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 50624
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:85fca7563933c3e18b8617e3f002bbe02e8f791207486aa98af646e097dcbda7
|
| 3 |
size 50624
|
last-checkpoint/optimizer.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 118090
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:af5624d6ca53d1eaa55883aa010611be99360bd545dff58eb71be4a4ab9a05aa
|
| 3 |
size 118090
|
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 sha256:3e765160d11b81e65ffe3d78d9c3e2c32c3dd5233c33521084cd844020498388
|
| 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:e70710c409284f74d525f8db5cfaccc22a8afd29416f19c595da9242ec92d936
|
| 3 |
size 1064
|
last-checkpoint/trainer_state.json
CHANGED
|
@@ -1,9 +1,9 @@
|
|
| 1 |
{
|
| 2 |
-
"best_metric": 10.
|
| 3 |
-
"best_model_checkpoint": "miner_id_24/checkpoint-
|
| 4 |
-
"epoch": 0.
|
| 5 |
"eval_steps": 25,
|
| 6 |
-
"global_step":
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
@@ -747,6 +747,372 @@
|
|
| 747 |
"eval_samples_per_second": 535.978,
|
| 748 |
"eval_steps_per_second": 75.037,
|
| 749 |
"step": 100
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 750 |
}
|
| 751 |
],
|
| 752 |
"logging_steps": 1,
|
|
@@ -775,7 +1141,7 @@
|
|
| 775 |
"attributes": {}
|
| 776 |
}
|
| 777 |
},
|
| 778 |
-
"total_flos":
|
| 779 |
"train_batch_size": 1,
|
| 780 |
"trial_name": null,
|
| 781 |
"trial_params": null
|
|
|
|
| 1 |
{
|
| 2 |
+
"best_metric": 10.295632362365723,
|
| 3 |
+
"best_model_checkpoint": "miner_id_24/checkpoint-150",
|
| 4 |
+
"epoch": 0.023816493914389628,
|
| 5 |
"eval_steps": 25,
|
| 6 |
+
"global_step": 150,
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
|
|
| 747 |
"eval_samples_per_second": 535.978,
|
| 748 |
"eval_steps_per_second": 75.037,
|
| 749 |
"step": 100
|
| 750 |
+
},
|
| 751 |
+
{
|
| 752 |
+
"epoch": 0.016036439235689016,
|
| 753 |
+
"grad_norm": 0.24361427128314972,
|
| 754 |
+
"learning_rate": 0.0001599135876488549,
|
| 755 |
+
"loss": 10.3028,
|
| 756 |
+
"step": 101
|
| 757 |
+
},
|
| 758 |
+
{
|
| 759 |
+
"epoch": 0.016195215861784948,
|
| 760 |
+
"grad_norm": 0.23904871940612793,
|
| 761 |
+
"learning_rate": 0.00015743756320098332,
|
| 762 |
+
"loss": 10.2934,
|
| 763 |
+
"step": 102
|
| 764 |
+
},
|
| 765 |
+
{
|
| 766 |
+
"epoch": 0.01635399248788088,
|
| 767 |
+
"grad_norm": 0.19915081560611725,
|
| 768 |
+
"learning_rate": 0.0001549595053975962,
|
| 769 |
+
"loss": 10.3083,
|
| 770 |
+
"step": 103
|
| 771 |
+
},
|
| 772 |
+
{
|
| 773 |
+
"epoch": 0.016512769113976807,
|
| 774 |
+
"grad_norm": 0.15487897396087646,
|
| 775 |
+
"learning_rate": 0.00015248009171495378,
|
| 776 |
+
"loss": 10.3023,
|
| 777 |
+
"step": 104
|
| 778 |
+
},
|
| 779 |
+
{
|
| 780 |
+
"epoch": 0.01667154574007274,
|
| 781 |
+
"grad_norm": 0.17113174498081207,
|
| 782 |
+
"learning_rate": 0.00015,
|
| 783 |
+
"loss": 10.2996,
|
| 784 |
+
"step": 105
|
| 785 |
+
},
|
| 786 |
+
{
|
| 787 |
+
"epoch": 0.01683032236616867,
|
| 788 |
+
"grad_norm": 0.12119137495756149,
|
| 789 |
+
"learning_rate": 0.00014751990828504622,
|
| 790 |
+
"loss": 10.3016,
|
| 791 |
+
"step": 106
|
| 792 |
+
},
|
| 793 |
+
{
|
| 794 |
+
"epoch": 0.0169890989922646,
|
| 795 |
+
"grad_norm": 0.15259002149105072,
|
| 796 |
+
"learning_rate": 0.00014504049460240375,
|
| 797 |
+
"loss": 10.2929,
|
| 798 |
+
"step": 107
|
| 799 |
+
},
|
| 800 |
+
{
|
| 801 |
+
"epoch": 0.017147875618360533,
|
| 802 |
+
"grad_norm": 0.11336802691221237,
|
| 803 |
+
"learning_rate": 0.00014256243679901663,
|
| 804 |
+
"loss": 10.2947,
|
| 805 |
+
"step": 108
|
| 806 |
+
},
|
| 807 |
+
{
|
| 808 |
+
"epoch": 0.017306652244456464,
|
| 809 |
+
"grad_norm": 0.10615869611501694,
|
| 810 |
+
"learning_rate": 0.00014008641235114508,
|
| 811 |
+
"loss": 10.2989,
|
| 812 |
+
"step": 109
|
| 813 |
+
},
|
| 814 |
+
{
|
| 815 |
+
"epoch": 0.017465428870552392,
|
| 816 |
+
"grad_norm": 0.10214941948652267,
|
| 817 |
+
"learning_rate": 0.00013761309817915014,
|
| 818 |
+
"loss": 10.2982,
|
| 819 |
+
"step": 110
|
| 820 |
+
},
|
| 821 |
+
{
|
| 822 |
+
"epoch": 0.017624205496648324,
|
| 823 |
+
"grad_norm": 0.07917061448097229,
|
| 824 |
+
"learning_rate": 0.00013514317046243058,
|
| 825 |
+
"loss": 10.2973,
|
| 826 |
+
"step": 111
|
| 827 |
+
},
|
| 828 |
+
{
|
| 829 |
+
"epoch": 0.017782982122744255,
|
| 830 |
+
"grad_norm": 0.11092983931303024,
|
| 831 |
+
"learning_rate": 0.00013267730445456208,
|
| 832 |
+
"loss": 10.2984,
|
| 833 |
+
"step": 112
|
| 834 |
+
},
|
| 835 |
+
{
|
| 836 |
+
"epoch": 0.017941758748840186,
|
| 837 |
+
"grad_norm": 0.08627593517303467,
|
| 838 |
+
"learning_rate": 0.00013021617429868963,
|
| 839 |
+
"loss": 10.2967,
|
| 840 |
+
"step": 113
|
| 841 |
+
},
|
| 842 |
+
{
|
| 843 |
+
"epoch": 0.018100535374936118,
|
| 844 |
+
"grad_norm": 0.10853682458400726,
|
| 845 |
+
"learning_rate": 0.00012776045284322368,
|
| 846 |
+
"loss": 10.3083,
|
| 847 |
+
"step": 114
|
| 848 |
+
},
|
| 849 |
+
{
|
| 850 |
+
"epoch": 0.01825931200103205,
|
| 851 |
+
"grad_norm": 0.08355826884508133,
|
| 852 |
+
"learning_rate": 0.00012531081145788987,
|
| 853 |
+
"loss": 10.3039,
|
| 854 |
+
"step": 115
|
| 855 |
+
},
|
| 856 |
+
{
|
| 857 |
+
"epoch": 0.01841808862712798,
|
| 858 |
+
"grad_norm": 0.12834160029888153,
|
| 859 |
+
"learning_rate": 0.00012286791985018355,
|
| 860 |
+
"loss": 10.3058,
|
| 861 |
+
"step": 116
|
| 862 |
+
},
|
| 863 |
+
{
|
| 864 |
+
"epoch": 0.01857686525322391,
|
| 865 |
+
"grad_norm": 0.10183387994766235,
|
| 866 |
+
"learning_rate": 0.00012043244588227796,
|
| 867 |
+
"loss": 10.3207,
|
| 868 |
+
"step": 117
|
| 869 |
+
},
|
| 870 |
+
{
|
| 871 |
+
"epoch": 0.01873564187931984,
|
| 872 |
+
"grad_norm": 0.0810333862900734,
|
| 873 |
+
"learning_rate": 0.00011800505538843798,
|
| 874 |
+
"loss": 10.3022,
|
| 875 |
+
"step": 118
|
| 876 |
+
},
|
| 877 |
+
{
|
| 878 |
+
"epoch": 0.01889441850541577,
|
| 879 |
+
"grad_norm": 0.11709333211183548,
|
| 880 |
+
"learning_rate": 0.00011558641199298727,
|
| 881 |
+
"loss": 10.3086,
|
| 882 |
+
"step": 119
|
| 883 |
+
},
|
| 884 |
+
{
|
| 885 |
+
"epoch": 0.019053195131511703,
|
| 886 |
+
"grad_norm": 0.06999313831329346,
|
| 887 |
+
"learning_rate": 0.00011317717692888012,
|
| 888 |
+
"loss": 10.3066,
|
| 889 |
+
"step": 120
|
| 890 |
+
},
|
| 891 |
+
{
|
| 892 |
+
"epoch": 0.019211971757607634,
|
| 893 |
+
"grad_norm": 0.0975615531206131,
|
| 894 |
+
"learning_rate": 0.00011077800885692702,
|
| 895 |
+
"loss": 10.3069,
|
| 896 |
+
"step": 121
|
| 897 |
+
},
|
| 898 |
+
{
|
| 899 |
+
"epoch": 0.019370748383703566,
|
| 900 |
+
"grad_norm": 0.08743242919445038,
|
| 901 |
+
"learning_rate": 0.00010838956368572334,
|
| 902 |
+
"loss": 10.3202,
|
| 903 |
+
"step": 122
|
| 904 |
+
},
|
| 905 |
+
{
|
| 906 |
+
"epoch": 0.019529525009799494,
|
| 907 |
+
"grad_norm": 0.12698757648468018,
|
| 908 |
+
"learning_rate": 0.0001060124943923303,
|
| 909 |
+
"loss": 10.2906,
|
| 910 |
+
"step": 123
|
| 911 |
+
},
|
| 912 |
+
{
|
| 913 |
+
"epoch": 0.019688301635895425,
|
| 914 |
+
"grad_norm": 0.06964551657438278,
|
| 915 |
+
"learning_rate": 0.0001036474508437579,
|
| 916 |
+
"loss": 10.3047,
|
| 917 |
+
"step": 124
|
| 918 |
+
},
|
| 919 |
+
{
|
| 920 |
+
"epoch": 0.019847078261991356,
|
| 921 |
+
"grad_norm": 0.0764366015791893,
|
| 922 |
+
"learning_rate": 0.00010129507961929748,
|
| 923 |
+
"loss": 10.2964,
|
| 924 |
+
"step": 125
|
| 925 |
+
},
|
| 926 |
+
{
|
| 927 |
+
"epoch": 0.019847078261991356,
|
| 928 |
+
"eval_loss": 10.29774284362793,
|
| 929 |
+
"eval_runtime": 0.0936,
|
| 930 |
+
"eval_samples_per_second": 534.393,
|
| 931 |
+
"eval_steps_per_second": 74.815,
|
| 932 |
+
"step": 125
|
| 933 |
+
},
|
| 934 |
+
{
|
| 935 |
+
"epoch": 0.020005854888087288,
|
| 936 |
+
"grad_norm": 0.08608151227235794,
|
| 937 |
+
"learning_rate": 9.895602383375353e-05,
|
| 938 |
+
"loss": 10.2955,
|
| 939 |
+
"step": 126
|
| 940 |
+
},
|
| 941 |
+
{
|
| 942 |
+
"epoch": 0.02016463151418322,
|
| 943 |
+
"grad_norm": 0.08954727649688721,
|
| 944 |
+
"learning_rate": 9.663092296162251e-05,
|
| 945 |
+
"loss": 10.3007,
|
| 946 |
+
"step": 127
|
| 947 |
+
},
|
| 948 |
+
{
|
| 949 |
+
"epoch": 0.02032340814027915,
|
| 950 |
+
"grad_norm": 0.10686002671718597,
|
| 951 |
+
"learning_rate": 9.432041266226686e-05,
|
| 952 |
+
"loss": 10.3021,
|
| 953 |
+
"step": 128
|
| 954 |
+
},
|
| 955 |
+
{
|
| 956 |
+
"epoch": 0.02048218476637508,
|
| 957 |
+
"grad_norm": 0.08813920617103577,
|
| 958 |
+
"learning_rate": 9.202512460613219e-05,
|
| 959 |
+
"loss": 10.3064,
|
| 960 |
+
"step": 129
|
| 961 |
+
},
|
| 962 |
+
{
|
| 963 |
+
"epoch": 0.02064096139247101,
|
| 964 |
+
"grad_norm": 0.07965640723705292,
|
| 965 |
+
"learning_rate": 8.97456863020546e-05,
|
| 966 |
+
"loss": 10.299,
|
| 967 |
+
"step": 130
|
| 968 |
+
},
|
| 969 |
+
{
|
| 970 |
+
"epoch": 0.02079973801856694,
|
| 971 |
+
"grad_norm": 0.09392104297876358,
|
| 972 |
+
"learning_rate": 8.748272092570646e-05,
|
| 973 |
+
"loss": 10.2917,
|
| 974 |
+
"step": 131
|
| 975 |
+
},
|
| 976 |
+
{
|
| 977 |
+
"epoch": 0.020958514644662873,
|
| 978 |
+
"grad_norm": 0.08309674263000488,
|
| 979 |
+
"learning_rate": 8.523684714922608e-05,
|
| 980 |
+
"loss": 10.2916,
|
| 981 |
+
"step": 132
|
| 982 |
+
},
|
| 983 |
+
{
|
| 984 |
+
"epoch": 0.021117291270758804,
|
| 985 |
+
"grad_norm": 0.09519728273153305,
|
| 986 |
+
"learning_rate": 8.300867897207903e-05,
|
| 987 |
+
"loss": 10.3004,
|
| 988 |
+
"step": 133
|
| 989 |
+
},
|
| 990 |
+
{
|
| 991 |
+
"epoch": 0.021276067896854736,
|
| 992 |
+
"grad_norm": 0.06975460052490234,
|
| 993 |
+
"learning_rate": 8.079882555319684e-05,
|
| 994 |
+
"loss": 10.3025,
|
| 995 |
+
"step": 134
|
| 996 |
+
},
|
| 997 |
+
{
|
| 998 |
+
"epoch": 0.021434844522950663,
|
| 999 |
+
"grad_norm": 0.07696589082479477,
|
| 1000 |
+
"learning_rate": 7.860789104443896e-05,
|
| 1001 |
+
"loss": 10.2976,
|
| 1002 |
+
"step": 135
|
| 1003 |
+
},
|
| 1004 |
+
{
|
| 1005 |
+
"epoch": 0.021593621149046595,
|
| 1006 |
+
"grad_norm": 0.10613939166069031,
|
| 1007 |
+
"learning_rate": 7.643647442542382e-05,
|
| 1008 |
+
"loss": 10.3022,
|
| 1009 |
+
"step": 136
|
| 1010 |
+
},
|
| 1011 |
+
{
|
| 1012 |
+
"epoch": 0.021752397775142526,
|
| 1013 |
+
"grad_norm": 0.11416519433259964,
|
| 1014 |
+
"learning_rate": 7.428516933977347e-05,
|
| 1015 |
+
"loss": 10.3094,
|
| 1016 |
+
"step": 137
|
| 1017 |
+
},
|
| 1018 |
+
{
|
| 1019 |
+
"epoch": 0.021911174401238458,
|
| 1020 |
+
"grad_norm": 0.0883881077170372,
|
| 1021 |
+
"learning_rate": 7.215456393281776e-05,
|
| 1022 |
+
"loss": 10.2996,
|
| 1023 |
+
"step": 138
|
| 1024 |
+
},
|
| 1025 |
+
{
|
| 1026 |
+
"epoch": 0.02206995102733439,
|
| 1027 |
+
"grad_norm": 0.10318446159362793,
|
| 1028 |
+
"learning_rate": 7.004524069080096e-05,
|
| 1029 |
+
"loss": 10.3031,
|
| 1030 |
+
"step": 139
|
| 1031 |
+
},
|
| 1032 |
+
{
|
| 1033 |
+
"epoch": 0.02222872765343032,
|
| 1034 |
+
"grad_norm": 0.10122831910848618,
|
| 1035 |
+
"learning_rate": 6.795777628163599e-05,
|
| 1036 |
+
"loss": 10.2992,
|
| 1037 |
+
"step": 140
|
| 1038 |
+
},
|
| 1039 |
+
{
|
| 1040 |
+
"epoch": 0.022387504279526252,
|
| 1041 |
+
"grad_norm": 0.0863846093416214,
|
| 1042 |
+
"learning_rate": 6.58927413972491e-05,
|
| 1043 |
+
"loss": 10.307,
|
| 1044 |
+
"step": 141
|
| 1045 |
+
},
|
| 1046 |
+
{
|
| 1047 |
+
"epoch": 0.02254628090562218,
|
| 1048 |
+
"grad_norm": 0.10809484869241714,
|
| 1049 |
+
"learning_rate": 6.385070059755846e-05,
|
| 1050 |
+
"loss": 10.3024,
|
| 1051 |
+
"step": 142
|
| 1052 |
+
},
|
| 1053 |
+
{
|
| 1054 |
+
"epoch": 0.02270505753171811,
|
| 1055 |
+
"grad_norm": 0.0929742157459259,
|
| 1056 |
+
"learning_rate": 6.183221215612904e-05,
|
| 1057 |
+
"loss": 10.2905,
|
| 1058 |
+
"step": 143
|
| 1059 |
+
},
|
| 1060 |
+
{
|
| 1061 |
+
"epoch": 0.022863834157814043,
|
| 1062 |
+
"grad_norm": 0.11549130827188492,
|
| 1063 |
+
"learning_rate": 5.983782790754623e-05,
|
| 1064 |
+
"loss": 10.2985,
|
| 1065 |
+
"step": 144
|
| 1066 |
+
},
|
| 1067 |
+
{
|
| 1068 |
+
"epoch": 0.023022610783909974,
|
| 1069 |
+
"grad_norm": 0.13555151224136353,
|
| 1070 |
+
"learning_rate": 5.786809309654982e-05,
|
| 1071 |
+
"loss": 10.2905,
|
| 1072 |
+
"step": 145
|
| 1073 |
+
},
|
| 1074 |
+
{
|
| 1075 |
+
"epoch": 0.023181387410005905,
|
| 1076 |
+
"grad_norm": 0.08594315499067307,
|
| 1077 |
+
"learning_rate": 5.592354622896944e-05,
|
| 1078 |
+
"loss": 10.2922,
|
| 1079 |
+
"step": 146
|
| 1080 |
+
},
|
| 1081 |
+
{
|
| 1082 |
+
"epoch": 0.023340164036101837,
|
| 1083 |
+
"grad_norm": 0.1327342391014099,
|
| 1084 |
+
"learning_rate": 5.40047189245025e-05,
|
| 1085 |
+
"loss": 10.2812,
|
| 1086 |
+
"step": 147
|
| 1087 |
+
},
|
| 1088 |
+
{
|
| 1089 |
+
"epoch": 0.023498940662197765,
|
| 1090 |
+
"grad_norm": 0.09222450852394104,
|
| 1091 |
+
"learning_rate": 5.211213577137469e-05,
|
| 1092 |
+
"loss": 10.2914,
|
| 1093 |
+
"step": 148
|
| 1094 |
+
},
|
| 1095 |
+
{
|
| 1096 |
+
"epoch": 0.023657717288293696,
|
| 1097 |
+
"grad_norm": 0.12407458573579788,
|
| 1098 |
+
"learning_rate": 5.024631418292274e-05,
|
| 1099 |
+
"loss": 10.2818,
|
| 1100 |
+
"step": 149
|
| 1101 |
+
},
|
| 1102 |
+
{
|
| 1103 |
+
"epoch": 0.023816493914389628,
|
| 1104 |
+
"grad_norm": 0.10363201797008514,
|
| 1105 |
+
"learning_rate": 4.840776425613886e-05,
|
| 1106 |
+
"loss": 10.2824,
|
| 1107 |
+
"step": 150
|
| 1108 |
+
},
|
| 1109 |
+
{
|
| 1110 |
+
"epoch": 0.023816493914389628,
|
| 1111 |
+
"eval_loss": 10.295632362365723,
|
| 1112 |
+
"eval_runtime": 0.093,
|
| 1113 |
+
"eval_samples_per_second": 537.649,
|
| 1114 |
+
"eval_steps_per_second": 75.271,
|
| 1115 |
+
"step": 150
|
| 1116 |
}
|
| 1117 |
],
|
| 1118 |
"logging_steps": 1,
|
|
|
|
| 1141 |
"attributes": {}
|
| 1142 |
}
|
| 1143 |
},
|
| 1144 |
+
"total_flos": 16671404851200.0,
|
| 1145 |
"train_batch_size": 1,
|
| 1146 |
"trial_name": null,
|
| 1147 |
"trial_params": null
|