Training in progress, epoch 10, 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 540026052
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:40e63c540803c66ab93614b2ff5d15de410719ac1be826b88bfe243577997d1e
|
| 3 |
size 540026052
|
last-checkpoint/optimizer.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 1075446330
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:1c62a36aee495e818da31750c0172ed44160dadea553535b4a8ae082c985e0a0
|
| 3 |
size 1075446330
|
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:c55c0792d1b7dccb405f01a97a5e512cae549864cac8915f93574637399af64a
|
| 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 sha256:46e3123fb72e92b10df5ef82d396a7b1f35005c16f55ce1c7706ce527f3809d4
|
| 3 |
size 1064
|
last-checkpoint/trainer_state.json
CHANGED
|
@@ -1,9 +1,9 @@
|
|
| 1 |
{
|
| 2 |
-
"best_metric": 10.
|
| 3 |
-
"best_model_checkpoint": "./results/checkpoint-
|
| 4 |
-
"epoch":
|
| 5 |
"eval_steps": 500,
|
| 6 |
-
"global_step":
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
@@ -3706,6 +3706,417 @@
|
|
| 3706 |
"eval_samples_per_second": 265.057,
|
| 3707 |
"eval_steps_per_second": 16.624,
|
| 3708 |
"step": 5130
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 3709 |
}
|
| 3710 |
],
|
| 3711 |
"logging_steps": 10,
|
|
@@ -3720,7 +4131,7 @@
|
|
| 3720 |
"should_evaluate": false,
|
| 3721 |
"should_log": false,
|
| 3722 |
"should_save": true,
|
| 3723 |
-
"should_training_stop":
|
| 3724 |
},
|
| 3725 |
"attributes": {}
|
| 3726 |
}
|
|
|
|
| 1 |
{
|
| 2 |
+
"best_metric": 10.266098292947513,
|
| 3 |
+
"best_model_checkpoint": "./results/checkpoint-5700",
|
| 4 |
+
"epoch": 10.0,
|
| 5 |
"eval_steps": 500,
|
| 6 |
+
"global_step": 5700,
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
|
|
| 3706 |
"eval_samples_per_second": 265.057,
|
| 3707 |
"eval_steps_per_second": 16.624,
|
| 3708 |
"step": 5130
|
| 3709 |
+
},
|
| 3710 |
+
{
|
| 3711 |
+
"epoch": 9.017543859649123,
|
| 3712 |
+
"grad_norm": 281.239013671875,
|
| 3713 |
+
"learning_rate": 4.912280701754386e-06,
|
| 3714 |
+
"loss": 166.2685,
|
| 3715 |
+
"step": 5140
|
| 3716 |
+
},
|
| 3717 |
+
{
|
| 3718 |
+
"epoch": 9.035087719298245,
|
| 3719 |
+
"grad_norm": 297.7884216308594,
|
| 3720 |
+
"learning_rate": 4.824561403508772e-06,
|
| 3721 |
+
"loss": 182.959,
|
| 3722 |
+
"step": 5150
|
| 3723 |
+
},
|
| 3724 |
+
{
|
| 3725 |
+
"epoch": 9.052631578947368,
|
| 3726 |
+
"grad_norm": 200.22116088867188,
|
| 3727 |
+
"learning_rate": 4.736842105263159e-06,
|
| 3728 |
+
"loss": 160.0898,
|
| 3729 |
+
"step": 5160
|
| 3730 |
+
},
|
| 3731 |
+
{
|
| 3732 |
+
"epoch": 9.070175438596491,
|
| 3733 |
+
"grad_norm": 729.8984375,
|
| 3734 |
+
"learning_rate": 4.649122807017544e-06,
|
| 3735 |
+
"loss": 162.7628,
|
| 3736 |
+
"step": 5170
|
| 3737 |
+
},
|
| 3738 |
+
{
|
| 3739 |
+
"epoch": 9.087719298245615,
|
| 3740 |
+
"grad_norm": 476.8837585449219,
|
| 3741 |
+
"learning_rate": 4.5614035087719304e-06,
|
| 3742 |
+
"loss": 186.9401,
|
| 3743 |
+
"step": 5180
|
| 3744 |
+
},
|
| 3745 |
+
{
|
| 3746 |
+
"epoch": 9.105263157894736,
|
| 3747 |
+
"grad_norm": 233.36331176757812,
|
| 3748 |
+
"learning_rate": 4.473684210526316e-06,
|
| 3749 |
+
"loss": 181.0498,
|
| 3750 |
+
"step": 5190
|
| 3751 |
+
},
|
| 3752 |
+
{
|
| 3753 |
+
"epoch": 9.12280701754386,
|
| 3754 |
+
"grad_norm": 313.01953125,
|
| 3755 |
+
"learning_rate": 4.3859649122807014e-06,
|
| 3756 |
+
"loss": 157.1372,
|
| 3757 |
+
"step": 5200
|
| 3758 |
+
},
|
| 3759 |
+
{
|
| 3760 |
+
"epoch": 9.140350877192983,
|
| 3761 |
+
"grad_norm": 337.495361328125,
|
| 3762 |
+
"learning_rate": 4.298245614035088e-06,
|
| 3763 |
+
"loss": 181.6845,
|
| 3764 |
+
"step": 5210
|
| 3765 |
+
},
|
| 3766 |
+
{
|
| 3767 |
+
"epoch": 9.157894736842104,
|
| 3768 |
+
"grad_norm": 218.62841796875,
|
| 3769 |
+
"learning_rate": 4.210526315789474e-06,
|
| 3770 |
+
"loss": 164.5618,
|
| 3771 |
+
"step": 5220
|
| 3772 |
+
},
|
| 3773 |
+
{
|
| 3774 |
+
"epoch": 9.175438596491228,
|
| 3775 |
+
"grad_norm": 239.92413330078125,
|
| 3776 |
+
"learning_rate": 4.1228070175438605e-06,
|
| 3777 |
+
"loss": 174.9292,
|
| 3778 |
+
"step": 5230
|
| 3779 |
+
},
|
| 3780 |
+
{
|
| 3781 |
+
"epoch": 9.192982456140351,
|
| 3782 |
+
"grad_norm": 390.8914794921875,
|
| 3783 |
+
"learning_rate": 4.035087719298246e-06,
|
| 3784 |
+
"loss": 213.891,
|
| 3785 |
+
"step": 5240
|
| 3786 |
+
},
|
| 3787 |
+
{
|
| 3788 |
+
"epoch": 9.210526315789474,
|
| 3789 |
+
"grad_norm": 416.03936767578125,
|
| 3790 |
+
"learning_rate": 3.9473684210526315e-06,
|
| 3791 |
+
"loss": 169.2472,
|
| 3792 |
+
"step": 5250
|
| 3793 |
+
},
|
| 3794 |
+
{
|
| 3795 |
+
"epoch": 9.228070175438596,
|
| 3796 |
+
"grad_norm": 304.3415832519531,
|
| 3797 |
+
"learning_rate": 3.859649122807018e-06,
|
| 3798 |
+
"loss": 150.4329,
|
| 3799 |
+
"step": 5260
|
| 3800 |
+
},
|
| 3801 |
+
{
|
| 3802 |
+
"epoch": 9.24561403508772,
|
| 3803 |
+
"grad_norm": 319.07684326171875,
|
| 3804 |
+
"learning_rate": 3.7719298245614037e-06,
|
| 3805 |
+
"loss": 194.0962,
|
| 3806 |
+
"step": 5270
|
| 3807 |
+
},
|
| 3808 |
+
{
|
| 3809 |
+
"epoch": 9.263157894736842,
|
| 3810 |
+
"grad_norm": 230.61453247070312,
|
| 3811 |
+
"learning_rate": 3.6842105263157892e-06,
|
| 3812 |
+
"loss": 154.6355,
|
| 3813 |
+
"step": 5280
|
| 3814 |
+
},
|
| 3815 |
+
{
|
| 3816 |
+
"epoch": 9.280701754385966,
|
| 3817 |
+
"grad_norm": 217.14149475097656,
|
| 3818 |
+
"learning_rate": 3.5964912280701756e-06,
|
| 3819 |
+
"loss": 162.511,
|
| 3820 |
+
"step": 5290
|
| 3821 |
+
},
|
| 3822 |
+
{
|
| 3823 |
+
"epoch": 9.298245614035087,
|
| 3824 |
+
"grad_norm": 256.52978515625,
|
| 3825 |
+
"learning_rate": 3.5087719298245615e-06,
|
| 3826 |
+
"loss": 168.9292,
|
| 3827 |
+
"step": 5300
|
| 3828 |
+
},
|
| 3829 |
+
{
|
| 3830 |
+
"epoch": 9.31578947368421,
|
| 3831 |
+
"grad_norm": 230.6572265625,
|
| 3832 |
+
"learning_rate": 3.421052631578948e-06,
|
| 3833 |
+
"loss": 170.788,
|
| 3834 |
+
"step": 5310
|
| 3835 |
+
},
|
| 3836 |
+
{
|
| 3837 |
+
"epoch": 9.333333333333334,
|
| 3838 |
+
"grad_norm": 226.62417602539062,
|
| 3839 |
+
"learning_rate": 3.3333333333333333e-06,
|
| 3840 |
+
"loss": 150.7565,
|
| 3841 |
+
"step": 5320
|
| 3842 |
+
},
|
| 3843 |
+
{
|
| 3844 |
+
"epoch": 9.350877192982455,
|
| 3845 |
+
"grad_norm": 335.15692138671875,
|
| 3846 |
+
"learning_rate": 3.2456140350877192e-06,
|
| 3847 |
+
"loss": 185.207,
|
| 3848 |
+
"step": 5330
|
| 3849 |
+
},
|
| 3850 |
+
{
|
| 3851 |
+
"epoch": 9.368421052631579,
|
| 3852 |
+
"grad_norm": 241.26544189453125,
|
| 3853 |
+
"learning_rate": 3.1578947368421056e-06,
|
| 3854 |
+
"loss": 169.805,
|
| 3855 |
+
"step": 5340
|
| 3856 |
+
},
|
| 3857 |
+
{
|
| 3858 |
+
"epoch": 9.385964912280702,
|
| 3859 |
+
"grad_norm": 196.6084747314453,
|
| 3860 |
+
"learning_rate": 3.070175438596491e-06,
|
| 3861 |
+
"loss": 148.7297,
|
| 3862 |
+
"step": 5350
|
| 3863 |
+
},
|
| 3864 |
+
{
|
| 3865 |
+
"epoch": 9.403508771929825,
|
| 3866 |
+
"grad_norm": 231.8640899658203,
|
| 3867 |
+
"learning_rate": 2.9824561403508774e-06,
|
| 3868 |
+
"loss": 149.0492,
|
| 3869 |
+
"step": 5360
|
| 3870 |
+
},
|
| 3871 |
+
{
|
| 3872 |
+
"epoch": 9.421052631578947,
|
| 3873 |
+
"grad_norm": 322.551025390625,
|
| 3874 |
+
"learning_rate": 2.8947368421052634e-06,
|
| 3875 |
+
"loss": 171.7541,
|
| 3876 |
+
"step": 5370
|
| 3877 |
+
},
|
| 3878 |
+
{
|
| 3879 |
+
"epoch": 9.43859649122807,
|
| 3880 |
+
"grad_norm": 359.8404541015625,
|
| 3881 |
+
"learning_rate": 2.8070175438596493e-06,
|
| 3882 |
+
"loss": 134.9022,
|
| 3883 |
+
"step": 5380
|
| 3884 |
+
},
|
| 3885 |
+
{
|
| 3886 |
+
"epoch": 9.456140350877194,
|
| 3887 |
+
"grad_norm": 393.9560241699219,
|
| 3888 |
+
"learning_rate": 2.719298245614035e-06,
|
| 3889 |
+
"loss": 160.3259,
|
| 3890 |
+
"step": 5390
|
| 3891 |
+
},
|
| 3892 |
+
{
|
| 3893 |
+
"epoch": 9.473684210526315,
|
| 3894 |
+
"grad_norm": 307.36712646484375,
|
| 3895 |
+
"learning_rate": 2.631578947368421e-06,
|
| 3896 |
+
"loss": 169.45,
|
| 3897 |
+
"step": 5400
|
| 3898 |
+
},
|
| 3899 |
+
{
|
| 3900 |
+
"epoch": 9.491228070175438,
|
| 3901 |
+
"grad_norm": 429.7928161621094,
|
| 3902 |
+
"learning_rate": 2.543859649122807e-06,
|
| 3903 |
+
"loss": 161.9689,
|
| 3904 |
+
"step": 5410
|
| 3905 |
+
},
|
| 3906 |
+
{
|
| 3907 |
+
"epoch": 9.508771929824562,
|
| 3908 |
+
"grad_norm": 260.0240173339844,
|
| 3909 |
+
"learning_rate": 2.456140350877193e-06,
|
| 3910 |
+
"loss": 182.8279,
|
| 3911 |
+
"step": 5420
|
| 3912 |
+
},
|
| 3913 |
+
{
|
| 3914 |
+
"epoch": 9.526315789473685,
|
| 3915 |
+
"grad_norm": 260.8311767578125,
|
| 3916 |
+
"learning_rate": 2.3684210526315793e-06,
|
| 3917 |
+
"loss": 140.941,
|
| 3918 |
+
"step": 5430
|
| 3919 |
+
},
|
| 3920 |
+
{
|
| 3921 |
+
"epoch": 9.543859649122806,
|
| 3922 |
+
"grad_norm": 189.70962524414062,
|
| 3923 |
+
"learning_rate": 2.2807017543859652e-06,
|
| 3924 |
+
"loss": 187.3211,
|
| 3925 |
+
"step": 5440
|
| 3926 |
+
},
|
| 3927 |
+
{
|
| 3928 |
+
"epoch": 9.56140350877193,
|
| 3929 |
+
"grad_norm": 364.1365966796875,
|
| 3930 |
+
"learning_rate": 2.1929824561403507e-06,
|
| 3931 |
+
"loss": 155.0701,
|
| 3932 |
+
"step": 5450
|
| 3933 |
+
},
|
| 3934 |
+
{
|
| 3935 |
+
"epoch": 9.578947368421053,
|
| 3936 |
+
"grad_norm": 244.2755126953125,
|
| 3937 |
+
"learning_rate": 2.105263157894737e-06,
|
| 3938 |
+
"loss": 182.5883,
|
| 3939 |
+
"step": 5460
|
| 3940 |
+
},
|
| 3941 |
+
{
|
| 3942 |
+
"epoch": 9.596491228070175,
|
| 3943 |
+
"grad_norm": 417.791015625,
|
| 3944 |
+
"learning_rate": 2.017543859649123e-06,
|
| 3945 |
+
"loss": 173.9106,
|
| 3946 |
+
"step": 5470
|
| 3947 |
+
},
|
| 3948 |
+
{
|
| 3949 |
+
"epoch": 9.614035087719298,
|
| 3950 |
+
"grad_norm": 373.1618347167969,
|
| 3951 |
+
"learning_rate": 1.929824561403509e-06,
|
| 3952 |
+
"loss": 178.0014,
|
| 3953 |
+
"step": 5480
|
| 3954 |
+
},
|
| 3955 |
+
{
|
| 3956 |
+
"epoch": 9.631578947368421,
|
| 3957 |
+
"grad_norm": 221.83309936523438,
|
| 3958 |
+
"learning_rate": 1.8421052631578946e-06,
|
| 3959 |
+
"loss": 182.6862,
|
| 3960 |
+
"step": 5490
|
| 3961 |
+
},
|
| 3962 |
+
{
|
| 3963 |
+
"epoch": 9.649122807017545,
|
| 3964 |
+
"grad_norm": 301.48382568359375,
|
| 3965 |
+
"learning_rate": 1.7543859649122807e-06,
|
| 3966 |
+
"loss": 159.5674,
|
| 3967 |
+
"step": 5500
|
| 3968 |
+
},
|
| 3969 |
+
{
|
| 3970 |
+
"epoch": 9.666666666666666,
|
| 3971 |
+
"grad_norm": 487.5570983886719,
|
| 3972 |
+
"learning_rate": 1.6666666666666667e-06,
|
| 3973 |
+
"loss": 164.2859,
|
| 3974 |
+
"step": 5510
|
| 3975 |
+
},
|
| 3976 |
+
{
|
| 3977 |
+
"epoch": 9.68421052631579,
|
| 3978 |
+
"grad_norm": 326.3913269042969,
|
| 3979 |
+
"learning_rate": 1.5789473684210528e-06,
|
| 3980 |
+
"loss": 188.0145,
|
| 3981 |
+
"step": 5520
|
| 3982 |
+
},
|
| 3983 |
+
{
|
| 3984 |
+
"epoch": 9.701754385964913,
|
| 3985 |
+
"grad_norm": 314.8993835449219,
|
| 3986 |
+
"learning_rate": 1.4912280701754387e-06,
|
| 3987 |
+
"loss": 156.2349,
|
| 3988 |
+
"step": 5530
|
| 3989 |
+
},
|
| 3990 |
+
{
|
| 3991 |
+
"epoch": 9.719298245614034,
|
| 3992 |
+
"grad_norm": 1071.3399658203125,
|
| 3993 |
+
"learning_rate": 1.4035087719298246e-06,
|
| 3994 |
+
"loss": 187.2437,
|
| 3995 |
+
"step": 5540
|
| 3996 |
+
},
|
| 3997 |
+
{
|
| 3998 |
+
"epoch": 9.736842105263158,
|
| 3999 |
+
"grad_norm": 228.32090759277344,
|
| 4000 |
+
"learning_rate": 1.3157894736842106e-06,
|
| 4001 |
+
"loss": 156.9955,
|
| 4002 |
+
"step": 5550
|
| 4003 |
+
},
|
| 4004 |
+
{
|
| 4005 |
+
"epoch": 9.75438596491228,
|
| 4006 |
+
"grad_norm": 291.3077697753906,
|
| 4007 |
+
"learning_rate": 1.2280701754385965e-06,
|
| 4008 |
+
"loss": 160.9737,
|
| 4009 |
+
"step": 5560
|
| 4010 |
+
},
|
| 4011 |
+
{
|
| 4012 |
+
"epoch": 9.771929824561404,
|
| 4013 |
+
"grad_norm": 530.8067626953125,
|
| 4014 |
+
"learning_rate": 1.1403508771929826e-06,
|
| 4015 |
+
"loss": 151.8838,
|
| 4016 |
+
"step": 5570
|
| 4017 |
+
},
|
| 4018 |
+
{
|
| 4019 |
+
"epoch": 9.789473684210526,
|
| 4020 |
+
"grad_norm": 337.1194763183594,
|
| 4021 |
+
"learning_rate": 1.0526315789473685e-06,
|
| 4022 |
+
"loss": 158.343,
|
| 4023 |
+
"step": 5580
|
| 4024 |
+
},
|
| 4025 |
+
{
|
| 4026 |
+
"epoch": 9.807017543859649,
|
| 4027 |
+
"grad_norm": 304.9252624511719,
|
| 4028 |
+
"learning_rate": 9.649122807017545e-07,
|
| 4029 |
+
"loss": 172.4637,
|
| 4030 |
+
"step": 5590
|
| 4031 |
+
},
|
| 4032 |
+
{
|
| 4033 |
+
"epoch": 9.824561403508772,
|
| 4034 |
+
"grad_norm": 236.6869354248047,
|
| 4035 |
+
"learning_rate": 8.771929824561404e-07,
|
| 4036 |
+
"loss": 156.1226,
|
| 4037 |
+
"step": 5600
|
| 4038 |
+
},
|
| 4039 |
+
{
|
| 4040 |
+
"epoch": 9.842105263157894,
|
| 4041 |
+
"grad_norm": 249.76043701171875,
|
| 4042 |
+
"learning_rate": 7.894736842105264e-07,
|
| 4043 |
+
"loss": 167.4769,
|
| 4044 |
+
"step": 5610
|
| 4045 |
+
},
|
| 4046 |
+
{
|
| 4047 |
+
"epoch": 9.859649122807017,
|
| 4048 |
+
"grad_norm": 296.2056884765625,
|
| 4049 |
+
"learning_rate": 7.017543859649123e-07,
|
| 4050 |
+
"loss": 155.2643,
|
| 4051 |
+
"step": 5620
|
| 4052 |
+
},
|
| 4053 |
+
{
|
| 4054 |
+
"epoch": 9.87719298245614,
|
| 4055 |
+
"grad_norm": 198.53219604492188,
|
| 4056 |
+
"learning_rate": 6.140350877192982e-07,
|
| 4057 |
+
"loss": 168.2789,
|
| 4058 |
+
"step": 5630
|
| 4059 |
+
},
|
| 4060 |
+
{
|
| 4061 |
+
"epoch": 9.894736842105264,
|
| 4062 |
+
"grad_norm": 714.234130859375,
|
| 4063 |
+
"learning_rate": 5.263157894736843e-07,
|
| 4064 |
+
"loss": 183.7942,
|
| 4065 |
+
"step": 5640
|
| 4066 |
+
},
|
| 4067 |
+
{
|
| 4068 |
+
"epoch": 9.912280701754385,
|
| 4069 |
+
"grad_norm": 229.8941192626953,
|
| 4070 |
+
"learning_rate": 4.385964912280702e-07,
|
| 4071 |
+
"loss": 147.4333,
|
| 4072 |
+
"step": 5650
|
| 4073 |
+
},
|
| 4074 |
+
{
|
| 4075 |
+
"epoch": 9.929824561403509,
|
| 4076 |
+
"grad_norm": 462.6064147949219,
|
| 4077 |
+
"learning_rate": 3.5087719298245616e-07,
|
| 4078 |
+
"loss": 154.1472,
|
| 4079 |
+
"step": 5660
|
| 4080 |
+
},
|
| 4081 |
+
{
|
| 4082 |
+
"epoch": 9.947368421052632,
|
| 4083 |
+
"grad_norm": 275.9501953125,
|
| 4084 |
+
"learning_rate": 2.6315789473684213e-07,
|
| 4085 |
+
"loss": 161.2116,
|
| 4086 |
+
"step": 5670
|
| 4087 |
+
},
|
| 4088 |
+
{
|
| 4089 |
+
"epoch": 9.964912280701755,
|
| 4090 |
+
"grad_norm": 696.220703125,
|
| 4091 |
+
"learning_rate": 1.7543859649122808e-07,
|
| 4092 |
+
"loss": 164.8297,
|
| 4093 |
+
"step": 5680
|
| 4094 |
+
},
|
| 4095 |
+
{
|
| 4096 |
+
"epoch": 9.982456140350877,
|
| 4097 |
+
"grad_norm": 498.91448974609375,
|
| 4098 |
+
"learning_rate": 8.771929824561404e-08,
|
| 4099 |
+
"loss": 153.6297,
|
| 4100 |
+
"step": 5690
|
| 4101 |
+
},
|
| 4102 |
+
{
|
| 4103 |
+
"epoch": 10.0,
|
| 4104 |
+
"grad_norm": 873.9326782226562,
|
| 4105 |
+
"learning_rate": 0.0,
|
| 4106 |
+
"loss": 155.5076,
|
| 4107 |
+
"step": 5700
|
| 4108 |
+
},
|
| 4109 |
+
{
|
| 4110 |
+
"epoch": 10.0,
|
| 4111 |
+
"eval_loss": 171.82171630859375,
|
| 4112 |
+
"eval_mae_lex": 11.493141912159167,
|
| 4113 |
+
"eval_mae_overall": 10.266098292947513,
|
| 4114 |
+
"eval_mae_sem": 7.863255273250111,
|
| 4115 |
+
"eval_mae_syn": 11.44189769343326,
|
| 4116 |
+
"eval_runtime": 8.6814,
|
| 4117 |
+
"eval_samples_per_second": 262.631,
|
| 4118 |
+
"eval_steps_per_second": 16.472,
|
| 4119 |
+
"step": 5700
|
| 4120 |
}
|
| 4121 |
],
|
| 4122 |
"logging_steps": 10,
|
|
|
|
| 4131 |
"should_evaluate": false,
|
| 4132 |
"should_log": false,
|
| 4133 |
"should_save": true,
|
| 4134 |
+
"should_training_stop": true
|
| 4135 |
},
|
| 4136 |
"attributes": {}
|
| 4137 |
}
|