Training in progress, step 138, 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 2436951232
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:f719a64470755771748becc2d79750c1dfc3c56ede76336615f2287e3183de55
|
| 3 |
size 2436951232
|
last-checkpoint/optimizer.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 2274077596
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:4b98a96d8f832a2978139f64d417323559b2351dd714affdd97f8810a53a1c30
|
| 3 |
size 2274077596
|
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 sha256:734e89105df261b24dbde0e06f6215c07754ac7f546e434ffc90b4c4aeac7c05
|
| 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 sha256:21d7f598026e67cff206e8b45c13b9fe02c682beb205dd1fd163b157e36c8d6c
|
| 3 |
size 1064
|
last-checkpoint/trainer_state.json
CHANGED
|
@@ -1,9 +1,9 @@
|
|
| 1 |
{
|
| 2 |
"best_metric": null,
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
-
"epoch":
|
| 5 |
"eval_steps": 12,
|
| 6 |
-
"global_step":
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
@@ -715,6 +715,360 @@
|
|
| 715 |
"learning_rate": 0.00010901530811120655,
|
| 716 |
"loss": 0.283,
|
| 717 |
"step": 92
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 718 |
}
|
| 719 |
],
|
| 720 |
"logging_steps": 1,
|
|
@@ -734,7 +1088,7 @@
|
|
| 734 |
"attributes": {}
|
| 735 |
}
|
| 736 |
},
|
| 737 |
-
"total_flos":
|
| 738 |
"train_batch_size": 2,
|
| 739 |
"trial_name": null,
|
| 740 |
"trial_params": null
|
|
|
|
| 1 |
{
|
| 2 |
"best_metric": null,
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 2.732620320855615,
|
| 5 |
"eval_steps": 12,
|
| 6 |
+
"global_step": 138,
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
|
|
| 715 |
"learning_rate": 0.00010901530811120655,
|
| 716 |
"loss": 0.283,
|
| 717 |
"step": 92
|
| 718 |
+
},
|
| 719 |
+
{
|
| 720 |
+
"epoch": 1.8770053475935828,
|
| 721 |
+
"grad_norm": 4.964606285095215,
|
| 722 |
+
"learning_rate": 0.00010721577558006164,
|
| 723 |
+
"loss": 0.2673,
|
| 724 |
+
"step": 93
|
| 725 |
+
},
|
| 726 |
+
{
|
| 727 |
+
"epoch": 1.8983957219251337,
|
| 728 |
+
"grad_norm": 8.798148155212402,
|
| 729 |
+
"learning_rate": 0.00010541389085854176,
|
| 730 |
+
"loss": 0.2055,
|
| 731 |
+
"step": 94
|
| 732 |
+
},
|
| 733 |
+
{
|
| 734 |
+
"epoch": 1.9197860962566846,
|
| 735 |
+
"grad_norm": 5.59722375869751,
|
| 736 |
+
"learning_rate": 0.00010361024132299364,
|
| 737 |
+
"loss": 0.209,
|
| 738 |
+
"step": 95
|
| 739 |
+
},
|
| 740 |
+
{
|
| 741 |
+
"epoch": 1.9411764705882353,
|
| 742 |
+
"grad_norm": 8.256745338439941,
|
| 743 |
+
"learning_rate": 0.00010180541492505604,
|
| 744 |
+
"loss": 0.1079,
|
| 745 |
+
"step": 96
|
| 746 |
+
},
|
| 747 |
+
{
|
| 748 |
+
"epoch": 1.9411764705882353,
|
| 749 |
+
"eval_loss": 0.7684900760650635,
|
| 750 |
+
"eval_runtime": 17.746,
|
| 751 |
+
"eval_samples_per_second": 16.905,
|
| 752 |
+
"eval_steps_per_second": 8.453,
|
| 753 |
+
"step": 96
|
| 754 |
+
},
|
| 755 |
+
{
|
| 756 |
+
"epoch": 1.962566844919786,
|
| 757 |
+
"grad_norm": 4.481995582580566,
|
| 758 |
+
"learning_rate": 0.0001,
|
| 759 |
+
"loss": 0.1824,
|
| 760 |
+
"step": 97
|
| 761 |
+
},
|
| 762 |
+
{
|
| 763 |
+
"epoch": 1.9839572192513368,
|
| 764 |
+
"grad_norm": 3.852792263031006,
|
| 765 |
+
"learning_rate": 9.819458507494394e-05,
|
| 766 |
+
"loss": 0.124,
|
| 767 |
+
"step": 98
|
| 768 |
+
},
|
| 769 |
+
{
|
| 770 |
+
"epoch": 2.0053475935828877,
|
| 771 |
+
"grad_norm": 29.360998153686523,
|
| 772 |
+
"learning_rate": 9.638975867700638e-05,
|
| 773 |
+
"loss": 0.1844,
|
| 774 |
+
"step": 99
|
| 775 |
+
},
|
| 776 |
+
{
|
| 777 |
+
"epoch": 2.0267379679144386,
|
| 778 |
+
"grad_norm": 4.80811071395874,
|
| 779 |
+
"learning_rate": 9.458610914145826e-05,
|
| 780 |
+
"loss": 0.1347,
|
| 781 |
+
"step": 100
|
| 782 |
+
},
|
| 783 |
+
{
|
| 784 |
+
"epoch": 2.048128342245989,
|
| 785 |
+
"grad_norm": 17.53403091430664,
|
| 786 |
+
"learning_rate": 9.27842244199384e-05,
|
| 787 |
+
"loss": 0.3378,
|
| 788 |
+
"step": 101
|
| 789 |
+
},
|
| 790 |
+
{
|
| 791 |
+
"epoch": 2.06951871657754,
|
| 792 |
+
"grad_norm": 7.34214973449707,
|
| 793 |
+
"learning_rate": 9.098469188879349e-05,
|
| 794 |
+
"loss": 0.2045,
|
| 795 |
+
"step": 102
|
| 796 |
+
},
|
| 797 |
+
{
|
| 798 |
+
"epoch": 2.090909090909091,
|
| 799 |
+
"grad_norm": 4.968944072723389,
|
| 800 |
+
"learning_rate": 8.918809815760585e-05,
|
| 801 |
+
"loss": 0.2284,
|
| 802 |
+
"step": 103
|
| 803 |
+
},
|
| 804 |
+
{
|
| 805 |
+
"epoch": 2.0053475935828877,
|
| 806 |
+
"grad_norm": 4.202042579650879,
|
| 807 |
+
"learning_rate": 8.739502887797107e-05,
|
| 808 |
+
"loss": 0.1655,
|
| 809 |
+
"step": 104
|
| 810 |
+
},
|
| 811 |
+
{
|
| 812 |
+
"epoch": 2.0267379679144386,
|
| 813 |
+
"grad_norm": 3.561790704727173,
|
| 814 |
+
"learning_rate": 8.560606855258808e-05,
|
| 815 |
+
"loss": 0.0442,
|
| 816 |
+
"step": 105
|
| 817 |
+
},
|
| 818 |
+
{
|
| 819 |
+
"epoch": 2.0481283422459895,
|
| 820 |
+
"grad_norm": 3.8292624950408936,
|
| 821 |
+
"learning_rate": 8.382180034472353e-05,
|
| 822 |
+
"loss": 0.0821,
|
| 823 |
+
"step": 106
|
| 824 |
+
},
|
| 825 |
+
{
|
| 826 |
+
"epoch": 2.06951871657754,
|
| 827 |
+
"grad_norm": 2.1650640964508057,
|
| 828 |
+
"learning_rate": 8.204280588811283e-05,
|
| 829 |
+
"loss": 0.0384,
|
| 830 |
+
"step": 107
|
| 831 |
+
},
|
| 832 |
+
{
|
| 833 |
+
"epoch": 2.090909090909091,
|
| 834 |
+
"grad_norm": 1.6922334432601929,
|
| 835 |
+
"learning_rate": 8.026966509736001e-05,
|
| 836 |
+
"loss": 0.0342,
|
| 837 |
+
"step": 108
|
| 838 |
+
},
|
| 839 |
+
{
|
| 840 |
+
"epoch": 2.090909090909091,
|
| 841 |
+
"eval_loss": 0.7716657519340515,
|
| 842 |
+
"eval_runtime": 17.9156,
|
| 843 |
+
"eval_samples_per_second": 16.745,
|
| 844 |
+
"eval_steps_per_second": 8.373,
|
| 845 |
+
"step": 108
|
| 846 |
+
},
|
| 847 |
+
{
|
| 848 |
+
"epoch": 2.1122994652406417,
|
| 849 |
+
"grad_norm": 1.0654356479644775,
|
| 850 |
+
"learning_rate": 7.85029559788976e-05,
|
| 851 |
+
"loss": 0.0184,
|
| 852 |
+
"step": 109
|
| 853 |
+
},
|
| 854 |
+
{
|
| 855 |
+
"epoch": 2.1336898395721926,
|
| 856 |
+
"grad_norm": 3.1057019233703613,
|
| 857 |
+
"learning_rate": 7.674325444256899e-05,
|
| 858 |
+
"loss": 0.0417,
|
| 859 |
+
"step": 110
|
| 860 |
+
},
|
| 861 |
+
{
|
| 862 |
+
"epoch": 2.1550802139037435,
|
| 863 |
+
"grad_norm": 0.19042205810546875,
|
| 864 |
+
"learning_rate": 7.499113411389371e-05,
|
| 865 |
+
"loss": 0.0026,
|
| 866 |
+
"step": 111
|
| 867 |
+
},
|
| 868 |
+
{
|
| 869 |
+
"epoch": 2.176470588235294,
|
| 870 |
+
"grad_norm": 1.5116851329803467,
|
| 871 |
+
"learning_rate": 7.324716614707793e-05,
|
| 872 |
+
"loss": 0.0089,
|
| 873 |
+
"step": 112
|
| 874 |
+
},
|
| 875 |
+
{
|
| 876 |
+
"epoch": 2.197860962566845,
|
| 877 |
+
"grad_norm": 2.5151679515838623,
|
| 878 |
+
"learning_rate": 7.151191903883001e-05,
|
| 879 |
+
"loss": 0.0357,
|
| 880 |
+
"step": 113
|
| 881 |
+
},
|
| 882 |
+
{
|
| 883 |
+
"epoch": 2.2192513368983957,
|
| 884 |
+
"grad_norm": 2.838503837585449,
|
| 885 |
+
"learning_rate": 6.978595844304271e-05,
|
| 886 |
+
"loss": 0.0366,
|
| 887 |
+
"step": 114
|
| 888 |
+
},
|
| 889 |
+
{
|
| 890 |
+
"epoch": 2.2406417112299466,
|
| 891 |
+
"grad_norm": 3.835000514984131,
|
| 892 |
+
"learning_rate": 6.806984698640202e-05,
|
| 893 |
+
"loss": 0.1412,
|
| 894 |
+
"step": 115
|
| 895 |
+
},
|
| 896 |
+
{
|
| 897 |
+
"epoch": 2.2620320855614975,
|
| 898 |
+
"grad_norm": 3.4443538188934326,
|
| 899 |
+
"learning_rate": 6.636414408498249e-05,
|
| 900 |
+
"loss": 0.0707,
|
| 901 |
+
"step": 116
|
| 902 |
+
},
|
| 903 |
+
{
|
| 904 |
+
"epoch": 2.283422459893048,
|
| 905 |
+
"grad_norm": 2.701524496078491,
|
| 906 |
+
"learning_rate": 6.466940576188977e-05,
|
| 907 |
+
"loss": 0.0497,
|
| 908 |
+
"step": 117
|
| 909 |
+
},
|
| 910 |
+
{
|
| 911 |
+
"epoch": 2.304812834224599,
|
| 912 |
+
"grad_norm": 2.612593412399292,
|
| 913 |
+
"learning_rate": 6.298618446600856e-05,
|
| 914 |
+
"loss": 0.052,
|
| 915 |
+
"step": 118
|
| 916 |
+
},
|
| 917 |
+
{
|
| 918 |
+
"epoch": 2.3262032085561497,
|
| 919 |
+
"grad_norm": 4.986962795257568,
|
| 920 |
+
"learning_rate": 6.13150288919161e-05,
|
| 921 |
+
"loss": 0.1255,
|
| 922 |
+
"step": 119
|
| 923 |
+
},
|
| 924 |
+
{
|
| 925 |
+
"epoch": 2.3475935828877006,
|
| 926 |
+
"grad_norm": 1.8598374128341675,
|
| 927 |
+
"learning_rate": 5.965648380101916e-05,
|
| 928 |
+
"loss": 0.0309,
|
| 929 |
+
"step": 120
|
| 930 |
+
},
|
| 931 |
+
{
|
| 932 |
+
"epoch": 2.3475935828877006,
|
| 933 |
+
"eval_loss": 0.785007119178772,
|
| 934 |
+
"eval_runtime": 17.7923,
|
| 935 |
+
"eval_samples_per_second": 16.861,
|
| 936 |
+
"eval_steps_per_second": 8.431,
|
| 937 |
+
"step": 120
|
| 938 |
+
},
|
| 939 |
+
{
|
| 940 |
+
"epoch": 2.3689839572192515,
|
| 941 |
+
"grad_norm": 1.5813214778900146,
|
| 942 |
+
"learning_rate": 5.801108984397354e-05,
|
| 943 |
+
"loss": 0.0201,
|
| 944 |
+
"step": 121
|
| 945 |
+
},
|
| 946 |
+
{
|
| 947 |
+
"epoch": 2.3903743315508024,
|
| 948 |
+
"grad_norm": 0.13843385875225067,
|
| 949 |
+
"learning_rate": 5.6379383384443255e-05,
|
| 950 |
+
"loss": 0.0018,
|
| 951 |
+
"step": 122
|
| 952 |
+
},
|
| 953 |
+
{
|
| 954 |
+
"epoch": 2.411764705882353,
|
| 955 |
+
"grad_norm": 4.4155707359313965,
|
| 956 |
+
"learning_rate": 5.476189632425732e-05,
|
| 957 |
+
"loss": 0.0326,
|
| 958 |
+
"step": 123
|
| 959 |
+
},
|
| 960 |
+
{
|
| 961 |
+
"epoch": 2.4331550802139037,
|
| 962 |
+
"grad_norm": 3.5101325511932373,
|
| 963 |
+
"learning_rate": 5.3159155930021e-05,
|
| 964 |
+
"loss": 0.0259,
|
| 965 |
+
"step": 124
|
| 966 |
+
},
|
| 967 |
+
{
|
| 968 |
+
"epoch": 2.4545454545454546,
|
| 969 |
+
"grad_norm": 5.201532363891602,
|
| 970 |
+
"learning_rate": 5.1571684661238075e-05,
|
| 971 |
+
"loss": 0.0761,
|
| 972 |
+
"step": 125
|
| 973 |
+
},
|
| 974 |
+
{
|
| 975 |
+
"epoch": 2.4759358288770055,
|
| 976 |
+
"grad_norm": 2.48543119430542,
|
| 977 |
+
"learning_rate": 5.000000000000002e-05,
|
| 978 |
+
"loss": 0.0587,
|
| 979 |
+
"step": 126
|
| 980 |
+
},
|
| 981 |
+
{
|
| 982 |
+
"epoch": 2.497326203208556,
|
| 983 |
+
"grad_norm": 7.39755916595459,
|
| 984 |
+
"learning_rate": 4.844461428229782e-05,
|
| 985 |
+
"loss": 0.0391,
|
| 986 |
+
"step": 127
|
| 987 |
+
},
|
| 988 |
+
{
|
| 989 |
+
"epoch": 2.518716577540107,
|
| 990 |
+
"grad_norm": 4.151485443115234,
|
| 991 |
+
"learning_rate": 4.6906034531011346e-05,
|
| 992 |
+
"loss": 0.0982,
|
| 993 |
+
"step": 128
|
| 994 |
+
},
|
| 995 |
+
{
|
| 996 |
+
"epoch": 2.5401069518716577,
|
| 997 |
+
"grad_norm": 4.144845485687256,
|
| 998 |
+
"learning_rate": 4.53847622906303e-05,
|
| 999 |
+
"loss": 0.0707,
|
| 1000 |
+
"step": 129
|
| 1001 |
+
},
|
| 1002 |
+
{
|
| 1003 |
+
"epoch": 2.5614973262032086,
|
| 1004 |
+
"grad_norm": 7.3682732582092285,
|
| 1005 |
+
"learning_rate": 4.388129346376178e-05,
|
| 1006 |
+
"loss": 0.0455,
|
| 1007 |
+
"step": 130
|
| 1008 |
+
},
|
| 1009 |
+
{
|
| 1010 |
+
"epoch": 2.5828877005347595,
|
| 1011 |
+
"grad_norm": 4.947929382324219,
|
| 1012 |
+
"learning_rate": 4.239611814947605e-05,
|
| 1013 |
+
"loss": 0.033,
|
| 1014 |
+
"step": 131
|
| 1015 |
+
},
|
| 1016 |
+
{
|
| 1017 |
+
"epoch": 2.6042780748663104,
|
| 1018 |
+
"grad_norm": 3.0208606719970703,
|
| 1019 |
+
"learning_rate": 4.092972048354491e-05,
|
| 1020 |
+
"loss": 0.0373,
|
| 1021 |
+
"step": 132
|
| 1022 |
+
},
|
| 1023 |
+
{
|
| 1024 |
+
"epoch": 2.6042780748663104,
|
| 1025 |
+
"eval_loss": 0.776565432548523,
|
| 1026 |
+
"eval_runtime": 17.3019,
|
| 1027 |
+
"eval_samples_per_second": 17.339,
|
| 1028 |
+
"eval_steps_per_second": 8.67,
|
| 1029 |
+
"step": 132
|
| 1030 |
+
},
|
| 1031 |
+
{
|
| 1032 |
+
"epoch": 2.625668449197861,
|
| 1033 |
+
"grad_norm": 7.514610290527344,
|
| 1034 |
+
"learning_rate": 3.948257848062351e-05,
|
| 1035 |
+
"loss": 0.0323,
|
| 1036 |
+
"step": 133
|
| 1037 |
+
},
|
| 1038 |
+
{
|
| 1039 |
+
"epoch": 2.6470588235294117,
|
| 1040 |
+
"grad_norm": 1.8352607488632202,
|
| 1041 |
+
"learning_rate": 3.80551638784277e-05,
|
| 1042 |
+
"loss": 0.043,
|
| 1043 |
+
"step": 134
|
| 1044 |
+
},
|
| 1045 |
+
{
|
| 1046 |
+
"epoch": 2.6684491978609626,
|
| 1047 |
+
"grad_norm": 3.525506019592285,
|
| 1048 |
+
"learning_rate": 3.664794198395764e-05,
|
| 1049 |
+
"loss": 0.0643,
|
| 1050 |
+
"step": 135
|
| 1051 |
+
},
|
| 1052 |
+
{
|
| 1053 |
+
"epoch": 2.6898395721925135,
|
| 1054 |
+
"grad_norm": 5.074891567230225,
|
| 1055 |
+
"learning_rate": 3.5261371521817244e-05,
|
| 1056 |
+
"loss": 0.0658,
|
| 1057 |
+
"step": 136
|
| 1058 |
+
},
|
| 1059 |
+
{
|
| 1060 |
+
"epoch": 2.711229946524064,
|
| 1061 |
+
"grad_norm": 3.6220922470092773,
|
| 1062 |
+
"learning_rate": 3.3895904484679984e-05,
|
| 1063 |
+
"loss": 0.1535,
|
| 1064 |
+
"step": 137
|
| 1065 |
+
},
|
| 1066 |
+
{
|
| 1067 |
+
"epoch": 2.732620320855615,
|
| 1068 |
+
"grad_norm": 3.9044840335845947,
|
| 1069 |
+
"learning_rate": 3.2551985985948616e-05,
|
| 1070 |
+
"loss": 0.0572,
|
| 1071 |
+
"step": 138
|
| 1072 |
}
|
| 1073 |
],
|
| 1074 |
"logging_steps": 1,
|
|
|
|
| 1088 |
"attributes": {}
|
| 1089 |
}
|
| 1090 |
},
|
| 1091 |
+
"total_flos": 1.3759516646572032e+16,
|
| 1092 |
"train_batch_size": 2,
|
| 1093 |
"trial_name": null,
|
| 1094 |
"trial_params": null
|