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 671149168
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:d469ea9df03972c343769dd3debf699740933ee629d873829a539d3c644ab553
|
3 |
size 671149168
|
last-checkpoint/optimizer.pt
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 341314196
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:8122a06e66c624c932a42178e5542adf5b26b376eb710b02aa90bcbdec3fb466
|
3 |
size 341314196
|
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 sha256:0cdb60e2481fd56b78600b940df2781df5f95c30c8817875fe3c80f69f6a8ca9
|
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 sha256:5de003083aa1583e9eb79bf332ee530b155db52c504be6ab21416c2870099d65
|
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": 12.684,
|
724 |
"eval_steps_per_second": 3.171,
|
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": 0.36784711480140686,
|
3 |
+
"best_model_checkpoint": "miner_id_24/checkpoint-200",
|
4 |
+
"epoch": 0.1568627450980392,
|
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": 12.684,
|
724 |
"eval_steps_per_second": 3.171,
|
725 |
"step": 100
|
726 |
+
},
|
727 |
+
{
|
728 |
+
"epoch": 0.0792156862745098,
|
729 |
+
"grad_norm": 4.533246040344238,
|
730 |
+
"learning_rate": 6.101006889170879e-05,
|
731 |
+
"loss": 3.0858,
|
732 |
+
"step": 101
|
733 |
+
},
|
734 |
+
{
|
735 |
+
"epoch": 0.08,
|
736 |
+
"grad_norm": 3.2407262325286865,
|
737 |
+
"learning_rate": 6.0820573841332556e-05,
|
738 |
+
"loss": 1.9589,
|
739 |
+
"step": 102
|
740 |
+
},
|
741 |
+
{
|
742 |
+
"epoch": 0.08078431372549019,
|
743 |
+
"grad_norm": 2.6930501461029053,
|
744 |
+
"learning_rate": 6.062940333092207e-05,
|
745 |
+
"loss": 1.269,
|
746 |
+
"step": 103
|
747 |
+
},
|
748 |
+
{
|
749 |
+
"epoch": 0.0815686274509804,
|
750 |
+
"grad_norm": 2.588240146636963,
|
751 |
+
"learning_rate": 6.043656976525781e-05,
|
752 |
+
"loss": 1.2504,
|
753 |
+
"step": 104
|
754 |
+
},
|
755 |
+
{
|
756 |
+
"epoch": 0.08235294117647059,
|
757 |
+
"grad_norm": 3.165269374847412,
|
758 |
+
"learning_rate": 6.0242085657033506e-05,
|
759 |
+
"loss": 1.6896,
|
760 |
+
"step": 105
|
761 |
+
},
|
762 |
+
{
|
763 |
+
"epoch": 0.08313725490196078,
|
764 |
+
"grad_norm": 2.964270830154419,
|
765 |
+
"learning_rate": 6.004596362604423e-05,
|
766 |
+
"loss": 1.2404,
|
767 |
+
"step": 106
|
768 |
+
},
|
769 |
+
{
|
770 |
+
"epoch": 0.08392156862745098,
|
771 |
+
"grad_norm": 3.3895821571350098,
|
772 |
+
"learning_rate": 5.9848216398367535e-05,
|
773 |
+
"loss": 1.2462,
|
774 |
+
"step": 107
|
775 |
+
},
|
776 |
+
{
|
777 |
+
"epoch": 0.08470588235294117,
|
778 |
+
"grad_norm": 3.5753889083862305,
|
779 |
+
"learning_rate": 5.964885680553763e-05,
|
780 |
+
"loss": 1.4962,
|
781 |
+
"step": 108
|
782 |
+
},
|
783 |
+
{
|
784 |
+
"epoch": 0.08549019607843138,
|
785 |
+
"grad_norm": 3.2786760330200195,
|
786 |
+
"learning_rate": 5.94478977837128e-05,
|
787 |
+
"loss": 1.5371,
|
788 |
+
"step": 109
|
789 |
+
},
|
790 |
+
{
|
791 |
+
"epoch": 0.08627450980392157,
|
792 |
+
"grad_norm": 3.7601850032806396,
|
793 |
+
"learning_rate": 5.9245352372835974e-05,
|
794 |
+
"loss": 1.1844,
|
795 |
+
"step": 110
|
796 |
+
},
|
797 |
+
{
|
798 |
+
"epoch": 0.08705882352941176,
|
799 |
+
"grad_norm": 3.1973512172698975,
|
800 |
+
"learning_rate": 5.9041233715788624e-05,
|
801 |
+
"loss": 1.1938,
|
802 |
+
"step": 111
|
803 |
+
},
|
804 |
+
{
|
805 |
+
"epoch": 0.08784313725490196,
|
806 |
+
"grad_norm": 3.321074962615967,
|
807 |
+
"learning_rate": 5.88355550575379e-05,
|
808 |
+
"loss": 1.3623,
|
809 |
+
"step": 112
|
810 |
+
},
|
811 |
+
{
|
812 |
+
"epoch": 0.08862745098039215,
|
813 |
+
"grad_norm": 3.9187910556793213,
|
814 |
+
"learning_rate": 5.862832974427716e-05,
|
815 |
+
"loss": 1.2505,
|
816 |
+
"step": 113
|
817 |
+
},
|
818 |
+
{
|
819 |
+
"epoch": 0.08941176470588236,
|
820 |
+
"grad_norm": 3.5713284015655518,
|
821 |
+
"learning_rate": 5.8419571222560034e-05,
|
822 |
+
"loss": 1.2605,
|
823 |
+
"step": 114
|
824 |
+
},
|
825 |
+
{
|
826 |
+
"epoch": 0.09019607843137255,
|
827 |
+
"grad_norm": 3.8370423316955566,
|
828 |
+
"learning_rate": 5.8209293038427824e-05,
|
829 |
+
"loss": 1.1408,
|
830 |
+
"step": 115
|
831 |
+
},
|
832 |
+
{
|
833 |
+
"epoch": 0.09098039215686274,
|
834 |
+
"grad_norm": 4.057117938995361,
|
835 |
+
"learning_rate": 5.7997508836530576e-05,
|
836 |
+
"loss": 1.4296,
|
837 |
+
"step": 116
|
838 |
+
},
|
839 |
+
{
|
840 |
+
"epoch": 0.09176470588235294,
|
841 |
+
"grad_norm": 5.31479024887085,
|
842 |
+
"learning_rate": 5.778423235924163e-05,
|
843 |
+
"loss": 1.3345,
|
844 |
+
"step": 117
|
845 |
+
},
|
846 |
+
{
|
847 |
+
"epoch": 0.09254901960784313,
|
848 |
+
"grad_norm": 4.572357177734375,
|
849 |
+
"learning_rate": 5.7569477445765964e-05,
|
850 |
+
"loss": 1.1513,
|
851 |
+
"step": 118
|
852 |
+
},
|
853 |
+
{
|
854 |
+
"epoch": 0.09333333333333334,
|
855 |
+
"grad_norm": 5.681028366088867,
|
856 |
+
"learning_rate": 5.735325803124217e-05,
|
857 |
+
"loss": 1.4338,
|
858 |
+
"step": 119
|
859 |
+
},
|
860 |
+
{
|
861 |
+
"epoch": 0.09411764705882353,
|
862 |
+
"grad_norm": 4.767769813537598,
|
863 |
+
"learning_rate": 5.71355881458382e-05,
|
864 |
+
"loss": 1.3424,
|
865 |
+
"step": 120
|
866 |
+
},
|
867 |
+
{
|
868 |
+
"epoch": 0.09490196078431372,
|
869 |
+
"grad_norm": 6.746960639953613,
|
870 |
+
"learning_rate": 5.6916481913840975e-05,
|
871 |
+
"loss": 1.3405,
|
872 |
+
"step": 121
|
873 |
+
},
|
874 |
+
{
|
875 |
+
"epoch": 0.09568627450980392,
|
876 |
+
"grad_norm": 4.564548492431641,
|
877 |
+
"learning_rate": 5.669595355273991e-05,
|
878 |
+
"loss": 1.2797,
|
879 |
+
"step": 122
|
880 |
+
},
|
881 |
+
{
|
882 |
+
"epoch": 0.09647058823529411,
|
883 |
+
"grad_norm": 4.359226703643799,
|
884 |
+
"learning_rate": 5.64740173723043e-05,
|
885 |
+
"loss": 1.006,
|
886 |
+
"step": 123
|
887 |
+
},
|
888 |
+
{
|
889 |
+
"epoch": 0.09725490196078432,
|
890 |
+
"grad_norm": 5.169023036956787,
|
891 |
+
"learning_rate": 5.625068777365489e-05,
|
892 |
+
"loss": 1.1459,
|
893 |
+
"step": 124
|
894 |
+
},
|
895 |
+
{
|
896 |
+
"epoch": 0.09803921568627451,
|
897 |
+
"grad_norm": 5.223190784454346,
|
898 |
+
"learning_rate": 5.6025979248329255e-05,
|
899 |
+
"loss": 1.2089,
|
900 |
+
"step": 125
|
901 |
+
},
|
902 |
+
{
|
903 |
+
"epoch": 0.0988235294117647,
|
904 |
+
"grad_norm": 6.859619617462158,
|
905 |
+
"learning_rate": 5.579990637734161e-05,
|
906 |
+
"loss": 0.9527,
|
907 |
+
"step": 126
|
908 |
+
},
|
909 |
+
{
|
910 |
+
"epoch": 0.0996078431372549,
|
911 |
+
"grad_norm": 4.413240432739258,
|
912 |
+
"learning_rate": 5.557248383023655e-05,
|
913 |
+
"loss": 0.9772,
|
914 |
+
"step": 127
|
915 |
+
},
|
916 |
+
{
|
917 |
+
"epoch": 0.1003921568627451,
|
918 |
+
"grad_norm": 6.224684238433838,
|
919 |
+
"learning_rate": 5.5343726364137276e-05,
|
920 |
+
"loss": 1.0011,
|
921 |
+
"step": 128
|
922 |
+
},
|
923 |
+
{
|
924 |
+
"epoch": 0.1011764705882353,
|
925 |
+
"grad_norm": 5.601330280303955,
|
926 |
+
"learning_rate": 5.511364882278793e-05,
|
927 |
+
"loss": 1.249,
|
928 |
+
"step": 129
|
929 |
+
},
|
930 |
+
{
|
931 |
+
"epoch": 0.10196078431372549,
|
932 |
+
"grad_norm": 5.103614807128906,
|
933 |
+
"learning_rate": 5.488226613559045e-05,
|
934 |
+
"loss": 1.1634,
|
935 |
+
"step": 130
|
936 |
+
},
|
937 |
+
{
|
938 |
+
"epoch": 0.1027450980392157,
|
939 |
+
"grad_norm": 7.077230453491211,
|
940 |
+
"learning_rate": 5.464959331663583e-05,
|
941 |
+
"loss": 1.5048,
|
942 |
+
"step": 131
|
943 |
+
},
|
944 |
+
{
|
945 |
+
"epoch": 0.10352941176470588,
|
946 |
+
"grad_norm": 5.399044990539551,
|
947 |
+
"learning_rate": 5.4415645463729855e-05,
|
948 |
+
"loss": 1.0851,
|
949 |
+
"step": 132
|
950 |
+
},
|
951 |
+
{
|
952 |
+
"epoch": 0.10431372549019607,
|
953 |
+
"grad_norm": 7.850146770477295,
|
954 |
+
"learning_rate": 5.4180437757413455e-05,
|
955 |
+
"loss": 1.1617,
|
956 |
+
"step": 133
|
957 |
+
},
|
958 |
+
{
|
959 |
+
"epoch": 0.10509803921568628,
|
960 |
+
"grad_norm": 5.438139915466309,
|
961 |
+
"learning_rate": 5.394398545997759e-05,
|
962 |
+
"loss": 0.9825,
|
963 |
+
"step": 134
|
964 |
+
},
|
965 |
+
{
|
966 |
+
"epoch": 0.10588235294117647,
|
967 |
+
"grad_norm": 5.945227146148682,
|
968 |
+
"learning_rate": 5.370630391447303e-05,
|
969 |
+
"loss": 1.165,
|
970 |
+
"step": 135
|
971 |
+
},
|
972 |
+
{
|
973 |
+
"epoch": 0.10666666666666667,
|
974 |
+
"grad_norm": 7.722451686859131,
|
975 |
+
"learning_rate": 5.346740854371464e-05,
|
976 |
+
"loss": 1.9443,
|
977 |
+
"step": 136
|
978 |
+
},
|
979 |
+
{
|
980 |
+
"epoch": 0.10745098039215686,
|
981 |
+
"grad_norm": 7.847848892211914,
|
982 |
+
"learning_rate": 5.322731484928068e-05,
|
983 |
+
"loss": 1.4477,
|
984 |
+
"step": 137
|
985 |
+
},
|
986 |
+
{
|
987 |
+
"epoch": 0.10823529411764705,
|
988 |
+
"grad_norm": 8.356393814086914,
|
989 |
+
"learning_rate": 5.29860384105069e-05,
|
990 |
+
"loss": 1.6004,
|
991 |
+
"step": 138
|
992 |
+
},
|
993 |
+
{
|
994 |
+
"epoch": 0.10901960784313726,
|
995 |
+
"grad_norm": 8.85940170288086,
|
996 |
+
"learning_rate": 5.2743594883475664e-05,
|
997 |
+
"loss": 1.4381,
|
998 |
+
"step": 139
|
999 |
+
},
|
1000 |
+
{
|
1001 |
+
"epoch": 0.10980392156862745,
|
1002 |
+
"grad_norm": 8.563812255859375,
|
1003 |
+
"learning_rate": 5.2499999999999995e-05,
|
1004 |
+
"loss": 1.5443,
|
1005 |
+
"step": 140
|
1006 |
+
},
|
1007 |
+
{
|
1008 |
+
"epoch": 0.11058823529411765,
|
1009 |
+
"grad_norm": 8.796785354614258,
|
1010 |
+
"learning_rate": 5.225526956660281e-05,
|
1011 |
+
"loss": 1.5716,
|
1012 |
+
"step": 141
|
1013 |
+
},
|
1014 |
+
{
|
1015 |
+
"epoch": 0.11137254901960784,
|
1016 |
+
"grad_norm": 11.239343643188477,
|
1017 |
+
"learning_rate": 5.200941946349121e-05,
|
1018 |
+
"loss": 2.2336,
|
1019 |
+
"step": 142
|
1020 |
+
},
|
1021 |
+
{
|
1022 |
+
"epoch": 0.11215686274509803,
|
1023 |
+
"grad_norm": 10.173134803771973,
|
1024 |
+
"learning_rate": 5.176246564352607e-05,
|
1025 |
+
"loss": 1.8621,
|
1026 |
+
"step": 143
|
1027 |
+
},
|
1028 |
+
{
|
1029 |
+
"epoch": 0.11294117647058824,
|
1030 |
+
"grad_norm": 7.808456897735596,
|
1031 |
+
"learning_rate": 5.1514424131186865e-05,
|
1032 |
+
"loss": 1.3717,
|
1033 |
+
"step": 144
|
1034 |
+
},
|
1035 |
+
{
|
1036 |
+
"epoch": 0.11372549019607843,
|
1037 |
+
"grad_norm": 11.386785507202148,
|
1038 |
+
"learning_rate": 5.12653110215319e-05,
|
1039 |
+
"loss": 1.9377,
|
1040 |
+
"step": 145
|
1041 |
+
},
|
1042 |
+
{
|
1043 |
+
"epoch": 0.11450980392156863,
|
1044 |
+
"grad_norm": 11.134692192077637,
|
1045 |
+
"learning_rate": 5.101514247915384e-05,
|
1046 |
+
"loss": 2.1051,
|
1047 |
+
"step": 146
|
1048 |
+
},
|
1049 |
+
{
|
1050 |
+
"epoch": 0.11529411764705882,
|
1051 |
+
"grad_norm": 11.531853675842285,
|
1052 |
+
"learning_rate": 5.0763934737130886e-05,
|
1053 |
+
"loss": 2.1194,
|
1054 |
+
"step": 147
|
1055 |
+
},
|
1056 |
+
{
|
1057 |
+
"epoch": 0.11607843137254902,
|
1058 |
+
"grad_norm": 12.823151588439941,
|
1059 |
+
"learning_rate": 5.0511704095973445e-05,
|
1060 |
+
"loss": 2.008,
|
1061 |
+
"step": 148
|
1062 |
+
},
|
1063 |
+
{
|
1064 |
+
"epoch": 0.11686274509803922,
|
1065 |
+
"grad_norm": 14.165589332580566,
|
1066 |
+
"learning_rate": 5.0258466922566345e-05,
|
1067 |
+
"loss": 2.579,
|
1068 |
+
"step": 149
|
1069 |
+
},
|
1070 |
+
{
|
1071 |
+
"epoch": 0.11764705882352941,
|
1072 |
+
"grad_norm": 22.23053741455078,
|
1073 |
+
"learning_rate": 5.000423964910689e-05,
|
1074 |
+
"loss": 2.1285,
|
1075 |
+
"step": 150
|
1076 |
+
},
|
1077 |
+
{
|
1078 |
+
"epoch": 0.11843137254901961,
|
1079 |
+
"grad_norm": 3.6735522747039795,
|
1080 |
+
"learning_rate": 4.974903877203852e-05,
|
1081 |
+
"loss": 3.2373,
|
1082 |
+
"step": 151
|
1083 |
+
},
|
1084 |
+
{
|
1085 |
+
"epoch": 0.1192156862745098,
|
1086 |
+
"grad_norm": 2.631013870239258,
|
1087 |
+
"learning_rate": 4.949288085098041e-05,
|
1088 |
+
"loss": 1.7498,
|
1089 |
+
"step": 152
|
1090 |
+
},
|
1091 |
+
{
|
1092 |
+
"epoch": 0.12,
|
1093 |
+
"grad_norm": 2.442901372909546,
|
1094 |
+
"learning_rate": 4.9235782507653006e-05,
|
1095 |
+
"loss": 1.7117,
|
1096 |
+
"step": 153
|
1097 |
+
},
|
1098 |
+
{
|
1099 |
+
"epoch": 0.1207843137254902,
|
1100 |
+
"grad_norm": 2.393778085708618,
|
1101 |
+
"learning_rate": 4.897776042479934e-05,
|
1102 |
+
"loss": 1.3329,
|
1103 |
+
"step": 154
|
1104 |
+
},
|
1105 |
+
{
|
1106 |
+
"epoch": 0.12156862745098039,
|
1107 |
+
"grad_norm": 2.4069385528564453,
|
1108 |
+
"learning_rate": 4.8718831345102625e-05,
|
1109 |
+
"loss": 1.3254,
|
1110 |
+
"step": 155
|
1111 |
+
},
|
1112 |
+
{
|
1113 |
+
"epoch": 0.1223529411764706,
|
1114 |
+
"grad_norm": 2.7312748432159424,
|
1115 |
+
"learning_rate": 4.845901207009978e-05,
|
1116 |
+
"loss": 1.1983,
|
1117 |
+
"step": 156
|
1118 |
+
},
|
1119 |
+
{
|
1120 |
+
"epoch": 0.12313725490196079,
|
1121 |
+
"grad_norm": 2.8144845962524414,
|
1122 |
+
"learning_rate": 4.819831945909121e-05,
|
1123 |
+
"loss": 1.1755,
|
1124 |
+
"step": 157
|
1125 |
+
},
|
1126 |
+
{
|
1127 |
+
"epoch": 0.12392156862745098,
|
1128 |
+
"grad_norm": 3.241198778152466,
|
1129 |
+
"learning_rate": 4.793677042804688e-05,
|
1130 |
+
"loss": 1.1232,
|
1131 |
+
"step": 158
|
1132 |
+
},
|
1133 |
+
{
|
1134 |
+
"epoch": 0.12470588235294118,
|
1135 |
+
"grad_norm": 3.7568423748016357,
|
1136 |
+
"learning_rate": 4.767438194850858e-05,
|
1137 |
+
"loss": 1.4093,
|
1138 |
+
"step": 159
|
1139 |
+
},
|
1140 |
+
{
|
1141 |
+
"epoch": 0.12549019607843137,
|
1142 |
+
"grad_norm": 3.8611888885498047,
|
1143 |
+
"learning_rate": 4.741117104648874e-05,
|
1144 |
+
"loss": 1.3189,
|
1145 |
+
"step": 160
|
1146 |
+
},
|
1147 |
+
{
|
1148 |
+
"epoch": 0.12627450980392158,
|
1149 |
+
"grad_norm": 3.7197554111480713,
|
1150 |
+
"learning_rate": 4.71471548013656e-05,
|
1151 |
+
"loss": 1.4241,
|
1152 |
+
"step": 161
|
1153 |
+
},
|
1154 |
+
{
|
1155 |
+
"epoch": 0.12705882352941175,
|
1156 |
+
"grad_norm": 3.673598527908325,
|
1157 |
+
"learning_rate": 4.688235034477495e-05,
|
1158 |
+
"loss": 1.2979,
|
1159 |
+
"step": 162
|
1160 |
+
},
|
1161 |
+
{
|
1162 |
+
"epoch": 0.12784313725490196,
|
1163 |
+
"grad_norm": 3.9346017837524414,
|
1164 |
+
"learning_rate": 4.6616774859498493e-05,
|
1165 |
+
"loss": 1.4983,
|
1166 |
+
"step": 163
|
1167 |
+
},
|
1168 |
+
{
|
1169 |
+
"epoch": 0.12862745098039216,
|
1170 |
+
"grad_norm": 3.7387478351593018,
|
1171 |
+
"learning_rate": 4.635044557834889e-05,
|
1172 |
+
"loss": 1.5969,
|
1173 |
+
"step": 164
|
1174 |
+
},
|
1175 |
+
{
|
1176 |
+
"epoch": 0.12941176470588237,
|
1177 |
+
"grad_norm": 4.286586284637451,
|
1178 |
+
"learning_rate": 4.6083379783051544e-05,
|
1179 |
+
"loss": 1.1618,
|
1180 |
+
"step": 165
|
1181 |
+
},
|
1182 |
+
{
|
1183 |
+
"epoch": 0.13019607843137254,
|
1184 |
+
"grad_norm": 3.6195132732391357,
|
1185 |
+
"learning_rate": 4.581559480312316e-05,
|
1186 |
+
"loss": 1.3013,
|
1187 |
+
"step": 166
|
1188 |
+
},
|
1189 |
+
{
|
1190 |
+
"epoch": 0.13098039215686275,
|
1191 |
+
"grad_norm": 3.9469423294067383,
|
1192 |
+
"learning_rate": 4.554710801474736e-05,
|
1193 |
+
"loss": 1.2563,
|
1194 |
+
"step": 167
|
1195 |
+
},
|
1196 |
+
{
|
1197 |
+
"epoch": 0.13176470588235295,
|
1198 |
+
"grad_norm": 3.9838595390319824,
|
1199 |
+
"learning_rate": 4.527793683964708e-05,
|
1200 |
+
"loss": 1.084,
|
1201 |
+
"step": 168
|
1202 |
+
},
|
1203 |
+
{
|
1204 |
+
"epoch": 0.13254901960784313,
|
1205 |
+
"grad_norm": 4.2628021240234375,
|
1206 |
+
"learning_rate": 4.5008098743954114e-05,
|
1207 |
+
"loss": 1.1541,
|
1208 |
+
"step": 169
|
1209 |
+
},
|
1210 |
+
{
|
1211 |
+
"epoch": 0.13333333333333333,
|
1212 |
+
"grad_norm": 4.49942684173584,
|
1213 |
+
"learning_rate": 4.473761123707584e-05,
|
1214 |
+
"loss": 1.4258,
|
1215 |
+
"step": 170
|
1216 |
+
},
|
1217 |
+
{
|
1218 |
+
"epoch": 0.13411764705882354,
|
1219 |
+
"grad_norm": 5.674098968505859,
|
1220 |
+
"learning_rate": 4.4466491870558936e-05,
|
1221 |
+
"loss": 1.8754,
|
1222 |
+
"step": 171
|
1223 |
+
},
|
1224 |
+
{
|
1225 |
+
"epoch": 0.1349019607843137,
|
1226 |
+
"grad_norm": 5.979763984680176,
|
1227 |
+
"learning_rate": 4.419475823695054e-05,
|
1228 |
+
"loss": 1.3147,
|
1229 |
+
"step": 172
|
1230 |
+
},
|
1231 |
+
{
|
1232 |
+
"epoch": 0.13568627450980392,
|
1233 |
+
"grad_norm": 4.832103252410889,
|
1234 |
+
"learning_rate": 4.392242796865676e-05,
|
1235 |
+
"loss": 1.2014,
|
1236 |
+
"step": 173
|
1237 |
+
},
|
1238 |
+
{
|
1239 |
+
"epoch": 0.13647058823529412,
|
1240 |
+
"grad_norm": 4.375277996063232,
|
1241 |
+
"learning_rate": 4.364951873679843e-05,
|
1242 |
+
"loss": 1.1941,
|
1243 |
+
"step": 174
|
1244 |
+
},
|
1245 |
+
{
|
1246 |
+
"epoch": 0.13725490196078433,
|
1247 |
+
"grad_norm": 4.7891526222229,
|
1248 |
+
"learning_rate": 4.337604825006452e-05,
|
1249 |
+
"loss": 1.1638,
|
1250 |
+
"step": 175
|
1251 |
+
},
|
1252 |
+
{
|
1253 |
+
"epoch": 0.1380392156862745,
|
1254 |
+
"grad_norm": 5.445965766906738,
|
1255 |
+
"learning_rate": 4.310203425356301e-05,
|
1256 |
+
"loss": 1.054,
|
1257 |
+
"step": 176
|
1258 |
+
},
|
1259 |
+
{
|
1260 |
+
"epoch": 0.1388235294117647,
|
1261 |
+
"grad_norm": 4.09723424911499,
|
1262 |
+
"learning_rate": 4.282749452766945e-05,
|
1263 |
+
"loss": 0.6077,
|
1264 |
+
"step": 177
|
1265 |
+
},
|
1266 |
+
{
|
1267 |
+
"epoch": 0.1396078431372549,
|
1268 |
+
"grad_norm": 5.525692939758301,
|
1269 |
+
"learning_rate": 4.255244688687321e-05,
|
1270 |
+
"loss": 1.0198,
|
1271 |
+
"step": 178
|
1272 |
+
},
|
1273 |
+
{
|
1274 |
+
"epoch": 0.1403921568627451,
|
1275 |
+
"grad_norm": 5.302620887756348,
|
1276 |
+
"learning_rate": 4.227690917862157e-05,
|
1277 |
+
"loss": 0.9727,
|
1278 |
+
"step": 179
|
1279 |
+
},
|
1280 |
+
{
|
1281 |
+
"epoch": 0.1411764705882353,
|
1282 |
+
"grad_norm": 4.293851375579834,
|
1283 |
+
"learning_rate": 4.200089928216155e-05,
|
1284 |
+
"loss": 0.7823,
|
1285 |
+
"step": 180
|
1286 |
+
},
|
1287 |
+
{
|
1288 |
+
"epoch": 0.1419607843137255,
|
1289 |
+
"grad_norm": 6.371180057525635,
|
1290 |
+
"learning_rate": 4.1724435107379795e-05,
|
1291 |
+
"loss": 1.2332,
|
1292 |
+
"step": 181
|
1293 |
+
},
|
1294 |
+
{
|
1295 |
+
"epoch": 0.14274509803921567,
|
1296 |
+
"grad_norm": 7.856477737426758,
|
1297 |
+
"learning_rate": 4.14475345936404e-05,
|
1298 |
+
"loss": 1.3729,
|
1299 |
+
"step": 182
|
1300 |
+
},
|
1301 |
+
{
|
1302 |
+
"epoch": 0.14352941176470588,
|
1303 |
+
"grad_norm": 4.569979190826416,
|
1304 |
+
"learning_rate": 4.1170215708620915e-05,
|
1305 |
+
"loss": 0.8416,
|
1306 |
+
"step": 183
|
1307 |
+
},
|
1308 |
+
{
|
1309 |
+
"epoch": 0.14431372549019608,
|
1310 |
+
"grad_norm": 6.07737398147583,
|
1311 |
+
"learning_rate": 4.089249644714634e-05,
|
1312 |
+
"loss": 1.4574,
|
1313 |
+
"step": 184
|
1314 |
+
},
|
1315 |
+
{
|
1316 |
+
"epoch": 0.1450980392156863,
|
1317 |
+
"grad_norm": 6.413763523101807,
|
1318 |
+
"learning_rate": 4.0614394830021604e-05,
|
1319 |
+
"loss": 1.5253,
|
1320 |
+
"step": 185
|
1321 |
+
},
|
1322 |
+
{
|
1323 |
+
"epoch": 0.14588235294117646,
|
1324 |
+
"grad_norm": 5.022174835205078,
|
1325 |
+
"learning_rate": 4.0335928902862086e-05,
|
1326 |
+
"loss": 0.9277,
|
1327 |
+
"step": 186
|
1328 |
+
},
|
1329 |
+
{
|
1330 |
+
"epoch": 0.14666666666666667,
|
1331 |
+
"grad_norm": 7.028177261352539,
|
1332 |
+
"learning_rate": 4.005711673492274e-05,
|
1333 |
+
"loss": 1.6335,
|
1334 |
+
"step": 187
|
1335 |
+
},
|
1336 |
+
{
|
1337 |
+
"epoch": 0.14745098039215687,
|
1338 |
+
"grad_norm": 7.276490211486816,
|
1339 |
+
"learning_rate": 3.977797641792561e-05,
|
1340 |
+
"loss": 1.3358,
|
1341 |
+
"step": 188
|
1342 |
+
},
|
1343 |
+
{
|
1344 |
+
"epoch": 0.14823529411764705,
|
1345 |
+
"grad_norm": 8.936336517333984,
|
1346 |
+
"learning_rate": 3.9498526064885806e-05,
|
1347 |
+
"loss": 1.8587,
|
1348 |
+
"step": 189
|
1349 |
+
},
|
1350 |
+
{
|
1351 |
+
"epoch": 0.14901960784313725,
|
1352 |
+
"grad_norm": 9.971092224121094,
|
1353 |
+
"learning_rate": 3.92187838089363e-05,
|
1354 |
+
"loss": 1.9026,
|
1355 |
+
"step": 190
|
1356 |
+
},
|
1357 |
+
{
|
1358 |
+
"epoch": 0.14980392156862746,
|
1359 |
+
"grad_norm": 8.014878273010254,
|
1360 |
+
"learning_rate": 3.893876780215119e-05,
|
1361 |
+
"loss": 1.6085,
|
1362 |
+
"step": 191
|
1363 |
+
},
|
1364 |
+
{
|
1365 |
+
"epoch": 0.15058823529411763,
|
1366 |
+
"grad_norm": 9.771968841552734,
|
1367 |
+
"learning_rate": 3.865849621436787e-05,
|
1368 |
+
"loss": 1.8793,
|
1369 |
+
"step": 192
|
1370 |
+
},
|
1371 |
+
{
|
1372 |
+
"epoch": 0.15137254901960784,
|
1373 |
+
"grad_norm": 11.60059928894043,
|
1374 |
+
"learning_rate": 3.8377987232008025e-05,
|
1375 |
+
"loss": 1.8012,
|
1376 |
+
"step": 193
|
1377 |
+
},
|
1378 |
+
{
|
1379 |
+
"epoch": 0.15215686274509804,
|
1380 |
+
"grad_norm": 9.739047050476074,
|
1381 |
+
"learning_rate": 3.809725905689754e-05,
|
1382 |
+
"loss": 1.9012,
|
1383 |
+
"step": 194
|
1384 |
+
},
|
1385 |
+
{
|
1386 |
+
"epoch": 0.15294117647058825,
|
1387 |
+
"grad_norm": 12.654723167419434,
|
1388 |
+
"learning_rate": 3.781632990508541e-05,
|
1389 |
+
"loss": 2.3507,
|
1390 |
+
"step": 195
|
1391 |
+
},
|
1392 |
+
{
|
1393 |
+
"epoch": 0.15372549019607842,
|
1394 |
+
"grad_norm": 11.12818717956543,
|
1395 |
+
"learning_rate": 3.7535218005661685e-05,
|
1396 |
+
"loss": 1.877,
|
1397 |
+
"step": 196
|
1398 |
+
},
|
1399 |
+
{
|
1400 |
+
"epoch": 0.15450980392156863,
|
1401 |
+
"grad_norm": 13.460821151733398,
|
1402 |
+
"learning_rate": 3.7253941599574675e-05,
|
1403 |
+
"loss": 2.0457,
|
1404 |
+
"step": 197
|
1405 |
+
},
|
1406 |
+
{
|
1407 |
+
"epoch": 0.15529411764705883,
|
1408 |
+
"grad_norm": 16.934213638305664,
|
1409 |
+
"learning_rate": 3.6972518938447306e-05,
|
1410 |
+
"loss": 2.5687,
|
1411 |
+
"step": 198
|
1412 |
+
},
|
1413 |
+
{
|
1414 |
+
"epoch": 0.156078431372549,
|
1415 |
+
"grad_norm": 17.341718673706055,
|
1416 |
+
"learning_rate": 3.6690968283392746e-05,
|
1417 |
+
"loss": 2.6656,
|
1418 |
+
"step": 199
|
1419 |
+
},
|
1420 |
+
{
|
1421 |
+
"epoch": 0.1568627450980392,
|
1422 |
+
"grad_norm": 13.909801483154297,
|
1423 |
+
"learning_rate": 3.640930790382953e-05,
|
1424 |
+
"loss": 2.0502,
|
1425 |
+
"step": 200
|
1426 |
+
},
|
1427 |
+
{
|
1428 |
+
"epoch": 0.1568627450980392,
|
1429 |
+
"eval_loss": 0.36784711480140686,
|
1430 |
+
"eval_runtime": 169.7291,
|
1431 |
+
"eval_samples_per_second": 12.655,
|
1432 |
+
"eval_steps_per_second": 3.164,
|
1433 |
+
"step": 200
|
1434 |
}
|
1435 |
],
|
1436 |
"logging_steps": 1,
|
|
|
1459 |
"attributes": {}
|
1460 |
}
|
1461 |
},
|
1462 |
+
"total_flos": 3.030141694503813e+17,
|
1463 |
"train_batch_size": 8,
|
1464 |
"trial_name": null,
|
1465 |
"trial_params": null
|