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 957942768
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:f67fb93d73ef689e29fced3646888b9acac71ded7ce8bdc2e47a329b3d916111
|
| 3 |
size 957942768
|
optimizer.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 1916174411
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:74d8c8563d3fd92da4fa183b9c5a3bef0b8fabc91f3062232d31df923404a061
|
| 3 |
size 1916174411
|
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:e5b517d1b8e2b0f837c8b00170b154961d4d989feba4326ac25583df7a55c57a
|
| 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 sha256:5c4e44404b58ce3af1b46c3d4a85a59edbbc386f340c476e894715a1199e1aed
|
| 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": 2.361471312710075e-05,
|
| 429 |
"loss": 0.5706,
|
| 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": 2.361471312710075e-05,
|
| 429 |
"loss": 0.5706,
|
| 430 |
"step": 60
|
| 431 |
+
},
|
| 432 |
+
{
|
| 433 |
+
"epoch": 3.050632911392405,
|
| 434 |
+
"grad_norm": 5.2546257972717285,
|
| 435 |
+
"learning_rate": 2.2693291013417453e-05,
|
| 436 |
+
"loss": 0.6983,
|
| 437 |
+
"step": 61
|
| 438 |
+
},
|
| 439 |
+
{
|
| 440 |
+
"epoch": 3.1012658227848102,
|
| 441 |
+
"grad_norm": 5.815437316894531,
|
| 442 |
+
"learning_rate": 2.1775019586744923e-05,
|
| 443 |
+
"loss": 0.9768,
|
| 444 |
+
"step": 62
|
| 445 |
+
},
|
| 446 |
+
{
|
| 447 |
+
"epoch": 3.151898734177215,
|
| 448 |
+
"grad_norm": 5.194660186767578,
|
| 449 |
+
"learning_rate": 2.0861153095396748e-05,
|
| 450 |
+
"loss": 0.6243,
|
| 451 |
+
"step": 63
|
| 452 |
+
},
|
| 453 |
+
{
|
| 454 |
+
"epoch": 3.2025316455696204,
|
| 455 |
+
"grad_norm": 4.012391567230225,
|
| 456 |
+
"learning_rate": 1.995293977107475e-05,
|
| 457 |
+
"loss": 0.469,
|
| 458 |
+
"step": 64
|
| 459 |
+
},
|
| 460 |
+
{
|
| 461 |
+
"epoch": 3.2531645569620253,
|
| 462 |
+
"grad_norm": 5.675468444824219,
|
| 463 |
+
"learning_rate": 1.9051620123934537e-05,
|
| 464 |
+
"loss": 0.6084,
|
| 465 |
+
"step": 65
|
| 466 |
+
},
|
| 467 |
+
{
|
| 468 |
+
"epoch": 3.3037974683544302,
|
| 469 |
+
"grad_norm": 5.8908209800720215,
|
| 470 |
+
"learning_rate": 1.815842524819793e-05,
|
| 471 |
+
"loss": 0.648,
|
| 472 |
+
"step": 66
|
| 473 |
+
},
|
| 474 |
+
{
|
| 475 |
+
"epoch": 3.3544303797468356,
|
| 476 |
+
"grad_norm": 7.725429534912109,
|
| 477 |
+
"learning_rate": 1.7274575140626318e-05,
|
| 478 |
+
"loss": 0.6949,
|
| 479 |
+
"step": 67
|
| 480 |
+
},
|
| 481 |
+
{
|
| 482 |
+
"epoch": 3.4050632911392404,
|
| 483 |
+
"grad_norm": 6.168173313140869,
|
| 484 |
+
"learning_rate": 1.6401277034151798e-05,
|
| 485 |
+
"loss": 0.9213,
|
| 486 |
+
"step": 68
|
| 487 |
+
},
|
| 488 |
+
{
|
| 489 |
+
"epoch": 3.4556962025316453,
|
| 490 |
+
"grad_norm": 6.947693347930908,
|
| 491 |
+
"learning_rate": 1.5539723748942245e-05,
|
| 492 |
+
"loss": 0.7397,
|
| 493 |
+
"step": 69
|
| 494 |
+
},
|
| 495 |
+
{
|
| 496 |
+
"epoch": 3.5063291139240507,
|
| 497 |
+
"grad_norm": 5.9794206619262695,
|
| 498 |
+
"learning_rate": 1.4691092063152417e-05,
|
| 499 |
+
"loss": 0.5009,
|
| 500 |
+
"step": 70
|
| 501 |
+
},
|
| 502 |
+
{
|
| 503 |
+
"epoch": 3.5569620253164556,
|
| 504 |
+
"grad_norm": 5.66774320602417,
|
| 505 |
+
"learning_rate": 1.3856541105586545e-05,
|
| 506 |
+
"loss": 0.5204,
|
| 507 |
+
"step": 71
|
| 508 |
+
},
|
| 509 |
+
{
|
| 510 |
+
"epoch": 3.607594936708861,
|
| 511 |
+
"grad_norm": 8.234807014465332,
|
| 512 |
+
"learning_rate": 1.303721077246784e-05,
|
| 513 |
+
"loss": 0.8793,
|
| 514 |
+
"step": 72
|
| 515 |
+
},
|
| 516 |
+
{
|
| 517 |
+
"epoch": 3.6582278481012658,
|
| 518 |
+
"grad_norm": 8.785400390625,
|
| 519 |
+
"learning_rate": 1.223422017047733e-05,
|
| 520 |
+
"loss": 0.6229,
|
| 521 |
+
"step": 73
|
| 522 |
+
},
|
| 523 |
+
{
|
| 524 |
+
"epoch": 3.708860759493671,
|
| 525 |
+
"grad_norm": 6.376526832580566,
|
| 526 |
+
"learning_rate": 1.1448666088188764e-05,
|
| 527 |
+
"loss": 0.6154,
|
| 528 |
+
"step": 74
|
| 529 |
+
},
|
| 530 |
+
{
|
| 531 |
+
"epoch": 3.759493670886076,
|
| 532 |
+
"grad_norm": 7.004448413848877,
|
| 533 |
+
"learning_rate": 1.068162149798737e-05,
|
| 534 |
+
"loss": 0.7203,
|
| 535 |
+
"step": 75
|
| 536 |
+
},
|
| 537 |
+
{
|
| 538 |
+
"epoch": 3.810126582278481,
|
| 539 |
+
"grad_norm": 5.858279705047607,
|
| 540 |
+
"learning_rate": 9.934134090518593e-06,
|
| 541 |
+
"loss": 0.5153,
|
| 542 |
+
"step": 76
|
| 543 |
+
},
|
| 544 |
+
{
|
| 545 |
+
"epoch": 3.8607594936708862,
|
| 546 |
+
"grad_norm": 7.578220844268799,
|
| 547 |
+
"learning_rate": 9.207224843668732e-06,
|
| 548 |
+
"loss": 0.7153,
|
| 549 |
+
"step": 77
|
| 550 |
+
},
|
| 551 |
+
{
|
| 552 |
+
"epoch": 3.911392405063291,
|
| 553 |
+
"grad_norm": 7.869601249694824,
|
| 554 |
+
"learning_rate": 8.50188662803194e-06,
|
| 555 |
+
"loss": 0.6988,
|
| 556 |
+
"step": 78
|
| 557 |
+
},
|
| 558 |
+
{
|
| 559 |
+
"epoch": 3.962025316455696,
|
| 560 |
+
"grad_norm": 6.777385234832764,
|
| 561 |
+
"learning_rate": 7.819082850768434e-06,
|
| 562 |
+
"loss": 0.6007,
|
| 563 |
+
"step": 79
|
| 564 |
+
},
|
| 565 |
+
{
|
| 566 |
+
"epoch": 4.0,
|
| 567 |
+
"grad_norm": 6.161752223968506,
|
| 568 |
+
"learning_rate": 7.159746139706194e-06,
|
| 569 |
+
"loss": 0.4779,
|
| 570 |
+
"step": 80
|
| 571 |
+
},
|
| 572 |
+
{
|
| 573 |
+
"epoch": 4.050632911392405,
|
| 574 |
+
"grad_norm": 5.206139087677002,
|
| 575 |
+
"learning_rate": 6.524777069483526e-06,
|
| 576 |
+
"loss": 0.4173,
|
| 577 |
+
"step": 81
|
| 578 |
+
},
|
| 579 |
+
{
|
| 580 |
+
"epoch": 4.10126582278481,
|
| 581 |
+
"grad_norm": 4.832441329956055,
|
| 582 |
+
"learning_rate": 5.915042931472425e-06,
|
| 583 |
+
"loss": 0.4491,
|
| 584 |
+
"step": 82
|
| 585 |
+
},
|
| 586 |
+
{
|
| 587 |
+
"epoch": 4.151898734177215,
|
| 588 |
+
"grad_norm": 4.783233165740967,
|
| 589 |
+
"learning_rate": 5.33137654916292e-06,
|
| 590 |
+
"loss": 0.3311,
|
| 591 |
+
"step": 83
|
| 592 |
+
},
|
| 593 |
+
{
|
| 594 |
+
"epoch": 4.2025316455696204,
|
| 595 |
+
"grad_norm": 3.099482536315918,
|
| 596 |
+
"learning_rate": 4.7745751406263165e-06,
|
| 597 |
+
"loss": 0.2116,
|
| 598 |
+
"step": 84
|
| 599 |
+
},
|
| 600 |
+
{
|
| 601 |
+
"epoch": 4.253164556962025,
|
| 602 |
+
"grad_norm": 5.326932907104492,
|
| 603 |
+
"learning_rate": 4.245399229611238e-06,
|
| 604 |
+
"loss": 0.3897,
|
| 605 |
+
"step": 85
|
| 606 |
+
},
|
| 607 |
+
{
|
| 608 |
+
"epoch": 4.30379746835443,
|
| 609 |
+
"grad_norm": 4.431222915649414,
|
| 610 |
+
"learning_rate": 3.7445716067596503e-06,
|
| 611 |
+
"loss": 0.4973,
|
| 612 |
+
"step": 86
|
| 613 |
+
},
|
| 614 |
+
{
|
| 615 |
+
"epoch": 4.3544303797468356,
|
| 616 |
+
"grad_norm": 4.217422008514404,
|
| 617 |
+
"learning_rate": 3.2727763423617913e-06,
|
| 618 |
+
"loss": 0.182,
|
| 619 |
+
"step": 87
|
| 620 |
+
},
|
| 621 |
+
{
|
| 622 |
+
"epoch": 4.405063291139241,
|
| 623 |
+
"grad_norm": 5.346303462982178,
|
| 624 |
+
"learning_rate": 2.8306578519984527e-06,
|
| 625 |
+
"loss": 0.5239,
|
| 626 |
+
"step": 88
|
| 627 |
+
},
|
| 628 |
+
{
|
| 629 |
+
"epoch": 4.455696202531645,
|
| 630 |
+
"grad_norm": 8.100042343139648,
|
| 631 |
+
"learning_rate": 2.418820016346779e-06,
|
| 632 |
+
"loss": 0.2284,
|
| 633 |
+
"step": 89
|
| 634 |
+
},
|
| 635 |
+
{
|
| 636 |
+
"epoch": 4.506329113924051,
|
| 637 |
+
"grad_norm": 4.507992267608643,
|
| 638 |
+
"learning_rate": 2.0378253563519247e-06,
|
| 639 |
+
"loss": 0.3344,
|
| 640 |
+
"step": 90
|
| 641 |
+
},
|
| 642 |
+
{
|
| 643 |
+
"epoch": 4.556962025316456,
|
| 644 |
+
"grad_norm": 4.841477394104004,
|
| 645 |
+
"learning_rate": 1.6881942648911076e-06,
|
| 646 |
+
"loss": 0.2872,
|
| 647 |
+
"step": 91
|
| 648 |
+
},
|
| 649 |
+
{
|
| 650 |
+
"epoch": 4.6075949367088604,
|
| 651 |
+
"grad_norm": 4.839809417724609,
|
| 652 |
+
"learning_rate": 1.3704042959795132e-06,
|
| 653 |
+
"loss": 0.5436,
|
| 654 |
+
"step": 92
|
| 655 |
+
},
|
| 656 |
+
{
|
| 657 |
+
"epoch": 4.658227848101266,
|
| 658 |
+
"grad_norm": 3.7410666942596436,
|
| 659 |
+
"learning_rate": 1.0848895124889818e-06,
|
| 660 |
+
"loss": 0.3488,
|
| 661 |
+
"step": 93
|
| 662 |
+
},
|
| 663 |
+
{
|
| 664 |
+
"epoch": 4.708860759493671,
|
| 665 |
+
"grad_norm": 5.837460041046143,
|
| 666 |
+
"learning_rate": 8.320398932703144e-07,
|
| 667 |
+
"loss": 0.458,
|
| 668 |
+
"step": 94
|
| 669 |
+
},
|
| 670 |
+
{
|
| 671 |
+
"epoch": 4.759493670886076,
|
| 672 |
+
"grad_norm": 5.102079391479492,
|
| 673 |
+
"learning_rate": 6.122008004890851e-07,
|
| 674 |
+
"loss": 0.2965,
|
| 675 |
+
"step": 95
|
| 676 |
+
},
|
| 677 |
+
{
|
| 678 |
+
"epoch": 4.810126582278481,
|
| 679 |
+
"grad_norm": 4.543964385986328,
|
| 680 |
+
"learning_rate": 4.256725079024554e-07,
|
| 681 |
+
"loss": 0.2352,
|
| 682 |
+
"step": 96
|
| 683 |
+
},
|
| 684 |
+
{
|
| 685 |
+
"epoch": 4.860759493670886,
|
| 686 |
+
"grad_norm": 4.671619415283203,
|
| 687 |
+
"learning_rate": 2.7270979072135104e-07,
|
| 688 |
+
"loss": 0.3958,
|
| 689 |
+
"step": 97
|
| 690 |
+
},
|
| 691 |
+
{
|
| 692 |
+
"epoch": 4.911392405063291,
|
| 693 |
+
"grad_norm": 5.004724979400635,
|
| 694 |
+
"learning_rate": 1.5352157761815977e-07,
|
| 695 |
+
"loss": 0.3075,
|
| 696 |
+
"step": 98
|
| 697 |
+
},
|
| 698 |
+
{
|
| 699 |
+
"epoch": 4.962025316455696,
|
| 700 |
+
"grad_norm": 4.00545597076416,
|
| 701 |
+
"learning_rate": 6.827066535529946e-08,
|
| 702 |
+
"loss": 0.3152,
|
| 703 |
+
"step": 99
|
| 704 |
+
},
|
| 705 |
+
{
|
| 706 |
+
"epoch": 5.0,
|
| 707 |
+
"grad_norm": 5.139050483703613,
|
| 708 |
+
"learning_rate": 1.7073496424427348e-08,
|
| 709 |
+
"loss": 0.1744,
|
| 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": 3.0794717131554816e+16,
|
| 731 |
"train_batch_size": 2,
|
| 732 |
"trial_name": null,
|
| 733 |
"trial_params": null
|