lole25 commited on
Commit
f8f0b71
1 Parent(s): 6625b8a

Model save

Browse files
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: mistralai/Mistral-7B-v0.1
9
+ model-index:
10
+ - name: zephyr-7b-gpo-update3-i1
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-7b-gpo-update3-i1
18
+
19
+ This model is a fine-tuned version of [mistralai/Mistral-7B-v0.1](https://huggingface.co/mistralai/Mistral-7B-v0.1) on the None dataset.
20
+ It achieves the following results on the evaluation set:
21
+ - Loss: 0.0574
22
+ - Rewards/chosen: -0.0003
23
+ - Rewards/rejected: 0.0038
24
+ - Rewards/accuracies: 0.3765
25
+ - Rewards/margins: -0.0041
26
+ - Logps/rejected: -254.1840
27
+ - Logps/chosen: -266.7596
28
+ - Logits/rejected: -1.8151
29
+ - Logits/chosen: -1.9709
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
+ | 0.0013 | 0.4 | 100 | 0.0537 | 0.0 | 0.0 | 0.0 | 0.0 | -254.9398 | -266.6976 | -1.8067 | -1.9618 |
65
+ | 0.0013 | 0.8 | 200 | 0.0575 | -0.0013 | 0.0029 | 0.3800 | -0.0041 | -254.3691 | -266.9557 | -1.8139 | -1.9695 |
66
+
67
+
68
+ ### Framework versions
69
+
70
+ - PEFT 0.7.1
71
+ - Transformers 4.36.2
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:ab63b8a7f3809338c650ccfd5464a84ac38fcd15e4aefbf270f009e2e8301b2c
3
  size 671150064
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d9d2f25b332c4650a67eac7e3ca2b1bf46300c95604dee259941128783de35e6
3
  size 671150064
all_results.json ADDED
@@ -0,0 +1,21 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 1.0,
3
+ "eval_logits/chosen": -1.9709315299987793,
4
+ "eval_logits/rejected": -1.8150750398635864,
5
+ "eval_logps/chosen": -266.7596130371094,
6
+ "eval_logps/rejected": -254.18397521972656,
7
+ "eval_loss": 0.057394202798604965,
8
+ "eval_rewards/accuracies": 0.3765000104904175,
9
+ "eval_rewards/chosen": -0.0003101456386502832,
10
+ "eval_rewards/margins": -0.004089393652975559,
11
+ "eval_rewards/rejected": 0.003779248334467411,
12
+ "eval_runtime": 709.1028,
13
+ "eval_samples": 2000,
14
+ "eval_samples_per_second": 2.82,
15
+ "eval_steps_per_second": 1.41,
16
+ "train_loss": 0.0016641309279948472,
17
+ "train_runtime": 2302.1859,
18
+ "train_samples": 61135,
19
+ "train_samples_per_second": 0.434,
20
+ "train_steps_per_second": 0.109
21
+ }
eval_results.json ADDED
@@ -0,0 +1,16 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 1.0,
3
+ "eval_logits/chosen": -1.9709315299987793,
4
+ "eval_logits/rejected": -1.8150750398635864,
5
+ "eval_logps/chosen": -266.7596130371094,
6
+ "eval_logps/rejected": -254.18397521972656,
7
+ "eval_loss": 0.057394202798604965,
8
+ "eval_rewards/accuracies": 0.3765000104904175,
9
+ "eval_rewards/chosen": -0.0003101456386502832,
10
+ "eval_rewards/margins": -0.004089393652975559,
11
+ "eval_rewards/rejected": 0.003779248334467411,
12
+ "eval_runtime": 709.1028,
13
+ "eval_samples": 2000,
14
+ "eval_samples_per_second": 2.82,
15
+ "eval_steps_per_second": 1.41
16
+ }
runs/Apr08_23-22-40_gpu4-119-5/events.out.tfevents.1712582624.gpu4-119-5.42432.0 CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:8cb34f378552a24a218061d0d4164fcdab48d7e751f857f22b3d490d56cbb57a
3
- size 19225
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a702b666d34efb4625a58f041788f819432b68e6a49c86b3acb93aa2d4f7bb4c
3
+ size 22749
runs/Apr08_23-22-40_gpu4-119-5/events.out.tfevents.1712585635.gpu4-119-5.42432.1 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:210bf1be03dda53fd4018949430cf6ddd5ab27ca2d6abb0cf8aae3e313086cb1
3
+ size 828
train_results.json ADDED
@@ -0,0 +1,8 @@
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 1.0,
3
+ "train_loss": 0.0016641309279948472,
4
+ "train_runtime": 2302.1859,
5
+ "train_samples": 61135,
6
+ "train_samples_per_second": 0.434,
7
+ "train_steps_per_second": 0.109
8
+ }
trainer_state.json ADDED
@@ -0,0 +1,426 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": null,
3
+ "best_model_checkpoint": null,
4
+ "epoch": 1.0,
5
+ "eval_steps": 100,
6
+ "global_step": 250,
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
+ "learning_rate": 2.0000000000000002e-07,
14
+ "logits/chosen": -1.8503975868225098,
15
+ "logits/rejected": -1.8503975868225098,
16
+ "logps/chosen": 0.0,
17
+ "logps/rejected": 0.0,
18
+ "loss": 0.0011,
19
+ "rewards/accuracies": 0.0,
20
+ "rewards/chosen": 0.0,
21
+ "rewards/margins": 0.0,
22
+ "rewards/rejected": 0.0,
23
+ "step": 1
24
+ },
25
+ {
26
+ "epoch": 0.04,
27
+ "learning_rate": 2.0000000000000003e-06,
28
+ "logits/chosen": -1.8588156700134277,
29
+ "logits/rejected": -1.8588156700134277,
30
+ "logps/chosen": 0.0,
31
+ "logps/rejected": 0.0,
32
+ "loss": 0.0019,
33
+ "rewards/accuracies": 0.0,
34
+ "rewards/chosen": 0.0,
35
+ "rewards/margins": 0.0,
36
+ "rewards/rejected": 0.0,
37
+ "step": 10
38
+ },
39
+ {
40
+ "epoch": 0.08,
41
+ "learning_rate": 4.000000000000001e-06,
42
+ "logits/chosen": -1.970517873764038,
43
+ "logits/rejected": -1.970517873764038,
44
+ "logps/chosen": 0.0,
45
+ "logps/rejected": 0.0,
46
+ "loss": 0.0018,
47
+ "rewards/accuracies": 0.0,
48
+ "rewards/chosen": 0.0,
49
+ "rewards/margins": 0.0,
50
+ "rewards/rejected": 0.0,
51
+ "step": 20
52
+ },
53
+ {
54
+ "epoch": 0.12,
55
+ "learning_rate": 4.993910125649561e-06,
56
+ "logits/chosen": -1.9209930896759033,
57
+ "logits/rejected": -1.9209930896759033,
58
+ "logps/chosen": 0.0,
59
+ "logps/rejected": 0.0,
60
+ "loss": 0.0013,
61
+ "rewards/accuracies": 0.0,
62
+ "rewards/chosen": 0.0,
63
+ "rewards/margins": 0.0,
64
+ "rewards/rejected": 0.0,
65
+ "step": 30
66
+ },
67
+ {
68
+ "epoch": 0.16,
69
+ "learning_rate": 4.9453690018345144e-06,
70
+ "logits/chosen": -1.883547067642212,
71
+ "logits/rejected": -1.883547067642212,
72
+ "logps/chosen": 0.0,
73
+ "logps/rejected": 0.0,
74
+ "loss": 0.0019,
75
+ "rewards/accuracies": 0.0,
76
+ "rewards/chosen": 0.0,
77
+ "rewards/margins": 0.0,
78
+ "rewards/rejected": 0.0,
79
+ "step": 40
80
+ },
81
+ {
82
+ "epoch": 0.2,
83
+ "learning_rate": 4.849231551964771e-06,
84
+ "logits/chosen": -1.9128715991973877,
85
+ "logits/rejected": -1.9128715991973877,
86
+ "logps/chosen": 0.0,
87
+ "logps/rejected": 0.0,
88
+ "loss": 0.0013,
89
+ "rewards/accuracies": 0.0,
90
+ "rewards/chosen": 0.0,
91
+ "rewards/margins": 0.0,
92
+ "rewards/rejected": 0.0,
93
+ "step": 50
94
+ },
95
+ {
96
+ "epoch": 0.24,
97
+ "learning_rate": 4.707368982147318e-06,
98
+ "logits/chosen": -2.0107295513153076,
99
+ "logits/rejected": -2.0107295513153076,
100
+ "logps/chosen": 0.0,
101
+ "logps/rejected": 0.0,
102
+ "loss": 0.0019,
103
+ "rewards/accuracies": 0.0,
104
+ "rewards/chosen": 0.0,
105
+ "rewards/margins": 0.0,
106
+ "rewards/rejected": 0.0,
107
+ "step": 60
108
+ },
109
+ {
110
+ "epoch": 0.28,
111
+ "learning_rate": 4.522542485937369e-06,
112
+ "logits/chosen": -1.9920228719711304,
113
+ "logits/rejected": -1.9920228719711304,
114
+ "logps/chosen": 0.0,
115
+ "logps/rejected": 0.0,
116
+ "loss": 0.0014,
117
+ "rewards/accuracies": 0.0,
118
+ "rewards/chosen": 0.0,
119
+ "rewards/margins": 0.0,
120
+ "rewards/rejected": 0.0,
121
+ "step": 70
122
+ },
123
+ {
124
+ "epoch": 0.32,
125
+ "learning_rate": 4.2983495008466285e-06,
126
+ "logits/chosen": -1.8801155090332031,
127
+ "logits/rejected": -1.8801155090332031,
128
+ "logps/chosen": 0.0,
129
+ "logps/rejected": 0.0,
130
+ "loss": 0.0024,
131
+ "rewards/accuracies": 0.0,
132
+ "rewards/chosen": 0.0,
133
+ "rewards/margins": 0.0,
134
+ "rewards/rejected": 0.0,
135
+ "step": 80
136
+ },
137
+ {
138
+ "epoch": 0.36,
139
+ "learning_rate": 4.039153688314146e-06,
140
+ "logits/chosen": -2.050198793411255,
141
+ "logits/rejected": -2.050198793411255,
142
+ "logps/chosen": 0.0,
143
+ "logps/rejected": 0.0,
144
+ "loss": 0.0021,
145
+ "rewards/accuracies": 0.0,
146
+ "rewards/chosen": 0.0,
147
+ "rewards/margins": 0.0,
148
+ "rewards/rejected": 0.0,
149
+ "step": 90
150
+ },
151
+ {
152
+ "epoch": 0.4,
153
+ "learning_rate": 3.7500000000000005e-06,
154
+ "logits/chosen": -1.8852717876434326,
155
+ "logits/rejected": -1.8852717876434326,
156
+ "logps/chosen": 0.0,
157
+ "logps/rejected": 0.0,
158
+ "loss": 0.0013,
159
+ "rewards/accuracies": 0.0,
160
+ "rewards/chosen": 0.0,
161
+ "rewards/margins": 0.0,
162
+ "rewards/rejected": 0.0,
163
+ "step": 100
164
+ },
165
+ {
166
+ "epoch": 0.4,
167
+ "eval_logits/chosen": -1.9617642164230347,
168
+ "eval_logits/rejected": -1.8066532611846924,
169
+ "eval_logps/chosen": -266.6976013183594,
170
+ "eval_logps/rejected": -254.9398193359375,
171
+ "eval_loss": 0.053734518587589264,
172
+ "eval_rewards/accuracies": 0.0,
173
+ "eval_rewards/chosen": 0.0,
174
+ "eval_rewards/margins": 0.0,
175
+ "eval_rewards/rejected": 0.0,
176
+ "eval_runtime": 702.6753,
177
+ "eval_samples_per_second": 2.846,
178
+ "eval_steps_per_second": 1.423,
179
+ "step": 100
180
+ },
181
+ {
182
+ "epoch": 0.44,
183
+ "learning_rate": 3.436516483539781e-06,
184
+ "logits/chosen": -1.731688141822815,
185
+ "logits/rejected": -1.731688141822815,
186
+ "logps/chosen": 0.0,
187
+ "logps/rejected": 0.0,
188
+ "loss": 0.0012,
189
+ "rewards/accuracies": 0.0,
190
+ "rewards/chosen": 0.0,
191
+ "rewards/margins": 0.0,
192
+ "rewards/rejected": 0.0,
193
+ "step": 110
194
+ },
195
+ {
196
+ "epoch": 0.48,
197
+ "learning_rate": 3.1048047389991693e-06,
198
+ "logits/chosen": -1.8530235290527344,
199
+ "logits/rejected": -1.8530235290527344,
200
+ "logps/chosen": 0.0,
201
+ "logps/rejected": 0.0,
202
+ "loss": 0.0013,
203
+ "rewards/accuracies": 0.0,
204
+ "rewards/chosen": 0.0,
205
+ "rewards/margins": 0.0,
206
+ "rewards/rejected": 0.0,
207
+ "step": 120
208
+ },
209
+ {
210
+ "epoch": 0.52,
211
+ "learning_rate": 2.761321158169134e-06,
212
+ "logits/chosen": -2.0225424766540527,
213
+ "logits/rejected": -2.0225424766540527,
214
+ "logps/chosen": 0.0,
215
+ "logps/rejected": 0.0,
216
+ "loss": 0.0016,
217
+ "rewards/accuracies": 0.0,
218
+ "rewards/chosen": 0.0,
219
+ "rewards/margins": 0.0,
220
+ "rewards/rejected": 0.0,
221
+ "step": 130
222
+ },
223
+ {
224
+ "epoch": 0.56,
225
+ "learning_rate": 2.4127512582437486e-06,
226
+ "logits/chosen": -1.8995482921600342,
227
+ "logits/rejected": -1.8995482921600342,
228
+ "logps/chosen": 0.0,
229
+ "logps/rejected": 0.0,
230
+ "loss": 0.0014,
231
+ "rewards/accuracies": 0.0,
232
+ "rewards/chosen": 0.0,
233
+ "rewards/margins": 0.0,
234
+ "rewards/rejected": 0.0,
235
+ "step": 140
236
+ },
237
+ {
238
+ "epoch": 0.6,
239
+ "learning_rate": 2.0658795558326745e-06,
240
+ "logits/chosen": -1.86004638671875,
241
+ "logits/rejected": -1.8391777276992798,
242
+ "logps/chosen": -4.896004676818848,
243
+ "logps/rejected": -1.6084611415863037,
244
+ "loss": 0.0016,
245
+ "rewards/accuracies": 0.0,
246
+ "rewards/chosen": 0.0,
247
+ "rewards/margins": 0.0,
248
+ "rewards/rejected": 0.0,
249
+ "step": 150
250
+ },
251
+ {
252
+ "epoch": 0.64,
253
+ "learning_rate": 1.7274575140626318e-06,
254
+ "logits/chosen": -2.013669490814209,
255
+ "logits/rejected": -2.013669490814209,
256
+ "logps/chosen": 0.0,
257
+ "logps/rejected": 0.0,
258
+ "loss": 0.0019,
259
+ "rewards/accuracies": 0.0,
260
+ "rewards/chosen": 0.0,
261
+ "rewards/margins": 0.0,
262
+ "rewards/rejected": 0.0,
263
+ "step": 160
264
+ },
265
+ {
266
+ "epoch": 0.68,
267
+ "learning_rate": 1.4040721330273063e-06,
268
+ "logits/chosen": -1.8206443786621094,
269
+ "logits/rejected": -1.8206443786621094,
270
+ "logps/chosen": 0.0,
271
+ "logps/rejected": 0.0,
272
+ "loss": 0.0019,
273
+ "rewards/accuracies": 0.0,
274
+ "rewards/chosen": 0.0,
275
+ "rewards/margins": 0.0,
276
+ "rewards/rejected": 0.0,
277
+ "step": 170
278
+ },
279
+ {
280
+ "epoch": 0.72,
281
+ "learning_rate": 1.1020177413231334e-06,
282
+ "logits/chosen": -1.9178645610809326,
283
+ "logits/rejected": -1.8847808837890625,
284
+ "logps/chosen": -10.10865306854248,
285
+ "logps/rejected": -3.267775297164917,
286
+ "loss": 0.0015,
287
+ "rewards/accuracies": 0.02500000037252903,
288
+ "rewards/chosen": -0.0017203291645273566,
289
+ "rewards/margins": 0.0016426773509010673,
290
+ "rewards/rejected": -0.003363006515428424,
291
+ "step": 180
292
+ },
293
+ {
294
+ "epoch": 0.76,
295
+ "learning_rate": 8.271734841028553e-07,
296
+ "logits/chosen": -1.8469321727752686,
297
+ "logits/rejected": -1.8469321727752686,
298
+ "logps/chosen": 0.0,
299
+ "logps/rejected": 0.0,
300
+ "loss": 0.0021,
301
+ "rewards/accuracies": 0.0,
302
+ "rewards/chosen": 0.0,
303
+ "rewards/margins": 0.0,
304
+ "rewards/rejected": 0.0,
305
+ "step": 190
306
+ },
307
+ {
308
+ "epoch": 0.8,
309
+ "learning_rate": 5.848888922025553e-07,
310
+ "logits/chosen": -1.9015337228775024,
311
+ "logits/rejected": -1.9031871557235718,
312
+ "logps/chosen": -5.545676231384277,
313
+ "logps/rejected": -6.33315372467041,
314
+ "loss": 0.0013,
315
+ "rewards/accuracies": 0.0,
316
+ "rewards/chosen": 0.0002652378170751035,
317
+ "rewards/margins": -0.0003098316374234855,
318
+ "rewards/rejected": 0.000575069454498589,
319
+ "step": 200
320
+ },
321
+ {
322
+ "epoch": 0.8,
323
+ "eval_logits/chosen": -1.969506859779358,
324
+ "eval_logits/rejected": -1.813860297203064,
325
+ "eval_logps/chosen": -266.95574951171875,
326
+ "eval_logps/rejected": -254.3690948486328,
327
+ "eval_loss": 0.057529229670763016,
328
+ "eval_rewards/accuracies": 0.3799999952316284,
329
+ "eval_rewards/chosen": -0.001290707616135478,
330
+ "eval_rewards/margins": -0.004144246224313974,
331
+ "eval_rewards/rejected": 0.00285353884100914,
332
+ "eval_runtime": 704.8428,
333
+ "eval_samples_per_second": 2.838,
334
+ "eval_steps_per_second": 1.419,
335
+ "step": 200
336
+ },
337
+ {
338
+ "epoch": 0.84,
339
+ "learning_rate": 3.798797596089351e-07,
340
+ "logits/chosen": -1.8618186712265015,
341
+ "logits/rejected": -1.8618186712265015,
342
+ "logps/chosen": 0.0,
343
+ "logps/rejected": 0.0,
344
+ "loss": 0.0012,
345
+ "rewards/accuracies": 0.0,
346
+ "rewards/chosen": 0.0,
347
+ "rewards/margins": 0.0,
348
+ "rewards/rejected": 0.0,
349
+ "step": 210
350
+ },
351
+ {
352
+ "epoch": 0.88,
353
+ "learning_rate": 2.1613635589349756e-07,
354
+ "logits/chosen": -2.0101757049560547,
355
+ "logits/rejected": -2.0101757049560547,
356
+ "logps/chosen": 0.0,
357
+ "logps/rejected": 0.0,
358
+ "loss": 0.0018,
359
+ "rewards/accuracies": 0.0,
360
+ "rewards/chosen": 0.0,
361
+ "rewards/margins": 0.0,
362
+ "rewards/rejected": 0.0,
363
+ "step": 220
364
+ },
365
+ {
366
+ "epoch": 0.92,
367
+ "learning_rate": 9.684576015420277e-08,
368
+ "logits/chosen": -1.938612699508667,
369
+ "logits/rejected": -1.9380161762237549,
370
+ "logps/chosen": -1.706319808959961,
371
+ "logps/rejected": -1.771810531616211,
372
+ "loss": 0.0022,
373
+ "rewards/accuracies": 0.0,
374
+ "rewards/chosen": -0.0014353947481140494,
375
+ "rewards/margins": -0.0004587741568684578,
376
+ "rewards/rejected": -0.0009766205912455916,
377
+ "step": 230
378
+ },
379
+ {
380
+ "epoch": 0.96,
381
+ "learning_rate": 2.4329828146074096e-08,
382
+ "logits/chosen": -2.1376256942749023,
383
+ "logits/rejected": -2.1376256942749023,
384
+ "logps/chosen": 0.0,
385
+ "logps/rejected": 0.0,
386
+ "loss": 0.0016,
387
+ "rewards/accuracies": 0.0,
388
+ "rewards/chosen": 0.0,
389
+ "rewards/margins": 0.0,
390
+ "rewards/rejected": 0.0,
391
+ "step": 240
392
+ },
393
+ {
394
+ "epoch": 1.0,
395
+ "learning_rate": 0.0,
396
+ "logits/chosen": -1.9790098667144775,
397
+ "logits/rejected": -1.9790098667144775,
398
+ "logps/chosen": 0.0,
399
+ "logps/rejected": 0.0,
400
+ "loss": 0.0018,
401
+ "rewards/accuracies": 0.0,
402
+ "rewards/chosen": 0.0,
403
+ "rewards/margins": 0.0,
404
+ "rewards/rejected": 0.0,
405
+ "step": 250
406
+ },
407
+ {
408
+ "epoch": 1.0,
409
+ "step": 250,
410
+ "total_flos": 0.0,
411
+ "train_loss": 0.0016641309279948472,
412
+ "train_runtime": 2302.1859,
413
+ "train_samples_per_second": 0.434,
414
+ "train_steps_per_second": 0.109
415
+ }
416
+ ],
417
+ "logging_steps": 10,
418
+ "max_steps": 250,
419
+ "num_input_tokens_seen": 0,
420
+ "num_train_epochs": 1,
421
+ "save_steps": 100,
422
+ "total_flos": 0.0,
423
+ "train_batch_size": 2,
424
+ "trial_name": null,
425
+ "trial_params": null
426
+ }