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
- trainer_state.json +284 -4
adapter_model.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 359270696
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:26c9ef3202a5fecea3cf4e8797de14edaf72f05c99e7339f204d4ae79fc4059e
|
| 3 |
size 359270696
|
optimizer.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 718831691
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:d57a71f32bc02f77680270598e164ef53cf521b9842a915012f167967134ca03
|
| 3 |
size 718831691
|
rng_state.pth
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 14645
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:a3d312928d0bb60518eb9856d5ab0ae1674bcb745294bf27f615cb6d07b0463e
|
| 3 |
size 14645
|
scheduler.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 1465
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:5418e0fc9a3a6a50ea3a7b440dfb8b2fa26686b28c8f28256150a09922035962
|
| 3 |
size 1465
|
trainer_state.json
CHANGED
|
@@ -2,9 +2,9 @@
|
|
| 2 |
"best_global_step": null,
|
| 3 |
"best_metric": null,
|
| 4 |
"best_model_checkpoint": null,
|
| 5 |
-
"epoch":
|
| 6 |
"eval_steps": 500,
|
| 7 |
-
"global_step":
|
| 8 |
"is_hyper_param_search": false,
|
| 9 |
"is_local_process_zero": true,
|
| 10 |
"is_world_process_zero": true,
|
|
@@ -428,6 +428,286 @@
|
|
| 428 |
"learning_rate": 0.00023614713127100752,
|
| 429 |
"loss": 0.5813,
|
| 430 |
"step": 60
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 431 |
}
|
| 432 |
],
|
| 433 |
"logging_steps": 1,
|
|
@@ -442,12 +722,12 @@
|
|
| 442 |
"should_evaluate": false,
|
| 443 |
"should_log": false,
|
| 444 |
"should_save": true,
|
| 445 |
-
"should_training_stop":
|
| 446 |
},
|
| 447 |
"attributes": {}
|
| 448 |
}
|
| 449 |
},
|
| 450 |
-
"total_flos":
|
| 451 |
"train_batch_size": 2,
|
| 452 |
"trial_name": null,
|
| 453 |
"trial_params": null
|
|
|
|
| 2 |
"best_global_step": null,
|
| 3 |
"best_metric": null,
|
| 4 |
"best_model_checkpoint": null,
|
| 5 |
+
"epoch": 5.0,
|
| 6 |
"eval_steps": 500,
|
| 7 |
+
"global_step": 100,
|
| 8 |
"is_hyper_param_search": false,
|
| 9 |
"is_local_process_zero": true,
|
| 10 |
"is_world_process_zero": true,
|
|
|
|
| 428 |
"learning_rate": 0.00023614713127100752,
|
| 429 |
"loss": 0.5813,
|
| 430 |
"step": 60
|
| 431 |
+
},
|
| 432 |
+
{
|
| 433 |
+
"epoch": 3.050632911392405,
|
| 434 |
+
"grad_norm": 2.65234112739563,
|
| 435 |
+
"learning_rate": 0.00022693291013417452,
|
| 436 |
+
"loss": 0.7218,
|
| 437 |
+
"step": 61
|
| 438 |
+
},
|
| 439 |
+
{
|
| 440 |
+
"epoch": 3.1012658227848102,
|
| 441 |
+
"grad_norm": 1.9882073402404785,
|
| 442 |
+
"learning_rate": 0.00021775019586744925,
|
| 443 |
+
"loss": 0.5474,
|
| 444 |
+
"step": 62
|
| 445 |
+
},
|
| 446 |
+
{
|
| 447 |
+
"epoch": 3.151898734177215,
|
| 448 |
+
"grad_norm": 2.5399203300476074,
|
| 449 |
+
"learning_rate": 0.0002086115309539675,
|
| 450 |
+
"loss": 0.7906,
|
| 451 |
+
"step": 63
|
| 452 |
+
},
|
| 453 |
+
{
|
| 454 |
+
"epoch": 3.2025316455696204,
|
| 455 |
+
"grad_norm": 1.6655718088150024,
|
| 456 |
+
"learning_rate": 0.0001995293977107475,
|
| 457 |
+
"loss": 0.3726,
|
| 458 |
+
"step": 64
|
| 459 |
+
},
|
| 460 |
+
{
|
| 461 |
+
"epoch": 3.2531645569620253,
|
| 462 |
+
"grad_norm": 2.330378770828247,
|
| 463 |
+
"learning_rate": 0.00019051620123934537,
|
| 464 |
+
"loss": 0.6493,
|
| 465 |
+
"step": 65
|
| 466 |
+
},
|
| 467 |
+
{
|
| 468 |
+
"epoch": 3.3037974683544302,
|
| 469 |
+
"grad_norm": 2.360882043838501,
|
| 470 |
+
"learning_rate": 0.0001815842524819793,
|
| 471 |
+
"loss": 0.6553,
|
| 472 |
+
"step": 66
|
| 473 |
+
},
|
| 474 |
+
{
|
| 475 |
+
"epoch": 3.3544303797468356,
|
| 476 |
+
"grad_norm": 2.726463556289673,
|
| 477 |
+
"learning_rate": 0.00017274575140626317,
|
| 478 |
+
"loss": 0.7101,
|
| 479 |
+
"step": 67
|
| 480 |
+
},
|
| 481 |
+
{
|
| 482 |
+
"epoch": 3.4050632911392404,
|
| 483 |
+
"grad_norm": 2.313028573989868,
|
| 484 |
+
"learning_rate": 0.00016401277034151795,
|
| 485 |
+
"loss": 0.773,
|
| 486 |
+
"step": 68
|
| 487 |
+
},
|
| 488 |
+
{
|
| 489 |
+
"epoch": 3.4556962025316453,
|
| 490 |
+
"grad_norm": 2.5557079315185547,
|
| 491 |
+
"learning_rate": 0.00015539723748942243,
|
| 492 |
+
"loss": 0.7098,
|
| 493 |
+
"step": 69
|
| 494 |
+
},
|
| 495 |
+
{
|
| 496 |
+
"epoch": 3.5063291139240507,
|
| 497 |
+
"grad_norm": 2.11527681350708,
|
| 498 |
+
"learning_rate": 0.00014691092063152418,
|
| 499 |
+
"loss": 0.5224,
|
| 500 |
+
"step": 70
|
| 501 |
+
},
|
| 502 |
+
{
|
| 503 |
+
"epoch": 3.5569620253164556,
|
| 504 |
+
"grad_norm": 2.06489634513855,
|
| 505 |
+
"learning_rate": 0.00013856541105586545,
|
| 506 |
+
"loss": 0.565,
|
| 507 |
+
"step": 71
|
| 508 |
+
},
|
| 509 |
+
{
|
| 510 |
+
"epoch": 3.607594936708861,
|
| 511 |
+
"grad_norm": 2.3769450187683105,
|
| 512 |
+
"learning_rate": 0.0001303721077246784,
|
| 513 |
+
"loss": 0.5167,
|
| 514 |
+
"step": 72
|
| 515 |
+
},
|
| 516 |
+
{
|
| 517 |
+
"epoch": 3.6582278481012658,
|
| 518 |
+
"grad_norm": 3.1446056365966797,
|
| 519 |
+
"learning_rate": 0.0001223422017047733,
|
| 520 |
+
"loss": 0.738,
|
| 521 |
+
"step": 73
|
| 522 |
+
},
|
| 523 |
+
{
|
| 524 |
+
"epoch": 3.708860759493671,
|
| 525 |
+
"grad_norm": 2.9032044410705566,
|
| 526 |
+
"learning_rate": 0.00011448666088188764,
|
| 527 |
+
"loss": 0.84,
|
| 528 |
+
"step": 74
|
| 529 |
+
},
|
| 530 |
+
{
|
| 531 |
+
"epoch": 3.759493670886076,
|
| 532 |
+
"grad_norm": 2.6407244205474854,
|
| 533 |
+
"learning_rate": 0.00010681621497987371,
|
| 534 |
+
"loss": 0.55,
|
| 535 |
+
"step": 75
|
| 536 |
+
},
|
| 537 |
+
{
|
| 538 |
+
"epoch": 3.810126582278481,
|
| 539 |
+
"grad_norm": 2.3889331817626953,
|
| 540 |
+
"learning_rate": 9.934134090518593e-05,
|
| 541 |
+
"loss": 0.6036,
|
| 542 |
+
"step": 76
|
| 543 |
+
},
|
| 544 |
+
{
|
| 545 |
+
"epoch": 3.8607594936708862,
|
| 546 |
+
"grad_norm": 2.591595411300659,
|
| 547 |
+
"learning_rate": 9.207224843668733e-05,
|
| 548 |
+
"loss": 0.4689,
|
| 549 |
+
"step": 77
|
| 550 |
+
},
|
| 551 |
+
{
|
| 552 |
+
"epoch": 3.911392405063291,
|
| 553 |
+
"grad_norm": 2.4760868549346924,
|
| 554 |
+
"learning_rate": 8.50188662803194e-05,
|
| 555 |
+
"loss": 0.4915,
|
| 556 |
+
"step": 78
|
| 557 |
+
},
|
| 558 |
+
{
|
| 559 |
+
"epoch": 3.962025316455696,
|
| 560 |
+
"grad_norm": 2.930954694747925,
|
| 561 |
+
"learning_rate": 7.819082850768433e-05,
|
| 562 |
+
"loss": 0.8307,
|
| 563 |
+
"step": 79
|
| 564 |
+
},
|
| 565 |
+
{
|
| 566 |
+
"epoch": 4.0,
|
| 567 |
+
"grad_norm": 2.1440188884735107,
|
| 568 |
+
"learning_rate": 7.159746139706194e-05,
|
| 569 |
+
"loss": 0.4358,
|
| 570 |
+
"step": 80
|
| 571 |
+
},
|
| 572 |
+
{
|
| 573 |
+
"epoch": 4.050632911392405,
|
| 574 |
+
"grad_norm": 1.4254531860351562,
|
| 575 |
+
"learning_rate": 6.524777069483526e-05,
|
| 576 |
+
"loss": 0.2303,
|
| 577 |
+
"step": 81
|
| 578 |
+
},
|
| 579 |
+
{
|
| 580 |
+
"epoch": 4.10126582278481,
|
| 581 |
+
"grad_norm": 2.0634138584136963,
|
| 582 |
+
"learning_rate": 5.9150429314724254e-05,
|
| 583 |
+
"loss": 0.5029,
|
| 584 |
+
"step": 82
|
| 585 |
+
},
|
| 586 |
+
{
|
| 587 |
+
"epoch": 4.151898734177215,
|
| 588 |
+
"grad_norm": 1.7933154106140137,
|
| 589 |
+
"learning_rate": 5.3313765491629194e-05,
|
| 590 |
+
"loss": 0.3326,
|
| 591 |
+
"step": 83
|
| 592 |
+
},
|
| 593 |
+
{
|
| 594 |
+
"epoch": 4.2025316455696204,
|
| 595 |
+
"grad_norm": 1.1411280632019043,
|
| 596 |
+
"learning_rate": 4.7745751406263163e-05,
|
| 597 |
+
"loss": 0.1651,
|
| 598 |
+
"step": 84
|
| 599 |
+
},
|
| 600 |
+
{
|
| 601 |
+
"epoch": 4.253164556962025,
|
| 602 |
+
"grad_norm": 1.6106775999069214,
|
| 603 |
+
"learning_rate": 4.245399229611238e-05,
|
| 604 |
+
"loss": 0.3276,
|
| 605 |
+
"step": 85
|
| 606 |
+
},
|
| 607 |
+
{
|
| 608 |
+
"epoch": 4.30379746835443,
|
| 609 |
+
"grad_norm": 1.6101868152618408,
|
| 610 |
+
"learning_rate": 3.7445716067596506e-05,
|
| 611 |
+
"loss": 0.3073,
|
| 612 |
+
"step": 86
|
| 613 |
+
},
|
| 614 |
+
{
|
| 615 |
+
"epoch": 4.3544303797468356,
|
| 616 |
+
"grad_norm": 1.1931700706481934,
|
| 617 |
+
"learning_rate": 3.2727763423617915e-05,
|
| 618 |
+
"loss": 0.1378,
|
| 619 |
+
"step": 87
|
| 620 |
+
},
|
| 621 |
+
{
|
| 622 |
+
"epoch": 4.405063291139241,
|
| 623 |
+
"grad_norm": 1.7510262727737427,
|
| 624 |
+
"learning_rate": 2.8306578519984528e-05,
|
| 625 |
+
"loss": 0.2117,
|
| 626 |
+
"step": 88
|
| 627 |
+
},
|
| 628 |
+
{
|
| 629 |
+
"epoch": 4.455696202531645,
|
| 630 |
+
"grad_norm": 1.7942403554916382,
|
| 631 |
+
"learning_rate": 2.4188200163467787e-05,
|
| 632 |
+
"loss": 0.1961,
|
| 633 |
+
"step": 89
|
| 634 |
+
},
|
| 635 |
+
{
|
| 636 |
+
"epoch": 4.506329113924051,
|
| 637 |
+
"grad_norm": 1.7418352365493774,
|
| 638 |
+
"learning_rate": 2.0378253563519245e-05,
|
| 639 |
+
"loss": 0.2915,
|
| 640 |
+
"step": 90
|
| 641 |
+
},
|
| 642 |
+
{
|
| 643 |
+
"epoch": 4.556962025316456,
|
| 644 |
+
"grad_norm": 1.7192072868347168,
|
| 645 |
+
"learning_rate": 1.6881942648911074e-05,
|
| 646 |
+
"loss": 0.1873,
|
| 647 |
+
"step": 91
|
| 648 |
+
},
|
| 649 |
+
{
|
| 650 |
+
"epoch": 4.6075949367088604,
|
| 651 |
+
"grad_norm": 1.5197887420654297,
|
| 652 |
+
"learning_rate": 1.3704042959795133e-05,
|
| 653 |
+
"loss": 0.3229,
|
| 654 |
+
"step": 92
|
| 655 |
+
},
|
| 656 |
+
{
|
| 657 |
+
"epoch": 4.658227848101266,
|
| 658 |
+
"grad_norm": 1.1745399236679077,
|
| 659 |
+
"learning_rate": 1.0848895124889818e-05,
|
| 660 |
+
"loss": 0.1945,
|
| 661 |
+
"step": 93
|
| 662 |
+
},
|
| 663 |
+
{
|
| 664 |
+
"epoch": 4.708860759493671,
|
| 665 |
+
"grad_norm": 1.7969136238098145,
|
| 666 |
+
"learning_rate": 8.320398932703144e-06,
|
| 667 |
+
"loss": 0.2555,
|
| 668 |
+
"step": 94
|
| 669 |
+
},
|
| 670 |
+
{
|
| 671 |
+
"epoch": 4.759493670886076,
|
| 672 |
+
"grad_norm": 1.9828176498413086,
|
| 673 |
+
"learning_rate": 6.12200800489085e-06,
|
| 674 |
+
"loss": 0.2931,
|
| 675 |
+
"step": 95
|
| 676 |
+
},
|
| 677 |
+
{
|
| 678 |
+
"epoch": 4.810126582278481,
|
| 679 |
+
"grad_norm": 1.6430472135543823,
|
| 680 |
+
"learning_rate": 4.256725079024554e-06,
|
| 681 |
+
"loss": 0.1705,
|
| 682 |
+
"step": 96
|
| 683 |
+
},
|
| 684 |
+
{
|
| 685 |
+
"epoch": 4.860759493670886,
|
| 686 |
+
"grad_norm": 1.5796780586242676,
|
| 687 |
+
"learning_rate": 2.7270979072135106e-06,
|
| 688 |
+
"loss": 0.2818,
|
| 689 |
+
"step": 97
|
| 690 |
+
},
|
| 691 |
+
{
|
| 692 |
+
"epoch": 4.911392405063291,
|
| 693 |
+
"grad_norm": 1.7523950338363647,
|
| 694 |
+
"learning_rate": 1.5352157761815977e-06,
|
| 695 |
+
"loss": 0.2395,
|
| 696 |
+
"step": 98
|
| 697 |
+
},
|
| 698 |
+
{
|
| 699 |
+
"epoch": 4.962025316455696,
|
| 700 |
+
"grad_norm": 1.9358466863632202,
|
| 701 |
+
"learning_rate": 6.827066535529947e-07,
|
| 702 |
+
"loss": 0.3112,
|
| 703 |
+
"step": 99
|
| 704 |
+
},
|
| 705 |
+
{
|
| 706 |
+
"epoch": 5.0,
|
| 707 |
+
"grad_norm": 2.1313531398773193,
|
| 708 |
+
"learning_rate": 1.7073496424427348e-07,
|
| 709 |
+
"loss": 0.102,
|
| 710 |
+
"step": 100
|
| 711 |
}
|
| 712 |
],
|
| 713 |
"logging_steps": 1,
|
|
|
|
| 722 |
"should_evaluate": false,
|
| 723 |
"should_log": false,
|
| 724 |
"should_save": true,
|
| 725 |
+
"should_training_stop": true
|
| 726 |
},
|
| 727 |
"attributes": {}
|
| 728 |
}
|
| 729 |
},
|
| 730 |
+
"total_flos": 2.9265650012307456e+16,
|
| 731 |
"train_batch_size": 2,
|
| 732 |
"trial_name": null,
|
| 733 |
"trial_params": null
|