Upload folder using huggingface_hub
Browse files- adapter_model.safetensors +1 -1
- optimizer.pt +1 -1
- rng_state.pth +1 -1
- scheduler.pt +1 -1
- tokenizer.json +13 -17
- tokenizer_config.json +32 -24
- trainer_state.json +4 -2139
adapter_model.safetensors
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 16794200
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:c1798137f652a8679bd8dff8868caa0c7d3650825dfad997314aaccf8253c226
|
3 |
size 16794200
|
optimizer.pt
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 33630266
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:9e0860fc0a59df18c6f13f01f83dcdb46be4e90b8d0f0ea7b080189423780876
|
3 |
size 33630266
|
rng_state.pth
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 14244
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:a1c06f6d1e961307af8d153c92a2d6f32b84473dc3657d383124bf630f1a8528
|
3 |
size 14244
|
scheduler.pt
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 1064
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:6c3d815b05ef65572f79e932f68d13fd57040d3c53068f392569be5b4b68acbb
|
3 |
size 1064
|
tokenizer.json
CHANGED
@@ -1,6 +1,11 @@
|
|
1 |
{
|
2 |
"version": "1.0",
|
3 |
-
"truncation":
|
|
|
|
|
|
|
|
|
|
|
4 |
"padding": null,
|
5 |
"added_tokens": [
|
6 |
{
|
@@ -31,23 +36,13 @@
|
|
31 |
"special": true
|
32 |
}
|
33 |
],
|
34 |
-
"normalizer":
|
35 |
-
|
36 |
-
"
|
37 |
-
|
38 |
-
|
39 |
-
|
40 |
-
},
|
41 |
-
{
|
42 |
-
"type": "Replace",
|
43 |
-
"pattern": {
|
44 |
-
"String": " "
|
45 |
-
},
|
46 |
-
"content": "▁"
|
47 |
-
}
|
48 |
-
]
|
49 |
},
|
50 |
-
"pre_tokenizer": null,
|
51 |
"post_processor": {
|
52 |
"type": "TemplateProcessing",
|
53 |
"single": [
|
@@ -134,6 +129,7 @@
|
|
134 |
"end_of_word_suffix": null,
|
135 |
"fuse_unk": true,
|
136 |
"byte_fallback": true,
|
|
|
137 |
"vocab": {
|
138 |
"<unk>": 0,
|
139 |
"<s>": 1,
|
|
|
1 |
{
|
2 |
"version": "1.0",
|
3 |
+
"truncation": {
|
4 |
+
"direction": "Right",
|
5 |
+
"max_length": 1024,
|
6 |
+
"strategy": "LongestFirst",
|
7 |
+
"stride": 0
|
8 |
+
},
|
9 |
"padding": null,
|
10 |
"added_tokens": [
|
11 |
{
|
|
|
36 |
"special": true
|
37 |
}
|
38 |
],
|
39 |
+
"normalizer": null,
|
40 |
+
"pre_tokenizer": {
|
41 |
+
"type": "Metaspace",
|
42 |
+
"replacement": "▁",
|
43 |
+
"prepend_scheme": "first",
|
44 |
+
"split": false
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
45 |
},
|
|
|
46 |
"post_processor": {
|
47 |
"type": "TemplateProcessing",
|
48 |
"single": [
|
|
|
129 |
"end_of_word_suffix": null,
|
130 |
"fuse_unk": true,
|
131 |
"byte_fallback": true,
|
132 |
+
"ignore_merges": false,
|
133 |
"vocab": {
|
134 |
"<unk>": 0,
|
135 |
"<s>": 1,
|
tokenizer_config.json
CHANGED
@@ -1,35 +1,43 @@
|
|
1 |
{
|
2 |
"add_bos_token": true,
|
3 |
"add_eos_token": false,
|
4 |
-
"
|
5 |
-
|
6 |
-
"
|
7 |
-
|
8 |
-
|
9 |
-
|
10 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
11 |
},
|
|
|
12 |
"clean_up_tokenization_spaces": false,
|
13 |
-
"eos_token":
|
14 |
-
"__type": "AddedToken",
|
15 |
-
"content": "</s>",
|
16 |
-
"lstrip": false,
|
17 |
-
"normalized": false,
|
18 |
-
"rstrip": false,
|
19 |
-
"single_word": false
|
20 |
-
},
|
21 |
"legacy": false,
|
22 |
"model_max_length": 4096,
|
23 |
-
"pad_token":
|
24 |
"padding_side": "right",
|
25 |
"sp_model_kwargs": {},
|
|
|
26 |
"tokenizer_class": "LlamaTokenizer",
|
27 |
-
"unk_token":
|
28 |
-
|
29 |
-
"content": "<unk>",
|
30 |
-
"lstrip": false,
|
31 |
-
"normalized": false,
|
32 |
-
"rstrip": false,
|
33 |
-
"single_word": false
|
34 |
-
}
|
35 |
}
|
|
|
1 |
{
|
2 |
"add_bos_token": true,
|
3 |
"add_eos_token": false,
|
4 |
+
"add_prefix_space": true,
|
5 |
+
"added_tokens_decoder": {
|
6 |
+
"0": {
|
7 |
+
"content": "<unk>",
|
8 |
+
"lstrip": false,
|
9 |
+
"normalized": false,
|
10 |
+
"rstrip": false,
|
11 |
+
"single_word": false,
|
12 |
+
"special": true
|
13 |
+
},
|
14 |
+
"1": {
|
15 |
+
"content": "<s>",
|
16 |
+
"lstrip": false,
|
17 |
+
"normalized": false,
|
18 |
+
"rstrip": false,
|
19 |
+
"single_word": false,
|
20 |
+
"special": true
|
21 |
+
},
|
22 |
+
"2": {
|
23 |
+
"content": "</s>",
|
24 |
+
"lstrip": false,
|
25 |
+
"normalized": false,
|
26 |
+
"rstrip": false,
|
27 |
+
"single_word": false,
|
28 |
+
"special": true
|
29 |
+
}
|
30 |
},
|
31 |
+
"bos_token": "<s>",
|
32 |
"clean_up_tokenization_spaces": false,
|
33 |
+
"eos_token": "</s>",
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
34 |
"legacy": false,
|
35 |
"model_max_length": 4096,
|
36 |
+
"pad_token": "</s>",
|
37 |
"padding_side": "right",
|
38 |
"sp_model_kwargs": {},
|
39 |
+
"spaces_between_special_tokens": false,
|
40 |
"tokenizer_class": "LlamaTokenizer",
|
41 |
+
"unk_token": "<unk>",
|
42 |
+
"use_default_system_prompt": false
|
|
|
|
|
|
|
|
|
|
|
|
|
43 |
}
|
trainer_state.json
CHANGED
@@ -1,9 +1,9 @@
|
|
1 |
{
|
2 |
"best_metric": null,
|
3 |
"best_model_checkpoint": null,
|
4 |
-
"epoch":
|
5 |
"eval_steps": 500,
|
6 |
-
"global_step":
|
7 |
"is_hyper_param_search": false,
|
8 |
"is_local_process_zero": true,
|
9 |
"is_world_process_zero": true,
|
@@ -336,2141 +336,6 @@
|
|
336 |
"learning_rate": 0.0001913545457642601,
|
337 |
"loss": 0.0283,
|
338 |
"step": 470
|
339 |
-
},
|
340 |
-
{
|
341 |
-
"epoch": 2.0425531914893615,
|
342 |
-
"grad_norm": 0.28837406635284424,
|
343 |
-
"learning_rate": 0.00019098842577186314,
|
344 |
-
"loss": 0.0137,
|
345 |
-
"step": 480
|
346 |
-
},
|
347 |
-
{
|
348 |
-
"epoch": 2.0851063829787235,
|
349 |
-
"grad_norm": 0.08533861488103867,
|
350 |
-
"learning_rate": 0.00019061507866756347,
|
351 |
-
"loss": 0.0182,
|
352 |
-
"step": 490
|
353 |
-
},
|
354 |
-
{
|
355 |
-
"epoch": 2.127659574468085,
|
356 |
-
"grad_norm": 0.3499375581741333,
|
357 |
-
"learning_rate": 0.00019023453410591635,
|
358 |
-
"loss": 0.0221,
|
359 |
-
"step": 500
|
360 |
-
},
|
361 |
-
{
|
362 |
-
"epoch": 2.1702127659574466,
|
363 |
-
"grad_norm": 0.07716694474220276,
|
364 |
-
"learning_rate": 0.00018984682231316333,
|
365 |
-
"loss": 0.0075,
|
366 |
-
"step": 510
|
367 |
-
},
|
368 |
-
{
|
369 |
-
"epoch": 2.2127659574468086,
|
370 |
-
"grad_norm": 0.3093757927417755,
|
371 |
-
"learning_rate": 0.00018945197408483123,
|
372 |
-
"loss": 0.0133,
|
373 |
-
"step": 520
|
374 |
-
},
|
375 |
-
{
|
376 |
-
"epoch": 2.25531914893617,
|
377 |
-
"grad_norm": 0.13492655754089355,
|
378 |
-
"learning_rate": 0.00018905002078328632,
|
379 |
-
"loss": 0.0184,
|
380 |
-
"step": 530
|
381 |
-
},
|
382 |
-
{
|
383 |
-
"epoch": 2.297872340425532,
|
384 |
-
"grad_norm": 0.07833054661750793,
|
385 |
-
"learning_rate": 0.000188640994335243,
|
386 |
-
"loss": 0.0109,
|
387 |
-
"step": 540
|
388 |
-
},
|
389 |
-
{
|
390 |
-
"epoch": 2.3404255319148937,
|
391 |
-
"grad_norm": 0.08865915983915329,
|
392 |
-
"learning_rate": 0.0001882249272292282,
|
393 |
-
"loss": 0.0121,
|
394 |
-
"step": 550
|
395 |
-
},
|
396 |
-
{
|
397 |
-
"epoch": 2.382978723404255,
|
398 |
-
"grad_norm": 0.31314581632614136,
|
399 |
-
"learning_rate": 0.00018780185251300046,
|
400 |
-
"loss": 0.0242,
|
401 |
-
"step": 560
|
402 |
-
},
|
403 |
-
{
|
404 |
-
"epoch": 2.425531914893617,
|
405 |
-
"grad_norm": 0.10387410968542099,
|
406 |
-
"learning_rate": 0.00018737180379092537,
|
407 |
-
"loss": 0.0285,
|
408 |
-
"step": 570
|
409 |
-
},
|
410 |
-
{
|
411 |
-
"epoch": 2.4680851063829787,
|
412 |
-
"grad_norm": 0.11755700409412384,
|
413 |
-
"learning_rate": 0.0001869348152213061,
|
414 |
-
"loss": 0.0281,
|
415 |
-
"step": 580
|
416 |
-
},
|
417 |
-
{
|
418 |
-
"epoch": 2.5106382978723403,
|
419 |
-
"grad_norm": 0.1438552439212799,
|
420 |
-
"learning_rate": 0.0001864909215136705,
|
421 |
-
"loss": 0.0216,
|
422 |
-
"step": 590
|
423 |
-
},
|
424 |
-
{
|
425 |
-
"epoch": 2.5531914893617023,
|
426 |
-
"grad_norm": 0.08940370380878448,
|
427 |
-
"learning_rate": 0.00018604015792601396,
|
428 |
-
"loss": 0.0226,
|
429 |
-
"step": 600
|
430 |
-
},
|
431 |
-
{
|
432 |
-
"epoch": 2.595744680851064,
|
433 |
-
"grad_norm": 0.19193390011787415,
|
434 |
-
"learning_rate": 0.00018558256026199896,
|
435 |
-
"loss": 0.0184,
|
436 |
-
"step": 610
|
437 |
-
},
|
438 |
-
{
|
439 |
-
"epoch": 2.6382978723404253,
|
440 |
-
"grad_norm": 0.4586654603481293,
|
441 |
-
"learning_rate": 0.00018511816486811134,
|
442 |
-
"loss": 0.0154,
|
443 |
-
"step": 620
|
444 |
-
},
|
445 |
-
{
|
446 |
-
"epoch": 2.6808510638297873,
|
447 |
-
"grad_norm": 0.05811255797743797,
|
448 |
-
"learning_rate": 0.00018464700863077312,
|
449 |
-
"loss": 0.0238,
|
450 |
-
"step": 630
|
451 |
-
},
|
452 |
-
{
|
453 |
-
"epoch": 2.723404255319149,
|
454 |
-
"grad_norm": 0.16280680894851685,
|
455 |
-
"learning_rate": 0.00018416912897341295,
|
456 |
-
"loss": 0.0253,
|
457 |
-
"step": 640
|
458 |
-
},
|
459 |
-
{
|
460 |
-
"epoch": 2.7659574468085104,
|
461 |
-
"grad_norm": 0.07031189650297165,
|
462 |
-
"learning_rate": 0.00018368456385349334,
|
463 |
-
"loss": 0.0146,
|
464 |
-
"step": 650
|
465 |
-
},
|
466 |
-
{
|
467 |
-
"epoch": 2.8085106382978724,
|
468 |
-
"grad_norm": 0.07851342856884003,
|
469 |
-
"learning_rate": 0.0001831933517594957,
|
470 |
-
"loss": 0.0301,
|
471 |
-
"step": 660
|
472 |
-
},
|
473 |
-
{
|
474 |
-
"epoch": 2.851063829787234,
|
475 |
-
"grad_norm": 0.11461298912763596,
|
476 |
-
"learning_rate": 0.0001826955317078636,
|
477 |
-
"loss": 0.0155,
|
478 |
-
"step": 670
|
479 |
-
},
|
480 |
-
{
|
481 |
-
"epoch": 2.8936170212765955,
|
482 |
-
"grad_norm": 0.1839868575334549,
|
483 |
-
"learning_rate": 0.00018219114323990345,
|
484 |
-
"loss": 0.0099,
|
485 |
-
"step": 680
|
486 |
-
},
|
487 |
-
{
|
488 |
-
"epoch": 2.9361702127659575,
|
489 |
-
"grad_norm": 0.04199373722076416,
|
490 |
-
"learning_rate": 0.00018168022641864377,
|
491 |
-
"loss": 0.0192,
|
492 |
-
"step": 690
|
493 |
-
},
|
494 |
-
{
|
495 |
-
"epoch": 2.978723404255319,
|
496 |
-
"grad_norm": 0.3203773498535156,
|
497 |
-
"learning_rate": 0.00018116282182565311,
|
498 |
-
"loss": 0.0272,
|
499 |
-
"step": 700
|
500 |
-
},
|
501 |
-
{
|
502 |
-
"epoch": 3.021276595744681,
|
503 |
-
"grad_norm": 0.327921599149704,
|
504 |
-
"learning_rate": 0.0001806389705578168,
|
505 |
-
"loss": 0.0081,
|
506 |
-
"step": 710
|
507 |
-
},
|
508 |
-
{
|
509 |
-
"epoch": 3.0638297872340425,
|
510 |
-
"grad_norm": 0.29155433177948,
|
511 |
-
"learning_rate": 0.00018010871422407236,
|
512 |
-
"loss": 0.0128,
|
513 |
-
"step": 720
|
514 |
-
},
|
515 |
-
{
|
516 |
-
"epoch": 3.106382978723404,
|
517 |
-
"grad_norm": 0.06926452368497849,
|
518 |
-
"learning_rate": 0.00017957209494210493,
|
519 |
-
"loss": 0.0171,
|
520 |
-
"step": 730
|
521 |
-
},
|
522 |
-
{
|
523 |
-
"epoch": 3.148936170212766,
|
524 |
-
"grad_norm": 0.08231089264154434,
|
525 |
-
"learning_rate": 0.0001790291553350016,
|
526 |
-
"loss": 0.0098,
|
527 |
-
"step": 740
|
528 |
-
},
|
529 |
-
{
|
530 |
-
"epoch": 3.1914893617021276,
|
531 |
-
"grad_norm": 0.14707215130329132,
|
532 |
-
"learning_rate": 0.0001784799385278661,
|
533 |
-
"loss": 0.0092,
|
534 |
-
"step": 750
|
535 |
-
},
|
536 |
-
{
|
537 |
-
"epoch": 3.2340425531914896,
|
538 |
-
"grad_norm": 0.2735896706581116,
|
539 |
-
"learning_rate": 0.00017792448814439333,
|
540 |
-
"loss": 0.0115,
|
541 |
-
"step": 760
|
542 |
-
},
|
543 |
-
{
|
544 |
-
"epoch": 3.276595744680851,
|
545 |
-
"grad_norm": 0.44960370659828186,
|
546 |
-
"learning_rate": 0.00017736284830340436,
|
547 |
-
"loss": 0.0195,
|
548 |
-
"step": 770
|
549 |
-
},
|
550 |
-
{
|
551 |
-
"epoch": 3.3191489361702127,
|
552 |
-
"grad_norm": 0.13413724303245544,
|
553 |
-
"learning_rate": 0.00017679506361534215,
|
554 |
-
"loss": 0.0187,
|
555 |
-
"step": 780
|
556 |
-
},
|
557 |
-
{
|
558 |
-
"epoch": 3.3617021276595747,
|
559 |
-
"grad_norm": 0.24698686599731445,
|
560 |
-
"learning_rate": 0.00017622117917872823,
|
561 |
-
"loss": 0.0125,
|
562 |
-
"step": 790
|
563 |
-
},
|
564 |
-
{
|
565 |
-
"epoch": 3.404255319148936,
|
566 |
-
"grad_norm": 0.48694342374801636,
|
567 |
-
"learning_rate": 0.00017564124057658056,
|
568 |
-
"loss": 0.0234,
|
569 |
-
"step": 800
|
570 |
-
},
|
571 |
-
{
|
572 |
-
"epoch": 3.4468085106382977,
|
573 |
-
"grad_norm": 0.6931429505348206,
|
574 |
-
"learning_rate": 0.00017505529387279277,
|
575 |
-
"loss": 0.0234,
|
576 |
-
"step": 810
|
577 |
-
},
|
578 |
-
{
|
579 |
-
"epoch": 3.4893617021276597,
|
580 |
-
"grad_norm": 0.13700473308563232,
|
581 |
-
"learning_rate": 0.00017446338560847568,
|
582 |
-
"loss": 0.0145,
|
583 |
-
"step": 820
|
584 |
-
},
|
585 |
-
{
|
586 |
-
"epoch": 3.5319148936170213,
|
587 |
-
"grad_norm": 0.3254775404930115,
|
588 |
-
"learning_rate": 0.00017386556279826021,
|
589 |
-
"loss": 0.0179,
|
590 |
-
"step": 830
|
591 |
-
},
|
592 |
-
{
|
593 |
-
"epoch": 3.574468085106383,
|
594 |
-
"grad_norm": 0.368379682302475,
|
595 |
-
"learning_rate": 0.00017326187292656333,
|
596 |
-
"loss": 0.013,
|
597 |
-
"step": 840
|
598 |
-
},
|
599 |
-
{
|
600 |
-
"epoch": 3.617021276595745,
|
601 |
-
"grad_norm": 0.2745888829231262,
|
602 |
-
"learning_rate": 0.00017265236394381633,
|
603 |
-
"loss": 0.0136,
|
604 |
-
"step": 850
|
605 |
-
},
|
606 |
-
{
|
607 |
-
"epoch": 3.6595744680851063,
|
608 |
-
"grad_norm": 0.0781714916229248,
|
609 |
-
"learning_rate": 0.00017203708426265614,
|
610 |
-
"loss": 0.0126,
|
611 |
-
"step": 860
|
612 |
-
},
|
613 |
-
{
|
614 |
-
"epoch": 3.702127659574468,
|
615 |
-
"grad_norm": 0.7518234848976135,
|
616 |
-
"learning_rate": 0.00017141608275408006,
|
617 |
-
"loss": 0.0134,
|
618 |
-
"step": 870
|
619 |
-
},
|
620 |
-
{
|
621 |
-
"epoch": 3.74468085106383,
|
622 |
-
"grad_norm": 0.4271518886089325,
|
623 |
-
"learning_rate": 0.00017078940874356392,
|
624 |
-
"loss": 0.0127,
|
625 |
-
"step": 880
|
626 |
-
},
|
627 |
-
{
|
628 |
-
"epoch": 3.7872340425531914,
|
629 |
-
"grad_norm": 0.34387120604515076,
|
630 |
-
"learning_rate": 0.00017015711200714414,
|
631 |
-
"loss": 0.0161,
|
632 |
-
"step": 890
|
633 |
-
},
|
634 |
-
{
|
635 |
-
"epoch": 3.829787234042553,
|
636 |
-
"grad_norm": 0.09641717374324799,
|
637 |
-
"learning_rate": 0.00016951924276746425,
|
638 |
-
"loss": 0.0185,
|
639 |
-
"step": 900
|
640 |
-
},
|
641 |
-
{
|
642 |
-
"epoch": 3.872340425531915,
|
643 |
-
"grad_norm": 0.44215765595436096,
|
644 |
-
"learning_rate": 0.00016887585168978562,
|
645 |
-
"loss": 0.0115,
|
646 |
-
"step": 910
|
647 |
-
},
|
648 |
-
{
|
649 |
-
"epoch": 3.9148936170212765,
|
650 |
-
"grad_norm": 0.06954783201217651,
|
651 |
-
"learning_rate": 0.0001682269898779632,
|
652 |
-
"loss": 0.0121,
|
653 |
-
"step": 920
|
654 |
-
},
|
655 |
-
{
|
656 |
-
"epoch": 3.9574468085106385,
|
657 |
-
"grad_norm": 0.21080243587493896,
|
658 |
-
"learning_rate": 0.00016757270887038654,
|
659 |
-
"loss": 0.0125,
|
660 |
-
"step": 930
|
661 |
-
},
|
662 |
-
{
|
663 |
-
"epoch": 4.0,
|
664 |
-
"grad_norm": 0.3382236063480377,
|
665 |
-
"learning_rate": 0.00016691306063588583,
|
666 |
-
"loss": 0.0109,
|
667 |
-
"step": 940
|
668 |
-
},
|
669 |
-
{
|
670 |
-
"epoch": 4.042553191489362,
|
671 |
-
"grad_norm": 0.36123183369636536,
|
672 |
-
"learning_rate": 0.00016624809756960444,
|
673 |
-
"loss": 0.0093,
|
674 |
-
"step": 950
|
675 |
-
},
|
676 |
-
{
|
677 |
-
"epoch": 4.085106382978723,
|
678 |
-
"grad_norm": 0.45228853821754456,
|
679 |
-
"learning_rate": 0.00016557787248883696,
|
680 |
-
"loss": 0.0117,
|
681 |
-
"step": 960
|
682 |
-
},
|
683 |
-
{
|
684 |
-
"epoch": 4.127659574468085,
|
685 |
-
"grad_norm": 0.2724202275276184,
|
686 |
-
"learning_rate": 0.00016490243862883413,
|
687 |
-
"loss": 0.0126,
|
688 |
-
"step": 970
|
689 |
-
},
|
690 |
-
{
|
691 |
-
"epoch": 4.170212765957447,
|
692 |
-
"grad_norm": 0.17904357612133026,
|
693 |
-
"learning_rate": 0.00016422184963857432,
|
694 |
-
"loss": 0.0103,
|
695 |
-
"step": 980
|
696 |
-
},
|
697 |
-
{
|
698 |
-
"epoch": 4.212765957446808,
|
699 |
-
"grad_norm": 0.4267734885215759,
|
700 |
-
"learning_rate": 0.00016353615957650236,
|
701 |
-
"loss": 0.0105,
|
702 |
-
"step": 990
|
703 |
-
},
|
704 |
-
{
|
705 |
-
"epoch": 4.25531914893617,
|
706 |
-
"grad_norm": 0.11816457659006119,
|
707 |
-
"learning_rate": 0.00016284542290623567,
|
708 |
-
"loss": 0.0097,
|
709 |
-
"step": 1000
|
710 |
-
},
|
711 |
-
{
|
712 |
-
"epoch": 4.297872340425532,
|
713 |
-
"grad_norm": 0.04166145250201225,
|
714 |
-
"learning_rate": 0.00016214969449223824,
|
715 |
-
"loss": 0.0116,
|
716 |
-
"step": 1010
|
717 |
-
},
|
718 |
-
{
|
719 |
-
"epoch": 4.340425531914893,
|
720 |
-
"grad_norm": 0.0687410831451416,
|
721 |
-
"learning_rate": 0.00016144902959546286,
|
722 |
-
"loss": 0.0088,
|
723 |
-
"step": 1020
|
724 |
-
},
|
725 |
-
{
|
726 |
-
"epoch": 4.382978723404255,
|
727 |
-
"grad_norm": 0.37416237592697144,
|
728 |
-
"learning_rate": 0.00016074348386896177,
|
729 |
-
"loss": 0.019,
|
730 |
-
"step": 1030
|
731 |
-
},
|
732 |
-
{
|
733 |
-
"epoch": 4.425531914893617,
|
734 |
-
"grad_norm": 0.06069188937544823,
|
735 |
-
"learning_rate": 0.00016003311335346636,
|
736 |
-
"loss": 0.0138,
|
737 |
-
"step": 1040
|
738 |
-
},
|
739 |
-
{
|
740 |
-
"epoch": 4.468085106382979,
|
741 |
-
"grad_norm": 0.0817495658993721,
|
742 |
-
"learning_rate": 0.00015931797447293552,
|
743 |
-
"loss": 0.0084,
|
744 |
-
"step": 1050
|
745 |
-
},
|
746 |
-
{
|
747 |
-
"epoch": 4.51063829787234,
|
748 |
-
"grad_norm": 0.09787007421255112,
|
749 |
-
"learning_rate": 0.00015859812403007443,
|
750 |
-
"loss": 0.009,
|
751 |
-
"step": 1060
|
752 |
-
},
|
753 |
-
{
|
754 |
-
"epoch": 4.553191489361702,
|
755 |
-
"grad_norm": 0.06179153174161911,
|
756 |
-
"learning_rate": 0.0001578736192018224,
|
757 |
-
"loss": 0.008,
|
758 |
-
"step": 1070
|
759 |
-
},
|
760 |
-
{
|
761 |
-
"epoch": 4.595744680851064,
|
762 |
-
"grad_norm": 0.3092339038848877,
|
763 |
-
"learning_rate": 0.00015714451753481168,
|
764 |
-
"loss": 0.0131,
|
765 |
-
"step": 1080
|
766 |
-
},
|
767 |
-
{
|
768 |
-
"epoch": 4.638297872340425,
|
769 |
-
"grad_norm": 0.06405780464410782,
|
770 |
-
"learning_rate": 0.0001564108769407962,
|
771 |
-
"loss": 0.0122,
|
772 |
-
"step": 1090
|
773 |
-
},
|
774 |
-
{
|
775 |
-
"epoch": 4.680851063829787,
|
776 |
-
"grad_norm": 0.21473410725593567,
|
777 |
-
"learning_rate": 0.00015567275569205218,
|
778 |
-
"loss": 0.0117,
|
779 |
-
"step": 1100
|
780 |
-
},
|
781 |
-
{
|
782 |
-
"epoch": 4.723404255319149,
|
783 |
-
"grad_norm": 0.3080317974090576,
|
784 |
-
"learning_rate": 0.00015493021241674918,
|
785 |
-
"loss": 0.011,
|
786 |
-
"step": 1110
|
787 |
-
},
|
788 |
-
{
|
789 |
-
"epoch": 4.76595744680851,
|
790 |
-
"grad_norm": 0.6501839756965637,
|
791 |
-
"learning_rate": 0.0001541833060942937,
|
792 |
-
"loss": 0.0124,
|
793 |
-
"step": 1120
|
794 |
-
},
|
795 |
-
{
|
796 |
-
"epoch": 4.808510638297872,
|
797 |
-
"grad_norm": 0.04628886282444,
|
798 |
-
"learning_rate": 0.00015343209605064422,
|
799 |
-
"loss": 0.0082,
|
800 |
-
"step": 1130
|
801 |
-
},
|
802 |
-
{
|
803 |
-
"epoch": 4.851063829787234,
|
804 |
-
"grad_norm": 0.059793177992105484,
|
805 |
-
"learning_rate": 0.00015267664195359917,
|
806 |
-
"loss": 0.0072,
|
807 |
-
"step": 1140
|
808 |
-
},
|
809 |
-
{
|
810 |
-
"epoch": 4.8936170212765955,
|
811 |
-
"grad_norm": 0.052162788808345795,
|
812 |
-
"learning_rate": 0.00015191700380805752,
|
813 |
-
"loss": 0.0095,
|
814 |
-
"step": 1150
|
815 |
-
},
|
816 |
-
{
|
817 |
-
"epoch": 4.9361702127659575,
|
818 |
-
"grad_norm": 0.06147542968392372,
|
819 |
-
"learning_rate": 0.00015115324195125274,
|
820 |
-
"loss": 0.0098,
|
821 |
-
"step": 1160
|
822 |
-
},
|
823 |
-
{
|
824 |
-
"epoch": 4.9787234042553195,
|
825 |
-
"grad_norm": 0.6586833000183105,
|
826 |
-
"learning_rate": 0.00015038541704796003,
|
827 |
-
"loss": 0.0139,
|
828 |
-
"step": 1170
|
829 |
-
},
|
830 |
-
{
|
831 |
-
"epoch": 5.0212765957446805,
|
832 |
-
"grad_norm": 0.2920892834663391,
|
833 |
-
"learning_rate": 0.0001496135900856782,
|
834 |
-
"loss": 0.0083,
|
835 |
-
"step": 1180
|
836 |
-
},
|
837 |
-
{
|
838 |
-
"epoch": 5.0638297872340425,
|
839 |
-
"grad_norm": 0.32393306493759155,
|
840 |
-
"learning_rate": 0.0001488378223697851,
|
841 |
-
"loss": 0.0155,
|
842 |
-
"step": 1190
|
843 |
-
},
|
844 |
-
{
|
845 |
-
"epoch": 5.1063829787234045,
|
846 |
-
"grad_norm": 0.49029844999313354,
|
847 |
-
"learning_rate": 0.00014805817551866838,
|
848 |
-
"loss": 0.0109,
|
849 |
-
"step": 1200
|
850 |
-
},
|
851 |
-
{
|
852 |
-
"epoch": 5.148936170212766,
|
853 |
-
"grad_norm": 0.05497799441218376,
|
854 |
-
"learning_rate": 0.00014727471145883127,
|
855 |
-
"loss": 0.0095,
|
856 |
-
"step": 1210
|
857 |
-
},
|
858 |
-
{
|
859 |
-
"epoch": 5.191489361702128,
|
860 |
-
"grad_norm": 0.4540445804595947,
|
861 |
-
"learning_rate": 0.00014648749241997363,
|
862 |
-
"loss": 0.0106,
|
863 |
-
"step": 1220
|
864 |
-
},
|
865 |
-
{
|
866 |
-
"epoch": 5.23404255319149,
|
867 |
-
"grad_norm": 0.16598157584667206,
|
868 |
-
"learning_rate": 0.00014569658093004935,
|
869 |
-
"loss": 0.0072,
|
870 |
-
"step": 1230
|
871 |
-
},
|
872 |
-
{
|
873 |
-
"epoch": 5.276595744680851,
|
874 |
-
"grad_norm": 0.07160704582929611,
|
875 |
-
"learning_rate": 0.0001449020398102996,
|
876 |
-
"loss": 0.0108,
|
877 |
-
"step": 1240
|
878 |
-
},
|
879 |
-
{
|
880 |
-
"epoch": 5.319148936170213,
|
881 |
-
"grad_norm": 0.197789266705513,
|
882 |
-
"learning_rate": 0.00014410393217026318,
|
883 |
-
"loss": 0.0118,
|
884 |
-
"step": 1250
|
885 |
-
},
|
886 |
-
{
|
887 |
-
"epoch": 5.361702127659575,
|
888 |
-
"grad_norm": 0.07983817905187607,
|
889 |
-
"learning_rate": 0.00014330232140276366,
|
890 |
-
"loss": 0.0076,
|
891 |
-
"step": 1260
|
892 |
-
},
|
893 |
-
{
|
894 |
-
"epoch": 5.404255319148936,
|
895 |
-
"grad_norm": 0.0746329054236412,
|
896 |
-
"learning_rate": 0.00014249727117887425,
|
897 |
-
"loss": 0.0089,
|
898 |
-
"step": 1270
|
899 |
-
},
|
900 |
-
{
|
901 |
-
"epoch": 5.446808510638298,
|
902 |
-
"grad_norm": 0.09392493963241577,
|
903 |
-
"learning_rate": 0.00014168884544286053,
|
904 |
-
"loss": 0.0103,
|
905 |
-
"step": 1280
|
906 |
-
},
|
907 |
-
{
|
908 |
-
"epoch": 5.48936170212766,
|
909 |
-
"grad_norm": 0.18386998772621155,
|
910 |
-
"learning_rate": 0.0001408771084071012,
|
911 |
-
"loss": 0.0096,
|
912 |
-
"step": 1290
|
913 |
-
},
|
914 |
-
{
|
915 |
-
"epoch": 5.531914893617021,
|
916 |
-
"grad_norm": 0.07400283962488174,
|
917 |
-
"learning_rate": 0.00014006212454698797,
|
918 |
-
"loss": 0.0083,
|
919 |
-
"step": 1300
|
920 |
-
},
|
921 |
-
{
|
922 |
-
"epoch": 5.574468085106383,
|
923 |
-
"grad_norm": 0.06513144075870514,
|
924 |
-
"learning_rate": 0.00013924395859580432,
|
925 |
-
"loss": 0.0093,
|
926 |
-
"step": 1310
|
927 |
-
},
|
928 |
-
{
|
929 |
-
"epoch": 5.617021276595745,
|
930 |
-
"grad_norm": 0.6950928568840027,
|
931 |
-
"learning_rate": 0.00013842267553958371,
|
932 |
-
"loss": 0.0073,
|
933 |
-
"step": 1320
|
934 |
-
},
|
935 |
-
{
|
936 |
-
"epoch": 5.659574468085106,
|
937 |
-
"grad_norm": 0.05320321023464203,
|
938 |
-
"learning_rate": 0.00013759834061194794,
|
939 |
-
"loss": 0.0098,
|
940 |
-
"step": 1330
|
941 |
-
},
|
942 |
-
{
|
943 |
-
"epoch": 5.702127659574468,
|
944 |
-
"grad_norm": 0.17569933831691742,
|
945 |
-
"learning_rate": 0.00013677101928892554,
|
946 |
-
"loss": 0.0075,
|
947 |
-
"step": 1340
|
948 |
-
},
|
949 |
-
{
|
950 |
-
"epoch": 5.74468085106383,
|
951 |
-
"grad_norm": 0.052122730761766434,
|
952 |
-
"learning_rate": 0.00013594077728375128,
|
953 |
-
"loss": 0.0107,
|
954 |
-
"step": 1350
|
955 |
-
},
|
956 |
-
{
|
957 |
-
"epoch": 5.787234042553192,
|
958 |
-
"grad_norm": 0.2108752280473709,
|
959 |
-
"learning_rate": 0.00013510768054164653,
|
960 |
-
"loss": 0.0119,
|
961 |
-
"step": 1360
|
962 |
-
},
|
963 |
-
{
|
964 |
-
"epoch": 5.829787234042553,
|
965 |
-
"grad_norm": 0.047486983239650726,
|
966 |
-
"learning_rate": 0.00013427179523458127,
|
967 |
-
"loss": 0.0092,
|
968 |
-
"step": 1370
|
969 |
-
},
|
970 |
-
{
|
971 |
-
"epoch": 5.872340425531915,
|
972 |
-
"grad_norm": 0.043320391327142715,
|
973 |
-
"learning_rate": 0.0001334331877560182,
|
974 |
-
"loss": 0.0081,
|
975 |
-
"step": 1380
|
976 |
-
},
|
977 |
-
{
|
978 |
-
"epoch": 5.914893617021277,
|
979 |
-
"grad_norm": 0.09155077487230301,
|
980 |
-
"learning_rate": 0.00013259192471563912,
|
981 |
-
"loss": 0.0091,
|
982 |
-
"step": 1390
|
983 |
-
},
|
984 |
-
{
|
985 |
-
"epoch": 5.957446808510638,
|
986 |
-
"grad_norm": 0.049143675714731216,
|
987 |
-
"learning_rate": 0.00013174807293405428,
|
988 |
-
"loss": 0.0089,
|
989 |
-
"step": 1400
|
990 |
-
},
|
991 |
-
{
|
992 |
-
"epoch": 6.0,
|
993 |
-
"grad_norm": 0.07365540415048599,
|
994 |
-
"learning_rate": 0.00013090169943749476,
|
995 |
-
"loss": 0.0068,
|
996 |
-
"step": 1410
|
997 |
-
},
|
998 |
-
{
|
999 |
-
"epoch": 6.042553191489362,
|
1000 |
-
"grad_norm": 0.11437718570232391,
|
1001 |
-
"learning_rate": 0.00013005287145248878,
|
1002 |
-
"loss": 0.0064,
|
1003 |
-
"step": 1420
|
1004 |
-
},
|
1005 |
-
{
|
1006 |
-
"epoch": 6.085106382978723,
|
1007 |
-
"grad_norm": 0.06010650098323822,
|
1008 |
-
"learning_rate": 0.0001292016564005219,
|
1009 |
-
"loss": 0.0074,
|
1010 |
-
"step": 1430
|
1011 |
-
},
|
1012 |
-
{
|
1013 |
-
"epoch": 6.127659574468085,
|
1014 |
-
"grad_norm": 0.04595167934894562,
|
1015 |
-
"learning_rate": 0.0001283481218926818,
|
1016 |
-
"loss": 0.0066,
|
1017 |
-
"step": 1440
|
1018 |
-
},
|
1019 |
-
{
|
1020 |
-
"epoch": 6.170212765957447,
|
1021 |
-
"grad_norm": 0.05351310595870018,
|
1022 |
-
"learning_rate": 0.00012749233572428804,
|
1023 |
-
"loss": 0.0097,
|
1024 |
-
"step": 1450
|
1025 |
-
},
|
1026 |
-
{
|
1027 |
-
"epoch": 6.212765957446808,
|
1028 |
-
"grad_norm": 0.13630953431129456,
|
1029 |
-
"learning_rate": 0.00012663436586950714,
|
1030 |
-
"loss": 0.0079,
|
1031 |
-
"step": 1460
|
1032 |
-
},
|
1033 |
-
{
|
1034 |
-
"epoch": 6.25531914893617,
|
1035 |
-
"grad_norm": 0.17622588574886322,
|
1036 |
-
"learning_rate": 0.00012577428047595344,
|
1037 |
-
"loss": 0.0084,
|
1038 |
-
"step": 1470
|
1039 |
-
},
|
1040 |
-
{
|
1041 |
-
"epoch": 6.297872340425532,
|
1042 |
-
"grad_norm": 0.050954531878232956,
|
1043 |
-
"learning_rate": 0.0001249121478592762,
|
1044 |
-
"loss": 0.0077,
|
1045 |
-
"step": 1480
|
1046 |
-
},
|
1047 |
-
{
|
1048 |
-
"epoch": 6.340425531914893,
|
1049 |
-
"grad_norm": 0.3051726818084717,
|
1050 |
-
"learning_rate": 0.0001240480364977335,
|
1051 |
-
"loss": 0.0085,
|
1052 |
-
"step": 1490
|
1053 |
-
},
|
1054 |
-
{
|
1055 |
-
"epoch": 6.382978723404255,
|
1056 |
-
"grad_norm": 0.30509302020072937,
|
1057 |
-
"learning_rate": 0.00012318201502675285,
|
1058 |
-
"loss": 0.0092,
|
1059 |
-
"step": 1500
|
1060 |
-
},
|
1061 |
-
{
|
1062 |
-
"epoch": 6.425531914893617,
|
1063 |
-
"grad_norm": 0.09164142608642578,
|
1064 |
-
"learning_rate": 0.00012231415223347972,
|
1065 |
-
"loss": 0.008,
|
1066 |
-
"step": 1510
|
1067 |
-
},
|
1068 |
-
{
|
1069 |
-
"epoch": 6.468085106382979,
|
1070 |
-
"grad_norm": 0.05406223237514496,
|
1071 |
-
"learning_rate": 0.0001214445170513139,
|
1072 |
-
"loss": 0.0078,
|
1073 |
-
"step": 1520
|
1074 |
-
},
|
1075 |
-
{
|
1076 |
-
"epoch": 6.51063829787234,
|
1077 |
-
"grad_norm": 0.05845744535326958,
|
1078 |
-
"learning_rate": 0.00012057317855443395,
|
1079 |
-
"loss": 0.0092,
|
1080 |
-
"step": 1530
|
1081 |
-
},
|
1082 |
-
{
|
1083 |
-
"epoch": 6.553191489361702,
|
1084 |
-
"grad_norm": 0.05021122843027115,
|
1085 |
-
"learning_rate": 0.00011970020595231101,
|
1086 |
-
"loss": 0.007,
|
1087 |
-
"step": 1540
|
1088 |
-
},
|
1089 |
-
{
|
1090 |
-
"epoch": 6.595744680851064,
|
1091 |
-
"grad_norm": 0.10315235704183578,
|
1092 |
-
"learning_rate": 0.00011882566858421135,
|
1093 |
-
"loss": 0.0068,
|
1094 |
-
"step": 1550
|
1095 |
-
},
|
1096 |
-
{
|
1097 |
-
"epoch": 6.638297872340425,
|
1098 |
-
"grad_norm": 0.08750782907009125,
|
1099 |
-
"learning_rate": 0.00011794963591368893,
|
1100 |
-
"loss": 0.009,
|
1101 |
-
"step": 1560
|
1102 |
-
},
|
1103 |
-
{
|
1104 |
-
"epoch": 6.680851063829787,
|
1105 |
-
"grad_norm": 0.05412838235497475,
|
1106 |
-
"learning_rate": 0.0001170721775230679,
|
1107 |
-
"loss": 0.0071,
|
1108 |
-
"step": 1570
|
1109 |
-
},
|
1110 |
-
{
|
1111 |
-
"epoch": 6.723404255319149,
|
1112 |
-
"grad_norm": 0.17292432487010956,
|
1113 |
-
"learning_rate": 0.00011619336310791586,
|
1114 |
-
"loss": 0.0091,
|
1115 |
-
"step": 1580
|
1116 |
-
},
|
1117 |
-
{
|
1118 |
-
"epoch": 6.76595744680851,
|
1119 |
-
"grad_norm": 0.05503688380122185,
|
1120 |
-
"learning_rate": 0.00011531326247150803,
|
1121 |
-
"loss": 0.0069,
|
1122 |
-
"step": 1590
|
1123 |
-
},
|
1124 |
-
{
|
1125 |
-
"epoch": 6.808510638297872,
|
1126 |
-
"grad_norm": 0.05121155083179474,
|
1127 |
-
"learning_rate": 0.00011443194551928266,
|
1128 |
-
"loss": 0.008,
|
1129 |
-
"step": 1600
|
1130 |
-
},
|
1131 |
-
{
|
1132 |
-
"epoch": 6.851063829787234,
|
1133 |
-
"grad_norm": 0.0626005157828331,
|
1134 |
-
"learning_rate": 0.00011354948225328877,
|
1135 |
-
"loss": 0.0065,
|
1136 |
-
"step": 1610
|
1137 |
-
},
|
1138 |
-
{
|
1139 |
-
"epoch": 6.8936170212765955,
|
1140 |
-
"grad_norm": 0.058921247720718384,
|
1141 |
-
"learning_rate": 0.0001126659427666257,
|
1142 |
-
"loss": 0.0078,
|
1143 |
-
"step": 1620
|
1144 |
-
},
|
1145 |
-
{
|
1146 |
-
"epoch": 6.9361702127659575,
|
1147 |
-
"grad_norm": 0.058523211628198624,
|
1148 |
-
"learning_rate": 0.00011178139723787597,
|
1149 |
-
"loss": 0.008,
|
1150 |
-
"step": 1630
|
1151 |
-
},
|
1152 |
-
{
|
1153 |
-
"epoch": 6.9787234042553195,
|
1154 |
-
"grad_norm": 0.18594586849212646,
|
1155 |
-
"learning_rate": 0.00011089591592553082,
|
1156 |
-
"loss": 0.0076,
|
1157 |
-
"step": 1640
|
1158 |
-
},
|
1159 |
-
{
|
1160 |
-
"epoch": 7.0212765957446805,
|
1161 |
-
"grad_norm": 0.053747180849313736,
|
1162 |
-
"learning_rate": 0.00011000956916240985,
|
1163 |
-
"loss": 0.0074,
|
1164 |
-
"step": 1650
|
1165 |
-
},
|
1166 |
-
{
|
1167 |
-
"epoch": 7.0638297872340425,
|
1168 |
-
"grad_norm": 0.03964696079492569,
|
1169 |
-
"learning_rate": 0.00010912242735007441,
|
1170 |
-
"loss": 0.0071,
|
1171 |
-
"step": 1660
|
1172 |
-
},
|
1173 |
-
{
|
1174 |
-
"epoch": 7.1063829787234045,
|
1175 |
-
"grad_norm": 0.05952566862106323,
|
1176 |
-
"learning_rate": 0.00010823456095323579,
|
1177 |
-
"loss": 0.0065,
|
1178 |
-
"step": 1670
|
1179 |
-
},
|
1180 |
-
{
|
1181 |
-
"epoch": 7.148936170212766,
|
1182 |
-
"grad_norm": 0.21424749493598938,
|
1183 |
-
"learning_rate": 0.00010734604049415822,
|
1184 |
-
"loss": 0.0075,
|
1185 |
-
"step": 1680
|
1186 |
-
},
|
1187 |
-
{
|
1188 |
-
"epoch": 7.191489361702128,
|
1189 |
-
"grad_norm": 0.03922433406114578,
|
1190 |
-
"learning_rate": 0.0001064569365470574,
|
1191 |
-
"loss": 0.0071,
|
1192 |
-
"step": 1690
|
1193 |
-
},
|
1194 |
-
{
|
1195 |
-
"epoch": 7.23404255319149,
|
1196 |
-
"grad_norm": 0.05505215749144554,
|
1197 |
-
"learning_rate": 0.00010556731973249485,
|
1198 |
-
"loss": 0.0061,
|
1199 |
-
"step": 1700
|
1200 |
-
},
|
1201 |
-
{
|
1202 |
-
"epoch": 7.276595744680851,
|
1203 |
-
"grad_norm": 0.03699268028140068,
|
1204 |
-
"learning_rate": 0.00010467726071176853,
|
1205 |
-
"loss": 0.0075,
|
1206 |
-
"step": 1710
|
1207 |
-
},
|
1208 |
-
{
|
1209 |
-
"epoch": 7.319148936170213,
|
1210 |
-
"grad_norm": 0.04546520113945007,
|
1211 |
-
"learning_rate": 0.00010378683018130047,
|
1212 |
-
"loss": 0.0072,
|
1213 |
-
"step": 1720
|
1214 |
-
},
|
1215 |
-
{
|
1216 |
-
"epoch": 7.361702127659575,
|
1217 |
-
"grad_norm": 0.056984953582286835,
|
1218 |
-
"learning_rate": 0.0001028960988670212,
|
1219 |
-
"loss": 0.007,
|
1220 |
-
"step": 1730
|
1221 |
-
},
|
1222 |
-
{
|
1223 |
-
"epoch": 7.404255319148936,
|
1224 |
-
"grad_norm": 0.24714773893356323,
|
1225 |
-
"learning_rate": 0.00010200513751875227,
|
1226 |
-
"loss": 0.0074,
|
1227 |
-
"step": 1740
|
1228 |
-
},
|
1229 |
-
{
|
1230 |
-
"epoch": 7.446808510638298,
|
1231 |
-
"grad_norm": 0.06558862328529358,
|
1232 |
-
"learning_rate": 0.00010111401690458654,
|
1233 |
-
"loss": 0.0064,
|
1234 |
-
"step": 1750
|
1235 |
-
},
|
1236 |
-
{
|
1237 |
-
"epoch": 7.48936170212766,
|
1238 |
-
"grad_norm": 0.06254340708255768,
|
1239 |
-
"learning_rate": 0.00010022280780526725,
|
1240 |
-
"loss": 0.0076,
|
1241 |
-
"step": 1760
|
1242 |
-
},
|
1243 |
-
{
|
1244 |
-
"epoch": 7.531914893617021,
|
1245 |
-
"grad_norm": 0.08606445789337158,
|
1246 |
-
"learning_rate": 9.93315810085658e-05,
|
1247 |
-
"loss": 0.0076,
|
1248 |
-
"step": 1770
|
1249 |
-
},
|
1250 |
-
{
|
1251 |
-
"epoch": 7.574468085106383,
|
1252 |
-
"grad_norm": 0.051956657320261,
|
1253 |
-
"learning_rate": 9.844040730365936e-05,
|
1254 |
-
"loss": 0.0073,
|
1255 |
-
"step": 1780
|
1256 |
-
},
|
1257 |
-
{
|
1258 |
-
"epoch": 7.617021276595745,
|
1259 |
-
"grad_norm": 0.24819940328598022,
|
1260 |
-
"learning_rate": 9.754935747550804e-05,
|
1261 |
-
"loss": 0.0077,
|
1262 |
-
"step": 1790
|
1263 |
-
},
|
1264 |
-
{
|
1265 |
-
"epoch": 7.659574468085106,
|
1266 |
-
"grad_norm": 0.039045918732881546,
|
1267 |
-
"learning_rate": 9.665850229923258e-05,
|
1268 |
-
"loss": 0.0071,
|
1269 |
-
"step": 1800
|
1270 |
-
},
|
1271 |
-
{
|
1272 |
-
"epoch": 7.702127659574468,
|
1273 |
-
"grad_norm": 0.04438428208231926,
|
1274 |
-
"learning_rate": 9.57679125344927e-05,
|
1275 |
-
"loss": 0.0074,
|
1276 |
-
"step": 1810
|
1277 |
-
},
|
1278 |
-
{
|
1279 |
-
"epoch": 7.74468085106383,
|
1280 |
-
"grad_norm": 0.31067439913749695,
|
1281 |
-
"learning_rate": 9.487765891986682e-05,
|
1282 |
-
"loss": 0.0087,
|
1283 |
-
"step": 1820
|
1284 |
-
},
|
1285 |
-
{
|
1286 |
-
"epoch": 7.787234042553192,
|
1287 |
-
"grad_norm": 0.06251853704452515,
|
1288 |
-
"learning_rate": 9.398781216723331e-05,
|
1289 |
-
"loss": 0.0069,
|
1290 |
-
"step": 1830
|
1291 |
-
},
|
1292 |
-
{
|
1293 |
-
"epoch": 7.829787234042553,
|
1294 |
-
"grad_norm": 0.04706185683608055,
|
1295 |
-
"learning_rate": 9.309844295615389e-05,
|
1296 |
-
"loss": 0.0072,
|
1297 |
-
"step": 1840
|
1298 |
-
},
|
1299 |
-
{
|
1300 |
-
"epoch": 7.872340425531915,
|
1301 |
-
"grad_norm": 0.03839968144893646,
|
1302 |
-
"learning_rate": 9.220962192825968e-05,
|
1303 |
-
"loss": 0.0069,
|
1304 |
-
"step": 1850
|
1305 |
-
},
|
1306 |
-
{
|
1307 |
-
"epoch": 7.914893617021277,
|
1308 |
-
"grad_norm": 0.03722318261861801,
|
1309 |
-
"learning_rate": 9.132141968164026e-05,
|
1310 |
-
"loss": 0.0069,
|
1311 |
-
"step": 1860
|
1312 |
-
},
|
1313 |
-
{
|
1314 |
-
"epoch": 7.957446808510638,
|
1315 |
-
"grad_norm": 0.0543668158352375,
|
1316 |
-
"learning_rate": 9.043390676523604e-05,
|
1317 |
-
"loss": 0.0076,
|
1318 |
-
"step": 1870
|
1319 |
-
},
|
1320 |
-
{
|
1321 |
-
"epoch": 8.0,
|
1322 |
-
"grad_norm": 0.051696889102458954,
|
1323 |
-
"learning_rate": 8.954715367323468e-05,
|
1324 |
-
"loss": 0.0066,
|
1325 |
-
"step": 1880
|
1326 |
-
},
|
1327 |
-
{
|
1328 |
-
"epoch": 8.042553191489361,
|
1329 |
-
"grad_norm": 0.08786992728710175,
|
1330 |
-
"learning_rate": 8.866123083947182e-05,
|
1331 |
-
"loss": 0.0062,
|
1332 |
-
"step": 1890
|
1333 |
-
},
|
1334 |
-
{
|
1335 |
-
"epoch": 8.085106382978724,
|
1336 |
-
"grad_norm": 0.3556790351867676,
|
1337 |
-
"learning_rate": 8.777620863183657e-05,
|
1338 |
-
"loss": 0.0079,
|
1339 |
-
"step": 1900
|
1340 |
-
},
|
1341 |
-
{
|
1342 |
-
"epoch": 8.127659574468085,
|
1343 |
-
"grad_norm": 0.08049122244119644,
|
1344 |
-
"learning_rate": 8.689215734668232e-05,
|
1345 |
-
"loss": 0.0064,
|
1346 |
-
"step": 1910
|
1347 |
-
},
|
1348 |
-
{
|
1349 |
-
"epoch": 8.170212765957446,
|
1350 |
-
"grad_norm": 0.05408492311835289,
|
1351 |
-
"learning_rate": 8.600914720324316e-05,
|
1352 |
-
"loss": 0.0077,
|
1353 |
-
"step": 1920
|
1354 |
-
},
|
1355 |
-
{
|
1356 |
-
"epoch": 8.212765957446809,
|
1357 |
-
"grad_norm": 0.034946855157613754,
|
1358 |
-
"learning_rate": 8.512724833805634e-05,
|
1359 |
-
"loss": 0.0067,
|
1360 |
-
"step": 1930
|
1361 |
-
},
|
1362 |
-
{
|
1363 |
-
"epoch": 8.25531914893617,
|
1364 |
-
"grad_norm": 0.05422681197524071,
|
1365 |
-
"learning_rate": 8.424653079939156e-05,
|
1366 |
-
"loss": 0.0062,
|
1367 |
-
"step": 1940
|
1368 |
-
},
|
1369 |
-
{
|
1370 |
-
"epoch": 8.297872340425531,
|
1371 |
-
"grad_norm": 0.06405791640281677,
|
1372 |
-
"learning_rate": 8.336706454168701e-05,
|
1373 |
-
"loss": 0.0064,
|
1374 |
-
"step": 1950
|
1375 |
-
},
|
1376 |
-
{
|
1377 |
-
"epoch": 8.340425531914894,
|
1378 |
-
"grad_norm": 0.08694509416818619,
|
1379 |
-
"learning_rate": 8.248891941999297e-05,
|
1380 |
-
"loss": 0.006,
|
1381 |
-
"step": 1960
|
1382 |
-
},
|
1383 |
-
{
|
1384 |
-
"epoch": 8.382978723404255,
|
1385 |
-
"grad_norm": 0.05589594319462776,
|
1386 |
-
"learning_rate": 8.161216518442334e-05,
|
1387 |
-
"loss": 0.0067,
|
1388 |
-
"step": 1970
|
1389 |
-
},
|
1390 |
-
{
|
1391 |
-
"epoch": 8.425531914893616,
|
1392 |
-
"grad_norm": 0.05914654955267906,
|
1393 |
-
"learning_rate": 8.073687147461547e-05,
|
1394 |
-
"loss": 0.0065,
|
1395 |
-
"step": 1980
|
1396 |
-
},
|
1397 |
-
{
|
1398 |
-
"epoch": 8.46808510638298,
|
1399 |
-
"grad_norm": 0.05237606540322304,
|
1400 |
-
"learning_rate": 7.98631078141987e-05,
|
1401 |
-
"loss": 0.0071,
|
1402 |
-
"step": 1990
|
1403 |
-
},
|
1404 |
-
{
|
1405 |
-
"epoch": 8.51063829787234,
|
1406 |
-
"grad_norm": 0.056138359010219574,
|
1407 |
-
"learning_rate": 7.89909436052722e-05,
|
1408 |
-
"loss": 0.0079,
|
1409 |
-
"step": 2000
|
1410 |
-
},
|
1411 |
-
{
|
1412 |
-
"epoch": 8.553191489361701,
|
1413 |
-
"grad_norm": 0.06285829097032547,
|
1414 |
-
"learning_rate": 7.812044812289249e-05,
|
1415 |
-
"loss": 0.0064,
|
1416 |
-
"step": 2010
|
1417 |
-
},
|
1418 |
-
{
|
1419 |
-
"epoch": 8.595744680851064,
|
1420 |
-
"grad_norm": 0.04014954715967178,
|
1421 |
-
"learning_rate": 7.72516905095709e-05,
|
1422 |
-
"loss": 0.0072,
|
1423 |
-
"step": 2020
|
1424 |
-
},
|
1425 |
-
{
|
1426 |
-
"epoch": 8.638297872340425,
|
1427 |
-
"grad_norm": 0.04423344135284424,
|
1428 |
-
"learning_rate": 7.638473976978177e-05,
|
1429 |
-
"loss": 0.0065,
|
1430 |
-
"step": 2030
|
1431 |
-
},
|
1432 |
-
{
|
1433 |
-
"epoch": 8.680851063829786,
|
1434 |
-
"grad_norm": 0.05287999287247658,
|
1435 |
-
"learning_rate": 7.55196647644814e-05,
|
1436 |
-
"loss": 0.0081,
|
1437 |
-
"step": 2040
|
1438 |
-
},
|
1439 |
-
{
|
1440 |
-
"epoch": 8.72340425531915,
|
1441 |
-
"grad_norm": 0.056493304669857025,
|
1442 |
-
"learning_rate": 7.465653420563845e-05,
|
1443 |
-
"loss": 0.0067,
|
1444 |
-
"step": 2050
|
1445 |
-
},
|
1446 |
-
{
|
1447 |
-
"epoch": 8.76595744680851,
|
1448 |
-
"grad_norm": 0.057281821966171265,
|
1449 |
-
"learning_rate": 7.379541665077643e-05,
|
1450 |
-
"loss": 0.0078,
|
1451 |
-
"step": 2060
|
1452 |
-
},
|
1453 |
-
{
|
1454 |
-
"epoch": 8.808510638297872,
|
1455 |
-
"grad_norm": 0.053731102496385574,
|
1456 |
-
"learning_rate": 7.293638049752812e-05,
|
1457 |
-
"loss": 0.0066,
|
1458 |
-
"step": 2070
|
1459 |
-
},
|
1460 |
-
{
|
1461 |
-
"epoch": 8.851063829787234,
|
1462 |
-
"grad_norm": 0.18697482347488403,
|
1463 |
-
"learning_rate": 7.207949397820278e-05,
|
1464 |
-
"loss": 0.0069,
|
1465 |
-
"step": 2080
|
1466 |
-
},
|
1467 |
-
{
|
1468 |
-
"epoch": 8.893617021276595,
|
1469 |
-
"grad_norm": 0.0387713797390461,
|
1470 |
-
"learning_rate": 7.122482515436661e-05,
|
1471 |
-
"loss": 0.007,
|
1472 |
-
"step": 2090
|
1473 |
-
},
|
1474 |
-
{
|
1475 |
-
"epoch": 8.936170212765958,
|
1476 |
-
"grad_norm": 0.04441935196518898,
|
1477 |
-
"learning_rate": 7.037244191143661e-05,
|
1478 |
-
"loss": 0.0067,
|
1479 |
-
"step": 2100
|
1480 |
-
},
|
1481 |
-
{
|
1482 |
-
"epoch": 8.97872340425532,
|
1483 |
-
"grad_norm": 0.05758345127105713,
|
1484 |
-
"learning_rate": 6.952241195328868e-05,
|
1485 |
-
"loss": 0.0065,
|
1486 |
-
"step": 2110
|
1487 |
-
},
|
1488 |
-
{
|
1489 |
-
"epoch": 9.02127659574468,
|
1490 |
-
"grad_norm": 0.050706226378679276,
|
1491 |
-
"learning_rate": 6.867480279687974e-05,
|
1492 |
-
"loss": 0.0063,
|
1493 |
-
"step": 2120
|
1494 |
-
},
|
1495 |
-
{
|
1496 |
-
"epoch": 9.063829787234043,
|
1497 |
-
"grad_norm": 0.05180887505412102,
|
1498 |
-
"learning_rate": 6.782968176688514e-05,
|
1499 |
-
"loss": 0.0062,
|
1500 |
-
"step": 2130
|
1501 |
-
},
|
1502 |
-
{
|
1503 |
-
"epoch": 9.106382978723405,
|
1504 |
-
"grad_norm": 0.05492401868104935,
|
1505 |
-
"learning_rate": 6.6987115990351e-05,
|
1506 |
-
"loss": 0.006,
|
1507 |
-
"step": 2140
|
1508 |
-
},
|
1509 |
-
{
|
1510 |
-
"epoch": 9.148936170212766,
|
1511 |
-
"grad_norm": 0.053439777344465256,
|
1512 |
-
"learning_rate": 6.614717239136246e-05,
|
1513 |
-
"loss": 0.0066,
|
1514 |
-
"step": 2150
|
1515 |
-
},
|
1516 |
-
{
|
1517 |
-
"epoch": 9.191489361702128,
|
1518 |
-
"grad_norm": 0.15650008618831635,
|
1519 |
-
"learning_rate": 6.530991768572794e-05,
|
1520 |
-
"loss": 0.006,
|
1521 |
-
"step": 2160
|
1522 |
-
},
|
1523 |
-
{
|
1524 |
-
"epoch": 9.23404255319149,
|
1525 |
-
"grad_norm": 0.04846300184726715,
|
1526 |
-
"learning_rate": 6.447541837568e-05,
|
1527 |
-
"loss": 0.0068,
|
1528 |
-
"step": 2170
|
1529 |
-
},
|
1530 |
-
{
|
1531 |
-
"epoch": 9.27659574468085,
|
1532 |
-
"grad_norm": 0.047157011926174164,
|
1533 |
-
"learning_rate": 6.364374074459307e-05,
|
1534 |
-
"loss": 0.006,
|
1535 |
-
"step": 2180
|
1536 |
-
},
|
1537 |
-
{
|
1538 |
-
"epoch": 9.319148936170214,
|
1539 |
-
"grad_norm": 0.047358132898807526,
|
1540 |
-
"learning_rate": 6.281495085171869e-05,
|
1541 |
-
"loss": 0.0058,
|
1542 |
-
"step": 2190
|
1543 |
-
},
|
1544 |
-
{
|
1545 |
-
"epoch": 9.361702127659575,
|
1546 |
-
"grad_norm": 0.11200093477964401,
|
1547 |
-
"learning_rate": 6.198911452693853e-05,
|
1548 |
-
"loss": 0.007,
|
1549 |
-
"step": 2200
|
1550 |
-
},
|
1551 |
-
{
|
1552 |
-
"epoch": 9.404255319148936,
|
1553 |
-
"grad_norm": 0.06481984257698059,
|
1554 |
-
"learning_rate": 6.116629736553552e-05,
|
1555 |
-
"loss": 0.0069,
|
1556 |
-
"step": 2210
|
1557 |
-
},
|
1558 |
-
{
|
1559 |
-
"epoch": 9.446808510638299,
|
1560 |
-
"grad_norm": 0.04743931442499161,
|
1561 |
-
"learning_rate": 6.0346564722983736e-05,
|
1562 |
-
"loss": 0.0072,
|
1563 |
-
"step": 2220
|
1564 |
-
},
|
1565 |
-
{
|
1566 |
-
"epoch": 9.48936170212766,
|
1567 |
-
"grad_norm": 0.07307924330234528,
|
1568 |
-
"learning_rate": 5.952998170975724e-05,
|
1569 |
-
"loss": 0.0062,
|
1570 |
-
"step": 2230
|
1571 |
-
},
|
1572 |
-
{
|
1573 |
-
"epoch": 9.53191489361702,
|
1574 |
-
"grad_norm": 0.05650079995393753,
|
1575 |
-
"learning_rate": 5.871661318615848e-05,
|
1576 |
-
"loss": 0.0061,
|
1577 |
-
"step": 2240
|
1578 |
-
},
|
1579 |
-
{
|
1580 |
-
"epoch": 9.574468085106384,
|
1581 |
-
"grad_norm": 0.057734813541173935,
|
1582 |
-
"learning_rate": 5.790652375716652e-05,
|
1583 |
-
"loss": 0.0068,
|
1584 |
-
"step": 2250
|
1585 |
-
},
|
1586 |
-
{
|
1587 |
-
"epoch": 9.617021276595745,
|
1588 |
-
"grad_norm": 0.05608903244137764,
|
1589 |
-
"learning_rate": 5.709977776730537e-05,
|
1590 |
-
"loss": 0.0071,
|
1591 |
-
"step": 2260
|
1592 |
-
},
|
1593 |
-
{
|
1594 |
-
"epoch": 9.659574468085106,
|
1595 |
-
"grad_norm": 0.07133158296346664,
|
1596 |
-
"learning_rate": 5.62964392955335e-05,
|
1597 |
-
"loss": 0.0067,
|
1598 |
-
"step": 2270
|
1599 |
-
},
|
1600 |
-
{
|
1601 |
-
"epoch": 9.702127659574469,
|
1602 |
-
"grad_norm": 0.2167043685913086,
|
1603 |
-
"learning_rate": 5.549657215015367e-05,
|
1604 |
-
"loss": 0.0067,
|
1605 |
-
"step": 2280
|
1606 |
-
},
|
1607 |
-
{
|
1608 |
-
"epoch": 9.74468085106383,
|
1609 |
-
"grad_norm": 0.06523007154464722,
|
1610 |
-
"learning_rate": 5.470023986374516e-05,
|
1611 |
-
"loss": 0.0068,
|
1612 |
-
"step": 2290
|
1613 |
-
},
|
1614 |
-
{
|
1615 |
-
"epoch": 9.787234042553191,
|
1616 |
-
"grad_norm": 0.04895370826125145,
|
1617 |
-
"learning_rate": 5.39075056881172e-05,
|
1618 |
-
"loss": 0.0058,
|
1619 |
-
"step": 2300
|
1620 |
-
},
|
1621 |
-
{
|
1622 |
-
"epoch": 9.829787234042554,
|
1623 |
-
"grad_norm": 0.056433409452438354,
|
1624 |
-
"learning_rate": 5.31184325892849e-05,
|
1625 |
-
"loss": 0.0067,
|
1626 |
-
"step": 2310
|
1627 |
-
},
|
1628 |
-
{
|
1629 |
-
"epoch": 9.872340425531915,
|
1630 |
-
"grad_norm": 0.049970466643571854,
|
1631 |
-
"learning_rate": 5.233308324246805e-05,
|
1632 |
-
"loss": 0.0058,
|
1633 |
-
"step": 2320
|
1634 |
-
},
|
1635 |
-
{
|
1636 |
-
"epoch": 9.914893617021276,
|
1637 |
-
"grad_norm": 0.08367093652486801,
|
1638 |
-
"learning_rate": 5.155152002711285e-05,
|
1639 |
-
"loss": 0.0063,
|
1640 |
-
"step": 2330
|
1641 |
-
},
|
1642 |
-
{
|
1643 |
-
"epoch": 9.957446808510639,
|
1644 |
-
"grad_norm": 0.07821632921695709,
|
1645 |
-
"learning_rate": 5.077380502193725e-05,
|
1646 |
-
"loss": 0.0071,
|
1647 |
-
"step": 2340
|
1648 |
-
},
|
1649 |
-
{
|
1650 |
-
"epoch": 10.0,
|
1651 |
-
"grad_norm": 0.04344159737229347,
|
1652 |
-
"learning_rate": 5.000000000000002e-05,
|
1653 |
-
"loss": 0.0078,
|
1654 |
-
"step": 2350
|
1655 |
-
},
|
1656 |
-
{
|
1657 |
-
"epoch": 10.042553191489361,
|
1658 |
-
"grad_norm": 0.06118736043572426,
|
1659 |
-
"learning_rate": 4.923016642379412e-05,
|
1660 |
-
"loss": 0.0061,
|
1661 |
-
"step": 2360
|
1662 |
-
},
|
1663 |
-
{
|
1664 |
-
"epoch": 10.085106382978724,
|
1665 |
-
"grad_norm": 0.07500752061605453,
|
1666 |
-
"learning_rate": 4.8464365440365044e-05,
|
1667 |
-
"loss": 0.006,
|
1668 |
-
"step": 2370
|
1669 |
-
},
|
1670 |
-
{
|
1671 |
-
"epoch": 10.127659574468085,
|
1672 |
-
"grad_norm": 0.05192973092198372,
|
1673 |
-
"learning_rate": 4.7702657876453616e-05,
|
1674 |
-
"loss": 0.0059,
|
1675 |
-
"step": 2380
|
1676 |
-
},
|
1677 |
-
{
|
1678 |
-
"epoch": 10.170212765957446,
|
1679 |
-
"grad_norm": 0.07265878468751907,
|
1680 |
-
"learning_rate": 4.6945104233665006e-05,
|
1681 |
-
"loss": 0.006,
|
1682 |
-
"step": 2390
|
1683 |
-
},
|
1684 |
-
{
|
1685 |
-
"epoch": 10.212765957446809,
|
1686 |
-
"grad_norm": 0.07713906466960907,
|
1687 |
-
"learning_rate": 4.6191764683662744e-05,
|
1688 |
-
"loss": 0.0065,
|
1689 |
-
"step": 2400
|
1690 |
-
},
|
1691 |
-
{
|
1692 |
-
"epoch": 10.25531914893617,
|
1693 |
-
"grad_norm": 0.1284199059009552,
|
1694 |
-
"learning_rate": 4.5442699063389705e-05,
|
1695 |
-
"loss": 0.0061,
|
1696 |
-
"step": 2410
|
1697 |
-
},
|
1698 |
-
{
|
1699 |
-
"epoch": 10.297872340425531,
|
1700 |
-
"grad_norm": 0.03467699885368347,
|
1701 |
-
"learning_rate": 4.469796687031502e-05,
|
1702 |
-
"loss": 0.0056,
|
1703 |
-
"step": 2420
|
1704 |
-
},
|
1705 |
-
{
|
1706 |
-
"epoch": 10.340425531914894,
|
1707 |
-
"grad_norm": 0.05337784066796303,
|
1708 |
-
"learning_rate": 4.395762725770852e-05,
|
1709 |
-
"loss": 0.006,
|
1710 |
-
"step": 2430
|
1711 |
-
},
|
1712 |
-
{
|
1713 |
-
"epoch": 10.382978723404255,
|
1714 |
-
"grad_norm": 0.054092224687337875,
|
1715 |
-
"learning_rate": 4.322173902994212e-05,
|
1716 |
-
"loss": 0.0065,
|
1717 |
-
"step": 2440
|
1718 |
-
},
|
1719 |
-
{
|
1720 |
-
"epoch": 10.425531914893616,
|
1721 |
-
"grad_norm": 0.044108908623456955,
|
1722 |
-
"learning_rate": 4.249036063781896e-05,
|
1723 |
-
"loss": 0.0058,
|
1724 |
-
"step": 2450
|
1725 |
-
},
|
1726 |
-
{
|
1727 |
-
"epoch": 10.46808510638298,
|
1728 |
-
"grad_norm": 0.1287708580493927,
|
1729 |
-
"learning_rate": 4.176355017393099e-05,
|
1730 |
-
"loss": 0.0062,
|
1731 |
-
"step": 2460
|
1732 |
-
},
|
1733 |
-
{
|
1734 |
-
"epoch": 10.51063829787234,
|
1735 |
-
"grad_norm": 0.03487354889512062,
|
1736 |
-
"learning_rate": 4.10413653680444e-05,
|
1737 |
-
"loss": 0.0065,
|
1738 |
-
"step": 2470
|
1739 |
-
},
|
1740 |
-
{
|
1741 |
-
"epoch": 10.553191489361701,
|
1742 |
-
"grad_norm": 0.06098255515098572,
|
1743 |
-
"learning_rate": 4.03238635825146e-05,
|
1744 |
-
"loss": 0.0058,
|
1745 |
-
"step": 2480
|
1746 |
-
},
|
1747 |
-
{
|
1748 |
-
"epoch": 10.595744680851064,
|
1749 |
-
"grad_norm": 0.04999317228794098,
|
1750 |
-
"learning_rate": 3.961110180772955e-05,
|
1751 |
-
"loss": 0.0059,
|
1752 |
-
"step": 2490
|
1753 |
-
},
|
1754 |
-
{
|
1755 |
-
"epoch": 10.638297872340425,
|
1756 |
-
"grad_norm": 0.05019402131438255,
|
1757 |
-
"learning_rate": 3.890313665758348e-05,
|
1758 |
-
"loss": 0.0061,
|
1759 |
-
"step": 2500
|
1760 |
-
},
|
1761 |
-
{
|
1762 |
-
"epoch": 10.680851063829786,
|
1763 |
-
"grad_norm": 0.0514717735350132,
|
1764 |
-
"learning_rate": 3.8200024364979815e-05,
|
1765 |
-
"loss": 0.006,
|
1766 |
-
"step": 2510
|
1767 |
-
},
|
1768 |
-
{
|
1769 |
-
"epoch": 10.72340425531915,
|
1770 |
-
"grad_norm": 0.048151805996894836,
|
1771 |
-
"learning_rate": 3.750182077736486e-05,
|
1772 |
-
"loss": 0.0064,
|
1773 |
-
"step": 2520
|
1774 |
-
},
|
1775 |
-
{
|
1776 |
-
"epoch": 10.76595744680851,
|
1777 |
-
"grad_norm": 0.08788628876209259,
|
1778 |
-
"learning_rate": 3.6808581352291716e-05,
|
1779 |
-
"loss": 0.006,
|
1780 |
-
"step": 2530
|
1781 |
-
},
|
1782 |
-
{
|
1783 |
-
"epoch": 10.808510638297872,
|
1784 |
-
"grad_norm": 0.08574005961418152,
|
1785 |
-
"learning_rate": 3.612036115301551e-05,
|
1786 |
-
"loss": 0.0064,
|
1787 |
-
"step": 2540
|
1788 |
-
},
|
1789 |
-
{
|
1790 |
-
"epoch": 10.851063829787234,
|
1791 |
-
"grad_norm": 0.05240662768483162,
|
1792 |
-
"learning_rate": 3.543721484411976e-05,
|
1793 |
-
"loss": 0.0065,
|
1794 |
-
"step": 2550
|
1795 |
-
},
|
1796 |
-
{
|
1797 |
-
"epoch": 10.893617021276595,
|
1798 |
-
"grad_norm": 0.05454961955547333,
|
1799 |
-
"learning_rate": 3.475919668717422e-05,
|
1800 |
-
"loss": 0.006,
|
1801 |
-
"step": 2560
|
1802 |
-
},
|
1803 |
-
{
|
1804 |
-
"epoch": 10.936170212765958,
|
1805 |
-
"grad_norm": 0.052435703575611115,
|
1806 |
-
"learning_rate": 3.408636053642528e-05,
|
1807 |
-
"loss": 0.0058,
|
1808 |
-
"step": 2570
|
1809 |
-
},
|
1810 |
-
{
|
1811 |
-
"epoch": 10.97872340425532,
|
1812 |
-
"grad_norm": 0.09059314429759979,
|
1813 |
-
"learning_rate": 3.3418759834518056e-05,
|
1814 |
-
"loss": 0.0061,
|
1815 |
-
"step": 2580
|
1816 |
-
},
|
1817 |
-
{
|
1818 |
-
"epoch": 11.02127659574468,
|
1819 |
-
"grad_norm": 0.04699910059571266,
|
1820 |
-
"learning_rate": 3.275644760825168e-05,
|
1821 |
-
"loss": 0.006,
|
1822 |
-
"step": 2590
|
1823 |
-
},
|
1824 |
-
{
|
1825 |
-
"epoch": 11.063829787234043,
|
1826 |
-
"grad_norm": 0.0523945577442646,
|
1827 |
-
"learning_rate": 3.209947646436752e-05,
|
1828 |
-
"loss": 0.0053,
|
1829 |
-
"step": 2600
|
1830 |
-
},
|
1831 |
-
{
|
1832 |
-
"epoch": 11.106382978723405,
|
1833 |
-
"grad_norm": 0.09489481896162033,
|
1834 |
-
"learning_rate": 3.1447898585370384e-05,
|
1835 |
-
"loss": 0.0061,
|
1836 |
-
"step": 2610
|
1837 |
-
},
|
1838 |
-
{
|
1839 |
-
"epoch": 11.148936170212766,
|
1840 |
-
"grad_norm": 0.04296072944998741,
|
1841 |
-
"learning_rate": 3.0801765725384066e-05,
|
1842 |
-
"loss": 0.006,
|
1843 |
-
"step": 2620
|
1844 |
-
},
|
1845 |
-
{
|
1846 |
-
"epoch": 11.191489361702128,
|
1847 |
-
"grad_norm": 0.05972497537732124,
|
1848 |
-
"learning_rate": 3.0161129206040284e-05,
|
1849 |
-
"loss": 0.0061,
|
1850 |
-
"step": 2630
|
1851 |
-
},
|
1852 |
-
{
|
1853 |
-
"epoch": 11.23404255319149,
|
1854 |
-
"grad_norm": 0.06595079600811005,
|
1855 |
-
"learning_rate": 2.9526039912402503e-05,
|
1856 |
-
"loss": 0.0058,
|
1857 |
-
"step": 2640
|
1858 |
-
},
|
1859 |
-
{
|
1860 |
-
"epoch": 11.27659574468085,
|
1861 |
-
"grad_norm": 0.04778233543038368,
|
1862 |
-
"learning_rate": 2.889654828892393e-05,
|
1863 |
-
"loss": 0.0059,
|
1864 |
-
"step": 2650
|
1865 |
-
},
|
1866 |
-
{
|
1867 |
-
"epoch": 11.319148936170214,
|
1868 |
-
"grad_norm": 0.06364385783672333,
|
1869 |
-
"learning_rate": 2.8272704335441058e-05,
|
1870 |
-
"loss": 0.0059,
|
1871 |
-
"step": 2660
|
1872 |
-
},
|
1873 |
-
{
|
1874 |
-
"epoch": 11.361702127659575,
|
1875 |
-
"grad_norm": 0.04678649827837944,
|
1876 |
-
"learning_rate": 2.7654557603201957e-05,
|
1877 |
-
"loss": 0.0058,
|
1878 |
-
"step": 2670
|
1879 |
-
},
|
1880 |
-
{
|
1881 |
-
"epoch": 11.404255319148936,
|
1882 |
-
"grad_norm": 0.06067119911313057,
|
1883 |
-
"learning_rate": 2.704215719093066e-05,
|
1884 |
-
"loss": 0.0062,
|
1885 |
-
"step": 2680
|
1886 |
-
},
|
1887 |
-
{
|
1888 |
-
"epoch": 11.446808510638299,
|
1889 |
-
"grad_norm": 0.05602847784757614,
|
1890 |
-
"learning_rate": 2.643555174092728e-05,
|
1891 |
-
"loss": 0.0064,
|
1892 |
-
"step": 2690
|
1893 |
-
},
|
1894 |
-
{
|
1895 |
-
"epoch": 11.48936170212766,
|
1896 |
-
"grad_norm": 0.06131875514984131,
|
1897 |
-
"learning_rate": 2.5834789435204243e-05,
|
1898 |
-
"loss": 0.0059,
|
1899 |
-
"step": 2700
|
1900 |
-
},
|
1901 |
-
{
|
1902 |
-
"epoch": 11.53191489361702,
|
1903 |
-
"grad_norm": 0.049852412194013596,
|
1904 |
-
"learning_rate": 2.5239917991659512e-05,
|
1905 |
-
"loss": 0.006,
|
1906 |
-
"step": 2710
|
1907 |
-
},
|
1908 |
-
{
|
1909 |
-
"epoch": 11.574468085106384,
|
1910 |
-
"grad_norm": 0.06347551941871643,
|
1911 |
-
"learning_rate": 2.4650984660286124e-05,
|
1912 |
-
"loss": 0.0058,
|
1913 |
-
"step": 2720
|
1914 |
-
},
|
1915 |
-
{
|
1916 |
-
"epoch": 11.617021276595745,
|
1917 |
-
"grad_norm": 0.0802001804113388,
|
1918 |
-
"learning_rate": 2.4068036219419432e-05,
|
1919 |
-
"loss": 0.0063,
|
1920 |
-
"step": 2730
|
1921 |
-
},
|
1922 |
-
{
|
1923 |
-
"epoch": 11.659574468085106,
|
1924 |
-
"grad_norm": 0.15571121871471405,
|
1925 |
-
"learning_rate": 2.349111897202134e-05,
|
1926 |
-
"loss": 0.0061,
|
1927 |
-
"step": 2740
|
1928 |
-
},
|
1929 |
-
{
|
1930 |
-
"epoch": 11.702127659574469,
|
1931 |
-
"grad_norm": 0.05023673549294472,
|
1932 |
-
"learning_rate": 2.2920278742002676e-05,
|
1933 |
-
"loss": 0.0058,
|
1934 |
-
"step": 2750
|
1935 |
-
},
|
1936 |
-
{
|
1937 |
-
"epoch": 11.74468085106383,
|
1938 |
-
"grad_norm": 0.06210224702954292,
|
1939 |
-
"learning_rate": 2.235556087058328e-05,
|
1940 |
-
"loss": 0.0057,
|
1941 |
-
"step": 2760
|
1942 |
-
},
|
1943 |
-
{
|
1944 |
-
"epoch": 11.787234042553191,
|
1945 |
-
"grad_norm": 0.054550446569919586,
|
1946 |
-
"learning_rate": 2.1797010212690795e-05,
|
1947 |
-
"loss": 0.0059,
|
1948 |
-
"step": 2770
|
1949 |
-
},
|
1950 |
-
{
|
1951 |
-
"epoch": 11.829787234042554,
|
1952 |
-
"grad_norm": 0.06770533323287964,
|
1953 |
-
"learning_rate": 2.1244671133397753e-05,
|
1954 |
-
"loss": 0.0058,
|
1955 |
-
"step": 2780
|
1956 |
-
},
|
1957 |
-
{
|
1958 |
-
"epoch": 11.872340425531915,
|
1959 |
-
"grad_norm": 0.07317465543746948,
|
1960 |
-
"learning_rate": 2.069858750439768e-05,
|
1961 |
-
"loss": 0.0061,
|
1962 |
-
"step": 2790
|
1963 |
-
},
|
1964 |
-
{
|
1965 |
-
"epoch": 11.914893617021276,
|
1966 |
-
"grad_norm": 0.04136810824275017,
|
1967 |
-
"learning_rate": 2.0158802700520574e-05,
|
1968 |
-
"loss": 0.0061,
|
1969 |
-
"step": 2800
|
1970 |
-
},
|
1971 |
-
{
|
1972 |
-
"epoch": 11.957446808510639,
|
1973 |
-
"grad_norm": 0.06249478831887245,
|
1974 |
-
"learning_rate": 1.9625359596287496e-05,
|
1975 |
-
"loss": 0.0061,
|
1976 |
-
"step": 2810
|
1977 |
-
},
|
1978 |
-
{
|
1979 |
-
"epoch": 12.0,
|
1980 |
-
"grad_norm": 0.07229502499103546,
|
1981 |
-
"learning_rate": 1.9098300562505266e-05,
|
1982 |
-
"loss": 0.0061,
|
1983 |
-
"step": 2820
|
1984 |
-
},
|
1985 |
-
{
|
1986 |
-
"epoch": 12.042553191489361,
|
1987 |
-
"grad_norm": 0.05772854760289192,
|
1988 |
-
"learning_rate": 1.857766746290084e-05,
|
1989 |
-
"loss": 0.0059,
|
1990 |
-
"step": 2830
|
1991 |
-
},
|
1992 |
-
{
|
1993 |
-
"epoch": 12.085106382978724,
|
1994 |
-
"grad_norm": 0.05857497826218605,
|
1995 |
-
"learning_rate": 1.8063501650796288e-05,
|
1996 |
-
"loss": 0.0053,
|
1997 |
-
"step": 2840
|
1998 |
-
},
|
1999 |
-
{
|
2000 |
-
"epoch": 12.127659574468085,
|
2001 |
-
"grad_norm": 0.0488237701356411,
|
2002 |
-
"learning_rate": 1.7555843965823992e-05,
|
2003 |
-
"loss": 0.006,
|
2004 |
-
"step": 2850
|
2005 |
-
},
|
2006 |
-
{
|
2007 |
-
"epoch": 12.170212765957446,
|
2008 |
-
"grad_norm": 0.04676933214068413,
|
2009 |
-
"learning_rate": 1.705473473068282e-05,
|
2010 |
-
"loss": 0.0056,
|
2011 |
-
"step": 2860
|
2012 |
-
},
|
2013 |
-
{
|
2014 |
-
"epoch": 12.212765957446809,
|
2015 |
-
"grad_norm": 0.04721114784479141,
|
2016 |
-
"learning_rate": 1.6560213747935503e-05,
|
2017 |
-
"loss": 0.0062,
|
2018 |
-
"step": 2870
|
2019 |
-
},
|
2020 |
-
{
|
2021 |
-
"epoch": 12.25531914893617,
|
2022 |
-
"grad_norm": 0.055576860904693604,
|
2023 |
-
"learning_rate": 1.6072320296846898e-05,
|
2024 |
-
"loss": 0.0058,
|
2025 |
-
"step": 2880
|
2026 |
-
},
|
2027 |
-
{
|
2028 |
-
"epoch": 12.297872340425531,
|
2029 |
-
"grad_norm": 0.06054997444152832,
|
2030 |
-
"learning_rate": 1.5591093130264334e-05,
|
2031 |
-
"loss": 0.0057,
|
2032 |
-
"step": 2890
|
2033 |
-
},
|
2034 |
-
{
|
2035 |
-
"epoch": 12.340425531914894,
|
2036 |
-
"grad_norm": 0.054234158247709274,
|
2037 |
-
"learning_rate": 1.5116570471539293e-05,
|
2038 |
-
"loss": 0.0057,
|
2039 |
-
"step": 2900
|
2040 |
-
},
|
2041 |
-
{
|
2042 |
-
"epoch": 12.382978723404255,
|
2043 |
-
"grad_norm": 0.05556226149201393,
|
2044 |
-
"learning_rate": 1.4648790011491542e-05,
|
2045 |
-
"loss": 0.0057,
|
2046 |
-
"step": 2910
|
2047 |
-
},
|
2048 |
-
{
|
2049 |
-
"epoch": 12.425531914893616,
|
2050 |
-
"grad_norm": 0.055104803293943405,
|
2051 |
-
"learning_rate": 1.4187788905415334e-05,
|
2052 |
-
"loss": 0.0059,
|
2053 |
-
"step": 2920
|
2054 |
-
},
|
2055 |
-
{
|
2056 |
-
"epoch": 12.46808510638298,
|
2057 |
-
"grad_norm": 0.04338991269469261,
|
2058 |
-
"learning_rate": 1.3733603770128112e-05,
|
2059 |
-
"loss": 0.0059,
|
2060 |
-
"step": 2930
|
2061 |
-
},
|
2062 |
-
{
|
2063 |
-
"epoch": 12.51063829787234,
|
2064 |
-
"grad_norm": 0.04588828980922699,
|
2065 |
-
"learning_rate": 1.3286270681062274e-05,
|
2066 |
-
"loss": 0.006,
|
2067 |
-
"step": 2940
|
2068 |
-
},
|
2069 |
-
{
|
2070 |
-
"epoch": 12.553191489361701,
|
2071 |
-
"grad_norm": 0.05913781002163887,
|
2072 |
-
"learning_rate": 1.2845825169399507e-05,
|
2073 |
-
"loss": 0.0063,
|
2074 |
-
"step": 2950
|
2075 |
-
},
|
2076 |
-
{
|
2077 |
-
"epoch": 12.595744680851064,
|
2078 |
-
"grad_norm": 0.056119441986083984,
|
2079 |
-
"learning_rate": 1.2412302219248805e-05,
|
2080 |
-
"loss": 0.0059,
|
2081 |
-
"step": 2960
|
2082 |
-
},
|
2083 |
-
{
|
2084 |
-
"epoch": 12.638297872340425,
|
2085 |
-
"grad_norm": 0.05296126753091812,
|
2086 |
-
"learning_rate": 1.1985736264867509e-05,
|
2087 |
-
"loss": 0.006,
|
2088 |
-
"step": 2970
|
2089 |
-
},
|
2090 |
-
{
|
2091 |
-
"epoch": 12.680851063829786,
|
2092 |
-
"grad_norm": 0.04919581860303879,
|
2093 |
-
"learning_rate": 1.1566161187926439e-05,
|
2094 |
-
"loss": 0.006,
|
2095 |
-
"step": 2980
|
2096 |
-
},
|
2097 |
-
{
|
2098 |
-
"epoch": 12.72340425531915,
|
2099 |
-
"grad_norm": 0.04949197918176651,
|
2100 |
-
"learning_rate": 1.115361031481853e-05,
|
2101 |
-
"loss": 0.0057,
|
2102 |
-
"step": 2990
|
2103 |
-
},
|
2104 |
-
{
|
2105 |
-
"epoch": 12.76595744680851,
|
2106 |
-
"grad_norm": 0.044637832790613174,
|
2107 |
-
"learning_rate": 1.0748116414011888e-05,
|
2108 |
-
"loss": 0.0055,
|
2109 |
-
"step": 3000
|
2110 |
-
},
|
2111 |
-
{
|
2112 |
-
"epoch": 12.808510638297872,
|
2113 |
-
"grad_norm": 0.04936418682336807,
|
2114 |
-
"learning_rate": 1.0349711693446918e-05,
|
2115 |
-
"loss": 0.0059,
|
2116 |
-
"step": 3010
|
2117 |
-
},
|
2118 |
-
{
|
2119 |
-
"epoch": 12.851063829787234,
|
2120 |
-
"grad_norm": 0.0362602099776268,
|
2121 |
-
"learning_rate": 9.958427797978166e-06,
|
2122 |
-
"loss": 0.0058,
|
2123 |
-
"step": 3020
|
2124 |
-
},
|
2125 |
-
{
|
2126 |
-
"epoch": 12.893617021276595,
|
2127 |
-
"grad_norm": 0.05616355687379837,
|
2128 |
-
"learning_rate": 9.574295806860767e-06,
|
2129 |
-
"loss": 0.0059,
|
2130 |
-
"step": 3030
|
2131 |
-
},
|
2132 |
-
{
|
2133 |
-
"epoch": 12.936170212765958,
|
2134 |
-
"grad_norm": 0.048922713845968246,
|
2135 |
-
"learning_rate": 9.197346231281845e-06,
|
2136 |
-
"loss": 0.0058,
|
2137 |
-
"step": 3040
|
2138 |
-
},
|
2139 |
-
{
|
2140 |
-
"epoch": 12.97872340425532,
|
2141 |
-
"grad_norm": 0.08081869035959244,
|
2142 |
-
"learning_rate": 8.827609011937066e-06,
|
2143 |
-
"loss": 0.006,
|
2144 |
-
"step": 3050
|
2145 |
-
},
|
2146 |
-
{
|
2147 |
-
"epoch": 13.02127659574468,
|
2148 |
-
"grad_norm": 0.06286204606294632,
|
2149 |
-
"learning_rate": 8.465113516652424e-06,
|
2150 |
-
"loss": 0.0057,
|
2151 |
-
"step": 3060
|
2152 |
-
},
|
2153 |
-
{
|
2154 |
-
"epoch": 13.063829787234043,
|
2155 |
-
"grad_norm": 0.05016550421714783,
|
2156 |
-
"learning_rate": 8.109888538051724e-06,
|
2157 |
-
"loss": 0.0056,
|
2158 |
-
"step": 3070
|
2159 |
-
},
|
2160 |
-
{
|
2161 |
-
"epoch": 13.106382978723405,
|
2162 |
-
"grad_norm": 0.04805108159780502,
|
2163 |
-
"learning_rate": 7.761962291269408e-06,
|
2164 |
-
"loss": 0.0056,
|
2165 |
-
"step": 3080
|
2166 |
-
},
|
2167 |
-
{
|
2168 |
-
"epoch": 13.148936170212766,
|
2169 |
-
"grad_norm": 0.08138354867696762,
|
2170 |
-
"learning_rate": 7.4213624117096755e-06,
|
2171 |
-
"loss": 0.0057,
|
2172 |
-
"step": 3090
|
2173 |
-
},
|
2174 |
-
{
|
2175 |
-
"epoch": 13.191489361702128,
|
2176 |
-
"grad_norm": 0.05341387912631035,
|
2177 |
-
"learning_rate": 7.088115952851238e-06,
|
2178 |
-
"loss": 0.0056,
|
2179 |
-
"step": 3100
|
2180 |
-
},
|
2181 |
-
{
|
2182 |
-
"epoch": 13.23404255319149,
|
2183 |
-
"grad_norm": 0.06248362362384796,
|
2184 |
-
"learning_rate": 6.762249384098662e-06,
|
2185 |
-
"loss": 0.0059,
|
2186 |
-
"step": 3110
|
2187 |
-
},
|
2188 |
-
{
|
2189 |
-
"epoch": 13.27659574468085,
|
2190 |
-
"grad_norm": 0.058334264904260635,
|
2191 |
-
"learning_rate": 6.4437885886798224e-06,
|
2192 |
-
"loss": 0.0057,
|
2193 |
-
"step": 3120
|
2194 |
-
},
|
2195 |
-
{
|
2196 |
-
"epoch": 13.319148936170214,
|
2197 |
-
"grad_norm": 0.07496757060289383,
|
2198 |
-
"learning_rate": 6.132758861590005e-06,
|
2199 |
-
"loss": 0.0057,
|
2200 |
-
"step": 3130
|
2201 |
-
},
|
2202 |
-
{
|
2203 |
-
"epoch": 13.361702127659575,
|
2204 |
-
"grad_norm": 0.04468085989356041,
|
2205 |
-
"learning_rate": 5.829184907582896e-06,
|
2206 |
-
"loss": 0.0058,
|
2207 |
-
"step": 3140
|
2208 |
-
},
|
2209 |
-
{
|
2210 |
-
"epoch": 13.404255319148936,
|
2211 |
-
"grad_norm": 0.06133480742573738,
|
2212 |
-
"learning_rate": 5.533090839208133e-06,
|
2213 |
-
"loss": 0.0058,
|
2214 |
-
"step": 3150
|
2215 |
-
},
|
2216 |
-
{
|
2217 |
-
"epoch": 13.446808510638299,
|
2218 |
-
"grad_norm": 0.056072745472192764,
|
2219 |
-
"learning_rate": 5.244500174896205e-06,
|
2220 |
-
"loss": 0.0057,
|
2221 |
-
"step": 3160
|
2222 |
-
},
|
2223 |
-
{
|
2224 |
-
"epoch": 13.48936170212766,
|
2225 |
-
"grad_norm": 0.07432785630226135,
|
2226 |
-
"learning_rate": 4.96343583709038e-06,
|
2227 |
-
"loss": 0.0057,
|
2228 |
-
"step": 3170
|
2229 |
-
},
|
2230 |
-
{
|
2231 |
-
"epoch": 13.53191489361702,
|
2232 |
-
"grad_norm": 0.051389969885349274,
|
2233 |
-
"learning_rate": 4.68992015042592e-06,
|
2234 |
-
"loss": 0.0056,
|
2235 |
-
"step": 3180
|
2236 |
-
},
|
2237 |
-
{
|
2238 |
-
"epoch": 13.574468085106384,
|
2239 |
-
"grad_norm": 0.0631624385714531,
|
2240 |
-
"learning_rate": 4.423974839956968e-06,
|
2241 |
-
"loss": 0.0059,
|
2242 |
-
"step": 3190
|
2243 |
-
},
|
2244 |
-
{
|
2245 |
-
"epoch": 13.617021276595745,
|
2246 |
-
"grad_norm": 0.0559026375412941,
|
2247 |
-
"learning_rate": 4.165621029430855e-06,
|
2248 |
-
"loss": 0.0057,
|
2249 |
-
"step": 3200
|
2250 |
-
},
|
2251 |
-
{
|
2252 |
-
"epoch": 13.659574468085106,
|
2253 |
-
"grad_norm": 0.05802515894174576,
|
2254 |
-
"learning_rate": 3.914879239610392e-06,
|
2255 |
-
"loss": 0.0059,
|
2256 |
-
"step": 3210
|
2257 |
-
},
|
2258 |
-
{
|
2259 |
-
"epoch": 13.702127659574469,
|
2260 |
-
"grad_norm": 0.06547726690769196,
|
2261 |
-
"learning_rate": 3.671769386643742e-06,
|
2262 |
-
"loss": 0.0058,
|
2263 |
-
"step": 3220
|
2264 |
-
},
|
2265 |
-
{
|
2266 |
-
"epoch": 13.74468085106383,
|
2267 |
-
"grad_norm": 0.08117694407701492,
|
2268 |
-
"learning_rate": 3.436310780482688e-06,
|
2269 |
-
"loss": 0.0061,
|
2270 |
-
"step": 3230
|
2271 |
-
},
|
2272 |
-
{
|
2273 |
-
"epoch": 13.787234042553191,
|
2274 |
-
"grad_norm": 0.059095997363328934,
|
2275 |
-
"learning_rate": 3.2085221233487562e-06,
|
2276 |
-
"loss": 0.0058,
|
2277 |
-
"step": 3240
|
2278 |
-
},
|
2279 |
-
{
|
2280 |
-
"epoch": 13.829787234042554,
|
2281 |
-
"grad_norm": 0.05627988278865814,
|
2282 |
-
"learning_rate": 2.9884215082477408e-06,
|
2283 |
-
"loss": 0.0056,
|
2284 |
-
"step": 3250
|
2285 |
-
},
|
2286 |
-
{
|
2287 |
-
"epoch": 13.872340425531915,
|
2288 |
-
"grad_norm": 0.04729423671960831,
|
2289 |
-
"learning_rate": 2.776026417532629e-06,
|
2290 |
-
"loss": 0.0058,
|
2291 |
-
"step": 3260
|
2292 |
-
},
|
2293 |
-
{
|
2294 |
-
"epoch": 13.914893617021276,
|
2295 |
-
"grad_norm": 0.04625534638762474,
|
2296 |
-
"learning_rate": 2.5713537215149132e-06,
|
2297 |
-
"loss": 0.0056,
|
2298 |
-
"step": 3270
|
2299 |
-
},
|
2300 |
-
{
|
2301 |
-
"epoch": 13.957446808510639,
|
2302 |
-
"grad_norm": 0.06575009226799011,
|
2303 |
-
"learning_rate": 2.3744196771247173e-06,
|
2304 |
-
"loss": 0.0057,
|
2305 |
-
"step": 3280
|
2306 |
-
},
|
2307 |
-
{
|
2308 |
-
"epoch": 14.0,
|
2309 |
-
"grad_norm": 0.04723024740815163,
|
2310 |
-
"learning_rate": 2.1852399266194314e-06,
|
2311 |
-
"loss": 0.0057,
|
2312 |
-
"step": 3290
|
2313 |
-
},
|
2314 |
-
{
|
2315 |
-
"epoch": 14.042553191489361,
|
2316 |
-
"grad_norm": 0.051644448190927505,
|
2317 |
-
"learning_rate": 2.003829496341325e-06,
|
2318 |
-
"loss": 0.0057,
|
2319 |
-
"step": 3300
|
2320 |
-
},
|
2321 |
-
{
|
2322 |
-
"epoch": 14.085106382978724,
|
2323 |
-
"grad_norm": 0.05777151137590408,
|
2324 |
-
"learning_rate": 1.8302027955239387e-06,
|
2325 |
-
"loss": 0.0055,
|
2326 |
-
"step": 3310
|
2327 |
-
},
|
2328 |
-
{
|
2329 |
-
"epoch": 14.127659574468085,
|
2330 |
-
"grad_norm": 0.06086675077676773,
|
2331 |
-
"learning_rate": 1.6643736151477185e-06,
|
2332 |
-
"loss": 0.0057,
|
2333 |
-
"step": 3320
|
2334 |
-
},
|
2335 |
-
{
|
2336 |
-
"epoch": 14.170212765957446,
|
2337 |
-
"grad_norm": 0.05680638551712036,
|
2338 |
-
"learning_rate": 1.5063551268444276e-06,
|
2339 |
-
"loss": 0.0057,
|
2340 |
-
"step": 3330
|
2341 |
-
},
|
2342 |
-
{
|
2343 |
-
"epoch": 14.212765957446809,
|
2344 |
-
"grad_norm": 0.05553250014781952,
|
2345 |
-
"learning_rate": 1.3561598818511045e-06,
|
2346 |
-
"loss": 0.0058,
|
2347 |
-
"step": 3340
|
2348 |
-
},
|
2349 |
-
{
|
2350 |
-
"epoch": 14.25531914893617,
|
2351 |
-
"grad_norm": 0.05792253836989403,
|
2352 |
-
"learning_rate": 1.21379981001305e-06,
|
2353 |
-
"loss": 0.0057,
|
2354 |
-
"step": 3350
|
2355 |
-
},
|
2356 |
-
{
|
2357 |
-
"epoch": 14.297872340425531,
|
2358 |
-
"grad_norm": 0.05049928277730942,
|
2359 |
-
"learning_rate": 1.0792862188362396e-06,
|
2360 |
-
"loss": 0.0059,
|
2361 |
-
"step": 3360
|
2362 |
-
},
|
2363 |
-
{
|
2364 |
-
"epoch": 14.340425531914894,
|
2365 |
-
"grad_norm": 0.04483390226960182,
|
2366 |
-
"learning_rate": 9.526297925892435e-07,
|
2367 |
-
"loss": 0.0056,
|
2368 |
-
"step": 3370
|
2369 |
-
},
|
2370 |
-
{
|
2371 |
-
"epoch": 14.382978723404255,
|
2372 |
-
"grad_norm": 0.052869729697704315,
|
2373 |
-
"learning_rate": 8.338405914545045e-07,
|
2374 |
-
"loss": 0.0058,
|
2375 |
-
"step": 3380
|
2376 |
-
},
|
2377 |
-
{
|
2378 |
-
"epoch": 14.425531914893616,
|
2379 |
-
"grad_norm": 0.07982738316059113,
|
2380 |
-
"learning_rate": 7.229280507293657e-07,
|
2381 |
-
"loss": 0.006,
|
2382 |
-
"step": 3390
|
2383 |
-
},
|
2384 |
-
{
|
2385 |
-
"epoch": 14.46808510638298,
|
2386 |
-
"grad_norm": 0.03856668993830681,
|
2387 |
-
"learning_rate": 6.199009800765265e-07,
|
2388 |
-
"loss": 0.0057,
|
2389 |
-
"step": 3400
|
2390 |
-
},
|
2391 |
-
{
|
2392 |
-
"epoch": 14.51063829787234,
|
2393 |
-
"grad_norm": 0.07019790261983871,
|
2394 |
-
"learning_rate": 5.24767562824402e-07,
|
2395 |
-
"loss": 0.0057,
|
2396 |
-
"step": 3410
|
2397 |
-
},
|
2398 |
-
{
|
2399 |
-
"epoch": 14.553191489361701,
|
2400 |
-
"grad_norm": 0.06207670271396637,
|
2401 |
-
"learning_rate": 4.375353553170647e-07,
|
2402 |
-
"loss": 0.0055,
|
2403 |
-
"step": 3420
|
2404 |
-
},
|
2405 |
-
{
|
2406 |
-
"epoch": 14.595744680851064,
|
2407 |
-
"grad_norm": 0.07232890278100967,
|
2408 |
-
"learning_rate": 3.5821128631408075e-07,
|
2409 |
-
"loss": 0.0056,
|
2410 |
-
"step": 3430
|
2411 |
-
},
|
2412 |
-
{
|
2413 |
-
"epoch": 14.638297872340425,
|
2414 |
-
"grad_norm": 0.05705394968390465,
|
2415 |
-
"learning_rate": 2.86801656440161e-07,
|
2416 |
-
"loss": 0.0059,
|
2417 |
-
"step": 3440
|
2418 |
-
},
|
2419 |
-
{
|
2420 |
-
"epoch": 14.680851063829786,
|
2421 |
-
"grad_norm": 0.049011845141649246,
|
2422 |
-
"learning_rate": 2.2331213768468363e-07,
|
2423 |
-
"loss": 0.0057,
|
2424 |
-
"step": 3450
|
2425 |
-
},
|
2426 |
-
{
|
2427 |
-
"epoch": 14.72340425531915,
|
2428 |
-
"grad_norm": 0.0473303347826004,
|
2429 |
-
"learning_rate": 1.6774777295123223e-07,
|
2430 |
-
"loss": 0.0056,
|
2431 |
-
"step": 3460
|
2432 |
-
},
|
2433 |
-
{
|
2434 |
-
"epoch": 14.76595744680851,
|
2435 |
-
"grad_norm": 0.06919126957654953,
|
2436 |
-
"learning_rate": 1.2011297565697188e-07,
|
2437 |
-
"loss": 0.0056,
|
2438 |
-
"step": 3470
|
2439 |
-
},
|
2440 |
-
{
|
2441 |
-
"epoch": 14.808510638297872,
|
2442 |
-
"grad_norm": 0.052880384027957916,
|
2443 |
-
"learning_rate": 8.041152938216278e-08,
|
2444 |
-
"loss": 0.0054,
|
2445 |
-
"step": 3480
|
2446 |
-
},
|
2447 |
-
{
|
2448 |
-
"epoch": 14.851063829787234,
|
2449 |
-
"grad_norm": 0.08039422333240509,
|
2450 |
-
"learning_rate": 4.8646587569578514e-08,
|
2451 |
-
"loss": 0.0059,
|
2452 |
-
"step": 3490
|
2453 |
-
},
|
2454 |
-
{
|
2455 |
-
"epoch": 14.893617021276595,
|
2456 |
-
"grad_norm": 0.060477934777736664,
|
2457 |
-
"learning_rate": 2.482067327409521e-08,
|
2458 |
-
"loss": 0.0056,
|
2459 |
-
"step": 3500
|
2460 |
-
},
|
2461 |
-
{
|
2462 |
-
"epoch": 14.936170212765958,
|
2463 |
-
"grad_norm": 0.053076595067977905,
|
2464 |
-
"learning_rate": 8.93567896219638e-09,
|
2465 |
-
"loss": 0.0055,
|
2466 |
-
"step": 3510
|
2467 |
-
},
|
2468 |
-
{
|
2469 |
-
"epoch": 14.97872340425532,
|
2470 |
-
"grad_norm": 0.0657854750752449,
|
2471 |
-
"learning_rate": 9.9286636175977e-10,
|
2472 |
-
"loss": 0.0057,
|
2473 |
-
"step": 3520
|
2474 |
}
|
2475 |
],
|
2476 |
"logging_steps": 10,
|
@@ -2485,12 +350,12 @@
|
|
2485 |
"should_evaluate": false,
|
2486 |
"should_log": false,
|
2487 |
"should_save": true,
|
2488 |
-
"should_training_stop":
|
2489 |
},
|
2490 |
"attributes": {}
|
2491 |
}
|
2492 |
},
|
2493 |
-
"total_flos": 1.
|
2494 |
"train_batch_size": 1,
|
2495 |
"trial_name": null,
|
2496 |
"trial_params": null
|
|
|
1 |
{
|
2 |
"best_metric": null,
|
3 |
"best_model_checkpoint": null,
|
4 |
+
"epoch": 2.0,
|
5 |
"eval_steps": 500,
|
6 |
+
"global_step": 470,
|
7 |
"is_hyper_param_search": false,
|
8 |
"is_local_process_zero": true,
|
9 |
"is_world_process_zero": true,
|
|
|
336 |
"learning_rate": 0.0001913545457642601,
|
337 |
"loss": 0.0283,
|
338 |
"step": 470
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
339 |
}
|
340 |
],
|
341 |
"logging_steps": 10,
|
|
|
350 |
"should_evaluate": false,
|
351 |
"should_log": false,
|
352 |
"should_save": true,
|
353 |
+
"should_training_stop": false
|
354 |
},
|
355 |
"attributes": {}
|
356 |
}
|
357 |
},
|
358 |
+
"total_flos": 1.889350100631552e+16,
|
359 |
"train_batch_size": 1,
|
360 |
"trial_name": null,
|
361 |
"trial_params": null
|