Training in progress, step 689, 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 125248064
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:041f64e371597393e019be4e6ff4350d5d401ed932c90bcf32d228f5048b36cb
|
3 |
size 125248064
|
last-checkpoint/optimizer.pt
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 64220436
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:82b2b3c85d6c13772f251cf444d87ea08f3e13e439374182d515b33b6bfa924e
|
3 |
size 64220436
|
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 sha256:733a8078a9a3de796ea1643fe169ffb5ef7502a8fc626cde056eafcb9b3baf9c
|
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 sha256:14bfcc404c13c41041f67f6bce66a78c5779da7ad0881e01dc99ce86c536e721
|
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": 173,
|
6 |
-
"global_step":
|
7 |
"is_hyper_param_search": false,
|
8 |
"is_local_process_zero": true,
|
9 |
"is_world_process_zero": true,
|
@@ -3672,6 +3672,1196 @@
|
|
3672 |
"eval_samples_per_second": 8.964,
|
3673 |
"eval_steps_per_second": 4.498,
|
3674 |
"step": 519
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
3675 |
}
|
3676 |
],
|
3677 |
"logging_steps": 1,
|
@@ -3686,12 +4876,12 @@
|
|
3686 |
"should_evaluate": false,
|
3687 |
"should_log": false,
|
3688 |
"should_save": true,
|
3689 |
-
"should_training_stop":
|
3690 |
},
|
3691 |
"attributes": {}
|
3692 |
}
|
3693 |
},
|
3694 |
-
"total_flos":
|
3695 |
"train_batch_size": 2,
|
3696 |
"trial_name": null,
|
3697 |
"trial_params": null
|
|
|
1 |
{
|
2 |
"best_metric": null,
|
3 |
"best_model_checkpoint": null,
|
4 |
+
"epoch": 1.0,
|
5 |
"eval_steps": 173,
|
6 |
+
"global_step": 689,
|
7 |
"is_hyper_param_search": false,
|
8 |
"is_local_process_zero": true,
|
9 |
"is_world_process_zero": true,
|
|
|
3672 |
"eval_samples_per_second": 8.964,
|
3673 |
"eval_steps_per_second": 4.498,
|
3674 |
"step": 519
|
3675 |
+
},
|
3676 |
+
{
|
3677 |
+
"epoch": 0.7547169811320755,
|
3678 |
+
"grad_norm": 0.24123501777648926,
|
3679 |
+
"learning_rate": 2.9044375337725593e-05,
|
3680 |
+
"loss": 0.4715,
|
3681 |
+
"step": 520
|
3682 |
+
},
|
3683 |
+
{
|
3684 |
+
"epoch": 0.7561683599419449,
|
3685 |
+
"grad_norm": 0.4682220220565796,
|
3686 |
+
"learning_rate": 2.8719109545317103e-05,
|
3687 |
+
"loss": 0.8464,
|
3688 |
+
"step": 521
|
3689 |
+
},
|
3690 |
+
{
|
3691 |
+
"epoch": 0.7576197387518142,
|
3692 |
+
"grad_norm": 0.3238670527935028,
|
3693 |
+
"learning_rate": 2.839536967562504e-05,
|
3694 |
+
"loss": 0.34,
|
3695 |
+
"step": 522
|
3696 |
+
},
|
3697 |
+
{
|
3698 |
+
"epoch": 0.7590711175616836,
|
3699 |
+
"grad_norm": 0.46153318881988525,
|
3700 |
+
"learning_rate": 2.807316265900649e-05,
|
3701 |
+
"loss": 1.183,
|
3702 |
+
"step": 523
|
3703 |
+
},
|
3704 |
+
{
|
3705 |
+
"epoch": 0.760522496371553,
|
3706 |
+
"grad_norm": 0.42253023386001587,
|
3707 |
+
"learning_rate": 2.7752495393004475e-05,
|
3708 |
+
"loss": 0.5708,
|
3709 |
+
"step": 524
|
3710 |
+
},
|
3711 |
+
{
|
3712 |
+
"epoch": 0.7619738751814223,
|
3713 |
+
"grad_norm": 0.4283324182033539,
|
3714 |
+
"learning_rate": 2.7433374742200192e-05,
|
3715 |
+
"loss": 1.2854,
|
3716 |
+
"step": 525
|
3717 |
+
},
|
3718 |
+
{
|
3719 |
+
"epoch": 0.7634252539912917,
|
3720 |
+
"grad_norm": 0.33337557315826416,
|
3721 |
+
"learning_rate": 2.7115807538066297e-05,
|
3722 |
+
"loss": 0.6406,
|
3723 |
+
"step": 526
|
3724 |
+
},
|
3725 |
+
{
|
3726 |
+
"epoch": 0.7648766328011611,
|
3727 |
+
"grad_norm": 0.33540481328964233,
|
3728 |
+
"learning_rate": 2.679980057882049e-05,
|
3729 |
+
"loss": 0.5988,
|
3730 |
+
"step": 527
|
3731 |
+
},
|
3732 |
+
{
|
3733 |
+
"epoch": 0.7663280116110305,
|
3734 |
+
"grad_norm": 0.5348614454269409,
|
3735 |
+
"learning_rate": 2.6485360629279987e-05,
|
3736 |
+
"loss": 1.6389,
|
3737 |
+
"step": 528
|
3738 |
+
},
|
3739 |
+
{
|
3740 |
+
"epoch": 0.7677793904208998,
|
3741 |
+
"grad_norm": 0.36704495549201965,
|
3742 |
+
"learning_rate": 2.6172494420716797e-05,
|
3743 |
+
"loss": 1.0625,
|
3744 |
+
"step": 529
|
3745 |
+
},
|
3746 |
+
{
|
3747 |
+
"epoch": 0.7692307692307693,
|
3748 |
+
"grad_norm": 0.3153030276298523,
|
3749 |
+
"learning_rate": 2.5861208650713588e-05,
|
3750 |
+
"loss": 0.3844,
|
3751 |
+
"step": 530
|
3752 |
+
},
|
3753 |
+
{
|
3754 |
+
"epoch": 0.7706821480406386,
|
3755 |
+
"grad_norm": 0.29406848549842834,
|
3756 |
+
"learning_rate": 2.555150998302026e-05,
|
3757 |
+
"loss": 0.575,
|
3758 |
+
"step": 531
|
3759 |
+
},
|
3760 |
+
{
|
3761 |
+
"epoch": 0.772133526850508,
|
3762 |
+
"grad_norm": 0.35533541440963745,
|
3763 |
+
"learning_rate": 2.524340504741135e-05,
|
3764 |
+
"loss": 0.9053,
|
3765 |
+
"step": 532
|
3766 |
+
},
|
3767 |
+
{
|
3768 |
+
"epoch": 0.7735849056603774,
|
3769 |
+
"grad_norm": 0.35642877221107483,
|
3770 |
+
"learning_rate": 2.4936900439544107e-05,
|
3771 |
+
"loss": 1.6089,
|
3772 |
+
"step": 533
|
3773 |
+
},
|
3774 |
+
{
|
3775 |
+
"epoch": 0.7750362844702468,
|
3776 |
+
"grad_norm": 0.5626897811889648,
|
3777 |
+
"learning_rate": 2.4632002720817228e-05,
|
3778 |
+
"loss": 1.136,
|
3779 |
+
"step": 534
|
3780 |
+
},
|
3781 |
+
{
|
3782 |
+
"epoch": 0.7764876632801161,
|
3783 |
+
"grad_norm": 0.3613468408584595,
|
3784 |
+
"learning_rate": 2.432871841823047e-05,
|
3785 |
+
"loss": 1.0527,
|
3786 |
+
"step": 535
|
3787 |
+
},
|
3788 |
+
{
|
3789 |
+
"epoch": 0.7779390420899854,
|
3790 |
+
"grad_norm": 0.5131534934043884,
|
3791 |
+
"learning_rate": 2.4027054024244956e-05,
|
3792 |
+
"loss": 1.0593,
|
3793 |
+
"step": 536
|
3794 |
+
},
|
3795 |
+
{
|
3796 |
+
"epoch": 0.7793904208998549,
|
3797 |
+
"grad_norm": 0.2713516652584076,
|
3798 |
+
"learning_rate": 2.372701599664404e-05,
|
3799 |
+
"loss": 0.3204,
|
3800 |
+
"step": 537
|
3801 |
+
},
|
3802 |
+
{
|
3803 |
+
"epoch": 0.7808417997097242,
|
3804 |
+
"grad_norm": 0.3534812033176422,
|
3805 |
+
"learning_rate": 2.342861075839523e-05,
|
3806 |
+
"loss": 0.6906,
|
3807 |
+
"step": 538
|
3808 |
+
},
|
3809 |
+
{
|
3810 |
+
"epoch": 0.7822931785195936,
|
3811 |
+
"grad_norm": 0.2669237554073334,
|
3812 |
+
"learning_rate": 2.313184469751265e-05,
|
3813 |
+
"loss": 0.4097,
|
3814 |
+
"step": 539
|
3815 |
+
},
|
3816 |
+
{
|
3817 |
+
"epoch": 0.783744557329463,
|
3818 |
+
"grad_norm": 0.26740333437919617,
|
3819 |
+
"learning_rate": 2.283672416692013e-05,
|
3820 |
+
"loss": 0.2641,
|
3821 |
+
"step": 540
|
3822 |
+
},
|
3823 |
+
{
|
3824 |
+
"epoch": 0.7851959361393324,
|
3825 |
+
"grad_norm": 0.4208941161632538,
|
3826 |
+
"learning_rate": 2.2543255484315484e-05,
|
3827 |
+
"loss": 0.9359,
|
3828 |
+
"step": 541
|
3829 |
+
},
|
3830 |
+
{
|
3831 |
+
"epoch": 0.7866473149492017,
|
3832 |
+
"grad_norm": 0.32194021344184875,
|
3833 |
+
"learning_rate": 2.2251444932035094e-05,
|
3834 |
+
"loss": 0.6239,
|
3835 |
+
"step": 542
|
3836 |
+
},
|
3837 |
+
{
|
3838 |
+
"epoch": 0.7880986937590712,
|
3839 |
+
"grad_norm": 0.4907798767089844,
|
3840 |
+
"learning_rate": 2.1961298756919335e-05,
|
3841 |
+
"loss": 1.76,
|
3842 |
+
"step": 543
|
3843 |
+
},
|
3844 |
+
{
|
3845 |
+
"epoch": 0.7895500725689405,
|
3846 |
+
"grad_norm": 0.45274341106414795,
|
3847 |
+
"learning_rate": 2.16728231701791e-05,
|
3848 |
+
"loss": 1.3965,
|
3849 |
+
"step": 544
|
3850 |
+
},
|
3851 |
+
{
|
3852 |
+
"epoch": 0.7910014513788098,
|
3853 |
+
"grad_norm": 0.3342875838279724,
|
3854 |
+
"learning_rate": 2.138602434726269e-05,
|
3855 |
+
"loss": 0.2511,
|
3856 |
+
"step": 545
|
3857 |
+
},
|
3858 |
+
{
|
3859 |
+
"epoch": 0.7924528301886793,
|
3860 |
+
"grad_norm": 0.37094488739967346,
|
3861 |
+
"learning_rate": 2.11009084277235e-05,
|
3862 |
+
"loss": 1.1932,
|
3863 |
+
"step": 546
|
3864 |
+
},
|
3865 |
+
{
|
3866 |
+
"epoch": 0.7939042089985486,
|
3867 |
+
"grad_norm": 0.36341187357902527,
|
3868 |
+
"learning_rate": 2.0817481515088833e-05,
|
3869 |
+
"loss": 0.7164,
|
3870 |
+
"step": 547
|
3871 |
+
},
|
3872 |
+
{
|
3873 |
+
"epoch": 0.795355587808418,
|
3874 |
+
"grad_norm": 0.4300606846809387,
|
3875 |
+
"learning_rate": 2.053574967672912e-05,
|
3876 |
+
"loss": 0.9558,
|
3877 |
+
"step": 548
|
3878 |
+
},
|
3879 |
+
{
|
3880 |
+
"epoch": 0.7968069666182874,
|
3881 |
+
"grad_norm": 0.4252452850341797,
|
3882 |
+
"learning_rate": 2.025571894372794e-05,
|
3883 |
+
"loss": 1.1288,
|
3884 |
+
"step": 549
|
3885 |
+
},
|
3886 |
+
{
|
3887 |
+
"epoch": 0.7982583454281568,
|
3888 |
+
"grad_norm": 0.5417127013206482,
|
3889 |
+
"learning_rate": 1.9977395310753087e-05,
|
3890 |
+
"loss": 2.1247,
|
3891 |
+
"step": 550
|
3892 |
+
},
|
3893 |
+
{
|
3894 |
+
"epoch": 0.7997097242380261,
|
3895 |
+
"grad_norm": 0.3363449275493622,
|
3896 |
+
"learning_rate": 1.970078473592817e-05,
|
3897 |
+
"loss": 0.5331,
|
3898 |
+
"step": 551
|
3899 |
+
},
|
3900 |
+
{
|
3901 |
+
"epoch": 0.8011611030478955,
|
3902 |
+
"grad_norm": 0.2750878632068634,
|
3903 |
+
"learning_rate": 1.9425893140704943e-05,
|
3904 |
+
"loss": 0.3197,
|
3905 |
+
"step": 552
|
3906 |
+
},
|
3907 |
+
{
|
3908 |
+
"epoch": 0.8026124818577649,
|
3909 |
+
"grad_norm": 0.2823106050491333,
|
3910 |
+
"learning_rate": 1.915272640973683e-05,
|
3911 |
+
"loss": 0.6639,
|
3912 |
+
"step": 553
|
3913 |
+
},
|
3914 |
+
{
|
3915 |
+
"epoch": 0.8040638606676342,
|
3916 |
+
"grad_norm": 0.41894766688346863,
|
3917 |
+
"learning_rate": 1.888129039075268e-05,
|
3918 |
+
"loss": 1.0442,
|
3919 |
+
"step": 554
|
3920 |
+
},
|
3921 |
+
{
|
3922 |
+
"epoch": 0.8055152394775036,
|
3923 |
+
"grad_norm": 0.3718733489513397,
|
3924 |
+
"learning_rate": 1.8611590894431653e-05,
|
3925 |
+
"loss": 0.7986,
|
3926 |
+
"step": 555
|
3927 |
+
},
|
3928 |
+
{
|
3929 |
+
"epoch": 0.806966618287373,
|
3930 |
+
"grad_norm": 0.3276025950908661,
|
3931 |
+
"learning_rate": 1.8343633694278895e-05,
|
3932 |
+
"loss": 0.7076,
|
3933 |
+
"step": 556
|
3934 |
+
},
|
3935 |
+
{
|
3936 |
+
"epoch": 0.8084179970972424,
|
3937 |
+
"grad_norm": 0.3458663523197174,
|
3938 |
+
"learning_rate": 1.807742452650196e-05,
|
3939 |
+
"loss": 0.8522,
|
3940 |
+
"step": 557
|
3941 |
+
},
|
3942 |
+
{
|
3943 |
+
"epoch": 0.8098693759071117,
|
3944 |
+
"grad_norm": 0.41551870107650757,
|
3945 |
+
"learning_rate": 1.7812969089887864e-05,
|
3946 |
+
"loss": 0.7282,
|
3947 |
+
"step": 558
|
3948 |
+
},
|
3949 |
+
{
|
3950 |
+
"epoch": 0.8113207547169812,
|
3951 |
+
"grad_norm": 0.36481913924217224,
|
3952 |
+
"learning_rate": 1.7550273045681242e-05,
|
3953 |
+
"loss": 0.5842,
|
3954 |
+
"step": 559
|
3955 |
+
},
|
3956 |
+
{
|
3957 |
+
"epoch": 0.8127721335268505,
|
3958 |
+
"grad_norm": 0.31234338879585266,
|
3959 |
+
"learning_rate": 1.7289342017463138e-05,
|
3960 |
+
"loss": 1.0055,
|
3961 |
+
"step": 560
|
3962 |
+
},
|
3963 |
+
{
|
3964 |
+
"epoch": 0.8142235123367199,
|
3965 |
+
"grad_norm": 0.29913321137428284,
|
3966 |
+
"learning_rate": 1.7030181591030493e-05,
|
3967 |
+
"loss": 0.4555,
|
3968 |
+
"step": 561
|
3969 |
+
},
|
3970 |
+
{
|
3971 |
+
"epoch": 0.8156748911465893,
|
3972 |
+
"grad_norm": 0.2602443993091583,
|
3973 |
+
"learning_rate": 1.6772797314276713e-05,
|
3974 |
+
"loss": 0.5486,
|
3975 |
+
"step": 562
|
3976 |
+
},
|
3977 |
+
{
|
3978 |
+
"epoch": 0.8171262699564587,
|
3979 |
+
"grad_norm": 0.43661990761756897,
|
3980 |
+
"learning_rate": 1.65171946970729e-05,
|
3981 |
+
"loss": 1.7129,
|
3982 |
+
"step": 563
|
3983 |
+
},
|
3984 |
+
{
|
3985 |
+
"epoch": 0.818577648766328,
|
3986 |
+
"grad_norm": 0.658896267414093,
|
3987 |
+
"learning_rate": 1.6263379211149732e-05,
|
3988 |
+
"loss": 2.026,
|
3989 |
+
"step": 564
|
3990 |
+
},
|
3991 |
+
{
|
3992 |
+
"epoch": 0.8200290275761973,
|
3993 |
+
"grad_norm": 0.48423880338668823,
|
3994 |
+
"learning_rate": 1.601135628998056e-05,
|
3995 |
+
"loss": 1.1105,
|
3996 |
+
"step": 565
|
3997 |
+
},
|
3998 |
+
{
|
3999 |
+
"epoch": 0.8214804063860668,
|
4000 |
+
"grad_norm": 0.3823278844356537,
|
4001 |
+
"learning_rate": 1.5761131328664947e-05,
|
4002 |
+
"loss": 0.9354,
|
4003 |
+
"step": 566
|
4004 |
+
},
|
4005 |
+
{
|
4006 |
+
"epoch": 0.8229317851959361,
|
4007 |
+
"grad_norm": 0.38709479570388794,
|
4008 |
+
"learning_rate": 1.5512709683813164e-05,
|
4009 |
+
"loss": 1.7732,
|
4010 |
+
"step": 567
|
4011 |
+
},
|
4012 |
+
{
|
4013 |
+
"epoch": 0.8243831640058055,
|
4014 |
+
"grad_norm": 0.4736160933971405,
|
4015 |
+
"learning_rate": 1.526609667343163e-05,
|
4016 |
+
"loss": 1.0733,
|
4017 |
+
"step": 568
|
4018 |
+
},
|
4019 |
+
{
|
4020 |
+
"epoch": 0.8258345428156749,
|
4021 |
+
"grad_norm": 0.4745509922504425,
|
4022 |
+
"learning_rate": 1.5021297576809001e-05,
|
4023 |
+
"loss": 1.4294,
|
4024 |
+
"step": 569
|
4025 |
+
},
|
4026 |
+
{
|
4027 |
+
"epoch": 0.8272859216255443,
|
4028 |
+
"grad_norm": 0.28139880299568176,
|
4029 |
+
"learning_rate": 1.4778317634403083e-05,
|
4030 |
+
"loss": 0.2901,
|
4031 |
+
"step": 570
|
4032 |
+
},
|
4033 |
+
{
|
4034 |
+
"epoch": 0.8287373004354136,
|
4035 |
+
"grad_norm": 0.39859816431999207,
|
4036 |
+
"learning_rate": 1.4537162047728804e-05,
|
4037 |
+
"loss": 0.7626,
|
4038 |
+
"step": 571
|
4039 |
+
},
|
4040 |
+
{
|
4041 |
+
"epoch": 0.8301886792452831,
|
4042 |
+
"grad_norm": 0.8511147499084473,
|
4043 |
+
"learning_rate": 1.4297835979246777e-05,
|
4044 |
+
"loss": 1.482,
|
4045 |
+
"step": 572
|
4046 |
+
},
|
4047 |
+
{
|
4048 |
+
"epoch": 0.8316400580551524,
|
4049 |
+
"grad_norm": 0.4506050646305084,
|
4050 |
+
"learning_rate": 1.4060344552252702e-05,
|
4051 |
+
"loss": 1.2382,
|
4052 |
+
"step": 573
|
4053 |
+
},
|
4054 |
+
{
|
4055 |
+
"epoch": 0.8330914368650217,
|
4056 |
+
"grad_norm": 0.22067701816558838,
|
4057 |
+
"learning_rate": 1.3824692850767929e-05,
|
4058 |
+
"loss": 0.571,
|
4059 |
+
"step": 574
|
4060 |
+
},
|
4061 |
+
{
|
4062 |
+
"epoch": 0.8345428156748912,
|
4063 |
+
"grad_norm": 0.3021042048931122,
|
4064 |
+
"learning_rate": 1.3590885919430384e-05,
|
4065 |
+
"loss": 0.522,
|
4066 |
+
"step": 575
|
4067 |
+
},
|
4068 |
+
{
|
4069 |
+
"epoch": 0.8359941944847605,
|
4070 |
+
"grad_norm": 0.49198654294013977,
|
4071 |
+
"learning_rate": 1.3358928763386646e-05,
|
4072 |
+
"loss": 0.9773,
|
4073 |
+
"step": 576
|
4074 |
+
},
|
4075 |
+
{
|
4076 |
+
"epoch": 0.8374455732946299,
|
4077 |
+
"grad_norm": 0.35230767726898193,
|
4078 |
+
"learning_rate": 1.3128826348184887e-05,
|
4079 |
+
"loss": 0.6312,
|
4080 |
+
"step": 577
|
4081 |
+
},
|
4082 |
+
{
|
4083 |
+
"epoch": 0.8388969521044993,
|
4084 |
+
"grad_norm": 0.3485874831676483,
|
4085 |
+
"learning_rate": 1.2900583599668515e-05,
|
4086 |
+
"loss": 0.6788,
|
4087 |
+
"step": 578
|
4088 |
+
},
|
4089 |
+
{
|
4090 |
+
"epoch": 0.8403483309143687,
|
4091 |
+
"grad_norm": 0.3603363335132599,
|
4092 |
+
"learning_rate": 1.2674205403870642e-05,
|
4093 |
+
"loss": 0.7338,
|
4094 |
+
"step": 579
|
4095 |
+
},
|
4096 |
+
{
|
4097 |
+
"epoch": 0.841799709724238,
|
4098 |
+
"grad_norm": 0.25826606154441833,
|
4099 |
+
"learning_rate": 1.2449696606909655e-05,
|
4100 |
+
"loss": 0.4404,
|
4101 |
+
"step": 580
|
4102 |
+
},
|
4103 |
+
{
|
4104 |
+
"epoch": 0.8432510885341074,
|
4105 |
+
"grad_norm": 0.3804086744785309,
|
4106 |
+
"learning_rate": 1.222706201488536e-05,
|
4107 |
+
"loss": 0.4955,
|
4108 |
+
"step": 581
|
4109 |
+
},
|
4110 |
+
{
|
4111 |
+
"epoch": 0.8447024673439768,
|
4112 |
+
"grad_norm": 0.47231999039649963,
|
4113 |
+
"learning_rate": 1.200630639377609e-05,
|
4114 |
+
"loss": 1.6225,
|
4115 |
+
"step": 582
|
4116 |
+
},
|
4117 |
+
{
|
4118 |
+
"epoch": 0.8461538461538461,
|
4119 |
+
"grad_norm": 0.44939887523651123,
|
4120 |
+
"learning_rate": 1.1787434469336766e-05,
|
4121 |
+
"loss": 1.252,
|
4122 |
+
"step": 583
|
4123 |
+
},
|
4124 |
+
{
|
4125 |
+
"epoch": 0.8476052249637155,
|
4126 |
+
"grad_norm": 0.3787010610103607,
|
4127 |
+
"learning_rate": 1.1570450926997655e-05,
|
4128 |
+
"loss": 0.544,
|
4129 |
+
"step": 584
|
4130 |
+
},
|
4131 |
+
{
|
4132 |
+
"epoch": 0.8490566037735849,
|
4133 |
+
"grad_norm": 0.4327794909477234,
|
4134 |
+
"learning_rate": 1.135536041176406e-05,
|
4135 |
+
"loss": 0.3797,
|
4136 |
+
"step": 585
|
4137 |
+
},
|
4138 |
+
{
|
4139 |
+
"epoch": 0.8505079825834543,
|
4140 |
+
"grad_norm": 0.35290834307670593,
|
4141 |
+
"learning_rate": 1.1142167528116965e-05,
|
4142 |
+
"loss": 0.8393,
|
4143 |
+
"step": 586
|
4144 |
+
},
|
4145 |
+
{
|
4146 |
+
"epoch": 0.8519593613933236,
|
4147 |
+
"grad_norm": 0.24693018198013306,
|
4148 |
+
"learning_rate": 1.0930876839914416e-05,
|
4149 |
+
"loss": 0.1381,
|
4150 |
+
"step": 587
|
4151 |
+
},
|
4152 |
+
{
|
4153 |
+
"epoch": 0.8534107402031931,
|
4154 |
+
"grad_norm": 0.3256937265396118,
|
4155 |
+
"learning_rate": 1.072149287029377e-05,
|
4156 |
+
"loss": 0.5073,
|
4157 |
+
"step": 588
|
4158 |
+
},
|
4159 |
+
{
|
4160 |
+
"epoch": 0.8548621190130624,
|
4161 |
+
"grad_norm": 0.32373976707458496,
|
4162 |
+
"learning_rate": 1.0514020101574972e-05,
|
4163 |
+
"loss": 0.4282,
|
4164 |
+
"step": 589
|
4165 |
+
},
|
4166 |
+
{
|
4167 |
+
"epoch": 0.8563134978229318,
|
4168 |
+
"grad_norm": 0.3258521854877472,
|
4169 |
+
"learning_rate": 1.0308462975164545e-05,
|
4170 |
+
"loss": 0.348,
|
4171 |
+
"step": 590
|
4172 |
+
},
|
4173 |
+
{
|
4174 |
+
"epoch": 0.8577648766328012,
|
4175 |
+
"grad_norm": 0.3401307165622711,
|
4176 |
+
"learning_rate": 1.010482589146048e-05,
|
4177 |
+
"loss": 1.1884,
|
4178 |
+
"step": 591
|
4179 |
+
},
|
4180 |
+
{
|
4181 |
+
"epoch": 0.8592162554426706,
|
4182 |
+
"grad_norm": 0.39415818452835083,
|
4183 |
+
"learning_rate": 9.903113209758096e-06,
|
4184 |
+
"loss": 1.2618,
|
4185 |
+
"step": 592
|
4186 |
+
},
|
4187 |
+
{
|
4188 |
+
"epoch": 0.8606676342525399,
|
4189 |
+
"grad_norm": 0.3865334689617157,
|
4190 |
+
"learning_rate": 9.703329248156711e-06,
|
4191 |
+
"loss": 0.9641,
|
4192 |
+
"step": 593
|
4193 |
+
},
|
4194 |
+
{
|
4195 |
+
"epoch": 0.8621190130624092,
|
4196 |
+
"grad_norm": 0.5107492208480835,
|
4197 |
+
"learning_rate": 9.505478283467128e-06,
|
4198 |
+
"loss": 0.6755,
|
4199 |
+
"step": 594
|
4200 |
+
},
|
4201 |
+
{
|
4202 |
+
"epoch": 0.8635703918722787,
|
4203 |
+
"grad_norm": 0.3883417248725891,
|
4204 |
+
"learning_rate": 9.309564551120254e-06,
|
4205 |
+
"loss": 1.4845,
|
4206 |
+
"step": 595
|
4207 |
+
},
|
4208 |
+
{
|
4209 |
+
"epoch": 0.865021770682148,
|
4210 |
+
"grad_norm": 0.5274085998535156,
|
4211 |
+
"learning_rate": 9.115592245076177e-06,
|
4212 |
+
"loss": 1.4281,
|
4213 |
+
"step": 596
|
4214 |
+
},
|
4215 |
+
{
|
4216 |
+
"epoch": 0.8664731494920174,
|
4217 |
+
"grad_norm": 0.4148600399494171,
|
4218 |
+
"learning_rate": 8.923565517734633e-06,
|
4219 |
+
"loss": 1.4102,
|
4220 |
+
"step": 597
|
4221 |
+
},
|
4222 |
+
{
|
4223 |
+
"epoch": 0.8679245283018868,
|
4224 |
+
"grad_norm": 0.3156782388687134,
|
4225 |
+
"learning_rate": 8.733488479845997e-06,
|
4226 |
+
"loss": 0.7565,
|
4227 |
+
"step": 598
|
4228 |
+
},
|
4229 |
+
{
|
4230 |
+
"epoch": 0.8693759071117562,
|
4231 |
+
"grad_norm": 0.41858044266700745,
|
4232 |
+
"learning_rate": 8.545365200423217e-06,
|
4233 |
+
"loss": 1.3413,
|
4234 |
+
"step": 599
|
4235 |
+
},
|
4236 |
+
{
|
4237 |
+
"epoch": 0.8708272859216255,
|
4238 |
+
"grad_norm": 0.3816291391849518,
|
4239 |
+
"learning_rate": 8.359199706654853e-06,
|
4240 |
+
"loss": 0.9125,
|
4241 |
+
"step": 600
|
4242 |
+
},
|
4243 |
+
{
|
4244 |
+
"epoch": 0.872278664731495,
|
4245 |
+
"grad_norm": 0.3754681944847107,
|
4246 |
+
"learning_rate": 8.1749959838188e-06,
|
4247 |
+
"loss": 0.5213,
|
4248 |
+
"step": 601
|
4249 |
+
},
|
4250 |
+
{
|
4251 |
+
"epoch": 0.8737300435413643,
|
4252 |
+
"grad_norm": 0.32858067750930786,
|
4253 |
+
"learning_rate": 7.992757975196974e-06,
|
4254 |
+
"loss": 0.9199,
|
4255 |
+
"step": 602
|
4256 |
+
},
|
4257 |
+
{
|
4258 |
+
"epoch": 0.8751814223512336,
|
4259 |
+
"grad_norm": 0.4272098243236542,
|
4260 |
+
"learning_rate": 7.812489581990844e-06,
|
4261 |
+
"loss": 1.4278,
|
4262 |
+
"step": 603
|
4263 |
+
},
|
4264 |
+
{
|
4265 |
+
"epoch": 0.8766328011611031,
|
4266 |
+
"grad_norm": 0.35721269249916077,
|
4267 |
+
"learning_rate": 7.634194663238015e-06,
|
4268 |
+
"loss": 0.4783,
|
4269 |
+
"step": 604
|
4270 |
+
},
|
4271 |
+
{
|
4272 |
+
"epoch": 0.8780841799709724,
|
4273 |
+
"grad_norm": 0.38133150339126587,
|
4274 |
+
"learning_rate": 7.457877035729588e-06,
|
4275 |
+
"loss": 0.6231,
|
4276 |
+
"step": 605
|
4277 |
+
},
|
4278 |
+
{
|
4279 |
+
"epoch": 0.8795355587808418,
|
4280 |
+
"grad_norm": 0.3652817904949188,
|
4281 |
+
"learning_rate": 7.283540473928385e-06,
|
4282 |
+
"loss": 0.7934,
|
4283 |
+
"step": 606
|
4284 |
+
},
|
4285 |
+
{
|
4286 |
+
"epoch": 0.8809869375907112,
|
4287 |
+
"grad_norm": 0.3659121096134186,
|
4288 |
+
"learning_rate": 7.11118870988825e-06,
|
4289 |
+
"loss": 0.9149,
|
4290 |
+
"step": 607
|
4291 |
+
},
|
4292 |
+
{
|
4293 |
+
"epoch": 0.8824383164005806,
|
4294 |
+
"grad_norm": 0.4516978859901428,
|
4295 |
+
"learning_rate": 6.940825433174103e-06,
|
4296 |
+
"loss": 1.2015,
|
4297 |
+
"step": 608
|
4298 |
+
},
|
4299 |
+
{
|
4300 |
+
"epoch": 0.8838896952104499,
|
4301 |
+
"grad_norm": 0.3403629660606384,
|
4302 |
+
"learning_rate": 6.772454290782926e-06,
|
4303 |
+
"loss": 0.552,
|
4304 |
+
"step": 609
|
4305 |
+
},
|
4306 |
+
{
|
4307 |
+
"epoch": 0.8853410740203193,
|
4308 |
+
"grad_norm": 0.24706141650676727,
|
4309 |
+
"learning_rate": 6.606078887065748e-06,
|
4310 |
+
"loss": 0.2273,
|
4311 |
+
"step": 610
|
4312 |
+
},
|
4313 |
+
{
|
4314 |
+
"epoch": 0.8867924528301887,
|
4315 |
+
"grad_norm": 0.3536119759082794,
|
4316 |
+
"learning_rate": 6.441702783650494e-06,
|
4317 |
+
"loss": 0.7869,
|
4318 |
+
"step": 611
|
4319 |
+
},
|
4320 |
+
{
|
4321 |
+
"epoch": 0.888243831640058,
|
4322 |
+
"grad_norm": 0.43990838527679443,
|
4323 |
+
"learning_rate": 6.2793294993656494e-06,
|
4324 |
+
"loss": 0.745,
|
4325 |
+
"step": 612
|
4326 |
+
},
|
4327 |
+
{
|
4328 |
+
"epoch": 0.8896952104499274,
|
4329 |
+
"grad_norm": 0.560270369052887,
|
4330 |
+
"learning_rate": 6.118962510165038e-06,
|
4331 |
+
"loss": 1.311,
|
4332 |
+
"step": 613
|
4333 |
+
},
|
4334 |
+
{
|
4335 |
+
"epoch": 0.8911465892597968,
|
4336 |
+
"grad_norm": 0.3678494691848755,
|
4337 |
+
"learning_rate": 5.960605249053386e-06,
|
4338 |
+
"loss": 1.182,
|
4339 |
+
"step": 614
|
4340 |
+
},
|
4341 |
+
{
|
4342 |
+
"epoch": 0.8925979680696662,
|
4343 |
+
"grad_norm": 0.36771297454833984,
|
4344 |
+
"learning_rate": 5.8042611060127315e-06,
|
4345 |
+
"loss": 0.7735,
|
4346 |
+
"step": 615
|
4347 |
+
},
|
4348 |
+
{
|
4349 |
+
"epoch": 0.8940493468795355,
|
4350 |
+
"grad_norm": 0.43662601709365845,
|
4351 |
+
"learning_rate": 5.649933427930043e-06,
|
4352 |
+
"loss": 0.9749,
|
4353 |
+
"step": 616
|
4354 |
+
},
|
4355 |
+
{
|
4356 |
+
"epoch": 0.895500725689405,
|
4357 |
+
"grad_norm": 0.4699442386627197,
|
4358 |
+
"learning_rate": 5.497625518525373e-06,
|
4359 |
+
"loss": 0.9527,
|
4360 |
+
"step": 617
|
4361 |
+
},
|
4362 |
+
{
|
4363 |
+
"epoch": 0.8969521044992743,
|
4364 |
+
"grad_norm": 0.3796102702617645,
|
4365 |
+
"learning_rate": 5.347340638281273e-06,
|
4366 |
+
"loss": 0.7223,
|
4367 |
+
"step": 618
|
4368 |
+
},
|
4369 |
+
{
|
4370 |
+
"epoch": 0.8984034833091437,
|
4371 |
+
"grad_norm": 0.3132692277431488,
|
4372 |
+
"learning_rate": 5.199082004372957e-06,
|
4373 |
+
"loss": 0.3637,
|
4374 |
+
"step": 619
|
4375 |
+
},
|
4376 |
+
{
|
4377 |
+
"epoch": 0.8998548621190131,
|
4378 |
+
"grad_norm": 0.46878132224082947,
|
4379 |
+
"learning_rate": 5.052852790599383e-06,
|
4380 |
+
"loss": 1.131,
|
4381 |
+
"step": 620
|
4382 |
+
},
|
4383 |
+
{
|
4384 |
+
"epoch": 0.9013062409288825,
|
4385 |
+
"grad_norm": 0.359684556722641,
|
4386 |
+
"learning_rate": 4.908656127315359e-06,
|
4387 |
+
"loss": 0.6985,
|
4388 |
+
"step": 621
|
4389 |
+
},
|
4390 |
+
{
|
4391 |
+
"epoch": 0.9027576197387518,
|
4392 |
+
"grad_norm": 0.32585737109184265,
|
4393 |
+
"learning_rate": 4.7664951013645875e-06,
|
4394 |
+
"loss": 0.6099,
|
4395 |
+
"step": 622
|
4396 |
+
},
|
4397 |
+
{
|
4398 |
+
"epoch": 0.9042089985486212,
|
4399 |
+
"grad_norm": 0.4055173397064209,
|
4400 |
+
"learning_rate": 4.626372756013409e-06,
|
4401 |
+
"loss": 0.6696,
|
4402 |
+
"step": 623
|
4403 |
+
},
|
4404 |
+
{
|
4405 |
+
"epoch": 0.9056603773584906,
|
4406 |
+
"grad_norm": 0.3793414235115051,
|
4407 |
+
"learning_rate": 4.48829209088586e-06,
|
4408 |
+
"loss": 1.2006,
|
4409 |
+
"step": 624
|
4410 |
+
},
|
4411 |
+
{
|
4412 |
+
"epoch": 0.9071117561683599,
|
4413 |
+
"grad_norm": 0.41631418466567993,
|
4414 |
+
"learning_rate": 4.352256061899329e-06,
|
4415 |
+
"loss": 0.6615,
|
4416 |
+
"step": 625
|
4417 |
+
},
|
4418 |
+
{
|
4419 |
+
"epoch": 0.9085631349782293,
|
4420 |
+
"grad_norm": 0.28770431876182556,
|
4421 |
+
"learning_rate": 4.2182675812012965e-06,
|
4422 |
+
"loss": 0.4437,
|
4423 |
+
"step": 626
|
4424 |
+
},
|
4425 |
+
{
|
4426 |
+
"epoch": 0.9100145137880987,
|
4427 |
+
"grad_norm": 0.4755733609199524,
|
4428 |
+
"learning_rate": 4.086329517107046e-06,
|
4429 |
+
"loss": 0.9437,
|
4430 |
+
"step": 627
|
4431 |
+
},
|
4432 |
+
{
|
4433 |
+
"epoch": 0.9114658925979681,
|
4434 |
+
"grad_norm": 0.4184425473213196,
|
4435 |
+
"learning_rate": 3.9564446940382084e-06,
|
4436 |
+
"loss": 1.8477,
|
4437 |
+
"step": 628
|
4438 |
+
},
|
4439 |
+
{
|
4440 |
+
"epoch": 0.9129172714078374,
|
4441 |
+
"grad_norm": 0.515277087688446,
|
4442 |
+
"learning_rate": 3.828615892462328e-06,
|
4443 |
+
"loss": 1.0525,
|
4444 |
+
"step": 629
|
4445 |
+
},
|
4446 |
+
{
|
4447 |
+
"epoch": 0.9143686502177069,
|
4448 |
+
"grad_norm": 0.35891175270080566,
|
4449 |
+
"learning_rate": 3.7028458488333138e-06,
|
4450 |
+
"loss": 1.4299,
|
4451 |
+
"step": 630
|
4452 |
+
},
|
4453 |
+
{
|
4454 |
+
"epoch": 0.9158200290275762,
|
4455 |
+
"grad_norm": 0.34797385334968567,
|
4456 |
+
"learning_rate": 3.5791372555328937e-06,
|
4457 |
+
"loss": 0.3932,
|
4458 |
+
"step": 631
|
4459 |
+
},
|
4460 |
+
{
|
4461 |
+
"epoch": 0.9172714078374455,
|
4462 |
+
"grad_norm": 0.4482119679450989,
|
4463 |
+
"learning_rate": 3.457492760812975e-06,
|
4464 |
+
"loss": 0.9157,
|
4465 |
+
"step": 632
|
4466 |
+
},
|
4467 |
+
{
|
4468 |
+
"epoch": 0.918722786647315,
|
4469 |
+
"grad_norm": 0.43506234884262085,
|
4470 |
+
"learning_rate": 3.3379149687388867e-06,
|
4471 |
+
"loss": 0.6012,
|
4472 |
+
"step": 633
|
4473 |
+
},
|
4474 |
+
{
|
4475 |
+
"epoch": 0.9201741654571843,
|
4476 |
+
"grad_norm": 0.539933443069458,
|
4477 |
+
"learning_rate": 3.2204064391337273e-06,
|
4478 |
+
"loss": 0.8629,
|
4479 |
+
"step": 634
|
4480 |
+
},
|
4481 |
+
{
|
4482 |
+
"epoch": 0.9216255442670537,
|
4483 |
+
"grad_norm": 0.32083696126937866,
|
4484 |
+
"learning_rate": 3.1049696875235512e-06,
|
4485 |
+
"loss": 0.4713,
|
4486 |
+
"step": 635
|
4487 |
+
},
|
4488 |
+
{
|
4489 |
+
"epoch": 0.9230769230769231,
|
4490 |
+
"grad_norm": 0.5421581864356995,
|
4491 |
+
"learning_rate": 2.9916071850834005e-06,
|
4492 |
+
"loss": 1.2597,
|
4493 |
+
"step": 636
|
4494 |
+
},
|
4495 |
+
{
|
4496 |
+
"epoch": 0.9245283018867925,
|
4497 |
+
"grad_norm": 0.2753216028213501,
|
4498 |
+
"learning_rate": 2.8803213585846035e-06,
|
4499 |
+
"loss": 0.4526,
|
4500 |
+
"step": 637
|
4501 |
+
},
|
4502 |
+
{
|
4503 |
+
"epoch": 0.9259796806966618,
|
4504 |
+
"grad_norm": 0.5353147983551025,
|
4505 |
+
"learning_rate": 2.7711145903426496e-06,
|
4506 |
+
"loss": 1.0714,
|
4507 |
+
"step": 638
|
4508 |
+
},
|
4509 |
+
{
|
4510 |
+
"epoch": 0.9274310595065312,
|
4511 |
+
"grad_norm": 0.26769256591796875,
|
4512 |
+
"learning_rate": 2.663989218166274e-06,
|
4513 |
+
"loss": 1.1174,
|
4514 |
+
"step": 639
|
4515 |
+
},
|
4516 |
+
{
|
4517 |
+
"epoch": 0.9288824383164006,
|
4518 |
+
"grad_norm": 0.2907368838787079,
|
4519 |
+
"learning_rate": 2.5589475353073988e-06,
|
4520 |
+
"loss": 1.2706,
|
4521 |
+
"step": 640
|
4522 |
+
},
|
4523 |
+
{
|
4524 |
+
"epoch": 0.93033381712627,
|
4525 |
+
"grad_norm": 0.49619778990745544,
|
4526 |
+
"learning_rate": 2.455991790412027e-06,
|
4527 |
+
"loss": 1.0534,
|
4528 |
+
"step": 641
|
4529 |
+
},
|
4530 |
+
{
|
4531 |
+
"epoch": 0.9317851959361393,
|
4532 |
+
"grad_norm": 0.32665136456489563,
|
4533 |
+
"learning_rate": 2.355124187472135e-06,
|
4534 |
+
"loss": 0.9813,
|
4535 |
+
"step": 642
|
4536 |
+
},
|
4537 |
+
{
|
4538 |
+
"epoch": 0.9332365747460087,
|
4539 |
+
"grad_norm": 0.24648268520832062,
|
4540 |
+
"learning_rate": 2.2563468857784818e-06,
|
4541 |
+
"loss": 0.3728,
|
4542 |
+
"step": 643
|
4543 |
+
},
|
4544 |
+
{
|
4545 |
+
"epoch": 0.9346879535558781,
|
4546 |
+
"grad_norm": 0.45106983184814453,
|
4547 |
+
"learning_rate": 2.159661999874307e-06,
|
4548 |
+
"loss": 0.6732,
|
4549 |
+
"step": 644
|
4550 |
+
},
|
4551 |
+
{
|
4552 |
+
"epoch": 0.9361393323657474,
|
4553 |
+
"grad_norm": 0.4321921467781067,
|
4554 |
+
"learning_rate": 2.0650715995102267e-06,
|
4555 |
+
"loss": 1.9884,
|
4556 |
+
"step": 645
|
4557 |
+
},
|
4558 |
+
{
|
4559 |
+
"epoch": 0.9375907111756169,
|
4560 |
+
"grad_norm": 0.33670032024383545,
|
4561 |
+
"learning_rate": 1.9725777095997657e-06,
|
4562 |
+
"loss": 0.6079,
|
4563 |
+
"step": 646
|
4564 |
+
},
|
4565 |
+
{
|
4566 |
+
"epoch": 0.9390420899854862,
|
4567 |
+
"grad_norm": 0.324014276266098,
|
4568 |
+
"learning_rate": 1.882182310176095e-06,
|
4569 |
+
"loss": 0.2433,
|
4570 |
+
"step": 647
|
4571 |
+
},
|
4572 |
+
{
|
4573 |
+
"epoch": 0.9404934687953556,
|
4574 |
+
"grad_norm": 0.41200318932533264,
|
4575 |
+
"learning_rate": 1.7938873363496422e-06,
|
4576 |
+
"loss": 1.2887,
|
4577 |
+
"step": 648
|
4578 |
+
},
|
4579 |
+
{
|
4580 |
+
"epoch": 0.941944847605225,
|
4581 |
+
"grad_norm": 0.4197370707988739,
|
4582 |
+
"learning_rate": 1.707694678266636e-06,
|
4583 |
+
"loss": 1.1368,
|
4584 |
+
"step": 649
|
4585 |
+
},
|
4586 |
+
{
|
4587 |
+
"epoch": 0.9433962264150944,
|
4588 |
+
"grad_norm": 0.5854864120483398,
|
4589 |
+
"learning_rate": 1.6236061810686487e-06,
|
4590 |
+
"loss": 0.5592,
|
4591 |
+
"step": 650
|
4592 |
+
},
|
4593 |
+
{
|
4594 |
+
"epoch": 0.9448476052249637,
|
4595 |
+
"grad_norm": 0.426692932844162,
|
4596 |
+
"learning_rate": 1.5416236448531074e-06,
|
4597 |
+
"loss": 1.1019,
|
4598 |
+
"step": 651
|
4599 |
+
},
|
4600 |
+
{
|
4601 |
+
"epoch": 0.9462989840348331,
|
4602 |
+
"grad_norm": 0.4369201362133026,
|
4603 |
+
"learning_rate": 1.461748824634801e-06,
|
4604 |
+
"loss": 1.0856,
|
4605 |
+
"step": 652
|
4606 |
+
},
|
4607 |
+
{
|
4608 |
+
"epoch": 0.9477503628447025,
|
4609 |
+
"grad_norm": 0.3605095148086548,
|
4610 |
+
"learning_rate": 1.383983430308189e-06,
|
4611 |
+
"loss": 1.3698,
|
4612 |
+
"step": 653
|
4613 |
+
},
|
4614 |
+
{
|
4615 |
+
"epoch": 0.9492017416545718,
|
4616 |
+
"grad_norm": 0.28937241435050964,
|
4617 |
+
"learning_rate": 1.30832912661093e-06,
|
4618 |
+
"loss": 0.4791,
|
4619 |
+
"step": 654
|
4620 |
+
},
|
4621 |
+
{
|
4622 |
+
"epoch": 0.9506531204644412,
|
4623 |
+
"grad_norm": 0.38452112674713135,
|
4624 |
+
"learning_rate": 1.2347875330881886e-06,
|
4625 |
+
"loss": 1.3005,
|
4626 |
+
"step": 655
|
4627 |
+
},
|
4628 |
+
{
|
4629 |
+
"epoch": 0.9521044992743106,
|
4630 |
+
"grad_norm": 0.44588330388069153,
|
4631 |
+
"learning_rate": 1.1633602240579522e-06,
|
4632 |
+
"loss": 1.1278,
|
4633 |
+
"step": 656
|
4634 |
+
},
|
4635 |
+
{
|
4636 |
+
"epoch": 0.95355587808418,
|
4637 |
+
"grad_norm": 0.34775638580322266,
|
4638 |
+
"learning_rate": 1.0940487285773459e-06,
|
4639 |
+
"loss": 0.6701,
|
4640 |
+
"step": 657
|
4641 |
+
},
|
4642 |
+
{
|
4643 |
+
"epoch": 0.9550072568940493,
|
4644 |
+
"grad_norm": 0.34269067645072937,
|
4645 |
+
"learning_rate": 1.026854530409882e-06,
|
4646 |
+
"loss": 1.3742,
|
4647 |
+
"step": 658
|
4648 |
+
},
|
4649 |
+
{
|
4650 |
+
"epoch": 0.9564586357039188,
|
4651 |
+
"grad_norm": 0.3833567500114441,
|
4652 |
+
"learning_rate": 9.61779067993729e-07,
|
4653 |
+
"loss": 1.387,
|
4654 |
+
"step": 659
|
4655 |
+
},
|
4656 |
+
{
|
4657 |
+
"epoch": 0.9579100145137881,
|
4658 |
+
"grad_norm": 0.37239694595336914,
|
4659 |
+
"learning_rate": 8.988237344109251e-07,
|
4660 |
+
"loss": 0.8695,
|
4661 |
+
"step": 660
|
4662 |
+
},
|
4663 |
+
{
|
4664 |
+
"epoch": 0.9593613933236574,
|
4665 |
+
"grad_norm": 0.30777066946029663,
|
4666 |
+
"learning_rate": 8.379898773574924e-07,
|
4667 |
+
"loss": 0.7924,
|
4668 |
+
"step": 661
|
4669 |
+
},
|
4670 |
+
{
|
4671 |
+
"epoch": 0.9608127721335269,
|
4672 |
+
"grad_norm": 0.38899528980255127,
|
4673 |
+
"learning_rate": 7.792787991146355e-07,
|
4674 |
+
"loss": 1.4396,
|
4675 |
+
"step": 662
|
4676 |
+
},
|
4677 |
+
{
|
4678 |
+
"epoch": 0.9622641509433962,
|
4679 |
+
"grad_norm": 0.4895128309726715,
|
4680 |
+
"learning_rate": 7.226917565208657e-07,
|
4681 |
+
"loss": 1.3093,
|
4682 |
+
"step": 663
|
4683 |
+
},
|
4684 |
+
{
|
4685 |
+
"epoch": 0.9637155297532656,
|
4686 |
+
"grad_norm": 0.5337472558021545,
|
4687 |
+
"learning_rate": 6.682299609450993e-07,
|
4688 |
+
"loss": 0.8694,
|
4689 |
+
"step": 664
|
4690 |
+
},
|
4691 |
+
{
|
4692 |
+
"epoch": 0.965166908563135,
|
4693 |
+
"grad_norm": 0.32183799147605896,
|
4694 |
+
"learning_rate": 6.158945782606673e-07,
|
4695 |
+
"loss": 0.5548,
|
4696 |
+
"step": 665
|
4697 |
+
},
|
4698 |
+
{
|
4699 |
+
"epoch": 0.9666182873730044,
|
4700 |
+
"grad_norm": 0.3950204849243164,
|
4701 |
+
"learning_rate": 5.65686728820436e-07,
|
4702 |
+
"loss": 0.4285,
|
4703 |
+
"step": 666
|
4704 |
+
},
|
4705 |
+
{
|
4706 |
+
"epoch": 0.9680696661828737,
|
4707 |
+
"grad_norm": 0.4570080041885376,
|
4708 |
+
"learning_rate": 5.176074874327919e-07,
|
4709 |
+
"loss": 1.5729,
|
4710 |
+
"step": 667
|
4711 |
+
},
|
4712 |
+
{
|
4713 |
+
"epoch": 0.969521044992743,
|
4714 |
+
"grad_norm": 0.42208755016326904,
|
4715 |
+
"learning_rate": 4.7165788333860536e-07,
|
4716 |
+
"loss": 0.8863,
|
4717 |
+
"step": 668
|
4718 |
+
},
|
4719 |
+
{
|
4720 |
+
"epoch": 0.9709724238026125,
|
4721 |
+
"grad_norm": 0.29929739236831665,
|
4722 |
+
"learning_rate": 4.278389001892369e-07,
|
4723 |
+
"loss": 0.7099,
|
4724 |
+
"step": 669
|
4725 |
+
},
|
4726 |
+
{
|
4727 |
+
"epoch": 0.9724238026124818,
|
4728 |
+
"grad_norm": 0.46631020307540894,
|
4729 |
+
"learning_rate": 3.8615147602546473e-07,
|
4730 |
+
"loss": 0.4506,
|
4731 |
+
"step": 670
|
4732 |
+
},
|
4733 |
+
{
|
4734 |
+
"epoch": 0.9738751814223512,
|
4735 |
+
"grad_norm": 0.4783049523830414,
|
4736 |
+
"learning_rate": 3.4659650325740147e-07,
|
4737 |
+
"loss": 1.1751,
|
4738 |
+
"step": 671
|
4739 |
+
},
|
4740 |
+
{
|
4741 |
+
"epoch": 0.9753265602322206,
|
4742 |
+
"grad_norm": 0.3503381907939911,
|
4743 |
+
"learning_rate": 3.091748286453866e-07,
|
4744 |
+
"loss": 0.9895,
|
4745 |
+
"step": 672
|
4746 |
+
},
|
4747 |
+
{
|
4748 |
+
"epoch": 0.97677793904209,
|
4749 |
+
"grad_norm": 0.28899291157722473,
|
4750 |
+
"learning_rate": 2.7388725328189036e-07,
|
4751 |
+
"loss": 0.5174,
|
4752 |
+
"step": 673
|
4753 |
+
},
|
4754 |
+
{
|
4755 |
+
"epoch": 0.9782293178519593,
|
4756 |
+
"grad_norm": 0.33355581760406494,
|
4757 |
+
"learning_rate": 2.407345325743049e-07,
|
4758 |
+
"loss": 0.4006,
|
4759 |
+
"step": 674
|
4760 |
+
},
|
4761 |
+
{
|
4762 |
+
"epoch": 0.9796806966618288,
|
4763 |
+
"grad_norm": 0.40867161750793457,
|
4764 |
+
"learning_rate": 2.0971737622883515e-07,
|
4765 |
+
"loss": 1.0604,
|
4766 |
+
"step": 675
|
4767 |
+
},
|
4768 |
+
{
|
4769 |
+
"epoch": 0.9811320754716981,
|
4770 |
+
"grad_norm": 0.40149986743927,
|
4771 |
+
"learning_rate": 1.8083644823526647e-07,
|
4772 |
+
"loss": 0.7342,
|
4773 |
+
"step": 676
|
4774 |
+
},
|
4775 |
+
{
|
4776 |
+
"epoch": 0.9825834542815675,
|
4777 |
+
"grad_norm": 0.4048071503639221,
|
4778 |
+
"learning_rate": 1.5409236685277605e-07,
|
4779 |
+
"loss": 0.966,
|
4780 |
+
"step": 677
|
4781 |
+
},
|
4782 |
+
{
|
4783 |
+
"epoch": 0.9840348330914369,
|
4784 |
+
"grad_norm": 0.4240354895591736,
|
4785 |
+
"learning_rate": 1.2948570459667686e-07,
|
4786 |
+
"loss": 1.29,
|
4787 |
+
"step": 678
|
4788 |
+
},
|
4789 |
+
{
|
4790 |
+
"epoch": 0.9854862119013063,
|
4791 |
+
"grad_norm": 0.429208904504776,
|
4792 |
+
"learning_rate": 1.0701698822614959e-07,
|
4793 |
+
"loss": 0.985,
|
4794 |
+
"step": 679
|
4795 |
+
},
|
4796 |
+
{
|
4797 |
+
"epoch": 0.9869375907111756,
|
4798 |
+
"grad_norm": 0.29551103711128235,
|
4799 |
+
"learning_rate": 8.668669873304058e-08,
|
4800 |
+
"loss": 0.6212,
|
4801 |
+
"step": 680
|
4802 |
+
},
|
4803 |
+
{
|
4804 |
+
"epoch": 0.988388969521045,
|
4805 |
+
"grad_norm": 0.4513908922672272,
|
4806 |
+
"learning_rate": 6.849527133144796e-08,
|
4807 |
+
"loss": 1.3823,
|
4808 |
+
"step": 681
|
4809 |
+
},
|
4810 |
+
{
|
4811 |
+
"epoch": 0.9898403483309144,
|
4812 |
+
"grad_norm": 0.38480904698371887,
|
4813 |
+
"learning_rate": 5.2443095448506674e-08,
|
4814 |
+
"loss": 0.8696,
|
4815 |
+
"step": 682
|
4816 |
+
},
|
4817 |
+
{
|
4818 |
+
"epoch": 0.9912917271407837,
|
4819 |
+
"grad_norm": 0.3505820035934448,
|
4820 |
+
"learning_rate": 3.8530514716017544e-08,
|
4821 |
+
"loss": 0.3473,
|
4822 |
+
"step": 683
|
4823 |
+
},
|
4824 |
+
{
|
4825 |
+
"epoch": 0.9927431059506531,
|
4826 |
+
"grad_norm": 0.37889888882637024,
|
4827 |
+
"learning_rate": 2.6757826963053066e-08,
|
4828 |
+
"loss": 0.3755,
|
4829 |
+
"step": 684
|
4830 |
+
},
|
4831 |
+
{
|
4832 |
+
"epoch": 0.9941944847605225,
|
4833 |
+
"grad_norm": 0.3586353063583374,
|
4834 |
+
"learning_rate": 1.712528420966253e-08,
|
4835 |
+
"loss": 1.2269,
|
4836 |
+
"step": 685
|
4837 |
+
},
|
4838 |
+
{
|
4839 |
+
"epoch": 0.9956458635703919,
|
4840 |
+
"grad_norm": 0.2834562361240387,
|
4841 |
+
"learning_rate": 9.633092661376352e-09,
|
4842 |
+
"loss": 0.1832,
|
4843 |
+
"step": 686
|
4844 |
+
},
|
4845 |
+
{
|
4846 |
+
"epoch": 0.9970972423802612,
|
4847 |
+
"grad_norm": 0.3964359760284424,
|
4848 |
+
"learning_rate": 4.281412704887356e-09,
|
4849 |
+
"loss": 1.9892,
|
4850 |
+
"step": 687
|
4851 |
+
},
|
4852 |
+
{
|
4853 |
+
"epoch": 0.9985486211901307,
|
4854 |
+
"grad_norm": 0.543126106262207,
|
4855 |
+
"learning_rate": 1.070358904564639e-09,
|
4856 |
+
"loss": 0.7731,
|
4857 |
+
"step": 688
|
4858 |
+
},
|
4859 |
+
{
|
4860 |
+
"epoch": 1.0,
|
4861 |
+
"grad_norm": 0.22489748895168304,
|
4862 |
+
"learning_rate": 0.0,
|
4863 |
+
"loss": 0.4556,
|
4864 |
+
"step": 689
|
4865 |
}
|
4866 |
],
|
4867 |
"logging_steps": 1,
|
|
|
4876 |
"should_evaluate": false,
|
4877 |
"should_log": false,
|
4878 |
"should_save": true,
|
4879 |
+
"should_training_stop": true
|
4880 |
},
|
4881 |
"attributes": {}
|
4882 |
}
|
4883 |
},
|
4884 |
+
"total_flos": 4.357516200247296e+17,
|
4885 |
"train_batch_size": 2,
|
4886 |
"trial_name": null,
|
4887 |
"trial_params": null
|