End of training
Browse files- README.md +10 -9
- all_results.json +10 -10
- eval_results.json +6 -6
- train_results.json +5 -5
- trainer_state.json +285 -280
README.md
CHANGED
@@ -2,38 +2,39 @@
|
|
2 |
license: apache-2.0
|
3 |
base_model: openai/whisper-small
|
4 |
tags:
|
|
|
5 |
- generated_from_trainer
|
6 |
datasets:
|
7 |
-
- fleurs
|
8 |
metrics:
|
9 |
- wer
|
10 |
model-index:
|
11 |
-
- name:
|
12 |
results:
|
13 |
- task:
|
14 |
name: Automatic Speech Recognition
|
15 |
type: automatic-speech-recognition
|
16 |
dataset:
|
17 |
-
name: fleurs
|
18 |
-
type: fleurs
|
19 |
config: sn_zw
|
20 |
split: test
|
21 |
args: sn_zw
|
22 |
metrics:
|
23 |
- name: Wer
|
24 |
type: wer
|
25 |
-
value: 45.
|
26 |
---
|
27 |
|
28 |
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
|
29 |
should probably proofread and complete it, then remove this comment. -->
|
30 |
|
31 |
-
#
|
32 |
|
33 |
-
This model is a fine-tuned version of [openai/whisper-small](https://huggingface.co/openai/whisper-small) on the fleurs dataset.
|
34 |
It achieves the following results on the evaluation set:
|
35 |
-
- Loss: 0.
|
36 |
-
- Wer: 45.
|
37 |
|
38 |
## Model description
|
39 |
|
|
|
2 |
license: apache-2.0
|
3 |
base_model: openai/whisper-small
|
4 |
tags:
|
5 |
+
- whisper-event
|
6 |
- generated_from_trainer
|
7 |
datasets:
|
8 |
+
- google/fleurs
|
9 |
metrics:
|
10 |
- wer
|
11 |
model-index:
|
12 |
+
- name: Whisper small shona
|
13 |
results:
|
14 |
- task:
|
15 |
name: Automatic Speech Recognition
|
16 |
type: automatic-speech-recognition
|
17 |
dataset:
|
18 |
+
name: google/fleurs sn_zw
|
19 |
+
type: google/fleurs
|
20 |
config: sn_zw
|
21 |
split: test
|
22 |
args: sn_zw
|
23 |
metrics:
|
24 |
- name: Wer
|
25 |
type: wer
|
26 |
+
value: 45.9
|
27 |
---
|
28 |
|
29 |
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
|
30 |
should probably proofread and complete it, then remove this comment. -->
|
31 |
|
32 |
+
# Whisper small shona
|
33 |
|
34 |
+
This model is a fine-tuned version of [openai/whisper-small](https://huggingface.co/openai/whisper-small) on the google/fleurs sn_zw dataset.
|
35 |
It achieves the following results on the evaluation set:
|
36 |
+
- Loss: 0.9653
|
37 |
+
- Wer: 45.9
|
38 |
|
39 |
## Model description
|
40 |
|
all_results.json
CHANGED
@@ -1,12 +1,12 @@
|
|
1 |
{
|
2 |
-
"epoch":
|
3 |
-
"eval_loss":
|
4 |
-
"eval_runtime":
|
5 |
-
"eval_samples_per_second":
|
6 |
-
"eval_steps_per_second": 0.
|
7 |
-
"eval_wer":
|
8 |
-
"train_loss": 0.
|
9 |
-
"train_runtime":
|
10 |
-
"train_samples_per_second":
|
11 |
-
"train_steps_per_second": 0.
|
12 |
}
|
|
|
1 |
{
|
2 |
+
"epoch": 40.82,
|
3 |
+
"eval_loss": 0.965289831161499,
|
4 |
+
"eval_runtime": 148.1672,
|
5 |
+
"eval_samples_per_second": 6.243,
|
6 |
+
"eval_steps_per_second": 0.391,
|
7 |
+
"eval_wer": 45.9,
|
8 |
+
"train_loss": 0.18361923832818866,
|
9 |
+
"train_runtime": 4653.6924,
|
10 |
+
"train_samples_per_second": 6.876,
|
11 |
+
"train_steps_per_second": 0.43
|
12 |
}
|
eval_results.json
CHANGED
@@ -1,8 +1,8 @@
|
|
1 |
{
|
2 |
-
"epoch":
|
3 |
-
"eval_loss":
|
4 |
-
"eval_runtime":
|
5 |
-
"eval_samples_per_second":
|
6 |
-
"eval_steps_per_second": 0.
|
7 |
-
"eval_wer":
|
8 |
}
|
|
|
1 |
{
|
2 |
+
"epoch": 40.82,
|
3 |
+
"eval_loss": 0.965289831161499,
|
4 |
+
"eval_runtime": 148.1672,
|
5 |
+
"eval_samples_per_second": 6.243,
|
6 |
+
"eval_steps_per_second": 0.391,
|
7 |
+
"eval_wer": 45.9
|
8 |
}
|
train_results.json
CHANGED
@@ -1,7 +1,7 @@
|
|
1 |
{
|
2 |
-
"epoch":
|
3 |
-
"train_loss": 0.
|
4 |
-
"train_runtime":
|
5 |
-
"train_samples_per_second":
|
6 |
-
"train_steps_per_second": 0.
|
7 |
}
|
|
|
1 |
{
|
2 |
+
"epoch": 40.82,
|
3 |
+
"train_loss": 0.18361923832818866,
|
4 |
+
"train_runtime": 4653.6924,
|
5 |
+
"train_samples_per_second": 6.876,
|
6 |
+
"train_steps_per_second": 0.43
|
7 |
}
|
trainer_state.json
CHANGED
@@ -1,550 +1,555 @@
|
|
1 |
{
|
2 |
-
"best_metric":
|
3 |
-
"best_model_checkpoint": "./whisper-small-shona/checkpoint-
|
4 |
-
"epoch":
|
|
|
5 |
"global_step": 2000,
|
6 |
"is_hyper_param_search": false,
|
7 |
"is_local_process_zero": true,
|
8 |
"is_world_process_zero": true,
|
9 |
"log_history": [
|
10 |
{
|
11 |
-
"epoch":
|
12 |
"learning_rate": 4.800000000000001e-07,
|
13 |
-
"loss":
|
14 |
"step": 25
|
15 |
},
|
16 |
{
|
17 |
-
"epoch":
|
18 |
-
"learning_rate": 9.
|
19 |
-
"loss": 2.
|
20 |
"step": 50
|
21 |
},
|
22 |
{
|
23 |
-
"epoch":
|
24 |
-
"learning_rate": 1.
|
25 |
-
"loss": 1.
|
26 |
"step": 75
|
27 |
},
|
28 |
{
|
29 |
-
"epoch":
|
30 |
-
"learning_rate": 1.
|
31 |
-
"loss": 1.
|
32 |
"step": 100
|
33 |
},
|
34 |
{
|
35 |
-
"epoch":
|
36 |
-
"learning_rate": 2.
|
37 |
-
"loss":
|
38 |
"step": 125
|
39 |
},
|
40 |
{
|
41 |
-
"epoch":
|
42 |
-
"learning_rate": 2.
|
43 |
-
"loss": 0.
|
44 |
"step": 150
|
45 |
},
|
46 |
{
|
47 |
-
"epoch":
|
48 |
-
"learning_rate": 3.
|
49 |
-
"loss": 0.
|
50 |
"step": 175
|
51 |
},
|
52 |
{
|
53 |
-
"epoch":
|
54 |
-
"learning_rate": 3.
|
55 |
-
"loss": 0.
|
56 |
"step": 200
|
57 |
},
|
58 |
{
|
59 |
-
"epoch":
|
60 |
-
"learning_rate": 4.
|
61 |
-
"loss": 0.
|
62 |
"step": 225
|
63 |
},
|
64 |
{
|
65 |
-
"epoch":
|
66 |
-
"learning_rate": 4.
|
67 |
-
"loss": 0.
|
68 |
"step": 250
|
69 |
},
|
70 |
{
|
71 |
-
"epoch":
|
72 |
-
"learning_rate": 5.
|
73 |
-
"loss": 0.
|
74 |
"step": 275
|
75 |
},
|
76 |
{
|
77 |
-
"epoch":
|
78 |
-
"learning_rate": 5.
|
79 |
-
"loss": 0.
|
80 |
"step": 300
|
81 |
},
|
82 |
{
|
83 |
-
"epoch":
|
84 |
-
"learning_rate": 6.
|
85 |
-
"loss": 0.
|
86 |
"step": 325
|
87 |
},
|
88 |
{
|
89 |
-
"epoch":
|
90 |
-
"learning_rate": 6.
|
91 |
-
"loss": 0.
|
92 |
"step": 350
|
93 |
},
|
94 |
{
|
95 |
-
"epoch":
|
96 |
-
"learning_rate": 7.
|
97 |
-
"loss": 0.
|
98 |
"step": 375
|
99 |
},
|
100 |
{
|
101 |
-
"epoch":
|
102 |
-
"learning_rate": 7.
|
103 |
-
"loss": 0.
|
104 |
"step": 400
|
105 |
},
|
106 |
{
|
107 |
-
"epoch":
|
108 |
-
"eval_loss": 0.
|
109 |
-
"eval_runtime":
|
110 |
-
"eval_samples_per_second":
|
111 |
-
"eval_steps_per_second": 0.
|
112 |
-
"eval_wer":
|
113 |
"step": 400
|
114 |
},
|
115 |
{
|
116 |
-
"epoch":
|
117 |
-
"learning_rate": 8.
|
118 |
-
"loss": 0.
|
119 |
"step": 425
|
120 |
},
|
121 |
{
|
122 |
-
"epoch":
|
123 |
-
"learning_rate": 8.
|
124 |
-
"loss": 0.
|
125 |
"step": 450
|
126 |
},
|
127 |
{
|
128 |
-
"epoch":
|
129 |
-
"learning_rate": 9.
|
130 |
-
"loss": 0.
|
131 |
"step": 475
|
132 |
},
|
133 |
{
|
134 |
-
"epoch":
|
135 |
-
"learning_rate": 9.
|
136 |
-
"loss": 0.
|
137 |
"step": 500
|
138 |
},
|
139 |
{
|
140 |
-
"epoch":
|
141 |
-
"learning_rate": 9.
|
142 |
-
"loss": 0.
|
143 |
"step": 525
|
144 |
},
|
145 |
{
|
146 |
-
"epoch":
|
147 |
-
"learning_rate": 9.
|
148 |
-
"loss": 0.
|
149 |
"step": 550
|
150 |
},
|
151 |
{
|
152 |
-
"epoch":
|
153 |
-
"learning_rate": 9.
|
154 |
-
"loss": 0.
|
155 |
"step": 575
|
156 |
},
|
157 |
{
|
158 |
-
"epoch":
|
159 |
-
"learning_rate": 9.
|
160 |
-
"loss": 0.
|
161 |
"step": 600
|
162 |
},
|
163 |
{
|
164 |
-
"epoch":
|
165 |
-
"learning_rate": 9.
|
166 |
-
"loss": 0.
|
167 |
"step": 625
|
168 |
},
|
169 |
{
|
170 |
-
"epoch":
|
171 |
-
"learning_rate": 9.
|
172 |
-
"loss": 0.
|
173 |
"step": 650
|
174 |
},
|
175 |
{
|
176 |
-
"epoch":
|
177 |
-
"learning_rate": 8.
|
178 |
-
"loss": 0.
|
179 |
"step": 675
|
180 |
},
|
181 |
{
|
182 |
-
"epoch":
|
183 |
-
"learning_rate": 8.
|
184 |
-
"loss": 0.
|
185 |
"step": 700
|
186 |
},
|
187 |
{
|
188 |
-
"epoch":
|
189 |
-
"learning_rate": 8.
|
190 |
-
"loss": 0.
|
191 |
"step": 725
|
192 |
},
|
193 |
{
|
194 |
-
"epoch":
|
195 |
-
"learning_rate": 8.
|
196 |
-
"loss": 0.
|
197 |
"step": 750
|
198 |
},
|
199 |
{
|
200 |
-
"epoch":
|
201 |
-
"learning_rate": 8.
|
202 |
-
"loss": 0.
|
203 |
"step": 775
|
204 |
},
|
205 |
{
|
206 |
-
"epoch":
|
207 |
-
"learning_rate": 8.
|
208 |
-
"loss": 0.
|
209 |
"step": 800
|
210 |
},
|
211 |
{
|
212 |
-
"epoch":
|
213 |
-
"eval_loss":
|
214 |
-
"eval_runtime":
|
215 |
-
"eval_samples_per_second":
|
216 |
-
"eval_steps_per_second": 0.
|
217 |
-
"eval_wer":
|
218 |
"step": 800
|
219 |
},
|
220 |
{
|
221 |
-
"epoch":
|
222 |
-
"learning_rate": 7.
|
223 |
-
"loss": 0.
|
224 |
"step": 825
|
225 |
},
|
226 |
{
|
227 |
-
"epoch":
|
228 |
-
"learning_rate": 7.
|
229 |
-
"loss": 0.
|
230 |
"step": 850
|
231 |
},
|
232 |
{
|
233 |
-
"epoch":
|
234 |
-
"learning_rate": 7.
|
235 |
-
"loss": 0.
|
236 |
"step": 875
|
237 |
},
|
238 |
{
|
239 |
-
"epoch":
|
240 |
-
"learning_rate": 7.
|
241 |
-
"loss": 0.
|
242 |
"step": 900
|
243 |
},
|
244 |
{
|
245 |
-
"epoch":
|
246 |
-
"learning_rate": 7.
|
247 |
-
"loss": 0.
|
248 |
"step": 925
|
249 |
},
|
250 |
{
|
251 |
-
"epoch":
|
252 |
-
"learning_rate": 7.
|
253 |
-
"loss": 0.
|
254 |
"step": 950
|
255 |
},
|
256 |
{
|
257 |
-
"epoch":
|
258 |
-
"learning_rate": 6.
|
259 |
-
"loss": 0.
|
260 |
"step": 975
|
261 |
},
|
262 |
{
|
263 |
-
"epoch":
|
264 |
-
"learning_rate": 6.
|
265 |
-
"loss": 0.
|
266 |
"step": 1000
|
267 |
},
|
268 |
{
|
269 |
-
"epoch":
|
270 |
-
"learning_rate": 6.
|
271 |
-
"loss": 0.
|
272 |
"step": 1025
|
273 |
},
|
274 |
{
|
275 |
-
"epoch":
|
276 |
-
"learning_rate": 6.
|
277 |
-
"loss": 0.
|
278 |
"step": 1050
|
279 |
},
|
280 |
{
|
281 |
-
"epoch":
|
282 |
-
"learning_rate": 6.
|
283 |
-
"loss": 0.
|
284 |
"step": 1075
|
285 |
},
|
286 |
{
|
287 |
-
"epoch":
|
288 |
-
"learning_rate": 6.
|
289 |
-
"loss": 0.
|
290 |
"step": 1100
|
291 |
},
|
292 |
{
|
293 |
-
"epoch":
|
294 |
-
"learning_rate": 5.
|
295 |
-
"loss": 0.
|
296 |
"step": 1125
|
297 |
},
|
298 |
{
|
299 |
-
"epoch":
|
300 |
-
"learning_rate": 5.
|
301 |
-
"loss": 0.
|
302 |
"step": 1150
|
303 |
},
|
304 |
{
|
305 |
-
"epoch":
|
306 |
-
"learning_rate": 5.
|
307 |
-
"loss": 0.
|
308 |
"step": 1175
|
309 |
},
|
310 |
{
|
311 |
-
"epoch":
|
312 |
-
"learning_rate": 5.
|
313 |
-
"loss": 0.
|
314 |
"step": 1200
|
315 |
},
|
316 |
{
|
317 |
-
"epoch":
|
318 |
-
"eval_loss":
|
319 |
-
"eval_runtime":
|
320 |
-
"eval_samples_per_second":
|
321 |
-
"eval_steps_per_second": 0.
|
322 |
-
"eval_wer":
|
323 |
"step": 1200
|
324 |
},
|
325 |
{
|
326 |
-
"epoch":
|
327 |
-
"learning_rate": 5.
|
328 |
-
"loss": 0.
|
329 |
"step": 1225
|
330 |
},
|
331 |
{
|
332 |
-
"epoch":
|
333 |
-
"learning_rate": 5.
|
334 |
-
"loss": 0.
|
335 |
"step": 1250
|
336 |
},
|
337 |
{
|
338 |
-
"epoch":
|
339 |
-
"learning_rate": 4.
|
340 |
-
"loss": 0.
|
341 |
"step": 1275
|
342 |
},
|
343 |
{
|
344 |
-
"epoch":
|
345 |
-
"learning_rate": 4.
|
346 |
-
"loss": 0.
|
347 |
"step": 1300
|
348 |
},
|
349 |
{
|
350 |
-
"epoch":
|
351 |
-
"learning_rate": 4.
|
352 |
-
"loss": 0.
|
353 |
"step": 1325
|
354 |
},
|
355 |
{
|
356 |
-
"epoch":
|
357 |
-
"learning_rate": 4.
|
358 |
-
"loss": 0.
|
359 |
"step": 1350
|
360 |
},
|
361 |
{
|
362 |
-
"epoch":
|
363 |
-
"learning_rate": 4.
|
364 |
-
"loss": 0.
|
365 |
"step": 1375
|
366 |
},
|
367 |
{
|
368 |
-
"epoch":
|
369 |
-
"learning_rate": 4.
|
370 |
-
"loss": 0.
|
371 |
"step": 1400
|
372 |
},
|
373 |
{
|
374 |
-
"epoch":
|
375 |
-
"learning_rate": 3.
|
376 |
-
"loss": 0.
|
377 |
"step": 1425
|
378 |
},
|
379 |
{
|
380 |
-
"epoch":
|
381 |
-
"learning_rate": 3.
|
382 |
-
"loss": 0.
|
383 |
"step": 1450
|
384 |
},
|
385 |
{
|
386 |
-
"epoch":
|
387 |
-
"learning_rate": 3.
|
388 |
-
"loss": 0.
|
389 |
"step": 1475
|
390 |
},
|
391 |
{
|
392 |
-
"epoch":
|
393 |
-
"learning_rate": 3.
|
394 |
-
"loss": 0.
|
395 |
"step": 1500
|
396 |
},
|
397 |
{
|
398 |
-
"epoch":
|
399 |
-
"learning_rate": 3.
|
400 |
-
"loss": 0.
|
401 |
"step": 1525
|
402 |
},
|
403 |
{
|
404 |
-
"epoch":
|
405 |
-
"learning_rate": 3.
|
406 |
-
"loss": 0.
|
407 |
"step": 1550
|
408 |
},
|
409 |
{
|
410 |
-
"epoch":
|
411 |
-
"learning_rate": 2.
|
412 |
-
"loss": 0.
|
413 |
"step": 1575
|
414 |
},
|
415 |
{
|
416 |
-
"epoch":
|
417 |
-
"learning_rate": 2.
|
418 |
-
"loss": 0.
|
419 |
"step": 1600
|
420 |
},
|
421 |
{
|
422 |
-
"epoch":
|
423 |
-
"eval_loss":
|
424 |
-
"eval_runtime":
|
425 |
-
"eval_samples_per_second":
|
426 |
-
"eval_steps_per_second": 0.
|
427 |
-
"eval_wer":
|
428 |
"step": 1600
|
429 |
},
|
430 |
{
|
431 |
-
"epoch":
|
432 |
-
"learning_rate": 2.
|
433 |
-
"loss": 0.
|
434 |
"step": 1625
|
435 |
},
|
436 |
{
|
437 |
-
"epoch":
|
438 |
-
"learning_rate": 2.
|
439 |
-
"loss": 0.
|
440 |
"step": 1650
|
441 |
},
|
442 |
{
|
443 |
-
"epoch":
|
444 |
-
"learning_rate": 2.
|
445 |
-
"loss": 0.
|
446 |
"step": 1675
|
447 |
},
|
448 |
{
|
449 |
-
"epoch":
|
450 |
-
"learning_rate": 2.
|
451 |
-
"loss": 0.
|
452 |
"step": 1700
|
453 |
},
|
454 |
{
|
455 |
-
"epoch":
|
456 |
-
"learning_rate": 1.
|
457 |
-
"loss": 0.
|
458 |
"step": 1725
|
459 |
},
|
460 |
{
|
461 |
-
"epoch":
|
462 |
-
"learning_rate": 1.
|
463 |
-
"loss": 0.
|
464 |
"step": 1750
|
465 |
},
|
466 |
{
|
467 |
-
"epoch":
|
468 |
-
"learning_rate": 1.
|
469 |
-
"loss": 0.
|
470 |
"step": 1775
|
471 |
},
|
472 |
{
|
473 |
-
"epoch":
|
474 |
-
"learning_rate": 1.
|
475 |
-
"loss": 0.
|
476 |
"step": 1800
|
477 |
},
|
478 |
{
|
479 |
-
"epoch":
|
480 |
-
"learning_rate": 1.
|
481 |
-
"loss": 0.
|
482 |
"step": 1825
|
483 |
},
|
484 |
{
|
485 |
-
"epoch":
|
486 |
-
"learning_rate": 1.
|
487 |
-
"loss": 0.
|
488 |
"step": 1850
|
489 |
},
|
490 |
{
|
491 |
-
"epoch":
|
492 |
-
"learning_rate": 8.
|
493 |
-
"loss": 0.
|
494 |
"step": 1875
|
495 |
},
|
496 |
{
|
497 |
-
"epoch":
|
498 |
-
"learning_rate": 6.
|
499 |
-
"loss": 0.
|
500 |
"step": 1900
|
501 |
},
|
502 |
{
|
503 |
-
"epoch":
|
504 |
-
"learning_rate": 5.
|
505 |
-
"loss": 0.
|
506 |
"step": 1925
|
507 |
},
|
508 |
{
|
509 |
-
"epoch":
|
510 |
-
"learning_rate": 3.
|
511 |
-
"loss": 0.
|
512 |
"step": 1950
|
513 |
},
|
514 |
{
|
515 |
-
"epoch":
|
516 |
-
"learning_rate": 1.
|
517 |
-
"loss": 0.
|
518 |
"step": 1975
|
519 |
},
|
520 |
{
|
521 |
-
"epoch":
|
522 |
-
"learning_rate":
|
523 |
-
"loss": 0.
|
524 |
"step": 2000
|
525 |
},
|
526 |
{
|
527 |
-
"epoch":
|
528 |
-
"eval_loss":
|
529 |
-
"eval_runtime":
|
530 |
-
"eval_samples_per_second":
|
531 |
-
"eval_steps_per_second": 0.
|
532 |
-
"eval_wer":
|
533 |
"step": 2000
|
534 |
},
|
535 |
{
|
536 |
-
"epoch":
|
537 |
"step": 2000,
|
538 |
-
"total_flos":
|
539 |
-
"train_loss": 0.
|
540 |
-
"train_runtime":
|
541 |
-
"train_samples_per_second":
|
542 |
-
"train_steps_per_second": 0.
|
543 |
}
|
544 |
],
|
|
|
545 |
"max_steps": 2000,
|
546 |
-
"
|
547 |
-
"
|
|
|
|
|
|
|
548 |
"trial_name": null,
|
549 |
"trial_params": null
|
550 |
}
|
|
|
1 |
{
|
2 |
+
"best_metric": 45.9,
|
3 |
+
"best_model_checkpoint": "./whisper-small-shona/checkpoint-1600",
|
4 |
+
"epoch": 40.816326530612244,
|
5 |
+
"eval_steps": 400,
|
6 |
"global_step": 2000,
|
7 |
"is_hyper_param_search": false,
|
8 |
"is_local_process_zero": true,
|
9 |
"is_world_process_zero": true,
|
10 |
"log_history": [
|
11 |
{
|
12 |
+
"epoch": 0.51,
|
13 |
"learning_rate": 4.800000000000001e-07,
|
14 |
+
"loss": 2.8886,
|
15 |
"step": 25
|
16 |
},
|
17 |
{
|
18 |
+
"epoch": 1.02,
|
19 |
+
"learning_rate": 9.800000000000001e-07,
|
20 |
+
"loss": 2.4034,
|
21 |
"step": 50
|
22 |
},
|
23 |
{
|
24 |
+
"epoch": 1.53,
|
25 |
+
"learning_rate": 1.48e-06,
|
26 |
+
"loss": 1.8889,
|
27 |
"step": 75
|
28 |
},
|
29 |
{
|
30 |
+
"epoch": 2.04,
|
31 |
+
"learning_rate": 1.98e-06,
|
32 |
+
"loss": 1.4574,
|
33 |
"step": 100
|
34 |
},
|
35 |
{
|
36 |
+
"epoch": 2.55,
|
37 |
+
"learning_rate": 2.4800000000000004e-06,
|
38 |
+
"loss": 1.1526,
|
39 |
"step": 125
|
40 |
},
|
41 |
{
|
42 |
+
"epoch": 3.06,
|
43 |
+
"learning_rate": 2.9800000000000003e-06,
|
44 |
+
"loss": 0.9975,
|
45 |
"step": 150
|
46 |
},
|
47 |
{
|
48 |
+
"epoch": 3.57,
|
49 |
+
"learning_rate": 3.48e-06,
|
50 |
+
"loss": 0.797,
|
51 |
"step": 175
|
52 |
},
|
53 |
{
|
54 |
+
"epoch": 4.08,
|
55 |
+
"learning_rate": 3.980000000000001e-06,
|
56 |
+
"loss": 0.6751,
|
57 |
"step": 200
|
58 |
},
|
59 |
{
|
60 |
+
"epoch": 4.59,
|
61 |
+
"learning_rate": 4.48e-06,
|
62 |
+
"loss": 0.5246,
|
63 |
"step": 225
|
64 |
},
|
65 |
{
|
66 |
+
"epoch": 5.1,
|
67 |
+
"learning_rate": 4.980000000000001e-06,
|
68 |
+
"loss": 0.4666,
|
69 |
"step": 250
|
70 |
},
|
71 |
{
|
72 |
+
"epoch": 5.61,
|
73 |
+
"learning_rate": 5.480000000000001e-06,
|
74 |
+
"loss": 0.3392,
|
75 |
"step": 275
|
76 |
},
|
77 |
{
|
78 |
+
"epoch": 6.12,
|
79 |
+
"learning_rate": 5.98e-06,
|
80 |
+
"loss": 0.2899,
|
81 |
"step": 300
|
82 |
},
|
83 |
{
|
84 |
+
"epoch": 6.63,
|
85 |
+
"learning_rate": 6.480000000000001e-06,
|
86 |
+
"loss": 0.1993,
|
87 |
"step": 325
|
88 |
},
|
89 |
{
|
90 |
+
"epoch": 7.14,
|
91 |
+
"learning_rate": 6.98e-06,
|
92 |
+
"loss": 0.1636,
|
93 |
"step": 350
|
94 |
},
|
95 |
{
|
96 |
+
"epoch": 7.65,
|
97 |
+
"learning_rate": 7.48e-06,
|
98 |
+
"loss": 0.1039,
|
99 |
"step": 375
|
100 |
},
|
101 |
{
|
102 |
+
"epoch": 8.16,
|
103 |
+
"learning_rate": 7.980000000000002e-06,
|
104 |
+
"loss": 0.09,
|
105 |
"step": 400
|
106 |
},
|
107 |
{
|
108 |
+
"epoch": 8.16,
|
109 |
+
"eval_loss": 0.857236921787262,
|
110 |
+
"eval_runtime": 149.0293,
|
111 |
+
"eval_samples_per_second": 6.207,
|
112 |
+
"eval_steps_per_second": 0.389,
|
113 |
+
"eval_wer": 50.6875,
|
114 |
"step": 400
|
115 |
},
|
116 |
{
|
117 |
+
"epoch": 8.67,
|
118 |
+
"learning_rate": 8.48e-06,
|
119 |
+
"loss": 0.0536,
|
120 |
"step": 425
|
121 |
},
|
122 |
{
|
123 |
+
"epoch": 9.18,
|
124 |
+
"learning_rate": 8.98e-06,
|
125 |
+
"loss": 0.0403,
|
126 |
"step": 450
|
127 |
},
|
128 |
{
|
129 |
+
"epoch": 9.69,
|
130 |
+
"learning_rate": 9.48e-06,
|
131 |
+
"loss": 0.0278,
|
132 |
"step": 475
|
133 |
},
|
134 |
{
|
135 |
+
"epoch": 10.2,
|
136 |
+
"learning_rate": 9.980000000000001e-06,
|
137 |
+
"loss": 0.0204,
|
138 |
"step": 500
|
139 |
},
|
140 |
{
|
141 |
+
"epoch": 10.71,
|
142 |
+
"learning_rate": 9.84e-06,
|
143 |
+
"loss": 0.0131,
|
144 |
"step": 525
|
145 |
},
|
146 |
{
|
147 |
+
"epoch": 11.22,
|
148 |
+
"learning_rate": 9.673333333333334e-06,
|
149 |
+
"loss": 0.0117,
|
150 |
"step": 550
|
151 |
},
|
152 |
{
|
153 |
+
"epoch": 11.73,
|
154 |
+
"learning_rate": 9.506666666666667e-06,
|
155 |
+
"loss": 0.0083,
|
156 |
"step": 575
|
157 |
},
|
158 |
{
|
159 |
+
"epoch": 12.24,
|
160 |
+
"learning_rate": 9.340000000000002e-06,
|
161 |
+
"loss": 0.0069,
|
162 |
"step": 600
|
163 |
},
|
164 |
{
|
165 |
+
"epoch": 12.76,
|
166 |
+
"learning_rate": 9.173333333333334e-06,
|
167 |
+
"loss": 0.0072,
|
168 |
"step": 625
|
169 |
},
|
170 |
{
|
171 |
+
"epoch": 13.27,
|
172 |
+
"learning_rate": 9.006666666666666e-06,
|
173 |
+
"loss": 0.006,
|
174 |
"step": 650
|
175 |
},
|
176 |
{
|
177 |
+
"epoch": 13.78,
|
178 |
+
"learning_rate": 8.84e-06,
|
179 |
+
"loss": 0.0047,
|
180 |
"step": 675
|
181 |
},
|
182 |
{
|
183 |
+
"epoch": 14.29,
|
184 |
+
"learning_rate": 8.673333333333334e-06,
|
185 |
+
"loss": 0.0034,
|
186 |
"step": 700
|
187 |
},
|
188 |
{
|
189 |
+
"epoch": 14.8,
|
190 |
+
"learning_rate": 8.506666666666668e-06,
|
191 |
+
"loss": 0.0029,
|
192 |
"step": 725
|
193 |
},
|
194 |
{
|
195 |
+
"epoch": 15.31,
|
196 |
+
"learning_rate": 8.34e-06,
|
197 |
+
"loss": 0.003,
|
198 |
"step": 750
|
199 |
},
|
200 |
{
|
201 |
+
"epoch": 15.82,
|
202 |
+
"learning_rate": 8.173333333333334e-06,
|
203 |
+
"loss": 0.0021,
|
204 |
"step": 775
|
205 |
},
|
206 |
{
|
207 |
+
"epoch": 16.33,
|
208 |
+
"learning_rate": 8.006666666666667e-06,
|
209 |
+
"loss": 0.0019,
|
210 |
"step": 800
|
211 |
},
|
212 |
{
|
213 |
+
"epoch": 16.33,
|
214 |
+
"eval_loss": 0.9069271087646484,
|
215 |
+
"eval_runtime": 147.3435,
|
216 |
+
"eval_samples_per_second": 6.278,
|
217 |
+
"eval_steps_per_second": 0.394,
|
218 |
+
"eval_wer": 46.44375,
|
219 |
"step": 800
|
220 |
},
|
221 |
{
|
222 |
+
"epoch": 16.84,
|
223 |
+
"learning_rate": 7.840000000000001e-06,
|
224 |
+
"loss": 0.0019,
|
225 |
"step": 825
|
226 |
},
|
227 |
{
|
228 |
+
"epoch": 17.35,
|
229 |
+
"learning_rate": 7.673333333333333e-06,
|
230 |
+
"loss": 0.0019,
|
231 |
"step": 850
|
232 |
},
|
233 |
{
|
234 |
+
"epoch": 17.86,
|
235 |
+
"learning_rate": 7.506666666666668e-06,
|
236 |
+
"loss": 0.0016,
|
237 |
"step": 875
|
238 |
},
|
239 |
{
|
240 |
+
"epoch": 18.37,
|
241 |
+
"learning_rate": 7.340000000000001e-06,
|
242 |
+
"loss": 0.0018,
|
243 |
"step": 900
|
244 |
},
|
245 |
{
|
246 |
+
"epoch": 18.88,
|
247 |
+
"learning_rate": 7.173333333333335e-06,
|
248 |
+
"loss": 0.0014,
|
249 |
"step": 925
|
250 |
},
|
251 |
{
|
252 |
+
"epoch": 19.39,
|
253 |
+
"learning_rate": 7.006666666666667e-06,
|
254 |
+
"loss": 0.0012,
|
255 |
"step": 950
|
256 |
},
|
257 |
{
|
258 |
+
"epoch": 19.9,
|
259 |
+
"learning_rate": 6.8400000000000014e-06,
|
260 |
+
"loss": 0.0013,
|
261 |
"step": 975
|
262 |
},
|
263 |
{
|
264 |
+
"epoch": 20.41,
|
265 |
+
"learning_rate": 6.6733333333333335e-06,
|
266 |
+
"loss": 0.0012,
|
267 |
"step": 1000
|
268 |
},
|
269 |
{
|
270 |
+
"epoch": 20.92,
|
271 |
+
"learning_rate": 6.5066666666666665e-06,
|
272 |
+
"loss": 0.0011,
|
273 |
"step": 1025
|
274 |
},
|
275 |
{
|
276 |
+
"epoch": 21.43,
|
277 |
+
"learning_rate": 6.34e-06,
|
278 |
+
"loss": 0.001,
|
279 |
"step": 1050
|
280 |
},
|
281 |
{
|
282 |
+
"epoch": 21.94,
|
283 |
+
"learning_rate": 6.173333333333333e-06,
|
284 |
+
"loss": 0.001,
|
285 |
"step": 1075
|
286 |
},
|
287 |
{
|
288 |
+
"epoch": 22.45,
|
289 |
+
"learning_rate": 6.006666666666667e-06,
|
290 |
+
"loss": 0.0009,
|
291 |
"step": 1100
|
292 |
},
|
293 |
{
|
294 |
+
"epoch": 22.96,
|
295 |
+
"learning_rate": 5.84e-06,
|
296 |
+
"loss": 0.0009,
|
297 |
"step": 1125
|
298 |
},
|
299 |
{
|
300 |
+
"epoch": 23.47,
|
301 |
+
"learning_rate": 5.673333333333334e-06,
|
302 |
+
"loss": 0.0009,
|
303 |
"step": 1150
|
304 |
},
|
305 |
{
|
306 |
+
"epoch": 23.98,
|
307 |
+
"learning_rate": 5.506666666666667e-06,
|
308 |
+
"loss": 0.0008,
|
309 |
"step": 1175
|
310 |
},
|
311 |
{
|
312 |
+
"epoch": 24.49,
|
313 |
+
"learning_rate": 5.3400000000000005e-06,
|
314 |
+
"loss": 0.0008,
|
315 |
"step": 1200
|
316 |
},
|
317 |
{
|
318 |
+
"epoch": 24.49,
|
319 |
+
"eval_loss": 0.9473171830177307,
|
320 |
+
"eval_runtime": 148.8306,
|
321 |
+
"eval_samples_per_second": 6.215,
|
322 |
+
"eval_steps_per_second": 0.39,
|
323 |
+
"eval_wer": 46.04375,
|
324 |
"step": 1200
|
325 |
},
|
326 |
{
|
327 |
+
"epoch": 25.0,
|
328 |
+
"learning_rate": 5.1733333333333335e-06,
|
329 |
+
"loss": 0.0008,
|
330 |
"step": 1225
|
331 |
},
|
332 |
{
|
333 |
+
"epoch": 25.51,
|
334 |
+
"learning_rate": 5.006666666666667e-06,
|
335 |
+
"loss": 0.0008,
|
336 |
"step": 1250
|
337 |
},
|
338 |
{
|
339 |
+
"epoch": 26.02,
|
340 |
+
"learning_rate": 4.84e-06,
|
341 |
+
"loss": 0.0007,
|
342 |
"step": 1275
|
343 |
},
|
344 |
{
|
345 |
+
"epoch": 26.53,
|
346 |
+
"learning_rate": 4.673333333333333e-06,
|
347 |
+
"loss": 0.0007,
|
348 |
"step": 1300
|
349 |
},
|
350 |
{
|
351 |
+
"epoch": 27.04,
|
352 |
+
"learning_rate": 4.506666666666667e-06,
|
353 |
+
"loss": 0.0007,
|
354 |
"step": 1325
|
355 |
},
|
356 |
{
|
357 |
+
"epoch": 27.55,
|
358 |
+
"learning_rate": 4.34e-06,
|
359 |
+
"loss": 0.0007,
|
360 |
"step": 1350
|
361 |
},
|
362 |
{
|
363 |
+
"epoch": 28.06,
|
364 |
+
"learning_rate": 4.173333333333334e-06,
|
365 |
+
"loss": 0.0007,
|
366 |
"step": 1375
|
367 |
},
|
368 |
{
|
369 |
+
"epoch": 28.57,
|
370 |
+
"learning_rate": 4.006666666666667e-06,
|
371 |
+
"loss": 0.0006,
|
372 |
"step": 1400
|
373 |
},
|
374 |
{
|
375 |
+
"epoch": 29.08,
|
376 |
+
"learning_rate": 3.8400000000000005e-06,
|
377 |
+
"loss": 0.0007,
|
378 |
"step": 1425
|
379 |
},
|
380 |
{
|
381 |
+
"epoch": 29.59,
|
382 |
+
"learning_rate": 3.673333333333334e-06,
|
383 |
+
"loss": 0.0006,
|
384 |
"step": 1450
|
385 |
},
|
386 |
{
|
387 |
+
"epoch": 30.1,
|
388 |
+
"learning_rate": 3.5066666666666673e-06,
|
389 |
+
"loss": 0.0006,
|
390 |
"step": 1475
|
391 |
},
|
392 |
{
|
393 |
+
"epoch": 30.61,
|
394 |
+
"learning_rate": 3.3400000000000006e-06,
|
395 |
+
"loss": 0.0006,
|
396 |
"step": 1500
|
397 |
},
|
398 |
{
|
399 |
+
"epoch": 31.12,
|
400 |
+
"learning_rate": 3.173333333333334e-06,
|
401 |
+
"loss": 0.0006,
|
402 |
"step": 1525
|
403 |
},
|
404 |
{
|
405 |
+
"epoch": 31.63,
|
406 |
+
"learning_rate": 3.0066666666666674e-06,
|
407 |
+
"loss": 0.0006,
|
408 |
"step": 1550
|
409 |
},
|
410 |
{
|
411 |
+
"epoch": 32.14,
|
412 |
+
"learning_rate": 2.84e-06,
|
413 |
+
"loss": 0.0006,
|
414 |
"step": 1575
|
415 |
},
|
416 |
{
|
417 |
+
"epoch": 32.65,
|
418 |
+
"learning_rate": 2.6733333333333333e-06,
|
419 |
+
"loss": 0.0006,
|
420 |
"step": 1600
|
421 |
},
|
422 |
{
|
423 |
+
"epoch": 32.65,
|
424 |
+
"eval_loss": 0.965289831161499,
|
425 |
+
"eval_runtime": 147.3092,
|
426 |
+
"eval_samples_per_second": 6.279,
|
427 |
+
"eval_steps_per_second": 0.394,
|
428 |
+
"eval_wer": 45.9,
|
429 |
"step": 1600
|
430 |
},
|
431 |
{
|
432 |
+
"epoch": 33.16,
|
433 |
+
"learning_rate": 2.5066666666666667e-06,
|
434 |
+
"loss": 0.0006,
|
435 |
"step": 1625
|
436 |
},
|
437 |
{
|
438 |
+
"epoch": 33.67,
|
439 |
+
"learning_rate": 2.3400000000000005e-06,
|
440 |
+
"loss": 0.0005,
|
441 |
"step": 1650
|
442 |
},
|
443 |
{
|
444 |
+
"epoch": 34.18,
|
445 |
+
"learning_rate": 2.1733333333333334e-06,
|
446 |
+
"loss": 0.0006,
|
447 |
"step": 1675
|
448 |
},
|
449 |
{
|
450 |
+
"epoch": 34.69,
|
451 |
+
"learning_rate": 2.006666666666667e-06,
|
452 |
+
"loss": 0.0005,
|
453 |
"step": 1700
|
454 |
},
|
455 |
{
|
456 |
+
"epoch": 35.2,
|
457 |
+
"learning_rate": 1.8400000000000002e-06,
|
458 |
+
"loss": 0.0006,
|
459 |
"step": 1725
|
460 |
},
|
461 |
{
|
462 |
+
"epoch": 35.71,
|
463 |
+
"learning_rate": 1.6733333333333335e-06,
|
464 |
+
"loss": 0.0005,
|
465 |
"step": 1750
|
466 |
},
|
467 |
{
|
468 |
+
"epoch": 36.22,
|
469 |
+
"learning_rate": 1.506666666666667e-06,
|
470 |
+
"loss": 0.0005,
|
471 |
"step": 1775
|
472 |
},
|
473 |
{
|
474 |
+
"epoch": 36.73,
|
475 |
+
"learning_rate": 1.34e-06,
|
476 |
+
"loss": 0.0005,
|
477 |
"step": 1800
|
478 |
},
|
479 |
{
|
480 |
+
"epoch": 37.24,
|
481 |
+
"learning_rate": 1.1733333333333335e-06,
|
482 |
+
"loss": 0.0005,
|
483 |
"step": 1825
|
484 |
},
|
485 |
{
|
486 |
+
"epoch": 37.76,
|
487 |
+
"learning_rate": 1.0066666666666668e-06,
|
488 |
+
"loss": 0.0005,
|
489 |
"step": 1850
|
490 |
},
|
491 |
{
|
492 |
+
"epoch": 38.27,
|
493 |
+
"learning_rate": 8.400000000000001e-07,
|
494 |
+
"loss": 0.0005,
|
495 |
"step": 1875
|
496 |
},
|
497 |
{
|
498 |
+
"epoch": 38.78,
|
499 |
+
"learning_rate": 6.733333333333334e-07,
|
500 |
+
"loss": 0.0005,
|
501 |
"step": 1900
|
502 |
},
|
503 |
{
|
504 |
+
"epoch": 39.29,
|
505 |
+
"learning_rate": 5.066666666666667e-07,
|
506 |
+
"loss": 0.0005,
|
507 |
"step": 1925
|
508 |
},
|
509 |
{
|
510 |
+
"epoch": 39.8,
|
511 |
+
"learning_rate": 3.4000000000000003e-07,
|
512 |
+
"loss": 0.0005,
|
513 |
"step": 1950
|
514 |
},
|
515 |
{
|
516 |
+
"epoch": 40.31,
|
517 |
+
"learning_rate": 1.7333333333333335e-07,
|
518 |
+
"loss": 0.0005,
|
519 |
"step": 1975
|
520 |
},
|
521 |
{
|
522 |
+
"epoch": 40.82,
|
523 |
+
"learning_rate": 6.666666666666667e-09,
|
524 |
+
"loss": 0.0005,
|
525 |
"step": 2000
|
526 |
},
|
527 |
{
|
528 |
+
"epoch": 40.82,
|
529 |
+
"eval_loss": 0.9719672799110413,
|
530 |
+
"eval_runtime": 147.2827,
|
531 |
+
"eval_samples_per_second": 6.28,
|
532 |
+
"eval_steps_per_second": 0.394,
|
533 |
+
"eval_wer": 45.94375,
|
534 |
"step": 2000
|
535 |
},
|
536 |
{
|
537 |
+
"epoch": 40.82,
|
538 |
"step": 2000,
|
539 |
+
"total_flos": 9.188559230976655e+18,
|
540 |
+
"train_loss": 0.18361923832818866,
|
541 |
+
"train_runtime": 4653.6924,
|
542 |
+
"train_samples_per_second": 6.876,
|
543 |
+
"train_steps_per_second": 0.43
|
544 |
}
|
545 |
],
|
546 |
+
"logging_steps": 25,
|
547 |
"max_steps": 2000,
|
548 |
+
"num_input_tokens_seen": 0,
|
549 |
+
"num_train_epochs": 41,
|
550 |
+
"save_steps": 400,
|
551 |
+
"total_flos": 9.188559230976655e+18,
|
552 |
+
"train_batch_size": 8,
|
553 |
"trial_name": null,
|
554 |
"trial_params": null
|
555 |
}
|