Training in progress, step 552, 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 144805440
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:9f5adb3a5e16f935dab431d5c7fc988b0cb1691a418940636f10652d62b82db7
|
| 3 |
size 144805440
|
last-checkpoint/optimizer.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 74292308
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:00a072bec46c4c9546123646ed3daf91f92526592704bf014cc7ef83d90b4303
|
| 3 |
size 74292308
|
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 sha256:8e1aa12142861e36681d375c8477ba6f862e4a421988518266fc08b6e26330f9
|
| 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:cf3863c8948e41c40f37017b57744f08214b2c38da69d49fe98001649774bc48
|
| 3 |
size 1064
|
last-checkpoint/trainer_state.json
CHANGED
|
@@ -1,9 +1,9 @@
|
|
| 1 |
{
|
| 2 |
"best_metric": 1.22141695022583,
|
| 3 |
"best_model_checkpoint": "miner_id_24/checkpoint-500",
|
| 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 +3555,370 @@
|
|
| 3555 |
"eval_samples_per_second": 3.977,
|
| 3556 |
"eval_steps_per_second": 0.994,
|
| 3557 |
"step": 500
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 3558 |
}
|
| 3559 |
],
|
| 3560 |
"logging_steps": 1,
|
|
@@ -3578,12 +3942,12 @@
|
|
| 3578 |
"should_evaluate": false,
|
| 3579 |
"should_log": false,
|
| 3580 |
"should_save": true,
|
| 3581 |
-
"should_training_stop":
|
| 3582 |
},
|
| 3583 |
"attributes": {}
|
| 3584 |
}
|
| 3585 |
},
|
| 3586 |
-
"total_flos": 5.
|
| 3587 |
"train_batch_size": 4,
|
| 3588 |
"trial_name": null,
|
| 3589 |
"trial_params": null
|
|
|
|
| 1 |
{
|
| 2 |
"best_metric": 1.22141695022583,
|
| 3 |
"best_model_checkpoint": "miner_id_24/checkpoint-500",
|
| 4 |
+
"epoch": 0.07017599618608716,
|
| 5 |
"eval_steps": 100,
|
| 6 |
+
"global_step": 552,
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
|
|
| 3555 |
"eval_samples_per_second": 3.977,
|
| 3556 |
"eval_steps_per_second": 0.994,
|
| 3557 |
"step": 500
|
| 3558 |
+
},
|
| 3559 |
+
{
|
| 3560 |
+
"epoch": 0.0636923443645465,
|
| 3561 |
+
"grad_norm": 0.20111538469791412,
|
| 3562 |
+
"learning_rate": 4.337571235498628e-06,
|
| 3563 |
+
"loss": 1.2477,
|
| 3564 |
+
"step": 501
|
| 3565 |
+
},
|
| 3566 |
+
{
|
| 3567 |
+
"epoch": 0.06381947479241985,
|
| 3568 |
+
"grad_norm": 0.19941484928131104,
|
| 3569 |
+
"learning_rate": 4.170318841070708e-06,
|
| 3570 |
+
"loss": 1.2335,
|
| 3571 |
+
"step": 502
|
| 3572 |
+
},
|
| 3573 |
+
{
|
| 3574 |
+
"epoch": 0.06394660522029319,
|
| 3575 |
+
"grad_norm": 0.19625377655029297,
|
| 3576 |
+
"learning_rate": 4.00628603241544e-06,
|
| 3577 |
+
"loss": 1.2302,
|
| 3578 |
+
"step": 503
|
| 3579 |
+
},
|
| 3580 |
+
{
|
| 3581 |
+
"epoch": 0.06407373564816654,
|
| 3582 |
+
"grad_norm": 0.2074848711490631,
|
| 3583 |
+
"learning_rate": 3.845478320536178e-06,
|
| 3584 |
+
"loss": 1.22,
|
| 3585 |
+
"step": 504
|
| 3586 |
+
},
|
| 3587 |
+
{
|
| 3588 |
+
"epoch": 0.06420086607603989,
|
| 3589 |
+
"grad_norm": 0.19643086194992065,
|
| 3590 |
+
"learning_rate": 3.687901108082892e-06,
|
| 3591 |
+
"loss": 1.2234,
|
| 3592 |
+
"step": 505
|
| 3593 |
+
},
|
| 3594 |
+
{
|
| 3595 |
+
"epoch": 0.06432799650391323,
|
| 3596 |
+
"grad_norm": 0.19488604366779327,
|
| 3597 |
+
"learning_rate": 3.53355968917054e-06,
|
| 3598 |
+
"loss": 1.2088,
|
| 3599 |
+
"step": 506
|
| 3600 |
+
},
|
| 3601 |
+
{
|
| 3602 |
+
"epoch": 0.06445512693178658,
|
| 3603 |
+
"grad_norm": 0.20913287997245789,
|
| 3604 |
+
"learning_rate": 3.3824592492013085e-06,
|
| 3605 |
+
"loss": 1.1734,
|
| 3606 |
+
"step": 507
|
| 3607 |
+
},
|
| 3608 |
+
{
|
| 3609 |
+
"epoch": 0.06458225735965993,
|
| 3610 |
+
"grad_norm": 0.19445528090000153,
|
| 3611 |
+
"learning_rate": 3.2346048646903494e-06,
|
| 3612 |
+
"loss": 1.2014,
|
| 3613 |
+
"step": 508
|
| 3614 |
+
},
|
| 3615 |
+
{
|
| 3616 |
+
"epoch": 0.06470938778753327,
|
| 3617 |
+
"grad_norm": 0.19821353256702423,
|
| 3618 |
+
"learning_rate": 3.0900015030951744e-06,
|
| 3619 |
+
"loss": 1.2446,
|
| 3620 |
+
"step": 509
|
| 3621 |
+
},
|
| 3622 |
+
{
|
| 3623 |
+
"epoch": 0.06483651821540662,
|
| 3624 |
+
"grad_norm": 0.20283670723438263,
|
| 3625 |
+
"learning_rate": 2.9486540226488557e-06,
|
| 3626 |
+
"loss": 1.1858,
|
| 3627 |
+
"step": 510
|
| 3628 |
+
},
|
| 3629 |
+
{
|
| 3630 |
+
"epoch": 0.06496364864327997,
|
| 3631 |
+
"grad_norm": 0.20396962761878967,
|
| 3632 |
+
"learning_rate": 2.8105671721967875e-06,
|
| 3633 |
+
"loss": 1.2166,
|
| 3634 |
+
"step": 511
|
| 3635 |
+
},
|
| 3636 |
+
{
|
| 3637 |
+
"epoch": 0.0650907790711533,
|
| 3638 |
+
"grad_norm": 0.2022467404603958,
|
| 3639 |
+
"learning_rate": 2.6757455910370488e-06,
|
| 3640 |
+
"loss": 1.1738,
|
| 3641 |
+
"step": 512
|
| 3642 |
+
},
|
| 3643 |
+
{
|
| 3644 |
+
"epoch": 0.06521790949902666,
|
| 3645 |
+
"grad_norm": 0.1929662674665451,
|
| 3646 |
+
"learning_rate": 2.5441938087646612e-06,
|
| 3647 |
+
"loss": 1.205,
|
| 3648 |
+
"step": 513
|
| 3649 |
+
},
|
| 3650 |
+
{
|
| 3651 |
+
"epoch": 0.06534503992690001,
|
| 3652 |
+
"grad_norm": 0.19997857511043549,
|
| 3653 |
+
"learning_rate": 2.4159162451193097e-06,
|
| 3654 |
+
"loss": 1.2103,
|
| 3655 |
+
"step": 514
|
| 3656 |
+
},
|
| 3657 |
+
{
|
| 3658 |
+
"epoch": 0.06547217035477335,
|
| 3659 |
+
"grad_norm": 0.1969158947467804,
|
| 3660 |
+
"learning_rate": 2.290917209836918e-06,
|
| 3661 |
+
"loss": 1.1912,
|
| 3662 |
+
"step": 515
|
| 3663 |
+
},
|
| 3664 |
+
{
|
| 3665 |
+
"epoch": 0.0655993007826467,
|
| 3666 |
+
"grad_norm": 0.19717784225940704,
|
| 3667 |
+
"learning_rate": 2.1692009025048422e-06,
|
| 3668 |
+
"loss": 1.2374,
|
| 3669 |
+
"step": 516
|
| 3670 |
+
},
|
| 3671 |
+
{
|
| 3672 |
+
"epoch": 0.06572643121052005,
|
| 3673 |
+
"grad_norm": 0.20149968564510345,
|
| 3674 |
+
"learning_rate": 2.0507714124207157e-06,
|
| 3675 |
+
"loss": 1.1857,
|
| 3676 |
+
"step": 517
|
| 3677 |
+
},
|
| 3678 |
+
{
|
| 3679 |
+
"epoch": 0.06585356163839338,
|
| 3680 |
+
"grad_norm": 0.19946229457855225,
|
| 3681 |
+
"learning_rate": 1.9356327184551714e-06,
|
| 3682 |
+
"loss": 1.1947,
|
| 3683 |
+
"step": 518
|
| 3684 |
+
},
|
| 3685 |
+
{
|
| 3686 |
+
"epoch": 0.06598069206626674,
|
| 3687 |
+
"grad_norm": 0.19773566722869873,
|
| 3688 |
+
"learning_rate": 1.8237886889180489e-06,
|
| 3689 |
+
"loss": 1.2825,
|
| 3690 |
+
"step": 519
|
| 3691 |
+
},
|
| 3692 |
+
{
|
| 3693 |
+
"epoch": 0.06610782249414009,
|
| 3694 |
+
"grad_norm": 0.20150107145309448,
|
| 3695 |
+
"learning_rate": 1.7152430814285303e-06,
|
| 3696 |
+
"loss": 1.2352,
|
| 3697 |
+
"step": 520
|
| 3698 |
+
},
|
| 3699 |
+
{
|
| 3700 |
+
"epoch": 0.06623495292201342,
|
| 3701 |
+
"grad_norm": 0.20660698413848877,
|
| 3702 |
+
"learning_rate": 1.6099995427888315e-06,
|
| 3703 |
+
"loss": 1.2127,
|
| 3704 |
+
"step": 521
|
| 3705 |
+
},
|
| 3706 |
+
{
|
| 3707 |
+
"epoch": 0.06636208334988677,
|
| 3708 |
+
"grad_norm": 0.19696985185146332,
|
| 3709 |
+
"learning_rate": 1.5080616088616884e-06,
|
| 3710 |
+
"loss": 1.2197,
|
| 3711 |
+
"step": 522
|
| 3712 |
+
},
|
| 3713 |
+
{
|
| 3714 |
+
"epoch": 0.06648921377776013,
|
| 3715 |
+
"grad_norm": 0.20150014758110046,
|
| 3716 |
+
"learning_rate": 1.4094327044515853e-06,
|
| 3717 |
+
"loss": 1.2534,
|
| 3718 |
+
"step": 523
|
| 3719 |
+
},
|
| 3720 |
+
{
|
| 3721 |
+
"epoch": 0.06661634420563346,
|
| 3722 |
+
"grad_norm": 0.1950562745332718,
|
| 3723 |
+
"learning_rate": 1.3141161431896808e-06,
|
| 3724 |
+
"loss": 1.165,
|
| 3725 |
+
"step": 524
|
| 3726 |
+
},
|
| 3727 |
+
{
|
| 3728 |
+
"epoch": 0.06674347463350681,
|
| 3729 |
+
"grad_norm": 0.20011726021766663,
|
| 3730 |
+
"learning_rate": 1.222115127422485e-06,
|
| 3731 |
+
"loss": 1.2179,
|
| 3732 |
+
"step": 525
|
| 3733 |
+
},
|
| 3734 |
+
{
|
| 3735 |
+
"epoch": 0.06687060506138016,
|
| 3736 |
+
"grad_norm": 0.20083405077457428,
|
| 3737 |
+
"learning_rate": 1.1334327481042573e-06,
|
| 3738 |
+
"loss": 1.305,
|
| 3739 |
+
"step": 526
|
| 3740 |
+
},
|
| 3741 |
+
{
|
| 3742 |
+
"epoch": 0.0669977354892535,
|
| 3743 |
+
"grad_norm": 0.20200292766094208,
|
| 3744 |
+
"learning_rate": 1.0480719846931774e-06,
|
| 3745 |
+
"loss": 1.2263,
|
| 3746 |
+
"step": 527
|
| 3747 |
+
},
|
| 3748 |
+
{
|
| 3749 |
+
"epoch": 0.06712486591712685,
|
| 3750 |
+
"grad_norm": 0.20572660863399506,
|
| 3751 |
+
"learning_rate": 9.660357050512158e-07,
|
| 3752 |
+
"loss": 1.2029,
|
| 3753 |
+
"step": 528
|
| 3754 |
+
},
|
| 3755 |
+
{
|
| 3756 |
+
"epoch": 0.0672519963450002,
|
| 3757 |
+
"grad_norm": 0.20432178676128387,
|
| 3758 |
+
"learning_rate": 8.873266653478208e-07,
|
| 3759 |
+
"loss": 1.2703,
|
| 3760 |
+
"step": 529
|
| 3761 |
+
},
|
| 3762 |
+
{
|
| 3763 |
+
"epoch": 0.06737912677287354,
|
| 3764 |
+
"grad_norm": 0.1951807290315628,
|
| 3765 |
+
"learning_rate": 8.119475099673036e-07,
|
| 3766 |
+
"loss": 1.2298,
|
| 3767 |
+
"step": 530
|
| 3768 |
+
},
|
| 3769 |
+
{
|
| 3770 |
+
"epoch": 0.06750625720074689,
|
| 3771 |
+
"grad_norm": 0.2002389281988144,
|
| 3772 |
+
"learning_rate": 7.399007714199658e-07,
|
| 3773 |
+
"loss": 1.2342,
|
| 3774 |
+
"step": 531
|
| 3775 |
+
},
|
| 3776 |
+
{
|
| 3777 |
+
"epoch": 0.06763338762862024,
|
| 3778 |
+
"grad_norm": 0.19941021502017975,
|
| 3779 |
+
"learning_rate": 6.711888702570556e-07,
|
| 3780 |
+
"loss": 1.146,
|
| 3781 |
+
"step": 532
|
| 3782 |
+
},
|
| 3783 |
+
{
|
| 3784 |
+
"epoch": 0.06776051805649358,
|
| 3785 |
+
"grad_norm": 0.19345982372760773,
|
| 3786 |
+
"learning_rate": 6.058141149894336e-07,
|
| 3787 |
+
"loss": 1.1954,
|
| 3788 |
+
"step": 533
|
| 3789 |
+
},
|
| 3790 |
+
{
|
| 3791 |
+
"epoch": 0.06788764848436693,
|
| 3792 |
+
"grad_norm": 0.1961802840232849,
|
| 3793 |
+
"learning_rate": 5.437787020100115e-07,
|
| 3794 |
+
"loss": 1.2165,
|
| 3795 |
+
"step": 534
|
| 3796 |
+
},
|
| 3797 |
+
{
|
| 3798 |
+
"epoch": 0.06801477891224028,
|
| 3799 |
+
"grad_norm": 0.20169439911842346,
|
| 3800 |
+
"learning_rate": 4.850847155199567e-07,
|
| 3801 |
+
"loss": 1.2445,
|
| 3802 |
+
"step": 535
|
| 3803 |
+
},
|
| 3804 |
+
{
|
| 3805 |
+
"epoch": 0.06814190934011362,
|
| 3806 |
+
"grad_norm": 0.19823016226291656,
|
| 3807 |
+
"learning_rate": 4.297341274586475e-07,
|
| 3808 |
+
"loss": 1.2371,
|
| 3809 |
+
"step": 536
|
| 3810 |
+
},
|
| 3811 |
+
{
|
| 3812 |
+
"epoch": 0.06826903976798697,
|
| 3813 |
+
"grad_norm": 0.2043391764163971,
|
| 3814 |
+
"learning_rate": 3.777287974374932e-07,
|
| 3815 |
+
"loss": 1.2942,
|
| 3816 |
+
"step": 537
|
| 3817 |
+
},
|
| 3818 |
+
{
|
| 3819 |
+
"epoch": 0.06839617019586032,
|
| 3820 |
+
"grad_norm": 0.20149071514606476,
|
| 3821 |
+
"learning_rate": 3.290704726773619e-07,
|
| 3822 |
+
"loss": 1.1842,
|
| 3823 |
+
"step": 538
|
| 3824 |
+
},
|
| 3825 |
+
{
|
| 3826 |
+
"epoch": 0.06852330062373366,
|
| 3827 |
+
"grad_norm": 0.20085620880126953,
|
| 3828 |
+
"learning_rate": 2.837607879499604e-07,
|
| 3829 |
+
"loss": 1.1982,
|
| 3830 |
+
"step": 539
|
| 3831 |
+
},
|
| 3832 |
+
{
|
| 3833 |
+
"epoch": 0.06865043105160701,
|
| 3834 |
+
"grad_norm": 0.2070370465517044,
|
| 3835 |
+
"learning_rate": 2.418012655228452e-07,
|
| 3836 |
+
"loss": 1.211,
|
| 3837 |
+
"step": 540
|
| 3838 |
+
},
|
| 3839 |
+
{
|
| 3840 |
+
"epoch": 0.06877756147948036,
|
| 3841 |
+
"grad_norm": 0.2030269056558609,
|
| 3842 |
+
"learning_rate": 2.0319331510835205e-07,
|
| 3843 |
+
"loss": 1.2534,
|
| 3844 |
+
"step": 541
|
| 3845 |
+
},
|
| 3846 |
+
{
|
| 3847 |
+
"epoch": 0.0689046919073537,
|
| 3848 |
+
"grad_norm": 0.1966077983379364,
|
| 3849 |
+
"learning_rate": 1.6793823381614505e-07,
|
| 3850 |
+
"loss": 1.1683,
|
| 3851 |
+
"step": 542
|
| 3852 |
+
},
|
| 3853 |
+
{
|
| 3854 |
+
"epoch": 0.06903182233522705,
|
| 3855 |
+
"grad_norm": 0.205659419298172,
|
| 3856 |
+
"learning_rate": 1.3603720610972925e-07,
|
| 3857 |
+
"loss": 1.141,
|
| 3858 |
+
"step": 543
|
| 3859 |
+
},
|
| 3860 |
+
{
|
| 3861 |
+
"epoch": 0.0691589527631004,
|
| 3862 |
+
"grad_norm": 0.2047136127948761,
|
| 3863 |
+
"learning_rate": 1.0749130376659366e-07,
|
| 3864 |
+
"loss": 1.2415,
|
| 3865 |
+
"step": 544
|
| 3866 |
+
},
|
| 3867 |
+
{
|
| 3868 |
+
"epoch": 0.06928608319097374,
|
| 3869 |
+
"grad_norm": 0.2038879543542862,
|
| 3870 |
+
"learning_rate": 8.230148584219554e-08,
|
| 3871 |
+
"loss": 1.2148,
|
| 3872 |
+
"step": 545
|
| 3873 |
+
},
|
| 3874 |
+
{
|
| 3875 |
+
"epoch": 0.06941321361884709,
|
| 3876 |
+
"grad_norm": 0.19984766840934753,
|
| 3877 |
+
"learning_rate": 6.046859863781951e-08,
|
| 3878 |
+
"loss": 1.1954,
|
| 3879 |
+
"step": 546
|
| 3880 |
+
},
|
| 3881 |
+
{
|
| 3882 |
+
"epoch": 0.06954034404672044,
|
| 3883 |
+
"grad_norm": 0.2025536447763443,
|
| 3884 |
+
"learning_rate": 4.199337567203365e-08,
|
| 3885 |
+
"loss": 1.2076,
|
| 3886 |
+
"step": 547
|
| 3887 |
+
},
|
| 3888 |
+
{
|
| 3889 |
+
"epoch": 0.06966747447459377,
|
| 3890 |
+
"grad_norm": 0.2022514045238495,
|
| 3891 |
+
"learning_rate": 2.6876437656153665e-08,
|
| 3892 |
+
"loss": 1.2369,
|
| 3893 |
+
"step": 548
|
| 3894 |
+
},
|
| 3895 |
+
{
|
| 3896 |
+
"epoch": 0.06979460490246713,
|
| 3897 |
+
"grad_norm": 0.21228355169296265,
|
| 3898 |
+
"learning_rate": 1.5118292473292885e-08,
|
| 3899 |
+
"loss": 1.2304,
|
| 3900 |
+
"step": 549
|
| 3901 |
+
},
|
| 3902 |
+
{
|
| 3903 |
+
"epoch": 0.06992173533034048,
|
| 3904 |
+
"grad_norm": 0.20524141192436218,
|
| 3905 |
+
"learning_rate": 6.719335161364804e-09,
|
| 3906 |
+
"loss": 1.2158,
|
| 3907 |
+
"step": 550
|
| 3908 |
+
},
|
| 3909 |
+
{
|
| 3910 |
+
"epoch": 0.07004886575821381,
|
| 3911 |
+
"grad_norm": 0.18994790315628052,
|
| 3912 |
+
"learning_rate": 1.6798478997825939e-09,
|
| 3913 |
+
"loss": 1.1945,
|
| 3914 |
+
"step": 551
|
| 3915 |
+
},
|
| 3916 |
+
{
|
| 3917 |
+
"epoch": 0.07017599618608716,
|
| 3918 |
+
"grad_norm": 0.20824959874153137,
|
| 3919 |
+
"learning_rate": 0.0,
|
| 3920 |
+
"loss": 1.241,
|
| 3921 |
+
"step": 552
|
| 3922 |
}
|
| 3923 |
],
|
| 3924 |
"logging_steps": 1,
|
|
|
|
| 3942 |
"should_evaluate": false,
|
| 3943 |
"should_log": false,
|
| 3944 |
"should_save": true,
|
| 3945 |
+
"should_training_stop": true
|
| 3946 |
},
|
| 3947 |
"attributes": {}
|
| 3948 |
}
|
| 3949 |
},
|
| 3950 |
+
"total_flos": 5.720928125466968e+18,
|
| 3951 |
"train_batch_size": 4,
|
| 3952 |
"trial_name": null,
|
| 3953 |
"trial_params": null
|