Training in progress, step 1700, 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 131146352
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:19c29d19ae137d8d9f6d075009b38f48ede5ee69b99c1d40e24e93602a5e42c2
|
| 3 |
size 131146352
|
last-checkpoint/optimizer.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 67210516
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:53f02bf536b4b6c5ec995aa54e017724d8fb20c583e62cdd2047ff066ff5e86d
|
| 3 |
size 67210516
|
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:a19b20244c2b147c4db6331bfcd526d49838499d3bb5d4e14d3a604b49d4cdc4
|
| 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:c12a87648583ab623d82ad394450eed417016fd7c996ca538887d063a9458228
|
| 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": 0.
|
| 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,
|
|
@@ -11557,6 +11557,356 @@
|
|
| 11557 |
"learning_rate": 9.999999833078691e-05,
|
| 11558 |
"loss": 3.3308,
|
| 11559 |
"step": 1650
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 11560 |
}
|
| 11561 |
],
|
| 11562 |
"logging_steps": 1,
|
|
@@ -11576,7 +11926,7 @@
|
|
| 11576 |
"attributes": {}
|
| 11577 |
}
|
| 11578 |
},
|
| 11579 |
-
"total_flos": 2.
|
| 11580 |
"train_batch_size": 4,
|
| 11581 |
"trial_name": null,
|
| 11582 |
"trial_params": null
|
|
|
|
| 1 |
{
|
| 2 |
"best_metric": null,
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 0.04566455356183518,
|
| 5 |
"eval_steps": 500,
|
| 6 |
+
"global_step": 1700,
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
|
|
| 11557 |
"learning_rate": 9.999999833078691e-05,
|
| 11558 |
"loss": 3.3308,
|
| 11559 |
"step": 1650
|
| 11560 |
+
},
|
| 11561 |
+
{
|
| 11562 |
+
"epoch": 0.044348339959170514,
|
| 11563 |
+
"grad_norm": 1.4079015254974365,
|
| 11564 |
+
"learning_rate": 9.999999832875686e-05,
|
| 11565 |
+
"loss": 2.9442,
|
| 11566 |
+
"step": 1651
|
| 11567 |
+
},
|
| 11568 |
+
{
|
| 11569 |
+
"epoch": 0.044375201461265716,
|
| 11570 |
+
"grad_norm": 1.4412444829940796,
|
| 11571 |
+
"learning_rate": 9.999999832672557e-05,
|
| 11572 |
+
"loss": 3.038,
|
| 11573 |
+
"step": 1652
|
| 11574 |
+
},
|
| 11575 |
+
{
|
| 11576 |
+
"epoch": 0.04440206296336091,
|
| 11577 |
+
"grad_norm": 1.4149682521820068,
|
| 11578 |
+
"learning_rate": 9.999999832469304e-05,
|
| 11579 |
+
"loss": 2.7261,
|
| 11580 |
+
"step": 1653
|
| 11581 |
+
},
|
| 11582 |
+
{
|
| 11583 |
+
"epoch": 0.044428924465456106,
|
| 11584 |
+
"grad_norm": 1.411676287651062,
|
| 11585 |
+
"learning_rate": 9.99999983226593e-05,
|
| 11586 |
+
"loss": 2.9101,
|
| 11587 |
+
"step": 1654
|
| 11588 |
+
},
|
| 11589 |
+
{
|
| 11590 |
+
"epoch": 0.04445578596755131,
|
| 11591 |
+
"grad_norm": 1.470566987991333,
|
| 11592 |
+
"learning_rate": 9.99999983206243e-05,
|
| 11593 |
+
"loss": 2.77,
|
| 11594 |
+
"step": 1655
|
| 11595 |
+
},
|
| 11596 |
+
{
|
| 11597 |
+
"epoch": 0.0444826474696465,
|
| 11598 |
+
"grad_norm": 1.397863507270813,
|
| 11599 |
+
"learning_rate": 9.999999831858808e-05,
|
| 11600 |
+
"loss": 2.9582,
|
| 11601 |
+
"step": 1656
|
| 11602 |
+
},
|
| 11603 |
+
{
|
| 11604 |
+
"epoch": 0.0445095089717417,
|
| 11605 |
+
"grad_norm": 1.3953698873519897,
|
| 11606 |
+
"learning_rate": 9.999999831655063e-05,
|
| 11607 |
+
"loss": 3.0912,
|
| 11608 |
+
"step": 1657
|
| 11609 |
+
},
|
| 11610 |
+
{
|
| 11611 |
+
"epoch": 0.0445363704738369,
|
| 11612 |
+
"grad_norm": 1.3832037448883057,
|
| 11613 |
+
"learning_rate": 9.999999831451192e-05,
|
| 11614 |
+
"loss": 2.8913,
|
| 11615 |
+
"step": 1658
|
| 11616 |
+
},
|
| 11617 |
+
{
|
| 11618 |
+
"epoch": 0.044563231975932094,
|
| 11619 |
+
"grad_norm": 1.3912192583084106,
|
| 11620 |
+
"learning_rate": 9.9999998312472e-05,
|
| 11621 |
+
"loss": 2.6594,
|
| 11622 |
+
"step": 1659
|
| 11623 |
+
},
|
| 11624 |
+
{
|
| 11625 |
+
"epoch": 0.04459009347802729,
|
| 11626 |
+
"grad_norm": 1.3151278495788574,
|
| 11627 |
+
"learning_rate": 9.999999831043084e-05,
|
| 11628 |
+
"loss": 2.6635,
|
| 11629 |
+
"step": 1660
|
| 11630 |
+
},
|
| 11631 |
+
{
|
| 11632 |
+
"epoch": 0.04461695498012249,
|
| 11633 |
+
"grad_norm": 1.2868192195892334,
|
| 11634 |
+
"learning_rate": 9.999999830838846e-05,
|
| 11635 |
+
"loss": 2.7065,
|
| 11636 |
+
"step": 1661
|
| 11637 |
+
},
|
| 11638 |
+
{
|
| 11639 |
+
"epoch": 0.044643816482217685,
|
| 11640 |
+
"grad_norm": 1.387132167816162,
|
| 11641 |
+
"learning_rate": 9.999999830634483e-05,
|
| 11642 |
+
"loss": 2.6766,
|
| 11643 |
+
"step": 1662
|
| 11644 |
+
},
|
| 11645 |
+
{
|
| 11646 |
+
"epoch": 0.04467067798431288,
|
| 11647 |
+
"grad_norm": 1.4079433679580688,
|
| 11648 |
+
"learning_rate": 9.999999830429997e-05,
|
| 11649 |
+
"loss": 2.8444,
|
| 11650 |
+
"step": 1663
|
| 11651 |
+
},
|
| 11652 |
+
{
|
| 11653 |
+
"epoch": 0.04469753948640808,
|
| 11654 |
+
"grad_norm": 1.4303300380706787,
|
| 11655 |
+
"learning_rate": 9.999999830225387e-05,
|
| 11656 |
+
"loss": 2.9663,
|
| 11657 |
+
"step": 1664
|
| 11658 |
+
},
|
| 11659 |
+
{
|
| 11660 |
+
"epoch": 0.04472440098850328,
|
| 11661 |
+
"grad_norm": 1.39901864528656,
|
| 11662 |
+
"learning_rate": 9.999999830020654e-05,
|
| 11663 |
+
"loss": 2.9923,
|
| 11664 |
+
"step": 1665
|
| 11665 |
+
},
|
| 11666 |
+
{
|
| 11667 |
+
"epoch": 0.04475126249059847,
|
| 11668 |
+
"grad_norm": 1.4272572994232178,
|
| 11669 |
+
"learning_rate": 9.999999829815798e-05,
|
| 11670 |
+
"loss": 2.899,
|
| 11671 |
+
"step": 1666
|
| 11672 |
+
},
|
| 11673 |
+
{
|
| 11674 |
+
"epoch": 0.04477812399269367,
|
| 11675 |
+
"grad_norm": 1.492218017578125,
|
| 11676 |
+
"learning_rate": 9.99999982961082e-05,
|
| 11677 |
+
"loss": 3.0391,
|
| 11678 |
+
"step": 1667
|
| 11679 |
+
},
|
| 11680 |
+
{
|
| 11681 |
+
"epoch": 0.04480498549478887,
|
| 11682 |
+
"grad_norm": 1.450843095779419,
|
| 11683 |
+
"learning_rate": 9.999999829405716e-05,
|
| 11684 |
+
"loss": 2.9127,
|
| 11685 |
+
"step": 1668
|
| 11686 |
+
},
|
| 11687 |
+
{
|
| 11688 |
+
"epoch": 0.04483184699688406,
|
| 11689 |
+
"grad_norm": 1.3440508842468262,
|
| 11690 |
+
"learning_rate": 9.99999982920049e-05,
|
| 11691 |
+
"loss": 3.0001,
|
| 11692 |
+
"step": 1669
|
| 11693 |
+
},
|
| 11694 |
+
{
|
| 11695 |
+
"epoch": 0.044858708498979265,
|
| 11696 |
+
"grad_norm": 1.4402127265930176,
|
| 11697 |
+
"learning_rate": 9.999999828995141e-05,
|
| 11698 |
+
"loss": 2.819,
|
| 11699 |
+
"step": 1670
|
| 11700 |
+
},
|
| 11701 |
+
{
|
| 11702 |
+
"epoch": 0.04488557000107446,
|
| 11703 |
+
"grad_norm": 1.4429413080215454,
|
| 11704 |
+
"learning_rate": 9.999999828789667e-05,
|
| 11705 |
+
"loss": 3.0246,
|
| 11706 |
+
"step": 1671
|
| 11707 |
+
},
|
| 11708 |
+
{
|
| 11709 |
+
"epoch": 0.044912431503169654,
|
| 11710 |
+
"grad_norm": 2.6085400581359863,
|
| 11711 |
+
"learning_rate": 9.999999828584071e-05,
|
| 11712 |
+
"loss": 2.9273,
|
| 11713 |
+
"step": 1672
|
| 11714 |
+
},
|
| 11715 |
+
{
|
| 11716 |
+
"epoch": 0.044939293005264856,
|
| 11717 |
+
"grad_norm": 1.5993056297302246,
|
| 11718 |
+
"learning_rate": 9.999999828378352e-05,
|
| 11719 |
+
"loss": 3.2987,
|
| 11720 |
+
"step": 1673
|
| 11721 |
+
},
|
| 11722 |
+
{
|
| 11723 |
+
"epoch": 0.04496615450736005,
|
| 11724 |
+
"grad_norm": 1.5255417823791504,
|
| 11725 |
+
"learning_rate": 9.999999828172508e-05,
|
| 11726 |
+
"loss": 2.9552,
|
| 11727 |
+
"step": 1674
|
| 11728 |
+
},
|
| 11729 |
+
{
|
| 11730 |
+
"epoch": 0.044993016009455246,
|
| 11731 |
+
"grad_norm": 1.557896614074707,
|
| 11732 |
+
"learning_rate": 9.999999827966542e-05,
|
| 11733 |
+
"loss": 3.123,
|
| 11734 |
+
"step": 1675
|
| 11735 |
+
},
|
| 11736 |
+
{
|
| 11737 |
+
"epoch": 0.04501987751155045,
|
| 11738 |
+
"grad_norm": 1.558122992515564,
|
| 11739 |
+
"learning_rate": 9.999999827760452e-05,
|
| 11740 |
+
"loss": 2.9187,
|
| 11741 |
+
"step": 1676
|
| 11742 |
+
},
|
| 11743 |
+
{
|
| 11744 |
+
"epoch": 0.04504673901364564,
|
| 11745 |
+
"grad_norm": 1.425349473953247,
|
| 11746 |
+
"learning_rate": 9.999999827554239e-05,
|
| 11747 |
+
"loss": 2.899,
|
| 11748 |
+
"step": 1677
|
| 11749 |
+
},
|
| 11750 |
+
{
|
| 11751 |
+
"epoch": 0.04507360051574084,
|
| 11752 |
+
"grad_norm": 1.5926076173782349,
|
| 11753 |
+
"learning_rate": 9.999999827347903e-05,
|
| 11754 |
+
"loss": 3.2055,
|
| 11755 |
+
"step": 1678
|
| 11756 |
+
},
|
| 11757 |
+
{
|
| 11758 |
+
"epoch": 0.04510046201783604,
|
| 11759 |
+
"grad_norm": 1.6083916425704956,
|
| 11760 |
+
"learning_rate": 9.999999827141443e-05,
|
| 11761 |
+
"loss": 2.8889,
|
| 11762 |
+
"step": 1679
|
| 11763 |
+
},
|
| 11764 |
+
{
|
| 11765 |
+
"epoch": 0.045127323519931234,
|
| 11766 |
+
"grad_norm": 1.4613057374954224,
|
| 11767 |
+
"learning_rate": 9.999999826934859e-05,
|
| 11768 |
+
"loss": 2.968,
|
| 11769 |
+
"step": 1680
|
| 11770 |
+
},
|
| 11771 |
+
{
|
| 11772 |
+
"epoch": 0.04515418502202643,
|
| 11773 |
+
"grad_norm": 1.591672420501709,
|
| 11774 |
+
"learning_rate": 9.999999826728153e-05,
|
| 11775 |
+
"loss": 3.0252,
|
| 11776 |
+
"step": 1681
|
| 11777 |
+
},
|
| 11778 |
+
{
|
| 11779 |
+
"epoch": 0.04518104652412163,
|
| 11780 |
+
"grad_norm": 1.539560317993164,
|
| 11781 |
+
"learning_rate": 9.999999826521322e-05,
|
| 11782 |
+
"loss": 2.8637,
|
| 11783 |
+
"step": 1682
|
| 11784 |
+
},
|
| 11785 |
+
{
|
| 11786 |
+
"epoch": 0.045207908026216825,
|
| 11787 |
+
"grad_norm": 1.4949159622192383,
|
| 11788 |
+
"learning_rate": 9.99999982631437e-05,
|
| 11789 |
+
"loss": 3.05,
|
| 11790 |
+
"step": 1683
|
| 11791 |
+
},
|
| 11792 |
+
{
|
| 11793 |
+
"epoch": 0.04523476952831202,
|
| 11794 |
+
"grad_norm": 1.5287278890609741,
|
| 11795 |
+
"learning_rate": 9.999999826107294e-05,
|
| 11796 |
+
"loss": 3.0176,
|
| 11797 |
+
"step": 1684
|
| 11798 |
+
},
|
| 11799 |
+
{
|
| 11800 |
+
"epoch": 0.04526163103040722,
|
| 11801 |
+
"grad_norm": 1.4513976573944092,
|
| 11802 |
+
"learning_rate": 9.999999825900092e-05,
|
| 11803 |
+
"loss": 3.0043,
|
| 11804 |
+
"step": 1685
|
| 11805 |
+
},
|
| 11806 |
+
{
|
| 11807 |
+
"epoch": 0.04528849253250242,
|
| 11808 |
+
"grad_norm": 1.4938799142837524,
|
| 11809 |
+
"learning_rate": 9.99999982569277e-05,
|
| 11810 |
+
"loss": 2.998,
|
| 11811 |
+
"step": 1686
|
| 11812 |
+
},
|
| 11813 |
+
{
|
| 11814 |
+
"epoch": 0.04531535403459761,
|
| 11815 |
+
"grad_norm": 1.6789308786392212,
|
| 11816 |
+
"learning_rate": 9.999999825485323e-05,
|
| 11817 |
+
"loss": 3.108,
|
| 11818 |
+
"step": 1687
|
| 11819 |
+
},
|
| 11820 |
+
{
|
| 11821 |
+
"epoch": 0.04534221553669281,
|
| 11822 |
+
"grad_norm": 1.4213138818740845,
|
| 11823 |
+
"learning_rate": 9.999999825277752e-05,
|
| 11824 |
+
"loss": 3.0397,
|
| 11825 |
+
"step": 1688
|
| 11826 |
+
},
|
| 11827 |
+
{
|
| 11828 |
+
"epoch": 0.04536907703878801,
|
| 11829 |
+
"grad_norm": 1.5696649551391602,
|
| 11830 |
+
"learning_rate": 9.999999825070058e-05,
|
| 11831 |
+
"loss": 3.1145,
|
| 11832 |
+
"step": 1689
|
| 11833 |
+
},
|
| 11834 |
+
{
|
| 11835 |
+
"epoch": 0.0453959385408832,
|
| 11836 |
+
"grad_norm": 1.6566909551620483,
|
| 11837 |
+
"learning_rate": 9.999999824862241e-05,
|
| 11838 |
+
"loss": 3.3197,
|
| 11839 |
+
"step": 1690
|
| 11840 |
+
},
|
| 11841 |
+
{
|
| 11842 |
+
"epoch": 0.045422800042978405,
|
| 11843 |
+
"grad_norm": 1.591908574104309,
|
| 11844 |
+
"learning_rate": 9.999999824654302e-05,
|
| 11845 |
+
"loss": 2.9155,
|
| 11846 |
+
"step": 1691
|
| 11847 |
+
},
|
| 11848 |
+
{
|
| 11849 |
+
"epoch": 0.0454496615450736,
|
| 11850 |
+
"grad_norm": 1.6618692874908447,
|
| 11851 |
+
"learning_rate": 9.999999824446238e-05,
|
| 11852 |
+
"loss": 3.3152,
|
| 11853 |
+
"step": 1692
|
| 11854 |
+
},
|
| 11855 |
+
{
|
| 11856 |
+
"epoch": 0.045476523047168795,
|
| 11857 |
+
"grad_norm": 1.660921573638916,
|
| 11858 |
+
"learning_rate": 9.999999824238051e-05,
|
| 11859 |
+
"loss": 3.1987,
|
| 11860 |
+
"step": 1693
|
| 11861 |
+
},
|
| 11862 |
+
{
|
| 11863 |
+
"epoch": 0.045503384549263996,
|
| 11864 |
+
"grad_norm": 1.680649757385254,
|
| 11865 |
+
"learning_rate": 9.99999982402974e-05,
|
| 11866 |
+
"loss": 3.2139,
|
| 11867 |
+
"step": 1694
|
| 11868 |
+
},
|
| 11869 |
+
{
|
| 11870 |
+
"epoch": 0.04553024605135919,
|
| 11871 |
+
"grad_norm": 1.7884151935577393,
|
| 11872 |
+
"learning_rate": 9.999999823821307e-05,
|
| 11873 |
+
"loss": 3.2991,
|
| 11874 |
+
"step": 1695
|
| 11875 |
+
},
|
| 11876 |
+
{
|
| 11877 |
+
"epoch": 0.045557107553454386,
|
| 11878 |
+
"grad_norm": 1.6756354570388794,
|
| 11879 |
+
"learning_rate": 9.99999982361275e-05,
|
| 11880 |
+
"loss": 3.1816,
|
| 11881 |
+
"step": 1696
|
| 11882 |
+
},
|
| 11883 |
+
{
|
| 11884 |
+
"epoch": 0.04558396905554959,
|
| 11885 |
+
"grad_norm": 1.6229280233383179,
|
| 11886 |
+
"learning_rate": 9.999999823404069e-05,
|
| 11887 |
+
"loss": 3.2563,
|
| 11888 |
+
"step": 1697
|
| 11889 |
+
},
|
| 11890 |
+
{
|
| 11891 |
+
"epoch": 0.04561083055764478,
|
| 11892 |
+
"grad_norm": 1.625819444656372,
|
| 11893 |
+
"learning_rate": 9.999999823195265e-05,
|
| 11894 |
+
"loss": 3.1226,
|
| 11895 |
+
"step": 1698
|
| 11896 |
+
},
|
| 11897 |
+
{
|
| 11898 |
+
"epoch": 0.04563769205973998,
|
| 11899 |
+
"grad_norm": 1.7141685485839844,
|
| 11900 |
+
"learning_rate": 9.999999822986337e-05,
|
| 11901 |
+
"loss": 3.3355,
|
| 11902 |
+
"step": 1699
|
| 11903 |
+
},
|
| 11904 |
+
{
|
| 11905 |
+
"epoch": 0.04566455356183518,
|
| 11906 |
+
"grad_norm": 1.7747595310211182,
|
| 11907 |
+
"learning_rate": 9.999999822777286e-05,
|
| 11908 |
+
"loss": 3.2331,
|
| 11909 |
+
"step": 1700
|
| 11910 |
}
|
| 11911 |
],
|
| 11912 |
"logging_steps": 1,
|
|
|
|
| 11926 |
"attributes": {}
|
| 11927 |
}
|
| 11928 |
},
|
| 11929 |
+
"total_flos": 2.320207374503117e+18,
|
| 11930 |
"train_batch_size": 4,
|
| 11931 |
"trial_name": null,
|
| 11932 |
"trial_params": null
|