Training in progress, step 600, 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 159967880
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:61fa39fe72d3b07877b75fe39479b59104182f209a2615a3e426dd500c69d610
|
3 |
size 159967880
|
last-checkpoint/optimizer.pt
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 81730644
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:b9b75154a8ac4a6cfadd72600ba7b4722dee87df5c3d0af9be47b403fa2dcf82
|
3 |
size 81730644
|
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 sha256:1d6b5d9c6f0c15be87f5391bf4c24c5abc24693b4a4ae1f1e316a59c22584e8b
|
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 sha256:3fee25a74f8614e14501bbc1e4dbf79c416bb01f3b851bb0f5f52f090982cd43
|
3 |
size 1064
|
last-checkpoint/trainer_state.json
CHANGED
@@ -1,9 +1,9 @@
|
|
1 |
{
|
2 |
-
"best_metric": 0.
|
3 |
-
"best_model_checkpoint": "miner_id_24/checkpoint-
|
4 |
-
"epoch": 0.
|
5 |
"eval_steps": 100,
|
6 |
-
"global_step":
|
7 |
"is_hyper_param_search": false,
|
8 |
"is_local_process_zero": true,
|
9 |
"is_world_process_zero": true,
|
@@ -3555,6 +3555,714 @@
|
|
3555 |
"eval_samples_per_second": 7.14,
|
3556 |
"eval_steps_per_second": 1.785,
|
3557 |
"step": 500
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
3558 |
}
|
3559 |
],
|
3560 |
"logging_steps": 1,
|
@@ -3583,7 +4291,7 @@
|
|
3583 |
"attributes": {}
|
3584 |
}
|
3585 |
},
|
3586 |
-
"total_flos": 1.
|
3587 |
"train_batch_size": 4,
|
3588 |
"trial_name": null,
|
3589 |
"trial_params": null
|
|
|
1 |
{
|
2 |
+
"best_metric": 0.7635987401008606,
|
3 |
+
"best_model_checkpoint": "miner_id_24/checkpoint-600",
|
4 |
+
"epoch": 0.041790005223750655,
|
5 |
"eval_steps": 100,
|
6 |
+
"global_step": 600,
|
7 |
"is_hyper_param_search": false,
|
8 |
"is_local_process_zero": true,
|
9 |
"is_world_process_zero": true,
|
|
|
3555 |
"eval_samples_per_second": 7.14,
|
3556 |
"eval_steps_per_second": 1.785,
|
3557 |
"step": 500
|
3558 |
+
},
|
3559 |
+
{
|
3560 |
+
"epoch": 0.0348946543618318,
|
3561 |
+
"grad_norm": 0.5795607566833496,
|
3562 |
+
"learning_rate": 9.572925909767412e-05,
|
3563 |
+
"loss": 0.4495,
|
3564 |
+
"step": 501
|
3565 |
+
},
|
3566 |
+
{
|
3567 |
+
"epoch": 0.034964304370538046,
|
3568 |
+
"grad_norm": 0.6874101161956787,
|
3569 |
+
"learning_rate": 9.540096409419296e-05,
|
3570 |
+
"loss": 0.8444,
|
3571 |
+
"step": 502
|
3572 |
+
},
|
3573 |
+
{
|
3574 |
+
"epoch": 0.035033954379244295,
|
3575 |
+
"grad_norm": 0.5595911145210266,
|
3576 |
+
"learning_rate": 9.507271875570381e-05,
|
3577 |
+
"loss": 0.9391,
|
3578 |
+
"step": 503
|
3579 |
+
},
|
3580 |
+
{
|
3581 |
+
"epoch": 0.03510360438795055,
|
3582 |
+
"grad_norm": 0.525644063949585,
|
3583 |
+
"learning_rate": 9.474452662692838e-05,
|
3584 |
+
"loss": 0.7833,
|
3585 |
+
"step": 504
|
3586 |
+
},
|
3587 |
+
{
|
3588 |
+
"epoch": 0.0351732543966568,
|
3589 |
+
"grad_norm": 0.6366891264915466,
|
3590 |
+
"learning_rate": 9.441639125201368e-05,
|
3591 |
+
"loss": 1.0472,
|
3592 |
+
"step": 505
|
3593 |
+
},
|
3594 |
+
{
|
3595 |
+
"epoch": 0.035242904405363054,
|
3596 |
+
"grad_norm": 0.8487269878387451,
|
3597 |
+
"learning_rate": 9.408831617449385e-05,
|
3598 |
+
"loss": 1.0513,
|
3599 |
+
"step": 506
|
3600 |
+
},
|
3601 |
+
{
|
3602 |
+
"epoch": 0.0353125544140693,
|
3603 |
+
"grad_norm": 0.7027648091316223,
|
3604 |
+
"learning_rate": 9.376030493725189e-05,
|
3605 |
+
"loss": 0.9505,
|
3606 |
+
"step": 507
|
3607 |
+
},
|
3608 |
+
{
|
3609 |
+
"epoch": 0.03538220442277555,
|
3610 |
+
"grad_norm": 0.6772575974464417,
|
3611 |
+
"learning_rate": 9.343236108248139e-05,
|
3612 |
+
"loss": 1.0417,
|
3613 |
+
"step": 508
|
3614 |
+
},
|
3615 |
+
{
|
3616 |
+
"epoch": 0.03545185443148181,
|
3617 |
+
"grad_norm": 0.5657368898391724,
|
3618 |
+
"learning_rate": 9.310448815164826e-05,
|
3619 |
+
"loss": 0.9236,
|
3620 |
+
"step": 509
|
3621 |
+
},
|
3622 |
+
{
|
3623 |
+
"epoch": 0.035521504440188055,
|
3624 |
+
"grad_norm": 0.64215087890625,
|
3625 |
+
"learning_rate": 9.277668968545253e-05,
|
3626 |
+
"loss": 1.0035,
|
3627 |
+
"step": 510
|
3628 |
+
},
|
3629 |
+
{
|
3630 |
+
"epoch": 0.035591154448894304,
|
3631 |
+
"grad_norm": 0.6276829242706299,
|
3632 |
+
"learning_rate": 9.244896922379007e-05,
|
3633 |
+
"loss": 0.8375,
|
3634 |
+
"step": 511
|
3635 |
+
},
|
3636 |
+
{
|
3637 |
+
"epoch": 0.03566080445760056,
|
3638 |
+
"grad_norm": 0.5804170966148376,
|
3639 |
+
"learning_rate": 9.212133030571437e-05,
|
3640 |
+
"loss": 0.4934,
|
3641 |
+
"step": 512
|
3642 |
+
},
|
3643 |
+
{
|
3644 |
+
"epoch": 0.03573045446630681,
|
3645 |
+
"grad_norm": 0.7230868935585022,
|
3646 |
+
"learning_rate": 9.17937764693983e-05,
|
3647 |
+
"loss": 0.9427,
|
3648 |
+
"step": 513
|
3649 |
+
},
|
3650 |
+
{
|
3651 |
+
"epoch": 0.035800104475013056,
|
3652 |
+
"grad_norm": 0.6632394194602966,
|
3653 |
+
"learning_rate": 9.146631125209607e-05,
|
3654 |
+
"loss": 0.4176,
|
3655 |
+
"step": 514
|
3656 |
+
},
|
3657 |
+
{
|
3658 |
+
"epoch": 0.03586975448371931,
|
3659 |
+
"grad_norm": 0.5885234475135803,
|
3660 |
+
"learning_rate": 9.113893819010475e-05,
|
3661 |
+
"loss": 0.6042,
|
3662 |
+
"step": 515
|
3663 |
+
},
|
3664 |
+
{
|
3665 |
+
"epoch": 0.03593940449242556,
|
3666 |
+
"grad_norm": 0.5666863322257996,
|
3667 |
+
"learning_rate": 9.081166081872626e-05,
|
3668 |
+
"loss": 1.5152,
|
3669 |
+
"step": 516
|
3670 |
+
},
|
3671 |
+
{
|
3672 |
+
"epoch": 0.036009054501131815,
|
3673 |
+
"grad_norm": 0.7007538676261902,
|
3674 |
+
"learning_rate": 9.048448267222918e-05,
|
3675 |
+
"loss": 0.9444,
|
3676 |
+
"step": 517
|
3677 |
+
},
|
3678 |
+
{
|
3679 |
+
"epoch": 0.036078704509838064,
|
3680 |
+
"grad_norm": 0.6212923526763916,
|
3681 |
+
"learning_rate": 9.015740728381054e-05,
|
3682 |
+
"loss": 0.634,
|
3683 |
+
"step": 518
|
3684 |
+
},
|
3685 |
+
{
|
3686 |
+
"epoch": 0.03614835451854431,
|
3687 |
+
"grad_norm": 0.6189596056938171,
|
3688 |
+
"learning_rate": 8.98304381855577e-05,
|
3689 |
+
"loss": 1.1091,
|
3690 |
+
"step": 519
|
3691 |
+
},
|
3692 |
+
{
|
3693 |
+
"epoch": 0.03621800452725057,
|
3694 |
+
"grad_norm": 0.6159670948982239,
|
3695 |
+
"learning_rate": 8.95035789084102e-05,
|
3696 |
+
"loss": 0.787,
|
3697 |
+
"step": 520
|
3698 |
+
},
|
3699 |
+
{
|
3700 |
+
"epoch": 0.036287654535956816,
|
3701 |
+
"grad_norm": 0.6371515989303589,
|
3702 |
+
"learning_rate": 8.917683298212158e-05,
|
3703 |
+
"loss": 0.6172,
|
3704 |
+
"step": 521
|
3705 |
+
},
|
3706 |
+
{
|
3707 |
+
"epoch": 0.036357304544663065,
|
3708 |
+
"grad_norm": 0.6314066052436829,
|
3709 |
+
"learning_rate": 8.885020393522135e-05,
|
3710 |
+
"loss": 0.9702,
|
3711 |
+
"step": 522
|
3712 |
+
},
|
3713 |
+
{
|
3714 |
+
"epoch": 0.03642695455336932,
|
3715 |
+
"grad_norm": 0.6285626888275146,
|
3716 |
+
"learning_rate": 8.852369529497679e-05,
|
3717 |
+
"loss": 0.9819,
|
3718 |
+
"step": 523
|
3719 |
+
},
|
3720 |
+
{
|
3721 |
+
"epoch": 0.03649660456207557,
|
3722 |
+
"grad_norm": 0.5257949233055115,
|
3723 |
+
"learning_rate": 8.819731058735501e-05,
|
3724 |
+
"loss": 0.8288,
|
3725 |
+
"step": 524
|
3726 |
+
},
|
3727 |
+
{
|
3728 |
+
"epoch": 0.036566254570781824,
|
3729 |
+
"grad_norm": 0.611438512802124,
|
3730 |
+
"learning_rate": 8.787105333698465e-05,
|
3731 |
+
"loss": 0.9246,
|
3732 |
+
"step": 525
|
3733 |
+
},
|
3734 |
+
{
|
3735 |
+
"epoch": 0.03663590457948807,
|
3736 |
+
"grad_norm": 0.5995710492134094,
|
3737 |
+
"learning_rate": 8.754492706711798e-05,
|
3738 |
+
"loss": 0.6855,
|
3739 |
+
"step": 526
|
3740 |
+
},
|
3741 |
+
{
|
3742 |
+
"epoch": 0.03670555458819432,
|
3743 |
+
"grad_norm": 0.681425154209137,
|
3744 |
+
"learning_rate": 8.721893529959287e-05,
|
3745 |
+
"loss": 1.1644,
|
3746 |
+
"step": 527
|
3747 |
+
},
|
3748 |
+
{
|
3749 |
+
"epoch": 0.036775204596900576,
|
3750 |
+
"grad_norm": 0.7111718654632568,
|
3751 |
+
"learning_rate": 8.68930815547946e-05,
|
3752 |
+
"loss": 0.9181,
|
3753 |
+
"step": 528
|
3754 |
+
},
|
3755 |
+
{
|
3756 |
+
"epoch": 0.036844854605606825,
|
3757 |
+
"grad_norm": 0.5794047713279724,
|
3758 |
+
"learning_rate": 8.656736935161802e-05,
|
3759 |
+
"loss": 1.061,
|
3760 |
+
"step": 529
|
3761 |
+
},
|
3762 |
+
{
|
3763 |
+
"epoch": 0.03691450461431307,
|
3764 |
+
"grad_norm": 0.5971503257751465,
|
3765 |
+
"learning_rate": 8.624180220742946e-05,
|
3766 |
+
"loss": 0.5903,
|
3767 |
+
"step": 530
|
3768 |
+
},
|
3769 |
+
{
|
3770 |
+
"epoch": 0.03698415462301933,
|
3771 |
+
"grad_norm": 0.7091482281684875,
|
3772 |
+
"learning_rate": 8.59163836380287e-05,
|
3773 |
+
"loss": 0.8907,
|
3774 |
+
"step": 531
|
3775 |
+
},
|
3776 |
+
{
|
3777 |
+
"epoch": 0.03705380463172558,
|
3778 |
+
"grad_norm": 0.6185580492019653,
|
3779 |
+
"learning_rate": 8.559111715761114e-05,
|
3780 |
+
"loss": 0.8452,
|
3781 |
+
"step": 532
|
3782 |
+
},
|
3783 |
+
{
|
3784 |
+
"epoch": 0.03712345464043183,
|
3785 |
+
"grad_norm": 0.68827223777771,
|
3786 |
+
"learning_rate": 8.52660062787297e-05,
|
3787 |
+
"loss": 0.8711,
|
3788 |
+
"step": 533
|
3789 |
+
},
|
3790 |
+
{
|
3791 |
+
"epoch": 0.03719310464913808,
|
3792 |
+
"grad_norm": 0.6279632449150085,
|
3793 |
+
"learning_rate": 8.494105451225704e-05,
|
3794 |
+
"loss": 0.6453,
|
3795 |
+
"step": 534
|
3796 |
+
},
|
3797 |
+
{
|
3798 |
+
"epoch": 0.03726275465784433,
|
3799 |
+
"grad_norm": 0.7252237200737,
|
3800 |
+
"learning_rate": 8.461626536734753e-05,
|
3801 |
+
"loss": 1.1148,
|
3802 |
+
"step": 535
|
3803 |
+
},
|
3804 |
+
{
|
3805 |
+
"epoch": 0.037332404666550585,
|
3806 |
+
"grad_norm": 0.6377342939376831,
|
3807 |
+
"learning_rate": 8.429164235139931e-05,
|
3808 |
+
"loss": 1.0532,
|
3809 |
+
"step": 536
|
3810 |
+
},
|
3811 |
+
{
|
3812 |
+
"epoch": 0.037402054675256834,
|
3813 |
+
"grad_norm": 0.7409278154373169,
|
3814 |
+
"learning_rate": 8.396718897001663e-05,
|
3815 |
+
"loss": 1.0161,
|
3816 |
+
"step": 537
|
3817 |
+
},
|
3818 |
+
{
|
3819 |
+
"epoch": 0.03747170468396308,
|
3820 |
+
"grad_norm": 0.6048555970191956,
|
3821 |
+
"learning_rate": 8.364290872697173e-05,
|
3822 |
+
"loss": 1.012,
|
3823 |
+
"step": 538
|
3824 |
+
},
|
3825 |
+
{
|
3826 |
+
"epoch": 0.03754135469266934,
|
3827 |
+
"grad_norm": 0.7676815390586853,
|
3828 |
+
"learning_rate": 8.331880512416724e-05,
|
3829 |
+
"loss": 0.9402,
|
3830 |
+
"step": 539
|
3831 |
+
},
|
3832 |
+
{
|
3833 |
+
"epoch": 0.037611004701375586,
|
3834 |
+
"grad_norm": 0.6360906958580017,
|
3835 |
+
"learning_rate": 8.299488166159817e-05,
|
3836 |
+
"loss": 0.4591,
|
3837 |
+
"step": 540
|
3838 |
+
},
|
3839 |
+
{
|
3840 |
+
"epoch": 0.03768065471008184,
|
3841 |
+
"grad_norm": 0.6816183924674988,
|
3842 |
+
"learning_rate": 8.267114183731421e-05,
|
3843 |
+
"loss": 0.661,
|
3844 |
+
"step": 541
|
3845 |
+
},
|
3846 |
+
{
|
3847 |
+
"epoch": 0.03775030471878809,
|
3848 |
+
"grad_norm": 0.6955873966217041,
|
3849 |
+
"learning_rate": 8.234758914738199e-05,
|
3850 |
+
"loss": 0.8015,
|
3851 |
+
"step": 542
|
3852 |
+
},
|
3853 |
+
{
|
3854 |
+
"epoch": 0.03781995472749434,
|
3855 |
+
"grad_norm": 0.787493884563446,
|
3856 |
+
"learning_rate": 8.20242270858472e-05,
|
3857 |
+
"loss": 0.6941,
|
3858 |
+
"step": 543
|
3859 |
+
},
|
3860 |
+
{
|
3861 |
+
"epoch": 0.037889604736200594,
|
3862 |
+
"grad_norm": 0.5939062833786011,
|
3863 |
+
"learning_rate": 8.170105914469702e-05,
|
3864 |
+
"loss": 0.9034,
|
3865 |
+
"step": 544
|
3866 |
+
},
|
3867 |
+
{
|
3868 |
+
"epoch": 0.03795925474490684,
|
3869 |
+
"grad_norm": 0.5235042572021484,
|
3870 |
+
"learning_rate": 8.137808881382226e-05,
|
3871 |
+
"loss": 1.0283,
|
3872 |
+
"step": 545
|
3873 |
+
},
|
3874 |
+
{
|
3875 |
+
"epoch": 0.03802890475361309,
|
3876 |
+
"grad_norm": 0.7017082571983337,
|
3877 |
+
"learning_rate": 8.105531958097972e-05,
|
3878 |
+
"loss": 1.0407,
|
3879 |
+
"step": 546
|
3880 |
+
},
|
3881 |
+
{
|
3882 |
+
"epoch": 0.038098554762319346,
|
3883 |
+
"grad_norm": 0.7762130498886108,
|
3884 |
+
"learning_rate": 8.073275493175464e-05,
|
3885 |
+
"loss": 0.7814,
|
3886 |
+
"step": 547
|
3887 |
+
},
|
3888 |
+
{
|
3889 |
+
"epoch": 0.038168204771025595,
|
3890 |
+
"grad_norm": 0.588405191898346,
|
3891 |
+
"learning_rate": 8.041039834952287e-05,
|
3892 |
+
"loss": 0.8832,
|
3893 |
+
"step": 548
|
3894 |
+
},
|
3895 |
+
{
|
3896 |
+
"epoch": 0.03823785477973185,
|
3897 |
+
"grad_norm": 0.7792285084724426,
|
3898 |
+
"learning_rate": 8.008825331541335e-05,
|
3899 |
+
"loss": 1.051,
|
3900 |
+
"step": 549
|
3901 |
+
},
|
3902 |
+
{
|
3903 |
+
"epoch": 0.0383075047884381,
|
3904 |
+
"grad_norm": 0.6209467649459839,
|
3905 |
+
"learning_rate": 7.976632330827056e-05,
|
3906 |
+
"loss": 0.8802,
|
3907 |
+
"step": 550
|
3908 |
+
},
|
3909 |
+
{
|
3910 |
+
"epoch": 0.03837715479714435,
|
3911 |
+
"grad_norm": 0.5231680274009705,
|
3912 |
+
"learning_rate": 7.944461180461686e-05,
|
3913 |
+
"loss": 0.7529,
|
3914 |
+
"step": 551
|
3915 |
+
},
|
3916 |
+
{
|
3917 |
+
"epoch": 0.0384468048058506,
|
3918 |
+
"grad_norm": 0.6021607518196106,
|
3919 |
+
"learning_rate": 7.912312227861503e-05,
|
3920 |
+
"loss": 1.1235,
|
3921 |
+
"step": 552
|
3922 |
+
},
|
3923 |
+
{
|
3924 |
+
"epoch": 0.03851645481455685,
|
3925 |
+
"grad_norm": 0.5573668479919434,
|
3926 |
+
"learning_rate": 7.880185820203065e-05,
|
3927 |
+
"loss": 0.6753,
|
3928 |
+
"step": 553
|
3929 |
+
},
|
3930 |
+
{
|
3931 |
+
"epoch": 0.0385861048232631,
|
3932 |
+
"grad_norm": 0.5354910492897034,
|
3933 |
+
"learning_rate": 7.848082304419478e-05,
|
3934 |
+
"loss": 0.6843,
|
3935 |
+
"step": 554
|
3936 |
+
},
|
3937 |
+
{
|
3938 |
+
"epoch": 0.038655754831969355,
|
3939 |
+
"grad_norm": 0.606436014175415,
|
3940 |
+
"learning_rate": 7.816002027196627e-05,
|
3941 |
+
"loss": 1.0557,
|
3942 |
+
"step": 555
|
3943 |
+
},
|
3944 |
+
{
|
3945 |
+
"epoch": 0.038725404840675604,
|
3946 |
+
"grad_norm": 0.6580552458763123,
|
3947 |
+
"learning_rate": 7.783945334969451e-05,
|
3948 |
+
"loss": 0.6222,
|
3949 |
+
"step": 556
|
3950 |
+
},
|
3951 |
+
{
|
3952 |
+
"epoch": 0.03879505484938186,
|
3953 |
+
"grad_norm": 0.6174128651618958,
|
3954 |
+
"learning_rate": 7.751912573918193e-05,
|
3955 |
+
"loss": 0.8194,
|
3956 |
+
"step": 557
|
3957 |
+
},
|
3958 |
+
{
|
3959 |
+
"epoch": 0.03886470485808811,
|
3960 |
+
"grad_norm": 0.6724019646644592,
|
3961 |
+
"learning_rate": 7.719904089964658e-05,
|
3962 |
+
"loss": 1.0095,
|
3963 |
+
"step": 558
|
3964 |
+
},
|
3965 |
+
{
|
3966 |
+
"epoch": 0.038934354866794356,
|
3967 |
+
"grad_norm": 0.7200993299484253,
|
3968 |
+
"learning_rate": 7.687920228768493e-05,
|
3969 |
+
"loss": 0.8115,
|
3970 |
+
"step": 559
|
3971 |
+
},
|
3972 |
+
{
|
3973 |
+
"epoch": 0.03900400487550061,
|
3974 |
+
"grad_norm": 0.5682472586631775,
|
3975 |
+
"learning_rate": 7.655961335723433e-05,
|
3976 |
+
"loss": 0.7034,
|
3977 |
+
"step": 560
|
3978 |
+
},
|
3979 |
+
{
|
3980 |
+
"epoch": 0.03907365488420686,
|
3981 |
+
"grad_norm": 0.7236086130142212,
|
3982 |
+
"learning_rate": 7.624027755953592e-05,
|
3983 |
+
"loss": 0.9028,
|
3984 |
+
"step": 561
|
3985 |
+
},
|
3986 |
+
{
|
3987 |
+
"epoch": 0.03914330489291311,
|
3988 |
+
"grad_norm": 0.5866789221763611,
|
3989 |
+
"learning_rate": 7.592119834309715e-05,
|
3990 |
+
"loss": 0.8919,
|
3991 |
+
"step": 562
|
3992 |
+
},
|
3993 |
+
{
|
3994 |
+
"epoch": 0.039212954901619364,
|
3995 |
+
"grad_norm": 0.6271937489509583,
|
3996 |
+
"learning_rate": 7.560237915365472e-05,
|
3997 |
+
"loss": 0.6447,
|
3998 |
+
"step": 563
|
3999 |
+
},
|
4000 |
+
{
|
4001 |
+
"epoch": 0.03928260491032561,
|
4002 |
+
"grad_norm": 0.5319473147392273,
|
4003 |
+
"learning_rate": 7.528382343413734e-05,
|
4004 |
+
"loss": 1.0977,
|
4005 |
+
"step": 564
|
4006 |
+
},
|
4007 |
+
{
|
4008 |
+
"epoch": 0.03935225491903187,
|
4009 |
+
"grad_norm": 0.673537015914917,
|
4010 |
+
"learning_rate": 7.49655346246284e-05,
|
4011 |
+
"loss": 0.6669,
|
4012 |
+
"step": 565
|
4013 |
+
},
|
4014 |
+
{
|
4015 |
+
"epoch": 0.039421904927738116,
|
4016 |
+
"grad_norm": 0.7043957114219666,
|
4017 |
+
"learning_rate": 7.464751616232902e-05,
|
4018 |
+
"loss": 0.6334,
|
4019 |
+
"step": 566
|
4020 |
+
},
|
4021 |
+
{
|
4022 |
+
"epoch": 0.039491554936444365,
|
4023 |
+
"grad_norm": 0.6532731652259827,
|
4024 |
+
"learning_rate": 7.432977148152074e-05,
|
4025 |
+
"loss": 0.659,
|
4026 |
+
"step": 567
|
4027 |
+
},
|
4028 |
+
{
|
4029 |
+
"epoch": 0.03956120494515062,
|
4030 |
+
"grad_norm": 0.6882482767105103,
|
4031 |
+
"learning_rate": 7.401230401352866e-05,
|
4032 |
+
"loss": 0.711,
|
4033 |
+
"step": 568
|
4034 |
+
},
|
4035 |
+
{
|
4036 |
+
"epoch": 0.03963085495385687,
|
4037 |
+
"grad_norm": 0.7171745896339417,
|
4038 |
+
"learning_rate": 7.369511718668418e-05,
|
4039 |
+
"loss": 0.941,
|
4040 |
+
"step": 569
|
4041 |
+
},
|
4042 |
+
{
|
4043 |
+
"epoch": 0.03970050496256312,
|
4044 |
+
"grad_norm": 0.6474679708480835,
|
4045 |
+
"learning_rate": 7.337821442628805e-05,
|
4046 |
+
"loss": 0.8192,
|
4047 |
+
"step": 570
|
4048 |
+
},
|
4049 |
+
{
|
4050 |
+
"epoch": 0.03977015497126937,
|
4051 |
+
"grad_norm": 0.7054280042648315,
|
4052 |
+
"learning_rate": 7.306159915457342e-05,
|
4053 |
+
"loss": 0.6327,
|
4054 |
+
"step": 571
|
4055 |
+
},
|
4056 |
+
{
|
4057 |
+
"epoch": 0.03983980497997562,
|
4058 |
+
"grad_norm": 0.7624709606170654,
|
4059 |
+
"learning_rate": 7.274527479066883e-05,
|
4060 |
+
"loss": 0.8132,
|
4061 |
+
"step": 572
|
4062 |
+
},
|
4063 |
+
{
|
4064 |
+
"epoch": 0.039909454988681876,
|
4065 |
+
"grad_norm": 0.6930527687072754,
|
4066 |
+
"learning_rate": 7.242924475056127e-05,
|
4067 |
+
"loss": 0.8482,
|
4068 |
+
"step": 573
|
4069 |
+
},
|
4070 |
+
{
|
4071 |
+
"epoch": 0.039979104997388125,
|
4072 |
+
"grad_norm": 0.6599513292312622,
|
4073 |
+
"learning_rate": 7.211351244705946e-05,
|
4074 |
+
"loss": 0.6787,
|
4075 |
+
"step": 574
|
4076 |
+
},
|
4077 |
+
{
|
4078 |
+
"epoch": 0.04004875500609437,
|
4079 |
+
"grad_norm": 0.7311400771141052,
|
4080 |
+
"learning_rate": 7.179808128975674e-05,
|
4081 |
+
"loss": 0.9747,
|
4082 |
+
"step": 575
|
4083 |
+
},
|
4084 |
+
{
|
4085 |
+
"epoch": 0.04011840501480063,
|
4086 |
+
"grad_norm": 0.615138828754425,
|
4087 |
+
"learning_rate": 7.148295468499438e-05,
|
4088 |
+
"loss": 0.9404,
|
4089 |
+
"step": 576
|
4090 |
+
},
|
4091 |
+
{
|
4092 |
+
"epoch": 0.04018805502350688,
|
4093 |
+
"grad_norm": 0.6401761174201965,
|
4094 |
+
"learning_rate": 7.116813603582482e-05,
|
4095 |
+
"loss": 0.4915,
|
4096 |
+
"step": 577
|
4097 |
+
},
|
4098 |
+
{
|
4099 |
+
"epoch": 0.040257705032213126,
|
4100 |
+
"grad_norm": 0.6191440224647522,
|
4101 |
+
"learning_rate": 7.08536287419749e-05,
|
4102 |
+
"loss": 0.6031,
|
4103 |
+
"step": 578
|
4104 |
+
},
|
4105 |
+
{
|
4106 |
+
"epoch": 0.04032735504091938,
|
4107 |
+
"grad_norm": 0.5751050710678101,
|
4108 |
+
"learning_rate": 7.053943619980907e-05,
|
4109 |
+
"loss": 0.8371,
|
4110 |
+
"step": 579
|
4111 |
+
},
|
4112 |
+
{
|
4113 |
+
"epoch": 0.04039700504962563,
|
4114 |
+
"grad_norm": 0.518409252166748,
|
4115 |
+
"learning_rate": 7.022556180229285e-05,
|
4116 |
+
"loss": 0.4333,
|
4117 |
+
"step": 580
|
4118 |
+
},
|
4119 |
+
{
|
4120 |
+
"epoch": 0.040466655058331885,
|
4121 |
+
"grad_norm": 0.5712803602218628,
|
4122 |
+
"learning_rate": 6.991200893895608e-05,
|
4123 |
+
"loss": 0.796,
|
4124 |
+
"step": 581
|
4125 |
+
},
|
4126 |
+
{
|
4127 |
+
"epoch": 0.040536305067038134,
|
4128 |
+
"grad_norm": 0.661482036113739,
|
4129 |
+
"learning_rate": 6.959878099585635e-05,
|
4130 |
+
"loss": 0.8585,
|
4131 |
+
"step": 582
|
4132 |
+
},
|
4133 |
+
{
|
4134 |
+
"epoch": 0.04060595507574438,
|
4135 |
+
"grad_norm": 0.6602011322975159,
|
4136 |
+
"learning_rate": 6.92858813555424e-05,
|
4137 |
+
"loss": 0.9474,
|
4138 |
+
"step": 583
|
4139 |
+
},
|
4140 |
+
{
|
4141 |
+
"epoch": 0.04067560508445064,
|
4142 |
+
"grad_norm": 0.5971815586090088,
|
4143 |
+
"learning_rate": 6.897331339701776e-05,
|
4144 |
+
"loss": 0.7689,
|
4145 |
+
"step": 584
|
4146 |
+
},
|
4147 |
+
{
|
4148 |
+
"epoch": 0.040745255093156886,
|
4149 |
+
"grad_norm": 0.571740448474884,
|
4150 |
+
"learning_rate": 6.866108049570397e-05,
|
4151 |
+
"loss": 0.9023,
|
4152 |
+
"step": 585
|
4153 |
+
},
|
4154 |
+
{
|
4155 |
+
"epoch": 0.040814905101863135,
|
4156 |
+
"grad_norm": 0.6928638219833374,
|
4157 |
+
"learning_rate": 6.834918602340438e-05,
|
4158 |
+
"loss": 0.8899,
|
4159 |
+
"step": 586
|
4160 |
+
},
|
4161 |
+
{
|
4162 |
+
"epoch": 0.04088455511056939,
|
4163 |
+
"grad_norm": 0.6468199491500854,
|
4164 |
+
"learning_rate": 6.803763334826763e-05,
|
4165 |
+
"loss": 0.8841,
|
4166 |
+
"step": 587
|
4167 |
+
},
|
4168 |
+
{
|
4169 |
+
"epoch": 0.04095420511927564,
|
4170 |
+
"grad_norm": 0.6777251362800598,
|
4171 |
+
"learning_rate": 6.772642583475126e-05,
|
4172 |
+
"loss": 0.8491,
|
4173 |
+
"step": 588
|
4174 |
+
},
|
4175 |
+
{
|
4176 |
+
"epoch": 0.041023855127981894,
|
4177 |
+
"grad_norm": 0.5866687297821045,
|
4178 |
+
"learning_rate": 6.741556684358545e-05,
|
4179 |
+
"loss": 0.6435,
|
4180 |
+
"step": 589
|
4181 |
+
},
|
4182 |
+
{
|
4183 |
+
"epoch": 0.04109350513668814,
|
4184 |
+
"grad_norm": 0.5522730350494385,
|
4185 |
+
"learning_rate": 6.710505973173664e-05,
|
4186 |
+
"loss": 0.9188,
|
4187 |
+
"step": 590
|
4188 |
+
},
|
4189 |
+
{
|
4190 |
+
"epoch": 0.04116315514539439,
|
4191 |
+
"grad_norm": 0.7048250436782837,
|
4192 |
+
"learning_rate": 6.679490785237137e-05,
|
4193 |
+
"loss": 0.911,
|
4194 |
+
"step": 591
|
4195 |
+
},
|
4196 |
+
{
|
4197 |
+
"epoch": 0.041232805154100646,
|
4198 |
+
"grad_norm": 0.849677324295044,
|
4199 |
+
"learning_rate": 6.648511455482003e-05,
|
4200 |
+
"loss": 1.0408,
|
4201 |
+
"step": 592
|
4202 |
+
},
|
4203 |
+
{
|
4204 |
+
"epoch": 0.041302455162806895,
|
4205 |
+
"grad_norm": 0.653287947177887,
|
4206 |
+
"learning_rate": 6.617568318454059e-05,
|
4207 |
+
"loss": 1.187,
|
4208 |
+
"step": 593
|
4209 |
+
},
|
4210 |
+
{
|
4211 |
+
"epoch": 0.04137210517151314,
|
4212 |
+
"grad_norm": 0.5278560519218445,
|
4213 |
+
"learning_rate": 6.586661708308272e-05,
|
4214 |
+
"loss": 0.8789,
|
4215 |
+
"step": 594
|
4216 |
+
},
|
4217 |
+
{
|
4218 |
+
"epoch": 0.0414417551802194,
|
4219 |
+
"grad_norm": 0.7803817987442017,
|
4220 |
+
"learning_rate": 6.555791958805147e-05,
|
4221 |
+
"loss": 0.8788,
|
4222 |
+
"step": 595
|
4223 |
+
},
|
4224 |
+
{
|
4225 |
+
"epoch": 0.04151140518892565,
|
4226 |
+
"grad_norm": 0.6425774097442627,
|
4227 |
+
"learning_rate": 6.524959403307125e-05,
|
4228 |
+
"loss": 0.9296,
|
4229 |
+
"step": 596
|
4230 |
+
},
|
4231 |
+
{
|
4232 |
+
"epoch": 0.0415810551976319,
|
4233 |
+
"grad_norm": 0.5787883400917053,
|
4234 |
+
"learning_rate": 6.494164374775e-05,
|
4235 |
+
"loss": 1.0127,
|
4236 |
+
"step": 597
|
4237 |
+
},
|
4238 |
+
{
|
4239 |
+
"epoch": 0.04165070520633815,
|
4240 |
+
"grad_norm": 0.5686517357826233,
|
4241 |
+
"learning_rate": 6.463407205764305e-05,
|
4242 |
+
"loss": 0.7869,
|
4243 |
+
"step": 598
|
4244 |
+
},
|
4245 |
+
{
|
4246 |
+
"epoch": 0.0417203552150444,
|
4247 |
+
"grad_norm": 0.5126462578773499,
|
4248 |
+
"learning_rate": 6.43268822842173e-05,
|
4249 |
+
"loss": 1.2029,
|
4250 |
+
"step": 599
|
4251 |
+
},
|
4252 |
+
{
|
4253 |
+
"epoch": 0.041790005223750655,
|
4254 |
+
"grad_norm": 0.5618976950645447,
|
4255 |
+
"learning_rate": 6.402007774481536e-05,
|
4256 |
+
"loss": 0.5725,
|
4257 |
+
"step": 600
|
4258 |
+
},
|
4259 |
+
{
|
4260 |
+
"epoch": 0.041790005223750655,
|
4261 |
+
"eval_loss": 0.7635987401008606,
|
4262 |
+
"eval_runtime": 701.6781,
|
4263 |
+
"eval_samples_per_second": 7.126,
|
4264 |
+
"eval_steps_per_second": 1.781,
|
4265 |
+
"step": 600
|
4266 |
}
|
4267 |
],
|
4268 |
"logging_steps": 1,
|
|
|
4291 |
"attributes": {}
|
4292 |
}
|
4293 |
},
|
4294 |
+
"total_flos": 1.6329142901994947e+18,
|
4295 |
"train_batch_size": 4,
|
4296 |
"trial_name": null,
|
4297 |
"trial_params": null
|