Hugging Face's logo

AdversarialRLHF
/
rloo_pythia410m_tldr6.9b_rm410mdata_mergedsft_prefix_nokl

Safetensors
gpt_neox
Model card Files
xet
Community
Muqeeth commited on
Commit
6173354
·
verified ·
1 Parent(s): fc6f9e7

Training in progress, step 26, checkpoint

Browse files
Files changed (11) hide show
  1. checkpoint-26/config.json +31 -0
  2. checkpoint-26/generation_config.json +5 -0
  3. checkpoint-26/model.safetensors +3 -0
  4. checkpoint-26/optimizer.pt +3 -0
  5. checkpoint-26/rng_state.pth +3 -0
  6. checkpoint-26/scheduler.pt +3 -0
  7. checkpoint-26/special_tokens_map.json +30 -0
  8. checkpoint-26/tokenizer.json +0 -0
  9. checkpoint-26/tokenizer_config.json +220 -0
  10. checkpoint-26/trainer_state.json +555 -0
  11. checkpoint-26/training_args.bin +3 -0
checkpoint-26/config.json ADDED
@@ -0,0 +1,31 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "architectures": [
3
+ "GPTNeoXForCausalLM"
4
+ ],
5
+ "attention_bias": true,
6
+ "attention_dropout": 0.0,
7
+ "bos_token_id": 0,
8
+ "classifier_dropout": 0.1,
9
+ "eos_token_id": 0,
10
+ "hidden_act": "gelu",
11
+ "hidden_dropout": 0.0,
12
+ "hidden_size": 1024,
13
+ "initializer_range": 0.02,
14
+ "intermediate_size": 4096,
15
+ "layer_norm_eps": 1e-05,
16
+ "max_position_embeddings": 2048,
17
+ "model_type": "gpt_neox",
18
+ "num_attention_heads": 16,
19
+ "num_hidden_layers": 24,
20
+ "partial_rotary_factor": 0.25,
21
+ "rope_scaling": null,
22
+ "rope_theta": 10000,
23
+ "rotary_emb_base": 10000,
24
+ "rotary_pct": 0.25,
25
+ "tie_word_embeddings": false,
26
+ "torch_dtype": "float32",
27
+ "transformers_version": "4.50.3",
28
+ "use_cache": true,
29
+ "use_parallel_residual": true,
30
+ "vocab_size": 50304
31
+ }
checkpoint-26/generation_config.json ADDED
@@ -0,0 +1,5 @@
 
 
 
 
 
 
1
+ {
2
+ "_from_model_config": true,
3
+ "bos_token_id": 0,
4
+ "transformers_version": "4.50.3"
5
+ }
checkpoint-26/model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:558057e409d98dd9008b52db550e18e66e122abfe1942f85a6609f55e2333130
3
+ size 1621370224
checkpoint-26/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d8430a69313a350ee6b97884ae4dd6ec5ffadfc69f998f6bcb292549ddbfd8db
3
+ size 3242916026
checkpoint-26/rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e379c8fcd0129e9287f2a2382c4d336eae7a403512ac6fc98509ec16c84e7ab4
3
+ size 14244
checkpoint-26/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8bcafb562e9b82d203a11d0302e8ec4228c3b9f47472a5ec3efb1cd79e8b6081
3
+ size 1064
checkpoint-26/special_tokens_map.json ADDED
@@ -0,0 +1,30 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "bos_token": {
3
+ "content": "<|endoftext|>",
4
+ "lstrip": false,
5
+ "normalized": false,
6
+ "rstrip": false,
7
+ "single_word": false
8
+ },
9
+ "eos_token": {
10
+ "content": "<|endoftext|>",
11
+ "lstrip": false,
12
+ "normalized": false,
13
+ "rstrip": false,
14
+ "single_word": false
15
+ },
16
+ "pad_token": {
17
+ "content": "<|padding|>",
18
+ "lstrip": false,
19
+ "normalized": false,
20
+ "rstrip": false,
21
+ "single_word": false
22
+ },
23
+ "unk_token": {
24
+ "content": "<|endoftext|>",
25
+ "lstrip": false,
26
+ "normalized": false,
27
+ "rstrip": false,
28
+ "single_word": false
29
+ }
30
+ }
checkpoint-26/tokenizer.json ADDED
The diff for this file is too large to render. See raw diff
 
checkpoint-26/tokenizer_config.json ADDED
@@ -0,0 +1,220 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "add_bos_token": false,
3
+ "add_eos_token": false,
4
+ "add_prefix_space": false,
5
+ "added_tokens_decoder": {
6
+ "0": {
7
+ "content": "<|endoftext|>",
8
+ "lstrip": false,
9
+ "normalized": false,
10
+ "rstrip": false,
11
+ "single_word": false,
12
+ "special": true
13
+ },
14
+ "1": {
15
+ "content": "<|padding|>",
16
+ "lstrip": false,
17
+ "normalized": false,
18
+ "rstrip": false,
19
+ "single_word": false,
20
+ "special": true
21
+ },
22
+ "50254": {
23
+ "content": " ",
24
+ "lstrip": false,
25
+ "normalized": true,
26
+ "rstrip": false,
27
+ "single_word": false,
28
+ "special": false
29
+ },
30
+ "50255": {
31
+ "content": " ",
32
+ "lstrip": false,
33
+ "normalized": true,
34
+ "rstrip": false,
35
+ "single_word": false,
36
+ "special": false
37
+ },
38
+ "50256": {
39
+ "content": " ",
40
+ "lstrip": false,
41
+ "normalized": true,
42
+ "rstrip": false,
43
+ "single_word": false,
44
+ "special": false
45
+ },
46
+ "50257": {
47
+ "content": " ",
48
+ "lstrip": false,
49
+ "normalized": true,
50
+ "rstrip": false,
51
+ "single_word": false,
52
+ "special": false
53
+ },
54
+ "50258": {
55
+ "content": " ",
56
+ "lstrip": false,
57
+ "normalized": true,
58
+ "rstrip": false,
59
+ "single_word": false,
60
+ "special": false
61
+ },
62
+ "50259": {
63
+ "content": " ",
64
+ "lstrip": false,
65
+ "normalized": true,
66
+ "rstrip": false,
67
+ "single_word": false,
68
+ "special": false
69
+ },
70
+ "50260": {
71
+ "content": " ",
72
+ "lstrip": false,
73
+ "normalized": true,
74
+ "rstrip": false,
75
+ "single_word": false,
76
+ "special": false
77
+ },
78
+ "50261": {
79
+ "content": " ",
80
+ "lstrip": false,
81
+ "normalized": true,
82
+ "rstrip": false,
83
+ "single_word": false,
84
+ "special": false
85
+ },
86
+ "50262": {
87
+ "content": " ",
88
+ "lstrip": false,
89
+ "normalized": true,
90
+ "rstrip": false,
91
+ "single_word": false,
92
+ "special": false
93
+ },
94
+ "50263": {
95
+ "content": " ",
96
+ "lstrip": false,
97
+ "normalized": true,
98
+ "rstrip": false,
99
+ "single_word": false,
100
+ "special": false
101
+ },
102
+ "50264": {
103
+ "content": " ",
104
+ "lstrip": false,
105
+ "normalized": true,
106
+ "rstrip": false,
107
+ "single_word": false,
108
+ "special": false
109
+ },
110
+ "50265": {
111
+ "content": " ",
112
+ "lstrip": false,
113
+ "normalized": true,
114
+ "rstrip": false,
115
+ "single_word": false,
116
+ "special": false
117
+ },
118
+ "50266": {
119
+ "content": " ",
120
+ "lstrip": false,
121
+ "normalized": true,
122
+ "rstrip": false,
123
+ "single_word": false,
124
+ "special": false
125
+ },
126
+ "50267": {
127
+ "content": " ",
128
+ "lstrip": false,
129
+ "normalized": true,
130
+ "rstrip": false,
131
+ "single_word": false,
132
+ "special": false
133
+ },
134
+ "50268": {
135
+ "content": " ",
136
+ "lstrip": false,
137
+ "normalized": true,
138
+ "rstrip": false,
139
+ "single_word": false,
140
+ "special": false
141
+ },
142
+ "50269": {
143
+ "content": " ",
144
+ "lstrip": false,
145
+ "normalized": true,
146
+ "rstrip": false,
147
+ "single_word": false,
148
+ "special": false
149
+ },
150
+ "50270": {
151
+ "content": " ",
152
+ "lstrip": false,
153
+ "normalized": true,
154
+ "rstrip": false,
155
+ "single_word": false,
156
+ "special": false
157
+ },
158
+ "50271": {
159
+ "content": " ",
160
+ "lstrip": false,
161
+ "normalized": true,
162
+ "rstrip": false,
163
+ "single_word": false,
164
+ "special": false
165
+ },
166
+ "50272": {
167
+ "content": " ",
168
+ "lstrip": false,
169
+ "normalized": true,
170
+ "rstrip": false,
171
+ "single_word": false,
172
+ "special": false
173
+ },
174
+ "50273": {
175
+ "content": " ",
176
+ "lstrip": false,
177
+ "normalized": true,
178
+ "rstrip": false,
179
+ "single_word": false,
180
+ "special": false
181
+ },
182
+ "50274": {
183
+ "content": " ",
184
+ "lstrip": false,
185
+ "normalized": true,
186
+ "rstrip": false,
187
+ "single_word": false,
188
+ "special": false
189
+ },
190
+ "50275": {
191
+ "content": " ",
192
+ "lstrip": false,
193
+ "normalized": true,
194
+ "rstrip": false,
195
+ "single_word": false,
196
+ "special": false
197
+ },
198
+ "50276": {
199
+ "content": " ",
200
+ "lstrip": false,
201
+ "normalized": true,
202
+ "rstrip": false,
203
+ "single_word": false,
204
+ "special": false
205
+ }
206
+ },
207
+ "bos_token": "<|endoftext|>",
208
+ "clean_up_tokenization_spaces": true,
209
+ "eos_token": "<|endoftext|>",
210
+ "extra_special_tokens": {},
211
+ "max_length": 580,
212
+ "model_max_length": 1000000000000000019884624838656,
213
+ "pad_token": "<|padding|>",
214
+ "padding_side": "left",
215
+ "stride": 0,
216
+ "tokenizer_class": "GPTNeoXTokenizer",
217
+ "truncation_side": "right",
218
+ "truncation_strategy": "longest_first",
219
+ "unk_token": "<|endoftext|>"
220
+ }
checkpoint-26/trainer_state.json ADDED
@@ -0,0 +1,555 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_global_step": null,
3
+ "best_metric": null,
4
+ "best_model_checkpoint": null,
5
+ "episode": 13312,
6
+ "epoch": 0.05702535983550377,
7
+ "eval_steps": 500,
8
+ "global_step": 26,
9
+ "is_hyper_param_search": false,
10
+ "is_local_process_zero": true,
11
+ "is_world_process_zero": true,
12
+ "log_history": [
13
+ {
14
+ "episode": 512,
15
+ "epoch": 0.0021932830705962986,
16
+ "eps": 6,
17
+ "loss/policy_avg": 0.15268513560295105,
18
+ "lr": 3e-06,
19
+ "objective/entropy": -45.407432556152344,
20
+ "objective/kl": 0.3935524821281433,
21
+ "objective/non_score_reward": 0.0,
22
+ "objective/rlhf_reward": 4.679054260253906,
23
+ "objective/scores": 4.679054260253906,
24
+ "policy/approxkl_avg": 0.11690396815538406,
25
+ "policy/clipfrac_avg": 0.34375,
26
+ "policy/entropy_avg": 0.8878304958343506,
27
+ "step": 0,
28
+ "val/clipfrac_avg": 0.0,
29
+ "val/num_eos_tokens": 2729,
30
+ "val/ratio": 1.000101089477539,
31
+ "val/ratio_var": 1.6687282595739816e-06
32
+ },
33
+ {
34
+ "episode": 1024,
35
+ "epoch": 0.004386566141192597,
36
+ "eps": 6,
37
+ "loss/policy_avg": 0.10589599609375,
38
+ "lr": 2.9882812500000002e-06,
39
+ "objective/entropy": -51.535125732421875,
40
+ "objective/kl": 0.6498137712478638,
41
+ "objective/non_score_reward": 0.0,
42
+ "objective/rlhf_reward": 4.541956901550293,
43
+ "objective/scores": 4.541956901550293,
44
+ "policy/approxkl_avg": 0.10543511807918549,
45
+ "policy/clipfrac_avg": 0.3203125,
46
+ "policy/entropy_avg": 0.8243188858032227,
47
+ "step": 1,
48
+ "val/clipfrac_avg": 0.0,
49
+ "val/num_eos_tokens": 3325,
50
+ "val/ratio": 1.0002458095550537,
51
+ "val/ratio_var": 1.5084623328220914e-06
52
+ },
53
+ {
54
+ "episode": 1536,
55
+ "epoch": 0.006579849211788897,
56
+ "eps": 6,
57
+ "loss/policy_avg": 0.13810233771800995,
58
+ "lr": 2.9765625e-06,
59
+ "objective/entropy": -55.89725112915039,
60
+ "objective/kl": 1.300195574760437,
61
+ "objective/non_score_reward": 0.0,
62
+ "objective/rlhf_reward": 4.745310306549072,
63
+ "objective/scores": 4.745310306549072,
64
+ "policy/approxkl_avg": 0.11424587666988373,
65
+ "policy/clipfrac_avg": 0.302734375,
66
+ "policy/entropy_avg": 0.760991096496582,
67
+ "step": 2,
68
+ "val/clipfrac_avg": 0.0,
69
+ "val/num_eos_tokens": 3801,
70
+ "val/ratio": 0.9999536275863647,
71
+ "val/ratio_var": 1.5804212125658523e-06
72
+ },
73
+ {
74
+ "episode": 2048,
75
+ "epoch": 0.008773132282385195,
76
+ "eps": 6,
77
+ "loss/policy_avg": 0.25361043214797974,
78
+ "lr": 2.96484375e-06,
79
+ "objective/entropy": -54.02920150756836,
80
+ "objective/kl": 1.721588373184204,
81
+ "objective/non_score_reward": 0.0,
82
+ "objective/rlhf_reward": 4.8178606033325195,
83
+ "objective/scores": 4.8178606033325195,
84
+ "policy/approxkl_avg": 0.12672686576843262,
85
+ "policy/clipfrac_avg": 0.310546875,
86
+ "policy/entropy_avg": 0.7554011344909668,
87
+ "step": 3,
88
+ "val/clipfrac_avg": 0.0,
89
+ "val/num_eos_tokens": 4010,
90
+ "val/ratio": 0.9999880790710449,
91
+ "val/ratio_var": 1.6228774484261521e-06
92
+ },
93
+ {
94
+ "episode": 2560,
95
+ "epoch": 0.010966415352981495,
96
+ "eps": 6,
97
+ "loss/policy_avg": 0.17445668578147888,
98
+ "lr": 2.953125e-06,
99
+ "objective/entropy": -60.463539123535156,
100
+ "objective/kl": 2.6447153091430664,
101
+ "objective/non_score_reward": 0.0,
102
+ "objective/rlhf_reward": 4.877408981323242,
103
+ "objective/scores": 4.877408981323242,
104
+ "policy/approxkl_avg": 0.13307543098926544,
105
+ "policy/clipfrac_avg": 0.275390625,
106
+ "policy/entropy_avg": 0.690821647644043,
107
+ "step": 4,
108
+ "val/clipfrac_avg": 0.0,
109
+ "val/num_eos_tokens": 4589,
110
+ "val/ratio": 1.0001537799835205,
111
+ "val/ratio_var": 1.283996084566752e-06
112
+ },
113
+ {
114
+ "episode": 3072,
115
+ "epoch": 0.013159698423577794,
116
+ "eps": 7,
117
+ "loss/policy_avg": 0.18251696228981018,
118
+ "lr": 2.94140625e-06,
119
+ "objective/entropy": -60.38776397705078,
120
+ "objective/kl": 3.3835220336914062,
121
+ "objective/non_score_reward": 0.0,
122
+ "objective/rlhf_reward": 5.029230117797852,
123
+ "objective/scores": 5.029230117797852,
124
+ "policy/approxkl_avg": 0.11075231432914734,
125
+ "policy/clipfrac_avg": 0.30859375,
126
+ "policy/entropy_avg": 0.6673464775085449,
127
+ "step": 5,
128
+ "val/clipfrac_avg": 0.0,
129
+ "val/num_eos_tokens": 4933,
130
+ "val/ratio": 1.0000388622283936,
131
+ "val/ratio_var": 1.886573841147765e-06
132
+ },
133
+ {
134
+ "episode": 3584,
135
+ "epoch": 0.015352981494174092,
136
+ "eps": 7,
137
+ "loss/policy_avg": 0.0862584114074707,
138
+ "lr": 2.9296875e-06,
139
+ "objective/entropy": -64.953857421875,
140
+ "objective/kl": 3.801461696624756,
141
+ "objective/non_score_reward": 0.0,
142
+ "objective/rlhf_reward": 5.2262468338012695,
143
+ "objective/scores": 5.2262468338012695,
144
+ "policy/approxkl_avg": 0.0921158567070961,
145
+ "policy/clipfrac_avg": 0.31640625,
146
+ "policy/entropy_avg": 0.6317387819290161,
147
+ "step": 6,
148
+ "val/clipfrac_avg": 0.0,
149
+ "val/num_eos_tokens": 5224,
150
+ "val/ratio": 1.0000205039978027,
151
+ "val/ratio_var": 1.223428625962697e-06
152
+ },
153
+ {
154
+ "episode": 4096,
155
+ "epoch": 0.01754626456477039,
156
+ "eps": 7,
157
+ "loss/policy_avg": 0.10839146375656128,
158
+ "lr": 2.91796875e-06,
159
+ "objective/entropy": -64.41149139404297,
160
+ "objective/kl": 4.314542770385742,
161
+ "objective/non_score_reward": 0.0,
162
+ "objective/rlhf_reward": 5.495314121246338,
163
+ "objective/scores": 5.495314121246338,
164
+ "policy/approxkl_avg": 0.11755571514368057,
165
+ "policy/clipfrac_avg": 0.28515625,
166
+ "policy/entropy_avg": 0.6287920475006104,
167
+ "step": 7,
168
+ "val/clipfrac_avg": 0.0,
169
+ "val/num_eos_tokens": 5463,
170
+ "val/ratio": 0.9998381733894348,
171
+ "val/ratio_var": 1.6926536545724957e-06
172
+ },
173
+ {
174
+ "episode": 4608,
175
+ "epoch": 0.01973954763536669,
176
+ "eps": 7,
177
+ "loss/policy_avg": 0.16045916080474854,
178
+ "lr": 2.90625e-06,
179
+ "objective/entropy": -65.33601379394531,
180
+ "objective/kl": 4.707500457763672,
181
+ "objective/non_score_reward": 0.0,
182
+ "objective/rlhf_reward": 5.608126640319824,
183
+ "objective/scores": 5.608126640319824,
184
+ "policy/approxkl_avg": 0.10813738405704498,
185
+ "policy/clipfrac_avg": 0.2890625,
186
+ "policy/entropy_avg": 0.6515289545059204,
187
+ "step": 8,
188
+ "val/clipfrac_avg": 0.0,
189
+ "val/num_eos_tokens": 5590,
190
+ "val/ratio": 1.0000832080841064,
191
+ "val/ratio_var": 1.661756641624379e-06
192
+ },
193
+ {
194
+ "episode": 5120,
195
+ "epoch": 0.02193283070596299,
196
+ "eps": 7,
197
+ "loss/policy_avg": 0.16000822186470032,
198
+ "lr": 2.89453125e-06,
199
+ "objective/entropy": -64.82266235351562,
200
+ "objective/kl": 5.513130187988281,
201
+ "objective/non_score_reward": 0.0,
202
+ "objective/rlhf_reward": 5.976546764373779,
203
+ "objective/scores": 5.976546764373779,
204
+ "policy/approxkl_avg": 0.12319613248109818,
205
+ "policy/clipfrac_avg": 0.29296875,
206
+ "policy/entropy_avg": 0.6444021463394165,
207
+ "step": 9,
208
+ "val/clipfrac_avg": 0.0,
209
+ "val/num_eos_tokens": 5330,
210
+ "val/ratio": 0.9997374415397644,
211
+ "val/ratio_var": 1.3334141613086103e-06
212
+ },
213
+ {
214
+ "episode": 5632,
215
+ "epoch": 0.024126113776559287,
216
+ "eps": 7,
217
+ "loss/policy_avg": 0.22769010066986084,
218
+ "lr": 2.8828125e-06,
219
+ "objective/entropy": -62.54362487792969,
220
+ "objective/kl": 6.375988960266113,
221
+ "objective/non_score_reward": 0.0,
222
+ "objective/rlhf_reward": 6.329231262207031,
223
+ "objective/scores": 6.329231262207031,
224
+ "policy/approxkl_avg": 0.133827343583107,
225
+ "policy/clipfrac_avg": 0.30078125,
226
+ "policy/entropy_avg": 0.6720603704452515,
227
+ "step": 10,
228
+ "val/clipfrac_avg": 0.0,
229
+ "val/num_eos_tokens": 4956,
230
+ "val/ratio": 0.9999898672103882,
231
+ "val/ratio_var": 2.208806336057023e-06
232
+ },
233
+ {
234
+ "episode": 6144,
235
+ "epoch": 0.026319396847155587,
236
+ "eps": 7,
237
+ "loss/policy_avg": 0.16269627213478088,
238
+ "lr": 2.87109375e-06,
239
+ "objective/entropy": -56.140655517578125,
240
+ "objective/kl": 7.935493469238281,
241
+ "objective/non_score_reward": 0.0,
242
+ "objective/rlhf_reward": 6.335080146789551,
243
+ "objective/scores": 6.335080146789551,
244
+ "policy/approxkl_avg": 0.13188697397708893,
245
+ "policy/clipfrac_avg": 0.35546875,
246
+ "policy/entropy_avg": 0.6935639381408691,
247
+ "step": 11,
248
+ "val/clipfrac_avg": 0.0,
249
+ "val/num_eos_tokens": 4904,
250
+ "val/ratio": 0.999968409538269,
251
+ "val/ratio_var": 1.8624212998474832e-06
252
+ },
253
+ {
254
+ "episode": 6656,
255
+ "epoch": 0.028512679917751884,
256
+ "eps": 7,
257
+ "loss/policy_avg": 0.1726413369178772,
258
+ "lr": 2.859375e-06,
259
+ "objective/entropy": -56.71982955932617,
260
+ "objective/kl": 10.505476951599121,
261
+ "objective/non_score_reward": 0.0,
262
+ "objective/rlhf_reward": 7.327908515930176,
263
+ "objective/scores": 7.327908515930176,
264
+ "policy/approxkl_avg": 0.13946227729320526,
265
+ "policy/clipfrac_avg": 0.306640625,
266
+ "policy/entropy_avg": 0.6527446508407593,
267
+ "step": 12,
268
+ "val/clipfrac_avg": 0.0,
269
+ "val/num_eos_tokens": 4210,
270
+ "val/ratio": 1.0002007484436035,
271
+ "val/ratio_var": 1.9585920654208167e-06
272
+ },
273
+ {
274
+ "episode": 7168,
275
+ "epoch": 0.030705962988348184,
276
+ "eps": 7,
277
+ "loss/policy_avg": 0.1594904661178589,
278
+ "lr": 2.84765625e-06,
279
+ "objective/entropy": -57.271263122558594,
280
+ "objective/kl": 12.708869934082031,
281
+ "objective/non_score_reward": 0.0,
282
+ "objective/rlhf_reward": 7.788185119628906,
283
+ "objective/scores": 7.788185119628906,
284
+ "policy/approxkl_avg": 0.14631032943725586,
285
+ "policy/clipfrac_avg": 0.33203125,
286
+ "policy/entropy_avg": 0.6090783476829529,
287
+ "step": 13,
288
+ "val/clipfrac_avg": 0.0,
289
+ "val/num_eos_tokens": 3340,
290
+ "val/ratio": 0.9998956918716431,
291
+ "val/ratio_var": 2.45040769186744e-06
292
+ },
293
+ {
294
+ "episode": 7680,
295
+ "epoch": 0.03289924605894448,
296
+ "eps": 7,
297
+ "loss/policy_avg": 0.15880730748176575,
298
+ "lr": 2.8359375e-06,
299
+ "objective/entropy": -53.81092071533203,
300
+ "objective/kl": 15.003689765930176,
301
+ "objective/non_score_reward": 0.0,
302
+ "objective/rlhf_reward": 7.487188339233398,
303
+ "objective/scores": 7.487188339233398,
304
+ "policy/approxkl_avg": 0.17456671595573425,
305
+ "policy/clipfrac_avg": 0.33203125,
306
+ "policy/entropy_avg": 0.5235726237297058,
307
+ "step": 14,
308
+ "val/clipfrac_avg": 0.0,
309
+ "val/num_eos_tokens": 2483,
310
+ "val/ratio": 1.0001764297485352,
311
+ "val/ratio_var": 1.739899175845494e-06
312
+ },
313
+ {
314
+ "episode": 8192,
315
+ "epoch": 0.03509252912954078,
316
+ "eps": 7,
317
+ "loss/policy_avg": 0.16226203739643097,
318
+ "lr": 2.82421875e-06,
319
+ "objective/entropy": -55.564083099365234,
320
+ "objective/kl": 14.753921508789062,
321
+ "objective/non_score_reward": 0.0,
322
+ "objective/rlhf_reward": 7.572874069213867,
323
+ "objective/scores": 7.572874069213867,
324
+ "policy/approxkl_avg": 0.16666629910469055,
325
+ "policy/clipfrac_avg": 0.302734375,
326
+ "policy/entropy_avg": 0.5176196098327637,
327
+ "step": 15,
328
+ "val/clipfrac_avg": 0.0,
329
+ "val/num_eos_tokens": 2405,
330
+ "val/ratio": 0.9999848008155823,
331
+ "val/ratio_var": 2.589519908724469e-06
332
+ },
333
+ {
334
+ "episode": 8704,
335
+ "epoch": 0.03728581220013708,
336
+ "eps": 7,
337
+ "loss/policy_avg": 0.21661071479320526,
338
+ "lr": 2.8125e-06,
339
+ "objective/entropy": -51.922306060791016,
340
+ "objective/kl": 16.144550323486328,
341
+ "objective/non_score_reward": 0.0,
342
+ "objective/rlhf_reward": 7.781338691711426,
343
+ "objective/scores": 7.781338691711426,
344
+ "policy/approxkl_avg": 0.2002604901790619,
345
+ "policy/clipfrac_avg": 0.337890625,
346
+ "policy/entropy_avg": 0.5265668630599976,
347
+ "step": 16,
348
+ "val/clipfrac_avg": 0.0,
349
+ "val/num_eos_tokens": 2439,
350
+ "val/ratio": 0.9997420907020569,
351
+ "val/ratio_var": 2.6209752377326367e-06
352
+ },
353
+ {
354
+ "episode": 9216,
355
+ "epoch": 0.03947909527073338,
356
+ "eps": 7,
357
+ "loss/policy_avg": 0.26599276065826416,
358
+ "lr": 2.80078125e-06,
359
+ "objective/entropy": -50.82372283935547,
360
+ "objective/kl": 17.150707244873047,
361
+ "objective/non_score_reward": 0.0,
362
+ "objective/rlhf_reward": 7.842694282531738,
363
+ "objective/scores": 7.842694282531738,
364
+ "policy/approxkl_avg": 0.19617530703544617,
365
+ "policy/clipfrac_avg": 0.345703125,
366
+ "policy/entropy_avg": 0.5609679222106934,
367
+ "step": 17,
368
+ "val/clipfrac_avg": 0.0,
369
+ "val/num_eos_tokens": 2719,
370
+ "val/ratio": 1.000065803527832,
371
+ "val/ratio_var": 2.653461933732615e-06
372
+ },
373
+ {
374
+ "episode": 9728,
375
+ "epoch": 0.041672378341329676,
376
+ "eps": 7,
377
+ "loss/policy_avg": 0.18615023791790009,
378
+ "lr": 2.7890625e-06,
379
+ "objective/entropy": -50.224365234375,
380
+ "objective/kl": 18.064285278320312,
381
+ "objective/non_score_reward": 0.0,
382
+ "objective/rlhf_reward": 8.09566879272461,
383
+ "objective/scores": 8.09566879272461,
384
+ "policy/approxkl_avg": 0.2271636724472046,
385
+ "policy/clipfrac_avg": 0.37890625,
386
+ "policy/entropy_avg": 0.5856865048408508,
387
+ "step": 18,
388
+ "val/clipfrac_avg": 0.0,
389
+ "val/num_eos_tokens": 2943,
390
+ "val/ratio": 1.0001342296600342,
391
+ "val/ratio_var": 3.7938905279588653e-06
392
+ },
393
+ {
394
+ "episode": 10240,
395
+ "epoch": 0.04386566141192598,
396
+ "eps": 7,
397
+ "loss/policy_avg": 0.20427988469600677,
398
+ "lr": 2.77734375e-06,
399
+ "objective/entropy": -50.77484893798828,
400
+ "objective/kl": 17.688308715820312,
401
+ "objective/non_score_reward": 0.0,
402
+ "objective/rlhf_reward": 7.970284461975098,
403
+ "objective/scores": 7.970284461975098,
404
+ "policy/approxkl_avg": 0.23018702864646912,
405
+ "policy/clipfrac_avg": 0.341796875,
406
+ "policy/entropy_avg": 0.5944638252258301,
407
+ "step": 19,
408
+ "val/clipfrac_avg": 0.0,
409
+ "val/num_eos_tokens": 3183,
410
+ "val/ratio": 1.000162124633789,
411
+ "val/ratio_var": 3.043409833480837e-06
412
+ },
413
+ {
414
+ "episode": 10752,
415
+ "epoch": 0.04605894448252228,
416
+ "eps": 7,
417
+ "loss/policy_avg": 0.282001793384552,
418
+ "lr": 2.765625e-06,
419
+ "objective/entropy": -44.80044937133789,
420
+ "objective/kl": 19.763790130615234,
421
+ "objective/non_score_reward": 0.0,
422
+ "objective/rlhf_reward": 8.012289047241211,
423
+ "objective/scores": 8.012289047241211,
424
+ "policy/approxkl_avg": 0.24563264846801758,
425
+ "policy/clipfrac_avg": 0.3359375,
426
+ "policy/entropy_avg": 0.6678606271743774,
427
+ "step": 20,
428
+ "val/clipfrac_avg": 0.0,
429
+ "val/num_eos_tokens": 3309,
430
+ "val/ratio": 0.9997632503509521,
431
+ "val/ratio_var": 3.618298023866373e-06
432
+ },
433
+ {
434
+ "episode": 11264,
435
+ "epoch": 0.048252227553118573,
436
+ "eps": 7,
437
+ "loss/policy_avg": 0.15888847410678864,
438
+ "lr": 2.75390625e-06,
439
+ "objective/entropy": -45.08258056640625,
440
+ "objective/kl": 19.886899948120117,
441
+ "objective/non_score_reward": 0.0,
442
+ "objective/rlhf_reward": 8.057466506958008,
443
+ "objective/scores": 8.057466506958008,
444
+ "policy/approxkl_avg": 0.2164287269115448,
445
+ "policy/clipfrac_avg": 0.3671875,
446
+ "policy/entropy_avg": 0.6539000272750854,
447
+ "step": 21,
448
+ "val/clipfrac_avg": 0.0,
449
+ "val/num_eos_tokens": 3399,
450
+ "val/ratio": 0.9997671842575073,
451
+ "val/ratio_var": 3.0791625249548815e-06
452
+ },
453
+ {
454
+ "episode": 11776,
455
+ "epoch": 0.05044551062371487,
456
+ "eps": 7,
457
+ "loss/policy_avg": 0.276736855506897,
458
+ "lr": 2.7421875e-06,
459
+ "objective/entropy": -40.99183654785156,
460
+ "objective/kl": 22.14881134033203,
461
+ "objective/non_score_reward": 0.0,
462
+ "objective/rlhf_reward": 8.060696601867676,
463
+ "objective/scores": 8.060696601867676,
464
+ "policy/approxkl_avg": 0.31746962666511536,
465
+ "policy/clipfrac_avg": 0.349609375,
466
+ "policy/entropy_avg": 0.7243172526359558,
467
+ "step": 22,
468
+ "val/clipfrac_avg": 0.0,
469
+ "val/num_eos_tokens": 3907,
470
+ "val/ratio": 0.9996672868728638,
471
+ "val/ratio_var": 4.220366463414393e-06
472
+ },
473
+ {
474
+ "episode": 12288,
475
+ "epoch": 0.052638793694311174,
476
+ "eps": 7,
477
+ "loss/policy_avg": 0.4665297865867615,
478
+ "lr": 2.73046875e-06,
479
+ "objective/entropy": -35.2369384765625,
480
+ "objective/kl": 23.011898040771484,
481
+ "objective/non_score_reward": 0.0,
482
+ "objective/rlhf_reward": 7.889699935913086,
483
+ "objective/scores": 7.889699935913086,
484
+ "policy/approxkl_avg": 0.3784940838813782,
485
+ "policy/clipfrac_avg": 0.3515625,
486
+ "policy/entropy_avg": 0.7655227184295654,
487
+ "step": 23,
488
+ "val/clipfrac_avg": 0.0,
489
+ "val/num_eos_tokens": 3972,
490
+ "val/ratio": 0.9998047351837158,
491
+ "val/ratio_var": 4.525894382823026e-06
492
+ },
493
+ {
494
+ "episode": 12800,
495
+ "epoch": 0.05483207676490747,
496
+ "eps": 7,
497
+ "loss/policy_avg": 0.43642693758010864,
498
+ "lr": 2.71875e-06,
499
+ "objective/entropy": -36.217777252197266,
500
+ "objective/kl": 22.847183227539062,
501
+ "objective/non_score_reward": 0.0,
502
+ "objective/rlhf_reward": 8.134607315063477,
503
+ "objective/scores": 8.134607315063477,
504
+ "policy/approxkl_avg": 0.330811470746994,
505
+ "policy/clipfrac_avg": 0.39453125,
506
+ "policy/entropy_avg": 0.7883225679397583,
507
+ "step": 24,
508
+ "val/clipfrac_avg": 0.0,
509
+ "val/num_eos_tokens": 4373,
510
+ "val/ratio": 0.9999172687530518,
511
+ "val/ratio_var": 5.261089881969383e-06
512
+ },
513
+ {
514
+ "episode": 13312,
515
+ "epoch": 0.05702535983550377,
516
+ "eps": 7,
517
+ "loss/policy_avg": 0.3709805905818939,
518
+ "lr": 2.70703125e-06,
519
+ "objective/entropy": -28.68521499633789,
520
+ "objective/kl": 24.381669998168945,
521
+ "objective/non_score_reward": 0.0,
522
+ "objective/rlhf_reward": 8.424071311950684,
523
+ "objective/scores": 8.424071311950684,
524
+ "policy/approxkl_avg": 0.26343879103660583,
525
+ "policy/clipfrac_avg": 0.359375,
526
+ "policy/entropy_avg": 0.7675451040267944,
527
+ "step": 25,
528
+ "val/clipfrac_avg": 0.0,
529
+ "val/num_eos_tokens": 4221,
530
+ "val/ratio": 1.0003796815872192,
531
+ "val/ratio_var": 4.443480065674521e-06
532
+ }
533
+ ],
534
+ "logging_steps": 1,
535
+ "max_steps": 128,
536
+ "num_input_tokens_seen": 0,
537
+ "num_train_epochs": 1.122960932145305,
538
+ "save_steps": 26,
539
+ "stateful_callbacks": {
540
+ "TrainerControl": {
541
+ "args": {
542
+ "should_epoch_stop": false,
543
+ "should_evaluate": false,
544
+ "should_log": true,
545
+ "should_save": true,
546
+ "should_training_stop": false
547
+ },
548
+ "attributes": {}
549
+ }
550
+ },
551
+ "total_flos": 0,
552
+ "train_batch_size": null,
553
+ "trial_name": null,
554
+ "trial_params": null
555
+ }
checkpoint-26/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a70ca9efe45d40957298b73ba73bfdfa16ec22c723e687576204531d74270deb
3
+ size 6584