Hugging Face's logo

Siddeshwar1625
/
osint-checkpoints

Safetensors
Model card Files
xet
 Community
Siddeshwar1625 commited on
Commit
edacd4f
·
verified ·
1 Parent(s): 76ec4f2

Intermediate checkpoint upload step=60 (answerer_train)

Browse files
Files changed (12) hide show
  1. .gitattributes +1 -0
  2. self_play_hf_l40s_full/round_006/answerer_train/checkpoint-60/chat_template.jinja +54 -0
  3. self_play_hf_l40s_full/round_006/answerer_train/checkpoint-60/config.json +57 -0
  4. self_play_hf_l40s_full/round_006/answerer_train/checkpoint-60/generation_config.json +13 -0
  5. self_play_hf_l40s_full/round_006/answerer_train/checkpoint-60/model.safetensors +3 -0
  6. self_play_hf_l40s_full/round_006/answerer_train/checkpoint-60/optimizer.pt +3 -0
  7. self_play_hf_l40s_full/round_006/answerer_train/checkpoint-60/rng_state.pth +3 -0
  8. self_play_hf_l40s_full/round_006/answerer_train/checkpoint-60/scheduler.pt +3 -0
  9. self_play_hf_l40s_full/round_006/answerer_train/checkpoint-60/tokenizer.json +3 -0
  10. self_play_hf_l40s_full/round_006/answerer_train/checkpoint-60/tokenizer_config.json +32 -0
  11. self_play_hf_l40s_full/round_006/answerer_train/checkpoint-60/trainer_state.json +1129 -0
  12. self_play_hf_l40s_full/round_006/answerer_train/checkpoint-60/training_args.bin +3 -0
.gitattributes CHANGED
@@ -98,3 +98,4 @@ self_play_hf_l40s_full/round_006/generator_train/checkpoint-90/tokenizer.json fi
98
  self_play_hf_l40s_full/round_006/generator_train/checkpoint-120/tokenizer.json filter=lfs diff=lfs merge=lfs -text
99
  self_play_hf_l40s_full/round_006/generator_train/final_model/tokenizer.json filter=lfs diff=lfs merge=lfs -text
100
  self_play_hf_l40s_full/round_006/answerer_train/checkpoint-30/tokenizer.json filter=lfs diff=lfs merge=lfs -text
 
 
98
  self_play_hf_l40s_full/round_006/generator_train/checkpoint-120/tokenizer.json filter=lfs diff=lfs merge=lfs -text
99
  self_play_hf_l40s_full/round_006/generator_train/final_model/tokenizer.json filter=lfs diff=lfs merge=lfs -text
100
  self_play_hf_l40s_full/round_006/answerer_train/checkpoint-30/tokenizer.json filter=lfs diff=lfs merge=lfs -text
101
+ self_play_hf_l40s_full/round_006/answerer_train/checkpoint-60/tokenizer.json filter=lfs diff=lfs merge=lfs -text
self_play_hf_l40s_full/round_006/answerer_train/checkpoint-60/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_006/answerer_train/checkpoint-60/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_006/answerer_train/checkpoint-60/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_006/answerer_train/checkpoint-60/model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:99ce61f73c62f5ad560a8112bc528e8edf12f8405cfc01c58f4d7ed92e8ca7f1
3
+ size 1976163472
self_play_hf_l40s_full/round_006/answerer_train/checkpoint-60/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4c1e57c8a52b9a9fb1f4396c52e3ca0e8c04077c2df9aab59bc1a388b622c76f
3
+ size 3952509771
self_play_hf_l40s_full/round_006/answerer_train/checkpoint-60/rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:2258c8f7e6687f5fa2dcc6cd5ebbafa0a88234fdbe6ce45ab0d52ed8947a2e83
3
+ size 14645
self_play_hf_l40s_full/round_006/answerer_train/checkpoint-60/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:aab883bcab393f0c916048be23a6c0284269524eb2950c4249e03acc7cac4f75
3
+ size 1465
self_play_hf_l40s_full/round_006/answerer_train/checkpoint-60/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_006/answerer_train/checkpoint-60/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_006/answerer_train/checkpoint-60/trainer_state.json ADDED
@@ -0,0 +1,1129 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_global_step": null,
3
+ "best_metric": null,
4
+ "best_model_checkpoint": null,
5
+ "epoch": 1.5,
6
+ "eval_steps": 500,
7
+ "global_step": 60,
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": 0.0,
14
+ "clip_ratio/high_mean": 0.0,
15
+ "clip_ratio/low_mean": 0.0,
16
+ "clip_ratio/low_min": 0.0,
17
+ "clip_ratio/region_mean": 0.0,
18
+ "completions/clipped_ratio": 1.0,
19
+ "completions/max_length": 256.0,
20
+ "completions/max_terminated_length": 0.0,
21
+ "completions/mean_length": 256.0,
22
+ "completions/mean_terminated_length": 0.0,
23
+ "completions/min_length": 256.0,
24
+ "completions/min_terminated_length": 0.0,
25
+ "entropy": 2.1889684200286865,
26
+ "epoch": 0.025,
27
+ "frac_reward_zero_std": 0.25,
28
+ "grad_norm": 1.8947820663452148,
29
+ "kl": 0.0,
30
+ "learning_rate": 3e-06,
31
+ "loss": -0.12490925192832947,
32
+ "num_tokens": 9876.0,
33
+ "reward": 0.6893415451049805,
34
+ "reward_std": 0.16361689567565918,
35
+ "rewards/AnswererRewardFunction/mean": 0.6893415451049805,
36
+ "rewards/AnswererRewardFunction/std": 0.16361691057682037,
37
+ "step": 1,
38
+ "step_time": 7.946243045999836
39
+ },
40
+ {
41
+ "clip_ratio/high_max": 0.021484375,
42
+ "clip_ratio/high_mean": 0.021484375,
43
+ "clip_ratio/low_mean": 0.0009765625,
44
+ "clip_ratio/low_min": 0.0009765625,
45
+ "clip_ratio/region_mean": 0.0224609375,
46
+ "entropy": 1.8370153903961182,
47
+ "epoch": 0.05,
48
+ "grad_norm": 4.110031604766846,
49
+ "kl": 0.6599398851394653,
50
+ "learning_rate": 2.9750000000000003e-06,
51
+ "loss": 0.0643470287322998,
52
+ "step": 2,
53
+ "step_time": 0.11823628799993458
54
+ },
55
+ {
56
+ "clip_ratio/high_max": 0.0595703125,
57
+ "clip_ratio/high_mean": 0.0595703125,
58
+ "clip_ratio/low_mean": 0.0263671875,
59
+ "clip_ratio/low_min": 0.0263671875,
60
+ "clip_ratio/region_mean": 0.0859375,
61
+ "entropy": 2.297727584838867,
62
+ "epoch": 0.075,
63
+ "grad_norm": 296.2247314453125,
64
+ "kl": 79.00199890136719,
65
+ "learning_rate": 2.9499999999999997e-06,
66
+ "loss": 0.23638710379600525,
67
+ "step": 3,
68
+ "step_time": 0.11235964000024978
69
+ },
70
+ {
71
+ "clip_ratio/high_max": 0.05859375,
72
+ "clip_ratio/high_mean": 0.05859375,
73
+ "clip_ratio/low_mean": 0.087890625,
74
+ "clip_ratio/low_min": 0.087890625,
75
+ "clip_ratio/region_mean": 0.146484375,
76
+ "entropy": 1.4839229583740234,
77
+ "epoch": 0.1,
78
+ "grad_norm": 10.855915069580078,
79
+ "kl": 3.744659900665283,
80
+ "learning_rate": 2.925e-06,
81
+ "loss": 0.0730188637971878,
82
+ "step": 4,
83
+ "step_time": 0.11265637799988326
84
+ },
85
+ {
86
+ "clip_ratio/high_max": 0.0,
87
+ "clip_ratio/high_mean": 0.0,
88
+ "clip_ratio/low_mean": 0.0,
89
+ "clip_ratio/low_min": 0.0,
90
+ "clip_ratio/region_mean": 0.0,
91
+ "completions/clipped_ratio": 1.0,
92
+ "completions/max_length": 256.0,
93
+ "completions/max_terminated_length": 0.0,
94
+ "completions/mean_length": 256.0,
95
+ "completions/mean_terminated_length": 0.0,
96
+ "completions/min_length": 256.0,
97
+ "completions/min_terminated_length": 0.0,
98
+ "entropy": 0.6120401620864868,
99
+ "epoch": 0.125,
100
+ "frac_reward_zero_std": 0.75,
101
+ "grad_norm": 0.009019882418215275,
102
+ "kl": 0.01311470940709114,
103
+ "learning_rate": 2.9e-06,
104
+ "loss": 3.278677468188107e-05,
105
+ "num_tokens": 19608.0,
106
+ "reward": 0.457000195980072,
107
+ "reward_std": 0.07295458763837814,
108
+ "rewards/AnswererRewardFunction/mean": 0.457000195980072,
109
+ "rewards/AnswererRewardFunction/std": 0.07295458763837814,
110
+ "step": 5,
111
+ "step_time": 7.926103121999404
112
+ },
113
+ {
114
+ "clip_ratio/high_max": 0.0,
115
+ "clip_ratio/high_mean": 0.0,
116
+ "clip_ratio/low_mean": 0.0,
117
+ "clip_ratio/low_min": 0.0,
118
+ "clip_ratio/region_mean": 0.0,
119
+ "entropy": 1.1489129066467285,
120
+ "epoch": 0.15,
121
+ "grad_norm": 0.015938131138682365,
122
+ "kl": 0.030105752870440483,
123
+ "learning_rate": 2.875e-06,
124
+ "loss": 7.526438275817782e-05,
125
+ "step": 6,
126
+ "step_time": 0.11689895999916189
127
+ },
128
+ {
129
+ "clip_ratio/high_max": 0.005859375,
130
+ "clip_ratio/high_mean": 0.005859375,
131
+ "clip_ratio/low_mean": 0.0048828125,
132
+ "clip_ratio/low_min": 0.0048828125,
133
+ "clip_ratio/region_mean": 0.0107421875,
134
+ "entropy": 0.6303037405014038,
135
+ "epoch": 0.175,
136
+ "grad_norm": 3.0208628177642822,
137
+ "kl": 0.02866116538643837,
138
+ "learning_rate": 2.85e-06,
139
+ "loss": -0.028981417417526245,
140
+ "step": 7,
141
+ "step_time": 0.10871063499962474
142
+ },
143
+ {
144
+ "clip_ratio/high_max": 0.0,
145
+ "clip_ratio/high_mean": 0.0,
146
+ "clip_ratio/low_mean": 0.005859375,
147
+ "clip_ratio/low_min": 0.005859375,
148
+ "clip_ratio/region_mean": 0.005859375,
149
+ "entropy": 2.0413095951080322,
150
+ "epoch": 0.2,
151
+ "grad_norm": 1.0012094974517822,
152
+ "kl": 0.05237983167171478,
153
+ "learning_rate": 2.825e-06,
154
+ "loss": 0.03189055621623993,
155
+ "step": 8,
156
+ "step_time": 0.10796216900052968
157
+ },
158
+ {
159
+ "clip_ratio/high_max": 0.0009765625,
160
+ "clip_ratio/high_mean": 0.0009765625,
161
+ "clip_ratio/low_mean": 0.0,
162
+ "clip_ratio/low_min": 0.0,
163
+ "clip_ratio/region_mean": 0.0009765625,
164
+ "completions/clipped_ratio": 1.0,
165
+ "completions/max_length": 256.0,
166
+ "completions/max_terminated_length": 0.0,
167
+ "completions/mean_length": 256.0,
168
+ "completions/mean_terminated_length": 0.0,
169
+ "completions/min_length": 256.0,
170
+ "completions/min_terminated_length": 0.0,
171
+ "entropy": 2.2237331867218018,
172
+ "epoch": 0.225,
173
+ "frac_reward_zero_std": 0.5,
174
+ "grad_norm": 1.620787501335144,
175
+ "kl": 0.0605711005628109,
176
+ "learning_rate": 2.8000000000000003e-06,
177
+ "loss": 0.02867729589343071,
178
+ "num_tokens": 29436.0,
179
+ "reward": 0.7435240745544434,
180
+ "reward_std": 0.23860794305801392,
181
+ "rewards/AnswererRewardFunction/mean": 0.7435240745544434,
182
+ "rewards/AnswererRewardFunction/std": 0.2386079579591751,
183
+ "step": 9,
184
+ "step_time": 7.88594858200031
185
+ },
186
+ {
187
+ "clip_ratio/high_max": 0.001953125,
188
+ "clip_ratio/high_mean": 0.001953125,
189
+ "clip_ratio/low_mean": 0.0009765625,
190
+ "clip_ratio/low_min": 0.0009765625,
191
+ "clip_ratio/region_mean": 0.0029296875,
192
+ "entropy": 2.133094310760498,
193
+ "epoch": 0.25,
194
+ "grad_norm": 3.653198003768921,
195
+ "kl": 0.07492633908987045,
196
+ "learning_rate": 2.775e-06,
197
+ "loss": 0.06285852193832397,
198
+ "step": 10,
199
+ "step_time": 0.11135919299977104
200
+ },
201
+ {
202
+ "clip_ratio/high_max": 0.0078125,
203
+ "clip_ratio/high_mean": 0.0078125,
204
+ "clip_ratio/low_mean": 0.0087890625,
205
+ "clip_ratio/low_min": 0.0087890625,
206
+ "clip_ratio/region_mean": 0.0166015625,
207
+ "entropy": 1.841313362121582,
208
+ "epoch": 0.275,
209
+ "grad_norm": 2.8710782527923584,
210
+ "kl": 0.05763581395149231,
211
+ "learning_rate": 2.75e-06,
212
+ "loss": -0.06052658334374428,
213
+ "step": 11,
214
+ "step_time": 0.11127804100033245
215
+ },
216
+ {
217
+ "clip_ratio/high_max": 0.0146484375,
218
+ "clip_ratio/high_mean": 0.0146484375,
219
+ "clip_ratio/low_mean": 0.0,
220
+ "clip_ratio/low_min": 0.0,
221
+ "clip_ratio/region_mean": 0.0146484375,
222
+ "entropy": 1.7089946269989014,
223
+ "epoch": 0.3,
224
+ "grad_norm": 0.9450397491455078,
225
+ "kl": 0.25390055775642395,
226
+ "learning_rate": 2.725e-06,
227
+ "loss": -0.02993064746260643,
228
+ "step": 12,
229
+ "step_time": 0.11213281900018046
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": 256.0,
239
+ "completions/max_terminated_length": 0.0,
240
+ "completions/mean_length": 256.0,
241
+ "completions/mean_terminated_length": 0.0,
242
+ "completions/min_length": 256.0,
243
+ "completions/min_terminated_length": 0.0,
244
+ "entropy": 2.0214595794677734,
245
+ "epoch": 0.325,
246
+ "frac_reward_zero_std": 0.5,
247
+ "grad_norm": 0.8579655289649963,
248
+ "kl": 0.08672940731048584,
249
+ "learning_rate": 2.7e-06,
250
+ "loss": -0.031010489910840988,
251
+ "num_tokens": 39540.0,
252
+ "reward": 0.8283045291900635,
253
+ "reward_std": 0.18679988384246826,
254
+ "rewards/AnswererRewardFunction/mean": 0.8283045291900635,
255
+ "rewards/AnswererRewardFunction/std": 0.18679988384246826,
256
+ "step": 13,
257
+ "step_time": 7.924993149999864
258
+ },
259
+ {
260
+ "clip_ratio/high_max": 0.0009765625,
261
+ "clip_ratio/high_mean": 0.0009765625,
262
+ "clip_ratio/low_mean": 0.0,
263
+ "clip_ratio/low_min": 0.0,
264
+ "clip_ratio/region_mean": 0.0009765625,
265
+ "entropy": 2.297172784805298,
266
+ "epoch": 0.35,
267
+ "grad_norm": 1.3188775777816772,
268
+ "kl": 0.13232484459877014,
269
+ "learning_rate": 2.6750000000000002e-06,
270
+ "loss": -0.06186162680387497,
271
+ "step": 14,
272
+ "step_time": 0.11182027799986827
273
+ },
274
+ {
275
+ "clip_ratio/high_max": 0.0068359375,
276
+ "clip_ratio/high_mean": 0.0068359375,
277
+ "clip_ratio/low_mean": 0.0341796875,
278
+ "clip_ratio/low_min": 0.0341796875,
279
+ "clip_ratio/region_mean": 0.041015625,
280
+ "entropy": 1.809241771697998,
281
+ "epoch": 0.375,
282
+ "grad_norm": 2.606775999069214,
283
+ "kl": 0.1521003544330597,
284
+ "learning_rate": 2.65e-06,
285
+ "loss": 0.06266739964485168,
286
+ "step": 15,
287
+ "step_time": 0.11220332800075994
288
+ },
289
+ {
290
+ "clip_ratio/high_max": 0.021484375,
291
+ "clip_ratio/high_mean": 0.021484375,
292
+ "clip_ratio/low_mean": 0.0,
293
+ "clip_ratio/low_min": 0.0,
294
+ "clip_ratio/region_mean": 0.021484375,
295
+ "entropy": 1.8247143030166626,
296
+ "epoch": 0.4,
297
+ "grad_norm": 3.210848808288574,
298
+ "kl": 0.1755138635635376,
299
+ "learning_rate": 2.6250000000000003e-06,
300
+ "loss": 0.03220568597316742,
301
+ "step": 16,
302
+ "step_time": 0.11238660199978767
303
+ },
304
+ {
305
+ "clip_ratio/high_max": 0.0009765625,
306
+ "clip_ratio/high_mean": 0.0009765625,
307
+ "clip_ratio/low_mean": 0.0,
308
+ "clip_ratio/low_min": 0.0,
309
+ "clip_ratio/region_mean": 0.0009765625,
310
+ "completions/clipped_ratio": 1.0,
311
+ "completions/max_length": 256.0,
312
+ "completions/max_terminated_length": 0.0,
313
+ "completions/mean_length": 256.0,
314
+ "completions/mean_terminated_length": 0.0,
315
+ "completions/min_length": 256.0,
316
+ "completions/min_terminated_length": 0.0,
317
+ "entropy": 2.2773776054382324,
318
+ "epoch": 0.425,
319
+ "frac_reward_zero_std": 0.5,
320
+ "grad_norm": 0.971729040145874,
321
+ "kl": 0.12777887284755707,
322
+ "learning_rate": 2.6e-06,
323
+ "loss": -0.030927438288927078,
324
+ "num_tokens": 49856.0,
325
+ "reward": 0.8496553301811218,
326
+ "reward_std": 0.16016095876693726,
327
+ "rewards/AnswererRewardFunction/mean": 0.8496553301811218,
328
+ "rewards/AnswererRewardFunction/std": 0.16016097366809845,
329
+ "step": 17,
330
+ "step_time": 7.993629739000426
331
+ },
332
+ {
333
+ "clip_ratio/high_max": 0.0,
334
+ "clip_ratio/high_mean": 0.0,
335
+ "clip_ratio/low_mean": 0.001953125,
336
+ "clip_ratio/low_min": 0.001953125,
337
+ "clip_ratio/region_mean": 0.001953125,
338
+ "entropy": 2.0068607330322266,
339
+ "epoch": 0.45,
340
+ "grad_norm": 2.8351244926452637,
341
+ "kl": 0.15364348888397217,
342
+ "learning_rate": 2.575e-06,
343
+ "loss": 0.15774129331111908,
344
+ "step": 18,
345
+ "step_time": 0.11176471900034812
346
+ },
347
+ {
348
+ "clip_ratio/high_max": 0.0029296875,
349
+ "clip_ratio/high_mean": 0.0029296875,
350
+ "clip_ratio/low_mean": 0.0,
351
+ "clip_ratio/low_min": 0.0,
352
+ "clip_ratio/region_mean": 0.0029296875,
353
+ "entropy": 2.258239269256592,
354
+ "epoch": 0.475,
355
+ "grad_norm": 1.3264741897583008,
356
+ "kl": 0.15728481113910675,
357
+ "learning_rate": 2.55e-06,
358
+ "loss": -0.061835240572690964,
359
+ "step": 19,
360
+ "step_time": 0.11123546799990436
361
+ },
362
+ {
363
+ "clip_ratio/high_max": 0.0029296875,
364
+ "clip_ratio/high_mean": 0.0029296875,
365
+ "clip_ratio/low_mean": 0.0068359375,
366
+ "clip_ratio/low_min": 0.0068359375,
367
+ "clip_ratio/region_mean": 0.009765625,
368
+ "entropy": 2.259558916091919,
369
+ "epoch": 0.5,
370
+ "grad_norm": 2.8228580951690674,
371
+ "kl": 0.17330552637577057,
372
+ "learning_rate": 2.525e-06,
373
+ "loss": -0.06122318655252457,
374
+ "step": 20,
375
+ "step_time": 0.1108333300007871
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": 256.0,
385
+ "completions/max_terminated_length": 0.0,
386
+ "completions/mean_length": 256.0,
387
+ "completions/mean_terminated_length": 0.0,
388
+ "completions/min_length": 256.0,
389
+ "completions/min_terminated_length": 0.0,
390
+ "entropy": 2.179208755493164,
391
+ "epoch": 0.525,
392
+ "frac_reward_zero_std": 1.0,
393
+ "grad_norm": 0.08803759515285492,
394
+ "kl": 0.12282326817512512,
395
+ "learning_rate": 2.5e-06,
396
+ "loss": 0.0003070581587962806,
397
+ "num_tokens": 59948.0,
398
+ "reward": 0.8848026990890503,
399
+ "reward_std": 0.08498957008123398,
400
+ "rewards/AnswererRewardFunction/mean": 0.8848026990890503,
401
+ "rewards/AnswererRewardFunction/std": 0.08498956263065338,
402
+ "step": 21,
403
+ "step_time": 7.9203589380003905
404
+ },
405
+ {
406
+ "clip_ratio/high_max": 0.0,
407
+ "clip_ratio/high_mean": 0.0,
408
+ "clip_ratio/low_mean": 0.0,
409
+ "clip_ratio/low_min": 0.0,
410
+ "clip_ratio/region_mean": 0.0,
411
+ "entropy": 2.306586980819702,
412
+ "epoch": 0.55,
413
+ "grad_norm": 0.052290912717580795,
414
+ "kl": 0.16986945271492004,
415
+ "learning_rate": 2.475e-06,
416
+ "loss": 0.0004246736061759293,
417
+ "step": 22,
418
+ "step_time": 0.11205570599940984
419
+ },
420
+ {
421
+ "clip_ratio/high_max": 0.0,
422
+ "clip_ratio/high_mean": 0.0,
423
+ "clip_ratio/low_mean": 0.0,
424
+ "clip_ratio/low_min": 0.0,
425
+ "clip_ratio/region_mean": 0.0,
426
+ "entropy": 2.19669508934021,
427
+ "epoch": 0.575,
428
+ "grad_norm": 0.05250624194741249,
429
+ "kl": 0.17336246371269226,
430
+ "learning_rate": 2.45e-06,
431
+ "loss": 0.00043340615229681134,
432
+ "step": 23,
433
+ "step_time": 0.11391655599982187
434
+ },
435
+ {
436
+ "clip_ratio/high_max": 0.0,
437
+ "clip_ratio/high_mean": 0.0,
438
+ "clip_ratio/low_mean": 0.0,
439
+ "clip_ratio/low_min": 0.0,
440
+ "clip_ratio/region_mean": 0.0,
441
+ "entropy": 2.0577125549316406,
442
+ "epoch": 0.6,
443
+ "grad_norm": 0.04398365691304207,
444
+ "kl": 0.0968913584947586,
445
+ "learning_rate": 2.425e-06,
446
+ "loss": 0.00024222838692367077,
447
+ "step": 24,
448
+ "step_time": 0.11333441400074662
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": 256.0,
458
+ "completions/max_terminated_length": 0.0,
459
+ "completions/mean_length": 256.0,
460
+ "completions/mean_terminated_length": 0.0,
461
+ "completions/min_length": 256.0,
462
+ "completions/min_terminated_length": 0.0,
463
+ "entropy": 2.0512969493865967,
464
+ "epoch": 0.625,
465
+ "frac_reward_zero_std": 1.0,
466
+ "grad_norm": 0.051768120378255844,
467
+ "kl": 0.16503679752349854,
468
+ "learning_rate": 2.4000000000000003e-06,
469
+ "loss": 0.00041259199497289956,
470
+ "num_tokens": 69888.0,
471
+ "reward": 0.8755199909210205,
472
+ "reward_std": 0.10159505903720856,
473
+ "rewards/AnswererRewardFunction/mean": 0.8755199909210205,
474
+ "rewards/AnswererRewardFunction/std": 0.10159505903720856,
475
+ "step": 25,
476
+ "step_time": 7.957063270000617
477
+ },
478
+ {
479
+ "clip_ratio/high_max": 0.0,
480
+ "clip_ratio/high_mean": 0.0,
481
+ "clip_ratio/low_mean": 0.0,
482
+ "clip_ratio/low_min": 0.0,
483
+ "clip_ratio/region_mean": 0.0,
484
+ "entropy": 2.2467222213745117,
485
+ "epoch": 0.65,
486
+ "grad_norm": 0.04632311314344406,
487
+ "kl": 0.13820040225982666,
488
+ "learning_rate": 2.375e-06,
489
+ "loss": 0.0003455010009929538,
490
+ "step": 26,
491
+ "step_time": 0.11183958200035704
492
+ },
493
+ {
494
+ "clip_ratio/high_max": 0.0,
495
+ "clip_ratio/high_mean": 0.0,
496
+ "clip_ratio/low_mean": 0.0,
497
+ "clip_ratio/low_min": 0.0,
498
+ "clip_ratio/region_mean": 0.0,
499
+ "entropy": 2.199760913848877,
500
+ "epoch": 0.675,
501
+ "grad_norm": 0.07714560627937317,
502
+ "kl": 0.16415069997310638,
503
+ "learning_rate": 2.35e-06,
504
+ "loss": 0.00041037675691768527,
505
+ "step": 27,
506
+ "step_time": 0.11124441800075147
507
+ },
508
+ {
509
+ "clip_ratio/high_max": 0.0,
510
+ "clip_ratio/high_mean": 0.0,
511
+ "clip_ratio/low_mean": 0.0,
512
+ "clip_ratio/low_min": 0.0,
513
+ "clip_ratio/region_mean": 0.0,
514
+ "entropy": 2.108006238937378,
515
+ "epoch": 0.7,
516
+ "grad_norm": 0.049742791801691055,
517
+ "kl": 0.16107158362865448,
518
+ "learning_rate": 2.325e-06,
519
+ "loss": 0.0004026789392810315,
520
+ "step": 28,
521
+ "step_time": 0.11063443700004427
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": 256.0,
531
+ "completions/max_terminated_length": 0.0,
532
+ "completions/mean_length": 256.0,
533
+ "completions/mean_terminated_length": 0.0,
534
+ "completions/min_length": 256.0,
535
+ "completions/min_terminated_length": 0.0,
536
+ "entropy": 2.3680968284606934,
537
+ "epoch": 0.725,
538
+ "frac_reward_zero_std": 1.0,
539
+ "grad_norm": 0.03891065716743469,
540
+ "kl": 0.1308700293302536,
541
+ "learning_rate": 2.3000000000000004e-06,
542
+ "loss": 0.00032717504655011,
543
+ "num_tokens": 80032.0,
544
+ "reward": 0.9358674883842468,
545
+ "reward_std": 0.0063578663393855095,
546
+ "rewards/AnswererRewardFunction/mean": 0.9358674883842468,
547
+ "rewards/AnswererRewardFunction/std": 0.006357857491821051,
548
+ "step": 29,
549
+ "step_time": 7.913411731999986
550
+ },
551
+ {
552
+ "clip_ratio/high_max": 0.0,
553
+ "clip_ratio/high_mean": 0.0,
554
+ "clip_ratio/low_mean": 0.0,
555
+ "clip_ratio/low_min": 0.0,
556
+ "clip_ratio/region_mean": 0.0,
557
+ "entropy": 2.180490016937256,
558
+ "epoch": 0.75,
559
+ "grad_norm": 0.039776675403118134,
560
+ "kl": 0.14568868279457092,
561
+ "learning_rate": 2.275e-06,
562
+ "loss": 0.00036422169068828225,
563
+ "step": 30,
564
+ "step_time": 0.11067452799943567
565
+ },
566
+ {
567
+ "clip_ratio/high_max": 0.0,
568
+ "clip_ratio/high_mean": 0.0,
569
+ "clip_ratio/low_mean": 0.0,
570
+ "clip_ratio/low_min": 0.0,
571
+ "clip_ratio/region_mean": 0.0,
572
+ "entropy": 2.4302010536193848,
573
+ "epoch": 0.775,
574
+ "grad_norm": 0.051406338810920715,
575
+ "kl": 0.16150793433189392,
576
+ "learning_rate": 2.25e-06,
577
+ "loss": 0.0004037698090542108,
578
+ "step": 31,
579
+ "step_time": 0.11291773700031627
580
+ },
581
+ {
582
+ "clip_ratio/high_max": 0.0,
583
+ "clip_ratio/high_mean": 0.0,
584
+ "clip_ratio/low_mean": 0.0,
585
+ "clip_ratio/low_min": 0.0,
586
+ "clip_ratio/region_mean": 0.0,
587
+ "entropy": 2.173776626586914,
588
+ "epoch": 0.8,
589
+ "grad_norm": 0.03287765383720398,
590
+ "kl": 0.131052628159523,
591
+ "learning_rate": 2.2250000000000003e-06,
592
+ "loss": 0.0003276315692346543,
593
+ "step": 32,
594
+ "step_time": 0.11095032400044147
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": 256.0,
604
+ "completions/max_terminated_length": 0.0,
605
+ "completions/mean_length": 256.0,
606
+ "completions/mean_terminated_length": 0.0,
607
+ "completions/min_length": 256.0,
608
+ "completions/min_terminated_length": 0.0,
609
+ "entropy": 2.1301045417785645,
610
+ "epoch": 0.825,
611
+ "frac_reward_zero_std": 1.0,
612
+ "grad_norm": 0.07125719636678696,
613
+ "kl": 0.14635741710662842,
614
+ "learning_rate": 2.1999999999999997e-06,
615
+ "loss": 0.00036589353112503886,
616
+ "num_tokens": 90360.0,
617
+ "reward": 0.9323133230209351,
618
+ "reward_std": 0.0,
619
+ "rewards/AnswererRewardFunction/mean": 0.9323133230209351,
620
+ "rewards/AnswererRewardFunction/std": 0.0,
621
+ "step": 33,
622
+ "step_time": 7.960296397000093
623
+ },
624
+ {
625
+ "clip_ratio/high_max": 0.0,
626
+ "clip_ratio/high_mean": 0.0,
627
+ "clip_ratio/low_mean": 0.0,
628
+ "clip_ratio/low_min": 0.0,
629
+ "clip_ratio/region_mean": 0.0,
630
+ "entropy": 2.0239813327789307,
631
+ "epoch": 0.85,
632
+ "grad_norm": 0.0397072434425354,
633
+ "kl": 0.13195659220218658,
634
+ "learning_rate": 2.175e-06,
635
+ "loss": 0.0003298914816696197,
636
+ "step": 34,
637
+ "step_time": 0.1131003039999996
638
+ },
639
+ {
640
+ "clip_ratio/high_max": 0.0,
641
+ "clip_ratio/high_mean": 0.0,
642
+ "clip_ratio/low_mean": 0.0,
643
+ "clip_ratio/low_min": 0.0,
644
+ "clip_ratio/region_mean": 0.0,
645
+ "entropy": 2.308377504348755,
646
+ "epoch": 0.875,
647
+ "grad_norm": 0.040582410991191864,
648
+ "kl": 0.13440832495689392,
649
+ "learning_rate": 2.15e-06,
650
+ "loss": 0.00033602083567529917,
651
+ "step": 35,
652
+ "step_time": 0.15085302500028774
653
+ },
654
+ {
655
+ "clip_ratio/high_max": 0.0,
656
+ "clip_ratio/high_mean": 0.0,
657
+ "clip_ratio/low_mean": 0.0,
658
+ "clip_ratio/low_min": 0.0,
659
+ "clip_ratio/region_mean": 0.0,
660
+ "entropy": 2.1328980922698975,
661
+ "epoch": 0.9,
662
+ "grad_norm": 0.034442681819200516,
663
+ "kl": 0.1075764000415802,
664
+ "learning_rate": 2.125e-06,
665
+ "loss": 0.0002689409884624183,
666
+ "step": 36,
667
+ "step_time": 0.1115491030004705
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": 256.0,
677
+ "completions/max_terminated_length": 0.0,
678
+ "completions/mean_length": 256.0,
679
+ "completions/mean_terminated_length": 0.0,
680
+ "completions/min_length": 256.0,
681
+ "completions/min_terminated_length": 0.0,
682
+ "entropy": 2.149538516998291,
683
+ "epoch": 0.925,
684
+ "frac_reward_zero_std": 1.0,
685
+ "grad_norm": 0.03783974051475525,
686
+ "kl": 0.1285218596458435,
687
+ "learning_rate": 2.1e-06,
688
+ "loss": 0.00032130462932400405,
689
+ "num_tokens": 100360.0,
690
+ "reward": 0.8372920751571655,
691
+ "reward_std": 0.09813749045133591,
692
+ "rewards/AnswererRewardFunction/mean": 0.8372920751571655,
693
+ "rewards/AnswererRewardFunction/std": 0.0981374979019165,
694
+ "step": 37,
695
+ "step_time": 7.983764245999737
696
+ },
697
+ {
698
+ "clip_ratio/high_max": 0.0,
699
+ "clip_ratio/high_mean": 0.0,
700
+ "clip_ratio/low_mean": 0.0,
701
+ "clip_ratio/low_min": 0.0,
702
+ "clip_ratio/region_mean": 0.0,
703
+ "entropy": 2.095836639404297,
704
+ "epoch": 0.95,
705
+ "grad_norm": 0.03456847742199898,
706
+ "kl": 0.11794041097164154,
707
+ "learning_rate": 2.075e-06,
708
+ "loss": 0.000294851022772491,
709
+ "step": 38,
710
+ "step_time": 0.11436837399924116
711
+ },
712
+ {
713
+ "clip_ratio/high_max": 0.0,
714
+ "clip_ratio/high_mean": 0.0,
715
+ "clip_ratio/low_mean": 0.0,
716
+ "clip_ratio/low_min": 0.0,
717
+ "clip_ratio/region_mean": 0.0,
718
+ "entropy": 2.0270957946777344,
719
+ "epoch": 0.975,
720
+ "grad_norm": 0.44049227237701416,
721
+ "kl": 0.21028028428554535,
722
+ "learning_rate": 2.0500000000000003e-06,
723
+ "loss": 0.0005257007433101535,
724
+ "step": 39,
725
+ "step_time": 0.11228540899992367
726
+ },
727
+ {
728
+ "clip_ratio/high_max": 0.0,
729
+ "clip_ratio/high_mean": 0.0,
730
+ "clip_ratio/low_mean": 0.0,
731
+ "clip_ratio/low_min": 0.0,
732
+ "clip_ratio/region_mean": 0.0,
733
+ "entropy": 2.257455587387085,
734
+ "epoch": 1.0,
735
+ "grad_norm": 0.0388871431350708,
736
+ "kl": 0.13575704395771027,
737
+ "learning_rate": 2.025e-06,
738
+ "loss": 0.00033939260174520314,
739
+ "step": 40,
740
+ "step_time": 0.11100837200046954
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": 256.0,
750
+ "completions/max_terminated_length": 0.0,
751
+ "completions/mean_length": 256.0,
752
+ "completions/mean_terminated_length": 0.0,
753
+ "completions/min_length": 256.0,
754
+ "completions/min_terminated_length": 0.0,
755
+ "entropy": 2.2085485458374023,
756
+ "epoch": 1.025,
757
+ "frac_reward_zero_std": 1.0,
758
+ "grad_norm": 0.03739944472908974,
759
+ "kl": 0.1360509842634201,
760
+ "learning_rate": 2e-06,
761
+ "loss": 0.0003401274443604052,
762
+ "num_tokens": 110016.0,
763
+ "reward": 0.8263232707977295,
764
+ "reward_std": 0.11883138865232468,
765
+ "rewards/AnswererRewardFunction/mean": 0.8263232707977295,
766
+ "rewards/AnswererRewardFunction/std": 0.11883135885000229,
767
+ "step": 41,
768
+ "step_time": 7.922437872999581
769
+ },
770
+ {
771
+ "clip_ratio/high_max": 0.0,
772
+ "clip_ratio/high_mean": 0.0,
773
+ "clip_ratio/low_mean": 0.0,
774
+ "clip_ratio/low_min": 0.0,
775
+ "clip_ratio/region_mean": 0.0,
776
+ "entropy": 2.4597482681274414,
777
+ "epoch": 1.05,
778
+ "grad_norm": 0.038012683391571045,
779
+ "kl": 0.13071982562541962,
780
+ "learning_rate": 1.975e-06,
781
+ "loss": 0.00032679951982572675,
782
+ "step": 42,
783
+ "step_time": 0.10853852999935043
784
+ },
785
+ {
786
+ "clip_ratio/high_max": 0.0,
787
+ "clip_ratio/high_mean": 0.0,
788
+ "clip_ratio/low_mean": 0.0,
789
+ "clip_ratio/low_min": 0.0,
790
+ "clip_ratio/region_mean": 0.0,
791
+ "entropy": 2.238304376602173,
792
+ "epoch": 1.075,
793
+ "grad_norm": 0.03296751156449318,
794
+ "kl": 0.11736348271369934,
795
+ "learning_rate": 1.95e-06,
796
+ "loss": 0.00029340869514271617,
797
+ "step": 43,
798
+ "step_time": 0.10875790699992649
799
+ },
800
+ {
801
+ "clip_ratio/high_max": 0.0,
802
+ "clip_ratio/high_mean": 0.0,
803
+ "clip_ratio/low_mean": 0.0,
804
+ "clip_ratio/low_min": 0.0,
805
+ "clip_ratio/region_mean": 0.0,
806
+ "entropy": 2.2012887001037598,
807
+ "epoch": 1.1,
808
+ "grad_norm": 0.03213638812303543,
809
+ "kl": 0.11174218356609344,
810
+ "learning_rate": 1.925e-06,
811
+ "loss": 0.000279355444945395,
812
+ "step": 44,
813
+ "step_time": 0.10765942200032441
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": 256.0,
823
+ "completions/max_terminated_length": 0.0,
824
+ "completions/mean_length": 256.0,
825
+ "completions/mean_terminated_length": 0.0,
826
+ "completions/min_length": 256.0,
827
+ "completions/min_terminated_length": 0.0,
828
+ "entropy": 2.246687889099121,
829
+ "epoch": 1.125,
830
+ "frac_reward_zero_std": 0.75,
831
+ "grad_norm": 3.2677783966064453,
832
+ "kl": 0.0954747349023819,
833
+ "learning_rate": 1.9e-06,
834
+ "loss": 0.06269100308418274,
835
+ "num_tokens": 119912.0,
836
+ "reward": 0.8023627996444702,
837
+ "reward_std": 0.11866423487663269,
838
+ "rewards/AnswererRewardFunction/mean": 0.8023627996444702,
839
+ "rewards/AnswererRewardFunction/std": 0.11866423487663269,
840
+ "step": 45,
841
+ "step_time": 7.962604862999797
842
+ },
843
+ {
844
+ "clip_ratio/high_max": 0.001953125,
845
+ "clip_ratio/high_mean": 0.001953125,
846
+ "clip_ratio/low_mean": 0.0,
847
+ "clip_ratio/low_min": 0.0,
848
+ "clip_ratio/region_mean": 0.001953125,
849
+ "entropy": 2.204995632171631,
850
+ "epoch": 1.15,
851
+ "grad_norm": 0.9533382058143616,
852
+ "kl": 0.11488857865333557,
853
+ "learning_rate": 1.875e-06,
854
+ "loss": -0.03102012351155281,
855
+ "step": 46,
856
+ "step_time": 0.11413234800056671
857
+ },
858
+ {
859
+ "clip_ratio/high_max": 0.0,
860
+ "clip_ratio/high_mean": 0.0,
861
+ "clip_ratio/low_mean": 0.0,
862
+ "clip_ratio/low_min": 0.0,
863
+ "clip_ratio/region_mean": 0.0,
864
+ "entropy": 2.193535804748535,
865
+ "epoch": 1.175,
866
+ "grad_norm": 0.028080500662326813,
867
+ "kl": 0.11596246063709259,
868
+ "learning_rate": 1.85e-06,
869
+ "loss": 0.0002899061073549092,
870
+ "step": 47,
871
+ "step_time": 0.1131547060003868
872
+ },
873
+ {
874
+ "clip_ratio/high_max": 0.00390625,
875
+ "clip_ratio/high_mean": 0.00390625,
876
+ "clip_ratio/low_mean": 0.0,
877
+ "clip_ratio/low_min": 0.0,
878
+ "clip_ratio/region_mean": 0.00390625,
879
+ "entropy": 2.2996320724487305,
880
+ "epoch": 1.2,
881
+ "grad_norm": 0.9624528884887695,
882
+ "kl": 0.09764513373374939,
883
+ "learning_rate": 1.8249999999999999e-06,
884
+ "loss": -0.030984658747911453,
885
+ "step": 48,
886
+ "step_time": 0.11348181499943166
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": 256.0,
896
+ "completions/max_terminated_length": 0.0,
897
+ "completions/mean_length": 256.0,
898
+ "completions/mean_terminated_length": 0.0,
899
+ "completions/min_length": 256.0,
900
+ "completions/min_terminated_length": 0.0,
901
+ "entropy": 2.1679866313934326,
902
+ "epoch": 1.225,
903
+ "frac_reward_zero_std": 1.0,
904
+ "grad_norm": 0.027085669338703156,
905
+ "kl": 0.10030621290206909,
906
+ "learning_rate": 1.8e-06,
907
+ "loss": 0.0002507655299268663,
908
+ "num_tokens": 129820.0,
909
+ "reward": 0.8280093669891357,
910
+ "reward_std": 0.10857454687356949,
911
+ "rewards/AnswererRewardFunction/mean": 0.8280093669891357,
912
+ "rewards/AnswererRewardFunction/std": 0.10857456177473068,
913
+ "step": 49,
914
+ "step_time": 7.983247337000648
915
+ },
916
+ {
917
+ "clip_ratio/high_max": 0.0,
918
+ "clip_ratio/high_mean": 0.0,
919
+ "clip_ratio/low_mean": 0.0,
920
+ "clip_ratio/low_min": 0.0,
921
+ "clip_ratio/region_mean": 0.0,
922
+ "entropy": 2.363748073577881,
923
+ "epoch": 1.25,
924
+ "grad_norm": 0.042082529515028,
925
+ "kl": 0.12735603749752045,
926
+ "learning_rate": 1.7750000000000002e-06,
927
+ "loss": 0.00031839008443057537,
928
+ "step": 50,
929
+ "step_time": 0.11199663499974122
930
+ },
931
+ {
932
+ "clip_ratio/high_max": 0.0,
933
+ "clip_ratio/high_mean": 0.0,
934
+ "clip_ratio/low_mean": 0.0,
935
+ "clip_ratio/low_min": 0.0,
936
+ "clip_ratio/region_mean": 0.0,
937
+ "entropy": 2.1276743412017822,
938
+ "epoch": 1.275,
939
+ "grad_norm": 0.03107348270714283,
940
+ "kl": 0.11797476559877396,
941
+ "learning_rate": 1.7500000000000002e-06,
942
+ "loss": 0.0002949369081761688,
943
+ "step": 51,
944
+ "step_time": 0.11148821400001907
945
+ },
946
+ {
947
+ "clip_ratio/high_max": 0.0,
948
+ "clip_ratio/high_mean": 0.0,
949
+ "clip_ratio/low_mean": 0.0,
950
+ "clip_ratio/low_min": 0.0,
951
+ "clip_ratio/region_mean": 0.0,
952
+ "entropy": 2.328831672668457,
953
+ "epoch": 1.3,
954
+ "grad_norm": 0.038491472601890564,
955
+ "kl": 0.13311246037483215,
956
+ "learning_rate": 1.725e-06,
957
+ "loss": 0.0003327811718918383,
958
+ "step": 52,
959
+ "step_time": 0.11236872300014511
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": 256.0,
969
+ "completions/max_terminated_length": 0.0,
970
+ "completions/mean_length": 256.0,
971
+ "completions/mean_terminated_length": 0.0,
972
+ "completions/min_length": 256.0,
973
+ "completions/min_terminated_length": 0.0,
974
+ "entropy": 2.4427859783172607,
975
+ "epoch": 1.325,
976
+ "frac_reward_zero_std": 1.0,
977
+ "grad_norm": 0.02960454300045967,
978
+ "kl": 0.10422872751951218,
979
+ "learning_rate": 1.7e-06,
980
+ "loss": 0.00026057183276861906,
981
+ "num_tokens": 139968.0,
982
+ "reward": 0.8848026990890503,
983
+ "reward_std": 0.08498957008123398,
984
+ "rewards/AnswererRewardFunction/mean": 0.8848026990890503,
985
+ "rewards/AnswererRewardFunction/std": 0.08498956263065338,
986
+ "step": 53,
987
+ "step_time": 7.788033633999476
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": 2.5127248764038086,
996
+ "epoch": 1.35,
997
+ "grad_norm": 0.04145759716629982,
998
+ "kl": 0.1229250580072403,
999
+ "learning_rate": 1.675e-06,
1000
+ "loss": 0.0003073126426897943,
1001
+ "step": 54,
1002
+ "step_time": 0.11092337199988833
1003
+ },
1004
+ {
1005
+ "clip_ratio/high_max": 0.0,
1006
+ "clip_ratio/high_mean": 0.0,
1007
+ "clip_ratio/low_mean": 0.0,
1008
+ "clip_ratio/low_min": 0.0,
1009
+ "clip_ratio/region_mean": 0.0,
1010
+ "entropy": 2.483670234680176,
1011
+ "epoch": 1.375,
1012
+ "grad_norm": 0.03221055120229721,
1013
+ "kl": 0.11193286627531052,
1014
+ "learning_rate": 1.65e-06,
1015
+ "loss": 0.0002798321656882763,
1016
+ "step": 55,
1017
+ "step_time": 0.11110094600007869
1018
+ },
1019
+ {
1020
+ "clip_ratio/high_max": 0.0,
1021
+ "clip_ratio/high_mean": 0.0,
1022
+ "clip_ratio/low_mean": 0.0,
1023
+ "clip_ratio/low_min": 0.0,
1024
+ "clip_ratio/region_mean": 0.0,
1025
+ "entropy": 2.4600026607513428,
1026
+ "epoch": 1.4,
1027
+ "grad_norm": 0.032304368913173676,
1028
+ "kl": 0.1237458735704422,
1029
+ "learning_rate": 1.625e-06,
1030
+ "loss": 0.0003093646664638072,
1031
+ "step": 56,
1032
+ "step_time": 0.1109393220003767
1033
+ },
1034
+ {
1035
+ "clip_ratio/high_max": 0.0,
1036
+ "clip_ratio/high_mean": 0.0,
1037
+ "clip_ratio/low_mean": 0.0,
1038
+ "clip_ratio/low_min": 0.0,
1039
+ "clip_ratio/region_mean": 0.0,
1040
+ "completions/clipped_ratio": 1.0,
1041
+ "completions/max_length": 256.0,
1042
+ "completions/max_terminated_length": 0.0,
1043
+ "completions/mean_length": 256.0,
1044
+ "completions/mean_terminated_length": 0.0,
1045
+ "completions/min_length": 256.0,
1046
+ "completions/min_terminated_length": 0.0,
1047
+ "entropy": 1.7456945180892944,
1048
+ "epoch": 1.425,
1049
+ "frac_reward_zero_std": 0.5,
1050
+ "grad_norm": 1.4578479528427124,
1051
+ "kl": 0.09936054795980453,
1052
+ "learning_rate": 1.6e-06,
1053
+ "loss": 0.031486742198467255,
1054
+ "num_tokens": 150328.0,
1055
+ "reward": 0.8651799559593201,
1056
+ "reward_std": 0.18344323337078094,
1057
+ "rewards/AnswererRewardFunction/mean": 0.8651799559593201,
1058
+ "rewards/AnswererRewardFunction/std": 0.18344323337078094,
1059
+ "step": 57,
1060
+ "step_time": 7.922721363000164
1061
+ },
1062
+ {
1063
+ "clip_ratio/high_max": 0.0,
1064
+ "clip_ratio/high_mean": 0.0,
1065
+ "clip_ratio/low_mean": 0.0009765625,
1066
+ "clip_ratio/low_min": 0.0009765625,
1067
+ "clip_ratio/region_mean": 0.0009765625,
1068
+ "entropy": 1.9535131454467773,
1069
+ "epoch": 1.45,
1070
+ "grad_norm": 1.1535450220108032,
1071
+ "kl": 0.12368227541446686,
1072
+ "learning_rate": 1.5750000000000002e-06,
1073
+ "loss": 0.06276790797710419,
1074
+ "step": 58,
1075
+ "step_time": 0.11167901599947072
1076
+ },
1077
+ {
1078
+ "clip_ratio/high_max": 0.005859375,
1079
+ "clip_ratio/high_mean": 0.005859375,
1080
+ "clip_ratio/low_mean": 0.0,
1081
+ "clip_ratio/low_min": 0.0,
1082
+ "clip_ratio/region_mean": 0.005859375,
1083
+ "entropy": 2.162888526916504,
1084
+ "epoch": 1.475,
1085
+ "grad_norm": 1.3076512813568115,
1086
+ "kl": 0.10642383247613907,
1087
+ "learning_rate": 1.5500000000000002e-06,
1088
+ "loss": -0.062260571867227554,
1089
+ "step": 59,
1090
+ "step_time": 0.1131628670000282
1091
+ },
1092
+ {
1093
+ "clip_ratio/high_max": 0.001953125,
1094
+ "clip_ratio/high_mean": 0.001953125,
1095
+ "clip_ratio/low_mean": 0.0,
1096
+ "clip_ratio/low_min": 0.0,
1097
+ "clip_ratio/region_mean": 0.001953125,
1098
+ "entropy": 2.2302868366241455,
1099
+ "epoch": 1.5,
1100
+ "grad_norm": 1.0681862831115723,
1101
+ "kl": 0.1137072965502739,
1102
+ "learning_rate": 1.525e-06,
1103
+ "loss": -0.030728042125701904,
1104
+ "step": 60,
1105
+ "step_time": 0.11138374100028159
1106
+ }
1107
+ ],
1108
+ "logging_steps": 1,
1109
+ "max_steps": 120,
1110
+ "num_input_tokens_seen": 150328,
1111
+ "num_train_epochs": 3,
1112
+ "save_steps": 30,
1113
+ "stateful_callbacks": {
1114
+ "TrainerControl": {
1115
+ "args": {
1116
+ "should_epoch_stop": false,
1117
+ "should_evaluate": false,
1118
+ "should_log": false,
1119
+ "should_save": true,
1120
+ "should_training_stop": false
1121
+ },
1122
+ "attributes": {}
1123
+ }
1124
+ },
1125
+ "total_flos": 0.0,
1126
+ "train_batch_size": 4,
1127
+ "trial_name": null,
1128
+ "trial_params": null
1129
+ }
self_play_hf_l40s_full/round_006/answerer_train/checkpoint-60/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:99b43de3228fc906dd533a65034a9a07e1b0017bc5ef4fa7868a1a78d1cee45d
3
+ size 7249