Training in progress, step 1500, 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 295488936
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:c3001072512cf0094aa413adef722b38d30d55f1d695532e69f11d0e79e17410
|
| 3 |
size 295488936
|
last-checkpoint/optimizer.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 591203178
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:2f6d21674325a8aaab36b8cf4642a6d5958787b319ca784ac8dfd0a1718a3756
|
| 3 |
size 591203178
|
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:c3fcb8b7132fdda989f7bbb14a5bf464435849629fe731ccbc64c4724068a57e
|
| 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:e579802271638ff75fe7ba64560b3e21e4f7e26236b794157498845ba12537a4
|
| 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": 100,
|
| 6 |
-
"global_step":
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
@@ -787,6 +787,396 @@
|
|
| 787 |
"eval_news_finetune_val_samples_per_second": 1.395,
|
| 788 |
"eval_news_finetune_val_steps_per_second": 1.395,
|
| 789 |
"step": 1000
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 790 |
}
|
| 791 |
],
|
| 792 |
"logging_steps": 10,
|
|
@@ -806,7 +1196,7 @@
|
|
| 806 |
"attributes": {}
|
| 807 |
}
|
| 808 |
},
|
| 809 |
-
"total_flos":
|
| 810 |
"train_batch_size": 1,
|
| 811 |
"trial_name": null,
|
| 812 |
"trial_params": null
|
|
|
|
| 1 |
{
|
| 2 |
"best_metric": null,
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 1.0714540907466952,
|
| 5 |
"eval_steps": 100,
|
| 6 |
+
"global_step": 1500,
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
|
|
| 787 |
"eval_news_finetune_val_samples_per_second": 1.395,
|
| 788 |
"eval_news_finetune_val_steps_per_second": 1.395,
|
| 789 |
"step": 1000
|
| 790 |
+
},
|
| 791 |
+
{
|
| 792 |
+
"epoch": 0.721686316541622,
|
| 793 |
+
"grad_norm": 1.5076738595962524,
|
| 794 |
+
"learning_rate": 9.409912607418172e-05,
|
| 795 |
+
"loss": 0.268,
|
| 796 |
+
"step": 1010
|
| 797 |
+
},
|
| 798 |
+
{
|
| 799 |
+
"epoch": 0.7288317256162915,
|
| 800 |
+
"grad_norm": 3.3230276107788086,
|
| 801 |
+
"learning_rate": 9.390160386775895e-05,
|
| 802 |
+
"loss": 0.3038,
|
| 803 |
+
"step": 1020
|
| 804 |
+
},
|
| 805 |
+
{
|
| 806 |
+
"epoch": 0.735977134690961,
|
| 807 |
+
"grad_norm": 1.699854850769043,
|
| 808 |
+
"learning_rate": 9.370104438953125e-05,
|
| 809 |
+
"loss": 0.2869,
|
| 810 |
+
"step": 1030
|
| 811 |
+
},
|
| 812 |
+
{
|
| 813 |
+
"epoch": 0.7431225437656306,
|
| 814 |
+
"grad_norm": 0.904507577419281,
|
| 815 |
+
"learning_rate": 9.349746151492902e-05,
|
| 816 |
+
"loss": 0.289,
|
| 817 |
+
"step": 1040
|
| 818 |
+
},
|
| 819 |
+
{
|
| 820 |
+
"epoch": 0.7502679528403001,
|
| 821 |
+
"grad_norm": 0.9463105201721191,
|
| 822 |
+
"learning_rate": 9.329086932855215e-05,
|
| 823 |
+
"loss": 0.3729,
|
| 824 |
+
"step": 1050
|
| 825 |
+
},
|
| 826 |
+
{
|
| 827 |
+
"epoch": 0.7574133619149697,
|
| 828 |
+
"grad_norm": 1.4746607542037964,
|
| 829 |
+
"learning_rate": 9.30812821231956e-05,
|
| 830 |
+
"loss": 0.2282,
|
| 831 |
+
"step": 1060
|
| 832 |
+
},
|
| 833 |
+
{
|
| 834 |
+
"epoch": 0.7645587709896392,
|
| 835 |
+
"grad_norm": 1.0270076990127563,
|
| 836 |
+
"learning_rate": 9.286871439886058e-05,
|
| 837 |
+
"loss": 0.3029,
|
| 838 |
+
"step": 1070
|
| 839 |
+
},
|
| 840 |
+
{
|
| 841 |
+
"epoch": 0.7717041800643086,
|
| 842 |
+
"grad_norm": 2.0656538009643555,
|
| 843 |
+
"learning_rate": 9.265318086175143e-05,
|
| 844 |
+
"loss": 0.3268,
|
| 845 |
+
"step": 1080
|
| 846 |
+
},
|
| 847 |
+
{
|
| 848 |
+
"epoch": 0.7788495891389782,
|
| 849 |
+
"grad_norm": 0.9798826575279236,
|
| 850 |
+
"learning_rate": 9.243469642325805e-05,
|
| 851 |
+
"loss": 0.2942,
|
| 852 |
+
"step": 1090
|
| 853 |
+
},
|
| 854 |
+
{
|
| 855 |
+
"epoch": 0.7859949982136477,
|
| 856 |
+
"grad_norm": 1.1419672966003418,
|
| 857 |
+
"learning_rate": 9.221327619892452e-05,
|
| 858 |
+
"loss": 0.3266,
|
| 859 |
+
"step": 1100
|
| 860 |
+
},
|
| 861 |
+
{
|
| 862 |
+
"epoch": 0.7859949982136477,
|
| 863 |
+
"eval_news_finetune_val_loss": 0.307956337928772,
|
| 864 |
+
"eval_news_finetune_val_runtime": 1003.1873,
|
| 865 |
+
"eval_news_finetune_val_samples_per_second": 1.396,
|
| 866 |
+
"eval_news_finetune_val_steps_per_second": 1.396,
|
| 867 |
+
"step": 1100
|
| 868 |
+
},
|
| 869 |
+
{
|
| 870 |
+
"epoch": 0.7931404072883173,
|
| 871 |
+
"grad_norm": 0.6810228228569031,
|
| 872 |
+
"learning_rate": 9.198893550740306e-05,
|
| 873 |
+
"loss": 0.3596,
|
| 874 |
+
"step": 1110
|
| 875 |
+
},
|
| 876 |
+
{
|
| 877 |
+
"epoch": 0.8002858163629868,
|
| 878 |
+
"grad_norm": 1.6553049087524414,
|
| 879 |
+
"learning_rate": 9.176168986939446e-05,
|
| 880 |
+
"loss": 0.3106,
|
| 881 |
+
"step": 1120
|
| 882 |
+
},
|
| 883 |
+
{
|
| 884 |
+
"epoch": 0.8074312254376563,
|
| 885 |
+
"grad_norm": 0.7749443650245667,
|
| 886 |
+
"learning_rate": 9.153155500657422e-05,
|
| 887 |
+
"loss": 0.3298,
|
| 888 |
+
"step": 1130
|
| 889 |
+
},
|
| 890 |
+
{
|
| 891 |
+
"epoch": 0.8145766345123259,
|
| 892 |
+
"grad_norm": 0.8693751096725464,
|
| 893 |
+
"learning_rate": 9.129854684050481e-05,
|
| 894 |
+
"loss": 0.279,
|
| 895 |
+
"step": 1140
|
| 896 |
+
},
|
| 897 |
+
{
|
| 898 |
+
"epoch": 0.8217220435869954,
|
| 899 |
+
"grad_norm": 1.1013332605361938,
|
| 900 |
+
"learning_rate": 9.10626814915343e-05,
|
| 901 |
+
"loss": 0.3195,
|
| 902 |
+
"step": 1150
|
| 903 |
+
},
|
| 904 |
+
{
|
| 905 |
+
"epoch": 0.8288674526616648,
|
| 906 |
+
"grad_norm": 1.2278695106506348,
|
| 907 |
+
"learning_rate": 9.082397527768092e-05,
|
| 908 |
+
"loss": 0.3027,
|
| 909 |
+
"step": 1160
|
| 910 |
+
},
|
| 911 |
+
{
|
| 912 |
+
"epoch": 0.8360128617363344,
|
| 913 |
+
"grad_norm": 2.173530101776123,
|
| 914 |
+
"learning_rate": 9.058244471350428e-05,
|
| 915 |
+
"loss": 0.2238,
|
| 916 |
+
"step": 1170
|
| 917 |
+
},
|
| 918 |
+
{
|
| 919 |
+
"epoch": 0.8431582708110039,
|
| 920 |
+
"grad_norm": 1.125986933708191,
|
| 921 |
+
"learning_rate": 9.033810650896274e-05,
|
| 922 |
+
"loss": 0.2399,
|
| 923 |
+
"step": 1180
|
| 924 |
+
},
|
| 925 |
+
{
|
| 926 |
+
"epoch": 0.8503036798856735,
|
| 927 |
+
"grad_norm": 0.6611151099205017,
|
| 928 |
+
"learning_rate": 9.009097756825737e-05,
|
| 929 |
+
"loss": 0.2736,
|
| 930 |
+
"step": 1190
|
| 931 |
+
},
|
| 932 |
+
{
|
| 933 |
+
"epoch": 0.857449088960343,
|
| 934 |
+
"grad_norm": 1.9068485498428345,
|
| 935 |
+
"learning_rate": 8.98410749886625e-05,
|
| 936 |
+
"loss": 0.2949,
|
| 937 |
+
"step": 1200
|
| 938 |
+
},
|
| 939 |
+
{
|
| 940 |
+
"epoch": 0.857449088960343,
|
| 941 |
+
"eval_news_finetune_val_loss": 0.31006094813346863,
|
| 942 |
+
"eval_news_finetune_val_runtime": 1002.7866,
|
| 943 |
+
"eval_news_finetune_val_samples_per_second": 1.396,
|
| 944 |
+
"eval_news_finetune_val_steps_per_second": 1.396,
|
| 945 |
+
"step": 1200
|
| 946 |
+
},
|
| 947 |
+
{
|
| 948 |
+
"epoch": 0.8645944980350125,
|
| 949 |
+
"grad_norm": 1.192031979560852,
|
| 950 |
+
"learning_rate": 8.958841605934278e-05,
|
| 951 |
+
"loss": 0.3657,
|
| 952 |
+
"step": 1210
|
| 953 |
+
},
|
| 954 |
+
{
|
| 955 |
+
"epoch": 0.8717399071096821,
|
| 956 |
+
"grad_norm": 1.2596725225448608,
|
| 957 |
+
"learning_rate": 8.933301826015715e-05,
|
| 958 |
+
"loss": 0.3068,
|
| 959 |
+
"step": 1220
|
| 960 |
+
},
|
| 961 |
+
{
|
| 962 |
+
"epoch": 0.8788853161843515,
|
| 963 |
+
"grad_norm": 1.4713683128356934,
|
| 964 |
+
"learning_rate": 8.907489926044945e-05,
|
| 965 |
+
"loss": 0.3122,
|
| 966 |
+
"step": 1230
|
| 967 |
+
},
|
| 968 |
+
{
|
| 969 |
+
"epoch": 0.886030725259021,
|
| 970 |
+
"grad_norm": 1.3583886623382568,
|
| 971 |
+
"learning_rate": 8.881407691782608e-05,
|
| 972 |
+
"loss": 0.2989,
|
| 973 |
+
"step": 1240
|
| 974 |
+
},
|
| 975 |
+
{
|
| 976 |
+
"epoch": 0.8931761343336906,
|
| 977 |
+
"grad_norm": 0.9863426089286804,
|
| 978 |
+
"learning_rate": 8.855056927692037e-05,
|
| 979 |
+
"loss": 0.2549,
|
| 980 |
+
"step": 1250
|
| 981 |
+
},
|
| 982 |
+
{
|
| 983 |
+
"epoch": 0.9003215434083601,
|
| 984 |
+
"grad_norm": 1.0579396486282349,
|
| 985 |
+
"learning_rate": 8.828439456814442e-05,
|
| 986 |
+
"loss": 0.2809,
|
| 987 |
+
"step": 1260
|
| 988 |
+
},
|
| 989 |
+
{
|
| 990 |
+
"epoch": 0.9074669524830297,
|
| 991 |
+
"grad_norm": 2.847482681274414,
|
| 992 |
+
"learning_rate": 8.801557120642766e-05,
|
| 993 |
+
"loss": 0.2933,
|
| 994 |
+
"step": 1270
|
| 995 |
+
},
|
| 996 |
+
{
|
| 997 |
+
"epoch": 0.9146123615576992,
|
| 998 |
+
"grad_norm": 0.8942415118217468,
|
| 999 |
+
"learning_rate": 8.774411778994295e-05,
|
| 1000 |
+
"loss": 0.2866,
|
| 1001 |
+
"step": 1280
|
| 1002 |
+
},
|
| 1003 |
+
{
|
| 1004 |
+
"epoch": 0.9217577706323687,
|
| 1005 |
+
"grad_norm": 1.297845721244812,
|
| 1006 |
+
"learning_rate": 8.747005309881984e-05,
|
| 1007 |
+
"loss": 0.2939,
|
| 1008 |
+
"step": 1290
|
| 1009 |
+
},
|
| 1010 |
+
{
|
| 1011 |
+
"epoch": 0.9289031797070382,
|
| 1012 |
+
"grad_norm": 1.2745181322097778,
|
| 1013 |
+
"learning_rate": 8.719339609384531e-05,
|
| 1014 |
+
"loss": 0.3018,
|
| 1015 |
+
"step": 1300
|
| 1016 |
+
},
|
| 1017 |
+
{
|
| 1018 |
+
"epoch": 0.9289031797070382,
|
| 1019 |
+
"eval_news_finetune_val_loss": 0.29822030663490295,
|
| 1020 |
+
"eval_news_finetune_val_runtime": 1002.5672,
|
| 1021 |
+
"eval_news_finetune_val_samples_per_second": 1.396,
|
| 1022 |
+
"eval_news_finetune_val_steps_per_second": 1.396,
|
| 1023 |
+
"step": 1300
|
| 1024 |
+
},
|
| 1025 |
+
{
|
| 1026 |
+
"epoch": 0.9360485887817077,
|
| 1027 |
+
"grad_norm": 1.3898978233337402,
|
| 1028 |
+
"learning_rate": 8.691416591515198e-05,
|
| 1029 |
+
"loss": 0.295,
|
| 1030 |
+
"step": 1310
|
| 1031 |
+
},
|
| 1032 |
+
{
|
| 1033 |
+
"epoch": 0.9431939978563773,
|
| 1034 |
+
"grad_norm": 1.1516591310501099,
|
| 1035 |
+
"learning_rate": 8.663238188089398e-05,
|
| 1036 |
+
"loss": 0.209,
|
| 1037 |
+
"step": 1320
|
| 1038 |
+
},
|
| 1039 |
+
{
|
| 1040 |
+
"epoch": 0.9503394069310468,
|
| 1041 |
+
"grad_norm": 0.9356768131256104,
|
| 1042 |
+
"learning_rate": 8.634806348591036e-05,
|
| 1043 |
+
"loss": 0.2904,
|
| 1044 |
+
"step": 1330
|
| 1045 |
+
},
|
| 1046 |
+
{
|
| 1047 |
+
"epoch": 0.9574848160057163,
|
| 1048 |
+
"grad_norm": 1.884950876235962,
|
| 1049 |
+
"learning_rate": 8.606123040037643e-05,
|
| 1050 |
+
"loss": 0.2607,
|
| 1051 |
+
"step": 1340
|
| 1052 |
+
},
|
| 1053 |
+
{
|
| 1054 |
+
"epoch": 0.9646302250803859,
|
| 1055 |
+
"grad_norm": 1.2719082832336426,
|
| 1056 |
+
"learning_rate": 8.577190246844291e-05,
|
| 1057 |
+
"loss": 0.3279,
|
| 1058 |
+
"step": 1350
|
| 1059 |
+
},
|
| 1060 |
+
{
|
| 1061 |
+
"epoch": 0.9717756341550554,
|
| 1062 |
+
"grad_norm": 0.935297429561615,
|
| 1063 |
+
"learning_rate": 8.548009970686302e-05,
|
| 1064 |
+
"loss": 0.3011,
|
| 1065 |
+
"step": 1360
|
| 1066 |
+
},
|
| 1067 |
+
{
|
| 1068 |
+
"epoch": 0.978921043229725,
|
| 1069 |
+
"grad_norm": 1.6732884645462036,
|
| 1070 |
+
"learning_rate": 8.51858423036076e-05,
|
| 1071 |
+
"loss": 0.2379,
|
| 1072 |
+
"step": 1370
|
| 1073 |
+
},
|
| 1074 |
+
{
|
| 1075 |
+
"epoch": 0.9860664523043944,
|
| 1076 |
+
"grad_norm": 0.6651692390441895,
|
| 1077 |
+
"learning_rate": 8.488915061646856e-05,
|
| 1078 |
+
"loss": 0.2599,
|
| 1079 |
+
"step": 1380
|
| 1080 |
+
},
|
| 1081 |
+
{
|
| 1082 |
+
"epoch": 0.9932118613790639,
|
| 1083 |
+
"grad_norm": 1.121752381324768,
|
| 1084 |
+
"learning_rate": 8.459004517165032e-05,
|
| 1085 |
+
"loss": 0.2265,
|
| 1086 |
+
"step": 1390
|
| 1087 |
+
},
|
| 1088 |
+
{
|
| 1089 |
+
"epoch": 1.0,
|
| 1090 |
+
"grad_norm": 0.5099928379058838,
|
| 1091 |
+
"learning_rate": 8.428854666234978e-05,
|
| 1092 |
+
"loss": 0.3301,
|
| 1093 |
+
"step": 1400
|
| 1094 |
+
},
|
| 1095 |
+
{
|
| 1096 |
+
"epoch": 1.0,
|
| 1097 |
+
"eval_news_finetune_val_loss": 0.28762951493263245,
|
| 1098 |
+
"eval_news_finetune_val_runtime": 1002.7793,
|
| 1099 |
+
"eval_news_finetune_val_samples_per_second": 1.396,
|
| 1100 |
+
"eval_news_finetune_val_steps_per_second": 1.396,
|
| 1101 |
+
"step": 1400
|
| 1102 |
+
},
|
| 1103 |
+
{
|
| 1104 |
+
"epoch": 1.0071454090746694,
|
| 1105 |
+
"grad_norm": 0.9986103177070618,
|
| 1106 |
+
"learning_rate": 8.398467594732478e-05,
|
| 1107 |
+
"loss": 0.2021,
|
| 1108 |
+
"step": 1410
|
| 1109 |
+
},
|
| 1110 |
+
{
|
| 1111 |
+
"epoch": 1.014290818149339,
|
| 1112 |
+
"grad_norm": 1.2675282955169678,
|
| 1113 |
+
"learning_rate": 8.367845404945084e-05,
|
| 1114 |
+
"loss": 0.2228,
|
| 1115 |
+
"step": 1420
|
| 1116 |
+
},
|
| 1117 |
+
{
|
| 1118 |
+
"epoch": 1.0214362272240085,
|
| 1119 |
+
"grad_norm": 0.8156709671020508,
|
| 1120 |
+
"learning_rate": 8.336990215426688e-05,
|
| 1121 |
+
"loss": 0.1947,
|
| 1122 |
+
"step": 1430
|
| 1123 |
+
},
|
| 1124 |
+
{
|
| 1125 |
+
"epoch": 1.0285816362986782,
|
| 1126 |
+
"grad_norm": 0.5374387502670288,
|
| 1127 |
+
"learning_rate": 8.305904160850941e-05,
|
| 1128 |
+
"loss": 0.2344,
|
| 1129 |
+
"step": 1440
|
| 1130 |
+
},
|
| 1131 |
+
{
|
| 1132 |
+
"epoch": 1.0357270453733476,
|
| 1133 |
+
"grad_norm": 0.6672261357307434,
|
| 1134 |
+
"learning_rate": 8.274589391863583e-05,
|
| 1135 |
+
"loss": 0.1919,
|
| 1136 |
+
"step": 1450
|
| 1137 |
+
},
|
| 1138 |
+
{
|
| 1139 |
+
"epoch": 1.0428724544480172,
|
| 1140 |
+
"grad_norm": 0.9803467988967896,
|
| 1141 |
+
"learning_rate": 8.243048074933634e-05,
|
| 1142 |
+
"loss": 0.2218,
|
| 1143 |
+
"step": 1460
|
| 1144 |
+
},
|
| 1145 |
+
{
|
| 1146 |
+
"epoch": 1.0500178635226867,
|
| 1147 |
+
"grad_norm": 1.482840657234192,
|
| 1148 |
+
"learning_rate": 8.21128239220353e-05,
|
| 1149 |
+
"loss": 0.2556,
|
| 1150 |
+
"step": 1470
|
| 1151 |
+
},
|
| 1152 |
+
{
|
| 1153 |
+
"epoch": 1.057163272597356,
|
| 1154 |
+
"grad_norm": 1.0589625835418701,
|
| 1155 |
+
"learning_rate": 8.179294541338135e-05,
|
| 1156 |
+
"loss": 0.2052,
|
| 1157 |
+
"step": 1480
|
| 1158 |
+
},
|
| 1159 |
+
{
|
| 1160 |
+
"epoch": 1.0643086816720257,
|
| 1161 |
+
"grad_norm": 0.8332052230834961,
|
| 1162 |
+
"learning_rate": 8.147086735372716e-05,
|
| 1163 |
+
"loss": 0.2386,
|
| 1164 |
+
"step": 1490
|
| 1165 |
+
},
|
| 1166 |
+
{
|
| 1167 |
+
"epoch": 1.0714540907466952,
|
| 1168 |
+
"grad_norm": 0.6018723845481873,
|
| 1169 |
+
"learning_rate": 8.114661202559828e-05,
|
| 1170 |
+
"loss": 0.1426,
|
| 1171 |
+
"step": 1500
|
| 1172 |
+
},
|
| 1173 |
+
{
|
| 1174 |
+
"epoch": 1.0714540907466952,
|
| 1175 |
+
"eval_news_finetune_val_loss": 0.30121028423309326,
|
| 1176 |
+
"eval_news_finetune_val_runtime": 1002.7457,
|
| 1177 |
+
"eval_news_finetune_val_samples_per_second": 1.396,
|
| 1178 |
+
"eval_news_finetune_val_steps_per_second": 1.396,
|
| 1179 |
+
"step": 1500
|
| 1180 |
}
|
| 1181 |
],
|
| 1182 |
"logging_steps": 10,
|
|
|
|
| 1196 |
"attributes": {}
|
| 1197 |
}
|
| 1198 |
},
|
| 1199 |
+
"total_flos": 4.14970883106816e+16,
|
| 1200 |
"train_batch_size": 1,
|
| 1201 |
"trial_name": null,
|
| 1202 |
"trial_params": null
|