Training in progress, step 6000, checkpoint
Browse files- last-checkpoint/model.safetensors +1 -1
- last-checkpoint/optimizer.pt +1 -1
- last-checkpoint/rng_state_0.pth +1 -1
- last-checkpoint/rng_state_1.pth +1 -1
- last-checkpoint/rng_state_2.pth +1 -1
- last-checkpoint/rng_state_3.pth +1 -1
- last-checkpoint/rng_state_4.pth +1 -1
- last-checkpoint/rng_state_5.pth +1 -1
- last-checkpoint/rng_state_6.pth +1 -1
- last-checkpoint/rng_state_7.pth +1 -1
- last-checkpoint/scheduler.pt +1 -1
- last-checkpoint/trainer_state.json +766 -2
last-checkpoint/model.safetensors
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 738367848
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:ee3ca4993c74fdba66a4a45cec937d889ea635d5a3363dbf41258ab3cdb82d1e
|
3 |
size 738367848
|
last-checkpoint/optimizer.pt
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 1476823354
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:703c818dd2c8e17f4fd71ce7085e8a08cd52a233c9ec0efa1da31b9b6249f59f
|
3 |
size 1476823354
|
last-checkpoint/rng_state_0.pth
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 15984
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:b92a5a9232e10290e92a7ee43e17a65c2d7fd5bd9b7fae4a78bb653de6ff7f1e
|
3 |
size 15984
|
last-checkpoint/rng_state_1.pth
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 15984
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:027e63f43c97b9a1e6e633ec27654b2d81e59843c5c61895f16184d95b5ecfce
|
3 |
size 15984
|
last-checkpoint/rng_state_2.pth
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 15984
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:5ce4c5bab7dcc40e8eda0dabcca4b51013677ae4eb8d8b9aae51fdbac3ff5302
|
3 |
size 15984
|
last-checkpoint/rng_state_3.pth
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 15984
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:3b04fd2cdaec73d40bd342c2736426a28ca23cb93fea46275f9c93f0355e8e51
|
3 |
size 15984
|
last-checkpoint/rng_state_4.pth
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 15984
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:6fb08899b31a5efc329181f0ecc59c2d36f1c1b6251e03bcab322df2bd5b23a5
|
3 |
size 15984
|
last-checkpoint/rng_state_5.pth
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 15984
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:e1f2ce10faab494375937f049ced3ce1d0fb669dd4ede3a5d75a6c0bc4eebe50
|
3 |
size 15984
|
last-checkpoint/rng_state_6.pth
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 15984
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:0fb2fd90ebb124ea406837f497911487ba5c20d875615f6a03594328a7dafc26
|
3 |
size 15984
|
last-checkpoint/rng_state_7.pth
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 15984
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:de98b2885a6933e6cd867d0a9af94f7453c3971e0017fa67668e1a0ca515fd9d
|
3 |
size 15984
|
last-checkpoint/scheduler.pt
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 1000
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:662f374a56de952606a2f764f88941c488163b14a1fd8282c0553ed7f96dbcfe
|
3 |
size 1000
|
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": 250,
|
6 |
-
"global_step":
|
7 |
"is_hyper_param_search": false,
|
8 |
"is_local_process_zero": true,
|
9 |
"is_world_process_zero": true,
|
@@ -3827,6 +3827,770 @@
|
|
3827 |
"eval_spearman_manhattan": 0.745568766414613,
|
3828 |
"eval_steps_per_second": 8.354,
|
3829 |
"step": 5000
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
3830 |
}
|
3831 |
],
|
3832 |
"logging_steps": 10,
|
|
|
1 |
{
|
2 |
"best_metric": null,
|
3 |
"best_model_checkpoint": null,
|
4 |
+
"epoch": 5.623242736644799,
|
5 |
"eval_steps": 250,
|
6 |
+
"global_step": 6000,
|
7 |
"is_hyper_param_search": false,
|
8 |
"is_local_process_zero": true,
|
9 |
"is_world_process_zero": true,
|
|
|
3827 |
"eval_spearman_manhattan": 0.745568766414613,
|
3828 |
"eval_steps_per_second": 8.354,
|
3829 |
"step": 5000
|
3830 |
+
},
|
3831 |
+
{
|
3832 |
+
"epoch": 4.695407685098407,
|
3833 |
+
"grad_norm": 1.6077407598495483,
|
3834 |
+
"learning_rate": 9.9633176646312e-06,
|
3835 |
+
"loss": 0.0993,
|
3836 |
+
"step": 5010
|
3837 |
+
},
|
3838 |
+
{
|
3839 |
+
"epoch": 4.704779756326148,
|
3840 |
+
"grad_norm": 1.206281065940857,
|
3841 |
+
"learning_rate": 9.963244446396931e-06,
|
3842 |
+
"loss": 0.082,
|
3843 |
+
"step": 5020
|
3844 |
+
},
|
3845 |
+
{
|
3846 |
+
"epoch": 4.71415182755389,
|
3847 |
+
"grad_norm": 1.168562650680542,
|
3848 |
+
"learning_rate": 9.963171228162662e-06,
|
3849 |
+
"loss": 0.075,
|
3850 |
+
"step": 5030
|
3851 |
+
},
|
3852 |
+
{
|
3853 |
+
"epoch": 4.723523898781631,
|
3854 |
+
"grad_norm": 1.0943313837051392,
|
3855 |
+
"learning_rate": 9.963098009928394e-06,
|
3856 |
+
"loss": 0.0907,
|
3857 |
+
"step": 5040
|
3858 |
+
},
|
3859 |
+
{
|
3860 |
+
"epoch": 4.7328959700093725,
|
3861 |
+
"grad_norm": 1.1832613945007324,
|
3862 |
+
"learning_rate": 9.963024791694125e-06,
|
3863 |
+
"loss": 0.0776,
|
3864 |
+
"step": 5050
|
3865 |
+
},
|
3866 |
+
{
|
3867 |
+
"epoch": 4.742268041237113,
|
3868 |
+
"grad_norm": 1.1568524837493896,
|
3869 |
+
"learning_rate": 9.962951573459856e-06,
|
3870 |
+
"loss": 0.0956,
|
3871 |
+
"step": 5060
|
3872 |
+
},
|
3873 |
+
{
|
3874 |
+
"epoch": 4.751640112464854,
|
3875 |
+
"grad_norm": 1.4179660081863403,
|
3876 |
+
"learning_rate": 9.962878355225586e-06,
|
3877 |
+
"loss": 0.079,
|
3878 |
+
"step": 5070
|
3879 |
+
},
|
3880 |
+
{
|
3881 |
+
"epoch": 4.761012183692596,
|
3882 |
+
"grad_norm": 1.56465744972229,
|
3883 |
+
"learning_rate": 9.962805136991317e-06,
|
3884 |
+
"loss": 0.0708,
|
3885 |
+
"step": 5080
|
3886 |
+
},
|
3887 |
+
{
|
3888 |
+
"epoch": 4.770384254920337,
|
3889 |
+
"grad_norm": 1.47963547706604,
|
3890 |
+
"learning_rate": 9.962731918757048e-06,
|
3891 |
+
"loss": 0.0817,
|
3892 |
+
"step": 5090
|
3893 |
+
},
|
3894 |
+
{
|
3895 |
+
"epoch": 4.779756326148079,
|
3896 |
+
"grad_norm": 1.4979149103164673,
|
3897 |
+
"learning_rate": 9.962658700522779e-06,
|
3898 |
+
"loss": 0.0859,
|
3899 |
+
"step": 5100
|
3900 |
+
},
|
3901 |
+
{
|
3902 |
+
"epoch": 4.78912839737582,
|
3903 |
+
"grad_norm": 1.0254287719726562,
|
3904 |
+
"learning_rate": 9.962585482288511e-06,
|
3905 |
+
"loss": 0.077,
|
3906 |
+
"step": 5110
|
3907 |
+
},
|
3908 |
+
{
|
3909 |
+
"epoch": 4.798500468603561,
|
3910 |
+
"grad_norm": 1.5644149780273438,
|
3911 |
+
"learning_rate": 9.96251226405424e-06,
|
3912 |
+
"loss": 0.0775,
|
3913 |
+
"step": 5120
|
3914 |
+
},
|
3915 |
+
{
|
3916 |
+
"epoch": 4.807872539831303,
|
3917 |
+
"grad_norm": 1.2777773141860962,
|
3918 |
+
"learning_rate": 9.962439045819971e-06,
|
3919 |
+
"loss": 0.0734,
|
3920 |
+
"step": 5130
|
3921 |
+
},
|
3922 |
+
{
|
3923 |
+
"epoch": 4.817244611059044,
|
3924 |
+
"grad_norm": 1.130614995956421,
|
3925 |
+
"learning_rate": 9.962365827585703e-06,
|
3926 |
+
"loss": 0.082,
|
3927 |
+
"step": 5140
|
3928 |
+
},
|
3929 |
+
{
|
3930 |
+
"epoch": 4.826616682286786,
|
3931 |
+
"grad_norm": 0.9016211032867432,
|
3932 |
+
"learning_rate": 9.962292609351434e-06,
|
3933 |
+
"loss": 0.08,
|
3934 |
+
"step": 5150
|
3935 |
+
},
|
3936 |
+
{
|
3937 |
+
"epoch": 4.835988753514527,
|
3938 |
+
"grad_norm": 1.4159069061279297,
|
3939 |
+
"learning_rate": 9.962219391117165e-06,
|
3940 |
+
"loss": 0.0841,
|
3941 |
+
"step": 5160
|
3942 |
+
},
|
3943 |
+
{
|
3944 |
+
"epoch": 4.845360824742268,
|
3945 |
+
"grad_norm": 1.600085973739624,
|
3946 |
+
"learning_rate": 9.962146172882896e-06,
|
3947 |
+
"loss": 0.0766,
|
3948 |
+
"step": 5170
|
3949 |
+
},
|
3950 |
+
{
|
3951 |
+
"epoch": 4.85473289597001,
|
3952 |
+
"grad_norm": 1.4401110410690308,
|
3953 |
+
"learning_rate": 9.962072954648626e-06,
|
3954 |
+
"loss": 0.0869,
|
3955 |
+
"step": 5180
|
3956 |
+
},
|
3957 |
+
{
|
3958 |
+
"epoch": 4.8641049671977505,
|
3959 |
+
"grad_norm": 1.4603939056396484,
|
3960 |
+
"learning_rate": 9.961999736414357e-06,
|
3961 |
+
"loss": 0.077,
|
3962 |
+
"step": 5190
|
3963 |
+
},
|
3964 |
+
{
|
3965 |
+
"epoch": 4.873477038425492,
|
3966 |
+
"grad_norm": 1.0498592853546143,
|
3967 |
+
"learning_rate": 9.961926518180088e-06,
|
3968 |
+
"loss": 0.0673,
|
3969 |
+
"step": 5200
|
3970 |
+
},
|
3971 |
+
{
|
3972 |
+
"epoch": 4.882849109653233,
|
3973 |
+
"grad_norm": 1.9157027006149292,
|
3974 |
+
"learning_rate": 9.96185329994582e-06,
|
3975 |
+
"loss": 0.0865,
|
3976 |
+
"step": 5210
|
3977 |
+
},
|
3978 |
+
{
|
3979 |
+
"epoch": 4.892221180880974,
|
3980 |
+
"grad_norm": 1.0183812379837036,
|
3981 |
+
"learning_rate": 9.961780081711551e-06,
|
3982 |
+
"loss": 0.0809,
|
3983 |
+
"step": 5220
|
3984 |
+
},
|
3985 |
+
{
|
3986 |
+
"epoch": 4.901593252108716,
|
3987 |
+
"grad_norm": 1.4563605785369873,
|
3988 |
+
"learning_rate": 9.96170686347728e-06,
|
3989 |
+
"loss": 0.086,
|
3990 |
+
"step": 5230
|
3991 |
+
},
|
3992 |
+
{
|
3993 |
+
"epoch": 4.910965323336457,
|
3994 |
+
"grad_norm": 1.1856083869934082,
|
3995 |
+
"learning_rate": 9.961633645243013e-06,
|
3996 |
+
"loss": 0.0802,
|
3997 |
+
"step": 5240
|
3998 |
+
},
|
3999 |
+
{
|
4000 |
+
"epoch": 4.920337394564199,
|
4001 |
+
"grad_norm": 1.3724653720855713,
|
4002 |
+
"learning_rate": 9.961560427008743e-06,
|
4003 |
+
"loss": 0.0839,
|
4004 |
+
"step": 5250
|
4005 |
+
},
|
4006 |
+
{
|
4007 |
+
"epoch": 4.920337394564199,
|
4008 |
+
"eval_loss": 0.04000931978225708,
|
4009 |
+
"eval_pearson_cosine": 0.7643105387687683,
|
4010 |
+
"eval_pearson_dot": 0.6954823732376099,
|
4011 |
+
"eval_pearson_euclidean": 0.7297146320343018,
|
4012 |
+
"eval_pearson_manhattan": 0.7310500144958496,
|
4013 |
+
"eval_runtime": 21.985,
|
4014 |
+
"eval_samples_per_second": 68.228,
|
4015 |
+
"eval_spearman_cosine": 0.7658903505068073,
|
4016 |
+
"eval_spearman_dot": 0.6968591888025883,
|
4017 |
+
"eval_spearman_euclidean": 0.7350736410651904,
|
4018 |
+
"eval_spearman_manhattan": 0.7366836781540181,
|
4019 |
+
"eval_steps_per_second": 8.551,
|
4020 |
+
"step": 5250
|
4021 |
+
},
|
4022 |
+
{
|
4023 |
+
"epoch": 4.92970946579194,
|
4024 |
+
"grad_norm": 1.7151585817337036,
|
4025 |
+
"learning_rate": 9.961487208774474e-06,
|
4026 |
+
"loss": 0.0791,
|
4027 |
+
"step": 5260
|
4028 |
+
},
|
4029 |
+
{
|
4030 |
+
"epoch": 4.939081537019681,
|
4031 |
+
"grad_norm": 1.6940653324127197,
|
4032 |
+
"learning_rate": 9.961413990540205e-06,
|
4033 |
+
"loss": 0.0893,
|
4034 |
+
"step": 5270
|
4035 |
+
},
|
4036 |
+
{
|
4037 |
+
"epoch": 4.948453608247423,
|
4038 |
+
"grad_norm": 1.5087528228759766,
|
4039 |
+
"learning_rate": 9.961340772305936e-06,
|
4040 |
+
"loss": 0.0801,
|
4041 |
+
"step": 5280
|
4042 |
+
},
|
4043 |
+
{
|
4044 |
+
"epoch": 4.957825679475164,
|
4045 |
+
"grad_norm": 1.2038474082946777,
|
4046 |
+
"learning_rate": 9.961267554071666e-06,
|
4047 |
+
"loss": 0.0791,
|
4048 |
+
"step": 5290
|
4049 |
+
},
|
4050 |
+
{
|
4051 |
+
"epoch": 4.967197750702906,
|
4052 |
+
"grad_norm": 1.4044734239578247,
|
4053 |
+
"learning_rate": 9.961194335837397e-06,
|
4054 |
+
"loss": 0.0832,
|
4055 |
+
"step": 5300
|
4056 |
+
},
|
4057 |
+
{
|
4058 |
+
"epoch": 4.976569821930647,
|
4059 |
+
"grad_norm": 1.057298183441162,
|
4060 |
+
"learning_rate": 9.96112111760313e-06,
|
4061 |
+
"loss": 0.0869,
|
4062 |
+
"step": 5310
|
4063 |
+
},
|
4064 |
+
{
|
4065 |
+
"epoch": 4.985941893158388,
|
4066 |
+
"grad_norm": 1.4192899465560913,
|
4067 |
+
"learning_rate": 9.96104789936886e-06,
|
4068 |
+
"loss": 0.0837,
|
4069 |
+
"step": 5320
|
4070 |
+
},
|
4071 |
+
{
|
4072 |
+
"epoch": 4.9953139643861295,
|
4073 |
+
"grad_norm": 1.7742289304733276,
|
4074 |
+
"learning_rate": 9.960974681134591e-06,
|
4075 |
+
"loss": 0.0858,
|
4076 |
+
"step": 5330
|
4077 |
+
},
|
4078 |
+
{
|
4079 |
+
"epoch": 5.0046860356138705,
|
4080 |
+
"grad_norm": 0.9188485741615295,
|
4081 |
+
"learning_rate": 9.960901462900322e-06,
|
4082 |
+
"loss": 0.0684,
|
4083 |
+
"step": 5340
|
4084 |
+
},
|
4085 |
+
{
|
4086 |
+
"epoch": 5.014058106841612,
|
4087 |
+
"grad_norm": 1.6541597843170166,
|
4088 |
+
"learning_rate": 9.960828244666052e-06,
|
4089 |
+
"loss": 0.0669,
|
4090 |
+
"step": 5350
|
4091 |
+
},
|
4092 |
+
{
|
4093 |
+
"epoch": 5.023430178069353,
|
4094 |
+
"grad_norm": 1.5705071687698364,
|
4095 |
+
"learning_rate": 9.960755026431783e-06,
|
4096 |
+
"loss": 0.0646,
|
4097 |
+
"step": 5360
|
4098 |
+
},
|
4099 |
+
{
|
4100 |
+
"epoch": 5.032802249297094,
|
4101 |
+
"grad_norm": 0.9007801413536072,
|
4102 |
+
"learning_rate": 9.960681808197514e-06,
|
4103 |
+
"loss": 0.0721,
|
4104 |
+
"step": 5370
|
4105 |
+
},
|
4106 |
+
{
|
4107 |
+
"epoch": 5.042174320524836,
|
4108 |
+
"grad_norm": 1.044138789176941,
|
4109 |
+
"learning_rate": 9.960608589963245e-06,
|
4110 |
+
"loss": 0.0585,
|
4111 |
+
"step": 5380
|
4112 |
+
},
|
4113 |
+
{
|
4114 |
+
"epoch": 5.051546391752577,
|
4115 |
+
"grad_norm": 1.455098032951355,
|
4116 |
+
"learning_rate": 9.960535371728977e-06,
|
4117 |
+
"loss": 0.0677,
|
4118 |
+
"step": 5390
|
4119 |
+
},
|
4120 |
+
{
|
4121 |
+
"epoch": 5.060918462980319,
|
4122 |
+
"grad_norm": 1.3480255603790283,
|
4123 |
+
"learning_rate": 9.960462153494708e-06,
|
4124 |
+
"loss": 0.0582,
|
4125 |
+
"step": 5400
|
4126 |
+
},
|
4127 |
+
{
|
4128 |
+
"epoch": 5.07029053420806,
|
4129 |
+
"grad_norm": 0.9733775854110718,
|
4130 |
+
"learning_rate": 9.960388935260437e-06,
|
4131 |
+
"loss": 0.057,
|
4132 |
+
"step": 5410
|
4133 |
+
},
|
4134 |
+
{
|
4135 |
+
"epoch": 5.079662605435801,
|
4136 |
+
"grad_norm": 1.202635645866394,
|
4137 |
+
"learning_rate": 9.96031571702617e-06,
|
4138 |
+
"loss": 0.0642,
|
4139 |
+
"step": 5420
|
4140 |
+
},
|
4141 |
+
{
|
4142 |
+
"epoch": 5.089034676663543,
|
4143 |
+
"grad_norm": 1.2410409450531006,
|
4144 |
+
"learning_rate": 9.9602424987919e-06,
|
4145 |
+
"loss": 0.055,
|
4146 |
+
"step": 5430
|
4147 |
+
},
|
4148 |
+
{
|
4149 |
+
"epoch": 5.098406747891284,
|
4150 |
+
"grad_norm": 1.341126799583435,
|
4151 |
+
"learning_rate": 9.960169280557631e-06,
|
4152 |
+
"loss": 0.066,
|
4153 |
+
"step": 5440
|
4154 |
+
},
|
4155 |
+
{
|
4156 |
+
"epoch": 5.107778819119026,
|
4157 |
+
"grad_norm": 1.070065975189209,
|
4158 |
+
"learning_rate": 9.960096062323362e-06,
|
4159 |
+
"loss": 0.0565,
|
4160 |
+
"step": 5450
|
4161 |
+
},
|
4162 |
+
{
|
4163 |
+
"epoch": 5.117150890346767,
|
4164 |
+
"grad_norm": 1.5855072736740112,
|
4165 |
+
"learning_rate": 9.960022844089092e-06,
|
4166 |
+
"loss": 0.0613,
|
4167 |
+
"step": 5460
|
4168 |
+
},
|
4169 |
+
{
|
4170 |
+
"epoch": 5.126522961574508,
|
4171 |
+
"grad_norm": 0.7614333629608154,
|
4172 |
+
"learning_rate": 9.959949625854823e-06,
|
4173 |
+
"loss": 0.0572,
|
4174 |
+
"step": 5470
|
4175 |
+
},
|
4176 |
+
{
|
4177 |
+
"epoch": 5.1358950328022495,
|
4178 |
+
"grad_norm": 1.0969761610031128,
|
4179 |
+
"learning_rate": 9.959876407620554e-06,
|
4180 |
+
"loss": 0.0557,
|
4181 |
+
"step": 5480
|
4182 |
+
},
|
4183 |
+
{
|
4184 |
+
"epoch": 5.14526710402999,
|
4185 |
+
"grad_norm": 1.7454636096954346,
|
4186 |
+
"learning_rate": 9.959803189386286e-06,
|
4187 |
+
"loss": 0.0647,
|
4188 |
+
"step": 5490
|
4189 |
+
},
|
4190 |
+
{
|
4191 |
+
"epoch": 5.154639175257732,
|
4192 |
+
"grad_norm": 0.9625281691551208,
|
4193 |
+
"learning_rate": 9.959729971152017e-06,
|
4194 |
+
"loss": 0.0499,
|
4195 |
+
"step": 5500
|
4196 |
+
},
|
4197 |
+
{
|
4198 |
+
"epoch": 5.154639175257732,
|
4199 |
+
"eval_loss": 0.03924967721104622,
|
4200 |
+
"eval_pearson_cosine": 0.7608553767204285,
|
4201 |
+
"eval_pearson_dot": 0.6993385553359985,
|
4202 |
+
"eval_pearson_euclidean": 0.732108473777771,
|
4203 |
+
"eval_pearson_manhattan": 0.7334935069084167,
|
4204 |
+
"eval_runtime": 28.2448,
|
4205 |
+
"eval_samples_per_second": 53.107,
|
4206 |
+
"eval_spearman_cosine": 0.7615678141531256,
|
4207 |
+
"eval_spearman_dot": 0.6999177956469285,
|
4208 |
+
"eval_spearman_euclidean": 0.7378738640113753,
|
4209 |
+
"eval_spearman_manhattan": 0.7392624046122273,
|
4210 |
+
"eval_steps_per_second": 6.656,
|
4211 |
+
"step": 5500
|
4212 |
+
},
|
4213 |
+
{
|
4214 |
+
"epoch": 5.164011246485473,
|
4215 |
+
"grad_norm": 1.4280071258544922,
|
4216 |
+
"learning_rate": 9.959656752917748e-06,
|
4217 |
+
"loss": 0.0557,
|
4218 |
+
"step": 5510
|
4219 |
+
},
|
4220 |
+
{
|
4221 |
+
"epoch": 5.173383317713214,
|
4222 |
+
"grad_norm": 1.6271259784698486,
|
4223 |
+
"learning_rate": 9.959583534683479e-06,
|
4224 |
+
"loss": 0.0602,
|
4225 |
+
"step": 5520
|
4226 |
+
},
|
4227 |
+
{
|
4228 |
+
"epoch": 5.182755388940956,
|
4229 |
+
"grad_norm": 1.2609021663665771,
|
4230 |
+
"learning_rate": 9.95951031644921e-06,
|
4231 |
+
"loss": 0.0545,
|
4232 |
+
"step": 5530
|
4233 |
+
},
|
4234 |
+
{
|
4235 |
+
"epoch": 5.192127460168697,
|
4236 |
+
"grad_norm": 1.2945165634155273,
|
4237 |
+
"learning_rate": 9.95943709821494e-06,
|
4238 |
+
"loss": 0.0592,
|
4239 |
+
"step": 5540
|
4240 |
+
},
|
4241 |
+
{
|
4242 |
+
"epoch": 5.201499531396439,
|
4243 |
+
"grad_norm": 1.3600184917449951,
|
4244 |
+
"learning_rate": 9.959363879980671e-06,
|
4245 |
+
"loss": 0.0492,
|
4246 |
+
"step": 5550
|
4247 |
+
},
|
4248 |
+
{
|
4249 |
+
"epoch": 5.21087160262418,
|
4250 |
+
"grad_norm": 1.3210471868515015,
|
4251 |
+
"learning_rate": 9.959290661746403e-06,
|
4252 |
+
"loss": 0.0558,
|
4253 |
+
"step": 5560
|
4254 |
+
},
|
4255 |
+
{
|
4256 |
+
"epoch": 5.220243673851921,
|
4257 |
+
"grad_norm": 0.8935280442237854,
|
4258 |
+
"learning_rate": 9.959217443512134e-06,
|
4259 |
+
"loss": 0.0566,
|
4260 |
+
"step": 5570
|
4261 |
+
},
|
4262 |
+
{
|
4263 |
+
"epoch": 5.229615745079663,
|
4264 |
+
"grad_norm": 0.9014615416526794,
|
4265 |
+
"learning_rate": 9.959144225277863e-06,
|
4266 |
+
"loss": 0.0578,
|
4267 |
+
"step": 5580
|
4268 |
+
},
|
4269 |
+
{
|
4270 |
+
"epoch": 5.238987816307404,
|
4271 |
+
"grad_norm": 0.9144461750984192,
|
4272 |
+
"learning_rate": 9.959071007043596e-06,
|
4273 |
+
"loss": 0.0642,
|
4274 |
+
"step": 5590
|
4275 |
+
},
|
4276 |
+
{
|
4277 |
+
"epoch": 5.248359887535146,
|
4278 |
+
"grad_norm": 1.1306620836257935,
|
4279 |
+
"learning_rate": 9.958997788809326e-06,
|
4280 |
+
"loss": 0.0645,
|
4281 |
+
"step": 5600
|
4282 |
+
},
|
4283 |
+
{
|
4284 |
+
"epoch": 5.257731958762887,
|
4285 |
+
"grad_norm": 1.6353179216384888,
|
4286 |
+
"learning_rate": 9.958924570575057e-06,
|
4287 |
+
"loss": 0.0563,
|
4288 |
+
"step": 5610
|
4289 |
+
},
|
4290 |
+
{
|
4291 |
+
"epoch": 5.2671040299906275,
|
4292 |
+
"grad_norm": 1.0438508987426758,
|
4293 |
+
"learning_rate": 9.958851352340788e-06,
|
4294 |
+
"loss": 0.0554,
|
4295 |
+
"step": 5620
|
4296 |
+
},
|
4297 |
+
{
|
4298 |
+
"epoch": 5.276476101218369,
|
4299 |
+
"grad_norm": 1.0287367105484009,
|
4300 |
+
"learning_rate": 9.958778134106519e-06,
|
4301 |
+
"loss": 0.0586,
|
4302 |
+
"step": 5630
|
4303 |
+
},
|
4304 |
+
{
|
4305 |
+
"epoch": 5.28584817244611,
|
4306 |
+
"grad_norm": 1.0613245964050293,
|
4307 |
+
"learning_rate": 9.95870491587225e-06,
|
4308 |
+
"loss": 0.0634,
|
4309 |
+
"step": 5640
|
4310 |
+
},
|
4311 |
+
{
|
4312 |
+
"epoch": 5.295220243673852,
|
4313 |
+
"grad_norm": 1.489405632019043,
|
4314 |
+
"learning_rate": 9.95863169763798e-06,
|
4315 |
+
"loss": 0.0474,
|
4316 |
+
"step": 5650
|
4317 |
+
},
|
4318 |
+
{
|
4319 |
+
"epoch": 5.304592314901593,
|
4320 |
+
"grad_norm": 1.4497292041778564,
|
4321 |
+
"learning_rate": 9.95855847940371e-06,
|
4322 |
+
"loss": 0.056,
|
4323 |
+
"step": 5660
|
4324 |
+
},
|
4325 |
+
{
|
4326 |
+
"epoch": 5.313964386129334,
|
4327 |
+
"grad_norm": 1.2881600856781006,
|
4328 |
+
"learning_rate": 9.958485261169443e-06,
|
4329 |
+
"loss": 0.0561,
|
4330 |
+
"step": 5670
|
4331 |
+
},
|
4332 |
+
{
|
4333 |
+
"epoch": 5.323336457357076,
|
4334 |
+
"grad_norm": 1.4863743782043457,
|
4335 |
+
"learning_rate": 9.958412042935174e-06,
|
4336 |
+
"loss": 0.0562,
|
4337 |
+
"step": 5680
|
4338 |
+
},
|
4339 |
+
{
|
4340 |
+
"epoch": 5.332708528584817,
|
4341 |
+
"grad_norm": 1.325191855430603,
|
4342 |
+
"learning_rate": 9.958338824700903e-06,
|
4343 |
+
"loss": 0.0569,
|
4344 |
+
"step": 5690
|
4345 |
+
},
|
4346 |
+
{
|
4347 |
+
"epoch": 5.342080599812559,
|
4348 |
+
"grad_norm": 1.0650861263275146,
|
4349 |
+
"learning_rate": 9.958265606466636e-06,
|
4350 |
+
"loss": 0.0574,
|
4351 |
+
"step": 5700
|
4352 |
+
},
|
4353 |
+
{
|
4354 |
+
"epoch": 5.3514526710403,
|
4355 |
+
"grad_norm": 1.7255184650421143,
|
4356 |
+
"learning_rate": 9.958192388232366e-06,
|
4357 |
+
"loss": 0.055,
|
4358 |
+
"step": 5710
|
4359 |
+
},
|
4360 |
+
{
|
4361 |
+
"epoch": 5.360824742268041,
|
4362 |
+
"grad_norm": 0.8258642554283142,
|
4363 |
+
"learning_rate": 9.958119169998097e-06,
|
4364 |
+
"loss": 0.0509,
|
4365 |
+
"step": 5720
|
4366 |
+
},
|
4367 |
+
{
|
4368 |
+
"epoch": 5.370196813495783,
|
4369 |
+
"grad_norm": 1.2811216115951538,
|
4370 |
+
"learning_rate": 9.958045951763828e-06,
|
4371 |
+
"loss": 0.0585,
|
4372 |
+
"step": 5730
|
4373 |
+
},
|
4374 |
+
{
|
4375 |
+
"epoch": 5.379568884723524,
|
4376 |
+
"grad_norm": 1.2582824230194092,
|
4377 |
+
"learning_rate": 9.95797273352956e-06,
|
4378 |
+
"loss": 0.0589,
|
4379 |
+
"step": 5740
|
4380 |
+
},
|
4381 |
+
{
|
4382 |
+
"epoch": 5.3889409559512655,
|
4383 |
+
"grad_norm": 1.3511929512023926,
|
4384 |
+
"learning_rate": 9.95789951529529e-06,
|
4385 |
+
"loss": 0.0542,
|
4386 |
+
"step": 5750
|
4387 |
+
},
|
4388 |
+
{
|
4389 |
+
"epoch": 5.3889409559512655,
|
4390 |
+
"eval_loss": 0.03850702941417694,
|
4391 |
+
"eval_pearson_cosine": 0.7663590312004089,
|
4392 |
+
"eval_pearson_dot": 0.7060524225234985,
|
4393 |
+
"eval_pearson_euclidean": 0.7385671734809875,
|
4394 |
+
"eval_pearson_manhattan": 0.7399072647094727,
|
4395 |
+
"eval_runtime": 27.6896,
|
4396 |
+
"eval_samples_per_second": 54.172,
|
4397 |
+
"eval_spearman_cosine": 0.7668814587849042,
|
4398 |
+
"eval_spearman_dot": 0.706466499232552,
|
4399 |
+
"eval_spearman_euclidean": 0.744533534662993,
|
4400 |
+
"eval_spearman_manhattan": 0.7454034343244123,
|
4401 |
+
"eval_steps_per_second": 6.79,
|
4402 |
+
"step": 5750
|
4403 |
+
},
|
4404 |
+
{
|
4405 |
+
"epoch": 5.3983130271790065,
|
4406 |
+
"grad_norm": 1.3905717134475708,
|
4407 |
+
"learning_rate": 9.95782629706102e-06,
|
4408 |
+
"loss": 0.0583,
|
4409 |
+
"step": 5760
|
4410 |
+
},
|
4411 |
+
{
|
4412 |
+
"epoch": 5.4076850984067475,
|
4413 |
+
"grad_norm": 1.5047788619995117,
|
4414 |
+
"learning_rate": 9.957753078826752e-06,
|
4415 |
+
"loss": 0.0605,
|
4416 |
+
"step": 5770
|
4417 |
+
},
|
4418 |
+
{
|
4419 |
+
"epoch": 5.417057169634489,
|
4420 |
+
"grad_norm": 1.280427098274231,
|
4421 |
+
"learning_rate": 9.957679860592483e-06,
|
4422 |
+
"loss": 0.0584,
|
4423 |
+
"step": 5780
|
4424 |
+
},
|
4425 |
+
{
|
4426 |
+
"epoch": 5.42642924086223,
|
4427 |
+
"grad_norm": 1.3530281782150269,
|
4428 |
+
"learning_rate": 9.957606642358214e-06,
|
4429 |
+
"loss": 0.0591,
|
4430 |
+
"step": 5790
|
4431 |
+
},
|
4432 |
+
{
|
4433 |
+
"epoch": 5.435801312089972,
|
4434 |
+
"grad_norm": 1.0610909461975098,
|
4435 |
+
"learning_rate": 9.957533424123945e-06,
|
4436 |
+
"loss": 0.0546,
|
4437 |
+
"step": 5800
|
4438 |
+
},
|
4439 |
+
{
|
4440 |
+
"epoch": 5.445173383317713,
|
4441 |
+
"grad_norm": 0.9637224674224854,
|
4442 |
+
"learning_rate": 9.957460205889675e-06,
|
4443 |
+
"loss": 0.0641,
|
4444 |
+
"step": 5810
|
4445 |
+
},
|
4446 |
+
{
|
4447 |
+
"epoch": 5.454545454545454,
|
4448 |
+
"grad_norm": 1.3324577808380127,
|
4449 |
+
"learning_rate": 9.957386987655406e-06,
|
4450 |
+
"loss": 0.0599,
|
4451 |
+
"step": 5820
|
4452 |
+
},
|
4453 |
+
{
|
4454 |
+
"epoch": 5.463917525773196,
|
4455 |
+
"grad_norm": 0.9660161137580872,
|
4456 |
+
"learning_rate": 9.957313769421137e-06,
|
4457 |
+
"loss": 0.0591,
|
4458 |
+
"step": 5830
|
4459 |
+
},
|
4460 |
+
{
|
4461 |
+
"epoch": 5.473289597000937,
|
4462 |
+
"grad_norm": 1.128570556640625,
|
4463 |
+
"learning_rate": 9.95724055118687e-06,
|
4464 |
+
"loss": 0.0579,
|
4465 |
+
"step": 5840
|
4466 |
+
},
|
4467 |
+
{
|
4468 |
+
"epoch": 5.482661668228679,
|
4469 |
+
"grad_norm": 1.444172739982605,
|
4470 |
+
"learning_rate": 9.9571673329526e-06,
|
4471 |
+
"loss": 0.0636,
|
4472 |
+
"step": 5850
|
4473 |
+
},
|
4474 |
+
{
|
4475 |
+
"epoch": 5.49203373945642,
|
4476 |
+
"grad_norm": 1.3510165214538574,
|
4477 |
+
"learning_rate": 9.95709411471833e-06,
|
4478 |
+
"loss": 0.0631,
|
4479 |
+
"step": 5860
|
4480 |
+
},
|
4481 |
+
{
|
4482 |
+
"epoch": 5.501405810684162,
|
4483 |
+
"grad_norm": 1.0439740419387817,
|
4484 |
+
"learning_rate": 9.957020896484062e-06,
|
4485 |
+
"loss": 0.0635,
|
4486 |
+
"step": 5870
|
4487 |
+
},
|
4488 |
+
{
|
4489 |
+
"epoch": 5.510777881911903,
|
4490 |
+
"grad_norm": 1.15412175655365,
|
4491 |
+
"learning_rate": 9.956947678249792e-06,
|
4492 |
+
"loss": 0.0595,
|
4493 |
+
"step": 5880
|
4494 |
+
},
|
4495 |
+
{
|
4496 |
+
"epoch": 5.520149953139644,
|
4497 |
+
"grad_norm": 1.221147894859314,
|
4498 |
+
"learning_rate": 9.956874460015523e-06,
|
4499 |
+
"loss": 0.0552,
|
4500 |
+
"step": 5890
|
4501 |
+
},
|
4502 |
+
{
|
4503 |
+
"epoch": 5.5295220243673855,
|
4504 |
+
"grad_norm": 1.4210234880447388,
|
4505 |
+
"learning_rate": 9.956801241781254e-06,
|
4506 |
+
"loss": 0.0593,
|
4507 |
+
"step": 5900
|
4508 |
+
},
|
4509 |
+
{
|
4510 |
+
"epoch": 5.5388940955951265,
|
4511 |
+
"grad_norm": 1.1082103252410889,
|
4512 |
+
"learning_rate": 9.956728023546985e-06,
|
4513 |
+
"loss": 0.0535,
|
4514 |
+
"step": 5910
|
4515 |
+
},
|
4516 |
+
{
|
4517 |
+
"epoch": 5.548266166822868,
|
4518 |
+
"grad_norm": 0.8931286334991455,
|
4519 |
+
"learning_rate": 9.956654805312715e-06,
|
4520 |
+
"loss": 0.0556,
|
4521 |
+
"step": 5920
|
4522 |
+
},
|
4523 |
+
{
|
4524 |
+
"epoch": 5.557638238050609,
|
4525 |
+
"grad_norm": 1.5182912349700928,
|
4526 |
+
"learning_rate": 9.956581587078446e-06,
|
4527 |
+
"loss": 0.0583,
|
4528 |
+
"step": 5930
|
4529 |
+
},
|
4530 |
+
{
|
4531 |
+
"epoch": 5.56701030927835,
|
4532 |
+
"grad_norm": 1.2056432962417603,
|
4533 |
+
"learning_rate": 9.956508368844177e-06,
|
4534 |
+
"loss": 0.064,
|
4535 |
+
"step": 5940
|
4536 |
+
},
|
4537 |
+
{
|
4538 |
+
"epoch": 5.576382380506092,
|
4539 |
+
"grad_norm": 1.5039522647857666,
|
4540 |
+
"learning_rate": 9.95643515060991e-06,
|
4541 |
+
"loss": 0.0708,
|
4542 |
+
"step": 5950
|
4543 |
+
},
|
4544 |
+
{
|
4545 |
+
"epoch": 5.585754451733833,
|
4546 |
+
"grad_norm": 1.2651883363723755,
|
4547 |
+
"learning_rate": 9.95636193237564e-06,
|
4548 |
+
"loss": 0.0596,
|
4549 |
+
"step": 5960
|
4550 |
+
},
|
4551 |
+
{
|
4552 |
+
"epoch": 5.595126522961575,
|
4553 |
+
"grad_norm": 1.317690134048462,
|
4554 |
+
"learning_rate": 9.956288714141371e-06,
|
4555 |
+
"loss": 0.0713,
|
4556 |
+
"step": 5970
|
4557 |
+
},
|
4558 |
+
{
|
4559 |
+
"epoch": 5.604498594189316,
|
4560 |
+
"grad_norm": 0.9705867767333984,
|
4561 |
+
"learning_rate": 9.956215495907102e-06,
|
4562 |
+
"loss": 0.0699,
|
4563 |
+
"step": 5980
|
4564 |
+
},
|
4565 |
+
{
|
4566 |
+
"epoch": 5.613870665417057,
|
4567 |
+
"grad_norm": 1.4250271320343018,
|
4568 |
+
"learning_rate": 9.956142277672832e-06,
|
4569 |
+
"loss": 0.0595,
|
4570 |
+
"step": 5990
|
4571 |
+
},
|
4572 |
+
{
|
4573 |
+
"epoch": 5.623242736644799,
|
4574 |
+
"grad_norm": 1.0857118368148804,
|
4575 |
+
"learning_rate": 9.956069059438563e-06,
|
4576 |
+
"loss": 0.0555,
|
4577 |
+
"step": 6000
|
4578 |
+
},
|
4579 |
+
{
|
4580 |
+
"epoch": 5.623242736644799,
|
4581 |
+
"eval_loss": 0.03963544964790344,
|
4582 |
+
"eval_pearson_cosine": 0.7571043968200684,
|
4583 |
+
"eval_pearson_dot": 0.700376570224762,
|
4584 |
+
"eval_pearson_euclidean": 0.7279260158538818,
|
4585 |
+
"eval_pearson_manhattan": 0.729307234287262,
|
4586 |
+
"eval_runtime": 25.5449,
|
4587 |
+
"eval_samples_per_second": 58.72,
|
4588 |
+
"eval_spearman_cosine": 0.7579022153365402,
|
4589 |
+
"eval_spearman_dot": 0.6992710065203335,
|
4590 |
+
"eval_spearman_euclidean": 0.7330627821557505,
|
4591 |
+
"eval_spearman_manhattan": 0.7343750357819732,
|
4592 |
+
"eval_steps_per_second": 7.36,
|
4593 |
+
"step": 6000
|
4594 |
}
|
4595 |
],
|
4596 |
"logging_steps": 10,
|