Training in progress, step 10000
Browse files- last-checkpoint/optimizer.pt +1 -1
- last-checkpoint/pytorch_model.bin +1 -1
- last-checkpoint/rng_state.pth +2 -2
- last-checkpoint/scheduler.pt +1 -1
- last-checkpoint/trainer_state.json +103 -703
- last-checkpoint/training_args.bin +1 -1
- pytorch_model.bin +1 -1
- training_args.bin +1 -1
last-checkpoint/optimizer.pt
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 100170757
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:c205aa1e4f5bca6eeb78422e0866131d9cc58c5a15fab7af01cc2505a790df94
|
3 |
size 100170757
|
last-checkpoint/pytorch_model.bin
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 146774203
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:f48f41b7b489ac24545b446eb3e28cf98daf67da888804a7513071641e3cb0ab
|
3 |
size 146774203
|
last-checkpoint/rng_state.pth
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
-
size
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:8b635e00d41e3e6e46449e8609b1f91fbb1205e9587df7d15d6fa3ee2cfe466d
|
3 |
+
size 14503
|
last-checkpoint/scheduler.pt
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 246897640
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:97030c307e5479014866fb832070891e4b01f3e0d70ac54668f29d2602b645e8
|
3 |
size 246897640
|
last-checkpoint/trainer_state.json
CHANGED
@@ -1,916 +1,316 @@
|
|
1 |
{
|
2 |
"best_metric": null,
|
3 |
"best_model_checkpoint": null,
|
4 |
-
"epoch": 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": 0.0,
|
12 |
-
"learning_rate":
|
13 |
-
"loss":
|
14 |
"step": 200
|
15 |
},
|
16 |
{
|
17 |
"epoch": 0.0,
|
18 |
-
"learning_rate":
|
19 |
-
"loss":
|
20 |
"step": 400
|
21 |
},
|
22 |
{
|
23 |
"epoch": 0.0,
|
24 |
-
"learning_rate":
|
25 |
-
"loss":
|
26 |
"step": 600
|
27 |
},
|
28 |
{
|
29 |
"epoch": 0.0,
|
30 |
-
"learning_rate":
|
31 |
-
"loss":
|
32 |
"step": 800
|
33 |
},
|
34 |
{
|
35 |
"epoch": 0.0,
|
36 |
-
"learning_rate":
|
37 |
-
"loss":
|
38 |
"step": 1000
|
39 |
},
|
40 |
{
|
41 |
"epoch": 0.01,
|
42 |
-
"learning_rate":
|
43 |
-
"loss":
|
44 |
"step": 1200
|
45 |
},
|
46 |
{
|
47 |
"epoch": 0.01,
|
48 |
-
"learning_rate":
|
49 |
-
"loss":
|
50 |
"step": 1400
|
51 |
},
|
52 |
{
|
53 |
"epoch": 0.01,
|
54 |
-
"learning_rate":
|
55 |
-
"loss":
|
56 |
"step": 1600
|
57 |
},
|
58 |
{
|
59 |
"epoch": 0.01,
|
60 |
-
"learning_rate":
|
61 |
-
"loss":
|
62 |
"step": 1800
|
63 |
},
|
64 |
{
|
65 |
"epoch": 0.01,
|
66 |
-
"learning_rate":
|
67 |
-
"loss":
|
68 |
"step": 2000
|
69 |
},
|
70 |
{
|
71 |
"epoch": 0.01,
|
72 |
-
"learning_rate":
|
73 |
-
"loss":
|
74 |
"step": 2200
|
75 |
},
|
76 |
{
|
77 |
"epoch": 0.01,
|
78 |
-
"learning_rate":
|
79 |
-
"loss":
|
80 |
"step": 2400
|
81 |
},
|
82 |
{
|
83 |
"epoch": 0.01,
|
84 |
-
"learning_rate":
|
85 |
-
"loss":
|
86 |
"step": 2600
|
87 |
},
|
88 |
{
|
89 |
"epoch": 0.01,
|
90 |
-
"learning_rate":
|
91 |
-
"loss":
|
92 |
"step": 2800
|
93 |
},
|
94 |
{
|
95 |
"epoch": 0.01,
|
96 |
-
"learning_rate":
|
97 |
-
"loss":
|
98 |
"step": 3000
|
99 |
},
|
100 |
{
|
101 |
"epoch": 0.01,
|
102 |
-
"learning_rate":
|
103 |
-
"loss":
|
104 |
"step": 3200
|
105 |
},
|
106 |
{
|
107 |
"epoch": 0.01,
|
108 |
-
"learning_rate":
|
109 |
-
"loss":
|
110 |
"step": 3400
|
111 |
},
|
112 |
{
|
113 |
"epoch": 0.02,
|
114 |
-
"learning_rate":
|
115 |
-
"loss":
|
116 |
"step": 3600
|
117 |
},
|
118 |
{
|
119 |
"epoch": 0.02,
|
120 |
-
"learning_rate":
|
121 |
-
"loss":
|
122 |
"step": 3800
|
123 |
},
|
124 |
{
|
125 |
"epoch": 0.02,
|
126 |
-
"learning_rate":
|
127 |
-
"loss":
|
128 |
"step": 4000
|
129 |
},
|
130 |
{
|
131 |
"epoch": 0.02,
|
132 |
-
"learning_rate":
|
133 |
-
"loss":
|
134 |
"step": 4200
|
135 |
},
|
136 |
{
|
137 |
"epoch": 0.02,
|
138 |
-
"learning_rate":
|
139 |
-
"loss":
|
140 |
"step": 4400
|
141 |
},
|
142 |
{
|
143 |
"epoch": 0.02,
|
144 |
-
"learning_rate":
|
145 |
-
"loss":
|
146 |
"step": 4600
|
147 |
},
|
148 |
{
|
149 |
"epoch": 0.02,
|
150 |
-
"learning_rate":
|
151 |
-
"loss":
|
152 |
"step": 4800
|
153 |
},
|
154 |
{
|
155 |
"epoch": 0.02,
|
156 |
-
"learning_rate":
|
157 |
-
"loss":
|
158 |
"step": 5000
|
159 |
},
|
160 |
{
|
161 |
"epoch": 0.02,
|
162 |
-
"learning_rate":
|
163 |
-
"loss":
|
164 |
"step": 5200
|
165 |
},
|
166 |
{
|
167 |
"epoch": 0.02,
|
168 |
-
"learning_rate":
|
169 |
-
"loss":
|
170 |
"step": 5400
|
171 |
},
|
172 |
{
|
173 |
"epoch": 0.02,
|
174 |
-
"learning_rate":
|
175 |
-
"loss":
|
176 |
"step": 5600
|
177 |
},
|
178 |
{
|
179 |
"epoch": 0.02,
|
180 |
-
"learning_rate":
|
181 |
-
"loss":
|
182 |
"step": 5800
|
183 |
},
|
184 |
{
|
185 |
"epoch": 0.03,
|
186 |
-
"learning_rate":
|
187 |
-
"loss":
|
188 |
"step": 6000
|
189 |
},
|
190 |
{
|
191 |
"epoch": 0.03,
|
192 |
-
"learning_rate":
|
193 |
-
"loss":
|
194 |
"step": 6200
|
195 |
},
|
196 |
{
|
197 |
"epoch": 0.03,
|
198 |
-
"learning_rate":
|
199 |
-
"loss":
|
200 |
"step": 6400
|
201 |
},
|
202 |
{
|
203 |
"epoch": 0.03,
|
204 |
-
"learning_rate":
|
205 |
-
"loss":
|
206 |
"step": 6600
|
207 |
},
|
208 |
{
|
209 |
"epoch": 0.03,
|
210 |
-
"learning_rate":
|
211 |
-
"loss":
|
212 |
"step": 6800
|
213 |
},
|
214 |
{
|
215 |
"epoch": 0.03,
|
216 |
-
"learning_rate":
|
217 |
-
"loss":
|
218 |
"step": 7000
|
219 |
},
|
220 |
{
|
221 |
"epoch": 0.03,
|
222 |
-
"learning_rate":
|
223 |
-
"loss":
|
224 |
"step": 7200
|
225 |
},
|
226 |
{
|
227 |
"epoch": 0.03,
|
228 |
-
"learning_rate":
|
229 |
-
"loss":
|
230 |
"step": 7400
|
231 |
},
|
232 |
{
|
233 |
"epoch": 0.03,
|
234 |
-
"learning_rate":
|
235 |
-
"loss":
|
236 |
"step": 7600
|
237 |
},
|
238 |
{
|
239 |
"epoch": 0.03,
|
240 |
-
"learning_rate":
|
241 |
-
"loss":
|
242 |
"step": 7800
|
243 |
},
|
244 |
{
|
245 |
"epoch": 0.03,
|
246 |
-
"learning_rate":
|
247 |
-
"loss":
|
248 |
"step": 8000
|
249 |
},
|
250 |
{
|
251 |
"epoch": 0.04,
|
252 |
-
"learning_rate":
|
253 |
-
"loss":
|
254 |
"step": 8200
|
255 |
},
|
256 |
{
|
257 |
"epoch": 0.04,
|
258 |
-
"learning_rate":
|
259 |
-
"loss":
|
260 |
"step": 8400
|
261 |
},
|
262 |
{
|
263 |
"epoch": 0.04,
|
264 |
-
"learning_rate":
|
265 |
-
"loss":
|
266 |
"step": 8600
|
267 |
},
|
268 |
{
|
269 |
"epoch": 0.04,
|
270 |
-
"learning_rate":
|
271 |
-
"loss":
|
272 |
"step": 8800
|
273 |
},
|
274 |
{
|
275 |
"epoch": 0.04,
|
276 |
-
"learning_rate":
|
277 |
-
"loss":
|
278 |
"step": 9000
|
279 |
},
|
280 |
{
|
281 |
"epoch": 0.04,
|
282 |
-
"learning_rate":
|
283 |
-
"loss":
|
284 |
"step": 9200
|
285 |
},
|
286 |
{
|
287 |
"epoch": 0.04,
|
288 |
-
"learning_rate":
|
289 |
-
"loss":
|
290 |
"step": 9400
|
291 |
},
|
292 |
{
|
293 |
"epoch": 0.04,
|
294 |
-
"learning_rate":
|
295 |
-
"loss":
|
296 |
"step": 9600
|
297 |
},
|
298 |
{
|
299 |
"epoch": 0.04,
|
300 |
-
"learning_rate":
|
301 |
-
"loss":
|
302 |
"step": 9800
|
303 |
},
|
304 |
{
|
305 |
"epoch": 0.04,
|
306 |
-
"learning_rate":
|
307 |
-
"loss":
|
308 |
"step": 10000
|
309 |
-
},
|
310 |
-
{
|
311 |
-
"epoch": 0.04,
|
312 |
-
"learning_rate": 0.0005091144561900837,
|
313 |
-
"loss": 5.0155,
|
314 |
-
"step": 10200
|
315 |
-
},
|
316 |
-
{
|
317 |
-
"epoch": 0.04,
|
318 |
-
"learning_rate": 0.0005134183095340927,
|
319 |
-
"loss": 4.9817,
|
320 |
-
"step": 10400
|
321 |
-
},
|
322 |
-
{
|
323 |
-
"epoch": 0.05,
|
324 |
-
"learning_rate": 0.0005178043944449977,
|
325 |
-
"loss": 4.9742,
|
326 |
-
"step": 10600
|
327 |
-
},
|
328 |
-
{
|
329 |
-
"epoch": 0.05,
|
330 |
-
"learning_rate": 0.0005222726339638023,
|
331 |
-
"loss": 4.9299,
|
332 |
-
"step": 10800
|
333 |
-
},
|
334 |
-
{
|
335 |
-
"epoch": 0.05,
|
336 |
-
"learning_rate": 0.0005268229496900086,
|
337 |
-
"loss": 4.9208,
|
338 |
-
"step": 11000
|
339 |
-
},
|
340 |
-
{
|
341 |
-
"epoch": 0.05,
|
342 |
-
"learning_rate": 0.0005314552617829947,
|
343 |
-
"loss": 4.8617,
|
344 |
-
"step": 11200
|
345 |
-
},
|
346 |
-
{
|
347 |
-
"epoch": 0.05,
|
348 |
-
"learning_rate": 0.0005361694889634196,
|
349 |
-
"loss": 4.7952,
|
350 |
-
"step": 11400
|
351 |
-
},
|
352 |
-
{
|
353 |
-
"epoch": 0.05,
|
354 |
-
"learning_rate": 0.0005409655485146408,
|
355 |
-
"loss": 4.7641,
|
356 |
-
"step": 11600
|
357 |
-
},
|
358 |
-
{
|
359 |
-
"epoch": 0.05,
|
360 |
-
"learning_rate": 0.0005458433562841782,
|
361 |
-
"loss": 4.7361,
|
362 |
-
"step": 11800
|
363 |
-
},
|
364 |
-
{
|
365 |
-
"epoch": 0.05,
|
366 |
-
"learning_rate": 0.0005508028266851747,
|
367 |
-
"loss": 4.7023,
|
368 |
-
"step": 12000
|
369 |
-
},
|
370 |
-
{
|
371 |
-
"epoch": 0.05,
|
372 |
-
"learning_rate": 0.000555843872697916,
|
373 |
-
"loss": 4.6561,
|
374 |
-
"step": 12200
|
375 |
-
},
|
376 |
-
{
|
377 |
-
"epoch": 0.05,
|
378 |
-
"learning_rate": 0.0005609664058713396,
|
379 |
-
"loss": 4.63,
|
380 |
-
"step": 12400
|
381 |
-
},
|
382 |
-
{
|
383 |
-
"epoch": 0.05,
|
384 |
-
"learning_rate": 0.0005661703363245996,
|
385 |
-
"loss": 4.6307,
|
386 |
-
"step": 12600
|
387 |
-
},
|
388 |
-
{
|
389 |
-
"epoch": 0.06,
|
390 |
-
"learning_rate": 0.0005714555727486404,
|
391 |
-
"loss": 4.5881,
|
392 |
-
"step": 12800
|
393 |
-
},
|
394 |
-
{
|
395 |
-
"epoch": 0.06,
|
396 |
-
"learning_rate": 0.0005768220224077955,
|
397 |
-
"loss": 4.5489,
|
398 |
-
"step": 13000
|
399 |
-
},
|
400 |
-
{
|
401 |
-
"epoch": 0.06,
|
402 |
-
"learning_rate": 0.0005822695911414169,
|
403 |
-
"loss": 4.5521,
|
404 |
-
"step": 13200
|
405 |
-
},
|
406 |
-
{
|
407 |
-
"epoch": 0.06,
|
408 |
-
"learning_rate": 0.0005877981833655298,
|
409 |
-
"loss": 4.5165,
|
410 |
-
"step": 13400
|
411 |
-
},
|
412 |
-
{
|
413 |
-
"epoch": 0.06,
|
414 |
-
"learning_rate": 0.0005934077020745051,
|
415 |
-
"loss": 4.505,
|
416 |
-
"step": 13600
|
417 |
-
},
|
418 |
-
{
|
419 |
-
"epoch": 0.06,
|
420 |
-
"learning_rate": 0.0005990980488427659,
|
421 |
-
"loss": 4.4863,
|
422 |
-
"step": 13800
|
423 |
-
},
|
424 |
-
{
|
425 |
-
"epoch": 0.06,
|
426 |
-
"learning_rate": 0.000604869123826509,
|
427 |
-
"loss": 4.5071,
|
428 |
-
"step": 14000
|
429 |
-
},
|
430 |
-
{
|
431 |
-
"epoch": 0.06,
|
432 |
-
"learning_rate": 0.0006107208257654633,
|
433 |
-
"loss": 4.4501,
|
434 |
-
"step": 14200
|
435 |
-
},
|
436 |
-
{
|
437 |
-
"epoch": 0.06,
|
438 |
-
"learning_rate": 0.0006166530519846631,
|
439 |
-
"loss": 4.4623,
|
440 |
-
"step": 14400
|
441 |
-
},
|
442 |
-
{
|
443 |
-
"epoch": 0.06,
|
444 |
-
"learning_rate": 0.0006226656983962468,
|
445 |
-
"loss": 4.4336,
|
446 |
-
"step": 14600
|
447 |
-
},
|
448 |
-
{
|
449 |
-
"epoch": 0.06,
|
450 |
-
"learning_rate": 0.0006287586595012887,
|
451 |
-
"loss": 4.4335,
|
452 |
-
"step": 14800
|
453 |
-
},
|
454 |
-
{
|
455 |
-
"epoch": 0.06,
|
456 |
-
"learning_rate": 0.000634931828391647,
|
457 |
-
"loss": 4.4142,
|
458 |
-
"step": 15000
|
459 |
-
},
|
460 |
-
{
|
461 |
-
"epoch": 0.07,
|
462 |
-
"learning_rate": 0.0006411850967518416,
|
463 |
-
"loss": 4.4145,
|
464 |
-
"step": 15200
|
465 |
-
},
|
466 |
-
{
|
467 |
-
"epoch": 0.07,
|
468 |
-
"learning_rate": 0.0006475183548609511,
|
469 |
-
"loss": 4.3842,
|
470 |
-
"step": 15400
|
471 |
-
},
|
472 |
-
{
|
473 |
-
"epoch": 0.07,
|
474 |
-
"learning_rate": 0.0006539314915945428,
|
475 |
-
"loss": 4.3748,
|
476 |
-
"step": 15600
|
477 |
-
},
|
478 |
-
{
|
479 |
-
"epoch": 0.07,
|
480 |
-
"learning_rate": 0.0006604243944266178,
|
481 |
-
"loss": 4.3815,
|
482 |
-
"step": 15800
|
483 |
-
},
|
484 |
-
{
|
485 |
-
"epoch": 0.07,
|
486 |
-
"learning_rate": 0.0006669969494315867,
|
487 |
-
"loss": 4.352,
|
488 |
-
"step": 16000
|
489 |
-
},
|
490 |
-
{
|
491 |
-
"epoch": 0.07,
|
492 |
-
"learning_rate": 0.0006736490412862749,
|
493 |
-
"loss": 4.3575,
|
494 |
-
"step": 16200
|
495 |
-
},
|
496 |
-
{
|
497 |
-
"epoch": 0.07,
|
498 |
-
"learning_rate": 0.000680380553271933,
|
499 |
-
"loss": 4.3416,
|
500 |
-
"step": 16400
|
501 |
-
},
|
502 |
-
{
|
503 |
-
"epoch": 0.07,
|
504 |
-
"learning_rate": 0.0006871913672762998,
|
505 |
-
"loss": 4.341,
|
506 |
-
"step": 16600
|
507 |
-
},
|
508 |
-
{
|
509 |
-
"epoch": 0.07,
|
510 |
-
"learning_rate": 0.0006940813637956594,
|
511 |
-
"loss": 4.3183,
|
512 |
-
"step": 16800
|
513 |
-
},
|
514 |
-
{
|
515 |
-
"epoch": 0.07,
|
516 |
-
"learning_rate": 0.0007010504219369541,
|
517 |
-
"loss": 4.3145,
|
518 |
-
"step": 17000
|
519 |
-
},
|
520 |
-
{
|
521 |
-
"epoch": 0.07,
|
522 |
-
"learning_rate": 0.0007080984194198885,
|
523 |
-
"loss": 4.3065,
|
524 |
-
"step": 17200
|
525 |
-
},
|
526 |
-
{
|
527 |
-
"epoch": 0.07,
|
528 |
-
"learning_rate": 0.0007152252325790948,
|
529 |
-
"loss": 4.2805,
|
530 |
-
"step": 17400
|
531 |
-
},
|
532 |
-
{
|
533 |
-
"epoch": 0.08,
|
534 |
-
"learning_rate": 0.0007224307363662818,
|
535 |
-
"loss": 4.2804,
|
536 |
-
"step": 17600
|
537 |
-
},
|
538 |
-
{
|
539 |
-
"epoch": 0.08,
|
540 |
-
"learning_rate": 0.0007297148043524434,
|
541 |
-
"loss": 4.2996,
|
542 |
-
"step": 17800
|
543 |
-
},
|
544 |
-
{
|
545 |
-
"epoch": 0.08,
|
546 |
-
"learning_rate": 0.0007370773087300737,
|
547 |
-
"loss": 4.2743,
|
548 |
-
"step": 18000
|
549 |
-
},
|
550 |
-
{
|
551 |
-
"epoch": 0.08,
|
552 |
-
"learning_rate": 0.0007445181203154048,
|
553 |
-
"loss": 4.2621,
|
554 |
-
"step": 18200
|
555 |
-
},
|
556 |
-
{
|
557 |
-
"epoch": 0.08,
|
558 |
-
"learning_rate": 0.0007520371085506811,
|
559 |
-
"loss": 4.2548,
|
560 |
-
"step": 18400
|
561 |
-
},
|
562 |
-
{
|
563 |
-
"epoch": 0.08,
|
564 |
-
"learning_rate": 0.0007596341415064441,
|
565 |
-
"loss": 4.2643,
|
566 |
-
"step": 18600
|
567 |
-
},
|
568 |
-
{
|
569 |
-
"epoch": 0.08,
|
570 |
-
"learning_rate": 0.0007673090858838494,
|
571 |
-
"loss": 4.266,
|
572 |
-
"step": 18800
|
573 |
-
},
|
574 |
-
{
|
575 |
-
"epoch": 0.08,
|
576 |
-
"learning_rate": 0.0007750618070170041,
|
577 |
-
"loss": 4.2503,
|
578 |
-
"step": 19000
|
579 |
-
},
|
580 |
-
{
|
581 |
-
"epoch": 0.08,
|
582 |
-
"learning_rate": 0.0007828921688753324,
|
583 |
-
"loss": 4.2093,
|
584 |
-
"step": 19200
|
585 |
-
},
|
586 |
-
{
|
587 |
-
"epoch": 0.08,
|
588 |
-
"learning_rate": 0.0007908000340659631,
|
589 |
-
"loss": 4.2449,
|
590 |
-
"step": 19400
|
591 |
-
},
|
592 |
-
{
|
593 |
-
"epoch": 0.08,
|
594 |
-
"learning_rate": 0.0007987852638361333,
|
595 |
-
"loss": 4.2158,
|
596 |
-
"step": 19600
|
597 |
-
},
|
598 |
-
{
|
599 |
-
"epoch": 0.09,
|
600 |
-
"learning_rate": 0.0008068477180756314,
|
601 |
-
"loss": 4.202,
|
602 |
-
"step": 19800
|
603 |
-
},
|
604 |
-
{
|
605 |
-
"epoch": 0.09,
|
606 |
-
"learning_rate": 0.0008149872553192515,
|
607 |
-
"loss": 4.2065,
|
608 |
-
"step": 20000
|
609 |
-
},
|
610 |
-
{
|
611 |
-
"epoch": 0.09,
|
612 |
-
"learning_rate": 0.0008232037327492777,
|
613 |
-
"loss": 4.1773,
|
614 |
-
"step": 20200
|
615 |
-
},
|
616 |
-
{
|
617 |
-
"epoch": 0.09,
|
618 |
-
"learning_rate": 0.0008314970061979818,
|
619 |
-
"loss": 4.1904,
|
620 |
-
"step": 20400
|
621 |
-
},
|
622 |
-
{
|
623 |
-
"epoch": 0.09,
|
624 |
-
"learning_rate": 0.0008398669301501703,
|
625 |
-
"loss": 4.1868,
|
626 |
-
"step": 20600
|
627 |
-
},
|
628 |
-
{
|
629 |
-
"epoch": 0.09,
|
630 |
-
"learning_rate": 0.0008483133577457148,
|
631 |
-
"loss": 4.2006,
|
632 |
-
"step": 20800
|
633 |
-
},
|
634 |
-
{
|
635 |
-
"epoch": 0.09,
|
636 |
-
"learning_rate": 0.0008568361407821495,
|
637 |
-
"loss": 4.1467,
|
638 |
-
"step": 21000
|
639 |
-
},
|
640 |
-
{
|
641 |
-
"epoch": 0.09,
|
642 |
-
"learning_rate": 0.0008654351297172607,
|
643 |
-
"loss": 4.1585,
|
644 |
-
"step": 21200
|
645 |
-
},
|
646 |
-
{
|
647 |
-
"epoch": 0.09,
|
648 |
-
"learning_rate": 0.0008741101736717116,
|
649 |
-
"loss": 4.1547,
|
650 |
-
"step": 21400
|
651 |
-
},
|
652 |
-
{
|
653 |
-
"epoch": 0.09,
|
654 |
-
"learning_rate": 0.0008828611204316911,
|
655 |
-
"loss": 4.1557,
|
656 |
-
"step": 21600
|
657 |
-
},
|
658 |
-
{
|
659 |
-
"epoch": 0.09,
|
660 |
-
"learning_rate": 0.0008916878164515838,
|
661 |
-
"loss": 4.1496,
|
662 |
-
"step": 21800
|
663 |
-
},
|
664 |
-
{
|
665 |
-
"epoch": 0.09,
|
666 |
-
"learning_rate": 0.0009005901068566691,
|
667 |
-
"loss": 4.1434,
|
668 |
-
"step": 22000
|
669 |
-
},
|
670 |
-
{
|
671 |
-
"epoch": 0.1,
|
672 |
-
"learning_rate": 0.0009095678354458306,
|
673 |
-
"loss": 4.1173,
|
674 |
-
"step": 22200
|
675 |
-
},
|
676 |
-
{
|
677 |
-
"epoch": 0.1,
|
678 |
-
"learning_rate": 0.0009186208446943008,
|
679 |
-
"loss": 4.1364,
|
680 |
-
"step": 22400
|
681 |
-
},
|
682 |
-
{
|
683 |
-
"epoch": 0.1,
|
684 |
-
"learning_rate": 0.0009277489757564244,
|
685 |
-
"loss": 4.1445,
|
686 |
-
"step": 22600
|
687 |
-
},
|
688 |
-
{
|
689 |
-
"epoch": 0.1,
|
690 |
-
"learning_rate": 0.0009369520684684475,
|
691 |
-
"loss": 4.1156,
|
692 |
-
"step": 22800
|
693 |
-
},
|
694 |
-
{
|
695 |
-
"epoch": 0.1,
|
696 |
-
"learning_rate": 0.0009462299613513248,
|
697 |
-
"loss": 4.1033,
|
698 |
-
"step": 23000
|
699 |
-
},
|
700 |
-
{
|
701 |
-
"epoch": 0.1,
|
702 |
-
"learning_rate": 0.0009555824916135536,
|
703 |
-
"loss": 4.1187,
|
704 |
-
"step": 23200
|
705 |
-
},
|
706 |
-
{
|
707 |
-
"epoch": 0.1,
|
708 |
-
"learning_rate": 0.0009650094951540386,
|
709 |
-
"loss": 4.0823,
|
710 |
-
"step": 23400
|
711 |
-
},
|
712 |
-
{
|
713 |
-
"epoch": 0.1,
|
714 |
-
"learning_rate": 0.0009745108065649499,
|
715 |
-
"loss": 4.0624,
|
716 |
-
"step": 23600
|
717 |
-
},
|
718 |
-
{
|
719 |
-
"epoch": 0.1,
|
720 |
-
"learning_rate": 0.0009840862591346498,
|
721 |
-
"loss": 4.0845,
|
722 |
-
"step": 23800
|
723 |
-
},
|
724 |
-
{
|
725 |
-
"epoch": 0.1,
|
726 |
-
"learning_rate": 0.0009937356848506058,
|
727 |
-
"loss": 4.0483,
|
728 |
-
"step": 24000
|
729 |
-
},
|
730 |
-
{
|
731 |
-
"epoch": 0.1,
|
732 |
-
"learning_rate": 0.001003458914402332,
|
733 |
-
"loss": 4.0512,
|
734 |
-
"step": 24200
|
735 |
-
},
|
736 |
-
{
|
737 |
-
"epoch": 0.1,
|
738 |
-
"learning_rate": 0.0010132557771843787,
|
739 |
-
"loss": 4.0606,
|
740 |
-
"step": 24400
|
741 |
-
},
|
742 |
-
{
|
743 |
-
"epoch": 0.11,
|
744 |
-
"learning_rate": 0.0010231261012993067,
|
745 |
-
"loss": 4.046,
|
746 |
-
"step": 24600
|
747 |
-
},
|
748 |
-
{
|
749 |
-
"epoch": 0.11,
|
750 |
-
"learning_rate": 0.0010330697135607168,
|
751 |
-
"loss": 4.0315,
|
752 |
-
"step": 24800
|
753 |
-
},
|
754 |
-
{
|
755 |
-
"epoch": 0.11,
|
756 |
-
"learning_rate": 0.00104308643949628,
|
757 |
-
"loss": 4.0179,
|
758 |
-
"step": 25000
|
759 |
-
},
|
760 |
-
{
|
761 |
-
"epoch": 0.11,
|
762 |
-
"learning_rate": 0.001053176103350803,
|
763 |
-
"loss": 4.0351,
|
764 |
-
"step": 25200
|
765 |
-
},
|
766 |
-
{
|
767 |
-
"epoch": 0.11,
|
768 |
-
"learning_rate": 0.0010633385280893123,
|
769 |
-
"loss": 4.02,
|
770 |
-
"step": 25400
|
771 |
-
},
|
772 |
-
{
|
773 |
-
"epoch": 0.11,
|
774 |
-
"learning_rate": 0.0010735735354001595,
|
775 |
-
"loss": 4.0201,
|
776 |
-
"step": 25600
|
777 |
-
},
|
778 |
-
{
|
779 |
-
"epoch": 0.11,
|
780 |
-
"learning_rate": 0.0010838809456981471,
|
781 |
-
"loss": 4.0044,
|
782 |
-
"step": 25800
|
783 |
-
},
|
784 |
-
{
|
785 |
-
"epoch": 0.11,
|
786 |
-
"learning_rate": 0.001094260578127686,
|
787 |
-
"loss": 3.9914,
|
788 |
-
"step": 26000
|
789 |
-
},
|
790 |
-
{
|
791 |
-
"epoch": 0.11,
|
792 |
-
"learning_rate": 0.0011047122505659646,
|
793 |
-
"loss": 3.9991,
|
794 |
-
"step": 26200
|
795 |
-
},
|
796 |
-
{
|
797 |
-
"epoch": 0.11,
|
798 |
-
"learning_rate": 0.0011152357796261423,
|
799 |
-
"loss": 4.0109,
|
800 |
-
"step": 26400
|
801 |
-
},
|
802 |
-
{
|
803 |
-
"epoch": 0.11,
|
804 |
-
"learning_rate": 0.0011258309806605731,
|
805 |
-
"loss": 4.0405,
|
806 |
-
"step": 26600
|
807 |
-
},
|
808 |
-
{
|
809 |
-
"epoch": 0.12,
|
810 |
-
"learning_rate": 0.0011364976677640387,
|
811 |
-
"loss": 4.0349,
|
812 |
-
"step": 26800
|
813 |
-
},
|
814 |
-
{
|
815 |
-
"epoch": 0.12,
|
816 |
-
"learning_rate": 0.0011472356537770186,
|
817 |
-
"loss": 4.0312,
|
818 |
-
"step": 27000
|
819 |
-
},
|
820 |
-
{
|
821 |
-
"epoch": 0.12,
|
822 |
-
"learning_rate": 0.0011580447502889633,
|
823 |
-
"loss": 4.0185,
|
824 |
-
"step": 27200
|
825 |
-
},
|
826 |
-
{
|
827 |
-
"epoch": 0.12,
|
828 |
-
"learning_rate": 0.0011689247676416152,
|
829 |
-
"loss": 4.011,
|
830 |
-
"step": 27400
|
831 |
-
},
|
832 |
-
{
|
833 |
-
"epoch": 0.12,
|
834 |
-
"learning_rate": 0.0011798755149323176,
|
835 |
-
"loss": 3.9898,
|
836 |
-
"step": 27600
|
837 |
-
},
|
838 |
-
{
|
839 |
-
"epoch": 0.12,
|
840 |
-
"learning_rate": 0.001190896800017379,
|
841 |
-
"loss": 3.981,
|
842 |
-
"step": 27800
|
843 |
-
},
|
844 |
-
{
|
845 |
-
"epoch": 0.12,
|
846 |
-
"learning_rate": 0.0012019884295154416,
|
847 |
-
"loss": 3.949,
|
848 |
-
"step": 28000
|
849 |
-
},
|
850 |
-
{
|
851 |
-
"epoch": 0.12,
|
852 |
-
"learning_rate": 0.0012131502088108658,
|
853 |
-
"loss": 3.9896,
|
854 |
-
"step": 28200
|
855 |
-
},
|
856 |
-
{
|
857 |
-
"epoch": 0.12,
|
858 |
-
"learning_rate": 0.0012243819420571598,
|
859 |
-
"loss": 3.9951,
|
860 |
-
"step": 28400
|
861 |
-
},
|
862 |
-
{
|
863 |
-
"epoch": 0.12,
|
864 |
-
"learning_rate": 0.0012356834321804039,
|
865 |
-
"loss": 4.0361,
|
866 |
-
"step": 28600
|
867 |
-
},
|
868 |
-
{
|
869 |
-
"epoch": 0.12,
|
870 |
-
"learning_rate": 0.0012470544808827113,
|
871 |
-
"loss": 4.1212,
|
872 |
-
"step": 28800
|
873 |
-
},
|
874 |
-
{
|
875 |
-
"epoch": 0.12,
|
876 |
-
"learning_rate": 0.001258494888645708,
|
877 |
-
"loss": 4.0721,
|
878 |
-
"step": 29000
|
879 |
-
},
|
880 |
-
{
|
881 |
-
"epoch": 0.13,
|
882 |
-
"learning_rate": 0.0012700044547340368,
|
883 |
-
"loss": 4.0311,
|
884 |
-
"step": 29200
|
885 |
-
},
|
886 |
-
{
|
887 |
-
"epoch": 0.13,
|
888 |
-
"learning_rate": 0.0012815829771988738,
|
889 |
-
"loss": 4.0114,
|
890 |
-
"step": 29400
|
891 |
-
},
|
892 |
-
{
|
893 |
-
"epoch": 0.13,
|
894 |
-
"learning_rate": 0.001293230252881479,
|
895 |
-
"loss": 3.9868,
|
896 |
-
"step": 29600
|
897 |
-
},
|
898 |
-
{
|
899 |
-
"epoch": 0.13,
|
900 |
-
"learning_rate": 0.0013049460774167514,
|
901 |
-
"loss": 3.9881,
|
902 |
-
"step": 29800
|
903 |
-
},
|
904 |
-
{
|
905 |
-
"epoch": 0.13,
|
906 |
-
"learning_rate": 0.0013167302452368242,
|
907 |
-
"loss": 3.9705,
|
908 |
-
"step": 30000
|
909 |
}
|
910 |
],
|
911 |
"max_steps": 500000,
|
912 |
"num_train_epochs": 3,
|
913 |
-
"total_flos":
|
914 |
"trial_name": null,
|
915 |
"trial_params": null
|
916 |
}
|
|
|
1 |
{
|
2 |
"best_metric": null,
|
3 |
"best_model_checkpoint": null,
|
4 |
+
"epoch": 0.042973785990545764,
|
5 |
+
"global_step": 10000,
|
6 |
"is_hyper_param_search": false,
|
7 |
"is_local_process_zero": true,
|
8 |
"is_world_process_zero": true,
|
9 |
"log_history": [
|
10 |
{
|
11 |
"epoch": 0.0,
|
12 |
+
"learning_rate": 4.0004211081201405e-05,
|
13 |
+
"loss": 3.9197,
|
14 |
"step": 200
|
15 |
},
|
16 |
{
|
17 |
"epoch": 0.0,
|
18 |
+
"learning_rate": 4.0016844250917146e-05,
|
19 |
+
"loss": 3.9108,
|
20 |
"step": 400
|
21 |
},
|
22 |
{
|
23 |
"epoch": 0.0,
|
24 |
+
"learning_rate": 4.003789928748371e-05,
|
25 |
+
"loss": 3.8655,
|
26 |
"step": 600
|
27 |
},
|
28 |
{
|
29 |
"epoch": 0.0,
|
30 |
+
"learning_rate": 4.006737582146571e-05,
|
31 |
+
"loss": 3.8602,
|
32 |
"step": 800
|
33 |
},
|
34 |
{
|
35 |
"epoch": 0.0,
|
36 |
+
"learning_rate": 4.010527333566261e-05,
|
37 |
+
"loss": 3.8478,
|
38 |
"step": 1000
|
39 |
},
|
40 |
{
|
41 |
"epoch": 0.01,
|
42 |
+
"learning_rate": 4.0151591165118474e-05,
|
43 |
+
"loss": 3.8608,
|
44 |
"step": 1200
|
45 |
},
|
46 |
{
|
47 |
"epoch": 0.01,
|
48 |
+
"learning_rate": 4.020632849713237e-05,
|
49 |
+
"loss": 3.8431,
|
50 |
"step": 1400
|
51 |
},
|
52 |
{
|
53 |
"epoch": 0.01,
|
54 |
+
"learning_rate": 4.0269484371273996e-05,
|
55 |
+
"loss": 3.8447,
|
56 |
"step": 1600
|
57 |
},
|
58 |
{
|
59 |
"epoch": 0.01,
|
60 |
+
"learning_rate": 4.034105767939918e-05,
|
61 |
+
"loss": 3.8422,
|
62 |
"step": 1800
|
63 |
},
|
64 |
{
|
65 |
"epoch": 0.01,
|
66 |
+
"learning_rate": 4.0421047165670686e-05,
|
67 |
+
"loss": 3.8019,
|
68 |
"step": 2000
|
69 |
},
|
70 |
{
|
71 |
"epoch": 0.01,
|
72 |
+
"learning_rate": 4.050945142657905e-05,
|
73 |
+
"loss": 3.8109,
|
74 |
"step": 2200
|
75 |
},
|
76 |
{
|
77 |
"epoch": 0.01,
|
78 |
+
"learning_rate": 4.0606268910968035e-05,
|
79 |
+
"loss": 3.8192,
|
80 |
"step": 2400
|
81 |
},
|
82 |
{
|
83 |
"epoch": 0.01,
|
84 |
+
"learning_rate": 4.071149792006154e-05,
|
85 |
+
"loss": 3.8037,
|
86 |
"step": 2600
|
87 |
},
|
88 |
{
|
89 |
"epoch": 0.01,
|
90 |
+
"learning_rate": 4.082513660749298e-05,
|
91 |
+
"loss": 3.8074,
|
92 |
"step": 2800
|
93 |
},
|
94 |
{
|
95 |
"epoch": 0.01,
|
96 |
+
"learning_rate": 4.0947182979338874e-05,
|
97 |
+
"loss": 3.801,
|
98 |
"step": 3000
|
99 |
},
|
100 |
{
|
101 |
"epoch": 0.01,
|
102 |
+
"learning_rate": 4.1077634894152375e-05,
|
103 |
+
"loss": 3.786,
|
104 |
"step": 3200
|
105 |
},
|
106 |
{
|
107 |
"epoch": 0.01,
|
108 |
+
"learning_rate": 4.121649006300163e-05,
|
109 |
+
"loss": 3.7799,
|
110 |
"step": 3400
|
111 |
},
|
112 |
{
|
113 |
"epoch": 0.02,
|
114 |
+
"learning_rate": 4.136374604951046e-05,
|
115 |
+
"loss": 3.7721,
|
116 |
"step": 3600
|
117 |
},
|
118 |
{
|
119 |
"epoch": 0.02,
|
120 |
+
"learning_rate": 4.1519400269899535e-05,
|
121 |
+
"loss": 3.7829,
|
122 |
"step": 3800
|
123 |
},
|
124 |
{
|
125 |
"epoch": 0.02,
|
126 |
+
"learning_rate": 4.1683449993033545e-05,
|
127 |
+
"loss": 3.7922,
|
128 |
"step": 4000
|
129 |
},
|
130 |
{
|
131 |
"epoch": 0.02,
|
132 |
+
"learning_rate": 4.185589234046783e-05,
|
133 |
+
"loss": 3.7686,
|
134 |
"step": 4200
|
135 |
},
|
136 |
{
|
137 |
"epoch": 0.02,
|
138 |
+
"learning_rate": 4.203672428649923e-05,
|
139 |
+
"loss": 3.7742,
|
140 |
"step": 4400
|
141 |
},
|
142 |
{
|
143 |
"epoch": 0.02,
|
144 |
+
"learning_rate": 4.2225942658219505e-05,
|
145 |
+
"loss": 3.753,
|
146 |
"step": 4600
|
147 |
},
|
148 |
{
|
149 |
"epoch": 0.02,
|
150 |
+
"learning_rate": 4.242354413557057e-05,
|
151 |
+
"loss": 3.7663,
|
152 |
"step": 4800
|
153 |
},
|
154 |
{
|
155 |
"epoch": 0.02,
|
156 |
+
"learning_rate": 4.262952525140289e-05,
|
157 |
+
"loss": 3.7589,
|
158 |
"step": 5000
|
159 |
},
|
160 |
{
|
161 |
"epoch": 0.02,
|
162 |
+
"learning_rate": 4.2843882391536666e-05,
|
163 |
+
"loss": 3.7652,
|
164 |
"step": 5200
|
165 |
},
|
166 |
{
|
167 |
"epoch": 0.02,
|
168 |
+
"learning_rate": 4.306661179482435e-05,
|
169 |
+
"loss": 3.7618,
|
170 |
"step": 5400
|
171 |
},
|
172 |
{
|
173 |
"epoch": 0.02,
|
174 |
+
"learning_rate": 4.329770955321802e-05,
|
175 |
+
"loss": 3.7441,
|
176 |
"step": 5600
|
177 |
},
|
178 |
{
|
179 |
"epoch": 0.02,
|
180 |
+
"learning_rate": 4.3537171611836355e-05,
|
181 |
+
"loss": 3.7364,
|
182 |
"step": 5800
|
183 |
},
|
184 |
{
|
185 |
"epoch": 0.03,
|
186 |
+
"learning_rate": 4.3784993769037384e-05,
|
187 |
+
"loss": 3.7501,
|
188 |
"step": 6000
|
189 |
},
|
190 |
{
|
191 |
"epoch": 0.03,
|
192 |
+
"learning_rate": 4.404117167649071e-05,
|
193 |
+
"loss": 3.7363,
|
194 |
"step": 6200
|
195 |
},
|
196 |
{
|
197 |
"epoch": 0.03,
|
198 |
+
"learning_rate": 4.43057008392547e-05,
|
199 |
+
"loss": 3.7279,
|
200 |
"step": 6400
|
201 |
},
|
202 |
{
|
203 |
"epoch": 0.03,
|
204 |
+
"learning_rate": 4.457857661585541e-05,
|
205 |
+
"loss": 3.7158,
|
206 |
"step": 6600
|
207 |
},
|
208 |
{
|
209 |
"epoch": 0.03,
|
210 |
+
"learning_rate": 4.485979421836779e-05,
|
211 |
+
"loss": 3.74,
|
212 |
"step": 6800
|
213 |
},
|
214 |
{
|
215 |
"epoch": 0.03,
|
216 |
+
"learning_rate": 4.514934871249906e-05,
|
217 |
+
"loss": 3.7441,
|
218 |
"step": 7000
|
219 |
},
|
220 |
{
|
221 |
"epoch": 0.03,
|
222 |
+
"learning_rate": 4.544723501767687e-05,
|
223 |
+
"loss": 3.7141,
|
224 |
"step": 7200
|
225 |
},
|
226 |
{
|
227 |
"epoch": 0.03,
|
228 |
+
"learning_rate": 4.575344790713656e-05,
|
229 |
+
"loss": 3.7264,
|
230 |
"step": 7400
|
231 |
},
|
232 |
{
|
233 |
"epoch": 0.03,
|
234 |
+
"learning_rate": 4.6067982008014406e-05,
|
235 |
+
"loss": 3.7221,
|
236 |
"step": 7600
|
237 |
},
|
238 |
{
|
239 |
"epoch": 0.03,
|
240 |
+
"learning_rate": 4.639083180144098e-05,
|
241 |
+
"loss": 3.7363,
|
242 |
"step": 7800
|
243 |
},
|
244 |
{
|
245 |
"epoch": 0.03,
|
246 |
+
"learning_rate": 4.672199162263839e-05,
|
247 |
+
"loss": 3.73,
|
248 |
"step": 8000
|
249 |
},
|
250 |
{
|
251 |
"epoch": 0.04,
|
252 |
+
"learning_rate": 4.706145566101981e-05,
|
253 |
+
"loss": 3.7205,
|
254 |
"step": 8200
|
255 |
},
|
256 |
{
|
257 |
"epoch": 0.04,
|
258 |
+
"learning_rate": 4.740921796029065e-05,
|
259 |
+
"loss": 3.7192,
|
260 |
"step": 8400
|
261 |
},
|
262 |
{
|
263 |
"epoch": 0.04,
|
264 |
+
"learning_rate": 4.776527241855393e-05,
|
265 |
+
"loss": 3.7095,
|
266 |
"step": 8600
|
267 |
},
|
268 |
{
|
269 |
"epoch": 0.04,
|
270 |
+
"learning_rate": 4.812961278841707e-05,
|
271 |
+
"loss": 3.7311,
|
272 |
"step": 8800
|
273 |
},
|
274 |
{
|
275 |
"epoch": 0.04,
|
276 |
+
"learning_rate": 4.85022326771014e-05,
|
277 |
+
"loss": 3.7031,
|
278 |
"step": 9000
|
279 |
},
|
280 |
{
|
281 |
"epoch": 0.04,
|
282 |
+
"learning_rate": 4.888312554655438e-05,
|
283 |
+
"loss": 3.6934,
|
284 |
"step": 9200
|
285 |
},
|
286 |
{
|
287 |
"epoch": 0.04,
|
288 |
+
"learning_rate": 4.92722847135643e-05,
|
289 |
+
"loss": 3.7218,
|
290 |
"step": 9400
|
291 |
},
|
292 |
{
|
293 |
"epoch": 0.04,
|
294 |
+
"learning_rate": 4.9669703349877704e-05,
|
295 |
+
"loss": 3.711,
|
296 |
"step": 9600
|
297 |
},
|
298 |
{
|
299 |
"epoch": 0.04,
|
300 |
+
"learning_rate": 5.0075374482318863e-05,
|
301 |
+
"loss": 3.7002,
|
302 |
"step": 9800
|
303 |
},
|
304 |
{
|
305 |
"epoch": 0.04,
|
306 |
+
"learning_rate": 5.0489290992912625e-05,
|
307 |
+
"loss": 3.698,
|
308 |
"step": 10000
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
309 |
}
|
310 |
],
|
311 |
"max_steps": 500000,
|
312 |
"num_train_epochs": 3,
|
313 |
+
"total_flos": 1.593829982208e+16,
|
314 |
"trial_name": null,
|
315 |
"trial_params": null
|
316 |
}
|
last-checkpoint/training_args.bin
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 3375
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:db1575e3c0c4f23842bbe1ba840257aad501aecba3d1800f59d7014cd73b0cee
|
3 |
size 3375
|
pytorch_model.bin
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 146774203
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:f48f41b7b489ac24545b446eb3e28cf98daf67da888804a7513071641e3cb0ab
|
3 |
size 146774203
|
training_args.bin
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 3375
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:db1575e3c0c4f23842bbe1ba840257aad501aecba3d1800f59d7014cd73b0cee
|
3 |
size 3375
|