Upload folder using huggingface_hub
Browse files- optimizer.pt +1 -1
- pytorch_model.bin +1 -1
- rng_state.pth +1 -1
- scheduler.pt +1 -1
- trainer_state.json +1402 -2
optimizer.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 1871270621
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:851551160902c1b40a2a1019a2d0b42ed6ad9135bf225ff47a98e00e3f971fb6
|
| 3 |
size 1871270621
|
pytorch_model.bin
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 950390010
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:54e00f0f811a1a92cb465ac006c6caabae5add3c2f6356fd8daebd03082a8fad
|
| 3 |
size 950390010
|
rng_state.pth
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 14244
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:44426bba1bda8a34d9528413baa4cda8985744fa7caff63fc73da84505a52f33
|
| 3 |
size 14244
|
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 sha256:0125f1f987952e9e21fcff9eb1765601ced4bad40bf35aa471e13ce184aba0c3
|
| 3 |
size 1064
|
trainer_state.json
CHANGED
|
@@ -1,9 +1,9 @@
|
|
| 1 |
{
|
| 2 |
"best_metric": null,
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
-
"epoch": 0.
|
| 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,
|
|
@@ -11207,6 +11207,1406 @@
|
|
| 11207 |
"learning_rate": 3.2000000000000005e-05,
|
| 11208 |
"loss": 174.3226,
|
| 11209 |
"step": 16000
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 11210 |
}
|
| 11211 |
],
|
| 11212 |
"logging_steps": 10,
|
|
|
|
| 1 |
{
|
| 2 |
"best_metric": null,
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 0.07272227766173636,
|
| 5 |
"eval_steps": 500,
|
| 6 |
+
"global_step": 18000,
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
|
|
| 11207 |
"learning_rate": 3.2000000000000005e-05,
|
| 11208 |
"loss": 174.3226,
|
| 11209 |
"step": 16000
|
| 11210 |
+
},
|
| 11211 |
+
{
|
| 11212 |
+
"epoch": 0.06468242585357774,
|
| 11213 |
+
"grad_norm": 12306.052734375,
|
| 11214 |
+
"learning_rate": 3.202e-05,
|
| 11215 |
+
"loss": 203.855,
|
| 11216 |
+
"step": 16010
|
| 11217 |
+
},
|
| 11218 |
+
{
|
| 11219 |
+
"epoch": 0.06472282711894536,
|
| 11220 |
+
"grad_norm": 2370.59765625,
|
| 11221 |
+
"learning_rate": 3.2040000000000005e-05,
|
| 11222 |
+
"loss": 144.5134,
|
| 11223 |
+
"step": 16020
|
| 11224 |
+
},
|
| 11225 |
+
{
|
| 11226 |
+
"epoch": 0.06476322838431299,
|
| 11227 |
+
"grad_norm": 1784.508056640625,
|
| 11228 |
+
"learning_rate": 3.206e-05,
|
| 11229 |
+
"loss": 159.992,
|
| 11230 |
+
"step": 16030
|
| 11231 |
+
},
|
| 11232 |
+
{
|
| 11233 |
+
"epoch": 0.06480362964968063,
|
| 11234 |
+
"grad_norm": 894.4352416992188,
|
| 11235 |
+
"learning_rate": 3.208e-05,
|
| 11236 |
+
"loss": 162.522,
|
| 11237 |
+
"step": 16040
|
| 11238 |
+
},
|
| 11239 |
+
{
|
| 11240 |
+
"epoch": 0.06484403091504826,
|
| 11241 |
+
"grad_norm": 1397.84814453125,
|
| 11242 |
+
"learning_rate": 3.21e-05,
|
| 11243 |
+
"loss": 113.1159,
|
| 11244 |
+
"step": 16050
|
| 11245 |
+
},
|
| 11246 |
+
{
|
| 11247 |
+
"epoch": 0.0648844321804159,
|
| 11248 |
+
"grad_norm": 2406.892578125,
|
| 11249 |
+
"learning_rate": 3.212e-05,
|
| 11250 |
+
"loss": 104.4846,
|
| 11251 |
+
"step": 16060
|
| 11252 |
+
},
|
| 11253 |
+
{
|
| 11254 |
+
"epoch": 0.06492483344578352,
|
| 11255 |
+
"grad_norm": 1434.818115234375,
|
| 11256 |
+
"learning_rate": 3.214e-05,
|
| 11257 |
+
"loss": 142.6461,
|
| 11258 |
+
"step": 16070
|
| 11259 |
+
},
|
| 11260 |
+
{
|
| 11261 |
+
"epoch": 0.06496523471115115,
|
| 11262 |
+
"grad_norm": 3077.646240234375,
|
| 11263 |
+
"learning_rate": 3.2160000000000004e-05,
|
| 11264 |
+
"loss": 101.5211,
|
| 11265 |
+
"step": 16080
|
| 11266 |
+
},
|
| 11267 |
+
{
|
| 11268 |
+
"epoch": 0.06500563597651879,
|
| 11269 |
+
"grad_norm": 1358.7418212890625,
|
| 11270 |
+
"learning_rate": 3.218e-05,
|
| 11271 |
+
"loss": 129.6278,
|
| 11272 |
+
"step": 16090
|
| 11273 |
+
},
|
| 11274 |
+
{
|
| 11275 |
+
"epoch": 0.06504603724188641,
|
| 11276 |
+
"grad_norm": 11825.310546875,
|
| 11277 |
+
"learning_rate": 3.2200000000000003e-05,
|
| 11278 |
+
"loss": 145.8424,
|
| 11279 |
+
"step": 16100
|
| 11280 |
+
},
|
| 11281 |
+
{
|
| 11282 |
+
"epoch": 0.06508643850725404,
|
| 11283 |
+
"grad_norm": 1621.6112060546875,
|
| 11284 |
+
"learning_rate": 3.222e-05,
|
| 11285 |
+
"loss": 117.4145,
|
| 11286 |
+
"step": 16110
|
| 11287 |
+
},
|
| 11288 |
+
{
|
| 11289 |
+
"epoch": 0.06512683977262168,
|
| 11290 |
+
"grad_norm": 1896.356689453125,
|
| 11291 |
+
"learning_rate": 3.224e-05,
|
| 11292 |
+
"loss": 103.7541,
|
| 11293 |
+
"step": 16120
|
| 11294 |
+
},
|
| 11295 |
+
{
|
| 11296 |
+
"epoch": 0.0651672410379893,
|
| 11297 |
+
"grad_norm": 8822.8671875,
|
| 11298 |
+
"learning_rate": 3.226e-05,
|
| 11299 |
+
"loss": 174.6907,
|
| 11300 |
+
"step": 16130
|
| 11301 |
+
},
|
| 11302 |
+
{
|
| 11303 |
+
"epoch": 0.06520764230335695,
|
| 11304 |
+
"grad_norm": 1795.896728515625,
|
| 11305 |
+
"learning_rate": 3.2279999999999996e-05,
|
| 11306 |
+
"loss": 187.9359,
|
| 11307 |
+
"step": 16140
|
| 11308 |
+
},
|
| 11309 |
+
{
|
| 11310 |
+
"epoch": 0.06524804356872457,
|
| 11311 |
+
"grad_norm": 870.3344116210938,
|
| 11312 |
+
"learning_rate": 3.2300000000000006e-05,
|
| 11313 |
+
"loss": 113.2782,
|
| 11314 |
+
"step": 16150
|
| 11315 |
+
},
|
| 11316 |
+
{
|
| 11317 |
+
"epoch": 0.0652884448340922,
|
| 11318 |
+
"grad_norm": 1710.718505859375,
|
| 11319 |
+
"learning_rate": 3.232e-05,
|
| 11320 |
+
"loss": 114.0318,
|
| 11321 |
+
"step": 16160
|
| 11322 |
+
},
|
| 11323 |
+
{
|
| 11324 |
+
"epoch": 0.06532884609945984,
|
| 11325 |
+
"grad_norm": 879.309326171875,
|
| 11326 |
+
"learning_rate": 3.2340000000000005e-05,
|
| 11327 |
+
"loss": 159.0129,
|
| 11328 |
+
"step": 16170
|
| 11329 |
+
},
|
| 11330 |
+
{
|
| 11331 |
+
"epoch": 0.06536924736482747,
|
| 11332 |
+
"grad_norm": 1179.8787841796875,
|
| 11333 |
+
"learning_rate": 3.236e-05,
|
| 11334 |
+
"loss": 108.0792,
|
| 11335 |
+
"step": 16180
|
| 11336 |
+
},
|
| 11337 |
+
{
|
| 11338 |
+
"epoch": 0.06540964863019509,
|
| 11339 |
+
"grad_norm": 1053.493896484375,
|
| 11340 |
+
"learning_rate": 3.238e-05,
|
| 11341 |
+
"loss": 162.0999,
|
| 11342 |
+
"step": 16190
|
| 11343 |
+
},
|
| 11344 |
+
{
|
| 11345 |
+
"epoch": 0.06545004989556273,
|
| 11346 |
+
"grad_norm": 1169.240478515625,
|
| 11347 |
+
"learning_rate": 3.24e-05,
|
| 11348 |
+
"loss": 117.8543,
|
| 11349 |
+
"step": 16200
|
| 11350 |
+
},
|
| 11351 |
+
{
|
| 11352 |
+
"epoch": 0.06549045116093036,
|
| 11353 |
+
"grad_norm": 1077.4576416015625,
|
| 11354 |
+
"learning_rate": 3.242e-05,
|
| 11355 |
+
"loss": 161.1472,
|
| 11356 |
+
"step": 16210
|
| 11357 |
+
},
|
| 11358 |
+
{
|
| 11359 |
+
"epoch": 0.065530852426298,
|
| 11360 |
+
"grad_norm": 827.2643432617188,
|
| 11361 |
+
"learning_rate": 3.244e-05,
|
| 11362 |
+
"loss": 122.5282,
|
| 11363 |
+
"step": 16220
|
| 11364 |
+
},
|
| 11365 |
+
{
|
| 11366 |
+
"epoch": 0.06557125369166562,
|
| 11367 |
+
"grad_norm": 1683.487060546875,
|
| 11368 |
+
"learning_rate": 3.2460000000000004e-05,
|
| 11369 |
+
"loss": 163.201,
|
| 11370 |
+
"step": 16230
|
| 11371 |
+
},
|
| 11372 |
+
{
|
| 11373 |
+
"epoch": 0.06561165495703325,
|
| 11374 |
+
"grad_norm": 1506.306884765625,
|
| 11375 |
+
"learning_rate": 3.248e-05,
|
| 11376 |
+
"loss": 123.3369,
|
| 11377 |
+
"step": 16240
|
| 11378 |
+
},
|
| 11379 |
+
{
|
| 11380 |
+
"epoch": 0.06565205622240089,
|
| 11381 |
+
"grad_norm": 410.9195861816406,
|
| 11382 |
+
"learning_rate": 3.2500000000000004e-05,
|
| 11383 |
+
"loss": 102.7201,
|
| 11384 |
+
"step": 16250
|
| 11385 |
+
},
|
| 11386 |
+
{
|
| 11387 |
+
"epoch": 0.06569245748776852,
|
| 11388 |
+
"grad_norm": 484.0295104980469,
|
| 11389 |
+
"learning_rate": 3.252e-05,
|
| 11390 |
+
"loss": 191.6438,
|
| 11391 |
+
"step": 16260
|
| 11392 |
+
},
|
| 11393 |
+
{
|
| 11394 |
+
"epoch": 0.06573285875313614,
|
| 11395 |
+
"grad_norm": 649.029541015625,
|
| 11396 |
+
"learning_rate": 3.2540000000000004e-05,
|
| 11397 |
+
"loss": 139.1572,
|
| 11398 |
+
"step": 16270
|
| 11399 |
+
},
|
| 11400 |
+
{
|
| 11401 |
+
"epoch": 0.06577326001850378,
|
| 11402 |
+
"grad_norm": 2727.947509765625,
|
| 11403 |
+
"learning_rate": 3.256e-05,
|
| 11404 |
+
"loss": 198.9751,
|
| 11405 |
+
"step": 16280
|
| 11406 |
+
},
|
| 11407 |
+
{
|
| 11408 |
+
"epoch": 0.06581366128387141,
|
| 11409 |
+
"grad_norm": 879.7364501953125,
|
| 11410 |
+
"learning_rate": 3.2579999999999996e-05,
|
| 11411 |
+
"loss": 119.9532,
|
| 11412 |
+
"step": 16290
|
| 11413 |
+
},
|
| 11414 |
+
{
|
| 11415 |
+
"epoch": 0.06585406254923905,
|
| 11416 |
+
"grad_norm": 2567.528564453125,
|
| 11417 |
+
"learning_rate": 3.26e-05,
|
| 11418 |
+
"loss": 182.2825,
|
| 11419 |
+
"step": 16300
|
| 11420 |
+
},
|
| 11421 |
+
{
|
| 11422 |
+
"epoch": 0.06589446381460667,
|
| 11423 |
+
"grad_norm": 1746.7139892578125,
|
| 11424 |
+
"learning_rate": 3.262e-05,
|
| 11425 |
+
"loss": 109.7434,
|
| 11426 |
+
"step": 16310
|
| 11427 |
+
},
|
| 11428 |
+
{
|
| 11429 |
+
"epoch": 0.0659348650799743,
|
| 11430 |
+
"grad_norm": 1511.91845703125,
|
| 11431 |
+
"learning_rate": 3.2640000000000006e-05,
|
| 11432 |
+
"loss": 117.7299,
|
| 11433 |
+
"step": 16320
|
| 11434 |
+
},
|
| 11435 |
+
{
|
| 11436 |
+
"epoch": 0.06597526634534194,
|
| 11437 |
+
"grad_norm": 626.0679321289062,
|
| 11438 |
+
"learning_rate": 3.266e-05,
|
| 11439 |
+
"loss": 154.5593,
|
| 11440 |
+
"step": 16330
|
| 11441 |
+
},
|
| 11442 |
+
{
|
| 11443 |
+
"epoch": 0.06601566761070957,
|
| 11444 |
+
"grad_norm": 1950.873779296875,
|
| 11445 |
+
"learning_rate": 3.268e-05,
|
| 11446 |
+
"loss": 134.659,
|
| 11447 |
+
"step": 16340
|
| 11448 |
+
},
|
| 11449 |
+
{
|
| 11450 |
+
"epoch": 0.06605606887607719,
|
| 11451 |
+
"grad_norm": 2815.42919921875,
|
| 11452 |
+
"learning_rate": 3.27e-05,
|
| 11453 |
+
"loss": 189.5417,
|
| 11454 |
+
"step": 16350
|
| 11455 |
+
},
|
| 11456 |
+
{
|
| 11457 |
+
"epoch": 0.06609647014144483,
|
| 11458 |
+
"grad_norm": 791.9197387695312,
|
| 11459 |
+
"learning_rate": 3.272e-05,
|
| 11460 |
+
"loss": 104.6571,
|
| 11461 |
+
"step": 16360
|
| 11462 |
+
},
|
| 11463 |
+
{
|
| 11464 |
+
"epoch": 0.06613687140681246,
|
| 11465 |
+
"grad_norm": 961.1111450195312,
|
| 11466 |
+
"learning_rate": 3.274e-05,
|
| 11467 |
+
"loss": 110.3847,
|
| 11468 |
+
"step": 16370
|
| 11469 |
+
},
|
| 11470 |
+
{
|
| 11471 |
+
"epoch": 0.0661772726721801,
|
| 11472 |
+
"grad_norm": 1224.700927734375,
|
| 11473 |
+
"learning_rate": 3.2760000000000005e-05,
|
| 11474 |
+
"loss": 129.7902,
|
| 11475 |
+
"step": 16380
|
| 11476 |
+
},
|
| 11477 |
+
{
|
| 11478 |
+
"epoch": 0.06621767393754772,
|
| 11479 |
+
"grad_norm": 892.2734985351562,
|
| 11480 |
+
"learning_rate": 3.278e-05,
|
| 11481 |
+
"loss": 131.0276,
|
| 11482 |
+
"step": 16390
|
| 11483 |
+
},
|
| 11484 |
+
{
|
| 11485 |
+
"epoch": 0.06625807520291535,
|
| 11486 |
+
"grad_norm": 2211.9521484375,
|
| 11487 |
+
"learning_rate": 3.2800000000000004e-05,
|
| 11488 |
+
"loss": 102.1123,
|
| 11489 |
+
"step": 16400
|
| 11490 |
+
},
|
| 11491 |
+
{
|
| 11492 |
+
"epoch": 0.06629847646828299,
|
| 11493 |
+
"grad_norm": 2223.035888671875,
|
| 11494 |
+
"learning_rate": 3.282e-05,
|
| 11495 |
+
"loss": 130.2351,
|
| 11496 |
+
"step": 16410
|
| 11497 |
+
},
|
| 11498 |
+
{
|
| 11499 |
+
"epoch": 0.06633887773365062,
|
| 11500 |
+
"grad_norm": 896.9553833007812,
|
| 11501 |
+
"learning_rate": 3.2840000000000004e-05,
|
| 11502 |
+
"loss": 113.7288,
|
| 11503 |
+
"step": 16420
|
| 11504 |
+
},
|
| 11505 |
+
{
|
| 11506 |
+
"epoch": 0.06637927899901824,
|
| 11507 |
+
"grad_norm": 0.0,
|
| 11508 |
+
"learning_rate": 3.286e-05,
|
| 11509 |
+
"loss": 70.063,
|
| 11510 |
+
"step": 16430
|
| 11511 |
+
},
|
| 11512 |
+
{
|
| 11513 |
+
"epoch": 0.06641968026438588,
|
| 11514 |
+
"grad_norm": 665.3837890625,
|
| 11515 |
+
"learning_rate": 3.288e-05,
|
| 11516 |
+
"loss": 102.8864,
|
| 11517 |
+
"step": 16440
|
| 11518 |
+
},
|
| 11519 |
+
{
|
| 11520 |
+
"epoch": 0.06646008152975351,
|
| 11521 |
+
"grad_norm": 529.5450439453125,
|
| 11522 |
+
"learning_rate": 3.29e-05,
|
| 11523 |
+
"loss": 107.0171,
|
| 11524 |
+
"step": 16450
|
| 11525 |
+
},
|
| 11526 |
+
{
|
| 11527 |
+
"epoch": 0.06650048279512115,
|
| 11528 |
+
"grad_norm": 445.30950927734375,
|
| 11529 |
+
"learning_rate": 3.292e-05,
|
| 11530 |
+
"loss": 118.4387,
|
| 11531 |
+
"step": 16460
|
| 11532 |
+
},
|
| 11533 |
+
{
|
| 11534 |
+
"epoch": 0.06654088406048878,
|
| 11535 |
+
"grad_norm": 1200.6650390625,
|
| 11536 |
+
"learning_rate": 3.2940000000000006e-05,
|
| 11537 |
+
"loss": 115.6011,
|
| 11538 |
+
"step": 16470
|
| 11539 |
+
},
|
| 11540 |
+
{
|
| 11541 |
+
"epoch": 0.0665812853258564,
|
| 11542 |
+
"grad_norm": 997.8995971679688,
|
| 11543 |
+
"learning_rate": 3.296e-05,
|
| 11544 |
+
"loss": 101.1121,
|
| 11545 |
+
"step": 16480
|
| 11546 |
+
},
|
| 11547 |
+
{
|
| 11548 |
+
"epoch": 0.06662168659122404,
|
| 11549 |
+
"grad_norm": 1181.9188232421875,
|
| 11550 |
+
"learning_rate": 3.298e-05,
|
| 11551 |
+
"loss": 94.8608,
|
| 11552 |
+
"step": 16490
|
| 11553 |
+
},
|
| 11554 |
+
{
|
| 11555 |
+
"epoch": 0.06666208785659167,
|
| 11556 |
+
"grad_norm": 816.2405395507812,
|
| 11557 |
+
"learning_rate": 3.3e-05,
|
| 11558 |
+
"loss": 125.1284,
|
| 11559 |
+
"step": 16500
|
| 11560 |
+
},
|
| 11561 |
+
{
|
| 11562 |
+
"epoch": 0.0667024891219593,
|
| 11563 |
+
"grad_norm": 703.4663696289062,
|
| 11564 |
+
"learning_rate": 3.302e-05,
|
| 11565 |
+
"loss": 123.1175,
|
| 11566 |
+
"step": 16510
|
| 11567 |
+
},
|
| 11568 |
+
{
|
| 11569 |
+
"epoch": 0.06674289038732693,
|
| 11570 |
+
"grad_norm": 671.1204833984375,
|
| 11571 |
+
"learning_rate": 3.304e-05,
|
| 11572 |
+
"loss": 133.0177,
|
| 11573 |
+
"step": 16520
|
| 11574 |
+
},
|
| 11575 |
+
{
|
| 11576 |
+
"epoch": 0.06678329165269456,
|
| 11577 |
+
"grad_norm": 635.96728515625,
|
| 11578 |
+
"learning_rate": 3.3060000000000005e-05,
|
| 11579 |
+
"loss": 117.4494,
|
| 11580 |
+
"step": 16530
|
| 11581 |
+
},
|
| 11582 |
+
{
|
| 11583 |
+
"epoch": 0.0668236929180622,
|
| 11584 |
+
"grad_norm": 0.0,
|
| 11585 |
+
"learning_rate": 3.308e-05,
|
| 11586 |
+
"loss": 84.0806,
|
| 11587 |
+
"step": 16540
|
| 11588 |
+
},
|
| 11589 |
+
{
|
| 11590 |
+
"epoch": 0.06686409418342983,
|
| 11591 |
+
"grad_norm": 1110.686279296875,
|
| 11592 |
+
"learning_rate": 3.3100000000000005e-05,
|
| 11593 |
+
"loss": 149.0371,
|
| 11594 |
+
"step": 16550
|
| 11595 |
+
},
|
| 11596 |
+
{
|
| 11597 |
+
"epoch": 0.06690449544879745,
|
| 11598 |
+
"grad_norm": 1048.002197265625,
|
| 11599 |
+
"learning_rate": 3.312e-05,
|
| 11600 |
+
"loss": 118.8765,
|
| 11601 |
+
"step": 16560
|
| 11602 |
+
},
|
| 11603 |
+
{
|
| 11604 |
+
"epoch": 0.06694489671416509,
|
| 11605 |
+
"grad_norm": 532.6653442382812,
|
| 11606 |
+
"learning_rate": 3.314e-05,
|
| 11607 |
+
"loss": 115.488,
|
| 11608 |
+
"step": 16570
|
| 11609 |
+
},
|
| 11610 |
+
{
|
| 11611 |
+
"epoch": 0.06698529797953272,
|
| 11612 |
+
"grad_norm": 1357.1689453125,
|
| 11613 |
+
"learning_rate": 3.316e-05,
|
| 11614 |
+
"loss": 121.7834,
|
| 11615 |
+
"step": 16580
|
| 11616 |
+
},
|
| 11617 |
+
{
|
| 11618 |
+
"epoch": 0.06702569924490034,
|
| 11619 |
+
"grad_norm": 3220.8486328125,
|
| 11620 |
+
"learning_rate": 3.318e-05,
|
| 11621 |
+
"loss": 146.6197,
|
| 11622 |
+
"step": 16590
|
| 11623 |
+
},
|
| 11624 |
+
{
|
| 11625 |
+
"epoch": 0.06706610051026798,
|
| 11626 |
+
"grad_norm": 809.4137573242188,
|
| 11627 |
+
"learning_rate": 3.32e-05,
|
| 11628 |
+
"loss": 157.1161,
|
| 11629 |
+
"step": 16600
|
| 11630 |
+
},
|
| 11631 |
+
{
|
| 11632 |
+
"epoch": 0.06710650177563561,
|
| 11633 |
+
"grad_norm": 479.6476135253906,
|
| 11634 |
+
"learning_rate": 3.3220000000000004e-05,
|
| 11635 |
+
"loss": 123.665,
|
| 11636 |
+
"step": 16610
|
| 11637 |
+
},
|
| 11638 |
+
{
|
| 11639 |
+
"epoch": 0.06714690304100325,
|
| 11640 |
+
"grad_norm": 1147.013916015625,
|
| 11641 |
+
"learning_rate": 3.324e-05,
|
| 11642 |
+
"loss": 151.3696,
|
| 11643 |
+
"step": 16620
|
| 11644 |
+
},
|
| 11645 |
+
{
|
| 11646 |
+
"epoch": 0.06718730430637088,
|
| 11647 |
+
"grad_norm": 1325.5657958984375,
|
| 11648 |
+
"learning_rate": 3.3260000000000003e-05,
|
| 11649 |
+
"loss": 228.0247,
|
| 11650 |
+
"step": 16630
|
| 11651 |
+
},
|
| 11652 |
+
{
|
| 11653 |
+
"epoch": 0.0672277055717385,
|
| 11654 |
+
"grad_norm": 1421.17578125,
|
| 11655 |
+
"learning_rate": 3.328e-05,
|
| 11656 |
+
"loss": 174.1318,
|
| 11657 |
+
"step": 16640
|
| 11658 |
+
},
|
| 11659 |
+
{
|
| 11660 |
+
"epoch": 0.06726810683710614,
|
| 11661 |
+
"grad_norm": 872.6542358398438,
|
| 11662 |
+
"learning_rate": 3.33e-05,
|
| 11663 |
+
"loss": 118.2037,
|
| 11664 |
+
"step": 16650
|
| 11665 |
+
},
|
| 11666 |
+
{
|
| 11667 |
+
"epoch": 0.06730850810247377,
|
| 11668 |
+
"grad_norm": 936.040283203125,
|
| 11669 |
+
"learning_rate": 3.332e-05,
|
| 11670 |
+
"loss": 113.6235,
|
| 11671 |
+
"step": 16660
|
| 11672 |
+
},
|
| 11673 |
+
{
|
| 11674 |
+
"epoch": 0.0673489093678414,
|
| 11675 |
+
"grad_norm": 829.9392700195312,
|
| 11676 |
+
"learning_rate": 3.3339999999999996e-05,
|
| 11677 |
+
"loss": 124.7012,
|
| 11678 |
+
"step": 16670
|
| 11679 |
+
},
|
| 11680 |
+
{
|
| 11681 |
+
"epoch": 0.06738931063320903,
|
| 11682 |
+
"grad_norm": 829.8602294921875,
|
| 11683 |
+
"learning_rate": 3.336e-05,
|
| 11684 |
+
"loss": 122.1225,
|
| 11685 |
+
"step": 16680
|
| 11686 |
+
},
|
| 11687 |
+
{
|
| 11688 |
+
"epoch": 0.06742971189857666,
|
| 11689 |
+
"grad_norm": 591.7499389648438,
|
| 11690 |
+
"learning_rate": 3.338e-05,
|
| 11691 |
+
"loss": 117.5648,
|
| 11692 |
+
"step": 16690
|
| 11693 |
+
},
|
| 11694 |
+
{
|
| 11695 |
+
"epoch": 0.0674701131639443,
|
| 11696 |
+
"grad_norm": 585.739013671875,
|
| 11697 |
+
"learning_rate": 3.3400000000000005e-05,
|
| 11698 |
+
"loss": 120.4557,
|
| 11699 |
+
"step": 16700
|
| 11700 |
+
},
|
| 11701 |
+
{
|
| 11702 |
+
"epoch": 0.06751051442931193,
|
| 11703 |
+
"grad_norm": 1151.9459228515625,
|
| 11704 |
+
"learning_rate": 3.342e-05,
|
| 11705 |
+
"loss": 122.6646,
|
| 11706 |
+
"step": 16710
|
| 11707 |
+
},
|
| 11708 |
+
{
|
| 11709 |
+
"epoch": 0.06755091569467955,
|
| 11710 |
+
"grad_norm": 1100.1302490234375,
|
| 11711 |
+
"learning_rate": 3.344e-05,
|
| 11712 |
+
"loss": 124.8112,
|
| 11713 |
+
"step": 16720
|
| 11714 |
+
},
|
| 11715 |
+
{
|
| 11716 |
+
"epoch": 0.06759131696004719,
|
| 11717 |
+
"grad_norm": 1049.39501953125,
|
| 11718 |
+
"learning_rate": 3.346e-05,
|
| 11719 |
+
"loss": 131.1703,
|
| 11720 |
+
"step": 16730
|
| 11721 |
+
},
|
| 11722 |
+
{
|
| 11723 |
+
"epoch": 0.06763171822541482,
|
| 11724 |
+
"grad_norm": 976.2811889648438,
|
| 11725 |
+
"learning_rate": 3.348e-05,
|
| 11726 |
+
"loss": 104.992,
|
| 11727 |
+
"step": 16740
|
| 11728 |
+
},
|
| 11729 |
+
{
|
| 11730 |
+
"epoch": 0.06767211949078245,
|
| 11731 |
+
"grad_norm": 1059.683837890625,
|
| 11732 |
+
"learning_rate": 3.35e-05,
|
| 11733 |
+
"loss": 109.1284,
|
| 11734 |
+
"step": 16750
|
| 11735 |
+
},
|
| 11736 |
+
{
|
| 11737 |
+
"epoch": 0.06771252075615009,
|
| 11738 |
+
"grad_norm": 1005.8837280273438,
|
| 11739 |
+
"learning_rate": 3.3520000000000004e-05,
|
| 11740 |
+
"loss": 107.3349,
|
| 11741 |
+
"step": 16760
|
| 11742 |
+
},
|
| 11743 |
+
{
|
| 11744 |
+
"epoch": 0.06775292202151771,
|
| 11745 |
+
"grad_norm": 850.0352783203125,
|
| 11746 |
+
"learning_rate": 3.354e-05,
|
| 11747 |
+
"loss": 126.588,
|
| 11748 |
+
"step": 16770
|
| 11749 |
+
},
|
| 11750 |
+
{
|
| 11751 |
+
"epoch": 0.06779332328688535,
|
| 11752 |
+
"grad_norm": 1083.05712890625,
|
| 11753 |
+
"learning_rate": 3.3560000000000004e-05,
|
| 11754 |
+
"loss": 139.6427,
|
| 11755 |
+
"step": 16780
|
| 11756 |
+
},
|
| 11757 |
+
{
|
| 11758 |
+
"epoch": 0.06783372455225298,
|
| 11759 |
+
"grad_norm": 1554.07080078125,
|
| 11760 |
+
"learning_rate": 3.358e-05,
|
| 11761 |
+
"loss": 145.7689,
|
| 11762 |
+
"step": 16790
|
| 11763 |
+
},
|
| 11764 |
+
{
|
| 11765 |
+
"epoch": 0.0678741258176206,
|
| 11766 |
+
"grad_norm": 3450.517333984375,
|
| 11767 |
+
"learning_rate": 3.3600000000000004e-05,
|
| 11768 |
+
"loss": 140.9184,
|
| 11769 |
+
"step": 16800
|
| 11770 |
+
},
|
| 11771 |
+
{
|
| 11772 |
+
"epoch": 0.06791452708298824,
|
| 11773 |
+
"grad_norm": 1706.325439453125,
|
| 11774 |
+
"learning_rate": 3.362e-05,
|
| 11775 |
+
"loss": 199.2792,
|
| 11776 |
+
"step": 16810
|
| 11777 |
+
},
|
| 11778 |
+
{
|
| 11779 |
+
"epoch": 0.06795492834835587,
|
| 11780 |
+
"grad_norm": 1185.1922607421875,
|
| 11781 |
+
"learning_rate": 3.3639999999999996e-05,
|
| 11782 |
+
"loss": 140.3882,
|
| 11783 |
+
"step": 16820
|
| 11784 |
+
},
|
| 11785 |
+
{
|
| 11786 |
+
"epoch": 0.0679953296137235,
|
| 11787 |
+
"grad_norm": 1169.992919921875,
|
| 11788 |
+
"learning_rate": 3.366e-05,
|
| 11789 |
+
"loss": 123.5538,
|
| 11790 |
+
"step": 16830
|
| 11791 |
+
},
|
| 11792 |
+
{
|
| 11793 |
+
"epoch": 0.06803573087909114,
|
| 11794 |
+
"grad_norm": 1750.4180908203125,
|
| 11795 |
+
"learning_rate": 3.368e-05,
|
| 11796 |
+
"loss": 159.0448,
|
| 11797 |
+
"step": 16840
|
| 11798 |
+
},
|
| 11799 |
+
{
|
| 11800 |
+
"epoch": 0.06807613214445876,
|
| 11801 |
+
"grad_norm": 1771.9908447265625,
|
| 11802 |
+
"learning_rate": 3.3700000000000006e-05,
|
| 11803 |
+
"loss": 153.4333,
|
| 11804 |
+
"step": 16850
|
| 11805 |
+
},
|
| 11806 |
+
{
|
| 11807 |
+
"epoch": 0.0681165334098264,
|
| 11808 |
+
"grad_norm": 2022.061279296875,
|
| 11809 |
+
"learning_rate": 3.372e-05,
|
| 11810 |
+
"loss": 139.1972,
|
| 11811 |
+
"step": 16860
|
| 11812 |
+
},
|
| 11813 |
+
{
|
| 11814 |
+
"epoch": 0.06815693467519403,
|
| 11815 |
+
"grad_norm": 1393.0201416015625,
|
| 11816 |
+
"learning_rate": 3.374e-05,
|
| 11817 |
+
"loss": 152.1366,
|
| 11818 |
+
"step": 16870
|
| 11819 |
+
},
|
| 11820 |
+
{
|
| 11821 |
+
"epoch": 0.06819733594056165,
|
| 11822 |
+
"grad_norm": 2371.81494140625,
|
| 11823 |
+
"learning_rate": 3.376e-05,
|
| 11824 |
+
"loss": 142.447,
|
| 11825 |
+
"step": 16880
|
| 11826 |
+
},
|
| 11827 |
+
{
|
| 11828 |
+
"epoch": 0.0682377372059293,
|
| 11829 |
+
"grad_norm": 1049.164794921875,
|
| 11830 |
+
"learning_rate": 3.378e-05,
|
| 11831 |
+
"loss": 115.6586,
|
| 11832 |
+
"step": 16890
|
| 11833 |
+
},
|
| 11834 |
+
{
|
| 11835 |
+
"epoch": 0.06827813847129692,
|
| 11836 |
+
"grad_norm": 1043.5615234375,
|
| 11837 |
+
"learning_rate": 3.38e-05,
|
| 11838 |
+
"loss": 174.1537,
|
| 11839 |
+
"step": 16900
|
| 11840 |
+
},
|
| 11841 |
+
{
|
| 11842 |
+
"epoch": 0.06831853973666455,
|
| 11843 |
+
"grad_norm": 1559.866455078125,
|
| 11844 |
+
"learning_rate": 3.3820000000000005e-05,
|
| 11845 |
+
"loss": 108.5037,
|
| 11846 |
+
"step": 16910
|
| 11847 |
+
},
|
| 11848 |
+
{
|
| 11849 |
+
"epoch": 0.06835894100203219,
|
| 11850 |
+
"grad_norm": 2484.740478515625,
|
| 11851 |
+
"learning_rate": 3.384e-05,
|
| 11852 |
+
"loss": 184.837,
|
| 11853 |
+
"step": 16920
|
| 11854 |
+
},
|
| 11855 |
+
{
|
| 11856 |
+
"epoch": 0.06839934226739981,
|
| 11857 |
+
"grad_norm": 775.7838745117188,
|
| 11858 |
+
"learning_rate": 3.3860000000000004e-05,
|
| 11859 |
+
"loss": 126.9209,
|
| 11860 |
+
"step": 16930
|
| 11861 |
+
},
|
| 11862 |
+
{
|
| 11863 |
+
"epoch": 0.06843974353276745,
|
| 11864 |
+
"grad_norm": 471.7488708496094,
|
| 11865 |
+
"learning_rate": 3.388e-05,
|
| 11866 |
+
"loss": 135.7176,
|
| 11867 |
+
"step": 16940
|
| 11868 |
+
},
|
| 11869 |
+
{
|
| 11870 |
+
"epoch": 0.06848014479813508,
|
| 11871 |
+
"grad_norm": 969.9213256835938,
|
| 11872 |
+
"learning_rate": 3.3900000000000004e-05,
|
| 11873 |
+
"loss": 95.2846,
|
| 11874 |
+
"step": 16950
|
| 11875 |
+
},
|
| 11876 |
+
{
|
| 11877 |
+
"epoch": 0.0685205460635027,
|
| 11878 |
+
"grad_norm": 1296.1826171875,
|
| 11879 |
+
"learning_rate": 3.392e-05,
|
| 11880 |
+
"loss": 159.3641,
|
| 11881 |
+
"step": 16960
|
| 11882 |
+
},
|
| 11883 |
+
{
|
| 11884 |
+
"epoch": 0.06856094732887034,
|
| 11885 |
+
"grad_norm": 2220.822021484375,
|
| 11886 |
+
"learning_rate": 3.394e-05,
|
| 11887 |
+
"loss": 171.076,
|
| 11888 |
+
"step": 16970
|
| 11889 |
+
},
|
| 11890 |
+
{
|
| 11891 |
+
"epoch": 0.06860134859423797,
|
| 11892 |
+
"grad_norm": 764.9531860351562,
|
| 11893 |
+
"learning_rate": 3.396e-05,
|
| 11894 |
+
"loss": 139.6084,
|
| 11895 |
+
"step": 16980
|
| 11896 |
+
},
|
| 11897 |
+
{
|
| 11898 |
+
"epoch": 0.0686417498596056,
|
| 11899 |
+
"grad_norm": 2453.2900390625,
|
| 11900 |
+
"learning_rate": 3.398e-05,
|
| 11901 |
+
"loss": 162.632,
|
| 11902 |
+
"step": 16990
|
| 11903 |
+
},
|
| 11904 |
+
{
|
| 11905 |
+
"epoch": 0.06868215112497324,
|
| 11906 |
+
"grad_norm": 1016.1632690429688,
|
| 11907 |
+
"learning_rate": 3.4000000000000007e-05,
|
| 11908 |
+
"loss": 236.2789,
|
| 11909 |
+
"step": 17000
|
| 11910 |
+
},
|
| 11911 |
+
{
|
| 11912 |
+
"epoch": 0.06872255239034086,
|
| 11913 |
+
"grad_norm": 1268.531494140625,
|
| 11914 |
+
"learning_rate": 3.402e-05,
|
| 11915 |
+
"loss": 119.6548,
|
| 11916 |
+
"step": 17010
|
| 11917 |
+
},
|
| 11918 |
+
{
|
| 11919 |
+
"epoch": 0.0687629536557085,
|
| 11920 |
+
"grad_norm": 0.0,
|
| 11921 |
+
"learning_rate": 3.404e-05,
|
| 11922 |
+
"loss": 115.3437,
|
| 11923 |
+
"step": 17020
|
| 11924 |
+
},
|
| 11925 |
+
{
|
| 11926 |
+
"epoch": 0.06880335492107613,
|
| 11927 |
+
"grad_norm": 867.4444580078125,
|
| 11928 |
+
"learning_rate": 3.406e-05,
|
| 11929 |
+
"loss": 173.6102,
|
| 11930 |
+
"step": 17030
|
| 11931 |
+
},
|
| 11932 |
+
{
|
| 11933 |
+
"epoch": 0.06884375618644376,
|
| 11934 |
+
"grad_norm": 1306.5975341796875,
|
| 11935 |
+
"learning_rate": 3.408e-05,
|
| 11936 |
+
"loss": 147.4384,
|
| 11937 |
+
"step": 17040
|
| 11938 |
+
},
|
| 11939 |
+
{
|
| 11940 |
+
"epoch": 0.0688841574518114,
|
| 11941 |
+
"grad_norm": 701.0490112304688,
|
| 11942 |
+
"learning_rate": 3.41e-05,
|
| 11943 |
+
"loss": 138.4343,
|
| 11944 |
+
"step": 17050
|
| 11945 |
+
},
|
| 11946 |
+
{
|
| 11947 |
+
"epoch": 0.06892455871717902,
|
| 11948 |
+
"grad_norm": 1092.9249267578125,
|
| 11949 |
+
"learning_rate": 3.412e-05,
|
| 11950 |
+
"loss": 112.9486,
|
| 11951 |
+
"step": 17060
|
| 11952 |
+
},
|
| 11953 |
+
{
|
| 11954 |
+
"epoch": 0.06896495998254665,
|
| 11955 |
+
"grad_norm": 1403.239990234375,
|
| 11956 |
+
"learning_rate": 3.414e-05,
|
| 11957 |
+
"loss": 106.8123,
|
| 11958 |
+
"step": 17070
|
| 11959 |
+
},
|
| 11960 |
+
{
|
| 11961 |
+
"epoch": 0.06900536124791429,
|
| 11962 |
+
"grad_norm": 2368.7109375,
|
| 11963 |
+
"learning_rate": 3.4160000000000005e-05,
|
| 11964 |
+
"loss": 90.9559,
|
| 11965 |
+
"step": 17080
|
| 11966 |
+
},
|
| 11967 |
+
{
|
| 11968 |
+
"epoch": 0.06904576251328191,
|
| 11969 |
+
"grad_norm": 723.0379028320312,
|
| 11970 |
+
"learning_rate": 3.418e-05,
|
| 11971 |
+
"loss": 92.075,
|
| 11972 |
+
"step": 17090
|
| 11973 |
+
},
|
| 11974 |
+
{
|
| 11975 |
+
"epoch": 0.06908616377864955,
|
| 11976 |
+
"grad_norm": 1454.345703125,
|
| 11977 |
+
"learning_rate": 3.4200000000000005e-05,
|
| 11978 |
+
"loss": 122.138,
|
| 11979 |
+
"step": 17100
|
| 11980 |
+
},
|
| 11981 |
+
{
|
| 11982 |
+
"epoch": 0.06912656504401718,
|
| 11983 |
+
"grad_norm": 849.4160766601562,
|
| 11984 |
+
"learning_rate": 3.422e-05,
|
| 11985 |
+
"loss": 132.3787,
|
| 11986 |
+
"step": 17110
|
| 11987 |
+
},
|
| 11988 |
+
{
|
| 11989 |
+
"epoch": 0.0691669663093848,
|
| 11990 |
+
"grad_norm": 1275.969482421875,
|
| 11991 |
+
"learning_rate": 3.424e-05,
|
| 11992 |
+
"loss": 80.5995,
|
| 11993 |
+
"step": 17120
|
| 11994 |
+
},
|
| 11995 |
+
{
|
| 11996 |
+
"epoch": 0.06920736757475245,
|
| 11997 |
+
"grad_norm": 977.9779052734375,
|
| 11998 |
+
"learning_rate": 3.426e-05,
|
| 11999 |
+
"loss": 126.0984,
|
| 12000 |
+
"step": 17130
|
| 12001 |
+
},
|
| 12002 |
+
{
|
| 12003 |
+
"epoch": 0.06924776884012007,
|
| 12004 |
+
"grad_norm": 1691.71826171875,
|
| 12005 |
+
"learning_rate": 3.4280000000000004e-05,
|
| 12006 |
+
"loss": 123.841,
|
| 12007 |
+
"step": 17140
|
| 12008 |
+
},
|
| 12009 |
+
{
|
| 12010 |
+
"epoch": 0.0692881701054877,
|
| 12011 |
+
"grad_norm": 3872.140869140625,
|
| 12012 |
+
"learning_rate": 3.430000000000001e-05,
|
| 12013 |
+
"loss": 148.4945,
|
| 12014 |
+
"step": 17150
|
| 12015 |
+
},
|
| 12016 |
+
{
|
| 12017 |
+
"epoch": 0.06932857137085534,
|
| 12018 |
+
"grad_norm": 3100.976318359375,
|
| 12019 |
+
"learning_rate": 3.4320000000000003e-05,
|
| 12020 |
+
"loss": 117.4309,
|
| 12021 |
+
"step": 17160
|
| 12022 |
+
},
|
| 12023 |
+
{
|
| 12024 |
+
"epoch": 0.06936897263622296,
|
| 12025 |
+
"grad_norm": 1637.591552734375,
|
| 12026 |
+
"learning_rate": 3.434e-05,
|
| 12027 |
+
"loss": 119.0649,
|
| 12028 |
+
"step": 17170
|
| 12029 |
+
},
|
| 12030 |
+
{
|
| 12031 |
+
"epoch": 0.0694093739015906,
|
| 12032 |
+
"grad_norm": 8878.0693359375,
|
| 12033 |
+
"learning_rate": 3.436e-05,
|
| 12034 |
+
"loss": 140.0768,
|
| 12035 |
+
"step": 17180
|
| 12036 |
+
},
|
| 12037 |
+
{
|
| 12038 |
+
"epoch": 0.06944977516695823,
|
| 12039 |
+
"grad_norm": 959.3743286132812,
|
| 12040 |
+
"learning_rate": 3.438e-05,
|
| 12041 |
+
"loss": 172.569,
|
| 12042 |
+
"step": 17190
|
| 12043 |
+
},
|
| 12044 |
+
{
|
| 12045 |
+
"epoch": 0.06949017643232586,
|
| 12046 |
+
"grad_norm": 520.9686889648438,
|
| 12047 |
+
"learning_rate": 3.4399999999999996e-05,
|
| 12048 |
+
"loss": 93.1335,
|
| 12049 |
+
"step": 17200
|
| 12050 |
+
},
|
| 12051 |
+
{
|
| 12052 |
+
"epoch": 0.0695305776976935,
|
| 12053 |
+
"grad_norm": 1169.4296875,
|
| 12054 |
+
"learning_rate": 3.442e-05,
|
| 12055 |
+
"loss": 109.4803,
|
| 12056 |
+
"step": 17210
|
| 12057 |
+
},
|
| 12058 |
+
{
|
| 12059 |
+
"epoch": 0.06957097896306112,
|
| 12060 |
+
"grad_norm": 2471.82568359375,
|
| 12061 |
+
"learning_rate": 3.444e-05,
|
| 12062 |
+
"loss": 149.8675,
|
| 12063 |
+
"step": 17220
|
| 12064 |
+
},
|
| 12065 |
+
{
|
| 12066 |
+
"epoch": 0.06961138022842875,
|
| 12067 |
+
"grad_norm": 1193.904541015625,
|
| 12068 |
+
"learning_rate": 3.4460000000000005e-05,
|
| 12069 |
+
"loss": 137.5065,
|
| 12070 |
+
"step": 17230
|
| 12071 |
+
},
|
| 12072 |
+
{
|
| 12073 |
+
"epoch": 0.06965178149379639,
|
| 12074 |
+
"grad_norm": 796.9132080078125,
|
| 12075 |
+
"learning_rate": 3.448e-05,
|
| 12076 |
+
"loss": 89.8475,
|
| 12077 |
+
"step": 17240
|
| 12078 |
+
},
|
| 12079 |
+
{
|
| 12080 |
+
"epoch": 0.06969218275916401,
|
| 12081 |
+
"grad_norm": 688.8414916992188,
|
| 12082 |
+
"learning_rate": 3.45e-05,
|
| 12083 |
+
"loss": 125.9576,
|
| 12084 |
+
"step": 17250
|
| 12085 |
+
},
|
| 12086 |
+
{
|
| 12087 |
+
"epoch": 0.06973258402453165,
|
| 12088 |
+
"grad_norm": 480.16009521484375,
|
| 12089 |
+
"learning_rate": 3.452e-05,
|
| 12090 |
+
"loss": 78.7464,
|
| 12091 |
+
"step": 17260
|
| 12092 |
+
},
|
| 12093 |
+
{
|
| 12094 |
+
"epoch": 0.06977298528989928,
|
| 12095 |
+
"grad_norm": 783.6867065429688,
|
| 12096 |
+
"learning_rate": 3.454e-05,
|
| 12097 |
+
"loss": 108.8815,
|
| 12098 |
+
"step": 17270
|
| 12099 |
+
},
|
| 12100 |
+
{
|
| 12101 |
+
"epoch": 0.06981338655526691,
|
| 12102 |
+
"grad_norm": 0.0,
|
| 12103 |
+
"learning_rate": 3.456e-05,
|
| 12104 |
+
"loss": 116.7318,
|
| 12105 |
+
"step": 17280
|
| 12106 |
+
},
|
| 12107 |
+
{
|
| 12108 |
+
"epoch": 0.06985378782063455,
|
| 12109 |
+
"grad_norm": 972.9898681640625,
|
| 12110 |
+
"learning_rate": 3.4580000000000004e-05,
|
| 12111 |
+
"loss": 163.7325,
|
| 12112 |
+
"step": 17290
|
| 12113 |
+
},
|
| 12114 |
+
{
|
| 12115 |
+
"epoch": 0.06989418908600217,
|
| 12116 |
+
"grad_norm": 2305.32470703125,
|
| 12117 |
+
"learning_rate": 3.46e-05,
|
| 12118 |
+
"loss": 102.6796,
|
| 12119 |
+
"step": 17300
|
| 12120 |
+
},
|
| 12121 |
+
{
|
| 12122 |
+
"epoch": 0.0699345903513698,
|
| 12123 |
+
"grad_norm": 985.5175170898438,
|
| 12124 |
+
"learning_rate": 3.4620000000000004e-05,
|
| 12125 |
+
"loss": 133.6453,
|
| 12126 |
+
"step": 17310
|
| 12127 |
+
},
|
| 12128 |
+
{
|
| 12129 |
+
"epoch": 0.06997499161673744,
|
| 12130 |
+
"grad_norm": 966.200439453125,
|
| 12131 |
+
"learning_rate": 3.464e-05,
|
| 12132 |
+
"loss": 133.9981,
|
| 12133 |
+
"step": 17320
|
| 12134 |
+
},
|
| 12135 |
+
{
|
| 12136 |
+
"epoch": 0.07001539288210507,
|
| 12137 |
+
"grad_norm": 582.17041015625,
|
| 12138 |
+
"learning_rate": 3.4660000000000004e-05,
|
| 12139 |
+
"loss": 119.5202,
|
| 12140 |
+
"step": 17330
|
| 12141 |
+
},
|
| 12142 |
+
{
|
| 12143 |
+
"epoch": 0.0700557941474727,
|
| 12144 |
+
"grad_norm": 841.1046752929688,
|
| 12145 |
+
"learning_rate": 3.468e-05,
|
| 12146 |
+
"loss": 133.1583,
|
| 12147 |
+
"step": 17340
|
| 12148 |
+
},
|
| 12149 |
+
{
|
| 12150 |
+
"epoch": 0.07009619541284033,
|
| 12151 |
+
"grad_norm": 675.8413696289062,
|
| 12152 |
+
"learning_rate": 3.4699999999999996e-05,
|
| 12153 |
+
"loss": 139.8855,
|
| 12154 |
+
"step": 17350
|
| 12155 |
+
},
|
| 12156 |
+
{
|
| 12157 |
+
"epoch": 0.07013659667820796,
|
| 12158 |
+
"grad_norm": 957.8898315429688,
|
| 12159 |
+
"learning_rate": 3.472e-05,
|
| 12160 |
+
"loss": 137.3781,
|
| 12161 |
+
"step": 17360
|
| 12162 |
+
},
|
| 12163 |
+
{
|
| 12164 |
+
"epoch": 0.0701769979435756,
|
| 12165 |
+
"grad_norm": 3106.9208984375,
|
| 12166 |
+
"learning_rate": 3.474e-05,
|
| 12167 |
+
"loss": 167.9274,
|
| 12168 |
+
"step": 17370
|
| 12169 |
+
},
|
| 12170 |
+
{
|
| 12171 |
+
"epoch": 0.07021739920894322,
|
| 12172 |
+
"grad_norm": 1493.3551025390625,
|
| 12173 |
+
"learning_rate": 3.4760000000000006e-05,
|
| 12174 |
+
"loss": 163.2431,
|
| 12175 |
+
"step": 17380
|
| 12176 |
+
},
|
| 12177 |
+
{
|
| 12178 |
+
"epoch": 0.07025780047431085,
|
| 12179 |
+
"grad_norm": 993.3938598632812,
|
| 12180 |
+
"learning_rate": 3.478e-05,
|
| 12181 |
+
"loss": 114.7312,
|
| 12182 |
+
"step": 17390
|
| 12183 |
+
},
|
| 12184 |
+
{
|
| 12185 |
+
"epoch": 0.07029820173967849,
|
| 12186 |
+
"grad_norm": 1034.116943359375,
|
| 12187 |
+
"learning_rate": 3.48e-05,
|
| 12188 |
+
"loss": 70.717,
|
| 12189 |
+
"step": 17400
|
| 12190 |
+
},
|
| 12191 |
+
{
|
| 12192 |
+
"epoch": 0.07033860300504612,
|
| 12193 |
+
"grad_norm": 921.6234741210938,
|
| 12194 |
+
"learning_rate": 3.482e-05,
|
| 12195 |
+
"loss": 185.619,
|
| 12196 |
+
"step": 17410
|
| 12197 |
+
},
|
| 12198 |
+
{
|
| 12199 |
+
"epoch": 0.07037900427041376,
|
| 12200 |
+
"grad_norm": 1703.0343017578125,
|
| 12201 |
+
"learning_rate": 3.484e-05,
|
| 12202 |
+
"loss": 142.7865,
|
| 12203 |
+
"step": 17420
|
| 12204 |
+
},
|
| 12205 |
+
{
|
| 12206 |
+
"epoch": 0.07041940553578138,
|
| 12207 |
+
"grad_norm": 1368.336181640625,
|
| 12208 |
+
"learning_rate": 3.486e-05,
|
| 12209 |
+
"loss": 116.2519,
|
| 12210 |
+
"step": 17430
|
| 12211 |
+
},
|
| 12212 |
+
{
|
| 12213 |
+
"epoch": 0.07045980680114901,
|
| 12214 |
+
"grad_norm": 1559.697265625,
|
| 12215 |
+
"learning_rate": 3.4880000000000005e-05,
|
| 12216 |
+
"loss": 110.2219,
|
| 12217 |
+
"step": 17440
|
| 12218 |
+
},
|
| 12219 |
+
{
|
| 12220 |
+
"epoch": 0.07050020806651665,
|
| 12221 |
+
"grad_norm": 3143.30517578125,
|
| 12222 |
+
"learning_rate": 3.49e-05,
|
| 12223 |
+
"loss": 91.7808,
|
| 12224 |
+
"step": 17450
|
| 12225 |
+
},
|
| 12226 |
+
{
|
| 12227 |
+
"epoch": 0.07054060933188427,
|
| 12228 |
+
"grad_norm": 1277.235595703125,
|
| 12229 |
+
"learning_rate": 3.4920000000000004e-05,
|
| 12230 |
+
"loss": 143.4617,
|
| 12231 |
+
"step": 17460
|
| 12232 |
+
},
|
| 12233 |
+
{
|
| 12234 |
+
"epoch": 0.0705810105972519,
|
| 12235 |
+
"grad_norm": 2107.630615234375,
|
| 12236 |
+
"learning_rate": 3.494e-05,
|
| 12237 |
+
"loss": 141.62,
|
| 12238 |
+
"step": 17470
|
| 12239 |
+
},
|
| 12240 |
+
{
|
| 12241 |
+
"epoch": 0.07062141186261954,
|
| 12242 |
+
"grad_norm": 916.9269409179688,
|
| 12243 |
+
"learning_rate": 3.4960000000000004e-05,
|
| 12244 |
+
"loss": 124.4125,
|
| 12245 |
+
"step": 17480
|
| 12246 |
+
},
|
| 12247 |
+
{
|
| 12248 |
+
"epoch": 0.07066181312798717,
|
| 12249 |
+
"grad_norm": 495.5371398925781,
|
| 12250 |
+
"learning_rate": 3.498e-05,
|
| 12251 |
+
"loss": 113.1787,
|
| 12252 |
+
"step": 17490
|
| 12253 |
+
},
|
| 12254 |
+
{
|
| 12255 |
+
"epoch": 0.0707022143933548,
|
| 12256 |
+
"grad_norm": 722.7784423828125,
|
| 12257 |
+
"learning_rate": 3.5e-05,
|
| 12258 |
+
"loss": 113.6147,
|
| 12259 |
+
"step": 17500
|
| 12260 |
+
},
|
| 12261 |
+
{
|
| 12262 |
+
"epoch": 0.07074261565872243,
|
| 12263 |
+
"grad_norm": 8005.68017578125,
|
| 12264 |
+
"learning_rate": 3.502e-05,
|
| 12265 |
+
"loss": 160.4929,
|
| 12266 |
+
"step": 17510
|
| 12267 |
+
},
|
| 12268 |
+
{
|
| 12269 |
+
"epoch": 0.07078301692409006,
|
| 12270 |
+
"grad_norm": 836.4596557617188,
|
| 12271 |
+
"learning_rate": 3.504e-05,
|
| 12272 |
+
"loss": 93.0578,
|
| 12273 |
+
"step": 17520
|
| 12274 |
+
},
|
| 12275 |
+
{
|
| 12276 |
+
"epoch": 0.0708234181894577,
|
| 12277 |
+
"grad_norm": 1631.076171875,
|
| 12278 |
+
"learning_rate": 3.5060000000000007e-05,
|
| 12279 |
+
"loss": 126.6171,
|
| 12280 |
+
"step": 17530
|
| 12281 |
+
},
|
| 12282 |
+
{
|
| 12283 |
+
"epoch": 0.07086381945482532,
|
| 12284 |
+
"grad_norm": 639.418212890625,
|
| 12285 |
+
"learning_rate": 3.508e-05,
|
| 12286 |
+
"loss": 105.7042,
|
| 12287 |
+
"step": 17540
|
| 12288 |
+
},
|
| 12289 |
+
{
|
| 12290 |
+
"epoch": 0.07090422072019295,
|
| 12291 |
+
"grad_norm": 1241.2578125,
|
| 12292 |
+
"learning_rate": 3.51e-05,
|
| 12293 |
+
"loss": 201.5257,
|
| 12294 |
+
"step": 17550
|
| 12295 |
+
},
|
| 12296 |
+
{
|
| 12297 |
+
"epoch": 0.07094462198556059,
|
| 12298 |
+
"grad_norm": 835.3104248046875,
|
| 12299 |
+
"learning_rate": 3.512e-05,
|
| 12300 |
+
"loss": 159.556,
|
| 12301 |
+
"step": 17560
|
| 12302 |
+
},
|
| 12303 |
+
{
|
| 12304 |
+
"epoch": 0.07098502325092822,
|
| 12305 |
+
"grad_norm": 565.4751586914062,
|
| 12306 |
+
"learning_rate": 3.514e-05,
|
| 12307 |
+
"loss": 93.872,
|
| 12308 |
+
"step": 17570
|
| 12309 |
+
},
|
| 12310 |
+
{
|
| 12311 |
+
"epoch": 0.07102542451629586,
|
| 12312 |
+
"grad_norm": 599.105712890625,
|
| 12313 |
+
"learning_rate": 3.516e-05,
|
| 12314 |
+
"loss": 132.7931,
|
| 12315 |
+
"step": 17580
|
| 12316 |
+
},
|
| 12317 |
+
{
|
| 12318 |
+
"epoch": 0.07106582578166348,
|
| 12319 |
+
"grad_norm": 1342.3175048828125,
|
| 12320 |
+
"learning_rate": 3.518e-05,
|
| 12321 |
+
"loss": 121.4404,
|
| 12322 |
+
"step": 17590
|
| 12323 |
+
},
|
| 12324 |
+
{
|
| 12325 |
+
"epoch": 0.07110622704703111,
|
| 12326 |
+
"grad_norm": 853.9281616210938,
|
| 12327 |
+
"learning_rate": 3.52e-05,
|
| 12328 |
+
"loss": 185.0158,
|
| 12329 |
+
"step": 17600
|
| 12330 |
+
},
|
| 12331 |
+
{
|
| 12332 |
+
"epoch": 0.07114662831239875,
|
| 12333 |
+
"grad_norm": 1526.38427734375,
|
| 12334 |
+
"learning_rate": 3.5220000000000005e-05,
|
| 12335 |
+
"loss": 130.5014,
|
| 12336 |
+
"step": 17610
|
| 12337 |
+
},
|
| 12338 |
+
{
|
| 12339 |
+
"epoch": 0.07118702957776638,
|
| 12340 |
+
"grad_norm": 0.0,
|
| 12341 |
+
"learning_rate": 3.524e-05,
|
| 12342 |
+
"loss": 155.2073,
|
| 12343 |
+
"step": 17620
|
| 12344 |
+
},
|
| 12345 |
+
{
|
| 12346 |
+
"epoch": 0.071227430843134,
|
| 12347 |
+
"grad_norm": 722.3214111328125,
|
| 12348 |
+
"learning_rate": 3.5260000000000005e-05,
|
| 12349 |
+
"loss": 98.7197,
|
| 12350 |
+
"step": 17630
|
| 12351 |
+
},
|
| 12352 |
+
{
|
| 12353 |
+
"epoch": 0.07126783210850164,
|
| 12354 |
+
"grad_norm": 769.6561889648438,
|
| 12355 |
+
"learning_rate": 3.528e-05,
|
| 12356 |
+
"loss": 98.4344,
|
| 12357 |
+
"step": 17640
|
| 12358 |
+
},
|
| 12359 |
+
{
|
| 12360 |
+
"epoch": 0.07130823337386927,
|
| 12361 |
+
"grad_norm": 1916.5062255859375,
|
| 12362 |
+
"learning_rate": 3.53e-05,
|
| 12363 |
+
"loss": 169.392,
|
| 12364 |
+
"step": 17650
|
| 12365 |
+
},
|
| 12366 |
+
{
|
| 12367 |
+
"epoch": 0.07134863463923691,
|
| 12368 |
+
"grad_norm": 790.8156127929688,
|
| 12369 |
+
"learning_rate": 3.532e-05,
|
| 12370 |
+
"loss": 112.2699,
|
| 12371 |
+
"step": 17660
|
| 12372 |
+
},
|
| 12373 |
+
{
|
| 12374 |
+
"epoch": 0.07138903590460453,
|
| 12375 |
+
"grad_norm": 0.0,
|
| 12376 |
+
"learning_rate": 3.5340000000000004e-05,
|
| 12377 |
+
"loss": 153.1781,
|
| 12378 |
+
"step": 17670
|
| 12379 |
+
},
|
| 12380 |
+
{
|
| 12381 |
+
"epoch": 0.07142943716997216,
|
| 12382 |
+
"grad_norm": 1370.72216796875,
|
| 12383 |
+
"learning_rate": 3.536000000000001e-05,
|
| 12384 |
+
"loss": 170.2639,
|
| 12385 |
+
"step": 17680
|
| 12386 |
+
},
|
| 12387 |
+
{
|
| 12388 |
+
"epoch": 0.0714698384353398,
|
| 12389 |
+
"grad_norm": 466.6986999511719,
|
| 12390 |
+
"learning_rate": 3.5380000000000003e-05,
|
| 12391 |
+
"loss": 140.3446,
|
| 12392 |
+
"step": 17690
|
| 12393 |
+
},
|
| 12394 |
+
{
|
| 12395 |
+
"epoch": 0.07151023970070743,
|
| 12396 |
+
"grad_norm": 1512.586181640625,
|
| 12397 |
+
"learning_rate": 3.54e-05,
|
| 12398 |
+
"loss": 123.2379,
|
| 12399 |
+
"step": 17700
|
| 12400 |
+
},
|
| 12401 |
+
{
|
| 12402 |
+
"epoch": 0.07155064096607505,
|
| 12403 |
+
"grad_norm": 1187.423828125,
|
| 12404 |
+
"learning_rate": 3.542e-05,
|
| 12405 |
+
"loss": 121.9325,
|
| 12406 |
+
"step": 17710
|
| 12407 |
+
},
|
| 12408 |
+
{
|
| 12409 |
+
"epoch": 0.07159104223144269,
|
| 12410 |
+
"grad_norm": 1491.8526611328125,
|
| 12411 |
+
"learning_rate": 3.544e-05,
|
| 12412 |
+
"loss": 97.88,
|
| 12413 |
+
"step": 17720
|
| 12414 |
+
},
|
| 12415 |
+
{
|
| 12416 |
+
"epoch": 0.07163144349681032,
|
| 12417 |
+
"grad_norm": 819.60400390625,
|
| 12418 |
+
"learning_rate": 3.546e-05,
|
| 12419 |
+
"loss": 136.5916,
|
| 12420 |
+
"step": 17730
|
| 12421 |
+
},
|
| 12422 |
+
{
|
| 12423 |
+
"epoch": 0.07167184476217796,
|
| 12424 |
+
"grad_norm": 650.4725952148438,
|
| 12425 |
+
"learning_rate": 3.548e-05,
|
| 12426 |
+
"loss": 100.4026,
|
| 12427 |
+
"step": 17740
|
| 12428 |
+
},
|
| 12429 |
+
{
|
| 12430 |
+
"epoch": 0.07171224602754558,
|
| 12431 |
+
"grad_norm": 822.773681640625,
|
| 12432 |
+
"learning_rate": 3.55e-05,
|
| 12433 |
+
"loss": 103.9474,
|
| 12434 |
+
"step": 17750
|
| 12435 |
+
},
|
| 12436 |
+
{
|
| 12437 |
+
"epoch": 0.07175264729291321,
|
| 12438 |
+
"grad_norm": 879.6776123046875,
|
| 12439 |
+
"learning_rate": 3.5520000000000006e-05,
|
| 12440 |
+
"loss": 90.2404,
|
| 12441 |
+
"step": 17760
|
| 12442 |
+
},
|
| 12443 |
+
{
|
| 12444 |
+
"epoch": 0.07179304855828085,
|
| 12445 |
+
"grad_norm": 1487.169921875,
|
| 12446 |
+
"learning_rate": 3.554e-05,
|
| 12447 |
+
"loss": 172.1061,
|
| 12448 |
+
"step": 17770
|
| 12449 |
+
},
|
| 12450 |
+
{
|
| 12451 |
+
"epoch": 0.07183344982364848,
|
| 12452 |
+
"grad_norm": 953.589111328125,
|
| 12453 |
+
"learning_rate": 3.5560000000000005e-05,
|
| 12454 |
+
"loss": 150.0098,
|
| 12455 |
+
"step": 17780
|
| 12456 |
+
},
|
| 12457 |
+
{
|
| 12458 |
+
"epoch": 0.0718738510890161,
|
| 12459 |
+
"grad_norm": 688.1337890625,
|
| 12460 |
+
"learning_rate": 3.558e-05,
|
| 12461 |
+
"loss": 78.3569,
|
| 12462 |
+
"step": 17790
|
| 12463 |
+
},
|
| 12464 |
+
{
|
| 12465 |
+
"epoch": 0.07191425235438374,
|
| 12466 |
+
"grad_norm": 789.90283203125,
|
| 12467 |
+
"learning_rate": 3.56e-05,
|
| 12468 |
+
"loss": 106.5888,
|
| 12469 |
+
"step": 17800
|
| 12470 |
+
},
|
| 12471 |
+
{
|
| 12472 |
+
"epoch": 0.07195465361975137,
|
| 12473 |
+
"grad_norm": 934.8272094726562,
|
| 12474 |
+
"learning_rate": 3.562e-05,
|
| 12475 |
+
"loss": 111.0842,
|
| 12476 |
+
"step": 17810
|
| 12477 |
+
},
|
| 12478 |
+
{
|
| 12479 |
+
"epoch": 0.07199505488511901,
|
| 12480 |
+
"grad_norm": 336.4629211425781,
|
| 12481 |
+
"learning_rate": 3.5640000000000004e-05,
|
| 12482 |
+
"loss": 99.3697,
|
| 12483 |
+
"step": 17820
|
| 12484 |
+
},
|
| 12485 |
+
{
|
| 12486 |
+
"epoch": 0.07203545615048663,
|
| 12487 |
+
"grad_norm": 3090.107666015625,
|
| 12488 |
+
"learning_rate": 3.566e-05,
|
| 12489 |
+
"loss": 97.4285,
|
| 12490 |
+
"step": 17830
|
| 12491 |
+
},
|
| 12492 |
+
{
|
| 12493 |
+
"epoch": 0.07207585741585426,
|
| 12494 |
+
"grad_norm": 1428.7044677734375,
|
| 12495 |
+
"learning_rate": 3.5680000000000004e-05,
|
| 12496 |
+
"loss": 118.1064,
|
| 12497 |
+
"step": 17840
|
| 12498 |
+
},
|
| 12499 |
+
{
|
| 12500 |
+
"epoch": 0.0721162586812219,
|
| 12501 |
+
"grad_norm": 2355.272705078125,
|
| 12502 |
+
"learning_rate": 3.57e-05,
|
| 12503 |
+
"loss": 151.7737,
|
| 12504 |
+
"step": 17850
|
| 12505 |
+
},
|
| 12506 |
+
{
|
| 12507 |
+
"epoch": 0.07215665994658953,
|
| 12508 |
+
"grad_norm": 1052.816650390625,
|
| 12509 |
+
"learning_rate": 3.5720000000000004e-05,
|
| 12510 |
+
"loss": 86.3124,
|
| 12511 |
+
"step": 17860
|
| 12512 |
+
},
|
| 12513 |
+
{
|
| 12514 |
+
"epoch": 0.07219706121195715,
|
| 12515 |
+
"grad_norm": 683.0614624023438,
|
| 12516 |
+
"learning_rate": 3.574e-05,
|
| 12517 |
+
"loss": 120.8411,
|
| 12518 |
+
"step": 17870
|
| 12519 |
+
},
|
| 12520 |
+
{
|
| 12521 |
+
"epoch": 0.0722374624773248,
|
| 12522 |
+
"grad_norm": 1161.9776611328125,
|
| 12523 |
+
"learning_rate": 3.5759999999999996e-05,
|
| 12524 |
+
"loss": 145.4579,
|
| 12525 |
+
"step": 17880
|
| 12526 |
+
},
|
| 12527 |
+
{
|
| 12528 |
+
"epoch": 0.07227786374269242,
|
| 12529 |
+
"grad_norm": 1213.32861328125,
|
| 12530 |
+
"learning_rate": 3.578e-05,
|
| 12531 |
+
"loss": 160.1318,
|
| 12532 |
+
"step": 17890
|
| 12533 |
+
},
|
| 12534 |
+
{
|
| 12535 |
+
"epoch": 0.07231826500806006,
|
| 12536 |
+
"grad_norm": 1079.72119140625,
|
| 12537 |
+
"learning_rate": 3.58e-05,
|
| 12538 |
+
"loss": 125.1568,
|
| 12539 |
+
"step": 17900
|
| 12540 |
+
},
|
| 12541 |
+
{
|
| 12542 |
+
"epoch": 0.07235866627342769,
|
| 12543 |
+
"grad_norm": 960.3634643554688,
|
| 12544 |
+
"learning_rate": 3.5820000000000006e-05,
|
| 12545 |
+
"loss": 116.5516,
|
| 12546 |
+
"step": 17910
|
| 12547 |
+
},
|
| 12548 |
+
{
|
| 12549 |
+
"epoch": 0.07239906753879531,
|
| 12550 |
+
"grad_norm": 1030.158447265625,
|
| 12551 |
+
"learning_rate": 3.584e-05,
|
| 12552 |
+
"loss": 78.2271,
|
| 12553 |
+
"step": 17920
|
| 12554 |
+
},
|
| 12555 |
+
{
|
| 12556 |
+
"epoch": 0.07243946880416295,
|
| 12557 |
+
"grad_norm": 812.413818359375,
|
| 12558 |
+
"learning_rate": 3.586e-05,
|
| 12559 |
+
"loss": 122.26,
|
| 12560 |
+
"step": 17930
|
| 12561 |
+
},
|
| 12562 |
+
{
|
| 12563 |
+
"epoch": 0.07247987006953058,
|
| 12564 |
+
"grad_norm": 2242.89794921875,
|
| 12565 |
+
"learning_rate": 3.588e-05,
|
| 12566 |
+
"loss": 150.4362,
|
| 12567 |
+
"step": 17940
|
| 12568 |
+
},
|
| 12569 |
+
{
|
| 12570 |
+
"epoch": 0.0725202713348982,
|
| 12571 |
+
"grad_norm": 1707.64306640625,
|
| 12572 |
+
"learning_rate": 3.59e-05,
|
| 12573 |
+
"loss": 107.4575,
|
| 12574 |
+
"step": 17950
|
| 12575 |
+
},
|
| 12576 |
+
{
|
| 12577 |
+
"epoch": 0.07256067260026584,
|
| 12578 |
+
"grad_norm": 1091.1968994140625,
|
| 12579 |
+
"learning_rate": 3.592e-05,
|
| 12580 |
+
"loss": 176.9033,
|
| 12581 |
+
"step": 17960
|
| 12582 |
+
},
|
| 12583 |
+
{
|
| 12584 |
+
"epoch": 0.07260107386563347,
|
| 12585 |
+
"grad_norm": 660.7951049804688,
|
| 12586 |
+
"learning_rate": 3.594e-05,
|
| 12587 |
+
"loss": 120.4594,
|
| 12588 |
+
"step": 17970
|
| 12589 |
+
},
|
| 12590 |
+
{
|
| 12591 |
+
"epoch": 0.07264147513100111,
|
| 12592 |
+
"grad_norm": 1081.6932373046875,
|
| 12593 |
+
"learning_rate": 3.596e-05,
|
| 12594 |
+
"loss": 101.9581,
|
| 12595 |
+
"step": 17980
|
| 12596 |
+
},
|
| 12597 |
+
{
|
| 12598 |
+
"epoch": 0.07268187639636874,
|
| 12599 |
+
"grad_norm": 1247.1197509765625,
|
| 12600 |
+
"learning_rate": 3.5980000000000004e-05,
|
| 12601 |
+
"loss": 90.3047,
|
| 12602 |
+
"step": 17990
|
| 12603 |
+
},
|
| 12604 |
+
{
|
| 12605 |
+
"epoch": 0.07272227766173636,
|
| 12606 |
+
"grad_norm": 2352.89208984375,
|
| 12607 |
+
"learning_rate": 3.6e-05,
|
| 12608 |
+
"loss": 121.5598,
|
| 12609 |
+
"step": 18000
|
| 12610 |
}
|
| 12611 |
],
|
| 12612 |
"logging_steps": 10,
|