Training in progress, epoch 2, checkpoint
Browse files
last-checkpoint/adapter_model.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 37789864
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:9d59a1bdda24e44f2444043aebd61cb910e5ddc3d63b85b414b042a835e9b5e5
|
| 3 |
size 37789864
|
last-checkpoint/optimizer.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
-
size
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:9b8f327ba865f24dfea32c7122116687661ed960bb9cd3e50057b7cac4357fd7
|
| 3 |
+
size 2622266
|
last-checkpoint/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 sha256:36a7109c13b0c541d186aea78a4bfc1b1fd2a88b673eb5bdc0ce4cb92eeeacf6
|
| 3 |
size 14244
|
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 sha256:0c005869a93a2083e266662b043cce087b86c03eccc6fd40eecb44f82b7664e2
|
| 3 |
size 1064
|
last-checkpoint/trainer_state.json
CHANGED
|
@@ -1,9 +1,9 @@
|
|
| 1 |
{
|
| 2 |
-
"best_metric": 3.
|
| 3 |
-
"best_model_checkpoint": "dq158/pingusPongus/checkpoint-
|
| 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,
|
|
@@ -506,13 +506,512 @@
|
|
| 506 |
"eval_steps_per_second": 0.511,
|
| 507 |
"eval_translation_length": 4569600,
|
| 508 |
"step": 40162
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 509 |
}
|
| 510 |
],
|
| 511 |
"logging_steps": 500,
|
| 512 |
"max_steps": 1204860,
|
| 513 |
"num_train_epochs": 30,
|
| 514 |
"save_steps": 1000,
|
| 515 |
-
"total_flos":
|
| 516 |
"trial_name": null,
|
| 517 |
"trial_params": null
|
| 518 |
}
|
|
|
|
| 1 |
{
|
| 2 |
+
"best_metric": 3.0774757862091064,
|
| 3 |
+
"best_model_checkpoint": "dq158/pingusPongus/checkpoint-80324",
|
| 4 |
+
"epoch": 2.0,
|
| 5 |
"eval_steps": 500,
|
| 6 |
+
"global_step": 80324,
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
|
|
| 506 |
"eval_steps_per_second": 0.511,
|
| 507 |
"eval_translation_length": 4569600,
|
| 508 |
"step": 40162
|
| 509 |
+
},
|
| 510 |
+
{
|
| 511 |
+
"epoch": 1.01,
|
| 512 |
+
"learning_rate": 9.972807255487761e-05,
|
| 513 |
+
"loss": 3.2504,
|
| 514 |
+
"step": 40500
|
| 515 |
+
},
|
| 516 |
+
{
|
| 517 |
+
"epoch": 1.02,
|
| 518 |
+
"learning_rate": 9.972123824349316e-05,
|
| 519 |
+
"loss": 3.23,
|
| 520 |
+
"step": 41000
|
| 521 |
+
},
|
| 522 |
+
{
|
| 523 |
+
"epoch": 1.03,
|
| 524 |
+
"learning_rate": 9.971431935185526e-05,
|
| 525 |
+
"loss": 3.1975,
|
| 526 |
+
"step": 41500
|
| 527 |
+
},
|
| 528 |
+
{
|
| 529 |
+
"epoch": 1.05,
|
| 530 |
+
"learning_rate": 9.970731589173359e-05,
|
| 531 |
+
"loss": 3.2147,
|
| 532 |
+
"step": 42000
|
| 533 |
+
},
|
| 534 |
+
{
|
| 535 |
+
"epoch": 1.06,
|
| 536 |
+
"learning_rate": 9.970022787504163e-05,
|
| 537 |
+
"loss": 3.1061,
|
| 538 |
+
"step": 42500
|
| 539 |
+
},
|
| 540 |
+
{
|
| 541 |
+
"epoch": 1.07,
|
| 542 |
+
"learning_rate": 9.969305531383673e-05,
|
| 543 |
+
"loss": 3.1693,
|
| 544 |
+
"step": 43000
|
| 545 |
+
},
|
| 546 |
+
{
|
| 547 |
+
"epoch": 1.08,
|
| 548 |
+
"learning_rate": 9.968579822032009e-05,
|
| 549 |
+
"loss": 3.1406,
|
| 550 |
+
"step": 43500
|
| 551 |
+
},
|
| 552 |
+
{
|
| 553 |
+
"epoch": 1.1,
|
| 554 |
+
"learning_rate": 9.967845660683664e-05,
|
| 555 |
+
"loss": 3.28,
|
| 556 |
+
"step": 44000
|
| 557 |
+
},
|
| 558 |
+
{
|
| 559 |
+
"epoch": 1.11,
|
| 560 |
+
"learning_rate": 9.967103048587511e-05,
|
| 561 |
+
"loss": 3.1341,
|
| 562 |
+
"step": 44500
|
| 563 |
+
},
|
| 564 |
+
{
|
| 565 |
+
"epoch": 1.12,
|
| 566 |
+
"learning_rate": 9.966351987006803e-05,
|
| 567 |
+
"loss": 3.1911,
|
| 568 |
+
"step": 45000
|
| 569 |
+
},
|
| 570 |
+
{
|
| 571 |
+
"epoch": 1.13,
|
| 572 |
+
"learning_rate": 9.965592477219158e-05,
|
| 573 |
+
"loss": 3.2183,
|
| 574 |
+
"step": 45500
|
| 575 |
+
},
|
| 576 |
+
{
|
| 577 |
+
"epoch": 1.15,
|
| 578 |
+
"learning_rate": 9.964824520516576e-05,
|
| 579 |
+
"loss": 3.1452,
|
| 580 |
+
"step": 46000
|
| 581 |
+
},
|
| 582 |
+
{
|
| 583 |
+
"epoch": 1.16,
|
| 584 |
+
"learning_rate": 9.964048118205414e-05,
|
| 585 |
+
"loss": 3.2294,
|
| 586 |
+
"step": 46500
|
| 587 |
+
},
|
| 588 |
+
{
|
| 589 |
+
"epoch": 1.17,
|
| 590 |
+
"learning_rate": 9.963263271606403e-05,
|
| 591 |
+
"loss": 3.2148,
|
| 592 |
+
"step": 47000
|
| 593 |
+
},
|
| 594 |
+
{
|
| 595 |
+
"epoch": 1.18,
|
| 596 |
+
"learning_rate": 9.962469982054638e-05,
|
| 597 |
+
"loss": 3.2715,
|
| 598 |
+
"step": 47500
|
| 599 |
+
},
|
| 600 |
+
{
|
| 601 |
+
"epoch": 1.2,
|
| 602 |
+
"learning_rate": 9.961668250899575e-05,
|
| 603 |
+
"loss": 3.1949,
|
| 604 |
+
"step": 48000
|
| 605 |
+
},
|
| 606 |
+
{
|
| 607 |
+
"epoch": 1.21,
|
| 608 |
+
"learning_rate": 9.96085807950503e-05,
|
| 609 |
+
"loss": 3.1926,
|
| 610 |
+
"step": 48500
|
| 611 |
+
},
|
| 612 |
+
{
|
| 613 |
+
"epoch": 1.22,
|
| 614 |
+
"learning_rate": 9.960039469249177e-05,
|
| 615 |
+
"loss": 3.2328,
|
| 616 |
+
"step": 49000
|
| 617 |
+
},
|
| 618 |
+
{
|
| 619 |
+
"epoch": 1.23,
|
| 620 |
+
"learning_rate": 9.959212421524542e-05,
|
| 621 |
+
"loss": 3.121,
|
| 622 |
+
"step": 49500
|
| 623 |
+
},
|
| 624 |
+
{
|
| 625 |
+
"epoch": 1.24,
|
| 626 |
+
"learning_rate": 9.95837693773801e-05,
|
| 627 |
+
"loss": 3.1933,
|
| 628 |
+
"step": 50000
|
| 629 |
+
},
|
| 630 |
+
{
|
| 631 |
+
"epoch": 1.26,
|
| 632 |
+
"learning_rate": 9.957533019310813e-05,
|
| 633 |
+
"loss": 3.1807,
|
| 634 |
+
"step": 50500
|
| 635 |
+
},
|
| 636 |
+
{
|
| 637 |
+
"epoch": 1.27,
|
| 638 |
+
"learning_rate": 9.956680667678531e-05,
|
| 639 |
+
"loss": 3.2696,
|
| 640 |
+
"step": 51000
|
| 641 |
+
},
|
| 642 |
+
{
|
| 643 |
+
"epoch": 1.28,
|
| 644 |
+
"learning_rate": 9.955819884291088e-05,
|
| 645 |
+
"loss": 3.1498,
|
| 646 |
+
"step": 51500
|
| 647 |
+
},
|
| 648 |
+
{
|
| 649 |
+
"epoch": 1.29,
|
| 650 |
+
"learning_rate": 9.954950670612758e-05,
|
| 651 |
+
"loss": 3.1895,
|
| 652 |
+
"step": 52000
|
| 653 |
+
},
|
| 654 |
+
{
|
| 655 |
+
"epoch": 1.31,
|
| 656 |
+
"learning_rate": 9.954073028122147e-05,
|
| 657 |
+
"loss": 3.2206,
|
| 658 |
+
"step": 52500
|
| 659 |
+
},
|
| 660 |
+
{
|
| 661 |
+
"epoch": 1.32,
|
| 662 |
+
"learning_rate": 9.953186958312204e-05,
|
| 663 |
+
"loss": 3.1473,
|
| 664 |
+
"step": 53000
|
| 665 |
+
},
|
| 666 |
+
{
|
| 667 |
+
"epoch": 1.33,
|
| 668 |
+
"learning_rate": 9.952292462690212e-05,
|
| 669 |
+
"loss": 3.2062,
|
| 670 |
+
"step": 53500
|
| 671 |
+
},
|
| 672 |
+
{
|
| 673 |
+
"epoch": 1.34,
|
| 674 |
+
"learning_rate": 9.951389542777789e-05,
|
| 675 |
+
"loss": 3.2043,
|
| 676 |
+
"step": 54000
|
| 677 |
+
},
|
| 678 |
+
{
|
| 679 |
+
"epoch": 1.36,
|
| 680 |
+
"learning_rate": 9.95047820011088e-05,
|
| 681 |
+
"loss": 3.1293,
|
| 682 |
+
"step": 54500
|
| 683 |
+
},
|
| 684 |
+
{
|
| 685 |
+
"epoch": 1.37,
|
| 686 |
+
"learning_rate": 9.949558436239762e-05,
|
| 687 |
+
"loss": 3.2194,
|
| 688 |
+
"step": 55000
|
| 689 |
+
},
|
| 690 |
+
{
|
| 691 |
+
"epoch": 1.38,
|
| 692 |
+
"learning_rate": 9.948630252729036e-05,
|
| 693 |
+
"loss": 3.2101,
|
| 694 |
+
"step": 55500
|
| 695 |
+
},
|
| 696 |
+
{
|
| 697 |
+
"epoch": 1.39,
|
| 698 |
+
"learning_rate": 9.947693651157621e-05,
|
| 699 |
+
"loss": 3.2321,
|
| 700 |
+
"step": 56000
|
| 701 |
+
},
|
| 702 |
+
{
|
| 703 |
+
"epoch": 1.41,
|
| 704 |
+
"learning_rate": 9.946748633118766e-05,
|
| 705 |
+
"loss": 3.2132,
|
| 706 |
+
"step": 56500
|
| 707 |
+
},
|
| 708 |
+
{
|
| 709 |
+
"epoch": 1.42,
|
| 710 |
+
"learning_rate": 9.945795200220022e-05,
|
| 711 |
+
"loss": 3.2587,
|
| 712 |
+
"step": 57000
|
| 713 |
+
},
|
| 714 |
+
{
|
| 715 |
+
"epoch": 1.43,
|
| 716 |
+
"learning_rate": 9.944833354083273e-05,
|
| 717 |
+
"loss": 3.1798,
|
| 718 |
+
"step": 57500
|
| 719 |
+
},
|
| 720 |
+
{
|
| 721 |
+
"epoch": 1.44,
|
| 722 |
+
"learning_rate": 9.943863096344698e-05,
|
| 723 |
+
"loss": 3.2519,
|
| 724 |
+
"step": 58000
|
| 725 |
+
},
|
| 726 |
+
{
|
| 727 |
+
"epoch": 1.46,
|
| 728 |
+
"learning_rate": 9.942884428654794e-05,
|
| 729 |
+
"loss": 3.1065,
|
| 730 |
+
"step": 58500
|
| 731 |
+
},
|
| 732 |
+
{
|
| 733 |
+
"epoch": 1.47,
|
| 734 |
+
"learning_rate": 9.941897352678362e-05,
|
| 735 |
+
"loss": 3.3128,
|
| 736 |
+
"step": 59000
|
| 737 |
+
},
|
| 738 |
+
{
|
| 739 |
+
"epoch": 1.48,
|
| 740 |
+
"learning_rate": 9.940901870094506e-05,
|
| 741 |
+
"loss": 3.2352,
|
| 742 |
+
"step": 59500
|
| 743 |
+
},
|
| 744 |
+
{
|
| 745 |
+
"epoch": 1.49,
|
| 746 |
+
"learning_rate": 9.939897982596631e-05,
|
| 747 |
+
"loss": 3.1773,
|
| 748 |
+
"step": 60000
|
| 749 |
+
},
|
| 750 |
+
{
|
| 751 |
+
"epoch": 1.51,
|
| 752 |
+
"learning_rate": 9.938885691892437e-05,
|
| 753 |
+
"loss": 3.2472,
|
| 754 |
+
"step": 60500
|
| 755 |
+
},
|
| 756 |
+
{
|
| 757 |
+
"epoch": 1.52,
|
| 758 |
+
"learning_rate": 9.937864999703925e-05,
|
| 759 |
+
"loss": 3.0983,
|
| 760 |
+
"step": 61000
|
| 761 |
+
},
|
| 762 |
+
{
|
| 763 |
+
"epoch": 1.53,
|
| 764 |
+
"learning_rate": 9.936835907767378e-05,
|
| 765 |
+
"loss": 3.2046,
|
| 766 |
+
"step": 61500
|
| 767 |
+
},
|
| 768 |
+
{
|
| 769 |
+
"epoch": 1.54,
|
| 770 |
+
"learning_rate": 9.935798417833376e-05,
|
| 771 |
+
"loss": 3.1239,
|
| 772 |
+
"step": 62000
|
| 773 |
+
},
|
| 774 |
+
{
|
| 775 |
+
"epoch": 1.56,
|
| 776 |
+
"learning_rate": 9.934752531666782e-05,
|
| 777 |
+
"loss": 3.2518,
|
| 778 |
+
"step": 62500
|
| 779 |
+
},
|
| 780 |
+
{
|
| 781 |
+
"epoch": 1.57,
|
| 782 |
+
"learning_rate": 9.933698251046739e-05,
|
| 783 |
+
"loss": 3.2518,
|
| 784 |
+
"step": 63000
|
| 785 |
+
},
|
| 786 |
+
{
|
| 787 |
+
"epoch": 1.58,
|
| 788 |
+
"learning_rate": 9.932635577766676e-05,
|
| 789 |
+
"loss": 3.0939,
|
| 790 |
+
"step": 63500
|
| 791 |
+
},
|
| 792 |
+
{
|
| 793 |
+
"epoch": 1.59,
|
| 794 |
+
"learning_rate": 9.931564513634291e-05,
|
| 795 |
+
"loss": 3.2243,
|
| 796 |
+
"step": 64000
|
| 797 |
+
},
|
| 798 |
+
{
|
| 799 |
+
"epoch": 1.61,
|
| 800 |
+
"learning_rate": 9.930485060471562e-05,
|
| 801 |
+
"loss": 3.196,
|
| 802 |
+
"step": 64500
|
| 803 |
+
},
|
| 804 |
+
{
|
| 805 |
+
"epoch": 1.62,
|
| 806 |
+
"learning_rate": 9.929397220114736e-05,
|
| 807 |
+
"loss": 3.2016,
|
| 808 |
+
"step": 65000
|
| 809 |
+
},
|
| 810 |
+
{
|
| 811 |
+
"epoch": 1.63,
|
| 812 |
+
"learning_rate": 9.928300994414321e-05,
|
| 813 |
+
"loss": 3.1955,
|
| 814 |
+
"step": 65500
|
| 815 |
+
},
|
| 816 |
+
{
|
| 817 |
+
"epoch": 1.64,
|
| 818 |
+
"learning_rate": 9.9271963852351e-05,
|
| 819 |
+
"loss": 3.2093,
|
| 820 |
+
"step": 66000
|
| 821 |
+
},
|
| 822 |
+
{
|
| 823 |
+
"epoch": 1.66,
|
| 824 |
+
"learning_rate": 9.92608339445611e-05,
|
| 825 |
+
"loss": 3.2283,
|
| 826 |
+
"step": 66500
|
| 827 |
+
},
|
| 828 |
+
{
|
| 829 |
+
"epoch": 1.67,
|
| 830 |
+
"learning_rate": 9.924962023970646e-05,
|
| 831 |
+
"loss": 3.2368,
|
| 832 |
+
"step": 67000
|
| 833 |
+
},
|
| 834 |
+
{
|
| 835 |
+
"epoch": 1.68,
|
| 836 |
+
"learning_rate": 9.92383227568626e-05,
|
| 837 |
+
"loss": 3.1782,
|
| 838 |
+
"step": 67500
|
| 839 |
+
},
|
| 840 |
+
{
|
| 841 |
+
"epoch": 1.69,
|
| 842 |
+
"learning_rate": 9.922694151524756e-05,
|
| 843 |
+
"loss": 3.1762,
|
| 844 |
+
"step": 68000
|
| 845 |
+
},
|
| 846 |
+
{
|
| 847 |
+
"epoch": 1.71,
|
| 848 |
+
"learning_rate": 9.921547653422182e-05,
|
| 849 |
+
"loss": 3.2227,
|
| 850 |
+
"step": 68500
|
| 851 |
+
},
|
| 852 |
+
{
|
| 853 |
+
"epoch": 1.72,
|
| 854 |
+
"learning_rate": 9.920392783328834e-05,
|
| 855 |
+
"loss": 3.1744,
|
| 856 |
+
"step": 69000
|
| 857 |
+
},
|
| 858 |
+
{
|
| 859 |
+
"epoch": 1.73,
|
| 860 |
+
"learning_rate": 9.91922954320925e-05,
|
| 861 |
+
"loss": 3.1868,
|
| 862 |
+
"step": 69500
|
| 863 |
+
},
|
| 864 |
+
{
|
| 865 |
+
"epoch": 1.74,
|
| 866 |
+
"learning_rate": 9.918057935042204e-05,
|
| 867 |
+
"loss": 3.1313,
|
| 868 |
+
"step": 70000
|
| 869 |
+
},
|
| 870 |
+
{
|
| 871 |
+
"epoch": 1.76,
|
| 872 |
+
"learning_rate": 9.916877960820705e-05,
|
| 873 |
+
"loss": 3.1623,
|
| 874 |
+
"step": 70500
|
| 875 |
+
},
|
| 876 |
+
{
|
| 877 |
+
"epoch": 1.77,
|
| 878 |
+
"learning_rate": 9.915689622551996e-05,
|
| 879 |
+
"loss": 3.0594,
|
| 880 |
+
"step": 71000
|
| 881 |
+
},
|
| 882 |
+
{
|
| 883 |
+
"epoch": 1.78,
|
| 884 |
+
"learning_rate": 9.914492922257546e-05,
|
| 885 |
+
"loss": 3.2163,
|
| 886 |
+
"step": 71500
|
| 887 |
+
},
|
| 888 |
+
{
|
| 889 |
+
"epoch": 1.79,
|
| 890 |
+
"learning_rate": 9.913287861973049e-05,
|
| 891 |
+
"loss": 3.0811,
|
| 892 |
+
"step": 72000
|
| 893 |
+
},
|
| 894 |
+
{
|
| 895 |
+
"epoch": 1.81,
|
| 896 |
+
"learning_rate": 9.912074443748416e-05,
|
| 897 |
+
"loss": 3.2226,
|
| 898 |
+
"step": 72500
|
| 899 |
+
},
|
| 900 |
+
{
|
| 901 |
+
"epoch": 1.82,
|
| 902 |
+
"learning_rate": 9.910852669647785e-05,
|
| 903 |
+
"loss": 3.2876,
|
| 904 |
+
"step": 73000
|
| 905 |
+
},
|
| 906 |
+
{
|
| 907 |
+
"epoch": 1.83,
|
| 908 |
+
"learning_rate": 9.909622541749499e-05,
|
| 909 |
+
"loss": 3.2285,
|
| 910 |
+
"step": 73500
|
| 911 |
+
},
|
| 912 |
+
{
|
| 913 |
+
"epoch": 1.84,
|
| 914 |
+
"learning_rate": 9.908384062146118e-05,
|
| 915 |
+
"loss": 3.1274,
|
| 916 |
+
"step": 74000
|
| 917 |
+
},
|
| 918 |
+
{
|
| 919 |
+
"epoch": 1.85,
|
| 920 |
+
"learning_rate": 9.907137232944404e-05,
|
| 921 |
+
"loss": 3.1894,
|
| 922 |
+
"step": 74500
|
| 923 |
+
},
|
| 924 |
+
{
|
| 925 |
+
"epoch": 1.87,
|
| 926 |
+
"learning_rate": 9.905882056265323e-05,
|
| 927 |
+
"loss": 3.1468,
|
| 928 |
+
"step": 75000
|
| 929 |
+
},
|
| 930 |
+
{
|
| 931 |
+
"epoch": 1.88,
|
| 932 |
+
"learning_rate": 9.904618534244044e-05,
|
| 933 |
+
"loss": 3.2397,
|
| 934 |
+
"step": 75500
|
| 935 |
+
},
|
| 936 |
+
{
|
| 937 |
+
"epoch": 1.89,
|
| 938 |
+
"learning_rate": 9.903346669029932e-05,
|
| 939 |
+
"loss": 3.0135,
|
| 940 |
+
"step": 76000
|
| 941 |
+
},
|
| 942 |
+
{
|
| 943 |
+
"epoch": 1.9,
|
| 944 |
+
"learning_rate": 9.90206646278654e-05,
|
| 945 |
+
"loss": 3.2206,
|
| 946 |
+
"step": 76500
|
| 947 |
+
},
|
| 948 |
+
{
|
| 949 |
+
"epoch": 1.92,
|
| 950 |
+
"learning_rate": 9.900777917691615e-05,
|
| 951 |
+
"loss": 3.1868,
|
| 952 |
+
"step": 77000
|
| 953 |
+
},
|
| 954 |
+
{
|
| 955 |
+
"epoch": 1.93,
|
| 956 |
+
"learning_rate": 9.899481035937086e-05,
|
| 957 |
+
"loss": 3.2184,
|
| 958 |
+
"step": 77500
|
| 959 |
+
},
|
| 960 |
+
{
|
| 961 |
+
"epoch": 1.94,
|
| 962 |
+
"learning_rate": 9.898175819729063e-05,
|
| 963 |
+
"loss": 3.1739,
|
| 964 |
+
"step": 78000
|
| 965 |
+
},
|
| 966 |
+
{
|
| 967 |
+
"epoch": 1.95,
|
| 968 |
+
"learning_rate": 9.896862271287839e-05,
|
| 969 |
+
"loss": 3.2227,
|
| 970 |
+
"step": 78500
|
| 971 |
+
},
|
| 972 |
+
{
|
| 973 |
+
"epoch": 1.97,
|
| 974 |
+
"learning_rate": 9.895540392847874e-05,
|
| 975 |
+
"loss": 3.2236,
|
| 976 |
+
"step": 79000
|
| 977 |
+
},
|
| 978 |
+
{
|
| 979 |
+
"epoch": 1.98,
|
| 980 |
+
"learning_rate": 9.8942101866578e-05,
|
| 981 |
+
"loss": 3.1321,
|
| 982 |
+
"step": 79500
|
| 983 |
+
},
|
| 984 |
+
{
|
| 985 |
+
"epoch": 1.99,
|
| 986 |
+
"learning_rate": 9.892871654980418e-05,
|
| 987 |
+
"loss": 3.1849,
|
| 988 |
+
"step": 80000
|
| 989 |
+
},
|
| 990 |
+
{
|
| 991 |
+
"epoch": 2.0,
|
| 992 |
+
"eval_bleu": 1.0,
|
| 993 |
+
"eval_brevity_penalty": 1.0,
|
| 994 |
+
"eval_length_ratio": 1.0,
|
| 995 |
+
"eval_loss": 3.0774757862091064,
|
| 996 |
+
"eval_precisions": [
|
| 997 |
+
1.0,
|
| 998 |
+
1.0,
|
| 999 |
+
1.0,
|
| 1000 |
+
1.0
|
| 1001 |
+
],
|
| 1002 |
+
"eval_reference_length": 4569600,
|
| 1003 |
+
"eval_runtime": 8605.765,
|
| 1004 |
+
"eval_samples_per_second": 1.037,
|
| 1005 |
+
"eval_steps_per_second": 0.519,
|
| 1006 |
+
"eval_translation_length": 4569600,
|
| 1007 |
+
"step": 80324
|
| 1008 |
}
|
| 1009 |
],
|
| 1010 |
"logging_steps": 500,
|
| 1011 |
"max_steps": 1204860,
|
| 1012 |
"num_train_epochs": 30,
|
| 1013 |
"save_steps": 1000,
|
| 1014 |
+
"total_flos": 1.3785697922643395e+18,
|
| 1015 |
"trial_name": null,
|
| 1016 |
"trial_params": null
|
| 1017 |
}
|