Hugging Face's logo

EllieS
/
zephyr-dpo-timedial-selfgen

PEFT
TensorBoard
Safetensors
mistral
alignment-handbook
trl
dpo
Generated from Trainer
Model card Files Metrics Community
EllieS commited on
Commit
72085a6
•
1 Parent(s): 4fb4bf7

Model save

Browse files
Files changed (6) hide show
  1. README.md +74 -0
  2. adapter_model.safetensors +1 -1
  3. all_results.json +8 -0
  4. runs/Apr11_18-28-33_612e66badb5c/events.out.tfevents.1712860224.612e66badb5c.7512.0 +2 -2
  5. train_results.json +8 -0
  6. trainer_state.json +497 -0
README.md ADDED
@@ -0,0 +1,74 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ license: apache-2.0
3
+ library_name: peft
4
+ tags:
5
+ - trl
6
+ - dpo
7
+ - generated_from_trainer
8
+ base_model: alignment-handbook/zephyr-7b-sft-full
9
+ model-index:
10
+ - name: zephyr-dpo-timedial-selfgen
11
+ results: []
12
+ ---
13
+
14
+ <!-- This model card has been generated automatically according to the information the Trainer had access to. You
15
+ should probably proofread and complete it, then remove this comment. -->
16
+
17
+ # zephyr-dpo-timedial-selfgen
18
+
19
+ This model is a fine-tuned version of [alignment-handbook/zephyr-7b-sft-full](https://huggingface.co/alignment-handbook/zephyr-7b-sft-full) on the None dataset.
20
+ It achieves the following results on the evaluation set:
21
+ - Loss: 0.0055
22
+ - Rewards/chosen: 0.4843
23
+ - Rewards/rejected: -5.0618
24
+ - Rewards/accuracies: 1.0
25
+ - Rewards/margins: 5.5461
26
+ - Logps/rejected: -740.9361
27
+ - Logps/chosen: -96.9992
28
+ - Logits/rejected: -3.1155
29
+ - Logits/chosen: -2.6815
30
+
31
+ ## Model description
32
+
33
+ More information needed
34
+
35
+ ## Intended uses & limitations
36
+
37
+ More information needed
38
+
39
+ ## Training and evaluation data
40
+
41
+ More information needed
42
+
43
+ ## Training procedure
44
+
45
+ ### Training hyperparameters
46
+
47
+ The following hyperparameters were used during training:
48
+ - learning_rate: 5e-06
49
+ - train_batch_size: 2
50
+ - eval_batch_size: 2
51
+ - seed: 42
52
+ - distributed_type: multi-GPU
53
+ - gradient_accumulation_steps: 2
54
+ - total_train_batch_size: 4
55
+ - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
56
+ - lr_scheduler_type: cosine
57
+ - lr_scheduler_warmup_ratio: 0.1
58
+ - num_epochs: 1
59
+
60
+ ### Training results
61
+
62
+ | Training Loss | Epoch | Step | Validation Loss | Rewards/chosen | Rewards/rejected | Rewards/accuracies | Rewards/margins | Logps/rejected | Logps/chosen | Logits/rejected | Logits/chosen |
63
+ |:-------------:|:-----:|:----:|:---------------:|:--------------:|:----------------:|:------------------:|:---------------:|:--------------:|:------------:|:---------------:|:-------------:|
64
+ | 0.0166 | 0.35 | 100 | 0.0198 | 0.4885 | -3.4443 | 1.0 | 3.9328 | -579.1801 | -96.5815 | -3.1058 | -2.6808 |
65
+ | 0.0015 | 0.7 | 200 | 0.0055 | 0.4843 | -5.0618 | 1.0 | 5.5461 | -740.9361 | -96.9992 | -3.1155 | -2.6815 |
66
+
67
+
68
+ ### Framework versions
69
+
70
+ - PEFT 0.7.1
71
+ - Transformers 4.39.0.dev0
72
+ - Pytorch 2.1.2+cu121
73
+ - Datasets 2.14.6
74
+ - Tokenizers 0.15.2
adapter_model.safetensors CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:019c85d55a294e0759b2c27f21a10c7029538b30a20766a24de23a77786761ad
3
  size 83946192
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:2e1fbe9123ee4ff95df395f6035ad916e3320acc06314ad717a138d3ab771489
3
  size 83946192
all_results.json ADDED
@@ -0,0 +1,8 @@
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 1.0,
3
+ "train_loss": 0.11324987118704277,
4
+ "train_runtime": 748.1588,
5
+ "train_samples": 1139,
6
+ "train_samples_per_second": 1.522,
7
+ "train_steps_per_second": 0.381
8
+ }
runs/Apr11_18-28-33_612e66badb5c/events.out.tfevents.1712860224.612e66badb5c.7512.0 CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:fbee69c4e9bd8c84975cb7012e04d30b7d22ce085319fe2af00ba89acd7a4f37
3
- size 20557
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:dd585e5e8af5f19b7ded3aee22762b3edb07d6a71219b79b3b171eef3a64b5d7
3
+ size 26415
train_results.json ADDED
@@ -0,0 +1,8 @@
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 1.0,
3
+ "train_loss": 0.11324987118704277,
4
+ "train_runtime": 748.1588,
5
+ "train_samples": 1139,
6
+ "train_samples_per_second": 1.522,
7
+ "train_steps_per_second": 0.381
8
+ }
trainer_state.json ADDED
@@ -0,0 +1,497 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": null,
3
+ "best_model_checkpoint": null,
4
+ "epoch": 1.0,
5
+ "eval_steps": 100,
6
+ "global_step": 285,
7
+ "is_hyper_param_search": false,
8
+ "is_local_process_zero": true,
9
+ "is_world_process_zero": true,
10
+ "log_history": [
11
+ {
12
+ "epoch": 0.0,
13
+ "grad_norm": 2.671875,
14
+ "learning_rate": 1.7241379310344828e-07,
15
+ "logits/chosen": -2.715578079223633,
16
+ "logits/rejected": -3.198270320892334,
17
+ "logps/chosen": -181.673828125,
18
+ "logps/rejected": -250.83145141601562,
19
+ "loss": 0.6931,
20
+ "rewards/accuracies": 0.0,
21
+ "rewards/chosen": 0.0,
22
+ "rewards/margins": 0.0,
23
+ "rewards/rejected": 0.0,
24
+ "step": 1
25
+ },
26
+ {
27
+ "epoch": 0.04,
28
+ "grad_norm": 2.75,
29
+ "learning_rate": 1.724137931034483e-06,
30
+ "logits/chosen": -2.6499929428100586,
31
+ "logits/rejected": -3.116462230682373,
32
+ "logps/chosen": -179.62359619140625,
33
+ "logps/rejected": -252.53973388671875,
34
+ "loss": 0.6925,
35
+ "rewards/accuracies": 0.6111111044883728,
36
+ "rewards/chosen": 0.0010814073029905558,
37
+ "rewards/margins": 0.0013816155260428786,
38
+ "rewards/rejected": -0.00030020822305232286,
39
+ "step": 10
40
+ },
41
+ {
42
+ "epoch": 0.07,
43
+ "grad_norm": 2.5625,
44
+ "learning_rate": 3.448275862068966e-06,
45
+ "logits/chosen": -2.6711201667785645,
46
+ "logits/rejected": -3.1086363792419434,
47
+ "logps/chosen": -176.72164916992188,
48
+ "logps/rejected": -252.67080688476562,
49
+ "loss": 0.6743,
50
+ "rewards/accuracies": 1.0,
51
+ "rewards/chosen": 0.023805765435099602,
52
+ "rewards/margins": 0.03818682208657265,
53
+ "rewards/rejected": -0.014381052926182747,
54
+ "step": 20
55
+ },
56
+ {
57
+ "epoch": 0.11,
58
+ "grad_norm": 1.78125,
59
+ "learning_rate": 4.999811754597862e-06,
60
+ "logits/chosen": -2.642437696456909,
61
+ "logits/rejected": -3.1196322441101074,
62
+ "logps/chosen": -167.26516723632812,
63
+ "logps/rejected": -260.6128845214844,
64
+ "loss": 0.6003,
65
+ "rewards/accuracies": 1.0,
66
+ "rewards/chosen": 0.1062479019165039,
67
+ "rewards/margins": 0.19668832421302795,
68
+ "rewards/rejected": -0.09044040739536285,
69
+ "step": 30
70
+ },
71
+ {
72
+ "epoch": 0.14,
73
+ "grad_norm": 1.4921875,
74
+ "learning_rate": 4.97725658856945e-06,
75
+ "logits/chosen": -2.6548495292663574,
76
+ "logits/rejected": -3.105985403060913,
77
+ "logps/chosen": -157.27615356445312,
78
+ "logps/rejected": -282.587646484375,
79
+ "loss": 0.4653,
80
+ "rewards/accuracies": 1.0,
81
+ "rewards/chosen": 0.22060635685920715,
82
+ "rewards/margins": 0.5277924537658691,
83
+ "rewards/rejected": -0.30718618631362915,
84
+ "step": 40
85
+ },
86
+ {
87
+ "epoch": 0.18,
88
+ "grad_norm": 1.15625,
89
+ "learning_rate": 4.917441177612131e-06,
90
+ "logits/chosen": -2.6114723682403564,
91
+ "logits/rejected": -3.054443120956421,
92
+ "logps/chosen": -146.25564575195312,
93
+ "logps/rejected": -317.78265380859375,
94
+ "loss": 0.3246,
95
+ "rewards/accuracies": 1.0,
96
+ "rewards/chosen": 0.314788281917572,
97
+ "rewards/margins": 0.967236340045929,
98
+ "rewards/rejected": -0.6524480581283569,
99
+ "step": 50
100
+ },
101
+ {
102
+ "epoch": 0.21,
103
+ "grad_norm": 0.97265625,
104
+ "learning_rate": 4.82126520118304e-06,
105
+ "logits/chosen": -2.6283130645751953,
106
+ "logits/rejected": -3.036774158477783,
107
+ "logps/chosen": -138.53135681152344,
108
+ "logps/rejected": -358.7559509277344,
109
+ "loss": 0.2051,
110
+ "rewards/accuracies": 1.0,
111
+ "rewards/chosen": 0.41738367080688477,
112
+ "rewards/margins": 1.4921326637268066,
113
+ "rewards/rejected": -1.0747489929199219,
114
+ "step": 60
115
+ },
116
+ {
117
+ "epoch": 0.25,
118
+ "grad_norm": 0.69921875,
119
+ "learning_rate": 4.6901752354885166e-06,
120
+ "logits/chosen": -2.6115269660949707,
121
+ "logits/rejected": -3.0476138591766357,
122
+ "logps/chosen": -129.70721435546875,
123
+ "logps/rejected": -413.200927734375,
124
+ "loss": 0.1157,
125
+ "rewards/accuracies": 1.0,
126
+ "rewards/chosen": 0.5094510316848755,
127
+ "rewards/margins": 2.1198441982269287,
128
+ "rewards/rejected": -1.6103931665420532,
129
+ "step": 70
130
+ },
131
+ {
132
+ "epoch": 0.28,
133
+ "grad_norm": 0.439453125,
134
+ "learning_rate": 4.526142995631488e-06,
135
+ "logits/chosen": -2.642693281173706,
136
+ "logits/rejected": -3.0582263469696045,
137
+ "logps/chosen": -123.13417053222656,
138
+ "logps/rejected": -475.8509216308594,
139
+ "loss": 0.0593,
140
+ "rewards/accuracies": 1.0,
141
+ "rewards/chosen": 0.572167694568634,
142
+ "rewards/margins": 2.8161401748657227,
143
+ "rewards/rejected": -2.2439725399017334,
144
+ "step": 80
145
+ },
146
+ {
147
+ "epoch": 0.32,
148
+ "grad_norm": 0.306640625,
149
+ "learning_rate": 4.331635679181032e-06,
150
+ "logits/chosen": -2.616441249847412,
151
+ "logits/rejected": -3.0742900371551514,
152
+ "logps/chosen": -119.6812744140625,
153
+ "logps/rejected": -535.9256591796875,
154
+ "loss": 0.0314,
155
+ "rewards/accuracies": 1.0,
156
+ "rewards/chosen": 0.61542809009552,
157
+ "rewards/margins": 3.4634997844696045,
158
+ "rewards/rejected": -2.848071813583374,
159
+ "step": 90
160
+ },
161
+ {
162
+ "epoch": 0.35,
163
+ "grad_norm": 0.224609375,
164
+ "learning_rate": 4.109578857224478e-06,
165
+ "logits/chosen": -2.6391441822052,
166
+ "logits/rejected": -3.0866036415100098,
167
+ "logps/chosen": -115.92496490478516,
168
+ "logps/rejected": -598.8836059570312,
169
+ "loss": 0.0166,
170
+ "rewards/accuracies": 1.0,
171
+ "rewards/chosen": 0.6349981427192688,
172
+ "rewards/margins": 4.107515811920166,
173
+ "rewards/rejected": -3.472517728805542,
174
+ "step": 100
175
+ },
176
+ {
177
+ "epoch": 0.35,
178
+ "eval_logits/chosen": -2.680793523788452,
179
+ "eval_logits/rejected": -3.105761766433716,
180
+ "eval_logps/chosen": -96.58148956298828,
181
+ "eval_logps/rejected": -579.1800537109375,
182
+ "eval_loss": 0.01978849433362484,
183
+ "eval_rewards/accuracies": 1.0,
184
+ "eval_rewards/chosen": 0.4884725511074066,
185
+ "eval_rewards/margins": 3.9327571392059326,
186
+ "eval_rewards/rejected": -3.444284439086914,
187
+ "eval_runtime": 1.2795,
188
+ "eval_samples_per_second": 3.908,
189
+ "eval_steps_per_second": 2.345,
190
+ "step": 100
191
+ },
192
+ {
193
+ "epoch": 0.39,
194
+ "grad_norm": 0.1416015625,
195
+ "learning_rate": 3.863312471055116e-06,
196
+ "logits/chosen": -2.5967559814453125,
197
+ "logits/rejected": -3.022758722305298,
198
+ "logps/chosen": -115.83799743652344,
199
+ "logps/rejected": -666.7772827148438,
200
+ "loss": 0.0084,
201
+ "rewards/accuracies": 1.0,
202
+ "rewards/chosen": 0.6381937265396118,
203
+ "rewards/margins": 4.796027183532715,
204
+ "rewards/rejected": -4.157833099365234,
205
+ "step": 110
206
+ },
207
+ {
208
+ "epoch": 0.42,
209
+ "grad_norm": 0.10400390625,
210
+ "learning_rate": 3.5965405963463197e-06,
211
+ "logits/chosen": -2.6488442420959473,
212
+ "logits/rejected": -3.06384015083313,
213
+ "logps/chosen": -113.5273208618164,
214
+ "logps/rejected": -722.9757080078125,
215
+ "loss": 0.0047,
216
+ "rewards/accuracies": 1.0,
217
+ "rewards/chosen": 0.6482743620872498,
218
+ "rewards/margins": 5.373518943786621,
219
+ "rewards/rejected": -4.725245475769043,
220
+ "step": 120
221
+ },
222
+ {
223
+ "epoch": 0.46,
224
+ "grad_norm": 0.07958984375,
225
+ "learning_rate": 3.313275730405658e-06,
226
+ "logits/chosen": -2.6456308364868164,
227
+ "logits/rejected": -3.099287509918213,
228
+ "logps/chosen": -114.421630859375,
229
+ "logps/rejected": -765.133544921875,
230
+ "loss": 0.0031,
231
+ "rewards/accuracies": 1.0,
232
+ "rewards/chosen": 0.6431270837783813,
233
+ "rewards/margins": 5.790853977203369,
234
+ "rewards/rejected": -5.147727012634277,
235
+ "step": 130
236
+ },
237
+ {
238
+ "epoch": 0.49,
239
+ "grad_norm": 0.056396484375,
240
+ "learning_rate": 3.0177784404805466e-06,
241
+ "logits/chosen": -2.6465344429016113,
242
+ "logits/rejected": -3.0662121772766113,
243
+ "logps/chosen": -115.8419418334961,
244
+ "logps/rejected": -792.1008911132812,
245
+ "loss": 0.0024,
246
+ "rewards/accuracies": 1.0,
247
+ "rewards/chosen": 0.642734169960022,
248
+ "rewards/margins": 6.044482231140137,
249
+ "rewards/rejected": -5.401747703552246,
250
+ "step": 140
251
+ },
252
+ {
253
+ "epoch": 0.53,
254
+ "grad_norm": 0.06591796875,
255
+ "learning_rate": 2.7144932808611002e-06,
256
+ "logits/chosen": -2.597626209259033,
257
+ "logits/rejected": -3.0344064235687256,
258
+ "logps/chosen": -115.4371337890625,
259
+ "logps/rejected": -800.051025390625,
260
+ "loss": 0.0022,
261
+ "rewards/accuracies": 1.0,
262
+ "rewards/chosen": 0.6369476914405823,
263
+ "rewards/margins": 6.132936000823975,
264
+ "rewards/rejected": -5.495988368988037,
265
+ "step": 150
266
+ },
267
+ {
268
+ "epoch": 0.56,
269
+ "grad_norm": 0.057373046875,
270
+ "learning_rate": 2.407981942646603e-06,
271
+ "logits/chosen": -2.659702777862549,
272
+ "logits/rejected": -3.115243434906006,
273
+ "logps/chosen": -115.50006103515625,
274
+ "logps/rejected": -816.8475341796875,
275
+ "loss": 0.0019,
276
+ "rewards/accuracies": 1.0,
277
+ "rewards/chosen": 0.6319659948348999,
278
+ "rewards/margins": 6.285677909851074,
279
+ "rewards/rejected": -5.653711795806885,
280
+ "step": 160
281
+ },
282
+ {
283
+ "epoch": 0.6,
284
+ "grad_norm": 0.05029296875,
285
+ "learning_rate": 2.102854641665347e-06,
286
+ "logits/chosen": -2.6373775005340576,
287
+ "logits/rejected": -3.08939790725708,
288
+ "logps/chosen": -115.76806640625,
289
+ "logps/rejected": -823.41650390625,
290
+ "loss": 0.0018,
291
+ "rewards/accuracies": 1.0,
292
+ "rewards/chosen": 0.6388428807258606,
293
+ "rewards/margins": 6.353511810302734,
294
+ "rewards/rejected": -5.7146687507629395,
295
+ "step": 170
296
+ },
297
+ {
298
+ "epoch": 0.63,
299
+ "grad_norm": 0.052734375,
300
+ "learning_rate": 1.8037007765373677e-06,
301
+ "logits/chosen": -2.663148880004883,
302
+ "logits/rejected": -3.0792253017425537,
303
+ "logps/chosen": -116.19815826416016,
304
+ "logps/rejected": -830.9599609375,
305
+ "loss": 0.0016,
306
+ "rewards/accuracies": 1.0,
307
+ "rewards/chosen": 0.6323127150535583,
308
+ "rewards/margins": 6.430493354797363,
309
+ "rewards/rejected": -5.798180103302002,
310
+ "step": 180
311
+ },
312
+ {
313
+ "epoch": 0.67,
314
+ "grad_norm": 0.042724609375,
315
+ "learning_rate": 1.5150198998473802e-06,
316
+ "logits/chosen": -2.6086456775665283,
317
+ "logits/rejected": -3.0565147399902344,
318
+ "logps/chosen": -116.57249450683594,
319
+ "logps/rejected": -834.7310791015625,
320
+ "loss": 0.0016,
321
+ "rewards/accuracies": 1.0,
322
+ "rewards/chosen": 0.6343868970870972,
323
+ "rewards/margins": 6.452385902404785,
324
+ "rewards/rejected": -5.817999362945557,
325
+ "step": 190
326
+ },
327
+ {
328
+ "epoch": 0.7,
329
+ "grad_norm": 0.04443359375,
330
+ "learning_rate": 1.2411540406857064e-06,
331
+ "logits/chosen": -2.620572566986084,
332
+ "logits/rejected": -3.073166608810425,
333
+ "logps/chosen": -114.90836334228516,
334
+ "logps/rejected": -838.9874877929688,
335
+ "loss": 0.0015,
336
+ "rewards/accuracies": 1.0,
337
+ "rewards/chosen": 0.6308305263519287,
338
+ "rewards/margins": 6.5031633377075195,
339
+ "rewards/rejected": -5.872332572937012,
340
+ "step": 200
341
+ },
342
+ {
343
+ "epoch": 0.7,
344
+ "eval_logits/chosen": -2.6815271377563477,
345
+ "eval_logits/rejected": -3.1155121326446533,
346
+ "eval_logps/chosen": -96.9991683959961,
347
+ "eval_logps/rejected": -740.9360961914062,
348
+ "eval_loss": 0.005451836623251438,
349
+ "eval_rewards/accuracies": 1.0,
350
+ "eval_rewards/chosen": 0.4842957556247711,
351
+ "eval_rewards/margins": 5.546140670776367,
352
+ "eval_rewards/rejected": -5.061844825744629,
353
+ "eval_runtime": 1.2858,
354
+ "eval_samples_per_second": 3.889,
355
+ "eval_steps_per_second": 2.333,
356
+ "step": 200
357
+ },
358
+ {
359
+ "epoch": 0.74,
360
+ "grad_norm": 0.044921875,
361
+ "learning_rate": 9.862223964891864e-07,
362
+ "logits/chosen": -2.659987449645996,
363
+ "logits/rejected": -3.083799362182617,
364
+ "logps/chosen": -114.6219482421875,
365
+ "logps/rejected": -828.9215087890625,
366
+ "loss": 0.002,
367
+ "rewards/accuracies": 1.0,
368
+ "rewards/chosen": 0.6267526149749756,
369
+ "rewards/margins": 6.408988952636719,
370
+ "rewards/rejected": -5.782236576080322,
371
+ "step": 210
372
+ },
373
+ {
374
+ "epoch": 0.77,
375
+ "grad_norm": 0.041748046875,
376
+ "learning_rate": 7.54059376477568e-07,
377
+ "logits/chosen": -2.6489715576171875,
378
+ "logits/rejected": -3.0990288257598877,
379
+ "logps/chosen": -116.3112564086914,
380
+ "logps/rejected": -840.1389770507812,
381
+ "loss": 0.0015,
382
+ "rewards/accuracies": 1.0,
383
+ "rewards/chosen": 0.6445231437683105,
384
+ "rewards/margins": 6.526413917541504,
385
+ "rewards/rejected": -5.881890773773193,
386
+ "step": 220
387
+ },
388
+ {
389
+ "epoch": 0.81,
390
+ "grad_norm": 0.03955078125,
391
+ "learning_rate": 5.481569285697638e-07,
392
+ "logits/chosen": -2.6626932621002197,
393
+ "logits/rejected": -3.06781005859375,
394
+ "logps/chosen": -115.39896392822266,
395
+ "logps/rejected": -836.1140747070312,
396
+ "loss": 0.0016,
397
+ "rewards/accuracies": 1.0,
398
+ "rewards/chosen": 0.6213386058807373,
399
+ "rewards/margins": 6.4759931564331055,
400
+ "rewards/rejected": -5.854653835296631,
401
+ "step": 230
402
+ },
403
+ {
404
+ "epoch": 0.84,
405
+ "grad_norm": 0.038818359375,
406
+ "learning_rate": 3.71612017236837e-07,
407
+ "logits/chosen": -2.6279659271240234,
408
+ "logits/rejected": -3.0500528812408447,
409
+ "logps/chosen": -116.8084945678711,
410
+ "logps/rejected": -845.3555908203125,
411
+ "loss": 0.0014,
412
+ "rewards/accuracies": 1.0,
413
+ "rewards/chosen": 0.6305276155471802,
414
+ "rewards/margins": 6.566590309143066,
415
+ "rewards/rejected": -5.936062335968018,
416
+ "step": 240
417
+ },
418
+ {
419
+ "epoch": 0.88,
420
+ "grad_norm": 0.04443359375,
421
+ "learning_rate": 2.2708004227369407e-07,
422
+ "logits/chosen": -2.6384856700897217,
423
+ "logits/rejected": -3.0602006912231445,
424
+ "logps/chosen": -115.619140625,
425
+ "logps/rejected": -842.5126953125,
426
+ "loss": 0.0015,
427
+ "rewards/accuracies": 1.0,
428
+ "rewards/chosen": 0.6297703981399536,
429
+ "rewards/margins": 6.527515411376953,
430
+ "rewards/rejected": -5.897744655609131,
431
+ "step": 250
432
+ },
433
+ {
434
+ "epoch": 0.91,
435
+ "grad_norm": 0.041015625,
436
+ "learning_rate": 1.1673489911451536e-07,
437
+ "logits/chosen": -2.635697603225708,
438
+ "logits/rejected": -3.0929272174835205,
439
+ "logps/chosen": -116.23138427734375,
440
+ "logps/rejected": -843.4679565429688,
441
+ "loss": 0.0015,
442
+ "rewards/accuracies": 1.0,
443
+ "rewards/chosen": 0.6368875503540039,
444
+ "rewards/margins": 6.544775485992432,
445
+ "rewards/rejected": -5.907887935638428,
446
+ "step": 260
447
+ },
448
+ {
449
+ "epoch": 0.95,
450
+ "grad_norm": 0.037353515625,
451
+ "learning_rate": 4.223628142195929e-08,
452
+ "logits/chosen": -2.6346564292907715,
453
+ "logits/rejected": -3.0773911476135254,
454
+ "logps/chosen": -115.94686126708984,
455
+ "logps/rejected": -848.7951049804688,
456
+ "loss": 0.0014,
457
+ "rewards/accuracies": 1.0,
458
+ "rewards/chosen": 0.6426523923873901,
459
+ "rewards/margins": 6.600114345550537,
460
+ "rewards/rejected": -5.957463264465332,
461
+ "step": 270
462
+ },
463
+ {
464
+ "epoch": 0.98,
465
+ "grad_norm": 0.044677734375,
466
+ "learning_rate": 4.704717749627052e-09,
467
+ "logits/chosen": -2.6303486824035645,
468
+ "logits/rejected": -3.084428310394287,
469
+ "logps/chosen": -115.82963562011719,
470
+ "logps/rejected": -837.4007568359375,
471
+ "loss": 0.0015,
472
+ "rewards/accuracies": 1.0,
473
+ "rewards/chosen": 0.6384350061416626,
474
+ "rewards/margins": 6.50278377532959,
475
+ "rewards/rejected": -5.864348888397217,
476
+ "step": 280
477
+ },
478
+ {
479
+ "epoch": 1.0,
480
+ "step": 285,
481
+ "total_flos": 0.0,
482
+ "train_loss": 0.11324987118704277,
483
+ "train_runtime": 748.1588,
484
+ "train_samples_per_second": 1.522,
485
+ "train_steps_per_second": 0.381
486
+ }
487
+ ],
488
+ "logging_steps": 10,
489
+ "max_steps": 285,
490
+ "num_input_tokens_seen": 0,
491
+ "num_train_epochs": 1,
492
+ "save_steps": 200,
493
+ "total_flos": 0.0,
494
+ "train_batch_size": 2,
495
+ "trial_name": null,
496
+ "trial_params": null
497
+ }