pszemraj commited on
Commit
3bb18e9
1 Parent(s): 4b90790

Delete trainer_state.json

Browse files
Files changed (1) hide show
  1. trainer_state.json +0 -517
trainer_state.json DELETED
@@ -1,517 +0,0 @@
1
- {
2
- "best_metric": null,
3
- "best_model_checkpoint": null,
4
- "epoch": 1.9905956112852663,
5
- "global_step": 158,
6
- "is_hyper_param_search": false,
7
- "is_local_process_zero": true,
8
- "is_world_process_zero": true,
9
- "log_history": [
10
- {
11
- "epoch": 0.03,
12
- "learning_rate": 2.5e-05,
13
- "loss": 1.5684,
14
- "step": 2
15
- },
16
- {
17
- "epoch": 0.05,
18
- "learning_rate": 5e-05,
19
- "loss": 1.7603,
20
- "step": 4
21
- },
22
- {
23
- "epoch": 0.08,
24
- "learning_rate": 7.500000000000001e-05,
25
- "loss": 1.7578,
26
- "step": 6
27
- },
28
- {
29
- "epoch": 0.1,
30
- "learning_rate": 0.0001,
31
- "loss": 1.5327,
32
- "step": 8
33
- },
34
- {
35
- "epoch": 0.13,
36
- "learning_rate": 9.995614150494293e-05,
37
- "loss": 1.5737,
38
- "step": 10
39
- },
40
- {
41
- "epoch": 0.15,
42
- "learning_rate": 9.982464296247522e-05,
43
- "loss": 1.4917,
44
- "step": 12
45
- },
46
- {
47
- "epoch": 0.18,
48
- "learning_rate": 9.96057350657239e-05,
49
- "loss": 1.4312,
50
- "step": 14
51
- },
52
- {
53
- "epoch": 0.2,
54
- "learning_rate": 9.929980185352526e-05,
55
- "loss": 1.4136,
56
- "step": 16
57
- },
58
- {
59
- "epoch": 0.23,
60
- "learning_rate": 9.890738003669029e-05,
61
- "loss": 1.3789,
62
- "step": 18
63
- },
64
- {
65
- "epoch": 0.25,
66
- "learning_rate": 9.842915805643155e-05,
67
- "loss": 1.5537,
68
- "step": 20
69
- },
70
- {
71
- "epoch": 0.28,
72
- "learning_rate": 9.786597487660337e-05,
73
- "loss": 1.6504,
74
- "step": 22
75
- },
76
- {
77
- "epoch": 0.3,
78
- "learning_rate": 9.721881851187406e-05,
79
- "loss": 1.4834,
80
- "step": 24
81
- },
82
- {
83
- "epoch": 0.33,
84
- "learning_rate": 9.648882429441257e-05,
85
- "loss": 1.4912,
86
- "step": 26
87
- },
88
- {
89
- "epoch": 0.35,
90
- "learning_rate": 9.567727288213005e-05,
91
- "loss": 1.4409,
92
- "step": 28
93
- },
94
- {
95
- "epoch": 0.38,
96
- "learning_rate": 9.478558801197065e-05,
97
- "loss": 1.3872,
98
- "step": 30
99
- },
100
- {
101
- "epoch": 0.4,
102
- "learning_rate": 9.381533400219318e-05,
103
- "loss": 1.4736,
104
- "step": 32
105
- },
106
- {
107
- "epoch": 0.43,
108
- "learning_rate": 9.276821300802534e-05,
109
- "loss": 1.4048,
110
- "step": 34
111
- },
112
- {
113
- "epoch": 0.45,
114
- "learning_rate": 9.164606203550497e-05,
115
- "loss": 1.4502,
116
- "step": 36
117
- },
118
- {
119
- "epoch": 0.48,
120
- "learning_rate": 9.045084971874738e-05,
121
- "loss": 1.4141,
122
- "step": 38
123
- },
124
- {
125
- "epoch": 0.5,
126
- "learning_rate": 8.9184672866292e-05,
127
- "loss": 1.3911,
128
- "step": 40
129
- },
130
- {
131
- "epoch": 0.53,
132
- "learning_rate": 8.784975278258783e-05,
133
- "loss": 1.3379,
134
- "step": 42
135
- },
136
- {
137
- "epoch": 0.55,
138
- "learning_rate": 8.644843137107059e-05,
139
- "loss": 1.3379,
140
- "step": 44
141
- },
142
- {
143
- "epoch": 0.58,
144
- "learning_rate": 8.498316702566828e-05,
145
- "loss": 1.3242,
146
- "step": 46
147
- },
148
- {
149
- "epoch": 0.6,
150
- "learning_rate": 8.345653031794292e-05,
151
- "loss": 1.3091,
152
- "step": 48
153
- },
154
- {
155
- "epoch": 0.63,
156
- "learning_rate": 8.18711994874345e-05,
157
- "loss": 1.312,
158
- "step": 50
159
- },
160
- {
161
- "epoch": 0.65,
162
- "learning_rate": 8.022995574311876e-05,
163
- "loss": 1.3018,
164
- "step": 52
165
- },
166
- {
167
- "epoch": 0.68,
168
- "learning_rate": 7.85356783842216e-05,
169
- "loss": 1.3115,
170
- "step": 54
171
- },
172
- {
173
- "epoch": 0.7,
174
- "learning_rate": 7.679133974894983e-05,
175
- "loss": 1.3188,
176
- "step": 56
177
- },
178
- {
179
- "epoch": 0.73,
180
- "learning_rate": 7.500000000000001e-05,
181
- "loss": 1.314,
182
- "step": 58
183
- },
184
- {
185
- "epoch": 0.75,
186
- "learning_rate": 7.316480175599309e-05,
187
- "loss": 1.3062,
188
- "step": 60
189
- },
190
- {
191
- "epoch": 0.78,
192
- "learning_rate": 7.128896457825364e-05,
193
- "loss": 1.3076,
194
- "step": 62
195
- },
196
- {
197
- "epoch": 0.8,
198
- "learning_rate": 6.937577932260515e-05,
199
- "loss": 1.2847,
200
- "step": 64
201
- },
202
- {
203
- "epoch": 0.83,
204
- "learning_rate": 6.742860236609077e-05,
205
- "loss": 1.2837,
206
- "step": 66
207
- },
208
- {
209
- "epoch": 0.85,
210
- "learning_rate": 6.545084971874738e-05,
211
- "loss": 1.2798,
212
- "step": 68
213
- },
214
- {
215
- "epoch": 0.88,
216
- "learning_rate": 6.344599103076329e-05,
217
- "loss": 1.3247,
218
- "step": 70
219
- },
220
- {
221
- "epoch": 0.9,
222
- "learning_rate": 6.141754350553279e-05,
223
- "loss": 1.2842,
224
- "step": 72
225
- },
226
- {
227
- "epoch": 0.93,
228
- "learning_rate": 5.9369065729286245e-05,
229
- "loss": 1.2969,
230
- "step": 74
231
- },
232
- {
233
- "epoch": 0.95,
234
- "learning_rate": 5.730415142812059e-05,
235
- "loss": 1.2866,
236
- "step": 76
237
- },
238
- {
239
- "epoch": 0.98,
240
- "learning_rate": 5.522642316338268e-05,
241
- "loss": 1.2598,
242
- "step": 78
243
- },
244
- {
245
- "epoch": 0.99,
246
- "eval_accuracy": 0.6496387020106186,
247
- "eval_loss": 1.329074740409851,
248
- "eval_runtime": 71.5884,
249
- "eval_samples_per_second": 3.562,
250
- "eval_steps_per_second": 1.788,
251
- "step": 79
252
- },
253
- {
254
- "epoch": 1.01,
255
- "learning_rate": 5.313952597646568e-05,
256
- "loss": 1.6475,
257
- "step": 80
258
- },
259
- {
260
- "epoch": 1.04,
261
- "learning_rate": 5.104712099416785e-05,
262
- "loss": 0.9348,
263
- "step": 82
264
- },
265
- {
266
- "epoch": 1.06,
267
- "learning_rate": 4.895287900583216e-05,
268
- "loss": 0.8953,
269
- "step": 84
270
- },
271
- {
272
- "epoch": 1.09,
273
- "learning_rate": 4.6860474023534335e-05,
274
- "loss": 0.8994,
275
- "step": 86
276
- },
277
- {
278
- "epoch": 1.11,
279
- "learning_rate": 4.477357683661734e-05,
280
- "loss": 0.8591,
281
- "step": 88
282
- },
283
- {
284
- "epoch": 1.14,
285
- "learning_rate": 4.269584857187943e-05,
286
- "loss": 0.8354,
287
- "step": 90
288
- },
289
- {
290
- "epoch": 1.16,
291
- "learning_rate": 4.063093427071376e-05,
292
- "loss": 0.863,
293
- "step": 92
294
- },
295
- {
296
- "epoch": 1.19,
297
- "learning_rate": 3.858245649446721e-05,
298
- "loss": 0.8845,
299
- "step": 94
300
- },
301
- {
302
- "epoch": 1.21,
303
- "learning_rate": 3.655400896923672e-05,
304
- "loss": 0.8325,
305
- "step": 96
306
- },
307
- {
308
- "epoch": 1.24,
309
- "learning_rate": 3.4549150281252636e-05,
310
- "loss": 0.8989,
311
- "step": 98
312
- },
313
- {
314
- "epoch": 1.26,
315
- "learning_rate": 3.257139763390925e-05,
316
- "loss": 0.8782,
317
- "step": 100
318
- },
319
- {
320
- "epoch": 1.29,
321
- "learning_rate": 3.062422067739485e-05,
322
- "loss": 0.8052,
323
- "step": 102
324
- },
325
- {
326
- "epoch": 1.31,
327
- "learning_rate": 2.8711035421746367e-05,
328
- "loss": 0.8518,
329
- "step": 104
330
- },
331
- {
332
- "epoch": 1.34,
333
- "learning_rate": 2.6835198244006927e-05,
334
- "loss": 0.8516,
335
- "step": 106
336
- },
337
- {
338
- "epoch": 1.36,
339
- "learning_rate": 2.500000000000001e-05,
340
- "loss": 0.7979,
341
- "step": 108
342
- },
343
- {
344
- "epoch": 1.39,
345
- "learning_rate": 2.3208660251050158e-05,
346
- "loss": 0.8108,
347
- "step": 110
348
- },
349
- {
350
- "epoch": 1.41,
351
- "learning_rate": 2.1464321615778422e-05,
352
- "loss": 0.8547,
353
- "step": 112
354
- },
355
- {
356
- "epoch": 1.44,
357
- "learning_rate": 1.977004425688126e-05,
358
- "loss": 0.8684,
359
- "step": 114
360
- },
361
- {
362
- "epoch": 1.46,
363
- "learning_rate": 1.8128800512565513e-05,
364
- "loss": 0.8428,
365
- "step": 116
366
- },
367
- {
368
- "epoch": 1.49,
369
- "learning_rate": 1.6543469682057106e-05,
370
- "loss": 0.7878,
371
- "step": 118
372
- },
373
- {
374
- "epoch": 1.51,
375
- "learning_rate": 1.5016832974331724e-05,
376
- "loss": 0.822,
377
- "step": 120
378
- },
379
- {
380
- "epoch": 1.54,
381
- "learning_rate": 1.3551568628929434e-05,
382
- "loss": 0.79,
383
- "step": 122
384
- },
385
- {
386
- "epoch": 1.56,
387
- "learning_rate": 1.2150247217412186e-05,
388
- "loss": 0.7979,
389
- "step": 124
390
- },
391
- {
392
- "epoch": 1.59,
393
- "learning_rate": 1.0815327133708015e-05,
394
- "loss": 0.8081,
395
- "step": 126
396
- },
397
- {
398
- "epoch": 1.61,
399
- "learning_rate": 9.549150281252633e-06,
400
- "loss": 0.7751,
401
- "step": 128
402
- },
403
- {
404
- "epoch": 1.64,
405
- "learning_rate": 8.353937964495029e-06,
406
- "loss": 0.7871,
407
- "step": 130
408
- },
409
- {
410
- "epoch": 1.66,
411
- "learning_rate": 7.2317869919746705e-06,
412
- "loss": 0.7722,
413
- "step": 132
414
- },
415
- {
416
- "epoch": 1.69,
417
- "learning_rate": 6.184665997806832e-06,
418
- "loss": 0.8198,
419
- "step": 134
420
- },
421
- {
422
- "epoch": 1.71,
423
- "learning_rate": 5.214411988029355e-06,
424
- "loss": 0.7859,
425
- "step": 136
426
- },
427
- {
428
- "epoch": 1.74,
429
- "learning_rate": 4.322727117869951e-06,
430
- "loss": 0.7458,
431
- "step": 138
432
- },
433
- {
434
- "epoch": 1.76,
435
- "learning_rate": 3.511175705587433e-06,
436
- "loss": 0.7554,
437
- "step": 140
438
- },
439
- {
440
- "epoch": 1.79,
441
- "learning_rate": 2.7811814881259503e-06,
442
- "loss": 0.7783,
443
- "step": 142
444
- },
445
- {
446
- "epoch": 1.82,
447
- "learning_rate": 2.134025123396638e-06,
448
- "loss": 0.7783,
449
- "step": 144
450
- },
451
- {
452
- "epoch": 1.84,
453
- "learning_rate": 1.5708419435684462e-06,
454
- "loss": 0.7793,
455
- "step": 146
456
- },
457
- {
458
- "epoch": 1.87,
459
- "learning_rate": 1.0926199633097157e-06,
460
- "loss": 0.8064,
461
- "step": 148
462
- },
463
- {
464
- "epoch": 1.89,
465
- "learning_rate": 7.001981464747565e-07,
466
- "loss": 0.8042,
467
- "step": 150
468
- },
469
- {
470
- "epoch": 1.92,
471
- "learning_rate": 3.9426493427611177e-07,
472
- "loss": 0.791,
473
- "step": 152
474
- },
475
- {
476
- "epoch": 1.94,
477
- "learning_rate": 1.753570375247815e-07,
478
- "loss": 0.8147,
479
- "step": 154
480
- },
481
- {
482
- "epoch": 1.97,
483
- "learning_rate": 4.385849505708084e-08,
484
- "loss": 0.7617,
485
- "step": 156
486
- },
487
- {
488
- "epoch": 1.99,
489
- "learning_rate": 0.0,
490
- "loss": 0.7446,
491
- "step": 158
492
- },
493
- {
494
- "epoch": 1.99,
495
- "eval_accuracy": 0.6768941789814655,
496
- "eval_loss": 1.2372242212295532,
497
- "eval_runtime": 71.61,
498
- "eval_samples_per_second": 3.561,
499
- "eval_steps_per_second": 1.787,
500
- "step": 158
501
- },
502
- {
503
- "epoch": 1.99,
504
- "step": 158,
505
- "total_flos": 4.14872908733612e+17,
506
- "train_loss": 1.1195918215981013,
507
- "train_runtime": 35849.6176,
508
- "train_samples_per_second": 0.284,
509
- "train_steps_per_second": 0.004
510
- }
511
- ],
512
- "max_steps": 158,
513
- "num_train_epochs": 2,
514
- "total_flos": 4.14872908733612e+17,
515
- "trial_name": null,
516
- "trial_params": null
517
- }