Commit
•
93d2d61
1
Parent(s):
150d894
End of training
Browse files- README.md +3 -1
- all_results.json +12 -12
- eval_results.json +6 -6
- runs/May29_11-47-47_5cce808c63ba/events.out.tfevents.1717001109.5cce808c63ba.1059.1 +3 -0
- train_results.json +7 -7
- trainer_state.json +350 -870
README.md
CHANGED
@@ -2,6 +2,8 @@
|
|
2 |
license: apache-2.0
|
3 |
base_model: facebook/wav2vec2-large-xlsr-53
|
4 |
tags:
|
|
|
|
|
5 |
- generated_from_trainer
|
6 |
metrics:
|
7 |
- wer
|
@@ -15,7 +17,7 @@ should probably proofread and complete it, then remove this comment. -->
|
|
15 |
|
16 |
# wav2vec2-xlsr-53-ft-btb-ccv-cy
|
17 |
|
18 |
-
This model is a fine-tuned version of [facebook/wav2vec2-large-xlsr-53](https://huggingface.co/facebook/wav2vec2-large-xlsr-53) on
|
19 |
It achieves the following results on the evaluation set:
|
20 |
- Loss: 1.6956
|
21 |
- Wer: 0.7702
|
|
|
2 |
license: apache-2.0
|
3 |
base_model: facebook/wav2vec2-large-xlsr-53
|
4 |
tags:
|
5 |
+
- automatic-speech-recognition
|
6 |
+
- DewiBrynJones/banc-trawsgrifiadau-bangor-clean-with-ccv
|
7 |
- generated_from_trainer
|
8 |
metrics:
|
9 |
- wer
|
|
|
17 |
|
18 |
# wav2vec2-xlsr-53-ft-btb-ccv-cy
|
19 |
|
20 |
+
This model is a fine-tuned version of [facebook/wav2vec2-large-xlsr-53](https://huggingface.co/facebook/wav2vec2-large-xlsr-53) on the DEWIBRYNJONES/BANC-TRAWSGRIFIADAU-BANGOR-CLEAN-WITH-CCV - DEFAULT dataset.
|
21 |
It achieves the following results on the evaluation set:
|
22 |
- Loss: 1.6956
|
23 |
- Wer: 0.7702
|
all_results.json
CHANGED
@@ -1,15 +1,15 @@
|
|
1 |
{
|
2 |
-
"epoch":
|
3 |
-
"eval_loss":
|
4 |
-
"eval_runtime":
|
5 |
"eval_samples": 5656,
|
6 |
-
"eval_samples_per_second":
|
7 |
-
"eval_steps_per_second": 4.
|
8 |
-
"eval_wer": 0.
|
9 |
-
"total_flos":
|
10 |
-
"train_loss": 1.
|
11 |
-
"train_runtime":
|
12 |
-
"train_samples":
|
13 |
-
"train_samples_per_second":
|
14 |
-
"train_steps_per_second": 0.
|
15 |
}
|
|
|
1 |
{
|
2 |
+
"epoch": 7.739938080495356,
|
3 |
+
"eval_loss": 1.6955511569976807,
|
4 |
+
"eval_runtime": 160.3719,
|
5 |
"eval_samples": 5656,
|
6 |
+
"eval_samples_per_second": 35.268,
|
7 |
+
"eval_steps_per_second": 4.409,
|
8 |
+
"eval_wer": 0.7701529424981143,
|
9 |
+
"total_flos": 4.528229347801944e+19,
|
10 |
+
"train_loss": 1.3220068420410156,
|
11 |
+
"train_runtime": 20543.6538,
|
12 |
+
"train_samples": 41326,
|
13 |
+
"train_samples_per_second": 15.577,
|
14 |
+
"train_steps_per_second": 0.243
|
15 |
}
|
eval_results.json
CHANGED
@@ -1,9 +1,9 @@
|
|
1 |
{
|
2 |
-
"epoch":
|
3 |
-
"eval_loss":
|
4 |
-
"eval_runtime":
|
5 |
"eval_samples": 5656,
|
6 |
-
"eval_samples_per_second":
|
7 |
-
"eval_steps_per_second": 4.
|
8 |
-
"eval_wer": 0.
|
9 |
}
|
|
|
1 |
{
|
2 |
+
"epoch": 7.739938080495356,
|
3 |
+
"eval_loss": 1.6955511569976807,
|
4 |
+
"eval_runtime": 160.3719,
|
5 |
"eval_samples": 5656,
|
6 |
+
"eval_samples_per_second": 35.268,
|
7 |
+
"eval_steps_per_second": 4.409,
|
8 |
+
"eval_wer": 0.7701529424981143
|
9 |
}
|
runs/May29_11-47-47_5cce808c63ba/events.out.tfevents.1717001109.5cce808c63ba.1059.1
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:d66927dc0537cb97a59820dae65931aaf5cd40dc5586fdb88d230d7538c445ab
|
3 |
+
size 406
|
train_results.json
CHANGED
@@ -1,9 +1,9 @@
|
|
1 |
{
|
2 |
-
"epoch":
|
3 |
-
"total_flos":
|
4 |
-
"train_loss": 1.
|
5 |
-
"train_runtime":
|
6 |
-
"train_samples":
|
7 |
-
"train_samples_per_second":
|
8 |
-
"train_steps_per_second": 0.
|
9 |
}
|
|
|
1 |
{
|
2 |
+
"epoch": 7.739938080495356,
|
3 |
+
"total_flos": 4.528229347801944e+19,
|
4 |
+
"train_loss": 1.3220068420410156,
|
5 |
+
"train_runtime": 20543.6538,
|
6 |
+
"train_samples": 41326,
|
7 |
+
"train_samples_per_second": 15.577,
|
8 |
+
"train_steps_per_second": 0.243
|
9 |
}
|
trainer_state.json
CHANGED
@@ -1,1069 +1,549 @@
|
|
1 |
{
|
2 |
"best_metric": null,
|
3 |
"best_model_checkpoint": null,
|
4 |
-
"epoch":
|
5 |
"eval_steps": 100,
|
6 |
-
"global_step":
|
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.
|
13 |
-
"eval_loss":
|
14 |
-
"eval_runtime":
|
15 |
-
"eval_samples_per_second":
|
16 |
-
"eval_steps_per_second": 4.
|
17 |
"eval_wer": 1.0,
|
18 |
"step": 100
|
19 |
},
|
20 |
{
|
21 |
-
"epoch": 0.
|
22 |
-
"eval_loss": 3.
|
23 |
-
"eval_runtime":
|
24 |
-
"eval_samples_per_second":
|
25 |
-
"eval_steps_per_second": 4.
|
26 |
"eval_wer": 1.0,
|
27 |
"step": 200
|
28 |
},
|
29 |
{
|
30 |
-
"epoch": 0.
|
31 |
-
"eval_loss":
|
32 |
-
"eval_runtime":
|
33 |
-
"eval_samples_per_second":
|
34 |
-
"eval_steps_per_second": 4.
|
35 |
-
"eval_wer":
|
36 |
"step": 300
|
37 |
},
|
38 |
{
|
39 |
-
"epoch": 0.
|
40 |
-
"eval_loss":
|
41 |
-
"eval_runtime":
|
42 |
-
"eval_samples_per_second":
|
43 |
-
"eval_steps_per_second": 4.
|
44 |
-
"eval_wer": 0.
|
45 |
"step": 400
|
46 |
},
|
47 |
{
|
48 |
-
"epoch": 0.
|
49 |
-
"grad_norm":
|
50 |
-
"learning_rate": 0.
|
51 |
-
"loss":
|
52 |
"step": 500
|
53 |
},
|
54 |
{
|
55 |
-
"epoch": 0.
|
56 |
-
"eval_loss":
|
57 |
-
"eval_runtime":
|
58 |
-
"eval_samples_per_second":
|
59 |
-
"eval_steps_per_second": 4.
|
60 |
-
"eval_wer": 0.
|
61 |
"step": 500
|
62 |
},
|
63 |
{
|
64 |
-
"epoch": 0.
|
65 |
-
"eval_loss": 0.
|
66 |
-
"eval_runtime":
|
67 |
-
"eval_samples_per_second":
|
68 |
-
"eval_steps_per_second": 4.
|
69 |
-
"eval_wer": 0.
|
70 |
"step": 600
|
71 |
},
|
72 |
{
|
73 |
-
"epoch":
|
74 |
-
"eval_loss": 0.
|
75 |
-
"eval_runtime":
|
76 |
-
"eval_samples_per_second":
|
77 |
-
"eval_steps_per_second": 4.
|
78 |
-
"eval_wer": 0.
|
79 |
"step": 700
|
80 |
},
|
81 |
{
|
82 |
-
"epoch":
|
83 |
-
"eval_loss": 0.
|
84 |
-
"eval_runtime":
|
85 |
-
"eval_samples_per_second":
|
86 |
-
"eval_steps_per_second": 4.
|
87 |
-
"eval_wer": 0.
|
88 |
"step": 800
|
89 |
},
|
90 |
{
|
91 |
-
"epoch":
|
92 |
-
"eval_loss": 0.
|
93 |
-
"eval_runtime":
|
94 |
-
"eval_samples_per_second":
|
95 |
-
"eval_steps_per_second": 4.
|
96 |
-
"eval_wer": 0.
|
97 |
"step": 900
|
98 |
},
|
99 |
{
|
100 |
-
"epoch":
|
101 |
-
"grad_norm":
|
102 |
-
"learning_rate": 0.
|
103 |
-
"loss": 0.
|
104 |
"step": 1000
|
105 |
},
|
106 |
{
|
107 |
-
"epoch":
|
108 |
-
"eval_loss": 0.
|
109 |
-
"eval_runtime":
|
110 |
-
"eval_samples_per_second":
|
111 |
-
"eval_steps_per_second": 4.
|
112 |
-
"eval_wer": 0.
|
113 |
"step": 1000
|
114 |
},
|
115 |
{
|
116 |
-
"epoch":
|
117 |
-
"eval_loss": 0.
|
118 |
-
"eval_runtime":
|
119 |
-
"eval_samples_per_second":
|
120 |
-
"eval_steps_per_second": 4.
|
121 |
-
"eval_wer": 0.
|
122 |
"step": 1100
|
123 |
},
|
124 |
{
|
125 |
-
"epoch":
|
126 |
-
"eval_loss": 0.
|
127 |
-
"eval_runtime":
|
128 |
-
"eval_samples_per_second":
|
129 |
-
"eval_steps_per_second": 4.
|
130 |
-
"eval_wer": 0.
|
131 |
"step": 1200
|
132 |
},
|
133 |
{
|
134 |
-
"epoch":
|
135 |
-
"eval_loss": 0.
|
136 |
-
"eval_runtime":
|
137 |
-
"eval_samples_per_second":
|
138 |
-
"eval_steps_per_second": 4.
|
139 |
-
"eval_wer": 0.
|
140 |
"step": 1300
|
141 |
},
|
142 |
{
|
143 |
-
"epoch":
|
144 |
-
"eval_loss": 0.
|
145 |
-
"eval_runtime":
|
146 |
-
"eval_samples_per_second":
|
147 |
-
"eval_steps_per_second": 4.
|
148 |
-
"eval_wer": 0.
|
149 |
"step": 1400
|
150 |
},
|
151 |
{
|
152 |
-
"epoch":
|
153 |
-
"grad_norm":
|
154 |
-
"learning_rate": 0.
|
155 |
-
"loss": 0.
|
156 |
"step": 1500
|
157 |
},
|
158 |
{
|
159 |
-
"epoch":
|
160 |
-
"eval_loss": 0.
|
161 |
-
"eval_runtime":
|
162 |
-
"eval_samples_per_second":
|
163 |
-
"eval_steps_per_second": 4.
|
164 |
-
"eval_wer": 0.
|
165 |
"step": 1500
|
166 |
},
|
167 |
{
|
168 |
-
"epoch":
|
169 |
-
"eval_loss": 0.
|
170 |
-
"eval_runtime":
|
171 |
-
"eval_samples_per_second":
|
172 |
-
"eval_steps_per_second": 4.
|
173 |
-
"eval_wer": 0.
|
174 |
"step": 1600
|
175 |
},
|
176 |
{
|
177 |
-
"epoch":
|
178 |
-
"eval_loss": 0.
|
179 |
-
"eval_runtime":
|
180 |
-
"eval_samples_per_second":
|
181 |
-
"eval_steps_per_second": 4.
|
182 |
-
"eval_wer": 0.
|
183 |
"step": 1700
|
184 |
},
|
185 |
{
|
186 |
-
"epoch":
|
187 |
-
"eval_loss": 0.
|
188 |
-
"eval_runtime":
|
189 |
-
"eval_samples_per_second": 35.
|
190 |
-
"eval_steps_per_second": 4.
|
191 |
-
"eval_wer": 0.
|
192 |
"step": 1800
|
193 |
},
|
194 |
{
|
195 |
-
"epoch":
|
196 |
-
"eval_loss": 0.
|
197 |
-
"eval_runtime":
|
198 |
-
"eval_samples_per_second":
|
199 |
-
"eval_steps_per_second": 4.
|
200 |
-
"eval_wer": 0.
|
201 |
"step": 1900
|
202 |
},
|
203 |
{
|
204 |
-
"epoch":
|
205 |
-
"grad_norm": 0.
|
206 |
-
"learning_rate": 0.
|
207 |
-
"loss": 0.
|
208 |
"step": 2000
|
209 |
},
|
210 |
{
|
211 |
-
"epoch":
|
212 |
-
"eval_loss": 0.
|
213 |
-
"eval_runtime":
|
214 |
-
"eval_samples_per_second":
|
215 |
-
"eval_steps_per_second": 4.
|
216 |
-
"eval_wer": 0.
|
217 |
"step": 2000
|
218 |
},
|
219 |
{
|
220 |
-
"epoch":
|
221 |
-
"eval_loss": 0.
|
222 |
-
"eval_runtime":
|
223 |
-
"eval_samples_per_second":
|
224 |
-
"eval_steps_per_second": 4.
|
225 |
-
"eval_wer": 0.
|
226 |
"step": 2100
|
227 |
},
|
228 |
{
|
229 |
-
"epoch":
|
230 |
-
"eval_loss": 0.
|
231 |
-
"eval_runtime":
|
232 |
-
"eval_samples_per_second":
|
233 |
-
"eval_steps_per_second": 4.
|
234 |
-
"eval_wer": 0.
|
235 |
"step": 2200
|
236 |
},
|
237 |
{
|
238 |
-
"epoch":
|
239 |
-
"eval_loss": 0.
|
240 |
-
"eval_runtime":
|
241 |
-
"eval_samples_per_second":
|
242 |
-
"eval_steps_per_second": 4.
|
243 |
-
"eval_wer": 0.
|
244 |
"step": 2300
|
245 |
},
|
246 |
{
|
247 |
-
"epoch":
|
248 |
-
"eval_loss": 0.
|
249 |
-
"eval_runtime":
|
250 |
-
"eval_samples_per_second":
|
251 |
-
"eval_steps_per_second": 4.
|
252 |
-
"eval_wer": 0.
|
253 |
"step": 2400
|
254 |
},
|
255 |
{
|
256 |
-
"epoch":
|
257 |
-
"grad_norm":
|
258 |
-
"learning_rate": 0.
|
259 |
-
"loss": 0.
|
260 |
"step": 2500
|
261 |
},
|
262 |
{
|
263 |
-
"epoch":
|
264 |
-
"eval_loss": 0.
|
265 |
-
"eval_runtime":
|
266 |
-
"eval_samples_per_second":
|
267 |
-
"eval_steps_per_second": 4.
|
268 |
-
"eval_wer": 0.
|
269 |
"step": 2500
|
270 |
},
|
271 |
{
|
272 |
-
"epoch":
|
273 |
-
"eval_loss": 0.
|
274 |
-
"eval_runtime":
|
275 |
-
"eval_samples_per_second":
|
276 |
-
"eval_steps_per_second": 4.
|
277 |
-
"eval_wer": 0.
|
278 |
"step": 2600
|
279 |
},
|
280 |
{
|
281 |
-
"epoch":
|
282 |
-
"eval_loss": 0.
|
283 |
-
"eval_runtime":
|
284 |
-
"eval_samples_per_second":
|
285 |
-
"eval_steps_per_second": 4.
|
286 |
-
"eval_wer": 0.
|
287 |
"step": 2700
|
288 |
},
|
289 |
{
|
290 |
-
"epoch":
|
291 |
-
"eval_loss": 0.
|
292 |
-
"eval_runtime":
|
293 |
-
"eval_samples_per_second":
|
294 |
-
"eval_steps_per_second": 4.
|
295 |
-
"eval_wer": 0.
|
296 |
"step": 2800
|
297 |
},
|
298 |
{
|
299 |
-
"epoch":
|
300 |
-
"eval_loss": 0.
|
301 |
-
"eval_runtime":
|
302 |
-
"eval_samples_per_second":
|
303 |
-
"eval_steps_per_second": 4.
|
304 |
-
"eval_wer": 0.
|
305 |
"step": 2900
|
306 |
},
|
307 |
{
|
308 |
-
"epoch":
|
309 |
-
"grad_norm":
|
310 |
-
"learning_rate": 0.
|
311 |
-
"loss": 0.
|
312 |
"step": 3000
|
313 |
},
|
314 |
{
|
315 |
-
"epoch":
|
316 |
-
"eval_loss": 0.
|
317 |
-
"eval_runtime":
|
318 |
-
"eval_samples_per_second":
|
319 |
-
"eval_steps_per_second": 4.
|
320 |
-
"eval_wer": 0.
|
321 |
"step": 3000
|
322 |
},
|
323 |
{
|
324 |
-
"epoch":
|
325 |
-
"eval_loss":
|
326 |
-
"eval_runtime":
|
327 |
-
"eval_samples_per_second":
|
328 |
-
"eval_steps_per_second": 4.
|
329 |
-
"eval_wer": 0.
|
330 |
"step": 3100
|
331 |
},
|
332 |
{
|
333 |
-
"epoch":
|
334 |
-
"eval_loss":
|
335 |
-
"eval_runtime":
|
336 |
-
"eval_samples_per_second":
|
337 |
-
"eval_steps_per_second": 4.
|
338 |
-
"eval_wer": 0.
|
339 |
"step": 3200
|
340 |
},
|
341 |
{
|
342 |
-
"epoch":
|
343 |
-
"eval_loss":
|
344 |
-
"eval_runtime":
|
345 |
-
"eval_samples_per_second":
|
346 |
-
"eval_steps_per_second": 4.
|
347 |
-
"eval_wer": 0.
|
348 |
"step": 3300
|
349 |
},
|
350 |
{
|
351 |
-
"epoch":
|
352 |
-
"eval_loss":
|
353 |
-
"eval_runtime":
|
354 |
-
"eval_samples_per_second":
|
355 |
-
"eval_steps_per_second": 4.
|
356 |
-
"eval_wer": 0.
|
357 |
"step": 3400
|
358 |
},
|
359 |
{
|
360 |
-
"epoch":
|
361 |
-
"grad_norm":
|
362 |
-
"learning_rate": 0.
|
363 |
-
"loss":
|
364 |
"step": 3500
|
365 |
},
|
366 |
{
|
367 |
-
"epoch":
|
368 |
-
"eval_loss":
|
369 |
-
"eval_runtime":
|
370 |
-
"eval_samples_per_second":
|
371 |
-
"eval_steps_per_second": 4.
|
372 |
-
"eval_wer": 0.
|
373 |
"step": 3500
|
374 |
},
|
375 |
{
|
376 |
-
"epoch":
|
377 |
-
"eval_loss":
|
378 |
-
"eval_runtime":
|
379 |
-
"eval_samples_per_second":
|
380 |
-
"eval_steps_per_second": 4.
|
381 |
-
"eval_wer": 0.
|
382 |
"step": 3600
|
383 |
},
|
384 |
{
|
385 |
-
"epoch":
|
386 |
-
"eval_loss":
|
387 |
-
"eval_runtime":
|
388 |
-
"eval_samples_per_second":
|
389 |
-
"eval_steps_per_second": 4.
|
390 |
-
"eval_wer": 0.
|
391 |
"step": 3700
|
392 |
},
|
393 |
{
|
394 |
-
"epoch":
|
395 |
-
"eval_loss":
|
396 |
-
"eval_runtime":
|
397 |
-
"eval_samples_per_second":
|
398 |
-
"eval_steps_per_second": 4.
|
399 |
-
"eval_wer": 0.
|
400 |
"step": 3800
|
401 |
},
|
402 |
{
|
403 |
-
"epoch":
|
404 |
-
"eval_loss":
|
405 |
-
"eval_runtime":
|
406 |
-
"eval_samples_per_second":
|
407 |
-
"eval_steps_per_second": 4.
|
408 |
-
"eval_wer": 0.
|
409 |
"step": 3900
|
410 |
},
|
411 |
{
|
412 |
-
"epoch":
|
413 |
-
"grad_norm":
|
414 |
-
"learning_rate":
|
415 |
-
"loss":
|
416 |
"step": 4000
|
417 |
},
|
418 |
{
|
419 |
-
"epoch":
|
420 |
-
"eval_loss":
|
421 |
-
"eval_runtime":
|
422 |
-
"eval_samples_per_second":
|
423 |
-
"eval_steps_per_second": 4.
|
424 |
-
"eval_wer": 0.
|
425 |
"step": 4000
|
426 |
},
|
427 |
{
|
428 |
-
"epoch":
|
429 |
-
"eval_loss":
|
430 |
-
"eval_runtime":
|
431 |
-
"eval_samples_per_second":
|
432 |
-
"eval_steps_per_second": 4.
|
433 |
-
"eval_wer": 0.
|
434 |
"step": 4100
|
435 |
},
|
436 |
{
|
437 |
-
"epoch":
|
438 |
-
"eval_loss":
|
439 |
-
"eval_runtime":
|
440 |
-
"eval_samples_per_second":
|
441 |
-
"eval_steps_per_second": 4.
|
442 |
-
"eval_wer": 0.
|
443 |
"step": 4200
|
444 |
},
|
445 |
{
|
446 |
-
"epoch":
|
447 |
-
"eval_loss":
|
448 |
-
"eval_runtime":
|
449 |
-
"eval_samples_per_second":
|
450 |
-
"eval_steps_per_second": 4.
|
451 |
-
"eval_wer": 0.
|
452 |
"step": 4300
|
453 |
},
|
454 |
{
|
455 |
-
"epoch":
|
456 |
-
"eval_loss":
|
457 |
-
"eval_runtime":
|
458 |
-
"eval_samples_per_second":
|
459 |
-
"eval_steps_per_second": 4.
|
460 |
-
"eval_wer": 0.
|
461 |
"step": 4400
|
462 |
},
|
463 |
{
|
464 |
-
"epoch":
|
465 |
-
"grad_norm":
|
466 |
-
"learning_rate":
|
467 |
-
"loss":
|
468 |
"step": 4500
|
469 |
},
|
470 |
{
|
471 |
-
"epoch":
|
472 |
-
"eval_loss":
|
473 |
-
"eval_runtime":
|
474 |
-
"eval_samples_per_second":
|
475 |
-
"eval_steps_per_second": 4.
|
476 |
-
"eval_wer": 0.
|
477 |
"step": 4500
|
478 |
},
|
479 |
{
|
480 |
-
"epoch":
|
481 |
-
"eval_loss":
|
482 |
-
"eval_runtime":
|
483 |
-
"eval_samples_per_second":
|
484 |
-
"eval_steps_per_second": 4.
|
485 |
-
"eval_wer": 0.
|
486 |
"step": 4600
|
487 |
},
|
488 |
{
|
489 |
-
"epoch":
|
490 |
-
"eval_loss":
|
491 |
-
"eval_runtime":
|
492 |
-
"eval_samples_per_second":
|
493 |
-
"eval_steps_per_second": 4.
|
494 |
-
"eval_wer": 0.
|
495 |
"step": 4700
|
496 |
},
|
497 |
{
|
498 |
-
"epoch":
|
499 |
-
"eval_loss":
|
500 |
-
"eval_runtime":
|
501 |
-
"eval_samples_per_second":
|
502 |
-
"eval_steps_per_second": 4.
|
503 |
-
"eval_wer": 0.
|
504 |
"step": 4800
|
505 |
},
|
506 |
{
|
507 |
-
"epoch":
|
508 |
-
"eval_loss":
|
509 |
-
"eval_runtime":
|
510 |
-
"eval_samples_per_second":
|
511 |
-
"eval_steps_per_second": 4.
|
512 |
-
"eval_wer": 0.
|
513 |
"step": 4900
|
514 |
},
|
515 |
{
|
516 |
-
"epoch":
|
517 |
-
"grad_norm":
|
518 |
-
"learning_rate":
|
519 |
-
"loss":
|
520 |
"step": 5000
|
521 |
},
|
522 |
{
|
523 |
-
"epoch":
|
524 |
-
"eval_loss":
|
525 |
-
"eval_runtime":
|
526 |
-
"eval_samples_per_second":
|
527 |
-
"eval_steps_per_second": 4.
|
528 |
-
"eval_wer": 0.
|
529 |
"step": 5000
|
530 |
},
|
531 |
{
|
532 |
-
"epoch":
|
533 |
-
"
|
534 |
-
"
|
535 |
-
"
|
536 |
-
"
|
537 |
-
"
|
538 |
-
"
|
539 |
-
},
|
540 |
-
{
|
541 |
-
"epoch": 2.9445073612684034,
|
542 |
-
"eval_loss": 0.5343597531318665,
|
543 |
-
"eval_runtime": 163.0683,
|
544 |
-
"eval_samples_per_second": 34.685,
|
545 |
-
"eval_steps_per_second": 4.336,
|
546 |
-
"eval_wer": 0.38002920832597775,
|
547 |
-
"step": 5200
|
548 |
-
},
|
549 |
-
{
|
550 |
-
"epoch": 3.001132502831257,
|
551 |
-
"eval_loss": 0.534857451915741,
|
552 |
-
"eval_runtime": 153.2271,
|
553 |
-
"eval_samples_per_second": 36.913,
|
554 |
-
"eval_steps_per_second": 4.614,
|
555 |
-
"eval_wer": 0.38415368073052913,
|
556 |
-
"step": 5300
|
557 |
-
},
|
558 |
-
{
|
559 |
-
"epoch": 3.057757644394111,
|
560 |
-
"eval_loss": 0.5775672793388367,
|
561 |
-
"eval_runtime": 151.3492,
|
562 |
-
"eval_samples_per_second": 37.371,
|
563 |
-
"eval_steps_per_second": 4.671,
|
564 |
-
"eval_wer": 0.4182728571199307,
|
565 |
-
"step": 5400
|
566 |
-
},
|
567 |
-
{
|
568 |
-
"epoch": 3.114382785956965,
|
569 |
-
"grad_norm": 1.614545226097107,
|
570 |
-
"learning_rate": 0.00015039999999999997,
|
571 |
-
"loss": 0.2839,
|
572 |
-
"step": 5500
|
573 |
-
},
|
574 |
-
{
|
575 |
-
"epoch": 3.114382785956965,
|
576 |
-
"eval_loss": 0.5882839560508728,
|
577 |
-
"eval_runtime": 150.3863,
|
578 |
-
"eval_samples_per_second": 37.61,
|
579 |
-
"eval_steps_per_second": 4.701,
|
580 |
-
"eval_wer": 0.41002391231082796,
|
581 |
-
"step": 5500
|
582 |
-
},
|
583 |
-
{
|
584 |
-
"epoch": 3.1710079275198186,
|
585 |
-
"eval_loss": 0.5722731947898865,
|
586 |
-
"eval_runtime": 154.6093,
|
587 |
-
"eval_samples_per_second": 36.583,
|
588 |
-
"eval_steps_per_second": 4.573,
|
589 |
-
"eval_wer": 0.4043587809536037,
|
590 |
-
"step": 5600
|
591 |
-
},
|
592 |
-
{
|
593 |
-
"epoch": 3.227633069082673,
|
594 |
-
"eval_loss": 0.5630057454109192,
|
595 |
-
"eval_runtime": 153.3174,
|
596 |
-
"eval_samples_per_second": 36.891,
|
597 |
-
"eval_steps_per_second": 4.611,
|
598 |
-
"eval_wer": 0.40779316653560366,
|
599 |
-
"step": 5700
|
600 |
-
},
|
601 |
-
{
|
602 |
-
"epoch": 3.2842582106455267,
|
603 |
-
"eval_loss": 0.5810334086418152,
|
604 |
-
"eval_runtime": 153.0372,
|
605 |
-
"eval_samples_per_second": 36.958,
|
606 |
-
"eval_steps_per_second": 4.62,
|
607 |
-
"eval_wer": 0.4191394777808092,
|
608 |
-
"step": 5800
|
609 |
-
},
|
610 |
-
{
|
611 |
-
"epoch": 3.3408833522083805,
|
612 |
-
"eval_loss": 0.5995615720748901,
|
613 |
-
"eval_runtime": 152.3384,
|
614 |
-
"eval_samples_per_second": 37.128,
|
615 |
-
"eval_steps_per_second": 4.641,
|
616 |
-
"eval_wer": 0.4228306398549213,
|
617 |
-
"step": 5900
|
618 |
-
},
|
619 |
-
{
|
620 |
-
"epoch": 3.3975084937712343,
|
621 |
-
"grad_norm": 9.246959686279297,
|
622 |
-
"learning_rate": 0.00013373333333333332,
|
623 |
-
"loss": 0.3019,
|
624 |
-
"step": 6000
|
625 |
-
},
|
626 |
-
{
|
627 |
-
"epoch": 3.3975084937712343,
|
628 |
-
"eval_loss": 0.5681526064872742,
|
629 |
-
"eval_runtime": 153.1128,
|
630 |
-
"eval_samples_per_second": 36.94,
|
631 |
-
"eval_steps_per_second": 4.618,
|
632 |
-
"eval_wer": 0.4015502880711271,
|
633 |
-
"step": 6000
|
634 |
-
},
|
635 |
-
{
|
636 |
-
"epoch": 3.454133635334088,
|
637 |
-
"eval_loss": 0.5560505390167236,
|
638 |
-
"eval_runtime": 152.0848,
|
639 |
-
"eval_samples_per_second": 37.19,
|
640 |
-
"eval_steps_per_second": 4.649,
|
641 |
-
"eval_wer": 0.40569080900643545,
|
642 |
-
"step": 6100
|
643 |
-
},
|
644 |
-
{
|
645 |
-
"epoch": 3.5107587768969424,
|
646 |
-
"eval_loss": 0.5905264616012573,
|
647 |
-
"eval_runtime": 152.0038,
|
648 |
-
"eval_samples_per_second": 37.21,
|
649 |
-
"eval_steps_per_second": 4.651,
|
650 |
-
"eval_wer": 0.41458169504581854,
|
651 |
-
"step": 6200
|
652 |
-
},
|
653 |
-
{
|
654 |
-
"epoch": 3.567383918459796,
|
655 |
-
"eval_loss": 0.5875168442726135,
|
656 |
-
"eval_runtime": 151.4059,
|
657 |
-
"eval_samples_per_second": 37.357,
|
658 |
-
"eval_steps_per_second": 4.67,
|
659 |
-
"eval_wer": 0.4190271380655101,
|
660 |
-
"step": 6300
|
661 |
-
},
|
662 |
-
{
|
663 |
-
"epoch": 3.62400906002265,
|
664 |
-
"eval_loss": 0.5877885818481445,
|
665 |
-
"eval_runtime": 150.4489,
|
666 |
-
"eval_samples_per_second": 37.594,
|
667 |
-
"eval_steps_per_second": 4.699,
|
668 |
-
"eval_wer": 0.44462454462293977,
|
669 |
-
"step": 6400
|
670 |
-
},
|
671 |
-
{
|
672 |
-
"epoch": 3.680634201585504,
|
673 |
-
"grad_norm": 1.084346890449524,
|
674 |
-
"learning_rate": 0.00011709999999999999,
|
675 |
-
"loss": 0.2944,
|
676 |
-
"step": 6500
|
677 |
-
},
|
678 |
-
{
|
679 |
-
"epoch": 3.680634201585504,
|
680 |
-
"eval_loss": 0.5938708782196045,
|
681 |
-
"eval_runtime": 150.586,
|
682 |
-
"eval_samples_per_second": 37.56,
|
683 |
-
"eval_steps_per_second": 4.695,
|
684 |
-
"eval_wer": 0.4403877325030893,
|
685 |
-
"step": 6500
|
686 |
-
},
|
687 |
-
{
|
688 |
-
"epoch": 3.7372593431483576,
|
689 |
-
"eval_loss": 0.590270459651947,
|
690 |
-
"eval_runtime": 150.087,
|
691 |
-
"eval_samples_per_second": 37.685,
|
692 |
-
"eval_steps_per_second": 4.711,
|
693 |
-
"eval_wer": 0.4183049541814447,
|
694 |
-
"step": 6600
|
695 |
-
},
|
696 |
-
{
|
697 |
-
"epoch": 3.793884484711212,
|
698 |
-
"eval_loss": 0.5807533264160156,
|
699 |
-
"eval_runtime": 149.6534,
|
700 |
-
"eval_samples_per_second": 37.794,
|
701 |
-
"eval_steps_per_second": 4.724,
|
702 |
-
"eval_wer": 0.4059475854985476,
|
703 |
-
"step": 6700
|
704 |
-
},
|
705 |
-
{
|
706 |
-
"epoch": 3.8505096262740657,
|
707 |
-
"eval_loss": 0.6154611706733704,
|
708 |
-
"eval_runtime": 150.8194,
|
709 |
-
"eval_samples_per_second": 37.502,
|
710 |
-
"eval_steps_per_second": 4.688,
|
711 |
-
"eval_wer": 0.410056009372342,
|
712 |
-
"step": 6800
|
713 |
-
},
|
714 |
-
{
|
715 |
-
"epoch": 3.9071347678369195,
|
716 |
-
"eval_loss": 0.7987228631973267,
|
717 |
-
"eval_runtime": 151.5296,
|
718 |
-
"eval_samples_per_second": 37.326,
|
719 |
-
"eval_steps_per_second": 4.666,
|
720 |
-
"eval_wer": 0.5822727929258077,
|
721 |
-
"step": 6900
|
722 |
-
},
|
723 |
-
{
|
724 |
-
"epoch": 3.9637599093997737,
|
725 |
-
"grad_norm": 9.350592613220215,
|
726 |
-
"learning_rate": 0.0001005,
|
727 |
-
"loss": 0.3918,
|
728 |
-
"step": 7000
|
729 |
-
},
|
730 |
-
{
|
731 |
-
"epoch": 3.9637599093997737,
|
732 |
-
"eval_loss": 0.9750258326530457,
|
733 |
-
"eval_runtime": 151.1967,
|
734 |
-
"eval_samples_per_second": 37.408,
|
735 |
-
"eval_steps_per_second": 4.676,
|
736 |
-
"eval_wer": 0.5545409317776957,
|
737 |
-
"step": 7000
|
738 |
-
},
|
739 |
-
{
|
740 |
-
"epoch": 4.020385050962627,
|
741 |
-
"eval_loss": 1.0540127754211426,
|
742 |
-
"eval_runtime": 150.2681,
|
743 |
-
"eval_samples_per_second": 37.639,
|
744 |
-
"eval_steps_per_second": 4.705,
|
745 |
-
"eval_wer": 0.5689043668052189,
|
746 |
-
"step": 7100
|
747 |
-
},
|
748 |
-
{
|
749 |
-
"epoch": 4.077010192525481,
|
750 |
-
"eval_loss": 0.6850923299789429,
|
751 |
-
"eval_runtime": 150.3993,
|
752 |
-
"eval_samples_per_second": 37.607,
|
753 |
-
"eval_steps_per_second": 4.701,
|
754 |
-
"eval_wer": 0.4396013544959959,
|
755 |
-
"step": 7200
|
756 |
-
},
|
757 |
-
{
|
758 |
-
"epoch": 4.133635334088336,
|
759 |
-
"eval_loss": 0.7331786155700684,
|
760 |
-
"eval_runtime": 151.5017,
|
761 |
-
"eval_samples_per_second": 37.333,
|
762 |
-
"eval_steps_per_second": 4.667,
|
763 |
-
"eval_wer": 0.49731187109820096,
|
764 |
-
"step": 7300
|
765 |
-
},
|
766 |
-
{
|
767 |
-
"epoch": 4.190260475651189,
|
768 |
-
"eval_loss": 0.9466453194618225,
|
769 |
-
"eval_runtime": 150.2633,
|
770 |
-
"eval_samples_per_second": 37.641,
|
771 |
-
"eval_steps_per_second": 4.705,
|
772 |
-
"eval_wer": 0.6394697565437885,
|
773 |
-
"step": 7400
|
774 |
-
},
|
775 |
-
{
|
776 |
-
"epoch": 4.246885617214043,
|
777 |
-
"grad_norm": 4.335416793823242,
|
778 |
-
"learning_rate": 8.386666666666665e-05,
|
779 |
-
"loss": 0.5378,
|
780 |
-
"step": 7500
|
781 |
-
},
|
782 |
-
{
|
783 |
-
"epoch": 4.246885617214043,
|
784 |
-
"eval_loss": 0.8257068991661072,
|
785 |
-
"eval_runtime": 151.0198,
|
786 |
-
"eval_samples_per_second": 37.452,
|
787 |
-
"eval_steps_per_second": 4.682,
|
788 |
-
"eval_wer": 0.485147084784388,
|
789 |
-
"step": 7500
|
790 |
-
},
|
791 |
-
{
|
792 |
-
"epoch": 4.303510758776897,
|
793 |
-
"eval_loss": 0.8490071296691895,
|
794 |
-
"eval_runtime": 150.818,
|
795 |
-
"eval_samples_per_second": 37.502,
|
796 |
-
"eval_steps_per_second": 4.688,
|
797 |
-
"eval_wer": 0.4867037922678179,
|
798 |
-
"step": 7600
|
799 |
-
},
|
800 |
-
{
|
801 |
-
"epoch": 4.360135900339751,
|
802 |
-
"eval_loss": 0.8716742396354675,
|
803 |
-
"eval_runtime": 150.797,
|
804 |
-
"eval_samples_per_second": 37.507,
|
805 |
-
"eval_steps_per_second": 4.688,
|
806 |
-
"eval_wer": 0.47112066890276194,
|
807 |
-
"step": 7700
|
808 |
-
},
|
809 |
-
{
|
810 |
-
"epoch": 4.416761041902605,
|
811 |
-
"eval_loss": 0.883883535861969,
|
812 |
-
"eval_runtime": 150.6157,
|
813 |
-
"eval_samples_per_second": 37.553,
|
814 |
-
"eval_steps_per_second": 4.694,
|
815 |
-
"eval_wer": 0.5860281491229478,
|
816 |
-
"step": 7800
|
817 |
-
},
|
818 |
-
{
|
819 |
-
"epoch": 4.4733861834654585,
|
820 |
-
"eval_loss": 2.911261558532715,
|
821 |
-
"eval_runtime": 150.8548,
|
822 |
-
"eval_samples_per_second": 37.493,
|
823 |
-
"eval_steps_per_second": 4.687,
|
824 |
-
"eval_wer": 1.0,
|
825 |
-
"step": 7900
|
826 |
-
},
|
827 |
-
{
|
828 |
-
"epoch": 4.530011325028313,
|
829 |
-
"grad_norm": 0.2805318534374237,
|
830 |
-
"learning_rate": 6.723333333333333e-05,
|
831 |
-
"loss": 1.3847,
|
832 |
-
"step": 8000
|
833 |
-
},
|
834 |
-
{
|
835 |
-
"epoch": 4.530011325028313,
|
836 |
-
"eval_loss": 2.8575596809387207,
|
837 |
-
"eval_runtime": 151.2637,
|
838 |
-
"eval_samples_per_second": 37.392,
|
839 |
-
"eval_steps_per_second": 4.674,
|
840 |
-
"eval_wer": 1.0,
|
841 |
-
"step": 8000
|
842 |
-
},
|
843 |
-
{
|
844 |
-
"epoch": 4.586636466591166,
|
845 |
-
"eval_loss": 2.83913516998291,
|
846 |
-
"eval_runtime": 151.4914,
|
847 |
-
"eval_samples_per_second": 37.335,
|
848 |
-
"eval_steps_per_second": 4.667,
|
849 |
-
"eval_wer": 1.0,
|
850 |
-
"step": 8100
|
851 |
-
},
|
852 |
-
{
|
853 |
-
"epoch": 4.64326160815402,
|
854 |
-
"eval_loss": 2.840644121170044,
|
855 |
-
"eval_runtime": 150.8361,
|
856 |
-
"eval_samples_per_second": 37.498,
|
857 |
-
"eval_steps_per_second": 4.687,
|
858 |
-
"eval_wer": 1.0,
|
859 |
-
"step": 8200
|
860 |
-
},
|
861 |
-
{
|
862 |
-
"epoch": 4.699886749716875,
|
863 |
-
"eval_loss": 2.856635093688965,
|
864 |
-
"eval_runtime": 150.3023,
|
865 |
-
"eval_samples_per_second": 37.631,
|
866 |
-
"eval_steps_per_second": 4.704,
|
867 |
-
"eval_wer": 1.0,
|
868 |
-
"step": 8300
|
869 |
-
},
|
870 |
-
{
|
871 |
-
"epoch": 4.756511891279728,
|
872 |
-
"eval_loss": 2.8454244136810303,
|
873 |
-
"eval_runtime": 150.6924,
|
874 |
-
"eval_samples_per_second": 37.533,
|
875 |
-
"eval_steps_per_second": 4.692,
|
876 |
-
"eval_wer": 0.999823466161673,
|
877 |
-
"step": 8400
|
878 |
-
},
|
879 |
-
{
|
880 |
-
"epoch": 4.813137032842582,
|
881 |
-
"grad_norm": 0.4468824863433838,
|
882 |
-
"learning_rate": 5.06e-05,
|
883 |
-
"loss": 2.8136,
|
884 |
-
"step": 8500
|
885 |
-
},
|
886 |
-
{
|
887 |
-
"epoch": 4.813137032842582,
|
888 |
-
"eval_loss": 2.8339831829071045,
|
889 |
-
"eval_runtime": 151.2056,
|
890 |
-
"eval_samples_per_second": 37.406,
|
891 |
-
"eval_steps_per_second": 4.676,
|
892 |
-
"eval_wer": 0.999919757346215,
|
893 |
-
"step": 8500
|
894 |
-
},
|
895 |
-
{
|
896 |
-
"epoch": 4.869762174405436,
|
897 |
-
"eval_loss": 2.836662530899048,
|
898 |
-
"eval_runtime": 152.389,
|
899 |
-
"eval_samples_per_second": 37.116,
|
900 |
-
"eval_steps_per_second": 4.639,
|
901 |
-
"eval_wer": 0.999935805876972,
|
902 |
-
"step": 8600
|
903 |
-
},
|
904 |
-
{
|
905 |
-
"epoch": 4.92638731596829,
|
906 |
-
"eval_loss": 2.833369016647339,
|
907 |
-
"eval_runtime": 151.3614,
|
908 |
-
"eval_samples_per_second": 37.368,
|
909 |
-
"eval_steps_per_second": 4.671,
|
910 |
-
"eval_wer": 0.9998876602847009,
|
911 |
-
"step": 8700
|
912 |
-
},
|
913 |
-
{
|
914 |
-
"epoch": 4.983012457531144,
|
915 |
-
"eval_loss": 2.8321285247802734,
|
916 |
-
"eval_runtime": 157.5502,
|
917 |
-
"eval_samples_per_second": 35.9,
|
918 |
-
"eval_steps_per_second": 4.487,
|
919 |
-
"eval_wer": 0.999935805876972,
|
920 |
-
"step": 8800
|
921 |
-
},
|
922 |
-
{
|
923 |
-
"epoch": 5.0396375990939974,
|
924 |
-
"eval_loss": 2.802509307861328,
|
925 |
-
"eval_runtime": 151.5865,
|
926 |
-
"eval_samples_per_second": 37.312,
|
927 |
-
"eval_steps_per_second": 4.664,
|
928 |
-
"eval_wer": 0.9982186130859719,
|
929 |
-
"step": 8900
|
930 |
-
},
|
931 |
-
{
|
932 |
-
"epoch": 5.096262740656852,
|
933 |
-
"grad_norm": 0.3411979377269745,
|
934 |
-
"learning_rate": 3.393333333333333e-05,
|
935 |
-
"loss": 2.8007,
|
936 |
-
"step": 9000
|
937 |
-
},
|
938 |
-
{
|
939 |
-
"epoch": 5.096262740656852,
|
940 |
-
"eval_loss": 2.8024423122406006,
|
941 |
-
"eval_runtime": 151.6682,
|
942 |
-
"eval_samples_per_second": 37.292,
|
943 |
-
"eval_steps_per_second": 4.661,
|
944 |
-
"eval_wer": 0.9974964292019066,
|
945 |
-
"step": 9000
|
946 |
-
},
|
947 |
-
{
|
948 |
-
"epoch": 5.152887882219706,
|
949 |
-
"eval_loss": 2.8043477535247803,
|
950 |
-
"eval_runtime": 153.1811,
|
951 |
-
"eval_samples_per_second": 36.924,
|
952 |
-
"eval_steps_per_second": 4.615,
|
953 |
-
"eval_wer": 0.9981062733706729,
|
954 |
-
"step": 9100
|
955 |
-
},
|
956 |
-
{
|
957 |
-
"epoch": 5.209513023782559,
|
958 |
-
"eval_loss": 2.810584783554077,
|
959 |
-
"eval_runtime": 151.6694,
|
960 |
-
"eval_samples_per_second": 37.292,
|
961 |
-
"eval_steps_per_second": 4.661,
|
962 |
-
"eval_wer": 0.9992457190544206,
|
963 |
-
"step": 9200
|
964 |
-
},
|
965 |
-
{
|
966 |
-
"epoch": 5.266138165345414,
|
967 |
-
"eval_loss": 2.8066723346710205,
|
968 |
-
"eval_runtime": 151.6427,
|
969 |
-
"eval_samples_per_second": 37.298,
|
970 |
-
"eval_steps_per_second": 4.662,
|
971 |
-
"eval_wer": 0.9993259617082056,
|
972 |
-
"step": 9300
|
973 |
-
},
|
974 |
-
{
|
975 |
-
"epoch": 5.322763306908267,
|
976 |
-
"eval_loss": 2.805284023284912,
|
977 |
-
"eval_runtime": 151.8566,
|
978 |
-
"eval_samples_per_second": 37.246,
|
979 |
-
"eval_steps_per_second": 4.656,
|
980 |
-
"eval_wer": 0.9985877292933832,
|
981 |
-
"step": 9400
|
982 |
-
},
|
983 |
-
{
|
984 |
-
"epoch": 5.379388448471121,
|
985 |
-
"grad_norm": 0.15991327166557312,
|
986 |
-
"learning_rate": 1.7299999999999997e-05,
|
987 |
-
"loss": 2.7935,
|
988 |
-
"step": 9500
|
989 |
-
},
|
990 |
-
{
|
991 |
-
"epoch": 5.379388448471121,
|
992 |
-
"eval_loss": 2.807739734649658,
|
993 |
-
"eval_runtime": 152.8905,
|
994 |
-
"eval_samples_per_second": 36.994,
|
995 |
-
"eval_steps_per_second": 4.624,
|
996 |
-
"eval_wer": 0.9978173998170468,
|
997 |
-
"step": 9500
|
998 |
-
},
|
999 |
-
{
|
1000 |
-
"epoch": 5.436013590033975,
|
1001 |
-
"eval_loss": 2.8082854747772217,
|
1002 |
-
"eval_runtime": 151.7494,
|
1003 |
-
"eval_samples_per_second": 37.272,
|
1004 |
-
"eval_steps_per_second": 4.659,
|
1005 |
-
"eval_wer": 0.9987161175394392,
|
1006 |
-
"step": 9600
|
1007 |
-
},
|
1008 |
-
{
|
1009 |
-
"epoch": 5.492638731596829,
|
1010 |
-
"eval_loss": 2.8080484867095947,
|
1011 |
-
"eval_runtime": 151.6292,
|
1012 |
-
"eval_samples_per_second": 37.302,
|
1013 |
-
"eval_steps_per_second": 4.663,
|
1014 |
-
"eval_wer": 0.9988926513777664,
|
1015 |
-
"step": 9700
|
1016 |
-
},
|
1017 |
-
{
|
1018 |
-
"epoch": 5.549263873159683,
|
1019 |
-
"eval_loss": 2.808607578277588,
|
1020 |
-
"eval_runtime": 153.2376,
|
1021 |
-
"eval_samples_per_second": 36.91,
|
1022 |
-
"eval_steps_per_second": 4.614,
|
1023 |
-
"eval_wer": 0.9986037778241402,
|
1024 |
-
"step": 9800
|
1025 |
-
},
|
1026 |
-
{
|
1027 |
-
"epoch": 5.605889014722536,
|
1028 |
-
"eval_loss": 2.8079216480255127,
|
1029 |
-
"eval_runtime": 151.6181,
|
1030 |
-
"eval_samples_per_second": 37.304,
|
1031 |
-
"eval_steps_per_second": 4.663,
|
1032 |
-
"eval_wer": 0.998186516024458,
|
1033 |
-
"step": 9900
|
1034 |
-
},
|
1035 |
-
{
|
1036 |
-
"epoch": 5.662514156285391,
|
1037 |
-
"grad_norm": 0.3985973000526428,
|
1038 |
-
"learning_rate": 7e-07,
|
1039 |
-
"loss": 2.7861,
|
1040 |
-
"step": 10000
|
1041 |
-
},
|
1042 |
-
{
|
1043 |
-
"epoch": 5.662514156285391,
|
1044 |
-
"eval_loss": 2.807446241378784,
|
1045 |
-
"eval_runtime": 151.2977,
|
1046 |
-
"eval_samples_per_second": 37.383,
|
1047 |
-
"eval_steps_per_second": 4.673,
|
1048 |
-
"eval_wer": 0.998266758678243,
|
1049 |
-
"step": 10000
|
1050 |
-
},
|
1051 |
-
{
|
1052 |
-
"epoch": 5.662514156285391,
|
1053 |
-
"step": 10000,
|
1054 |
-
"total_flos": 9.138710330328565e+19,
|
1055 |
-
"train_loss": 1.1088516311645509,
|
1056 |
-
"train_runtime": 39423.4073,
|
1057 |
-
"train_samples_per_second": 16.234,
|
1058 |
-
"train_steps_per_second": 0.254
|
1059 |
}
|
1060 |
],
|
1061 |
"logging_steps": 500,
|
1062 |
-
"max_steps":
|
1063 |
"num_input_tokens_seen": 0,
|
1064 |
-
"num_train_epochs":
|
1065 |
"save_steps": 400,
|
1066 |
-
"total_flos":
|
1067 |
"train_batch_size": 64,
|
1068 |
"trial_name": null,
|
1069 |
"trial_params": null
|
|
|
1 |
{
|
2 |
"best_metric": null,
|
3 |
"best_model_checkpoint": null,
|
4 |
+
"epoch": 7.739938080495356,
|
5 |
"eval_steps": 100,
|
6 |
+
"global_step": 5000,
|
7 |
"is_hyper_param_search": false,
|
8 |
"is_local_process_zero": true,
|
9 |
"is_world_process_zero": true,
|
10 |
"log_history": [
|
11 |
{
|
12 |
+
"epoch": 0.15479876160990713,
|
13 |
+
"eval_loss": 3.5586698055267334,
|
14 |
+
"eval_runtime": 163.4718,
|
15 |
+
"eval_samples_per_second": 34.599,
|
16 |
+
"eval_steps_per_second": 4.325,
|
17 |
"eval_wer": 1.0,
|
18 |
"step": 100
|
19 |
},
|
20 |
{
|
21 |
+
"epoch": 0.30959752321981426,
|
22 |
+
"eval_loss": 3.2505505084991455,
|
23 |
+
"eval_runtime": 161.1897,
|
24 |
+
"eval_samples_per_second": 35.089,
|
25 |
+
"eval_steps_per_second": 4.386,
|
26 |
"eval_wer": 1.0,
|
27 |
"step": 200
|
28 |
},
|
29 |
{
|
30 |
+
"epoch": 0.46439628482972134,
|
31 |
+
"eval_loss": 2.774010181427002,
|
32 |
+
"eval_runtime": 161.9005,
|
33 |
+
"eval_samples_per_second": 34.935,
|
34 |
+
"eval_steps_per_second": 4.367,
|
35 |
+
"eval_wer": 0.9999518544077289,
|
36 |
"step": 300
|
37 |
},
|
38 |
{
|
39 |
+
"epoch": 0.6191950464396285,
|
40 |
+
"eval_loss": 1.11959707736969,
|
41 |
+
"eval_runtime": 162.1425,
|
42 |
+
"eval_samples_per_second": 34.883,
|
43 |
+
"eval_steps_per_second": 4.36,
|
44 |
+
"eval_wer": 0.7807449727977404,
|
45 |
"step": 400
|
46 |
},
|
47 |
{
|
48 |
+
"epoch": 0.7739938080495357,
|
49 |
+
"grad_norm": 2.676631212234497,
|
50 |
+
"learning_rate": 0.00029699999999999996,
|
51 |
+
"loss": 3.6484,
|
52 |
"step": 500
|
53 |
},
|
54 |
{
|
55 |
+
"epoch": 0.7739938080495357,
|
56 |
+
"eval_loss": 0.9134386777877808,
|
57 |
+
"eval_runtime": 162.7792,
|
58 |
+
"eval_samples_per_second": 34.746,
|
59 |
+
"eval_steps_per_second": 4.343,
|
60 |
+
"eval_wer": 0.6538813371635827,
|
61 |
"step": 500
|
62 |
},
|
63 |
{
|
64 |
+
"epoch": 0.9287925696594427,
|
65 |
+
"eval_loss": 0.7674785256385803,
|
66 |
+
"eval_runtime": 162.7057,
|
67 |
+
"eval_samples_per_second": 34.762,
|
68 |
+
"eval_steps_per_second": 4.345,
|
69 |
+
"eval_wer": 0.5923191731796954,
|
70 |
"step": 600
|
71 |
},
|
72 |
{
|
73 |
+
"epoch": 1.08359133126935,
|
74 |
+
"eval_loss": 0.7207810282707214,
|
75 |
+
"eval_runtime": 163.9828,
|
76 |
+
"eval_samples_per_second": 34.491,
|
77 |
+
"eval_steps_per_second": 4.311,
|
78 |
+
"eval_wer": 0.5289595737510231,
|
79 |
"step": 700
|
80 |
},
|
81 |
{
|
82 |
+
"epoch": 1.238390092879257,
|
83 |
+
"eval_loss": 0.6209472417831421,
|
84 |
+
"eval_runtime": 163.6218,
|
85 |
+
"eval_samples_per_second": 34.568,
|
86 |
+
"eval_steps_per_second": 4.321,
|
87 |
+
"eval_wer": 0.4744748118309769,
|
88 |
"step": 800
|
89 |
},
|
90 |
{
|
91 |
+
"epoch": 1.3931888544891642,
|
92 |
+
"eval_loss": 0.6220189332962036,
|
93 |
+
"eval_runtime": 162.6081,
|
94 |
+
"eval_samples_per_second": 34.783,
|
95 |
+
"eval_steps_per_second": 4.348,
|
96 |
+
"eval_wer": 0.47879186660461237,
|
97 |
"step": 900
|
98 |
},
|
99 |
{
|
100 |
+
"epoch": 1.5479876160990713,
|
101 |
+
"grad_norm": 0.3950090706348419,
|
102 |
+
"learning_rate": 0.0002672,
|
103 |
+
"loss": 0.6286,
|
104 |
"step": 1000
|
105 |
},
|
106 |
{
|
107 |
+
"epoch": 1.5479876160990713,
|
108 |
+
"eval_loss": 0.5738953351974487,
|
109 |
+
"eval_runtime": 162.8483,
|
110 |
+
"eval_samples_per_second": 34.732,
|
111 |
+
"eval_steps_per_second": 4.341,
|
112 |
+
"eval_wer": 0.4588114458121359,
|
113 |
"step": 1000
|
114 |
},
|
115 |
{
|
116 |
+
"epoch": 1.7027863777089784,
|
117 |
+
"eval_loss": 0.564153790473938,
|
118 |
+
"eval_runtime": 164.0752,
|
119 |
+
"eval_samples_per_second": 34.472,
|
120 |
+
"eval_steps_per_second": 4.309,
|
121 |
+
"eval_wer": 0.4262008313138932,
|
122 |
"step": 1100
|
123 |
},
|
124 |
{
|
125 |
+
"epoch": 1.8575851393188856,
|
126 |
+
"eval_loss": 0.5511888265609741,
|
127 |
+
"eval_runtime": 164.5846,
|
128 |
+
"eval_samples_per_second": 34.365,
|
129 |
+
"eval_steps_per_second": 4.296,
|
130 |
+
"eval_wer": 0.42080852497953813,
|
131 |
"step": 1200
|
132 |
},
|
133 |
{
|
134 |
+
"epoch": 2.0123839009287927,
|
135 |
+
"eval_loss": 0.527522623538971,
|
136 |
+
"eval_runtime": 162.1572,
|
137 |
+
"eval_samples_per_second": 34.88,
|
138 |
+
"eval_steps_per_second": 4.36,
|
139 |
+
"eval_wer": 0.38652886328256647,
|
140 |
"step": 1300
|
141 |
},
|
142 |
{
|
143 |
+
"epoch": 2.1671826625387,
|
144 |
+
"eval_loss": 0.4955059587955475,
|
145 |
+
"eval_runtime": 160.6015,
|
146 |
+
"eval_samples_per_second": 35.218,
|
147 |
+
"eval_steps_per_second": 4.402,
|
148 |
+
"eval_wer": 0.37545537706023013,
|
149 |
"step": 1400
|
150 |
},
|
151 |
{
|
152 |
+
"epoch": 2.321981424148607,
|
153 |
+
"grad_norm": 0.4153783619403839,
|
154 |
+
"learning_rate": 0.00023393333333333332,
|
155 |
+
"loss": 0.4816,
|
156 |
"step": 1500
|
157 |
},
|
158 |
{
|
159 |
+
"epoch": 2.321981424148607,
|
160 |
+
"eval_loss": 0.4908938407897949,
|
161 |
+
"eval_runtime": 160.2948,
|
162 |
+
"eval_samples_per_second": 35.285,
|
163 |
+
"eval_steps_per_second": 4.411,
|
164 |
+
"eval_wer": 0.37325672834651985,
|
165 |
"step": 1500
|
166 |
},
|
167 |
{
|
168 |
+
"epoch": 2.476780185758514,
|
169 |
+
"eval_loss": 0.4982919991016388,
|
170 |
+
"eval_runtime": 162.1379,
|
171 |
+
"eval_samples_per_second": 34.884,
|
172 |
+
"eval_steps_per_second": 4.36,
|
173 |
+
"eval_wer": 0.37279132095456663,
|
174 |
"step": 1600
|
175 |
},
|
176 |
{
|
177 |
+
"epoch": 2.6315789473684212,
|
178 |
+
"eval_loss": 0.48909762501716614,
|
179 |
+
"eval_runtime": 161.7559,
|
180 |
+
"eval_samples_per_second": 34.966,
|
181 |
+
"eval_steps_per_second": 4.371,
|
182 |
+
"eval_wer": 0.36550528799088444,
|
183 |
"step": 1700
|
184 |
},
|
185 |
{
|
186 |
+
"epoch": 2.7863777089783284,
|
187 |
+
"eval_loss": 0.47961312532424927,
|
188 |
+
"eval_runtime": 160.2886,
|
189 |
+
"eval_samples_per_second": 35.286,
|
190 |
+
"eval_steps_per_second": 4.411,
|
191 |
+
"eval_wer": 0.3570958578742116,
|
192 |
"step": 1800
|
193 |
},
|
194 |
{
|
195 |
+
"epoch": 2.9411764705882355,
|
196 |
+
"eval_loss": 0.46432051062583923,
|
197 |
+
"eval_runtime": 162.1185,
|
198 |
+
"eval_samples_per_second": 34.888,
|
199 |
+
"eval_steps_per_second": 4.361,
|
200 |
+
"eval_wer": 0.3591982154033798,
|
201 |
"step": 1900
|
202 |
},
|
203 |
{
|
204 |
+
"epoch": 3.0959752321981426,
|
205 |
+
"grad_norm": 0.8931769728660583,
|
206 |
+
"learning_rate": 0.00020079999999999997,
|
207 |
+
"loss": 0.4017,
|
208 |
"step": 2000
|
209 |
},
|
210 |
{
|
211 |
+
"epoch": 3.0959752321981426,
|
212 |
+
"eval_loss": 0.5084750652313232,
|
213 |
+
"eval_runtime": 162.3398,
|
214 |
+
"eval_samples_per_second": 34.841,
|
215 |
+
"eval_steps_per_second": 4.355,
|
216 |
+
"eval_wer": 0.3697902457030059,
|
217 |
"step": 2000
|
218 |
},
|
219 |
{
|
220 |
+
"epoch": 3.2507739938080498,
|
221 |
+
"eval_loss": 0.6755269169807434,
|
222 |
+
"eval_runtime": 163.0508,
|
223 |
+
"eval_samples_per_second": 34.689,
|
224 |
+
"eval_steps_per_second": 4.336,
|
225 |
+
"eval_wer": 0.4530018776780986,
|
226 |
"step": 2100
|
227 |
},
|
228 |
{
|
229 |
+
"epoch": 3.405572755417957,
|
230 |
+
"eval_loss": 0.710012674331665,
|
231 |
+
"eval_runtime": 161.8251,
|
232 |
+
"eval_samples_per_second": 34.951,
|
233 |
+
"eval_steps_per_second": 4.369,
|
234 |
+
"eval_wer": 0.5108247339956027,
|
235 |
"step": 2200
|
236 |
},
|
237 |
{
|
238 |
+
"epoch": 3.560371517027864,
|
239 |
+
"eval_loss": 0.8310704231262207,
|
240 |
+
"eval_runtime": 162.1264,
|
241 |
+
"eval_samples_per_second": 34.886,
|
242 |
+
"eval_steps_per_second": 4.361,
|
243 |
+
"eval_wer": 0.5642984384779574,
|
244 |
"step": 2300
|
245 |
},
|
246 |
{
|
247 |
+
"epoch": 3.715170278637771,
|
248 |
+
"eval_loss": 0.7031980156898499,
|
249 |
+
"eval_runtime": 166.0393,
|
250 |
+
"eval_samples_per_second": 34.064,
|
251 |
+
"eval_steps_per_second": 4.258,
|
252 |
+
"eval_wer": 0.5028807112708832,
|
253 |
"step": 2400
|
254 |
},
|
255 |
{
|
256 |
+
"epoch": 3.8699690402476783,
|
257 |
+
"grad_norm": 4.817399024963379,
|
258 |
+
"learning_rate": 0.00016766666666666666,
|
259 |
+
"loss": 0.6839,
|
260 |
"step": 2500
|
261 |
},
|
262 |
{
|
263 |
+
"epoch": 3.8699690402476783,
|
264 |
+
"eval_loss": 0.7070674896240234,
|
265 |
+
"eval_runtime": 164.0683,
|
266 |
+
"eval_samples_per_second": 34.473,
|
267 |
+
"eval_steps_per_second": 4.309,
|
268 |
+
"eval_wer": 0.5006660140264159,
|
269 |
"step": 2500
|
270 |
},
|
271 |
{
|
272 |
+
"epoch": 4.024767801857585,
|
273 |
+
"eval_loss": 0.8223607540130615,
|
274 |
+
"eval_runtime": 164.3114,
|
275 |
+
"eval_samples_per_second": 34.422,
|
276 |
+
"eval_steps_per_second": 4.303,
|
277 |
+
"eval_wer": 0.5069409895524065,
|
278 |
"step": 2600
|
279 |
},
|
280 |
{
|
281 |
+
"epoch": 4.179566563467493,
|
282 |
+
"eval_loss": 0.8343736529350281,
|
283 |
+
"eval_runtime": 163.8342,
|
284 |
+
"eval_samples_per_second": 34.523,
|
285 |
+
"eval_steps_per_second": 4.315,
|
286 |
+
"eval_wer": 0.5162491373914718,
|
287 |
"step": 2700
|
288 |
},
|
289 |
{
|
290 |
+
"epoch": 4.3343653250774,
|
291 |
+
"eval_loss": 0.9089197516441345,
|
292 |
+
"eval_runtime": 162.2031,
|
293 |
+
"eval_samples_per_second": 34.87,
|
294 |
+
"eval_steps_per_second": 4.359,
|
295 |
+
"eval_wer": 0.5620195471104621,
|
296 |
"step": 2800
|
297 |
},
|
298 |
{
|
299 |
+
"epoch": 4.489164086687307,
|
300 |
+
"eval_loss": 0.9664539098739624,
|
301 |
+
"eval_runtime": 163.3849,
|
302 |
+
"eval_samples_per_second": 34.618,
|
303 |
+
"eval_steps_per_second": 4.327,
|
304 |
+
"eval_wer": 0.5640095649243312,
|
305 |
"step": 2900
|
306 |
},
|
307 |
{
|
308 |
+
"epoch": 4.643962848297214,
|
309 |
+
"grad_norm": 7.185929775238037,
|
310 |
+
"learning_rate": 0.00013446666666666666,
|
311 |
+
"loss": 0.8292,
|
312 |
"step": 3000
|
313 |
},
|
314 |
{
|
315 |
+
"epoch": 4.643962848297214,
|
316 |
+
"eval_loss": 0.9127740859985352,
|
317 |
+
"eval_runtime": 162.7094,
|
318 |
+
"eval_samples_per_second": 34.761,
|
319 |
+
"eval_steps_per_second": 4.345,
|
320 |
+
"eval_wer": 0.5414774277414902,
|
321 |
"step": 3000
|
322 |
},
|
323 |
{
|
324 |
+
"epoch": 4.798761609907121,
|
325 |
+
"eval_loss": 1.1924536228179932,
|
326 |
+
"eval_runtime": 161.8545,
|
327 |
+
"eval_samples_per_second": 34.945,
|
328 |
+
"eval_steps_per_second": 4.368,
|
329 |
+
"eval_wer": 0.5938598321323683,
|
330 |
"step": 3100
|
331 |
},
|
332 |
{
|
333 |
+
"epoch": 4.953560371517028,
|
334 |
+
"eval_loss": 1.4327375888824463,
|
335 |
+
"eval_runtime": 167.5589,
|
336 |
+
"eval_samples_per_second": 33.755,
|
337 |
+
"eval_steps_per_second": 4.219,
|
338 |
+
"eval_wer": 0.6999406204361991,
|
339 |
"step": 3200
|
340 |
},
|
341 |
{
|
342 |
+
"epoch": 5.108359133126935,
|
343 |
+
"eval_loss": 1.2741221189498901,
|
344 |
+
"eval_runtime": 166.0465,
|
345 |
+
"eval_samples_per_second": 34.063,
|
346 |
+
"eval_steps_per_second": 4.258,
|
347 |
+
"eval_wer": 0.7826707964885815,
|
348 |
"step": 3300
|
349 |
},
|
350 |
{
|
351 |
+
"epoch": 5.2631578947368425,
|
352 |
+
"eval_loss": 1.9348175525665283,
|
353 |
+
"eval_runtime": 163.4154,
|
354 |
+
"eval_samples_per_second": 34.611,
|
355 |
+
"eval_steps_per_second": 4.326,
|
356 |
+
"eval_wer": 0.8741795188650479,
|
357 |
"step": 3400
|
358 |
},
|
359 |
{
|
360 |
+
"epoch": 5.41795665634675,
|
361 |
+
"grad_norm": 0.6057500839233398,
|
362 |
+
"learning_rate": 0.00010126666666666666,
|
363 |
+
"loss": 1.4131,
|
364 |
"step": 3500
|
365 |
},
|
366 |
{
|
367 |
+
"epoch": 5.41795665634675,
|
368 |
+
"eval_loss": 1.9216177463531494,
|
369 |
+
"eval_runtime": 161.6687,
|
370 |
+
"eval_samples_per_second": 34.985,
|
371 |
+
"eval_steps_per_second": 4.373,
|
372 |
+
"eval_wer": 0.9870167386175795,
|
373 |
"step": 3500
|
374 |
},
|
375 |
{
|
376 |
+
"epoch": 5.572755417956657,
|
377 |
+
"eval_loss": 1.8565247058868408,
|
378 |
+
"eval_runtime": 164.0672,
|
379 |
+
"eval_samples_per_second": 34.474,
|
380 |
+
"eval_steps_per_second": 4.309,
|
381 |
+
"eval_wer": 0.9367045946943557,
|
382 |
"step": 3600
|
383 |
},
|
384 |
{
|
385 |
+
"epoch": 5.727554179566564,
|
386 |
+
"eval_loss": 1.7827845811843872,
|
387 |
+
"eval_runtime": 163.0587,
|
388 |
+
"eval_samples_per_second": 34.687,
|
389 |
+
"eval_steps_per_second": 4.336,
|
390 |
+
"eval_wer": 0.8240278602493941,
|
391 |
"step": 3700
|
392 |
},
|
393 |
{
|
394 |
+
"epoch": 5.882352941176471,
|
395 |
+
"eval_loss": 1.6846531629562378,
|
396 |
+
"eval_runtime": 162.8654,
|
397 |
+
"eval_samples_per_second": 34.728,
|
398 |
+
"eval_steps_per_second": 4.341,
|
399 |
+
"eval_wer": 0.8059090690247308,
|
400 |
"step": 3800
|
401 |
},
|
402 |
{
|
403 |
+
"epoch": 6.037151702786378,
|
404 |
+
"eval_loss": 1.6440324783325195,
|
405 |
+
"eval_runtime": 163.0292,
|
406 |
+
"eval_samples_per_second": 34.693,
|
407 |
+
"eval_steps_per_second": 4.337,
|
408 |
+
"eval_wer": 0.7983823080996935,
|
409 |
"step": 3900
|
410 |
},
|
411 |
{
|
412 |
+
"epoch": 6.191950464396285,
|
413 |
+
"grad_norm": 0.5905019640922546,
|
414 |
+
"learning_rate": 6.806666666666666e-05,
|
415 |
+
"loss": 1.7728,
|
416 |
"step": 4000
|
417 |
},
|
418 |
{
|
419 |
+
"epoch": 6.191950464396285,
|
420 |
+
"eval_loss": 1.6765395402908325,
|
421 |
+
"eval_runtime": 166.4929,
|
422 |
+
"eval_samples_per_second": 33.971,
|
423 |
+
"eval_steps_per_second": 4.246,
|
424 |
+
"eval_wer": 0.8053313219174785,
|
425 |
"step": 4000
|
426 |
},
|
427 |
{
|
428 |
+
"epoch": 6.346749226006192,
|
429 |
+
"eval_loss": 1.6733070611953735,
|
430 |
+
"eval_runtime": 163.1125,
|
431 |
+
"eval_samples_per_second": 34.675,
|
432 |
+
"eval_steps_per_second": 4.334,
|
433 |
+
"eval_wer": 0.8024265378504598,
|
434 |
"step": 4100
|
435 |
},
|
436 |
{
|
437 |
+
"epoch": 6.5015479876160995,
|
438 |
+
"eval_loss": 1.6601324081420898,
|
439 |
+
"eval_runtime": 163.6995,
|
440 |
+
"eval_samples_per_second": 34.551,
|
441 |
+
"eval_steps_per_second": 4.319,
|
442 |
+
"eval_wer": 0.7899568294522636,
|
443 |
"step": 4200
|
444 |
},
|
445 |
{
|
446 |
+
"epoch": 6.656346749226007,
|
447 |
+
"eval_loss": 1.6604827642440796,
|
448 |
+
"eval_runtime": 162.8495,
|
449 |
+
"eval_samples_per_second": 34.731,
|
450 |
+
"eval_steps_per_second": 4.341,
|
451 |
+
"eval_wer": 0.7972910080082168,
|
452 |
"step": 4300
|
453 |
},
|
454 |
{
|
455 |
+
"epoch": 6.811145510835914,
|
456 |
+
"eval_loss": 1.6598834991455078,
|
457 |
+
"eval_runtime": 163.8109,
|
458 |
+
"eval_samples_per_second": 34.528,
|
459 |
+
"eval_steps_per_second": 4.316,
|
460 |
+
"eval_wer": 0.7804560992441142,
|
461 |
"step": 4400
|
462 |
},
|
463 |
{
|
464 |
+
"epoch": 6.965944272445821,
|
465 |
+
"grad_norm": 0.9840993881225586,
|
466 |
+
"learning_rate": 3.493333333333333e-05,
|
467 |
+
"loss": 1.6777,
|
468 |
"step": 4500
|
469 |
},
|
470 |
{
|
471 |
+
"epoch": 6.965944272445821,
|
472 |
+
"eval_loss": 1.635949730873108,
|
473 |
+
"eval_runtime": 163.8566,
|
474 |
+
"eval_samples_per_second": 34.518,
|
475 |
+
"eval_steps_per_second": 4.315,
|
476 |
+
"eval_wer": 0.7693184188987499,
|
477 |
"step": 4500
|
478 |
},
|
479 |
{
|
480 |
+
"epoch": 7.120743034055727,
|
481 |
+
"eval_loss": 1.6399564743041992,
|
482 |
+
"eval_runtime": 164.6612,
|
483 |
+
"eval_samples_per_second": 34.349,
|
484 |
+
"eval_steps_per_second": 4.294,
|
485 |
+
"eval_wer": 0.7651137038404134,
|
486 |
"step": 4600
|
487 |
},
|
488 |
{
|
489 |
+
"epoch": 7.275541795665634,
|
490 |
+
"eval_loss": 1.6759086847305298,
|
491 |
+
"eval_runtime": 161.5503,
|
492 |
+
"eval_samples_per_second": 35.011,
|
493 |
+
"eval_steps_per_second": 4.376,
|
494 |
+
"eval_wer": 0.7672481584310956,
|
495 |
"step": 4700
|
496 |
},
|
497 |
{
|
498 |
+
"epoch": 7.430340557275541,
|
499 |
+
"eval_loss": 1.6848710775375366,
|
500 |
+
"eval_runtime": 162.0136,
|
501 |
+
"eval_samples_per_second": 34.911,
|
502 |
+
"eval_steps_per_second": 4.364,
|
503 |
+
"eval_wer": 0.7686283320761984,
|
504 |
"step": 4800
|
505 |
},
|
506 |
{
|
507 |
+
"epoch": 7.585139318885449,
|
508 |
+
"eval_loss": 1.6857717037200928,
|
509 |
+
"eval_runtime": 161.3723,
|
510 |
+
"eval_samples_per_second": 35.049,
|
511 |
+
"eval_steps_per_second": 4.381,
|
512 |
+
"eval_wer": 0.7689974482836096,
|
513 |
"step": 4900
|
514 |
},
|
515 |
{
|
516 |
+
"epoch": 7.739938080495356,
|
517 |
+
"grad_norm": 1.753746747970581,
|
518 |
+
"learning_rate": 1.8e-06,
|
519 |
+
"loss": 1.683,
|
520 |
"step": 5000
|
521 |
},
|
522 |
{
|
523 |
+
"epoch": 7.739938080495356,
|
524 |
+
"eval_loss": 1.6955511569976807,
|
525 |
+
"eval_runtime": 160.8529,
|
526 |
+
"eval_samples_per_second": 35.163,
|
527 |
+
"eval_steps_per_second": 4.395,
|
528 |
+
"eval_wer": 0.7701529424981143,
|
529 |
"step": 5000
|
530 |
},
|
531 |
{
|
532 |
+
"epoch": 7.739938080495356,
|
533 |
+
"step": 5000,
|
534 |
+
"total_flos": 4.528229347801944e+19,
|
535 |
+
"train_loss": 1.3220068420410156,
|
536 |
+
"train_runtime": 20543.6538,
|
537 |
+
"train_samples_per_second": 15.577,
|
538 |
+
"train_steps_per_second": 0.243
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
539 |
}
|
540 |
],
|
541 |
"logging_steps": 500,
|
542 |
+
"max_steps": 5000,
|
543 |
"num_input_tokens_seen": 0,
|
544 |
+
"num_train_epochs": 8,
|
545 |
"save_steps": 400,
|
546 |
+
"total_flos": 4.528229347801944e+19,
|
547 |
"train_batch_size": 64,
|
548 |
"trial_name": null,
|
549 |
"trial_params": null
|