Training in progress, step 1200, 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 35237104
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:6a8ea02f63b71e9ecc281715285720652652fca4f282320a62cc716249a6ef2c
|
3 |
size 35237104
|
last-checkpoint/optimizer.pt
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 18810356
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:8dae95bb49116e81bed3c68d86278f66e9b0452f468bf29d48fecf5e625ba7e0
|
3 |
size 18810356
|
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 sha256:2ea4a928811810f84d0f8b5936c44f2709f0f2cadb34cf4fa9a714f98b7ed5c0
|
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:b30780423ff45c94e8ad2d6a438363a868ec26f078372884947da902f24979de
|
3 |
size 1064
|
last-checkpoint/trainer_state.json
CHANGED
@@ -1,9 +1,9 @@
|
|
1 |
{
|
2 |
"best_metric": 1.8883090019226074,
|
3 |
"best_model_checkpoint": "miner_id_24/checkpoint-1000",
|
4 |
-
"epoch": 1.
|
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,
|
@@ -7803,6 +7803,714 @@
|
|
7803 |
"eval_samples_per_second": 58.854,
|
7804 |
"eval_steps_per_second": 14.724,
|
7805 |
"step": 1100
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
7806 |
}
|
7807 |
],
|
7808 |
"logging_steps": 1,
|
@@ -7817,7 +8525,7 @@
|
|
7817 |
"early_stopping_threshold": 0.0
|
7818 |
},
|
7819 |
"attributes": {
|
7820 |
-
"early_stopping_patience_counter":
|
7821 |
}
|
7822 |
},
|
7823 |
"TrainerControl": {
|
@@ -7826,12 +8534,12 @@
|
|
7826 |
"should_evaluate": false,
|
7827 |
"should_log": false,
|
7828 |
"should_save": true,
|
7829 |
-
"should_training_stop":
|
7830 |
},
|
7831 |
"attributes": {}
|
7832 |
}
|
7833 |
},
|
7834 |
-
"total_flos":
|
7835 |
"train_batch_size": 4,
|
7836 |
"trial_name": null,
|
7837 |
"trial_params": null
|
|
|
1 |
{
|
2 |
"best_metric": 1.8883090019226074,
|
3 |
"best_model_checkpoint": "miner_id_24/checkpoint-1000",
|
4 |
+
"epoch": 1.1774806819575616,
|
5 |
"eval_steps": 100,
|
6 |
+
"global_step": 1200,
|
7 |
"is_hyper_param_search": false,
|
8 |
"is_local_process_zero": true,
|
9 |
"is_world_process_zero": true,
|
|
|
7803 |
"eval_samples_per_second": 58.854,
|
7804 |
"eval_steps_per_second": 14.724,
|
7805 |
"step": 1100
|
7806 |
+
},
|
7807 |
+
{
|
7808 |
+
"epoch": 1.0803385256960627,
|
7809 |
+
"grad_norm": 2.066479444503784,
|
7810 |
+
"learning_rate": 8.818283612384952e-05,
|
7811 |
+
"loss": 1.6725,
|
7812 |
+
"step": 1101
|
7813 |
+
},
|
7814 |
+
{
|
7815 |
+
"epoch": 1.081319759597694,
|
7816 |
+
"grad_norm": 2.083176851272583,
|
7817 |
+
"learning_rate": 8.80291007154402e-05,
|
7818 |
+
"loss": 1.5404,
|
7819 |
+
"step": 1102
|
7820 |
+
},
|
7821 |
+
{
|
7822 |
+
"epoch": 1.0823009934993255,
|
7823 |
+
"grad_norm": 2.1391541957855225,
|
7824 |
+
"learning_rate": 8.787539400573867e-05,
|
7825 |
+
"loss": 1.6664,
|
7826 |
+
"step": 1103
|
7827 |
+
},
|
7828 |
+
{
|
7829 |
+
"epoch": 1.0832822274009568,
|
7830 |
+
"grad_norm": 1.8939257860183716,
|
7831 |
+
"learning_rate": 8.772171636323714e-05,
|
7832 |
+
"loss": 1.6494,
|
7833 |
+
"step": 1104
|
7834 |
+
},
|
7835 |
+
{
|
7836 |
+
"epoch": 1.084263461302588,
|
7837 |
+
"grad_norm": 2.128265619277954,
|
7838 |
+
"learning_rate": 8.75680681563582e-05,
|
7839 |
+
"loss": 1.7697,
|
7840 |
+
"step": 1105
|
7841 |
+
},
|
7842 |
+
{
|
7843 |
+
"epoch": 1.0852446952042194,
|
7844 |
+
"grad_norm": 1.9691377878189087,
|
7845 |
+
"learning_rate": 8.74144497534539e-05,
|
7846 |
+
"loss": 1.6618,
|
7847 |
+
"step": 1106
|
7848 |
+
},
|
7849 |
+
{
|
7850 |
+
"epoch": 1.0862259291058507,
|
7851 |
+
"grad_norm": 2.111213445663452,
|
7852 |
+
"learning_rate": 8.726086152280483e-05,
|
7853 |
+
"loss": 1.6478,
|
7854 |
+
"step": 1107
|
7855 |
+
},
|
7856 |
+
{
|
7857 |
+
"epoch": 1.087207163007482,
|
7858 |
+
"grad_norm": 2.051668882369995,
|
7859 |
+
"learning_rate": 8.710730383261916e-05,
|
7860 |
+
"loss": 1.7054,
|
7861 |
+
"step": 1108
|
7862 |
+
},
|
7863 |
+
{
|
7864 |
+
"epoch": 1.0881883969091133,
|
7865 |
+
"grad_norm": 1.7977656126022339,
|
7866 |
+
"learning_rate": 8.695377705103199e-05,
|
7867 |
+
"loss": 1.4778,
|
7868 |
+
"step": 1109
|
7869 |
+
},
|
7870 |
+
{
|
7871 |
+
"epoch": 1.0891696308107446,
|
7872 |
+
"grad_norm": 2.13419246673584,
|
7873 |
+
"learning_rate": 8.680028154610419e-05,
|
7874 |
+
"loss": 1.8192,
|
7875 |
+
"step": 1110
|
7876 |
+
},
|
7877 |
+
{
|
7878 |
+
"epoch": 1.0901508647123759,
|
7879 |
+
"grad_norm": 1.9920395612716675,
|
7880 |
+
"learning_rate": 8.66468176858217e-05,
|
7881 |
+
"loss": 1.6159,
|
7882 |
+
"step": 1111
|
7883 |
+
},
|
7884 |
+
{
|
7885 |
+
"epoch": 1.0911320986140072,
|
7886 |
+
"grad_norm": 1.8859739303588867,
|
7887 |
+
"learning_rate": 8.649338583809466e-05,
|
7888 |
+
"loss": 1.4988,
|
7889 |
+
"step": 1112
|
7890 |
+
},
|
7891 |
+
{
|
7892 |
+
"epoch": 1.0921133325156385,
|
7893 |
+
"grad_norm": 1.9767767190933228,
|
7894 |
+
"learning_rate": 8.633998637075634e-05,
|
7895 |
+
"loss": 1.6233,
|
7896 |
+
"step": 1113
|
7897 |
+
},
|
7898 |
+
{
|
7899 |
+
"epoch": 1.0930945664172698,
|
7900 |
+
"grad_norm": 1.9266538619995117,
|
7901 |
+
"learning_rate": 8.618661965156243e-05,
|
7902 |
+
"loss": 1.6353,
|
7903 |
+
"step": 1114
|
7904 |
+
},
|
7905 |
+
{
|
7906 |
+
"epoch": 1.094075800318901,
|
7907 |
+
"grad_norm": 2.1085169315338135,
|
7908 |
+
"learning_rate": 8.60332860481902e-05,
|
7909 |
+
"loss": 1.5874,
|
7910 |
+
"step": 1115
|
7911 |
+
},
|
7912 |
+
{
|
7913 |
+
"epoch": 1.0950570342205324,
|
7914 |
+
"grad_norm": 1.8933839797973633,
|
7915 |
+
"learning_rate": 8.587998592823738e-05,
|
7916 |
+
"loss": 1.574,
|
7917 |
+
"step": 1116
|
7918 |
+
},
|
7919 |
+
{
|
7920 |
+
"epoch": 1.0960382681221637,
|
7921 |
+
"grad_norm": 2.0371503829956055,
|
7922 |
+
"learning_rate": 8.572671965922148e-05,
|
7923 |
+
"loss": 1.7796,
|
7924 |
+
"step": 1117
|
7925 |
+
},
|
7926 |
+
{
|
7927 |
+
"epoch": 1.097019502023795,
|
7928 |
+
"grad_norm": 1.9261376857757568,
|
7929 |
+
"learning_rate": 8.557348760857899e-05,
|
7930 |
+
"loss": 1.4353,
|
7931 |
+
"step": 1118
|
7932 |
+
},
|
7933 |
+
{
|
7934 |
+
"epoch": 1.0980007359254262,
|
7935 |
+
"grad_norm": 2.089020252227783,
|
7936 |
+
"learning_rate": 8.542029014366419e-05,
|
7937 |
+
"loss": 1.7129,
|
7938 |
+
"step": 1119
|
7939 |
+
},
|
7940 |
+
{
|
7941 |
+
"epoch": 1.0989819698270575,
|
7942 |
+
"grad_norm": 1.8985182046890259,
|
7943 |
+
"learning_rate": 8.526712763174844e-05,
|
7944 |
+
"loss": 1.5255,
|
7945 |
+
"step": 1120
|
7946 |
+
},
|
7947 |
+
{
|
7948 |
+
"epoch": 1.0999632037286888,
|
7949 |
+
"grad_norm": 2.0458781719207764,
|
7950 |
+
"learning_rate": 8.511400044001948e-05,
|
7951 |
+
"loss": 1.742,
|
7952 |
+
"step": 1121
|
7953 |
+
},
|
7954 |
+
{
|
7955 |
+
"epoch": 1.1009444376303201,
|
7956 |
+
"grad_norm": 1.925270438194275,
|
7957 |
+
"learning_rate": 8.496090893558019e-05,
|
7958 |
+
"loss": 1.5204,
|
7959 |
+
"step": 1122
|
7960 |
+
},
|
7961 |
+
{
|
7962 |
+
"epoch": 1.1019256715319514,
|
7963 |
+
"grad_norm": 1.7764109373092651,
|
7964 |
+
"learning_rate": 8.480785348544802e-05,
|
7965 |
+
"loss": 1.4883,
|
7966 |
+
"step": 1123
|
7967 |
+
},
|
7968 |
+
{
|
7969 |
+
"epoch": 1.1029069054335827,
|
7970 |
+
"grad_norm": 2.034646511077881,
|
7971 |
+
"learning_rate": 8.465483445655394e-05,
|
7972 |
+
"loss": 1.7298,
|
7973 |
+
"step": 1124
|
7974 |
+
},
|
7975 |
+
{
|
7976 |
+
"epoch": 1.103888139335214,
|
7977 |
+
"grad_norm": 1.8433154821395874,
|
7978 |
+
"learning_rate": 8.450185221574152e-05,
|
7979 |
+
"loss": 1.5534,
|
7980 |
+
"step": 1125
|
7981 |
+
},
|
7982 |
+
{
|
7983 |
+
"epoch": 1.1048693732368453,
|
7984 |
+
"grad_norm": 2.088686943054199,
|
7985 |
+
"learning_rate": 8.434890712976632e-05,
|
7986 |
+
"loss": 1.5524,
|
7987 |
+
"step": 1126
|
7988 |
+
},
|
7989 |
+
{
|
7990 |
+
"epoch": 1.1058506071384766,
|
7991 |
+
"grad_norm": 2.0298330783843994,
|
7992 |
+
"learning_rate": 8.419599956529466e-05,
|
7993 |
+
"loss": 1.7555,
|
7994 |
+
"step": 1127
|
7995 |
+
},
|
7996 |
+
{
|
7997 |
+
"epoch": 1.106831841040108,
|
7998 |
+
"grad_norm": 1.9726227521896362,
|
7999 |
+
"learning_rate": 8.404312988890299e-05,
|
8000 |
+
"loss": 1.6304,
|
8001 |
+
"step": 1128
|
8002 |
+
},
|
8003 |
+
{
|
8004 |
+
"epoch": 1.1078130749417392,
|
8005 |
+
"grad_norm": 1.998617172241211,
|
8006 |
+
"learning_rate": 8.389029846707688e-05,
|
8007 |
+
"loss": 1.6792,
|
8008 |
+
"step": 1129
|
8009 |
+
},
|
8010 |
+
{
|
8011 |
+
"epoch": 1.1087943088433705,
|
8012 |
+
"grad_norm": 1.983628749847412,
|
8013 |
+
"learning_rate": 8.373750566621027e-05,
|
8014 |
+
"loss": 1.5545,
|
8015 |
+
"step": 1130
|
8016 |
+
},
|
8017 |
+
{
|
8018 |
+
"epoch": 1.1097755427450018,
|
8019 |
+
"grad_norm": 1.913836121559143,
|
8020 |
+
"learning_rate": 8.358475185260438e-05,
|
8021 |
+
"loss": 1.5676,
|
8022 |
+
"step": 1131
|
8023 |
+
},
|
8024 |
+
{
|
8025 |
+
"epoch": 1.110756776646633,
|
8026 |
+
"grad_norm": 1.8630105257034302,
|
8027 |
+
"learning_rate": 8.343203739246712e-05,
|
8028 |
+
"loss": 1.4564,
|
8029 |
+
"step": 1132
|
8030 |
+
},
|
8031 |
+
{
|
8032 |
+
"epoch": 1.1117380105482644,
|
8033 |
+
"grad_norm": 2.2467024326324463,
|
8034 |
+
"learning_rate": 8.32793626519119e-05,
|
8035 |
+
"loss": 1.7565,
|
8036 |
+
"step": 1133
|
8037 |
+
},
|
8038 |
+
{
|
8039 |
+
"epoch": 1.1127192444498957,
|
8040 |
+
"grad_norm": 1.967323899269104,
|
8041 |
+
"learning_rate": 8.312672799695702e-05,
|
8042 |
+
"loss": 1.5484,
|
8043 |
+
"step": 1134
|
8044 |
+
},
|
8045 |
+
{
|
8046 |
+
"epoch": 1.113700478351527,
|
8047 |
+
"grad_norm": 1.8815773725509644,
|
8048 |
+
"learning_rate": 8.297413379352469e-05,
|
8049 |
+
"loss": 1.5544,
|
8050 |
+
"step": 1135
|
8051 |
+
},
|
8052 |
+
{
|
8053 |
+
"epoch": 1.1146817122531583,
|
8054 |
+
"grad_norm": 2.087390422821045,
|
8055 |
+
"learning_rate": 8.282158040744003e-05,
|
8056 |
+
"loss": 1.5492,
|
8057 |
+
"step": 1136
|
8058 |
+
},
|
8059 |
+
{
|
8060 |
+
"epoch": 1.1156629461547896,
|
8061 |
+
"grad_norm": 2.0206923484802246,
|
8062 |
+
"learning_rate": 8.266906820443036e-05,
|
8063 |
+
"loss": 1.5578,
|
8064 |
+
"step": 1137
|
8065 |
+
},
|
8066 |
+
{
|
8067 |
+
"epoch": 1.1166441800564209,
|
8068 |
+
"grad_norm": 2.1264116764068604,
|
8069 |
+
"learning_rate": 8.251659755012435e-05,
|
8070 |
+
"loss": 1.7562,
|
8071 |
+
"step": 1138
|
8072 |
+
},
|
8073 |
+
{
|
8074 |
+
"epoch": 1.1176254139580521,
|
8075 |
+
"grad_norm": 2.1084797382354736,
|
8076 |
+
"learning_rate": 8.236416881005093e-05,
|
8077 |
+
"loss": 1.8296,
|
8078 |
+
"step": 1139
|
8079 |
+
},
|
8080 |
+
{
|
8081 |
+
"epoch": 1.1186066478596834,
|
8082 |
+
"grad_norm": 2.130791425704956,
|
8083 |
+
"learning_rate": 8.22117823496386e-05,
|
8084 |
+
"loss": 1.6879,
|
8085 |
+
"step": 1140
|
8086 |
+
},
|
8087 |
+
{
|
8088 |
+
"epoch": 1.1195878817613147,
|
8089 |
+
"grad_norm": 1.9630494117736816,
|
8090 |
+
"learning_rate": 8.205943853421457e-05,
|
8091 |
+
"loss": 1.6107,
|
8092 |
+
"step": 1141
|
8093 |
+
},
|
8094 |
+
{
|
8095 |
+
"epoch": 1.1205691156629463,
|
8096 |
+
"grad_norm": 2.180497407913208,
|
8097 |
+
"learning_rate": 8.190713772900374e-05,
|
8098 |
+
"loss": 1.7211,
|
8099 |
+
"step": 1142
|
8100 |
+
},
|
8101 |
+
{
|
8102 |
+
"epoch": 1.1215503495645776,
|
8103 |
+
"grad_norm": 1.846765160560608,
|
8104 |
+
"learning_rate": 8.175488029912783e-05,
|
8105 |
+
"loss": 1.5607,
|
8106 |
+
"step": 1143
|
8107 |
+
},
|
8108 |
+
{
|
8109 |
+
"epoch": 1.1225315834662088,
|
8110 |
+
"grad_norm": 2.244673252105713,
|
8111 |
+
"learning_rate": 8.160266660960472e-05,
|
8112 |
+
"loss": 1.8069,
|
8113 |
+
"step": 1144
|
8114 |
+
},
|
8115 |
+
{
|
8116 |
+
"epoch": 1.1235128173678401,
|
8117 |
+
"grad_norm": 2.002246141433716,
|
8118 |
+
"learning_rate": 8.14504970253474e-05,
|
8119 |
+
"loss": 1.6429,
|
8120 |
+
"step": 1145
|
8121 |
+
},
|
8122 |
+
{
|
8123 |
+
"epoch": 1.1244940512694714,
|
8124 |
+
"grad_norm": 1.974992275238037,
|
8125 |
+
"learning_rate": 8.129837191116298e-05,
|
8126 |
+
"loss": 1.5133,
|
8127 |
+
"step": 1146
|
8128 |
+
},
|
8129 |
+
{
|
8130 |
+
"epoch": 1.1254752851711027,
|
8131 |
+
"grad_norm": 1.9683125019073486,
|
8132 |
+
"learning_rate": 8.114629163175215e-05,
|
8133 |
+
"loss": 1.5906,
|
8134 |
+
"step": 1147
|
8135 |
+
},
|
8136 |
+
{
|
8137 |
+
"epoch": 1.126456519072734,
|
8138 |
+
"grad_norm": 1.8641992807388306,
|
8139 |
+
"learning_rate": 8.099425655170801e-05,
|
8140 |
+
"loss": 1.4253,
|
8141 |
+
"step": 1148
|
8142 |
+
},
|
8143 |
+
{
|
8144 |
+
"epoch": 1.1274377529743653,
|
8145 |
+
"grad_norm": 1.923235535621643,
|
8146 |
+
"learning_rate": 8.084226703551528e-05,
|
8147 |
+
"loss": 1.591,
|
8148 |
+
"step": 1149
|
8149 |
+
},
|
8150 |
+
{
|
8151 |
+
"epoch": 1.1284189868759966,
|
8152 |
+
"grad_norm": 1.962015151977539,
|
8153 |
+
"learning_rate": 8.06903234475495e-05,
|
8154 |
+
"loss": 1.573,
|
8155 |
+
"step": 1150
|
8156 |
+
},
|
8157 |
+
{
|
8158 |
+
"epoch": 1.129400220777628,
|
8159 |
+
"grad_norm": 1.9990873336791992,
|
8160 |
+
"learning_rate": 8.053842615207615e-05,
|
8161 |
+
"loss": 1.7404,
|
8162 |
+
"step": 1151
|
8163 |
+
},
|
8164 |
+
{
|
8165 |
+
"epoch": 1.1303814546792592,
|
8166 |
+
"grad_norm": 2.225425958633423,
|
8167 |
+
"learning_rate": 8.038657551324955e-05,
|
8168 |
+
"loss": 1.8477,
|
8169 |
+
"step": 1152
|
8170 |
+
},
|
8171 |
+
{
|
8172 |
+
"epoch": 1.1313626885808905,
|
8173 |
+
"grad_norm": 1.7779438495635986,
|
8174 |
+
"learning_rate": 8.02347718951124e-05,
|
8175 |
+
"loss": 1.5567,
|
8176 |
+
"step": 1153
|
8177 |
+
},
|
8178 |
+
{
|
8179 |
+
"epoch": 1.1323439224825218,
|
8180 |
+
"grad_norm": 2.0244076251983643,
|
8181 |
+
"learning_rate": 8.008301566159447e-05,
|
8182 |
+
"loss": 1.6891,
|
8183 |
+
"step": 1154
|
8184 |
+
},
|
8185 |
+
{
|
8186 |
+
"epoch": 1.133325156384153,
|
8187 |
+
"grad_norm": 1.987186074256897,
|
8188 |
+
"learning_rate": 7.993130717651207e-05,
|
8189 |
+
"loss": 1.7452,
|
8190 |
+
"step": 1155
|
8191 |
+
},
|
8192 |
+
{
|
8193 |
+
"epoch": 1.1343063902857844,
|
8194 |
+
"grad_norm": 2.076939344406128,
|
8195 |
+
"learning_rate": 7.977964680356696e-05,
|
8196 |
+
"loss": 1.7669,
|
8197 |
+
"step": 1156
|
8198 |
+
},
|
8199 |
+
{
|
8200 |
+
"epoch": 1.1352876241874157,
|
8201 |
+
"grad_norm": 1.9476886987686157,
|
8202 |
+
"learning_rate": 7.962803490634563e-05,
|
8203 |
+
"loss": 1.6541,
|
8204 |
+
"step": 1157
|
8205 |
+
},
|
8206 |
+
{
|
8207 |
+
"epoch": 1.136268858089047,
|
8208 |
+
"grad_norm": 2.047985792160034,
|
8209 |
+
"learning_rate": 7.947647184831824e-05,
|
8210 |
+
"loss": 1.486,
|
8211 |
+
"step": 1158
|
8212 |
+
},
|
8213 |
+
{
|
8214 |
+
"epoch": 1.1372500919906783,
|
8215 |
+
"grad_norm": 2.015349864959717,
|
8216 |
+
"learning_rate": 7.932495799283801e-05,
|
8217 |
+
"loss": 1.5786,
|
8218 |
+
"step": 1159
|
8219 |
+
},
|
8220 |
+
{
|
8221 |
+
"epoch": 1.1382313258923096,
|
8222 |
+
"grad_norm": 1.8462163209915161,
|
8223 |
+
"learning_rate": 7.917349370314007e-05,
|
8224 |
+
"loss": 1.483,
|
8225 |
+
"step": 1160
|
8226 |
+
},
|
8227 |
+
{
|
8228 |
+
"epoch": 1.1392125597939409,
|
8229 |
+
"grad_norm": 1.9125447273254395,
|
8230 |
+
"learning_rate": 7.902207934234078e-05,
|
8231 |
+
"loss": 1.6106,
|
8232 |
+
"step": 1161
|
8233 |
+
},
|
8234 |
+
{
|
8235 |
+
"epoch": 1.1401937936955722,
|
8236 |
+
"grad_norm": 1.9936769008636475,
|
8237 |
+
"learning_rate": 7.887071527343687e-05,
|
8238 |
+
"loss": 1.748,
|
8239 |
+
"step": 1162
|
8240 |
+
},
|
8241 |
+
{
|
8242 |
+
"epoch": 1.1411750275972035,
|
8243 |
+
"grad_norm": 2.2277896404266357,
|
8244 |
+
"learning_rate": 7.871940185930438e-05,
|
8245 |
+
"loss": 1.9264,
|
8246 |
+
"step": 1163
|
8247 |
+
},
|
8248 |
+
{
|
8249 |
+
"epoch": 1.1421562614988348,
|
8250 |
+
"grad_norm": 2.0859293937683105,
|
8251 |
+
"learning_rate": 7.856813946269795e-05,
|
8252 |
+
"loss": 1.5817,
|
8253 |
+
"step": 1164
|
8254 |
+
},
|
8255 |
+
{
|
8256 |
+
"epoch": 1.143137495400466,
|
8257 |
+
"grad_norm": 1.9917875528335571,
|
8258 |
+
"learning_rate": 7.841692844624999e-05,
|
8259 |
+
"loss": 1.7425,
|
8260 |
+
"step": 1165
|
8261 |
+
},
|
8262 |
+
{
|
8263 |
+
"epoch": 1.1441187293020973,
|
8264 |
+
"grad_norm": 1.9697751998901367,
|
8265 |
+
"learning_rate": 7.826576917246961e-05,
|
8266 |
+
"loss": 1.4807,
|
8267 |
+
"step": 1166
|
8268 |
+
},
|
8269 |
+
{
|
8270 |
+
"epoch": 1.1450999632037286,
|
8271 |
+
"grad_norm": 2.149662733078003,
|
8272 |
+
"learning_rate": 7.811466200374194e-05,
|
8273 |
+
"loss": 1.902,
|
8274 |
+
"step": 1167
|
8275 |
+
},
|
8276 |
+
{
|
8277 |
+
"epoch": 1.14608119710536,
|
8278 |
+
"grad_norm": 2.1379075050354004,
|
8279 |
+
"learning_rate": 7.796360730232724e-05,
|
8280 |
+
"loss": 1.7285,
|
8281 |
+
"step": 1168
|
8282 |
+
},
|
8283 |
+
{
|
8284 |
+
"epoch": 1.1470624310069912,
|
8285 |
+
"grad_norm": 1.9367038011550903,
|
8286 |
+
"learning_rate": 7.78126054303599e-05,
|
8287 |
+
"loss": 1.5475,
|
8288 |
+
"step": 1169
|
8289 |
+
},
|
8290 |
+
{
|
8291 |
+
"epoch": 1.1480436649086225,
|
8292 |
+
"grad_norm": 1.949225902557373,
|
8293 |
+
"learning_rate": 7.766165674984766e-05,
|
8294 |
+
"loss": 1.6411,
|
8295 |
+
"step": 1170
|
8296 |
+
},
|
8297 |
+
{
|
8298 |
+
"epoch": 1.1490248988102538,
|
8299 |
+
"grad_norm": 1.9248489141464233,
|
8300 |
+
"learning_rate": 7.751076162267086e-05,
|
8301 |
+
"loss": 1.4931,
|
8302 |
+
"step": 1171
|
8303 |
+
},
|
8304 |
+
{
|
8305 |
+
"epoch": 1.1500061327118851,
|
8306 |
+
"grad_norm": 2.1530492305755615,
|
8307 |
+
"learning_rate": 7.735992041058126e-05,
|
8308 |
+
"loss": 1.7015,
|
8309 |
+
"step": 1172
|
8310 |
+
},
|
8311 |
+
{
|
8312 |
+
"epoch": 1.1509873666135164,
|
8313 |
+
"grad_norm": 1.9067368507385254,
|
8314 |
+
"learning_rate": 7.720913347520154e-05,
|
8315 |
+
"loss": 1.723,
|
8316 |
+
"step": 1173
|
8317 |
+
},
|
8318 |
+
{
|
8319 |
+
"epoch": 1.1519686005151477,
|
8320 |
+
"grad_norm": 2.1236000061035156,
|
8321 |
+
"learning_rate": 7.705840117802418e-05,
|
8322 |
+
"loss": 1.4677,
|
8323 |
+
"step": 1174
|
8324 |
+
},
|
8325 |
+
{
|
8326 |
+
"epoch": 1.152949834416779,
|
8327 |
+
"grad_norm": 2.022383213043213,
|
8328 |
+
"learning_rate": 7.69077238804107e-05,
|
8329 |
+
"loss": 1.6665,
|
8330 |
+
"step": 1175
|
8331 |
+
},
|
8332 |
+
{
|
8333 |
+
"epoch": 1.1539310683184105,
|
8334 |
+
"grad_norm": 2.1001341342926025,
|
8335 |
+
"learning_rate": 7.675710194359069e-05,
|
8336 |
+
"loss": 1.832,
|
8337 |
+
"step": 1176
|
8338 |
+
},
|
8339 |
+
{
|
8340 |
+
"epoch": 1.1549123022200418,
|
8341 |
+
"grad_norm": 1.9974193572998047,
|
8342 |
+
"learning_rate": 7.660653572866112e-05,
|
8343 |
+
"loss": 1.4889,
|
8344 |
+
"step": 1177
|
8345 |
+
},
|
8346 |
+
{
|
8347 |
+
"epoch": 1.1558935361216731,
|
8348 |
+
"grad_norm": 1.9368098974227905,
|
8349 |
+
"learning_rate": 7.645602559658533e-05,
|
8350 |
+
"loss": 1.471,
|
8351 |
+
"step": 1178
|
8352 |
+
},
|
8353 |
+
{
|
8354 |
+
"epoch": 1.1568747700233044,
|
8355 |
+
"grad_norm": 2.0265512466430664,
|
8356 |
+
"learning_rate": 7.630557190819217e-05,
|
8357 |
+
"loss": 1.5049,
|
8358 |
+
"step": 1179
|
8359 |
+
},
|
8360 |
+
{
|
8361 |
+
"epoch": 1.1578560039249357,
|
8362 |
+
"grad_norm": 1.9389729499816895,
|
8363 |
+
"learning_rate": 7.61551750241753e-05,
|
8364 |
+
"loss": 1.591,
|
8365 |
+
"step": 1180
|
8366 |
+
},
|
8367 |
+
{
|
8368 |
+
"epoch": 1.158837237826567,
|
8369 |
+
"grad_norm": 2.064378023147583,
|
8370 |
+
"learning_rate": 7.600483530509204e-05,
|
8371 |
+
"loss": 1.6401,
|
8372 |
+
"step": 1181
|
8373 |
+
},
|
8374 |
+
{
|
8375 |
+
"epoch": 1.1598184717281983,
|
8376 |
+
"grad_norm": 1.7993282079696655,
|
8377 |
+
"learning_rate": 7.58545531113627e-05,
|
8378 |
+
"loss": 1.4952,
|
8379 |
+
"step": 1182
|
8380 |
+
},
|
8381 |
+
{
|
8382 |
+
"epoch": 1.1607997056298296,
|
8383 |
+
"grad_norm": 1.8726584911346436,
|
8384 |
+
"learning_rate": 7.57043288032698e-05,
|
8385 |
+
"loss": 1.5525,
|
8386 |
+
"step": 1183
|
8387 |
+
},
|
8388 |
+
{
|
8389 |
+
"epoch": 1.1617809395314609,
|
8390 |
+
"grad_norm": 1.9391907453536987,
|
8391 |
+
"learning_rate": 7.555416274095694e-05,
|
8392 |
+
"loss": 1.6255,
|
8393 |
+
"step": 1184
|
8394 |
+
},
|
8395 |
+
{
|
8396 |
+
"epoch": 1.1627621734330922,
|
8397 |
+
"grad_norm": 2.114257574081421,
|
8398 |
+
"learning_rate": 7.540405528442822e-05,
|
8399 |
+
"loss": 1.7399,
|
8400 |
+
"step": 1185
|
8401 |
+
},
|
8402 |
+
{
|
8403 |
+
"epoch": 1.1637434073347235,
|
8404 |
+
"grad_norm": 2.2467081546783447,
|
8405 |
+
"learning_rate": 7.525400679354712e-05,
|
8406 |
+
"loss": 1.7055,
|
8407 |
+
"step": 1186
|
8408 |
+
},
|
8409 |
+
{
|
8410 |
+
"epoch": 1.1647246412363548,
|
8411 |
+
"grad_norm": 2.285236120223999,
|
8412 |
+
"learning_rate": 7.510401762803575e-05,
|
8413 |
+
"loss": 1.783,
|
8414 |
+
"step": 1187
|
8415 |
+
},
|
8416 |
+
{
|
8417 |
+
"epoch": 1.165705875137986,
|
8418 |
+
"grad_norm": 2.042032480239868,
|
8419 |
+
"learning_rate": 7.495408814747418e-05,
|
8420 |
+
"loss": 1.5541,
|
8421 |
+
"step": 1188
|
8422 |
+
},
|
8423 |
+
{
|
8424 |
+
"epoch": 1.1666871090396174,
|
8425 |
+
"grad_norm": 2.073132276535034,
|
8426 |
+
"learning_rate": 7.480421871129914e-05,
|
8427 |
+
"loss": 1.7099,
|
8428 |
+
"step": 1189
|
8429 |
+
},
|
8430 |
+
{
|
8431 |
+
"epoch": 1.1676683429412487,
|
8432 |
+
"grad_norm": 1.807121992111206,
|
8433 |
+
"learning_rate": 7.465440967880354e-05,
|
8434 |
+
"loss": 1.3538,
|
8435 |
+
"step": 1190
|
8436 |
+
},
|
8437 |
+
{
|
8438 |
+
"epoch": 1.16864957684288,
|
8439 |
+
"grad_norm": 2.3562142848968506,
|
8440 |
+
"learning_rate": 7.450466140913557e-05,
|
8441 |
+
"loss": 1.8335,
|
8442 |
+
"step": 1191
|
8443 |
+
},
|
8444 |
+
{
|
8445 |
+
"epoch": 1.1696308107445113,
|
8446 |
+
"grad_norm": 2.075679302215576,
|
8447 |
+
"learning_rate": 7.435497426129759e-05,
|
8448 |
+
"loss": 1.6448,
|
8449 |
+
"step": 1192
|
8450 |
+
},
|
8451 |
+
{
|
8452 |
+
"epoch": 1.1706120446461425,
|
8453 |
+
"grad_norm": 2.0627026557922363,
|
8454 |
+
"learning_rate": 7.420534859414542e-05,
|
8455 |
+
"loss": 1.5626,
|
8456 |
+
"step": 1193
|
8457 |
+
},
|
8458 |
+
{
|
8459 |
+
"epoch": 1.1715932785477738,
|
8460 |
+
"grad_norm": 2.0521657466888428,
|
8461 |
+
"learning_rate": 7.405578476638768e-05,
|
8462 |
+
"loss": 1.7099,
|
8463 |
+
"step": 1194
|
8464 |
+
},
|
8465 |
+
{
|
8466 |
+
"epoch": 1.1725745124494051,
|
8467 |
+
"grad_norm": 1.9551303386688232,
|
8468 |
+
"learning_rate": 7.390628313658457e-05,
|
8469 |
+
"loss": 1.5361,
|
8470 |
+
"step": 1195
|
8471 |
+
},
|
8472 |
+
{
|
8473 |
+
"epoch": 1.1735557463510364,
|
8474 |
+
"grad_norm": 1.9095375537872314,
|
8475 |
+
"learning_rate": 7.375684406314715e-05,
|
8476 |
+
"loss": 1.4968,
|
8477 |
+
"step": 1196
|
8478 |
+
},
|
8479 |
+
{
|
8480 |
+
"epoch": 1.1745369802526677,
|
8481 |
+
"grad_norm": 2.150273323059082,
|
8482 |
+
"learning_rate": 7.360746790433672e-05,
|
8483 |
+
"loss": 1.6411,
|
8484 |
+
"step": 1197
|
8485 |
+
},
|
8486 |
+
{
|
8487 |
+
"epoch": 1.175518214154299,
|
8488 |
+
"grad_norm": 1.9941962957382202,
|
8489 |
+
"learning_rate": 7.345815501826353e-05,
|
8490 |
+
"loss": 1.6179,
|
8491 |
+
"step": 1198
|
8492 |
+
},
|
8493 |
+
{
|
8494 |
+
"epoch": 1.1764994480559303,
|
8495 |
+
"grad_norm": 2.2665584087371826,
|
8496 |
+
"learning_rate": 7.330890576288619e-05,
|
8497 |
+
"loss": 1.8879,
|
8498 |
+
"step": 1199
|
8499 |
+
},
|
8500 |
+
{
|
8501 |
+
"epoch": 1.1774806819575616,
|
8502 |
+
"grad_norm": 2.202577829360962,
|
8503 |
+
"learning_rate": 7.315972049601086e-05,
|
8504 |
+
"loss": 1.7252,
|
8505 |
+
"step": 1200
|
8506 |
+
},
|
8507 |
+
{
|
8508 |
+
"epoch": 1.1774806819575616,
|
8509 |
+
"eval_loss": 1.8954843282699585,
|
8510 |
+
"eval_runtime": 23.0768,
|
8511 |
+
"eval_samples_per_second": 58.89,
|
8512 |
+
"eval_steps_per_second": 14.733,
|
8513 |
+
"step": 1200
|
8514 |
}
|
8515 |
],
|
8516 |
"logging_steps": 1,
|
|
|
8525 |
"early_stopping_threshold": 0.0
|
8526 |
},
|
8527 |
"attributes": {
|
8528 |
+
"early_stopping_patience_counter": 2
|
8529 |
}
|
8530 |
},
|
8531 |
"TrainerControl": {
|
|
|
8534 |
"should_evaluate": false,
|
8535 |
"should_log": false,
|
8536 |
"should_save": true,
|
8537 |
+
"should_training_stop": true
|
8538 |
},
|
8539 |
"attributes": {}
|
8540 |
}
|
8541 |
},
|
8542 |
+
"total_flos": 8.650775715250176e+16,
|
8543 |
"train_batch_size": 4,
|
8544 |
"trial_name": null,
|
8545 |
"trial_params": null
|