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 289512208
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:cc7f9f12a7d975eceff199ac96f1fb4abb58016ab42da13fe25b1845a1e0b3ce
|
3 |
size 289512208
|
last-checkpoint/optimizer.pt
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 147781972
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:37a24bd2239b20338e528442d83d2139315c8017ad79ac6456a27ebc2a7c4982
|
3 |
size 147781972
|
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 sha256:28ace9fc649252ea1299cd2d9b1953184b717d1b1778bd2d51cf81f8fdd857fb
|
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 sha256:14de197ce4fca667a77214b11d375124cfec5ed9c075fb60180e734827aaa864
|
3 |
size 1064
|
last-checkpoint/trainer_state.json
CHANGED
@@ -1,9 +1,9 @@
|
|
1 |
{
|
2 |
-
"best_metric": 1.
|
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": 4.035,
|
3556 |
"eval_steps_per_second": 1.009,
|
3557 |
"step": 500
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
3558 |
}
|
3559 |
],
|
3560 |
"logging_steps": 1,
|
@@ -3583,7 +4291,7 @@
|
|
3583 |
"attributes": {}
|
3584 |
}
|
3585 |
},
|
3586 |
-
"total_flos":
|
3587 |
"train_batch_size": 4,
|
3588 |
"trial_name": null,
|
3589 |
"trial_params": null
|
|
|
1 |
{
|
2 |
+
"best_metric": 1.2089511156082153,
|
3 |
+
"best_model_checkpoint": "miner_id_24/checkpoint-600",
|
4 |
+
"epoch": 0.03813912836200389,
|
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": 4.035,
|
3556 |
"eval_steps_per_second": 1.009,
|
3557 |
"step": 500
|
3558 |
+
},
|
3559 |
+
{
|
3560 |
+
"epoch": 0.03184617218227325,
|
3561 |
+
"grad_norm": 0.26652681827545166,
|
3562 |
+
"learning_rate": 7.162871012503003e-05,
|
3563 |
+
"loss": 1.2841,
|
3564 |
+
"step": 501
|
3565 |
+
},
|
3566 |
+
{
|
3567 |
+
"epoch": 0.031909737396209926,
|
3568 |
+
"grad_norm": 0.258722722530365,
|
3569 |
+
"learning_rate": 7.126596213658488e-05,
|
3570 |
+
"loss": 1.3076,
|
3571 |
+
"step": 502
|
3572 |
+
},
|
3573 |
+
{
|
3574 |
+
"epoch": 0.031973302610146595,
|
3575 |
+
"grad_norm": 0.2519145607948303,
|
3576 |
+
"learning_rate": 7.090362580911808e-05,
|
3577 |
+
"loss": 1.3191,
|
3578 |
+
"step": 503
|
3579 |
+
},
|
3580 |
+
{
|
3581 |
+
"epoch": 0.03203686782408327,
|
3582 |
+
"grad_norm": 0.24408216774463654,
|
3583 |
+
"learning_rate": 7.05417063336761e-05,
|
3584 |
+
"loss": 1.2056,
|
3585 |
+
"step": 504
|
3586 |
+
},
|
3587 |
+
{
|
3588 |
+
"epoch": 0.032100433038019946,
|
3589 |
+
"grad_norm": 0.25859251618385315,
|
3590 |
+
"learning_rate": 7.018020889533348e-05,
|
3591 |
+
"loss": 1.2509,
|
3592 |
+
"step": 505
|
3593 |
+
},
|
3594 |
+
{
|
3595 |
+
"epoch": 0.032163998251956614,
|
3596 |
+
"grad_norm": 0.274295836687088,
|
3597 |
+
"learning_rate": 6.981913867311819e-05,
|
3598 |
+
"loss": 1.1457,
|
3599 |
+
"step": 506
|
3600 |
+
},
|
3601 |
+
{
|
3602 |
+
"epoch": 0.03222756346589329,
|
3603 |
+
"grad_norm": 0.25392991304397583,
|
3604 |
+
"learning_rate": 6.945850083993781e-05,
|
3605 |
+
"loss": 1.1389,
|
3606 |
+
"step": 507
|
3607 |
+
},
|
3608 |
+
{
|
3609 |
+
"epoch": 0.032291128679829965,
|
3610 |
+
"grad_norm": 0.2518613338470459,
|
3611 |
+
"learning_rate": 6.909830056250527e-05,
|
3612 |
+
"loss": 1.1288,
|
3613 |
+
"step": 508
|
3614 |
+
},
|
3615 |
+
{
|
3616 |
+
"epoch": 0.032354693893766634,
|
3617 |
+
"grad_norm": 0.25153687596321106,
|
3618 |
+
"learning_rate": 6.873854300126467e-05,
|
3619 |
+
"loss": 1.1106,
|
3620 |
+
"step": 509
|
3621 |
+
},
|
3622 |
+
{
|
3623 |
+
"epoch": 0.03241825910770331,
|
3624 |
+
"grad_norm": 0.2594717741012573,
|
3625 |
+
"learning_rate": 6.83792333103176e-05,
|
3626 |
+
"loss": 1.2556,
|
3627 |
+
"step": 510
|
3628 |
+
},
|
3629 |
+
{
|
3630 |
+
"epoch": 0.032481824321639985,
|
3631 |
+
"grad_norm": 0.2567812204360962,
|
3632 |
+
"learning_rate": 6.802037663734923e-05,
|
3633 |
+
"loss": 1.1773,
|
3634 |
+
"step": 511
|
3635 |
+
},
|
3636 |
+
{
|
3637 |
+
"epoch": 0.03254538953557665,
|
3638 |
+
"grad_norm": 0.2533416748046875,
|
3639 |
+
"learning_rate": 6.766197812355438e-05,
|
3640 |
+
"loss": 1.2053,
|
3641 |
+
"step": 512
|
3642 |
+
},
|
3643 |
+
{
|
3644 |
+
"epoch": 0.03260895474951333,
|
3645 |
+
"grad_norm": 0.2625131607055664,
|
3646 |
+
"learning_rate": 6.73040429035641e-05,
|
3647 |
+
"loss": 1.2467,
|
3648 |
+
"step": 513
|
3649 |
+
},
|
3650 |
+
{
|
3651 |
+
"epoch": 0.032672519963450004,
|
3652 |
+
"grad_norm": 0.2665756344795227,
|
3653 |
+
"learning_rate": 6.69465761053721e-05,
|
3654 |
+
"loss": 1.2737,
|
3655 |
+
"step": 514
|
3656 |
+
},
|
3657 |
+
{
|
3658 |
+
"epoch": 0.03273608517738667,
|
3659 |
+
"grad_norm": 0.2558302581310272,
|
3660 |
+
"learning_rate": 6.658958285026102e-05,
|
3661 |
+
"loss": 1.2684,
|
3662 |
+
"step": 515
|
3663 |
+
},
|
3664 |
+
{
|
3665 |
+
"epoch": 0.03279965039132335,
|
3666 |
+
"grad_norm": 0.24979649484157562,
|
3667 |
+
"learning_rate": 6.623306825272937e-05,
|
3668 |
+
"loss": 1.2144,
|
3669 |
+
"step": 516
|
3670 |
+
},
|
3671 |
+
{
|
3672 |
+
"epoch": 0.032863215605260024,
|
3673 |
+
"grad_norm": 0.25010988116264343,
|
3674 |
+
"learning_rate": 6.58770374204181e-05,
|
3675 |
+
"loss": 1.2679,
|
3676 |
+
"step": 517
|
3677 |
+
},
|
3678 |
+
{
|
3679 |
+
"epoch": 0.03292678081919669,
|
3680 |
+
"grad_norm": 0.2494756430387497,
|
3681 |
+
"learning_rate": 6.552149545403739e-05,
|
3682 |
+
"loss": 1.2291,
|
3683 |
+
"step": 518
|
3684 |
+
},
|
3685 |
+
{
|
3686 |
+
"epoch": 0.03299034603313337,
|
3687 |
+
"grad_norm": 0.2528434097766876,
|
3688 |
+
"learning_rate": 6.516644744729367e-05,
|
3689 |
+
"loss": 1.2519,
|
3690 |
+
"step": 519
|
3691 |
+
},
|
3692 |
+
{
|
3693 |
+
"epoch": 0.03305391124707004,
|
3694 |
+
"grad_norm": 0.2631024718284607,
|
3695 |
+
"learning_rate": 6.48118984868167e-05,
|
3696 |
+
"loss": 1.2086,
|
3697 |
+
"step": 520
|
3698 |
+
},
|
3699 |
+
{
|
3700 |
+
"epoch": 0.03311747646100671,
|
3701 |
+
"grad_norm": 0.2514025866985321,
|
3702 |
+
"learning_rate": 6.445785365208645e-05,
|
3703 |
+
"loss": 1.2098,
|
3704 |
+
"step": 521
|
3705 |
+
},
|
3706 |
+
{
|
3707 |
+
"epoch": 0.03318104167494339,
|
3708 |
+
"grad_norm": 0.25096845626831055,
|
3709 |
+
"learning_rate": 6.410431801536058e-05,
|
3710 |
+
"loss": 1.2062,
|
3711 |
+
"step": 522
|
3712 |
+
},
|
3713 |
+
{
|
3714 |
+
"epoch": 0.03324460688888006,
|
3715 |
+
"grad_norm": 0.24552412331104279,
|
3716 |
+
"learning_rate": 6.375129664160168e-05,
|
3717 |
+
"loss": 1.139,
|
3718 |
+
"step": 523
|
3719 |
+
},
|
3720 |
+
{
|
3721 |
+
"epoch": 0.03330817210281673,
|
3722 |
+
"grad_norm": 0.257235586643219,
|
3723 |
+
"learning_rate": 6.339879458840465e-05,
|
3724 |
+
"loss": 1.3102,
|
3725 |
+
"step": 524
|
3726 |
+
},
|
3727 |
+
{
|
3728 |
+
"epoch": 0.03337173731675341,
|
3729 |
+
"grad_norm": 0.24831882119178772,
|
3730 |
+
"learning_rate": 6.304681690592431e-05,
|
3731 |
+
"loss": 1.1848,
|
3732 |
+
"step": 525
|
3733 |
+
},
|
3734 |
+
{
|
3735 |
+
"epoch": 0.03343530253069008,
|
3736 |
+
"grad_norm": 0.2522595524787903,
|
3737 |
+
"learning_rate": 6.269536863680307e-05,
|
3738 |
+
"loss": 1.228,
|
3739 |
+
"step": 526
|
3740 |
+
},
|
3741 |
+
{
|
3742 |
+
"epoch": 0.03349886774462675,
|
3743 |
+
"grad_norm": 0.25006258487701416,
|
3744 |
+
"learning_rate": 6.23444548160986e-05,
|
3745 |
+
"loss": 1.1059,
|
3746 |
+
"step": 527
|
3747 |
+
},
|
3748 |
+
{
|
3749 |
+
"epoch": 0.033562432958563426,
|
3750 |
+
"grad_norm": 0.2498662769794464,
|
3751 |
+
"learning_rate": 6.199408047121174e-05,
|
3752 |
+
"loss": 1.2354,
|
3753 |
+
"step": 528
|
3754 |
+
},
|
3755 |
+
{
|
3756 |
+
"epoch": 0.0336259981725001,
|
3757 |
+
"grad_norm": 0.25638020038604736,
|
3758 |
+
"learning_rate": 6.16442506218146e-05,
|
3759 |
+
"loss": 1.2332,
|
3760 |
+
"step": 529
|
3761 |
+
},
|
3762 |
+
{
|
3763 |
+
"epoch": 0.03368956338643677,
|
3764 |
+
"grad_norm": 0.26172903180122375,
|
3765 |
+
"learning_rate": 6.129497027977829e-05,
|
3766 |
+
"loss": 1.2062,
|
3767 |
+
"step": 530
|
3768 |
+
},
|
3769 |
+
{
|
3770 |
+
"epoch": 0.033753128600373446,
|
3771 |
+
"grad_norm": 0.2451334148645401,
|
3772 |
+
"learning_rate": 6.0946244449101574e-05,
|
3773 |
+
"loss": 1.2513,
|
3774 |
+
"step": 531
|
3775 |
+
},
|
3776 |
+
{
|
3777 |
+
"epoch": 0.03381669381431012,
|
3778 |
+
"grad_norm": 0.2550620436668396,
|
3779 |
+
"learning_rate": 6.059807812583883e-05,
|
3780 |
+
"loss": 1.08,
|
3781 |
+
"step": 532
|
3782 |
+
},
|
3783 |
+
{
|
3784 |
+
"epoch": 0.03388025902824679,
|
3785 |
+
"grad_norm": 0.2467305064201355,
|
3786 |
+
"learning_rate": 6.02504762980286e-05,
|
3787 |
+
"loss": 1.1822,
|
3788 |
+
"step": 533
|
3789 |
+
},
|
3790 |
+
{
|
3791 |
+
"epoch": 0.033943824242183465,
|
3792 |
+
"grad_norm": 0.2539975345134735,
|
3793 |
+
"learning_rate": 5.990344394562226e-05,
|
3794 |
+
"loss": 1.2334,
|
3795 |
+
"step": 534
|
3796 |
+
},
|
3797 |
+
{
|
3798 |
+
"epoch": 0.03400738945612014,
|
3799 |
+
"grad_norm": 0.26874592900276184,
|
3800 |
+
"learning_rate": 5.955698604041231e-05,
|
3801 |
+
"loss": 1.2736,
|
3802 |
+
"step": 535
|
3803 |
+
},
|
3804 |
+
{
|
3805 |
+
"epoch": 0.03407095467005681,
|
3806 |
+
"grad_norm": 0.25802844762802124,
|
3807 |
+
"learning_rate": 5.92111075459616e-05,
|
3808 |
+
"loss": 1.1959,
|
3809 |
+
"step": 536
|
3810 |
+
},
|
3811 |
+
{
|
3812 |
+
"epoch": 0.034134519883993485,
|
3813 |
+
"grad_norm": 0.2507290542125702,
|
3814 |
+
"learning_rate": 5.88658134175319e-05,
|
3815 |
+
"loss": 1.2045,
|
3816 |
+
"step": 537
|
3817 |
+
},
|
3818 |
+
{
|
3819 |
+
"epoch": 0.03419808509793016,
|
3820 |
+
"grad_norm": 0.25168654322624207,
|
3821 |
+
"learning_rate": 5.852110860201294e-05,
|
3822 |
+
"loss": 1.259,
|
3823 |
+
"step": 538
|
3824 |
+
},
|
3825 |
+
{
|
3826 |
+
"epoch": 0.03426165031186683,
|
3827 |
+
"grad_norm": 0.25015729665756226,
|
3828 |
+
"learning_rate": 5.817699803785174e-05,
|
3829 |
+
"loss": 1.0865,
|
3830 |
+
"step": 539
|
3831 |
+
},
|
3832 |
+
{
|
3833 |
+
"epoch": 0.034325215525803504,
|
3834 |
+
"grad_norm": 0.24665945768356323,
|
3835 |
+
"learning_rate": 5.7833486654981606e-05,
|
3836 |
+
"loss": 1.1366,
|
3837 |
+
"step": 540
|
3838 |
+
},
|
3839 |
+
{
|
3840 |
+
"epoch": 0.03438878073974018,
|
3841 |
+
"grad_norm": 0.2521713972091675,
|
3842 |
+
"learning_rate": 5.7490579374751686e-05,
|
3843 |
+
"loss": 1.2052,
|
3844 |
+
"step": 541
|
3845 |
+
},
|
3846 |
+
{
|
3847 |
+
"epoch": 0.03445234595367685,
|
3848 |
+
"grad_norm": 0.24369041621685028,
|
3849 |
+
"learning_rate": 5.714828110985635e-05,
|
3850 |
+
"loss": 1.1564,
|
3851 |
+
"step": 542
|
3852 |
+
},
|
3853 |
+
{
|
3854 |
+
"epoch": 0.034515911167613524,
|
3855 |
+
"grad_norm": 0.2600708603858948,
|
3856 |
+
"learning_rate": 5.6806596764264874e-05,
|
3857 |
+
"loss": 1.1852,
|
3858 |
+
"step": 543
|
3859 |
+
},
|
3860 |
+
{
|
3861 |
+
"epoch": 0.0345794763815502,
|
3862 |
+
"grad_norm": 0.25693967938423157,
|
3863 |
+
"learning_rate": 5.6465531233151126e-05,
|
3864 |
+
"loss": 1.1887,
|
3865 |
+
"step": 544
|
3866 |
+
},
|
3867 |
+
{
|
3868 |
+
"epoch": 0.03464304159548687,
|
3869 |
+
"grad_norm": 0.24953435361385345,
|
3870 |
+
"learning_rate": 5.6125089402823485e-05,
|
3871 |
+
"loss": 1.0897,
|
3872 |
+
"step": 545
|
3873 |
+
},
|
3874 |
+
{
|
3875 |
+
"epoch": 0.03470660680942354,
|
3876 |
+
"grad_norm": 0.2587718665599823,
|
3877 |
+
"learning_rate": 5.578527615065492e-05,
|
3878 |
+
"loss": 1.1345,
|
3879 |
+
"step": 546
|
3880 |
+
},
|
3881 |
+
{
|
3882 |
+
"epoch": 0.03477017202336022,
|
3883 |
+
"grad_norm": 0.2567615807056427,
|
3884 |
+
"learning_rate": 5.544609634501279e-05,
|
3885 |
+
"loss": 1.277,
|
3886 |
+
"step": 547
|
3887 |
+
},
|
3888 |
+
{
|
3889 |
+
"epoch": 0.03483373723729689,
|
3890 |
+
"grad_norm": 0.25938692688941956,
|
3891 |
+
"learning_rate": 5.510755484518955e-05,
|
3892 |
+
"loss": 1.2087,
|
3893 |
+
"step": 548
|
3894 |
+
},
|
3895 |
+
{
|
3896 |
+
"epoch": 0.03489730245123356,
|
3897 |
+
"grad_norm": 0.25982430577278137,
|
3898 |
+
"learning_rate": 5.476965650133279e-05,
|
3899 |
+
"loss": 1.2554,
|
3900 |
+
"step": 549
|
3901 |
+
},
|
3902 |
+
{
|
3903 |
+
"epoch": 0.03496086766517024,
|
3904 |
+
"grad_norm": 0.25441205501556396,
|
3905 |
+
"learning_rate": 5.443240615437586e-05,
|
3906 |
+
"loss": 1.1437,
|
3907 |
+
"step": 550
|
3908 |
+
},
|
3909 |
+
{
|
3910 |
+
"epoch": 0.035024432879106907,
|
3911 |
+
"grad_norm": 0.2553006410598755,
|
3912 |
+
"learning_rate": 5.4095808635968546e-05,
|
3913 |
+
"loss": 1.2544,
|
3914 |
+
"step": 551
|
3915 |
+
},
|
3916 |
+
{
|
3917 |
+
"epoch": 0.03508799809304358,
|
3918 |
+
"grad_norm": 0.2520417273044586,
|
3919 |
+
"learning_rate": 5.375986876840784e-05,
|
3920 |
+
"loss": 1.0661,
|
3921 |
+
"step": 552
|
3922 |
+
},
|
3923 |
+
{
|
3924 |
+
"epoch": 0.03515156330698026,
|
3925 |
+
"grad_norm": 0.2597469091415405,
|
3926 |
+
"learning_rate": 5.342459136456881e-05,
|
3927 |
+
"loss": 1.1732,
|
3928 |
+
"step": 553
|
3929 |
+
},
|
3930 |
+
{
|
3931 |
+
"epoch": 0.035215128520916926,
|
3932 |
+
"grad_norm": 0.2597144544124603,
|
3933 |
+
"learning_rate": 5.30899812278356e-05,
|
3934 |
+
"loss": 1.2128,
|
3935 |
+
"step": 554
|
3936 |
+
},
|
3937 |
+
{
|
3938 |
+
"epoch": 0.0352786937348536,
|
3939 |
+
"grad_norm": 0.2565470039844513,
|
3940 |
+
"learning_rate": 5.275604315203293e-05,
|
3941 |
+
"loss": 1.1585,
|
3942 |
+
"step": 555
|
3943 |
+
},
|
3944 |
+
{
|
3945 |
+
"epoch": 0.03534225894879028,
|
3946 |
+
"grad_norm": 0.24314232170581818,
|
3947 |
+
"learning_rate": 5.2422781921356826e-05,
|
3948 |
+
"loss": 1.0955,
|
3949 |
+
"step": 556
|
3950 |
+
},
|
3951 |
+
{
|
3952 |
+
"epoch": 0.035405824162726945,
|
3953 |
+
"grad_norm": 0.24443262815475464,
|
3954 |
+
"learning_rate": 5.209020231030672e-05,
|
3955 |
+
"loss": 1.1649,
|
3956 |
+
"step": 557
|
3957 |
+
},
|
3958 |
+
{
|
3959 |
+
"epoch": 0.03546938937666362,
|
3960 |
+
"grad_norm": 0.26083680987358093,
|
3961 |
+
"learning_rate": 5.1758309083616673e-05,
|
3962 |
+
"loss": 1.2195,
|
3963 |
+
"step": 558
|
3964 |
+
},
|
3965 |
+
{
|
3966 |
+
"epoch": 0.035532954590600296,
|
3967 |
+
"grad_norm": 0.25181084871292114,
|
3968 |
+
"learning_rate": 5.142710699618701e-05,
|
3969 |
+
"loss": 1.2411,
|
3970 |
+
"step": 559
|
3971 |
+
},
|
3972 |
+
{
|
3973 |
+
"epoch": 0.035596519804536965,
|
3974 |
+
"grad_norm": 0.24913780391216278,
|
3975 |
+
"learning_rate": 5.109660079301668e-05,
|
3976 |
+
"loss": 1.2251,
|
3977 |
+
"step": 560
|
3978 |
+
},
|
3979 |
+
{
|
3980 |
+
"epoch": 0.03566008501847364,
|
3981 |
+
"grad_norm": 0.2578226625919342,
|
3982 |
+
"learning_rate": 5.076679520913479e-05,
|
3983 |
+
"loss": 1.1685,
|
3984 |
+
"step": 561
|
3985 |
+
},
|
3986 |
+
{
|
3987 |
+
"epoch": 0.035723650232410316,
|
3988 |
+
"grad_norm": 0.25115758180618286,
|
3989 |
+
"learning_rate": 5.043769496953299e-05,
|
3990 |
+
"loss": 1.0909,
|
3991 |
+
"step": 562
|
3992 |
+
},
|
3993 |
+
{
|
3994 |
+
"epoch": 0.035787215446346984,
|
3995 |
+
"grad_norm": 0.24539780616760254,
|
3996 |
+
"learning_rate": 5.010930478909779e-05,
|
3997 |
+
"loss": 1.221,
|
3998 |
+
"step": 563
|
3999 |
+
},
|
4000 |
+
{
|
4001 |
+
"epoch": 0.03585078066028366,
|
4002 |
+
"grad_norm": 0.26085364818573,
|
4003 |
+
"learning_rate": 4.9781629372542895e-05,
|
4004 |
+
"loss": 1.2141,
|
4005 |
+
"step": 564
|
4006 |
+
},
|
4007 |
+
{
|
4008 |
+
"epoch": 0.035914345874220335,
|
4009 |
+
"grad_norm": 0.25367042422294617,
|
4010 |
+
"learning_rate": 4.945467341434195e-05,
|
4011 |
+
"loss": 1.2999,
|
4012 |
+
"step": 565
|
4013 |
+
},
|
4014 |
+
{
|
4015 |
+
"epoch": 0.035977911088157004,
|
4016 |
+
"grad_norm": 0.27679672837257385,
|
4017 |
+
"learning_rate": 4.912844159866112e-05,
|
4018 |
+
"loss": 1.2494,
|
4019 |
+
"step": 566
|
4020 |
+
},
|
4021 |
+
{
|
4022 |
+
"epoch": 0.03604147630209368,
|
4023 |
+
"grad_norm": 0.2423914223909378,
|
4024 |
+
"learning_rate": 4.880293859929227e-05,
|
4025 |
+
"loss": 1.1681,
|
4026 |
+
"step": 567
|
4027 |
+
},
|
4028 |
+
{
|
4029 |
+
"epoch": 0.036105041516030355,
|
4030 |
+
"grad_norm": 0.24873730540275574,
|
4031 |
+
"learning_rate": 4.847816907958549e-05,
|
4032 |
+
"loss": 1.2964,
|
4033 |
+
"step": 568
|
4034 |
+
},
|
4035 |
+
{
|
4036 |
+
"epoch": 0.03616860672996702,
|
4037 |
+
"grad_norm": 0.2618268132209778,
|
4038 |
+
"learning_rate": 4.8154137692382907e-05,
|
4039 |
+
"loss": 1.2184,
|
4040 |
+
"step": 569
|
4041 |
+
},
|
4042 |
+
{
|
4043 |
+
"epoch": 0.0362321719439037,
|
4044 |
+
"grad_norm": 0.2559436857700348,
|
4045 |
+
"learning_rate": 4.783084907995156e-05,
|
4046 |
+
"loss": 1.216,
|
4047 |
+
"step": 570
|
4048 |
+
},
|
4049 |
+
{
|
4050 |
+
"epoch": 0.036295737157840374,
|
4051 |
+
"grad_norm": 0.2557474970817566,
|
4052 |
+
"learning_rate": 4.750830787391708e-05,
|
4053 |
+
"loss": 1.2993,
|
4054 |
+
"step": 571
|
4055 |
+
},
|
4056 |
+
{
|
4057 |
+
"epoch": 0.03635930237177704,
|
4058 |
+
"grad_norm": 0.25075972080230713,
|
4059 |
+
"learning_rate": 4.718651869519731e-05,
|
4060 |
+
"loss": 1.1178,
|
4061 |
+
"step": 572
|
4062 |
+
},
|
4063 |
+
{
|
4064 |
+
"epoch": 0.03642286758571372,
|
4065 |
+
"grad_norm": 0.2534898817539215,
|
4066 |
+
"learning_rate": 4.686548615393613e-05,
|
4067 |
+
"loss": 1.1891,
|
4068 |
+
"step": 573
|
4069 |
+
},
|
4070 |
+
{
|
4071 |
+
"epoch": 0.036486432799650394,
|
4072 |
+
"grad_norm": 0.26913774013519287,
|
4073 |
+
"learning_rate": 4.654521484943735e-05,
|
4074 |
+
"loss": 1.2552,
|
4075 |
+
"step": 574
|
4076 |
+
},
|
4077 |
+
{
|
4078 |
+
"epoch": 0.03654999801358706,
|
4079 |
+
"grad_norm": 0.2566869556903839,
|
4080 |
+
"learning_rate": 4.622570937009879e-05,
|
4081 |
+
"loss": 1.3527,
|
4082 |
+
"step": 575
|
4083 |
+
},
|
4084 |
+
{
|
4085 |
+
"epoch": 0.03661356322752374,
|
4086 |
+
"grad_norm": 0.26296791434288025,
|
4087 |
+
"learning_rate": 4.59069742933468e-05,
|
4088 |
+
"loss": 1.2464,
|
4089 |
+
"step": 576
|
4090 |
+
},
|
4091 |
+
{
|
4092 |
+
"epoch": 0.03667712844146041,
|
4093 |
+
"grad_norm": 0.261929988861084,
|
4094 |
+
"learning_rate": 4.558901418557021e-05,
|
4095 |
+
"loss": 1.2744,
|
4096 |
+
"step": 577
|
4097 |
+
},
|
4098 |
+
{
|
4099 |
+
"epoch": 0.03674069365539708,
|
4100 |
+
"grad_norm": 0.252572238445282,
|
4101 |
+
"learning_rate": 4.527183360205541e-05,
|
4102 |
+
"loss": 1.112,
|
4103 |
+
"step": 578
|
4104 |
+
},
|
4105 |
+
{
|
4106 |
+
"epoch": 0.03680425886933376,
|
4107 |
+
"grad_norm": 0.27232858538627625,
|
4108 |
+
"learning_rate": 4.495543708692075e-05,
|
4109 |
+
"loss": 1.2743,
|
4110 |
+
"step": 579
|
4111 |
+
},
|
4112 |
+
{
|
4113 |
+
"epoch": 0.03686782408327043,
|
4114 |
+
"grad_norm": 0.2609007656574249,
|
4115 |
+
"learning_rate": 4.4639829173051554e-05,
|
4116 |
+
"loss": 1.3188,
|
4117 |
+
"step": 580
|
4118 |
+
},
|
4119 |
+
{
|
4120 |
+
"epoch": 0.0369313892972071,
|
4121 |
+
"grad_norm": 0.2632231116294861,
|
4122 |
+
"learning_rate": 4.43250143820352e-05,
|
4123 |
+
"loss": 1.1243,
|
4124 |
+
"step": 581
|
4125 |
+
},
|
4126 |
+
{
|
4127 |
+
"epoch": 0.03699495451114378,
|
4128 |
+
"grad_norm": 0.2551668882369995,
|
4129 |
+
"learning_rate": 4.401099722409631e-05,
|
4130 |
+
"loss": 1.1864,
|
4131 |
+
"step": 582
|
4132 |
+
},
|
4133 |
+
{
|
4134 |
+
"epoch": 0.03705851972508045,
|
4135 |
+
"grad_norm": 0.2566946744918823,
|
4136 |
+
"learning_rate": 4.369778219803211e-05,
|
4137 |
+
"loss": 1.2117,
|
4138 |
+
"step": 583
|
4139 |
+
},
|
4140 |
+
{
|
4141 |
+
"epoch": 0.03712208493901712,
|
4142 |
+
"grad_norm": 0.2648017704486847,
|
4143 |
+
"learning_rate": 4.338537379114801e-05,
|
4144 |
+
"loss": 1.2357,
|
4145 |
+
"step": 584
|
4146 |
+
},
|
4147 |
+
{
|
4148 |
+
"epoch": 0.037185650152953796,
|
4149 |
+
"grad_norm": 0.24553461372852325,
|
4150 |
+
"learning_rate": 4.307377647919343e-05,
|
4151 |
+
"loss": 1.1774,
|
4152 |
+
"step": 585
|
4153 |
+
},
|
4154 |
+
{
|
4155 |
+
"epoch": 0.03724921536689047,
|
4156 |
+
"grad_norm": 0.24831490218639374,
|
4157 |
+
"learning_rate": 4.2762994726297346e-05,
|
4158 |
+
"loss": 1.1336,
|
4159 |
+
"step": 586
|
4160 |
+
},
|
4161 |
+
{
|
4162 |
+
"epoch": 0.03731278058082714,
|
4163 |
+
"grad_norm": 0.2644321024417877,
|
4164 |
+
"learning_rate": 4.245303298490467e-05,
|
4165 |
+
"loss": 1.1865,
|
4166 |
+
"step": 587
|
4167 |
+
},
|
4168 |
+
{
|
4169 |
+
"epoch": 0.037376345794763816,
|
4170 |
+
"grad_norm": 0.2654249370098114,
|
4171 |
+
"learning_rate": 4.2143895695712444e-05,
|
4172 |
+
"loss": 1.1872,
|
4173 |
+
"step": 588
|
4174 |
+
},
|
4175 |
+
{
|
4176 |
+
"epoch": 0.03743991100870049,
|
4177 |
+
"grad_norm": 0.24722780287265778,
|
4178 |
+
"learning_rate": 4.183558728760586e-05,
|
4179 |
+
"loss": 1.1609,
|
4180 |
+
"step": 589
|
4181 |
+
},
|
4182 |
+
{
|
4183 |
+
"epoch": 0.03750347622263716,
|
4184 |
+
"grad_norm": 0.2560960054397583,
|
4185 |
+
"learning_rate": 4.152811217759529e-05,
|
4186 |
+
"loss": 1.1897,
|
4187 |
+
"step": 590
|
4188 |
+
},
|
4189 |
+
{
|
4190 |
+
"epoch": 0.037567041436573835,
|
4191 |
+
"grad_norm": 0.24647943675518036,
|
4192 |
+
"learning_rate": 4.12214747707527e-05,
|
4193 |
+
"loss": 1.1098,
|
4194 |
+
"step": 591
|
4195 |
+
},
|
4196 |
+
{
|
4197 |
+
"epoch": 0.03763060665051051,
|
4198 |
+
"grad_norm": 0.2763408422470093,
|
4199 |
+
"learning_rate": 4.091567946014858e-05,
|
4200 |
+
"loss": 1.2313,
|
4201 |
+
"step": 592
|
4202 |
+
},
|
4203 |
+
{
|
4204 |
+
"epoch": 0.03769417186444718,
|
4205 |
+
"grad_norm": 0.254727303981781,
|
4206 |
+
"learning_rate": 4.061073062678912e-05,
|
4207 |
+
"loss": 1.2363,
|
4208 |
+
"step": 593
|
4209 |
+
},
|
4210 |
+
{
|
4211 |
+
"epoch": 0.037757737078383855,
|
4212 |
+
"grad_norm": 0.2722640037536621,
|
4213 |
+
"learning_rate": 4.0306632639553323e-05,
|
4214 |
+
"loss": 1.1633,
|
4215 |
+
"step": 594
|
4216 |
+
},
|
4217 |
+
{
|
4218 |
+
"epoch": 0.03782130229232053,
|
4219 |
+
"grad_norm": 0.2594556212425232,
|
4220 |
+
"learning_rate": 4.000338985513046e-05,
|
4221 |
+
"loss": 1.224,
|
4222 |
+
"step": 595
|
4223 |
+
},
|
4224 |
+
{
|
4225 |
+
"epoch": 0.0378848675062572,
|
4226 |
+
"grad_norm": 0.25068074464797974,
|
4227 |
+
"learning_rate": 3.970100661795766e-05,
|
4228 |
+
"loss": 1.1809,
|
4229 |
+
"step": 596
|
4230 |
+
},
|
4231 |
+
{
|
4232 |
+
"epoch": 0.037948432720193874,
|
4233 |
+
"grad_norm": 0.2550009787082672,
|
4234 |
+
"learning_rate": 3.9399487260157766e-05,
|
4235 |
+
"loss": 1.2022,
|
4236 |
+
"step": 597
|
4237 |
+
},
|
4238 |
+
{
|
4239 |
+
"epoch": 0.03801199793413055,
|
4240 |
+
"grad_norm": 0.2593897879123688,
|
4241 |
+
"learning_rate": 3.909883610147696e-05,
|
4242 |
+
"loss": 1.3124,
|
4243 |
+
"step": 598
|
4244 |
+
},
|
4245 |
+
{
|
4246 |
+
"epoch": 0.03807556314806722,
|
4247 |
+
"grad_norm": 0.2558402717113495,
|
4248 |
+
"learning_rate": 3.879905744922329e-05,
|
4249 |
+
"loss": 1.1618,
|
4250 |
+
"step": 599
|
4251 |
+
},
|
4252 |
+
{
|
4253 |
+
"epoch": 0.03813912836200389,
|
4254 |
+
"grad_norm": 0.25934746861457825,
|
4255 |
+
"learning_rate": 3.8500155598204644e-05,
|
4256 |
+
"loss": 1.0083,
|
4257 |
+
"step": 600
|
4258 |
+
},
|
4259 |
+
{
|
4260 |
+
"epoch": 0.03813912836200389,
|
4261 |
+
"eval_loss": 1.2089511156082153,
|
4262 |
+
"eval_runtime": 1239.249,
|
4263 |
+
"eval_samples_per_second": 4.035,
|
4264 |
+
"eval_steps_per_second": 1.009,
|
4265 |
+
"step": 600
|
4266 |
}
|
4267 |
],
|
4268 |
"logging_steps": 1,
|
|
|
4291 |
"attributes": {}
|
4292 |
}
|
4293 |
},
|
4294 |
+
"total_flos": 3.117735147995136e+18,
|
4295 |
"train_batch_size": 4,
|
4296 |
"trial_name": null,
|
4297 |
"trial_params": null
|