christineyu commited on
Commit
31fb9cb
·
verified ·
1 Parent(s): 6a7f11e

Model save

Browse files
Files changed (4) hide show
  1. README.md +57 -0
  2. all_results.json +9 -0
  3. train_results.json +9 -0
  4. trainer_state.json +933 -0
README.md ADDED
@@ -0,0 +1,57 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ base_model: mistralai/Mistral-7B-Instruct-v0.2
3
+ library_name: transformers
4
+ model_name: prometheus-7b-v1.5-beta-1-over1212-v2
5
+ tags:
6
+ - generated_from_trainer
7
+ - trl
8
+ - sft
9
+ licence: license
10
+ ---
11
+
12
+ # Model Card for prometheus-7b-v1.5-beta-1-over1212-v2
13
+
14
+ This model is a fine-tuned version of [mistralai/Mistral-7B-Instruct-v0.2](https://huggingface.co/mistralai/Mistral-7B-Instruct-v0.2).
15
+ It has been trained using [TRL](https://github.com/huggingface/trl).
16
+
17
+ ## Quick start
18
+
19
+ ```python
20
+ from transformers import pipeline
21
+
22
+ question = "If you had a time machine, but could only go to the past or the future once and never return, which would you choose and why?"
23
+ generator = pipeline("text-generation", model="christineyu/prometheus-7b-v1.5-beta-1-over1212-v2", device="cuda")
24
+ output = generator([{"role": "user", "content": question}], max_new_tokens=128, return_full_text=False)[0]
25
+ print(output["generated_text"])
26
+ ```
27
+
28
+ ## Training procedure
29
+
30
+ [<img src="https://raw.githubusercontent.com/wandb/assets/main/wandb-github-badge-28.svg" alt="Visualize in Weights & Biases" width="150" height="24"/>](https://wandb.ai/myexp/huggingface/runs/hhbwr73y)
31
+
32
+ This model was trained with SFT.
33
+
34
+ ### Framework versions
35
+
36
+ - TRL: 0.12.2
37
+ - Transformers: 4.46.3
38
+ - Pytorch: 2.3.0
39
+ - Datasets: 3.2.0
40
+ - Tokenizers: 0.20.3
41
+
42
+ ## Citations
43
+
44
+
45
+
46
+ Cite TRL as:
47
+
48
+ ```bibtex
49
+ @misc{vonwerra2022trl,
50
+ title = {{TRL: Transformer Reinforcement Learning}},
51
+ author = {Leandro von Werra and Younes Belkada and Lewis Tunstall and Edward Beeching and Tristan Thrush and Nathan Lambert and Shengyi Huang and Kashif Rasul and Quentin Gallouédec},
52
+ year = 2020,
53
+ journal = {GitHub repository},
54
+ publisher = {GitHub},
55
+ howpublished = {\url{https://github.com/huggingface/trl}}
56
+ }
57
+ ```
all_results.json ADDED
@@ -0,0 +1,9 @@
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 2.0,
3
+ "total_flos": 1.1181005894949274e+17,
4
+ "train_loss": 0.03632537112270873,
5
+ "train_runtime": 2624.7531,
6
+ "train_samples": 990,
7
+ "train_samples_per_second": 0.236,
8
+ "train_steps_per_second": 0.236
9
+ }
train_results.json ADDED
@@ -0,0 +1,9 @@
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 2.0,
3
+ "total_flos": 1.1181005894949274e+17,
4
+ "train_loss": 0.03632537112270873,
5
+ "train_runtime": 2624.7531,
6
+ "train_samples": 990,
7
+ "train_samples_per_second": 0.236,
8
+ "train_steps_per_second": 0.236
9
+ }
trainer_state.json ADDED
@@ -0,0 +1,933 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": null,
3
+ "best_model_checkpoint": null,
4
+ "epoch": 2.0,
5
+ "eval_steps": 500,
6
+ "global_step": 620,
7
+ "is_hyper_param_search": false,
8
+ "is_local_process_zero": true,
9
+ "is_world_process_zero": true,
10
+ "log_history": [
11
+ {
12
+ "epoch": 0.0032258064516129032,
13
+ "grad_norm": 3.971249580383301,
14
+ "learning_rate": 1.6129032258064518e-07,
15
+ "loss": 0.6284,
16
+ "step": 1
17
+ },
18
+ {
19
+ "epoch": 0.016129032258064516,
20
+ "grad_norm": 5.593815326690674,
21
+ "learning_rate": 8.064516129032258e-07,
22
+ "loss": 0.6406,
23
+ "step": 5
24
+ },
25
+ {
26
+ "epoch": 0.03225806451612903,
27
+ "grad_norm": 3.338022232055664,
28
+ "learning_rate": 1.6129032258064516e-06,
29
+ "loss": 0.6179,
30
+ "step": 10
31
+ },
32
+ {
33
+ "epoch": 0.04838709677419355,
34
+ "grad_norm": 2.572110891342163,
35
+ "learning_rate": 2.4193548387096776e-06,
36
+ "loss": 0.5695,
37
+ "step": 15
38
+ },
39
+ {
40
+ "epoch": 0.06451612903225806,
41
+ "grad_norm": 2.096151113510132,
42
+ "learning_rate": 3.225806451612903e-06,
43
+ "loss": 0.5072,
44
+ "step": 20
45
+ },
46
+ {
47
+ "epoch": 0.08064516129032258,
48
+ "grad_norm": 2.135874032974243,
49
+ "learning_rate": 4.032258064516129e-06,
50
+ "loss": 0.4539,
51
+ "step": 25
52
+ },
53
+ {
54
+ "epoch": 0.0967741935483871,
55
+ "grad_norm": 1.8665850162506104,
56
+ "learning_rate": 4.838709677419355e-06,
57
+ "loss": 0.394,
58
+ "step": 30
59
+ },
60
+ {
61
+ "epoch": 0.11290322580645161,
62
+ "grad_norm": 1.8506455421447754,
63
+ "learning_rate": 5.645161290322582e-06,
64
+ "loss": 0.3099,
65
+ "step": 35
66
+ },
67
+ {
68
+ "epoch": 0.12903225806451613,
69
+ "grad_norm": 1.882623314857483,
70
+ "learning_rate": 6.451612903225806e-06,
71
+ "loss": 0.2164,
72
+ "step": 40
73
+ },
74
+ {
75
+ "epoch": 0.14516129032258066,
76
+ "grad_norm": 2.4444594383239746,
77
+ "learning_rate": 7.258064516129033e-06,
78
+ "loss": 0.1262,
79
+ "step": 45
80
+ },
81
+ {
82
+ "epoch": 0.16129032258064516,
83
+ "grad_norm": 1.5979326963424683,
84
+ "learning_rate": 8.064516129032258e-06,
85
+ "loss": 0.068,
86
+ "step": 50
87
+ },
88
+ {
89
+ "epoch": 0.1774193548387097,
90
+ "grad_norm": 3.323439359664917,
91
+ "learning_rate": 8.870967741935484e-06,
92
+ "loss": 0.0337,
93
+ "step": 55
94
+ },
95
+ {
96
+ "epoch": 0.1935483870967742,
97
+ "grad_norm": 1.5633885860443115,
98
+ "learning_rate": 9.67741935483871e-06,
99
+ "loss": 0.0272,
100
+ "step": 60
101
+ },
102
+ {
103
+ "epoch": 0.20967741935483872,
104
+ "grad_norm": 1.1317998170852661,
105
+ "learning_rate": 9.99928681279855e-06,
106
+ "loss": 0.0208,
107
+ "step": 65
108
+ },
109
+ {
110
+ "epoch": 0.22580645161290322,
111
+ "grad_norm": 0.4592137932777405,
112
+ "learning_rate": 9.994929183335237e-06,
113
+ "loss": 0.0158,
114
+ "step": 70
115
+ },
116
+ {
117
+ "epoch": 0.24193548387096775,
118
+ "grad_norm": 0.7832735180854797,
119
+ "learning_rate": 9.986613588305435e-06,
120
+ "loss": 0.0145,
121
+ "step": 75
122
+ },
123
+ {
124
+ "epoch": 0.25806451612903225,
125
+ "grad_norm": 0.8761097192764282,
126
+ "learning_rate": 9.974346616959476e-06,
127
+ "loss": 0.0104,
128
+ "step": 80
129
+ },
130
+ {
131
+ "epoch": 0.27419354838709675,
132
+ "grad_norm": 0.9911714196205139,
133
+ "learning_rate": 9.95813798960538e-06,
134
+ "loss": 0.0115,
135
+ "step": 85
136
+ },
137
+ {
138
+ "epoch": 0.2903225806451613,
139
+ "grad_norm": 0.48290345072746277,
140
+ "learning_rate": 9.938000549906509e-06,
141
+ "loss": 0.0092,
142
+ "step": 90
143
+ },
144
+ {
145
+ "epoch": 0.3064516129032258,
146
+ "grad_norm": 0.8782544136047363,
147
+ "learning_rate": 9.913950254704291e-06,
148
+ "loss": 0.0167,
149
+ "step": 95
150
+ },
151
+ {
152
+ "epoch": 0.3225806451612903,
153
+ "grad_norm": 0.827684223651886,
154
+ "learning_rate": 9.88600616137407e-06,
155
+ "loss": 0.0087,
156
+ "step": 100
157
+ },
158
+ {
159
+ "epoch": 0.3387096774193548,
160
+ "grad_norm": 0.4351416826248169,
161
+ "learning_rate": 9.854190412724114e-06,
162
+ "loss": 0.0092,
163
+ "step": 105
164
+ },
165
+ {
166
+ "epoch": 0.3548387096774194,
167
+ "grad_norm": 0.617416262626648,
168
+ "learning_rate": 9.818528219449705e-06,
169
+ "loss": 0.0106,
170
+ "step": 110
171
+ },
172
+ {
173
+ "epoch": 0.3709677419354839,
174
+ "grad_norm": 0.3410218060016632,
175
+ "learning_rate": 9.779047840156288e-06,
176
+ "loss": 0.0092,
177
+ "step": 115
178
+ },
179
+ {
180
+ "epoch": 0.3870967741935484,
181
+ "grad_norm": 0.6016420722007751,
182
+ "learning_rate": 9.735780558967434e-06,
183
+ "loss": 0.0081,
184
+ "step": 120
185
+ },
186
+ {
187
+ "epoch": 0.4032258064516129,
188
+ "grad_norm": 0.34005704522132874,
189
+ "learning_rate": 9.688760660735403e-06,
190
+ "loss": 0.0075,
191
+ "step": 125
192
+ },
193
+ {
194
+ "epoch": 0.41935483870967744,
195
+ "grad_norm": 1.3894850015640259,
196
+ "learning_rate": 9.638025403873939e-06,
197
+ "loss": 0.0074,
198
+ "step": 130
199
+ },
200
+ {
201
+ "epoch": 0.43548387096774194,
202
+ "grad_norm": 0.3171682059764862,
203
+ "learning_rate": 9.58361499083483e-06,
204
+ "loss": 0.007,
205
+ "step": 135
206
+ },
207
+ {
208
+ "epoch": 0.45161290322580644,
209
+ "grad_norm": 0.5876194834709167,
210
+ "learning_rate": 9.525572536251608e-06,
211
+ "loss": 0.0085,
212
+ "step": 140
213
+ },
214
+ {
215
+ "epoch": 0.46774193548387094,
216
+ "grad_norm": 0.47260814905166626,
217
+ "learning_rate": 9.46394403277566e-06,
218
+ "loss": 0.0067,
219
+ "step": 145
220
+ },
221
+ {
222
+ "epoch": 0.4838709677419355,
223
+ "grad_norm": 0.6115548014640808,
224
+ "learning_rate": 9.398778314631801e-06,
225
+ "loss": 0.0084,
226
+ "step": 150
227
+ },
228
+ {
229
+ "epoch": 0.5,
230
+ "grad_norm": 0.44270601868629456,
231
+ "learning_rate": 9.330127018922195e-06,
232
+ "loss": 0.0063,
233
+ "step": 155
234
+ },
235
+ {
236
+ "epoch": 0.5161290322580645,
237
+ "grad_norm": 0.5065975785255432,
238
+ "learning_rate": 9.258044544709276e-06,
239
+ "loss": 0.0079,
240
+ "step": 160
241
+ },
242
+ {
243
+ "epoch": 0.532258064516129,
244
+ "grad_norm": 0.9441617131233215,
245
+ "learning_rate": 9.182588009910119e-06,
246
+ "loss": 0.0075,
247
+ "step": 165
248
+ },
249
+ {
250
+ "epoch": 0.5483870967741935,
251
+ "grad_norm": 0.7031365036964417,
252
+ "learning_rate": 9.103817206036383e-06,
253
+ "loss": 0.0067,
254
+ "step": 170
255
+ },
256
+ {
257
+ "epoch": 0.5645161290322581,
258
+ "grad_norm": 0.6422185301780701,
259
+ "learning_rate": 9.021794550815713e-06,
260
+ "loss": 0.0088,
261
+ "step": 175
262
+ },
263
+ {
264
+ "epoch": 0.5806451612903226,
265
+ "grad_norm": 0.547900378704071,
266
+ "learning_rate": 8.936585038732143e-06,
267
+ "loss": 0.006,
268
+ "step": 180
269
+ },
270
+ {
271
+ "epoch": 0.5967741935483871,
272
+ "grad_norm": 0.6972706913948059,
273
+ "learning_rate": 8.848256189524661e-06,
274
+ "loss": 0.0057,
275
+ "step": 185
276
+ },
277
+ {
278
+ "epoch": 0.6129032258064516,
279
+ "grad_norm": 0.37197422981262207,
280
+ "learning_rate": 8.756877994684818e-06,
281
+ "loss": 0.0049,
282
+ "step": 190
283
+ },
284
+ {
285
+ "epoch": 0.6290322580645161,
286
+ "grad_norm": 0.2147480994462967,
287
+ "learning_rate": 8.66252286199567e-06,
288
+ "loss": 0.0081,
289
+ "step": 195
290
+ },
291
+ {
292
+ "epoch": 0.6451612903225806,
293
+ "grad_norm": 0.44903531670570374,
294
+ "learning_rate": 8.565265558156101e-06,
295
+ "loss": 0.0055,
296
+ "step": 200
297
+ },
298
+ {
299
+ "epoch": 0.6612903225806451,
300
+ "grad_norm": 1.0128227472305298,
301
+ "learning_rate": 8.465183149535939e-06,
302
+ "loss": 0.0054,
303
+ "step": 205
304
+ },
305
+ {
306
+ "epoch": 0.6774193548387096,
307
+ "grad_norm": 0.4877210557460785,
308
+ "learning_rate": 8.362354941108803e-06,
309
+ "loss": 0.0057,
310
+ "step": 210
311
+ },
312
+ {
313
+ "epoch": 0.6935483870967742,
314
+ "grad_norm": 0.2644139230251312,
315
+ "learning_rate": 8.256862413611113e-06,
316
+ "loss": 0.005,
317
+ "step": 215
318
+ },
319
+ {
320
+ "epoch": 0.7096774193548387,
321
+ "grad_norm": 0.6257811784744263,
322
+ "learning_rate": 8.148789158977012e-06,
323
+ "loss": 0.0055,
324
+ "step": 220
325
+ },
326
+ {
327
+ "epoch": 0.7258064516129032,
328
+ "grad_norm": 0.19504040479660034,
329
+ "learning_rate": 8.038220814100403e-06,
330
+ "loss": 0.005,
331
+ "step": 225
332
+ },
333
+ {
334
+ "epoch": 0.7419354838709677,
335
+ "grad_norm": 0.5125346779823303,
336
+ "learning_rate": 7.925244992976538e-06,
337
+ "loss": 0.0055,
338
+ "step": 230
339
+ },
340
+ {
341
+ "epoch": 0.7580645161290323,
342
+ "grad_norm": 0.46777665615081787,
343
+ "learning_rate": 7.809951217276986e-06,
344
+ "loss": 0.0052,
345
+ "step": 235
346
+ },
347
+ {
348
+ "epoch": 0.7741935483870968,
349
+ "grad_norm": 0.5273892283439636,
350
+ "learning_rate": 7.692430845412946e-06,
351
+ "loss": 0.006,
352
+ "step": 240
353
+ },
354
+ {
355
+ "epoch": 0.7903225806451613,
356
+ "grad_norm": 0.5478794574737549,
357
+ "learning_rate": 7.572777000143145e-06,
358
+ "loss": 0.0078,
359
+ "step": 245
360
+ },
361
+ {
362
+ "epoch": 0.8064516129032258,
363
+ "grad_norm": 0.35614484548568726,
364
+ "learning_rate": 7.451084494783668e-06,
365
+ "loss": 0.0051,
366
+ "step": 250
367
+ },
368
+ {
369
+ "epoch": 0.8225806451612904,
370
+ "grad_norm": 0.42776939272880554,
371
+ "learning_rate": 7.327449758078194e-06,
372
+ "loss": 0.0057,
373
+ "step": 255
374
+ },
375
+ {
376
+ "epoch": 0.8387096774193549,
377
+ "grad_norm": 0.19282685220241547,
378
+ "learning_rate": 7.201970757788172e-06,
379
+ "loss": 0.0039,
380
+ "step": 260
381
+ },
382
+ {
383
+ "epoch": 0.8548387096774194,
384
+ "grad_norm": 0.7167965173721313,
385
+ "learning_rate": 7.074746923063497e-06,
386
+ "loss": 0.0042,
387
+ "step": 265
388
+ },
389
+ {
390
+ "epoch": 0.8709677419354839,
391
+ "grad_norm": 3.5436365604400635,
392
+ "learning_rate": 6.945879065655164e-06,
393
+ "loss": 0.0042,
394
+ "step": 270
395
+ },
396
+ {
397
+ "epoch": 0.8870967741935484,
398
+ "grad_norm": 0.3570103347301483,
399
+ "learning_rate": 6.815469300032374e-06,
400
+ "loss": 0.0043,
401
+ "step": 275
402
+ },
403
+ {
404
+ "epoch": 0.9032258064516129,
405
+ "grad_norm": 0.252421110868454,
406
+ "learning_rate": 6.6836209624673575e-06,
407
+ "loss": 0.0056,
408
+ "step": 280
409
+ },
410
+ {
411
+ "epoch": 0.9193548387096774,
412
+ "grad_norm": 0.2856729030609131,
413
+ "learning_rate": 6.5504385291520554e-06,
414
+ "loss": 0.0052,
415
+ "step": 285
416
+ },
417
+ {
418
+ "epoch": 0.9354838709677419,
419
+ "grad_norm": 0.25228211283683777,
420
+ "learning_rate": 6.41602753341152e-06,
421
+ "loss": 0.0055,
422
+ "step": 290
423
+ },
424
+ {
425
+ "epoch": 0.9516129032258065,
426
+ "grad_norm": 0.29258614778518677,
427
+ "learning_rate": 6.2804944820796596e-06,
428
+ "loss": 0.0037,
429
+ "step": 295
430
+ },
431
+ {
432
+ "epoch": 0.967741935483871,
433
+ "grad_norm": 0.1690083146095276,
434
+ "learning_rate": 6.143946771103561e-06,
435
+ "loss": 0.0032,
436
+ "step": 300
437
+ },
438
+ {
439
+ "epoch": 0.9838709677419355,
440
+ "grad_norm": 0.45113542675971985,
441
+ "learning_rate": 6.006492600443301e-06,
442
+ "loss": 0.0036,
443
+ "step": 305
444
+ },
445
+ {
446
+ "epoch": 1.0,
447
+ "grad_norm": 0.34520477056503296,
448
+ "learning_rate": 5.8682408883346535e-06,
449
+ "loss": 0.0049,
450
+ "step": 310
451
+ },
452
+ {
453
+ "epoch": 1.0,
454
+ "eval_loss": 0.004980116616934538,
455
+ "eval_runtime": 3.8044,
456
+ "eval_samples_per_second": 0.789,
457
+ "eval_steps_per_second": 0.789,
458
+ "step": 310
459
+ },
460
+ {
461
+ "epoch": 1.0161290322580645,
462
+ "grad_norm": 0.19288980960845947,
463
+ "learning_rate": 5.729301184982622e-06,
464
+ "loss": 0.0034,
465
+ "step": 315
466
+ },
467
+ {
468
+ "epoch": 1.032258064516129,
469
+ "grad_norm": 0.19823557138442993,
470
+ "learning_rate": 5.5897835857542315e-06,
471
+ "loss": 0.0026,
472
+ "step": 320
473
+ },
474
+ {
475
+ "epoch": 1.0483870967741935,
476
+ "grad_norm": 0.09697025269269943,
477
+ "learning_rate": 5.449798643939305e-06,
478
+ "loss": 0.0024,
479
+ "step": 325
480
+ },
481
+ {
482
+ "epoch": 1.064516129032258,
483
+ "grad_norm": 0.23041202127933502,
484
+ "learning_rate": 5.30945728314841e-06,
485
+ "loss": 0.0028,
486
+ "step": 330
487
+ },
488
+ {
489
+ "epoch": 1.0806451612903225,
490
+ "grad_norm": 0.27025359869003296,
491
+ "learning_rate": 5.168870709417342e-06,
492
+ "loss": 0.0034,
493
+ "step": 335
494
+ },
495
+ {
496
+ "epoch": 1.096774193548387,
497
+ "grad_norm": 0.3000248968601227,
498
+ "learning_rate": 5.0281503230878304e-06,
499
+ "loss": 0.0033,
500
+ "step": 340
501
+ },
502
+ {
503
+ "epoch": 1.1129032258064515,
504
+ "grad_norm": 0.22583173215389252,
505
+ "learning_rate": 4.887407630534271e-06,
506
+ "loss": 0.0027,
507
+ "step": 345
508
+ },
509
+ {
510
+ "epoch": 1.129032258064516,
511
+ "grad_norm": 0.21011537313461304,
512
+ "learning_rate": 4.746754155806437e-06,
513
+ "loss": 0.0026,
514
+ "step": 350
515
+ },
516
+ {
517
+ "epoch": 1.1451612903225807,
518
+ "grad_norm": 0.13884232938289642,
519
+ "learning_rate": 4.606301352258192e-06,
520
+ "loss": 0.0028,
521
+ "step": 355
522
+ },
523
+ {
524
+ "epoch": 1.1612903225806452,
525
+ "grad_norm": 0.25201186537742615,
526
+ "learning_rate": 4.466160514232206e-06,
527
+ "loss": 0.0027,
528
+ "step": 360
529
+ },
530
+ {
531
+ "epoch": 1.1774193548387097,
532
+ "grad_norm": 0.22268442809581757,
533
+ "learning_rate": 4.326442688870697e-06,
534
+ "loss": 0.0028,
535
+ "step": 365
536
+ },
537
+ {
538
+ "epoch": 1.1935483870967742,
539
+ "grad_norm": 0.21970783174037933,
540
+ "learning_rate": 4.187258588122019e-06,
541
+ "loss": 0.0031,
542
+ "step": 370
543
+ },
544
+ {
545
+ "epoch": 1.2096774193548387,
546
+ "grad_norm": 0.26229333877563477,
547
+ "learning_rate": 4.048718501012895e-06,
548
+ "loss": 0.0026,
549
+ "step": 375
550
+ },
551
+ {
552
+ "epoch": 1.2258064516129032,
553
+ "grad_norm": 0.25772520899772644,
554
+ "learning_rate": 3.910932206255742e-06,
555
+ "loss": 0.0022,
556
+ "step": 380
557
+ },
558
+ {
559
+ "epoch": 1.2419354838709677,
560
+ "grad_norm": 0.4050711691379547,
561
+ "learning_rate": 3.77400888526038e-06,
562
+ "loss": 0.0023,
563
+ "step": 385
564
+ },
565
+ {
566
+ "epoch": 1.2580645161290323,
567
+ "grad_norm": 0.2486872673034668,
568
+ "learning_rate": 3.6380570356190346e-06,
569
+ "loss": 0.0036,
570
+ "step": 390
571
+ },
572
+ {
573
+ "epoch": 1.2741935483870968,
574
+ "grad_norm": 0.20115166902542114,
575
+ "learning_rate": 3.5031843851332105e-06,
576
+ "loss": 0.0017,
577
+ "step": 395
578
+ },
579
+ {
580
+ "epoch": 1.2903225806451613,
581
+ "grad_norm": 0.19394518435001373,
582
+ "learning_rate": 3.3694978064505258e-06,
583
+ "loss": 0.0029,
584
+ "step": 400
585
+ },
586
+ {
587
+ "epoch": 1.3064516129032258,
588
+ "grad_norm": 0.25958162546157837,
589
+ "learning_rate": 3.2371032323791757e-06,
590
+ "loss": 0.003,
591
+ "step": 405
592
+ },
593
+ {
594
+ "epoch": 1.3225806451612903,
595
+ "grad_norm": 0.27680277824401855,
596
+ "learning_rate": 3.10610557194712e-06,
597
+ "loss": 0.002,
598
+ "step": 410
599
+ },
600
+ {
601
+ "epoch": 1.3387096774193548,
602
+ "grad_norm": 0.31953132152557373,
603
+ "learning_rate": 2.97660862727252e-06,
604
+ "loss": 0.0039,
605
+ "step": 415
606
+ },
607
+ {
608
+ "epoch": 1.3548387096774195,
609
+ "grad_norm": 0.3594481647014618,
610
+ "learning_rate": 2.848715011311271e-06,
611
+ "loss": 0.0031,
612
+ "step": 420
613
+ },
614
+ {
615
+ "epoch": 1.370967741935484,
616
+ "grad_norm": 0.15407758951187134,
617
+ "learning_rate": 2.72252606654683e-06,
618
+ "loss": 0.0032,
619
+ "step": 425
620
+ },
621
+ {
622
+ "epoch": 1.3870967741935485,
623
+ "grad_norm": 0.25268280506134033,
624
+ "learning_rate": 2.5981417846867753e-06,
625
+ "loss": 0.0029,
626
+ "step": 430
627
+ },
628
+ {
629
+ "epoch": 1.403225806451613,
630
+ "grad_norm": 0.19317582249641418,
631
+ "learning_rate": 2.4756607274296844e-06,
632
+ "loss": 0.0035,
633
+ "step": 435
634
+ },
635
+ {
636
+ "epoch": 1.4193548387096775,
637
+ "grad_norm": 0.15693552792072296,
638
+ "learning_rate": 2.3551799483651894e-06,
639
+ "loss": 0.0026,
640
+ "step": 440
641
+ },
642
+ {
643
+ "epoch": 1.435483870967742,
644
+ "grad_norm": 0.19608916342258453,
645
+ "learning_rate": 2.236794916069007e-06,
646
+ "loss": 0.0028,
647
+ "step": 445
648
+ },
649
+ {
650
+ "epoch": 1.4516129032258065,
651
+ "grad_norm": 0.403870552778244,
652
+ "learning_rate": 2.120599438453968e-06,
653
+ "loss": 0.0032,
654
+ "step": 450
655
+ },
656
+ {
657
+ "epoch": 1.467741935483871,
658
+ "grad_norm": 0.16400307416915894,
659
+ "learning_rate": 2.0066855884369246e-06,
660
+ "loss": 0.0025,
661
+ "step": 455
662
+ },
663
+ {
664
+ "epoch": 1.4838709677419355,
665
+ "grad_norm": 0.954981803894043,
666
+ "learning_rate": 1.8951436309804766e-06,
667
+ "loss": 0.0024,
668
+ "step": 460
669
+ },
670
+ {
671
+ "epoch": 1.5,
672
+ "grad_norm": 0.23671939969062805,
673
+ "learning_rate": 1.7860619515673034e-06,
674
+ "loss": 0.0029,
675
+ "step": 465
676
+ },
677
+ {
678
+ "epoch": 1.5161290322580645,
679
+ "grad_norm": 0.1934683918952942,
680
+ "learning_rate": 1.6795269861638041e-06,
681
+ "loss": 0.0033,
682
+ "step": 470
683
+ },
684
+ {
685
+ "epoch": 1.532258064516129,
686
+ "grad_norm": 0.12341190874576569,
687
+ "learning_rate": 1.5756231527285181e-06,
688
+ "loss": 0.0027,
689
+ "step": 475
690
+ },
691
+ {
692
+ "epoch": 1.5483870967741935,
693
+ "grad_norm": 0.44817498326301575,
694
+ "learning_rate": 1.4744327843196043e-06,
695
+ "loss": 0.0034,
696
+ "step": 480
697
+ },
698
+ {
699
+ "epoch": 1.564516129032258,
700
+ "grad_norm": 0.29806482791900635,
701
+ "learning_rate": 1.3760360638544012e-06,
702
+ "loss": 0.0031,
703
+ "step": 485
704
+ },
705
+ {
706
+ "epoch": 1.5806451612903225,
707
+ "grad_norm": 0.19129404425621033,
708
+ "learning_rate": 1.280510960572745e-06,
709
+ "loss": 0.0017,
710
+ "step": 490
711
+ },
712
+ {
713
+ "epoch": 1.596774193548387,
714
+ "grad_norm": 0.21360230445861816,
715
+ "learning_rate": 1.1879331682543972e-06,
716
+ "loss": 0.0034,
717
+ "step": 495
718
+ },
719
+ {
720
+ "epoch": 1.6129032258064515,
721
+ "grad_norm": 0.2776956260204315,
722
+ "learning_rate": 1.0983760452395415e-06,
723
+ "loss": 0.0021,
724
+ "step": 500
725
+ },
726
+ {
727
+ "epoch": 1.629032258064516,
728
+ "grad_norm": 0.15755566954612732,
729
+ "learning_rate": 1.01191055629987e-06,
730
+ "loss": 0.0019,
731
+ "step": 505
732
+ },
733
+ {
734
+ "epoch": 1.6451612903225805,
735
+ "grad_norm": 0.3595326840877533,
736
+ "learning_rate": 9.286052164063369e-07,
737
+ "loss": 0.0023,
738
+ "step": 510
739
+ },
740
+ {
741
+ "epoch": 1.661290322580645,
742
+ "grad_norm": 0.19125986099243164,
743
+ "learning_rate": 8.485260364381187e-07,
744
+ "loss": 0.0041,
745
+ "step": 515
746
+ },
747
+ {
748
+ "epoch": 1.6774193548387095,
749
+ "grad_norm": 0.3362486958503723,
750
+ "learning_rate": 7.717364708758024e-07,
751
+ "loss": 0.002,
752
+ "step": 520
753
+ },
754
+ {
755
+ "epoch": 1.6935483870967742,
756
+ "grad_norm": 0.23303864896297455,
757
+ "learning_rate": 6.982973675202676e-07,
758
+ "loss": 0.0017,
759
+ "step": 525
760
+ },
761
+ {
762
+ "epoch": 1.7096774193548387,
763
+ "grad_norm": 0.13133780658245087,
764
+ "learning_rate": 6.282669192770896e-07,
765
+ "loss": 0.0024,
766
+ "step": 530
767
+ },
768
+ {
769
+ "epoch": 1.7258064516129032,
770
+ "grad_norm": 0.35981041193008423,
771
+ "learning_rate": 5.617006180446688e-07,
772
+ "loss": 0.0033,
773
+ "step": 535
774
+ },
775
+ {
776
+ "epoch": 1.7419354838709677,
777
+ "grad_norm": 0.41394540667533875,
778
+ "learning_rate": 4.986512107426283e-07,
779
+ "loss": 0.003,
780
+ "step": 540
781
+ },
782
+ {
783
+ "epoch": 1.7580645161290323,
784
+ "grad_norm": 0.24306242167949677,
785
+ "learning_rate": 4.3916865751533313e-07,
786
+ "loss": 0.003,
787
+ "step": 545
788
+ },
789
+ {
790
+ "epoch": 1.7741935483870968,
791
+ "grad_norm": 0.1294822096824646,
792
+ "learning_rate": 3.8330009214363197e-07,
793
+ "loss": 0.0028,
794
+ "step": 550
795
+ },
796
+ {
797
+ "epoch": 1.7903225806451613,
798
+ "grad_norm": 0.12696751952171326,
799
+ "learning_rate": 3.310897846962041e-07,
800
+ "loss": 0.0027,
801
+ "step": 555
802
+ },
803
+ {
804
+ "epoch": 1.8064516129032258,
805
+ "grad_norm": 0.27052465081214905,
806
+ "learning_rate": 2.8257910645009935e-07,
807
+ "loss": 0.0034,
808
+ "step": 560
809
+ },
810
+ {
811
+ "epoch": 1.8225806451612905,
812
+ "grad_norm": 0.21007302403450012,
813
+ "learning_rate": 2.3780649710827552e-07,
814
+ "loss": 0.0029,
815
+ "step": 565
816
+ },
817
+ {
818
+ "epoch": 1.838709677419355,
819
+ "grad_norm": 0.20920442044734955,
820
+ "learning_rate": 1.9680743434010385e-07,
821
+ "loss": 0.0026,
822
+ "step": 570
823
+ },
824
+ {
825
+ "epoch": 1.8548387096774195,
826
+ "grad_norm": 0.2021329551935196,
827
+ "learning_rate": 1.5961440566897913e-07,
828
+ "loss": 0.0026,
829
+ "step": 575
830
+ },
831
+ {
832
+ "epoch": 1.870967741935484,
833
+ "grad_norm": 0.23637109994888306,
834
+ "learning_rate": 1.2625688272930925e-07,
835
+ "loss": 0.0036,
836
+ "step": 580
837
+ },
838
+ {
839
+ "epoch": 1.8870967741935485,
840
+ "grad_norm": 0.17650996148586273,
841
+ "learning_rate": 9.676129791329481e-08,
842
+ "loss": 0.0024,
843
+ "step": 585
844
+ },
845
+ {
846
+ "epoch": 1.903225806451613,
847
+ "grad_norm": 0.19747471809387207,
848
+ "learning_rate": 7.115102342598101e-08,
849
+ "loss": 0.0035,
850
+ "step": 590
851
+ },
852
+ {
853
+ "epoch": 1.9193548387096775,
854
+ "grad_norm": 0.6039708852767944,
855
+ "learning_rate": 4.944635276520393e-08,
856
+ "loss": 0.0029,
857
+ "step": 595
858
+ },
859
+ {
860
+ "epoch": 1.935483870967742,
861
+ "grad_norm": 0.3589779734611511,
862
+ "learning_rate": 3.166448464108629e-08,
863
+ "loss": 0.0049,
864
+ "step": 600
865
+ },
866
+ {
867
+ "epoch": 1.9516129032258065,
868
+ "grad_norm": 0.12553617358207703,
869
+ "learning_rate": 1.781950934783505e-08,
870
+ "loss": 0.0027,
871
+ "step": 605
872
+ },
873
+ {
874
+ "epoch": 1.967741935483871,
875
+ "grad_norm": 0.1768651306629181,
876
+ "learning_rate": 7.922397598642551e-09,
877
+ "loss": 0.0018,
878
+ "step": 610
879
+ },
880
+ {
881
+ "epoch": 1.9838709677419355,
882
+ "grad_norm": 0.24496199190616608,
883
+ "learning_rate": 1.980991832524759e-09,
884
+ "loss": 0.0021,
885
+ "step": 615
886
+ },
887
+ {
888
+ "epoch": 2.0,
889
+ "grad_norm": 0.2647761106491089,
890
+ "learning_rate": 0.0,
891
+ "loss": 0.0026,
892
+ "step": 620
893
+ },
894
+ {
895
+ "epoch": 2.0,
896
+ "eval_loss": 0.004049910232424736,
897
+ "eval_runtime": 3.8097,
898
+ "eval_samples_per_second": 0.787,
899
+ "eval_steps_per_second": 0.787,
900
+ "step": 620
901
+ },
902
+ {
903
+ "epoch": 2.0,
904
+ "step": 620,
905
+ "total_flos": 1.1181005894949274e+17,
906
+ "train_loss": 0.03632537112270873,
907
+ "train_runtime": 2624.7531,
908
+ "train_samples_per_second": 0.236,
909
+ "train_steps_per_second": 0.236
910
+ }
911
+ ],
912
+ "logging_steps": 5,
913
+ "max_steps": 620,
914
+ "num_input_tokens_seen": 0,
915
+ "num_train_epochs": 2,
916
+ "save_steps": 100,
917
+ "stateful_callbacks": {
918
+ "TrainerControl": {
919
+ "args": {
920
+ "should_epoch_stop": false,
921
+ "should_evaluate": false,
922
+ "should_log": false,
923
+ "should_save": true,
924
+ "should_training_stop": true
925
+ },
926
+ "attributes": {}
927
+ }
928
+ },
929
+ "total_flos": 1.1181005894949274e+17,
930
+ "train_batch_size": 1,
931
+ "trial_name": null,
932
+ "trial_params": null
933
+ }