Zaid commited on
Commit
243b790
1 Parent(s): f4cfa7e

End of training

Browse files
all_results.json ADDED
@@ -0,0 +1,12 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 1.0,
3
+ "eval_loss": 0.8486328125,
4
+ "eval_runtime": 2157.5812,
5
+ "eval_samples_per_second": 2.621,
6
+ "eval_steps_per_second": 0.164,
7
+ "eval_wer": 20.214938516576623,
8
+ "train_loss": 0.27437216796875,
9
+ "train_runtime": 51922.2504,
10
+ "train_samples_per_second": 3.082,
11
+ "train_steps_per_second": 0.096
12
+ }
eval_results.json ADDED
@@ -0,0 +1,8 @@
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 1.0,
3
+ "eval_loss": 0.8486328125,
4
+ "eval_runtime": 2157.5812,
5
+ "eval_samples_per_second": 2.621,
6
+ "eval_steps_per_second": 0.164,
7
+ "eval_wer": 20.214938516576623
8
+ }
runs/Dec14_18-42-08_129-146-181-180/events.out.tfevents.1671097691.129-146-181-180.74119.2 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:967636ea132ec850c732093664d1f4cabc58d6e790561d337567fc5a16397212
3
+ size 358
train_results.json ADDED
@@ -0,0 +1,7 @@
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 1.0,
3
+ "train_loss": 0.27437216796875,
4
+ "train_runtime": 51922.2504,
5
+ "train_samples_per_second": 3.082,
6
+ "train_steps_per_second": 0.096
7
+ }
trainer_state.json ADDED
@@ -0,0 +1,1270 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": 20.214938516576623,
3
+ "best_model_checkpoint": "./checkpoint-5000",
4
+ "epoch": 1.0,
5
+ "global_step": 5000,
6
+ "is_hyper_param_search": false,
7
+ "is_local_process_zero": true,
8
+ "is_world_process_zero": true,
9
+ "log_history": [
10
+ {
11
+ "epoch": 0.01,
12
+ "learning_rate": 4.973833272194737e-06,
13
+ "loss": 0.4248,
14
+ "step": 25
15
+ },
16
+ {
17
+ "epoch": 0.01,
18
+ "learning_rate": 6.195318418690893e-06,
19
+ "loss": 0.3353,
20
+ "step": 50
21
+ },
22
+ {
23
+ "epoch": 0.01,
24
+ "learning_rate": 6.881634451095711e-06,
25
+ "loss": 0.3227,
26
+ "step": 75
27
+ },
28
+ {
29
+ "epoch": 0.02,
30
+ "learning_rate": 7.344547104469332e-06,
31
+ "loss": 0.2826,
32
+ "step": 100
33
+ },
34
+ {
35
+ "epoch": 0.03,
36
+ "learning_rate": 7.716963756434345e-06,
37
+ "loss": 0.3717,
38
+ "step": 125
39
+ },
40
+ {
41
+ "epoch": 0.03,
42
+ "learning_rate": 8.019180844200955e-06,
43
+ "loss": 0.3568,
44
+ "step": 150
45
+ },
46
+ {
47
+ "epoch": 0.04,
48
+ "learning_rate": 8.27351214279797e-06,
49
+ "loss": 0.3498,
50
+ "step": 175
51
+ },
52
+ {
53
+ "epoch": 0.04,
54
+ "learning_rate": 8.49307723936858e-06,
55
+ "loss": 0.3809,
56
+ "step": 200
57
+ },
58
+ {
59
+ "epoch": 0.04,
60
+ "learning_rate": 8.686247975778677e-06,
61
+ "loss": 0.3468,
62
+ "step": 225
63
+ },
64
+ {
65
+ "epoch": 0.05,
66
+ "learning_rate": 8.858694625217149e-06,
67
+ "loss": 0.3089,
68
+ "step": 250
69
+ },
70
+ {
71
+ "epoch": 0.06,
72
+ "learning_rate": 9.014436199608479e-06,
73
+ "loss": 0.3316,
74
+ "step": 275
75
+ },
76
+ {
77
+ "epoch": 0.06,
78
+ "learning_rate": 9.156425255148058e-06,
79
+ "loss": 0.3564,
80
+ "step": 300
81
+ },
82
+ {
83
+ "epoch": 0.07,
84
+ "learning_rate": 9.28689473531776e-06,
85
+ "loss": 0.2979,
86
+ "step": 325
87
+ },
88
+ {
89
+ "epoch": 0.07,
90
+ "learning_rate": 9.407574351377137e-06,
91
+ "loss": 0.321,
92
+ "step": 350
93
+ },
94
+ {
95
+ "epoch": 0.07,
96
+ "learning_rate": 9.519831289296397e-06,
97
+ "loss": 0.2998,
98
+ "step": 375
99
+ },
100
+ {
101
+ "epoch": 0.08,
102
+ "learning_rate": 9.624764935335318e-06,
103
+ "loss": 0.2896,
104
+ "step": 400
105
+ },
106
+ {
107
+ "epoch": 0.09,
108
+ "learning_rate": 9.723272550712454e-06,
109
+ "loss": 0.2802,
110
+ "step": 425
111
+ },
112
+ {
113
+ "epoch": 0.09,
114
+ "learning_rate": 9.816095971633122e-06,
115
+ "loss": 0.3408,
116
+ "step": 450
117
+ },
118
+ {
119
+ "epoch": 0.1,
120
+ "learning_rate": 9.900435550016748e-06,
121
+ "loss": 0.3171,
122
+ "step": 475
123
+ },
124
+ {
125
+ "epoch": 0.1,
126
+ "learning_rate": 9.98382788472848e-06,
127
+ "loss": 0.3302,
128
+ "step": 500
129
+ },
130
+ {
131
+ "epoch": 0.1,
132
+ "learning_rate": 1e-05,
133
+ "loss": 0.277,
134
+ "step": 525
135
+ },
136
+ {
137
+ "epoch": 0.11,
138
+ "learning_rate": 1e-05,
139
+ "loss": 0.2987,
140
+ "step": 550
141
+ },
142
+ {
143
+ "epoch": 0.12,
144
+ "learning_rate": 1e-05,
145
+ "loss": 0.2941,
146
+ "step": 575
147
+ },
148
+ {
149
+ "epoch": 0.12,
150
+ "learning_rate": 1e-05,
151
+ "loss": 0.2529,
152
+ "step": 600
153
+ },
154
+ {
155
+ "epoch": 0.12,
156
+ "learning_rate": 1e-05,
157
+ "loss": 0.339,
158
+ "step": 625
159
+ },
160
+ {
161
+ "epoch": 0.13,
162
+ "learning_rate": 1e-05,
163
+ "loss": 0.2919,
164
+ "step": 650
165
+ },
166
+ {
167
+ "epoch": 0.14,
168
+ "learning_rate": 1e-05,
169
+ "loss": 0.3202,
170
+ "step": 675
171
+ },
172
+ {
173
+ "epoch": 0.14,
174
+ "learning_rate": 1e-05,
175
+ "loss": 0.3154,
176
+ "step": 700
177
+ },
178
+ {
179
+ "epoch": 0.14,
180
+ "learning_rate": 1e-05,
181
+ "loss": 0.3017,
182
+ "step": 725
183
+ },
184
+ {
185
+ "epoch": 0.15,
186
+ "learning_rate": 1e-05,
187
+ "loss": 0.3514,
188
+ "step": 750
189
+ },
190
+ {
191
+ "epoch": 0.15,
192
+ "learning_rate": 1e-05,
193
+ "loss": 0.3026,
194
+ "step": 775
195
+ },
196
+ {
197
+ "epoch": 0.16,
198
+ "learning_rate": 1e-05,
199
+ "loss": 0.302,
200
+ "step": 800
201
+ },
202
+ {
203
+ "epoch": 0.17,
204
+ "learning_rate": 1e-05,
205
+ "loss": 0.2944,
206
+ "step": 825
207
+ },
208
+ {
209
+ "epoch": 0.17,
210
+ "learning_rate": 1e-05,
211
+ "loss": 0.3156,
212
+ "step": 850
213
+ },
214
+ {
215
+ "epoch": 0.17,
216
+ "learning_rate": 1e-05,
217
+ "loss": 0.3407,
218
+ "step": 875
219
+ },
220
+ {
221
+ "epoch": 0.18,
222
+ "learning_rate": 1e-05,
223
+ "loss": 0.302,
224
+ "step": 900
225
+ },
226
+ {
227
+ "epoch": 0.18,
228
+ "learning_rate": 1e-05,
229
+ "loss": 0.2982,
230
+ "step": 925
231
+ },
232
+ {
233
+ "epoch": 0.19,
234
+ "learning_rate": 1e-05,
235
+ "loss": 0.294,
236
+ "step": 950
237
+ },
238
+ {
239
+ "epoch": 0.2,
240
+ "learning_rate": 1e-05,
241
+ "loss": 0.2874,
242
+ "step": 975
243
+ },
244
+ {
245
+ "epoch": 0.2,
246
+ "learning_rate": 1e-05,
247
+ "loss": 0.2767,
248
+ "step": 1000
249
+ },
250
+ {
251
+ "epoch": 0.2,
252
+ "eval_loss": 1.0654296875,
253
+ "eval_runtime": 2194.1241,
254
+ "eval_samples_per_second": 2.577,
255
+ "eval_steps_per_second": 0.161,
256
+ "eval_wer": 25.397216227568325,
257
+ "step": 1000
258
+ },
259
+ {
260
+ "epoch": 0.2,
261
+ "learning_rate": 1e-05,
262
+ "loss": 0.2915,
263
+ "step": 1025
264
+ },
265
+ {
266
+ "epoch": 0.21,
267
+ "learning_rate": 1e-05,
268
+ "loss": 0.2937,
269
+ "step": 1050
270
+ },
271
+ {
272
+ "epoch": 0.21,
273
+ "learning_rate": 1e-05,
274
+ "loss": 0.2843,
275
+ "step": 1075
276
+ },
277
+ {
278
+ "epoch": 0.22,
279
+ "learning_rate": 1e-05,
280
+ "loss": 0.2789,
281
+ "step": 1100
282
+ },
283
+ {
284
+ "epoch": 0.23,
285
+ "learning_rate": 1e-05,
286
+ "loss": 0.28,
287
+ "step": 1125
288
+ },
289
+ {
290
+ "epoch": 0.23,
291
+ "learning_rate": 1e-05,
292
+ "loss": 0.2801,
293
+ "step": 1150
294
+ },
295
+ {
296
+ "epoch": 0.23,
297
+ "learning_rate": 1e-05,
298
+ "loss": 0.3144,
299
+ "step": 1175
300
+ },
301
+ {
302
+ "epoch": 0.24,
303
+ "learning_rate": 1e-05,
304
+ "loss": 0.3067,
305
+ "step": 1200
306
+ },
307
+ {
308
+ "epoch": 0.24,
309
+ "learning_rate": 1e-05,
310
+ "loss": 0.3055,
311
+ "step": 1225
312
+ },
313
+ {
314
+ "epoch": 0.25,
315
+ "learning_rate": 1e-05,
316
+ "loss": 0.345,
317
+ "step": 1250
318
+ },
319
+ {
320
+ "epoch": 0.26,
321
+ "learning_rate": 1e-05,
322
+ "loss": 0.2893,
323
+ "step": 1275
324
+ },
325
+ {
326
+ "epoch": 0.26,
327
+ "learning_rate": 1e-05,
328
+ "loss": 0.2749,
329
+ "step": 1300
330
+ },
331
+ {
332
+ "epoch": 0.27,
333
+ "learning_rate": 1e-05,
334
+ "loss": 0.3303,
335
+ "step": 1325
336
+ },
337
+ {
338
+ "epoch": 0.27,
339
+ "learning_rate": 1e-05,
340
+ "loss": 0.3312,
341
+ "step": 1350
342
+ },
343
+ {
344
+ "epoch": 0.28,
345
+ "learning_rate": 1e-05,
346
+ "loss": 0.2868,
347
+ "step": 1375
348
+ },
349
+ {
350
+ "epoch": 0.28,
351
+ "learning_rate": 1e-05,
352
+ "loss": 0.2923,
353
+ "step": 1400
354
+ },
355
+ {
356
+ "epoch": 0.28,
357
+ "learning_rate": 1e-05,
358
+ "loss": 0.2963,
359
+ "step": 1425
360
+ },
361
+ {
362
+ "epoch": 0.29,
363
+ "learning_rate": 1e-05,
364
+ "loss": 0.321,
365
+ "step": 1450
366
+ },
367
+ {
368
+ "epoch": 0.29,
369
+ "learning_rate": 1e-05,
370
+ "loss": 0.2929,
371
+ "step": 1475
372
+ },
373
+ {
374
+ "epoch": 0.3,
375
+ "learning_rate": 1e-05,
376
+ "loss": 0.2385,
377
+ "step": 1500
378
+ },
379
+ {
380
+ "epoch": 0.3,
381
+ "learning_rate": 1e-05,
382
+ "loss": 0.2804,
383
+ "step": 1525
384
+ },
385
+ {
386
+ "epoch": 0.31,
387
+ "learning_rate": 1e-05,
388
+ "loss": 0.2676,
389
+ "step": 1550
390
+ },
391
+ {
392
+ "epoch": 0.32,
393
+ "learning_rate": 1e-05,
394
+ "loss": 0.2758,
395
+ "step": 1575
396
+ },
397
+ {
398
+ "epoch": 0.32,
399
+ "learning_rate": 1e-05,
400
+ "loss": 0.2623,
401
+ "step": 1600
402
+ },
403
+ {
404
+ "epoch": 0.33,
405
+ "learning_rate": 1e-05,
406
+ "loss": 0.3237,
407
+ "step": 1625
408
+ },
409
+ {
410
+ "epoch": 0.33,
411
+ "learning_rate": 1e-05,
412
+ "loss": 0.2559,
413
+ "step": 1650
414
+ },
415
+ {
416
+ "epoch": 0.34,
417
+ "learning_rate": 1e-05,
418
+ "loss": 0.2841,
419
+ "step": 1675
420
+ },
421
+ {
422
+ "epoch": 0.34,
423
+ "learning_rate": 1e-05,
424
+ "loss": 0.2503,
425
+ "step": 1700
426
+ },
427
+ {
428
+ "epoch": 0.34,
429
+ "learning_rate": 1e-05,
430
+ "loss": 0.3199,
431
+ "step": 1725
432
+ },
433
+ {
434
+ "epoch": 0.35,
435
+ "learning_rate": 1e-05,
436
+ "loss": 0.2686,
437
+ "step": 1750
438
+ },
439
+ {
440
+ "epoch": 0.35,
441
+ "learning_rate": 1e-05,
442
+ "loss": 0.2463,
443
+ "step": 1775
444
+ },
445
+ {
446
+ "epoch": 0.36,
447
+ "learning_rate": 1e-05,
448
+ "loss": 0.2853,
449
+ "step": 1800
450
+ },
451
+ {
452
+ "epoch": 0.36,
453
+ "learning_rate": 1e-05,
454
+ "loss": 0.272,
455
+ "step": 1825
456
+ },
457
+ {
458
+ "epoch": 0.37,
459
+ "learning_rate": 1e-05,
460
+ "loss": 0.2859,
461
+ "step": 1850
462
+ },
463
+ {
464
+ "epoch": 0.38,
465
+ "learning_rate": 1e-05,
466
+ "loss": 0.2862,
467
+ "step": 1875
468
+ },
469
+ {
470
+ "epoch": 0.38,
471
+ "learning_rate": 1e-05,
472
+ "loss": 0.3104,
473
+ "step": 1900
474
+ },
475
+ {
476
+ "epoch": 0.39,
477
+ "learning_rate": 1e-05,
478
+ "loss": 0.2786,
479
+ "step": 1925
480
+ },
481
+ {
482
+ "epoch": 0.39,
483
+ "learning_rate": 1e-05,
484
+ "loss": 0.2828,
485
+ "step": 1950
486
+ },
487
+ {
488
+ "epoch": 0.4,
489
+ "learning_rate": 1e-05,
490
+ "loss": 0.2732,
491
+ "step": 1975
492
+ },
493
+ {
494
+ "epoch": 0.4,
495
+ "learning_rate": 1e-05,
496
+ "loss": 0.2528,
497
+ "step": 2000
498
+ },
499
+ {
500
+ "epoch": 0.4,
501
+ "eval_loss": 0.93701171875,
502
+ "eval_runtime": 2189.9196,
503
+ "eval_samples_per_second": 2.582,
504
+ "eval_steps_per_second": 0.162,
505
+ "eval_wer": 22.131135669075793,
506
+ "step": 2000
507
+ },
508
+ {
509
+ "epoch": 0.41,
510
+ "learning_rate": 1e-05,
511
+ "loss": 0.2352,
512
+ "step": 2025
513
+ },
514
+ {
515
+ "epoch": 0.41,
516
+ "learning_rate": 1e-05,
517
+ "loss": 0.2332,
518
+ "step": 2050
519
+ },
520
+ {
521
+ "epoch": 0.41,
522
+ "learning_rate": 1e-05,
523
+ "loss": 0.2303,
524
+ "step": 2075
525
+ },
526
+ {
527
+ "epoch": 0.42,
528
+ "learning_rate": 1e-05,
529
+ "loss": 0.2777,
530
+ "step": 2100
531
+ },
532
+ {
533
+ "epoch": 0.42,
534
+ "learning_rate": 1e-05,
535
+ "loss": 0.306,
536
+ "step": 2125
537
+ },
538
+ {
539
+ "epoch": 0.43,
540
+ "learning_rate": 1e-05,
541
+ "loss": 0.2996,
542
+ "step": 2150
543
+ },
544
+ {
545
+ "epoch": 0.43,
546
+ "learning_rate": 1e-05,
547
+ "loss": 0.2405,
548
+ "step": 2175
549
+ },
550
+ {
551
+ "epoch": 0.44,
552
+ "learning_rate": 1e-05,
553
+ "loss": 0.2519,
554
+ "step": 2200
555
+ },
556
+ {
557
+ "epoch": 0.45,
558
+ "learning_rate": 1e-05,
559
+ "loss": 0.3032,
560
+ "step": 2225
561
+ },
562
+ {
563
+ "epoch": 0.45,
564
+ "learning_rate": 1e-05,
565
+ "loss": 0.2881,
566
+ "step": 2250
567
+ },
568
+ {
569
+ "epoch": 0.46,
570
+ "learning_rate": 1e-05,
571
+ "loss": 0.2801,
572
+ "step": 2275
573
+ },
574
+ {
575
+ "epoch": 0.46,
576
+ "learning_rate": 1e-05,
577
+ "loss": 0.2945,
578
+ "step": 2300
579
+ },
580
+ {
581
+ "epoch": 0.47,
582
+ "learning_rate": 1e-05,
583
+ "loss": 0.2836,
584
+ "step": 2325
585
+ },
586
+ {
587
+ "epoch": 0.47,
588
+ "learning_rate": 1e-05,
589
+ "loss": 0.2651,
590
+ "step": 2350
591
+ },
592
+ {
593
+ "epoch": 0.47,
594
+ "learning_rate": 1e-05,
595
+ "loss": 0.2837,
596
+ "step": 2375
597
+ },
598
+ {
599
+ "epoch": 0.48,
600
+ "learning_rate": 1e-05,
601
+ "loss": 0.2498,
602
+ "step": 2400
603
+ },
604
+ {
605
+ "epoch": 0.48,
606
+ "learning_rate": 1e-05,
607
+ "loss": 0.2304,
608
+ "step": 2425
609
+ },
610
+ {
611
+ "epoch": 0.49,
612
+ "learning_rate": 1e-05,
613
+ "loss": 0.2475,
614
+ "step": 2450
615
+ },
616
+ {
617
+ "epoch": 0.49,
618
+ "learning_rate": 1e-05,
619
+ "loss": 0.2568,
620
+ "step": 2475
621
+ },
622
+ {
623
+ "epoch": 0.5,
624
+ "learning_rate": 1e-05,
625
+ "loss": 0.2431,
626
+ "step": 2500
627
+ },
628
+ {
629
+ "epoch": 0.51,
630
+ "learning_rate": 1e-05,
631
+ "loss": 0.2973,
632
+ "step": 2525
633
+ },
634
+ {
635
+ "epoch": 0.51,
636
+ "learning_rate": 1e-05,
637
+ "loss": 0.2509,
638
+ "step": 2550
639
+ },
640
+ {
641
+ "epoch": 0.52,
642
+ "learning_rate": 1e-05,
643
+ "loss": 0.2662,
644
+ "step": 2575
645
+ },
646
+ {
647
+ "epoch": 0.52,
648
+ "learning_rate": 1e-05,
649
+ "loss": 0.2174,
650
+ "step": 2600
651
+ },
652
+ {
653
+ "epoch": 0.53,
654
+ "learning_rate": 1e-05,
655
+ "loss": 0.2375,
656
+ "step": 2625
657
+ },
658
+ {
659
+ "epoch": 0.53,
660
+ "learning_rate": 1e-05,
661
+ "loss": 0.3,
662
+ "step": 2650
663
+ },
664
+ {
665
+ "epoch": 0.54,
666
+ "learning_rate": 1e-05,
667
+ "loss": 0.2586,
668
+ "step": 2675
669
+ },
670
+ {
671
+ "epoch": 0.54,
672
+ "learning_rate": 1e-05,
673
+ "loss": 0.2233,
674
+ "step": 2700
675
+ },
676
+ {
677
+ "epoch": 0.55,
678
+ "learning_rate": 1e-05,
679
+ "loss": 0.2458,
680
+ "step": 2725
681
+ },
682
+ {
683
+ "epoch": 0.55,
684
+ "learning_rate": 1e-05,
685
+ "loss": 0.2033,
686
+ "step": 2750
687
+ },
688
+ {
689
+ "epoch": 0.56,
690
+ "learning_rate": 1e-05,
691
+ "loss": 0.2896,
692
+ "step": 2775
693
+ },
694
+ {
695
+ "epoch": 0.56,
696
+ "learning_rate": 1e-05,
697
+ "loss": 0.269,
698
+ "step": 2800
699
+ },
700
+ {
701
+ "epoch": 0.56,
702
+ "learning_rate": 1e-05,
703
+ "loss": 0.2408,
704
+ "step": 2825
705
+ },
706
+ {
707
+ "epoch": 0.57,
708
+ "learning_rate": 1e-05,
709
+ "loss": 0.2474,
710
+ "step": 2850
711
+ },
712
+ {
713
+ "epoch": 0.57,
714
+ "learning_rate": 1e-05,
715
+ "loss": 0.2357,
716
+ "step": 2875
717
+ },
718
+ {
719
+ "epoch": 0.58,
720
+ "learning_rate": 1e-05,
721
+ "loss": 0.3061,
722
+ "step": 2900
723
+ },
724
+ {
725
+ "epoch": 0.58,
726
+ "learning_rate": 1e-05,
727
+ "loss": 0.3008,
728
+ "step": 2925
729
+ },
730
+ {
731
+ "epoch": 0.59,
732
+ "learning_rate": 1e-05,
733
+ "loss": 0.2849,
734
+ "step": 2950
735
+ },
736
+ {
737
+ "epoch": 0.59,
738
+ "learning_rate": 1e-05,
739
+ "loss": 0.2692,
740
+ "step": 2975
741
+ },
742
+ {
743
+ "epoch": 0.6,
744
+ "learning_rate": 1e-05,
745
+ "loss": 0.3038,
746
+ "step": 3000
747
+ },
748
+ {
749
+ "epoch": 0.6,
750
+ "eval_loss": 0.99658203125,
751
+ "eval_runtime": 2194.0772,
752
+ "eval_samples_per_second": 2.577,
753
+ "eval_steps_per_second": 0.161,
754
+ "eval_wer": 20.575583332126357,
755
+ "step": 3000
756
+ },
757
+ {
758
+ "epoch": 0.6,
759
+ "learning_rate": 1e-05,
760
+ "loss": 0.2707,
761
+ "step": 3025
762
+ },
763
+ {
764
+ "epoch": 0.61,
765
+ "learning_rate": 1e-05,
766
+ "loss": 0.2503,
767
+ "step": 3050
768
+ },
769
+ {
770
+ "epoch": 0.61,
771
+ "learning_rate": 1e-05,
772
+ "loss": 0.2505,
773
+ "step": 3075
774
+ },
775
+ {
776
+ "epoch": 0.62,
777
+ "learning_rate": 1e-05,
778
+ "loss": 0.2751,
779
+ "step": 3100
780
+ },
781
+ {
782
+ "epoch": 0.62,
783
+ "learning_rate": 1e-05,
784
+ "loss": 0.2501,
785
+ "step": 3125
786
+ },
787
+ {
788
+ "epoch": 0.63,
789
+ "learning_rate": 1e-05,
790
+ "loss": 0.2805,
791
+ "step": 3150
792
+ },
793
+ {
794
+ "epoch": 0.64,
795
+ "learning_rate": 1e-05,
796
+ "loss": 0.2512,
797
+ "step": 3175
798
+ },
799
+ {
800
+ "epoch": 0.64,
801
+ "learning_rate": 1e-05,
802
+ "loss": 0.2928,
803
+ "step": 3200
804
+ },
805
+ {
806
+ "epoch": 0.65,
807
+ "learning_rate": 1e-05,
808
+ "loss": 0.2776,
809
+ "step": 3225
810
+ },
811
+ {
812
+ "epoch": 0.65,
813
+ "learning_rate": 1e-05,
814
+ "loss": 0.2497,
815
+ "step": 3250
816
+ },
817
+ {
818
+ "epoch": 0.66,
819
+ "learning_rate": 1e-05,
820
+ "loss": 0.2434,
821
+ "step": 3275
822
+ },
823
+ {
824
+ "epoch": 0.66,
825
+ "learning_rate": 1e-05,
826
+ "loss": 0.2577,
827
+ "step": 3300
828
+ },
829
+ {
830
+ "epoch": 0.67,
831
+ "learning_rate": 1e-05,
832
+ "loss": 0.241,
833
+ "step": 3325
834
+ },
835
+ {
836
+ "epoch": 0.67,
837
+ "learning_rate": 1e-05,
838
+ "loss": 0.2555,
839
+ "step": 3350
840
+ },
841
+ {
842
+ "epoch": 0.68,
843
+ "learning_rate": 1e-05,
844
+ "loss": 0.2963,
845
+ "step": 3375
846
+ },
847
+ {
848
+ "epoch": 0.68,
849
+ "learning_rate": 1e-05,
850
+ "loss": 0.2876,
851
+ "step": 3400
852
+ },
853
+ {
854
+ "epoch": 0.69,
855
+ "learning_rate": 1e-05,
856
+ "loss": 0.2444,
857
+ "step": 3425
858
+ },
859
+ {
860
+ "epoch": 0.69,
861
+ "learning_rate": 1e-05,
862
+ "loss": 0.2383,
863
+ "step": 3450
864
+ },
865
+ {
866
+ "epoch": 0.69,
867
+ "learning_rate": 1e-05,
868
+ "loss": 0.246,
869
+ "step": 3475
870
+ },
871
+ {
872
+ "epoch": 0.7,
873
+ "learning_rate": 1e-05,
874
+ "loss": 0.2611,
875
+ "step": 3500
876
+ },
877
+ {
878
+ "epoch": 0.7,
879
+ "learning_rate": 1e-05,
880
+ "loss": 0.2633,
881
+ "step": 3525
882
+ },
883
+ {
884
+ "epoch": 0.71,
885
+ "learning_rate": 1e-05,
886
+ "loss": 0.261,
887
+ "step": 3550
888
+ },
889
+ {
890
+ "epoch": 0.71,
891
+ "learning_rate": 1e-05,
892
+ "loss": 0.236,
893
+ "step": 3575
894
+ },
895
+ {
896
+ "epoch": 0.72,
897
+ "learning_rate": 1e-05,
898
+ "loss": 0.2509,
899
+ "step": 3600
900
+ },
901
+ {
902
+ "epoch": 0.72,
903
+ "learning_rate": 1e-05,
904
+ "loss": 0.2725,
905
+ "step": 3625
906
+ },
907
+ {
908
+ "epoch": 0.73,
909
+ "learning_rate": 1e-05,
910
+ "loss": 0.2433,
911
+ "step": 3650
912
+ },
913
+ {
914
+ "epoch": 0.73,
915
+ "learning_rate": 1e-05,
916
+ "loss": 0.2287,
917
+ "step": 3675
918
+ },
919
+ {
920
+ "epoch": 0.74,
921
+ "learning_rate": 1e-05,
922
+ "loss": 0.2596,
923
+ "step": 3700
924
+ },
925
+ {
926
+ "epoch": 0.74,
927
+ "learning_rate": 1e-05,
928
+ "loss": 0.247,
929
+ "step": 3725
930
+ },
931
+ {
932
+ "epoch": 0.75,
933
+ "learning_rate": 1e-05,
934
+ "loss": 0.2559,
935
+ "step": 3750
936
+ },
937
+ {
938
+ "epoch": 0.76,
939
+ "learning_rate": 1e-05,
940
+ "loss": 0.2356,
941
+ "step": 3775
942
+ },
943
+ {
944
+ "epoch": 0.76,
945
+ "learning_rate": 1e-05,
946
+ "loss": 0.2166,
947
+ "step": 3800
948
+ },
949
+ {
950
+ "epoch": 0.77,
951
+ "learning_rate": 1e-05,
952
+ "loss": 0.2397,
953
+ "step": 3825
954
+ },
955
+ {
956
+ "epoch": 0.77,
957
+ "learning_rate": 1e-05,
958
+ "loss": 0.2638,
959
+ "step": 3850
960
+ },
961
+ {
962
+ "epoch": 0.78,
963
+ "learning_rate": 1e-05,
964
+ "loss": 0.2146,
965
+ "step": 3875
966
+ },
967
+ {
968
+ "epoch": 0.78,
969
+ "learning_rate": 1e-05,
970
+ "loss": 0.2261,
971
+ "step": 3900
972
+ },
973
+ {
974
+ "epoch": 0.79,
975
+ "learning_rate": 1e-05,
976
+ "loss": 0.2542,
977
+ "step": 3925
978
+ },
979
+ {
980
+ "epoch": 0.79,
981
+ "learning_rate": 1e-05,
982
+ "loss": 0.2548,
983
+ "step": 3950
984
+ },
985
+ {
986
+ "epoch": 0.8,
987
+ "learning_rate": 1e-05,
988
+ "loss": 0.2075,
989
+ "step": 3975
990
+ },
991
+ {
992
+ "epoch": 0.8,
993
+ "learning_rate": 1e-05,
994
+ "loss": 0.2718,
995
+ "step": 4000
996
+ },
997
+ {
998
+ "epoch": 0.8,
999
+ "eval_loss": 0.8720703125,
1000
+ "eval_runtime": 2119.3543,
1001
+ "eval_samples_per_second": 2.668,
1002
+ "eval_steps_per_second": 0.167,
1003
+ "eval_wer": 24.929391828280924,
1004
+ "step": 4000
1005
+ },
1006
+ {
1007
+ "epoch": 0.81,
1008
+ "learning_rate": 1e-05,
1009
+ "loss": 0.2896,
1010
+ "step": 4025
1011
+ },
1012
+ {
1013
+ "epoch": 0.81,
1014
+ "learning_rate": 1e-05,
1015
+ "loss": 0.22,
1016
+ "step": 4050
1017
+ },
1018
+ {
1019
+ "epoch": 0.81,
1020
+ "learning_rate": 1e-05,
1021
+ "loss": 0.2534,
1022
+ "step": 4075
1023
+ },
1024
+ {
1025
+ "epoch": 0.82,
1026
+ "learning_rate": 1e-05,
1027
+ "loss": 0.251,
1028
+ "step": 4100
1029
+ },
1030
+ {
1031
+ "epoch": 0.82,
1032
+ "learning_rate": 1e-05,
1033
+ "loss": 0.2616,
1034
+ "step": 4125
1035
+ },
1036
+ {
1037
+ "epoch": 0.83,
1038
+ "learning_rate": 1e-05,
1039
+ "loss": 0.267,
1040
+ "step": 4150
1041
+ },
1042
+ {
1043
+ "epoch": 0.83,
1044
+ "learning_rate": 1e-05,
1045
+ "loss": 0.2211,
1046
+ "step": 4175
1047
+ },
1048
+ {
1049
+ "epoch": 0.84,
1050
+ "learning_rate": 1e-05,
1051
+ "loss": 0.2264,
1052
+ "step": 4200
1053
+ },
1054
+ {
1055
+ "epoch": 0.84,
1056
+ "learning_rate": 1e-05,
1057
+ "loss": 0.2683,
1058
+ "step": 4225
1059
+ },
1060
+ {
1061
+ "epoch": 0.85,
1062
+ "learning_rate": 1e-05,
1063
+ "loss": 0.2711,
1064
+ "step": 4250
1065
+ },
1066
+ {
1067
+ "epoch": 0.85,
1068
+ "learning_rate": 1e-05,
1069
+ "loss": 0.2582,
1070
+ "step": 4275
1071
+ },
1072
+ {
1073
+ "epoch": 0.86,
1074
+ "learning_rate": 1e-05,
1075
+ "loss": 0.2348,
1076
+ "step": 4300
1077
+ },
1078
+ {
1079
+ "epoch": 0.86,
1080
+ "learning_rate": 1e-05,
1081
+ "loss": 0.2457,
1082
+ "step": 4325
1083
+ },
1084
+ {
1085
+ "epoch": 0.87,
1086
+ "learning_rate": 1e-05,
1087
+ "loss": 0.2883,
1088
+ "step": 4350
1089
+ },
1090
+ {
1091
+ "epoch": 0.88,
1092
+ "learning_rate": 1e-05,
1093
+ "loss": 0.2408,
1094
+ "step": 4375
1095
+ },
1096
+ {
1097
+ "epoch": 0.88,
1098
+ "learning_rate": 1e-05,
1099
+ "loss": 0.2509,
1100
+ "step": 4400
1101
+ },
1102
+ {
1103
+ "epoch": 0.89,
1104
+ "learning_rate": 1e-05,
1105
+ "loss": 0.2546,
1106
+ "step": 4425
1107
+ },
1108
+ {
1109
+ "epoch": 0.89,
1110
+ "learning_rate": 1e-05,
1111
+ "loss": 0.2427,
1112
+ "step": 4450
1113
+ },
1114
+ {
1115
+ "epoch": 0.9,
1116
+ "learning_rate": 1e-05,
1117
+ "loss": 0.2382,
1118
+ "step": 4475
1119
+ },
1120
+ {
1121
+ "epoch": 0.9,
1122
+ "learning_rate": 1e-05,
1123
+ "loss": 0.2371,
1124
+ "step": 4500
1125
+ },
1126
+ {
1127
+ "epoch": 0.91,
1128
+ "learning_rate": 1e-05,
1129
+ "loss": 0.2422,
1130
+ "step": 4525
1131
+ },
1132
+ {
1133
+ "epoch": 0.91,
1134
+ "learning_rate": 1e-05,
1135
+ "loss": 0.2686,
1136
+ "step": 4550
1137
+ },
1138
+ {
1139
+ "epoch": 0.92,
1140
+ "learning_rate": 1e-05,
1141
+ "loss": 0.2689,
1142
+ "step": 4575
1143
+ },
1144
+ {
1145
+ "epoch": 0.92,
1146
+ "learning_rate": 1e-05,
1147
+ "loss": 0.2693,
1148
+ "step": 4600
1149
+ },
1150
+ {
1151
+ "epoch": 0.93,
1152
+ "learning_rate": 1e-05,
1153
+ "loss": 0.2359,
1154
+ "step": 4625
1155
+ },
1156
+ {
1157
+ "epoch": 0.93,
1158
+ "learning_rate": 1e-05,
1159
+ "loss": 0.2611,
1160
+ "step": 4650
1161
+ },
1162
+ {
1163
+ "epoch": 0.94,
1164
+ "learning_rate": 1e-05,
1165
+ "loss": 0.2249,
1166
+ "step": 4675
1167
+ },
1168
+ {
1169
+ "epoch": 0.94,
1170
+ "learning_rate": 1e-05,
1171
+ "loss": 0.2564,
1172
+ "step": 4700
1173
+ },
1174
+ {
1175
+ "epoch": 0.94,
1176
+ "learning_rate": 1e-05,
1177
+ "loss": 0.2328,
1178
+ "step": 4725
1179
+ },
1180
+ {
1181
+ "epoch": 0.95,
1182
+ "learning_rate": 1e-05,
1183
+ "loss": 0.2576,
1184
+ "step": 4750
1185
+ },
1186
+ {
1187
+ "epoch": 0.95,
1188
+ "learning_rate": 1e-05,
1189
+ "loss": 0.2797,
1190
+ "step": 4775
1191
+ },
1192
+ {
1193
+ "epoch": 0.96,
1194
+ "learning_rate": 1e-05,
1195
+ "loss": 0.2583,
1196
+ "step": 4800
1197
+ },
1198
+ {
1199
+ "epoch": 0.96,
1200
+ "learning_rate": 1e-05,
1201
+ "loss": 0.2488,
1202
+ "step": 4825
1203
+ },
1204
+ {
1205
+ "epoch": 0.97,
1206
+ "learning_rate": 1e-05,
1207
+ "loss": 0.2767,
1208
+ "step": 4850
1209
+ },
1210
+ {
1211
+ "epoch": 0.97,
1212
+ "learning_rate": 1e-05,
1213
+ "loss": 0.2527,
1214
+ "step": 4875
1215
+ },
1216
+ {
1217
+ "epoch": 0.98,
1218
+ "learning_rate": 1e-05,
1219
+ "loss": 0.2304,
1220
+ "step": 4900
1221
+ },
1222
+ {
1223
+ "epoch": 0.98,
1224
+ "learning_rate": 1e-05,
1225
+ "loss": 0.23,
1226
+ "step": 4925
1227
+ },
1228
+ {
1229
+ "epoch": 0.99,
1230
+ "learning_rate": 1e-05,
1231
+ "loss": 0.2285,
1232
+ "step": 4950
1233
+ },
1234
+ {
1235
+ "epoch": 0.99,
1236
+ "learning_rate": 1e-05,
1237
+ "loss": 0.2367,
1238
+ "step": 4975
1239
+ },
1240
+ {
1241
+ "epoch": 1.0,
1242
+ "learning_rate": 1e-05,
1243
+ "loss": 0.2269,
1244
+ "step": 5000
1245
+ },
1246
+ {
1247
+ "epoch": 1.0,
1248
+ "eval_loss": 0.8486328125,
1249
+ "eval_runtime": 2156.4186,
1250
+ "eval_samples_per_second": 2.622,
1251
+ "eval_steps_per_second": 0.164,
1252
+ "eval_wer": 20.214938516576623,
1253
+ "step": 5000
1254
+ },
1255
+ {
1256
+ "epoch": 1.0,
1257
+ "step": 5000,
1258
+ "total_flos": 3.397073665165099e+20,
1259
+ "train_loss": 0.27437216796875,
1260
+ "train_runtime": 51922.2504,
1261
+ "train_samples_per_second": 3.082,
1262
+ "train_steps_per_second": 0.096
1263
+ }
1264
+ ],
1265
+ "max_steps": 5000,
1266
+ "num_train_epochs": 9223372036854775807,
1267
+ "total_flos": 3.397073665165099e+20,
1268
+ "trial_name": null,
1269
+ "trial_params": null
1270
+ }