Training in progress, step 600, checkpoint
Browse files- last-checkpoint/optimizer_0/.metadata +0 -0
- last-checkpoint/optimizer_0/__0_0.distcp +1 -1
- last-checkpoint/optimizer_0/__1_0.distcp +1 -1
- last-checkpoint/optimizer_0/__2_0.distcp +1 -1
- last-checkpoint/optimizer_0/__3_0.distcp +1 -1
- last-checkpoint/pytorch_model_fsdp_0/.metadata +0 -0
- last-checkpoint/pytorch_model_fsdp_0/__0_0.distcp +1 -1
- last-checkpoint/pytorch_model_fsdp_0/__1_0.distcp +1 -1
- last-checkpoint/pytorch_model_fsdp_0/__2_0.distcp +1 -1
- last-checkpoint/pytorch_model_fsdp_0/__3_0.distcp +1 -1
- last-checkpoint/rng_state_0.pth +1 -1
- last-checkpoint/scheduler.pt +1 -1
- last-checkpoint/trainer_state.json +763 -3
last-checkpoint/optimizer_0/.metadata
CHANGED
|
Binary files a/last-checkpoint/optimizer_0/.metadata and b/last-checkpoint/optimizer_0/.metadata differ
|
|
|
last-checkpoint/optimizer_0/__0_0.distcp
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 13934748
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:d887d13337c308c5ca733d6323a9864415d8d7fdb689e7054b8b65fc58af94a6
|
| 3 |
size 13934748
|
last-checkpoint/optimizer_0/__1_0.distcp
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 13999412
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:b06daf412411fb1721690a1827cb6e1038a59c2d2dc7e7999d0561604a9799f2
|
| 3 |
size 13999412
|
last-checkpoint/optimizer_0/__2_0.distcp
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 13990904
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:f7aff80990d5d3dea0a8ba059b313ef8993ab8cddb3b3079270f3d150397961f
|
| 3 |
size 13990904
|
last-checkpoint/optimizer_0/__3_0.distcp
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 13990904
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:1b2cfeefe6c41b4091cb2d5b0fdfa5ff9456b7db0f81ce256bdc2e5c5cfd9de5
|
| 3 |
size 13990904
|
last-checkpoint/pytorch_model_fsdp_0/.metadata
CHANGED
|
Binary files a/last-checkpoint/pytorch_model_fsdp_0/.metadata and b/last-checkpoint/pytorch_model_fsdp_0/.metadata differ
|
|
|
last-checkpoint/pytorch_model_fsdp_0/__0_0.distcp
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 6966784
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:e6e7f5bf95c7024205a8119ce9bcf03a4fdd51aae9f436a0a35acb375bb3884b
|
| 3 |
size 6966784
|
last-checkpoint/pytorch_model_fsdp_0/__1_0.distcp
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 6966784
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:3b48bfb689ef000812053e33e2dc970f52c5dd0b8bea9cda6777c2c298951a80
|
| 3 |
size 6966784
|
last-checkpoint/pytorch_model_fsdp_0/__2_0.distcp
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 6966784
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:516752f1fbf9ee4f98d0697eb89075aa40a9ed64fdd213cf582c9b1a40cf8d2d
|
| 3 |
size 6966784
|
last-checkpoint/pytorch_model_fsdp_0/__3_0.distcp
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 6966784
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:f5b0bde13b911b7337906544cf9c5bd2107b176fc7783ebbbad8f40ece936477
|
| 3 |
size 6966784
|
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 14960
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:4cedc08212b07c1673536f40c11159242b59a9f0aadf57db9bad6f13ac81a6af
|
| 3 |
size 14960
|
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 sha256:6cbab2b6a44c6ac01a6a8e4c5859b2a55eaebd5654b8efdeeaf69b4dbf690320
|
| 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": 0.
|
| 5 |
"eval_steps": 20,
|
| 6 |
-
"global_step":
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
@@ -3819,6 +3819,766 @@
|
|
| 3819 |
"eval_samples_per_second": 5.626,
|
| 3820 |
"eval_steps_per_second": 0.189,
|
| 3821 |
"step": 500
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 3822 |
}
|
| 3823 |
],
|
| 3824 |
"logging_steps": 1,
|
|
@@ -3838,7 +4598,7 @@
|
|
| 3838 |
"attributes": {}
|
| 3839 |
}
|
| 3840 |
},
|
| 3841 |
-
"total_flos": 1.
|
| 3842 |
"train_batch_size": 8,
|
| 3843 |
"trial_name": null,
|
| 3844 |
"trial_params": null
|
|
|
|
| 1 |
{
|
| 2 |
"best_metric": null,
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 0.4059539918809202,
|
| 5 |
"eval_steps": 20,
|
| 6 |
+
"global_step": 600,
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
|
|
| 3819 |
"eval_samples_per_second": 5.626,
|
| 3820 |
"eval_steps_per_second": 0.189,
|
| 3821 |
"step": 500
|
| 3822 |
+
},
|
| 3823 |
+
{
|
| 3824 |
+
"epoch": 0.33897158322056836,
|
| 3825 |
+
"grad_norm": 4.322254180908203,
|
| 3826 |
+
"learning_rate": 1.672051050494526e-05,
|
| 3827 |
+
"loss": 0.2227,
|
| 3828 |
+
"step": 501
|
| 3829 |
+
},
|
| 3830 |
+
{
|
| 3831 |
+
"epoch": 0.33964817320703655,
|
| 3832 |
+
"grad_norm": 3.6232991218566895,
|
| 3833 |
+
"learning_rate": 1.67030003102464e-05,
|
| 3834 |
+
"loss": 0.1609,
|
| 3835 |
+
"step": 502
|
| 3836 |
+
},
|
| 3837 |
+
{
|
| 3838 |
+
"epoch": 0.34032476319350474,
|
| 3839 |
+
"grad_norm": 6.037874221801758,
|
| 3840 |
+
"learning_rate": 1.6685452716072946e-05,
|
| 3841 |
+
"loss": 0.144,
|
| 3842 |
+
"step": 503
|
| 3843 |
+
},
|
| 3844 |
+
{
|
| 3845 |
+
"epoch": 0.34100135317997293,
|
| 3846 |
+
"grad_norm": 3.3179101943969727,
|
| 3847 |
+
"learning_rate": 1.6667867820331927e-05,
|
| 3848 |
+
"loss": 0.1325,
|
| 3849 |
+
"step": 504
|
| 3850 |
+
},
|
| 3851 |
+
{
|
| 3852 |
+
"epoch": 0.3416779431664411,
|
| 3853 |
+
"grad_norm": 3.1885428428649902,
|
| 3854 |
+
"learning_rate": 1.6650245721138483e-05,
|
| 3855 |
+
"loss": 0.1493,
|
| 3856 |
+
"step": 505
|
| 3857 |
+
},
|
| 3858 |
+
{
|
| 3859 |
+
"epoch": 0.3423545331529093,
|
| 3860 |
+
"grad_norm": 3.5949137210845947,
|
| 3861 |
+
"learning_rate": 1.6632586516815346e-05,
|
| 3862 |
+
"loss": 0.1273,
|
| 3863 |
+
"step": 506
|
| 3864 |
+
},
|
| 3865 |
+
{
|
| 3866 |
+
"epoch": 0.34303112313937756,
|
| 3867 |
+
"grad_norm": 2.8679418563842773,
|
| 3868 |
+
"learning_rate": 1.6614890305892266e-05,
|
| 3869 |
+
"loss": 0.0887,
|
| 3870 |
+
"step": 507
|
| 3871 |
+
},
|
| 3872 |
+
{
|
| 3873 |
+
"epoch": 0.34370771312584575,
|
| 3874 |
+
"grad_norm": 2.384528160095215,
|
| 3875 |
+
"learning_rate": 1.6597157187105475e-05,
|
| 3876 |
+
"loss": 0.0974,
|
| 3877 |
+
"step": 508
|
| 3878 |
+
},
|
| 3879 |
+
{
|
| 3880 |
+
"epoch": 0.34438430311231394,
|
| 3881 |
+
"grad_norm": 3.2372498512268066,
|
| 3882 |
+
"learning_rate": 1.657938725939713e-05,
|
| 3883 |
+
"loss": 0.1175,
|
| 3884 |
+
"step": 509
|
| 3885 |
+
},
|
| 3886 |
+
{
|
| 3887 |
+
"epoch": 0.34506089309878213,
|
| 3888 |
+
"grad_norm": 2.4635872840881348,
|
| 3889 |
+
"learning_rate": 1.6561580621914764e-05,
|
| 3890 |
+
"loss": 0.0602,
|
| 3891 |
+
"step": 510
|
| 3892 |
+
},
|
| 3893 |
+
{
|
| 3894 |
+
"epoch": 0.3457374830852503,
|
| 3895 |
+
"grad_norm": 4.7463531494140625,
|
| 3896 |
+
"learning_rate": 1.6543737374010742e-05,
|
| 3897 |
+
"loss": 0.1404,
|
| 3898 |
+
"step": 511
|
| 3899 |
+
},
|
| 3900 |
+
{
|
| 3901 |
+
"epoch": 0.34641407307171856,
|
| 3902 |
+
"grad_norm": 3.910125255584717,
|
| 3903 |
+
"learning_rate": 1.6525857615241686e-05,
|
| 3904 |
+
"loss": 0.1732,
|
| 3905 |
+
"step": 512
|
| 3906 |
+
},
|
| 3907 |
+
{
|
| 3908 |
+
"epoch": 0.34709066305818675,
|
| 3909 |
+
"grad_norm": 3.2249362468719482,
|
| 3910 |
+
"learning_rate": 1.6507941445367935e-05,
|
| 3911 |
+
"loss": 0.1706,
|
| 3912 |
+
"step": 513
|
| 3913 |
+
},
|
| 3914 |
+
{
|
| 3915 |
+
"epoch": 0.34776725304465494,
|
| 3916 |
+
"grad_norm": 3.5670406818389893,
|
| 3917 |
+
"learning_rate": 1.648998896435299e-05,
|
| 3918 |
+
"loss": 0.1288,
|
| 3919 |
+
"step": 514
|
| 3920 |
+
},
|
| 3921 |
+
{
|
| 3922 |
+
"epoch": 0.34844384303112313,
|
| 3923 |
+
"grad_norm": 2.954425096511841,
|
| 3924 |
+
"learning_rate": 1.6472000272362937e-05,
|
| 3925 |
+
"loss": 0.1691,
|
| 3926 |
+
"step": 515
|
| 3927 |
+
},
|
| 3928 |
+
{
|
| 3929 |
+
"epoch": 0.3491204330175913,
|
| 3930 |
+
"grad_norm": 3.0924575328826904,
|
| 3931 |
+
"learning_rate": 1.6453975469765913e-05,
|
| 3932 |
+
"loss": 0.1445,
|
| 3933 |
+
"step": 516
|
| 3934 |
+
},
|
| 3935 |
+
{
|
| 3936 |
+
"epoch": 0.3497970230040595,
|
| 3937 |
+
"grad_norm": 3.242204427719116,
|
| 3938 |
+
"learning_rate": 1.643591465713153e-05,
|
| 3939 |
+
"loss": 0.113,
|
| 3940 |
+
"step": 517
|
| 3941 |
+
},
|
| 3942 |
+
{
|
| 3943 |
+
"epoch": 0.35047361299052776,
|
| 3944 |
+
"grad_norm": 3.513796806335449,
|
| 3945 |
+
"learning_rate": 1.6417817935230318e-05,
|
| 3946 |
+
"loss": 0.1342,
|
| 3947 |
+
"step": 518
|
| 3948 |
+
},
|
| 3949 |
+
{
|
| 3950 |
+
"epoch": 0.35115020297699595,
|
| 3951 |
+
"grad_norm": 3.459606409072876,
|
| 3952 |
+
"learning_rate": 1.6399685405033168e-05,
|
| 3953 |
+
"loss": 0.167,
|
| 3954 |
+
"step": 519
|
| 3955 |
+
},
|
| 3956 |
+
{
|
| 3957 |
+
"epoch": 0.35182679296346414,
|
| 3958 |
+
"grad_norm": 3.4279625415802,
|
| 3959 |
+
"learning_rate": 1.6381517167710757e-05,
|
| 3960 |
+
"loss": 0.1466,
|
| 3961 |
+
"step": 520
|
| 3962 |
+
},
|
| 3963 |
+
{
|
| 3964 |
+
"epoch": 0.35182679296346414,
|
| 3965 |
+
"eval_accuracy": 0.8027522935779816,
|
| 3966 |
+
"eval_f1": 0.48687350835322196,
|
| 3967 |
+
"eval_loss": 0.4396270513534546,
|
| 3968 |
+
"eval_precision": 0.8429752066115702,
|
| 3969 |
+
"eval_recall": 0.3422818791946309,
|
| 3970 |
+
"eval_runtime": 53.1809,
|
| 3971 |
+
"eval_samples_per_second": 5.604,
|
| 3972 |
+
"eval_steps_per_second": 0.188,
|
| 3973 |
+
"step": 520
|
| 3974 |
+
},
|
| 3975 |
+
{
|
| 3976 |
+
"epoch": 0.35250338294993233,
|
| 3977 |
+
"grad_norm": 3.181802988052368,
|
| 3978 |
+
"learning_rate": 1.6363313324632995e-05,
|
| 3979 |
+
"loss": 0.1381,
|
| 3980 |
+
"step": 521
|
| 3981 |
+
},
|
| 3982 |
+
{
|
| 3983 |
+
"epoch": 0.3531799729364005,
|
| 3984 |
+
"grad_norm": 2.620626449584961,
|
| 3985 |
+
"learning_rate": 1.6345073977368455e-05,
|
| 3986 |
+
"loss": 0.1523,
|
| 3987 |
+
"step": 522
|
| 3988 |
+
},
|
| 3989 |
+
{
|
| 3990 |
+
"epoch": 0.35385656292286877,
|
| 3991 |
+
"grad_norm": 4.116923809051514,
|
| 3992 |
+
"learning_rate": 1.6326799227683806e-05,
|
| 3993 |
+
"loss": 0.0602,
|
| 3994 |
+
"step": 523
|
| 3995 |
+
},
|
| 3996 |
+
{
|
| 3997 |
+
"epoch": 0.35453315290933696,
|
| 3998 |
+
"grad_norm": 3.4836175441741943,
|
| 3999 |
+
"learning_rate": 1.630848917754324e-05,
|
| 4000 |
+
"loss": 0.0969,
|
| 4001 |
+
"step": 524
|
| 4002 |
+
},
|
| 4003 |
+
{
|
| 4004 |
+
"epoch": 0.35520974289580515,
|
| 4005 |
+
"grad_norm": 3.9089815616607666,
|
| 4006 |
+
"learning_rate": 1.629014392910791e-05,
|
| 4007 |
+
"loss": 0.1509,
|
| 4008 |
+
"step": 525
|
| 4009 |
+
},
|
| 4010 |
+
{
|
| 4011 |
+
"epoch": 0.35588633288227334,
|
| 4012 |
+
"grad_norm": 3.1335699558258057,
|
| 4013 |
+
"learning_rate": 1.6271763584735373e-05,
|
| 4014 |
+
"loss": 0.1366,
|
| 4015 |
+
"step": 526
|
| 4016 |
+
},
|
| 4017 |
+
{
|
| 4018 |
+
"epoch": 0.3565629228687415,
|
| 4019 |
+
"grad_norm": 3.3636960983276367,
|
| 4020 |
+
"learning_rate": 1.625334824697898e-05,
|
| 4021 |
+
"loss": 0.1233,
|
| 4022 |
+
"step": 527
|
| 4023 |
+
},
|
| 4024 |
+
{
|
| 4025 |
+
"epoch": 0.3572395128552097,
|
| 4026 |
+
"grad_norm": 3.7551486492156982,
|
| 4027 |
+
"learning_rate": 1.6234898018587336e-05,
|
| 4028 |
+
"loss": 0.1541,
|
| 4029 |
+
"step": 528
|
| 4030 |
+
},
|
| 4031 |
+
{
|
| 4032 |
+
"epoch": 0.35791610284167796,
|
| 4033 |
+
"grad_norm": 6.929388046264648,
|
| 4034 |
+
"learning_rate": 1.6216413002503736e-05,
|
| 4035 |
+
"loss": 0.2057,
|
| 4036 |
+
"step": 529
|
| 4037 |
+
},
|
| 4038 |
+
{
|
| 4039 |
+
"epoch": 0.35859269282814615,
|
| 4040 |
+
"grad_norm": 4.0752763748168945,
|
| 4041 |
+
"learning_rate": 1.619789330186555e-05,
|
| 4042 |
+
"loss": 0.1008,
|
| 4043 |
+
"step": 530
|
| 4044 |
+
},
|
| 4045 |
+
{
|
| 4046 |
+
"epoch": 0.35926928281461434,
|
| 4047 |
+
"grad_norm": 3.1588234901428223,
|
| 4048 |
+
"learning_rate": 1.6179339020003685e-05,
|
| 4049 |
+
"loss": 0.1454,
|
| 4050 |
+
"step": 531
|
| 4051 |
+
},
|
| 4052 |
+
{
|
| 4053 |
+
"epoch": 0.35994587280108253,
|
| 4054 |
+
"grad_norm": 6.536987781524658,
|
| 4055 |
+
"learning_rate": 1.616075026044199e-05,
|
| 4056 |
+
"loss": 0.1461,
|
| 4057 |
+
"step": 532
|
| 4058 |
+
},
|
| 4059 |
+
{
|
| 4060 |
+
"epoch": 0.3606224627875507,
|
| 4061 |
+
"grad_norm": 3.1867458820343018,
|
| 4062 |
+
"learning_rate": 1.6142127126896682e-05,
|
| 4063 |
+
"loss": 0.182,
|
| 4064 |
+
"step": 533
|
| 4065 |
+
},
|
| 4066 |
+
{
|
| 4067 |
+
"epoch": 0.36129905277401897,
|
| 4068 |
+
"grad_norm": 3.9853105545043945,
|
| 4069 |
+
"learning_rate": 1.6123469723275766e-05,
|
| 4070 |
+
"loss": 0.1525,
|
| 4071 |
+
"step": 534
|
| 4072 |
+
},
|
| 4073 |
+
{
|
| 4074 |
+
"epoch": 0.36197564276048716,
|
| 4075 |
+
"grad_norm": 2.4770116806030273,
|
| 4076 |
+
"learning_rate": 1.6104778153678467e-05,
|
| 4077 |
+
"loss": 0.1789,
|
| 4078 |
+
"step": 535
|
| 4079 |
+
},
|
| 4080 |
+
{
|
| 4081 |
+
"epoch": 0.36265223274695535,
|
| 4082 |
+
"grad_norm": 4.895524024963379,
|
| 4083 |
+
"learning_rate": 1.6086052522394625e-05,
|
| 4084 |
+
"loss": 0.1909,
|
| 4085 |
+
"step": 536
|
| 4086 |
+
},
|
| 4087 |
+
{
|
| 4088 |
+
"epoch": 0.36332882273342354,
|
| 4089 |
+
"grad_norm": 7.819604873657227,
|
| 4090 |
+
"learning_rate": 1.6067292933904144e-05,
|
| 4091 |
+
"loss": 0.1793,
|
| 4092 |
+
"step": 537
|
| 4093 |
+
},
|
| 4094 |
+
{
|
| 4095 |
+
"epoch": 0.36400541271989173,
|
| 4096 |
+
"grad_norm": 5.251774311065674,
|
| 4097 |
+
"learning_rate": 1.6048499492876378e-05,
|
| 4098 |
+
"loss": 0.1829,
|
| 4099 |
+
"step": 538
|
| 4100 |
+
},
|
| 4101 |
+
{
|
| 4102 |
+
"epoch": 0.3646820027063599,
|
| 4103 |
+
"grad_norm": 5.605532646179199,
|
| 4104 |
+
"learning_rate": 1.602967230416957e-05,
|
| 4105 |
+
"loss": 0.273,
|
| 4106 |
+
"step": 539
|
| 4107 |
+
},
|
| 4108 |
+
{
|
| 4109 |
+
"epoch": 0.36535859269282817,
|
| 4110 |
+
"grad_norm": 3.302903175354004,
|
| 4111 |
+
"learning_rate": 1.6010811472830253e-05,
|
| 4112 |
+
"loss": 0.1608,
|
| 4113 |
+
"step": 540
|
| 4114 |
+
},
|
| 4115 |
+
{
|
| 4116 |
+
"epoch": 0.36535859269282817,
|
| 4117 |
+
"eval_accuracy": 0.8165137614678899,
|
| 4118 |
+
"eval_f1": 0.5412844036697247,
|
| 4119 |
+
"eval_loss": 0.4029388427734375,
|
| 4120 |
+
"eval_precision": 0.855072463768116,
|
| 4121 |
+
"eval_recall": 0.3959731543624161,
|
| 4122 |
+
"eval_runtime": 54.3008,
|
| 4123 |
+
"eval_samples_per_second": 5.488,
|
| 4124 |
+
"eval_steps_per_second": 0.184,
|
| 4125 |
+
"step": 540
|
| 4126 |
+
},
|
| 4127 |
+
{
|
| 4128 |
+
"epoch": 0.36603518267929636,
|
| 4129 |
+
"grad_norm": 3.057288408279419,
|
| 4130 |
+
"learning_rate": 1.5991917104092677e-05,
|
| 4131 |
+
"loss": 0.1671,
|
| 4132 |
+
"step": 541
|
| 4133 |
+
},
|
| 4134 |
+
{
|
| 4135 |
+
"epoch": 0.36671177266576455,
|
| 4136 |
+
"grad_norm": 4.837218761444092,
|
| 4137 |
+
"learning_rate": 1.5972989303378207e-05,
|
| 4138 |
+
"loss": 0.1425,
|
| 4139 |
+
"step": 542
|
| 4140 |
+
},
|
| 4141 |
+
{
|
| 4142 |
+
"epoch": 0.36738836265223274,
|
| 4143 |
+
"grad_norm": 2.922201633453369,
|
| 4144 |
+
"learning_rate": 1.595402817629475e-05,
|
| 4145 |
+
"loss": 0.2097,
|
| 4146 |
+
"step": 543
|
| 4147 |
+
},
|
| 4148 |
+
{
|
| 4149 |
+
"epoch": 0.3680649526387009,
|
| 4150 |
+
"grad_norm": 8.20699691772461,
|
| 4151 |
+
"learning_rate": 1.593503382863615e-05,
|
| 4152 |
+
"loss": 0.1657,
|
| 4153 |
+
"step": 544
|
| 4154 |
+
},
|
| 4155 |
+
{
|
| 4156 |
+
"epoch": 0.36874154262516917,
|
| 4157 |
+
"grad_norm": 3.043370246887207,
|
| 4158 |
+
"learning_rate": 1.591600636638161e-05,
|
| 4159 |
+
"loss": 0.1568,
|
| 4160 |
+
"step": 545
|
| 4161 |
+
},
|
| 4162 |
+
{
|
| 4163 |
+
"epoch": 0.36941813261163736,
|
| 4164 |
+
"grad_norm": 6.523357391357422,
|
| 4165 |
+
"learning_rate": 1.589694589569509e-05,
|
| 4166 |
+
"loss": 0.1299,
|
| 4167 |
+
"step": 546
|
| 4168 |
+
},
|
| 4169 |
+
{
|
| 4170 |
+
"epoch": 0.37009472259810555,
|
| 4171 |
+
"grad_norm": 3.4266302585601807,
|
| 4172 |
+
"learning_rate": 1.5877852522924733e-05,
|
| 4173 |
+
"loss": 0.1608,
|
| 4174 |
+
"step": 547
|
| 4175 |
+
},
|
| 4176 |
+
{
|
| 4177 |
+
"epoch": 0.37077131258457374,
|
| 4178 |
+
"grad_norm": 4.111809253692627,
|
| 4179 |
+
"learning_rate": 1.5858726354602248e-05,
|
| 4180 |
+
"loss": 0.1975,
|
| 4181 |
+
"step": 548
|
| 4182 |
+
},
|
| 4183 |
+
{
|
| 4184 |
+
"epoch": 0.37144790257104193,
|
| 4185 |
+
"grad_norm": 3.3651816844940186,
|
| 4186 |
+
"learning_rate": 1.5839567497442338e-05,
|
| 4187 |
+
"loss": 0.171,
|
| 4188 |
+
"step": 549
|
| 4189 |
+
},
|
| 4190 |
+
{
|
| 4191 |
+
"epoch": 0.3721244925575101,
|
| 4192 |
+
"grad_norm": 3.0030105113983154,
|
| 4193 |
+
"learning_rate": 1.5820376058342077e-05,
|
| 4194 |
+
"loss": 0.1365,
|
| 4195 |
+
"step": 550
|
| 4196 |
+
},
|
| 4197 |
+
{
|
| 4198 |
+
"epoch": 0.37280108254397837,
|
| 4199 |
+
"grad_norm": 3.5873923301696777,
|
| 4200 |
+
"learning_rate": 1.5801152144380353e-05,
|
| 4201 |
+
"loss": 0.1745,
|
| 4202 |
+
"step": 551
|
| 4203 |
+
},
|
| 4204 |
+
{
|
| 4205 |
+
"epoch": 0.37347767253044656,
|
| 4206 |
+
"grad_norm": 3.0994861125946045,
|
| 4207 |
+
"learning_rate": 1.578189586281723e-05,
|
| 4208 |
+
"loss": 0.1407,
|
| 4209 |
+
"step": 552
|
| 4210 |
+
},
|
| 4211 |
+
{
|
| 4212 |
+
"epoch": 0.37415426251691475,
|
| 4213 |
+
"grad_norm": 2.788184642791748,
|
| 4214 |
+
"learning_rate": 1.5762607321093368e-05,
|
| 4215 |
+
"loss": 0.134,
|
| 4216 |
+
"step": 553
|
| 4217 |
+
},
|
| 4218 |
+
{
|
| 4219 |
+
"epoch": 0.37483085250338294,
|
| 4220 |
+
"grad_norm": 2.5775389671325684,
|
| 4221 |
+
"learning_rate": 1.5743286626829437e-05,
|
| 4222 |
+
"loss": 0.1075,
|
| 4223 |
+
"step": 554
|
| 4224 |
+
},
|
| 4225 |
+
{
|
| 4226 |
+
"epoch": 0.37550744248985113,
|
| 4227 |
+
"grad_norm": 3.5334537029266357,
|
| 4228 |
+
"learning_rate": 1.5723933887825492e-05,
|
| 4229 |
+
"loss": 0.1165,
|
| 4230 |
+
"step": 555
|
| 4231 |
+
},
|
| 4232 |
+
{
|
| 4233 |
+
"epoch": 0.3761840324763194,
|
| 4234 |
+
"grad_norm": 3.544222116470337,
|
| 4235 |
+
"learning_rate": 1.5704549212060383e-05,
|
| 4236 |
+
"loss": 0.1739,
|
| 4237 |
+
"step": 556
|
| 4238 |
+
},
|
| 4239 |
+
{
|
| 4240 |
+
"epoch": 0.37686062246278756,
|
| 4241 |
+
"grad_norm": 3.89497709274292,
|
| 4242 |
+
"learning_rate": 1.568513270769115e-05,
|
| 4243 |
+
"loss": 0.1549,
|
| 4244 |
+
"step": 557
|
| 4245 |
+
},
|
| 4246 |
+
{
|
| 4247 |
+
"epoch": 0.37753721244925575,
|
| 4248 |
+
"grad_norm": 2.996244192123413,
|
| 4249 |
+
"learning_rate": 1.5665684483052425e-05,
|
| 4250 |
+
"loss": 0.1742,
|
| 4251 |
+
"step": 558
|
| 4252 |
+
},
|
| 4253 |
+
{
|
| 4254 |
+
"epoch": 0.37821380243572394,
|
| 4255 |
+
"grad_norm": 4.149686336517334,
|
| 4256 |
+
"learning_rate": 1.564620464665582e-05,
|
| 4257 |
+
"loss": 0.1655,
|
| 4258 |
+
"step": 559
|
| 4259 |
+
},
|
| 4260 |
+
{
|
| 4261 |
+
"epoch": 0.37889039242219213,
|
| 4262 |
+
"grad_norm": 3.9218225479125977,
|
| 4263 |
+
"learning_rate": 1.5626693307189334e-05,
|
| 4264 |
+
"loss": 0.1472,
|
| 4265 |
+
"step": 560
|
| 4266 |
+
},
|
| 4267 |
+
{
|
| 4268 |
+
"epoch": 0.37889039242219213,
|
| 4269 |
+
"eval_accuracy": 0.8064220183486238,
|
| 4270 |
+
"eval_f1": 0.5104408352668214,
|
| 4271 |
+
"eval_loss": 0.44546324014663696,
|
| 4272 |
+
"eval_precision": 0.8270676691729323,
|
| 4273 |
+
"eval_recall": 0.3691275167785235,
|
| 4274 |
+
"eval_runtime": 52.4956,
|
| 4275 |
+
"eval_samples_per_second": 5.677,
|
| 4276 |
+
"eval_steps_per_second": 0.19,
|
| 4277 |
+
"step": 560
|
| 4278 |
+
},
|
| 4279 |
+
{
|
| 4280 |
+
"epoch": 0.3795669824086603,
|
| 4281 |
+
"grad_norm": 7.877570629119873,
|
| 4282 |
+
"learning_rate": 1.560715057351673e-05,
|
| 4283 |
+
"loss": 0.0964,
|
| 4284 |
+
"step": 561
|
| 4285 |
+
},
|
| 4286 |
+
{
|
| 4287 |
+
"epoch": 0.38024357239512857,
|
| 4288 |
+
"grad_norm": 3.73523211479187,
|
| 4289 |
+
"learning_rate": 1.5587576554676927e-05,
|
| 4290 |
+
"loss": 0.1444,
|
| 4291 |
+
"step": 562
|
| 4292 |
+
},
|
| 4293 |
+
{
|
| 4294 |
+
"epoch": 0.38092016238159676,
|
| 4295 |
+
"grad_norm": 10.825215339660645,
|
| 4296 |
+
"learning_rate": 1.556797135988342e-05,
|
| 4297 |
+
"loss": 0.151,
|
| 4298 |
+
"step": 563
|
| 4299 |
+
},
|
| 4300 |
+
{
|
| 4301 |
+
"epoch": 0.38159675236806495,
|
| 4302 |
+
"grad_norm": 6.863844871520996,
|
| 4303 |
+
"learning_rate": 1.5548335098523634e-05,
|
| 4304 |
+
"loss": 0.2423,
|
| 4305 |
+
"step": 564
|
| 4306 |
+
},
|
| 4307 |
+
{
|
| 4308 |
+
"epoch": 0.38227334235453314,
|
| 4309 |
+
"grad_norm": 3.01707124710083,
|
| 4310 |
+
"learning_rate": 1.5528667880158338e-05,
|
| 4311 |
+
"loss": 0.1483,
|
| 4312 |
+
"step": 565
|
| 4313 |
+
},
|
| 4314 |
+
{
|
| 4315 |
+
"epoch": 0.38294993234100133,
|
| 4316 |
+
"grad_norm": 3.50577712059021,
|
| 4317 |
+
"learning_rate": 1.5508969814521026e-05,
|
| 4318 |
+
"loss": 0.1359,
|
| 4319 |
+
"step": 566
|
| 4320 |
+
},
|
| 4321 |
+
{
|
| 4322 |
+
"epoch": 0.3836265223274696,
|
| 4323 |
+
"grad_norm": 3.558225631713867,
|
| 4324 |
+
"learning_rate": 1.5489241011517303e-05,
|
| 4325 |
+
"loss": 0.0951,
|
| 4326 |
+
"step": 567
|
| 4327 |
+
},
|
| 4328 |
+
{
|
| 4329 |
+
"epoch": 0.38430311231393777,
|
| 4330 |
+
"grad_norm": 7.069665431976318,
|
| 4331 |
+
"learning_rate": 1.5469481581224274e-05,
|
| 4332 |
+
"loss": 0.0979,
|
| 4333 |
+
"step": 568
|
| 4334 |
+
},
|
| 4335 |
+
{
|
| 4336 |
+
"epoch": 0.38497970230040596,
|
| 4337 |
+
"grad_norm": 4.208998680114746,
|
| 4338 |
+
"learning_rate": 1.5449691633889924e-05,
|
| 4339 |
+
"loss": 0.1451,
|
| 4340 |
+
"step": 569
|
| 4341 |
+
},
|
| 4342 |
+
{
|
| 4343 |
+
"epoch": 0.38565629228687415,
|
| 4344 |
+
"grad_norm": 3.042346477508545,
|
| 4345 |
+
"learning_rate": 1.5429871279932514e-05,
|
| 4346 |
+
"loss": 0.1555,
|
| 4347 |
+
"step": 570
|
| 4348 |
+
},
|
| 4349 |
+
{
|
| 4350 |
+
"epoch": 0.38633288227334234,
|
| 4351 |
+
"grad_norm": 6.0393595695495605,
|
| 4352 |
+
"learning_rate": 1.5410020629939966e-05,
|
| 4353 |
+
"loss": 0.1965,
|
| 4354 |
+
"step": 571
|
| 4355 |
+
},
|
| 4356 |
+
{
|
| 4357 |
+
"epoch": 0.3870094722598105,
|
| 4358 |
+
"grad_norm": 3.022724151611328,
|
| 4359 |
+
"learning_rate": 1.5390139794669225e-05,
|
| 4360 |
+
"loss": 0.1219,
|
| 4361 |
+
"step": 572
|
| 4362 |
+
},
|
| 4363 |
+
{
|
| 4364 |
+
"epoch": 0.3876860622462788,
|
| 4365 |
+
"grad_norm": 2.8513102531433105,
|
| 4366 |
+
"learning_rate": 1.5370228885045662e-05,
|
| 4367 |
+
"loss": 0.1634,
|
| 4368 |
+
"step": 573
|
| 4369 |
+
},
|
| 4370 |
+
{
|
| 4371 |
+
"epoch": 0.38836265223274696,
|
| 4372 |
+
"grad_norm": 3.42635440826416,
|
| 4373 |
+
"learning_rate": 1.535028801216245e-05,
|
| 4374 |
+
"loss": 0.1874,
|
| 4375 |
+
"step": 574
|
| 4376 |
+
},
|
| 4377 |
+
{
|
| 4378 |
+
"epoch": 0.38903924221921515,
|
| 4379 |
+
"grad_norm": 6.154781341552734,
|
| 4380 |
+
"learning_rate": 1.533031728727994e-05,
|
| 4381 |
+
"loss": 0.1979,
|
| 4382 |
+
"step": 575
|
| 4383 |
+
},
|
| 4384 |
+
{
|
| 4385 |
+
"epoch": 0.38971583220568334,
|
| 4386 |
+
"grad_norm": 4.227107524871826,
|
| 4387 |
+
"learning_rate": 1.531031682182504e-05,
|
| 4388 |
+
"loss": 0.2125,
|
| 4389 |
+
"step": 576
|
| 4390 |
+
},
|
| 4391 |
+
{
|
| 4392 |
+
"epoch": 0.39039242219215153,
|
| 4393 |
+
"grad_norm": 3.389040946960449,
|
| 4394 |
+
"learning_rate": 1.5290286727390604e-05,
|
| 4395 |
+
"loss": 0.1554,
|
| 4396 |
+
"step": 577
|
| 4397 |
+
},
|
| 4398 |
+
{
|
| 4399 |
+
"epoch": 0.3910690121786198,
|
| 4400 |
+
"grad_norm": 3.34859299659729,
|
| 4401 |
+
"learning_rate": 1.527022711573479e-05,
|
| 4402 |
+
"loss": 0.1859,
|
| 4403 |
+
"step": 578
|
| 4404 |
+
},
|
| 4405 |
+
{
|
| 4406 |
+
"epoch": 0.39174560216508797,
|
| 4407 |
+
"grad_norm": 2.8738038539886475,
|
| 4408 |
+
"learning_rate": 1.5250138098780456e-05,
|
| 4409 |
+
"loss": 0.105,
|
| 4410 |
+
"step": 579
|
| 4411 |
+
},
|
| 4412 |
+
{
|
| 4413 |
+
"epoch": 0.39242219215155616,
|
| 4414 |
+
"grad_norm": 3.0525081157684326,
|
| 4415 |
+
"learning_rate": 1.5230019788614527e-05,
|
| 4416 |
+
"loss": 0.1437,
|
| 4417 |
+
"step": 580
|
| 4418 |
+
},
|
| 4419 |
+
{
|
| 4420 |
+
"epoch": 0.39242219215155616,
|
| 4421 |
+
"eval_accuracy": 0.8036697247706422,
|
| 4422 |
+
"eval_f1": 0.4928909952606635,
|
| 4423 |
+
"eval_loss": 0.4023875296115875,
|
| 4424 |
+
"eval_precision": 0.8387096774193549,
|
| 4425 |
+
"eval_recall": 0.348993288590604,
|
| 4426 |
+
"eval_runtime": 53.2583,
|
| 4427 |
+
"eval_samples_per_second": 5.595,
|
| 4428 |
+
"eval_steps_per_second": 0.188,
|
| 4429 |
+
"step": 580
|
| 4430 |
+
},
|
| 4431 |
+
{
|
| 4432 |
+
"epoch": 0.39309878213802435,
|
| 4433 |
+
"grad_norm": 2.1808993816375732,
|
| 4434 |
+
"learning_rate": 1.5209872297487365e-05,
|
| 4435 |
+
"loss": 0.158,
|
| 4436 |
+
"step": 581
|
| 4437 |
+
},
|
| 4438 |
+
{
|
| 4439 |
+
"epoch": 0.39377537212449254,
|
| 4440 |
+
"grad_norm": 2.2843339443206787,
|
| 4441 |
+
"learning_rate": 1.5189695737812153e-05,
|
| 4442 |
+
"loss": 0.0944,
|
| 4443 |
+
"step": 582
|
| 4444 |
+
},
|
| 4445 |
+
{
|
| 4446 |
+
"epoch": 0.3944519621109608,
|
| 4447 |
+
"grad_norm": 3.0277621746063232,
|
| 4448 |
+
"learning_rate": 1.5169490222164255e-05,
|
| 4449 |
+
"loss": 0.1253,
|
| 4450 |
+
"step": 583
|
| 4451 |
+
},
|
| 4452 |
+
{
|
| 4453 |
+
"epoch": 0.395128552097429,
|
| 4454 |
+
"grad_norm": 3.6869609355926514,
|
| 4455 |
+
"learning_rate": 1.5149255863280607e-05,
|
| 4456 |
+
"loss": 0.1293,
|
| 4457 |
+
"step": 584
|
| 4458 |
+
},
|
| 4459 |
+
{
|
| 4460 |
+
"epoch": 0.39580514208389717,
|
| 4461 |
+
"grad_norm": 2.9757912158966064,
|
| 4462 |
+
"learning_rate": 1.5128992774059063e-05,
|
| 4463 |
+
"loss": 0.1344,
|
| 4464 |
+
"step": 585
|
| 4465 |
+
},
|
| 4466 |
+
{
|
| 4467 |
+
"epoch": 0.39648173207036536,
|
| 4468 |
+
"grad_norm": 3.4651436805725098,
|
| 4469 |
+
"learning_rate": 1.5108701067557787e-05,
|
| 4470 |
+
"loss": 0.1131,
|
| 4471 |
+
"step": 586
|
| 4472 |
+
},
|
| 4473 |
+
{
|
| 4474 |
+
"epoch": 0.39715832205683355,
|
| 4475 |
+
"grad_norm": 3.6160385608673096,
|
| 4476 |
+
"learning_rate": 1.5088380856994608e-05,
|
| 4477 |
+
"loss": 0.1679,
|
| 4478 |
+
"step": 587
|
| 4479 |
+
},
|
| 4480 |
+
{
|
| 4481 |
+
"epoch": 0.39783491204330174,
|
| 4482 |
+
"grad_norm": 2.459801197052002,
|
| 4483 |
+
"learning_rate": 1.50680322557464e-05,
|
| 4484 |
+
"loss": 0.0886,
|
| 4485 |
+
"step": 588
|
| 4486 |
+
},
|
| 4487 |
+
{
|
| 4488 |
+
"epoch": 0.39851150202977,
|
| 4489 |
+
"grad_norm": 3.2416129112243652,
|
| 4490 |
+
"learning_rate": 1.504765537734844e-05,
|
| 4491 |
+
"loss": 0.1534,
|
| 4492 |
+
"step": 589
|
| 4493 |
+
},
|
| 4494 |
+
{
|
| 4495 |
+
"epoch": 0.39918809201623817,
|
| 4496 |
+
"grad_norm": 2.9892609119415283,
|
| 4497 |
+
"learning_rate": 1.5027250335493771e-05,
|
| 4498 |
+
"loss": 0.0943,
|
| 4499 |
+
"step": 590
|
| 4500 |
+
},
|
| 4501 |
+
{
|
| 4502 |
+
"epoch": 0.39986468200270636,
|
| 4503 |
+
"grad_norm": 5.113293647766113,
|
| 4504 |
+
"learning_rate": 1.5006817244032589e-05,
|
| 4505 |
+
"loss": 0.1717,
|
| 4506 |
+
"step": 591
|
| 4507 |
+
},
|
| 4508 |
+
{
|
| 4509 |
+
"epoch": 0.40054127198917455,
|
| 4510 |
+
"grad_norm": 6.7918195724487305,
|
| 4511 |
+
"learning_rate": 1.4986356216971583e-05,
|
| 4512 |
+
"loss": 0.1747,
|
| 4513 |
+
"step": 592
|
| 4514 |
+
},
|
| 4515 |
+
{
|
| 4516 |
+
"epoch": 0.40121786197564274,
|
| 4517 |
+
"grad_norm": 3.726599931716919,
|
| 4518 |
+
"learning_rate": 1.4965867368473308e-05,
|
| 4519 |
+
"loss": 0.1416,
|
| 4520 |
+
"step": 593
|
| 4521 |
+
},
|
| 4522 |
+
{
|
| 4523 |
+
"epoch": 0.401894451962111,
|
| 4524 |
+
"grad_norm": 6.388960361480713,
|
| 4525 |
+
"learning_rate": 1.4945350812855555e-05,
|
| 4526 |
+
"loss": 0.1946,
|
| 4527 |
+
"step": 594
|
| 4528 |
+
},
|
| 4529 |
+
{
|
| 4530 |
+
"epoch": 0.4025710419485792,
|
| 4531 |
+
"grad_norm": 3.8928306102752686,
|
| 4532 |
+
"learning_rate": 1.4924806664590702e-05,
|
| 4533 |
+
"loss": 0.1622,
|
| 4534 |
+
"step": 595
|
| 4535 |
+
},
|
| 4536 |
+
{
|
| 4537 |
+
"epoch": 0.40324763193504737,
|
| 4538 |
+
"grad_norm": 3.4860315322875977,
|
| 4539 |
+
"learning_rate": 1.4904235038305084e-05,
|
| 4540 |
+
"loss": 0.1549,
|
| 4541 |
+
"step": 596
|
| 4542 |
+
},
|
| 4543 |
+
{
|
| 4544 |
+
"epoch": 0.40392422192151556,
|
| 4545 |
+
"grad_norm": 4.641502857208252,
|
| 4546 |
+
"learning_rate": 1.4883636048778347e-05,
|
| 4547 |
+
"loss": 0.1789,
|
| 4548 |
+
"step": 597
|
| 4549 |
+
},
|
| 4550 |
+
{
|
| 4551 |
+
"epoch": 0.40460081190798375,
|
| 4552 |
+
"grad_norm": 4.543725967407227,
|
| 4553 |
+
"learning_rate": 1.4863009810942814e-05,
|
| 4554 |
+
"loss": 0.1651,
|
| 4555 |
+
"step": 598
|
| 4556 |
+
},
|
| 4557 |
+
{
|
| 4558 |
+
"epoch": 0.40527740189445194,
|
| 4559 |
+
"grad_norm": 10.844779014587402,
|
| 4560 |
+
"learning_rate": 1.4842356439882841e-05,
|
| 4561 |
+
"loss": 0.2244,
|
| 4562 |
+
"step": 599
|
| 4563 |
+
},
|
| 4564 |
+
{
|
| 4565 |
+
"epoch": 0.4059539918809202,
|
| 4566 |
+
"grad_norm": 7.078522205352783,
|
| 4567 |
+
"learning_rate": 1.4821676050834166e-05,
|
| 4568 |
+
"loss": 0.2055,
|
| 4569 |
+
"step": 600
|
| 4570 |
+
},
|
| 4571 |
+
{
|
| 4572 |
+
"epoch": 0.4059539918809202,
|
| 4573 |
+
"eval_accuracy": 0.8229357798165138,
|
| 4574 |
+
"eval_f1": 0.5758241758241758,
|
| 4575 |
+
"eval_loss": 0.40693244338035583,
|
| 4576 |
+
"eval_precision": 0.8343949044585988,
|
| 4577 |
+
"eval_recall": 0.4395973154362416,
|
| 4578 |
+
"eval_runtime": 53.515,
|
| 4579 |
+
"eval_samples_per_second": 5.569,
|
| 4580 |
+
"eval_steps_per_second": 0.187,
|
| 4581 |
+
"step": 600
|
| 4582 |
}
|
| 4583 |
],
|
| 4584 |
"logging_steps": 1,
|
|
|
|
| 4598 |
"attributes": {}
|
| 4599 |
}
|
| 4600 |
},
|
| 4601 |
+
"total_flos": 1.81775286403072e+17,
|
| 4602 |
"train_batch_size": 8,
|
| 4603 |
"trial_name": null,
|
| 4604 |
"trial_params": null
|