imvladikon commited on
Commit
e678519
1 Parent(s): fa6809f

End of training

Browse files
all_results.json CHANGED
@@ -1,14 +1,14 @@
1
  {
2
- "epoch": 100.0,
3
- "eval_loss": 0.5438345074653625,
4
- "eval_runtime": 140.268,
5
- "eval_samples": 5076,
6
- "eval_samples_per_second": 36.188,
7
- "eval_steps_per_second": 2.267,
8
- "eval_wer": 0.177349387392344,
9
- "train_loss": 0.8928292760036721,
10
- "train_runtime": 80759.6589,
11
- "train_samples": 20306,
12
- "train_samples_per_second": 25.144,
13
- "train_steps_per_second": 0.393
14
  }
 
1
  {
2
+ "epoch": 60.0,
3
+ "eval_loss": 0.4502493739128113,
4
+ "eval_runtime": 418.5824,
5
+ "eval_samples": 20246,
6
+ "eval_samples_per_second": 48.368,
7
+ "eval_steps_per_second": 3.024,
8
+ "eval_wer": 0.23182961772311134,
9
+ "train_loss": 1.0617500136590194,
10
+ "train_runtime": 191647.3893,
11
+ "train_samples": 90777,
12
+ "train_samples_per_second": 28.42,
13
+ "train_steps_per_second": 0.444
14
  }
train_results.json CHANGED
@@ -1,8 +1,8 @@
1
  {
2
- "epoch": 100.0,
3
- "train_loss": 0.8928292760036721,
4
- "train_runtime": 80759.6589,
5
- "train_samples": 20306,
6
- "train_samples_per_second": 25.144,
7
- "train_steps_per_second": 0.393
8
  }
 
1
  {
2
+ "epoch": 60.0,
3
+ "train_loss": 1.0617500136590194,
4
+ "train_runtime": 191647.3893,
5
+ "train_samples": 90777,
6
+ "train_samples_per_second": 28.42,
7
+ "train_steps_per_second": 0.444
8
  }
trainer_state.json CHANGED
@@ -1,394 +1,1042 @@
1
  {
2
  "best_metric": null,
3
  "best_model_checkpoint": null,
4
- "epoch": 99.99842519685039,
5
- "global_step": 31700,
6
  "is_hyper_param_search": false,
7
  "is_local_process_zero": true,
8
  "is_world_process_zero": true,
9
  "log_history": [
10
  {
11
- "epoch": 3.15,
12
- "eval_loss": 0.5203462243080139,
13
- "eval_runtime": 141.5768,
14
- "eval_samples_per_second": 35.853,
15
- "eval_steps_per_second": 2.246,
16
- "eval_wer": 0.4333326279704594,
17
  "step": 1000
18
  },
19
  {
20
- "epoch": 6.31,
21
- "learning_rate": 0.0009674592833876221,
22
- "loss": 1.4284,
23
  "step": 2000
24
  },
25
  {
26
- "epoch": 6.31,
27
- "eval_loss": 0.48156219720840454,
28
- "eval_runtime": 142.5636,
29
- "eval_samples_per_second": 35.605,
30
- "eval_steps_per_second": 2.231,
31
- "eval_wer": 0.3950949065746873,
32
  "step": 2000
33
  },
34
  {
35
- "epoch": 9.46,
36
- "eval_loss": 0.4314565062522888,
37
- "eval_runtime": 139.8322,
38
- "eval_samples_per_second": 36.301,
39
- "eval_steps_per_second": 2.274,
40
- "eval_wer": 0.3545506485811626,
41
  "step": 3000
42
  },
43
  {
44
- "epoch": 12.62,
45
- "learning_rate": 0.0009023452768729642,
46
- "loss": 1.283,
47
  "step": 4000
48
  },
49
  {
50
- "epoch": 12.62,
51
- "eval_loss": 0.42783403396606445,
52
- "eval_runtime": 141.0912,
53
- "eval_samples_per_second": 35.977,
54
- "eval_steps_per_second": 2.254,
55
- "eval_wer": 0.34039401570137756,
56
  "step": 4000
57
  },
58
  {
59
- "epoch": 15.77,
60
- "eval_loss": 0.40902915596961975,
61
- "eval_runtime": 140.9561,
62
- "eval_samples_per_second": 36.011,
63
- "eval_steps_per_second": 2.256,
64
- "eval_wer": 0.3053939098969465,
65
  "step": 5000
66
  },
67
  {
68
- "epoch": 18.93,
69
- "learning_rate": 0.0008372312703583062,
70
- "loss": 1.1777,
71
  "step": 6000
72
  },
73
  {
74
- "epoch": 18.93,
75
- "eval_loss": 0.3892641067504883,
76
- "eval_runtime": 139.9163,
77
- "eval_samples_per_second": 36.279,
78
- "eval_steps_per_second": 2.273,
79
- "eval_wer": 0.30056922783926193,
80
  "step": 6000
81
  },
82
  {
83
- "epoch": 22.08,
84
- "eval_loss": 0.3967570960521698,
85
- "eval_runtime": 139.2725,
86
- "eval_samples_per_second": 36.447,
87
- "eval_steps_per_second": 2.283,
88
- "eval_wer": 0.28565080305563195,
89
  "step": 7000
90
  },
91
  {
92
- "epoch": 25.24,
93
- "learning_rate": 0.0007720846905537459,
94
- "loss": 1.0994,
95
  "step": 8000
96
  },
97
  {
98
- "epoch": 25.24,
99
- "eval_loss": 0.3892391324043274,
100
- "eval_runtime": 138.7844,
101
- "eval_samples_per_second": 36.575,
102
- "eval_steps_per_second": 2.291,
103
- "eval_wer": 0.27509152083289246,
104
  "step": 8000
105
  },
106
  {
107
- "epoch": 28.39,
108
- "eval_loss": 0.4061281681060791,
109
- "eval_runtime": 139.2312,
110
- "eval_samples_per_second": 36.457,
111
- "eval_steps_per_second": 2.284,
112
- "eval_wer": 0.2689760247159151,
113
  "step": 9000
114
  },
115
  {
116
- "epoch": 31.54,
117
- "learning_rate": 0.000706970684039088,
118
- "loss": 1.0323,
119
  "step": 10000
120
  },
121
  {
122
- "epoch": 31.54,
123
- "eval_loss": 0.41136494278907776,
124
- "eval_runtime": 139.4432,
125
- "eval_samples_per_second": 36.402,
126
- "eval_steps_per_second": 2.28,
127
- "eval_wer": 0.25065069725120087,
128
  "step": 10000
129
  },
130
  {
131
- "epoch": 34.7,
132
- "eval_loss": 0.40214526653289795,
133
- "eval_runtime": 139.4093,
134
- "eval_samples_per_second": 36.411,
135
- "eval_steps_per_second": 2.281,
136
- "eval_wer": 0.2508411452271621,
137
  "step": 11000
138
  },
139
  {
140
- "epoch": 37.85,
141
- "learning_rate": 0.00064185667752443,
142
- "loss": 0.9623,
143
  "step": 12000
144
  },
145
  {
146
- "epoch": 37.85,
147
- "eval_loss": 0.40321338176727295,
148
- "eval_runtime": 139.9917,
149
- "eval_samples_per_second": 36.259,
150
- "eval_steps_per_second": 2.272,
151
- "eval_wer": 0.2378060393169266,
152
  "step": 12000
153
  },
154
  {
155
- "epoch": 41.01,
156
- "eval_loss": 0.4147748053073883,
157
- "eval_runtime": 139.3612,
158
- "eval_samples_per_second": 36.423,
159
- "eval_steps_per_second": 2.282,
160
- "eval_wer": 0.23744630425122204,
161
  "step": 13000
162
  },
163
  {
164
- "epoch": 44.16,
165
- "learning_rate": 0.0005767100977198697,
166
- "loss": 0.9077,
167
  "step": 14000
168
  },
169
  {
170
- "epoch": 44.16,
171
- "eval_loss": 0.4350396394729614,
172
- "eval_runtime": 138.8108,
173
- "eval_samples_per_second": 36.568,
174
- "eval_steps_per_second": 2.291,
175
- "eval_wer": 0.23230420890026873,
176
  "step": 14000
177
  },
178
  {
179
- "epoch": 47.32,
180
- "eval_loss": 0.4514589309692383,
181
- "eval_runtime": 138.719,
182
- "eval_samples_per_second": 36.592,
183
- "eval_steps_per_second": 2.292,
184
- "eval_wer": 0.22464396808938358,
185
  "step": 15000
186
  },
187
  {
188
- "epoch": 50.47,
189
- "learning_rate": 0.0005115960912052118,
190
- "loss": 0.8573,
191
  "step": 16000
192
  },
193
  {
194
- "epoch": 50.47,
195
- "eval_loss": 0.4473990499973297,
196
- "eval_runtime": 140.3605,
197
- "eval_samples_per_second": 36.164,
198
- "eval_steps_per_second": 2.266,
199
- "eval_wer": 0.21797828893074042,
200
  "step": 16000
201
  },
202
  {
203
- "epoch": 53.63,
204
- "eval_loss": 0.4649062752723694,
205
- "eval_runtime": 137.3039,
206
- "eval_samples_per_second": 36.969,
207
- "eval_steps_per_second": 2.316,
208
- "eval_wer": 0.21713185348202382,
209
  "step": 17000
210
  },
211
  {
212
- "epoch": 56.78,
213
- "learning_rate": 0.00044651465798045605,
214
- "loss": 0.8083,
215
  "step": 18000
216
  },
217
  {
218
- "epoch": 56.78,
219
- "eval_loss": 0.44551119208335876,
220
- "eval_runtime": 139.4699,
221
- "eval_samples_per_second": 36.395,
222
- "eval_steps_per_second": 2.28,
223
- "eval_wer": 0.2102334045749836,
224
  "step": 18000
225
  },
226
  {
227
- "epoch": 59.94,
228
- "eval_loss": 0.4586869478225708,
229
- "eval_runtime": 139.1403,
230
- "eval_samples_per_second": 36.481,
231
- "eval_steps_per_second": 2.285,
232
- "eval_wer": 0.20917536026408787,
233
  "step": 19000
234
  },
235
  {
236
- "epoch": 63.09,
237
- "learning_rate": 0.00038140065146579803,
238
- "loss": 0.769,
239
  "step": 20000
240
  },
241
  {
242
- "epoch": 63.09,
243
- "eval_loss": 0.4793929159641266,
244
- "eval_runtime": 139.562,
245
- "eval_samples_per_second": 36.371,
246
- "eval_steps_per_second": 2.279,
247
- "eval_wer": 0.20117654527371606,
248
  "step": 20000
249
  },
250
  {
251
- "epoch": 66.25,
252
- "eval_loss": 0.4844733476638794,
253
- "eval_runtime": 138.9678,
254
- "eval_samples_per_second": 36.526,
255
- "eval_steps_per_second": 2.288,
256
- "eval_wer": 0.20073216666313987,
257
  "step": 21000
258
  },
259
  {
260
- "epoch": 69.4,
261
- "learning_rate": 0.00031628664495114006,
262
- "loss": 0.7308,
263
  "step": 22000
264
  },
265
  {
266
- "epoch": 69.4,
267
- "eval_loss": 0.49372631311416626,
268
- "eval_runtime": 139.7927,
269
- "eval_samples_per_second": 36.311,
270
- "eval_steps_per_second": 2.275,
271
- "eval_wer": 0.20075332754935776,
272
  "step": 22000
273
  },
274
  {
275
- "epoch": 72.55,
276
- "eval_loss": 0.4920376241207123,
277
- "eval_runtime": 138.7644,
278
- "eval_samples_per_second": 36.58,
279
- "eval_steps_per_second": 2.292,
280
- "eval_wer": 0.1894745751952092,
281
  "step": 23000
282
  },
283
  {
284
- "epoch": 75.71,
285
- "learning_rate": 0.0002511400651465798,
286
- "loss": 0.6927,
287
  "step": 24000
288
  },
289
  {
290
- "epoch": 75.71,
291
- "eval_loss": 0.5178954005241394,
292
- "eval_runtime": 139.799,
293
- "eval_samples_per_second": 36.309,
294
- "eval_steps_per_second": 2.275,
295
- "eval_wer": 0.19114628520642443,
296
  "step": 24000
297
  },
298
  {
299
- "epoch": 78.86,
300
- "eval_loss": 0.520152747631073,
301
- "eval_runtime": 140.3812,
302
- "eval_samples_per_second": 36.159,
303
- "eval_steps_per_second": 2.265,
304
- "eval_wer": 0.18767589986668642,
305
  "step": 25000
306
  },
307
  {
308
- "epoch": 82.02,
309
- "learning_rate": 0.00018602605863192182,
310
- "loss": 0.6622,
311
  "step": 26000
312
  },
313
  {
314
- "epoch": 82.02,
315
- "eval_loss": 0.5265706181526184,
316
- "eval_runtime": 138.1289,
317
- "eval_samples_per_second": 36.748,
318
- "eval_steps_per_second": 2.302,
319
- "eval_wer": 0.18401506655098715,
320
  "step": 26000
321
  },
322
  {
323
- "epoch": 85.17,
324
- "eval_loss": 0.5350863933563232,
325
- "eval_runtime": 140.1605,
326
- "eval_samples_per_second": 36.216,
327
- "eval_steps_per_second": 2.269,
328
- "eval_wer": 0.18541168504136954,
329
  "step": 27000
330
  },
331
  {
332
- "epoch": 88.33,
333
- "learning_rate": 0.00012091205211726384,
334
- "loss": 0.6315,
335
  "step": 28000
336
  },
337
  {
338
- "epoch": 88.33,
339
- "eval_loss": 0.5373002290725708,
340
- "eval_runtime": 138.125,
341
- "eval_samples_per_second": 36.749,
342
- "eval_steps_per_second": 2.302,
343
- "eval_wer": 0.18113718602535075,
344
  "step": 28000
345
  },
346
  {
347
- "epoch": 91.48,
348
- "eval_loss": 0.5330832600593567,
349
- "eval_runtime": 139.0156,
350
- "eval_samples_per_second": 36.514,
351
- "eval_steps_per_second": 2.288,
352
- "eval_wer": 0.17923270626573842,
353
  "step": 29000
354
  },
355
  {
356
- "epoch": 94.64,
357
- "learning_rate": 5.576547231270358e-05,
358
- "loss": 0.6075,
359
  "step": 30000
360
  },
361
  {
362
- "epoch": 94.64,
363
- "eval_loss": 0.538992166519165,
364
- "eval_runtime": 138.185,
365
- "eval_samples_per_second": 36.733,
366
- "eval_steps_per_second": 2.301,
367
- "eval_wer": 0.17787840954779185,
368
  "step": 30000
369
  },
370
  {
371
- "epoch": 97.79,
372
- "eval_loss": 0.5459240078926086,
373
- "eval_runtime": 137.8608,
374
- "eval_samples_per_second": 36.82,
375
- "eval_steps_per_second": 2.307,
376
- "eval_wer": 0.17730706561990817,
377
  "step": 31000
378
  },
379
  {
380
- "epoch": 100.0,
381
- "step": 31700,
382
- "total_flos": 3.173184730349909e+20,
383
- "train_loss": 0.8928292760036721,
384
- "train_runtime": 80759.6589,
385
- "train_samples_per_second": 25.144,
386
- "train_steps_per_second": 0.393
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
387
  }
388
  ],
389
- "max_steps": 31700,
390
- "num_train_epochs": 100,
391
- "total_flos": 3.173184730349909e+20,
392
  "trial_name": null,
393
  "trial_params": null
394
  }
 
1
  {
2
  "best_metric": null,
3
  "best_model_checkpoint": null,
4
+ "epoch": 59.999647514980616,
5
+ "global_step": 85080,
6
  "is_hyper_param_search": false,
7
  "is_local_process_zero": true,
8
  "is_world_process_zero": true,
9
  "log_history": [
10
  {
11
+ "epoch": 0.7,
12
+ "eval_loss": 0.5371278524398804,
13
+ "eval_runtime": 423.1427,
14
+ "eval_samples_per_second": 47.847,
15
+ "eval_steps_per_second": 2.992,
16
+ "eval_wer": 0.3810800894541897,
17
  "step": 1000
18
  },
19
  {
20
+ "epoch": 1.41,
21
+ "learning_rate": 0.0009881184586108469,
22
+ "loss": 1.3606,
23
  "step": 2000
24
  },
25
  {
26
+ "epoch": 1.41,
27
+ "eval_loss": 0.5247300267219543,
28
+ "eval_runtime": 419.7941,
29
+ "eval_samples_per_second": 48.228,
30
+ "eval_steps_per_second": 3.016,
31
+ "eval_wer": 0.3902264309176043,
32
  "step": 2000
33
  },
34
  {
35
+ "epoch": 2.12,
36
+ "eval_loss": 0.512639582157135,
37
+ "eval_runtime": 421.2981,
38
+ "eval_samples_per_second": 48.056,
39
+ "eval_steps_per_second": 3.005,
40
+ "eval_wer": 0.38586728632329304,
41
  "step": 3000
42
  },
43
  {
44
+ "epoch": 2.82,
45
+ "learning_rate": 0.0009643553758325404,
46
+ "loss": 1.3671,
47
  "step": 4000
48
  },
49
  {
50
+ "epoch": 2.82,
51
+ "eval_loss": 0.5062423348426819,
52
+ "eval_runtime": 423.4242,
53
+ "eval_samples_per_second": 47.815,
54
+ "eval_steps_per_second": 2.99,
55
+ "eval_wer": 0.3827660912712279,
56
  "step": 4000
57
  },
58
  {
59
+ "epoch": 3.53,
60
+ "eval_loss": 0.49790534377098083,
61
+ "eval_runtime": 423.631,
62
+ "eval_samples_per_second": 47.792,
63
+ "eval_steps_per_second": 2.988,
64
+ "eval_wer": 0.3671902299252219,
65
  "step": 5000
66
  },
67
  {
68
+ "epoch": 4.23,
69
+ "learning_rate": 0.0009405803996194102,
70
+ "loss": 1.3421,
71
  "step": 6000
72
  },
73
  {
74
+ "epoch": 4.23,
75
+ "eval_loss": 0.4905886650085449,
76
+ "eval_runtime": 422.3048,
77
+ "eval_samples_per_second": 47.942,
78
+ "eval_steps_per_second": 2.998,
79
+ "eval_wer": 0.38159549933608217,
80
  "step": 6000
81
  },
82
  {
83
+ "epoch": 4.94,
84
+ "eval_loss": 0.4783899188041687,
85
+ "eval_runtime": 421.2427,
86
+ "eval_samples_per_second": 48.063,
87
+ "eval_steps_per_second": 3.005,
88
+ "eval_wer": 0.36512859039765183,
89
  "step": 7000
90
  },
91
  {
92
+ "epoch": 5.64,
93
+ "learning_rate": 0.0009168054234062797,
94
+ "loss": 1.328,
95
  "step": 8000
96
  },
97
  {
98
+ "epoch": 5.64,
99
+ "eval_loss": 0.48102879524230957,
100
+ "eval_runtime": 420.6976,
101
+ "eval_samples_per_second": 48.125,
102
+ "eval_steps_per_second": 3.009,
103
+ "eval_wer": 0.3669106855825005,
104
  "step": 8000
105
  },
106
  {
107
+ "epoch": 6.35,
108
+ "eval_loss": 0.47466185688972473,
109
+ "eval_runtime": 421.0932,
110
+ "eval_samples_per_second": 48.08,
111
+ "eval_steps_per_second": 3.006,
112
+ "eval_wer": 0.35974736180026556,
113
  "step": 9000
114
  },
115
  {
116
+ "epoch": 7.05,
117
+ "learning_rate": 0.0008930185537583254,
118
+ "loss": 1.3109,
119
  "step": 10000
120
  },
121
  {
122
+ "epoch": 7.05,
123
+ "eval_loss": 0.4812825620174408,
124
+ "eval_runtime": 417.8509,
125
+ "eval_samples_per_second": 48.453,
126
+ "eval_steps_per_second": 3.03,
127
+ "eval_wer": 0.3808267523935984,
128
  "step": 10000
129
  },
130
  {
131
+ "epoch": 7.76,
132
+ "eval_loss": 0.46314355731010437,
133
+ "eval_runtime": 421.3095,
134
+ "eval_samples_per_second": 48.055,
135
+ "eval_steps_per_second": 3.005,
136
+ "eval_wer": 0.3560696065413376,
137
  "step": 11000
138
  },
139
  {
140
+ "epoch": 8.46,
141
+ "learning_rate": 0.0008692435775451952,
142
+ "loss": 1.2873,
143
  "step": 12000
144
  },
145
  {
146
+ "epoch": 8.46,
147
+ "eval_loss": 0.4602561295032501,
148
+ "eval_runtime": 419.2403,
149
+ "eval_samples_per_second": 48.292,
150
+ "eval_steps_per_second": 3.02,
151
+ "eval_wer": 0.3430533230833741,
152
  "step": 12000
153
  },
154
  {
155
+ "epoch": 9.17,
156
+ "eval_loss": 0.4578526020050049,
157
+ "eval_runtime": 418.3262,
158
+ "eval_samples_per_second": 48.398,
159
+ "eval_steps_per_second": 3.026,
160
+ "eval_wer": 0.3532916346355441,
161
  "step": 13000
162
  },
163
  {
164
+ "epoch": 9.87,
165
+ "learning_rate": 0.0008454567078972407,
166
+ "loss": 1.2661,
167
  "step": 14000
168
  },
169
  {
170
+ "epoch": 9.87,
171
+ "eval_loss": 0.44712555408477783,
172
+ "eval_runtime": 420.1714,
173
+ "eval_samples_per_second": 48.185,
174
+ "eval_steps_per_second": 3.013,
175
+ "eval_wer": 0.336475295268712,
176
  "step": 14000
177
  },
178
  {
179
+ "epoch": 10.58,
180
+ "eval_loss": 0.45836716890335083,
181
+ "eval_runtime": 415.4093,
182
+ "eval_samples_per_second": 48.737,
183
+ "eval_steps_per_second": 3.048,
184
+ "eval_wer": 0.3436910336152072,
185
  "step": 15000
186
  },
187
  {
188
+ "epoch": 11.28,
189
+ "learning_rate": 0.0008216817316841104,
190
+ "loss": 1.249,
191
  "step": 16000
192
  },
193
  {
194
+ "epoch": 11.28,
195
+ "eval_loss": 0.4460853040218353,
196
+ "eval_runtime": 419.1988,
197
+ "eval_samples_per_second": 48.297,
198
+ "eval_steps_per_second": 3.02,
199
+ "eval_wer": 0.3454294499965057,
200
  "step": 16000
201
  },
202
  {
203
+ "epoch": 11.99,
204
+ "eval_loss": 0.44824984669685364,
205
+ "eval_runtime": 419.1569,
206
+ "eval_samples_per_second": 48.302,
207
+ "eval_steps_per_second": 3.02,
208
+ "eval_wer": 0.3367024250471731,
209
  "step": 17000
210
  },
211
  {
212
+ "epoch": 12.69,
213
+ "learning_rate": 0.00079790675547098,
214
+ "loss": 1.2322,
215
  "step": 18000
216
  },
217
  {
218
+ "epoch": 12.69,
219
+ "eval_loss": 0.44639432430267334,
220
+ "eval_runtime": 417.2494,
221
+ "eval_samples_per_second": 48.523,
222
+ "eval_steps_per_second": 3.034,
223
+ "eval_wer": 0.33347019358445734,
224
  "step": 18000
225
  },
226
  {
227
+ "epoch": 13.4,
228
+ "eval_loss": 0.4426889717578888,
229
+ "eval_runtime": 417.2512,
230
+ "eval_samples_per_second": 48.522,
231
+ "eval_steps_per_second": 3.034,
232
+ "eval_wer": 0.3454469215179258,
233
  "step": 19000
234
  },
235
  {
236
+ "epoch": 14.1,
237
+ "learning_rate": 0.0007741317792578497,
238
+ "loss": 1.22,
239
  "step": 20000
240
  },
241
  {
242
+ "epoch": 14.1,
243
+ "eval_loss": 0.44404253363609314,
244
+ "eval_runtime": 419.2873,
245
+ "eval_samples_per_second": 48.287,
246
+ "eval_steps_per_second": 3.019,
247
+ "eval_wer": 0.33950660423509676,
248
  "step": 20000
249
  },
250
  {
251
+ "epoch": 14.81,
252
+ "eval_loss": 0.44594520330429077,
253
+ "eval_runtime": 418.1805,
254
+ "eval_samples_per_second": 48.415,
255
+ "eval_steps_per_second": 3.027,
256
+ "eval_wer": 0.3378468097001887,
257
  "step": 21000
258
  },
259
  {
260
+ "epoch": 15.51,
261
+ "learning_rate": 0.0007503449096098953,
262
+ "loss": 1.2044,
263
  "step": 22000
264
  },
265
  {
266
+ "epoch": 15.51,
267
+ "eval_loss": 0.4406070411205292,
268
+ "eval_runtime": 418.2792,
269
+ "eval_samples_per_second": 48.403,
270
+ "eval_steps_per_second": 3.027,
271
+ "eval_wer": 0.3199035572017611,
272
  "step": 22000
273
  },
274
  {
275
+ "epoch": 16.22,
276
+ "eval_loss": 0.4397943317890167,
277
+ "eval_runtime": 419.2827,
278
+ "eval_samples_per_second": 48.287,
279
+ "eval_steps_per_second": 3.019,
280
+ "eval_wer": 0.31545705500034943,
281
  "step": 23000
282
  },
283
  {
284
+ "epoch": 16.92,
285
+ "learning_rate": 0.000726581826831589,
286
+ "loss": 1.1913,
287
  "step": 24000
288
  },
289
  {
290
+ "epoch": 16.92,
291
+ "eval_loss": 0.42369282245635986,
292
+ "eval_runtime": 417.1191,
293
+ "eval_samples_per_second": 48.538,
294
+ "eval_steps_per_second": 3.035,
295
+ "eval_wer": 0.314959116639877,
296
  "step": 24000
297
  },
298
  {
299
+ "epoch": 17.63,
300
+ "eval_loss": 0.4286690652370453,
301
+ "eval_runtime": 422.4918,
302
+ "eval_samples_per_second": 47.92,
303
+ "eval_steps_per_second": 2.997,
304
+ "eval_wer": 0.3278880424907401,
305
  "step": 25000
306
  },
307
  {
308
+ "epoch": 18.34,
309
+ "learning_rate": 0.0007028068506184586,
310
+ "loss": 1.1705,
311
  "step": 26000
312
  },
313
  {
314
+ "epoch": 18.34,
315
+ "eval_loss": 0.4252525866031647,
316
+ "eval_runtime": 419.581,
317
+ "eval_samples_per_second": 48.253,
318
+ "eval_steps_per_second": 3.017,
319
+ "eval_wer": 0.31029422042071425,
320
  "step": 26000
321
  },
322
  {
323
+ "epoch": 19.04,
324
+ "eval_loss": 0.42342108488082886,
325
+ "eval_runtime": 420.8706,
326
+ "eval_samples_per_second": 48.105,
327
+ "eval_steps_per_second": 3.008,
328
+ "eval_wer": 0.30976133901740166,
329
  "step": 27000
330
  },
331
  {
332
+ "epoch": 19.75,
333
+ "learning_rate": 0.0006790199809705044,
334
+ "loss": 1.1564,
335
  "step": 28000
336
  },
337
  {
338
+ "epoch": 19.75,
339
+ "eval_loss": 0.4174347519874573,
340
+ "eval_runtime": 419.264,
341
+ "eval_samples_per_second": 48.289,
342
+ "eval_steps_per_second": 3.02,
343
+ "eval_wer": 0.3076472849255713,
344
  "step": 28000
345
  },
346
  {
347
+ "epoch": 20.45,
348
+ "eval_loss": 0.42600032687187195,
349
+ "eval_runtime": 415.7754,
350
+ "eval_samples_per_second": 48.695,
351
+ "eval_steps_per_second": 3.045,
352
+ "eval_wer": 0.31604235096792227,
353
  "step": 29000
354
  },
355
  {
356
+ "epoch": 21.16,
357
+ "learning_rate": 0.0006552450047573739,
358
+ "loss": 1.1461,
359
  "step": 30000
360
  },
361
  {
362
+ "epoch": 21.16,
363
+ "eval_loss": 0.42350149154663086,
364
+ "eval_runtime": 419.5333,
365
+ "eval_samples_per_second": 48.258,
366
+ "eval_steps_per_second": 3.018,
367
+ "eval_wer": 0.30363757075966175,
368
  "step": 30000
369
  },
370
  {
371
+ "epoch": 21.86,
372
+ "eval_loss": 0.43086713552474976,
373
+ "eval_runtime": 421.0996,
374
+ "eval_samples_per_second": 48.079,
375
+ "eval_steps_per_second": 3.006,
376
+ "eval_wer": 0.30550702355161086,
377
  "step": 31000
378
  },
379
  {
380
+ "epoch": 22.57,
381
+ "learning_rate": 0.0006314581351094196,
382
+ "loss": 1.1285,
383
+ "step": 32000
384
+ },
385
+ {
386
+ "epoch": 22.57,
387
+ "eval_loss": 0.4263738989830017,
388
+ "eval_runtime": 418.0963,
389
+ "eval_samples_per_second": 48.424,
390
+ "eval_steps_per_second": 3.028,
391
+ "eval_wer": 0.3006324690754071,
392
+ "step": 32000
393
+ },
394
+ {
395
+ "epoch": 23.27,
396
+ "eval_loss": 0.420136034488678,
397
+ "eval_runtime": 418.959,
398
+ "eval_samples_per_second": 48.325,
399
+ "eval_steps_per_second": 3.022,
400
+ "eval_wer": 0.28796561604584525,
401
+ "step": 33000
402
+ },
403
+ {
404
+ "epoch": 23.98,
405
+ "learning_rate": 0.0006076950523311132,
406
+ "loss": 1.1135,
407
+ "step": 34000
408
+ },
409
+ {
410
+ "epoch": 23.98,
411
+ "eval_loss": 0.41308358311653137,
412
+ "eval_runtime": 419.6417,
413
+ "eval_samples_per_second": 48.246,
414
+ "eval_steps_per_second": 3.017,
415
+ "eval_wer": 0.29749633098050177,
416
+ "step": 34000
417
+ },
418
+ {
419
+ "epoch": 24.68,
420
+ "eval_loss": 0.42022430896759033,
421
+ "eval_runtime": 415.5643,
422
+ "eval_samples_per_second": 48.719,
423
+ "eval_steps_per_second": 3.046,
424
+ "eval_wer": 0.2848556852330701,
425
+ "step": 35000
426
+ },
427
+ {
428
+ "epoch": 25.39,
429
+ "learning_rate": 0.0005839200761179829,
430
+ "loss": 1.0968,
431
+ "step": 36000
432
+ },
433
+ {
434
+ "epoch": 25.39,
435
+ "eval_loss": 0.41045960783958435,
436
+ "eval_runtime": 416.7195,
437
+ "eval_samples_per_second": 48.584,
438
+ "eval_steps_per_second": 3.038,
439
+ "eval_wer": 0.2887867775525893,
440
+ "step": 36000
441
+ },
442
+ {
443
+ "epoch": 26.09,
444
+ "eval_loss": 0.4209502339363098,
445
+ "eval_runtime": 419.7061,
446
+ "eval_samples_per_second": 48.239,
447
+ "eval_steps_per_second": 3.016,
448
+ "eval_wer": 0.28344922775875325,
449
+ "step": 37000
450
+ },
451
+ {
452
+ "epoch": 26.8,
453
+ "learning_rate": 0.0005601332064700286,
454
+ "loss": 1.087,
455
+ "step": 38000
456
+ },
457
+ {
458
+ "epoch": 26.8,
459
+ "eval_loss": 0.4122714400291443,
460
+ "eval_runtime": 417.6093,
461
+ "eval_samples_per_second": 48.481,
462
+ "eval_steps_per_second": 3.032,
463
+ "eval_wer": 0.2843228038297575,
464
+ "step": 38000
465
+ },
466
+ {
467
+ "epoch": 27.5,
468
+ "eval_loss": 0.42156872153282166,
469
+ "eval_runtime": 415.6358,
470
+ "eval_samples_per_second": 48.711,
471
+ "eval_steps_per_second": 3.046,
472
+ "eval_wer": 0.2802694108602977,
473
+ "step": 39000
474
+ },
475
+ {
476
+ "epoch": 28.21,
477
+ "learning_rate": 0.0005363582302568982,
478
+ "loss": 1.0707,
479
+ "step": 40000
480
+ },
481
+ {
482
+ "epoch": 28.21,
483
+ "eval_loss": 0.4161020517349243,
484
+ "eval_runtime": 421.5813,
485
+ "eval_samples_per_second": 48.024,
486
+ "eval_steps_per_second": 3.003,
487
+ "eval_wer": 0.2786707666503599,
488
+ "step": 40000
489
+ },
490
+ {
491
+ "epoch": 28.91,
492
+ "eval_loss": 0.4186325967311859,
493
+ "eval_runtime": 419.2011,
494
+ "eval_samples_per_second": 48.297,
495
+ "eval_steps_per_second": 3.02,
496
+ "eval_wer": 0.27398839890977705,
497
+ "step": 41000
498
+ },
499
+ {
500
+ "epoch": 29.62,
501
+ "learning_rate": 0.0005125713606089438,
502
+ "loss": 1.0575,
503
+ "step": 42000
504
+ },
505
+ {
506
+ "epoch": 29.62,
507
+ "eval_loss": 0.41177135705947876,
508
+ "eval_runtime": 420.5227,
509
+ "eval_samples_per_second": 48.145,
510
+ "eval_steps_per_second": 3.011,
511
+ "eval_wer": 0.2844625760011182,
512
+ "step": 42000
513
+ },
514
+ {
515
+ "epoch": 30.32,
516
+ "eval_loss": 0.4242798388004303,
517
+ "eval_runtime": 415.2438,
518
+ "eval_samples_per_second": 48.757,
519
+ "eval_steps_per_second": 3.049,
520
+ "eval_wer": 0.27729925221888324,
521
+ "step": 43000
522
+ },
523
+ {
524
+ "epoch": 31.03,
525
+ "learning_rate": 0.0004888082778306375,
526
+ "loss": 1.0474,
527
+ "step": 44000
528
+ },
529
+ {
530
+ "epoch": 31.03,
531
+ "eval_loss": 0.4220682382583618,
532
+ "eval_runtime": 420.3236,
533
+ "eval_samples_per_second": 48.168,
534
+ "eval_steps_per_second": 3.012,
535
+ "eval_wer": 0.27072122440422114,
536
+ "step": 44000
537
+ },
538
+ {
539
+ "epoch": 31.73,
540
+ "eval_loss": 0.4138353765010834,
541
+ "eval_runtime": 416.304,
542
+ "eval_samples_per_second": 48.633,
543
+ "eval_steps_per_second": 3.041,
544
+ "eval_wer": 0.2699524774617374,
545
+ "step": 45000
546
+ },
547
+ {
548
+ "epoch": 32.44,
549
+ "learning_rate": 0.00046502140818268316,
550
+ "loss": 1.0333,
551
+ "step": 46000
552
+ },
553
+ {
554
+ "epoch": 32.44,
555
+ "eval_loss": 0.41024765372276306,
556
+ "eval_runtime": 419.6952,
557
+ "eval_samples_per_second": 48.24,
558
+ "eval_steps_per_second": 3.016,
559
+ "eval_wer": 0.26381997344328745,
560
+ "step": 46000
561
+ },
562
+ {
563
+ "epoch": 33.15,
564
+ "eval_loss": 0.4162220358848572,
565
+ "eval_runtime": 421.7395,
566
+ "eval_samples_per_second": 48.006,
567
+ "eval_steps_per_second": 3.002,
568
+ "eval_wer": 0.26496435809630303,
569
+ "step": 47000
570
+ },
571
+ {
572
+ "epoch": 33.85,
573
+ "learning_rate": 0.00044123453853472884,
574
+ "loss": 1.0191,
575
+ "step": 48000
576
+ },
577
+ {
578
+ "epoch": 33.85,
579
+ "eval_loss": 0.4154505133628845,
580
+ "eval_runtime": 415.0181,
581
+ "eval_samples_per_second": 48.783,
582
+ "eval_steps_per_second": 3.05,
583
+ "eval_wer": 0.2636452582290866,
584
+ "step": 48000
585
+ },
586
+ {
587
+ "epoch": 34.56,
588
+ "eval_loss": 0.41287967562675476,
589
+ "eval_runtime": 419.9379,
590
+ "eval_samples_per_second": 48.212,
591
+ "eval_steps_per_second": 3.015,
592
+ "eval_wer": 0.2655933328674261,
593
+ "step": 49000
594
+ },
595
+ {
596
+ "epoch": 35.26,
597
+ "learning_rate": 0.0004174595623215985,
598
+ "loss": 1.0087,
599
+ "step": 50000
600
+ },
601
+ {
602
+ "epoch": 35.26,
603
+ "eval_loss": 0.4157230257987976,
604
+ "eval_runtime": 417.0542,
605
+ "eval_samples_per_second": 48.545,
606
+ "eval_steps_per_second": 3.036,
607
+ "eval_wer": 0.2631909986721644,
608
+ "step": 50000
609
+ },
610
+ {
611
+ "epoch": 35.97,
612
+ "eval_loss": 0.40904876589775085,
613
+ "eval_runtime": 417.7439,
614
+ "eval_samples_per_second": 48.465,
615
+ "eval_steps_per_second": 3.031,
616
+ "eval_wer": 0.26537493884967506,
617
+ "step": 51000
618
+ },
619
+ {
620
+ "epoch": 36.67,
621
+ "learning_rate": 0.0003936845861084681,
622
+ "loss": 0.9901,
623
+ "step": 52000
624
+ },
625
+ {
626
+ "epoch": 36.67,
627
+ "eval_loss": 0.41830816864967346,
628
+ "eval_runtime": 417.9487,
629
+ "eval_samples_per_second": 48.441,
630
+ "eval_steps_per_second": 3.029,
631
+ "eval_wer": 0.25867461038507233,
632
+ "step": 52000
633
+ },
634
+ {
635
+ "epoch": 37.38,
636
+ "eval_loss": 0.4250655770301819,
637
+ "eval_runtime": 417.46,
638
+ "eval_samples_per_second": 48.498,
639
+ "eval_steps_per_second": 3.033,
640
+ "eval_wer": 0.2648420574463624,
641
+ "step": 53000
642
+ },
643
+ {
644
+ "epoch": 38.08,
645
+ "learning_rate": 0.00036990960989533776,
646
+ "loss": 0.9795,
647
+ "step": 54000
648
+ },
649
+ {
650
+ "epoch": 38.08,
651
+ "eval_loss": 0.4228881299495697,
652
+ "eval_runtime": 417.6821,
653
+ "eval_samples_per_second": 48.472,
654
+ "eval_steps_per_second": 3.031,
655
+ "eval_wer": 0.25547732196519674,
656
+ "step": 54000
657
+ },
658
+ {
659
+ "epoch": 38.79,
660
+ "eval_loss": 0.4176000952720642,
661
+ "eval_runtime": 418.3393,
662
+ "eval_samples_per_second": 48.396,
663
+ "eval_steps_per_second": 3.026,
664
+ "eval_wer": 0.2545513313299322,
665
+ "step": 55000
666
+ },
667
+ {
668
+ "epoch": 39.49,
669
+ "learning_rate": 0.00034613463368220746,
670
+ "loss": 0.9644,
671
+ "step": 56000
672
+ },
673
+ {
674
+ "epoch": 39.49,
675
+ "eval_loss": 0.4222715497016907,
676
+ "eval_runtime": 423.2092,
677
+ "eval_samples_per_second": 47.839,
678
+ "eval_steps_per_second": 2.991,
679
+ "eval_wer": 0.25131036410650637,
680
+ "step": 56000
681
+ },
682
+ {
683
+ "epoch": 40.2,
684
+ "eval_loss": 0.4243711531162262,
685
+ "eval_runtime": 419.2699,
686
+ "eval_samples_per_second": 48.289,
687
+ "eval_steps_per_second": 3.02,
688
+ "eval_wer": 0.2530138374449647,
689
+ "step": 57000
690
+ },
691
+ {
692
+ "epoch": 40.9,
693
+ "learning_rate": 0.0003223477640342531,
694
+ "loss": 0.9534,
695
+ "step": 58000
696
+ },
697
+ {
698
+ "epoch": 40.9,
699
+ "eval_loss": 0.4174785912036896,
700
+ "eval_runtime": 417.6661,
701
+ "eval_samples_per_second": 48.474,
702
+ "eval_steps_per_second": 3.031,
703
+ "eval_wer": 0.2538349989517087,
704
+ "step": 58000
705
+ },
706
+ {
707
+ "epoch": 41.61,
708
+ "eval_loss": 0.4212724566459656,
709
+ "eval_runtime": 425.7723,
710
+ "eval_samples_per_second": 47.551,
711
+ "eval_steps_per_second": 2.973,
712
+ "eval_wer": 0.2505416171640226,
713
+ "step": 59000
714
+ },
715
+ {
716
+ "epoch": 42.31,
717
+ "learning_rate": 0.00029856089438629876,
718
+ "loss": 0.9397,
719
+ "step": 60000
720
+ },
721
+ {
722
+ "epoch": 42.31,
723
+ "eval_loss": 0.4275393486022949,
724
+ "eval_runtime": 419.6181,
725
+ "eval_samples_per_second": 48.249,
726
+ "eval_steps_per_second": 3.017,
727
+ "eval_wer": 0.2565343490111119,
728
+ "step": 60000
729
+ },
730
+ {
731
+ "epoch": 43.02,
732
+ "eval_loss": 0.4315040111541748,
733
+ "eval_runtime": 420.9924,
734
+ "eval_samples_per_second": 48.091,
735
+ "eval_steps_per_second": 3.007,
736
+ "eval_wer": 0.2528129149486337,
737
+ "step": 61000
738
+ },
739
+ {
740
+ "epoch": 43.72,
741
+ "learning_rate": 0.00027479781160799237,
742
+ "loss": 0.9269,
743
+ "step": 62000
744
+ },
745
+ {
746
+ "epoch": 43.72,
747
+ "eval_loss": 0.4316493570804596,
748
+ "eval_runtime": 417.0076,
749
+ "eval_samples_per_second": 48.551,
750
+ "eval_steps_per_second": 3.036,
751
+ "eval_wer": 0.2501048291285205,
752
+ "step": 62000
753
+ },
754
+ {
755
+ "epoch": 44.43,
756
+ "eval_loss": 0.42470675706863403,
757
+ "eval_runtime": 418.3345,
758
+ "eval_samples_per_second": 48.397,
759
+ "eval_steps_per_second": 3.026,
760
+ "eval_wer": 0.24705604864071562,
761
+ "step": 63000
762
+ },
763
+ {
764
+ "epoch": 45.13,
765
+ "learning_rate": 0.00025101094196003804,
766
+ "loss": 0.9175,
767
+ "step": 64000
768
+ },
769
+ {
770
+ "epoch": 45.13,
771
+ "eval_loss": 0.43763282895088196,
772
+ "eval_runtime": 418.9083,
773
+ "eval_samples_per_second": 48.33,
774
+ "eval_steps_per_second": 3.022,
775
+ "eval_wer": 0.24685512614438465,
776
+ "step": 64000
777
+ },
778
+ {
779
+ "epoch": 45.84,
780
+ "eval_loss": 0.4334784150123596,
781
+ "eval_runtime": 418.8375,
782
+ "eval_samples_per_second": 48.339,
783
+ "eval_steps_per_second": 3.023,
784
+ "eval_wer": 0.24501188063456567,
785
+ "step": 65000
786
+ },
787
+ {
788
+ "epoch": 46.54,
789
+ "learning_rate": 0.00022724785918173168,
790
+ "loss": 0.9026,
791
+ "step": 66000
792
+ },
793
+ {
794
+ "epoch": 46.54,
795
+ "eval_loss": 0.4336349070072174,
796
+ "eval_runtime": 419.7669,
797
+ "eval_samples_per_second": 48.232,
798
+ "eval_steps_per_second": 3.016,
799
+ "eval_wer": 0.24517786008805648,
800
+ "step": 66000
801
+ },
802
+ {
803
+ "epoch": 47.25,
804
+ "eval_loss": 0.4399877190589905,
805
+ "eval_runtime": 419.3411,
806
+ "eval_samples_per_second": 48.281,
807
+ "eval_steps_per_second": 3.019,
808
+ "eval_wer": 0.24265322524285415,
809
+ "step": 67000
810
+ },
811
+ {
812
+ "epoch": 47.95,
813
+ "learning_rate": 0.00020346098953377738,
814
+ "loss": 0.8929,
815
+ "step": 68000
816
+ },
817
+ {
818
+ "epoch": 47.95,
819
+ "eval_loss": 0.43824830651283264,
820
+ "eval_runtime": 419.5549,
821
+ "eval_samples_per_second": 48.256,
822
+ "eval_steps_per_second": 3.017,
823
+ "eval_wer": 0.24285414773918512,
824
+ "step": 68000
825
+ },
826
+ {
827
+ "epoch": 48.66,
828
+ "eval_loss": 0.43613117933273315,
829
+ "eval_runtime": 418.6436,
830
+ "eval_samples_per_second": 48.361,
831
+ "eval_steps_per_second": 3.024,
832
+ "eval_wer": 0.24154378363267873,
833
+ "step": 69000
834
+ },
835
+ {
836
+ "epoch": 49.37,
837
+ "learning_rate": 0.00017967411988582303,
838
+ "loss": 0.8786,
839
+ "step": 70000
840
+ },
841
+ {
842
+ "epoch": 49.37,
843
+ "eval_loss": 0.44130945205688477,
844
+ "eval_runtime": 418.5492,
845
+ "eval_samples_per_second": 48.372,
846
+ "eval_steps_per_second": 3.025,
847
+ "eval_wer": 0.23977915996925012,
848
+ "step": 70000
849
+ },
850
+ {
851
+ "epoch": 50.07,
852
+ "eval_loss": 0.43924885988235474,
853
+ "eval_runtime": 419.6699,
854
+ "eval_samples_per_second": 48.243,
855
+ "eval_steps_per_second": 3.017,
856
+ "eval_wer": 0.2415001048291285,
857
+ "step": 71000
858
+ },
859
+ {
860
+ "epoch": 50.78,
861
+ "learning_rate": 0.00015591103710751666,
862
+ "loss": 0.8714,
863
+ "step": 72000
864
+ },
865
+ {
866
+ "epoch": 50.78,
867
+ "eval_loss": 0.4345008134841919,
868
+ "eval_runtime": 418.4339,
869
+ "eval_samples_per_second": 48.385,
870
+ "eval_steps_per_second": 3.026,
871
+ "eval_wer": 0.24062652875812426,
872
+ "step": 72000
873
+ },
874
+ {
875
+ "epoch": 51.48,
876
+ "eval_loss": 0.44752031564712524,
877
+ "eval_runtime": 417.7023,
878
+ "eval_samples_per_second": 48.47,
879
+ "eval_steps_per_second": 3.031,
880
+ "eval_wer": 0.24017226920120205,
881
+ "step": 73000
882
+ },
883
+ {
884
+ "epoch": 52.19,
885
+ "learning_rate": 0.00013212416745956231,
886
+ "loss": 0.8589,
887
+ "step": 74000
888
+ },
889
+ {
890
+ "epoch": 52.19,
891
+ "eval_loss": 0.4473401606082916,
892
+ "eval_runtime": 419.491,
893
+ "eval_samples_per_second": 48.263,
894
+ "eval_steps_per_second": 3.018,
895
+ "eval_wer": 0.23740303305611854,
896
+ "step": 74000
897
+ },
898
+ {
899
+ "epoch": 52.89,
900
+ "eval_loss": 0.4457215368747711,
901
+ "eval_runtime": 416.4935,
902
+ "eval_samples_per_second": 48.611,
903
+ "eval_steps_per_second": 3.04,
904
+ "eval_wer": 0.23568208819624012,
905
+ "step": 75000
906
+ },
907
+ {
908
+ "epoch": 53.6,
909
+ "learning_rate": 0.00010834919124643197,
910
+ "loss": 0.8493,
911
+ "step": 76000
912
+ },
913
+ {
914
+ "epoch": 53.6,
915
+ "eval_loss": 0.44615164399147034,
916
+ "eval_runtime": 416.8896,
917
+ "eval_samples_per_second": 48.564,
918
+ "eval_steps_per_second": 3.037,
919
+ "eval_wer": 0.2365556642672444,
920
+ "step": 76000
921
+ },
922
+ {
923
+ "epoch": 54.3,
924
+ "eval_loss": 0.44939640164375305,
925
+ "eval_runtime": 416.8835,
926
+ "eval_samples_per_second": 48.565,
927
+ "eval_steps_per_second": 3.037,
928
+ "eval_wer": 0.2355947305891397,
929
+ "step": 77000
930
+ },
931
+ {
932
+ "epoch": 55.01,
933
+ "learning_rate": 8.456232159847763e-05,
934
+ "loss": 0.8395,
935
+ "step": 78000
936
+ },
937
+ {
938
+ "epoch": 55.01,
939
+ "eval_loss": 0.44722679257392883,
940
+ "eval_runtime": 419.1529,
941
+ "eval_samples_per_second": 48.302,
942
+ "eval_steps_per_second": 3.02,
943
+ "eval_wer": 0.23519288559647775,
944
+ "step": 78000
945
+ },
946
+ {
947
+ "epoch": 55.71,
948
+ "eval_loss": 0.44897979497909546,
949
+ "eval_runtime": 419.2062,
950
+ "eval_samples_per_second": 48.296,
951
+ "eval_steps_per_second": 3.02,
952
+ "eval_wer": 0.23388252148997135,
953
+ "step": 79000
954
+ },
955
+ {
956
+ "epoch": 56.42,
957
+ "learning_rate": 6.0799238820171265e-05,
958
+ "loss": 0.8295,
959
+ "step": 80000
960
+ },
961
+ {
962
+ "epoch": 56.42,
963
+ "eval_loss": 0.4489339590072632,
964
+ "eval_runtime": 420.6011,
965
+ "eval_samples_per_second": 48.136,
966
+ "eval_steps_per_second": 3.01,
967
+ "eval_wer": 0.23176846739814103,
968
+ "step": 80000
969
+ },
970
+ {
971
+ "epoch": 57.12,
972
+ "eval_loss": 0.4468826949596405,
973
+ "eval_runtime": 417.4618,
974
+ "eval_samples_per_second": 48.498,
975
+ "eval_steps_per_second": 3.033,
976
+ "eval_wer": 0.23203927598015237,
977
+ "step": 81000
978
+ },
979
+ {
980
+ "epoch": 57.83,
981
+ "learning_rate": 3.7012369172216936e-05,
982
+ "loss": 0.8225,
983
+ "step": 82000
984
+ },
985
+ {
986
+ "epoch": 57.83,
987
+ "eval_loss": 0.4478228688240051,
988
+ "eval_runtime": 417.4954,
989
+ "eval_samples_per_second": 48.494,
990
+ "eval_steps_per_second": 3.032,
991
+ "eval_wer": 0.23214410510867287,
992
+ "step": 82000
993
+ },
994
+ {
995
+ "epoch": 58.53,
996
+ "eval_loss": 0.4525238573551178,
997
+ "eval_runtime": 415.49,
998
+ "eval_samples_per_second": 48.728,
999
+ "eval_steps_per_second": 3.047,
1000
+ "eval_wer": 0.2326071004263051,
1001
+ "step": 83000
1002
+ },
1003
+ {
1004
+ "epoch": 59.24,
1005
+ "learning_rate": 1.3225499524262608e-05,
1006
+ "loss": 0.816,
1007
+ "step": 84000
1008
+ },
1009
+ {
1010
+ "epoch": 59.24,
1011
+ "eval_loss": 0.4532177150249481,
1012
+ "eval_runtime": 419.041,
1013
+ "eval_samples_per_second": 48.315,
1014
+ "eval_steps_per_second": 3.021,
1015
+ "eval_wer": 0.2315588091411,
1016
+ "step": 84000
1017
+ },
1018
+ {
1019
+ "epoch": 59.94,
1020
+ "eval_loss": 0.45018914341926575,
1021
+ "eval_runtime": 418.7584,
1022
+ "eval_samples_per_second": 48.348,
1023
+ "eval_steps_per_second": 3.023,
1024
+ "eval_wer": 0.23179467468027115,
1025
+ "step": 85000
1026
+ },
1027
+ {
1028
+ "epoch": 60.0,
1029
+ "step": 85080,
1030
+ "total_flos": 4.6401496923493315e+20,
1031
+ "train_loss": 1.0617500136590194,
1032
+ "train_runtime": 191647.3893,
1033
+ "train_samples_per_second": 28.42,
1034
+ "train_steps_per_second": 0.444
1035
  }
1036
  ],
1037
+ "max_steps": 85080,
1038
+ "num_train_epochs": 60,
1039
+ "total_flos": 4.6401496923493315e+20,
1040
  "trial_name": null,
1041
  "trial_params": null
1042
  }
validation_results.json CHANGED
@@ -1,9 +1,9 @@
1
  {
2
- "epoch": 100.0,
3
- "eval_loss": 0.5438345074653625,
4
- "eval_runtime": 140.268,
5
- "eval_samples": 5076,
6
- "eval_samples_per_second": 36.188,
7
- "eval_steps_per_second": 2.267,
8
- "eval_wer": 0.177349387392344
9
  }
 
1
  {
2
+ "epoch": 60.0,
3
+ "eval_loss": 0.4502493739128113,
4
+ "eval_runtime": 418.5824,
5
+ "eval_samples": 20246,
6
+ "eval_samples_per_second": 48.368,
7
+ "eval_steps_per_second": 3.024,
8
+ "eval_wer": 0.23182961772311134
9
  }