Training in progress, step 200, 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 645975704
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:90cf86756df572135f03c0f7b02e368ed11704950657d207cb96e852a67974c0
|
| 3 |
size 645975704
|
last-checkpoint/optimizer.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 328468404
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:c49898add5a21782109bcea17e094f097a2730fa5548b1a51e188fe8213625b0
|
| 3 |
size 328468404
|
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 sha256:9c1ef41deb3863cb4e4ca9f914e8469725fb857615262228d60028b63b24c217
|
| 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 sha256:5a94566b02bf28490606f2fb9cb895b81b322144d6c373560bb6cc082ffc05f1
|
| 3 |
size 1064
|
last-checkpoint/trainer_state.json
CHANGED
|
@@ -1,9 +1,9 @@
|
|
| 1 |
{
|
| 2 |
-
"best_metric": 0.
|
| 3 |
-
"best_model_checkpoint": "miner_id_24/checkpoint-
|
| 4 |
-
"epoch": 0.
|
| 5 |
"eval_steps": 100,
|
| 6 |
-
"global_step":
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
@@ -723,6 +723,714 @@
|
|
| 723 |
"eval_samples_per_second": 12.054,
|
| 724 |
"eval_steps_per_second": 3.021,
|
| 725 |
"step": 100
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 726 |
}
|
| 727 |
],
|
| 728 |
"logging_steps": 1,
|
|
@@ -751,7 +1459,7 @@
|
|
| 751 |
"attributes": {}
|
| 752 |
}
|
| 753 |
},
|
| 754 |
-
"total_flos":
|
| 755 |
"train_batch_size": 8,
|
| 756 |
"trial_name": null,
|
| 757 |
"trial_params": null
|
|
|
|
| 1 |
{
|
| 2 |
+
"best_metric": 0.10709268599748611,
|
| 3 |
+
"best_model_checkpoint": "miner_id_24/checkpoint-200",
|
| 4 |
+
"epoch": 0.8705114254624592,
|
| 5 |
"eval_steps": 100,
|
| 6 |
+
"global_step": 200,
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
|
|
| 723 |
"eval_samples_per_second": 12.054,
|
| 724 |
"eval_steps_per_second": 3.021,
|
| 725 |
"step": 100
|
| 726 |
+
},
|
| 727 |
+
{
|
| 728 |
+
"epoch": 0.4396082698585419,
|
| 729 |
+
"grad_norm": 0.08145804703235626,
|
| 730 |
+
"learning_rate": 6.756874120406714e-05,
|
| 731 |
+
"loss": 0.0543,
|
| 732 |
+
"step": 101
|
| 733 |
+
},
|
| 734 |
+
{
|
| 735 |
+
"epoch": 0.4439608269858542,
|
| 736 |
+
"grad_norm": 0.08369223773479462,
|
| 737 |
+
"learning_rate": 6.686649936914152e-05,
|
| 738 |
+
"loss": 0.0668,
|
| 739 |
+
"step": 102
|
| 740 |
+
},
|
| 741 |
+
{
|
| 742 |
+
"epoch": 0.44831338411316646,
|
| 743 |
+
"grad_norm": 0.07429920136928558,
|
| 744 |
+
"learning_rate": 6.616048287272301e-05,
|
| 745 |
+
"loss": 0.0694,
|
| 746 |
+
"step": 103
|
| 747 |
+
},
|
| 748 |
+
{
|
| 749 |
+
"epoch": 0.45266594124047876,
|
| 750 |
+
"grad_norm": 0.07733328640460968,
|
| 751 |
+
"learning_rate": 6.545084971874738e-05,
|
| 752 |
+
"loss": 0.0881,
|
| 753 |
+
"step": 104
|
| 754 |
+
},
|
| 755 |
+
{
|
| 756 |
+
"epoch": 0.45701849836779107,
|
| 757 |
+
"grad_norm": 0.07064362615346909,
|
| 758 |
+
"learning_rate": 6.473775872054521e-05,
|
| 759 |
+
"loss": 0.0777,
|
| 760 |
+
"step": 105
|
| 761 |
+
},
|
| 762 |
+
{
|
| 763 |
+
"epoch": 0.4613710554951034,
|
| 764 |
+
"grad_norm": 0.06992431730031967,
|
| 765 |
+
"learning_rate": 6.402136946530014e-05,
|
| 766 |
+
"loss": 0.0802,
|
| 767 |
+
"step": 106
|
| 768 |
+
},
|
| 769 |
+
{
|
| 770 |
+
"epoch": 0.4657236126224157,
|
| 771 |
+
"grad_norm": 0.0750981792807579,
|
| 772 |
+
"learning_rate": 6.330184227833376e-05,
|
| 773 |
+
"loss": 0.0868,
|
| 774 |
+
"step": 107
|
| 775 |
+
},
|
| 776 |
+
{
|
| 777 |
+
"epoch": 0.470076169749728,
|
| 778 |
+
"grad_norm": 0.06660962849855423,
|
| 779 |
+
"learning_rate": 6.257933818722543e-05,
|
| 780 |
+
"loss": 0.0796,
|
| 781 |
+
"step": 108
|
| 782 |
+
},
|
| 783 |
+
{
|
| 784 |
+
"epoch": 0.47442872687704024,
|
| 785 |
+
"grad_norm": 0.06947064399719238,
|
| 786 |
+
"learning_rate": 6.185401888577488e-05,
|
| 787 |
+
"loss": 0.0816,
|
| 788 |
+
"step": 109
|
| 789 |
+
},
|
| 790 |
+
{
|
| 791 |
+
"epoch": 0.47878128400435255,
|
| 792 |
+
"grad_norm": 0.07036959379911423,
|
| 793 |
+
"learning_rate": 6.112604669781572e-05,
|
| 794 |
+
"loss": 0.0802,
|
| 795 |
+
"step": 110
|
| 796 |
+
},
|
| 797 |
+
{
|
| 798 |
+
"epoch": 0.48313384113166485,
|
| 799 |
+
"grad_norm": 0.0723167285323143,
|
| 800 |
+
"learning_rate": 6.0395584540887963e-05,
|
| 801 |
+
"loss": 0.0793,
|
| 802 |
+
"step": 111
|
| 803 |
+
},
|
| 804 |
+
{
|
| 805 |
+
"epoch": 0.48748639825897716,
|
| 806 |
+
"grad_norm": 0.06801431626081467,
|
| 807 |
+
"learning_rate": 5.9662795889777666e-05,
|
| 808 |
+
"loss": 0.0785,
|
| 809 |
+
"step": 112
|
| 810 |
+
},
|
| 811 |
+
{
|
| 812 |
+
"epoch": 0.49183895538628947,
|
| 813 |
+
"grad_norm": 0.06911415606737137,
|
| 814 |
+
"learning_rate": 5.8927844739931834e-05,
|
| 815 |
+
"loss": 0.0724,
|
| 816 |
+
"step": 113
|
| 817 |
+
},
|
| 818 |
+
{
|
| 819 |
+
"epoch": 0.4961915125136017,
|
| 820 |
+
"grad_norm": 0.07668229192495346,
|
| 821 |
+
"learning_rate": 5.819089557075689e-05,
|
| 822 |
+
"loss": 0.0836,
|
| 823 |
+
"step": 114
|
| 824 |
+
},
|
| 825 |
+
{
|
| 826 |
+
"epoch": 0.500544069640914,
|
| 827 |
+
"grad_norm": 0.08050959557294846,
|
| 828 |
+
"learning_rate": 5.745211330880872e-05,
|
| 829 |
+
"loss": 0.0801,
|
| 830 |
+
"step": 115
|
| 831 |
+
},
|
| 832 |
+
{
|
| 833 |
+
"epoch": 0.5048966267682263,
|
| 834 |
+
"grad_norm": 0.07494529336690903,
|
| 835 |
+
"learning_rate": 5.6711663290882776e-05,
|
| 836 |
+
"loss": 0.0771,
|
| 837 |
+
"step": 116
|
| 838 |
+
},
|
| 839 |
+
{
|
| 840 |
+
"epoch": 0.5092491838955386,
|
| 841 |
+
"grad_norm": 0.08858672529459,
|
| 842 |
+
"learning_rate": 5.596971122701221e-05,
|
| 843 |
+
"loss": 0.0914,
|
| 844 |
+
"step": 117
|
| 845 |
+
},
|
| 846 |
+
{
|
| 847 |
+
"epoch": 0.5136017410228509,
|
| 848 |
+
"grad_norm": 0.07994210720062256,
|
| 849 |
+
"learning_rate": 5.522642316338268e-05,
|
| 850 |
+
"loss": 0.0827,
|
| 851 |
+
"step": 118
|
| 852 |
+
},
|
| 853 |
+
{
|
| 854 |
+
"epoch": 0.5179542981501633,
|
| 855 |
+
"grad_norm": 0.09033293277025223,
|
| 856 |
+
"learning_rate": 5.448196544517168e-05,
|
| 857 |
+
"loss": 0.0911,
|
| 858 |
+
"step": 119
|
| 859 |
+
},
|
| 860 |
+
{
|
| 861 |
+
"epoch": 0.5223068552774756,
|
| 862 |
+
"grad_norm": 0.08702769875526428,
|
| 863 |
+
"learning_rate": 5.373650467932122e-05,
|
| 864 |
+
"loss": 0.095,
|
| 865 |
+
"step": 120
|
| 866 |
+
},
|
| 867 |
+
{
|
| 868 |
+
"epoch": 0.5266594124047879,
|
| 869 |
+
"grad_norm": 0.08354932814836502,
|
| 870 |
+
"learning_rate": 5.299020769725172e-05,
|
| 871 |
+
"loss": 0.0756,
|
| 872 |
+
"step": 121
|
| 873 |
+
},
|
| 874 |
+
{
|
| 875 |
+
"epoch": 0.5310119695321001,
|
| 876 |
+
"grad_norm": 0.08543534576892853,
|
| 877 |
+
"learning_rate": 5.2243241517525754e-05,
|
| 878 |
+
"loss": 0.0846,
|
| 879 |
+
"step": 122
|
| 880 |
+
},
|
| 881 |
+
{
|
| 882 |
+
"epoch": 0.5353645266594124,
|
| 883 |
+
"grad_norm": 0.082368865609169,
|
| 884 |
+
"learning_rate": 5.149577330846993e-05,
|
| 885 |
+
"loss": 0.0827,
|
| 886 |
+
"step": 123
|
| 887 |
+
},
|
| 888 |
+
{
|
| 889 |
+
"epoch": 0.5397170837867247,
|
| 890 |
+
"grad_norm": 0.10136285424232483,
|
| 891 |
+
"learning_rate": 5.074797035076319e-05,
|
| 892 |
+
"loss": 0.1007,
|
| 893 |
+
"step": 124
|
| 894 |
+
},
|
| 895 |
+
{
|
| 896 |
+
"epoch": 0.544069640914037,
|
| 897 |
+
"grad_norm": 0.09355714172124863,
|
| 898 |
+
"learning_rate": 5e-05,
|
| 899 |
+
"loss": 0.0842,
|
| 900 |
+
"step": 125
|
| 901 |
+
},
|
| 902 |
+
{
|
| 903 |
+
"epoch": 0.5484221980413493,
|
| 904 |
+
"grad_norm": 0.0911218672990799,
|
| 905 |
+
"learning_rate": 4.925202964923683e-05,
|
| 906 |
+
"loss": 0.0938,
|
| 907 |
+
"step": 126
|
| 908 |
+
},
|
| 909 |
+
{
|
| 910 |
+
"epoch": 0.5527747551686616,
|
| 911 |
+
"grad_norm": 0.10014523565769196,
|
| 912 |
+
"learning_rate": 4.850422669153009e-05,
|
| 913 |
+
"loss": 0.0899,
|
| 914 |
+
"step": 127
|
| 915 |
+
},
|
| 916 |
+
{
|
| 917 |
+
"epoch": 0.5571273122959739,
|
| 918 |
+
"grad_norm": 0.11333120614290237,
|
| 919 |
+
"learning_rate": 4.775675848247427e-05,
|
| 920 |
+
"loss": 0.0994,
|
| 921 |
+
"step": 128
|
| 922 |
+
},
|
| 923 |
+
{
|
| 924 |
+
"epoch": 0.5614798694232862,
|
| 925 |
+
"grad_norm": 0.10408790409564972,
|
| 926 |
+
"learning_rate": 4.700979230274829e-05,
|
| 927 |
+
"loss": 0.0997,
|
| 928 |
+
"step": 129
|
| 929 |
+
},
|
| 930 |
+
{
|
| 931 |
+
"epoch": 0.5658324265505985,
|
| 932 |
+
"grad_norm": 0.10925531387329102,
|
| 933 |
+
"learning_rate": 4.626349532067879e-05,
|
| 934 |
+
"loss": 0.1019,
|
| 935 |
+
"step": 130
|
| 936 |
+
},
|
| 937 |
+
{
|
| 938 |
+
"epoch": 0.5701849836779108,
|
| 939 |
+
"grad_norm": 0.12819620966911316,
|
| 940 |
+
"learning_rate": 4.551803455482833e-05,
|
| 941 |
+
"loss": 0.1163,
|
| 942 |
+
"step": 131
|
| 943 |
+
},
|
| 944 |
+
{
|
| 945 |
+
"epoch": 0.5745375408052231,
|
| 946 |
+
"grad_norm": 0.10823093354701996,
|
| 947 |
+
"learning_rate": 4.477357683661734e-05,
|
| 948 |
+
"loss": 0.0873,
|
| 949 |
+
"step": 132
|
| 950 |
+
},
|
| 951 |
+
{
|
| 952 |
+
"epoch": 0.5788900979325353,
|
| 953 |
+
"grad_norm": 0.13952915370464325,
|
| 954 |
+
"learning_rate": 4.403028877298779e-05,
|
| 955 |
+
"loss": 0.1062,
|
| 956 |
+
"step": 133
|
| 957 |
+
},
|
| 958 |
+
{
|
| 959 |
+
"epoch": 0.5832426550598476,
|
| 960 |
+
"grad_norm": 0.14679069817066193,
|
| 961 |
+
"learning_rate": 4.328833670911724e-05,
|
| 962 |
+
"loss": 0.1156,
|
| 963 |
+
"step": 134
|
| 964 |
+
},
|
| 965 |
+
{
|
| 966 |
+
"epoch": 0.5875952121871599,
|
| 967 |
+
"grad_norm": 0.14709697663784027,
|
| 968 |
+
"learning_rate": 4.254788669119127e-05,
|
| 969 |
+
"loss": 0.1068,
|
| 970 |
+
"step": 135
|
| 971 |
+
},
|
| 972 |
+
{
|
| 973 |
+
"epoch": 0.5919477693144722,
|
| 974 |
+
"grad_norm": 0.1794285625219345,
|
| 975 |
+
"learning_rate": 4.180910442924312e-05,
|
| 976 |
+
"loss": 0.1481,
|
| 977 |
+
"step": 136
|
| 978 |
+
},
|
| 979 |
+
{
|
| 980 |
+
"epoch": 0.5963003264417845,
|
| 981 |
+
"grad_norm": 0.16199396550655365,
|
| 982 |
+
"learning_rate": 4.107215526006817e-05,
|
| 983 |
+
"loss": 0.1008,
|
| 984 |
+
"step": 137
|
| 985 |
+
},
|
| 986 |
+
{
|
| 987 |
+
"epoch": 0.6006528835690969,
|
| 988 |
+
"grad_norm": 0.20580187439918518,
|
| 989 |
+
"learning_rate": 4.0337204110222346e-05,
|
| 990 |
+
"loss": 0.1383,
|
| 991 |
+
"step": 138
|
| 992 |
+
},
|
| 993 |
+
{
|
| 994 |
+
"epoch": 0.6050054406964092,
|
| 995 |
+
"grad_norm": 0.2128140777349472,
|
| 996 |
+
"learning_rate": 3.960441545911204e-05,
|
| 997 |
+
"loss": 0.1014,
|
| 998 |
+
"step": 139
|
| 999 |
+
},
|
| 1000 |
+
{
|
| 1001 |
+
"epoch": 0.6093579978237215,
|
| 1002 |
+
"grad_norm": 0.2550433874130249,
|
| 1003 |
+
"learning_rate": 3.887395330218429e-05,
|
| 1004 |
+
"loss": 0.1657,
|
| 1005 |
+
"step": 140
|
| 1006 |
+
},
|
| 1007 |
+
{
|
| 1008 |
+
"epoch": 0.6137105549510338,
|
| 1009 |
+
"grad_norm": 0.23781460523605347,
|
| 1010 |
+
"learning_rate": 3.814598111422513e-05,
|
| 1011 |
+
"loss": 0.1541,
|
| 1012 |
+
"step": 141
|
| 1013 |
+
},
|
| 1014 |
+
{
|
| 1015 |
+
"epoch": 0.6180631120783461,
|
| 1016 |
+
"grad_norm": 0.2926751375198364,
|
| 1017 |
+
"learning_rate": 3.742066181277458e-05,
|
| 1018 |
+
"loss": 0.1675,
|
| 1019 |
+
"step": 142
|
| 1020 |
+
},
|
| 1021 |
+
{
|
| 1022 |
+
"epoch": 0.6224156692056583,
|
| 1023 |
+
"grad_norm": 0.3387477993965149,
|
| 1024 |
+
"learning_rate": 3.6698157721666246e-05,
|
| 1025 |
+
"loss": 0.2082,
|
| 1026 |
+
"step": 143
|
| 1027 |
+
},
|
| 1028 |
+
{
|
| 1029 |
+
"epoch": 0.6267682263329706,
|
| 1030 |
+
"grad_norm": 0.28316712379455566,
|
| 1031 |
+
"learning_rate": 3.597863053469987e-05,
|
| 1032 |
+
"loss": 0.1789,
|
| 1033 |
+
"step": 144
|
| 1034 |
+
},
|
| 1035 |
+
{
|
| 1036 |
+
"epoch": 0.6311207834602829,
|
| 1037 |
+
"grad_norm": 0.3266338109970093,
|
| 1038 |
+
"learning_rate": 3.5262241279454785e-05,
|
| 1039 |
+
"loss": 0.2029,
|
| 1040 |
+
"step": 145
|
| 1041 |
+
},
|
| 1042 |
+
{
|
| 1043 |
+
"epoch": 0.6354733405875952,
|
| 1044 |
+
"grad_norm": 0.3332321345806122,
|
| 1045 |
+
"learning_rate": 3.4549150281252636e-05,
|
| 1046 |
+
"loss": 0.186,
|
| 1047 |
+
"step": 146
|
| 1048 |
+
},
|
| 1049 |
+
{
|
| 1050 |
+
"epoch": 0.6398258977149075,
|
| 1051 |
+
"grad_norm": 0.3611805737018585,
|
| 1052 |
+
"learning_rate": 3.383951712727701e-05,
|
| 1053 |
+
"loss": 0.1816,
|
| 1054 |
+
"step": 147
|
| 1055 |
+
},
|
| 1056 |
+
{
|
| 1057 |
+
"epoch": 0.6441784548422198,
|
| 1058 |
+
"grad_norm": 0.4147430956363678,
|
| 1059 |
+
"learning_rate": 3.313350063085851e-05,
|
| 1060 |
+
"loss": 0.2133,
|
| 1061 |
+
"step": 148
|
| 1062 |
+
},
|
| 1063 |
+
{
|
| 1064 |
+
"epoch": 0.6485310119695321,
|
| 1065 |
+
"grad_norm": 0.48492154479026794,
|
| 1066 |
+
"learning_rate": 3.243125879593286e-05,
|
| 1067 |
+
"loss": 0.2802,
|
| 1068 |
+
"step": 149
|
| 1069 |
+
},
|
| 1070 |
+
{
|
| 1071 |
+
"epoch": 0.6528835690968444,
|
| 1072 |
+
"grad_norm": 0.7098520994186401,
|
| 1073 |
+
"learning_rate": 3.173294878168025e-05,
|
| 1074 |
+
"loss": 0.2229,
|
| 1075 |
+
"step": 150
|
| 1076 |
+
},
|
| 1077 |
+
{
|
| 1078 |
+
"epoch": 0.6572361262241567,
|
| 1079 |
+
"grad_norm": 0.06466802954673767,
|
| 1080 |
+
"learning_rate": 3.103872686735358e-05,
|
| 1081 |
+
"loss": 0.0642,
|
| 1082 |
+
"step": 151
|
| 1083 |
+
},
|
| 1084 |
+
{
|
| 1085 |
+
"epoch": 0.661588683351469,
|
| 1086 |
+
"grad_norm": 0.07565873116254807,
|
| 1087 |
+
"learning_rate": 3.0348748417303823e-05,
|
| 1088 |
+
"loss": 0.0699,
|
| 1089 |
+
"step": 152
|
| 1090 |
+
},
|
| 1091 |
+
{
|
| 1092 |
+
"epoch": 0.6659412404787813,
|
| 1093 |
+
"grad_norm": 0.07711576670408249,
|
| 1094 |
+
"learning_rate": 2.9663167846209998e-05,
|
| 1095 |
+
"loss": 0.0826,
|
| 1096 |
+
"step": 153
|
| 1097 |
+
},
|
| 1098 |
+
{
|
| 1099 |
+
"epoch": 0.6702937976060935,
|
| 1100 |
+
"grad_norm": 0.06918355822563171,
|
| 1101 |
+
"learning_rate": 2.8982138584521735e-05,
|
| 1102 |
+
"loss": 0.0698,
|
| 1103 |
+
"step": 154
|
| 1104 |
+
},
|
| 1105 |
+
{
|
| 1106 |
+
"epoch": 0.6746463547334058,
|
| 1107 |
+
"grad_norm": 0.07764638215303421,
|
| 1108 |
+
"learning_rate": 2.8305813044122097e-05,
|
| 1109 |
+
"loss": 0.0778,
|
| 1110 |
+
"step": 155
|
| 1111 |
+
},
|
| 1112 |
+
{
|
| 1113 |
+
"epoch": 0.6789989118607181,
|
| 1114 |
+
"grad_norm": 0.06955543160438538,
|
| 1115 |
+
"learning_rate": 2.7634342584218365e-05,
|
| 1116 |
+
"loss": 0.0755,
|
| 1117 |
+
"step": 156
|
| 1118 |
+
},
|
| 1119 |
+
{
|
| 1120 |
+
"epoch": 0.6833514689880305,
|
| 1121 |
+
"grad_norm": 0.07822169363498688,
|
| 1122 |
+
"learning_rate": 2.6967877477468397e-05,
|
| 1123 |
+
"loss": 0.0896,
|
| 1124 |
+
"step": 157
|
| 1125 |
+
},
|
| 1126 |
+
{
|
| 1127 |
+
"epoch": 0.6877040261153428,
|
| 1128 |
+
"grad_norm": 0.07690484076738358,
|
| 1129 |
+
"learning_rate": 2.630656687635007e-05,
|
| 1130 |
+
"loss": 0.0682,
|
| 1131 |
+
"step": 158
|
| 1132 |
+
},
|
| 1133 |
+
{
|
| 1134 |
+
"epoch": 0.6920565832426551,
|
| 1135 |
+
"grad_norm": 0.07423487305641174,
|
| 1136 |
+
"learning_rate": 2.5650558779781635e-05,
|
| 1137 |
+
"loss": 0.0728,
|
| 1138 |
+
"step": 159
|
| 1139 |
+
},
|
| 1140 |
+
{
|
| 1141 |
+
"epoch": 0.6964091403699674,
|
| 1142 |
+
"grad_norm": 0.07923182845115662,
|
| 1143 |
+
"learning_rate": 2.500000000000001e-05,
|
| 1144 |
+
"loss": 0.0781,
|
| 1145 |
+
"step": 160
|
| 1146 |
+
},
|
| 1147 |
+
{
|
| 1148 |
+
"epoch": 0.7007616974972797,
|
| 1149 |
+
"grad_norm": 0.07322314381599426,
|
| 1150 |
+
"learning_rate": 2.43550361297047e-05,
|
| 1151 |
+
"loss": 0.0846,
|
| 1152 |
+
"step": 161
|
| 1153 |
+
},
|
| 1154 |
+
{
|
| 1155 |
+
"epoch": 0.705114254624592,
|
| 1156 |
+
"grad_norm": 0.07659196853637695,
|
| 1157 |
+
"learning_rate": 2.371581150947476e-05,
|
| 1158 |
+
"loss": 0.0837,
|
| 1159 |
+
"step": 162
|
| 1160 |
+
},
|
| 1161 |
+
{
|
| 1162 |
+
"epoch": 0.7094668117519043,
|
| 1163 |
+
"grad_norm": 0.0735752210021019,
|
| 1164 |
+
"learning_rate": 2.3082469195465893e-05,
|
| 1165 |
+
"loss": 0.0797,
|
| 1166 |
+
"step": 163
|
| 1167 |
+
},
|
| 1168 |
+
{
|
| 1169 |
+
"epoch": 0.7138193688792165,
|
| 1170 |
+
"grad_norm": 0.07915199548006058,
|
| 1171 |
+
"learning_rate": 2.245515092739488e-05,
|
| 1172 |
+
"loss": 0.0855,
|
| 1173 |
+
"step": 164
|
| 1174 |
+
},
|
| 1175 |
+
{
|
| 1176 |
+
"epoch": 0.7181719260065288,
|
| 1177 |
+
"grad_norm": 0.07221949100494385,
|
| 1178 |
+
"learning_rate": 2.1833997096818898e-05,
|
| 1179 |
+
"loss": 0.0744,
|
| 1180 |
+
"step": 165
|
| 1181 |
+
},
|
| 1182 |
+
{
|
| 1183 |
+
"epoch": 0.7225244831338411,
|
| 1184 |
+
"grad_norm": 0.08134427666664124,
|
| 1185 |
+
"learning_rate": 2.1219146715716332e-05,
|
| 1186 |
+
"loss": 0.0931,
|
| 1187 |
+
"step": 166
|
| 1188 |
+
},
|
| 1189 |
+
{
|
| 1190 |
+
"epoch": 0.7268770402611534,
|
| 1191 |
+
"grad_norm": 0.08584997057914734,
|
| 1192 |
+
"learning_rate": 2.061073738537635e-05,
|
| 1193 |
+
"loss": 0.0956,
|
| 1194 |
+
"step": 167
|
| 1195 |
+
},
|
| 1196 |
+
{
|
| 1197 |
+
"epoch": 0.7312295973884657,
|
| 1198 |
+
"grad_norm": 0.08483581990003586,
|
| 1199 |
+
"learning_rate": 2.0008905265604316e-05,
|
| 1200 |
+
"loss": 0.0951,
|
| 1201 |
+
"step": 168
|
| 1202 |
+
},
|
| 1203 |
+
{
|
| 1204 |
+
"epoch": 0.735582154515778,
|
| 1205 |
+
"grad_norm": 0.08276212960481644,
|
| 1206 |
+
"learning_rate": 1.9413785044249678e-05,
|
| 1207 |
+
"loss": 0.0957,
|
| 1208 |
+
"step": 169
|
| 1209 |
+
},
|
| 1210 |
+
{
|
| 1211 |
+
"epoch": 0.7399347116430903,
|
| 1212 |
+
"grad_norm": 0.08514475077390671,
|
| 1213 |
+
"learning_rate": 1.8825509907063327e-05,
|
| 1214 |
+
"loss": 0.0888,
|
| 1215 |
+
"step": 170
|
| 1216 |
+
},
|
| 1217 |
+
{
|
| 1218 |
+
"epoch": 0.7442872687704026,
|
| 1219 |
+
"grad_norm": 0.07838297635316849,
|
| 1220 |
+
"learning_rate": 1.8244211507891063e-05,
|
| 1221 |
+
"loss": 0.0737,
|
| 1222 |
+
"step": 171
|
| 1223 |
+
},
|
| 1224 |
+
{
|
| 1225 |
+
"epoch": 0.7486398258977149,
|
| 1226 |
+
"grad_norm": 0.08513329923152924,
|
| 1227 |
+
"learning_rate": 1.7670019939210024e-05,
|
| 1228 |
+
"loss": 0.0859,
|
| 1229 |
+
"step": 172
|
| 1230 |
+
},
|
| 1231 |
+
{
|
| 1232 |
+
"epoch": 0.7529923830250272,
|
| 1233 |
+
"grad_norm": 0.08508553355932236,
|
| 1234 |
+
"learning_rate": 1.7103063703014372e-05,
|
| 1235 |
+
"loss": 0.0912,
|
| 1236 |
+
"step": 173
|
| 1237 |
+
},
|
| 1238 |
+
{
|
| 1239 |
+
"epoch": 0.7573449401523396,
|
| 1240 |
+
"grad_norm": 0.09576783329248428,
|
| 1241 |
+
"learning_rate": 1.6543469682057106e-05,
|
| 1242 |
+
"loss": 0.1088,
|
| 1243 |
+
"step": 174
|
| 1244 |
+
},
|
| 1245 |
+
{
|
| 1246 |
+
"epoch": 0.7616974972796517,
|
| 1247 |
+
"grad_norm": 0.09258586913347244,
|
| 1248 |
+
"learning_rate": 1.599136311145402e-05,
|
| 1249 |
+
"loss": 0.0915,
|
| 1250 |
+
"step": 175
|
| 1251 |
+
},
|
| 1252 |
+
{
|
| 1253 |
+
"epoch": 0.766050054406964,
|
| 1254 |
+
"grad_norm": 0.10096845775842667,
|
| 1255 |
+
"learning_rate": 1.544686755065677e-05,
|
| 1256 |
+
"loss": 0.0967,
|
| 1257 |
+
"step": 176
|
| 1258 |
+
},
|
| 1259 |
+
{
|
| 1260 |
+
"epoch": 0.7704026115342764,
|
| 1261 |
+
"grad_norm": 0.10371614992618561,
|
| 1262 |
+
"learning_rate": 1.4910104855800427e-05,
|
| 1263 |
+
"loss": 0.1021,
|
| 1264 |
+
"step": 177
|
| 1265 |
+
},
|
| 1266 |
+
{
|
| 1267 |
+
"epoch": 0.7747551686615887,
|
| 1268 |
+
"grad_norm": 0.09611905366182327,
|
| 1269 |
+
"learning_rate": 1.438119515243277e-05,
|
| 1270 |
+
"loss": 0.0893,
|
| 1271 |
+
"step": 178
|
| 1272 |
+
},
|
| 1273 |
+
{
|
| 1274 |
+
"epoch": 0.779107725788901,
|
| 1275 |
+
"grad_norm": 0.09496966004371643,
|
| 1276 |
+
"learning_rate": 1.3860256808630428e-05,
|
| 1277 |
+
"loss": 0.0871,
|
| 1278 |
+
"step": 179
|
| 1279 |
+
},
|
| 1280 |
+
{
|
| 1281 |
+
"epoch": 0.7834602829162133,
|
| 1282 |
+
"grad_norm": 0.10661742091178894,
|
| 1283 |
+
"learning_rate": 1.3347406408508695e-05,
|
| 1284 |
+
"loss": 0.0838,
|
| 1285 |
+
"step": 180
|
| 1286 |
+
},
|
| 1287 |
+
{
|
| 1288 |
+
"epoch": 0.7878128400435256,
|
| 1289 |
+
"grad_norm": 0.12279023230075836,
|
| 1290 |
+
"learning_rate": 1.2842758726130283e-05,
|
| 1291 |
+
"loss": 0.0967,
|
| 1292 |
+
"step": 181
|
| 1293 |
+
},
|
| 1294 |
+
{
|
| 1295 |
+
"epoch": 0.7921653971708379,
|
| 1296 |
+
"grad_norm": 0.11622121185064316,
|
| 1297 |
+
"learning_rate": 1.2346426699819458e-05,
|
| 1298 |
+
"loss": 0.0953,
|
| 1299 |
+
"step": 182
|
| 1300 |
+
},
|
| 1301 |
+
{
|
| 1302 |
+
"epoch": 0.7965179542981502,
|
| 1303 |
+
"grad_norm": 0.15825197100639343,
|
| 1304 |
+
"learning_rate": 1.1858521406886675e-05,
|
| 1305 |
+
"loss": 0.1235,
|
| 1306 |
+
"step": 183
|
| 1307 |
+
},
|
| 1308 |
+
{
|
| 1309 |
+
"epoch": 0.8008705114254625,
|
| 1310 |
+
"grad_norm": 0.15962378680706024,
|
| 1311 |
+
"learning_rate": 1.137915203877003e-05,
|
| 1312 |
+
"loss": 0.1179,
|
| 1313 |
+
"step": 184
|
| 1314 |
+
},
|
| 1315 |
+
{
|
| 1316 |
+
"epoch": 0.8052230685527747,
|
| 1317 |
+
"grad_norm": 0.17821140587329865,
|
| 1318 |
+
"learning_rate": 1.090842587659851e-05,
|
| 1319 |
+
"loss": 0.1458,
|
| 1320 |
+
"step": 185
|
| 1321 |
+
},
|
| 1322 |
+
{
|
| 1323 |
+
"epoch": 0.809575625680087,
|
| 1324 |
+
"grad_norm": 0.1665080189704895,
|
| 1325 |
+
"learning_rate": 1.0446448267182952e-05,
|
| 1326 |
+
"loss": 0.1239,
|
| 1327 |
+
"step": 186
|
| 1328 |
+
},
|
| 1329 |
+
{
|
| 1330 |
+
"epoch": 0.8139281828073993,
|
| 1331 |
+
"grad_norm": 0.1886962503194809,
|
| 1332 |
+
"learning_rate": 9.993322599439692e-06,
|
| 1333 |
+
"loss": 0.1275,
|
| 1334 |
+
"step": 187
|
| 1335 |
+
},
|
| 1336 |
+
{
|
| 1337 |
+
"epoch": 0.8182807399347116,
|
| 1338 |
+
"grad_norm": 0.18211908638477325,
|
| 1339 |
+
"learning_rate": 9.549150281252633e-06,
|
| 1340 |
+
"loss": 0.1026,
|
| 1341 |
+
"step": 188
|
| 1342 |
+
},
|
| 1343 |
+
{
|
| 1344 |
+
"epoch": 0.8226332970620239,
|
| 1345 |
+
"grad_norm": 0.24852445721626282,
|
| 1346 |
+
"learning_rate": 9.114030716778433e-06,
|
| 1347 |
+
"loss": 0.1498,
|
| 1348 |
+
"step": 189
|
| 1349 |
+
},
|
| 1350 |
+
{
|
| 1351 |
+
"epoch": 0.8269858541893362,
|
| 1352 |
+
"grad_norm": 0.231397345662117,
|
| 1353 |
+
"learning_rate": 8.688061284200266e-06,
|
| 1354 |
+
"loss": 0.1242,
|
| 1355 |
+
"step": 190
|
| 1356 |
+
},
|
| 1357 |
+
{
|
| 1358 |
+
"epoch": 0.8313384113166485,
|
| 1359 |
+
"grad_norm": 0.26934152841567993,
|
| 1360 |
+
"learning_rate": 8.271337313934869e-06,
|
| 1361 |
+
"loss": 0.1469,
|
| 1362 |
+
"step": 191
|
| 1363 |
+
},
|
| 1364 |
+
{
|
| 1365 |
+
"epoch": 0.8356909684439608,
|
| 1366 |
+
"grad_norm": 0.2680892050266266,
|
| 1367 |
+
"learning_rate": 7.863952067298042e-06,
|
| 1368 |
+
"loss": 0.1441,
|
| 1369 |
+
"step": 192
|
| 1370 |
+
},
|
| 1371 |
+
{
|
| 1372 |
+
"epoch": 0.8400435255712732,
|
| 1373 |
+
"grad_norm": 0.33063918352127075,
|
| 1374 |
+
"learning_rate": 7.465996715633028e-06,
|
| 1375 |
+
"loss": 0.1923,
|
| 1376 |
+
"step": 193
|
| 1377 |
+
},
|
| 1378 |
+
{
|
| 1379 |
+
"epoch": 0.8443960826985855,
|
| 1380 |
+
"grad_norm": 0.2599483132362366,
|
| 1381 |
+
"learning_rate": 7.077560319906695e-06,
|
| 1382 |
+
"loss": 0.1467,
|
| 1383 |
+
"step": 194
|
| 1384 |
+
},
|
| 1385 |
+
{
|
| 1386 |
+
"epoch": 0.8487486398258978,
|
| 1387 |
+
"grad_norm": 0.30372336506843567,
|
| 1388 |
+
"learning_rate": 6.698729810778065e-06,
|
| 1389 |
+
"loss": 0.1614,
|
| 1390 |
+
"step": 195
|
| 1391 |
+
},
|
| 1392 |
+
{
|
| 1393 |
+
"epoch": 0.85310119695321,
|
| 1394 |
+
"grad_norm": 0.36303284764289856,
|
| 1395 |
+
"learning_rate": 6.329589969143518e-06,
|
| 1396 |
+
"loss": 0.2805,
|
| 1397 |
+
"step": 196
|
| 1398 |
+
},
|
| 1399 |
+
{
|
| 1400 |
+
"epoch": 0.8574537540805223,
|
| 1401 |
+
"grad_norm": 0.37215545773506165,
|
| 1402 |
+
"learning_rate": 5.9702234071631e-06,
|
| 1403 |
+
"loss": 0.2837,
|
| 1404 |
+
"step": 197
|
| 1405 |
+
},
|
| 1406 |
+
{
|
| 1407 |
+
"epoch": 0.8618063112078346,
|
| 1408 |
+
"grad_norm": 0.4205995500087738,
|
| 1409 |
+
"learning_rate": 5.620710549772295e-06,
|
| 1410 |
+
"loss": 0.312,
|
| 1411 |
+
"step": 198
|
| 1412 |
+
},
|
| 1413 |
+
{
|
| 1414 |
+
"epoch": 0.8661588683351469,
|
| 1415 |
+
"grad_norm": 0.5164013504981995,
|
| 1416 |
+
"learning_rate": 5.281129616683167e-06,
|
| 1417 |
+
"loss": 0.2912,
|
| 1418 |
+
"step": 199
|
| 1419 |
+
},
|
| 1420 |
+
{
|
| 1421 |
+
"epoch": 0.8705114254624592,
|
| 1422 |
+
"grad_norm": 0.5630713105201721,
|
| 1423 |
+
"learning_rate": 4.951556604879048e-06,
|
| 1424 |
+
"loss": 0.2237,
|
| 1425 |
+
"step": 200
|
| 1426 |
+
},
|
| 1427 |
+
{
|
| 1428 |
+
"epoch": 0.8705114254624592,
|
| 1429 |
+
"eval_loss": 0.10709268599748611,
|
| 1430 |
+
"eval_runtime": 32.1216,
|
| 1431 |
+
"eval_samples_per_second": 12.048,
|
| 1432 |
+
"eval_steps_per_second": 3.02,
|
| 1433 |
+
"step": 200
|
| 1434 |
}
|
| 1435 |
],
|
| 1436 |
"logging_steps": 1,
|
|
|
|
| 1459 |
"attributes": {}
|
| 1460 |
}
|
| 1461 |
},
|
| 1462 |
+
"total_flos": 3.238351480868045e+17,
|
| 1463 |
"train_batch_size": 8,
|
| 1464 |
"trial_name": null,
|
| 1465 |
"trial_params": null
|