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 34456
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:88445dbb7e20143fb2df5f5dec8fd33cb9d160ee172d6e6c7ba158e20db7fb57
|
3 |
size 34456
|
last-checkpoint/optimizer.pt
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 73222
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:68f8668d449a4aa038e29f380206ac99b14e4e110d76326e57ca117a273ffb08
|
3 |
size 73222
|
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 sha256:620c0f0601732d3fb735a43013e0b9fe901a08ce6988cf173943896b7cdd3449
|
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:d2d754412c61116546142914503e7369d0cc35d3c380a07e5218f595d76b6d96
|
3 |
size 1064
|
last-checkpoint/trainer_state.json
CHANGED
@@ -1,9 +1,9 @@
|
|
1 |
{
|
2 |
"best_metric": 11.5,
|
3 |
"best_model_checkpoint": "miner_id_24/checkpoint-50",
|
4 |
-
"epoch": 0.
|
5 |
"eval_steps": 50,
|
6 |
-
"global_step":
|
7 |
"is_hyper_param_search": false,
|
8 |
"is_local_process_zero": true,
|
9 |
"is_world_process_zero": true,
|
@@ -731,6 +731,722 @@
|
|
731 |
"eval_samples_per_second": 217.255,
|
732 |
"eval_steps_per_second": 54.396,
|
733 |
"step": 100
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
734 |
}
|
735 |
],
|
736 |
"logging_steps": 1,
|
@@ -745,7 +1461,7 @@
|
|
745 |
"early_stopping_threshold": 0.0
|
746 |
},
|
747 |
"attributes": {
|
748 |
-
"early_stopping_patience_counter":
|
749 |
}
|
750 |
},
|
751 |
"TrainerControl": {
|
@@ -754,12 +1470,12 @@
|
|
754 |
"should_evaluate": false,
|
755 |
"should_log": false,
|
756 |
"should_save": true,
|
757 |
-
"should_training_stop":
|
758 |
},
|
759 |
"attributes": {}
|
760 |
}
|
761 |
},
|
762 |
-
"total_flos":
|
763 |
"train_batch_size": 8,
|
764 |
"trial_name": null,
|
765 |
"trial_params": null
|
|
|
1 |
{
|
2 |
"best_metric": 11.5,
|
3 |
"best_model_checkpoint": "miner_id_24/checkpoint-50",
|
4 |
+
"epoch": 0.5079365079365079,
|
5 |
"eval_steps": 50,
|
6 |
+
"global_step": 200,
|
7 |
"is_hyper_param_search": false,
|
8 |
"is_local_process_zero": true,
|
9 |
"is_world_process_zero": true,
|
|
|
731 |
"eval_samples_per_second": 217.255,
|
732 |
"eval_steps_per_second": 54.396,
|
733 |
"step": 100
|
734 |
+
},
|
735 |
+
{
|
736 |
+
"epoch": 0.2565079365079365,
|
737 |
+
"grad_norm": 0.00015683531819377095,
|
738 |
+
"learning_rate": 5.330452921628497e-05,
|
739 |
+
"loss": 46.0,
|
740 |
+
"step": 101
|
741 |
+
},
|
742 |
+
{
|
743 |
+
"epoch": 0.259047619047619,
|
744 |
+
"grad_norm": 0.00012049138604197651,
|
745 |
+
"learning_rate": 5.247918773366112e-05,
|
746 |
+
"loss": 46.0,
|
747 |
+
"step": 102
|
748 |
+
},
|
749 |
+
{
|
750 |
+
"epoch": 0.2615873015873016,
|
751 |
+
"grad_norm": 0.00017800556088332087,
|
752 |
+
"learning_rate": 5.165316846586541e-05,
|
753 |
+
"loss": 46.0,
|
754 |
+
"step": 103
|
755 |
+
},
|
756 |
+
{
|
757 |
+
"epoch": 0.26412698412698415,
|
758 |
+
"grad_norm": 0.00026861418155021966,
|
759 |
+
"learning_rate": 5.0826697238317935e-05,
|
760 |
+
"loss": 46.0,
|
761 |
+
"step": 104
|
762 |
+
},
|
763 |
+
{
|
764 |
+
"epoch": 0.26666666666666666,
|
765 |
+
"grad_norm": 0.0003716795181389898,
|
766 |
+
"learning_rate": 5e-05,
|
767 |
+
"loss": 46.0,
|
768 |
+
"step": 105
|
769 |
+
},
|
770 |
+
{
|
771 |
+
"epoch": 0.2692063492063492,
|
772 |
+
"grad_norm": 0.0002833975595422089,
|
773 |
+
"learning_rate": 4.917330276168208e-05,
|
774 |
+
"loss": 46.0,
|
775 |
+
"step": 106
|
776 |
+
},
|
777 |
+
{
|
778 |
+
"epoch": 0.27174603174603174,
|
779 |
+
"grad_norm": 0.0003513269766699523,
|
780 |
+
"learning_rate": 4.834683153413459e-05,
|
781 |
+
"loss": 46.0,
|
782 |
+
"step": 107
|
783 |
+
},
|
784 |
+
{
|
785 |
+
"epoch": 0.2742857142857143,
|
786 |
+
"grad_norm": 0.00023855481413193047,
|
787 |
+
"learning_rate": 4.7520812266338885e-05,
|
788 |
+
"loss": 46.0,
|
789 |
+
"step": 108
|
790 |
+
},
|
791 |
+
{
|
792 |
+
"epoch": 0.2768253968253968,
|
793 |
+
"grad_norm": 0.00023193252854980528,
|
794 |
+
"learning_rate": 4.669547078371504e-05,
|
795 |
+
"loss": 46.0,
|
796 |
+
"step": 109
|
797 |
+
},
|
798 |
+
{
|
799 |
+
"epoch": 0.27936507936507937,
|
800 |
+
"grad_norm": 0.00029786513186991215,
|
801 |
+
"learning_rate": 4.5871032726383386e-05,
|
802 |
+
"loss": 46.0,
|
803 |
+
"step": 110
|
804 |
+
},
|
805 |
+
{
|
806 |
+
"epoch": 0.2819047619047619,
|
807 |
+
"grad_norm": 0.00025254423962906003,
|
808 |
+
"learning_rate": 4.504772348747687e-05,
|
809 |
+
"loss": 46.0,
|
810 |
+
"step": 111
|
811 |
+
},
|
812 |
+
{
|
813 |
+
"epoch": 0.28444444444444444,
|
814 |
+
"grad_norm": 0.000365566520486027,
|
815 |
+
"learning_rate": 4.4225768151520694e-05,
|
816 |
+
"loss": 46.0,
|
817 |
+
"step": 112
|
818 |
+
},
|
819 |
+
{
|
820 |
+
"epoch": 0.286984126984127,
|
821 |
+
"grad_norm": 0.00027666729874908924,
|
822 |
+
"learning_rate": 4.3405391432896555e-05,
|
823 |
+
"loss": 46.0,
|
824 |
+
"step": 113
|
825 |
+
},
|
826 |
+
{
|
827 |
+
"epoch": 0.2895238095238095,
|
828 |
+
"grad_norm": 0.00039132856181822717,
|
829 |
+
"learning_rate": 4.2586817614407895e-05,
|
830 |
+
"loss": 46.0,
|
831 |
+
"step": 114
|
832 |
+
},
|
833 |
+
{
|
834 |
+
"epoch": 0.2920634920634921,
|
835 |
+
"grad_norm": 0.00044664432061836123,
|
836 |
+
"learning_rate": 4.17702704859633e-05,
|
837 |
+
"loss": 46.0,
|
838 |
+
"step": 115
|
839 |
+
},
|
840 |
+
{
|
841 |
+
"epoch": 0.2946031746031746,
|
842 |
+
"grad_norm": 0.00042760182986967266,
|
843 |
+
"learning_rate": 4.095597328339452e-05,
|
844 |
+
"loss": 46.0,
|
845 |
+
"step": 116
|
846 |
+
},
|
847 |
+
{
|
848 |
+
"epoch": 0.29714285714285715,
|
849 |
+
"grad_norm": 0.0003109555982518941,
|
850 |
+
"learning_rate": 4.0144148627425993e-05,
|
851 |
+
"loss": 46.0,
|
852 |
+
"step": 117
|
853 |
+
},
|
854 |
+
{
|
855 |
+
"epoch": 0.29968253968253966,
|
856 |
+
"grad_norm": 0.0005549000925384462,
|
857 |
+
"learning_rate": 3.933501846281267e-05,
|
858 |
+
"loss": 46.0,
|
859 |
+
"step": 118
|
860 |
+
},
|
861 |
+
{
|
862 |
+
"epoch": 0.3022222222222222,
|
863 |
+
"grad_norm": 0.00037709763273596764,
|
864 |
+
"learning_rate": 3.852880399766243e-05,
|
865 |
+
"loss": 46.0,
|
866 |
+
"step": 119
|
867 |
+
},
|
868 |
+
{
|
869 |
+
"epoch": 0.3047619047619048,
|
870 |
+
"grad_norm": 0.0005275582661852241,
|
871 |
+
"learning_rate": 3.772572564296005e-05,
|
872 |
+
"loss": 46.0,
|
873 |
+
"step": 120
|
874 |
+
},
|
875 |
+
{
|
876 |
+
"epoch": 0.3073015873015873,
|
877 |
+
"grad_norm": 0.0006868797354400158,
|
878 |
+
"learning_rate": 3.6926002952309016e-05,
|
879 |
+
"loss": 46.0,
|
880 |
+
"step": 121
|
881 |
+
},
|
882 |
+
{
|
883 |
+
"epoch": 0.30984126984126986,
|
884 |
+
"grad_norm": 0.0003916198038496077,
|
885 |
+
"learning_rate": 3.612985456190778e-05,
|
886 |
+
"loss": 46.0,
|
887 |
+
"step": 122
|
888 |
+
},
|
889 |
+
{
|
890 |
+
"epoch": 0.31238095238095237,
|
891 |
+
"grad_norm": 0.00041374217835254967,
|
892 |
+
"learning_rate": 3.533749813077677e-05,
|
893 |
+
"loss": 46.0,
|
894 |
+
"step": 123
|
895 |
+
},
|
896 |
+
{
|
897 |
+
"epoch": 0.31492063492063493,
|
898 |
+
"grad_norm": 0.00048302829964086413,
|
899 |
+
"learning_rate": 3.4549150281252636e-05,
|
900 |
+
"loss": 46.0,
|
901 |
+
"step": 124
|
902 |
+
},
|
903 |
+
{
|
904 |
+
"epoch": 0.31746031746031744,
|
905 |
+
"grad_norm": 0.000454338121926412,
|
906 |
+
"learning_rate": 3.3765026539765834e-05,
|
907 |
+
"loss": 46.0,
|
908 |
+
"step": 125
|
909 |
+
},
|
910 |
+
{
|
911 |
+
"epoch": 0.32,
|
912 |
+
"grad_norm": 0.000542081892490387,
|
913 |
+
"learning_rate": 3.298534127791785e-05,
|
914 |
+
"loss": 46.0,
|
915 |
+
"step": 126
|
916 |
+
},
|
917 |
+
{
|
918 |
+
"epoch": 0.3225396825396825,
|
919 |
+
"grad_norm": 0.000495821819640696,
|
920 |
+
"learning_rate": 3.221030765387417e-05,
|
921 |
+
"loss": 46.0,
|
922 |
+
"step": 127
|
923 |
+
},
|
924 |
+
{
|
925 |
+
"epoch": 0.3250793650793651,
|
926 |
+
"grad_norm": 0.0005824922118335962,
|
927 |
+
"learning_rate": 3.144013755408895e-05,
|
928 |
+
"loss": 46.0,
|
929 |
+
"step": 128
|
930 |
+
},
|
931 |
+
{
|
932 |
+
"epoch": 0.32761904761904764,
|
933 |
+
"grad_norm": 0.0007732115336693823,
|
934 |
+
"learning_rate": 3.0675041535377405e-05,
|
935 |
+
"loss": 46.0,
|
936 |
+
"step": 129
|
937 |
+
},
|
938 |
+
{
|
939 |
+
"epoch": 0.33015873015873015,
|
940 |
+
"grad_norm": 0.0003601844946388155,
|
941 |
+
"learning_rate": 2.991522876735154e-05,
|
942 |
+
"loss": 46.0,
|
943 |
+
"step": 130
|
944 |
+
},
|
945 |
+
{
|
946 |
+
"epoch": 0.3326984126984127,
|
947 |
+
"grad_norm": 0.0007275976240634918,
|
948 |
+
"learning_rate": 2.916090697523549e-05,
|
949 |
+
"loss": 46.0,
|
950 |
+
"step": 131
|
951 |
+
},
|
952 |
+
{
|
953 |
+
"epoch": 0.3352380952380952,
|
954 |
+
"grad_norm": 0.0007252413779497147,
|
955 |
+
"learning_rate": 2.8412282383075363e-05,
|
956 |
+
"loss": 46.0,
|
957 |
+
"step": 132
|
958 |
+
},
|
959 |
+
{
|
960 |
+
"epoch": 0.3377777777777778,
|
961 |
+
"grad_norm": 0.0006040868465788662,
|
962 |
+
"learning_rate": 2.766955965735968e-05,
|
963 |
+
"loss": 46.0,
|
964 |
+
"step": 133
|
965 |
+
},
|
966 |
+
{
|
967 |
+
"epoch": 0.3403174603174603,
|
968 |
+
"grad_norm": 0.0007312872330658138,
|
969 |
+
"learning_rate": 2.693294185106562e-05,
|
970 |
+
"loss": 46.0,
|
971 |
+
"step": 134
|
972 |
+
},
|
973 |
+
{
|
974 |
+
"epoch": 0.34285714285714286,
|
975 |
+
"grad_norm": 0.0006050340598449111,
|
976 |
+
"learning_rate": 2.6202630348146324e-05,
|
977 |
+
"loss": 46.0,
|
978 |
+
"step": 135
|
979 |
+
},
|
980 |
+
{
|
981 |
+
"epoch": 0.3453968253968254,
|
982 |
+
"grad_norm": 0.0005079463589936495,
|
983 |
+
"learning_rate": 2.547882480847461e-05,
|
984 |
+
"loss": 46.0,
|
985 |
+
"step": 136
|
986 |
+
},
|
987 |
+
{
|
988 |
+
"epoch": 0.34793650793650793,
|
989 |
+
"grad_norm": 0.0007020235061645508,
|
990 |
+
"learning_rate": 2.476172311325783e-05,
|
991 |
+
"loss": 46.0,
|
992 |
+
"step": 137
|
993 |
+
},
|
994 |
+
{
|
995 |
+
"epoch": 0.3504761904761905,
|
996 |
+
"grad_norm": 0.00045382286771200597,
|
997 |
+
"learning_rate": 2.405152131093926e-05,
|
998 |
+
"loss": 46.0,
|
999 |
+
"step": 138
|
1000 |
+
},
|
1001 |
+
{
|
1002 |
+
"epoch": 0.353015873015873,
|
1003 |
+
"grad_norm": 0.000577699625864625,
|
1004 |
+
"learning_rate": 2.3348413563600325e-05,
|
1005 |
+
"loss": 46.0,
|
1006 |
+
"step": 139
|
1007 |
+
},
|
1008 |
+
{
|
1009 |
+
"epoch": 0.35555555555555557,
|
1010 |
+
"grad_norm": 0.0006426837062463164,
|
1011 |
+
"learning_rate": 2.2652592093878666e-05,
|
1012 |
+
"loss": 46.0,
|
1013 |
+
"step": 140
|
1014 |
+
},
|
1015 |
+
{
|
1016 |
+
"epoch": 0.3580952380952381,
|
1017 |
+
"grad_norm": 0.0007479987107217312,
|
1018 |
+
"learning_rate": 2.196424713241637e-05,
|
1019 |
+
"loss": 46.0,
|
1020 |
+
"step": 141
|
1021 |
+
},
|
1022 |
+
{
|
1023 |
+
"epoch": 0.36063492063492064,
|
1024 |
+
"grad_norm": 0.0011289599351584911,
|
1025 |
+
"learning_rate": 2.128356686585282e-05,
|
1026 |
+
"loss": 46.0,
|
1027 |
+
"step": 142
|
1028 |
+
},
|
1029 |
+
{
|
1030 |
+
"epoch": 0.36317460317460315,
|
1031 |
+
"grad_norm": 0.0007505420944653451,
|
1032 |
+
"learning_rate": 2.061073738537635e-05,
|
1033 |
+
"loss": 46.0,
|
1034 |
+
"step": 143
|
1035 |
+
},
|
1036 |
+
{
|
1037 |
+
"epoch": 0.3657142857142857,
|
1038 |
+
"grad_norm": 0.0009757123771123588,
|
1039 |
+
"learning_rate": 1.9945942635848748e-05,
|
1040 |
+
"loss": 46.0,
|
1041 |
+
"step": 144
|
1042 |
+
},
|
1043 |
+
{
|
1044 |
+
"epoch": 0.3682539682539683,
|
1045 |
+
"grad_norm": 0.0010634862119331956,
|
1046 |
+
"learning_rate": 1.928936436551661e-05,
|
1047 |
+
"loss": 46.0,
|
1048 |
+
"step": 145
|
1049 |
+
},
|
1050 |
+
{
|
1051 |
+
"epoch": 0.3707936507936508,
|
1052 |
+
"grad_norm": 0.0010427199304103851,
|
1053 |
+
"learning_rate": 1.8641182076323148e-05,
|
1054 |
+
"loss": 46.0,
|
1055 |
+
"step": 146
|
1056 |
+
},
|
1057 |
+
{
|
1058 |
+
"epoch": 0.37333333333333335,
|
1059 |
+
"grad_norm": 0.0013985201949253678,
|
1060 |
+
"learning_rate": 1.800157297483417e-05,
|
1061 |
+
"loss": 46.0,
|
1062 |
+
"step": 147
|
1063 |
+
},
|
1064 |
+
{
|
1065 |
+
"epoch": 0.37587301587301586,
|
1066 |
+
"grad_norm": 0.0008832611492834985,
|
1067 |
+
"learning_rate": 1.7370711923791567e-05,
|
1068 |
+
"loss": 46.0,
|
1069 |
+
"step": 148
|
1070 |
+
},
|
1071 |
+
{
|
1072 |
+
"epoch": 0.3784126984126984,
|
1073 |
+
"grad_norm": 0.001664287643507123,
|
1074 |
+
"learning_rate": 1.6748771394307585e-05,
|
1075 |
+
"loss": 46.0,
|
1076 |
+
"step": 149
|
1077 |
+
},
|
1078 |
+
{
|
1079 |
+
"epoch": 0.38095238095238093,
|
1080 |
+
"grad_norm": 0.0016788316424936056,
|
1081 |
+
"learning_rate": 1.6135921418712956e-05,
|
1082 |
+
"loss": 46.0,
|
1083 |
+
"step": 150
|
1084 |
+
},
|
1085 |
+
{
|
1086 |
+
"epoch": 0.38095238095238093,
|
1087 |
+
"eval_loss": 11.5,
|
1088 |
+
"eval_runtime": 3.0583,
|
1089 |
+
"eval_samples_per_second": 216.79,
|
1090 |
+
"eval_steps_per_second": 54.279,
|
1091 |
+
"step": 150
|
1092 |
+
},
|
1093 |
+
{
|
1094 |
+
"epoch": 0.3834920634920635,
|
1095 |
+
"grad_norm": 0.0001507675915490836,
|
1096 |
+
"learning_rate": 1.553232954407171e-05,
|
1097 |
+
"loss": 46.0,
|
1098 |
+
"step": 151
|
1099 |
+
},
|
1100 |
+
{
|
1101 |
+
"epoch": 0.38603174603174606,
|
1102 |
+
"grad_norm": 0.00015799202083144337,
|
1103 |
+
"learning_rate": 1.4938160786375572e-05,
|
1104 |
+
"loss": 46.0,
|
1105 |
+
"step": 152
|
1106 |
+
},
|
1107 |
+
{
|
1108 |
+
"epoch": 0.38857142857142857,
|
1109 |
+
"grad_norm": 0.0002981535217259079,
|
1110 |
+
"learning_rate": 1.435357758543015e-05,
|
1111 |
+
"loss": 46.0,
|
1112 |
+
"step": 153
|
1113 |
+
},
|
1114 |
+
{
|
1115 |
+
"epoch": 0.39111111111111113,
|
1116 |
+
"grad_norm": 0.0005736402818001807,
|
1117 |
+
"learning_rate": 1.3778739760445552e-05,
|
1118 |
+
"loss": 46.0,
|
1119 |
+
"step": 154
|
1120 |
+
},
|
1121 |
+
{
|
1122 |
+
"epoch": 0.39365079365079364,
|
1123 |
+
"grad_norm": 0.00046769826440140605,
|
1124 |
+
"learning_rate": 1.3213804466343421e-05,
|
1125 |
+
"loss": 46.0,
|
1126 |
+
"step": 155
|
1127 |
+
},
|
1128 |
+
{
|
1129 |
+
"epoch": 0.3961904761904762,
|
1130 |
+
"grad_norm": 0.0005333162262104452,
|
1131 |
+
"learning_rate": 1.2658926150792322e-05,
|
1132 |
+
"loss": 46.0,
|
1133 |
+
"step": 156
|
1134 |
+
},
|
1135 |
+
{
|
1136 |
+
"epoch": 0.3987301587301587,
|
1137 |
+
"grad_norm": 0.0004069180286023766,
|
1138 |
+
"learning_rate": 1.2114256511983274e-05,
|
1139 |
+
"loss": 46.0,
|
1140 |
+
"step": 157
|
1141 |
+
},
|
1142 |
+
{
|
1143 |
+
"epoch": 0.4012698412698413,
|
1144 |
+
"grad_norm": 0.00045734853483736515,
|
1145 |
+
"learning_rate": 1.157994445715706e-05,
|
1146 |
+
"loss": 46.0,
|
1147 |
+
"step": 158
|
1148 |
+
},
|
1149 |
+
{
|
1150 |
+
"epoch": 0.4038095238095238,
|
1151 |
+
"grad_norm": 0.00038881177897565067,
|
1152 |
+
"learning_rate": 1.1056136061894384e-05,
|
1153 |
+
"loss": 46.0,
|
1154 |
+
"step": 159
|
1155 |
+
},
|
1156 |
+
{
|
1157 |
+
"epoch": 0.40634920634920635,
|
1158 |
+
"grad_norm": 0.0006046105409041047,
|
1159 |
+
"learning_rate": 1.0542974530180327e-05,
|
1160 |
+
"loss": 46.0,
|
1161 |
+
"step": 160
|
1162 |
+
},
|
1163 |
+
{
|
1164 |
+
"epoch": 0.4088888888888889,
|
1165 |
+
"grad_norm": 0.0008934616926126182,
|
1166 |
+
"learning_rate": 1.0040600155253765e-05,
|
1167 |
+
"loss": 46.0,
|
1168 |
+
"step": 161
|
1169 |
+
},
|
1170 |
+
{
|
1171 |
+
"epoch": 0.4114285714285714,
|
1172 |
+
"grad_norm": 0.0006283086840994656,
|
1173 |
+
"learning_rate": 9.549150281252633e-06,
|
1174 |
+
"loss": 46.0,
|
1175 |
+
"step": 162
|
1176 |
+
},
|
1177 |
+
{
|
1178 |
+
"epoch": 0.413968253968254,
|
1179 |
+
"grad_norm": 0.0002501502458471805,
|
1180 |
+
"learning_rate": 9.068759265665384e-06,
|
1181 |
+
"loss": 46.0,
|
1182 |
+
"step": 163
|
1183 |
+
},
|
1184 |
+
{
|
1185 |
+
"epoch": 0.4165079365079365,
|
1186 |
+
"grad_norm": 0.000643797277007252,
|
1187 |
+
"learning_rate": 8.599558442598998e-06,
|
1188 |
+
"loss": 46.0,
|
1189 |
+
"step": 164
|
1190 |
+
},
|
1191 |
+
{
|
1192 |
+
"epoch": 0.41904761904761906,
|
1193 |
+
"grad_norm": 0.00046031048987060785,
|
1194 |
+
"learning_rate": 8.141676086873572e-06,
|
1195 |
+
"loss": 46.0,
|
1196 |
+
"step": 165
|
1197 |
+
},
|
1198 |
+
{
|
1199 |
+
"epoch": 0.42158730158730157,
|
1200 |
+
"grad_norm": 0.00045485360897146165,
|
1201 |
+
"learning_rate": 7.695237378953223e-06,
|
1202 |
+
"loss": 46.0,
|
1203 |
+
"step": 166
|
1204 |
+
},
|
1205 |
+
{
|
1206 |
+
"epoch": 0.42412698412698413,
|
1207 |
+
"grad_norm": 0.00041288204374723136,
|
1208 |
+
"learning_rate": 7.260364370723044e-06,
|
1209 |
+
"loss": 46.0,
|
1210 |
+
"step": 167
|
1211 |
+
},
|
1212 |
+
{
|
1213 |
+
"epoch": 0.4266666666666667,
|
1214 |
+
"grad_norm": 0.0003486561181489378,
|
1215 |
+
"learning_rate": 6.837175952121306e-06,
|
1216 |
+
"loss": 46.0,
|
1217 |
+
"step": 168
|
1218 |
+
},
|
1219 |
+
{
|
1220 |
+
"epoch": 0.4292063492063492,
|
1221 |
+
"grad_norm": 0.00033542406163178384,
|
1222 |
+
"learning_rate": 6.425787818636131e-06,
|
1223 |
+
"loss": 46.0,
|
1224 |
+
"step": 169
|
1225 |
+
},
|
1226 |
+
{
|
1227 |
+
"epoch": 0.43174603174603177,
|
1228 |
+
"grad_norm": 0.00043678382644429803,
|
1229 |
+
"learning_rate": 6.026312439675552e-06,
|
1230 |
+
"loss": 46.0,
|
1231 |
+
"step": 170
|
1232 |
+
},
|
1233 |
+
{
|
1234 |
+
"epoch": 0.4342857142857143,
|
1235 |
+
"grad_norm": 0.0004820231406483799,
|
1236 |
+
"learning_rate": 5.6388590278194096e-06,
|
1237 |
+
"loss": 46.0,
|
1238 |
+
"step": 171
|
1239 |
+
},
|
1240 |
+
{
|
1241 |
+
"epoch": 0.43682539682539684,
|
1242 |
+
"grad_norm": 0.0006722785765305161,
|
1243 |
+
"learning_rate": 5.263533508961827e-06,
|
1244 |
+
"loss": 46.0,
|
1245 |
+
"step": 172
|
1246 |
+
},
|
1247 |
+
{
|
1248 |
+
"epoch": 0.43936507936507935,
|
1249 |
+
"grad_norm": 0.0005217442521825433,
|
1250 |
+
"learning_rate": 4.900438493352055e-06,
|
1251 |
+
"loss": 46.0,
|
1252 |
+
"step": 173
|
1253 |
+
},
|
1254 |
+
{
|
1255 |
+
"epoch": 0.4419047619047619,
|
1256 |
+
"grad_norm": 0.0005299354088492692,
|
1257 |
+
"learning_rate": 4.549673247541875e-06,
|
1258 |
+
"loss": 46.0,
|
1259 |
+
"step": 174
|
1260 |
+
},
|
1261 |
+
{
|
1262 |
+
"epoch": 0.4444444444444444,
|
1263 |
+
"grad_norm": 0.0006642525549978018,
|
1264 |
+
"learning_rate": 4.2113336672471245e-06,
|
1265 |
+
"loss": 46.0,
|
1266 |
+
"step": 175
|
1267 |
+
},
|
1268 |
+
{
|
1269 |
+
"epoch": 0.446984126984127,
|
1270 |
+
"grad_norm": 0.0006187099497765303,
|
1271 |
+
"learning_rate": 3.885512251130763e-06,
|
1272 |
+
"loss": 46.0,
|
1273 |
+
"step": 176
|
1274 |
+
},
|
1275 |
+
{
|
1276 |
+
"epoch": 0.44952380952380955,
|
1277 |
+
"grad_norm": 0.0008217595750465989,
|
1278 |
+
"learning_rate": 3.5722980755146517e-06,
|
1279 |
+
"loss": 46.0,
|
1280 |
+
"step": 177
|
1281 |
+
},
|
1282 |
+
{
|
1283 |
+
"epoch": 0.45206349206349206,
|
1284 |
+
"grad_norm": 0.0006930683157406747,
|
1285 |
+
"learning_rate": 3.271776770026963e-06,
|
1286 |
+
"loss": 46.0,
|
1287 |
+
"step": 178
|
1288 |
+
},
|
1289 |
+
{
|
1290 |
+
"epoch": 0.4546031746031746,
|
1291 |
+
"grad_norm": 0.0007860871846787632,
|
1292 |
+
"learning_rate": 2.9840304941919415e-06,
|
1293 |
+
"loss": 46.0,
|
1294 |
+
"step": 179
|
1295 |
+
},
|
1296 |
+
{
|
1297 |
+
"epoch": 0.45714285714285713,
|
1298 |
+
"grad_norm": 0.0006492060492746532,
|
1299 |
+
"learning_rate": 2.7091379149682685e-06,
|
1300 |
+
"loss": 46.0,
|
1301 |
+
"step": 180
|
1302 |
+
},
|
1303 |
+
{
|
1304 |
+
"epoch": 0.4596825396825397,
|
1305 |
+
"grad_norm": 0.0005901999538764358,
|
1306 |
+
"learning_rate": 2.4471741852423237e-06,
|
1307 |
+
"loss": 46.0,
|
1308 |
+
"step": 181
|
1309 |
+
},
|
1310 |
+
{
|
1311 |
+
"epoch": 0.4622222222222222,
|
1312 |
+
"grad_norm": 0.0006238595233298838,
|
1313 |
+
"learning_rate": 2.1982109232821178e-06,
|
1314 |
+
"loss": 46.0,
|
1315 |
+
"step": 182
|
1316 |
+
},
|
1317 |
+
{
|
1318 |
+
"epoch": 0.46476190476190476,
|
1319 |
+
"grad_norm": 0.0008841100498102605,
|
1320 |
+
"learning_rate": 1.962316193157593e-06,
|
1321 |
+
"loss": 46.0,
|
1322 |
+
"step": 183
|
1323 |
+
},
|
1324 |
+
{
|
1325 |
+
"epoch": 0.46730158730158733,
|
1326 |
+
"grad_norm": 0.0005608927458524704,
|
1327 |
+
"learning_rate": 1.7395544861325718e-06,
|
1328 |
+
"loss": 46.0,
|
1329 |
+
"step": 184
|
1330 |
+
},
|
1331 |
+
{
|
1332 |
+
"epoch": 0.46984126984126984,
|
1333 |
+
"grad_norm": 0.0006603376241400838,
|
1334 |
+
"learning_rate": 1.5299867030334814e-06,
|
1335 |
+
"loss": 46.0,
|
1336 |
+
"step": 185
|
1337 |
+
},
|
1338 |
+
{
|
1339 |
+
"epoch": 0.4723809523809524,
|
1340 |
+
"grad_norm": 0.0008104827720671892,
|
1341 |
+
"learning_rate": 1.333670137599713e-06,
|
1342 |
+
"loss": 46.0,
|
1343 |
+
"step": 186
|
1344 |
+
},
|
1345 |
+
{
|
1346 |
+
"epoch": 0.4749206349206349,
|
1347 |
+
"grad_norm": 0.0008827261626720428,
|
1348 |
+
"learning_rate": 1.1506584608200367e-06,
|
1349 |
+
"loss": 46.0,
|
1350 |
+
"step": 187
|
1351 |
+
},
|
1352 |
+
{
|
1353 |
+
"epoch": 0.4774603174603175,
|
1354 |
+
"grad_norm": 0.000377334508812055,
|
1355 |
+
"learning_rate": 9.810017062595322e-07,
|
1356 |
+
"loss": 46.0,
|
1357 |
+
"step": 188
|
1358 |
+
},
|
1359 |
+
{
|
1360 |
+
"epoch": 0.48,
|
1361 |
+
"grad_norm": 0.0011284862412139773,
|
1362 |
+
"learning_rate": 8.247462563808817e-07,
|
1363 |
+
"loss": 46.0,
|
1364 |
+
"step": 189
|
1365 |
+
},
|
1366 |
+
{
|
1367 |
+
"epoch": 0.48253968253968255,
|
1368 |
+
"grad_norm": 0.0005800533108413219,
|
1369 |
+
"learning_rate": 6.819348298638839e-07,
|
1370 |
+
"loss": 46.0,
|
1371 |
+
"step": 190
|
1372 |
+
},
|
1373 |
+
{
|
1374 |
+
"epoch": 0.48507936507936505,
|
1375 |
+
"grad_norm": 0.0010874561266973615,
|
1376 |
+
"learning_rate": 5.526064699265753e-07,
|
1377 |
+
"loss": 46.0,
|
1378 |
+
"step": 191
|
1379 |
+
},
|
1380 |
+
{
|
1381 |
+
"epoch": 0.4876190476190476,
|
1382 |
+
"grad_norm": 0.0008321768254972994,
|
1383 |
+
"learning_rate": 4.367965336512403e-07,
|
1384 |
+
"loss": 46.0,
|
1385 |
+
"step": 192
|
1386 |
+
},
|
1387 |
+
{
|
1388 |
+
"epoch": 0.4901587301587302,
|
1389 |
+
"grad_norm": 0.0008640267769806087,
|
1390 |
+
"learning_rate": 3.3453668231809286e-07,
|
1391 |
+
"loss": 46.0,
|
1392 |
+
"step": 193
|
1393 |
+
},
|
1394 |
+
{
|
1395 |
+
"epoch": 0.4926984126984127,
|
1396 |
+
"grad_norm": 0.001046935678459704,
|
1397 |
+
"learning_rate": 2.458548727494292e-07,
|
1398 |
+
"loss": 46.0,
|
1399 |
+
"step": 194
|
1400 |
+
},
|
1401 |
+
{
|
1402 |
+
"epoch": 0.49523809523809526,
|
1403 |
+
"grad_norm": 0.0009591786074452102,
|
1404 |
+
"learning_rate": 1.7077534966650766e-07,
|
1405 |
+
"loss": 46.0,
|
1406 |
+
"step": 195
|
1407 |
+
},
|
1408 |
+
{
|
1409 |
+
"epoch": 0.49777777777777776,
|
1410 |
+
"grad_norm": 0.00164878461509943,
|
1411 |
+
"learning_rate": 1.0931863906127327e-07,
|
1412 |
+
"loss": 46.0,
|
1413 |
+
"step": 196
|
1414 |
+
},
|
1415 |
+
{
|
1416 |
+
"epoch": 0.5003174603174603,
|
1417 |
+
"grad_norm": 0.001384081901051104,
|
1418 |
+
"learning_rate": 6.150154258476315e-08,
|
1419 |
+
"loss": 46.0,
|
1420 |
+
"step": 197
|
1421 |
+
},
|
1422 |
+
{
|
1423 |
+
"epoch": 0.5028571428571429,
|
1424 |
+
"grad_norm": 0.001185267698019743,
|
1425 |
+
"learning_rate": 2.7337132953697554e-08,
|
1426 |
+
"loss": 46.0,
|
1427 |
+
"step": 198
|
1428 |
+
},
|
1429 |
+
{
|
1430 |
+
"epoch": 0.5053968253968254,
|
1431 |
+
"grad_norm": 0.0015531462850049138,
|
1432 |
+
"learning_rate": 6.834750376549792e-09,
|
1433 |
+
"loss": 46.0,
|
1434 |
+
"step": 199
|
1435 |
+
},
|
1436 |
+
{
|
1437 |
+
"epoch": 0.5079365079365079,
|
1438 |
+
"grad_norm": 0.0017203431343659759,
|
1439 |
+
"learning_rate": 0.0,
|
1440 |
+
"loss": 46.0,
|
1441 |
+
"step": 200
|
1442 |
+
},
|
1443 |
+
{
|
1444 |
+
"epoch": 0.5079365079365079,
|
1445 |
+
"eval_loss": 11.5,
|
1446 |
+
"eval_runtime": 3.0566,
|
1447 |
+
"eval_samples_per_second": 216.91,
|
1448 |
+
"eval_steps_per_second": 54.309,
|
1449 |
+
"step": 200
|
1450 |
}
|
1451 |
],
|
1452 |
"logging_steps": 1,
|
|
|
1461 |
"early_stopping_threshold": 0.0
|
1462 |
},
|
1463 |
"attributes": {
|
1464 |
+
"early_stopping_patience_counter": 3
|
1465 |
}
|
1466 |
},
|
1467 |
"TrainerControl": {
|
|
|
1470 |
"should_evaluate": false,
|
1471 |
"should_log": false,
|
1472 |
"should_save": true,
|
1473 |
+
"should_training_stop": true
|
1474 |
},
|
1475 |
"attributes": {}
|
1476 |
}
|
1477 |
},
|
1478 |
+
"total_flos": 32158462771200.0,
|
1479 |
"train_batch_size": 8,
|
1480 |
"trial_name": null,
|
1481 |
"trial_params": null
|