fkov commited on
Commit
3bba2d4
1 Parent(s): e2813c2

End of training

Browse files
Files changed (4) hide show
  1. all_results.json +12 -0
  2. eval_results.json +8 -0
  3. train_results.json +7 -0
  4. trainer_state.json +1264 -0
all_results.json ADDED
@@ -0,0 +1,12 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 5.0,
3
+ "eval_accuracy": 0.980435422182995,
4
+ "eval_loss": 0.08353430777788162,
5
+ "eval_runtime": 5.3428,
6
+ "eval_samples_per_second": 1272.355,
7
+ "eval_steps_per_second": 39.866,
8
+ "train_loss": 0.4383530242102487,
9
+ "train_runtime": 692.5799,
10
+ "train_samples_per_second": 368.867,
11
+ "train_steps_per_second": 2.881
12
+ }
eval_results.json ADDED
@@ -0,0 +1,8 @@
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 5.0,
3
+ "eval_accuracy": 0.980435422182995,
4
+ "eval_loss": 0.08353430777788162,
5
+ "eval_runtime": 5.3428,
6
+ "eval_samples_per_second": 1272.355,
7
+ "eval_steps_per_second": 39.866
8
+ }
train_results.json ADDED
@@ -0,0 +1,7 @@
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 5.0,
3
+ "train_loss": 0.4383530242102487,
4
+ "train_runtime": 692.5799,
5
+ "train_samples_per_second": 368.867,
6
+ "train_steps_per_second": 2.881
7
+ }
trainer_state.json ADDED
@@ -0,0 +1,1264 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": 0.980435422182995,
3
+ "best_model_checkpoint": "l/users/felicia.kovacs/outputs/wav2vec2-base-ft-keyword-spotting/checkpoint-1995",
4
+ "epoch": 4.996869129618034,
5
+ "global_step": 1995,
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": 1.5e-06,
13
+ "loss": 2.495,
14
+ "step": 10
15
+ },
16
+ {
17
+ "epoch": 0.05,
18
+ "learning_rate": 3e-06,
19
+ "loss": 2.4758,
20
+ "step": 20
21
+ },
22
+ {
23
+ "epoch": 0.08,
24
+ "learning_rate": 4.5e-06,
25
+ "loss": 2.4193,
26
+ "step": 30
27
+ },
28
+ {
29
+ "epoch": 0.1,
30
+ "learning_rate": 6e-06,
31
+ "loss": 2.3206,
32
+ "step": 40
33
+ },
34
+ {
35
+ "epoch": 0.13,
36
+ "learning_rate": 7.5e-06,
37
+ "loss": 2.1873,
38
+ "step": 50
39
+ },
40
+ {
41
+ "epoch": 0.15,
42
+ "learning_rate": 9e-06,
43
+ "loss": 1.9877,
44
+ "step": 60
45
+ },
46
+ {
47
+ "epoch": 0.18,
48
+ "learning_rate": 1.05e-05,
49
+ "loss": 1.8641,
50
+ "step": 70
51
+ },
52
+ {
53
+ "epoch": 0.2,
54
+ "learning_rate": 1.2e-05,
55
+ "loss": 1.7812,
56
+ "step": 80
57
+ },
58
+ {
59
+ "epoch": 0.23,
60
+ "learning_rate": 1.3500000000000001e-05,
61
+ "loss": 1.6689,
62
+ "step": 90
63
+ },
64
+ {
65
+ "epoch": 0.25,
66
+ "learning_rate": 1.5e-05,
67
+ "loss": 1.6289,
68
+ "step": 100
69
+ },
70
+ {
71
+ "epoch": 0.28,
72
+ "learning_rate": 1.65e-05,
73
+ "loss": 1.5056,
74
+ "step": 110
75
+ },
76
+ {
77
+ "epoch": 0.3,
78
+ "learning_rate": 1.8e-05,
79
+ "loss": 1.4966,
80
+ "step": 120
81
+ },
82
+ {
83
+ "epoch": 0.33,
84
+ "learning_rate": 1.95e-05,
85
+ "loss": 1.4948,
86
+ "step": 130
87
+ },
88
+ {
89
+ "epoch": 0.35,
90
+ "learning_rate": 2.1e-05,
91
+ "loss": 1.4834,
92
+ "step": 140
93
+ },
94
+ {
95
+ "epoch": 0.38,
96
+ "learning_rate": 2.25e-05,
97
+ "loss": 1.5055,
98
+ "step": 150
99
+ },
100
+ {
101
+ "epoch": 0.4,
102
+ "learning_rate": 2.4e-05,
103
+ "loss": 1.4917,
104
+ "step": 160
105
+ },
106
+ {
107
+ "epoch": 0.43,
108
+ "learning_rate": 2.55e-05,
109
+ "loss": 1.3775,
110
+ "step": 170
111
+ },
112
+ {
113
+ "epoch": 0.45,
114
+ "learning_rate": 2.7000000000000002e-05,
115
+ "loss": 1.4123,
116
+ "step": 180
117
+ },
118
+ {
119
+ "epoch": 0.48,
120
+ "learning_rate": 2.8349999999999998e-05,
121
+ "loss": 1.3893,
122
+ "step": 190
123
+ },
124
+ {
125
+ "epoch": 0.5,
126
+ "learning_rate": 2.985e-05,
127
+ "loss": 1.3646,
128
+ "step": 200
129
+ },
130
+ {
131
+ "epoch": 0.53,
132
+ "learning_rate": 2.984958217270195e-05,
133
+ "loss": 1.3373,
134
+ "step": 210
135
+ },
136
+ {
137
+ "epoch": 0.55,
138
+ "learning_rate": 2.9682451253481892e-05,
139
+ "loss": 1.2649,
140
+ "step": 220
141
+ },
142
+ {
143
+ "epoch": 0.58,
144
+ "learning_rate": 2.951532033426184e-05,
145
+ "loss": 1.2243,
146
+ "step": 230
147
+ },
148
+ {
149
+ "epoch": 0.6,
150
+ "learning_rate": 2.9348189415041785e-05,
151
+ "loss": 1.1601,
152
+ "step": 240
153
+ },
154
+ {
155
+ "epoch": 0.63,
156
+ "learning_rate": 2.9181058495821726e-05,
157
+ "loss": 1.0979,
158
+ "step": 250
159
+ },
160
+ {
161
+ "epoch": 0.65,
162
+ "learning_rate": 2.901392757660167e-05,
163
+ "loss": 1.0832,
164
+ "step": 260
165
+ },
166
+ {
167
+ "epoch": 0.68,
168
+ "learning_rate": 2.884679665738162e-05,
169
+ "loss": 1.0624,
170
+ "step": 270
171
+ },
172
+ {
173
+ "epoch": 0.7,
174
+ "learning_rate": 2.867966573816156e-05,
175
+ "loss": 0.9944,
176
+ "step": 280
177
+ },
178
+ {
179
+ "epoch": 0.73,
180
+ "learning_rate": 2.8512534818941504e-05,
181
+ "loss": 0.9941,
182
+ "step": 290
183
+ },
184
+ {
185
+ "epoch": 0.75,
186
+ "learning_rate": 2.834540389972145e-05,
187
+ "loss": 0.9083,
188
+ "step": 300
189
+ },
190
+ {
191
+ "epoch": 0.78,
192
+ "learning_rate": 2.8178272980501393e-05,
193
+ "loss": 0.9283,
194
+ "step": 310
195
+ },
196
+ {
197
+ "epoch": 0.8,
198
+ "learning_rate": 2.8011142061281338e-05,
199
+ "loss": 0.8759,
200
+ "step": 320
201
+ },
202
+ {
203
+ "epoch": 0.83,
204
+ "learning_rate": 2.7844011142061283e-05,
205
+ "loss": 0.8799,
206
+ "step": 330
207
+ },
208
+ {
209
+ "epoch": 0.85,
210
+ "learning_rate": 2.7676880222841227e-05,
211
+ "loss": 0.8552,
212
+ "step": 340
213
+ },
214
+ {
215
+ "epoch": 0.88,
216
+ "learning_rate": 2.750974930362117e-05,
217
+ "loss": 0.7627,
218
+ "step": 350
219
+ },
220
+ {
221
+ "epoch": 0.9,
222
+ "learning_rate": 2.7342618384401116e-05,
223
+ "loss": 0.7589,
224
+ "step": 360
225
+ },
226
+ {
227
+ "epoch": 0.93,
228
+ "learning_rate": 2.717548746518106e-05,
229
+ "loss": 0.763,
230
+ "step": 370
231
+ },
232
+ {
233
+ "epoch": 0.95,
234
+ "learning_rate": 2.7008356545961002e-05,
235
+ "loss": 0.6831,
236
+ "step": 380
237
+ },
238
+ {
239
+ "epoch": 0.98,
240
+ "learning_rate": 2.6841225626740946e-05,
241
+ "loss": 0.6324,
242
+ "step": 390
243
+ },
244
+ {
245
+ "epoch": 1.0,
246
+ "eval_accuracy": 0.9626360694321859,
247
+ "eval_loss": 0.4707280397415161,
248
+ "eval_runtime": 5.218,
249
+ "eval_samples_per_second": 1302.81,
250
+ "eval_steps_per_second": 40.821,
251
+ "step": 399
252
+ },
253
+ {
254
+ "epoch": 1.0,
255
+ "learning_rate": 2.6674094707520894e-05,
256
+ "loss": 0.6469,
257
+ "step": 400
258
+ },
259
+ {
260
+ "epoch": 1.03,
261
+ "learning_rate": 2.6506963788300836e-05,
262
+ "loss": 0.6207,
263
+ "step": 410
264
+ },
265
+ {
266
+ "epoch": 1.05,
267
+ "learning_rate": 2.633983286908078e-05,
268
+ "loss": 0.5212,
269
+ "step": 420
270
+ },
271
+ {
272
+ "epoch": 1.08,
273
+ "learning_rate": 2.6172701949860725e-05,
274
+ "loss": 0.5533,
275
+ "step": 430
276
+ },
277
+ {
278
+ "epoch": 1.1,
279
+ "learning_rate": 2.600557103064067e-05,
280
+ "loss": 0.5268,
281
+ "step": 440
282
+ },
283
+ {
284
+ "epoch": 1.13,
285
+ "learning_rate": 2.5838440111420614e-05,
286
+ "loss": 0.499,
287
+ "step": 450
288
+ },
289
+ {
290
+ "epoch": 1.15,
291
+ "learning_rate": 2.567130919220056e-05,
292
+ "loss": 0.4704,
293
+ "step": 460
294
+ },
295
+ {
296
+ "epoch": 1.18,
297
+ "learning_rate": 2.55041782729805e-05,
298
+ "loss": 0.4469,
299
+ "step": 470
300
+ },
301
+ {
302
+ "epoch": 1.2,
303
+ "learning_rate": 2.5337047353760448e-05,
304
+ "loss": 0.468,
305
+ "step": 480
306
+ },
307
+ {
308
+ "epoch": 1.23,
309
+ "learning_rate": 2.5169916434540392e-05,
310
+ "loss": 0.406,
311
+ "step": 490
312
+ },
313
+ {
314
+ "epoch": 1.25,
315
+ "learning_rate": 2.5002785515320333e-05,
316
+ "loss": 0.3866,
317
+ "step": 500
318
+ },
319
+ {
320
+ "epoch": 1.28,
321
+ "learning_rate": 2.4835654596100278e-05,
322
+ "loss": 0.4289,
323
+ "step": 510
324
+ },
325
+ {
326
+ "epoch": 1.3,
327
+ "learning_rate": 2.4668523676880226e-05,
328
+ "loss": 0.366,
329
+ "step": 520
330
+ },
331
+ {
332
+ "epoch": 1.33,
333
+ "learning_rate": 2.4501392757660167e-05,
334
+ "loss": 0.408,
335
+ "step": 530
336
+ },
337
+ {
338
+ "epoch": 1.35,
339
+ "learning_rate": 2.433426183844011e-05,
340
+ "loss": 0.3556,
341
+ "step": 540
342
+ },
343
+ {
344
+ "epoch": 1.38,
345
+ "learning_rate": 2.4167130919220056e-05,
346
+ "loss": 0.3739,
347
+ "step": 550
348
+ },
349
+ {
350
+ "epoch": 1.4,
351
+ "learning_rate": 2.4e-05,
352
+ "loss": 0.3544,
353
+ "step": 560
354
+ },
355
+ {
356
+ "epoch": 1.43,
357
+ "learning_rate": 2.3832869080779945e-05,
358
+ "loss": 0.318,
359
+ "step": 570
360
+ },
361
+ {
362
+ "epoch": 1.45,
363
+ "learning_rate": 2.366573816155989e-05,
364
+ "loss": 0.3164,
365
+ "step": 580
366
+ },
367
+ {
368
+ "epoch": 1.48,
369
+ "learning_rate": 2.349860724233983e-05,
370
+ "loss": 0.3514,
371
+ "step": 590
372
+ },
373
+ {
374
+ "epoch": 1.5,
375
+ "learning_rate": 2.333147632311978e-05,
376
+ "loss": 0.3087,
377
+ "step": 600
378
+ },
379
+ {
380
+ "epoch": 1.53,
381
+ "learning_rate": 2.3164345403899723e-05,
382
+ "loss": 0.3007,
383
+ "step": 610
384
+ },
385
+ {
386
+ "epoch": 1.55,
387
+ "learning_rate": 2.2997214484679665e-05,
388
+ "loss": 0.3113,
389
+ "step": 620
390
+ },
391
+ {
392
+ "epoch": 1.58,
393
+ "learning_rate": 2.283008356545961e-05,
394
+ "loss": 0.3276,
395
+ "step": 630
396
+ },
397
+ {
398
+ "epoch": 1.6,
399
+ "learning_rate": 2.2662952646239557e-05,
400
+ "loss": 0.2916,
401
+ "step": 640
402
+ },
403
+ {
404
+ "epoch": 1.63,
405
+ "learning_rate": 2.2495821727019498e-05,
406
+ "loss": 0.2433,
407
+ "step": 650
408
+ },
409
+ {
410
+ "epoch": 1.65,
411
+ "learning_rate": 2.2328690807799443e-05,
412
+ "loss": 0.2566,
413
+ "step": 660
414
+ },
415
+ {
416
+ "epoch": 1.68,
417
+ "learning_rate": 2.2161559888579387e-05,
418
+ "loss": 0.2889,
419
+ "step": 670
420
+ },
421
+ {
422
+ "epoch": 1.7,
423
+ "learning_rate": 2.1994428969359335e-05,
424
+ "loss": 0.2469,
425
+ "step": 680
426
+ },
427
+ {
428
+ "epoch": 1.73,
429
+ "learning_rate": 2.1827298050139276e-05,
430
+ "loss": 0.2597,
431
+ "step": 690
432
+ },
433
+ {
434
+ "epoch": 1.75,
435
+ "learning_rate": 2.166016713091922e-05,
436
+ "loss": 0.2915,
437
+ "step": 700
438
+ },
439
+ {
440
+ "epoch": 1.78,
441
+ "learning_rate": 2.1493036211699166e-05,
442
+ "loss": 0.2798,
443
+ "step": 710
444
+ },
445
+ {
446
+ "epoch": 1.8,
447
+ "learning_rate": 2.1325905292479107e-05,
448
+ "loss": 0.2484,
449
+ "step": 720
450
+ },
451
+ {
452
+ "epoch": 1.83,
453
+ "learning_rate": 2.1158774373259055e-05,
454
+ "loss": 0.2467,
455
+ "step": 730
456
+ },
457
+ {
458
+ "epoch": 1.85,
459
+ "learning_rate": 2.0991643454039e-05,
460
+ "loss": 0.2608,
461
+ "step": 740
462
+ },
463
+ {
464
+ "epoch": 1.88,
465
+ "learning_rate": 2.082451253481894e-05,
466
+ "loss": 0.2467,
467
+ "step": 750
468
+ },
469
+ {
470
+ "epoch": 1.9,
471
+ "learning_rate": 2.0657381615598885e-05,
472
+ "loss": 0.245,
473
+ "step": 760
474
+ },
475
+ {
476
+ "epoch": 1.93,
477
+ "learning_rate": 2.0490250696378833e-05,
478
+ "loss": 0.2417,
479
+ "step": 770
480
+ },
481
+ {
482
+ "epoch": 1.95,
483
+ "learning_rate": 2.0323119777158774e-05,
484
+ "loss": 0.2396,
485
+ "step": 780
486
+ },
487
+ {
488
+ "epoch": 1.98,
489
+ "learning_rate": 2.015598885793872e-05,
490
+ "loss": 0.2367,
491
+ "step": 790
492
+ },
493
+ {
494
+ "epoch": 2.0,
495
+ "eval_accuracy": 0.9741100323624595,
496
+ "eval_loss": 0.13766124844551086,
497
+ "eval_runtime": 5.1822,
498
+ "eval_samples_per_second": 1311.81,
499
+ "eval_steps_per_second": 41.103,
500
+ "step": 798
501
+ },
502
+ {
503
+ "epoch": 2.0,
504
+ "learning_rate": 1.9988857938718663e-05,
505
+ "loss": 0.2507,
506
+ "step": 800
507
+ },
508
+ {
509
+ "epoch": 2.03,
510
+ "learning_rate": 1.9821727019498608e-05,
511
+ "loss": 0.2567,
512
+ "step": 810
513
+ },
514
+ {
515
+ "epoch": 2.05,
516
+ "learning_rate": 1.9654596100278552e-05,
517
+ "loss": 0.2209,
518
+ "step": 820
519
+ },
520
+ {
521
+ "epoch": 2.08,
522
+ "learning_rate": 1.9487465181058497e-05,
523
+ "loss": 0.2021,
524
+ "step": 830
525
+ },
526
+ {
527
+ "epoch": 2.1,
528
+ "learning_rate": 1.9320334261838438e-05,
529
+ "loss": 0.1968,
530
+ "step": 840
531
+ },
532
+ {
533
+ "epoch": 2.13,
534
+ "learning_rate": 1.9153203342618386e-05,
535
+ "loss": 0.1992,
536
+ "step": 850
537
+ },
538
+ {
539
+ "epoch": 2.15,
540
+ "learning_rate": 1.898607242339833e-05,
541
+ "loss": 0.2283,
542
+ "step": 860
543
+ },
544
+ {
545
+ "epoch": 2.18,
546
+ "learning_rate": 1.8818941504178272e-05,
547
+ "loss": 0.201,
548
+ "step": 870
549
+ },
550
+ {
551
+ "epoch": 2.2,
552
+ "learning_rate": 1.8651810584958216e-05,
553
+ "loss": 0.203,
554
+ "step": 880
555
+ },
556
+ {
557
+ "epoch": 2.23,
558
+ "learning_rate": 1.8484679665738164e-05,
559
+ "loss": 0.1997,
560
+ "step": 890
561
+ },
562
+ {
563
+ "epoch": 2.25,
564
+ "learning_rate": 1.8317548746518105e-05,
565
+ "loss": 0.2453,
566
+ "step": 900
567
+ },
568
+ {
569
+ "epoch": 2.28,
570
+ "learning_rate": 1.815041782729805e-05,
571
+ "loss": 0.1963,
572
+ "step": 910
573
+ },
574
+ {
575
+ "epoch": 2.3,
576
+ "learning_rate": 1.7983286908077995e-05,
577
+ "loss": 0.1799,
578
+ "step": 920
579
+ },
580
+ {
581
+ "epoch": 2.33,
582
+ "learning_rate": 1.781615598885794e-05,
583
+ "loss": 0.179,
584
+ "step": 930
585
+ },
586
+ {
587
+ "epoch": 2.35,
588
+ "learning_rate": 1.7649025069637884e-05,
589
+ "loss": 0.1908,
590
+ "step": 940
591
+ },
592
+ {
593
+ "epoch": 2.38,
594
+ "learning_rate": 1.7481894150417828e-05,
595
+ "loss": 0.2037,
596
+ "step": 950
597
+ },
598
+ {
599
+ "epoch": 2.4,
600
+ "learning_rate": 1.731476323119777e-05,
601
+ "loss": 0.1902,
602
+ "step": 960
603
+ },
604
+ {
605
+ "epoch": 2.43,
606
+ "learning_rate": 1.7147632311977717e-05,
607
+ "loss": 0.1849,
608
+ "step": 970
609
+ },
610
+ {
611
+ "epoch": 2.45,
612
+ "learning_rate": 1.6980501392757662e-05,
613
+ "loss": 0.2228,
614
+ "step": 980
615
+ },
616
+ {
617
+ "epoch": 2.48,
618
+ "learning_rate": 1.6813370473537606e-05,
619
+ "loss": 0.1879,
620
+ "step": 990
621
+ },
622
+ {
623
+ "epoch": 2.5,
624
+ "learning_rate": 1.6646239554317548e-05,
625
+ "loss": 0.186,
626
+ "step": 1000
627
+ },
628
+ {
629
+ "epoch": 2.53,
630
+ "learning_rate": 1.6479108635097496e-05,
631
+ "loss": 0.2408,
632
+ "step": 1010
633
+ },
634
+ {
635
+ "epoch": 2.55,
636
+ "learning_rate": 1.631197771587744e-05,
637
+ "loss": 0.2052,
638
+ "step": 1020
639
+ },
640
+ {
641
+ "epoch": 2.58,
642
+ "learning_rate": 1.614484679665738e-05,
643
+ "loss": 0.1984,
644
+ "step": 1030
645
+ },
646
+ {
647
+ "epoch": 2.6,
648
+ "learning_rate": 1.5977715877437326e-05,
649
+ "loss": 0.1707,
650
+ "step": 1040
651
+ },
652
+ {
653
+ "epoch": 2.63,
654
+ "learning_rate": 1.581058495821727e-05,
655
+ "loss": 0.178,
656
+ "step": 1050
657
+ },
658
+ {
659
+ "epoch": 2.65,
660
+ "learning_rate": 1.5643454038997215e-05,
661
+ "loss": 0.1721,
662
+ "step": 1060
663
+ },
664
+ {
665
+ "epoch": 2.68,
666
+ "learning_rate": 1.547632311977716e-05,
667
+ "loss": 0.188,
668
+ "step": 1070
669
+ },
670
+ {
671
+ "epoch": 2.71,
672
+ "learning_rate": 1.5309192200557104e-05,
673
+ "loss": 0.1876,
674
+ "step": 1080
675
+ },
676
+ {
677
+ "epoch": 2.73,
678
+ "learning_rate": 1.5142061281337047e-05,
679
+ "loss": 0.1792,
680
+ "step": 1090
681
+ },
682
+ {
683
+ "epoch": 2.76,
684
+ "learning_rate": 1.4974930362116992e-05,
685
+ "loss": 0.182,
686
+ "step": 1100
687
+ },
688
+ {
689
+ "epoch": 2.78,
690
+ "learning_rate": 1.4807799442896936e-05,
691
+ "loss": 0.228,
692
+ "step": 1110
693
+ },
694
+ {
695
+ "epoch": 2.81,
696
+ "learning_rate": 1.464066852367688e-05,
697
+ "loss": 0.1858,
698
+ "step": 1120
699
+ },
700
+ {
701
+ "epoch": 2.83,
702
+ "learning_rate": 1.4473537604456825e-05,
703
+ "loss": 0.1908,
704
+ "step": 1130
705
+ },
706
+ {
707
+ "epoch": 2.86,
708
+ "learning_rate": 1.4306406685236768e-05,
709
+ "loss": 0.1689,
710
+ "step": 1140
711
+ },
712
+ {
713
+ "epoch": 2.88,
714
+ "learning_rate": 1.4139275766016714e-05,
715
+ "loss": 0.1797,
716
+ "step": 1150
717
+ },
718
+ {
719
+ "epoch": 2.91,
720
+ "learning_rate": 1.3972144846796657e-05,
721
+ "loss": 0.1665,
722
+ "step": 1160
723
+ },
724
+ {
725
+ "epoch": 2.93,
726
+ "learning_rate": 1.3805013927576602e-05,
727
+ "loss": 0.1914,
728
+ "step": 1170
729
+ },
730
+ {
731
+ "epoch": 2.96,
732
+ "learning_rate": 1.3637883008356546e-05,
733
+ "loss": 0.1706,
734
+ "step": 1180
735
+ },
736
+ {
737
+ "epoch": 2.98,
738
+ "learning_rate": 1.3470752089136491e-05,
739
+ "loss": 0.1685,
740
+ "step": 1190
741
+ },
742
+ {
743
+ "epoch": 3.0,
744
+ "eval_accuracy": 0.9788172992056487,
745
+ "eval_loss": 0.1008954793214798,
746
+ "eval_runtime": 5.2829,
747
+ "eval_samples_per_second": 1286.803,
748
+ "eval_steps_per_second": 40.319,
749
+ "step": 1197
750
+ },
751
+ {
752
+ "epoch": 3.01,
753
+ "learning_rate": 1.3303621169916434e-05,
754
+ "loss": 0.1766,
755
+ "step": 1200
756
+ },
757
+ {
758
+ "epoch": 3.03,
759
+ "learning_rate": 1.313649025069638e-05,
760
+ "loss": 0.1852,
761
+ "step": 1210
762
+ },
763
+ {
764
+ "epoch": 3.06,
765
+ "learning_rate": 1.2969359331476323e-05,
766
+ "loss": 0.1602,
767
+ "step": 1220
768
+ },
769
+ {
770
+ "epoch": 3.08,
771
+ "learning_rate": 1.2802228412256267e-05,
772
+ "loss": 0.1492,
773
+ "step": 1230
774
+ },
775
+ {
776
+ "epoch": 3.11,
777
+ "learning_rate": 1.2635097493036212e-05,
778
+ "loss": 0.1385,
779
+ "step": 1240
780
+ },
781
+ {
782
+ "epoch": 3.13,
783
+ "learning_rate": 1.2467966573816157e-05,
784
+ "loss": 0.2077,
785
+ "step": 1250
786
+ },
787
+ {
788
+ "epoch": 3.16,
789
+ "learning_rate": 1.2300835654596101e-05,
790
+ "loss": 0.1484,
791
+ "step": 1260
792
+ },
793
+ {
794
+ "epoch": 3.18,
795
+ "learning_rate": 1.2133704735376046e-05,
796
+ "loss": 0.1479,
797
+ "step": 1270
798
+ },
799
+ {
800
+ "epoch": 3.21,
801
+ "learning_rate": 1.1966573816155989e-05,
802
+ "loss": 0.1695,
803
+ "step": 1280
804
+ },
805
+ {
806
+ "epoch": 3.23,
807
+ "learning_rate": 1.1799442896935935e-05,
808
+ "loss": 0.1682,
809
+ "step": 1290
810
+ },
811
+ {
812
+ "epoch": 3.26,
813
+ "learning_rate": 1.1632311977715878e-05,
814
+ "loss": 0.1538,
815
+ "step": 1300
816
+ },
817
+ {
818
+ "epoch": 3.28,
819
+ "learning_rate": 1.1465181058495822e-05,
820
+ "loss": 0.1565,
821
+ "step": 1310
822
+ },
823
+ {
824
+ "epoch": 3.31,
825
+ "learning_rate": 1.1298050139275767e-05,
826
+ "loss": 0.168,
827
+ "step": 1320
828
+ },
829
+ {
830
+ "epoch": 3.33,
831
+ "learning_rate": 1.1130919220055711e-05,
832
+ "loss": 0.1762,
833
+ "step": 1330
834
+ },
835
+ {
836
+ "epoch": 3.36,
837
+ "learning_rate": 1.0963788300835654e-05,
838
+ "loss": 0.1653,
839
+ "step": 1340
840
+ },
841
+ {
842
+ "epoch": 3.38,
843
+ "learning_rate": 1.0796657381615599e-05,
844
+ "loss": 0.1748,
845
+ "step": 1350
846
+ },
847
+ {
848
+ "epoch": 3.41,
849
+ "learning_rate": 1.0629526462395543e-05,
850
+ "loss": 0.1555,
851
+ "step": 1360
852
+ },
853
+ {
854
+ "epoch": 3.43,
855
+ "learning_rate": 1.0462395543175486e-05,
856
+ "loss": 0.1451,
857
+ "step": 1370
858
+ },
859
+ {
860
+ "epoch": 3.46,
861
+ "learning_rate": 1.0295264623955432e-05,
862
+ "loss": 0.1546,
863
+ "step": 1380
864
+ },
865
+ {
866
+ "epoch": 3.48,
867
+ "learning_rate": 1.0128133704735375e-05,
868
+ "loss": 0.1292,
869
+ "step": 1390
870
+ },
871
+ {
872
+ "epoch": 3.51,
873
+ "learning_rate": 9.961002785515322e-06,
874
+ "loss": 0.1152,
875
+ "step": 1400
876
+ },
877
+ {
878
+ "epoch": 3.53,
879
+ "learning_rate": 9.793871866295264e-06,
880
+ "loss": 0.1341,
881
+ "step": 1410
882
+ },
883
+ {
884
+ "epoch": 3.56,
885
+ "learning_rate": 9.626740947075209e-06,
886
+ "loss": 0.1535,
887
+ "step": 1420
888
+ },
889
+ {
890
+ "epoch": 3.58,
891
+ "learning_rate": 9.459610027855154e-06,
892
+ "loss": 0.1696,
893
+ "step": 1430
894
+ },
895
+ {
896
+ "epoch": 3.61,
897
+ "learning_rate": 9.292479108635098e-06,
898
+ "loss": 0.1472,
899
+ "step": 1440
900
+ },
901
+ {
902
+ "epoch": 3.63,
903
+ "learning_rate": 9.125348189415041e-06,
904
+ "loss": 0.1653,
905
+ "step": 1450
906
+ },
907
+ {
908
+ "epoch": 3.66,
909
+ "learning_rate": 8.958217270194987e-06,
910
+ "loss": 0.1567,
911
+ "step": 1460
912
+ },
913
+ {
914
+ "epoch": 3.68,
915
+ "learning_rate": 8.79108635097493e-06,
916
+ "loss": 0.1468,
917
+ "step": 1470
918
+ },
919
+ {
920
+ "epoch": 3.71,
921
+ "learning_rate": 8.623955431754875e-06,
922
+ "loss": 0.147,
923
+ "step": 1480
924
+ },
925
+ {
926
+ "epoch": 3.73,
927
+ "learning_rate": 8.45682451253482e-06,
928
+ "loss": 0.132,
929
+ "step": 1490
930
+ },
931
+ {
932
+ "epoch": 3.76,
933
+ "learning_rate": 8.289693593314764e-06,
934
+ "loss": 0.1695,
935
+ "step": 1500
936
+ },
937
+ {
938
+ "epoch": 3.78,
939
+ "learning_rate": 8.122562674094707e-06,
940
+ "loss": 0.1632,
941
+ "step": 1510
942
+ },
943
+ {
944
+ "epoch": 3.81,
945
+ "learning_rate": 7.955431754874653e-06,
946
+ "loss": 0.1513,
947
+ "step": 1520
948
+ },
949
+ {
950
+ "epoch": 3.83,
951
+ "learning_rate": 7.788300835654596e-06,
952
+ "loss": 0.1703,
953
+ "step": 1530
954
+ },
955
+ {
956
+ "epoch": 3.86,
957
+ "learning_rate": 7.621169916434541e-06,
958
+ "loss": 0.1257,
959
+ "step": 1540
960
+ },
961
+ {
962
+ "epoch": 3.88,
963
+ "learning_rate": 7.454038997214485e-06,
964
+ "loss": 0.1792,
965
+ "step": 1550
966
+ },
967
+ {
968
+ "epoch": 3.91,
969
+ "learning_rate": 7.2869080779944286e-06,
970
+ "loss": 0.171,
971
+ "step": 1560
972
+ },
973
+ {
974
+ "epoch": 3.93,
975
+ "learning_rate": 7.119777158774373e-06,
976
+ "loss": 0.163,
977
+ "step": 1570
978
+ },
979
+ {
980
+ "epoch": 3.96,
981
+ "learning_rate": 6.952646239554318e-06,
982
+ "loss": 0.1427,
983
+ "step": 1580
984
+ },
985
+ {
986
+ "epoch": 3.98,
987
+ "learning_rate": 6.785515320334261e-06,
988
+ "loss": 0.1091,
989
+ "step": 1590
990
+ },
991
+ {
992
+ "epoch": 4.0,
993
+ "eval_accuracy": 0.9791115033833481,
994
+ "eval_loss": 0.09216145426034927,
995
+ "eval_runtime": 5.0576,
996
+ "eval_samples_per_second": 1344.125,
997
+ "eval_steps_per_second": 42.115,
998
+ "step": 1597
999
+ },
1000
+ {
1001
+ "epoch": 4.01,
1002
+ "learning_rate": 6.618384401114206e-06,
1003
+ "loss": 0.1405,
1004
+ "step": 1600
1005
+ },
1006
+ {
1007
+ "epoch": 4.03,
1008
+ "learning_rate": 6.4512534818941505e-06,
1009
+ "loss": 0.1493,
1010
+ "step": 1610
1011
+ },
1012
+ {
1013
+ "epoch": 4.06,
1014
+ "learning_rate": 6.284122562674095e-06,
1015
+ "loss": 0.1404,
1016
+ "step": 1620
1017
+ },
1018
+ {
1019
+ "epoch": 4.08,
1020
+ "learning_rate": 6.116991643454039e-06,
1021
+ "loss": 0.1462,
1022
+ "step": 1630
1023
+ },
1024
+ {
1025
+ "epoch": 4.11,
1026
+ "learning_rate": 5.949860724233983e-06,
1027
+ "loss": 0.1531,
1028
+ "step": 1640
1029
+ },
1030
+ {
1031
+ "epoch": 4.13,
1032
+ "learning_rate": 5.782729805013928e-06,
1033
+ "loss": 0.1229,
1034
+ "step": 1650
1035
+ },
1036
+ {
1037
+ "epoch": 4.16,
1038
+ "learning_rate": 5.615598885793872e-06,
1039
+ "loss": 0.157,
1040
+ "step": 1660
1041
+ },
1042
+ {
1043
+ "epoch": 4.18,
1044
+ "learning_rate": 5.448467966573816e-06,
1045
+ "loss": 0.1508,
1046
+ "step": 1670
1047
+ },
1048
+ {
1049
+ "epoch": 4.21,
1050
+ "learning_rate": 5.281337047353761e-06,
1051
+ "loss": 0.1298,
1052
+ "step": 1680
1053
+ },
1054
+ {
1055
+ "epoch": 4.23,
1056
+ "learning_rate": 5.114206128133705e-06,
1057
+ "loss": 0.1359,
1058
+ "step": 1690
1059
+ },
1060
+ {
1061
+ "epoch": 4.26,
1062
+ "learning_rate": 4.947075208913649e-06,
1063
+ "loss": 0.1389,
1064
+ "step": 1700
1065
+ },
1066
+ {
1067
+ "epoch": 4.28,
1068
+ "learning_rate": 4.7799442896935936e-06,
1069
+ "loss": 0.1285,
1070
+ "step": 1710
1071
+ },
1072
+ {
1073
+ "epoch": 4.31,
1074
+ "learning_rate": 4.612813370473538e-06,
1075
+ "loss": 0.169,
1076
+ "step": 1720
1077
+ },
1078
+ {
1079
+ "epoch": 4.33,
1080
+ "learning_rate": 4.445682451253482e-06,
1081
+ "loss": 0.1267,
1082
+ "step": 1730
1083
+ },
1084
+ {
1085
+ "epoch": 4.36,
1086
+ "learning_rate": 4.278551532033426e-06,
1087
+ "loss": 0.1429,
1088
+ "step": 1740
1089
+ },
1090
+ {
1091
+ "epoch": 4.38,
1092
+ "learning_rate": 4.111420612813371e-06,
1093
+ "loss": 0.1394,
1094
+ "step": 1750
1095
+ },
1096
+ {
1097
+ "epoch": 4.41,
1098
+ "learning_rate": 3.944289693593315e-06,
1099
+ "loss": 0.1446,
1100
+ "step": 1760
1101
+ },
1102
+ {
1103
+ "epoch": 4.43,
1104
+ "learning_rate": 3.7771587743732592e-06,
1105
+ "loss": 0.1559,
1106
+ "step": 1770
1107
+ },
1108
+ {
1109
+ "epoch": 4.46,
1110
+ "learning_rate": 3.6100278551532034e-06,
1111
+ "loss": 0.1332,
1112
+ "step": 1780
1113
+ },
1114
+ {
1115
+ "epoch": 4.48,
1116
+ "learning_rate": 3.4428969359331475e-06,
1117
+ "loss": 0.1545,
1118
+ "step": 1790
1119
+ },
1120
+ {
1121
+ "epoch": 4.51,
1122
+ "learning_rate": 3.275766016713092e-06,
1123
+ "loss": 0.1185,
1124
+ "step": 1800
1125
+ },
1126
+ {
1127
+ "epoch": 4.53,
1128
+ "learning_rate": 3.108635097493036e-06,
1129
+ "loss": 0.1313,
1130
+ "step": 1810
1131
+ },
1132
+ {
1133
+ "epoch": 4.56,
1134
+ "learning_rate": 2.9415041782729803e-06,
1135
+ "loss": 0.1348,
1136
+ "step": 1820
1137
+ },
1138
+ {
1139
+ "epoch": 4.58,
1140
+ "learning_rate": 2.774373259052925e-06,
1141
+ "loss": 0.1158,
1142
+ "step": 1830
1143
+ },
1144
+ {
1145
+ "epoch": 4.61,
1146
+ "learning_rate": 2.607242339832869e-06,
1147
+ "loss": 0.1306,
1148
+ "step": 1840
1149
+ },
1150
+ {
1151
+ "epoch": 4.63,
1152
+ "learning_rate": 2.4401114206128136e-06,
1153
+ "loss": 0.1407,
1154
+ "step": 1850
1155
+ },
1156
+ {
1157
+ "epoch": 4.66,
1158
+ "learning_rate": 2.2729805013927577e-06,
1159
+ "loss": 0.1462,
1160
+ "step": 1860
1161
+ },
1162
+ {
1163
+ "epoch": 4.68,
1164
+ "learning_rate": 2.1058495821727023e-06,
1165
+ "loss": 0.1356,
1166
+ "step": 1870
1167
+ },
1168
+ {
1169
+ "epoch": 4.71,
1170
+ "learning_rate": 1.9387186629526464e-06,
1171
+ "loss": 0.1376,
1172
+ "step": 1880
1173
+ },
1174
+ {
1175
+ "epoch": 4.73,
1176
+ "learning_rate": 1.7715877437325906e-06,
1177
+ "loss": 0.1217,
1178
+ "step": 1890
1179
+ },
1180
+ {
1181
+ "epoch": 4.76,
1182
+ "learning_rate": 1.604456824512535e-06,
1183
+ "loss": 0.156,
1184
+ "step": 1900
1185
+ },
1186
+ {
1187
+ "epoch": 4.78,
1188
+ "learning_rate": 1.437325905292479e-06,
1189
+ "loss": 0.1365,
1190
+ "step": 1910
1191
+ },
1192
+ {
1193
+ "epoch": 4.81,
1194
+ "learning_rate": 1.2701949860724234e-06,
1195
+ "loss": 0.1381,
1196
+ "step": 1920
1197
+ },
1198
+ {
1199
+ "epoch": 4.83,
1200
+ "learning_rate": 1.1030640668523677e-06,
1201
+ "loss": 0.1114,
1202
+ "step": 1930
1203
+ },
1204
+ {
1205
+ "epoch": 4.86,
1206
+ "learning_rate": 9.35933147632312e-07,
1207
+ "loss": 0.1233,
1208
+ "step": 1940
1209
+ },
1210
+ {
1211
+ "epoch": 4.88,
1212
+ "learning_rate": 7.688022284122563e-07,
1213
+ "loss": 0.1034,
1214
+ "step": 1950
1215
+ },
1216
+ {
1217
+ "epoch": 4.91,
1218
+ "learning_rate": 6.016713091922006e-07,
1219
+ "loss": 0.1258,
1220
+ "step": 1960
1221
+ },
1222
+ {
1223
+ "epoch": 4.93,
1224
+ "learning_rate": 4.3454038997214486e-07,
1225
+ "loss": 0.1236,
1226
+ "step": 1970
1227
+ },
1228
+ {
1229
+ "epoch": 4.96,
1230
+ "learning_rate": 2.6740947075208915e-07,
1231
+ "loss": 0.1206,
1232
+ "step": 1980
1233
+ },
1234
+ {
1235
+ "epoch": 4.98,
1236
+ "learning_rate": 1.0027855153203343e-07,
1237
+ "loss": 0.1297,
1238
+ "step": 1990
1239
+ },
1240
+ {
1241
+ "epoch": 5.0,
1242
+ "eval_accuracy": 0.980435422182995,
1243
+ "eval_loss": 0.08353430777788162,
1244
+ "eval_runtime": 5.4193,
1245
+ "eval_samples_per_second": 1254.406,
1246
+ "eval_steps_per_second": 39.304,
1247
+ "step": 1995
1248
+ },
1249
+ {
1250
+ "epoch": 5.0,
1251
+ "step": 1995,
1252
+ "total_flos": 2.3180194781952e+18,
1253
+ "train_loss": 0.4383530242102487,
1254
+ "train_runtime": 692.5799,
1255
+ "train_samples_per_second": 368.867,
1256
+ "train_steps_per_second": 2.881
1257
+ }
1258
+ ],
1259
+ "max_steps": 1995,
1260
+ "num_train_epochs": 5,
1261
+ "total_flos": 2.3180194781952e+18,
1262
+ "trial_name": null,
1263
+ "trial_params": null
1264
+ }