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 +327 -7
- training_args.bin +1 -1
adapter_model.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 201361312
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:8ba743cbcbe7a17a13ffee64e044e449254882634e848aa631f63e6778810b27
|
| 3 |
size 201361312
|
optimizer.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 402868986
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:aa8c590b229a780debdb448bc28cb8f79b28f2ec2c6ea5636a4abf950ae5a038
|
| 3 |
size 402868986
|
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 sha256:78412adf2dda42daa646069b544a18df9b06cb455b0068bb5473d031abd28e97
|
| 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 sha256:45e12526c8172a948234d8cb869935e517c484d36da5eb6ac9a7382e7d268eff
|
| 3 |
size 1064
|
trainer_state.json
CHANGED
|
@@ -1,9 +1,9 @@
|
|
| 1 |
{
|
| 2 |
-
"best_metric": 0.
|
| 3 |
-
"best_model_checkpoint": "//outputs/task7_microsoft/Phi-3.5-mini-instruct/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,
|
|
@@ -471,12 +471,332 @@
|
|
| 471 |
"eval_samples_per_second": 4.46,
|
| 472 |
"eval_steps_per_second": 0.595,
|
| 473 |
"step": 200
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 474 |
}
|
| 475 |
],
|
| 476 |
"logging_steps": 5,
|
| 477 |
-
"max_steps":
|
| 478 |
"num_input_tokens_seen": 0,
|
| 479 |
-
"num_train_epochs":
|
| 480 |
"save_steps": 500,
|
| 481 |
"stateful_callbacks": {
|
| 482 |
"TrainerControl": {
|
|
@@ -490,7 +810,7 @@
|
|
| 490 |
"attributes": {}
|
| 491 |
}
|
| 492 |
},
|
| 493 |
-
"total_flos": 1.
|
| 494 |
"train_batch_size": 2,
|
| 495 |
"trial_name": null,
|
| 496 |
"trial_params": null
|
|
|
|
| 1 |
{
|
| 2 |
+
"best_metric": 0.5424160957336426,
|
| 3 |
+
"best_model_checkpoint": "//outputs/task7_microsoft/Phi-3.5-mini-instruct/checkpoint-400",
|
| 4 |
+
"epoch": 8.0,
|
| 5 |
"eval_steps": 500,
|
| 6 |
+
"global_step": 400,
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
|
|
| 471 |
"eval_samples_per_second": 4.46,
|
| 472 |
"eval_steps_per_second": 0.595,
|
| 473 |
"step": 200
|
| 474 |
+
},
|
| 475 |
+
{
|
| 476 |
+
"epoch": 3.586666666666667,
|
| 477 |
+
"eval_loss": 0.5931960940361023,
|
| 478 |
+
"eval_runtime": 28.5506,
|
| 479 |
+
"eval_samples_per_second": 3.503,
|
| 480 |
+
"eval_steps_per_second": 0.455,
|
| 481 |
+
"step": 201
|
| 482 |
+
},
|
| 483 |
+
{
|
| 484 |
+
"epoch": 4.1,
|
| 485 |
+
"grad_norm": 0.8143700957298279,
|
| 486 |
+
"learning_rate": 5.206624871244066e-06,
|
| 487 |
+
"loss": 0.9672,
|
| 488 |
+
"step": 205
|
| 489 |
+
},
|
| 490 |
+
{
|
| 491 |
+
"epoch": 4.2,
|
| 492 |
+
"grad_norm": 0.670274019241333,
|
| 493 |
+
"learning_rate": 5e-06,
|
| 494 |
+
"loss": 0.9171,
|
| 495 |
+
"step": 210
|
| 496 |
+
},
|
| 497 |
+
{
|
| 498 |
+
"epoch": 4.3,
|
| 499 |
+
"grad_norm": 0.5900228023529053,
|
| 500 |
+
"learning_rate": 4.793375128755934e-06,
|
| 501 |
+
"loss": 0.8865,
|
| 502 |
+
"step": 215
|
| 503 |
+
},
|
| 504 |
+
{
|
| 505 |
+
"epoch": 4.4,
|
| 506 |
+
"grad_norm": 0.5981155633926392,
|
| 507 |
+
"learning_rate": 4.587103272638339e-06,
|
| 508 |
+
"loss": 1.1775,
|
| 509 |
+
"step": 220
|
| 510 |
+
},
|
| 511 |
+
{
|
| 512 |
+
"epoch": 4.5,
|
| 513 |
+
"grad_norm": 0.5991724729537964,
|
| 514 |
+
"learning_rate": 4.381536843653262e-06,
|
| 515 |
+
"loss": 0.7489,
|
| 516 |
+
"step": 225
|
| 517 |
+
},
|
| 518 |
+
{
|
| 519 |
+
"epoch": 4.6,
|
| 520 |
+
"grad_norm": 0.5450884103775024,
|
| 521 |
+
"learning_rate": 4.17702704859633e-06,
|
| 522 |
+
"loss": 0.8612,
|
| 523 |
+
"step": 230
|
| 524 |
+
},
|
| 525 |
+
{
|
| 526 |
+
"epoch": 4.7,
|
| 527 |
+
"grad_norm": 0.444416344165802,
|
| 528 |
+
"learning_rate": 3.973923289021829e-06,
|
| 529 |
+
"loss": 0.7293,
|
| 530 |
+
"step": 235
|
| 531 |
+
},
|
| 532 |
+
{
|
| 533 |
+
"epoch": 4.8,
|
| 534 |
+
"grad_norm": 0.3834201395511627,
|
| 535 |
+
"learning_rate": 3.7725725642960047e-06,
|
| 536 |
+
"loss": 0.7699,
|
| 537 |
+
"step": 240
|
| 538 |
+
},
|
| 539 |
+
{
|
| 540 |
+
"epoch": 4.9,
|
| 541 |
+
"grad_norm": 0.3441762924194336,
|
| 542 |
+
"learning_rate": 3.573318878754475e-06,
|
| 543 |
+
"loss": 0.8972,
|
| 544 |
+
"step": 245
|
| 545 |
+
},
|
| 546 |
+
{
|
| 547 |
+
"epoch": 5.0,
|
| 548 |
+
"grad_norm": 0.5351847410202026,
|
| 549 |
+
"learning_rate": 3.3765026539765832e-06,
|
| 550 |
+
"loss": 0.6602,
|
| 551 |
+
"step": 250
|
| 552 |
+
},
|
| 553 |
+
{
|
| 554 |
+
"epoch": 5.0,
|
| 555 |
+
"eval_loss": 0.5578957200050354,
|
| 556 |
+
"eval_runtime": 52.5326,
|
| 557 |
+
"eval_samples_per_second": 3.807,
|
| 558 |
+
"eval_steps_per_second": 0.476,
|
| 559 |
+
"step": 250
|
| 560 |
+
},
|
| 561 |
+
{
|
| 562 |
+
"epoch": 5.1,
|
| 563 |
+
"grad_norm": 0.37455469369888306,
|
| 564 |
+
"learning_rate": 3.1824601471808504e-06,
|
| 565 |
+
"loss": 0.884,
|
| 566 |
+
"step": 255
|
| 567 |
+
},
|
| 568 |
+
{
|
| 569 |
+
"epoch": 5.2,
|
| 570 |
+
"grad_norm": 0.6285215020179749,
|
| 571 |
+
"learning_rate": 2.991522876735154e-06,
|
| 572 |
+
"loss": 0.8042,
|
| 573 |
+
"step": 260
|
| 574 |
+
},
|
| 575 |
+
{
|
| 576 |
+
"epoch": 5.3,
|
| 577 |
+
"grad_norm": 0.37903887033462524,
|
| 578 |
+
"learning_rate": 2.804017055763149e-06,
|
| 579 |
+
"loss": 0.6865,
|
| 580 |
+
"step": 265
|
| 581 |
+
},
|
| 582 |
+
{
|
| 583 |
+
"epoch": 5.4,
|
| 584 |
+
"grad_norm": 0.4468790292739868,
|
| 585 |
+
"learning_rate": 2.6202630348146323e-06,
|
| 586 |
+
"loss": 0.9571,
|
| 587 |
+
"step": 270
|
| 588 |
+
},
|
| 589 |
+
{
|
| 590 |
+
"epoch": 5.5,
|
| 591 |
+
"grad_norm": 2.321368932723999,
|
| 592 |
+
"learning_rate": 2.4405747545519966e-06,
|
| 593 |
+
"loss": 0.7722,
|
| 594 |
+
"step": 275
|
| 595 |
+
},
|
| 596 |
+
{
|
| 597 |
+
"epoch": 5.6,
|
| 598 |
+
"grad_norm": 0.3462996482849121,
|
| 599 |
+
"learning_rate": 2.265259209387867e-06,
|
| 600 |
+
"loss": 0.6575,
|
| 601 |
+
"step": 280
|
| 602 |
+
},
|
| 603 |
+
{
|
| 604 |
+
"epoch": 5.7,
|
| 605 |
+
"grad_norm": 0.7634517550468445,
|
| 606 |
+
"learning_rate": 2.094615922990309e-06,
|
| 607 |
+
"loss": 0.7036,
|
| 608 |
+
"step": 285
|
| 609 |
+
},
|
| 610 |
+
{
|
| 611 |
+
"epoch": 5.8,
|
| 612 |
+
"grad_norm": 0.33972227573394775,
|
| 613 |
+
"learning_rate": 1.928936436551661e-06,
|
| 614 |
+
"loss": 0.6193,
|
| 615 |
+
"step": 290
|
| 616 |
+
},
|
| 617 |
+
{
|
| 618 |
+
"epoch": 5.9,
|
| 619 |
+
"grad_norm": 0.863368570804596,
|
| 620 |
+
"learning_rate": 1.7685038106952952e-06,
|
| 621 |
+
"loss": 0.7429,
|
| 622 |
+
"step": 295
|
| 623 |
+
},
|
| 624 |
+
{
|
| 625 |
+
"epoch": 6.0,
|
| 626 |
+
"grad_norm": 0.8421957492828369,
|
| 627 |
+
"learning_rate": 1.6135921418712959e-06,
|
| 628 |
+
"loss": 0.6177,
|
| 629 |
+
"step": 300
|
| 630 |
+
},
|
| 631 |
+
{
|
| 632 |
+
"epoch": 6.0,
|
| 633 |
+
"eval_loss": 0.5471388697624207,
|
| 634 |
+
"eval_runtime": 52.1971,
|
| 635 |
+
"eval_samples_per_second": 3.832,
|
| 636 |
+
"eval_steps_per_second": 0.479,
|
| 637 |
+
"step": 300
|
| 638 |
+
},
|
| 639 |
+
{
|
| 640 |
+
"epoch": 6.1,
|
| 641 |
+
"grad_norm": 0.42387768626213074,
|
| 642 |
+
"learning_rate": 1.4644660940672628e-06,
|
| 643 |
+
"loss": 0.7107,
|
| 644 |
+
"step": 305
|
| 645 |
+
},
|
| 646 |
+
{
|
| 647 |
+
"epoch": 6.2,
|
| 648 |
+
"grad_norm": 0.40212640166282654,
|
| 649 |
+
"learning_rate": 1.321380446634342e-06,
|
| 650 |
+
"loss": 0.6465,
|
| 651 |
+
"step": 310
|
| 652 |
+
},
|
| 653 |
+
{
|
| 654 |
+
"epoch": 6.3,
|
| 655 |
+
"grad_norm": 0.38275906443595886,
|
| 656 |
+
"learning_rate": 1.1845796590009684e-06,
|
| 657 |
+
"loss": 0.7838,
|
| 658 |
+
"step": 315
|
| 659 |
+
},
|
| 660 |
+
{
|
| 661 |
+
"epoch": 6.4,
|
| 662 |
+
"grad_norm": 0.517331063747406,
|
| 663 |
+
"learning_rate": 1.0542974530180327e-06,
|
| 664 |
+
"loss": 0.6743,
|
| 665 |
+
"step": 320
|
| 666 |
+
},
|
| 667 |
+
{
|
| 668 |
+
"epoch": 6.5,
|
| 669 |
+
"grad_norm": 0.4819343388080597,
|
| 670 |
+
"learning_rate": 9.307564136490255e-07,
|
| 671 |
+
"loss": 0.6544,
|
| 672 |
+
"step": 325
|
| 673 |
+
},
|
| 674 |
+
{
|
| 675 |
+
"epoch": 6.6,
|
| 676 |
+
"grad_norm": 0.5918112397193909,
|
| 677 |
+
"learning_rate": 8.141676086873574e-07,
|
| 678 |
+
"loss": 0.6178,
|
| 679 |
+
"step": 330
|
| 680 |
+
},
|
| 681 |
+
{
|
| 682 |
+
"epoch": 6.7,
|
| 683 |
+
"grad_norm": 0.3847924768924713,
|
| 684 |
+
"learning_rate": 7.047302281505735e-07,
|
| 685 |
+
"loss": 0.5631,
|
| 686 |
+
"step": 335
|
| 687 |
+
},
|
| 688 |
+
{
|
| 689 |
+
"epoch": 6.8,
|
| 690 |
+
"grad_norm": 0.43630239367485046,
|
| 691 |
+
"learning_rate": 6.026312439675553e-07,
|
| 692 |
+
"loss": 0.5709,
|
| 693 |
+
"step": 340
|
| 694 |
+
},
|
| 695 |
+
{
|
| 696 |
+
"epoch": 6.9,
|
| 697 |
+
"grad_norm": 0.6350282430648804,
|
| 698 |
+
"learning_rate": 5.080450905401057e-07,
|
| 699 |
+
"loss": 0.7065,
|
| 700 |
+
"step": 345
|
| 701 |
+
},
|
| 702 |
+
{
|
| 703 |
+
"epoch": 7.0,
|
| 704 |
+
"grad_norm": 0.5881220102310181,
|
| 705 |
+
"learning_rate": 4.211333667247125e-07,
|
| 706 |
+
"loss": 0.6102,
|
| 707 |
+
"step": 350
|
| 708 |
+
},
|
| 709 |
+
{
|
| 710 |
+
"epoch": 7.0,
|
| 711 |
+
"eval_loss": 0.5426855683326721,
|
| 712 |
+
"eval_runtime": 52.2072,
|
| 713 |
+
"eval_samples_per_second": 3.831,
|
| 714 |
+
"eval_steps_per_second": 0.479,
|
| 715 |
+
"step": 350
|
| 716 |
+
},
|
| 717 |
+
{
|
| 718 |
+
"epoch": 7.1,
|
| 719 |
+
"grad_norm": 0.5317939519882202,
|
| 720 |
+
"learning_rate": 3.420445597436056e-07,
|
| 721 |
+
"loss": 0.6632,
|
| 722 |
+
"step": 355
|
| 723 |
+
},
|
| 724 |
+
{
|
| 725 |
+
"epoch": 7.2,
|
| 726 |
+
"grad_norm": 0.5702535510063171,
|
| 727 |
+
"learning_rate": 2.7091379149682683e-07,
|
| 728 |
+
"loss": 0.5992,
|
| 729 |
+
"step": 360
|
| 730 |
+
},
|
| 731 |
+
{
|
| 732 |
+
"epoch": 7.3,
|
| 733 |
+
"grad_norm": 0.6872391104698181,
|
| 734 |
+
"learning_rate": 2.0786258770873647e-07,
|
| 735 |
+
"loss": 0.6422,
|
| 736 |
+
"step": 365
|
| 737 |
+
},
|
| 738 |
+
{
|
| 739 |
+
"epoch": 7.4,
|
| 740 |
+
"grad_norm": 0.32829490303993225,
|
| 741 |
+
"learning_rate": 1.5299867030334815e-07,
|
| 742 |
+
"loss": 0.6811,
|
| 743 |
+
"step": 370
|
| 744 |
+
},
|
| 745 |
+
{
|
| 746 |
+
"epoch": 7.5,
|
| 747 |
+
"grad_norm": 0.5375828742980957,
|
| 748 |
+
"learning_rate": 1.0641577336322761e-07,
|
| 749 |
+
"loss": 0.8423,
|
| 750 |
+
"step": 375
|
| 751 |
+
},
|
| 752 |
+
{
|
| 753 |
+
"epoch": 7.6,
|
| 754 |
+
"grad_norm": 0.6306584477424622,
|
| 755 |
+
"learning_rate": 6.819348298638839e-08,
|
| 756 |
+
"loss": 0.5899,
|
| 757 |
+
"step": 380
|
| 758 |
+
},
|
| 759 |
+
{
|
| 760 |
+
"epoch": 7.7,
|
| 761 |
+
"grad_norm": 0.44418570399284363,
|
| 762 |
+
"learning_rate": 3.839710131477492e-08,
|
| 763 |
+
"loss": 0.6571,
|
| 764 |
+
"step": 385
|
| 765 |
+
},
|
| 766 |
+
{
|
| 767 |
+
"epoch": 7.8,
|
| 768 |
+
"grad_norm": 0.49700650572776794,
|
| 769 |
+
"learning_rate": 1.7077534966650767e-08,
|
| 770 |
+
"loss": 0.6561,
|
| 771 |
+
"step": 390
|
| 772 |
+
},
|
| 773 |
+
{
|
| 774 |
+
"epoch": 7.9,
|
| 775 |
+
"grad_norm": 0.3311610519886017,
|
| 776 |
+
"learning_rate": 4.2712080634949024e-09,
|
| 777 |
+
"loss": 0.6226,
|
| 778 |
+
"step": 395
|
| 779 |
+
},
|
| 780 |
+
{
|
| 781 |
+
"epoch": 8.0,
|
| 782 |
+
"grad_norm": 1.5899903774261475,
|
| 783 |
+
"learning_rate": 0.0,
|
| 784 |
+
"loss": 0.6762,
|
| 785 |
+
"step": 400
|
| 786 |
+
},
|
| 787 |
+
{
|
| 788 |
+
"epoch": 8.0,
|
| 789 |
+
"eval_loss": 0.5424160957336426,
|
| 790 |
+
"eval_runtime": 52.193,
|
| 791 |
+
"eval_samples_per_second": 3.832,
|
| 792 |
+
"eval_steps_per_second": 0.479,
|
| 793 |
+
"step": 400
|
| 794 |
}
|
| 795 |
],
|
| 796 |
"logging_steps": 5,
|
| 797 |
+
"max_steps": 400,
|
| 798 |
"num_input_tokens_seen": 0,
|
| 799 |
+
"num_train_epochs": 8,
|
| 800 |
"save_steps": 500,
|
| 801 |
"stateful_callbacks": {
|
| 802 |
"TrainerControl": {
|
|
|
|
| 810 |
"attributes": {}
|
| 811 |
}
|
| 812 |
},
|
| 813 |
+
"total_flos": 1.4231605134807245e+17,
|
| 814 |
"train_batch_size": 2,
|
| 815 |
"trial_name": null,
|
| 816 |
"trial_params": null
|
training_args.bin
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 5624
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:4d7ef1ca84158a115fb2ab949b3f781c814c5ef428f591fc8d6d01108daabb83
|
| 3 |
size 5624
|