AndrewMcDowell
commited on
Commit
·
7dfcfaf
1
Parent(s):
6c0e054
End of training
Browse files- all_results.json +11 -11
- eval_results.json +6 -6
- train_results.json +5 -5
- trainer_state.json +367 -406
all_results.json
CHANGED
@@ -1,14 +1,14 @@
|
|
1 |
{
|
2 |
"epoch": 50.0,
|
3 |
-
"eval_loss": 0.
|
4 |
-
"eval_runtime":
|
5 |
-
"eval_samples":
|
6 |
-
"eval_samples_per_second": 18.
|
7 |
-
"eval_steps_per_second": 2.
|
8 |
-
"eval_wer":
|
9 |
-
"train_loss": 2.
|
10 |
-
"train_runtime":
|
11 |
-
"train_samples":
|
12 |
-
"train_samples_per_second":
|
13 |
-
"train_steps_per_second": 0.
|
14 |
}
|
|
|
1 |
{
|
2 |
"epoch": 50.0,
|
3 |
+
"eval_loss": 0.5212448239326477,
|
4 |
+
"eval_runtime": 235.8799,
|
5 |
+
"eval_samples": 4466,
|
6 |
+
"eval_samples_per_second": 18.933,
|
7 |
+
"eval_steps_per_second": 2.37,
|
8 |
+
"eval_wer": 1.306780040277467,
|
9 |
+
"train_loss": 2.5914736377068284,
|
10 |
+
"train_runtime": 29653.6134,
|
11 |
+
"train_samples": 10152,
|
12 |
+
"train_samples_per_second": 17.118,
|
13 |
+
"train_steps_per_second": 0.357
|
14 |
}
|
eval_results.json
CHANGED
@@ -1,9 +1,9 @@
|
|
1 |
{
|
2 |
"epoch": 50.0,
|
3 |
-
"eval_loss": 0.
|
4 |
-
"eval_runtime":
|
5 |
-
"eval_samples":
|
6 |
-
"eval_samples_per_second": 18.
|
7 |
-
"eval_steps_per_second": 2.
|
8 |
-
"eval_wer":
|
9 |
}
|
|
|
1 |
{
|
2 |
"epoch": 50.0,
|
3 |
+
"eval_loss": 0.5212448239326477,
|
4 |
+
"eval_runtime": 235.8799,
|
5 |
+
"eval_samples": 4466,
|
6 |
+
"eval_samples_per_second": 18.933,
|
7 |
+
"eval_steps_per_second": 2.37,
|
8 |
+
"eval_wer": 1.306780040277467
|
9 |
}
|
train_results.json
CHANGED
@@ -1,8 +1,8 @@
|
|
1 |
{
|
2 |
"epoch": 50.0,
|
3 |
-
"train_loss": 2.
|
4 |
-
"train_runtime":
|
5 |
-
"train_samples":
|
6 |
-
"train_samples_per_second":
|
7 |
-
"train_steps_per_second": 0.
|
8 |
}
|
|
|
1 |
{
|
2 |
"epoch": 50.0,
|
3 |
+
"train_loss": 2.5914736377068284,
|
4 |
+
"train_runtime": 29653.6134,
|
5 |
+
"train_samples": 10152,
|
6 |
+
"train_samples_per_second": 17.118,
|
7 |
+
"train_steps_per_second": 0.357
|
8 |
}
|
trainer_state.json
CHANGED
@@ -2,789 +2,750 @@
|
|
2 |
"best_metric": null,
|
3 |
"best_model_checkpoint": null,
|
4 |
"epoch": 50.0,
|
5 |
-
"global_step":
|
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.
|
12 |
"learning_rate": 3.6e-06,
|
13 |
-
"loss":
|
14 |
"step": 100
|
15 |
},
|
16 |
{
|
17 |
-
"epoch": 0.
|
18 |
"learning_rate": 7.35e-06,
|
19 |
-
"loss":
|
20 |
"step": 200
|
21 |
},
|
22 |
{
|
23 |
-
"epoch": 1.
|
24 |
"learning_rate": 1.1099999999999999e-05,
|
25 |
-
"loss":
|
26 |
"step": 300
|
27 |
},
|
28 |
{
|
29 |
-
"epoch": 1.
|
30 |
"learning_rate": 1.485e-05,
|
31 |
-
"loss":
|
32 |
"step": 400
|
33 |
},
|
34 |
{
|
35 |
-
"epoch": 2.
|
36 |
"learning_rate": 1.8599999999999998e-05,
|
37 |
-
"loss":
|
38 |
"step": 500
|
39 |
},
|
40 |
{
|
41 |
-
"epoch": 2.
|
42 |
"learning_rate": 2.2349999999999998e-05,
|
43 |
-
"loss":
|
44 |
"step": 600
|
45 |
},
|
46 |
{
|
47 |
-
"epoch": 3.
|
48 |
"learning_rate": 2.6099999999999997e-05,
|
49 |
-
"loss":
|
50 |
"step": 700
|
51 |
},
|
52 |
{
|
53 |
-
"epoch": 3.
|
54 |
"learning_rate": 2.985e-05,
|
55 |
-
"loss": 4.
|
56 |
"step": 800
|
57 |
},
|
58 |
{
|
59 |
-
"epoch": 4.
|
60 |
"learning_rate": 3.36e-05,
|
61 |
-
"loss": 4.
|
62 |
"step": 900
|
63 |
},
|
64 |
{
|
65 |
-
"epoch": 4.
|
66 |
"learning_rate": 3.735e-05,
|
67 |
-
"loss": 4.
|
68 |
"step": 1000
|
69 |
},
|
70 |
{
|
71 |
-
"epoch": 4.
|
72 |
-
"eval_loss": 4.
|
73 |
-
"eval_runtime":
|
74 |
-
"eval_samples_per_second":
|
75 |
-
"eval_steps_per_second": 2.
|
76 |
-
"eval_wer":
|
77 |
"step": 1000
|
78 |
},
|
79 |
{
|
80 |
-
"epoch":
|
81 |
"learning_rate": 4.11e-05,
|
82 |
-
"loss": 4.
|
83 |
"step": 1100
|
84 |
},
|
85 |
{
|
86 |
-
"epoch": 5.
|
87 |
"learning_rate": 4.484999999999999e-05,
|
88 |
-
"loss":
|
89 |
"step": 1200
|
90 |
},
|
91 |
{
|
92 |
-
"epoch":
|
93 |
"learning_rate": 4.8599999999999995e-05,
|
94 |
-
"loss":
|
95 |
"step": 1300
|
96 |
},
|
97 |
{
|
98 |
-
"epoch": 6.
|
99 |
"learning_rate": 5.234999999999999e-05,
|
100 |
-
"loss":
|
101 |
"step": 1400
|
102 |
},
|
103 |
{
|
104 |
-
"epoch":
|
105 |
"learning_rate": 5.6099999999999995e-05,
|
106 |
-
"loss": 3.
|
107 |
"step": 1500
|
108 |
},
|
109 |
{
|
110 |
-
"epoch": 7.
|
111 |
"learning_rate": 5.985e-05,
|
112 |
-
"loss":
|
113 |
"step": 1600
|
114 |
},
|
115 |
{
|
116 |
-
"epoch":
|
117 |
"learning_rate": 6.359999999999999e-05,
|
118 |
-
"loss":
|
119 |
"step": 1700
|
120 |
},
|
121 |
{
|
122 |
-
"epoch": 8.
|
123 |
"learning_rate": 6.735e-05,
|
124 |
-
"loss": 2.
|
125 |
"step": 1800
|
126 |
},
|
127 |
{
|
128 |
-
"epoch": 8.
|
129 |
"learning_rate": 7.11e-05,
|
130 |
-
"loss": 2.
|
131 |
"step": 1900
|
132 |
},
|
133 |
{
|
134 |
-
"epoch": 9.
|
135 |
"learning_rate": 7.484999999999999e-05,
|
136 |
-
"loss": 2.
|
137 |
"step": 2000
|
138 |
},
|
139 |
{
|
140 |
-
"epoch": 9.
|
141 |
-
"eval_loss":
|
142 |
-
"eval_runtime":
|
143 |
-
"eval_samples_per_second":
|
144 |
-
"eval_steps_per_second": 2.
|
145 |
-
"eval_wer":
|
146 |
"step": 2000
|
147 |
},
|
148 |
{
|
149 |
-
"epoch": 9.
|
150 |
-
"learning_rate": 7.
|
151 |
-
"loss": 2.
|
152 |
"step": 2100
|
153 |
},
|
154 |
{
|
155 |
-
"epoch":
|
156 |
-
"learning_rate": 7.
|
157 |
-
"loss": 2.
|
158 |
"step": 2200
|
159 |
},
|
160 |
{
|
161 |
-
"epoch": 10.
|
162 |
-
"learning_rate": 7.
|
163 |
-
"loss":
|
164 |
"step": 2300
|
165 |
},
|
166 |
{
|
167 |
-
"epoch":
|
168 |
-
"learning_rate": 7.
|
169 |
-
"loss":
|
170 |
"step": 2400
|
171 |
},
|
172 |
{
|
173 |
-
"epoch": 11.
|
174 |
-
"learning_rate": 7.
|
175 |
-
"loss":
|
176 |
"step": 2500
|
177 |
},
|
178 |
{
|
179 |
-
"epoch":
|
180 |
-
"learning_rate":
|
181 |
-
"loss": 1.
|
182 |
"step": 2600
|
183 |
},
|
184 |
{
|
185 |
-
"epoch": 12.
|
186 |
-
"learning_rate": 6.
|
187 |
-
"loss": 1.
|
188 |
"step": 2700
|
189 |
},
|
190 |
{
|
191 |
-
"epoch":
|
192 |
-
"learning_rate": 6.
|
193 |
-
"loss": 1.
|
194 |
"step": 2800
|
195 |
},
|
196 |
{
|
197 |
-
"epoch": 13.
|
198 |
-
"learning_rate": 6.
|
199 |
-
"loss": 1.
|
200 |
"step": 2900
|
201 |
},
|
202 |
{
|
203 |
-
"epoch":
|
204 |
-
"learning_rate": 6.
|
205 |
-
"loss": 1.
|
206 |
"step": 3000
|
207 |
},
|
208 |
{
|
209 |
-
"epoch":
|
210 |
-
"eval_loss": 0.
|
211 |
-
"eval_runtime":
|
212 |
-
"eval_samples_per_second":
|
213 |
-
"eval_steps_per_second": 2.
|
214 |
-
"eval_wer":
|
215 |
"step": 3000
|
216 |
},
|
217 |
{
|
218 |
-
"epoch":
|
219 |
-
"learning_rate": 6.
|
220 |
-
"loss": 1.
|
221 |
"step": 3100
|
222 |
},
|
223 |
{
|
224 |
-
"epoch":
|
225 |
-
"learning_rate": 6.
|
226 |
-
"loss": 1.
|
227 |
"step": 3200
|
228 |
},
|
229 |
{
|
230 |
-
"epoch":
|
231 |
-
"learning_rate": 6.
|
232 |
-
"loss": 1.
|
233 |
"step": 3300
|
234 |
},
|
235 |
{
|
236 |
-
"epoch":
|
237 |
-
"learning_rate": 6.
|
238 |
-
"loss": 1.
|
239 |
"step": 3400
|
240 |
},
|
241 |
{
|
242 |
-
"epoch":
|
243 |
-
"learning_rate": 6.
|
244 |
-
"loss": 1.
|
245 |
"step": 3500
|
246 |
},
|
247 |
{
|
248 |
-
"epoch": 16.
|
249 |
-
"learning_rate": 6.
|
250 |
-
"loss": 1.
|
251 |
"step": 3600
|
252 |
},
|
253 |
{
|
254 |
-
"epoch":
|
255 |
-
"learning_rate": 6.
|
256 |
-
"loss": 1.
|
257 |
"step": 3700
|
258 |
},
|
259 |
{
|
260 |
-
"epoch": 17.
|
261 |
-
"learning_rate":
|
262 |
-
"loss": 1.
|
263 |
"step": 3800
|
264 |
},
|
265 |
{
|
266 |
-
"epoch":
|
267 |
-
"learning_rate": 5.
|
268 |
-
"loss": 1.
|
269 |
"step": 3900
|
270 |
},
|
271 |
{
|
272 |
-
"epoch": 18.
|
273 |
-
"learning_rate": 5.
|
274 |
-
"loss": 1.
|
275 |
"step": 4000
|
276 |
},
|
277 |
{
|
278 |
-
"epoch": 18.
|
279 |
-
"eval_loss": 0.
|
280 |
-
"eval_runtime":
|
281 |
-
"eval_samples_per_second":
|
282 |
-
"eval_steps_per_second": 2.
|
283 |
-
"eval_wer":
|
284 |
"step": 4000
|
285 |
},
|
286 |
{
|
287 |
-
"epoch":
|
288 |
-
"learning_rate": 5.
|
289 |
-
"loss": 1.
|
290 |
"step": 4100
|
291 |
},
|
292 |
{
|
293 |
-
"epoch":
|
294 |
-
"learning_rate": 5.
|
295 |
-
"loss": 1.
|
296 |
"step": 4200
|
297 |
},
|
298 |
{
|
299 |
-
"epoch":
|
300 |
-
"learning_rate": 5.
|
301 |
-
"loss": 1.
|
302 |
"step": 4300
|
303 |
},
|
304 |
{
|
305 |
-
"epoch":
|
306 |
-
"learning_rate": 5.
|
307 |
-
"loss": 1.
|
308 |
"step": 4400
|
309 |
},
|
310 |
{
|
311 |
-
"epoch":
|
312 |
-
"learning_rate": 5.
|
313 |
-
"loss": 1.
|
314 |
"step": 4500
|
315 |
},
|
316 |
{
|
317 |
-
"epoch":
|
318 |
-
"learning_rate": 5.
|
319 |
-
"loss": 1.
|
320 |
"step": 4600
|
321 |
},
|
322 |
{
|
323 |
-
"epoch":
|
324 |
-
"learning_rate": 5.
|
325 |
-
"loss": 1.
|
326 |
"step": 4700
|
327 |
},
|
328 |
{
|
329 |
-
"epoch":
|
330 |
-
"learning_rate": 5.
|
331 |
-
"loss": 1.
|
332 |
"step": 4800
|
333 |
},
|
334 |
{
|
335 |
-
"epoch":
|
336 |
-
"learning_rate":
|
337 |
-
"loss": 1.
|
338 |
"step": 4900
|
339 |
},
|
340 |
{
|
341 |
-
"epoch":
|
342 |
-
"learning_rate":
|
343 |
-
"loss": 1.
|
344 |
"step": 5000
|
345 |
},
|
346 |
{
|
347 |
-
"epoch":
|
348 |
-
"eval_loss": 0.
|
349 |
-
"eval_runtime":
|
350 |
-
"eval_samples_per_second": 18.
|
351 |
-
"eval_steps_per_second": 2.
|
352 |
-
"eval_wer":
|
353 |
"step": 5000
|
354 |
},
|
355 |
{
|
356 |
-
"epoch":
|
357 |
-
"learning_rate": 4.
|
358 |
-
"loss": 1.
|
359 |
"step": 5100
|
360 |
},
|
361 |
{
|
362 |
-
"epoch":
|
363 |
-
"learning_rate": 4.
|
364 |
-
"loss": 1.
|
365 |
"step": 5200
|
366 |
},
|
367 |
{
|
368 |
-
"epoch":
|
369 |
-
"learning_rate": 4.
|
370 |
-
"loss": 1.
|
371 |
"step": 5300
|
372 |
},
|
373 |
{
|
374 |
-
"epoch":
|
375 |
-
"learning_rate": 4.
|
376 |
-
"loss": 1.
|
377 |
"step": 5400
|
378 |
},
|
379 |
{
|
380 |
-
"epoch":
|
381 |
-
"learning_rate": 4.
|
382 |
-
"loss": 1.
|
383 |
"step": 5500
|
384 |
},
|
385 |
{
|
386 |
-
"epoch":
|
387 |
-
"learning_rate": 4.
|
388 |
-
"loss": 1.
|
389 |
"step": 5600
|
390 |
},
|
391 |
{
|
392 |
-
"epoch":
|
393 |
-
"learning_rate": 4.
|
394 |
-
"loss": 1.
|
395 |
"step": 5700
|
396 |
},
|
397 |
{
|
398 |
-
"epoch":
|
399 |
-
"learning_rate": 4.
|
400 |
-
"loss": 1.
|
401 |
"step": 5800
|
402 |
},
|
403 |
{
|
404 |
-
"epoch":
|
405 |
-
"learning_rate": 4.
|
406 |
-
"loss": 1.
|
407 |
"step": 5900
|
408 |
},
|
409 |
{
|
410 |
-
"epoch":
|
411 |
-
"learning_rate": 4.
|
412 |
-
"loss": 1.
|
413 |
"step": 6000
|
414 |
},
|
415 |
{
|
416 |
-
"epoch":
|
417 |
-
"eval_loss": 0.
|
418 |
-
"eval_runtime":
|
419 |
-
"eval_samples_per_second":
|
420 |
-
"eval_steps_per_second": 2.
|
421 |
-
"eval_wer":
|
422 |
"step": 6000
|
423 |
},
|
424 |
{
|
425 |
-
"epoch":
|
426 |
-
"learning_rate":
|
427 |
-
"loss": 1.
|
428 |
"step": 6100
|
429 |
},
|
430 |
{
|
431 |
-
"epoch":
|
432 |
-
"learning_rate":
|
433 |
-
"loss": 1.
|
434 |
"step": 6200
|
435 |
},
|
436 |
{
|
437 |
-
"epoch":
|
438 |
-
"learning_rate": 3.
|
439 |
-
"loss": 1.
|
440 |
"step": 6300
|
441 |
},
|
442 |
{
|
443 |
-
"epoch":
|
444 |
-
"learning_rate": 3.
|
445 |
-
"loss": 1.
|
446 |
"step": 6400
|
447 |
},
|
448 |
{
|
449 |
-
"epoch":
|
450 |
-
"learning_rate": 3.
|
451 |
-
"loss": 1.
|
452 |
"step": 6500
|
453 |
},
|
454 |
{
|
455 |
-
"epoch":
|
456 |
-
"learning_rate": 3.
|
457 |
-
"loss": 1.
|
458 |
"step": 6600
|
459 |
},
|
460 |
{
|
461 |
-
"epoch":
|
462 |
-
"learning_rate": 3.
|
463 |
-
"loss": 1.
|
464 |
"step": 6700
|
465 |
},
|
466 |
{
|
467 |
-
"epoch":
|
468 |
-
"learning_rate": 3.
|
469 |
-
"loss": 1.
|
470 |
"step": 6800
|
471 |
},
|
472 |
{
|
473 |
-
"epoch":
|
474 |
-
"learning_rate": 3.
|
475 |
-
"loss": 1.
|
476 |
"step": 6900
|
477 |
},
|
478 |
{
|
479 |
-
"epoch":
|
480 |
-
"learning_rate": 3.
|
481 |
-
"loss": 1.
|
482 |
"step": 7000
|
483 |
},
|
484 |
{
|
485 |
-
"epoch":
|
486 |
-
"eval_loss": 0.
|
487 |
-
"eval_runtime":
|
488 |
-
"eval_samples_per_second":
|
489 |
-
"eval_steps_per_second": 2.
|
490 |
-
"eval_wer":
|
491 |
"step": 7000
|
492 |
},
|
493 |
{
|
494 |
-
"epoch":
|
495 |
-
"learning_rate": 3.
|
496 |
-
"loss": 1.
|
497 |
"step": 7100
|
498 |
},
|
499 |
{
|
500 |
-
"epoch":
|
501 |
-
"learning_rate":
|
502 |
-
"loss": 1.
|
503 |
"step": 7200
|
504 |
},
|
505 |
{
|
506 |
-
"epoch":
|
507 |
-
"learning_rate":
|
508 |
-
"loss": 1.
|
509 |
"step": 7300
|
510 |
},
|
511 |
{
|
512 |
-
"epoch":
|
513 |
-
"learning_rate":
|
514 |
-
"loss": 1.
|
515 |
"step": 7400
|
516 |
},
|
517 |
{
|
518 |
-
"epoch":
|
519 |
-
"learning_rate": 2.
|
520 |
-
"loss": 1.
|
521 |
"step": 7500
|
522 |
},
|
523 |
{
|
524 |
-
"epoch":
|
525 |
-
"learning_rate": 2.
|
526 |
-
"loss": 1.
|
527 |
"step": 7600
|
528 |
},
|
529 |
{
|
530 |
-
"epoch":
|
531 |
-
"learning_rate": 2.
|
532 |
-
"loss": 1.
|
533 |
"step": 7700
|
534 |
},
|
535 |
{
|
536 |
-
"epoch":
|
537 |
-
"learning_rate": 2.
|
538 |
-
"loss": 1.
|
539 |
"step": 7800
|
540 |
},
|
541 |
{
|
542 |
-
"epoch":
|
543 |
-
"learning_rate": 2.
|
544 |
-
"loss": 1.
|
545 |
"step": 7900
|
546 |
},
|
547 |
{
|
548 |
-
"epoch":
|
549 |
-
"learning_rate": 2.
|
550 |
-
"loss": 1.
|
551 |
"step": 8000
|
552 |
},
|
553 |
{
|
554 |
-
"epoch":
|
555 |
-
"eval_loss": 0.
|
556 |
-
"eval_runtime":
|
557 |
-
"eval_samples_per_second":
|
558 |
-
"eval_steps_per_second": 2.
|
559 |
-
"eval_wer":
|
560 |
"step": 8000
|
561 |
},
|
562 |
{
|
563 |
-
"epoch":
|
564 |
-
"learning_rate": 2.
|
565 |
-
"loss": 1.
|
566 |
"step": 8100
|
567 |
},
|
568 |
{
|
569 |
-
"epoch":
|
570 |
-
"learning_rate": 2.
|
571 |
-
"loss": 1.
|
572 |
"step": 8200
|
573 |
},
|
574 |
{
|
575 |
-
"epoch":
|
576 |
-
"learning_rate": 2.
|
577 |
-
"loss": 1.
|
578 |
"step": 8300
|
579 |
},
|
580 |
{
|
581 |
-
"epoch":
|
582 |
-
"learning_rate":
|
583 |
-
"loss": 1.
|
584 |
"step": 8400
|
585 |
},
|
586 |
{
|
587 |
-
"epoch":
|
588 |
-
"learning_rate":
|
589 |
-
"loss": 1.
|
590 |
"step": 8500
|
591 |
},
|
592 |
{
|
593 |
-
"epoch":
|
594 |
-
"learning_rate":
|
595 |
-
"loss": 1.
|
596 |
"step": 8600
|
597 |
},
|
598 |
{
|
599 |
-
"epoch":
|
600 |
-
"learning_rate": 1.
|
601 |
-
"loss": 1.
|
602 |
"step": 8700
|
603 |
},
|
604 |
{
|
605 |
-
"epoch":
|
606 |
-
"learning_rate": 1.
|
607 |
-
"loss": 1.
|
608 |
"step": 8800
|
609 |
},
|
610 |
{
|
611 |
-
"epoch":
|
612 |
-
"learning_rate": 1.
|
613 |
-
"loss": 1.
|
614 |
"step": 8900
|
615 |
},
|
616 |
{
|
617 |
-
"epoch":
|
618 |
-
"learning_rate": 1.
|
619 |
-
"loss": 1.
|
620 |
"step": 9000
|
621 |
},
|
622 |
{
|
623 |
-
"epoch":
|
624 |
-
"eval_loss": 0.
|
625 |
-
"eval_runtime":
|
626 |
-
"eval_samples_per_second":
|
627 |
-
"eval_steps_per_second": 2.
|
628 |
-
"eval_wer":
|
629 |
"step": 9000
|
630 |
},
|
631 |
{
|
632 |
-
"epoch":
|
633 |
-
"learning_rate": 1.
|
634 |
-
"loss": 1.
|
635 |
"step": 9100
|
636 |
},
|
637 |
{
|
638 |
-
"epoch":
|
639 |
-
"learning_rate": 1.
|
640 |
-
"loss": 1.
|
641 |
"step": 9200
|
642 |
},
|
643 |
{
|
644 |
-
"epoch":
|
645 |
-
"learning_rate": 1.
|
646 |
-
"loss": 1.
|
647 |
"step": 9300
|
648 |
},
|
649 |
{
|
650 |
-
"epoch":
|
651 |
-
"learning_rate": 1.
|
652 |
-
"loss": 1.
|
653 |
"step": 9400
|
654 |
},
|
655 |
{
|
656 |
-
"epoch":
|
657 |
-
"learning_rate":
|
658 |
-
"loss": 1.
|
659 |
"step": 9500
|
660 |
},
|
661 |
{
|
662 |
-
"epoch":
|
663 |
-
"learning_rate":
|
664 |
-
"loss": 1.
|
665 |
"step": 9600
|
666 |
},
|
667 |
{
|
668 |
-
"epoch":
|
669 |
-
"learning_rate":
|
670 |
-
"loss": 1.
|
671 |
"step": 9700
|
672 |
},
|
673 |
{
|
674 |
-
"epoch":
|
675 |
-
"learning_rate":
|
676 |
-
"loss": 1.
|
677 |
"step": 9800
|
678 |
},
|
679 |
{
|
680 |
-
"epoch":
|
681 |
-
"learning_rate":
|
682 |
-
"loss": 1.
|
683 |
"step": 9900
|
684 |
},
|
685 |
{
|
686 |
-
"epoch":
|
687 |
-
"learning_rate":
|
688 |
-
"loss": 1.
|
689 |
"step": 10000
|
690 |
},
|
691 |
{
|
692 |
-
"epoch":
|
693 |
-
"eval_loss": 0.
|
694 |
-
"eval_runtime":
|
695 |
-
"eval_samples_per_second":
|
696 |
-
"eval_steps_per_second": 2.
|
697 |
-
"eval_wer":
|
698 |
"step": 10000
|
699 |
},
|
700 |
{
|
701 |
-
"epoch":
|
702 |
-
"learning_rate":
|
703 |
-
"loss": 1.
|
704 |
"step": 10100
|
705 |
},
|
706 |
{
|
707 |
-
"epoch":
|
708 |
-
"learning_rate":
|
709 |
-
"loss": 1.
|
710 |
"step": 10200
|
711 |
},
|
712 |
{
|
713 |
-
"epoch":
|
714 |
-
"learning_rate":
|
715 |
-
"loss": 1.
|
716 |
"step": 10300
|
717 |
},
|
718 |
{
|
719 |
-
"epoch":
|
720 |
-
"learning_rate":
|
721 |
-
"loss": 1.
|
722 |
"step": 10400
|
723 |
},
|
724 |
{
|
725 |
-
"epoch":
|
726 |
-
"learning_rate":
|
727 |
-
"loss": 1.
|
728 |
"step": 10500
|
729 |
},
|
730 |
-
{
|
731 |
-
"epoch": 47.75,
|
732 |
-
"learning_rate": 4.17032967032967e-06,
|
733 |
-
"loss": 1.3462,
|
734 |
-
"step": 10600
|
735 |
-
},
|
736 |
-
{
|
737 |
-
"epoch": 48.2,
|
738 |
-
"learning_rate": 3.3461538461538457e-06,
|
739 |
-
"loss": 1.3636,
|
740 |
-
"step": 10700
|
741 |
-
},
|
742 |
-
{
|
743 |
-
"epoch": 48.65,
|
744 |
-
"learning_rate": 2.521978021978022e-06,
|
745 |
-
"loss": 1.3366,
|
746 |
-
"step": 10800
|
747 |
-
},
|
748 |
-
{
|
749 |
-
"epoch": 49.1,
|
750 |
-
"learning_rate": 1.6978021978021975e-06,
|
751 |
-
"loss": 1.337,
|
752 |
-
"step": 10900
|
753 |
-
},
|
754 |
-
{
|
755 |
-
"epoch": 49.55,
|
756 |
-
"learning_rate": 8.818681318681318e-07,
|
757 |
-
"loss": 1.3298,
|
758 |
-
"step": 11000
|
759 |
-
},
|
760 |
-
{
|
761 |
-
"epoch": 49.55,
|
762 |
-
"eval_loss": 0.5348710417747498,
|
763 |
-
"eval_runtime": 238.2007,
|
764 |
-
"eval_samples_per_second": 18.82,
|
765 |
-
"eval_steps_per_second": 2.355,
|
766 |
-
"eval_wer": 2.627446619217082,
|
767 |
-
"step": 11000
|
768 |
-
},
|
769 |
{
|
770 |
"epoch": 50.0,
|
771 |
-
"learning_rate":
|
772 |
-
"loss": 1.
|
773 |
-
"step":
|
774 |
},
|
775 |
{
|
776 |
"epoch": 50.0,
|
777 |
-
"step":
|
778 |
-
"total_flos":
|
779 |
-
"train_loss": 2.
|
780 |
-
"train_runtime":
|
781 |
-
"train_samples_per_second":
|
782 |
-
"train_steps_per_second": 0.
|
783 |
}
|
784 |
],
|
785 |
-
"max_steps":
|
786 |
"num_train_epochs": 50,
|
787 |
-
"total_flos":
|
788 |
"trial_name": null,
|
789 |
"trial_params": null
|
790 |
}
|
|
|
2 |
"best_metric": null,
|
3 |
"best_model_checkpoint": null,
|
4 |
"epoch": 50.0,
|
5 |
+
"global_step": 10600,
|
6 |
"is_hyper_param_search": false,
|
7 |
"is_local_process_zero": true,
|
8 |
"is_world_process_zero": true,
|
9 |
"log_history": [
|
10 |
{
|
11 |
+
"epoch": 0.47,
|
12 |
"learning_rate": 3.6e-06,
|
13 |
+
"loss": 36.2649,
|
14 |
"step": 100
|
15 |
},
|
16 |
{
|
17 |
+
"epoch": 0.94,
|
18 |
"learning_rate": 7.35e-06,
|
19 |
+
"loss": 21.3706,
|
20 |
"step": 200
|
21 |
},
|
22 |
{
|
23 |
+
"epoch": 1.42,
|
24 |
"learning_rate": 1.1099999999999999e-05,
|
25 |
+
"loss": 13.7098,
|
26 |
"step": 300
|
27 |
},
|
28 |
{
|
29 |
+
"epoch": 1.89,
|
30 |
"learning_rate": 1.485e-05,
|
31 |
+
"loss": 10.7365,
|
32 |
"step": 400
|
33 |
},
|
34 |
{
|
35 |
+
"epoch": 2.36,
|
36 |
"learning_rate": 1.8599999999999998e-05,
|
37 |
+
"loss": 7.9817,
|
38 |
"step": 500
|
39 |
},
|
40 |
{
|
41 |
+
"epoch": 2.83,
|
42 |
"learning_rate": 2.2349999999999998e-05,
|
43 |
+
"loss": 5.8439,
|
44 |
"step": 600
|
45 |
},
|
46 |
{
|
47 |
+
"epoch": 3.3,
|
48 |
"learning_rate": 2.6099999999999997e-05,
|
49 |
+
"loss": 4.7275,
|
50 |
"step": 700
|
51 |
},
|
52 |
{
|
53 |
+
"epoch": 3.77,
|
54 |
"learning_rate": 2.985e-05,
|
55 |
+
"loss": 4.3969,
|
56 |
"step": 800
|
57 |
},
|
58 |
{
|
59 |
+
"epoch": 4.25,
|
60 |
"learning_rate": 3.36e-05,
|
61 |
+
"loss": 4.1733,
|
62 |
"step": 900
|
63 |
},
|
64 |
{
|
65 |
+
"epoch": 4.72,
|
66 |
"learning_rate": 3.735e-05,
|
67 |
+
"loss": 4.0974,
|
68 |
"step": 1000
|
69 |
},
|
70 |
{
|
71 |
+
"epoch": 4.72,
|
72 |
+
"eval_loss": 4.017753601074219,
|
73 |
+
"eval_runtime": 233.3211,
|
74 |
+
"eval_samples_per_second": 19.141,
|
75 |
+
"eval_steps_per_second": 2.396,
|
76 |
+
"eval_wer": 1.9534571492503916,
|
77 |
"step": 1000
|
78 |
},
|
79 |
{
|
80 |
+
"epoch": 5.19,
|
81 |
"learning_rate": 4.11e-05,
|
82 |
+
"loss": 4.0623,
|
83 |
"step": 1100
|
84 |
},
|
85 |
{
|
86 |
+
"epoch": 5.66,
|
87 |
"learning_rate": 4.484999999999999e-05,
|
88 |
+
"loss": 3.9986,
|
89 |
"step": 1200
|
90 |
},
|
91 |
{
|
92 |
+
"epoch": 6.13,
|
93 |
"learning_rate": 4.8599999999999995e-05,
|
94 |
+
"loss": 3.9302,
|
95 |
"step": 1300
|
96 |
},
|
97 |
{
|
98 |
+
"epoch": 6.6,
|
99 |
"learning_rate": 5.234999999999999e-05,
|
100 |
+
"loss": 3.7211,
|
101 |
"step": 1400
|
102 |
},
|
103 |
{
|
104 |
+
"epoch": 7.08,
|
105 |
"learning_rate": 5.6099999999999995e-05,
|
106 |
+
"loss": 3.162,
|
107 |
"step": 1500
|
108 |
},
|
109 |
{
|
110 |
+
"epoch": 7.55,
|
111 |
"learning_rate": 5.985e-05,
|
112 |
+
"loss": 2.6718,
|
113 |
"step": 1600
|
114 |
},
|
115 |
{
|
116 |
+
"epoch": 8.02,
|
117 |
"learning_rate": 6.359999999999999e-05,
|
118 |
+
"loss": 2.4628,
|
119 |
"step": 1700
|
120 |
},
|
121 |
{
|
122 |
+
"epoch": 8.49,
|
123 |
"learning_rate": 6.735e-05,
|
124 |
+
"loss": 2.2766,
|
125 |
"step": 1800
|
126 |
},
|
127 |
{
|
128 |
+
"epoch": 8.96,
|
129 |
"learning_rate": 7.11e-05,
|
130 |
+
"loss": 2.2276,
|
131 |
"step": 1900
|
132 |
},
|
133 |
{
|
134 |
+
"epoch": 9.43,
|
135 |
"learning_rate": 7.484999999999999e-05,
|
136 |
+
"loss": 2.1276,
|
137 |
"step": 2000
|
138 |
},
|
139 |
{
|
140 |
+
"epoch": 9.43,
|
141 |
+
"eval_loss": 0.930120587348938,
|
142 |
+
"eval_runtime": 231.4711,
|
143 |
+
"eval_samples_per_second": 19.294,
|
144 |
+
"eval_steps_per_second": 2.415,
|
145 |
+
"eval_wer": 1.2127992839561423,
|
146 |
"step": 2000
|
147 |
},
|
148 |
{
|
149 |
+
"epoch": 9.91,
|
150 |
+
"learning_rate": 7.416279069767441e-05,
|
151 |
+
"loss": 2.0853,
|
152 |
"step": 2100
|
153 |
},
|
154 |
{
|
155 |
+
"epoch": 10.38,
|
156 |
+
"learning_rate": 7.329069767441859e-05,
|
157 |
+
"loss": 2.0115,
|
158 |
"step": 2200
|
159 |
},
|
160 |
{
|
161 |
+
"epoch": 10.85,
|
162 |
+
"learning_rate": 7.241860465116278e-05,
|
163 |
+
"loss": 1.9745,
|
164 |
"step": 2300
|
165 |
},
|
166 |
{
|
167 |
+
"epoch": 11.32,
|
168 |
+
"learning_rate": 7.154651162790697e-05,
|
169 |
+
"loss": 1.9235,
|
170 |
"step": 2400
|
171 |
},
|
172 |
{
|
173 |
+
"epoch": 11.79,
|
174 |
+
"learning_rate": 7.067441860465115e-05,
|
175 |
+
"loss": 1.9008,
|
176 |
"step": 2500
|
177 |
},
|
178 |
{
|
179 |
+
"epoch": 12.26,
|
180 |
+
"learning_rate": 6.980232558139535e-05,
|
181 |
+
"loss": 1.878,
|
182 |
"step": 2600
|
183 |
},
|
184 |
{
|
185 |
+
"epoch": 12.74,
|
186 |
+
"learning_rate": 6.893023255813953e-05,
|
187 |
+
"loss": 1.8435,
|
188 |
"step": 2700
|
189 |
},
|
190 |
{
|
191 |
+
"epoch": 13.21,
|
192 |
+
"learning_rate": 6.805813953488372e-05,
|
193 |
+
"loss": 1.8357,
|
194 |
"step": 2800
|
195 |
},
|
196 |
{
|
197 |
+
"epoch": 13.68,
|
198 |
+
"learning_rate": 6.71860465116279e-05,
|
199 |
+
"loss": 1.8002,
|
200 |
"step": 2900
|
201 |
},
|
202 |
{
|
203 |
+
"epoch": 14.15,
|
204 |
+
"learning_rate": 6.631395348837209e-05,
|
205 |
+
"loss": 1.7622,
|
206 |
"step": 3000
|
207 |
},
|
208 |
{
|
209 |
+
"epoch": 14.15,
|
210 |
+
"eval_loss": 0.7103351354598999,
|
211 |
+
"eval_runtime": 235.0291,
|
212 |
+
"eval_samples_per_second": 19.002,
|
213 |
+
"eval_steps_per_second": 2.378,
|
214 |
+
"eval_wer": 1.5526963526516,
|
215 |
"step": 3000
|
216 |
},
|
217 |
{
|
218 |
+
"epoch": 14.62,
|
219 |
+
"learning_rate": 6.544186046511627e-05,
|
220 |
+
"loss": 1.7696,
|
221 |
"step": 3100
|
222 |
},
|
223 |
{
|
224 |
+
"epoch": 15.09,
|
225 |
+
"learning_rate": 6.456976744186045e-05,
|
226 |
+
"loss": 1.732,
|
227 |
"step": 3200
|
228 |
},
|
229 |
{
|
230 |
+
"epoch": 15.57,
|
231 |
+
"learning_rate": 6.369767441860465e-05,
|
232 |
+
"loss": 1.7099,
|
233 |
"step": 3300
|
234 |
},
|
235 |
{
|
236 |
+
"epoch": 16.04,
|
237 |
+
"learning_rate": 6.282558139534884e-05,
|
238 |
+
"loss": 1.7177,
|
239 |
"step": 3400
|
240 |
},
|
241 |
{
|
242 |
+
"epoch": 16.51,
|
243 |
+
"learning_rate": 6.195348837209302e-05,
|
244 |
+
"loss": 1.6886,
|
245 |
"step": 3500
|
246 |
},
|
247 |
{
|
248 |
+
"epoch": 16.98,
|
249 |
+
"learning_rate": 6.10813953488372e-05,
|
250 |
+
"loss": 1.7001,
|
251 |
"step": 3600
|
252 |
},
|
253 |
{
|
254 |
+
"epoch": 17.45,
|
255 |
+
"learning_rate": 6.020930232558139e-05,
|
256 |
+
"loss": 1.6783,
|
257 |
"step": 3700
|
258 |
},
|
259 |
{
|
260 |
+
"epoch": 17.92,
|
261 |
+
"learning_rate": 5.933720930232557e-05,
|
262 |
+
"loss": 1.6688,
|
263 |
"step": 3800
|
264 |
},
|
265 |
{
|
266 |
+
"epoch": 18.4,
|
267 |
+
"learning_rate": 5.8465116279069764e-05,
|
268 |
+
"loss": 1.6316,
|
269 |
"step": 3900
|
270 |
},
|
271 |
{
|
272 |
+
"epoch": 18.87,
|
273 |
+
"learning_rate": 5.759302325581395e-05,
|
274 |
+
"loss": 1.6397,
|
275 |
"step": 4000
|
276 |
},
|
277 |
{
|
278 |
+
"epoch": 18.87,
|
279 |
+
"eval_loss": 0.6728963851928711,
|
280 |
+
"eval_runtime": 232.2669,
|
281 |
+
"eval_samples_per_second": 19.228,
|
282 |
+
"eval_steps_per_second": 2.407,
|
283 |
+
"eval_wer": 1.4269411501454463,
|
284 |
"step": 4000
|
285 |
},
|
286 |
{
|
287 |
+
"epoch": 19.34,
|
288 |
+
"learning_rate": 5.672093023255814e-05,
|
289 |
+
"loss": 1.6214,
|
290 |
"step": 4100
|
291 |
},
|
292 |
{
|
293 |
+
"epoch": 19.81,
|
294 |
+
"learning_rate": 5.5848837209302316e-05,
|
295 |
+
"loss": 1.6283,
|
296 |
"step": 4200
|
297 |
},
|
298 |
{
|
299 |
+
"epoch": 20.28,
|
300 |
+
"learning_rate": 5.49767441860465e-05,
|
301 |
+
"loss": 1.6226,
|
302 |
"step": 4300
|
303 |
},
|
304 |
{
|
305 |
+
"epoch": 20.75,
|
306 |
+
"learning_rate": 5.411337209302325e-05,
|
307 |
+
"loss": 1.5882,
|
308 |
"step": 4400
|
309 |
},
|
310 |
{
|
311 |
+
"epoch": 21.23,
|
312 |
+
"learning_rate": 5.324127906976744e-05,
|
313 |
+
"loss": 1.6129,
|
314 |
"step": 4500
|
315 |
},
|
316 |
{
|
317 |
+
"epoch": 21.7,
|
318 |
+
"learning_rate": 5.2369186046511624e-05,
|
319 |
+
"loss": 1.5853,
|
320 |
"step": 4600
|
321 |
},
|
322 |
{
|
323 |
+
"epoch": 22.17,
|
324 |
+
"learning_rate": 5.1497093023255804e-05,
|
325 |
+
"loss": 1.553,
|
326 |
"step": 4700
|
327 |
},
|
328 |
{
|
329 |
+
"epoch": 22.64,
|
330 |
+
"learning_rate": 5.0625e-05,
|
331 |
+
"loss": 1.5719,
|
332 |
"step": 4800
|
333 |
},
|
334 |
{
|
335 |
+
"epoch": 23.11,
|
336 |
+
"learning_rate": 4.9752906976744184e-05,
|
337 |
+
"loss": 1.5471,
|
338 |
"step": 4900
|
339 |
},
|
340 |
{
|
341 |
+
"epoch": 23.58,
|
342 |
+
"learning_rate": 4.888081395348837e-05,
|
343 |
+
"loss": 1.5468,
|
344 |
"step": 5000
|
345 |
},
|
346 |
{
|
347 |
+
"epoch": 23.58,
|
348 |
+
"eval_loss": 0.6087133884429932,
|
349 |
+
"eval_runtime": 236.0684,
|
350 |
+
"eval_samples_per_second": 18.918,
|
351 |
+
"eval_steps_per_second": 2.368,
|
352 |
+
"eval_wer": 1.2497202953680913,
|
353 |
"step": 5000
|
354 |
},
|
355 |
{
|
356 |
+
"epoch": 24.06,
|
357 |
+
"learning_rate": 4.800872093023255e-05,
|
358 |
+
"loss": 1.5584,
|
359 |
"step": 5100
|
360 |
},
|
361 |
{
|
362 |
+
"epoch": 24.53,
|
363 |
+
"learning_rate": 4.7136627906976736e-05,
|
364 |
+
"loss": 1.5359,
|
365 |
"step": 5200
|
366 |
},
|
367 |
{
|
368 |
+
"epoch": 25.0,
|
369 |
+
"learning_rate": 4.626453488372093e-05,
|
370 |
+
"loss": 1.5546,
|
371 |
"step": 5300
|
372 |
},
|
373 |
{
|
374 |
+
"epoch": 25.47,
|
375 |
+
"learning_rate": 4.5392441860465116e-05,
|
376 |
+
"loss": 1.5263,
|
377 |
"step": 5400
|
378 |
},
|
379 |
{
|
380 |
+
"epoch": 25.94,
|
381 |
+
"learning_rate": 4.45203488372093e-05,
|
382 |
+
"loss": 1.5168,
|
383 |
"step": 5500
|
384 |
},
|
385 |
{
|
386 |
+
"epoch": 26.42,
|
387 |
+
"learning_rate": 4.364825581395348e-05,
|
388 |
+
"loss": 1.489,
|
389 |
"step": 5600
|
390 |
},
|
391 |
{
|
392 |
+
"epoch": 26.89,
|
393 |
+
"learning_rate": 4.277616279069767e-05,
|
394 |
+
"loss": 1.5108,
|
395 |
"step": 5700
|
396 |
},
|
397 |
{
|
398 |
+
"epoch": 27.36,
|
399 |
+
"learning_rate": 4.1904069767441855e-05,
|
400 |
+
"loss": 1.4882,
|
401 |
"step": 5800
|
402 |
},
|
403 |
{
|
404 |
+
"epoch": 27.83,
|
405 |
+
"learning_rate": 4.103197674418605e-05,
|
406 |
+
"loss": 1.5027,
|
407 |
"step": 5900
|
408 |
},
|
409 |
{
|
410 |
+
"epoch": 28.3,
|
411 |
+
"learning_rate": 4.015988372093023e-05,
|
412 |
+
"loss": 1.4885,
|
413 |
"step": 6000
|
414 |
},
|
415 |
{
|
416 |
+
"epoch": 28.3,
|
417 |
+
"eval_loss": 0.5786022543907166,
|
418 |
+
"eval_runtime": 230.8558,
|
419 |
+
"eval_samples_per_second": 19.345,
|
420 |
+
"eval_steps_per_second": 2.421,
|
421 |
+
"eval_wer": 1.3222197359588275,
|
422 |
"step": 6000
|
423 |
},
|
424 |
{
|
425 |
+
"epoch": 28.77,
|
426 |
+
"learning_rate": 3.9287790697674414e-05,
|
427 |
+
"loss": 1.4804,
|
428 |
"step": 6100
|
429 |
},
|
430 |
{
|
431 |
+
"epoch": 29.25,
|
432 |
+
"learning_rate": 3.84156976744186e-05,
|
433 |
+
"loss": 1.5115,
|
434 |
"step": 6200
|
435 |
},
|
436 |
{
|
437 |
+
"epoch": 29.72,
|
438 |
+
"learning_rate": 3.754360465116279e-05,
|
439 |
+
"loss": 1.4561,
|
440 |
"step": 6300
|
441 |
},
|
442 |
{
|
443 |
+
"epoch": 30.19,
|
444 |
+
"learning_rate": 3.6671511627906973e-05,
|
445 |
+
"loss": 1.4786,
|
446 |
"step": 6400
|
447 |
},
|
448 |
{
|
449 |
+
"epoch": 30.66,
|
450 |
+
"learning_rate": 3.579941860465116e-05,
|
451 |
+
"loss": 1.4522,
|
452 |
"step": 6500
|
453 |
},
|
454 |
{
|
455 |
+
"epoch": 31.13,
|
456 |
+
"learning_rate": 3.4927325581395346e-05,
|
457 |
+
"loss": 1.4499,
|
458 |
"step": 6600
|
459 |
},
|
460 |
{
|
461 |
+
"epoch": 31.6,
|
462 |
+
"learning_rate": 3.406395348837209e-05,
|
463 |
+
"loss": 1.4398,
|
464 |
"step": 6700
|
465 |
},
|
466 |
{
|
467 |
+
"epoch": 32.08,
|
468 |
+
"learning_rate": 3.3191860465116275e-05,
|
469 |
+
"loss": 1.4364,
|
470 |
"step": 6800
|
471 |
},
|
472 |
{
|
473 |
+
"epoch": 32.55,
|
474 |
+
"learning_rate": 3.231976744186046e-05,
|
475 |
+
"loss": 1.4247,
|
476 |
"step": 6900
|
477 |
},
|
478 |
{
|
479 |
+
"epoch": 33.02,
|
480 |
+
"learning_rate": 3.144767441860465e-05,
|
481 |
+
"loss": 1.451,
|
482 |
"step": 7000
|
483 |
},
|
484 |
{
|
485 |
+
"epoch": 33.02,
|
486 |
+
"eval_loss": 0.5726000070571899,
|
487 |
+
"eval_runtime": 233.1518,
|
488 |
+
"eval_samples_per_second": 19.155,
|
489 |
+
"eval_steps_per_second": 2.398,
|
490 |
+
"eval_wer": 1.3768180801074066,
|
491 |
"step": 7000
|
492 |
},
|
493 |
{
|
494 |
+
"epoch": 33.49,
|
495 |
+
"learning_rate": 3.0575581395348834e-05,
|
496 |
+
"loss": 1.4259,
|
497 |
"step": 7100
|
498 |
},
|
499 |
{
|
500 |
+
"epoch": 33.96,
|
501 |
+
"learning_rate": 2.970348837209302e-05,
|
502 |
+
"loss": 1.4252,
|
503 |
"step": 7200
|
504 |
},
|
505 |
{
|
506 |
+
"epoch": 34.43,
|
507 |
+
"learning_rate": 2.8831395348837207e-05,
|
508 |
+
"loss": 1.4214,
|
509 |
"step": 7300
|
510 |
},
|
511 |
{
|
512 |
+
"epoch": 34.91,
|
513 |
+
"learning_rate": 2.795930232558139e-05,
|
514 |
+
"loss": 1.4222,
|
515 |
"step": 7400
|
516 |
},
|
517 |
{
|
518 |
+
"epoch": 35.38,
|
519 |
+
"learning_rate": 2.708720930232558e-05,
|
520 |
+
"loss": 1.3976,
|
521 |
"step": 7500
|
522 |
},
|
523 |
{
|
524 |
+
"epoch": 35.85,
|
525 |
+
"learning_rate": 2.6215116279069766e-05,
|
526 |
+
"loss": 1.4026,
|
527 |
"step": 7600
|
528 |
},
|
529 |
{
|
530 |
+
"epoch": 36.32,
|
531 |
+
"learning_rate": 2.5343023255813953e-05,
|
532 |
+
"loss": 1.4031,
|
533 |
"step": 7700
|
534 |
},
|
535 |
{
|
536 |
+
"epoch": 36.79,
|
537 |
+
"learning_rate": 2.447093023255814e-05,
|
538 |
+
"loss": 1.406,
|
539 |
"step": 7800
|
540 |
},
|
541 |
{
|
542 |
+
"epoch": 37.26,
|
543 |
+
"learning_rate": 2.3598837209302322e-05,
|
544 |
+
"loss": 1.4022,
|
545 |
"step": 7900
|
546 |
},
|
547 |
{
|
548 |
+
"epoch": 37.74,
|
549 |
+
"learning_rate": 2.2726744186046512e-05,
|
550 |
+
"loss": 1.3912,
|
551 |
"step": 8000
|
552 |
},
|
553 |
{
|
554 |
+
"epoch": 37.74,
|
555 |
+
"eval_loss": 0.551810622215271,
|
556 |
+
"eval_runtime": 230.6724,
|
557 |
+
"eval_samples_per_second": 19.361,
|
558 |
+
"eval_steps_per_second": 2.423,
|
559 |
+
"eval_wer": 1.2497202953680913,
|
560 |
"step": 8000
|
561 |
},
|
562 |
{
|
563 |
+
"epoch": 38.21,
|
564 |
+
"learning_rate": 2.1854651162790695e-05,
|
565 |
+
"loss": 1.4014,
|
566 |
"step": 8100
|
567 |
},
|
568 |
{
|
569 |
+
"epoch": 38.68,
|
570 |
+
"learning_rate": 2.098255813953488e-05,
|
571 |
+
"loss": 1.3701,
|
572 |
"step": 8200
|
573 |
},
|
574 |
{
|
575 |
+
"epoch": 39.15,
|
576 |
+
"learning_rate": 2.0110465116279068e-05,
|
577 |
+
"loss": 1.3704,
|
578 |
"step": 8300
|
579 |
},
|
580 |
{
|
581 |
+
"epoch": 39.62,
|
582 |
+
"learning_rate": 1.9238372093023254e-05,
|
583 |
+
"loss": 1.3746,
|
584 |
"step": 8400
|
585 |
},
|
586 |
{
|
587 |
+
"epoch": 40.09,
|
588 |
+
"learning_rate": 1.836627906976744e-05,
|
589 |
+
"loss": 1.3738,
|
590 |
"step": 8500
|
591 |
},
|
592 |
{
|
593 |
+
"epoch": 40.57,
|
594 |
+
"learning_rate": 1.7494186046511627e-05,
|
595 |
+
"loss": 1.3657,
|
596 |
"step": 8600
|
597 |
},
|
598 |
{
|
599 |
+
"epoch": 41.04,
|
600 |
+
"learning_rate": 1.6622093023255814e-05,
|
601 |
+
"loss": 1.3782,
|
602 |
"step": 8700
|
603 |
},
|
604 |
{
|
605 |
+
"epoch": 41.51,
|
606 |
+
"learning_rate": 1.5749999999999997e-05,
|
607 |
+
"loss": 1.3504,
|
608 |
"step": 8800
|
609 |
},
|
610 |
{
|
611 |
+
"epoch": 41.98,
|
612 |
+
"learning_rate": 1.4886627906976742e-05,
|
613 |
+
"loss": 1.3802,
|
614 |
"step": 8900
|
615 |
},
|
616 |
{
|
617 |
+
"epoch": 42.45,
|
618 |
+
"learning_rate": 1.401453488372093e-05,
|
619 |
+
"loss": 1.3617,
|
620 |
"step": 9000
|
621 |
},
|
622 |
{
|
623 |
+
"epoch": 42.45,
|
624 |
+
"eval_loss": 0.5352301597595215,
|
625 |
+
"eval_runtime": 231.9967,
|
626 |
+
"eval_samples_per_second": 19.25,
|
627 |
+
"eval_steps_per_second": 2.41,
|
628 |
+
"eval_wer": 1.269411501454464,
|
629 |
"step": 9000
|
630 |
},
|
631 |
{
|
632 |
+
"epoch": 42.92,
|
633 |
+
"learning_rate": 1.3142441860465116e-05,
|
634 |
+
"loss": 1.357,
|
635 |
"step": 9100
|
636 |
},
|
637 |
{
|
638 |
+
"epoch": 43.4,
|
639 |
+
"learning_rate": 1.2270348837209301e-05,
|
640 |
+
"loss": 1.332,
|
641 |
"step": 9200
|
642 |
},
|
643 |
{
|
644 |
+
"epoch": 43.87,
|
645 |
+
"learning_rate": 1.1398255813953488e-05,
|
646 |
+
"loss": 1.3447,
|
647 |
"step": 9300
|
648 |
},
|
649 |
{
|
650 |
+
"epoch": 44.34,
|
651 |
+
"learning_rate": 1.0526162790697674e-05,
|
652 |
+
"loss": 1.3431,
|
653 |
"step": 9400
|
654 |
},
|
655 |
{
|
656 |
+
"epoch": 44.81,
|
657 |
+
"learning_rate": 9.65406976744186e-06,
|
658 |
+
"loss": 1.3337,
|
659 |
"step": 9500
|
660 |
},
|
661 |
{
|
662 |
+
"epoch": 45.28,
|
663 |
+
"learning_rate": 8.781976744186047e-06,
|
664 |
+
"loss": 1.3505,
|
665 |
"step": 9600
|
666 |
},
|
667 |
{
|
668 |
+
"epoch": 45.75,
|
669 |
+
"learning_rate": 7.909883720930232e-06,
|
670 |
+
"loss": 1.3325,
|
671 |
"step": 9700
|
672 |
},
|
673 |
{
|
674 |
+
"epoch": 46.23,
|
675 |
+
"learning_rate": 7.037790697674418e-06,
|
676 |
+
"loss": 1.3522,
|
677 |
"step": 9800
|
678 |
},
|
679 |
{
|
680 |
+
"epoch": 46.7,
|
681 |
+
"learning_rate": 6.165697674418604e-06,
|
682 |
+
"loss": 1.3287,
|
683 |
"step": 9900
|
684 |
},
|
685 |
{
|
686 |
+
"epoch": 47.17,
|
687 |
+
"learning_rate": 5.293604651162791e-06,
|
688 |
+
"loss": 1.3113,
|
689 |
"step": 10000
|
690 |
},
|
691 |
{
|
692 |
+
"epoch": 47.17,
|
693 |
+
"eval_loss": 0.5227699279785156,
|
694 |
+
"eval_runtime": 235.0378,
|
695 |
+
"eval_samples_per_second": 19.001,
|
696 |
+
"eval_steps_per_second": 2.378,
|
697 |
+
"eval_wer": 1.2781382859700157,
|
698 |
"step": 10000
|
699 |
},
|
700 |
{
|
701 |
+
"epoch": 47.64,
|
702 |
+
"learning_rate": 4.4215116279069765e-06,
|
703 |
+
"loss": 1.3283,
|
704 |
"step": 10100
|
705 |
},
|
706 |
{
|
707 |
+
"epoch": 48.11,
|
708 |
+
"learning_rate": 3.5494186046511625e-06,
|
709 |
+
"loss": 1.3183,
|
710 |
"step": 10200
|
711 |
},
|
712 |
{
|
713 |
+
"epoch": 48.58,
|
714 |
+
"learning_rate": 2.6773255813953485e-06,
|
715 |
+
"loss": 1.3157,
|
716 |
"step": 10300
|
717 |
},
|
718 |
{
|
719 |
+
"epoch": 49.06,
|
720 |
+
"learning_rate": 1.8052325581395345e-06,
|
721 |
+
"loss": 1.3313,
|
722 |
"step": 10400
|
723 |
},
|
724 |
{
|
725 |
+
"epoch": 49.53,
|
726 |
+
"learning_rate": 9.331395348837209e-07,
|
727 |
+
"loss": 1.3258,
|
728 |
"step": 10500
|
729 |
},
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
730 |
{
|
731 |
"epoch": 50.0,
|
732 |
+
"learning_rate": 6.104651162790697e-08,
|
733 |
+
"loss": 1.3432,
|
734 |
+
"step": 10600
|
735 |
},
|
736 |
{
|
737 |
"epoch": 50.0,
|
738 |
+
"step": 10600,
|
739 |
+
"total_flos": 7.86273636478683e+19,
|
740 |
+
"train_loss": 2.5914736377068284,
|
741 |
+
"train_runtime": 29653.6134,
|
742 |
+
"train_samples_per_second": 17.118,
|
743 |
+
"train_steps_per_second": 0.357
|
744 |
}
|
745 |
],
|
746 |
+
"max_steps": 10600,
|
747 |
"num_train_epochs": 50,
|
748 |
+
"total_flos": 7.86273636478683e+19,
|
749 |
"trial_name": null,
|
750 |
"trial_params": null
|
751 |
}
|