yiran-wang3 commited on
Commit
647285a
1 Parent(s): 97903c8

End of training

Browse files
README.md ADDED
@@ -0,0 +1,64 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ library_name: transformers
3
+ license: apache-2.0
4
+ base_model: yiran-wang3/qwen2_coder_reflct_adamw_iter3
5
+ tags:
6
+ - alignment-handbook
7
+ - generated_from_trainer
8
+ - trl
9
+ - dpo
10
+ datasets:
11
+ - self-generate/qwcoder2_reflct_sppo_hard_new_cn_mining_oj_iter3-binarized-reflection-scored
12
+ model-index:
13
+ - name: qwen2_coder_reflct_adamw_iter4
14
+ results: []
15
+ ---
16
+
17
+ <!-- This model card has been generated automatically according to the information the Trainer had access to. You
18
+ should probably proofread and complete it, then remove this comment. -->
19
+
20
+ # qwen2_coder_reflct_adamw_iter4
21
+
22
+ This model is a fine-tuned version of [yiran-wang3/qwen2_coder_reflct_adamw_iter3](https://huggingface.co/yiran-wang3/qwen2_coder_reflct_adamw_iter3) on the self-generate/qwcoder2_reflct_sppo_hard_new_cn_mining_oj_iter3-binarized-reflection-scored dataset.
23
+
24
+ ## Model description
25
+
26
+ More information needed
27
+
28
+ ## Intended uses & limitations
29
+
30
+ More information needed
31
+
32
+ ## Training and evaluation data
33
+
34
+ More information needed
35
+
36
+ ## Training procedure
37
+
38
+ ### Training hyperparameters
39
+
40
+ The following hyperparameters were used during training:
41
+ - learning_rate: 1e-06
42
+ - train_batch_size: 8
43
+ - eval_batch_size: 4
44
+ - seed: 42
45
+ - distributed_type: multi-GPU
46
+ - num_devices: 8
47
+ - total_train_batch_size: 64
48
+ - total_eval_batch_size: 32
49
+ - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
50
+ - lr_scheduler_type: constant
51
+ - lr_scheduler_warmup_ratio: 0.1
52
+ - lr_scheduler_warmup_steps: 100
53
+ - num_epochs: 1.0
54
+
55
+ ### Training results
56
+
57
+
58
+
59
+ ### Framework versions
60
+
61
+ - Transformers 4.45.0
62
+ - Pytorch 2.4.0+cu121
63
+ - Datasets 2.14.6
64
+ - Tokenizers 0.20.2
all_results.json ADDED
@@ -0,0 +1,9 @@
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 1.0,
3
+ "total_flos": 0.0,
4
+ "train_loss": 0.4622946350198043,
5
+ "train_runtime": 144.3728,
6
+ "train_samples": 2428,
7
+ "train_samples_per_second": 16.818,
8
+ "train_steps_per_second": 0.263
9
+ }
config.json CHANGED
@@ -23,7 +23,7 @@
23
  "tie_word_embeddings": true,
24
  "torch_dtype": "bfloat16",
25
  "transformers_version": "4.45.0",
26
- "use_cache": false,
27
  "use_sliding_window": false,
28
  "vocab_size": 151936
29
  }
 
23
  "tie_word_embeddings": true,
24
  "torch_dtype": "bfloat16",
25
  "transformers_version": "4.45.0",
26
+ "use_cache": true,
27
  "use_sliding_window": false,
28
  "vocab_size": 151936
29
  }
generation_config.json ADDED
@@ -0,0 +1,14 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "bos_token_id": 151643,
3
+ "do_sample": true,
4
+ "eos_token_id": [
5
+ 151645,
6
+ 151643
7
+ ],
8
+ "pad_token_id": 151643,
9
+ "repetition_penalty": 1.1,
10
+ "temperature": 0.7,
11
+ "top_k": 20,
12
+ "top_p": 0.8,
13
+ "transformers_version": "4.45.0"
14
+ }
train_results.json ADDED
@@ -0,0 +1,9 @@
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 1.0,
3
+ "total_flos": 0.0,
4
+ "train_loss": 0.4622946350198043,
5
+ "train_runtime": 144.3728,
6
+ "train_samples": 2428,
7
+ "train_samples_per_second": 16.818,
8
+ "train_steps_per_second": 0.263
9
+ }
trainer_state.json ADDED
@@ -0,0 +1,840 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": null,
3
+ "best_model_checkpoint": null,
4
+ "epoch": 1.0,
5
+ "eval_steps": 100,
6
+ "global_step": 38,
7
+ "is_hyper_param_search": false,
8
+ "is_local_process_zero": true,
9
+ "is_world_process_zero": true,
10
+ "log_history": [
11
+ {
12
+ "debug/policy_chosen_logits": -3.0018529891967773,
13
+ "debug/policy_chosen_logps": -178.11398315429688,
14
+ "debug/policy_rejected_logits": -3.148291826248169,
15
+ "debug/policy_rejected_logps": -211.41403198242188,
16
+ "debug/reference_chosen_logps": -178.11398315429688,
17
+ "debug/reference_rejected_logps": -211.41403198242188,
18
+ "epoch": 0.02631578947368421,
19
+ "grad_norm": 7.001357289258973,
20
+ "learning_rate": 1e-06,
21
+ "logits/chosen": -3.0018529891967773,
22
+ "logits/rejected": -3.148291826248169,
23
+ "logps/chosen": -178.11398315429688,
24
+ "logps/rejected": -211.41403198242188,
25
+ "loss": 0.5,
26
+ "rewards/accuracies": 0.0,
27
+ "rewards/chosen": 0.0,
28
+ "rewards/margins": 0.0,
29
+ "rewards/rejected": 0.0,
30
+ "step": 1
31
+ },
32
+ {
33
+ "debug/policy_chosen_logits": -3.109767436981201,
34
+ "debug/policy_chosen_logps": -181.03724670410156,
35
+ "debug/policy_rejected_logits": -3.0829992294311523,
36
+ "debug/policy_rejected_logps": -192.5626220703125,
37
+ "debug/reference_chosen_logps": -180.47354125976562,
38
+ "debug/reference_rejected_logps": -192.10348510742188,
39
+ "epoch": 0.05263157894736842,
40
+ "grad_norm": 6.690576807834117,
41
+ "learning_rate": 1e-06,
42
+ "logits/chosen": -3.109767436981201,
43
+ "logits/rejected": -3.0829992294311523,
44
+ "logps/chosen": -181.03724670410156,
45
+ "logps/rejected": -192.5626220703125,
46
+ "loss": 0.5012,
47
+ "rewards/accuracies": 0.5,
48
+ "rewards/chosen": -0.005636939778923988,
49
+ "rewards/margins": -0.0010454560397192836,
50
+ "rewards/rejected": -0.004591483622789383,
51
+ "step": 2
52
+ },
53
+ {
54
+ "debug/policy_chosen_logits": -3.1953020095825195,
55
+ "debug/policy_chosen_logps": -168.2648162841797,
56
+ "debug/policy_rejected_logits": -3.0766544342041016,
57
+ "debug/policy_rejected_logps": -181.56866455078125,
58
+ "debug/reference_chosen_logps": -168.64642333984375,
59
+ "debug/reference_rejected_logps": -181.52621459960938,
60
+ "epoch": 0.07894736842105263,
61
+ "grad_norm": 10.217915675316835,
62
+ "learning_rate": 1e-06,
63
+ "logits/chosen": -3.1953020095825195,
64
+ "logits/rejected": -3.0766544342041016,
65
+ "logps/chosen": -168.2648162841797,
66
+ "logps/rejected": -181.56866455078125,
67
+ "loss": 0.4988,
68
+ "rewards/accuracies": 0.5,
69
+ "rewards/chosen": 0.0038160705007612705,
70
+ "rewards/margins": 0.004240531474351883,
71
+ "rewards/rejected": -0.00042446135194040835,
72
+ "step": 3
73
+ },
74
+ {
75
+ "debug/policy_chosen_logits": -2.936859607696533,
76
+ "debug/policy_chosen_logps": -196.4380645751953,
77
+ "debug/policy_rejected_logits": -2.9797654151916504,
78
+ "debug/policy_rejected_logps": -223.4413299560547,
79
+ "debug/reference_chosen_logps": -196.45535278320312,
80
+ "debug/reference_rejected_logps": -223.3170166015625,
81
+ "epoch": 0.10526315789473684,
82
+ "grad_norm": 6.408309822092249,
83
+ "learning_rate": 1e-06,
84
+ "logits/chosen": -2.936859607696533,
85
+ "logits/rejected": -2.9797654151916504,
86
+ "logps/chosen": -196.4380645751953,
87
+ "logps/rejected": -223.4413299560547,
88
+ "loss": 0.4949,
89
+ "rewards/accuracies": 0.375,
90
+ "rewards/chosen": 0.00017278676386922598,
91
+ "rewards/margins": 0.001416072715073824,
92
+ "rewards/rejected": -0.0012432858347892761,
93
+ "step": 4
94
+ },
95
+ {
96
+ "debug/policy_chosen_logits": -3.322523593902588,
97
+ "debug/policy_chosen_logps": -167.5500946044922,
98
+ "debug/policy_rejected_logits": -3.356478214263916,
99
+ "debug/policy_rejected_logps": -200.8585205078125,
100
+ "debug/reference_chosen_logps": -167.9774169921875,
101
+ "debug/reference_rejected_logps": -200.06298828125,
102
+ "epoch": 0.13157894736842105,
103
+ "grad_norm": 7.788497859306542,
104
+ "learning_rate": 1e-06,
105
+ "logits/chosen": -3.322523593902588,
106
+ "logits/rejected": -3.356478214263916,
107
+ "logps/chosen": -167.5500946044922,
108
+ "logps/rejected": -200.8585205078125,
109
+ "loss": 0.4933,
110
+ "rewards/accuracies": 0.75,
111
+ "rewards/chosen": 0.00427322369068861,
112
+ "rewards/margins": 0.012228774838149548,
113
+ "rewards/rejected": -0.007955551147460938,
114
+ "step": 5
115
+ },
116
+ {
117
+ "debug/policy_chosen_logits": -2.9735865592956543,
118
+ "debug/policy_chosen_logps": -180.93771362304688,
119
+ "debug/policy_rejected_logits": -3.0403549671173096,
120
+ "debug/policy_rejected_logps": -221.18209838867188,
121
+ "debug/reference_chosen_logps": -181.0188751220703,
122
+ "debug/reference_rejected_logps": -219.86424255371094,
123
+ "epoch": 0.15789473684210525,
124
+ "grad_norm": 7.672338278520194,
125
+ "learning_rate": 1e-06,
126
+ "logits/chosen": -2.9735865592956543,
127
+ "logits/rejected": -3.0403549671173096,
128
+ "logps/chosen": -180.93771362304688,
129
+ "logps/rejected": -221.18209838867188,
130
+ "loss": 0.491,
131
+ "rewards/accuracies": 0.875,
132
+ "rewards/chosen": 0.0008116818498820066,
133
+ "rewards/margins": 0.013990296050906181,
134
+ "rewards/rejected": -0.013178614899516106,
135
+ "step": 6
136
+ },
137
+ {
138
+ "debug/policy_chosen_logits": -3.1537790298461914,
139
+ "debug/policy_chosen_logps": -170.55799865722656,
140
+ "debug/policy_rejected_logits": -3.094151020050049,
141
+ "debug/policy_rejected_logps": -186.5414276123047,
142
+ "debug/reference_chosen_logps": -170.28248596191406,
143
+ "debug/reference_rejected_logps": -185.3914794921875,
144
+ "epoch": 0.18421052631578946,
145
+ "grad_norm": 9.985482485638771,
146
+ "learning_rate": 1e-06,
147
+ "logits/chosen": -3.1537790298461914,
148
+ "logits/rejected": -3.094151020050049,
149
+ "logps/chosen": -170.55799865722656,
150
+ "logps/rejected": -186.5414276123047,
151
+ "loss": 0.4911,
152
+ "rewards/accuracies": 0.625,
153
+ "rewards/chosen": -0.0027551264502108097,
154
+ "rewards/margins": 0.008744277991354465,
155
+ "rewards/rejected": -0.011499403975903988,
156
+ "step": 7
157
+ },
158
+ {
159
+ "debug/policy_chosen_logits": -3.1766419410705566,
160
+ "debug/policy_chosen_logps": -162.15562438964844,
161
+ "debug/policy_rejected_logits": -3.1496686935424805,
162
+ "debug/policy_rejected_logps": -205.495361328125,
163
+ "debug/reference_chosen_logps": -162.75245666503906,
164
+ "debug/reference_rejected_logps": -204.7869873046875,
165
+ "epoch": 0.21052631578947367,
166
+ "grad_norm": 9.955698852386796,
167
+ "learning_rate": 1e-06,
168
+ "logits/chosen": -3.1766419410705566,
169
+ "logits/rejected": -3.1496686935424805,
170
+ "logps/chosen": -162.15562438964844,
171
+ "logps/rejected": -205.495361328125,
172
+ "loss": 0.4907,
173
+ "rewards/accuracies": 0.625,
174
+ "rewards/chosen": 0.005968417041003704,
175
+ "rewards/margins": 0.013052292168140411,
176
+ "rewards/rejected": -0.007083873264491558,
177
+ "step": 8
178
+ },
179
+ {
180
+ "debug/policy_chosen_logits": -3.072287082672119,
181
+ "debug/policy_chosen_logps": -169.07131958007812,
182
+ "debug/policy_rejected_logits": -3.13765287399292,
183
+ "debug/policy_rejected_logps": -239.11859130859375,
184
+ "debug/reference_chosen_logps": -170.79885864257812,
185
+ "debug/reference_rejected_logps": -237.47584533691406,
186
+ "epoch": 0.23684210526315788,
187
+ "grad_norm": 7.624438441931389,
188
+ "learning_rate": 1e-06,
189
+ "logits/chosen": -3.072287082672119,
190
+ "logits/rejected": -3.13765287399292,
191
+ "logps/chosen": -169.07131958007812,
192
+ "logps/rejected": -239.11859130859375,
193
+ "loss": 0.4938,
194
+ "rewards/accuracies": 0.75,
195
+ "rewards/chosen": 0.017275407910346985,
196
+ "rewards/margins": 0.033702850341796875,
197
+ "rewards/rejected": -0.01642744056880474,
198
+ "step": 9
199
+ },
200
+ {
201
+ "debug/policy_chosen_logits": -3.0883498191833496,
202
+ "debug/policy_chosen_logps": -162.46302795410156,
203
+ "debug/policy_rejected_logits": -3.048661231994629,
204
+ "debug/policy_rejected_logps": -195.8798828125,
205
+ "debug/reference_chosen_logps": -166.1143798828125,
206
+ "debug/reference_rejected_logps": -195.7821807861328,
207
+ "epoch": 0.2631578947368421,
208
+ "grad_norm": 6.502437447513198,
209
+ "learning_rate": 1e-06,
210
+ "logits/chosen": -3.0883498191833496,
211
+ "logits/rejected": -3.048661231994629,
212
+ "logps/chosen": -162.46302795410156,
213
+ "logps/rejected": -195.8798828125,
214
+ "loss": 0.474,
215
+ "rewards/accuracies": 0.75,
216
+ "rewards/chosen": 0.03651344031095505,
217
+ "rewards/margins": 0.037490349262952805,
218
+ "rewards/rejected": -0.0009769070893526077,
219
+ "step": 10
220
+ },
221
+ {
222
+ "debug/policy_chosen_logits": -3.0253381729125977,
223
+ "debug/policy_chosen_logps": -188.37547302246094,
224
+ "debug/policy_rejected_logits": -2.9573111534118652,
225
+ "debug/policy_rejected_logps": -183.5548553466797,
226
+ "debug/reference_chosen_logps": -190.44561767578125,
227
+ "debug/reference_rejected_logps": -185.97744750976562,
228
+ "epoch": 0.2894736842105263,
229
+ "grad_norm": 6.481854486966455,
230
+ "learning_rate": 1e-06,
231
+ "logits/chosen": -3.0253381729125977,
232
+ "logits/rejected": -2.9573111534118652,
233
+ "logps/chosen": -188.37547302246094,
234
+ "logps/rejected": -183.5548553466797,
235
+ "loss": 0.491,
236
+ "rewards/accuracies": 0.5,
237
+ "rewards/chosen": 0.02070142701268196,
238
+ "rewards/margins": -0.0035245325416326523,
239
+ "rewards/rejected": 0.024225957691669464,
240
+ "step": 11
241
+ },
242
+ {
243
+ "debug/policy_chosen_logits": -3.051511287689209,
244
+ "debug/policy_chosen_logps": -179.14675903320312,
245
+ "debug/policy_rejected_logits": -3.232536792755127,
246
+ "debug/policy_rejected_logps": -183.38912963867188,
247
+ "debug/reference_chosen_logps": -184.552490234375,
248
+ "debug/reference_rejected_logps": -185.21734619140625,
249
+ "epoch": 0.3157894736842105,
250
+ "grad_norm": 7.869767154728272,
251
+ "learning_rate": 1e-06,
252
+ "logits/chosen": -3.051511287689209,
253
+ "logits/rejected": -3.232536792755127,
254
+ "logps/chosen": -179.14675903320312,
255
+ "logps/rejected": -183.38912963867188,
256
+ "loss": 0.4665,
257
+ "rewards/accuracies": 0.875,
258
+ "rewards/chosen": 0.0540575236082077,
259
+ "rewards/margins": 0.035775475203990936,
260
+ "rewards/rejected": 0.018282050266861916,
261
+ "step": 12
262
+ },
263
+ {
264
+ "debug/policy_chosen_logits": -3.1170427799224854,
265
+ "debug/policy_chosen_logps": -175.19915771484375,
266
+ "debug/policy_rejected_logits": -3.182417869567871,
267
+ "debug/policy_rejected_logps": -176.825927734375,
268
+ "debug/reference_chosen_logps": -177.46597290039062,
269
+ "debug/reference_rejected_logps": -179.0623779296875,
270
+ "epoch": 0.34210526315789475,
271
+ "grad_norm": 12.633470392723837,
272
+ "learning_rate": 1e-06,
273
+ "logits/chosen": -3.1170427799224854,
274
+ "logits/rejected": -3.182417869567871,
275
+ "logps/chosen": -175.19915771484375,
276
+ "logps/rejected": -176.825927734375,
277
+ "loss": 0.4774,
278
+ "rewards/accuracies": 0.5,
279
+ "rewards/chosen": 0.022668171674013138,
280
+ "rewards/margins": 0.0003036591224372387,
281
+ "rewards/rejected": 0.022364510223269463,
282
+ "step": 13
283
+ },
284
+ {
285
+ "debug/policy_chosen_logits": -3.0715718269348145,
286
+ "debug/policy_chosen_logps": -196.95872497558594,
287
+ "debug/policy_rejected_logits": -3.0461950302124023,
288
+ "debug/policy_rejected_logps": -234.42738342285156,
289
+ "debug/reference_chosen_logps": -198.927001953125,
290
+ "debug/reference_rejected_logps": -231.70632934570312,
291
+ "epoch": 0.3684210526315789,
292
+ "grad_norm": 7.86574979608913,
293
+ "learning_rate": 1e-06,
294
+ "logits/chosen": -3.0715718269348145,
295
+ "logits/rejected": -3.0461950302124023,
296
+ "logps/chosen": -196.95872497558594,
297
+ "logps/rejected": -234.42738342285156,
298
+ "loss": 0.4723,
299
+ "rewards/accuracies": 0.625,
300
+ "rewards/chosen": 0.019682805985212326,
301
+ "rewards/margins": 0.046893440186977386,
302
+ "rewards/rejected": -0.02721063420176506,
303
+ "step": 14
304
+ },
305
+ {
306
+ "debug/policy_chosen_logits": -3.148141622543335,
307
+ "debug/policy_chosen_logps": -174.8628692626953,
308
+ "debug/policy_rejected_logits": -3.151831865310669,
309
+ "debug/policy_rejected_logps": -239.0707244873047,
310
+ "debug/reference_chosen_logps": -179.73025512695312,
311
+ "debug/reference_rejected_logps": -234.18270874023438,
312
+ "epoch": 0.39473684210526316,
313
+ "grad_norm": 7.551670325207112,
314
+ "learning_rate": 1e-06,
315
+ "logits/chosen": -3.148141622543335,
316
+ "logits/rejected": -3.151831865310669,
317
+ "logps/chosen": -174.8628692626953,
318
+ "logps/rejected": -239.0707244873047,
319
+ "loss": 0.4394,
320
+ "rewards/accuracies": 1.0,
321
+ "rewards/chosen": 0.0486738383769989,
322
+ "rewards/margins": 0.09755405783653259,
323
+ "rewards/rejected": -0.048880212008953094,
324
+ "step": 15
325
+ },
326
+ {
327
+ "debug/policy_chosen_logits": -3.2363226413726807,
328
+ "debug/policy_chosen_logps": -180.26939392089844,
329
+ "debug/policy_rejected_logits": -3.2402045726776123,
330
+ "debug/policy_rejected_logps": -191.19422912597656,
331
+ "debug/reference_chosen_logps": -182.0409698486328,
332
+ "debug/reference_rejected_logps": -188.7088623046875,
333
+ "epoch": 0.42105263157894735,
334
+ "grad_norm": 7.971499111246793,
335
+ "learning_rate": 1e-06,
336
+ "logits/chosen": -3.2363226413726807,
337
+ "logits/rejected": -3.2402045726776123,
338
+ "logps/chosen": -180.26939392089844,
339
+ "logps/rejected": -191.19422912597656,
340
+ "loss": 0.4732,
341
+ "rewards/accuracies": 0.75,
342
+ "rewards/chosen": 0.01771572232246399,
343
+ "rewards/margins": 0.042569201439619064,
344
+ "rewards/rejected": -0.024853479117155075,
345
+ "step": 16
346
+ },
347
+ {
348
+ "debug/policy_chosen_logits": -3.1031575202941895,
349
+ "debug/policy_chosen_logps": -181.3292999267578,
350
+ "debug/policy_rejected_logits": -3.198836326599121,
351
+ "debug/policy_rejected_logps": -221.01437377929688,
352
+ "debug/reference_chosen_logps": -179.9342041015625,
353
+ "debug/reference_rejected_logps": -214.72239685058594,
354
+ "epoch": 0.4473684210526316,
355
+ "grad_norm": 10.584039363754105,
356
+ "learning_rate": 1e-06,
357
+ "logits/chosen": -3.1031575202941895,
358
+ "logits/rejected": -3.198836326599121,
359
+ "logps/chosen": -181.3292999267578,
360
+ "logps/rejected": -221.01437377929688,
361
+ "loss": 0.4709,
362
+ "rewards/accuracies": 0.875,
363
+ "rewards/chosen": -0.013950958847999573,
364
+ "rewards/margins": 0.04896876960992813,
365
+ "rewards/rejected": -0.0629197359085083,
366
+ "step": 17
367
+ },
368
+ {
369
+ "debug/policy_chosen_logits": -3.0805013179779053,
370
+ "debug/policy_chosen_logps": -217.13873291015625,
371
+ "debug/policy_rejected_logits": -3.088634967803955,
372
+ "debug/policy_rejected_logps": -210.9022674560547,
373
+ "debug/reference_chosen_logps": -213.02728271484375,
374
+ "debug/reference_rejected_logps": -202.1422576904297,
375
+ "epoch": 0.47368421052631576,
376
+ "grad_norm": 15.605225931338353,
377
+ "learning_rate": 1e-06,
378
+ "logits/chosen": -3.0805013179779053,
379
+ "logits/rejected": -3.088634967803955,
380
+ "logps/chosen": -217.13873291015625,
381
+ "logps/rejected": -210.9022674560547,
382
+ "loss": 0.4666,
383
+ "rewards/accuracies": 0.5,
384
+ "rewards/chosen": -0.041114576160907745,
385
+ "rewards/margins": 0.046485595405101776,
386
+ "rewards/rejected": -0.08760017156600952,
387
+ "step": 18
388
+ },
389
+ {
390
+ "debug/policy_chosen_logits": -3.227140426635742,
391
+ "debug/policy_chosen_logps": -176.44325256347656,
392
+ "debug/policy_rejected_logits": -3.1130940914154053,
393
+ "debug/policy_rejected_logps": -197.5587921142578,
394
+ "debug/reference_chosen_logps": -171.158203125,
395
+ "debug/reference_rejected_logps": -189.12246704101562,
396
+ "epoch": 0.5,
397
+ "grad_norm": 20.388524998578973,
398
+ "learning_rate": 1e-06,
399
+ "logits/chosen": -3.227140426635742,
400
+ "logits/rejected": -3.1130940914154053,
401
+ "logps/chosen": -176.44325256347656,
402
+ "logps/rejected": -197.5587921142578,
403
+ "loss": 0.4711,
404
+ "rewards/accuracies": 0.75,
405
+ "rewards/chosen": -0.052850380539894104,
406
+ "rewards/margins": 0.031512755900621414,
407
+ "rewards/rejected": -0.08436313271522522,
408
+ "step": 19
409
+ },
410
+ {
411
+ "debug/policy_chosen_logits": -3.2375006675720215,
412
+ "debug/policy_chosen_logps": -178.27442932128906,
413
+ "debug/policy_rejected_logits": -3.2282068729400635,
414
+ "debug/policy_rejected_logps": -236.70098876953125,
415
+ "debug/reference_chosen_logps": -175.0205078125,
416
+ "debug/reference_rejected_logps": -222.80160522460938,
417
+ "epoch": 0.5263157894736842,
418
+ "grad_norm": 19.93135079488986,
419
+ "learning_rate": 1e-06,
420
+ "logits/chosen": -3.2375006675720215,
421
+ "logits/rejected": -3.2282068729400635,
422
+ "logps/chosen": -178.27442932128906,
423
+ "logps/rejected": -236.70098876953125,
424
+ "loss": 0.4549,
425
+ "rewards/accuracies": 0.75,
426
+ "rewards/chosen": -0.03253912180662155,
427
+ "rewards/margins": 0.10645455867052078,
428
+ "rewards/rejected": -0.13899368047714233,
429
+ "step": 20
430
+ },
431
+ {
432
+ "debug/policy_chosen_logits": -3.1386053562164307,
433
+ "debug/policy_chosen_logps": -197.573486328125,
434
+ "debug/policy_rejected_logits": -3.2233846187591553,
435
+ "debug/policy_rejected_logps": -214.78076171875,
436
+ "debug/reference_chosen_logps": -199.56964111328125,
437
+ "debug/reference_rejected_logps": -205.45248413085938,
438
+ "epoch": 0.5526315789473685,
439
+ "grad_norm": 7.07361132167883,
440
+ "learning_rate": 1e-06,
441
+ "logits/chosen": -3.1386053562164307,
442
+ "logits/rejected": -3.2233846187591553,
443
+ "logps/chosen": -197.573486328125,
444
+ "logps/rejected": -214.78076171875,
445
+ "loss": 0.4536,
446
+ "rewards/accuracies": 1.0,
447
+ "rewards/chosen": 0.019961394369602203,
448
+ "rewards/margins": 0.11324435472488403,
449
+ "rewards/rejected": -0.09328296780586243,
450
+ "step": 21
451
+ },
452
+ {
453
+ "debug/policy_chosen_logits": -3.124323844909668,
454
+ "debug/policy_chosen_logps": -199.28509521484375,
455
+ "debug/policy_rejected_logits": -3.118931770324707,
456
+ "debug/policy_rejected_logps": -212.33740234375,
457
+ "debug/reference_chosen_logps": -196.1334686279297,
458
+ "debug/reference_rejected_logps": -212.9163360595703,
459
+ "epoch": 0.5789473684210527,
460
+ "grad_norm": 12.998049377939719,
461
+ "learning_rate": 1e-06,
462
+ "logits/chosen": -3.124323844909668,
463
+ "logits/rejected": -3.118931770324707,
464
+ "logps/chosen": -199.28509521484375,
465
+ "logps/rejected": -212.33740234375,
466
+ "loss": 0.4404,
467
+ "rewards/accuracies": 0.5,
468
+ "rewards/chosen": -0.03151614964008331,
469
+ "rewards/margins": -0.037305522710084915,
470
+ "rewards/rejected": 0.005789373070001602,
471
+ "step": 22
472
+ },
473
+ {
474
+ "debug/policy_chosen_logits": -3.2582762241363525,
475
+ "debug/policy_chosen_logps": -173.89572143554688,
476
+ "debug/policy_rejected_logits": -3.2754862308502197,
477
+ "debug/policy_rejected_logps": -203.75881958007812,
478
+ "debug/reference_chosen_logps": -178.55136108398438,
479
+ "debug/reference_rejected_logps": -199.2237548828125,
480
+ "epoch": 0.6052631578947368,
481
+ "grad_norm": 16.668545244677244,
482
+ "learning_rate": 1e-06,
483
+ "logits/chosen": -3.2582762241363525,
484
+ "logits/rejected": -3.2754862308502197,
485
+ "logps/chosen": -173.89572143554688,
486
+ "logps/rejected": -203.75881958007812,
487
+ "loss": 0.4502,
488
+ "rewards/accuracies": 0.75,
489
+ "rewards/chosen": 0.04655652865767479,
490
+ "rewards/margins": 0.09190709888935089,
491
+ "rewards/rejected": -0.0453505702316761,
492
+ "step": 23
493
+ },
494
+ {
495
+ "debug/policy_chosen_logits": -3.2273776531219482,
496
+ "debug/policy_chosen_logps": -150.5181121826172,
497
+ "debug/policy_rejected_logits": -3.2226057052612305,
498
+ "debug/policy_rejected_logps": -208.79214477539062,
499
+ "debug/reference_chosen_logps": -158.41482543945312,
500
+ "debug/reference_rejected_logps": -212.96640014648438,
501
+ "epoch": 0.631578947368421,
502
+ "grad_norm": 13.427151271997946,
503
+ "learning_rate": 1e-06,
504
+ "logits/chosen": -3.2273776531219482,
505
+ "logits/rejected": -3.2226057052612305,
506
+ "logps/chosen": -150.5181121826172,
507
+ "logps/rejected": -208.79214477539062,
508
+ "loss": 0.4678,
509
+ "rewards/accuracies": 0.75,
510
+ "rewards/chosen": 0.07896722853183746,
511
+ "rewards/margins": 0.0372246578335762,
512
+ "rewards/rejected": 0.04174257069826126,
513
+ "step": 24
514
+ },
515
+ {
516
+ "debug/policy_chosen_logits": -3.165872573852539,
517
+ "debug/policy_chosen_logps": -172.341064453125,
518
+ "debug/policy_rejected_logits": -3.2021644115448,
519
+ "debug/policy_rejected_logps": -201.33978271484375,
520
+ "debug/reference_chosen_logps": -186.38406372070312,
521
+ "debug/reference_rejected_logps": -198.79547119140625,
522
+ "epoch": 0.6578947368421053,
523
+ "grad_norm": 18.736524383787305,
524
+ "learning_rate": 1e-06,
525
+ "logits/chosen": -3.165872573852539,
526
+ "logits/rejected": -3.2021644115448,
527
+ "logps/chosen": -172.341064453125,
528
+ "logps/rejected": -201.33978271484375,
529
+ "loss": 0.4624,
530
+ "rewards/accuracies": 0.75,
531
+ "rewards/chosen": 0.14042991399765015,
532
+ "rewards/margins": 0.1658729612827301,
533
+ "rewards/rejected": -0.02544303983449936,
534
+ "step": 25
535
+ },
536
+ {
537
+ "debug/policy_chosen_logits": -3.1740562915802,
538
+ "debug/policy_chosen_logps": -177.84255981445312,
539
+ "debug/policy_rejected_logits": -3.151543140411377,
540
+ "debug/policy_rejected_logps": -191.53330993652344,
541
+ "debug/reference_chosen_logps": -182.20455932617188,
542
+ "debug/reference_rejected_logps": -196.35205078125,
543
+ "epoch": 0.6842105263157895,
544
+ "grad_norm": 7.938159632089657,
545
+ "learning_rate": 1e-06,
546
+ "logits/chosen": -3.1740562915802,
547
+ "logits/rejected": -3.151543140411377,
548
+ "logps/chosen": -177.84255981445312,
549
+ "logps/rejected": -191.53330993652344,
550
+ "loss": 0.4525,
551
+ "rewards/accuracies": 0.5,
552
+ "rewards/chosen": 0.043619830161333084,
553
+ "rewards/margins": -0.004567727446556091,
554
+ "rewards/rejected": 0.048187561333179474,
555
+ "step": 26
556
+ },
557
+ {
558
+ "debug/policy_chosen_logits": -3.146756649017334,
559
+ "debug/policy_chosen_logps": -178.16688537597656,
560
+ "debug/policy_rejected_logits": -3.192675828933716,
561
+ "debug/policy_rejected_logps": -213.1707763671875,
562
+ "debug/reference_chosen_logps": -184.39260864257812,
563
+ "debug/reference_rejected_logps": -201.81924438476562,
564
+ "epoch": 0.7105263157894737,
565
+ "grad_norm": 10.884065391328518,
566
+ "learning_rate": 1e-06,
567
+ "logits/chosen": -3.146756649017334,
568
+ "logits/rejected": -3.192675828933716,
569
+ "logps/chosen": -178.16688537597656,
570
+ "logps/rejected": -213.1707763671875,
571
+ "loss": 0.4459,
572
+ "rewards/accuracies": 0.875,
573
+ "rewards/chosen": 0.06225726753473282,
574
+ "rewards/margins": 0.17577265202999115,
575
+ "rewards/rejected": -0.11351536959409714,
576
+ "step": 27
577
+ },
578
+ {
579
+ "debug/policy_chosen_logits": -3.281649589538574,
580
+ "debug/policy_chosen_logps": -199.9445343017578,
581
+ "debug/policy_rejected_logits": -3.2994844913482666,
582
+ "debug/policy_rejected_logps": -219.6336669921875,
583
+ "debug/reference_chosen_logps": -196.85604858398438,
584
+ "debug/reference_rejected_logps": -210.05245971679688,
585
+ "epoch": 0.7368421052631579,
586
+ "grad_norm": 6.010691347934726,
587
+ "learning_rate": 1e-06,
588
+ "logits/chosen": -3.281649589538574,
589
+ "logits/rejected": -3.2994844913482666,
590
+ "logps/chosen": -199.9445343017578,
591
+ "logps/rejected": -219.6336669921875,
592
+ "loss": 0.4625,
593
+ "rewards/accuracies": 0.25,
594
+ "rewards/chosen": -0.03088477998971939,
595
+ "rewards/margins": 0.06492738425731659,
596
+ "rewards/rejected": -0.09581217169761658,
597
+ "step": 28
598
+ },
599
+ {
600
+ "debug/policy_chosen_logits": -3.335418939590454,
601
+ "debug/policy_chosen_logps": -163.25552368164062,
602
+ "debug/policy_rejected_logits": -3.3753366470336914,
603
+ "debug/policy_rejected_logps": -188.67388916015625,
604
+ "debug/reference_chosen_logps": -169.972412109375,
605
+ "debug/reference_rejected_logps": -187.3748321533203,
606
+ "epoch": 0.7631578947368421,
607
+ "grad_norm": 9.368732561204727,
608
+ "learning_rate": 1e-06,
609
+ "logits/chosen": -3.335418939590454,
610
+ "logits/rejected": -3.3753366470336914,
611
+ "logps/chosen": -163.25552368164062,
612
+ "logps/rejected": -188.67388916015625,
613
+ "loss": 0.4284,
614
+ "rewards/accuracies": 0.625,
615
+ "rewards/chosen": 0.06716887652873993,
616
+ "rewards/margins": 0.08015935868024826,
617
+ "rewards/rejected": -0.012990472838282585,
618
+ "step": 29
619
+ },
620
+ {
621
+ "debug/policy_chosen_logits": -3.1683101654052734,
622
+ "debug/policy_chosen_logps": -170.42584228515625,
623
+ "debug/policy_rejected_logits": -3.232621908187866,
624
+ "debug/policy_rejected_logps": -190.65835571289062,
625
+ "debug/reference_chosen_logps": -174.41644287109375,
626
+ "debug/reference_rejected_logps": -181.63363647460938,
627
+ "epoch": 0.7894736842105263,
628
+ "grad_norm": 8.036358997916743,
629
+ "learning_rate": 1e-06,
630
+ "logits/chosen": -3.1683101654052734,
631
+ "logits/rejected": -3.232621908187866,
632
+ "logps/chosen": -170.42584228515625,
633
+ "logps/rejected": -190.65835571289062,
634
+ "loss": 0.3995,
635
+ "rewards/accuracies": 0.75,
636
+ "rewards/chosen": 0.03990606591105461,
637
+ "rewards/margins": 0.13015341758728027,
638
+ "rewards/rejected": -0.09024736285209656,
639
+ "step": 30
640
+ },
641
+ {
642
+ "debug/policy_chosen_logits": -3.2613167762756348,
643
+ "debug/policy_chosen_logps": -181.21347045898438,
644
+ "debug/policy_rejected_logits": -3.233255386352539,
645
+ "debug/policy_rejected_logps": -194.60850524902344,
646
+ "debug/reference_chosen_logps": -173.81005859375,
647
+ "debug/reference_rejected_logps": -193.19393920898438,
648
+ "epoch": 0.8157894736842105,
649
+ "grad_norm": 21.754243234566374,
650
+ "learning_rate": 1e-06,
651
+ "logits/chosen": -3.2613167762756348,
652
+ "logits/rejected": -3.233255386352539,
653
+ "logps/chosen": -181.21347045898438,
654
+ "logps/rejected": -194.60850524902344,
655
+ "loss": 0.4758,
656
+ "rewards/accuracies": 0.375,
657
+ "rewards/chosen": -0.07403402030467987,
658
+ "rewards/margins": -0.059888288378715515,
659
+ "rewards/rejected": -0.014145737513899803,
660
+ "step": 31
661
+ },
662
+ {
663
+ "debug/policy_chosen_logits": -3.1505517959594727,
664
+ "debug/policy_chosen_logps": -178.563720703125,
665
+ "debug/policy_rejected_logits": -3.3182621002197266,
666
+ "debug/policy_rejected_logps": -229.11212158203125,
667
+ "debug/reference_chosen_logps": -185.50924682617188,
668
+ "debug/reference_rejected_logps": -218.29916381835938,
669
+ "epoch": 0.8421052631578947,
670
+ "grad_norm": 34.79466400268049,
671
+ "learning_rate": 1e-06,
672
+ "logits/chosen": -3.1505517959594727,
673
+ "logits/rejected": -3.3182621002197266,
674
+ "logps/chosen": -178.563720703125,
675
+ "logps/rejected": -229.11212158203125,
676
+ "loss": 0.4946,
677
+ "rewards/accuracies": 0.75,
678
+ "rewards/chosen": 0.06945528090000153,
679
+ "rewards/margins": 0.17758482694625854,
680
+ "rewards/rejected": -0.10812953114509583,
681
+ "step": 32
682
+ },
683
+ {
684
+ "debug/policy_chosen_logits": -3.2886526584625244,
685
+ "debug/policy_chosen_logps": -183.58511352539062,
686
+ "debug/policy_rejected_logits": -3.3318674564361572,
687
+ "debug/policy_rejected_logps": -178.8451690673828,
688
+ "debug/reference_chosen_logps": -192.07305908203125,
689
+ "debug/reference_rejected_logps": -176.05474853515625,
690
+ "epoch": 0.868421052631579,
691
+ "grad_norm": 13.98393503620536,
692
+ "learning_rate": 1e-06,
693
+ "logits/chosen": -3.2886526584625244,
694
+ "logits/rejected": -3.3318674564361572,
695
+ "logps/chosen": -183.58511352539062,
696
+ "logps/rejected": -178.8451690673828,
697
+ "loss": 0.4312,
698
+ "rewards/accuracies": 0.875,
699
+ "rewards/chosen": 0.08487950265407562,
700
+ "rewards/margins": 0.11278380453586578,
701
+ "rewards/rejected": -0.027904313057661057,
702
+ "step": 33
703
+ },
704
+ {
705
+ "debug/policy_chosen_logits": -3.1525185108184814,
706
+ "debug/policy_chosen_logps": -180.24249267578125,
707
+ "debug/policy_rejected_logits": -3.143476963043213,
708
+ "debug/policy_rejected_logps": -204.80078125,
709
+ "debug/reference_chosen_logps": -180.59463500976562,
710
+ "debug/reference_rejected_logps": -201.7593536376953,
711
+ "epoch": 0.8947368421052632,
712
+ "grad_norm": 28.559943285925073,
713
+ "learning_rate": 1e-06,
714
+ "logits/chosen": -3.1525185108184814,
715
+ "logits/rejected": -3.143476963043213,
716
+ "logps/chosen": -180.24249267578125,
717
+ "logps/rejected": -204.80078125,
718
+ "loss": 0.4658,
719
+ "rewards/accuracies": 0.625,
720
+ "rewards/chosen": 0.0035214386880397797,
721
+ "rewards/margins": 0.03393552824854851,
722
+ "rewards/rejected": -0.03041408583521843,
723
+ "step": 34
724
+ },
725
+ {
726
+ "debug/policy_chosen_logits": -3.2434704303741455,
727
+ "debug/policy_chosen_logps": -153.05389404296875,
728
+ "debug/policy_rejected_logits": -3.1566455364227295,
729
+ "debug/policy_rejected_logps": -205.80572509765625,
730
+ "debug/reference_chosen_logps": -168.76995849609375,
731
+ "debug/reference_rejected_logps": -203.63229370117188,
732
+ "epoch": 0.9210526315789473,
733
+ "grad_norm": 12.367548900105426,
734
+ "learning_rate": 1e-06,
735
+ "logits/chosen": -3.2434704303741455,
736
+ "logits/rejected": -3.1566455364227295,
737
+ "logps/chosen": -153.05389404296875,
738
+ "logps/rejected": -205.80572509765625,
739
+ "loss": 0.3944,
740
+ "rewards/accuracies": 0.75,
741
+ "rewards/chosen": 0.15716060996055603,
742
+ "rewards/margins": 0.17889487743377686,
743
+ "rewards/rejected": -0.02173428051173687,
744
+ "step": 35
745
+ },
746
+ {
747
+ "debug/policy_chosen_logits": -3.1490585803985596,
748
+ "debug/policy_chosen_logps": -184.06216430664062,
749
+ "debug/policy_rejected_logits": -3.078645706176758,
750
+ "debug/policy_rejected_logps": -192.70101928710938,
751
+ "debug/reference_chosen_logps": -187.65826416015625,
752
+ "debug/reference_rejected_logps": -196.99972534179688,
753
+ "epoch": 0.9473684210526315,
754
+ "grad_norm": 6.171234961762308,
755
+ "learning_rate": 1e-06,
756
+ "logits/chosen": -3.1490585803985596,
757
+ "logits/rejected": -3.078645706176758,
758
+ "logps/chosen": -184.06216430664062,
759
+ "logps/rejected": -192.70101928710938,
760
+ "loss": 0.4299,
761
+ "rewards/accuracies": 0.375,
762
+ "rewards/chosen": 0.0359608456492424,
763
+ "rewards/margins": -0.0070261843502521515,
764
+ "rewards/rejected": 0.04298701882362366,
765
+ "step": 36
766
+ },
767
+ {
768
+ "debug/policy_chosen_logits": -3.260350227355957,
769
+ "debug/policy_chosen_logps": -186.05783081054688,
770
+ "debug/policy_rejected_logits": -3.428133964538574,
771
+ "debug/policy_rejected_logps": -219.6826934814453,
772
+ "debug/reference_chosen_logps": -198.08245849609375,
773
+ "debug/reference_rejected_logps": -213.1524658203125,
774
+ "epoch": 0.9736842105263158,
775
+ "grad_norm": 7.8362879657016995,
776
+ "learning_rate": 1e-06,
777
+ "logits/chosen": -3.260350227355957,
778
+ "logits/rejected": -3.428133964538574,
779
+ "logps/chosen": -186.05783081054688,
780
+ "logps/rejected": -219.6826934814453,
781
+ "loss": 0.4041,
782
+ "rewards/accuracies": 1.0,
783
+ "rewards/chosen": 0.12024623155593872,
784
+ "rewards/margins": 0.18554842472076416,
785
+ "rewards/rejected": -0.06530220061540604,
786
+ "step": 37
787
+ },
788
+ {
789
+ "debug/policy_chosen_logits": -3.2531356811523438,
790
+ "debug/policy_chosen_logps": -186.32887268066406,
791
+ "debug/policy_rejected_logits": -3.2301692962646484,
792
+ "debug/policy_rejected_logps": -201.42446899414062,
793
+ "debug/reference_chosen_logps": -190.052490234375,
794
+ "debug/reference_rejected_logps": -197.30398559570312,
795
+ "epoch": 1.0,
796
+ "grad_norm": 10.428660286754967,
797
+ "learning_rate": 1e-06,
798
+ "logits/chosen": -3.2531356811523438,
799
+ "logits/rejected": -3.2301692962646484,
800
+ "logps/chosen": -186.32887268066406,
801
+ "logps/rejected": -201.42446899414062,
802
+ "loss": 0.3962,
803
+ "rewards/accuracies": 0.75,
804
+ "rewards/chosen": 0.03723614662885666,
805
+ "rewards/margins": 0.07844101637601852,
806
+ "rewards/rejected": -0.041204869747161865,
807
+ "step": 38
808
+ },
809
+ {
810
+ "epoch": 1.0,
811
+ "step": 38,
812
+ "total_flos": 0.0,
813
+ "train_loss": 0.4622946350198043,
814
+ "train_runtime": 144.3728,
815
+ "train_samples_per_second": 16.818,
816
+ "train_steps_per_second": 0.263
817
+ }
818
+ ],
819
+ "logging_steps": 1,
820
+ "max_steps": 38,
821
+ "num_input_tokens_seen": 0,
822
+ "num_train_epochs": 1,
823
+ "save_steps": 500,
824
+ "stateful_callbacks": {
825
+ "TrainerControl": {
826
+ "args": {
827
+ "should_epoch_stop": false,
828
+ "should_evaluate": false,
829
+ "should_log": false,
830
+ "should_save": true,
831
+ "should_training_stop": true
832
+ },
833
+ "attributes": {}
834
+ }
835
+ },
836
+ "total_flos": 0.0,
837
+ "train_batch_size": 8,
838
+ "trial_name": null,
839
+ "trial_params": null
840
+ }