infinitejoy commited on
Commit
48d4f70
1 Parent(s): 26d583d

End of training

Browse files
Files changed (4) hide show
  1. all_results.json +14 -0
  2. eval_results.json +9 -0
  3. train_results.json +8 -0
  4. trainer_state.json +493 -0
all_results.json ADDED
@@ -0,0 +1,14 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 500.0,
3
+ "eval_loss": 2.165587902069092,
4
+ "eval_runtime": 5.9037,
5
+ "eval_samples": 142,
6
+ "eval_samples_per_second": 24.053,
7
+ "eval_steps_per_second": 0.847,
8
+ "eval_wer": 0.8352223190932868,
9
+ "train_loss": 0.5668576599160831,
10
+ "train_runtime": 11637.7564,
11
+ "train_samples": 378,
12
+ "train_samples_per_second": 16.24,
13
+ "train_steps_per_second": 0.516
14
+ }
eval_results.json ADDED
@@ -0,0 +1,9 @@
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 500.0,
3
+ "eval_loss": 2.165587902069092,
4
+ "eval_runtime": 5.9037,
5
+ "eval_samples": 142,
6
+ "eval_samples_per_second": 24.053,
7
+ "eval_steps_per_second": 0.847,
8
+ "eval_wer": 0.8352223190932868
9
+ }
train_results.json ADDED
@@ -0,0 +1,8 @@
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 500.0,
3
+ "train_loss": 0.5668576599160831,
4
+ "train_runtime": 11637.7564,
5
+ "train_samples": 378,
6
+ "train_samples_per_second": 16.24,
7
+ "train_steps_per_second": 0.516
8
+ }
trainer_state.json ADDED
@@ -0,0 +1,493 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": null,
3
+ "best_model_checkpoint": null,
4
+ "epoch": 500.0,
5
+ "global_step": 6000,
6
+ "is_hyper_param_search": false,
7
+ "is_local_process_zero": true,
8
+ "is_world_process_zero": true,
9
+ "log_history": [
10
+ {
11
+ "epoch": 8.33,
12
+ "learning_rate": 5.82e-05,
13
+ "loss": 11.9523,
14
+ "step": 100
15
+ },
16
+ {
17
+ "epoch": 16.67,
18
+ "learning_rate": 0.0001182,
19
+ "loss": 3.7399,
20
+ "step": 200
21
+ },
22
+ {
23
+ "epoch": 25.0,
24
+ "learning_rate": 0.00017819999999999997,
25
+ "loss": 3.074,
26
+ "step": 300
27
+ },
28
+ {
29
+ "epoch": 33.33,
30
+ "learning_rate": 0.0002382,
31
+ "loss": 2.8995,
32
+ "step": 400
33
+ },
34
+ {
35
+ "epoch": 41.67,
36
+ "learning_rate": 0.0002982,
37
+ "loss": 2.2627,
38
+ "step": 500
39
+ },
40
+ {
41
+ "epoch": 41.67,
42
+ "eval_loss": 1.4175914525985718,
43
+ "eval_runtime": 5.6701,
44
+ "eval_samples_per_second": 25.044,
45
+ "eval_steps_per_second": 0.882,
46
+ "eval_wer": 0.8282476024411508,
47
+ "step": 500
48
+ },
49
+ {
50
+ "epoch": 50.0,
51
+ "learning_rate": 0.0002947090909090909,
52
+ "loss": 1.5352,
53
+ "step": 600
54
+ },
55
+ {
56
+ "epoch": 58.33,
57
+ "learning_rate": 0.00028925454545454543,
58
+ "loss": 1.2177,
59
+ "step": 700
60
+ },
61
+ {
62
+ "epoch": 66.67,
63
+ "learning_rate": 0.00028379999999999996,
64
+ "loss": 0.9821,
65
+ "step": 800
66
+ },
67
+ {
68
+ "epoch": 75.0,
69
+ "learning_rate": 0.0002783454545454545,
70
+ "loss": 0.8097,
71
+ "step": 900
72
+ },
73
+ {
74
+ "epoch": 83.33,
75
+ "learning_rate": 0.0002728909090909091,
76
+ "loss": 0.6692,
77
+ "step": 1000
78
+ },
79
+ {
80
+ "epoch": 83.33,
81
+ "eval_loss": 1.3905961513519287,
82
+ "eval_runtime": 5.5949,
83
+ "eval_samples_per_second": 25.38,
84
+ "eval_steps_per_second": 0.894,
85
+ "eval_wer": 0.7375762859633828,
86
+ "step": 1000
87
+ },
88
+ {
89
+ "epoch": 91.67,
90
+ "learning_rate": 0.0002674363636363636,
91
+ "loss": 0.5587,
92
+ "step": 1100
93
+ },
94
+ {
95
+ "epoch": 100.0,
96
+ "learning_rate": 0.0002619818181818182,
97
+ "loss": 0.4692,
98
+ "step": 1200
99
+ },
100
+ {
101
+ "epoch": 108.33,
102
+ "learning_rate": 0.0002565272727272727,
103
+ "loss": 0.3949,
104
+ "step": 1300
105
+ },
106
+ {
107
+ "epoch": 116.67,
108
+ "learning_rate": 0.00025107272727272725,
109
+ "loss": 0.3484,
110
+ "step": 1400
111
+ },
112
+ {
113
+ "epoch": 125.0,
114
+ "learning_rate": 0.0002456181818181818,
115
+ "loss": 0.2874,
116
+ "step": 1500
117
+ },
118
+ {
119
+ "epoch": 125.0,
120
+ "eval_loss": 1.791098952293396,
121
+ "eval_runtime": 5.4161,
122
+ "eval_samples_per_second": 26.218,
123
+ "eval_steps_per_second": 0.923,
124
+ "eval_wer": 0.7733217088055798,
125
+ "step": 1500
126
+ },
127
+ {
128
+ "epoch": 133.33,
129
+ "learning_rate": 0.00024016363636363634,
130
+ "loss": 0.2444,
131
+ "step": 1600
132
+ },
133
+ {
134
+ "epoch": 141.67,
135
+ "learning_rate": 0.0002347090909090909,
136
+ "loss": 0.2249,
137
+ "step": 1700
138
+ },
139
+ {
140
+ "epoch": 150.0,
141
+ "learning_rate": 0.00022925454545454543,
142
+ "loss": 0.1996,
143
+ "step": 1800
144
+ },
145
+ {
146
+ "epoch": 158.33,
147
+ "learning_rate": 0.0002238,
148
+ "loss": 0.1676,
149
+ "step": 1900
150
+ },
151
+ {
152
+ "epoch": 166.67,
153
+ "learning_rate": 0.00021834545454545452,
154
+ "loss": 0.1504,
155
+ "step": 2000
156
+ },
157
+ {
158
+ "epoch": 166.67,
159
+ "eval_loss": 1.761537790298462,
160
+ "eval_runtime": 5.7426,
161
+ "eval_samples_per_second": 24.728,
162
+ "eval_steps_per_second": 0.871,
163
+ "eval_wer": 0.8169136878814298,
164
+ "step": 2000
165
+ },
166
+ {
167
+ "epoch": 175.0,
168
+ "learning_rate": 0.00021289090909090905,
169
+ "loss": 0.1356,
170
+ "step": 2100
171
+ },
172
+ {
173
+ "epoch": 183.33,
174
+ "learning_rate": 0.0002074363636363636,
175
+ "loss": 0.1207,
176
+ "step": 2200
177
+ },
178
+ {
179
+ "epoch": 191.67,
180
+ "learning_rate": 0.00020198181818181814,
181
+ "loss": 0.119,
182
+ "step": 2300
183
+ },
184
+ {
185
+ "epoch": 200.0,
186
+ "learning_rate": 0.0001965272727272727,
187
+ "loss": 0.1085,
188
+ "step": 2400
189
+ },
190
+ {
191
+ "epoch": 208.33,
192
+ "learning_rate": 0.00019107272727272723,
193
+ "loss": 0.102,
194
+ "step": 2500
195
+ },
196
+ {
197
+ "epoch": 208.33,
198
+ "eval_loss": 1.9240303039550781,
199
+ "eval_runtime": 5.3949,
200
+ "eval_samples_per_second": 26.321,
201
+ "eval_steps_per_second": 0.927,
202
+ "eval_wer": 0.8456843940714909,
203
+ "step": 2500
204
+ },
205
+ {
206
+ "epoch": 216.67,
207
+ "learning_rate": 0.0001856181818181818,
208
+ "loss": 0.0899,
209
+ "step": 2600
210
+ },
211
+ {
212
+ "epoch": 225.0,
213
+ "learning_rate": 0.00018016363636363632,
214
+ "loss": 0.0838,
215
+ "step": 2700
216
+ },
217
+ {
218
+ "epoch": 233.33,
219
+ "learning_rate": 0.0001747090909090909,
220
+ "loss": 0.0825,
221
+ "step": 2800
222
+ },
223
+ {
224
+ "epoch": 241.67,
225
+ "learning_rate": 0.00016925454545454547,
226
+ "loss": 0.0699,
227
+ "step": 2900
228
+ },
229
+ {
230
+ "epoch": 250.0,
231
+ "learning_rate": 0.0001638,
232
+ "loss": 0.071,
233
+ "step": 3000
234
+ },
235
+ {
236
+ "epoch": 250.0,
237
+ "eval_loss": 1.9720656871795654,
238
+ "eval_runtime": 5.5146,
239
+ "eval_samples_per_second": 25.75,
240
+ "eval_steps_per_second": 0.907,
241
+ "eval_wer": 0.7741935483870968,
242
+ "step": 3000
243
+ },
244
+ {
245
+ "epoch": 258.33,
246
+ "learning_rate": 0.00015834545454545456,
247
+ "loss": 0.0616,
248
+ "step": 3100
249
+ },
250
+ {
251
+ "epoch": 266.67,
252
+ "learning_rate": 0.00015289090909090909,
253
+ "loss": 0.0564,
254
+ "step": 3200
255
+ },
256
+ {
257
+ "epoch": 275.0,
258
+ "learning_rate": 0.00014743636363636362,
259
+ "loss": 0.0538,
260
+ "step": 3300
261
+ },
262
+ {
263
+ "epoch": 283.33,
264
+ "learning_rate": 0.00014198181818181818,
265
+ "loss": 0.047,
266
+ "step": 3400
267
+ },
268
+ {
269
+ "epoch": 291.67,
270
+ "learning_rate": 0.0001365272727272727,
271
+ "loss": 0.046,
272
+ "step": 3500
273
+ },
274
+ {
275
+ "epoch": 291.67,
276
+ "eval_loss": 2.0821120738983154,
277
+ "eval_runtime": 5.458,
278
+ "eval_samples_per_second": 26.017,
279
+ "eval_steps_per_second": 0.916,
280
+ "eval_wer": 0.8326068003487358,
281
+ "step": 3500
282
+ },
283
+ {
284
+ "epoch": 300.0,
285
+ "learning_rate": 0.00013107272727272727,
286
+ "loss": 0.0408,
287
+ "step": 3600
288
+ },
289
+ {
290
+ "epoch": 308.33,
291
+ "learning_rate": 0.0001256181818181818,
292
+ "loss": 0.0401,
293
+ "step": 3700
294
+ },
295
+ {
296
+ "epoch": 316.67,
297
+ "learning_rate": 0.00012016363636363635,
298
+ "loss": 0.037,
299
+ "step": 3800
300
+ },
301
+ {
302
+ "epoch": 325.0,
303
+ "learning_rate": 0.0001147090909090909,
304
+ "loss": 0.0362,
305
+ "step": 3900
306
+ },
307
+ {
308
+ "epoch": 333.33,
309
+ "learning_rate": 0.00010925454545454544,
310
+ "loss": 0.0357,
311
+ "step": 4000
312
+ },
313
+ {
314
+ "epoch": 333.33,
315
+ "eval_loss": 2.052865505218506,
316
+ "eval_runtime": 5.4898,
317
+ "eval_samples_per_second": 25.866,
318
+ "eval_steps_per_second": 0.911,
319
+ "eval_wer": 0.8308631211857018,
320
+ "step": 4000
321
+ },
322
+ {
323
+ "epoch": 341.67,
324
+ "learning_rate": 0.00010379999999999999,
325
+ "loss": 0.0336,
326
+ "step": 4100
327
+ },
328
+ {
329
+ "epoch": 350.0,
330
+ "learning_rate": 9.834545454545455e-05,
331
+ "loss": 0.0283,
332
+ "step": 4200
333
+ },
334
+ {
335
+ "epoch": 358.33,
336
+ "learning_rate": 9.289090909090909e-05,
337
+ "loss": 0.0274,
338
+ "step": 4300
339
+ },
340
+ {
341
+ "epoch": 366.67,
342
+ "learning_rate": 8.743636363636364e-05,
343
+ "loss": 0.0256,
344
+ "step": 4400
345
+ },
346
+ {
347
+ "epoch": 375.0,
348
+ "learning_rate": 8.198181818181818e-05,
349
+ "loss": 0.0241,
350
+ "step": 4500
351
+ },
352
+ {
353
+ "epoch": 375.0,
354
+ "eval_loss": 2.0744874477386475,
355
+ "eval_runtime": 5.4725,
356
+ "eval_samples_per_second": 25.948,
357
+ "eval_steps_per_second": 0.914,
358
+ "eval_wer": 0.8029642545771578,
359
+ "step": 4500
360
+ },
361
+ {
362
+ "epoch": 383.33,
363
+ "learning_rate": 7.652727272727273e-05,
364
+ "loss": 0.0266,
365
+ "step": 4600
366
+ },
367
+ {
368
+ "epoch": 391.67,
369
+ "learning_rate": 7.107272727272727e-05,
370
+ "loss": 0.0243,
371
+ "step": 4700
372
+ },
373
+ {
374
+ "epoch": 400.0,
375
+ "learning_rate": 6.561818181818182e-05,
376
+ "loss": 0.0191,
377
+ "step": 4800
378
+ },
379
+ {
380
+ "epoch": 408.33,
381
+ "learning_rate": 6.016363636363636e-05,
382
+ "loss": 0.0193,
383
+ "step": 4900
384
+ },
385
+ {
386
+ "epoch": 416.67,
387
+ "learning_rate": 5.4709090909090905e-05,
388
+ "loss": 0.0165,
389
+ "step": 5000
390
+ },
391
+ {
392
+ "epoch": 416.67,
393
+ "eval_loss": 2.2303550243377686,
394
+ "eval_runtime": 5.5806,
395
+ "eval_samples_per_second": 25.445,
396
+ "eval_steps_per_second": 0.896,
397
+ "eval_wer": 0.8238884045335658,
398
+ "step": 5000
399
+ },
400
+ {
401
+ "epoch": 425.0,
402
+ "learning_rate": 4.925454545454545e-05,
403
+ "loss": 0.0195,
404
+ "step": 5100
405
+ },
406
+ {
407
+ "epoch": 433.33,
408
+ "learning_rate": 4.3799999999999994e-05,
409
+ "loss": 0.0159,
410
+ "step": 5200
411
+ },
412
+ {
413
+ "epoch": 441.67,
414
+ "learning_rate": 3.84e-05,
415
+ "loss": 0.0146,
416
+ "step": 5300
417
+ },
418
+ {
419
+ "epoch": 450.0,
420
+ "learning_rate": 3.294545454545454e-05,
421
+ "loss": 0.0145,
422
+ "step": 5400
423
+ },
424
+ {
425
+ "epoch": 458.33,
426
+ "learning_rate": 2.749090909090909e-05,
427
+ "loss": 0.0114,
428
+ "step": 5500
429
+ },
430
+ {
431
+ "epoch": 458.33,
432
+ "eval_loss": 2.157787561416626,
433
+ "eval_runtime": 5.9361,
434
+ "eval_samples_per_second": 23.921,
435
+ "eval_steps_per_second": 0.842,
436
+ "eval_wer": 0.8430688753269399,
437
+ "step": 5500
438
+ },
439
+ {
440
+ "epoch": 466.67,
441
+ "learning_rate": 2.2036363636363632e-05,
442
+ "loss": 0.0132,
443
+ "step": 5600
444
+ },
445
+ {
446
+ "epoch": 475.0,
447
+ "learning_rate": 1.658181818181818e-05,
448
+ "loss": 0.0142,
449
+ "step": 5700
450
+ },
451
+ {
452
+ "epoch": 483.33,
453
+ "learning_rate": 1.1127272727272727e-05,
454
+ "loss": 0.0132,
455
+ "step": 5800
456
+ },
457
+ {
458
+ "epoch": 491.67,
459
+ "learning_rate": 5.672727272727272e-06,
460
+ "loss": 0.0128,
461
+ "step": 5900
462
+ },
463
+ {
464
+ "epoch": 500.0,
465
+ "learning_rate": 2.1818181818181815e-07,
466
+ "loss": 0.012,
467
+ "step": 6000
468
+ },
469
+ {
470
+ "epoch": 500.0,
471
+ "eval_loss": 2.165587902069092,
472
+ "eval_runtime": 5.7782,
473
+ "eval_samples_per_second": 24.575,
474
+ "eval_steps_per_second": 0.865,
475
+ "eval_wer": 0.8352223190932868,
476
+ "step": 6000
477
+ },
478
+ {
479
+ "epoch": 500.0,
480
+ "step": 6000,
481
+ "total_flos": 3.5552286921447256e+19,
482
+ "train_loss": 0.5668576599160831,
483
+ "train_runtime": 11637.7564,
484
+ "train_samples_per_second": 16.24,
485
+ "train_steps_per_second": 0.516
486
+ }
487
+ ],
488
+ "max_steps": 6000,
489
+ "num_train_epochs": 500,
490
+ "total_flos": 3.5552286921447256e+19,
491
+ "trial_name": null,
492
+ "trial_params": null
493
+ }