
904
+ "grad_norm": 3.3418045043945312,
905
+ "kl": 0.2323862910270691,
906
+ "learning_rate": 3e-06,
907
+ "loss": 0.12015211582183838,
908
+ "num_tokens": 329212.0,
909
+ "reward": -0.09343749284744263,
910
+ "reward_std": 0.14919470250606537,
911
+ "rewards/GeneratorRewardFunction/mean": -0.09343749284744263,
912
+ "rewards/GeneratorRewardFunction/std": 0.14919471740722656,
913
+ "step": 49,
914
+ "step_time": 12.334705416999896
915
+ },
916
+ {
917
+ "clip_ratio/high_max": 0.0026041667442768812,
918
+ "clip_ratio/high_mean": 0.0026041667442768812,
919
+ "clip_ratio/low_mean": 0.0,
920
+ "clip_ratio/low_min": 0.0,
921
+ "clip_ratio/region_mean": 0.0026041667442768812,
922
+ "entropy": 2.0504140853881836,
923
+ "epoch": 2.0833333333333335,
924
+ "grad_norm": 1.1310595273971558,
925
+ "kl": 0.21666662395000458,
926
+ "learning_rate": 2.9583333333333335e-06,
927
+ "loss": -0.06185801699757576,
928
+ "step": 50,
929
+ "step_time": 0.22278345899985652
930
+ },
931
+ {
932
+ "clip_ratio/high_max": 0.0032552082557231188,
933
+ "clip_ratio/high_mean": 0.0032552082557231188,
934
+ "clip_ratio/low_mean": 0.0,
935
+ "clip_ratio/low_min": 0.0,
936
+ "clip_ratio/region_mean": 0.0032552082557231188,
937
+ "entropy": 2.090566635131836,
938
+ "epoch": 2.125,
939
+ "grad_norm": 0.8605477213859558,
940
+ "kl": 0.19241023063659668,
941
+ "learning_rate": 2.916666666666667e-06,
942
+ "loss": -0.05228700116276741,
943
+ "step": 51,
944
+ "step_time": 0.22478994999983115
945
+ },
946
+ {
947
+ "clip_ratio/high_max": 0.01888020895421505,
948
+ "clip_ratio/high_mean": 0.01888020895421505,
949
+ "clip_ratio/low_mean": 0.0026041667442768812,
950
+ "clip_ratio/low_min": 0.0026041667442768812,
951
+ "clip_ratio/region_mean": 0.021484375,
952
+ "entropy": 1.8061343431472778,
953
+ "epoch": 2.1666666666666665,
954
+ "grad_norm": 2.3467957973480225,
955
+ "kl": 0.20559708774089813,
956
+ "learning_rate": 2.875e-06,
957
+ "loss": -0.004945727530866861,
958
+ "step": 52,
959
+ "step_time": 0.2233186100002058
960
+ },
961
+ {
962
+ "clip_ratio/high_max": 0.0,
963
+ "clip_ratio/high_mean": 0.0,
964
+ "clip_ratio/low_mean": 0.0,
965
+ "clip_ratio/low_min": 0.0,
966
+ "clip_ratio/region_mean": 0.0,
967
+ "completions/clipped_ratio": 1.0,
968
+ "completions/max_length": 384.0,
969
+ "completions/max_terminated_length": 0.0,
970
+ "completions/mean_length": 384.0,
971
+ "completions/mean_terminated_length": 0.0,
972
+ "completions/min_length": 384.0,
973
+ "completions/min_terminated_length": 0.0,
974
+ "entropy": 2.0986216068267822,
975
+ "epoch": 2.2083333333333335,
976
+ "frac_reward_zero_std": 0.75,
977
+ "grad_norm": 0.6935897469520569,
978
+ "kl": 0.2905842065811157,
979
+ "learning_rate": 2.8333333333333335e-06,
980
+ "loss": -0.030335674062371254,
981
+ "num_tokens": 354388.0,
982
+ "reward": -0.05218750238418579,
983
+ "reward_std": 0.008750000968575478,
984
+ "rewards/GeneratorRewardFunction/mean": -0.05218750238418579,
985
+ "rewards/GeneratorRewardFunction/std": 0.008750000037252903,
986
+ "step": 53,
987
+ "step_time": 12.331060373999662
988
+ },
989
+ {
990
+ "clip_ratio/high_max": 0.0,
991
+ "clip_ratio/high_mean": 0.0,
992
+ "clip_ratio/low_mean": 0.0,
993
+ "clip_ratio/low_min": 0.0,
994
+ "clip_ratio/region_mean": 0.0,
995
+ "entropy": 1.8389402627944946,
996
+ "epoch": 2.25,
997
+ "grad_norm": 0.056280333548784256,
998
+ "kl": 0.3166055381298065,
999
+ "learning_rate": 2.791666666666667e-06,
1000
+ "loss": 0.0007915138266980648,
1001
+ "step": 54,
1002
+ "step_time": 0.22823307299995577
1003
+ },
1004
+ {
1005
+ "clip_ratio/high_max": 0.0026041667442768812,
1006
+ "clip_ratio/high_mean": 0.0026041667442768812,
1007
+ "clip_ratio/low_mean": 0.0,
1008
+ "clip_ratio/low_min": 0.0,
1009
+ "clip_ratio/region_mean": 0.0026041667442768812,
1010
+ "entropy": 2.0672857761383057,
1011
+ "epoch": 2.2916666666666665,
1012
+ "grad_norm": 0.6982938051223755,
1013
+ "kl": 0.255536288022995,
1014
+ "learning_rate": 2.7500000000000004e-06,
1015
+ "loss": -0.030553629621863365,
1016
+ "step": 55,
1017
+ "step_time": 0.2274559459997363
1018
+ },
1019
+ {
1020
+ "clip_ratio/high_max": 0.014322916977107525,
1021
+ "clip_ratio/high_mean": 0.014322916977107525,
1022
+ "clip_ratio/low_mean": 0.0006510416860692203,
1023
+ "clip_ratio/low_min": 0.0006510416860692203,
1024
+ "clip_ratio/region_mean": 0.014973958022892475,
1025
+ "entropy": 1.800961971282959,
1026
+ "epoch": 2.3333333333333335,
1027
+ "grad_norm": 1.8801226615905762,
1028
+ "kl": 0.20345354080200195,
1029
+ "learning_rate": 2.7083333333333334e-06,
1030
+ "loss": 0.06233185529708862,
1031
+ "step": 56,
1032
+ "step_time": 0.2291451330002019
1033
+ },
1034
+ {
1035
+ "clip_ratio/high_max": 0.0006510416860692203,
1036
+ "clip_ratio/high_mean": 0.0006510416860692203,
1037
+ "clip_ratio/low_mean": 0.0,
1038
+ "clip_ratio/low_min": 0.0,
1039
+ "clip_ratio/region_mean": 0.0006510416860692203,
1040
+ "completions/clipped_ratio": 1.0,
1041
+ "completions/max_length": 384.0,
1042
+ "completions/max_terminated_length": 0.0,
1043
+ "completions/mean_length": 384.0,
1044
+ "completions/mean_terminated_length": 0.0,
1045
+ "completions/min_length": 384.0,
1046
+ "completions/min_terminated_length": 0.0,
1047
+ "entropy": 2.1910431385040283,
1048
+ "epoch": 2.375,
1049
+ "frac_reward_zero_std": 0.75,
1050
+ "grad_norm": 0.6606095433235168,
1051
+ "kl": 0.2666972875595093,
1052
+ "learning_rate": 2.666666666666667e-06,
1053
+ "loss": -0.029419520869851112,
1054
+ "num_tokens": 379740.0,
1055
+ "reward": -0.050312504172325134,
1056
+ "reward_std": 0.0012500007869675756,
1057
+ "rewards/GeneratorRewardFunction/mean": -0.050312504172325134,
1058
+ "rewards/GeneratorRewardFunction/std": 0.001249999739229679,
1059
+ "step": 57,
1060
+ "step_time": 12.390588525000112
1061
+ },
1062
+ {
1063
+ "clip_ratio/high_max": 0.0,
1064
+ "clip_ratio/high_mean": 0.0,
1065
+ "clip_ratio/low_mean": 0.0,
1066
+ "clip_ratio/low_min": 0.0,
1067
+ "clip_ratio/region_mean": 0.0,
1068
+ "entropy": 2.6509644985198975,
1069
+ "epoch": 2.4166666666666665,
1070
+ "grad_norm": 0.7344714403152466,
1071
+ "kl": 0.27095064520835876,
1072
+ "learning_rate": 2.6250000000000003e-06,
1073
+ "loss": -0.02949327416718006,
1074
+ "step": 58,
1075
+ "step_time": 0.22986735099993894
1076
+ },
1077
+ {
1078
+ "clip_ratio/high_max": 0.0,
1079
+ "clip_ratio/high_mean": 0.0,
1080
+ "clip_ratio/low_mean": 0.0006510416860692203,
1081
+ "clip_ratio/low_min": 0.0006510416860692203,
1082
+ "clip_ratio/region_mean": 0.0006510416860692203,
1083
+ "entropy": 1.9508165121078491,
1084
+ "epoch": 2.4583333333333335,
1085
+ "grad_norm": 2.0245089530944824,
1086
+ "kl": 0.24057161808013916,
1087
+ "learning_rate": 2.5833333333333337e-06,
1088
+ "loss": 0.09111762046813965,
1089
+ "step": 59,
1090
+ "step_time": 0.22998812399964663
1091
+ },
1092
+ {
1093
+ "clip_ratio/high_max": 0.0071614584885537624,
1094
+ "clip_ratio/high_mean": 0.0071614584885537624,
1095
+ "clip_ratio/low_mean": 0.0,
1096
+ "clip_ratio/low_min": 0.0,
1097
+ "clip_ratio/region_mean": 0.0071614584885537624,
1098
+ "entropy": 2.048118829727173,
1099
+ "epoch": 2.5,
1100
+ "grad_norm": 0.641863226890564,
1101
+ "kl": 0.30453336238861084,
1102
+ "learning_rate": 2.5416666666666668e-06,
1103
+ "loss": -0.02960226871073246,
1104
+ "step": 60,
1105
+ "step_time": 0.2306224280000606
1106
+ },
1107
+ {
1108
+ "clip_ratio/high_max": 0.0,
1109
+ "clip_ratio/high_mean": 0.0,
1110
+ "clip_ratio/low_mean": 0.0,
1111
+ "clip_ratio/low_min": 0.0,
1112
+ "clip_ratio/region_mean": 0.0,
1113
+ "completions/clipped_ratio": 1.0,
1114
+ "completions/max_length": 384.0,
1115
+ "completions/max_terminated_length": 0.0,
1116
+ "completions/mean_length": 384.0,
1117
+ "completions/mean_terminated_length": 0.0,
1118
+ "completions/min_length": 384.0,
1119
+ "completions/min_terminated_length": 0.0,
1120
+ "entropy": 2.2969672679901123,
1121
+ "epoch": 2.5416666666666665,
1122
+ "frac_reward_zero_std": 0.5,
1123
+ "grad_norm": 0.7382878661155701,
1124
+ "kl": 0.296942800283432,
1125
+ "learning_rate": 2.5e-06,
1126
+ "loss": -0.030259618535637856,
1127
+ "num_tokens": 405216.0,
1128
+ "reward": -0.05218750238418579,
1129
+ "reward_std": 0.006574888247996569,
1130
+ "rewards/GeneratorRewardFunction/mean": -0.05218750238418579,
1131
+ "rewards/GeneratorRewardFunction/std": 0.006574889644980431,
1132
+ "step": 61,
1133
+ "step_time": 12.311173853000128
1134
+ },
1135
+ {
1136
+ "clip_ratio/high_max": 0.0006510416860692203,
1137
+ "clip_ratio/high_mean": 0.0006510416860692203,
1138
+ "clip_ratio/low_mean": 0.0,
1139
+ "clip_ratio/low_min": 0.0,
1140
+ "clip_ratio/region_mean": 0.0006510416860692203,
1141
+ "entropy": 2.1786930561065674,
1142
+ "epoch": 2.5833333333333335,
1143
+ "grad_norm": 2.3152403831481934,
1144
+ "kl": 0.3079628050327301,
1145
+ "learning_rate": 2.4583333333333332e-06,
1146
+ "loss": 0.06318899989128113,
1147
+ "step": 62,
1148
+ "step_time": 0.22889465099979134
1149
+ },
1150
+ {
1151
+ "clip_ratio/high_max": 0.0071614584885537624,
1152
+ "clip_ratio/high_mean": 0.0071614584885537624,
1153
+ "clip_ratio/low_mean": 0.0,
1154
+ "clip_ratio/low_min": 0.0,
1155
+ "clip_ratio/region_mean": 0.0071614584885537624,
1156
+ "entropy": 2.3085124492645264,
1157
+ "epoch": 2.625,
1158
+ "grad_norm": 1.1132763624191284,
1159
+ "kl": 0.30134913325309753,
1160
+ "learning_rate": 2.4166666666666667e-06,
1161
+ "loss": -0.06140845641493797,
1162
+ "step": 63,
1163
+ "step_time": 0.22866916800012405
1164
+ },
1165
+ {
1166
+ "clip_ratio/high_max": 0.0071614584885537624,
1167
+ "clip_ratio/high_mean": 0.0071614584885537624,
1168
+ "clip_ratio/low_mean": 0.0006510416860692203,
1169
+ "clip_ratio/low_min": 0.0006510416860692203,
1170
+ "clip_ratio/region_mean": 0.0078125,
1171
+ "entropy": 2.3601491451263428,
1172
+ "epoch": 2.6666666666666665,
1173
+ "grad_norm": 2.493839740753174,
1174
+ "kl": 0.3362214267253876,
1175
+ "learning_rate": 2.375e-06,
1176
+ "loss": 0.030866017565131187,
1177
+ "step": 64,
1178
+ "step_time": 0.2289339420003671
1179
+ },
1180
+ {
1181
+ "clip_ratio/high_max": 0.0,
1182
+ "clip_ratio/high_mean": 0.0,
1183
+ "clip_ratio/low_mean": 0.0,
1184
+ "clip_ratio/low_min": 0.0,
1185
+ "clip_ratio/region_mean": 0.0,
1186
+ "completions/clipped_ratio": 1.0,
1187
+ "completions/max_length": 384.0,
1188
+ "completions/max_terminated_length": 0.0,
1189
+ "completions/mean_length": 384.0,
1190
+ "completions/mean_terminated_length": 0.0,
1191
+ "completions/min_length": 384.0,
1192
+ "completions/min_terminated_length": 0.0,
1193
+ "entropy": 2.1568443775177,
1194
+ "epoch": 2.7083333333333335,
1195
+ "frac_reward_zero_std": 0.5,
1196
+ "grad_norm": 1.3307244777679443,
1197
+ "kl": 0.34504207968711853,
1198
+ "learning_rate": 2.3333333333333336e-06,
1199
+ "loss": -0.09026249498128891,
1200
+ "num_tokens": 430464.0,
1201
+ "reward": -0.051875002682209015,
1202
+ "reward_std": 0.006291529163718224,
1203
+ "rewards/GeneratorRewardFunction/mean": -0.051875002682209015,
1204
+ "rewards/GeneratorRewardFunction/std": 0.006291529163718224,
1205
+ "step": 65,
1206
+ "step_time": 12.406595935000041
1207
+ },
1208
+ {
1209
+ "clip_ratio/high_max": 0.0013020833721384406,
1210
+ "clip_ratio/high_mean": 0.0013020833721384406,
1211
+ "clip_ratio/low_mean": 0.0,
1212
+ "clip_ratio/low_min": 0.0,
1213
+ "clip_ratio/region_mean": 0.0013020833721384406,
1214
+ "entropy": 2.1828343868255615,
1215
+ "epoch": 2.75,
1216
+ "grad_norm": 0.7081091403961182,
1217
+ "kl": 0.29117047786712646,
1218
+ "learning_rate": 2.2916666666666666e-06,
1219
+ "loss": -0.029572701081633568,
1220
+ "step": 66,
1221
+ "step_time": 0.21885032299996965
1222
+ },
1223
+ {
1224
+ "clip_ratio/high_max": 0.0032552082557231188,
1225
+ "clip_ratio/high_mean": 0.0032552082557231188,
1226
+ "clip_ratio/low_mean": 0.0,
1227
+ "clip_ratio/low_min": 0.0,
1228
+ "clip_ratio/region_mean": 0.0032552082557231188,
1229
+ "entropy": 2.5054445266723633,
1230
+ "epoch": 2.7916666666666665,
1231
+ "grad_norm": 2.2795095443725586,
1232
+ "kl": 0.2908760607242584,
1233
+ "learning_rate": 2.25e-06,
1234
+ "loss": 0.06366726756095886,
1235
+ "step": 67,
1236
+ "step_time": 0.21955299000001105
1237
+ },
1238
+ {
1239
+ "clip_ratio/high_max": 0.0071614584885537624,
1240
+ "clip_ratio/high_mean": 0.0071614584885537624,
1241
+ "clip_ratio/low_mean": 0.0013020833721384406,
1242
+ "clip_ratio/low_min": 0.0013020833721384406,
1243
+ "clip_ratio/region_mean": 0.008463541977107525,
1244
+ "entropy": 2.5866119861602783,
1245
+ "epoch": 2.8333333333333335,
1246
+ "grad_norm": 2.5023953914642334,
1247
+ "kl": 0.32891178131103516,
1248
+ "learning_rate": 2.2083333333333335e-06,
1249
+ "loss": 0.06120677292346954,
1250
+ "step": 68,
1251
+ "step_time": 0.2204980289998275
1252
+ },
1253
+ {
1254
+ "clip_ratio/high_max": 0.0,
1255
+ "clip_ratio/high_mean": 0.0,
1256
+ "clip_ratio/low_mean": 0.0,
1257
+ "clip_ratio/low_min": 0.0,
1258
+ "clip_ratio/region_mean": 0.0,
1259
+ "completions/clipped_ratio": 1.0,
1260
+ "completions/max_length": 384.0,
1261
+ "completions/max_terminated_length": 0.0,
1262
+ "completions/mean_length": 384.0,
1263
+ "completions/mean_terminated_length": 0.0,
1264
+ "completions/min_length": 384.0,
1265
+ "completions/min_terminated_length": 0.0,
1266
+ "entropy": 2.6892755031585693,
1267
+ "epoch": 2.875,
1268
+ "frac_reward_zero_std": 0.75,
1269
+ "grad_norm": 2.4671924114227295,
1270
+ "kl": 0.4024137556552887,
1271
+ "learning_rate": 2.166666666666667e-06,
1272
+ "loss": 0.0010057635372504592,
1273
+ "num_tokens": 455748.0,
1274
+ "reward": -0.050312504172325134,
1275
+ "reward_std": 0.0012500007869675756,
1276
+ "rewards/GeneratorRewardFunction/mean": -0.050312504172325134,
1277
+ "rewards/GeneratorRewardFunction/std": 0.001249999739229679,
1278
+ "step": 69,
1279
+ "step_time": 12.337975863999873
1280
+ },
1281
+ {
1282
+ "clip_ratio/high_max": 0.0,
1283
+ "clip_ratio/high_mean": 0.0,
1284
+ "clip_ratio/low_mean": 0.0,
1285
+ "clip_ratio/low_min": 0.0,
1286
+ "clip_ratio/region_mean": 0.0,
1287
+ "entropy": 2.2920382022857666,
1288
+ "epoch": 2.9166666666666665,
1289
+ "grad_norm": 0.050838783383369446,
1290
+ "kl": 0.3717615604400635,
1291
+ "learning_rate": 2.125e-06,
1292
+ "loss": 0.0009294038754887879,
1293
+ "step": 70,
1294
+ "step_time": 0.22696916699987923
1295
+ },
1296
+ {
1297
+ "clip_ratio/high_max": 0.0,
1298
+ "clip_ratio/high_mean": 0.0,
1299
+ "clip_ratio/low_mean": 0.0,
1300
+ "clip_ratio/low_min": 0.0,
1301
+ "clip_ratio/region_mean": 0.0,
1302
+ "entropy": 2.742776870727539,
1303
+ "epoch": 2.9583333333333335,
1304
+ "grad_norm": 0.05694516375660896,
1305
+ "kl": 0.32230833172798157,
1306
+ "learning_rate": 2.0833333333333334e-06,
1307
+ "loss": 0.0008057708037085831,
1308
+ "step": 71,
1309
+ "step_time": 0.226955056999941
1310
+ },
1311
+ {
1312
+ "clip_ratio/high_max": 0.0,
1313
+ "clip_ratio/high_mean": 0.0,
1314
+ "clip_ratio/low_mean": 0.0,
1315
+ "clip_ratio/low_min": 0.0,
1316
+ "clip_ratio/region_mean": 0.0,
1317
+ "entropy": 2.6485788822174072,
1318
+ "epoch": 3.0,
1319
+ "grad_norm": 0.04752276465296745,
1320
+ "kl": 0.3216642141342163,
1321
+ "learning_rate": 2.041666666666667e-06,
1322
+ "loss": 0.000804160488769412,
1323
+ "step": 72,
1324
+ "step_time": 0.22688675500012323
1325
+ },
1326
+ {
1327
+ "clip_ratio/high_max": 0.0,
1328
+ "clip_ratio/high_mean": 0.0,
1329
+ "clip_ratio/low_mean": 0.0,
1330
+ "clip_ratio/low_min": 0.0,
1331
+ "clip_ratio/region_mean": 0.0,
1332
+ "completions/clipped_ratio": 1.0,
1333
+ "completions/max_length": 384.0,
1334
+ "completions/max_terminated_length": 0.0,
1335
+ "completions/mean_length": 384.0,
1336
+ "completions/mean_terminated_length": 0.0,
1337
+ "completions/min_length": 384.0,
1338
+ "completions/min_terminated_length": 0.0,
1339
+ "entropy": 3.031399965286255,
1340
+ "epoch": 3.0416666666666665,
1341
+ "frac_reward_zero_std": 1.0,
1342
+ "grad_norm": 0.04752670228481293,
1343
+ "kl": 0.4015932083129883,
1344
+ "learning_rate": 2.0000000000000003e-06,
1345
+ "loss": 0.0010039829649031162,
1346
+ "num_tokens": 481324.0,
1347
+ "reward": -0.05000000074505806,
1348
+ "reward_std": 0.0,
1349
+ "rewards/GeneratorRewardFunction/mean": -0.05000000074505806,
1350
+ "rewards/GeneratorRewardFunction/std": 0.0,
1351
+ "step": 73,
1352
+ "step_time": 12.331203256000208
1353
+ },
1354
+ {
1355
+ "clip_ratio/high_max": 0.0,
1356
+ "clip_ratio/high_mean": 0.0,
1357
+ "clip_ratio/low_mean": 0.0,
1358
+ "clip_ratio/low_min": 0.0,
1359
+ "clip_ratio/region_mean": 0.0,
1360
+ "entropy": 2.7322070598602295,
1361
+ "epoch": 3.0833333333333335,
1362
+ "grad_norm": 0.0566370002925396,
1363
+ "kl": 0.4283183515071869,
1364
+ "learning_rate": 1.9583333333333334e-06,
1365
+ "loss": 0.0010707959299907088,
1366
+ "step": 74,
1367
+ "step_time": 0.23062532400035707
1368
+ },
1369
+ {
1370
+ "clip_ratio/high_max": 0.0,
1371
+ "clip_ratio/high_mean": 0.0,
1372
+ "clip_ratio/low_mean": 0.0,
1373
+ "clip_ratio/low_min": 0.0,
1374
+ "clip_ratio/region_mean": 0.0,
1375
+ "entropy": 2.501298189163208,
1376
+ "epoch": 3.125,
1377
+ "grad_norm": 0.10160665214061737,
1378
+ "kl": 0.4141906201839447,
1379
+ "learning_rate": 1.916666666666667e-06,
1380
+ "loss": 0.0010354764526709914,
1381
+ "step": 75,
1382
+ "step_time": 0.23127166999984183
1383
+ },
1384
+ {
1385
+ "clip_ratio/high_max": 0.0,
1386
+ "clip_ratio/high_mean": 0.0,
1387
+ "clip_ratio/low_mean": 0.0,
1388
+ "clip_ratio/low_min": 0.0,
1389
+ "clip_ratio/region_mean": 0.0,
1390
+ "entropy": 2.740558385848999,
1391
+ "epoch": 3.1666666666666665,
1392
+ "grad_norm": 0.05294595658779144,
1393
+ "kl": 0.3685241937637329,
1394
+ "learning_rate": 1.8750000000000003e-06,
1395
+ "loss": 0.0009213103912770748,
1396
+ "step": 76,
1397
+ "step_time": 0.23059906500020588
1398
+ },
1399
+ {
1400
+ "clip_ratio/high_max": 0.0,
1401
+ "clip_ratio/high_mean": 0.0,
1402
+ "clip_ratio/low_mean": 0.0,
1403
+ "clip_ratio/low_min": 0.0,
1404
+ "clip_ratio/region_mean": 0.0,
1405
+ "completions/clipped_ratio": 1.0,
1406
+ "completions/max_length": 384.0,
1407
+ "completions/max_terminated_length": 0.0,
1408
+ "completions/mean_length": 384.0,
1409
+ "completions/mean_terminated_length": 0.0,
1410
+ "completions/min_length": 384.0,
1411
+ "completions/min_terminated_length": 0.0,
1412
+ "entropy": 2.565075159072876,
1413
+ "epoch": 3.2083333333333335,
1414
+ "frac_reward_zero_std": 0.5,
1415
+ "grad_norm": 2.2318193912506104,
1416
+ "kl": 0.39917126297950745,
1417
+ "learning_rate": 1.8333333333333333e-06,
1418
+ "loss": 0.06347710639238358,
1419
+ "num_tokens": 506544.0,
1420
+ "reward": -0.1251562535762787,
1421
+ "reward_std": 0.20536647737026215,
1422
+ "rewards/GeneratorRewardFunction/mean": -0.1251562535762787,
1423
+ "rewards/GeneratorRewardFunction/std": 0.20536647737026215,
1424
+ "step": 77,
1425
+ "step_time": 12.451436804999958
1426
+ },
1427
+ {
1428
+ "clip_ratio/high_max": 0.0013020833721384406,
1429
+ "clip_ratio/high_mean": 0.0013020833721384406,
1430
+ "clip_ratio/low_mean": 0.0,
1431
+ "clip_ratio/low_min": 0.0,
1432
+ "clip_ratio/region_mean": 0.0013020833721384406,
1433
+ "entropy": 2.5695462226867676,
1434
+ "epoch": 3.25,
1435
+ "grad_norm": 1.0660792589187622,
1436
+ "kl": 0.43529757857322693,
1437
+ "learning_rate": 1.7916666666666667e-06,
1438
+ "loss": -0.06140017509460449,
1439
+ "step": 78,
1440
+ "step_time": 0.223440125000252
1441
+ },
1442
+ {
1443
+ "clip_ratio/high_max": 0.0006510416860692203,
1444
+ "clip_ratio/high_mean": 0.0006510416860692203,
1445
+ "clip_ratio/low_mean": 0.0,
1446
+ "clip_ratio/low_min": 0.0,
1447
+ "clip_ratio/region_mean": 0.0006510416860692203,
1448
+ "entropy": 2.771979570388794,
1449
+ "epoch": 3.2916666666666665,
1450
+ "grad_norm": 2.520599365234375,
1451
+ "kl": 0.3895336091518402,
1452
+ "learning_rate": 1.75e-06,
1453
+ "loss": 0.06315714865922928,
1454
+ "step": 79,
1455
+ "step_time": 0.22384726399968713
1456
+ },
1457
+ {
1458
+ "clip_ratio/high_max": 0.0032552082557231188,
1459
+ "clip_ratio/high_mean": 0.0032552082557231188,
1460
+ "clip_ratio/low_mean": 0.0,
1461
+ "clip_ratio/low_min": 0.0,
1462
+ "clip_ratio/region_mean": 0.0032552082557231188,
1463
+ "entropy": 2.956615447998047,
1464
+ "epoch": 3.3333333333333335,
1465
+ "grad_norm": 1.1839433908462524,
1466
+ "kl": 0.42561575770378113,
1467
+ "learning_rate": 1.7083333333333334e-06,
1468
+ "loss": -0.061606038361787796,
1469
+ "step": 80,
1470
+ "step_time": 0.22456051800008936
1471
+ },
1472
+ {
1473
+ "clip_ratio/high_max": 0.0,
1474
+ "clip_ratio/high_mean": 0.0,
1475
+ "clip_ratio/low_mean": 0.0,
1476
+ "clip_ratio/low_min": 0.0,
1477
+ "clip_ratio/region_mean": 0.0,
1478
+ "completions/clipped_ratio": 1.0,
1479
+ "completions/max_length": 384.0,
1480
+ "completions/max_terminated_length": 0.0,
1481
+ "completions/mean_length": 384.0,
1482
+ "completions/mean_terminated_length": 0.0,
1483
+ "completions/min_length": 384.0,
1484
+ "completions/min_terminated_length": 0.0,
1485
+ "entropy": 2.8945653438568115,
1486
+ "epoch": 3.375,
1487
+ "frac_reward_zero_std": 0.75,
1488
+ "grad_norm": 1.9572227001190186,
1489
+ "kl": 0.3704819977283478,
1490
+ "learning_rate": 1.6666666666666667e-06,
1491
+ "loss": 0.0934426486492157,
1492
+ "num_tokens": 531820.0,
1493
+ "reward": -0.050937503576278687,
1494
+ "reward_std": 0.003749999450519681,
1495
+ "rewards/GeneratorRewardFunction/mean": -0.050937503576278687,
1496
+ "rewards/GeneratorRewardFunction/std": 0.0037499992176890373,
1497
+ "step": 81,
1498
+ "step_time": 12.363581040999634
1499
+ },
1500
+ {
1501
+ "clip_ratio/high_max": 0.0013020833721384406,
1502
+ "clip_ratio/high_mean": 0.0013020833721384406,
1503
+ "clip_ratio/low_mean": 0.0,
1504
+ "clip_ratio/low_min": 0.0,
1505
+ "clip_ratio/region_mean": 0.0013020833721384406,
1506
+ "entropy": 2.7152912616729736,
1507
+ "epoch": 3.4166666666666665,
1508
+ "grad_norm": 0.7033668756484985,
1509
+ "kl": 0.41299042105674744,
1510
+ "learning_rate": 1.6250000000000001e-06,
1511
+ "loss": -0.029737064614892006,
1512
+ "step": 82,
1513
+ "step_time": 0.22236188099986975
1514
+ },
1515
+ {
1516
+ "clip_ratio/high_max": 0.001953125,
1517
+ "clip_ratio/high_mean": 0.001953125,
1518
+ "clip_ratio/low_mean": 0.0,
1519
+ "clip_ratio/low_min": 0.0,
1520
+ "clip_ratio/region_mean": 0.001953125,
1521
+ "entropy": 2.211116313934326,
1522
+ "epoch": 3.4583333333333335,
1523
+ "grad_norm": 0.657794177532196,
1524
+ "kl": 0.3911992013454437,
1525
+ "learning_rate": 1.5833333333333333e-06,
1526
+ "loss": -0.029873592779040337,
1527
+ "step": 83,
1528
+ "step_time": 0.22224327899994023
1529
+ },
1530
+ {
1531
+ "clip_ratio/high_max": 0.009765625,
1532
+ "clip_ratio/high_mean": 0.009765625,
1533
+ "clip_ratio/low_mean": 0.0,
1534
+ "clip_ratio/low_min": 0.0,
1535
+ "clip_ratio/region_mean": 0.009765625,
1536
+ "entropy": 2.4470269680023193,
1537
+ "epoch": 3.5,
1538
+ "grad_norm": 0.6856485605239868,
1539
+ "kl": 0.41081511974334717,
1540
+ "learning_rate": 1.5416666666666668e-06,
1541
+ "loss": -0.029741326346993446,
1542
+ "step": 84,
1543
+ "step_time": 0.2234124139999949
1544
+ },
1545
+ {
1546
+ "clip_ratio/high_max": 0.0,
1547
+ "clip_ratio/high_mean": 0.0,
1548
+ "clip_ratio/low_mean": 0.0,
1549
+ "clip_ratio/low_min": 0.0,
1550
+ "clip_ratio/region_mean": 0.0,
1551
+ "completions/clipped_ratio": 1.0,
1552
+ "completions/max_length": 384.0,
1553
+ "completions/max_terminated_length": 0.0,
1554
+ "completions/mean_length": 384.0,
1555
+ "completions/mean_terminated_length": 0.0,
1556
+ "completions/min_length": 384.0,
1557
+ "completions/min_terminated_length": 0.0,
1558
+ "entropy": 2.9690322875976562,
1559
+ "epoch": 3.5416666666666665,
1560
+ "frac_reward_zero_std": 0.75,
1561
+ "grad_norm": 0.06701389700174332,
1562
+ "kl": 0.4637357294559479,
1563
+ "learning_rate": 1.5e-06,
1564
+ "loss": 0.0011593393282964826,
1565
+ "num_tokens": 556888.0,
1566
+ "reward": -0.08843749761581421,
1567
+ "reward_std": 0.1497967392206192,
1568
+ "rewards/GeneratorRewardFunction/mean": -0.08843749761581421,
1569
+ "rewards/GeneratorRewardFunction/std": 0.1497967392206192,
1570
+ "step": 85,
1571
+ "step_time": 12.362771884000267
1572
+ },
1573
+ {
1574
+ "clip_ratio/high_max": 0.0,
1575
+ "clip_ratio/high_mean": 0.0,
1576
+ "clip_ratio/low_mean": 0.0,
1577
+ "clip_ratio/low_min": 0.0,
1578
+ "clip_ratio/region_mean": 0.0,
1579
+ "entropy": 2.651390552520752,
1580
+ "epoch": 3.5833333333333335,
1581
+ "grad_norm": 2.273343086242676,
1582
+ "kl": 0.30297160148620605,
1583
+ "learning_rate": 1.4583333333333335e-06,
1584
+ "loss": 0.06558748334646225,
1585
+ "step": 86,
1586
+ "step_time": 0.22204756400014958
1587
+ },
1588
+ {
1589
+ "clip_ratio/high_max": 0.0,
1590
+ "clip_ratio/high_mean": 0.0,
1591
+ "clip_ratio/low_mean": 0.0,
1592
+ "clip_ratio/low_min": 0.0,
1593
+ "clip_ratio/region_mean": 0.0,
1594
+ "entropy": 2.563995599746704,
1595
+ "epoch": 3.625,
1596
+ "grad_norm": 0.8811482191085815,
1597
+ "kl": 0.34387049078941345,
1598
+ "learning_rate": 1.4166666666666667e-06,
1599
+ "loss": -0.03156151995062828,
1600
+ "step": 87,
1601
+ "step_time": 0.22124292899979991
1602
+ },
1603
+ {
1604
+ "clip_ratio/high_max": 0.0026041667442768812,
1605
+ "clip_ratio/high_mean": 0.0026041667442768812,
1606
+ "clip_ratio/low_mean": 0.0,
1607
+ "clip_ratio/low_min": 0.0,
1608
+ "clip_ratio/region_mean": 0.0026041667442768812,
1609
+ "entropy": 2.9350931644439697,
1610
+ "epoch": 3.6666666666666665,
1611
+ "grad_norm": 0.7562029957771301,
1612
+ "kl": 0.43041226267814636,
1613
+ "learning_rate": 1.3750000000000002e-06,
1614
+ "loss": -0.03143274411559105,
1615
+ "step": 88,
1616
+ "step_time": 0.22161645599999247
1617
+ },
1618
+ {
1619
+ "clip_ratio/high_max": 0.0,
1620
+ "clip_ratio/high_mean": 0.0,
1621
+ "clip_ratio/low_mean": 0.0,
1622
+ "clip_ratio/low_min": 0.0,
1623
+ "clip_ratio/region_mean": 0.0,
1624
+ "completions/clipped_ratio": 1.0,
1625
+ "completions/max_length": 384.0,
1626
+ "completions/max_terminated_length": 0.0,
1627
+ "completions/mean_length": 384.0,
1628
+ "completions/mean_terminated_length": 0.0,
1629
+ "completions/min_length": 384.0,
1630
+ "completions/min_terminated_length": 0.0,
1631
+ "entropy": 3.0377445220947266,
1632
+ "epoch": 3.7083333333333335,
1633
+ "frac_reward_zero_std": 0.5,
1634
+ "grad_norm": 2.229877233505249,
1635
+ "kl": 0.4358494281768799,
1636
+ "learning_rate": 1.3333333333333334e-06,
1637
+ "loss": 0.06356880068778992,
1638
+ "num_tokens": 582232.0,
1639
+ "reward": -0.09531249850988388,
1640
+ "reward_std": 0.15116733312606812,
1641
+ "rewards/GeneratorRewardFunction/mean": -0.09531249850988388,
1642
+ "rewards/GeneratorRewardFunction/std": 0.15116733312606812,
1643
+ "step": 89,
1644
+ "step_time": 12.385518950000005
1645
+ },
1646
+ {
1647
+ "clip_ratio/high_max": 0.0,
1648
+ "clip_ratio/high_mean": 0.0,
1649
+ "clip_ratio/low_mean": 0.0,
1650
+ "clip_ratio/low_min": 0.0,
1651
+ "clip_ratio/region_mean": 0.0,
1652
+ "entropy": 3.346175193786621,
1653
+ "epoch": 3.75,
1654
+ "grad_norm": 2.875476360321045,
1655
+ "kl": 0.29328086972236633,
1656
+ "learning_rate": 1.2916666666666669e-06,
1657
+ "loss": 0.06316754966974258,
1658
+ "step": 90,
1659
+ "step_time": 0.2198658039997099
1660
+ }
1661
+ ],
1662
+ "logging_steps": 1,
1663
+ "max_steps": 120,
1664
+ "num_input_tokens_seen": 582232,
1665
+ "num_train_epochs": 5,
1666
+ "save_steps": 30,
1667
+ "stateful_callbacks": {
1668
+ "TrainerControl": {
1669
+ "args": {
1670
+ "should_epoch_stop": false,
1671
+ "should_evaluate": false,
1672
+ "should_log": false,
1673
+ "should_save": true,
1674
+ "should_training_stop": false
1675
+ },
1676
+ "attributes": {}
1677
+ }
1678
+ },
1679
+ "total_flos": 0.0,
1680
+ "train_batch_size": 4,
1681
+ "trial_name": null,
1682
+ "trial_params": null
1683
+ }
self_play_hf_l40s_full/round_002/generator_train/checkpoint-90/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c88c3b33b27b27c1cc7746fe61baacbbd09e1573fd273a1e9122cb89b0d1870e
3
+ size 7249