Training in progress, epoch 2, 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 37789864
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:0d3e6d023e00071389b82d517a75652364fc7f67b1df3b1b6865dc6676f02b02
|
| 3 |
size 37789864
|
last-checkpoint/optimizer.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 75739130
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:94acb3ad391a9e09fce184ac9a6598bf155e8ca9c19b1a3e2c1d99ab825dab1c
|
| 3 |
size 75739130
|
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 sha256:696f0f4ebcda175eb794a5ee308d8e807d6e9ab28887c87361abef72f59c031a
|
| 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 sha256:735cb01abe84c010cb8b13b9f6930c5e7f959fe15fd3ad50df2af5c065f9fb35
|
| 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": 500,
|
| 6 |
-
"global_step":
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
@@ -842,13 +842,848 @@
|
|
| 842 |
"eval_steps_per_second": 0.366,
|
| 843 |
"eval_translation_length": 7761920,
|
| 844 |
"step": 68219
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 845 |
}
|
| 846 |
],
|
| 847 |
"logging_steps": 500,
|
| 848 |
"max_steps": 341095,
|
| 849 |
"num_train_epochs": 5,
|
| 850 |
"save_steps": 5000,
|
| 851 |
-
"total_flos":
|
| 852 |
"trial_name": null,
|
| 853 |
"trial_params": null
|
| 854 |
}
|
|
|
|
| 1 |
{
|
| 2 |
"best_metric": null,
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 2.0,
|
| 5 |
"eval_steps": 500,
|
| 6 |
+
"global_step": 136438,
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
|
|
| 842 |
"eval_steps_per_second": 0.366,
|
| 843 |
"eval_translation_length": 7761920,
|
| 844 |
"step": 68219
|
| 845 |
+
},
|
| 846 |
+
{
|
| 847 |
+
"epoch": 1.0,
|
| 848 |
+
"learning_rate": 7.238646725905267e-05,
|
| 849 |
+
"loss": 3.1948,
|
| 850 |
+
"step": 68500
|
| 851 |
+
},
|
| 852 |
+
{
|
| 853 |
+
"epoch": 1.01,
|
| 854 |
+
"learning_rate": 7.227785439524523e-05,
|
| 855 |
+
"loss": 3.1433,
|
| 856 |
+
"step": 69000
|
| 857 |
+
},
|
| 858 |
+
{
|
| 859 |
+
"epoch": 1.02,
|
| 860 |
+
"learning_rate": 7.216855498944522e-05,
|
| 861 |
+
"loss": 3.083,
|
| 862 |
+
"step": 69500
|
| 863 |
+
},
|
| 864 |
+
{
|
| 865 |
+
"epoch": 1.03,
|
| 866 |
+
"learning_rate": 7.205857136642402e-05,
|
| 867 |
+
"loss": 3.0778,
|
| 868 |
+
"step": 70000
|
| 869 |
+
},
|
| 870 |
+
{
|
| 871 |
+
"epoch": 1.03,
|
| 872 |
+
"learning_rate": 7.194790586550611e-05,
|
| 873 |
+
"loss": 3.1369,
|
| 874 |
+
"step": 70500
|
| 875 |
+
},
|
| 876 |
+
{
|
| 877 |
+
"epoch": 1.04,
|
| 878 |
+
"learning_rate": 7.183656084051938e-05,
|
| 879 |
+
"loss": 3.2123,
|
| 880 |
+
"step": 71000
|
| 881 |
+
},
|
| 882 |
+
{
|
| 883 |
+
"epoch": 1.05,
|
| 884 |
+
"learning_rate": 7.172453865974503e-05,
|
| 885 |
+
"loss": 3.1238,
|
| 886 |
+
"step": 71500
|
| 887 |
+
},
|
| 888 |
+
{
|
| 889 |
+
"epoch": 1.06,
|
| 890 |
+
"learning_rate": 7.161184170586717e-05,
|
| 891 |
+
"loss": 3.1365,
|
| 892 |
+
"step": 72000
|
| 893 |
+
},
|
| 894 |
+
{
|
| 895 |
+
"epoch": 1.06,
|
| 896 |
+
"learning_rate": 7.149847237592218e-05,
|
| 897 |
+
"loss": 3.1618,
|
| 898 |
+
"step": 72500
|
| 899 |
+
},
|
| 900 |
+
{
|
| 901 |
+
"epoch": 1.07,
|
| 902 |
+
"learning_rate": 7.138443308124771e-05,
|
| 903 |
+
"loss": 3.1355,
|
| 904 |
+
"step": 73000
|
| 905 |
+
},
|
| 906 |
+
{
|
| 907 |
+
"epoch": 1.08,
|
| 908 |
+
"learning_rate": 7.12697262474314e-05,
|
| 909 |
+
"loss": 3.1671,
|
| 910 |
+
"step": 73500
|
| 911 |
+
},
|
| 912 |
+
{
|
| 913 |
+
"epoch": 1.08,
|
| 914 |
+
"learning_rate": 7.115435431425928e-05,
|
| 915 |
+
"loss": 3.1321,
|
| 916 |
+
"step": 74000
|
| 917 |
+
},
|
| 918 |
+
{
|
| 919 |
+
"epoch": 1.09,
|
| 920 |
+
"learning_rate": 7.103831973566391e-05,
|
| 921 |
+
"loss": 3.2061,
|
| 922 |
+
"step": 74500
|
| 923 |
+
},
|
| 924 |
+
{
|
| 925 |
+
"epoch": 1.1,
|
| 926 |
+
"learning_rate": 7.092162497967207e-05,
|
| 927 |
+
"loss": 3.1906,
|
| 928 |
+
"step": 75000
|
| 929 |
+
},
|
| 930 |
+
{
|
| 931 |
+
"epoch": 1.11,
|
| 932 |
+
"learning_rate": 7.080427252835243e-05,
|
| 933 |
+
"loss": 3.0249,
|
| 934 |
+
"step": 75500
|
| 935 |
+
},
|
| 936 |
+
{
|
| 937 |
+
"epoch": 1.11,
|
| 938 |
+
"learning_rate": 7.068626487776266e-05,
|
| 939 |
+
"loss": 3.1191,
|
| 940 |
+
"step": 76000
|
| 941 |
+
},
|
| 942 |
+
{
|
| 943 |
+
"epoch": 1.12,
|
| 944 |
+
"learning_rate": 7.056760453789635e-05,
|
| 945 |
+
"loss": 3.0911,
|
| 946 |
+
"step": 76500
|
| 947 |
+
},
|
| 948 |
+
{
|
| 949 |
+
"epoch": 1.13,
|
| 950 |
+
"learning_rate": 7.044829403262961e-05,
|
| 951 |
+
"loss": 3.1361,
|
| 952 |
+
"step": 77000
|
| 953 |
+
},
|
| 954 |
+
{
|
| 955 |
+
"epoch": 1.14,
|
| 956 |
+
"learning_rate": 7.032833589966745e-05,
|
| 957 |
+
"loss": 3.0885,
|
| 958 |
+
"step": 77500
|
| 959 |
+
},
|
| 960 |
+
{
|
| 961 |
+
"epoch": 1.14,
|
| 962 |
+
"learning_rate": 7.02077326904897e-05,
|
| 963 |
+
"loss": 3.117,
|
| 964 |
+
"step": 78000
|
| 965 |
+
},
|
| 966 |
+
{
|
| 967 |
+
"epoch": 1.15,
|
| 968 |
+
"learning_rate": 7.008648697029687e-05,
|
| 969 |
+
"loss": 3.145,
|
| 970 |
+
"step": 78500
|
| 971 |
+
},
|
| 972 |
+
{
|
| 973 |
+
"epoch": 1.16,
|
| 974 |
+
"learning_rate": 6.996460131795548e-05,
|
| 975 |
+
"loss": 3.1472,
|
| 976 |
+
"step": 79000
|
| 977 |
+
},
|
| 978 |
+
{
|
| 979 |
+
"epoch": 1.17,
|
| 980 |
+
"learning_rate": 6.984207832594325e-05,
|
| 981 |
+
"loss": 3.055,
|
| 982 |
+
"step": 79500
|
| 983 |
+
},
|
| 984 |
+
{
|
| 985 |
+
"epoch": 1.17,
|
| 986 |
+
"learning_rate": 6.971892060029398e-05,
|
| 987 |
+
"loss": 3.1593,
|
| 988 |
+
"step": 80000
|
| 989 |
+
},
|
| 990 |
+
{
|
| 991 |
+
"epoch": 1.18,
|
| 992 |
+
"learning_rate": 6.959513076054207e-05,
|
| 993 |
+
"loss": 3.1755,
|
| 994 |
+
"step": 80500
|
| 995 |
+
},
|
| 996 |
+
{
|
| 997 |
+
"epoch": 1.19,
|
| 998 |
+
"learning_rate": 6.947071143966686e-05,
|
| 999 |
+
"loss": 3.1139,
|
| 1000 |
+
"step": 81000
|
| 1001 |
+
},
|
| 1002 |
+
{
|
| 1003 |
+
"epoch": 1.19,
|
| 1004 |
+
"learning_rate": 6.934566528403655e-05,
|
| 1005 |
+
"loss": 3.0516,
|
| 1006 |
+
"step": 81500
|
| 1007 |
+
},
|
| 1008 |
+
{
|
| 1009 |
+
"epoch": 1.2,
|
| 1010 |
+
"learning_rate": 6.921999495335204e-05,
|
| 1011 |
+
"loss": 3.1973,
|
| 1012 |
+
"step": 82000
|
| 1013 |
+
},
|
| 1014 |
+
{
|
| 1015 |
+
"epoch": 1.21,
|
| 1016 |
+
"learning_rate": 6.909370312059018e-05,
|
| 1017 |
+
"loss": 3.0881,
|
| 1018 |
+
"step": 82500
|
| 1019 |
+
},
|
| 1020 |
+
{
|
| 1021 |
+
"epoch": 1.22,
|
| 1022 |
+
"learning_rate": 6.896679247194712e-05,
|
| 1023 |
+
"loss": 3.0801,
|
| 1024 |
+
"step": 83000
|
| 1025 |
+
},
|
| 1026 |
+
{
|
| 1027 |
+
"epoch": 1.22,
|
| 1028 |
+
"learning_rate": 6.883926570678098e-05,
|
| 1029 |
+
"loss": 3.1369,
|
| 1030 |
+
"step": 83500
|
| 1031 |
+
},
|
| 1032 |
+
{
|
| 1033 |
+
"epoch": 1.23,
|
| 1034 |
+
"learning_rate": 6.871112553755457e-05,
|
| 1035 |
+
"loss": 3.1119,
|
| 1036 |
+
"step": 84000
|
| 1037 |
+
},
|
| 1038 |
+
{
|
| 1039 |
+
"epoch": 1.24,
|
| 1040 |
+
"learning_rate": 6.858237468977765e-05,
|
| 1041 |
+
"loss": 3.189,
|
| 1042 |
+
"step": 84500
|
| 1043 |
+
},
|
| 1044 |
+
{
|
| 1045 |
+
"epoch": 1.25,
|
| 1046 |
+
"learning_rate": 6.845301590194898e-05,
|
| 1047 |
+
"loss": 3.1466,
|
| 1048 |
+
"step": 85000
|
| 1049 |
+
},
|
| 1050 |
+
{
|
| 1051 |
+
"epoch": 1.25,
|
| 1052 |
+
"learning_rate": 6.8323051925498e-05,
|
| 1053 |
+
"loss": 3.0994,
|
| 1054 |
+
"step": 85500
|
| 1055 |
+
},
|
| 1056 |
+
{
|
| 1057 |
+
"epoch": 1.26,
|
| 1058 |
+
"learning_rate": 6.819248552472642e-05,
|
| 1059 |
+
"loss": 3.0293,
|
| 1060 |
+
"step": 86000
|
| 1061 |
+
},
|
| 1062 |
+
{
|
| 1063 |
+
"epoch": 1.27,
|
| 1064 |
+
"learning_rate": 6.806131947674935e-05,
|
| 1065 |
+
"loss": 3.2261,
|
| 1066 |
+
"step": 86500
|
| 1067 |
+
},
|
| 1068 |
+
{
|
| 1069 |
+
"epoch": 1.28,
|
| 1070 |
+
"learning_rate": 6.792955657143622e-05,
|
| 1071 |
+
"loss": 3.1032,
|
| 1072 |
+
"step": 87000
|
| 1073 |
+
},
|
| 1074 |
+
{
|
| 1075 |
+
"epoch": 1.28,
|
| 1076 |
+
"learning_rate": 6.779719961135148e-05,
|
| 1077 |
+
"loss": 3.2075,
|
| 1078 |
+
"step": 87500
|
| 1079 |
+
},
|
| 1080 |
+
{
|
| 1081 |
+
"epoch": 1.29,
|
| 1082 |
+
"learning_rate": 6.766425141169503e-05,
|
| 1083 |
+
"loss": 3.1075,
|
| 1084 |
+
"step": 88000
|
| 1085 |
+
},
|
| 1086 |
+
{
|
| 1087 |
+
"epoch": 1.3,
|
| 1088 |
+
"learning_rate": 6.753071480024223e-05,
|
| 1089 |
+
"loss": 3.0876,
|
| 1090 |
+
"step": 88500
|
| 1091 |
+
},
|
| 1092 |
+
{
|
| 1093 |
+
"epoch": 1.3,
|
| 1094 |
+
"learning_rate": 6.739659261728381e-05,
|
| 1095 |
+
"loss": 3.1328,
|
| 1096 |
+
"step": 89000
|
| 1097 |
+
},
|
| 1098 |
+
{
|
| 1099 |
+
"epoch": 1.31,
|
| 1100 |
+
"learning_rate": 6.726188771556552e-05,
|
| 1101 |
+
"loss": 3.1133,
|
| 1102 |
+
"step": 89500
|
| 1103 |
+
},
|
| 1104 |
+
{
|
| 1105 |
+
"epoch": 1.32,
|
| 1106 |
+
"learning_rate": 6.712660296022731e-05,
|
| 1107 |
+
"loss": 3.0909,
|
| 1108 |
+
"step": 90000
|
| 1109 |
+
},
|
| 1110 |
+
{
|
| 1111 |
+
"epoch": 1.33,
|
| 1112 |
+
"learning_rate": 6.699074122874255e-05,
|
| 1113 |
+
"loss": 3.0745,
|
| 1114 |
+
"step": 90500
|
| 1115 |
+
},
|
| 1116 |
+
{
|
| 1117 |
+
"epoch": 1.33,
|
| 1118 |
+
"learning_rate": 6.685430541085673e-05,
|
| 1119 |
+
"loss": 3.2101,
|
| 1120 |
+
"step": 91000
|
| 1121 |
+
},
|
| 1122 |
+
{
|
| 1123 |
+
"epoch": 1.34,
|
| 1124 |
+
"learning_rate": 6.671729840852598e-05,
|
| 1125 |
+
"loss": 3.1616,
|
| 1126 |
+
"step": 91500
|
| 1127 |
+
},
|
| 1128 |
+
{
|
| 1129 |
+
"epoch": 1.35,
|
| 1130 |
+
"learning_rate": 6.657972313585541e-05,
|
| 1131 |
+
"loss": 3.0461,
|
| 1132 |
+
"step": 92000
|
| 1133 |
+
},
|
| 1134 |
+
{
|
| 1135 |
+
"epoch": 1.36,
|
| 1136 |
+
"learning_rate": 6.644158251903713e-05,
|
| 1137 |
+
"loss": 3.0758,
|
| 1138 |
+
"step": 92500
|
| 1139 |
+
},
|
| 1140 |
+
{
|
| 1141 |
+
"epoch": 1.36,
|
| 1142 |
+
"learning_rate": 6.630287949628791e-05,
|
| 1143 |
+
"loss": 3.0974,
|
| 1144 |
+
"step": 93000
|
| 1145 |
+
},
|
| 1146 |
+
{
|
| 1147 |
+
"epoch": 1.37,
|
| 1148 |
+
"learning_rate": 6.616361701778681e-05,
|
| 1149 |
+
"loss": 3.0996,
|
| 1150 |
+
"step": 93500
|
| 1151 |
+
},
|
| 1152 |
+
{
|
| 1153 |
+
"epoch": 1.38,
|
| 1154 |
+
"learning_rate": 6.602379804561237e-05,
|
| 1155 |
+
"loss": 3.0394,
|
| 1156 |
+
"step": 94000
|
| 1157 |
+
},
|
| 1158 |
+
{
|
| 1159 |
+
"epoch": 1.39,
|
| 1160 |
+
"learning_rate": 6.588342555367958e-05,
|
| 1161 |
+
"loss": 3.1698,
|
| 1162 |
+
"step": 94500
|
| 1163 |
+
},
|
| 1164 |
+
{
|
| 1165 |
+
"epoch": 1.39,
|
| 1166 |
+
"learning_rate": 6.57425025276767e-05,
|
| 1167 |
+
"loss": 3.1192,
|
| 1168 |
+
"step": 95000
|
| 1169 |
+
},
|
| 1170 |
+
{
|
| 1171 |
+
"epoch": 1.4,
|
| 1172 |
+
"learning_rate": 6.560103196500169e-05,
|
| 1173 |
+
"loss": 3.225,
|
| 1174 |
+
"step": 95500
|
| 1175 |
+
},
|
| 1176 |
+
{
|
| 1177 |
+
"epoch": 1.41,
|
| 1178 |
+
"learning_rate": 6.545901687469848e-05,
|
| 1179 |
+
"loss": 3.1941,
|
| 1180 |
+
"step": 96000
|
| 1181 |
+
},
|
| 1182 |
+
{
|
| 1183 |
+
"epoch": 1.41,
|
| 1184 |
+
"learning_rate": 6.531646027739298e-05,
|
| 1185 |
+
"loss": 3.1165,
|
| 1186 |
+
"step": 96500
|
| 1187 |
+
},
|
| 1188 |
+
{
|
| 1189 |
+
"epoch": 1.42,
|
| 1190 |
+
"learning_rate": 6.517336520522879e-05,
|
| 1191 |
+
"loss": 3.1859,
|
| 1192 |
+
"step": 97000
|
| 1193 |
+
},
|
| 1194 |
+
{
|
| 1195 |
+
"epoch": 1.43,
|
| 1196 |
+
"learning_rate": 6.50297347018028e-05,
|
| 1197 |
+
"loss": 2.9972,
|
| 1198 |
+
"step": 97500
|
| 1199 |
+
},
|
| 1200 |
+
{
|
| 1201 |
+
"epoch": 1.44,
|
| 1202 |
+
"learning_rate": 6.48855718221003e-05,
|
| 1203 |
+
"loss": 3.146,
|
| 1204 |
+
"step": 98000
|
| 1205 |
+
},
|
| 1206 |
+
{
|
| 1207 |
+
"epoch": 1.44,
|
| 1208 |
+
"learning_rate": 6.47408796324302e-05,
|
| 1209 |
+
"loss": 3.1514,
|
| 1210 |
+
"step": 98500
|
| 1211 |
+
},
|
| 1212 |
+
{
|
| 1213 |
+
"epoch": 1.45,
|
| 1214 |
+
"learning_rate": 6.45956612103596e-05,
|
| 1215 |
+
"loss": 3.1352,
|
| 1216 |
+
"step": 99000
|
| 1217 |
+
},
|
| 1218 |
+
{
|
| 1219 |
+
"epoch": 1.46,
|
| 1220 |
+
"learning_rate": 6.444991964464851e-05,
|
| 1221 |
+
"loss": 3.2001,
|
| 1222 |
+
"step": 99500
|
| 1223 |
+
},
|
| 1224 |
+
{
|
| 1225 |
+
"epoch": 1.47,
|
| 1226 |
+
"learning_rate": 6.430365803518404e-05,
|
| 1227 |
+
"loss": 3.1473,
|
| 1228 |
+
"step": 100000
|
| 1229 |
+
},
|
| 1230 |
+
{
|
| 1231 |
+
"epoch": 1.47,
|
| 1232 |
+
"learning_rate": 6.415687949291448e-05,
|
| 1233 |
+
"loss": 3.1511,
|
| 1234 |
+
"step": 100500
|
| 1235 |
+
},
|
| 1236 |
+
{
|
| 1237 |
+
"epoch": 1.48,
|
| 1238 |
+
"learning_rate": 6.40095871397832e-05,
|
| 1239 |
+
"loss": 3.0987,
|
| 1240 |
+
"step": 101000
|
| 1241 |
+
},
|
| 1242 |
+
{
|
| 1243 |
+
"epoch": 1.49,
|
| 1244 |
+
"learning_rate": 6.386178410866217e-05,
|
| 1245 |
+
"loss": 3.0302,
|
| 1246 |
+
"step": 101500
|
| 1247 |
+
},
|
| 1248 |
+
{
|
| 1249 |
+
"epoch": 1.5,
|
| 1250 |
+
"learning_rate": 6.371347354328537e-05,
|
| 1251 |
+
"loss": 3.1538,
|
| 1252 |
+
"step": 102000
|
| 1253 |
+
},
|
| 1254 |
+
{
|
| 1255 |
+
"epoch": 1.5,
|
| 1256 |
+
"learning_rate": 6.356465859818192e-05,
|
| 1257 |
+
"loss": 3.0881,
|
| 1258 |
+
"step": 102500
|
| 1259 |
+
},
|
| 1260 |
+
{
|
| 1261 |
+
"epoch": 1.51,
|
| 1262 |
+
"learning_rate": 6.341534243860894e-05,
|
| 1263 |
+
"loss": 3.0747,
|
| 1264 |
+
"step": 103000
|
| 1265 |
+
},
|
| 1266 |
+
{
|
| 1267 |
+
"epoch": 1.52,
|
| 1268 |
+
"learning_rate": 6.326552824048426e-05,
|
| 1269 |
+
"loss": 3.0888,
|
| 1270 |
+
"step": 103500
|
| 1271 |
+
},
|
| 1272 |
+
{
|
| 1273 |
+
"epoch": 1.52,
|
| 1274 |
+
"learning_rate": 6.31152191903189e-05,
|
| 1275 |
+
"loss": 3.0922,
|
| 1276 |
+
"step": 104000
|
| 1277 |
+
},
|
| 1278 |
+
{
|
| 1279 |
+
"epoch": 1.53,
|
| 1280 |
+
"learning_rate": 6.296441848514925e-05,
|
| 1281 |
+
"loss": 3.0606,
|
| 1282 |
+
"step": 104500
|
| 1283 |
+
},
|
| 1284 |
+
{
|
| 1285 |
+
"epoch": 1.54,
|
| 1286 |
+
"learning_rate": 6.281312933246902e-05,
|
| 1287 |
+
"loss": 3.0063,
|
| 1288 |
+
"step": 105000
|
| 1289 |
+
},
|
| 1290 |
+
{
|
| 1291 |
+
"epoch": 1.55,
|
| 1292 |
+
"learning_rate": 6.26613549501612e-05,
|
| 1293 |
+
"loss": 3.0931,
|
| 1294 |
+
"step": 105500
|
| 1295 |
+
},
|
| 1296 |
+
{
|
| 1297 |
+
"epoch": 1.55,
|
| 1298 |
+
"learning_rate": 6.250909856642937e-05,
|
| 1299 |
+
"loss": 3.0515,
|
| 1300 |
+
"step": 106000
|
| 1301 |
+
},
|
| 1302 |
+
{
|
| 1303 |
+
"epoch": 1.56,
|
| 1304 |
+
"learning_rate": 6.235636341972923e-05,
|
| 1305 |
+
"loss": 3.1833,
|
| 1306 |
+
"step": 106500
|
| 1307 |
+
},
|
| 1308 |
+
{
|
| 1309 |
+
"epoch": 1.57,
|
| 1310 |
+
"learning_rate": 6.220315275869965e-05,
|
| 1311 |
+
"loss": 3.1109,
|
| 1312 |
+
"step": 107000
|
| 1313 |
+
},
|
| 1314 |
+
{
|
| 1315 |
+
"epoch": 1.58,
|
| 1316 |
+
"learning_rate": 6.204946984209355e-05,
|
| 1317 |
+
"loss": 3.0665,
|
| 1318 |
+
"step": 107500
|
| 1319 |
+
},
|
| 1320 |
+
{
|
| 1321 |
+
"epoch": 1.58,
|
| 1322 |
+
"learning_rate": 6.189531793870862e-05,
|
| 1323 |
+
"loss": 3.0685,
|
| 1324 |
+
"step": 108000
|
| 1325 |
+
},
|
| 1326 |
+
{
|
| 1327 |
+
"epoch": 1.59,
|
| 1328 |
+
"learning_rate": 6.174070032731776e-05,
|
| 1329 |
+
"loss": 3.0161,
|
| 1330 |
+
"step": 108500
|
| 1331 |
+
},
|
| 1332 |
+
{
|
| 1333 |
+
"epoch": 1.6,
|
| 1334 |
+
"learning_rate": 6.15856202965994e-05,
|
| 1335 |
+
"loss": 3.0533,
|
| 1336 |
+
"step": 109000
|
| 1337 |
+
},
|
| 1338 |
+
{
|
| 1339 |
+
"epoch": 1.61,
|
| 1340 |
+
"learning_rate": 6.14300811450675e-05,
|
| 1341 |
+
"loss": 3.1078,
|
| 1342 |
+
"step": 109500
|
| 1343 |
+
},
|
| 1344 |
+
{
|
| 1345 |
+
"epoch": 1.61,
|
| 1346 |
+
"learning_rate": 6.12740861810014e-05,
|
| 1347 |
+
"loss": 3.0947,
|
| 1348 |
+
"step": 110000
|
| 1349 |
+
},
|
| 1350 |
+
{
|
| 1351 |
+
"epoch": 1.62,
|
| 1352 |
+
"learning_rate": 6.111763872237548e-05,
|
| 1353 |
+
"loss": 3.1501,
|
| 1354 |
+
"step": 110500
|
| 1355 |
+
},
|
| 1356 |
+
{
|
| 1357 |
+
"epoch": 1.63,
|
| 1358 |
+
"learning_rate": 6.096074209678854e-05,
|
| 1359 |
+
"loss": 3.1103,
|
| 1360 |
+
"step": 111000
|
| 1361 |
+
},
|
| 1362 |
+
{
|
| 1363 |
+
"epoch": 1.63,
|
| 1364 |
+
"learning_rate": 6.080339964139306e-05,
|
| 1365 |
+
"loss": 3.0426,
|
| 1366 |
+
"step": 111500
|
| 1367 |
+
},
|
| 1368 |
+
{
|
| 1369 |
+
"epoch": 1.64,
|
| 1370 |
+
"learning_rate": 6.0645614702824225e-05,
|
| 1371 |
+
"loss": 3.1161,
|
| 1372 |
+
"step": 112000
|
| 1373 |
+
},
|
| 1374 |
+
{
|
| 1375 |
+
"epoch": 1.65,
|
| 1376 |
+
"learning_rate": 6.0487390637128715e-05,
|
| 1377 |
+
"loss": 3.0774,
|
| 1378 |
+
"step": 112500
|
| 1379 |
+
},
|
| 1380 |
+
{
|
| 1381 |
+
"epoch": 1.66,
|
| 1382 |
+
"learning_rate": 6.032873080969333e-05,
|
| 1383 |
+
"loss": 3.115,
|
| 1384 |
+
"step": 113000
|
| 1385 |
+
},
|
| 1386 |
+
{
|
| 1387 |
+
"epoch": 1.66,
|
| 1388 |
+
"learning_rate": 6.016963859517342e-05,
|
| 1389 |
+
"loss": 3.0771,
|
| 1390 |
+
"step": 113500
|
| 1391 |
+
},
|
| 1392 |
+
{
|
| 1393 |
+
"epoch": 1.67,
|
| 1394 |
+
"learning_rate": 6.00101173774211e-05,
|
| 1395 |
+
"loss": 3.189,
|
| 1396 |
+
"step": 114000
|
| 1397 |
+
},
|
| 1398 |
+
{
|
| 1399 |
+
"epoch": 1.68,
|
| 1400 |
+
"learning_rate": 5.9850170549413294e-05,
|
| 1401 |
+
"loss": 3.1125,
|
| 1402 |
+
"step": 114500
|
| 1403 |
+
},
|
| 1404 |
+
{
|
| 1405 |
+
"epoch": 1.69,
|
| 1406 |
+
"learning_rate": 5.968980151317953e-05,
|
| 1407 |
+
"loss": 3.1025,
|
| 1408 |
+
"step": 115000
|
| 1409 |
+
},
|
| 1410 |
+
{
|
| 1411 |
+
"epoch": 1.69,
|
| 1412 |
+
"learning_rate": 5.9529013679729616e-05,
|
| 1413 |
+
"loss": 3.0638,
|
| 1414 |
+
"step": 115500
|
| 1415 |
+
},
|
| 1416 |
+
{
|
| 1417 |
+
"epoch": 1.7,
|
| 1418 |
+
"learning_rate": 5.936781046898106e-05,
|
| 1419 |
+
"loss": 3.0991,
|
| 1420 |
+
"step": 116000
|
| 1421 |
+
},
|
| 1422 |
+
{
|
| 1423 |
+
"epoch": 1.71,
|
| 1424 |
+
"learning_rate": 5.920619530968634e-05,
|
| 1425 |
+
"loss": 3.1114,
|
| 1426 |
+
"step": 116500
|
| 1427 |
+
},
|
| 1428 |
+
{
|
| 1429 |
+
"epoch": 1.72,
|
| 1430 |
+
"learning_rate": 5.904417163936001e-05,
|
| 1431 |
+
"loss": 3.1139,
|
| 1432 |
+
"step": 117000
|
| 1433 |
+
},
|
| 1434 |
+
{
|
| 1435 |
+
"epoch": 1.72,
|
| 1436 |
+
"learning_rate": 5.88817429042055e-05,
|
| 1437 |
+
"loss": 3.1221,
|
| 1438 |
+
"step": 117500
|
| 1439 |
+
},
|
| 1440 |
+
{
|
| 1441 |
+
"epoch": 1.73,
|
| 1442 |
+
"learning_rate": 5.871891255904191e-05,
|
| 1443 |
+
"loss": 3.1329,
|
| 1444 |
+
"step": 118000
|
| 1445 |
+
},
|
| 1446 |
+
{
|
| 1447 |
+
"epoch": 1.74,
|
| 1448 |
+
"learning_rate": 5.855568406723049e-05,
|
| 1449 |
+
"loss": 3.0913,
|
| 1450 |
+
"step": 118500
|
| 1451 |
+
},
|
| 1452 |
+
{
|
| 1453 |
+
"epoch": 1.74,
|
| 1454 |
+
"learning_rate": 5.839206090060094e-05,
|
| 1455 |
+
"loss": 3.1303,
|
| 1456 |
+
"step": 119000
|
| 1457 |
+
},
|
| 1458 |
+
{
|
| 1459 |
+
"epoch": 1.75,
|
| 1460 |
+
"learning_rate": 5.82280465393776e-05,
|
| 1461 |
+
"loss": 3.1551,
|
| 1462 |
+
"step": 119500
|
| 1463 |
+
},
|
| 1464 |
+
{
|
| 1465 |
+
"epoch": 1.76,
|
| 1466 |
+
"learning_rate": 5.806364447210545e-05,
|
| 1467 |
+
"loss": 3.138,
|
| 1468 |
+
"step": 120000
|
| 1469 |
+
},
|
| 1470 |
+
{
|
| 1471 |
+
"epoch": 1.77,
|
| 1472 |
+
"learning_rate": 5.789885819557585e-05,
|
| 1473 |
+
"loss": 3.0969,
|
| 1474 |
+
"step": 120500
|
| 1475 |
+
},
|
| 1476 |
+
{
|
| 1477 |
+
"epoch": 1.77,
|
| 1478 |
+
"learning_rate": 5.773369121475223e-05,
|
| 1479 |
+
"loss": 3.1819,
|
| 1480 |
+
"step": 121000
|
| 1481 |
+
},
|
| 1482 |
+
{
|
| 1483 |
+
"epoch": 1.78,
|
| 1484 |
+
"learning_rate": 5.756814704269547e-05,
|
| 1485 |
+
"loss": 3.153,
|
| 1486 |
+
"step": 121500
|
| 1487 |
+
},
|
| 1488 |
+
{
|
| 1489 |
+
"epoch": 1.79,
|
| 1490 |
+
"learning_rate": 5.7402229200489246e-05,
|
| 1491 |
+
"loss": 3.2158,
|
| 1492 |
+
"step": 122000
|
| 1493 |
+
},
|
| 1494 |
+
{
|
| 1495 |
+
"epoch": 1.8,
|
| 1496 |
+
"learning_rate": 5.7235941217165076e-05,
|
| 1497 |
+
"loss": 3.1272,
|
| 1498 |
+
"step": 122500
|
| 1499 |
+
},
|
| 1500 |
+
{
|
| 1501 |
+
"epoch": 1.8,
|
| 1502 |
+
"learning_rate": 5.706928662962732e-05,
|
| 1503 |
+
"loss": 3.0045,
|
| 1504 |
+
"step": 123000
|
| 1505 |
+
},
|
| 1506 |
+
{
|
| 1507 |
+
"epoch": 1.81,
|
| 1508 |
+
"learning_rate": 5.690226898257789e-05,
|
| 1509 |
+
"loss": 3.0725,
|
| 1510 |
+
"step": 123500
|
| 1511 |
+
},
|
| 1512 |
+
{
|
| 1513 |
+
"epoch": 1.82,
|
| 1514 |
+
"learning_rate": 5.6734891828440906e-05,
|
| 1515 |
+
"loss": 3.1955,
|
| 1516 |
+
"step": 124000
|
| 1517 |
+
},
|
| 1518 |
+
{
|
| 1519 |
+
"epoch": 1.83,
|
| 1520 |
+
"learning_rate": 5.6567158727287094e-05,
|
| 1521 |
+
"loss": 3.0665,
|
| 1522 |
+
"step": 124500
|
| 1523 |
+
},
|
| 1524 |
+
{
|
| 1525 |
+
"epoch": 1.83,
|
| 1526 |
+
"learning_rate": 5.639907324675809e-05,
|
| 1527 |
+
"loss": 3.1695,
|
| 1528 |
+
"step": 125000
|
| 1529 |
+
},
|
| 1530 |
+
{
|
| 1531 |
+
"epoch": 1.84,
|
| 1532 |
+
"learning_rate": 5.623063896199059e-05,
|
| 1533 |
+
"loss": 3.119,
|
| 1534 |
+
"step": 125500
|
| 1535 |
+
},
|
| 1536 |
+
{
|
| 1537 |
+
"epoch": 1.85,
|
| 1538 |
+
"learning_rate": 5.606185945554023e-05,
|
| 1539 |
+
"loss": 3.0608,
|
| 1540 |
+
"step": 126000
|
| 1541 |
+
},
|
| 1542 |
+
{
|
| 1543 |
+
"epoch": 1.85,
|
| 1544 |
+
"learning_rate": 5.589273831730541e-05,
|
| 1545 |
+
"loss": 3.0697,
|
| 1546 |
+
"step": 126500
|
| 1547 |
+
},
|
| 1548 |
+
{
|
| 1549 |
+
"epoch": 1.86,
|
| 1550 |
+
"learning_rate": 5.572327914445101e-05,
|
| 1551 |
+
"loss": 3.1139,
|
| 1552 |
+
"step": 127000
|
| 1553 |
+
},
|
| 1554 |
+
{
|
| 1555 |
+
"epoch": 1.87,
|
| 1556 |
+
"learning_rate": 5.555348554133178e-05,
|
| 1557 |
+
"loss": 3.0539,
|
| 1558 |
+
"step": 127500
|
| 1559 |
+
},
|
| 1560 |
+
{
|
| 1561 |
+
"epoch": 1.88,
|
| 1562 |
+
"learning_rate": 5.538336111941571e-05,
|
| 1563 |
+
"loss": 3.0316,
|
| 1564 |
+
"step": 128000
|
| 1565 |
+
},
|
| 1566 |
+
{
|
| 1567 |
+
"epoch": 1.88,
|
| 1568 |
+
"learning_rate": 5.521290949720728e-05,
|
| 1569 |
+
"loss": 3.0949,
|
| 1570 |
+
"step": 128500
|
| 1571 |
+
},
|
| 1572 |
+
{
|
| 1573 |
+
"epoch": 1.89,
|
| 1574 |
+
"learning_rate": 5.504213430017038e-05,
|
| 1575 |
+
"loss": 3.0412,
|
| 1576 |
+
"step": 129000
|
| 1577 |
+
},
|
| 1578 |
+
{
|
| 1579 |
+
"epoch": 1.9,
|
| 1580 |
+
"learning_rate": 5.4871039160651295e-05,
|
| 1581 |
+
"loss": 3.2267,
|
| 1582 |
+
"step": 129500
|
| 1583 |
+
},
|
| 1584 |
+
{
|
| 1585 |
+
"epoch": 1.91,
|
| 1586 |
+
"learning_rate": 5.4699627717801396e-05,
|
| 1587 |
+
"loss": 3.0183,
|
| 1588 |
+
"step": 130000
|
| 1589 |
+
},
|
| 1590 |
+
{
|
| 1591 |
+
"epoch": 1.91,
|
| 1592 |
+
"learning_rate": 5.452790361749973e-05,
|
| 1593 |
+
"loss": 3.105,
|
| 1594 |
+
"step": 130500
|
| 1595 |
+
},
|
| 1596 |
+
{
|
| 1597 |
+
"epoch": 1.92,
|
| 1598 |
+
"learning_rate": 5.4355870512275515e-05,
|
| 1599 |
+
"loss": 3.0873,
|
| 1600 |
+
"step": 131000
|
| 1601 |
+
},
|
| 1602 |
+
{
|
| 1603 |
+
"epoch": 1.93,
|
| 1604 |
+
"learning_rate": 5.418353206123042e-05,
|
| 1605 |
+
"loss": 3.1381,
|
| 1606 |
+
"step": 131500
|
| 1607 |
+
},
|
| 1608 |
+
{
|
| 1609 |
+
"epoch": 1.93,
|
| 1610 |
+
"learning_rate": 5.401089192996074e-05,
|
| 1611 |
+
"loss": 3.137,
|
| 1612 |
+
"step": 132000
|
| 1613 |
+
},
|
| 1614 |
+
{
|
| 1615 |
+
"epoch": 1.94,
|
| 1616 |
+
"learning_rate": 5.383795379047944e-05,
|
| 1617 |
+
"loss": 3.0532,
|
| 1618 |
+
"step": 132500
|
| 1619 |
+
},
|
| 1620 |
+
{
|
| 1621 |
+
"epoch": 1.95,
|
| 1622 |
+
"learning_rate": 5.3664721321138046e-05,
|
| 1623 |
+
"loss": 3.0882,
|
| 1624 |
+
"step": 133000
|
| 1625 |
+
},
|
| 1626 |
+
{
|
| 1627 |
+
"epoch": 1.96,
|
| 1628 |
+
"learning_rate": 5.349119820654839e-05,
|
| 1629 |
+
"loss": 3.1207,
|
| 1630 |
+
"step": 133500
|
| 1631 |
+
},
|
| 1632 |
+
{
|
| 1633 |
+
"epoch": 1.96,
|
| 1634 |
+
"learning_rate": 5.331738813750429e-05,
|
| 1635 |
+
"loss": 3.1349,
|
| 1636 |
+
"step": 134000
|
| 1637 |
+
},
|
| 1638 |
+
{
|
| 1639 |
+
"epoch": 1.97,
|
| 1640 |
+
"learning_rate": 5.314329481090299e-05,
|
| 1641 |
+
"loss": 3.0082,
|
| 1642 |
+
"step": 134500
|
| 1643 |
+
},
|
| 1644 |
+
{
|
| 1645 |
+
"epoch": 1.98,
|
| 1646 |
+
"learning_rate": 5.296892192966653e-05,
|
| 1647 |
+
"loss": 3.0684,
|
| 1648 |
+
"step": 135000
|
| 1649 |
+
},
|
| 1650 |
+
{
|
| 1651 |
+
"epoch": 1.99,
|
| 1652 |
+
"learning_rate": 5.2794273202663085e-05,
|
| 1653 |
+
"loss": 3.088,
|
| 1654 |
+
"step": 135500
|
| 1655 |
+
},
|
| 1656 |
+
{
|
| 1657 |
+
"epoch": 1.99,
|
| 1658 |
+
"learning_rate": 5.261935234462791e-05,
|
| 1659 |
+
"loss": 3.0847,
|
| 1660 |
+
"step": 136000
|
| 1661 |
+
},
|
| 1662 |
+
{
|
| 1663 |
+
"epoch": 2.0,
|
| 1664 |
+
"eval_bleu": 1.0,
|
| 1665 |
+
"eval_brevity_penalty": 1.0,
|
| 1666 |
+
"eval_length_ratio": 1.0,
|
| 1667 |
+
"eval_loss": 2.969712018966675,
|
| 1668 |
+
"eval_precisions": [
|
| 1669 |
+
1.0,
|
| 1670 |
+
1.0,
|
| 1671 |
+
1.0,
|
| 1672 |
+
1.0
|
| 1673 |
+
],
|
| 1674 |
+
"eval_reference_length": 7761920,
|
| 1675 |
+
"eval_runtime": 20440.6168,
|
| 1676 |
+
"eval_samples_per_second": 0.742,
|
| 1677 |
+
"eval_steps_per_second": 0.371,
|
| 1678 |
+
"eval_translation_length": 7761920,
|
| 1679 |
+
"step": 136438
|
| 1680 |
}
|
| 1681 |
],
|
| 1682 |
"logging_steps": 500,
|
| 1683 |
"max_steps": 341095,
|
| 1684 |
"num_train_epochs": 5,
|
| 1685 |
"save_steps": 5000,
|
| 1686 |
+
"total_flos": 2.3416327040107807e+18,
|
| 1687 |
"trial_name": null,
|
| 1688 |
"trial_params": null
|
| 1689 |
}
|