Training in progress, epoch 1, checkpoint
Browse files- last-checkpoint/model-00001-of-00004.safetensors +1 -1
- last-checkpoint/model-00002-of-00004.safetensors +1 -1
- last-checkpoint/model-00003-of-00004.safetensors +1 -1
- last-checkpoint/model-00004-of-00004.safetensors +1 -1
- last-checkpoint/optimizer.pt +1 -1
- last-checkpoint/rng_state_0.pth +1 -1
- last-checkpoint/rng_state_1.pth +1 -1
- last-checkpoint/rng_state_2.pth +1 -1
- last-checkpoint/rng_state_3.pth +1 -1
- last-checkpoint/rng_state_4.pth +1 -1
- last-checkpoint/rng_state_5.pth +1 -1
- last-checkpoint/rng_state_6.pth +1 -1
- last-checkpoint/rng_state_7.pth +1 -1
- last-checkpoint/scheduler.pt +1 -1
- last-checkpoint/trainer_state.json +8 -1046
- last-checkpoint/training_args.bin +1 -1
last-checkpoint/model-00001-of-00004.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 4903351912
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:74e57f92c1ec268c49ead693c84d11168cf6c84e28d329fed6b2cb7a5048df2a
|
| 3 |
size 4903351912
|
last-checkpoint/model-00002-of-00004.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 4947570872
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:c3a38d3bf8a2f9e0982939f4c5439179b4fdc7144dd0403a272e15ec4a84db7f
|
| 3 |
size 4947570872
|
last-checkpoint/model-00003-of-00004.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 4962221464
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:c62cfb4125f8f62e4faaff99381d17665887359a89ab6dd9a6d039b621e57efa
|
| 3 |
size 4962221464
|
last-checkpoint/model-00004-of-00004.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 3670322200
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:538fb12da9c3be08625c4eb663aa308960933dbedb34c04f8a002b4731a461dd
|
| 3 |
size 3670322200
|
last-checkpoint/optimizer.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 36967230034
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:9673ca6d5acaf9929eb1ce5f92ad68f06f1201c34e9c3438a6363aac70b67a0c
|
| 3 |
size 36967230034
|
last-checkpoint/rng_state_0.pth
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 15984
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:78d3f197f6c6558fa8056324f1563ab9e957255f5a1a959362aa4eed7a9545db
|
| 3 |
size 15984
|
last-checkpoint/rng_state_1.pth
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 15984
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:1c1a9c65c2869356282cad6b4a0f7dff7f4dd68ab3d9d216c72b7d6cb524f860
|
| 3 |
size 15984
|
last-checkpoint/rng_state_2.pth
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 15984
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:896febe768e17bae5022a95960c041f6425783774ec8859d99d3b149063b1bf9
|
| 3 |
size 15984
|
last-checkpoint/rng_state_3.pth
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 15984
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:eac482d57e966585467c8ef44dae2869bf7e5d92886f69c11ed7bccc34c07efe
|
| 3 |
size 15984
|
last-checkpoint/rng_state_4.pth
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 15984
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:e1f27d227a20dc320ac283e0938fb2f6e5b475829a583f8c44d1a16a8c828307
|
| 3 |
size 15984
|
last-checkpoint/rng_state_5.pth
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 15984
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:d05a7106aaeaec4b81704e3f4a998b5123cf9342a6733bd9fd2d578e99108c3b
|
| 3 |
size 15984
|
last-checkpoint/rng_state_6.pth
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 15984
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:b94120d8d88502ec8d8b623ec7550315caca003b44fcffbb5767ab0de91baefe
|
| 3 |
size 15984
|
last-checkpoint/rng_state_7.pth
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 15984
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:332e4d901be380f740b5d8578f7b80ef1865c7fba83bc288c8a35852205cc668
|
| 3 |
size 15984
|
last-checkpoint/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 sha256:46406e88400d513ea1c0eb4f27e1146b6ce63027c7c32b44eef432a0d86d55f0
|
| 3 |
size 1064
|
last-checkpoint/trainer_state.json
CHANGED
|
@@ -1,9 +1,9 @@
|
|
| 1 |
{
|
| 2 |
"best_metric": null,
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
-
"epoch":
|
| 5 |
"eval_steps": 300,
|
| 6 |
-
"global_step":
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
@@ -437,10 +437,10 @@
|
|
| 437 |
},
|
| 438 |
{
|
| 439 |
"epoch": 0.819672131147541,
|
| 440 |
-
"eval_loss": 1.
|
| 441 |
-
"eval_runtime": 19.
|
| 442 |
-
"eval_samples_per_second":
|
| 443 |
-
"eval_steps_per_second": 1.
|
| 444 |
"step": 300
|
| 445 |
},
|
| 446 |
{
|
|
@@ -533,1044 +533,6 @@
|
|
| 533 |
"learning_rate": 1.5057810159948714e-06,
|
| 534 |
"loss": 1.5148,
|
| 535 |
"step": 365
|
| 536 |
-
},
|
| 537 |
-
{
|
| 538 |
-
"epoch": 1.010928961748634,
|
| 539 |
-
"grad_norm": 1.6328125,
|
| 540 |
-
"learning_rate": 1.493365751989454e-06,
|
| 541 |
-
"loss": 1.5175,
|
| 542 |
-
"step": 370
|
| 543 |
-
},
|
| 544 |
-
{
|
| 545 |
-
"epoch": 1.0245901639344261,
|
| 546 |
-
"grad_norm": 1.609375,
|
| 547 |
-
"learning_rate": 1.4808491481449143e-06,
|
| 548 |
-
"loss": 1.5084,
|
| 549 |
-
"step": 375
|
| 550 |
-
},
|
| 551 |
-
{
|
| 552 |
-
"epoch": 1.0382513661202186,
|
| 553 |
-
"grad_norm": 1.625,
|
| 554 |
-
"learning_rate": 1.4682337754354533e-06,
|
| 555 |
-
"loss": 1.5141,
|
| 556 |
-
"step": 380
|
| 557 |
-
},
|
| 558 |
-
{
|
| 559 |
-
"epoch": 1.0519125683060109,
|
| 560 |
-
"grad_norm": 1.6796875,
|
| 561 |
-
"learning_rate": 1.4555222251228995e-06,
|
| 562 |
-
"loss": 1.52,
|
| 563 |
-
"step": 385
|
| 564 |
-
},
|
| 565 |
-
{
|
| 566 |
-
"epoch": 1.0655737704918034,
|
| 567 |
-
"grad_norm": 1.671875,
|
| 568 |
-
"learning_rate": 1.442717108224452e-06,
|
| 569 |
-
"loss": 1.5167,
|
| 570 |
-
"step": 390
|
| 571 |
-
},
|
| 572 |
-
{
|
| 573 |
-
"epoch": 1.0792349726775956,
|
| 574 |
-
"grad_norm": 1.6640625,
|
| 575 |
-
"learning_rate": 1.4298210549763628e-06,
|
| 576 |
-
"loss": 1.5087,
|
| 577 |
-
"step": 395
|
| 578 |
-
},
|
| 579 |
-
{
|
| 580 |
-
"epoch": 1.092896174863388,
|
| 581 |
-
"grad_norm": 1.6875,
|
| 582 |
-
"learning_rate": 1.4168367142936734e-06,
|
| 583 |
-
"loss": 1.5049,
|
| 584 |
-
"step": 400
|
| 585 |
-
},
|
| 586 |
-
{
|
| 587 |
-
"epoch": 1.1065573770491803,
|
| 588 |
-
"grad_norm": 1.6796875,
|
| 589 |
-
"learning_rate": 1.4037667532261142e-06,
|
| 590 |
-
"loss": 1.5102,
|
| 591 |
-
"step": 405
|
| 592 |
-
},
|
| 593 |
-
{
|
| 594 |
-
"epoch": 1.1202185792349726,
|
| 595 |
-
"grad_norm": 1.671875,
|
| 596 |
-
"learning_rate": 1.3906138564102792e-06,
|
| 597 |
-
"loss": 1.5156,
|
| 598 |
-
"step": 410
|
| 599 |
-
},
|
| 600 |
-
{
|
| 601 |
-
"epoch": 1.133879781420765,
|
| 602 |
-
"grad_norm": 1.609375,
|
| 603 |
-
"learning_rate": 1.3773807255181874e-06,
|
| 604 |
-
"loss": 1.5162,
|
| 605 |
-
"step": 415
|
| 606 |
-
},
|
| 607 |
-
{
|
| 608 |
-
"epoch": 1.1475409836065573,
|
| 609 |
-
"grad_norm": 1.6171875,
|
| 610 |
-
"learning_rate": 1.3640700787023462e-06,
|
| 611 |
-
"loss": 1.5099,
|
| 612 |
-
"step": 420
|
| 613 |
-
},
|
| 614 |
-
{
|
| 615 |
-
"epoch": 1.1612021857923498,
|
| 616 |
-
"grad_norm": 1.6640625,
|
| 617 |
-
"learning_rate": 1.3506846500374283e-06,
|
| 618 |
-
"loss": 1.5159,
|
| 619 |
-
"step": 425
|
| 620 |
-
},
|
| 621 |
-
{
|
| 622 |
-
"epoch": 1.174863387978142,
|
| 623 |
-
"grad_norm": 1.671875,
|
| 624 |
-
"learning_rate": 1.337227188958679e-06,
|
| 625 |
-
"loss": 1.5145,
|
| 626 |
-
"step": 430
|
| 627 |
-
},
|
| 628 |
-
{
|
| 629 |
-
"epoch": 1.1885245901639343,
|
| 630 |
-
"grad_norm": 3.078125,
|
| 631 |
-
"learning_rate": 1.3237004596971687e-06,
|
| 632 |
-
"loss": 1.5066,
|
| 633 |
-
"step": 435
|
| 634 |
-
},
|
| 635 |
-
{
|
| 636 |
-
"epoch": 1.2021857923497268,
|
| 637 |
-
"grad_norm": 1.6484375,
|
| 638 |
-
"learning_rate": 1.3101072407120055e-06,
|
| 639 |
-
"loss": 1.5144,
|
| 640 |
-
"step": 440
|
| 641 |
-
},
|
| 642 |
-
{
|
| 643 |
-
"epoch": 1.215846994535519,
|
| 644 |
-
"grad_norm": 1.6953125,
|
| 645 |
-
"learning_rate": 1.2964503241196256e-06,
|
| 646 |
-
"loss": 1.5124,
|
| 647 |
-
"step": 445
|
| 648 |
-
},
|
| 649 |
-
{
|
| 650 |
-
"epoch": 1.2295081967213115,
|
| 651 |
-
"grad_norm": 1.65625,
|
| 652 |
-
"learning_rate": 1.2827325151202782e-06,
|
| 653 |
-
"loss": 1.5052,
|
| 654 |
-
"step": 450
|
| 655 |
-
},
|
| 656 |
-
{
|
| 657 |
-
"epoch": 1.2431693989071038,
|
| 658 |
-
"grad_norm": 1.7734375,
|
| 659 |
-
"learning_rate": 1.2689566314218228e-06,
|
| 660 |
-
"loss": 1.5103,
|
| 661 |
-
"step": 455
|
| 662 |
-
},
|
| 663 |
-
{
|
| 664 |
-
"epoch": 1.2568306010928962,
|
| 665 |
-
"grad_norm": 1.6953125,
|
| 666 |
-
"learning_rate": 1.255125502660958e-06,
|
| 667 |
-
"loss": 1.5078,
|
| 668 |
-
"step": 460
|
| 669 |
-
},
|
| 670 |
-
{
|
| 671 |
-
"epoch": 1.2704918032786885,
|
| 672 |
-
"grad_norm": 1.625,
|
| 673 |
-
"learning_rate": 1.2412419698220001e-06,
|
| 674 |
-
"loss": 1.5134,
|
| 675 |
-
"step": 465
|
| 676 |
-
},
|
| 677 |
-
{
|
| 678 |
-
"epoch": 1.2841530054644807,
|
| 679 |
-
"grad_norm": 1.640625,
|
| 680 |
-
"learning_rate": 1.2273088846533302e-06,
|
| 681 |
-
"loss": 1.5064,
|
| 682 |
-
"step": 470
|
| 683 |
-
},
|
| 684 |
-
{
|
| 685 |
-
"epoch": 1.2978142076502732,
|
| 686 |
-
"grad_norm": 1.6640625,
|
| 687 |
-
"learning_rate": 1.2133291090816295e-06,
|
| 688 |
-
"loss": 1.5124,
|
| 689 |
-
"step": 475
|
| 690 |
-
},
|
| 691 |
-
{
|
| 692 |
-
"epoch": 1.3114754098360657,
|
| 693 |
-
"grad_norm": 1.609375,
|
| 694 |
-
"learning_rate": 1.1993055146240272e-06,
|
| 695 |
-
"loss": 1.5144,
|
| 696 |
-
"step": 480
|
| 697 |
-
},
|
| 698 |
-
{
|
| 699 |
-
"epoch": 1.325136612021858,
|
| 700 |
-
"grad_norm": 1.65625,
|
| 701 |
-
"learning_rate": 1.185240981798273e-06,
|
| 702 |
-
"loss": 1.5221,
|
| 703 |
-
"step": 485
|
| 704 |
-
},
|
| 705 |
-
{
|
| 706 |
-
"epoch": 1.3387978142076502,
|
| 707 |
-
"grad_norm": 1.671875,
|
| 708 |
-
"learning_rate": 1.171138399531068e-06,
|
| 709 |
-
"loss": 1.5079,
|
| 710 |
-
"step": 490
|
| 711 |
-
},
|
| 712 |
-
{
|
| 713 |
-
"epoch": 1.3524590163934427,
|
| 714 |
-
"grad_norm": 1.6640625,
|
| 715 |
-
"learning_rate": 1.1570006645646614e-06,
|
| 716 |
-
"loss": 1.5145,
|
| 717 |
-
"step": 495
|
| 718 |
-
},
|
| 719 |
-
{
|
| 720 |
-
"epoch": 1.366120218579235,
|
| 721 |
-
"grad_norm": 1.640625,
|
| 722 |
-
"learning_rate": 1.1428306808618454e-06,
|
| 723 |
-
"loss": 1.5043,
|
| 724 |
-
"step": 500
|
| 725 |
-
},
|
| 726 |
-
{
|
| 727 |
-
"epoch": 1.3797814207650272,
|
| 728 |
-
"grad_norm": 1.671875,
|
| 729 |
-
"learning_rate": 1.1286313590094686e-06,
|
| 730 |
-
"loss": 1.5171,
|
| 731 |
-
"step": 505
|
| 732 |
-
},
|
| 733 |
-
{
|
| 734 |
-
"epoch": 1.3934426229508197,
|
| 735 |
-
"grad_norm": 1.640625,
|
| 736 |
-
"learning_rate": 1.1144056156205831e-06,
|
| 737 |
-
"loss": 1.5114,
|
| 738 |
-
"step": 510
|
| 739 |
-
},
|
| 740 |
-
{
|
| 741 |
-
"epoch": 1.4071038251366121,
|
| 742 |
-
"grad_norm": 1.6484375,
|
| 743 |
-
"learning_rate": 1.100156372735361e-06,
|
| 744 |
-
"loss": 1.5032,
|
| 745 |
-
"step": 515
|
| 746 |
-
},
|
| 747 |
-
{
|
| 748 |
-
"epoch": 1.4207650273224044,
|
| 749 |
-
"grad_norm": 1.6328125,
|
| 750 |
-
"learning_rate": 1.0858865572208891e-06,
|
| 751 |
-
"loss": 1.5151,
|
| 752 |
-
"step": 520
|
| 753 |
-
},
|
| 754 |
-
{
|
| 755 |
-
"epoch": 1.4344262295081966,
|
| 756 |
-
"grad_norm": 1.625,
|
| 757 |
-
"learning_rate": 1.071599100169978e-06,
|
| 758 |
-
"loss": 1.5015,
|
| 759 |
-
"step": 525
|
| 760 |
-
},
|
| 761 |
-
{
|
| 762 |
-
"epoch": 1.4480874316939891,
|
| 763 |
-
"grad_norm": 1.65625,
|
| 764 |
-
"learning_rate": 1.0572969362990997e-06,
|
| 765 |
-
"loss": 1.5075,
|
| 766 |
-
"step": 530
|
| 767 |
-
},
|
| 768 |
-
{
|
| 769 |
-
"epoch": 1.4617486338797814,
|
| 770 |
-
"grad_norm": 2.65625,
|
| 771 |
-
"learning_rate": 1.042983003345582e-06,
|
| 772 |
-
"loss": 1.5117,
|
| 773 |
-
"step": 535
|
| 774 |
-
},
|
| 775 |
-
{
|
| 776 |
-
"epoch": 1.4754098360655736,
|
| 777 |
-
"grad_norm": 1.6328125,
|
| 778 |
-
"learning_rate": 1.0286602414641815e-06,
|
| 779 |
-
"loss": 1.5036,
|
| 780 |
-
"step": 540
|
| 781 |
-
},
|
| 782 |
-
{
|
| 783 |
-
"epoch": 1.489071038251366,
|
| 784 |
-
"grad_norm": 1.703125,
|
| 785 |
-
"learning_rate": 1.0143315926231624e-06,
|
| 786 |
-
"loss": 1.527,
|
| 787 |
-
"step": 545
|
| 788 |
-
},
|
| 789 |
-
{
|
| 790 |
-
"epoch": 1.5027322404371586,
|
| 791 |
-
"grad_norm": 1.6484375,
|
| 792 |
-
"learning_rate": 1e-06,
|
| 793 |
-
"loss": 1.5065,
|
| 794 |
-
"step": 550
|
| 795 |
-
},
|
| 796 |
-
{
|
| 797 |
-
"epoch": 1.5163934426229508,
|
| 798 |
-
"grad_norm": 1.65625,
|
| 799 |
-
"learning_rate": 9.856684073768378e-07,
|
| 800 |
-
"loss": 1.4989,
|
| 801 |
-
"step": 555
|
| 802 |
-
},
|
| 803 |
-
{
|
| 804 |
-
"epoch": 1.530054644808743,
|
| 805 |
-
"grad_norm": 1.6640625,
|
| 806 |
-
"learning_rate": 9.713397585358188e-07,
|
| 807 |
-
"loss": 1.5066,
|
| 808 |
-
"step": 560
|
| 809 |
-
},
|
| 810 |
-
{
|
| 811 |
-
"epoch": 1.5437158469945356,
|
| 812 |
-
"grad_norm": 1.6484375,
|
| 813 |
-
"learning_rate": 9.57016996654418e-07,
|
| 814 |
-
"loss": 1.5024,
|
| 815 |
-
"step": 565
|
| 816 |
-
},
|
| 817 |
-
{
|
| 818 |
-
"epoch": 1.5573770491803278,
|
| 819 |
-
"grad_norm": 1.6484375,
|
| 820 |
-
"learning_rate": 9.427030637009002e-07,
|
| 821 |
-
"loss": 1.5095,
|
| 822 |
-
"step": 570
|
| 823 |
-
},
|
| 824 |
-
{
|
| 825 |
-
"epoch": 1.57103825136612,
|
| 826 |
-
"grad_norm": 1.671875,
|
| 827 |
-
"learning_rate": 9.28400899830022e-07,
|
| 828 |
-
"loss": 1.5144,
|
| 829 |
-
"step": 575
|
| 830 |
-
},
|
| 831 |
-
{
|
| 832 |
-
"epoch": 1.5846994535519126,
|
| 833 |
-
"grad_norm": 1.640625,
|
| 834 |
-
"learning_rate": 9.141134427791109e-07,
|
| 835 |
-
"loss": 1.5191,
|
| 836 |
-
"step": 580
|
| 837 |
-
},
|
| 838 |
-
{
|
| 839 |
-
"epoch": 1.598360655737705,
|
| 840 |
-
"grad_norm": 1.6640625,
|
| 841 |
-
"learning_rate": 8.998436272646393e-07,
|
| 842 |
-
"loss": 1.5177,
|
| 843 |
-
"step": 585
|
| 844 |
-
},
|
| 845 |
-
{
|
| 846 |
-
"epoch": 1.6120218579234973,
|
| 847 |
-
"grad_norm": 1.6796875,
|
| 848 |
-
"learning_rate": 8.85594384379417e-07,
|
| 849 |
-
"loss": 1.5098,
|
| 850 |
-
"step": 590
|
| 851 |
-
},
|
| 852 |
-
{
|
| 853 |
-
"epoch": 1.6256830601092895,
|
| 854 |
-
"grad_norm": 1.6875,
|
| 855 |
-
"learning_rate": 8.713686409905313e-07,
|
| 856 |
-
"loss": 1.503,
|
| 857 |
-
"step": 595
|
| 858 |
-
},
|
| 859 |
-
{
|
| 860 |
-
"epoch": 1.639344262295082,
|
| 861 |
-
"grad_norm": 1.6640625,
|
| 862 |
-
"learning_rate": 8.571693191381544e-07,
|
| 863 |
-
"loss": 1.5091,
|
| 864 |
-
"step": 600
|
| 865 |
-
},
|
| 866 |
-
{
|
| 867 |
-
"epoch": 1.639344262295082,
|
| 868 |
-
"eval_loss": 1.5181070566177368,
|
| 869 |
-
"eval_runtime": 19.9478,
|
| 870 |
-
"eval_samples_per_second": 35.493,
|
| 871 |
-
"eval_steps_per_second": 1.153,
|
| 872 |
-
"step": 600
|
| 873 |
-
},
|
| 874 |
-
{
|
| 875 |
-
"epoch": 1.6530054644808743,
|
| 876 |
-
"grad_norm": 1.6640625,
|
| 877 |
-
"learning_rate": 8.429993354353388e-07,
|
| 878 |
-
"loss": 1.5003,
|
| 879 |
-
"step": 605
|
| 880 |
-
},
|
| 881 |
-
{
|
| 882 |
-
"epoch": 1.6666666666666665,
|
| 883 |
-
"grad_norm": 1.625,
|
| 884 |
-
"learning_rate": 8.288616004689319e-07,
|
| 885 |
-
"loss": 1.5068,
|
| 886 |
-
"step": 610
|
| 887 |
-
},
|
| 888 |
-
{
|
| 889 |
-
"epoch": 1.680327868852459,
|
| 890 |
-
"grad_norm": 1.6953125,
|
| 891 |
-
"learning_rate": 8.147590182017269e-07,
|
| 892 |
-
"loss": 1.5067,
|
| 893 |
-
"step": 615
|
| 894 |
-
},
|
| 895 |
-
{
|
| 896 |
-
"epoch": 1.6939890710382515,
|
| 897 |
-
"grad_norm": 1.640625,
|
| 898 |
-
"learning_rate": 8.006944853759732e-07,
|
| 899 |
-
"loss": 1.5069,
|
| 900 |
-
"step": 620
|
| 901 |
-
},
|
| 902 |
-
{
|
| 903 |
-
"epoch": 1.7076502732240437,
|
| 904 |
-
"grad_norm": 1.7734375,
|
| 905 |
-
"learning_rate": 7.866708909183702e-07,
|
| 906 |
-
"loss": 1.4937,
|
| 907 |
-
"step": 625
|
| 908 |
-
},
|
| 909 |
-
{
|
| 910 |
-
"epoch": 1.721311475409836,
|
| 911 |
-
"grad_norm": 1.640625,
|
| 912 |
-
"learning_rate": 7.726911153466697e-07,
|
| 913 |
-
"loss": 1.5089,
|
| 914 |
-
"step": 630
|
| 915 |
-
},
|
| 916 |
-
{
|
| 917 |
-
"epoch": 1.7349726775956285,
|
| 918 |
-
"grad_norm": 1.6484375,
|
| 919 |
-
"learning_rate": 7.587580301779999e-07,
|
| 920 |
-
"loss": 1.511,
|
| 921 |
-
"step": 635
|
| 922 |
-
},
|
| 923 |
-
{
|
| 924 |
-
"epoch": 1.748633879781421,
|
| 925 |
-
"grad_norm": 1.6875,
|
| 926 |
-
"learning_rate": 7.448744973390422e-07,
|
| 927 |
-
"loss": 1.5025,
|
| 928 |
-
"step": 640
|
| 929 |
-
},
|
| 930 |
-
{
|
| 931 |
-
"epoch": 1.762295081967213,
|
| 932 |
-
"grad_norm": 1.640625,
|
| 933 |
-
"learning_rate": 7.310433685781777e-07,
|
| 934 |
-
"loss": 1.5007,
|
| 935 |
-
"step": 645
|
| 936 |
-
},
|
| 937 |
-
{
|
| 938 |
-
"epoch": 1.7759562841530054,
|
| 939 |
-
"grad_norm": 1.671875,
|
| 940 |
-
"learning_rate": 7.172674848797217e-07,
|
| 941 |
-
"loss": 1.4922,
|
| 942 |
-
"step": 650
|
| 943 |
-
},
|
| 944 |
-
{
|
| 945 |
-
"epoch": 1.789617486338798,
|
| 946 |
-
"grad_norm": 1.640625,
|
| 947 |
-
"learning_rate": 7.035496758803743e-07,
|
| 948 |
-
"loss": 1.505,
|
| 949 |
-
"step": 655
|
| 950 |
-
},
|
| 951 |
-
{
|
| 952 |
-
"epoch": 1.8032786885245902,
|
| 953 |
-
"grad_norm": 1.625,
|
| 954 |
-
"learning_rate": 6.898927592879944e-07,
|
| 955 |
-
"loss": 1.5111,
|
| 956 |
-
"step": 660
|
| 957 |
-
},
|
| 958 |
-
{
|
| 959 |
-
"epoch": 1.8169398907103824,
|
| 960 |
-
"grad_norm": 1.640625,
|
| 961 |
-
"learning_rate": 6.762995403028314e-07,
|
| 962 |
-
"loss": 1.5145,
|
| 963 |
-
"step": 665
|
| 964 |
-
},
|
| 965 |
-
{
|
| 966 |
-
"epoch": 1.830601092896175,
|
| 967 |
-
"grad_norm": 1.8359375,
|
| 968 |
-
"learning_rate": 6.627728110413213e-07,
|
| 969 |
-
"loss": 1.5097,
|
| 970 |
-
"step": 670
|
| 971 |
-
},
|
| 972 |
-
{
|
| 973 |
-
"epoch": 1.8442622950819674,
|
| 974 |
-
"grad_norm": 1.609375,
|
| 975 |
-
"learning_rate": 6.493153499625719e-07,
|
| 976 |
-
"loss": 1.5087,
|
| 977 |
-
"step": 675
|
| 978 |
-
},
|
| 979 |
-
{
|
| 980 |
-
"epoch": 1.8579234972677594,
|
| 981 |
-
"grad_norm": 2.0625,
|
| 982 |
-
"learning_rate": 6.359299212976534e-07,
|
| 983 |
-
"loss": 1.5044,
|
| 984 |
-
"step": 680
|
| 985 |
-
},
|
| 986 |
-
{
|
| 987 |
-
"epoch": 1.8715846994535519,
|
| 988 |
-
"grad_norm": 1.640625,
|
| 989 |
-
"learning_rate": 6.226192744818124e-07,
|
| 990 |
-
"loss": 1.5105,
|
| 991 |
-
"step": 685
|
| 992 |
-
},
|
| 993 |
-
{
|
| 994 |
-
"epoch": 1.8852459016393444,
|
| 995 |
-
"grad_norm": 1.6484375,
|
| 996 |
-
"learning_rate": 6.093861435897207e-07,
|
| 997 |
-
"loss": 1.5109,
|
| 998 |
-
"step": 690
|
| 999 |
-
},
|
| 1000 |
-
{
|
| 1001 |
-
"epoch": 1.8989071038251366,
|
| 1002 |
-
"grad_norm": 1.6640625,
|
| 1003 |
-
"learning_rate": 5.962332467738857e-07,
|
| 1004 |
-
"loss": 1.5103,
|
| 1005 |
-
"step": 695
|
| 1006 |
-
},
|
| 1007 |
-
{
|
| 1008 |
-
"epoch": 1.9125683060109289,
|
| 1009 |
-
"grad_norm": 1.6796875,
|
| 1010 |
-
"learning_rate": 5.83163285706327e-07,
|
| 1011 |
-
"loss": 1.5023,
|
| 1012 |
-
"step": 700
|
| 1013 |
-
},
|
| 1014 |
-
{
|
| 1015 |
-
"epoch": 1.9262295081967213,
|
| 1016 |
-
"grad_norm": 1.6640625,
|
| 1017 |
-
"learning_rate": 5.701789450236376e-07,
|
| 1018 |
-
"loss": 1.5084,
|
| 1019 |
-
"step": 705
|
| 1020 |
-
},
|
| 1021 |
-
{
|
| 1022 |
-
"epoch": 1.9398907103825138,
|
| 1023 |
-
"grad_norm": 1.6484375,
|
| 1024 |
-
"learning_rate": 5.57282891775548e-07,
|
| 1025 |
-
"loss": 1.5063,
|
| 1026 |
-
"step": 710
|
| 1027 |
-
},
|
| 1028 |
-
{
|
| 1029 |
-
"epoch": 1.9535519125683058,
|
| 1030 |
-
"grad_norm": 1.6328125,
|
| 1031 |
-
"learning_rate": 5.444777748771006e-07,
|
| 1032 |
-
"loss": 1.5074,
|
| 1033 |
-
"step": 715
|
| 1034 |
-
},
|
| 1035 |
-
{
|
| 1036 |
-
"epoch": 1.9672131147540983,
|
| 1037 |
-
"grad_norm": 1.6640625,
|
| 1038 |
-
"learning_rate": 5.317662245645469e-07,
|
| 1039 |
-
"loss": 1.5037,
|
| 1040 |
-
"step": 720
|
| 1041 |
-
},
|
| 1042 |
-
{
|
| 1043 |
-
"epoch": 1.9808743169398908,
|
| 1044 |
-
"grad_norm": 1.6796875,
|
| 1045 |
-
"learning_rate": 5.191508518550855e-07,
|
| 1046 |
-
"loss": 1.5085,
|
| 1047 |
-
"step": 725
|
| 1048 |
-
},
|
| 1049 |
-
{
|
| 1050 |
-
"epoch": 1.994535519125683,
|
| 1051 |
-
"grad_norm": 1.6015625,
|
| 1052 |
-
"learning_rate": 5.066342480105459e-07,
|
| 1053 |
-
"loss": 1.5082,
|
| 1054 |
-
"step": 730
|
| 1055 |
-
},
|
| 1056 |
-
{
|
| 1057 |
-
"epoch": 2.0081967213114753,
|
| 1058 |
-
"grad_norm": 1.6484375,
|
| 1059 |
-
"learning_rate": 4.942189840051287e-07,
|
| 1060 |
-
"loss": 1.5054,
|
| 1061 |
-
"step": 735
|
| 1062 |
-
},
|
| 1063 |
-
{
|
| 1064 |
-
"epoch": 2.021857923497268,
|
| 1065 |
-
"grad_norm": 1.84375,
|
| 1066 |
-
"learning_rate": 4.819076099973152e-07,
|
| 1067 |
-
"loss": 1.4999,
|
| 1068 |
-
"step": 740
|
| 1069 |
-
},
|
| 1070 |
-
{
|
| 1071 |
-
"epoch": 2.0355191256830603,
|
| 1072 |
-
"grad_norm": 1.6171875,
|
| 1073 |
-
"learning_rate": 4.697026548060528e-07,
|
| 1074 |
-
"loss": 1.5094,
|
| 1075 |
-
"step": 745
|
| 1076 |
-
},
|
| 1077 |
-
{
|
| 1078 |
-
"epoch": 2.0491803278688523,
|
| 1079 |
-
"grad_norm": 1.703125,
|
| 1080 |
-
"learning_rate": 4.5760662539132077e-07,
|
| 1081 |
-
"loss": 1.5065,
|
| 1082 |
-
"step": 750
|
| 1083 |
-
},
|
| 1084 |
-
{
|
| 1085 |
-
"epoch": 2.0628415300546448,
|
| 1086 |
-
"grad_norm": 1.921875,
|
| 1087 |
-
"learning_rate": 4.4562200633918943e-07,
|
| 1088 |
-
"loss": 1.5013,
|
| 1089 |
-
"step": 755
|
| 1090 |
-
},
|
| 1091 |
-
{
|
| 1092 |
-
"epoch": 2.0765027322404372,
|
| 1093 |
-
"grad_norm": 1.625,
|
| 1094 |
-
"learning_rate": 4.337512593514728e-07,
|
| 1095 |
-
"loss": 1.5041,
|
| 1096 |
-
"step": 760
|
| 1097 |
-
},
|
| 1098 |
-
{
|
| 1099 |
-
"epoch": 2.0901639344262297,
|
| 1100 |
-
"grad_norm": 1.6640625,
|
| 1101 |
-
"learning_rate": 4.2199682274008255e-07,
|
| 1102 |
-
"loss": 1.5083,
|
| 1103 |
-
"step": 765
|
| 1104 |
-
},
|
| 1105 |
-
{
|
| 1106 |
-
"epoch": 2.1038251366120218,
|
| 1107 |
-
"grad_norm": 1.6171875,
|
| 1108 |
-
"learning_rate": 4.103611109261872e-07,
|
| 1109 |
-
"loss": 1.4982,
|
| 1110 |
-
"step": 770
|
| 1111 |
-
},
|
| 1112 |
-
{
|
| 1113 |
-
"epoch": 2.1174863387978142,
|
| 1114 |
-
"grad_norm": 1.6328125,
|
| 1115 |
-
"learning_rate": 3.9884651394427625e-07,
|
| 1116 |
-
"loss": 1.5051,
|
| 1117 |
-
"step": 775
|
| 1118 |
-
},
|
| 1119 |
-
{
|
| 1120 |
-
"epoch": 2.1311475409836067,
|
| 1121 |
-
"grad_norm": 1.671875,
|
| 1122 |
-
"learning_rate": 3.8745539695123577e-07,
|
| 1123 |
-
"loss": 1.4924,
|
| 1124 |
-
"step": 780
|
| 1125 |
-
},
|
| 1126 |
-
{
|
| 1127 |
-
"epoch": 2.1448087431693987,
|
| 1128 |
-
"grad_norm": 1.65625,
|
| 1129 |
-
"learning_rate": 3.761900997405332e-07,
|
| 1130 |
-
"loss": 1.5113,
|
| 1131 |
-
"step": 785
|
| 1132 |
-
},
|
| 1133 |
-
{
|
| 1134 |
-
"epoch": 2.158469945355191,
|
| 1135 |
-
"grad_norm": 1.6484375,
|
| 1136 |
-
"learning_rate": 3.6505293626161127e-07,
|
| 1137 |
-
"loss": 1.5058,
|
| 1138 |
-
"step": 790
|
| 1139 |
-
},
|
| 1140 |
-
{
|
| 1141 |
-
"epoch": 2.1721311475409837,
|
| 1142 |
-
"grad_norm": 1.625,
|
| 1143 |
-
"learning_rate": 3.5404619414459147e-07,
|
| 1144 |
-
"loss": 1.5019,
|
| 1145 |
-
"step": 795
|
| 1146 |
-
},
|
| 1147 |
-
{
|
| 1148 |
-
"epoch": 2.185792349726776,
|
| 1149 |
-
"grad_norm": 1.640625,
|
| 1150 |
-
"learning_rate": 3.4317213423038384e-07,
|
| 1151 |
-
"loss": 1.4948,
|
| 1152 |
-
"step": 800
|
| 1153 |
-
},
|
| 1154 |
-
{
|
| 1155 |
-
"epoch": 2.199453551912568,
|
| 1156 |
-
"grad_norm": 1.625,
|
| 1157 |
-
"learning_rate": 3.32432990106297e-07,
|
| 1158 |
-
"loss": 1.5092,
|
| 1159 |
-
"step": 805
|
| 1160 |
-
},
|
| 1161 |
-
{
|
| 1162 |
-
"epoch": 2.2131147540983607,
|
| 1163 |
-
"grad_norm": 1.734375,
|
| 1164 |
-
"learning_rate": 3.2183096764724914e-07,
|
| 1165 |
-
"loss": 1.5048,
|
| 1166 |
-
"step": 810
|
| 1167 |
-
},
|
| 1168 |
-
{
|
| 1169 |
-
"epoch": 2.226775956284153,
|
| 1170 |
-
"grad_norm": 1.8984375,
|
| 1171 |
-
"learning_rate": 3.1136824456267006e-07,
|
| 1172 |
-
"loss": 1.5005,
|
| 1173 |
-
"step": 815
|
| 1174 |
-
},
|
| 1175 |
-
{
|
| 1176 |
-
"epoch": 2.240437158469945,
|
| 1177 |
-
"grad_norm": 1.7421875,
|
| 1178 |
-
"learning_rate": 3.01046969949188e-07,
|
| 1179 |
-
"loss": 1.5097,
|
| 1180 |
-
"step": 820
|
| 1181 |
-
},
|
| 1182 |
-
{
|
| 1183 |
-
"epoch": 2.2540983606557377,
|
| 1184 |
-
"grad_norm": 1.6640625,
|
| 1185 |
-
"learning_rate": 2.908692638491945e-07,
|
| 1186 |
-
"loss": 1.5054,
|
| 1187 |
-
"step": 825
|
| 1188 |
-
},
|
| 1189 |
-
{
|
| 1190 |
-
"epoch": 2.26775956284153,
|
| 1191 |
-
"grad_norm": 1.65625,
|
| 1192 |
-
"learning_rate": 2.80837216815378e-07,
|
| 1193 |
-
"loss": 1.5131,
|
| 1194 |
-
"step": 830
|
| 1195 |
-
},
|
| 1196 |
-
{
|
| 1197 |
-
"epoch": 2.281420765027322,
|
| 1198 |
-
"grad_norm": 1.671875,
|
| 1199 |
-
"learning_rate": 2.7095288948131114e-07,
|
| 1200 |
-
"loss": 1.4999,
|
| 1201 |
-
"step": 835
|
| 1202 |
-
},
|
| 1203 |
-
{
|
| 1204 |
-
"epoch": 2.2950819672131146,
|
| 1205 |
-
"grad_norm": 1.6328125,
|
| 1206 |
-
"learning_rate": 2.6121831213818826e-07,
|
| 1207 |
-
"loss": 1.4989,
|
| 1208 |
-
"step": 840
|
| 1209 |
-
},
|
| 1210 |
-
{
|
| 1211 |
-
"epoch": 2.308743169398907,
|
| 1212 |
-
"grad_norm": 1.6953125,
|
| 1213 |
-
"learning_rate": 2.51635484317793e-07,
|
| 1214 |
-
"loss": 1.495,
|
| 1215 |
-
"step": 845
|
| 1216 |
-
},
|
| 1217 |
-
{
|
| 1218 |
-
"epoch": 2.3224043715846996,
|
| 1219 |
-
"grad_norm": 1.703125,
|
| 1220 |
-
"learning_rate": 2.4220637438178313e-07,
|
| 1221 |
-
"loss": 1.5125,
|
| 1222 |
-
"step": 850
|
| 1223 |
-
},
|
| 1224 |
-
{
|
| 1225 |
-
"epoch": 2.3360655737704916,
|
| 1226 |
-
"grad_norm": 1.703125,
|
| 1227 |
-
"learning_rate": 2.3293291911738078e-07,
|
| 1228 |
-
"loss": 1.5092,
|
| 1229 |
-
"step": 855
|
| 1230 |
-
},
|
| 1231 |
-
{
|
| 1232 |
-
"epoch": 2.349726775956284,
|
| 1233 |
-
"grad_norm": 1.6640625,
|
| 1234 |
-
"learning_rate": 2.2381702333954433e-07,
|
| 1235 |
-
"loss": 1.4965,
|
| 1236 |
-
"step": 860
|
| 1237 |
-
},
|
| 1238 |
-
{
|
| 1239 |
-
"epoch": 2.3633879781420766,
|
| 1240 |
-
"grad_norm": 1.6328125,
|
| 1241 |
-
"learning_rate": 2.148605594997115e-07,
|
| 1242 |
-
"loss": 1.497,
|
| 1243 |
-
"step": 865
|
| 1244 |
-
},
|
| 1245 |
-
{
|
| 1246 |
-
"epoch": 2.3770491803278686,
|
| 1247 |
-
"grad_norm": 1.765625,
|
| 1248 |
-
"learning_rate": 2.0606536730118763e-07,
|
| 1249 |
-
"loss": 1.5081,
|
| 1250 |
-
"step": 870
|
| 1251 |
-
},
|
| 1252 |
-
{
|
| 1253 |
-
"epoch": 2.390710382513661,
|
| 1254 |
-
"grad_norm": 1.625,
|
| 1255 |
-
"learning_rate": 1.9743325332126105e-07,
|
| 1256 |
-
"loss": 1.5091,
|
| 1257 |
-
"step": 875
|
| 1258 |
-
},
|
| 1259 |
-
{
|
| 1260 |
-
"epoch": 2.4043715846994536,
|
| 1261 |
-
"grad_norm": 1.6171875,
|
| 1262 |
-
"learning_rate": 1.8896599064012298e-07,
|
| 1263 |
-
"loss": 1.5045,
|
| 1264 |
-
"step": 880
|
| 1265 |
-
},
|
| 1266 |
-
{
|
| 1267 |
-
"epoch": 2.418032786885246,
|
| 1268 |
-
"grad_norm": 1.640625,
|
| 1269 |
-
"learning_rate": 1.8066531847666888e-07,
|
| 1270 |
-
"loss": 1.5008,
|
| 1271 |
-
"step": 885
|
| 1272 |
-
},
|
| 1273 |
-
{
|
| 1274 |
-
"epoch": 2.431693989071038,
|
| 1275 |
-
"grad_norm": 1.65625,
|
| 1276 |
-
"learning_rate": 1.7253294183125222e-07,
|
| 1277 |
-
"loss": 1.511,
|
| 1278 |
-
"step": 890
|
| 1279 |
-
},
|
| 1280 |
-
{
|
| 1281 |
-
"epoch": 2.4453551912568305,
|
| 1282 |
-
"grad_norm": 1.6328125,
|
| 1283 |
-
"learning_rate": 1.645705311354697e-07,
|
| 1284 |
-
"loss": 1.4998,
|
| 1285 |
-
"step": 895
|
| 1286 |
-
},
|
| 1287 |
-
{
|
| 1288 |
-
"epoch": 2.459016393442623,
|
| 1289 |
-
"grad_norm": 1.65625,
|
| 1290 |
-
"learning_rate": 1.5677972190904621e-07,
|
| 1291 |
-
"loss": 1.5021,
|
| 1292 |
-
"step": 900
|
| 1293 |
-
},
|
| 1294 |
-
{
|
| 1295 |
-
"epoch": 2.459016393442623,
|
| 1296 |
-
"eval_loss": 1.5161317586898804,
|
| 1297 |
-
"eval_runtime": 19.9284,
|
| 1298 |
-
"eval_samples_per_second": 35.527,
|
| 1299 |
-
"eval_steps_per_second": 1.154,
|
| 1300 |
-
"step": 900
|
| 1301 |
-
},
|
| 1302 |
-
{
|
| 1303 |
-
"epoch": 2.4726775956284155,
|
| 1304 |
-
"grad_norm": 1.6328125,
|
| 1305 |
-
"learning_rate": 1.4916211442389048e-07,
|
| 1306 |
-
"loss": 1.502,
|
| 1307 |
-
"step": 905
|
| 1308 |
-
},
|
| 1309 |
-
{
|
| 1310 |
-
"epoch": 2.4863387978142075,
|
| 1311 |
-
"grad_norm": 1.625,
|
| 1312 |
-
"learning_rate": 1.4171927337539104e-07,
|
| 1313 |
-
"loss": 1.5012,
|
| 1314 |
-
"step": 910
|
| 1315 |
-
},
|
| 1316 |
-
{
|
| 1317 |
-
"epoch": 2.5,
|
| 1318 |
-
"grad_norm": 1.640625,
|
| 1319 |
-
"learning_rate": 1.344527275610202e-07,
|
| 1320 |
-
"loss": 1.5019,
|
| 1321 |
-
"step": 915
|
| 1322 |
-
},
|
| 1323 |
-
{
|
| 1324 |
-
"epoch": 2.5136612021857925,
|
| 1325 |
-
"grad_norm": 1.640625,
|
| 1326 |
-
"learning_rate": 1.273639695663108e-07,
|
| 1327 |
-
"loss": 1.5085,
|
| 1328 |
-
"step": 920
|
| 1329 |
-
},
|
| 1330 |
-
{
|
| 1331 |
-
"epoch": 2.527322404371585,
|
| 1332 |
-
"grad_norm": 1.75,
|
| 1333 |
-
"learning_rate": 1.204544554582716e-07,
|
| 1334 |
-
"loss": 1.4973,
|
| 1335 |
-
"step": 925
|
| 1336 |
-
},
|
| 1337 |
-
{
|
| 1338 |
-
"epoch": 2.540983606557377,
|
| 1339 |
-
"grad_norm": 1.671875,
|
| 1340 |
-
"learning_rate": 1.1372560448630375e-07,
|
| 1341 |
-
"loss": 1.5037,
|
| 1342 |
-
"step": 930
|
| 1343 |
-
},
|
| 1344 |
-
{
|
| 1345 |
-
"epoch": 2.5546448087431695,
|
| 1346 |
-
"grad_norm": 1.65625,
|
| 1347 |
-
"learning_rate": 1.0717879879068004e-07,
|
| 1348 |
-
"loss": 1.5005,
|
| 1349 |
-
"step": 935
|
| 1350 |
-
},
|
| 1351 |
-
{
|
| 1352 |
-
"epoch": 2.5683060109289615,
|
| 1353 |
-
"grad_norm": 1.6796875,
|
| 1354 |
-
"learning_rate": 1.0081538311864568e-07,
|
| 1355 |
-
"loss": 1.5024,
|
| 1356 |
-
"step": 940
|
| 1357 |
-
},
|
| 1358 |
-
{
|
| 1359 |
-
"epoch": 2.581967213114754,
|
| 1360 |
-
"grad_norm": 1.625,
|
| 1361 |
-
"learning_rate": 9.463666454820118e-08,
|
| 1362 |
-
"loss": 1.4988,
|
| 1363 |
-
"step": 945
|
| 1364 |
-
},
|
| 1365 |
-
{
|
| 1366 |
-
"epoch": 2.5956284153005464,
|
| 1367 |
-
"grad_norm": 1.625,
|
| 1368 |
-
"learning_rate": 8.864391221962064e-08,
|
| 1369 |
-
"loss": 1.5053,
|
| 1370 |
-
"step": 950
|
| 1371 |
-
},
|
| 1372 |
-
{
|
| 1373 |
-
"epoch": 2.609289617486339,
|
| 1374 |
-
"grad_norm": 1.6328125,
|
| 1375 |
-
"learning_rate": 8.28383570747655e-08,
|
| 1376 |
-
"loss": 1.5044,
|
| 1377 |
-
"step": 955
|
| 1378 |
-
},
|
| 1379 |
-
{
|
| 1380 |
-
"epoch": 2.6229508196721314,
|
| 1381 |
-
"grad_norm": 1.609375,
|
| 1382 |
-
"learning_rate": 7.722119160424112e-08,
|
| 1383 |
-
"loss": 1.4995,
|
| 1384 |
-
"step": 960
|
| 1385 |
-
},
|
| 1386 |
-
{
|
| 1387 |
-
"epoch": 2.6366120218579234,
|
| 1388 |
-
"grad_norm": 1.640625,
|
| 1389 |
-
"learning_rate": 7.179356960245409e-08,
|
| 1390 |
-
"loss": 1.5122,
|
| 1391 |
-
"step": 965
|
| 1392 |
-
},
|
| 1393 |
-
{
|
| 1394 |
-
"epoch": 2.650273224043716,
|
| 1395 |
-
"grad_norm": 1.7890625,
|
| 1396 |
-
"learning_rate": 6.655660593061718e-08,
|
| 1397 |
-
"loss": 1.5054,
|
| 1398 |
-
"step": 970
|
| 1399 |
-
},
|
| 1400 |
-
{
|
| 1401 |
-
"epoch": 2.663934426229508,
|
| 1402 |
-
"grad_norm": 1.6484375,
|
| 1403 |
-
"learning_rate": 6.151137628775049e-08,
|
| 1404 |
-
"loss": 1.5108,
|
| 1405 |
-
"step": 975
|
| 1406 |
-
},
|
| 1407 |
-
{
|
| 1408 |
-
"epoch": 2.6775956284153004,
|
| 1409 |
-
"grad_norm": 1.609375,
|
| 1410 |
-
"learning_rate": 5.665891698972769e-08,
|
| 1411 |
-
"loss": 1.5003,
|
| 1412 |
-
"step": 980
|
| 1413 |
-
},
|
| 1414 |
-
{
|
| 1415 |
-
"epoch": 2.691256830601093,
|
| 1416 |
-
"grad_norm": 1.640625,
|
| 1417 |
-
"learning_rate": 5.200022475641153e-08,
|
| 1418 |
-
"loss": 1.5015,
|
| 1419 |
-
"step": 985
|
| 1420 |
-
},
|
| 1421 |
-
{
|
| 1422 |
-
"epoch": 2.7049180327868854,
|
| 1423 |
-
"grad_norm": 1.609375,
|
| 1424 |
-
"learning_rate": 4.75362565069225e-08,
|
| 1425 |
-
"loss": 1.5002,
|
| 1426 |
-
"step": 990
|
| 1427 |
-
},
|
| 1428 |
-
{
|
| 1429 |
-
"epoch": 2.718579234972678,
|
| 1430 |
-
"grad_norm": 1.640625,
|
| 1431 |
-
"learning_rate": 4.326792916308242e-08,
|
| 1432 |
-
"loss": 1.5029,
|
| 1433 |
-
"step": 995
|
| 1434 |
-
},
|
| 1435 |
-
{
|
| 1436 |
-
"epoch": 2.73224043715847,
|
| 1437 |
-
"grad_norm": 1.7421875,
|
| 1438 |
-
"learning_rate": 3.919611946107493e-08,
|
| 1439 |
-
"loss": 1.5068,
|
| 1440 |
-
"step": 1000
|
| 1441 |
-
},
|
| 1442 |
-
{
|
| 1443 |
-
"epoch": 2.7459016393442623,
|
| 1444 |
-
"grad_norm": 1.921875,
|
| 1445 |
-
"learning_rate": 3.532166377135814e-08,
|
| 1446 |
-
"loss": 1.4961,
|
| 1447 |
-
"step": 1005
|
| 1448 |
-
},
|
| 1449 |
-
{
|
| 1450 |
-
"epoch": 2.7595628415300544,
|
| 1451 |
-
"grad_norm": 1.6953125,
|
| 1452 |
-
"learning_rate": 3.164535792687095e-08,
|
| 1453 |
-
"loss": 1.5,
|
| 1454 |
-
"step": 1010
|
| 1455 |
-
},
|
| 1456 |
-
{
|
| 1457 |
-
"epoch": 2.773224043715847,
|
| 1458 |
-
"grad_norm": 1.6171875,
|
| 1459 |
-
"learning_rate": 2.8167957059564095e-08,
|
| 1460 |
-
"loss": 1.5035,
|
| 1461 |
-
"step": 1015
|
| 1462 |
-
},
|
| 1463 |
-
{
|
| 1464 |
-
"epoch": 2.7868852459016393,
|
| 1465 |
-
"grad_norm": 1.6484375,
|
| 1466 |
-
"learning_rate": 2.4890175445293147e-08,
|
| 1467 |
-
"loss": 1.5007,
|
| 1468 |
-
"step": 1020
|
| 1469 |
-
},
|
| 1470 |
-
{
|
| 1471 |
-
"epoch": 2.800546448087432,
|
| 1472 |
-
"grad_norm": 1.6328125,
|
| 1473 |
-
"learning_rate": 2.1812686357101428e-08,
|
| 1474 |
-
"loss": 1.505,
|
| 1475 |
-
"step": 1025
|
| 1476 |
-
},
|
| 1477 |
-
{
|
| 1478 |
-
"epoch": 2.8142076502732243,
|
| 1479 |
-
"grad_norm": 1.6484375,
|
| 1480 |
-
"learning_rate": 1.8936121926927507e-08,
|
| 1481 |
-
"loss": 1.5066,
|
| 1482 |
-
"step": 1030
|
| 1483 |
-
},
|
| 1484 |
-
{
|
| 1485 |
-
"epoch": 2.8278688524590163,
|
| 1486 |
-
"grad_norm": 1.640625,
|
| 1487 |
-
"learning_rate": 1.6261073015761072e-08,
|
| 1488 |
-
"loss": 1.502,
|
| 1489 |
-
"step": 1035
|
| 1490 |
-
},
|
| 1491 |
-
{
|
| 1492 |
-
"epoch": 2.841530054644809,
|
| 1493 |
-
"grad_norm": 1.75,
|
| 1494 |
-
"learning_rate": 1.3788089092277688e-08,
|
| 1495 |
-
"loss": 1.5016,
|
| 1496 |
-
"step": 1040
|
| 1497 |
-
},
|
| 1498 |
-
{
|
| 1499 |
-
"epoch": 2.855191256830601,
|
| 1500 |
-
"grad_norm": 1.65625,
|
| 1501 |
-
"learning_rate": 1.1517678119975061e-08,
|
| 1502 |
-
"loss": 1.5024,
|
| 1503 |
-
"step": 1045
|
| 1504 |
-
},
|
| 1505 |
-
{
|
| 1506 |
-
"epoch": 2.8688524590163933,
|
| 1507 |
-
"grad_norm": 1.609375,
|
| 1508 |
-
"learning_rate": 9.450306452834178e-09,
|
| 1509 |
-
"loss": 1.5,
|
| 1510 |
-
"step": 1050
|
| 1511 |
-
},
|
| 1512 |
-
{
|
| 1513 |
-
"epoch": 2.8825136612021858,
|
| 1514 |
-
"grad_norm": 1.734375,
|
| 1515 |
-
"learning_rate": 7.586398739528932e-09,
|
| 1516 |
-
"loss": 1.4989,
|
| 1517 |
-
"step": 1055
|
| 1518 |
-
},
|
| 1519 |
-
{
|
| 1520 |
-
"epoch": 2.8961748633879782,
|
| 1521 |
-
"grad_norm": 1.7890625,
|
| 1522 |
-
"learning_rate": 5.926337836199891e-09,
|
| 1523 |
-
"loss": 1.507,
|
| 1524 |
-
"step": 1060
|
| 1525 |
-
},
|
| 1526 |
-
{
|
| 1527 |
-
"epoch": 2.9098360655737707,
|
| 1528 |
-
"grad_norm": 1.7734375,
|
| 1529 |
-
"learning_rate": 4.470464727814538e-09,
|
| 1530 |
-
"loss": 1.5119,
|
| 1531 |
-
"step": 1065
|
| 1532 |
-
},
|
| 1533 |
-
{
|
| 1534 |
-
"epoch": 2.9234972677595628,
|
| 1535 |
-
"grad_norm": 1.7734375,
|
| 1536 |
-
"learning_rate": 3.219078458127078e-09,
|
| 1537 |
-
"loss": 1.5072,
|
| 1538 |
-
"step": 1070
|
| 1539 |
-
},
|
| 1540 |
-
{
|
| 1541 |
-
"epoch": 2.9371584699453552,
|
| 1542 |
-
"grad_norm": 1.6328125,
|
| 1543 |
-
"learning_rate": 2.172436068252792e-09,
|
| 1544 |
-
"loss": 1.5074,
|
| 1545 |
-
"step": 1075
|
| 1546 |
-
},
|
| 1547 |
-
{
|
| 1548 |
-
"epoch": 2.9508196721311473,
|
| 1549 |
-
"grad_norm": 1.6171875,
|
| 1550 |
-
"learning_rate": 1.330752543871161e-09,
|
| 1551 |
-
"loss": 1.4947,
|
| 1552 |
-
"step": 1080
|
| 1553 |
-
},
|
| 1554 |
-
{
|
| 1555 |
-
"epoch": 2.9644808743169397,
|
| 1556 |
-
"grad_norm": 1.671875,
|
| 1557 |
-
"learning_rate": 6.942007710665221e-10,
|
| 1558 |
-
"loss": 1.5025,
|
| 1559 |
-
"step": 1085
|
| 1560 |
-
},
|
| 1561 |
-
{
|
| 1562 |
-
"epoch": 2.978142076502732,
|
| 1563 |
-
"grad_norm": 1.6640625,
|
| 1564 |
-
"learning_rate": 2.6291150081603207e-10,
|
| 1565 |
-
"loss": 1.5014,
|
| 1566 |
-
"step": 1090
|
| 1567 |
-
},
|
| 1568 |
-
{
|
| 1569 |
-
"epoch": 2.9918032786885247,
|
| 1570 |
-
"grad_norm": 1.609375,
|
| 1571 |
-
"learning_rate": 3.697332213348225e-11,
|
| 1572 |
-
"loss": 1.4964,
|
| 1573 |
-
"step": 1095
|
| 1574 |
}
|
| 1575 |
],
|
| 1576 |
"logging_steps": 5,
|
|
@@ -1585,12 +547,12 @@
|
|
| 1585 |
"should_evaluate": false,
|
| 1586 |
"should_log": false,
|
| 1587 |
"should_save": true,
|
| 1588 |
-
"should_training_stop":
|
| 1589 |
},
|
| 1590 |
"attributes": {}
|
| 1591 |
}
|
| 1592 |
},
|
| 1593 |
-
"total_flos":
|
| 1594 |
"train_batch_size": 8,
|
| 1595 |
"trial_name": null,
|
| 1596 |
"trial_params": null
|
|
|
|
| 1 |
{
|
| 2 |
"best_metric": null,
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 1.0,
|
| 5 |
"eval_steps": 300,
|
| 6 |
+
"global_step": 366,
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
|
|
| 437 |
},
|
| 438 |
{
|
| 439 |
"epoch": 0.819672131147541,
|
| 440 |
+
"eval_loss": 1.533558964729309,
|
| 441 |
+
"eval_runtime": 19.6321,
|
| 442 |
+
"eval_samples_per_second": 36.063,
|
| 443 |
+
"eval_steps_per_second": 1.172,
|
| 444 |
"step": 300
|
| 445 |
},
|
| 446 |
{
|
|
|
|
| 533 |
"learning_rate": 1.5057810159948714e-06,
|
| 534 |
"loss": 1.5148,
|
| 535 |
"step": 365
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 536 |
}
|
| 537 |
],
|
| 538 |
"logging_steps": 5,
|
|
|
|
| 547 |
"should_evaluate": false,
|
| 548 |
"should_log": false,
|
| 549 |
"should_save": true,
|
| 550 |
+
"should_training_stop": false
|
| 551 |
},
|
| 552 |
"attributes": {}
|
| 553 |
}
|
| 554 |
},
|
| 555 |
+
"total_flos": 2.3959892757042954e+18,
|
| 556 |
"train_batch_size": 8,
|
| 557 |
"trial_name": null,
|
| 558 |
"trial_params": null
|
last-checkpoint/training_args.bin
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 5880
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:9ac8158cbd21120f6861c2f8267bdae636a8018f1490c0ef0fd32a29cb938908
|
| 3 |
size 5880
|