dkqjrm commited on
Commit
ba3024b
1 Parent(s): 8229252

End of training

Browse files
Files changed (3) hide show
  1. all_results.json +7 -0
  2. train_results.json +7 -0
  3. trainer_state.json +1147 -0
all_results.json ADDED
@@ -0,0 +1,7 @@
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 60.0,
3
+ "train_loss": 0.17944894468682443,
4
+ "train_runtime": 4049.329,
5
+ "train_samples_per_second": 36.895,
6
+ "train_steps_per_second": 4.623
7
+ }
train_results.json ADDED
@@ -0,0 +1,7 @@
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 60.0,
3
+ "train_loss": 0.17944894468682443,
4
+ "train_runtime": 4049.329,
5
+ "train_samples_per_second": 36.895,
6
+ "train_steps_per_second": 4.623
7
+ }
trainer_state.json ADDED
@@ -0,0 +1,1147 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": null,
3
+ "best_model_checkpoint": null,
4
+ "epoch": 60.0,
5
+ "global_step": 18720,
6
+ "is_hyper_param_search": false,
7
+ "is_local_process_zero": true,
8
+ "is_world_process_zero": true,
9
+ "log_history": [
10
+ {
11
+ "epoch": 1.0,
12
+ "eval_accuracy": 0.5306859205776173,
13
+ "eval_loss": 0.17852023243904114,
14
+ "eval_runtime": 4.2873,
15
+ "eval_samples_per_second": 64.61,
16
+ "eval_steps_per_second": 8.164,
17
+ "step": 312
18
+ },
19
+ {
20
+ "best_epoch": 0,
21
+ "best_eval_accuracy": 0.5306859205776173,
22
+ "epoch": 1.0,
23
+ "step": 312
24
+ },
25
+ {
26
+ "epoch": 1.6,
27
+ "learning_rate": 0.002919871794871795,
28
+ "loss": 0.2552,
29
+ "step": 500
30
+ },
31
+ {
32
+ "epoch": 2.0,
33
+ "eval_accuracy": 0.5054151624548736,
34
+ "eval_loss": 0.18264690041542053,
35
+ "eval_runtime": 4.3896,
36
+ "eval_samples_per_second": 63.104,
37
+ "eval_steps_per_second": 7.973,
38
+ "step": 624
39
+ },
40
+ {
41
+ "best_epoch": 0,
42
+ "best_eval_accuracy": 0.5306859205776173,
43
+ "epoch": 2.0,
44
+ "step": 624
45
+ },
46
+ {
47
+ "epoch": 3.0,
48
+ "eval_accuracy": 0.4729241877256318,
49
+ "eval_loss": 0.33281320333480835,
50
+ "eval_runtime": 4.433,
51
+ "eval_samples_per_second": 62.486,
52
+ "eval_steps_per_second": 7.895,
53
+ "step": 936
54
+ },
55
+ {
56
+ "best_epoch": 0,
57
+ "best_eval_accuracy": 0.5306859205776173,
58
+ "epoch": 3.0,
59
+ "step": 936
60
+ },
61
+ {
62
+ "epoch": 3.21,
63
+ "learning_rate": 0.0028397435897435895,
64
+ "loss": 0.24,
65
+ "step": 1000
66
+ },
67
+ {
68
+ "epoch": 4.0,
69
+ "eval_accuracy": 0.4729241877256318,
70
+ "eval_loss": 0.2049672156572342,
71
+ "eval_runtime": 4.4533,
72
+ "eval_samples_per_second": 62.201,
73
+ "eval_steps_per_second": 7.859,
74
+ "step": 1248
75
+ },
76
+ {
77
+ "best_epoch": 0,
78
+ "best_eval_accuracy": 0.5306859205776173,
79
+ "epoch": 4.0,
80
+ "step": 1248
81
+ },
82
+ {
83
+ "epoch": 4.81,
84
+ "learning_rate": 0.0027596153846153847,
85
+ "loss": 0.2369,
86
+ "step": 1500
87
+ },
88
+ {
89
+ "epoch": 5.0,
90
+ "eval_accuracy": 0.6064981949458483,
91
+ "eval_loss": 0.17499716579914093,
92
+ "eval_runtime": 4.4615,
93
+ "eval_samples_per_second": 62.087,
94
+ "eval_steps_per_second": 7.845,
95
+ "step": 1560
96
+ },
97
+ {
98
+ "best_epoch": 4,
99
+ "best_eval_accuracy": 0.6064981949458483,
100
+ "epoch": 5.0,
101
+ "step": 1560
102
+ },
103
+ {
104
+ "epoch": 6.0,
105
+ "eval_accuracy": 0.47653429602888087,
106
+ "eval_loss": 0.17517220973968506,
107
+ "eval_runtime": 4.4386,
108
+ "eval_samples_per_second": 62.407,
109
+ "eval_steps_per_second": 7.885,
110
+ "step": 1872
111
+ },
112
+ {
113
+ "best_epoch": 4,
114
+ "best_eval_accuracy": 0.6064981949458483,
115
+ "epoch": 6.0,
116
+ "step": 1872
117
+ },
118
+ {
119
+ "epoch": 6.41,
120
+ "learning_rate": 0.00267948717948718,
121
+ "loss": 0.2199,
122
+ "step": 2000
123
+ },
124
+ {
125
+ "epoch": 7.0,
126
+ "eval_accuracy": 0.592057761732852,
127
+ "eval_loss": 0.17986096441745758,
128
+ "eval_runtime": 4.441,
129
+ "eval_samples_per_second": 62.373,
130
+ "eval_steps_per_second": 7.881,
131
+ "step": 2184
132
+ },
133
+ {
134
+ "best_epoch": 4,
135
+ "best_eval_accuracy": 0.6064981949458483,
136
+ "epoch": 7.0,
137
+ "step": 2184
138
+ },
139
+ {
140
+ "epoch": 8.0,
141
+ "eval_accuracy": 0.4729241877256318,
142
+ "eval_loss": 0.18959033489227295,
143
+ "eval_runtime": 4.4419,
144
+ "eval_samples_per_second": 62.361,
145
+ "eval_steps_per_second": 7.88,
146
+ "step": 2496
147
+ },
148
+ {
149
+ "best_epoch": 4,
150
+ "best_eval_accuracy": 0.6064981949458483,
151
+ "epoch": 8.0,
152
+ "step": 2496
153
+ },
154
+ {
155
+ "epoch": 8.01,
156
+ "learning_rate": 0.0025993589743589745,
157
+ "loss": 0.1955,
158
+ "step": 2500
159
+ },
160
+ {
161
+ "epoch": 9.0,
162
+ "eval_accuracy": 0.6245487364620939,
163
+ "eval_loss": 0.1726665496826172,
164
+ "eval_runtime": 4.4446,
165
+ "eval_samples_per_second": 62.322,
166
+ "eval_steps_per_second": 7.875,
167
+ "step": 2808
168
+ },
169
+ {
170
+ "best_epoch": 8,
171
+ "best_eval_accuracy": 0.6245487364620939,
172
+ "epoch": 9.0,
173
+ "step": 2808
174
+ },
175
+ {
176
+ "epoch": 9.62,
177
+ "learning_rate": 0.0025192307692307693,
178
+ "loss": 0.185,
179
+ "step": 3000
180
+ },
181
+ {
182
+ "epoch": 10.0,
183
+ "eval_accuracy": 0.5667870036101083,
184
+ "eval_loss": 0.1733800321817398,
185
+ "eval_runtime": 4.4393,
186
+ "eval_samples_per_second": 62.397,
187
+ "eval_steps_per_second": 7.884,
188
+ "step": 3120
189
+ },
190
+ {
191
+ "best_epoch": 8,
192
+ "best_eval_accuracy": 0.6245487364620939,
193
+ "epoch": 10.0,
194
+ "step": 3120
195
+ },
196
+ {
197
+ "epoch": 11.0,
198
+ "eval_accuracy": 0.5812274368231047,
199
+ "eval_loss": 0.1781184822320938,
200
+ "eval_runtime": 4.4403,
201
+ "eval_samples_per_second": 62.383,
202
+ "eval_steps_per_second": 7.882,
203
+ "step": 3432
204
+ },
205
+ {
206
+ "best_epoch": 8,
207
+ "best_eval_accuracy": 0.6245487364620939,
208
+ "epoch": 11.0,
209
+ "step": 3432
210
+ },
211
+ {
212
+ "epoch": 11.22,
213
+ "learning_rate": 0.002439102564102564,
214
+ "loss": 0.184,
215
+ "step": 3500
216
+ },
217
+ {
218
+ "epoch": 12.0,
219
+ "eval_accuracy": 0.631768953068592,
220
+ "eval_loss": 0.17108257114887238,
221
+ "eval_runtime": 4.4449,
222
+ "eval_samples_per_second": 62.319,
223
+ "eval_steps_per_second": 7.874,
224
+ "step": 3744
225
+ },
226
+ {
227
+ "best_epoch": 11,
228
+ "best_eval_accuracy": 0.631768953068592,
229
+ "epoch": 12.0,
230
+ "step": 3744
231
+ },
232
+ {
233
+ "epoch": 12.82,
234
+ "learning_rate": 0.002358974358974359,
235
+ "loss": 0.1819,
236
+ "step": 4000
237
+ },
238
+ {
239
+ "epoch": 13.0,
240
+ "eval_accuracy": 0.49097472924187724,
241
+ "eval_loss": 0.17825525999069214,
242
+ "eval_runtime": 4.4419,
243
+ "eval_samples_per_second": 62.361,
244
+ "eval_steps_per_second": 7.88,
245
+ "step": 4056
246
+ },
247
+ {
248
+ "best_epoch": 11,
249
+ "best_eval_accuracy": 0.631768953068592,
250
+ "epoch": 13.0,
251
+ "step": 4056
252
+ },
253
+ {
254
+ "epoch": 14.0,
255
+ "eval_accuracy": 0.6534296028880866,
256
+ "eval_loss": 0.1702543944120407,
257
+ "eval_runtime": 4.4448,
258
+ "eval_samples_per_second": 62.32,
259
+ "eval_steps_per_second": 7.874,
260
+ "step": 4368
261
+ },
262
+ {
263
+ "best_epoch": 13,
264
+ "best_eval_accuracy": 0.6534296028880866,
265
+ "epoch": 14.0,
266
+ "step": 4368
267
+ },
268
+ {
269
+ "epoch": 14.42,
270
+ "learning_rate": 0.002278846153846154,
271
+ "loss": 0.1793,
272
+ "step": 4500
273
+ },
274
+ {
275
+ "epoch": 15.0,
276
+ "eval_accuracy": 0.6931407942238267,
277
+ "eval_loss": 0.16970041394233704,
278
+ "eval_runtime": 4.4427,
279
+ "eval_samples_per_second": 62.35,
280
+ "eval_steps_per_second": 7.878,
281
+ "step": 4680
282
+ },
283
+ {
284
+ "best_epoch": 14,
285
+ "best_eval_accuracy": 0.6931407942238267,
286
+ "epoch": 15.0,
287
+ "step": 4680
288
+ },
289
+ {
290
+ "epoch": 16.0,
291
+ "eval_accuracy": 0.6642599277978339,
292
+ "eval_loss": 0.1709691882133484,
293
+ "eval_runtime": 4.4418,
294
+ "eval_samples_per_second": 62.362,
295
+ "eval_steps_per_second": 7.88,
296
+ "step": 4992
297
+ },
298
+ {
299
+ "best_epoch": 14,
300
+ "best_eval_accuracy": 0.6931407942238267,
301
+ "epoch": 16.0,
302
+ "step": 4992
303
+ },
304
+ {
305
+ "epoch": 16.03,
306
+ "learning_rate": 0.0021987179487179486,
307
+ "loss": 0.179,
308
+ "step": 5000
309
+ },
310
+ {
311
+ "epoch": 17.0,
312
+ "eval_accuracy": 0.6534296028880866,
313
+ "eval_loss": 0.17275716364383698,
314
+ "eval_runtime": 4.4437,
315
+ "eval_samples_per_second": 62.335,
316
+ "eval_steps_per_second": 7.876,
317
+ "step": 5304
318
+ },
319
+ {
320
+ "best_epoch": 14,
321
+ "best_eval_accuracy": 0.6931407942238267,
322
+ "epoch": 17.0,
323
+ "step": 5304
324
+ },
325
+ {
326
+ "epoch": 17.63,
327
+ "learning_rate": 0.0021185897435897437,
328
+ "loss": 0.1784,
329
+ "step": 5500
330
+ },
331
+ {
332
+ "epoch": 18.0,
333
+ "eval_accuracy": 0.6498194945848376,
334
+ "eval_loss": 0.1711890995502472,
335
+ "eval_runtime": 4.4513,
336
+ "eval_samples_per_second": 62.229,
337
+ "eval_steps_per_second": 7.863,
338
+ "step": 5616
339
+ },
340
+ {
341
+ "best_epoch": 14,
342
+ "best_eval_accuracy": 0.6931407942238267,
343
+ "epoch": 18.0,
344
+ "step": 5616
345
+ },
346
+ {
347
+ "epoch": 19.0,
348
+ "eval_accuracy": 0.6064981949458483,
349
+ "eval_loss": 0.1725841462612152,
350
+ "eval_runtime": 4.4464,
351
+ "eval_samples_per_second": 62.297,
352
+ "eval_steps_per_second": 7.872,
353
+ "step": 5928
354
+ },
355
+ {
356
+ "best_epoch": 14,
357
+ "best_eval_accuracy": 0.6931407942238267,
358
+ "epoch": 19.0,
359
+ "step": 5928
360
+ },
361
+ {
362
+ "epoch": 19.23,
363
+ "learning_rate": 0.0020384615384615385,
364
+ "loss": 0.1778,
365
+ "step": 6000
366
+ },
367
+ {
368
+ "epoch": 20.0,
369
+ "eval_accuracy": 0.6678700361010831,
370
+ "eval_loss": 0.17198872566223145,
371
+ "eval_runtime": 4.4507,
372
+ "eval_samples_per_second": 62.237,
373
+ "eval_steps_per_second": 7.864,
374
+ "step": 6240
375
+ },
376
+ {
377
+ "best_epoch": 14,
378
+ "best_eval_accuracy": 0.6931407942238267,
379
+ "epoch": 20.0,
380
+ "step": 6240
381
+ },
382
+ {
383
+ "epoch": 20.83,
384
+ "learning_rate": 0.0019583333333333336,
385
+ "loss": 0.1761,
386
+ "step": 6500
387
+ },
388
+ {
389
+ "epoch": 21.0,
390
+ "eval_accuracy": 0.6606498194945848,
391
+ "eval_loss": 0.17243142426013947,
392
+ "eval_runtime": 4.4385,
393
+ "eval_samples_per_second": 62.408,
394
+ "eval_steps_per_second": 7.886,
395
+ "step": 6552
396
+ },
397
+ {
398
+ "best_epoch": 14,
399
+ "best_eval_accuracy": 0.6931407942238267,
400
+ "epoch": 21.0,
401
+ "step": 6552
402
+ },
403
+ {
404
+ "epoch": 22.0,
405
+ "eval_accuracy": 0.6534296028880866,
406
+ "eval_loss": 0.17924970388412476,
407
+ "eval_runtime": 4.4358,
408
+ "eval_samples_per_second": 62.447,
409
+ "eval_steps_per_second": 7.89,
410
+ "step": 6864
411
+ },
412
+ {
413
+ "best_epoch": 14,
414
+ "best_eval_accuracy": 0.6931407942238267,
415
+ "epoch": 22.0,
416
+ "step": 6864
417
+ },
418
+ {
419
+ "epoch": 22.44,
420
+ "learning_rate": 0.0018782051282051281,
421
+ "loss": 0.1761,
422
+ "step": 7000
423
+ },
424
+ {
425
+ "epoch": 23.0,
426
+ "eval_accuracy": 0.6714801444043321,
427
+ "eval_loss": 0.1700066775083542,
428
+ "eval_runtime": 4.4501,
429
+ "eval_samples_per_second": 62.246,
430
+ "eval_steps_per_second": 7.865,
431
+ "step": 7176
432
+ },
433
+ {
434
+ "best_epoch": 14,
435
+ "best_eval_accuracy": 0.6931407942238267,
436
+ "epoch": 23.0,
437
+ "step": 7176
438
+ },
439
+ {
440
+ "epoch": 24.0,
441
+ "eval_accuracy": 0.6678700361010831,
442
+ "eval_loss": 0.16977772116661072,
443
+ "eval_runtime": 4.4334,
444
+ "eval_samples_per_second": 62.48,
445
+ "eval_steps_per_second": 7.895,
446
+ "step": 7488
447
+ },
448
+ {
449
+ "best_epoch": 14,
450
+ "best_eval_accuracy": 0.6931407942238267,
451
+ "epoch": 24.0,
452
+ "step": 7488
453
+ },
454
+ {
455
+ "epoch": 24.04,
456
+ "learning_rate": 0.001798076923076923,
457
+ "loss": 0.1748,
458
+ "step": 7500
459
+ },
460
+ {
461
+ "epoch": 25.0,
462
+ "eval_accuracy": 0.6967509025270758,
463
+ "eval_loss": 0.16966477036476135,
464
+ "eval_runtime": 4.4363,
465
+ "eval_samples_per_second": 62.44,
466
+ "eval_steps_per_second": 7.89,
467
+ "step": 7800
468
+ },
469
+ {
470
+ "best_epoch": 24,
471
+ "best_eval_accuracy": 0.6967509025270758,
472
+ "epoch": 25.0,
473
+ "step": 7800
474
+ },
475
+ {
476
+ "epoch": 25.64,
477
+ "learning_rate": 0.0017179487179487178,
478
+ "loss": 0.1744,
479
+ "step": 8000
480
+ },
481
+ {
482
+ "epoch": 26.0,
483
+ "eval_accuracy": 0.6859205776173285,
484
+ "eval_loss": 0.17290985584259033,
485
+ "eval_runtime": 4.4401,
486
+ "eval_samples_per_second": 62.387,
487
+ "eval_steps_per_second": 7.883,
488
+ "step": 8112
489
+ },
490
+ {
491
+ "best_epoch": 24,
492
+ "best_eval_accuracy": 0.6967509025270758,
493
+ "epoch": 26.0,
494
+ "step": 8112
495
+ },
496
+ {
497
+ "epoch": 27.0,
498
+ "eval_accuracy": 0.6570397111913358,
499
+ "eval_loss": 0.1702173352241516,
500
+ "eval_runtime": 4.4312,
501
+ "eval_samples_per_second": 62.511,
502
+ "eval_steps_per_second": 7.898,
503
+ "step": 8424
504
+ },
505
+ {
506
+ "best_epoch": 24,
507
+ "best_eval_accuracy": 0.6967509025270758,
508
+ "epoch": 27.0,
509
+ "step": 8424
510
+ },
511
+ {
512
+ "epoch": 27.24,
513
+ "learning_rate": 0.0016378205128205127,
514
+ "loss": 0.1736,
515
+ "step": 8500
516
+ },
517
+ {
518
+ "epoch": 28.0,
519
+ "eval_accuracy": 0.6931407942238267,
520
+ "eval_loss": 0.1707705557346344,
521
+ "eval_runtime": 4.4336,
522
+ "eval_samples_per_second": 62.477,
523
+ "eval_steps_per_second": 7.894,
524
+ "step": 8736
525
+ },
526
+ {
527
+ "best_epoch": 24,
528
+ "best_eval_accuracy": 0.6967509025270758,
529
+ "epoch": 28.0,
530
+ "step": 8736
531
+ },
532
+ {
533
+ "epoch": 28.85,
534
+ "learning_rate": 0.0015576923076923079,
535
+ "loss": 0.1723,
536
+ "step": 9000
537
+ },
538
+ {
539
+ "epoch": 29.0,
540
+ "eval_accuracy": 0.6787003610108303,
541
+ "eval_loss": 0.16975490748882294,
542
+ "eval_runtime": 4.4304,
543
+ "eval_samples_per_second": 62.523,
544
+ "eval_steps_per_second": 7.9,
545
+ "step": 9048
546
+ },
547
+ {
548
+ "best_epoch": 24,
549
+ "best_eval_accuracy": 0.6967509025270758,
550
+ "epoch": 29.0,
551
+ "step": 9048
552
+ },
553
+ {
554
+ "epoch": 30.0,
555
+ "eval_accuracy": 0.6462093862815884,
556
+ "eval_loss": 0.17993663251399994,
557
+ "eval_runtime": 4.437,
558
+ "eval_samples_per_second": 62.429,
559
+ "eval_steps_per_second": 7.888,
560
+ "step": 9360
561
+ },
562
+ {
563
+ "best_epoch": 24,
564
+ "best_eval_accuracy": 0.6967509025270758,
565
+ "epoch": 30.0,
566
+ "step": 9360
567
+ },
568
+ {
569
+ "epoch": 30.45,
570
+ "learning_rate": 0.0014775641025641026,
571
+ "loss": 0.1735,
572
+ "step": 9500
573
+ },
574
+ {
575
+ "epoch": 31.0,
576
+ "eval_accuracy": 0.6750902527075813,
577
+ "eval_loss": 0.17266085743904114,
578
+ "eval_runtime": 4.4358,
579
+ "eval_samples_per_second": 62.447,
580
+ "eval_steps_per_second": 7.89,
581
+ "step": 9672
582
+ },
583
+ {
584
+ "best_epoch": 24,
585
+ "best_eval_accuracy": 0.6967509025270758,
586
+ "epoch": 31.0,
587
+ "step": 9672
588
+ },
589
+ {
590
+ "epoch": 32.0,
591
+ "eval_accuracy": 0.6498194945848376,
592
+ "eval_loss": 0.17320023477077484,
593
+ "eval_runtime": 4.4326,
594
+ "eval_samples_per_second": 62.491,
595
+ "eval_steps_per_second": 7.896,
596
+ "step": 9984
597
+ },
598
+ {
599
+ "best_epoch": 24,
600
+ "best_eval_accuracy": 0.6967509025270758,
601
+ "epoch": 32.0,
602
+ "step": 9984
603
+ },
604
+ {
605
+ "epoch": 32.05,
606
+ "learning_rate": 0.0013974358974358976,
607
+ "loss": 0.1722,
608
+ "step": 10000
609
+ },
610
+ {
611
+ "epoch": 33.0,
612
+ "eval_accuracy": 0.6750902527075813,
613
+ "eval_loss": 0.1702079027891159,
614
+ "eval_runtime": 4.4347,
615
+ "eval_samples_per_second": 62.463,
616
+ "eval_steps_per_second": 7.892,
617
+ "step": 10296
618
+ },
619
+ {
620
+ "best_epoch": 24,
621
+ "best_eval_accuracy": 0.6967509025270758,
622
+ "epoch": 33.0,
623
+ "step": 10296
624
+ },
625
+ {
626
+ "epoch": 33.65,
627
+ "learning_rate": 0.0013173076923076923,
628
+ "loss": 0.1709,
629
+ "step": 10500
630
+ },
631
+ {
632
+ "epoch": 34.0,
633
+ "eval_accuracy": 0.6967509025270758,
634
+ "eval_loss": 0.17065663635730743,
635
+ "eval_runtime": 4.4365,
636
+ "eval_samples_per_second": 62.437,
637
+ "eval_steps_per_second": 7.889,
638
+ "step": 10608
639
+ },
640
+ {
641
+ "best_epoch": 24,
642
+ "best_eval_accuracy": 0.6967509025270758,
643
+ "epoch": 34.0,
644
+ "step": 10608
645
+ },
646
+ {
647
+ "epoch": 35.0,
648
+ "eval_accuracy": 0.6967509025270758,
649
+ "eval_loss": 0.17139047384262085,
650
+ "eval_runtime": 4.4388,
651
+ "eval_samples_per_second": 62.405,
652
+ "eval_steps_per_second": 7.885,
653
+ "step": 10920
654
+ },
655
+ {
656
+ "best_epoch": 24,
657
+ "best_eval_accuracy": 0.6967509025270758,
658
+ "epoch": 35.0,
659
+ "step": 10920
660
+ },
661
+ {
662
+ "epoch": 35.26,
663
+ "learning_rate": 0.0012371794871794872,
664
+ "loss": 0.1697,
665
+ "step": 11000
666
+ },
667
+ {
668
+ "epoch": 36.0,
669
+ "eval_accuracy": 0.6750902527075813,
670
+ "eval_loss": 0.17118841409683228,
671
+ "eval_runtime": 4.4402,
672
+ "eval_samples_per_second": 62.385,
673
+ "eval_steps_per_second": 7.883,
674
+ "step": 11232
675
+ },
676
+ {
677
+ "best_epoch": 24,
678
+ "best_eval_accuracy": 0.6967509025270758,
679
+ "epoch": 36.0,
680
+ "step": 11232
681
+ },
682
+ {
683
+ "epoch": 36.86,
684
+ "learning_rate": 0.0011570512820512822,
685
+ "loss": 0.1696,
686
+ "step": 11500
687
+ },
688
+ {
689
+ "epoch": 37.0,
690
+ "eval_accuracy": 0.6570397111913358,
691
+ "eval_loss": 0.17884960770606995,
692
+ "eval_runtime": 4.4456,
693
+ "eval_samples_per_second": 62.309,
694
+ "eval_steps_per_second": 7.873,
695
+ "step": 11544
696
+ },
697
+ {
698
+ "best_epoch": 24,
699
+ "best_eval_accuracy": 0.6967509025270758,
700
+ "epoch": 37.0,
701
+ "step": 11544
702
+ },
703
+ {
704
+ "epoch": 38.0,
705
+ "eval_accuracy": 0.6787003610108303,
706
+ "eval_loss": 0.17026345431804657,
707
+ "eval_runtime": 4.4452,
708
+ "eval_samples_per_second": 62.315,
709
+ "eval_steps_per_second": 7.874,
710
+ "step": 11856
711
+ },
712
+ {
713
+ "best_epoch": 24,
714
+ "best_eval_accuracy": 0.6967509025270758,
715
+ "epoch": 38.0,
716
+ "step": 11856
717
+ },
718
+ {
719
+ "epoch": 38.46,
720
+ "learning_rate": 0.0010769230769230769,
721
+ "loss": 0.1697,
722
+ "step": 12000
723
+ },
724
+ {
725
+ "epoch": 39.0,
726
+ "eval_accuracy": 0.6750902527075813,
727
+ "eval_loss": 0.17346355319023132,
728
+ "eval_runtime": 4.4459,
729
+ "eval_samples_per_second": 62.305,
730
+ "eval_steps_per_second": 7.872,
731
+ "step": 12168
732
+ },
733
+ {
734
+ "best_epoch": 24,
735
+ "best_eval_accuracy": 0.6967509025270758,
736
+ "epoch": 39.0,
737
+ "step": 12168
738
+ },
739
+ {
740
+ "epoch": 40.0,
741
+ "eval_accuracy": 0.6787003610108303,
742
+ "eval_loss": 0.17404569685459137,
743
+ "eval_runtime": 4.4468,
744
+ "eval_samples_per_second": 62.292,
745
+ "eval_steps_per_second": 7.871,
746
+ "step": 12480
747
+ },
748
+ {
749
+ "best_epoch": 24,
750
+ "best_eval_accuracy": 0.6967509025270758,
751
+ "epoch": 40.0,
752
+ "step": 12480
753
+ },
754
+ {
755
+ "epoch": 40.06,
756
+ "learning_rate": 0.0009967948717948718,
757
+ "loss": 0.1683,
758
+ "step": 12500
759
+ },
760
+ {
761
+ "epoch": 41.0,
762
+ "eval_accuracy": 0.6895306859205776,
763
+ "eval_loss": 0.1709529161453247,
764
+ "eval_runtime": 4.4491,
765
+ "eval_samples_per_second": 62.26,
766
+ "eval_steps_per_second": 7.867,
767
+ "step": 12792
768
+ },
769
+ {
770
+ "best_epoch": 24,
771
+ "best_eval_accuracy": 0.6967509025270758,
772
+ "epoch": 41.0,
773
+ "step": 12792
774
+ },
775
+ {
776
+ "epoch": 41.67,
777
+ "learning_rate": 0.0009166666666666668,
778
+ "loss": 0.1688,
779
+ "step": 13000
780
+ },
781
+ {
782
+ "epoch": 42.0,
783
+ "eval_accuracy": 0.7075812274368231,
784
+ "eval_loss": 0.17236575484275818,
785
+ "eval_runtime": 4.4523,
786
+ "eval_samples_per_second": 62.215,
787
+ "eval_steps_per_second": 7.861,
788
+ "step": 13104
789
+ },
790
+ {
791
+ "best_epoch": 41,
792
+ "best_eval_accuracy": 0.7075812274368231,
793
+ "epoch": 42.0,
794
+ "step": 13104
795
+ },
796
+ {
797
+ "epoch": 43.0,
798
+ "eval_accuracy": 0.7003610108303249,
799
+ "eval_loss": 0.17182697355747223,
800
+ "eval_runtime": 4.4464,
801
+ "eval_samples_per_second": 62.298,
802
+ "eval_steps_per_second": 7.872,
803
+ "step": 13416
804
+ },
805
+ {
806
+ "best_epoch": 41,
807
+ "best_eval_accuracy": 0.7075812274368231,
808
+ "epoch": 43.0,
809
+ "step": 13416
810
+ },
811
+ {
812
+ "epoch": 43.27,
813
+ "learning_rate": 0.0008365384615384616,
814
+ "loss": 0.1679,
815
+ "step": 13500
816
+ },
817
+ {
818
+ "epoch": 44.0,
819
+ "eval_accuracy": 0.703971119133574,
820
+ "eval_loss": 0.17357730865478516,
821
+ "eval_runtime": 4.4496,
822
+ "eval_samples_per_second": 62.253,
823
+ "eval_steps_per_second": 7.866,
824
+ "step": 13728
825
+ },
826
+ {
827
+ "best_epoch": 41,
828
+ "best_eval_accuracy": 0.7075812274368231,
829
+ "epoch": 44.0,
830
+ "step": 13728
831
+ },
832
+ {
833
+ "epoch": 44.87,
834
+ "learning_rate": 0.0007564102564102564,
835
+ "loss": 0.1681,
836
+ "step": 14000
837
+ },
838
+ {
839
+ "epoch": 45.0,
840
+ "eval_accuracy": 0.703971119133574,
841
+ "eval_loss": 0.17195428907871246,
842
+ "eval_runtime": 4.4504,
843
+ "eval_samples_per_second": 62.242,
844
+ "eval_steps_per_second": 7.865,
845
+ "step": 14040
846
+ },
847
+ {
848
+ "best_epoch": 41,
849
+ "best_eval_accuracy": 0.7075812274368231,
850
+ "epoch": 45.0,
851
+ "step": 14040
852
+ },
853
+ {
854
+ "epoch": 46.0,
855
+ "eval_accuracy": 0.7075812274368231,
856
+ "eval_loss": 0.17170363664627075,
857
+ "eval_runtime": 4.4496,
858
+ "eval_samples_per_second": 62.253,
859
+ "eval_steps_per_second": 7.866,
860
+ "step": 14352
861
+ },
862
+ {
863
+ "best_epoch": 41,
864
+ "best_eval_accuracy": 0.7075812274368231,
865
+ "epoch": 46.0,
866
+ "step": 14352
867
+ },
868
+ {
869
+ "epoch": 46.47,
870
+ "learning_rate": 0.0006762820512820514,
871
+ "loss": 0.1664,
872
+ "step": 14500
873
+ },
874
+ {
875
+ "epoch": 47.0,
876
+ "eval_accuracy": 0.6895306859205776,
877
+ "eval_loss": 0.17101722955703735,
878
+ "eval_runtime": 4.454,
879
+ "eval_samples_per_second": 62.192,
880
+ "eval_steps_per_second": 7.858,
881
+ "step": 14664
882
+ },
883
+ {
884
+ "best_epoch": 41,
885
+ "best_eval_accuracy": 0.7075812274368231,
886
+ "epoch": 47.0,
887
+ "step": 14664
888
+ },
889
+ {
890
+ "epoch": 48.0,
891
+ "eval_accuracy": 0.6895306859205776,
892
+ "eval_loss": 0.17659461498260498,
893
+ "eval_runtime": 4.4486,
894
+ "eval_samples_per_second": 62.267,
895
+ "eval_steps_per_second": 7.868,
896
+ "step": 14976
897
+ },
898
+ {
899
+ "best_epoch": 41,
900
+ "best_eval_accuracy": 0.7075812274368231,
901
+ "epoch": 48.0,
902
+ "step": 14976
903
+ },
904
+ {
905
+ "epoch": 48.08,
906
+ "learning_rate": 0.0005961538461538461,
907
+ "loss": 0.1662,
908
+ "step": 15000
909
+ },
910
+ {
911
+ "epoch": 49.0,
912
+ "eval_accuracy": 0.703971119133574,
913
+ "eval_loss": 0.1728653609752655,
914
+ "eval_runtime": 4.4498,
915
+ "eval_samples_per_second": 62.25,
916
+ "eval_steps_per_second": 7.865,
917
+ "step": 15288
918
+ },
919
+ {
920
+ "best_epoch": 41,
921
+ "best_eval_accuracy": 0.7075812274368231,
922
+ "epoch": 49.0,
923
+ "step": 15288
924
+ },
925
+ {
926
+ "epoch": 49.68,
927
+ "learning_rate": 0.000516025641025641,
928
+ "loss": 0.1655,
929
+ "step": 15500
930
+ },
931
+ {
932
+ "epoch": 50.0,
933
+ "eval_accuracy": 0.7075812274368231,
934
+ "eval_loss": 0.17041344940662384,
935
+ "eval_runtime": 4.4544,
936
+ "eval_samples_per_second": 62.185,
937
+ "eval_steps_per_second": 7.857,
938
+ "step": 15600
939
+ },
940
+ {
941
+ "best_epoch": 41,
942
+ "best_eval_accuracy": 0.7075812274368231,
943
+ "epoch": 50.0,
944
+ "step": 15600
945
+ },
946
+ {
947
+ "epoch": 51.0,
948
+ "eval_accuracy": 0.7184115523465704,
949
+ "eval_loss": 0.1710704118013382,
950
+ "eval_runtime": 4.4496,
951
+ "eval_samples_per_second": 62.252,
952
+ "eval_steps_per_second": 7.866,
953
+ "step": 15912
954
+ },
955
+ {
956
+ "best_epoch": 50,
957
+ "best_eval_accuracy": 0.7184115523465704,
958
+ "epoch": 51.0,
959
+ "step": 15912
960
+ },
961
+ {
962
+ "epoch": 51.28,
963
+ "learning_rate": 0.00043589743589743596,
964
+ "loss": 0.1665,
965
+ "step": 16000
966
+ },
967
+ {
968
+ "epoch": 52.0,
969
+ "eval_accuracy": 0.703971119133574,
970
+ "eval_loss": 0.1708567589521408,
971
+ "eval_runtime": 4.4539,
972
+ "eval_samples_per_second": 62.192,
973
+ "eval_steps_per_second": 7.858,
974
+ "step": 16224
975
+ },
976
+ {
977
+ "best_epoch": 50,
978
+ "best_eval_accuracy": 0.7184115523465704,
979
+ "epoch": 52.0,
980
+ "step": 16224
981
+ },
982
+ {
983
+ "epoch": 52.88,
984
+ "learning_rate": 0.00035576923076923074,
985
+ "loss": 0.1651,
986
+ "step": 16500
987
+ },
988
+ {
989
+ "epoch": 53.0,
990
+ "eval_accuracy": 0.6931407942238267,
991
+ "eval_loss": 0.1710604727268219,
992
+ "eval_runtime": 4.4518,
993
+ "eval_samples_per_second": 62.222,
994
+ "eval_steps_per_second": 7.862,
995
+ "step": 16536
996
+ },
997
+ {
998
+ "best_epoch": 50,
999
+ "best_eval_accuracy": 0.7184115523465704,
1000
+ "epoch": 53.0,
1001
+ "step": 16536
1002
+ },
1003
+ {
1004
+ "epoch": 54.0,
1005
+ "eval_accuracy": 0.703971119133574,
1006
+ "eval_loss": 0.173600971698761,
1007
+ "eval_runtime": 4.4493,
1008
+ "eval_samples_per_second": 62.257,
1009
+ "eval_steps_per_second": 7.866,
1010
+ "step": 16848
1011
+ },
1012
+ {
1013
+ "best_epoch": 50,
1014
+ "best_eval_accuracy": 0.7184115523465704,
1015
+ "epoch": 54.0,
1016
+ "step": 16848
1017
+ },
1018
+ {
1019
+ "epoch": 54.49,
1020
+ "learning_rate": 0.0002756410256410257,
1021
+ "loss": 0.1646,
1022
+ "step": 17000
1023
+ },
1024
+ {
1025
+ "epoch": 55.0,
1026
+ "eval_accuracy": 0.7111913357400722,
1027
+ "eval_loss": 0.17123006284236908,
1028
+ "eval_runtime": 4.452,
1029
+ "eval_samples_per_second": 62.219,
1030
+ "eval_steps_per_second": 7.862,
1031
+ "step": 17160
1032
+ },
1033
+ {
1034
+ "best_epoch": 50,
1035
+ "best_eval_accuracy": 0.7184115523465704,
1036
+ "epoch": 55.0,
1037
+ "step": 17160
1038
+ },
1039
+ {
1040
+ "epoch": 56.0,
1041
+ "eval_accuracy": 0.7075812274368231,
1042
+ "eval_loss": 0.17398422956466675,
1043
+ "eval_runtime": 4.4571,
1044
+ "eval_samples_per_second": 62.149,
1045
+ "eval_steps_per_second": 7.853,
1046
+ "step": 17472
1047
+ },
1048
+ {
1049
+ "best_epoch": 50,
1050
+ "best_eval_accuracy": 0.7184115523465704,
1051
+ "epoch": 56.0,
1052
+ "step": 17472
1053
+ },
1054
+ {
1055
+ "epoch": 56.09,
1056
+ "learning_rate": 0.00019551282051282054,
1057
+ "loss": 0.1647,
1058
+ "step": 17500
1059
+ },
1060
+ {
1061
+ "epoch": 57.0,
1062
+ "eval_accuracy": 0.7075812274368231,
1063
+ "eval_loss": 0.17230337858200073,
1064
+ "eval_runtime": 4.446,
1065
+ "eval_samples_per_second": 62.303,
1066
+ "eval_steps_per_second": 7.872,
1067
+ "step": 17784
1068
+ },
1069
+ {
1070
+ "best_epoch": 50,
1071
+ "best_eval_accuracy": 0.7184115523465704,
1072
+ "epoch": 57.0,
1073
+ "step": 17784
1074
+ },
1075
+ {
1076
+ "epoch": 57.69,
1077
+ "learning_rate": 0.0001153846153846154,
1078
+ "loss": 0.1642,
1079
+ "step": 18000
1080
+ },
1081
+ {
1082
+ "epoch": 58.0,
1083
+ "eval_accuracy": 0.7003610108303249,
1084
+ "eval_loss": 0.1714901179075241,
1085
+ "eval_runtime": 4.4636,
1086
+ "eval_samples_per_second": 62.057,
1087
+ "eval_steps_per_second": 7.841,
1088
+ "step": 18096
1089
+ },
1090
+ {
1091
+ "best_epoch": 50,
1092
+ "best_eval_accuracy": 0.7184115523465704,
1093
+ "epoch": 58.0,
1094
+ "step": 18096
1095
+ },
1096
+ {
1097
+ "epoch": 59.0,
1098
+ "eval_accuracy": 0.7075812274368231,
1099
+ "eval_loss": 0.17266467213630676,
1100
+ "eval_runtime": 4.446,
1101
+ "eval_samples_per_second": 62.303,
1102
+ "eval_steps_per_second": 7.872,
1103
+ "step": 18408
1104
+ },
1105
+ {
1106
+ "best_epoch": 50,
1107
+ "best_eval_accuracy": 0.7184115523465704,
1108
+ "epoch": 59.0,
1109
+ "step": 18408
1110
+ },
1111
+ {
1112
+ "epoch": 59.29,
1113
+ "learning_rate": 3.5256410256410254e-05,
1114
+ "loss": 0.1643,
1115
+ "step": 18500
1116
+ },
1117
+ {
1118
+ "epoch": 60.0,
1119
+ "eval_accuracy": 0.7111913357400722,
1120
+ "eval_loss": 0.1723625361919403,
1121
+ "eval_runtime": 4.4235,
1122
+ "eval_samples_per_second": 62.62,
1123
+ "eval_steps_per_second": 7.912,
1124
+ "step": 18720
1125
+ },
1126
+ {
1127
+ "best_epoch": 50,
1128
+ "best_eval_accuracy": 0.7184115523465704,
1129
+ "epoch": 60.0,
1130
+ "step": 18720
1131
+ },
1132
+ {
1133
+ "epoch": 60.0,
1134
+ "step": 18720,
1135
+ "total_flos": 6.96152728406016e+16,
1136
+ "train_loss": 0.17944894468682443,
1137
+ "train_runtime": 4049.329,
1138
+ "train_samples_per_second": 36.895,
1139
+ "train_steps_per_second": 4.623
1140
+ }
1141
+ ],
1142
+ "max_steps": 18720,
1143
+ "num_train_epochs": 60,
1144
+ "total_flos": 6.96152728406016e+16,
1145
+ "trial_name": null,
1146
+ "trial_params": null
1147
+ }