Training in progress, step 150, 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 639691872
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:ab04bf573bb548b14bc1e4c23c22e82709e2045080c9240f74bca41fed448b35
|
| 3 |
size 639691872
|
last-checkpoint/optimizer.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 325339796
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:b164e1a32ffbf1d41ebd3d733213f5731c6b474578ae5e8f5250d25bbd5cb190
|
| 3 |
size 325339796
|
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 sha256:2f5c150069ac4edb3b4fa5932ee41620b748c614c83563cbba626c7e7fbf790a
|
| 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 sha256:61638fd1c8ee876b663b88bf9611e79d360655e5839812fcd5fc993296ace5e9
|
| 3 |
size 1064
|
last-checkpoint/trainer_state.json
CHANGED
|
@@ -1,9 +1,9 @@
|
|
| 1 |
{
|
| 2 |
-
"best_metric": 1.
|
| 3 |
-
"best_model_checkpoint": "miner_id_24/checkpoint-
|
| 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 +731,364 @@
|
|
| 731 |
"eval_samples_per_second": 9.262,
|
| 732 |
"eval_steps_per_second": 2.335,
|
| 733 |
"step": 100
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 734 |
}
|
| 735 |
],
|
| 736 |
"logging_steps": 1,
|
|
@@ -759,7 +1117,7 @@
|
|
| 759 |
"attributes": {}
|
| 760 |
}
|
| 761 |
},
|
| 762 |
-
"total_flos":
|
| 763 |
"train_batch_size": 8,
|
| 764 |
"trial_name": null,
|
| 765 |
"trial_params": null
|
|
|
|
| 1 |
{
|
| 2 |
+
"best_metric": 1.1165207624435425,
|
| 3 |
+
"best_model_checkpoint": "miner_id_24/checkpoint-150",
|
| 4 |
+
"epoch": 0.7092198581560284,
|
| 5 |
"eval_steps": 50,
|
| 6 |
+
"global_step": 150,
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
|
|
| 731 |
"eval_samples_per_second": 9.262,
|
| 732 |
"eval_steps_per_second": 2.335,
|
| 733 |
"step": 100
|
| 734 |
+
},
|
| 735 |
+
{
|
| 736 |
+
"epoch": 0.47754137115839246,
|
| 737 |
+
"grad_norm": 0.6622450351715088,
|
| 738 |
+
"learning_rate": 5.774490901070424e-05,
|
| 739 |
+
"loss": 1.2761,
|
| 740 |
+
"step": 101
|
| 741 |
+
},
|
| 742 |
+
{
|
| 743 |
+
"epoch": 0.48226950354609927,
|
| 744 |
+
"grad_norm": 0.49634912610054016,
|
| 745 |
+
"learning_rate": 5.697576694719616e-05,
|
| 746 |
+
"loss": 1.2634,
|
| 747 |
+
"step": 102
|
| 748 |
+
},
|
| 749 |
+
{
|
| 750 |
+
"epoch": 0.48699763593380613,
|
| 751 |
+
"grad_norm": 0.41559910774230957,
|
| 752 |
+
"learning_rate": 5.620493763066297e-05,
|
| 753 |
+
"loss": 1.2799,
|
| 754 |
+
"step": 103
|
| 755 |
+
},
|
| 756 |
+
{
|
| 757 |
+
"epoch": 0.491725768321513,
|
| 758 |
+
"grad_norm": 0.38408902287483215,
|
| 759 |
+
"learning_rate": 5.543260750427373e-05,
|
| 760 |
+
"loss": 1.1719,
|
| 761 |
+
"step": 104
|
| 762 |
+
},
|
| 763 |
+
{
|
| 764 |
+
"epoch": 0.49645390070921985,
|
| 765 |
+
"grad_norm": 0.3431536555290222,
|
| 766 |
+
"learning_rate": 5.465896337420359e-05,
|
| 767 |
+
"loss": 1.1056,
|
| 768 |
+
"step": 105
|
| 769 |
+
},
|
| 770 |
+
{
|
| 771 |
+
"epoch": 0.5011820330969267,
|
| 772 |
+
"grad_norm": 0.342176228761673,
|
| 773 |
+
"learning_rate": 5.3884192364450325e-05,
|
| 774 |
+
"loss": 1.2039,
|
| 775 |
+
"step": 106
|
| 776 |
+
},
|
| 777 |
+
{
|
| 778 |
+
"epoch": 0.5059101654846335,
|
| 779 |
+
"grad_norm": 0.37251392006874084,
|
| 780 |
+
"learning_rate": 5.3108481871574036e-05,
|
| 781 |
+
"loss": 1.223,
|
| 782 |
+
"step": 107
|
| 783 |
+
},
|
| 784 |
+
{
|
| 785 |
+
"epoch": 0.5106382978723404,
|
| 786 |
+
"grad_norm": 0.3797588646411896,
|
| 787 |
+
"learning_rate": 5.233201951937088e-05,
|
| 788 |
+
"loss": 1.1501,
|
| 789 |
+
"step": 108
|
| 790 |
+
},
|
| 791 |
+
{
|
| 792 |
+
"epoch": 0.5153664302600472,
|
| 793 |
+
"grad_norm": 0.3910278081893921,
|
| 794 |
+
"learning_rate": 5.155499311349185e-05,
|
| 795 |
+
"loss": 1.2336,
|
| 796 |
+
"step": 109
|
| 797 |
+
},
|
| 798 |
+
{
|
| 799 |
+
"epoch": 0.5200945626477541,
|
| 800 |
+
"grad_norm": 0.3525397479534149,
|
| 801 |
+
"learning_rate": 5.077759059601755e-05,
|
| 802 |
+
"loss": 1.1213,
|
| 803 |
+
"step": 110
|
| 804 |
+
},
|
| 805 |
+
{
|
| 806 |
+
"epoch": 0.524822695035461,
|
| 807 |
+
"grad_norm": 0.3806985020637512,
|
| 808 |
+
"learning_rate": 5e-05,
|
| 809 |
+
"loss": 1.2013,
|
| 810 |
+
"step": 111
|
| 811 |
+
},
|
| 812 |
+
{
|
| 813 |
+
"epoch": 0.5295508274231678,
|
| 814 |
+
"grad_norm": 0.38082411885261536,
|
| 815 |
+
"learning_rate": 4.9222409403982453e-05,
|
| 816 |
+
"loss": 1.1766,
|
| 817 |
+
"step": 112
|
| 818 |
+
},
|
| 819 |
+
{
|
| 820 |
+
"epoch": 0.5342789598108747,
|
| 821 |
+
"grad_norm": 0.35277897119522095,
|
| 822 |
+
"learning_rate": 4.844500688650816e-05,
|
| 823 |
+
"loss": 1.1419,
|
| 824 |
+
"step": 113
|
| 825 |
+
},
|
| 826 |
+
{
|
| 827 |
+
"epoch": 0.5390070921985816,
|
| 828 |
+
"grad_norm": 0.3725706934928894,
|
| 829 |
+
"learning_rate": 4.7667980480629124e-05,
|
| 830 |
+
"loss": 1.1276,
|
| 831 |
+
"step": 114
|
| 832 |
+
},
|
| 833 |
+
{
|
| 834 |
+
"epoch": 0.5437352245862884,
|
| 835 |
+
"grad_norm": 0.3541959226131439,
|
| 836 |
+
"learning_rate": 4.6891518128425976e-05,
|
| 837 |
+
"loss": 1.0653,
|
| 838 |
+
"step": 115
|
| 839 |
+
},
|
| 840 |
+
{
|
| 841 |
+
"epoch": 0.5484633569739953,
|
| 842 |
+
"grad_norm": 0.371341347694397,
|
| 843 |
+
"learning_rate": 4.611580763554969e-05,
|
| 844 |
+
"loss": 1.1224,
|
| 845 |
+
"step": 116
|
| 846 |
+
},
|
| 847 |
+
{
|
| 848 |
+
"epoch": 0.5531914893617021,
|
| 849 |
+
"grad_norm": 0.38338702917099,
|
| 850 |
+
"learning_rate": 4.534103662579642e-05,
|
| 851 |
+
"loss": 1.1465,
|
| 852 |
+
"step": 117
|
| 853 |
+
},
|
| 854 |
+
{
|
| 855 |
+
"epoch": 0.557919621749409,
|
| 856 |
+
"grad_norm": 0.3807014226913452,
|
| 857 |
+
"learning_rate": 4.456739249572627e-05,
|
| 858 |
+
"loss": 1.1659,
|
| 859 |
+
"step": 118
|
| 860 |
+
},
|
| 861 |
+
{
|
| 862 |
+
"epoch": 0.5626477541371159,
|
| 863 |
+
"grad_norm": 0.38342806696891785,
|
| 864 |
+
"learning_rate": 4.3795062369337034e-05,
|
| 865 |
+
"loss": 1.1965,
|
| 866 |
+
"step": 119
|
| 867 |
+
},
|
| 868 |
+
{
|
| 869 |
+
"epoch": 0.5673758865248227,
|
| 870 |
+
"grad_norm": 0.39216044545173645,
|
| 871 |
+
"learning_rate": 4.3024233052803855e-05,
|
| 872 |
+
"loss": 1.2172,
|
| 873 |
+
"step": 120
|
| 874 |
+
},
|
| 875 |
+
{
|
| 876 |
+
"epoch": 0.5721040189125296,
|
| 877 |
+
"grad_norm": 0.383404016494751,
|
| 878 |
+
"learning_rate": 4.2255090989295764e-05,
|
| 879 |
+
"loss": 1.1739,
|
| 880 |
+
"step": 121
|
| 881 |
+
},
|
| 882 |
+
{
|
| 883 |
+
"epoch": 0.5768321513002365,
|
| 884 |
+
"grad_norm": 0.38151639699935913,
|
| 885 |
+
"learning_rate": 4.148782221388007e-05,
|
| 886 |
+
"loss": 1.0698,
|
| 887 |
+
"step": 122
|
| 888 |
+
},
|
| 889 |
+
{
|
| 890 |
+
"epoch": 0.5815602836879432,
|
| 891 |
+
"grad_norm": 0.35033509135246277,
|
| 892 |
+
"learning_rate": 4.0722612308525335e-05,
|
| 893 |
+
"loss": 1.121,
|
| 894 |
+
"step": 123
|
| 895 |
+
},
|
| 896 |
+
{
|
| 897 |
+
"epoch": 0.5862884160756501,
|
| 898 |
+
"grad_norm": 0.36325815320014954,
|
| 899 |
+
"learning_rate": 3.9959646357214084e-05,
|
| 900 |
+
"loss": 1.0989,
|
| 901 |
+
"step": 124
|
| 902 |
+
},
|
| 903 |
+
{
|
| 904 |
+
"epoch": 0.5910165484633569,
|
| 905 |
+
"grad_norm": 0.38052505254745483,
|
| 906 |
+
"learning_rate": 3.919910890117584e-05,
|
| 907 |
+
"loss": 1.0776,
|
| 908 |
+
"step": 125
|
| 909 |
+
},
|
| 910 |
+
{
|
| 911 |
+
"epoch": 0.5957446808510638,
|
| 912 |
+
"grad_norm": 0.3674011826515198,
|
| 913 |
+
"learning_rate": 3.844118389425153e-05,
|
| 914 |
+
"loss": 1.12,
|
| 915 |
+
"step": 126
|
| 916 |
+
},
|
| 917 |
+
{
|
| 918 |
+
"epoch": 0.6004728132387707,
|
| 919 |
+
"grad_norm": 0.3708929717540741,
|
| 920 |
+
"learning_rate": 3.7686054658399935e-05,
|
| 921 |
+
"loss": 1.1564,
|
| 922 |
+
"step": 127
|
| 923 |
+
},
|
| 924 |
+
{
|
| 925 |
+
"epoch": 0.6052009456264775,
|
| 926 |
+
"grad_norm": 0.37105265259742737,
|
| 927 |
+
"learning_rate": 3.693390383935698e-05,
|
| 928 |
+
"loss": 1.1331,
|
| 929 |
+
"step": 128
|
| 930 |
+
},
|
| 931 |
+
{
|
| 932 |
+
"epoch": 0.6099290780141844,
|
| 933 |
+
"grad_norm": 0.38267219066619873,
|
| 934 |
+
"learning_rate": 3.618491336245849e-05,
|
| 935 |
+
"loss": 1.0379,
|
| 936 |
+
"step": 129
|
| 937 |
+
},
|
| 938 |
+
{
|
| 939 |
+
"epoch": 0.6146572104018913,
|
| 940 |
+
"grad_norm": 0.3863324522972107,
|
| 941 |
+
"learning_rate": 3.5439264388637405e-05,
|
| 942 |
+
"loss": 1.1361,
|
| 943 |
+
"step": 130
|
| 944 |
+
},
|
| 945 |
+
{
|
| 946 |
+
"epoch": 0.6193853427895981,
|
| 947 |
+
"grad_norm": 0.3742581307888031,
|
| 948 |
+
"learning_rate": 3.469713727060564e-05,
|
| 949 |
+
"loss": 1.0579,
|
| 950 |
+
"step": 131
|
| 951 |
+
},
|
| 952 |
+
{
|
| 953 |
+
"epoch": 0.624113475177305,
|
| 954 |
+
"grad_norm": 0.3845396041870117,
|
| 955 |
+
"learning_rate": 3.395871150923163e-05,
|
| 956 |
+
"loss": 1.0839,
|
| 957 |
+
"step": 132
|
| 958 |
+
},
|
| 959 |
+
{
|
| 960 |
+
"epoch": 0.6288416075650118,
|
| 961 |
+
"grad_norm": 0.4115036129951477,
|
| 962 |
+
"learning_rate": 3.3224165710123756e-05,
|
| 963 |
+
"loss": 1.1757,
|
| 964 |
+
"step": 133
|
| 965 |
+
},
|
| 966 |
+
{
|
| 967 |
+
"epoch": 0.6335697399527187,
|
| 968 |
+
"grad_norm": 0.39382243156433105,
|
| 969 |
+
"learning_rate": 3.249367754043047e-05,
|
| 970 |
+
"loss": 1.1352,
|
| 971 |
+
"step": 134
|
| 972 |
+
},
|
| 973 |
+
{
|
| 974 |
+
"epoch": 0.6382978723404256,
|
| 975 |
+
"grad_norm": 0.39766818284988403,
|
| 976 |
+
"learning_rate": 3.176742368586725e-05,
|
| 977 |
+
"loss": 1.1353,
|
| 978 |
+
"step": 135
|
| 979 |
+
},
|
| 980 |
+
{
|
| 981 |
+
"epoch": 0.6430260047281324,
|
| 982 |
+
"grad_norm": 0.3955276310443878,
|
| 983 |
+
"learning_rate": 3.104557980798104e-05,
|
| 984 |
+
"loss": 1.1329,
|
| 985 |
+
"step": 136
|
| 986 |
+
},
|
| 987 |
+
{
|
| 988 |
+
"epoch": 0.6477541371158393,
|
| 989 |
+
"grad_norm": 0.4006507396697998,
|
| 990 |
+
"learning_rate": 3.032832050166239e-05,
|
| 991 |
+
"loss": 1.0925,
|
| 992 |
+
"step": 137
|
| 993 |
+
},
|
| 994 |
+
{
|
| 995 |
+
"epoch": 0.6524822695035462,
|
| 996 |
+
"grad_norm": 0.37227386236190796,
|
| 997 |
+
"learning_rate": 2.9615819252915565e-05,
|
| 998 |
+
"loss": 1.0568,
|
| 999 |
+
"step": 138
|
| 1000 |
+
},
|
| 1001 |
+
{
|
| 1002 |
+
"epoch": 0.6572104018912529,
|
| 1003 |
+
"grad_norm": 0.4012226462364197,
|
| 1004 |
+
"learning_rate": 2.890824839689689e-05,
|
| 1005 |
+
"loss": 1.1103,
|
| 1006 |
+
"step": 139
|
| 1007 |
+
},
|
| 1008 |
+
{
|
| 1009 |
+
"epoch": 0.6619385342789598,
|
| 1010 |
+
"grad_norm": 0.398710697889328,
|
| 1011 |
+
"learning_rate": 2.8205779076231447e-05,
|
| 1012 |
+
"loss": 1.086,
|
| 1013 |
+
"step": 140
|
| 1014 |
+
},
|
| 1015 |
+
{
|
| 1016 |
+
"epoch": 0.6666666666666666,
|
| 1017 |
+
"grad_norm": 0.38928738236427307,
|
| 1018 |
+
"learning_rate": 2.750858119961821e-05,
|
| 1019 |
+
"loss": 1.1066,
|
| 1020 |
+
"step": 141
|
| 1021 |
+
},
|
| 1022 |
+
{
|
| 1023 |
+
"epoch": 0.6713947990543735,
|
| 1024 |
+
"grad_norm": 0.41606059670448303,
|
| 1025 |
+
"learning_rate": 2.6816823400733625e-05,
|
| 1026 |
+
"loss": 1.1212,
|
| 1027 |
+
"step": 142
|
| 1028 |
+
},
|
| 1029 |
+
{
|
| 1030 |
+
"epoch": 0.6761229314420804,
|
| 1031 |
+
"grad_norm": 0.4255363345146179,
|
| 1032 |
+
"learning_rate": 2.613067299744364e-05,
|
| 1033 |
+
"loss": 1.0511,
|
| 1034 |
+
"step": 143
|
| 1035 |
+
},
|
| 1036 |
+
{
|
| 1037 |
+
"epoch": 0.6808510638297872,
|
| 1038 |
+
"grad_norm": 0.422050416469574,
|
| 1039 |
+
"learning_rate": 2.5450295951333896e-05,
|
| 1040 |
+
"loss": 1.1598,
|
| 1041 |
+
"step": 144
|
| 1042 |
+
},
|
| 1043 |
+
{
|
| 1044 |
+
"epoch": 0.6855791962174941,
|
| 1045 |
+
"grad_norm": 0.46749165654182434,
|
| 1046 |
+
"learning_rate": 2.4775856827568016e-05,
|
| 1047 |
+
"loss": 1.111,
|
| 1048 |
+
"step": 145
|
| 1049 |
+
},
|
| 1050 |
+
{
|
| 1051 |
+
"epoch": 0.6903073286052009,
|
| 1052 |
+
"grad_norm": 0.4487079083919525,
|
| 1053 |
+
"learning_rate": 2.410751875508373e-05,
|
| 1054 |
+
"loss": 1.0855,
|
| 1055 |
+
"step": 146
|
| 1056 |
+
},
|
| 1057 |
+
{
|
| 1058 |
+
"epoch": 0.6950354609929078,
|
| 1059 |
+
"grad_norm": 0.45914727449417114,
|
| 1060 |
+
"learning_rate": 2.3445443387136244e-05,
|
| 1061 |
+
"loss": 1.0781,
|
| 1062 |
+
"step": 147
|
| 1063 |
+
},
|
| 1064 |
+
{
|
| 1065 |
+
"epoch": 0.6997635933806147,
|
| 1066 |
+
"grad_norm": 0.4794670045375824,
|
| 1067 |
+
"learning_rate": 2.2789790862198628e-05,
|
| 1068 |
+
"loss": 1.0906,
|
| 1069 |
+
"step": 148
|
| 1070 |
+
},
|
| 1071 |
+
{
|
| 1072 |
+
"epoch": 0.7044917257683215,
|
| 1073 |
+
"grad_norm": 0.49251869320869446,
|
| 1074 |
+
"learning_rate": 2.2140719765228584e-05,
|
| 1075 |
+
"loss": 1.1217,
|
| 1076 |
+
"step": 149
|
| 1077 |
+
},
|
| 1078 |
+
{
|
| 1079 |
+
"epoch": 0.7092198581560284,
|
| 1080 |
+
"grad_norm": 0.6213652491569519,
|
| 1081 |
+
"learning_rate": 2.1498387089310868e-05,
|
| 1082 |
+
"loss": 1.1458,
|
| 1083 |
+
"step": 150
|
| 1084 |
+
},
|
| 1085 |
+
{
|
| 1086 |
+
"epoch": 0.7092198581560284,
|
| 1087 |
+
"eval_loss": 1.1165207624435425,
|
| 1088 |
+
"eval_runtime": 38.5292,
|
| 1089 |
+
"eval_samples_per_second": 9.266,
|
| 1090 |
+
"eval_steps_per_second": 2.336,
|
| 1091 |
+
"step": 150
|
| 1092 |
}
|
| 1093 |
],
|
| 1094 |
"logging_steps": 1,
|
|
|
|
| 1117 |
"attributes": {}
|
| 1118 |
}
|
| 1119 |
},
|
| 1120 |
+
"total_flos": 2.0057223796988314e+17,
|
| 1121 |
"train_batch_size": 8,
|
| 1122 |
"trial_name": null,
|
| 1123 |
"trial_params": null
|