Training in progress, step 5500, checkpoint
Browse files
last-checkpoint/model.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 328277848
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:e8ae4d439763bede675a7bb8407ca626ba1a1ca1d28d508145ff27990bcdfd60
|
| 3 |
size 328277848
|
last-checkpoint/optimizer.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 318646859
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:8df4f8c8c0f93c7a4647906cc1e5f85c72386b1b581eb687df3d305abbdc44a7
|
| 3 |
size 318646859
|
last-checkpoint/rng_state.pth
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 14645
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:8af3cc3f1560f815527e73bcdf0bbfb03998a87b5067ff9928ca94f46e638231
|
| 3 |
size 14645
|
last-checkpoint/scheduler.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 1465
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:bda4b56b57284b5d776cea834f86539fa062d5e046885e07dcb7516921ccd6ee
|
| 3 |
size 1465
|
last-checkpoint/trainer_state.json
CHANGED
|
@@ -2,9 +2,9 @@
|
|
| 2 |
"best_global_step": null,
|
| 3 |
"best_metric": null,
|
| 4 |
"best_model_checkpoint": null,
|
| 5 |
-
"epoch": 0.
|
| 6 |
"eval_steps": 500,
|
| 7 |
-
"global_step":
|
| 8 |
"is_hyper_param_search": false,
|
| 9 |
"is_local_process_zero": true,
|
| 10 |
"is_world_process_zero": true,
|
|
@@ -3595,6 +3595,364 @@
|
|
| 3595 |
"eval_samples_per_second": 279.83,
|
| 3596 |
"eval_steps_per_second": 5.876,
|
| 3597 |
"step": 5000
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 3598 |
}
|
| 3599 |
],
|
| 3600 |
"logging_steps": 10,
|
|
@@ -3614,7 +3972,7 @@
|
|
| 3614 |
"attributes": {}
|
| 3615 |
}
|
| 3616 |
},
|
| 3617 |
-
"total_flos": 1.
|
| 3618 |
"train_batch_size": 48,
|
| 3619 |
"trial_name": null,
|
| 3620 |
"trial_params": null
|
|
|
|
| 2 |
"best_global_step": null,
|
| 3 |
"best_metric": null,
|
| 4 |
"best_model_checkpoint": null,
|
| 5 |
+
"epoch": 0.9292110153742186,
|
| 6 |
"eval_steps": 500,
|
| 7 |
+
"global_step": 5500,
|
| 8 |
"is_hyper_param_search": false,
|
| 9 |
"is_local_process_zero": true,
|
| 10 |
"is_world_process_zero": true,
|
|
|
|
| 3595 |
"eval_samples_per_second": 279.83,
|
| 3596 |
"eval_steps_per_second": 5.876,
|
| 3597 |
"step": 5000
|
| 3598 |
+
},
|
| 3599 |
+
{
|
| 3600 |
+
"epoch": 0.8464267612772428,
|
| 3601 |
+
"grad_norm": 0.559013307094574,
|
| 3602 |
+
"learning_rate": 0.00023592089546305216,
|
| 3603 |
+
"loss": 4.576456832885742,
|
| 3604 |
+
"step": 5010
|
| 3605 |
+
},
|
| 3606 |
+
{
|
| 3607 |
+
"epoch": 0.8481162358506504,
|
| 3608 |
+
"grad_norm": 0.49753278493881226,
|
| 3609 |
+
"learning_rate": 0.00023552782714923343,
|
| 3610 |
+
"loss": 4.614764404296875,
|
| 3611 |
+
"step": 5020
|
| 3612 |
+
},
|
| 3613 |
+
{
|
| 3614 |
+
"epoch": 0.8498057104240582,
|
| 3615 |
+
"grad_norm": 0.56475430727005,
|
| 3616 |
+
"learning_rate": 0.00023513388668141118,
|
| 3617 |
+
"loss": 4.58197250366211,
|
| 3618 |
+
"step": 5030
|
| 3619 |
+
},
|
| 3620 |
+
{
|
| 3621 |
+
"epoch": 0.8514951849974658,
|
| 3622 |
+
"grad_norm": 0.4953176975250244,
|
| 3623 |
+
"learning_rate": 0.00023473907807671952,
|
| 3624 |
+
"loss": 4.605267333984375,
|
| 3625 |
+
"step": 5040
|
| 3626 |
+
},
|
| 3627 |
+
{
|
| 3628 |
+
"epoch": 0.8531846595708734,
|
| 3629 |
+
"grad_norm": 0.5137485861778259,
|
| 3630 |
+
"learning_rate": 0.00023434340536114531,
|
| 3631 |
+
"loss": 4.596157836914062,
|
| 3632 |
+
"step": 5050
|
| 3633 |
+
},
|
| 3634 |
+
{
|
| 3635 |
+
"epoch": 0.8548741341442812,
|
| 3636 |
+
"grad_norm": 0.5363683104515076,
|
| 3637 |
+
"learning_rate": 0.00023394687256948697,
|
| 3638 |
+
"loss": 4.594855499267578,
|
| 3639 |
+
"step": 5060
|
| 3640 |
+
},
|
| 3641 |
+
{
|
| 3642 |
+
"epoch": 0.8565636087176888,
|
| 3643 |
+
"grad_norm": 0.5634586811065674,
|
| 3644 |
+
"learning_rate": 0.00023354948374531344,
|
| 3645 |
+
"loss": 4.589244842529297,
|
| 3646 |
+
"step": 5070
|
| 3647 |
+
},
|
| 3648 |
+
{
|
| 3649 |
+
"epoch": 0.8582530832910965,
|
| 3650 |
+
"grad_norm": 0.5043785572052002,
|
| 3651 |
+
"learning_rate": 0.00023315124294092277,
|
| 3652 |
+
"loss": 4.576361083984375,
|
| 3653 |
+
"step": 5080
|
| 3654 |
+
},
|
| 3655 |
+
{
|
| 3656 |
+
"epoch": 0.8599425578645041,
|
| 3657 |
+
"grad_norm": 0.5204640626907349,
|
| 3658 |
+
"learning_rate": 0.000232752154217301,
|
| 3659 |
+
"loss": 4.570015716552734,
|
| 3660 |
+
"step": 5090
|
| 3661 |
+
},
|
| 3662 |
+
{
|
| 3663 |
+
"epoch": 0.8616320324379118,
|
| 3664 |
+
"grad_norm": 0.5251067280769348,
|
| 3665 |
+
"learning_rate": 0.00023235222164408076,
|
| 3666 |
+
"loss": 4.598841857910156,
|
| 3667 |
+
"step": 5100
|
| 3668 |
+
},
|
| 3669 |
+
{
|
| 3670 |
+
"epoch": 0.8633215070113195,
|
| 3671 |
+
"grad_norm": 0.5268970131874084,
|
| 3672 |
+
"learning_rate": 0.00023195144929949953,
|
| 3673 |
+
"loss": 4.574850082397461,
|
| 3674 |
+
"step": 5110
|
| 3675 |
+
},
|
| 3676 |
+
{
|
| 3677 |
+
"epoch": 0.8650109815847271,
|
| 3678 |
+
"grad_norm": 0.5099704265594482,
|
| 3679 |
+
"learning_rate": 0.00023154984127035823,
|
| 3680 |
+
"loss": 4.602288436889649,
|
| 3681 |
+
"step": 5120
|
| 3682 |
+
},
|
| 3683 |
+
{
|
| 3684 |
+
"epoch": 0.8667004561581348,
|
| 3685 |
+
"grad_norm": 0.49661147594451904,
|
| 3686 |
+
"learning_rate": 0.00023114740165197957,
|
| 3687 |
+
"loss": 4.56927604675293,
|
| 3688 |
+
"step": 5130
|
| 3689 |
+
},
|
| 3690 |
+
{
|
| 3691 |
+
"epoch": 0.8683899307315425,
|
| 3692 |
+
"grad_norm": 0.5453396439552307,
|
| 3693 |
+
"learning_rate": 0.00023074413454816619,
|
| 3694 |
+
"loss": 4.587471771240234,
|
| 3695 |
+
"step": 5140
|
| 3696 |
+
},
|
| 3697 |
+
{
|
| 3698 |
+
"epoch": 0.8700794053049502,
|
| 3699 |
+
"grad_norm": 0.49157091975212097,
|
| 3700 |
+
"learning_rate": 0.0002303400440711589,
|
| 3701 |
+
"loss": 4.580040740966797,
|
| 3702 |
+
"step": 5150
|
| 3703 |
+
},
|
| 3704 |
+
{
|
| 3705 |
+
"epoch": 0.8717688798783578,
|
| 3706 |
+
"grad_norm": 0.5203030705451965,
|
| 3707 |
+
"learning_rate": 0.00022993513434159464,
|
| 3708 |
+
"loss": 4.604449462890625,
|
| 3709 |
+
"step": 5160
|
| 3710 |
+
},
|
| 3711 |
+
{
|
| 3712 |
+
"epoch": 0.8734583544517655,
|
| 3713 |
+
"grad_norm": 0.49596408009529114,
|
| 3714 |
+
"learning_rate": 0.0002295294094884646,
|
| 3715 |
+
"loss": 4.597255706787109,
|
| 3716 |
+
"step": 5170
|
| 3717 |
+
},
|
| 3718 |
+
{
|
| 3719 |
+
"epoch": 0.8751478290251732,
|
| 3720 |
+
"grad_norm": 0.482197642326355,
|
| 3721 |
+
"learning_rate": 0.00022912287364907204,
|
| 3722 |
+
"loss": 4.575711822509765,
|
| 3723 |
+
"step": 5180
|
| 3724 |
+
},
|
| 3725 |
+
{
|
| 3726 |
+
"epoch": 0.8768373035985808,
|
| 3727 |
+
"grad_norm": 0.49015575647354126,
|
| 3728 |
+
"learning_rate": 0.00022871553096899,
|
| 3729 |
+
"loss": 4.620565032958984,
|
| 3730 |
+
"step": 5190
|
| 3731 |
+
},
|
| 3732 |
+
{
|
| 3733 |
+
"epoch": 0.8785267781719885,
|
| 3734 |
+
"grad_norm": 0.49912846088409424,
|
| 3735 |
+
"learning_rate": 0.00022830738560201911,
|
| 3736 |
+
"loss": 4.575767898559571,
|
| 3737 |
+
"step": 5200
|
| 3738 |
+
},
|
| 3739 |
+
{
|
| 3740 |
+
"epoch": 0.8802162527453962,
|
| 3741 |
+
"grad_norm": 0.5013103485107422,
|
| 3742 |
+
"learning_rate": 0.00022789844171014557,
|
| 3743 |
+
"loss": 4.5688629150390625,
|
| 3744 |
+
"step": 5210
|
| 3745 |
+
},
|
| 3746 |
+
{
|
| 3747 |
+
"epoch": 0.8819057273188039,
|
| 3748 |
+
"grad_norm": 0.5322986245155334,
|
| 3749 |
+
"learning_rate": 0.00022748870346349796,
|
| 3750 |
+
"loss": 4.590381622314453,
|
| 3751 |
+
"step": 5220
|
| 3752 |
+
},
|
| 3753 |
+
{
|
| 3754 |
+
"epoch": 0.8835952018922115,
|
| 3755 |
+
"grad_norm": 0.5467557311058044,
|
| 3756 |
+
"learning_rate": 0.00022707817504030538,
|
| 3757 |
+
"loss": 4.598742294311523,
|
| 3758 |
+
"step": 5230
|
| 3759 |
+
},
|
| 3760 |
+
{
|
| 3761 |
+
"epoch": 0.8852846764656191,
|
| 3762 |
+
"grad_norm": 0.5180667638778687,
|
| 3763 |
+
"learning_rate": 0.0002266668606268545,
|
| 3764 |
+
"loss": 4.550464630126953,
|
| 3765 |
+
"step": 5240
|
| 3766 |
+
},
|
| 3767 |
+
{
|
| 3768 |
+
"epoch": 0.8869741510390269,
|
| 3769 |
+
"grad_norm": 0.5265566110610962,
|
| 3770 |
+
"learning_rate": 0.00022625476441744715,
|
| 3771 |
+
"loss": 4.595706176757813,
|
| 3772 |
+
"step": 5250
|
| 3773 |
+
},
|
| 3774 |
+
{
|
| 3775 |
+
"epoch": 0.8886636256124345,
|
| 3776 |
+
"grad_norm": 0.5108802318572998,
|
| 3777 |
+
"learning_rate": 0.00022584189061435725,
|
| 3778 |
+
"loss": 4.564280700683594,
|
| 3779 |
+
"step": 5260
|
| 3780 |
+
},
|
| 3781 |
+
{
|
| 3782 |
+
"epoch": 0.8903531001858422,
|
| 3783 |
+
"grad_norm": 0.5016060471534729,
|
| 3784 |
+
"learning_rate": 0.00022542824342778806,
|
| 3785 |
+
"loss": 4.561073303222656,
|
| 3786 |
+
"step": 5270
|
| 3787 |
+
},
|
| 3788 |
+
{
|
| 3789 |
+
"epoch": 0.8920425747592499,
|
| 3790 |
+
"grad_norm": 0.531934916973114,
|
| 3791 |
+
"learning_rate": 0.0002250138270758293,
|
| 3792 |
+
"loss": 4.576354598999023,
|
| 3793 |
+
"step": 5280
|
| 3794 |
+
},
|
| 3795 |
+
{
|
| 3796 |
+
"epoch": 0.8937320493326576,
|
| 3797 |
+
"grad_norm": 0.5116508603096008,
|
| 3798 |
+
"learning_rate": 0.00022459864578441415,
|
| 3799 |
+
"loss": 4.586645889282226,
|
| 3800 |
+
"step": 5290
|
| 3801 |
+
},
|
| 3802 |
+
{
|
| 3803 |
+
"epoch": 0.8954215239060652,
|
| 3804 |
+
"grad_norm": 0.5074120163917542,
|
| 3805 |
+
"learning_rate": 0.0002241827037872761,
|
| 3806 |
+
"loss": 4.564454650878906,
|
| 3807 |
+
"step": 5300
|
| 3808 |
+
},
|
| 3809 |
+
{
|
| 3810 |
+
"epoch": 0.8971109984794728,
|
| 3811 |
+
"grad_norm": 0.47892510890960693,
|
| 3812 |
+
"learning_rate": 0.00022376600532590578,
|
| 3813 |
+
"loss": 4.534092712402344,
|
| 3814 |
+
"step": 5310
|
| 3815 |
+
},
|
| 3816 |
+
{
|
| 3817 |
+
"epoch": 0.8988004730528806,
|
| 3818 |
+
"grad_norm": 0.4657728970050812,
|
| 3819 |
+
"learning_rate": 0.00022334855464950775,
|
| 3820 |
+
"loss": 4.5831245422363285,
|
| 3821 |
+
"step": 5320
|
| 3822 |
+
},
|
| 3823 |
+
{
|
| 3824 |
+
"epoch": 0.9004899476262882,
|
| 3825 |
+
"grad_norm": 0.4809263348579407,
|
| 3826 |
+
"learning_rate": 0.00022293035601495708,
|
| 3827 |
+
"loss": 4.578067398071289,
|
| 3828 |
+
"step": 5330
|
| 3829 |
+
},
|
| 3830 |
+
{
|
| 3831 |
+
"epoch": 0.9021794221996959,
|
| 3832 |
+
"grad_norm": 0.5298095941543579,
|
| 3833 |
+
"learning_rate": 0.00022251141368675607,
|
| 3834 |
+
"loss": 4.576302719116211,
|
| 3835 |
+
"step": 5340
|
| 3836 |
+
},
|
| 3837 |
+
{
|
| 3838 |
+
"epoch": 0.9038688967731036,
|
| 3839 |
+
"grad_norm": 0.5315806865692139,
|
| 3840 |
+
"learning_rate": 0.00022209173193699067,
|
| 3841 |
+
"loss": 4.582082748413086,
|
| 3842 |
+
"step": 5350
|
| 3843 |
+
},
|
| 3844 |
+
{
|
| 3845 |
+
"epoch": 0.9055583713465112,
|
| 3846 |
+
"grad_norm": 0.5084795355796814,
|
| 3847 |
+
"learning_rate": 0.00022167131504528695,
|
| 3848 |
+
"loss": 4.5921672821044925,
|
| 3849 |
+
"step": 5360
|
| 3850 |
+
},
|
| 3851 |
+
{
|
| 3852 |
+
"epoch": 0.9072478459199189,
|
| 3853 |
+
"grad_norm": 0.4921436011791229,
|
| 3854 |
+
"learning_rate": 0.00022125016729876743,
|
| 3855 |
+
"loss": 4.573263931274414,
|
| 3856 |
+
"step": 5370
|
| 3857 |
+
},
|
| 3858 |
+
{
|
| 3859 |
+
"epoch": 0.9089373204933265,
|
| 3860 |
+
"grad_norm": 0.4985114336013794,
|
| 3861 |
+
"learning_rate": 0.00022082829299200743,
|
| 3862 |
+
"loss": 4.580036163330078,
|
| 3863 |
+
"step": 5380
|
| 3864 |
+
},
|
| 3865 |
+
{
|
| 3866 |
+
"epoch": 0.9106267950667343,
|
| 3867 |
+
"grad_norm": 0.486751526594162,
|
| 3868 |
+
"learning_rate": 0.00022040569642699112,
|
| 3869 |
+
"loss": 4.554470062255859,
|
| 3870 |
+
"step": 5390
|
| 3871 |
+
},
|
| 3872 |
+
{
|
| 3873 |
+
"epoch": 0.9123162696401419,
|
| 3874 |
+
"grad_norm": 0.49089571833610535,
|
| 3875 |
+
"learning_rate": 0.00021998238191306798,
|
| 3876 |
+
"loss": 4.53393783569336,
|
| 3877 |
+
"step": 5400
|
| 3878 |
+
},
|
| 3879 |
+
{
|
| 3880 |
+
"epoch": 0.9140057442135496,
|
| 3881 |
+
"grad_norm": 0.5415358543395996,
|
| 3882 |
+
"learning_rate": 0.00021955835376690841,
|
| 3883 |
+
"loss": 4.545674514770508,
|
| 3884 |
+
"step": 5410
|
| 3885 |
+
},
|
| 3886 |
+
{
|
| 3887 |
+
"epoch": 0.9156952187869573,
|
| 3888 |
+
"grad_norm": 0.5360416173934937,
|
| 3889 |
+
"learning_rate": 0.00021913361631246004,
|
| 3890 |
+
"loss": 4.559771728515625,
|
| 3891 |
+
"step": 5420
|
| 3892 |
+
},
|
| 3893 |
+
{
|
| 3894 |
+
"epoch": 0.9173846933603649,
|
| 3895 |
+
"grad_norm": 0.5081850290298462,
|
| 3896 |
+
"learning_rate": 0.0002187081738809036,
|
| 3897 |
+
"loss": 4.550452804565429,
|
| 3898 |
+
"step": 5430
|
| 3899 |
+
},
|
| 3900 |
+
{
|
| 3901 |
+
"epoch": 0.9190741679337726,
|
| 3902 |
+
"grad_norm": 0.5538184642791748,
|
| 3903 |
+
"learning_rate": 0.00021828203081060858,
|
| 3904 |
+
"loss": 4.558630752563476,
|
| 3905 |
+
"step": 5440
|
| 3906 |
+
},
|
| 3907 |
+
{
|
| 3908 |
+
"epoch": 0.9207636425071802,
|
| 3909 |
+
"grad_norm": 0.4913816452026367,
|
| 3910 |
+
"learning_rate": 0.00021785519144708912,
|
| 3911 |
+
"loss": 4.532632446289062,
|
| 3912 |
+
"step": 5450
|
| 3913 |
+
},
|
| 3914 |
+
{
|
| 3915 |
+
"epoch": 0.922453117080588,
|
| 3916 |
+
"grad_norm": 0.4864713251590729,
|
| 3917 |
+
"learning_rate": 0.00021742766014295976,
|
| 3918 |
+
"loss": 4.546042251586914,
|
| 3919 |
+
"step": 5460
|
| 3920 |
+
},
|
| 3921 |
+
{
|
| 3922 |
+
"epoch": 0.9241425916539956,
|
| 3923 |
+
"grad_norm": 0.5309963822364807,
|
| 3924 |
+
"learning_rate": 0.00021699944125789096,
|
| 3925 |
+
"loss": 4.532254791259765,
|
| 3926 |
+
"step": 5470
|
| 3927 |
+
},
|
| 3928 |
+
{
|
| 3929 |
+
"epoch": 0.9258320662274033,
|
| 3930 |
+
"grad_norm": 0.47752541303634644,
|
| 3931 |
+
"learning_rate": 0.00021657053915856455,
|
| 3932 |
+
"loss": 4.558838272094727,
|
| 3933 |
+
"step": 5480
|
| 3934 |
+
},
|
| 3935 |
+
{
|
| 3936 |
+
"epoch": 0.927521540800811,
|
| 3937 |
+
"grad_norm": 0.47722741961479187,
|
| 3938 |
+
"learning_rate": 0.0002161409582186294,
|
| 3939 |
+
"loss": 4.583000183105469,
|
| 3940 |
+
"step": 5490
|
| 3941 |
+
},
|
| 3942 |
+
{
|
| 3943 |
+
"epoch": 0.9292110153742186,
|
| 3944 |
+
"grad_norm": 0.5001320838928223,
|
| 3945 |
+
"learning_rate": 0.0002157107028186567,
|
| 3946 |
+
"loss": 4.5466560363769535,
|
| 3947 |
+
"step": 5500
|
| 3948 |
+
},
|
| 3949 |
+
{
|
| 3950 |
+
"epoch": 0.9292110153742186,
|
| 3951 |
+
"eval_loss": 4.510837554931641,
|
| 3952 |
+
"eval_runtime": 3.6228,
|
| 3953 |
+
"eval_samples_per_second": 276.03,
|
| 3954 |
+
"eval_steps_per_second": 5.797,
|
| 3955 |
+
"step": 5500
|
| 3956 |
}
|
| 3957 |
],
|
| 3958 |
"logging_steps": 10,
|
|
|
|
| 3972 |
"attributes": {}
|
| 3973 |
}
|
| 3974 |
},
|
| 3975 |
+
"total_flos": 1.83951251472384e+17,
|
| 3976 |
"train_batch_size": 48,
|
| 3977 |
"trial_name": null,
|
| 3978 |
"trial_params": null
|