Hugging Face's logo

jz666
/
simpo-train-small-wrong

Text Generation
Transformers
Safetensors
gemma2
alignment-handbook
trl
simpo
Generated from Trainer
conversational
text-generation-inference
Model card Files
xet
Community
jz666 commited on
Commit
8d4d182
·
verified ·
1 Parent(s): 0001d49

Model save

Browse files
Files changed (5) hide show
  1. README.md +61 -0
  2. all_results.json +9 -0
  3. generation_config.json +8 -0
  4. train_results.json +9 -0
  5. trainer_state.json +447 -0
README.md ADDED
@@ -0,0 +1,61 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ library_name: transformers
3
+ license: gemma
4
+ base_model: google/gemma-2-9b-it
5
+ tags:
6
+ - trl
7
+ - simpo
8
+ - generated_from_trainer
9
+ model-index:
10
+ - name: simpo-train-small-wrong
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
+ # simpo-train-small-wrong
18
+
19
+ This model is a fine-tuned version of [google/gemma-2-9b-it](https://huggingface.co/google/gemma-2-9b-it) on an unknown dataset.
20
+
21
+ ## Model description
22
+
23
+ More information needed
24
+
25
+ ## Intended uses & limitations
26
+
27
+ More information needed
28
+
29
+ ## Training and evaluation data
30
+
31
+ More information needed
32
+
33
+ ## Training procedure
34
+
35
+ ### Training hyperparameters
36
+
37
+ The following hyperparameters were used during training:
38
+ - learning_rate: 8e-07
39
+ - train_batch_size: 2
40
+ - eval_batch_size: 4
41
+ - seed: 42
42
+ - distributed_type: multi-GPU
43
+ - num_devices: 4
44
+ - gradient_accumulation_steps: 16
45
+ - total_train_batch_size: 128
46
+ - total_eval_batch_size: 16
47
+ - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
48
+ - lr_scheduler_type: cosine
49
+ - lr_scheduler_warmup_ratio: 0.1
50
+ - num_epochs: 1
51
+
52
+ ### Training results
53
+
54
+
55
+
56
+ ### Framework versions
57
+
58
+ - Transformers 4.44.2
59
+ - Pytorch 2.7.0+cu128
60
+ - Datasets 2.18.0
61
+ - Tokenizers 0.19.1
all_results.json ADDED
@@ -0,0 +1,9 @@
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 0.995910949568378,
3
+ "total_flos": 0.0,
4
+ "train_loss": 4.865393144370866,
5
+ "train_runtime": 3024.5765,
6
+ "train_samples": 17605,
7
+ "train_samples_per_second": 5.821,
8
+ "train_steps_per_second": 0.045
9
+ }
generation_config.json ADDED
@@ -0,0 +1,8 @@
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_from_model_config": true,
3
+ "bos_token_id": 2,
4
+ "cache_implementation": "hybrid",
5
+ "eos_token_id": 1,
6
+ "pad_token_id": 0,
7
+ "transformers_version": "4.44.2"
8
+ }
train_results.json ADDED
@@ -0,0 +1,9 @@
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 0.995910949568378,
3
+ "total_flos": 0.0,
4
+ "train_loss": 4.865393144370866,
5
+ "train_runtime": 3024.5765,
6
+ "train_samples": 17605,
7
+ "train_samples_per_second": 5.821,
8
+ "train_steps_per_second": 0.045
9
+ }
trainer_state.json ADDED
@@ -0,0 +1,447 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": null,
3
+ "best_model_checkpoint": null,
4
+ "epoch": 0.995910949568378,
5
+ "eval_steps": 400,
6
+ "global_step": 137,
7
+ "is_hyper_param_search": false,
8
+ "is_local_process_zero": true,
9
+ "is_world_process_zero": true,
10
+ "log_history": [
11
+ {
12
+ "epoch": 0.03634711494775102,
13
+ "grad_norm": 42.18056900627306,
14
+ "learning_rate": 2.857142857142857e-07,
15
+ "logits/chosen": -10.936590194702148,
16
+ "logits/rejected": -11.232436180114746,
17
+ "logps/chosen": -0.6059076189994812,
18
+ "logps/rejected": -0.5706321001052856,
19
+ "loss": 5.4752,
20
+ "rewards/accuracies": 0.3125,
21
+ "rewards/chosen": -6.05907678604126,
22
+ "rewards/margins": -0.352755606174469,
23
+ "rewards/rejected": -5.7063212394714355,
24
+ "step": 5
25
+ },
26
+ {
27
+ "epoch": 0.07269422989550205,
28
+ "grad_norm": 33.128716826319426,
29
+ "learning_rate": 5.714285714285714e-07,
30
+ "logits/chosen": -10.911026954650879,
31
+ "logits/rejected": -11.181764602661133,
32
+ "logps/chosen": -0.6222352385520935,
33
+ "logps/rejected": -0.5984331965446472,
34
+ "loss": 5.5068,
35
+ "rewards/accuracies": 0.3187499940395355,
36
+ "rewards/chosen": -6.222352504730225,
37
+ "rewards/margins": -0.23802027106285095,
38
+ "rewards/rejected": -5.984332084655762,
39
+ "step": 10
40
+ },
41
+ {
42
+ "epoch": 0.10904134484325306,
43
+ "grad_norm": 50.848111590558474,
44
+ "learning_rate": 7.998695344323425e-07,
45
+ "logits/chosen": -10.209379196166992,
46
+ "logits/rejected": -10.468820571899414,
47
+ "logps/chosen": -0.6080077290534973,
48
+ "logps/rejected": -0.5699423551559448,
49
+ "loss": 5.3874,
50
+ "rewards/accuracies": 0.2874999940395355,
51
+ "rewards/chosen": -6.080076694488525,
52
+ "rewards/margins": -0.3806529641151428,
53
+ "rewards/rejected": -5.699423789978027,
54
+ "step": 15
55
+ },
56
+ {
57
+ "epoch": 0.1453884597910041,
58
+ "grad_norm": 35.145738479410355,
59
+ "learning_rate": 7.953121695121394e-07,
60
+ "logits/chosen": -9.792947769165039,
61
+ "logits/rejected": -10.061820030212402,
62
+ "logps/chosen": -0.5358205437660217,
63
+ "logps/rejected": -0.5112504363059998,
64
+ "loss": 5.3654,
65
+ "rewards/accuracies": 0.3499999940395355,
66
+ "rewards/chosen": -5.358205318450928,
67
+ "rewards/margins": -0.24570095539093018,
68
+ "rewards/rejected": -5.112504482269287,
69
+ "step": 20
70
+ },
71
+ {
72
+ "epoch": 0.18173557473875512,
73
+ "grad_norm": 35.43905001965967,
74
+ "learning_rate": 7.843163833184991e-07,
75
+ "logits/chosen": -9.215612411499023,
76
+ "logits/rejected": -9.5372314453125,
77
+ "logps/chosen": -0.47783246636390686,
78
+ "logps/rejected": -0.4458232522010803,
79
+ "loss": 5.2677,
80
+ "rewards/accuracies": 0.33125001192092896,
81
+ "rewards/chosen": -4.778324604034424,
82
+ "rewards/margins": -0.32009226083755493,
83
+ "rewards/rejected": -4.4582319259643555,
84
+ "step": 25
85
+ },
86
+ {
87
+ "epoch": 0.21808268968650613,
88
+ "grad_norm": 34.872469426364724,
89
+ "learning_rate": 7.670612634414511e-07,
90
+ "logits/chosen": -8.819177627563477,
91
+ "logits/rejected": -8.965360641479492,
92
+ "logps/chosen": -0.49288082122802734,
93
+ "logps/rejected": -0.4752427041530609,
94
+ "loss": 5.1863,
95
+ "rewards/accuracies": 0.34375,
96
+ "rewards/chosen": -4.928807735443115,
97
+ "rewards/margins": -0.17638036608695984,
98
+ "rewards/rejected": -4.752427577972412,
99
+ "step": 30
100
+ },
101
+ {
102
+ "epoch": 0.25442980463425713,
103
+ "grad_norm": 41.683131083762184,
104
+ "learning_rate": 7.438278427948805e-07,
105
+ "logits/chosen": -8.983338356018066,
106
+ "logits/rejected": -9.358332633972168,
107
+ "logps/chosen": -0.45715832710266113,
108
+ "logps/rejected": -0.4428199231624603,
109
+ "loss": 5.1479,
110
+ "rewards/accuracies": 0.3187499940395355,
111
+ "rewards/chosen": -4.571582794189453,
112
+ "rewards/margins": -0.14338405430316925,
113
+ "rewards/rejected": -4.428199291229248,
114
+ "step": 35
115
+ },
116
+ {
117
+ "epoch": 0.2907769195820082,
118
+ "grad_norm": 35.49350724475253,
119
+ "learning_rate": 7.149945224533862e-07,
120
+ "logits/chosen": -8.473846435546875,
121
+ "logits/rejected": -8.634288787841797,
122
+ "logps/chosen": -0.49668779969215393,
123
+ "logps/rejected": -0.4822346568107605,
124
+ "loss": 5.1059,
125
+ "rewards/accuracies": 0.4124999940395355,
126
+ "rewards/chosen": -4.9668779373168945,
127
+ "rewards/margins": -0.14453084766864777,
128
+ "rewards/rejected": -4.8223466873168945,
129
+ "step": 40
130
+ },
131
+ {
132
+ "epoch": 0.3271240345297592,
133
+ "grad_norm": 35.201815027229166,
134
+ "learning_rate": 6.810309086608129e-07,
135
+ "logits/chosen": -9.485041618347168,
136
+ "logits/rejected": -9.641453742980957,
137
+ "logps/chosen": -0.4795761704444885,
138
+ "logps/rejected": -0.4812702536582947,
139
+ "loss": 5.0687,
140
+ "rewards/accuracies": 0.3812499940395355,
141
+ "rewards/chosen": -4.795762062072754,
142
+ "rewards/margins": 0.016940664499998093,
143
+ "rewards/rejected": -4.812702655792236,
144
+ "step": 45
145
+ },
146
+ {
147
+ "epoch": 0.36347114947751025,
148
+ "grad_norm": 41.62812201333068,
149
+ "learning_rate": 6.424901643866552e-07,
150
+ "logits/chosen": -9.312422752380371,
151
+ "logits/rejected": -9.527371406555176,
152
+ "logps/chosen": -0.5089246034622192,
153
+ "logps/rejected": -0.5194448232650757,
154
+ "loss": 5.0107,
155
+ "rewards/accuracies": 0.4124999940395355,
156
+ "rewards/chosen": -5.0892462730407715,
157
+ "rewards/margins": 0.10520193725824356,
158
+ "rewards/rejected": -5.194448471069336,
159
+ "step": 50
160
+ },
161
+ {
162
+ "epoch": 0.39981826442526125,
163
+ "grad_norm": 34.69469413711179,
164
+ "learning_rate": 6e-07,
165
+ "logits/chosen": -9.82011604309082,
166
+ "logits/rejected": -9.918914794921875,
167
+ "logps/chosen": -0.5091612935066223,
168
+ "logps/rejected": -0.5071789026260376,
169
+ "loss": 4.9611,
170
+ "rewards/accuracies": 0.4749999940395355,
171
+ "rewards/chosen": -5.091612815856934,
172
+ "rewards/margins": -0.019824281334877014,
173
+ "rewards/rejected": -5.071788311004639,
174
+ "step": 55
175
+ },
176
+ {
177
+ "epoch": 0.43616537937301225,
178
+ "grad_norm": 39.20092819974629,
179
+ "learning_rate": 5.542524497952543e-07,
180
+ "logits/chosen": -10.392576217651367,
181
+ "logits/rejected": -10.515856742858887,
182
+ "logps/chosen": -0.49750223755836487,
183
+ "logps/rejected": -0.5498560070991516,
184
+ "loss": 4.7725,
185
+ "rewards/accuracies": 0.4625000059604645,
186
+ "rewards/chosen": -4.975022315979004,
187
+ "rewards/margins": 0.5235382318496704,
188
+ "rewards/rejected": -5.498560428619385,
189
+ "step": 60
190
+ },
191
+ {
192
+ "epoch": 0.4725124943207633,
193
+ "grad_norm": 41.76932417294435,
194
+ "learning_rate": 5.059926008786647e-07,
195
+ "logits/chosen": -9.944269180297852,
196
+ "logits/rejected": -10.23642635345459,
197
+ "logps/chosen": -0.5527420043945312,
198
+ "logps/rejected": -0.6039702296257019,
199
+ "loss": 4.9196,
200
+ "rewards/accuracies": 0.5,
201
+ "rewards/chosen": -5.5274200439453125,
202
+ "rewards/margins": 0.5122829675674438,
203
+ "rewards/rejected": -6.039702415466309,
204
+ "step": 65
205
+ },
206
+ {
207
+ "epoch": 0.5088596092685143,
208
+ "grad_norm": 42.78642756929462,
209
+ "learning_rate": 4.5600645798745166e-07,
210
+ "logits/chosen": -10.92486572265625,
211
+ "logits/rejected": -11.130925178527832,
212
+ "logps/chosen": -0.5507891774177551,
213
+ "logps/rejected": -0.5964936017990112,
214
+ "loss": 4.7685,
215
+ "rewards/accuracies": 0.5062500238418579,
216
+ "rewards/chosen": -5.50789213180542,
217
+ "rewards/margins": 0.4570443034172058,
218
+ "rewards/rejected": -5.964935779571533,
219
+ "step": 70
220
+ },
221
+ {
222
+ "epoch": 0.5452067242162654,
223
+ "grad_norm": 43.736655032116566,
224
+ "learning_rate": 4.051081418863895e-07,
225
+ "logits/chosen": -10.675444602966309,
226
+ "logits/rejected": -10.829668045043945,
227
+ "logps/chosen": -0.6147274374961853,
228
+ "logps/rejected": -0.6446117758750916,
229
+ "loss": 4.8607,
230
+ "rewards/accuracies": 0.581250011920929,
231
+ "rewards/chosen": -6.147274494171143,
232
+ "rewards/margins": 0.29884377121925354,
233
+ "rewards/rejected": -6.446118354797363,
234
+ "step": 75
235
+ },
236
+ {
237
+ "epoch": 0.5815538391640164,
238
+ "grad_norm": 50.47780593376421,
239
+ "learning_rate": 3.541266298406398e-07,
240
+ "logits/chosen": -11.719624519348145,
241
+ "logits/rejected": -11.85960578918457,
242
+ "logps/chosen": -0.6017196774482727,
243
+ "logps/rejected": -0.6400548815727234,
244
+ "loss": 4.7166,
245
+ "rewards/accuracies": 0.5,
246
+ "rewards/chosen": -6.017197608947754,
247
+ "rewards/margins": 0.38335245847702026,
248
+ "rewards/rejected": -6.400549411773682,
249
+ "step": 80
250
+ },
251
+ {
252
+ "epoch": 0.6179009541117674,
253
+ "grad_norm": 49.58320001105474,
254
+ "learning_rate": 3.0389225412181565e-07,
255
+ "logits/chosen": -11.57774543762207,
256
+ "logits/rejected": -11.787096977233887,
257
+ "logps/chosen": -0.639508843421936,
258
+ "logps/rejected": -0.7087122797966003,
259
+ "loss": 4.6635,
260
+ "rewards/accuracies": 0.59375,
261
+ "rewards/chosen": -6.395088195800781,
262
+ "rewards/margins": 0.6920341849327087,
263
+ "rewards/rejected": -7.087122917175293,
264
+ "step": 85
265
+ },
266
+ {
267
+ "epoch": 0.6542480690595184,
268
+ "grad_norm": 60.72521586951035,
269
+ "learning_rate": 2.5522317844515273e-07,
270
+ "logits/chosen": -11.539664268493652,
271
+ "logits/rejected": -11.73334789276123,
272
+ "logps/chosen": -0.6763519644737244,
273
+ "logps/rejected": -0.719671905040741,
274
+ "loss": 4.6877,
275
+ "rewards/accuracies": 0.543749988079071,
276
+ "rewards/chosen": -6.763520240783691,
277
+ "rewards/margins": 0.4331996440887451,
278
+ "rewards/rejected": -7.196720123291016,
279
+ "step": 90
280
+ },
281
+ {
282
+ "epoch": 0.6905951840072694,
283
+ "grad_norm": 73.18876407608641,
284
+ "learning_rate": 2.0891207259509476e-07,
285
+ "logits/chosen": -11.891935348510742,
286
+ "logits/rejected": -11.978799819946289,
287
+ "logps/chosen": -0.6321767568588257,
288
+ "logps/rejected": -0.7210510969161987,
289
+ "loss": 4.6661,
290
+ "rewards/accuracies": 0.5625,
291
+ "rewards/chosen": -6.321768283843994,
292
+ "rewards/margins": 0.8887429237365723,
293
+ "rewards/rejected": -7.210511207580566,
294
+ "step": 95
295
+ },
296
+ {
297
+ "epoch": 0.7269422989550205,
298
+ "grad_norm": 58.29758563166122,
299
+ "learning_rate": 1.6571320226872206e-07,
300
+ "logits/chosen": -12.378867149353027,
301
+ "logits/rejected": -12.450654983520508,
302
+ "logps/chosen": -0.7091072797775269,
303
+ "logps/rejected": -0.7949134111404419,
304
+ "loss": 4.5318,
305
+ "rewards/accuracies": 0.6499999761581421,
306
+ "rewards/chosen": -7.091072082519531,
307
+ "rewards/margins": 0.8580614924430847,
308
+ "rewards/rejected": -7.949133396148682,
309
+ "step": 100
310
+ },
311
+ {
312
+ "epoch": 0.7632894139027715,
313
+ "grad_norm": 65.14186021974389,
314
+ "learning_rate": 1.2633014440382787e-07,
315
+ "logits/chosen": -12.120596885681152,
316
+ "logits/rejected": -12.198526382446289,
317
+ "logps/chosen": -0.7462261319160461,
318
+ "logps/rejected": -0.8084705471992493,
319
+ "loss": 4.5075,
320
+ "rewards/accuracies": 0.5375000238418579,
321
+ "rewards/chosen": -7.462261199951172,
322
+ "rewards/margins": 0.622443437576294,
323
+ "rewards/rejected": -8.084705352783203,
324
+ "step": 105
325
+ },
326
+ {
327
+ "epoch": 0.7996365288505225,
328
+ "grad_norm": 66.21038830050585,
329
+ "learning_rate": 9.14043280712228e-08,
330
+ "logits/chosen": -12.642437934875488,
331
+ "logits/rejected": -12.874890327453613,
332
+ "logps/chosen": -0.7560291290283203,
333
+ "logps/rejected": -0.8245932459831238,
334
+ "loss": 4.6235,
335
+ "rewards/accuracies": 0.5874999761581421,
336
+ "rewards/chosen": -7.560290336608887,
337
+ "rewards/margins": 0.6856421828269958,
338
+ "rewards/rejected": -8.245932579040527,
339
+ "step": 110
340
+ },
341
+ {
342
+ "epoch": 0.8359836437982735,
343
+ "grad_norm": 58.78661492817544,
344
+ "learning_rate": 6.150458756494239e-08,
345
+ "logits/chosen": -12.5349760055542,
346
+ "logits/rejected": -12.633232116699219,
347
+ "logps/chosen": -0.7407893538475037,
348
+ "logps/rejected": -0.8003439903259277,
349
+ "loss": 4.537,
350
+ "rewards/accuracies": 0.5375000238418579,
351
+ "rewards/chosen": -7.407894134521484,
352
+ "rewards/margins": 0.5955458879470825,
353
+ "rewards/rejected": -8.003438949584961,
354
+ "step": 115
355
+ },
356
+ {
357
+ "epoch": 0.8723307587460245,
358
+ "grad_norm": 72.85901824273444,
359
+ "learning_rate": 3.711789783843522e-08,
360
+ "logits/chosen": -12.179682731628418,
361
+ "logits/rejected": -12.421493530273438,
362
+ "logps/chosen": -0.7077151536941528,
363
+ "logps/rejected": -0.8436015248298645,
364
+ "loss": 4.41,
365
+ "rewards/accuracies": 0.6187499761581421,
366
+ "rewards/chosen": -7.077151298522949,
367
+ "rewards/margins": 1.358863115310669,
368
+ "rewards/rejected": -8.436014175415039,
369
+ "step": 120
370
+ },
371
+ {
372
+ "epoch": 0.9086778736937755,
373
+ "grad_norm": 81.44112780212329,
374
+ "learning_rate": 1.8641443178027784e-08,
375
+ "logits/chosen": -12.844386100769043,
376
+ "logits/rejected": -13.204521179199219,
377
+ "logps/chosen": -0.7584520578384399,
378
+ "logps/rejected": -0.8550151586532593,
379
+ "loss": 4.4737,
380
+ "rewards/accuracies": 0.6312500238418579,
381
+ "rewards/chosen": -7.584519863128662,
382
+ "rewards/margins": 0.965630829334259,
383
+ "rewards/rejected": -8.550151824951172,
384
+ "step": 125
385
+ },
386
+ {
387
+ "epoch": 0.9450249886415266,
388
+ "grad_norm": 63.15208705033616,
389
+ "learning_rate": 6.376148290617145e-09,
390
+ "logits/chosen": -13.006174087524414,
391
+ "logits/rejected": -13.145883560180664,
392
+ "logps/chosen": -0.7715423703193665,
393
+ "logps/rejected": -0.8375405073165894,
394
+ "loss": 4.6746,
395
+ "rewards/accuracies": 0.6187499761581421,
396
+ "rewards/chosen": -7.715423583984375,
397
+ "rewards/margins": 0.6599816679954529,
398
+ "rewards/rejected": -8.375406265258789,
399
+ "step": 130
400
+ },
401
+ {
402
+ "epoch": 0.9813721035892776,
403
+ "grad_norm": 69.30502594164692,
404
+ "learning_rate": 5.217771643080127e-10,
405
+ "logits/chosen": -12.759674072265625,
406
+ "logits/rejected": -12.864163398742676,
407
+ "logps/chosen": -0.7980765700340271,
408
+ "logps/rejected": -0.8657040596008301,
409
+ "loss": 4.3661,
410
+ "rewards/accuracies": 0.612500011920929,
411
+ "rewards/chosen": -7.980767250061035,
412
+ "rewards/margins": 0.6762741804122925,
413
+ "rewards/rejected": -8.6570405960083,
414
+ "step": 135
415
+ },
416
+ {
417
+ "epoch": 0.995910949568378,
418
+ "step": 137,
419
+ "total_flos": 0.0,
420
+ "train_loss": 4.865393144370866,
421
+ "train_runtime": 3024.5765,
422
+ "train_samples_per_second": 5.821,
423
+ "train_steps_per_second": 0.045
424
+ }
425
+ ],
426
+ "logging_steps": 5,
427
+ "max_steps": 137,
428
+ "num_input_tokens_seen": 0,
429
+ "num_train_epochs": 1,
430
+ "save_steps": 1000000,
431
+ "stateful_callbacks": {
432
+ "TrainerControl": {
433
+ "args": {
434
+ "should_epoch_stop": false,
435
+ "should_evaluate": false,
436
+ "should_log": false,
437
+ "should_save": true,
438
+ "should_training_stop": true
439
+ },
440
+ "attributes": {}
441
+ }
442
+ },
443
+ "total_flos": 0.0,
444
+ "train_batch_size": 2,
445
+ "trial_name": null,
446
+ "trial_params": null
447
+ }