Training in progress, epoch 4, checkpoint
Browse files- last-checkpoint/adapter_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 +934 -3
last-checkpoint/adapter_model.safetensors
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 97307544
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:a29ac7ba6f53f92462f92ab317dac22f616978288aa0a0bf5eb7b73a93e8623b
|
3 |
size 97307544
|
last-checkpoint/optimizer.pt
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 49846644
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:87a7f4d414d1b3ef9f1439d2a5c336e25a673350c1f7c1a36bae8691277656ef
|
3 |
size 49846644
|
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:b34c69f6d6c32635ff50b0a1faed18792675796ed61b4917357f7fa723205e2f
|
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:f00b65fe360af6cac79aca5d512b7d43a8aea8109a2419986cb6ec4493d90572
|
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 sha256:3338a4e0a24655344a2e4fe71d8047a4351bfcfb4de0042932ed9ef74e6a9a04
|
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:a28ffcb98b6e03917e67aa0d155cd2bcb5bba7adff5ab9f65b2a312750f07526
|
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 sha256:1f34acd8813d6b9c98be80489c6f644fc7604245d4167815c19292650ca21464
|
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 sha256:359fcea393c9e357ac9329090b6808db28e869fcad41f7d4e9d5f4ea556aac70
|
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:d9ec9b30c1aa25c7c28c666461ba2a0d9c5781d925d9a3f624531498737156b1
|
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 sha256:1b73a0419b8ce412c9ab0e7b0fff0b7e12dd9a30c2645954ee9b29ede96197d8
|
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 1064
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:f3b5e41a3049ed9c6f775655d186804bf7846bdec4fbbd638a092baab2602b8a
|
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": 4.
|
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,
|
@@ -3738,6 +3738,937 @@
|
|
3738 |
"learning_rate": 3.621358721151505e-05,
|
3739 |
"loss": 1.8768,
|
3740 |
"step": 533
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
3741 |
}
|
3742 |
],
|
3743 |
"logging_steps": 1,
|
@@ -3757,7 +4688,7 @@
|
|
3757 |
"attributes": {}
|
3758 |
}
|
3759 |
},
|
3760 |
-
"total_flos": 1.
|
3761 |
"train_batch_size": 6,
|
3762 |
"trial_name": null,
|
3763 |
"trial_params": null
|
|
|
1 |
{
|
2 |
"best_metric": null,
|
3 |
"best_model_checkpoint": null,
|
4 |
+
"epoch": 4.99812382739212,
|
5 |
"eval_steps": 500,
|
6 |
+
"global_step": 666,
|
7 |
"is_hyper_param_search": false,
|
8 |
"is_local_process_zero": true,
|
9 |
"is_world_process_zero": true,
|
|
|
3738 |
"learning_rate": 3.621358721151505e-05,
|
3739 |
"loss": 1.8768,
|
3740 |
"step": 533
|
3741 |
+
},
|
3742 |
+
{
|
3743 |
+
"epoch": 4.00750469043152,
|
3744 |
+
"grad_norm": 0.23720650374889374,
|
3745 |
+
"learning_rate": 3.5808773539880973e-05,
|
3746 |
+
"loss": 1.0723,
|
3747 |
+
"step": 534
|
3748 |
+
},
|
3749 |
+
{
|
3750 |
+
"epoch": 4.01500938086304,
|
3751 |
+
"grad_norm": 0.24943062663078308,
|
3752 |
+
"learning_rate": 3.540574711865146e-05,
|
3753 |
+
"loss": 0.9952,
|
3754 |
+
"step": 535
|
3755 |
+
},
|
3756 |
+
{
|
3757 |
+
"epoch": 4.022514071294559,
|
3758 |
+
"grad_norm": 0.25107982754707336,
|
3759 |
+
"learning_rate": 3.500451899699935e-05,
|
3760 |
+
"loss": 1.0325,
|
3761 |
+
"step": 536
|
3762 |
+
},
|
3763 |
+
{
|
3764 |
+
"epoch": 4.030018761726079,
|
3765 |
+
"grad_norm": 0.2480771690607071,
|
3766 |
+
"learning_rate": 3.460510017479631e-05,
|
3767 |
+
"loss": 1.0892,
|
3768 |
+
"step": 537
|
3769 |
+
},
|
3770 |
+
{
|
3771 |
+
"epoch": 4.037523452157599,
|
3772 |
+
"grad_norm": 0.2446199208498001,
|
3773 |
+
"learning_rate": 3.420750160231118e-05,
|
3774 |
+
"loss": 1.1082,
|
3775 |
+
"step": 538
|
3776 |
+
},
|
3777 |
+
{
|
3778 |
+
"epoch": 4.045028142589119,
|
3779 |
+
"grad_norm": 0.2848975956439972,
|
3780 |
+
"learning_rate": 3.381173417990957e-05,
|
3781 |
+
"loss": 1.0322,
|
3782 |
+
"step": 539
|
3783 |
+
},
|
3784 |
+
{
|
3785 |
+
"epoch": 4.052532833020638,
|
3786 |
+
"grad_norm": 0.2551771104335785,
|
3787 |
+
"learning_rate": 3.3417808757755355e-05,
|
3788 |
+
"loss": 1.029,
|
3789 |
+
"step": 540
|
3790 |
+
},
|
3791 |
+
{
|
3792 |
+
"epoch": 4.0600375234521575,
|
3793 |
+
"grad_norm": 0.2457280158996582,
|
3794 |
+
"learning_rate": 3.302573613551292e-05,
|
3795 |
+
"loss": 1.0687,
|
3796 |
+
"step": 541
|
3797 |
+
},
|
3798 |
+
{
|
3799 |
+
"epoch": 4.067542213883677,
|
3800 |
+
"grad_norm": 0.23483915627002716,
|
3801 |
+
"learning_rate": 3.263552706205128e-05,
|
3802 |
+
"loss": 1.0712,
|
3803 |
+
"step": 542
|
3804 |
+
},
|
3805 |
+
{
|
3806 |
+
"epoch": 4.075046904315197,
|
3807 |
+
"grad_norm": 0.2625795602798462,
|
3808 |
+
"learning_rate": 3.22471922351493e-05,
|
3809 |
+
"loss": 1.0424,
|
3810 |
+
"step": 543
|
3811 |
+
},
|
3812 |
+
{
|
3813 |
+
"epoch": 4.082551594746716,
|
3814 |
+
"grad_norm": 0.25196078419685364,
|
3815 |
+
"learning_rate": 3.186074230120244e-05,
|
3816 |
+
"loss": 0.9866,
|
3817 |
+
"step": 544
|
3818 |
+
},
|
3819 |
+
{
|
3820 |
+
"epoch": 4.090056285178236,
|
3821 |
+
"grad_norm": 0.26508569717407227,
|
3822 |
+
"learning_rate": 3.147618785493083e-05,
|
3823 |
+
"loss": 1.0696,
|
3824 |
+
"step": 545
|
3825 |
+
},
|
3826 |
+
{
|
3827 |
+
"epoch": 4.097560975609756,
|
3828 |
+
"grad_norm": 0.26318731904029846,
|
3829 |
+
"learning_rate": 3.109353943908893e-05,
|
3830 |
+
"loss": 1.1073,
|
3831 |
+
"step": 546
|
3832 |
+
},
|
3833 |
+
{
|
3834 |
+
"epoch": 4.105065666041276,
|
3835 |
+
"grad_norm": 0.2586771249771118,
|
3836 |
+
"learning_rate": 3.071280754417626e-05,
|
3837 |
+
"loss": 1.0968,
|
3838 |
+
"step": 547
|
3839 |
+
},
|
3840 |
+
{
|
3841 |
+
"epoch": 4.112570356472795,
|
3842 |
+
"grad_norm": 0.26417478919029236,
|
3843 |
+
"learning_rate": 3.033400260815008e-05,
|
3844 |
+
"loss": 1.0892,
|
3845 |
+
"step": 548
|
3846 |
+
},
|
3847 |
+
{
|
3848 |
+
"epoch": 4.120075046904315,
|
3849 |
+
"grad_norm": 0.25766685605049133,
|
3850 |
+
"learning_rate": 2.9957135016139122e-05,
|
3851 |
+
"loss": 0.9827,
|
3852 |
+
"step": 549
|
3853 |
+
},
|
3854 |
+
{
|
3855 |
+
"epoch": 4.127579737335835,
|
3856 |
+
"grad_norm": 0.2724906802177429,
|
3857 |
+
"learning_rate": 2.9582215100158706e-05,
|
3858 |
+
"loss": 1.1577,
|
3859 |
+
"step": 550
|
3860 |
+
},
|
3861 |
+
{
|
3862 |
+
"epoch": 4.135084427767355,
|
3863 |
+
"grad_norm": 0.24692752957344055,
|
3864 |
+
"learning_rate": 2.920925313882776e-05,
|
3865 |
+
"loss": 1.1081,
|
3866 |
+
"step": 551
|
3867 |
+
},
|
3868 |
+
{
|
3869 |
+
"epoch": 4.142589118198874,
|
3870 |
+
"grad_norm": 0.27086886763572693,
|
3871 |
+
"learning_rate": 2.8838259357086884e-05,
|
3872 |
+
"loss": 0.9835,
|
3873 |
+
"step": 552
|
3874 |
+
},
|
3875 |
+
{
|
3876 |
+
"epoch": 4.150093808630394,
|
3877 |
+
"grad_norm": 0.25520050525665283,
|
3878 |
+
"learning_rate": 2.846924392591794e-05,
|
3879 |
+
"loss": 1.0608,
|
3880 |
+
"step": 553
|
3881 |
+
},
|
3882 |
+
{
|
3883 |
+
"epoch": 4.157598499061914,
|
3884 |
+
"grad_norm": 0.2637060582637787,
|
3885 |
+
"learning_rate": 2.8102216962065423e-05,
|
3886 |
+
"loss": 0.9214,
|
3887 |
+
"step": 554
|
3888 |
+
},
|
3889 |
+
{
|
3890 |
+
"epoch": 4.165103189493434,
|
3891 |
+
"grad_norm": 0.2559458613395691,
|
3892 |
+
"learning_rate": 2.7737188527758972e-05,
|
3893 |
+
"loss": 0.9785,
|
3894 |
+
"step": 555
|
3895 |
+
},
|
3896 |
+
{
|
3897 |
+
"epoch": 4.1726078799249535,
|
3898 |
+
"grad_norm": 0.2692560851573944,
|
3899 |
+
"learning_rate": 2.7374168630437456e-05,
|
3900 |
+
"loss": 1.1126,
|
3901 |
+
"step": 556
|
3902 |
+
},
|
3903 |
+
{
|
3904 |
+
"epoch": 4.1801125703564725,
|
3905 |
+
"grad_norm": 0.26712581515312195,
|
3906 |
+
"learning_rate": 2.7013167222474756e-05,
|
3907 |
+
"loss": 0.9964,
|
3908 |
+
"step": 557
|
3909 |
+
},
|
3910 |
+
{
|
3911 |
+
"epoch": 4.1876172607879925,
|
3912 |
+
"grad_norm": 0.269401878118515,
|
3913 |
+
"learning_rate": 2.6654194200906833e-05,
|
3914 |
+
"loss": 1.0647,
|
3915 |
+
"step": 558
|
3916 |
+
},
|
3917 |
+
{
|
3918 |
+
"epoch": 4.195121951219512,
|
3919 |
+
"grad_norm": 0.2693142294883728,
|
3920 |
+
"learning_rate": 2.629725940716041e-05,
|
3921 |
+
"loss": 1.0686,
|
3922 |
+
"step": 559
|
3923 |
+
},
|
3924 |
+
{
|
3925 |
+
"epoch": 4.202626641651032,
|
3926 |
+
"grad_norm": 0.23360580205917358,
|
3927 |
+
"learning_rate": 2.5942372626783172e-05,
|
3928 |
+
"loss": 0.9899,
|
3929 |
+
"step": 560
|
3930 |
+
},
|
3931 |
+
{
|
3932 |
+
"epoch": 4.210131332082551,
|
3933 |
+
"grad_norm": 0.27039438486099243,
|
3934 |
+
"learning_rate": 2.5589543589175485e-05,
|
3935 |
+
"loss": 1.1598,
|
3936 |
+
"step": 561
|
3937 |
+
},
|
3938 |
+
{
|
3939 |
+
"epoch": 4.217636022514071,
|
3940 |
+
"grad_norm": 0.2767198979854584,
|
3941 |
+
"learning_rate": 2.523878196732358e-05,
|
3942 |
+
"loss": 0.9577,
|
3943 |
+
"step": 562
|
3944 |
+
},
|
3945 |
+
{
|
3946 |
+
"epoch": 4.225140712945591,
|
3947 |
+
"grad_norm": 0.27451032400131226,
|
3948 |
+
"learning_rate": 2.489009737753459e-05,
|
3949 |
+
"loss": 1.2496,
|
3950 |
+
"step": 563
|
3951 |
+
},
|
3952 |
+
{
|
3953 |
+
"epoch": 4.232645403377111,
|
3954 |
+
"grad_norm": 0.24222443997859955,
|
3955 |
+
"learning_rate": 2.4543499379172615e-05,
|
3956 |
+
"loss": 1.018,
|
3957 |
+
"step": 564
|
3958 |
+
},
|
3959 |
+
{
|
3960 |
+
"epoch": 4.24015009380863,
|
3961 |
+
"grad_norm": 0.26674020290374756,
|
3962 |
+
"learning_rate": 2.4198997474396877e-05,
|
3963 |
+
"loss": 1.0977,
|
3964 |
+
"step": 565
|
3965 |
+
},
|
3966 |
+
{
|
3967 |
+
"epoch": 4.24765478424015,
|
3968 |
+
"grad_norm": 0.2519540786743164,
|
3969 |
+
"learning_rate": 2.3856601107901166e-05,
|
3970 |
+
"loss": 1.0182,
|
3971 |
+
"step": 566
|
3972 |
+
},
|
3973 |
+
{
|
3974 |
+
"epoch": 4.25515947467167,
|
3975 |
+
"grad_norm": 0.2805282771587372,
|
3976 |
+
"learning_rate": 2.351631966665476e-05,
|
3977 |
+
"loss": 0.9412,
|
3978 |
+
"step": 567
|
3979 |
+
},
|
3980 |
+
{
|
3981 |
+
"epoch": 4.26266416510319,
|
3982 |
+
"grad_norm": 0.25769346952438354,
|
3983 |
+
"learning_rate": 2.31781624796453e-05,
|
3984 |
+
"loss": 1.0087,
|
3985 |
+
"step": 568
|
3986 |
+
},
|
3987 |
+
{
|
3988 |
+
"epoch": 4.270168855534709,
|
3989 |
+
"grad_norm": 0.25440049171447754,
|
3990 |
+
"learning_rate": 2.2842138817622883e-05,
|
3991 |
+
"loss": 1.0862,
|
3992 |
+
"step": 569
|
3993 |
+
},
|
3994 |
+
{
|
3995 |
+
"epoch": 4.277673545966229,
|
3996 |
+
"grad_norm": 0.25586551427841187,
|
3997 |
+
"learning_rate": 2.250825789284594e-05,
|
3998 |
+
"loss": 1.0173,
|
3999 |
+
"step": 570
|
4000 |
+
},
|
4001 |
+
{
|
4002 |
+
"epoch": 4.285178236397749,
|
4003 |
+
"grad_norm": 0.2734562158584595,
|
4004 |
+
"learning_rate": 2.217652885882869e-05,
|
4005 |
+
"loss": 1.0565,
|
4006 |
+
"step": 571
|
4007 |
+
},
|
4008 |
+
{
|
4009 |
+
"epoch": 4.2926829268292686,
|
4010 |
+
"grad_norm": 0.2658538818359375,
|
4011 |
+
"learning_rate": 2.1846960810090188e-05,
|
4012 |
+
"loss": 1.0235,
|
4013 |
+
"step": 572
|
4014 |
+
},
|
4015 |
+
{
|
4016 |
+
"epoch": 4.300187617260788,
|
4017 |
+
"grad_norm": 0.2710430920124054,
|
4018 |
+
"learning_rate": 2.151956278190494e-05,
|
4019 |
+
"loss": 1.0054,
|
4020 |
+
"step": 573
|
4021 |
+
},
|
4022 |
+
{
|
4023 |
+
"epoch": 4.3076923076923075,
|
4024 |
+
"grad_norm": 0.2484230250120163,
|
4025 |
+
"learning_rate": 2.119434375005527e-05,
|
4026 |
+
"loss": 0.9814,
|
4027 |
+
"step": 574
|
4028 |
+
},
|
4029 |
+
{
|
4030 |
+
"epoch": 4.315196998123827,
|
4031 |
+
"grad_norm": 0.27289608120918274,
|
4032 |
+
"learning_rate": 2.087131263058526e-05,
|
4033 |
+
"loss": 1.0434,
|
4034 |
+
"step": 575
|
4035 |
+
},
|
4036 |
+
{
|
4037 |
+
"epoch": 4.322701688555347,
|
4038 |
+
"grad_norm": 0.25142380595207214,
|
4039 |
+
"learning_rate": 2.055047827955618e-05,
|
4040 |
+
"loss": 0.996,
|
4041 |
+
"step": 576
|
4042 |
+
},
|
4043 |
+
{
|
4044 |
+
"epoch": 4.330206378986867,
|
4045 |
+
"grad_norm": 0.281368225812912,
|
4046 |
+
"learning_rate": 2.0231849492803852e-05,
|
4047 |
+
"loss": 1.0575,
|
4048 |
+
"step": 577
|
4049 |
+
},
|
4050 |
+
{
|
4051 |
+
"epoch": 4.337711069418386,
|
4052 |
+
"grad_norm": 0.28785380721092224,
|
4053 |
+
"learning_rate": 1.991543500569745e-05,
|
4054 |
+
"loss": 1.0859,
|
4055 |
+
"step": 578
|
4056 |
+
},
|
4057 |
+
{
|
4058 |
+
"epoch": 4.345215759849906,
|
4059 |
+
"grad_norm": 0.28092002868652344,
|
4060 |
+
"learning_rate": 1.960124349289992e-05,
|
4061 |
+
"loss": 1.1349,
|
4062 |
+
"step": 579
|
4063 |
+
},
|
4064 |
+
{
|
4065 |
+
"epoch": 4.352720450281426,
|
4066 |
+
"grad_norm": 0.2787632644176483,
|
4067 |
+
"learning_rate": 1.928928356813032e-05,
|
4068 |
+
"loss": 1.0366,
|
4069 |
+
"step": 580
|
4070 |
+
},
|
4071 |
+
{
|
4072 |
+
"epoch": 4.360225140712946,
|
4073 |
+
"grad_norm": 0.2454954981803894,
|
4074 |
+
"learning_rate": 1.8979563783927565e-05,
|
4075 |
+
"loss": 0.9675,
|
4076 |
+
"step": 581
|
4077 |
+
},
|
4078 |
+
{
|
4079 |
+
"epoch": 4.367729831144465,
|
4080 |
+
"grad_norm": 0.2587539851665497,
|
4081 |
+
"learning_rate": 1.8672092631416013e-05,
|
4082 |
+
"loss": 1.0678,
|
4083 |
+
"step": 582
|
4084 |
+
},
|
4085 |
+
{
|
4086 |
+
"epoch": 4.375234521575985,
|
4087 |
+
"grad_norm": 0.2558777928352356,
|
4088 |
+
"learning_rate": 1.8366878540072614e-05,
|
4089 |
+
"loss": 0.9883,
|
4090 |
+
"step": 583
|
4091 |
+
},
|
4092 |
+
{
|
4093 |
+
"epoch": 4.382739212007505,
|
4094 |
+
"grad_norm": 0.2781766355037689,
|
4095 |
+
"learning_rate": 1.8063929877495892e-05,
|
4096 |
+
"loss": 0.9614,
|
4097 |
+
"step": 584
|
4098 |
+
},
|
4099 |
+
{
|
4100 |
+
"epoch": 4.390243902439025,
|
4101 |
+
"grad_norm": 0.3783913254737854,
|
4102 |
+
"learning_rate": 1.7763254949176414e-05,
|
4103 |
+
"loss": 1.0883,
|
4104 |
+
"step": 585
|
4105 |
+
},
|
4106 |
+
{
|
4107 |
+
"epoch": 4.397748592870544,
|
4108 |
+
"grad_norm": 0.25674161314964294,
|
4109 |
+
"learning_rate": 1.7464861998269243e-05,
|
4110 |
+
"loss": 1.03,
|
4111 |
+
"step": 586
|
4112 |
+
},
|
4113 |
+
{
|
4114 |
+
"epoch": 4.405253283302064,
|
4115 |
+
"grad_norm": 0.2544950842857361,
|
4116 |
+
"learning_rate": 1.7168759205367893e-05,
|
4117 |
+
"loss": 1.0044,
|
4118 |
+
"step": 587
|
4119 |
+
},
|
4120 |
+
{
|
4121 |
+
"epoch": 4.412757973733584,
|
4122 |
+
"grad_norm": 0.2635541558265686,
|
4123 |
+
"learning_rate": 1.6874954688279956e-05,
|
4124 |
+
"loss": 0.9767,
|
4125 |
+
"step": 588
|
4126 |
+
},
|
4127 |
+
{
|
4128 |
+
"epoch": 4.4202626641651035,
|
4129 |
+
"grad_norm": 0.27090808749198914,
|
4130 |
+
"learning_rate": 1.6583456501804725e-05,
|
4131 |
+
"loss": 0.9826,
|
4132 |
+
"step": 589
|
4133 |
+
},
|
4134 |
+
{
|
4135 |
+
"epoch": 4.4277673545966225,
|
4136 |
+
"grad_norm": 0.2597501277923584,
|
4137 |
+
"learning_rate": 1.6294272637512183e-05,
|
4138 |
+
"loss": 1.0187,
|
4139 |
+
"step": 590
|
4140 |
+
},
|
4141 |
+
{
|
4142 |
+
"epoch": 4.435272045028142,
|
4143 |
+
"grad_norm": 0.28622573614120483,
|
4144 |
+
"learning_rate": 1.600741102352409e-05,
|
4145 |
+
"loss": 1.0623,
|
4146 |
+
"step": 591
|
4147 |
+
},
|
4148 |
+
{
|
4149 |
+
"epoch": 4.442776735459662,
|
4150 |
+
"grad_norm": 0.25954049825668335,
|
4151 |
+
"learning_rate": 1.57228795242965e-05,
|
4152 |
+
"loss": 1.0268,
|
4153 |
+
"step": 592
|
4154 |
+
},
|
4155 |
+
{
|
4156 |
+
"epoch": 4.450281425891182,
|
4157 |
+
"grad_norm": 0.2589193284511566,
|
4158 |
+
"learning_rate": 1.544068594040417e-05,
|
4159 |
+
"loss": 0.9672,
|
4160 |
+
"step": 593
|
4161 |
+
},
|
4162 |
+
{
|
4163 |
+
"epoch": 4.457786116322701,
|
4164 |
+
"grad_norm": 0.2728925943374634,
|
4165 |
+
"learning_rate": 1.516083800832676e-05,
|
4166 |
+
"loss": 1.0571,
|
4167 |
+
"step": 594
|
4168 |
+
},
|
4169 |
+
{
|
4170 |
+
"epoch": 4.465290806754221,
|
4171 |
+
"grad_norm": 0.2503071129322052,
|
4172 |
+
"learning_rate": 1.488334340023669e-05,
|
4173 |
+
"loss": 1.0172,
|
4174 |
+
"step": 595
|
4175 |
+
},
|
4176 |
+
{
|
4177 |
+
"epoch": 4.472795497185741,
|
4178 |
+
"grad_norm": 0.2626873552799225,
|
4179 |
+
"learning_rate": 1.4608209723788835e-05,
|
4180 |
+
"loss": 1.036,
|
4181 |
+
"step": 596
|
4182 |
+
},
|
4183 |
+
{
|
4184 |
+
"epoch": 4.480300187617261,
|
4185 |
+
"grad_norm": 0.3006345331668854,
|
4186 |
+
"learning_rate": 1.4335444521911899e-05,
|
4187 |
+
"loss": 1.1155,
|
4188 |
+
"step": 597
|
4189 |
+
},
|
4190 |
+
{
|
4191 |
+
"epoch": 4.487804878048781,
|
4192 |
+
"grad_norm": 0.2929830849170685,
|
4193 |
+
"learning_rate": 1.4065055272601703e-05,
|
4194 |
+
"loss": 1.0768,
|
4195 |
+
"step": 598
|
4196 |
+
},
|
4197 |
+
{
|
4198 |
+
"epoch": 4.4953095684803,
|
4199 |
+
"grad_norm": 0.25911423563957214,
|
4200 |
+
"learning_rate": 1.3797049388716065e-05,
|
4201 |
+
"loss": 1.0368,
|
4202 |
+
"step": 599
|
4203 |
+
},
|
4204 |
+
{
|
4205 |
+
"epoch": 4.50281425891182,
|
4206 |
+
"grad_norm": 0.2785639464855194,
|
4207 |
+
"learning_rate": 1.3531434217771692e-05,
|
4208 |
+
"loss": 1.0747,
|
4209 |
+
"step": 600
|
4210 |
+
},
|
4211 |
+
{
|
4212 |
+
"epoch": 4.51031894934334,
|
4213 |
+
"grad_norm": 0.27841538190841675,
|
4214 |
+
"learning_rate": 1.3268217041742701e-05,
|
4215 |
+
"loss": 1.0116,
|
4216 |
+
"step": 601
|
4217 |
+
},
|
4218 |
+
{
|
4219 |
+
"epoch": 4.517823639774859,
|
4220 |
+
"grad_norm": 0.27080854773521423,
|
4221 |
+
"learning_rate": 1.3007405076860875e-05,
|
4222 |
+
"loss": 0.9667,
|
4223 |
+
"step": 602
|
4224 |
+
},
|
4225 |
+
{
|
4226 |
+
"epoch": 4.525328330206379,
|
4227 |
+
"grad_norm": 0.25426366925239563,
|
4228 |
+
"learning_rate": 1.2749005473418015e-05,
|
4229 |
+
"loss": 0.9367,
|
4230 |
+
"step": 603
|
4231 |
+
},
|
4232 |
+
{
|
4233 |
+
"epoch": 4.532833020637899,
|
4234 |
+
"grad_norm": 0.24661357700824738,
|
4235 |
+
"learning_rate": 1.2493025315569801e-05,
|
4236 |
+
"loss": 0.9841,
|
4237 |
+
"step": 604
|
4238 |
+
},
|
4239 |
+
{
|
4240 |
+
"epoch": 4.5403377110694185,
|
4241 |
+
"grad_norm": 0.2595573365688324,
|
4242 |
+
"learning_rate": 1.2239471621141508e-05,
|
4243 |
+
"loss": 0.9494,
|
4244 |
+
"step": 605
|
4245 |
+
},
|
4246 |
+
{
|
4247 |
+
"epoch": 4.547842401500938,
|
4248 |
+
"grad_norm": 0.27605584263801575,
|
4249 |
+
"learning_rate": 1.1988351341435792e-05,
|
4250 |
+
"loss": 1.0487,
|
4251 |
+
"step": 606
|
4252 |
+
},
|
4253 |
+
{
|
4254 |
+
"epoch": 4.5553470919324575,
|
4255 |
+
"grad_norm": 0.2632475197315216,
|
4256 |
+
"learning_rate": 1.173967136104196e-05,
|
4257 |
+
"loss": 1.0948,
|
4258 |
+
"step": 607
|
4259 |
+
},
|
4260 |
+
{
|
4261 |
+
"epoch": 4.562851782363977,
|
4262 |
+
"grad_norm": 0.2705388367176056,
|
4263 |
+
"learning_rate": 1.1493438497647313e-05,
|
4264 |
+
"loss": 1.0667,
|
4265 |
+
"step": 608
|
4266 |
+
},
|
4267 |
+
{
|
4268 |
+
"epoch": 4.570356472795497,
|
4269 |
+
"grad_norm": 0.25751256942749023,
|
4270 |
+
"learning_rate": 1.1249659501850155e-05,
|
4271 |
+
"loss": 1.1783,
|
4272 |
+
"step": 609
|
4273 |
+
},
|
4274 |
+
{
|
4275 |
+
"epoch": 4.577861163227017,
|
4276 |
+
"grad_norm": 0.2475949227809906,
|
4277 |
+
"learning_rate": 1.1008341056974854e-05,
|
4278 |
+
"loss": 1.0667,
|
4279 |
+
"step": 610
|
4280 |
+
},
|
4281 |
+
{
|
4282 |
+
"epoch": 4.585365853658536,
|
4283 |
+
"grad_norm": 0.26045867800712585,
|
4284 |
+
"learning_rate": 1.0769489778888405e-05,
|
4285 |
+
"loss": 0.9714,
|
4286 |
+
"step": 611
|
4287 |
+
},
|
4288 |
+
{
|
4289 |
+
"epoch": 4.592870544090056,
|
4290 |
+
"grad_norm": 0.2527018189430237,
|
4291 |
+
"learning_rate": 1.0533112215819298e-05,
|
4292 |
+
"loss": 1.0692,
|
4293 |
+
"step": 612
|
4294 |
+
},
|
4295 |
+
{
|
4296 |
+
"epoch": 4.600375234521576,
|
4297 |
+
"grad_norm": 0.25061407685279846,
|
4298 |
+
"learning_rate": 1.029921484817783e-05,
|
4299 |
+
"loss": 0.9363,
|
4300 |
+
"step": 613
|
4301 |
+
},
|
4302 |
+
{
|
4303 |
+
"epoch": 4.607879924953096,
|
4304 |
+
"grad_norm": 0.24939140677452087,
|
4305 |
+
"learning_rate": 1.0067804088378455e-05,
|
4306 |
+
"loss": 1.038,
|
4307 |
+
"step": 614
|
4308 |
+
},
|
4309 |
+
{
|
4310 |
+
"epoch": 4.615384615384615,
|
4311 |
+
"grad_norm": 0.2591976523399353,
|
4312 |
+
"learning_rate": 9.8388862806641e-06,
|
4313 |
+
"loss": 0.9335,
|
4314 |
+
"step": 615
|
4315 |
+
},
|
4316 |
+
{
|
4317 |
+
"epoch": 4.622889305816135,
|
4318 |
+
"grad_norm": 0.26032814383506775,
|
4319 |
+
"learning_rate": 9.612467700932045e-06,
|
4320 |
+
"loss": 1.0562,
|
4321 |
+
"step": 616
|
4322 |
+
},
|
4323 |
+
{
|
4324 |
+
"epoch": 4.630393996247655,
|
4325 |
+
"grad_norm": 0.2716333270072937,
|
4326 |
+
"learning_rate": 9.388554556562049e-06,
|
4327 |
+
"loss": 0.9728,
|
4328 |
+
"step": 617
|
4329 |
+
},
|
4330 |
+
{
|
4331 |
+
"epoch": 4.637898686679175,
|
4332 |
+
"grad_norm": 0.26602107286453247,
|
4333 |
+
"learning_rate": 9.167152986246078e-06,
|
4334 |
+
"loss": 0.9198,
|
4335 |
+
"step": 618
|
4336 |
+
},
|
4337 |
+
{
|
4338 |
+
"epoch": 4.645403377110695,
|
4339 |
+
"grad_norm": 0.27277711033821106,
|
4340 |
+
"learning_rate": 8.948269059820025e-06,
|
4341 |
+
"loss": 1.0677,
|
4342 |
+
"step": 619
|
4343 |
+
},
|
4344 |
+
{
|
4345 |
+
"epoch": 4.652908067542214,
|
4346 |
+
"grad_norm": 0.2933201491832733,
|
4347 |
+
"learning_rate": 8.731908778097302e-06,
|
4348 |
+
"loss": 1.0856,
|
4349 |
+
"step": 620
|
4350 |
+
},
|
4351 |
+
{
|
4352 |
+
"epoch": 4.6604127579737336,
|
4353 |
+
"grad_norm": 0.26761552691459656,
|
4354 |
+
"learning_rate": 8.518078072704338e-06,
|
4355 |
+
"loss": 0.9855,
|
4356 |
+
"step": 621
|
4357 |
+
},
|
4358 |
+
{
|
4359 |
+
"epoch": 4.6679174484052535,
|
4360 |
+
"grad_norm": 0.26589658856391907,
|
4361 |
+
"learning_rate": 8.306782805917904e-06,
|
4362 |
+
"loss": 1.1391,
|
4363 |
+
"step": 622
|
4364 |
+
},
|
4365 |
+
{
|
4366 |
+
"epoch": 4.6754221388367725,
|
4367 |
+
"grad_norm": 0.245664581656456,
|
4368 |
+
"learning_rate": 8.098028770504494e-06,
|
4369 |
+
"loss": 1.0441,
|
4370 |
+
"step": 623
|
4371 |
+
},
|
4372 |
+
{
|
4373 |
+
"epoch": 4.682926829268292,
|
4374 |
+
"grad_norm": 0.25962963700294495,
|
4375 |
+
"learning_rate": 7.891821689561459e-06,
|
4376 |
+
"loss": 0.9938,
|
4377 |
+
"step": 624
|
4378 |
+
},
|
4379 |
+
{
|
4380 |
+
"epoch": 4.690431519699812,
|
4381 |
+
"grad_norm": 0.2487301230430603,
|
4382 |
+
"learning_rate": 7.68816721636004e-06,
|
4383 |
+
"loss": 1.0108,
|
4384 |
+
"step": 625
|
4385 |
+
},
|
4386 |
+
{
|
4387 |
+
"epoch": 4.697936210131332,
|
4388 |
+
"grad_norm": 0.2539227604866028,
|
4389 |
+
"learning_rate": 7.487070934190532e-06,
|
4390 |
+
"loss": 1.0537,
|
4391 |
+
"step": 626
|
4392 |
+
},
|
4393 |
+
{
|
4394 |
+
"epoch": 4.705440900562852,
|
4395 |
+
"grad_norm": 0.2824779152870178,
|
4396 |
+
"learning_rate": 7.288538356209092e-06,
|
4397 |
+
"loss": 1.0141,
|
4398 |
+
"step": 627
|
4399 |
+
},
|
4400 |
+
{
|
4401 |
+
"epoch": 4.712945590994371,
|
4402 |
+
"grad_norm": 0.24700817465782166,
|
4403 |
+
"learning_rate": 7.092574925286614e-06,
|
4404 |
+
"loss": 0.8957,
|
4405 |
+
"step": 628
|
4406 |
+
},
|
4407 |
+
{
|
4408 |
+
"epoch": 4.720450281425891,
|
4409 |
+
"grad_norm": 0.23364083468914032,
|
4410 |
+
"learning_rate": 6.899186013859561e-06,
|
4411 |
+
"loss": 1.0811,
|
4412 |
+
"step": 629
|
4413 |
+
},
|
4414 |
+
{
|
4415 |
+
"epoch": 4.727954971857411,
|
4416 |
+
"grad_norm": 0.25310617685317993,
|
4417 |
+
"learning_rate": 6.708376923782635e-06,
|
4418 |
+
"loss": 1.0577,
|
4419 |
+
"step": 630
|
4420 |
+
},
|
4421 |
+
{
|
4422 |
+
"epoch": 4.735459662288931,
|
4423 |
+
"grad_norm": 0.2497323900461197,
|
4424 |
+
"learning_rate": 6.520152886183406e-06,
|
4425 |
+
"loss": 0.9827,
|
4426 |
+
"step": 631
|
4427 |
+
},
|
4428 |
+
{
|
4429 |
+
"epoch": 4.74296435272045,
|
4430 |
+
"grad_norm": 0.27253395318984985,
|
4431 |
+
"learning_rate": 6.3345190613189635e-06,
|
4432 |
+
"loss": 0.963,
|
4433 |
+
"step": 632
|
4434 |
+
},
|
4435 |
+
{
|
4436 |
+
"epoch": 4.75046904315197,
|
4437 |
+
"grad_norm": 0.25061896443367004,
|
4438 |
+
"learning_rate": 6.151480538434382e-06,
|
4439 |
+
"loss": 1.1357,
|
4440 |
+
"step": 633
|
4441 |
+
},
|
4442 |
+
{
|
4443 |
+
"epoch": 4.75797373358349,
|
4444 |
+
"grad_norm": 0.25422045588493347,
|
4445 |
+
"learning_rate": 5.971042335623229e-06,
|
4446 |
+
"loss": 0.9789,
|
4447 |
+
"step": 634
|
4448 |
+
},
|
4449 |
+
{
|
4450 |
+
"epoch": 4.76547842401501,
|
4451 |
+
"grad_norm": 0.28153881430625916,
|
4452 |
+
"learning_rate": 5.793209399689978e-06,
|
4453 |
+
"loss": 1.0069,
|
4454 |
+
"step": 635
|
4455 |
+
},
|
4456 |
+
{
|
4457 |
+
"epoch": 4.772983114446529,
|
4458 |
+
"grad_norm": 0.26730191707611084,
|
4459 |
+
"learning_rate": 5.617986606014419e-06,
|
4460 |
+
"loss": 0.9413,
|
4461 |
+
"step": 636
|
4462 |
+
},
|
4463 |
+
{
|
4464 |
+
"epoch": 4.780487804878049,
|
4465 |
+
"grad_norm": 0.25754016637802124,
|
4466 |
+
"learning_rate": 5.445378758417925e-06,
|
4467 |
+
"loss": 0.967,
|
4468 |
+
"step": 637
|
4469 |
+
},
|
4470 |
+
{
|
4471 |
+
"epoch": 4.7879924953095685,
|
4472 |
+
"grad_norm": 0.26099568605422974,
|
4473 |
+
"learning_rate": 5.275390589031859e-06,
|
4474 |
+
"loss": 1.0104,
|
4475 |
+
"step": 638
|
4476 |
+
},
|
4477 |
+
{
|
4478 |
+
"epoch": 4.795497185741088,
|
4479 |
+
"grad_norm": 0.27538684010505676,
|
4480 |
+
"learning_rate": 5.108026758167719e-06,
|
4481 |
+
"loss": 1.0233,
|
4482 |
+
"step": 639
|
4483 |
+
},
|
4484 |
+
{
|
4485 |
+
"epoch": 4.803001876172608,
|
4486 |
+
"grad_norm": 0.26298317313194275,
|
4487 |
+
"learning_rate": 4.943291854189493e-06,
|
4488 |
+
"loss": 0.9756,
|
4489 |
+
"step": 640
|
4490 |
+
},
|
4491 |
+
{
|
4492 |
+
"epoch": 4.810506566604127,
|
4493 |
+
"grad_norm": 0.2920237183570862,
|
4494 |
+
"learning_rate": 4.781190393387796e-06,
|
4495 |
+
"loss": 0.985,
|
4496 |
+
"step": 641
|
4497 |
+
},
|
4498 |
+
{
|
4499 |
+
"epoch": 4.818011257035647,
|
4500 |
+
"grad_norm": 0.2614547908306122,
|
4501 |
+
"learning_rate": 4.6217268198560404e-06,
|
4502 |
+
"loss": 1.0421,
|
4503 |
+
"step": 642
|
4504 |
+
},
|
4505 |
+
{
|
4506 |
+
"epoch": 4.825515947467167,
|
4507 |
+
"grad_norm": 0.2709522843360901,
|
4508 |
+
"learning_rate": 4.464905505368658e-06,
|
4509 |
+
"loss": 1.0479,
|
4510 |
+
"step": 643
|
4511 |
+
},
|
4512 |
+
{
|
4513 |
+
"epoch": 4.833020637898686,
|
4514 |
+
"grad_norm": 0.2588852047920227,
|
4515 |
+
"learning_rate": 4.3107307492612086e-06,
|
4516 |
+
"loss": 0.8586,
|
4517 |
+
"step": 644
|
4518 |
+
},
|
4519 |
+
{
|
4520 |
+
"epoch": 4.840525328330206,
|
4521 |
+
"grad_norm": 0.2764037549495697,
|
4522 |
+
"learning_rate": 4.1592067783125015e-06,
|
4523 |
+
"loss": 1.0139,
|
4524 |
+
"step": 645
|
4525 |
+
},
|
4526 |
+
{
|
4527 |
+
"epoch": 4.848030018761726,
|
4528 |
+
"grad_norm": 0.2646092474460602,
|
4529 |
+
"learning_rate": 4.010337746628751e-06,
|
4530 |
+
"loss": 0.8966,
|
4531 |
+
"step": 646
|
4532 |
+
},
|
4533 |
+
{
|
4534 |
+
"epoch": 4.855534709193246,
|
4535 |
+
"grad_norm": 0.26426824927330017,
|
4536 |
+
"learning_rate": 3.864127735529656e-06,
|
4537 |
+
"loss": 1.0248,
|
4538 |
+
"step": 647
|
4539 |
+
},
|
4540 |
+
{
|
4541 |
+
"epoch": 4.863039399624766,
|
4542 |
+
"grad_norm": 0.2680165767669678,
|
4543 |
+
"learning_rate": 3.7205807534365315e-06,
|
4544 |
+
"loss": 1.002,
|
4545 |
+
"step": 648
|
4546 |
+
},
|
4547 |
+
{
|
4548 |
+
"epoch": 4.870544090056285,
|
4549 |
+
"grad_norm": 0.27381497621536255,
|
4550 |
+
"learning_rate": 3.5797007357623945e-06,
|
4551 |
+
"loss": 1.005,
|
4552 |
+
"step": 649
|
4553 |
+
},
|
4554 |
+
{
|
4555 |
+
"epoch": 4.878048780487805,
|
4556 |
+
"grad_norm": 0.24401965737342834,
|
4557 |
+
"learning_rate": 3.441491544804112e-06,
|
4558 |
+
"loss": 0.9504,
|
4559 |
+
"step": 650
|
4560 |
+
},
|
4561 |
+
{
|
4562 |
+
"epoch": 4.885553470919325,
|
4563 |
+
"grad_norm": 0.28096142411231995,
|
4564 |
+
"learning_rate": 3.3059569696364502e-06,
|
4565 |
+
"loss": 1.0629,
|
4566 |
+
"step": 651
|
4567 |
+
},
|
4568 |
+
{
|
4569 |
+
"epoch": 4.893058161350845,
|
4570 |
+
"grad_norm": 0.25856783986091614,
|
4571 |
+
"learning_rate": 3.1731007260082616e-06,
|
4572 |
+
"loss": 1.0072,
|
4573 |
+
"step": 652
|
4574 |
+
},
|
4575 |
+
{
|
4576 |
+
"epoch": 4.900562851782364,
|
4577 |
+
"grad_norm": 0.25341498851776123,
|
4578 |
+
"learning_rate": 3.0429264562405776e-06,
|
4579 |
+
"loss": 1.0113,
|
4580 |
+
"step": 653
|
4581 |
+
},
|
4582 |
+
{
|
4583 |
+
"epoch": 4.9080675422138835,
|
4584 |
+
"grad_norm": 0.2537418007850647,
|
4585 |
+
"learning_rate": 2.9154377291267674e-06,
|
4586 |
+
"loss": 1.0907,
|
4587 |
+
"step": 654
|
4588 |
+
},
|
4589 |
+
{
|
4590 |
+
"epoch": 4.915572232645403,
|
4591 |
+
"grad_norm": 0.252174973487854,
|
4592 |
+
"learning_rate": 2.790638039834668e-06,
|
4593 |
+
"loss": 0.9687,
|
4594 |
+
"step": 655
|
4595 |
+
},
|
4596 |
+
{
|
4597 |
+
"epoch": 4.923076923076923,
|
4598 |
+
"grad_norm": 0.26291623711586,
|
4599 |
+
"learning_rate": 2.6685308098108106e-06,
|
4600 |
+
"loss": 1.0686,
|
4601 |
+
"step": 656
|
4602 |
+
},
|
4603 |
+
{
|
4604 |
+
"epoch": 4.930581613508442,
|
4605 |
+
"grad_norm": 0.2872166037559509,
|
4606 |
+
"learning_rate": 2.5491193866866025e-06,
|
4607 |
+
"loss": 0.9811,
|
4608 |
+
"step": 657
|
4609 |
+
},
|
4610 |
+
{
|
4611 |
+
"epoch": 4.938086303939962,
|
4612 |
+
"grad_norm": 0.28239676356315613,
|
4613 |
+
"learning_rate": 2.432407044186509e-06,
|
4614 |
+
"loss": 1.0442,
|
4615 |
+
"step": 658
|
4616 |
+
},
|
4617 |
+
{
|
4618 |
+
"epoch": 4.945590994371482,
|
4619 |
+
"grad_norm": 0.2630953788757324,
|
4620 |
+
"learning_rate": 2.3183969820383735e-06,
|
4621 |
+
"loss": 0.9368,
|
4622 |
+
"step": 659
|
4623 |
+
},
|
4624 |
+
{
|
4625 |
+
"epoch": 4.953095684803002,
|
4626 |
+
"grad_norm": 0.2792162597179413,
|
4627 |
+
"learning_rate": 2.2070923258856255e-06,
|
4628 |
+
"loss": 1.1048,
|
4629 |
+
"step": 660
|
4630 |
+
},
|
4631 |
+
{
|
4632 |
+
"epoch": 4.960600375234522,
|
4633 |
+
"grad_norm": 0.2646627724170685,
|
4634 |
+
"learning_rate": 2.098496127201648e-06,
|
4635 |
+
"loss": 0.9354,
|
4636 |
+
"step": 661
|
4637 |
+
},
|
4638 |
+
{
|
4639 |
+
"epoch": 4.968105065666041,
|
4640 |
+
"grad_norm": 0.2521073818206787,
|
4641 |
+
"learning_rate": 1.992611363206103e-06,
|
4642 |
+
"loss": 1.0318,
|
4643 |
+
"step": 662
|
4644 |
+
},
|
4645 |
+
{
|
4646 |
+
"epoch": 4.975609756097561,
|
4647 |
+
"grad_norm": 0.27099478244781494,
|
4648 |
+
"learning_rate": 1.889440936783242e-06,
|
4649 |
+
"loss": 1.0485,
|
4650 |
+
"step": 663
|
4651 |
+
},
|
4652 |
+
{
|
4653 |
+
"epoch": 4.983114446529081,
|
4654 |
+
"grad_norm": 0.2716807425022125,
|
4655 |
+
"learning_rate": 1.7889876764024505e-06,
|
4656 |
+
"loss": 0.9949,
|
4657 |
+
"step": 664
|
4658 |
+
},
|
4659 |
+
{
|
4660 |
+
"epoch": 4.9906191369606,
|
4661 |
+
"grad_norm": 0.2702957093715668,
|
4662 |
+
"learning_rate": 1.691254336040595e-06,
|
4663 |
+
"loss": 1.11,
|
4664 |
+
"step": 665
|
4665 |
+
},
|
4666 |
+
{
|
4667 |
+
"epoch": 4.99812382739212,
|
4668 |
+
"grad_norm": 0.2564277648925781,
|
4669 |
+
"learning_rate": 1.59624359510657e-06,
|
4670 |
+
"loss": 0.8788,
|
4671 |
+
"step": 666
|
4672 |
}
|
4673 |
],
|
4674 |
"logging_steps": 1,
|
|
|
4688 |
"attributes": {}
|
4689 |
}
|
4690 |
},
|
4691 |
+
"total_flos": 1.4625303974966723e+18,
|
4692 |
"train_batch_size": 6,
|
4693 |
"trial_name": null,
|
4694 |
"trial_params": null
|