Training in progress, step 696, checkpoint
Browse files
last-checkpoint/adapter_model.safetensors
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 83945296
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:69108424b1bcd90145c81dcab6b81a281f0479689a5a19bbcaa1c5436862d67a
|
3 |
size 83945296
|
last-checkpoint/optimizer.pt
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 43123028
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:1f0425bbfddd7ab5e730d9e74c85b37d66ef7b02895991a9af0b26751f7bdf7d
|
3 |
size 43123028
|
last-checkpoint/rng_state.pth
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 14244
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:69850639cee8f2685525ebf1af181c92827de6e02461bad526ac2c8862741977
|
3 |
size 14244
|
last-checkpoint/scheduler.pt
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 1064
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:e9c4dc7ed2ce313c1651dba3a80399685ba74c1e69b5f37bde89bd99b81a9c1e
|
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": 348,
|
6 |
-
"global_step":
|
7 |
"is_hyper_param_search": false,
|
8 |
"is_local_process_zero": true,
|
9 |
"is_world_process_zero": true,
|
@@ -2459,6 +2459,2450 @@
|
|
2459 |
"eval_samples_per_second": 27.284,
|
2460 |
"eval_steps_per_second": 13.642,
|
2461 |
"step": 348
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2462 |
}
|
2463 |
],
|
2464 |
"logging_steps": 1,
|
@@ -2478,7 +4922,7 @@
|
|
2478 |
"attributes": {}
|
2479 |
}
|
2480 |
},
|
2481 |
-
"total_flos":
|
2482 |
"train_batch_size": 2,
|
2483 |
"trial_name": null,
|
2484 |
"trial_params": null
|
|
|
1 |
{
|
2 |
"best_metric": null,
|
3 |
"best_model_checkpoint": null,
|
4 |
+
"epoch": 0.13648396901657026,
|
5 |
"eval_steps": 348,
|
6 |
+
"global_step": 696,
|
7 |
"is_hyper_param_search": false,
|
8 |
"is_local_process_zero": true,
|
9 |
"is_world_process_zero": true,
|
|
|
2459 |
"eval_samples_per_second": 27.284,
|
2460 |
"eval_steps_per_second": 13.642,
|
2461 |
"step": 348
|
2462 |
+
},
|
2463 |
+
{
|
2464 |
+
"epoch": 0.06843808216491813,
|
2465 |
+
"grad_norm": 7.237430095672607,
|
2466 |
+
"learning_rate": 0.00017163085670731371,
|
2467 |
+
"loss": 1.6659,
|
2468 |
+
"step": 349
|
2469 |
+
},
|
2470 |
+
{
|
2471 |
+
"epoch": 0.06863417982155114,
|
2472 |
+
"grad_norm": 10.189397811889648,
|
2473 |
+
"learning_rate": 0.00017147170417210333,
|
2474 |
+
"loss": 1.5962,
|
2475 |
+
"step": 350
|
2476 |
+
},
|
2477 |
+
{
|
2478 |
+
"epoch": 0.06883027747818414,
|
2479 |
+
"grad_norm": 11.650528907775879,
|
2480 |
+
"learning_rate": 0.00017131218069561593,
|
2481 |
+
"loss": 2.8224,
|
2482 |
+
"step": 351
|
2483 |
+
},
|
2484 |
+
{
|
2485 |
+
"epoch": 0.06902637513481714,
|
2486 |
+
"grad_norm": 5.0007147789001465,
|
2487 |
+
"learning_rate": 0.00017115228710578534,
|
2488 |
+
"loss": 1.8055,
|
2489 |
+
"step": 352
|
2490 |
+
},
|
2491 |
+
{
|
2492 |
+
"epoch": 0.06922247279145015,
|
2493 |
+
"grad_norm": 4.950777530670166,
|
2494 |
+
"learning_rate": 0.0001709920242324663,
|
2495 |
+
"loss": 1.1707,
|
2496 |
+
"step": 353
|
2497 |
+
},
|
2498 |
+
{
|
2499 |
+
"epoch": 0.06941857044808314,
|
2500 |
+
"grad_norm": 11.658537864685059,
|
2501 |
+
"learning_rate": 0.0001708313929074302,
|
2502 |
+
"loss": 1.6848,
|
2503 |
+
"step": 354
|
2504 |
+
},
|
2505 |
+
{
|
2506 |
+
"epoch": 0.06961466810471614,
|
2507 |
+
"grad_norm": 6.6815009117126465,
|
2508 |
+
"learning_rate": 0.00017067039396436058,
|
2509 |
+
"loss": 1.6768,
|
2510 |
+
"step": 355
|
2511 |
+
},
|
2512 |
+
{
|
2513 |
+
"epoch": 0.06981076576134915,
|
2514 |
+
"grad_norm": 7.522084712982178,
|
2515 |
+
"learning_rate": 0.00017050902823884903,
|
2516 |
+
"loss": 1.2074,
|
2517 |
+
"step": 356
|
2518 |
+
},
|
2519 |
+
{
|
2520 |
+
"epoch": 0.07000686341798215,
|
2521 |
+
"grad_norm": 9.470911026000977,
|
2522 |
+
"learning_rate": 0.00017034729656839078,
|
2523 |
+
"loss": 0.8594,
|
2524 |
+
"step": 357
|
2525 |
+
},
|
2526 |
+
{
|
2527 |
+
"epoch": 0.07020296107461516,
|
2528 |
+
"grad_norm": 4.798830032348633,
|
2529 |
+
"learning_rate": 0.00017018519979238023,
|
2530 |
+
"loss": 1.6142,
|
2531 |
+
"step": 358
|
2532 |
+
},
|
2533 |
+
{
|
2534 |
+
"epoch": 0.07039905873124816,
|
2535 |
+
"grad_norm": 4.626184940338135,
|
2536 |
+
"learning_rate": 0.0001700227387521068,
|
2537 |
+
"loss": 1.6397,
|
2538 |
+
"step": 359
|
2539 |
+
},
|
2540 |
+
{
|
2541 |
+
"epoch": 0.07059515638788116,
|
2542 |
+
"grad_norm": 6.68535041809082,
|
2543 |
+
"learning_rate": 0.00016985991429075036,
|
2544 |
+
"loss": 0.623,
|
2545 |
+
"step": 360
|
2546 |
+
},
|
2547 |
+
{
|
2548 |
+
"epoch": 0.07079125404451417,
|
2549 |
+
"grad_norm": 3.028799533843994,
|
2550 |
+
"learning_rate": 0.00016969672725337706,
|
2551 |
+
"loss": 0.7686,
|
2552 |
+
"step": 361
|
2553 |
+
},
|
2554 |
+
{
|
2555 |
+
"epoch": 0.07098735170114717,
|
2556 |
+
"grad_norm": 14.592923164367676,
|
2557 |
+
"learning_rate": 0.00016953317848693474,
|
2558 |
+
"loss": 1.5171,
|
2559 |
+
"step": 362
|
2560 |
+
},
|
2561 |
+
{
|
2562 |
+
"epoch": 0.07118344935778018,
|
2563 |
+
"grad_norm": 9.847641944885254,
|
2564 |
+
"learning_rate": 0.00016936926884024864,
|
2565 |
+
"loss": 2.5365,
|
2566 |
+
"step": 363
|
2567 |
+
},
|
2568 |
+
{
|
2569 |
+
"epoch": 0.07137954701441318,
|
2570 |
+
"grad_norm": 5.795401573181152,
|
2571 |
+
"learning_rate": 0.00016920499916401707,
|
2572 |
+
"loss": 1.2201,
|
2573 |
+
"step": 364
|
2574 |
+
},
|
2575 |
+
{
|
2576 |
+
"epoch": 0.07157564467104618,
|
2577 |
+
"grad_norm": 10.178254127502441,
|
2578 |
+
"learning_rate": 0.0001690403703108068,
|
2579 |
+
"loss": 2.3089,
|
2580 |
+
"step": 365
|
2581 |
+
},
|
2582 |
+
{
|
2583 |
+
"epoch": 0.07177174232767919,
|
2584 |
+
"grad_norm": 5.105298042297363,
|
2585 |
+
"learning_rate": 0.00016887538313504883,
|
2586 |
+
"loss": 1.3503,
|
2587 |
+
"step": 366
|
2588 |
+
},
|
2589 |
+
{
|
2590 |
+
"epoch": 0.07196783998431219,
|
2591 |
+
"grad_norm": 5.165500164031982,
|
2592 |
+
"learning_rate": 0.00016871003849303382,
|
2593 |
+
"loss": 1.9088,
|
2594 |
+
"step": 367
|
2595 |
+
},
|
2596 |
+
{
|
2597 |
+
"epoch": 0.0721639376409452,
|
2598 |
+
"grad_norm": 6.532622337341309,
|
2599 |
+
"learning_rate": 0.0001685443372429077,
|
2600 |
+
"loss": 1.2754,
|
2601 |
+
"step": 368
|
2602 |
+
},
|
2603 |
+
{
|
2604 |
+
"epoch": 0.07236003529757819,
|
2605 |
+
"grad_norm": 6.758758068084717,
|
2606 |
+
"learning_rate": 0.0001683782802446672,
|
2607 |
+
"loss": 1.6721,
|
2608 |
+
"step": 369
|
2609 |
+
},
|
2610 |
+
{
|
2611 |
+
"epoch": 0.07255613295421119,
|
2612 |
+
"grad_norm": 4.5956902503967285,
|
2613 |
+
"learning_rate": 0.0001682118683601555,
|
2614 |
+
"loss": 1.7976,
|
2615 |
+
"step": 370
|
2616 |
+
},
|
2617 |
+
{
|
2618 |
+
"epoch": 0.0727522306108442,
|
2619 |
+
"grad_norm": 8.587203025817871,
|
2620 |
+
"learning_rate": 0.00016804510245305745,
|
2621 |
+
"loss": 1.7933,
|
2622 |
+
"step": 371
|
2623 |
+
},
|
2624 |
+
{
|
2625 |
+
"epoch": 0.0729483282674772,
|
2626 |
+
"grad_norm": 8.018026351928711,
|
2627 |
+
"learning_rate": 0.00016787798338889552,
|
2628 |
+
"loss": 2.4964,
|
2629 |
+
"step": 372
|
2630 |
+
},
|
2631 |
+
{
|
2632 |
+
"epoch": 0.0731444259241102,
|
2633 |
+
"grad_norm": 8.505922317504883,
|
2634 |
+
"learning_rate": 0.00016771051203502493,
|
2635 |
+
"loss": 2.9766,
|
2636 |
+
"step": 373
|
2637 |
+
},
|
2638 |
+
{
|
2639 |
+
"epoch": 0.07334052358074321,
|
2640 |
+
"grad_norm": 7.051482200622559,
|
2641 |
+
"learning_rate": 0.00016754268926062938,
|
2642 |
+
"loss": 1.9882,
|
2643 |
+
"step": 374
|
2644 |
+
},
|
2645 |
+
{
|
2646 |
+
"epoch": 0.07353662123737621,
|
2647 |
+
"grad_norm": 4.620705604553223,
|
2648 |
+
"learning_rate": 0.00016737451593671636,
|
2649 |
+
"loss": 2.3198,
|
2650 |
+
"step": 375
|
2651 |
+
},
|
2652 |
+
{
|
2653 |
+
"epoch": 0.07373271889400922,
|
2654 |
+
"grad_norm": 8.614477157592773,
|
2655 |
+
"learning_rate": 0.00016720599293611286,
|
2656 |
+
"loss": 2.0611,
|
2657 |
+
"step": 376
|
2658 |
+
},
|
2659 |
+
{
|
2660 |
+
"epoch": 0.07392881655064222,
|
2661 |
+
"grad_norm": 5.840315818786621,
|
2662 |
+
"learning_rate": 0.0001670371211334606,
|
2663 |
+
"loss": 1.5961,
|
2664 |
+
"step": 377
|
2665 |
+
},
|
2666 |
+
{
|
2667 |
+
"epoch": 0.07412491420727522,
|
2668 |
+
"grad_norm": 6.045500755310059,
|
2669 |
+
"learning_rate": 0.00016686790140521164,
|
2670 |
+
"loss": 1.5132,
|
2671 |
+
"step": 378
|
2672 |
+
},
|
2673 |
+
{
|
2674 |
+
"epoch": 0.07432101186390823,
|
2675 |
+
"grad_norm": 11.521757125854492,
|
2676 |
+
"learning_rate": 0.0001666983346296238,
|
2677 |
+
"loss": 2.0003,
|
2678 |
+
"step": 379
|
2679 |
+
},
|
2680 |
+
{
|
2681 |
+
"epoch": 0.07451710952054123,
|
2682 |
+
"grad_norm": 7.7568230628967285,
|
2683 |
+
"learning_rate": 0.0001665284216867561,
|
2684 |
+
"loss": 2.4289,
|
2685 |
+
"step": 380
|
2686 |
+
},
|
2687 |
+
{
|
2688 |
+
"epoch": 0.07471320717717424,
|
2689 |
+
"grad_norm": 9.928805351257324,
|
2690 |
+
"learning_rate": 0.0001663581634584641,
|
2691 |
+
"loss": 2.2129,
|
2692 |
+
"step": 381
|
2693 |
+
},
|
2694 |
+
{
|
2695 |
+
"epoch": 0.07490930483380724,
|
2696 |
+
"grad_norm": 6.011744022369385,
|
2697 |
+
"learning_rate": 0.00016618756082839554,
|
2698 |
+
"loss": 1.8396,
|
2699 |
+
"step": 382
|
2700 |
+
},
|
2701 |
+
{
|
2702 |
+
"epoch": 0.07510540249044025,
|
2703 |
+
"grad_norm": 7.308403968811035,
|
2704 |
+
"learning_rate": 0.0001660166146819855,
|
2705 |
+
"loss": 0.9335,
|
2706 |
+
"step": 383
|
2707 |
+
},
|
2708 |
+
{
|
2709 |
+
"epoch": 0.07530150014707324,
|
2710 |
+
"grad_norm": 10.000845909118652,
|
2711 |
+
"learning_rate": 0.0001658453259064519,
|
2712 |
+
"loss": 1.7884,
|
2713 |
+
"step": 384
|
2714 |
+
},
|
2715 |
+
{
|
2716 |
+
"epoch": 0.07549759780370624,
|
2717 |
+
"grad_norm": 16.585180282592773,
|
2718 |
+
"learning_rate": 0.00016567369539079114,
|
2719 |
+
"loss": 3.0675,
|
2720 |
+
"step": 385
|
2721 |
+
},
|
2722 |
+
{
|
2723 |
+
"epoch": 0.07569369546033924,
|
2724 |
+
"grad_norm": 10.055068969726562,
|
2725 |
+
"learning_rate": 0.000165501724025773,
|
2726 |
+
"loss": 2.0398,
|
2727 |
+
"step": 386
|
2728 |
+
},
|
2729 |
+
{
|
2730 |
+
"epoch": 0.07588979311697225,
|
2731 |
+
"grad_norm": 9.56424331665039,
|
2732 |
+
"learning_rate": 0.0001653294127039365,
|
2733 |
+
"loss": 3.1253,
|
2734 |
+
"step": 387
|
2735 |
+
},
|
2736 |
+
{
|
2737 |
+
"epoch": 0.07608589077360525,
|
2738 |
+
"grad_norm": 6.628775119781494,
|
2739 |
+
"learning_rate": 0.0001651567623195849,
|
2740 |
+
"loss": 0.869,
|
2741 |
+
"step": 388
|
2742 |
+
},
|
2743 |
+
{
|
2744 |
+
"epoch": 0.07628198843023826,
|
2745 |
+
"grad_norm": 8.333356857299805,
|
2746 |
+
"learning_rate": 0.00016498377376878126,
|
2747 |
+
"loss": 2.1941,
|
2748 |
+
"step": 389
|
2749 |
+
},
|
2750 |
+
{
|
2751 |
+
"epoch": 0.07647808608687126,
|
2752 |
+
"grad_norm": 6.76808500289917,
|
2753 |
+
"learning_rate": 0.0001648104479493437,
|
2754 |
+
"loss": 1.3802,
|
2755 |
+
"step": 390
|
2756 |
+
},
|
2757 |
+
{
|
2758 |
+
"epoch": 0.07667418374350427,
|
2759 |
+
"grad_norm": 7.221822738647461,
|
2760 |
+
"learning_rate": 0.0001646367857608409,
|
2761 |
+
"loss": 1.0955,
|
2762 |
+
"step": 391
|
2763 |
+
},
|
2764 |
+
{
|
2765 |
+
"epoch": 0.07687028140013727,
|
2766 |
+
"grad_norm": 31.469024658203125,
|
2767 |
+
"learning_rate": 0.00016446278810458716,
|
2768 |
+
"loss": 2.4294,
|
2769 |
+
"step": 392
|
2770 |
+
},
|
2771 |
+
{
|
2772 |
+
"epoch": 0.07706637905677027,
|
2773 |
+
"grad_norm": 5.735899448394775,
|
2774 |
+
"learning_rate": 0.000164288455883638,
|
2775 |
+
"loss": 2.1574,
|
2776 |
+
"step": 393
|
2777 |
+
},
|
2778 |
+
{
|
2779 |
+
"epoch": 0.07726247671340328,
|
2780 |
+
"grad_norm": 6.462769031524658,
|
2781 |
+
"learning_rate": 0.00016411379000278524,
|
2782 |
+
"loss": 1.8524,
|
2783 |
+
"step": 394
|
2784 |
+
},
|
2785 |
+
{
|
2786 |
+
"epoch": 0.07745857437003628,
|
2787 |
+
"grad_norm": 12.341320037841797,
|
2788 |
+
"learning_rate": 0.00016393879136855248,
|
2789 |
+
"loss": 2.2439,
|
2790 |
+
"step": 395
|
2791 |
+
},
|
2792 |
+
{
|
2793 |
+
"epoch": 0.07765467202666929,
|
2794 |
+
"grad_norm": 6.433187007904053,
|
2795 |
+
"learning_rate": 0.0001637634608891903,
|
2796 |
+
"loss": 1.2662,
|
2797 |
+
"step": 396
|
2798 |
+
},
|
2799 |
+
{
|
2800 |
+
"epoch": 0.07785076968330229,
|
2801 |
+
"grad_norm": 5.184267997741699,
|
2802 |
+
"learning_rate": 0.00016358779947467158,
|
2803 |
+
"loss": 1.0905,
|
2804 |
+
"step": 397
|
2805 |
+
},
|
2806 |
+
{
|
2807 |
+
"epoch": 0.0780468673399353,
|
2808 |
+
"grad_norm": 8.464557647705078,
|
2809 |
+
"learning_rate": 0.00016341180803668674,
|
2810 |
+
"loss": 2.0666,
|
2811 |
+
"step": 398
|
2812 |
+
},
|
2813 |
+
{
|
2814 |
+
"epoch": 0.07824296499656828,
|
2815 |
+
"grad_norm": 9.498639106750488,
|
2816 |
+
"learning_rate": 0.00016323548748863907,
|
2817 |
+
"loss": 2.2394,
|
2818 |
+
"step": 399
|
2819 |
+
},
|
2820 |
+
{
|
2821 |
+
"epoch": 0.07843906265320129,
|
2822 |
+
"grad_norm": 7.747230052947998,
|
2823 |
+
"learning_rate": 0.00016305883874563994,
|
2824 |
+
"loss": 1.4313,
|
2825 |
+
"step": 400
|
2826 |
+
},
|
2827 |
+
{
|
2828 |
+
"epoch": 0.07863516030983429,
|
2829 |
+
"grad_norm": 8.759115219116211,
|
2830 |
+
"learning_rate": 0.0001628818627245041,
|
2831 |
+
"loss": 1.6428,
|
2832 |
+
"step": 401
|
2833 |
+
},
|
2834 |
+
{
|
2835 |
+
"epoch": 0.0788312579664673,
|
2836 |
+
"grad_norm": 5.250945568084717,
|
2837 |
+
"learning_rate": 0.00016270456034374474,
|
2838 |
+
"loss": 0.7777,
|
2839 |
+
"step": 402
|
2840 |
+
},
|
2841 |
+
{
|
2842 |
+
"epoch": 0.0790273556231003,
|
2843 |
+
"grad_norm": 9.001906394958496,
|
2844 |
+
"learning_rate": 0.00016252693252356916,
|
2845 |
+
"loss": 1.678,
|
2846 |
+
"step": 403
|
2847 |
+
},
|
2848 |
+
{
|
2849 |
+
"epoch": 0.0792234532797333,
|
2850 |
+
"grad_norm": 7.177880764007568,
|
2851 |
+
"learning_rate": 0.00016234898018587337,
|
2852 |
+
"loss": 1.4463,
|
2853 |
+
"step": 404
|
2854 |
+
},
|
2855 |
+
{
|
2856 |
+
"epoch": 0.07941955093636631,
|
2857 |
+
"grad_norm": 10.035067558288574,
|
2858 |
+
"learning_rate": 0.00016217070425423788,
|
2859 |
+
"loss": 2.3035,
|
2860 |
+
"step": 405
|
2861 |
+
},
|
2862 |
+
{
|
2863 |
+
"epoch": 0.07961564859299931,
|
2864 |
+
"grad_norm": 4.974843978881836,
|
2865 |
+
"learning_rate": 0.0001619921056539226,
|
2866 |
+
"loss": 1.2633,
|
2867 |
+
"step": 406
|
2868 |
+
},
|
2869 |
+
{
|
2870 |
+
"epoch": 0.07981174624963232,
|
2871 |
+
"grad_norm": 7.599184036254883,
|
2872 |
+
"learning_rate": 0.00016181318531186206,
|
2873 |
+
"loss": 0.8719,
|
2874 |
+
"step": 407
|
2875 |
+
},
|
2876 |
+
{
|
2877 |
+
"epoch": 0.08000784390626532,
|
2878 |
+
"grad_norm": 11.066450119018555,
|
2879 |
+
"learning_rate": 0.0001616339441566607,
|
2880 |
+
"loss": 2.2938,
|
2881 |
+
"step": 408
|
2882 |
+
},
|
2883 |
+
{
|
2884 |
+
"epoch": 0.08020394156289833,
|
2885 |
+
"grad_norm": 6.441056251525879,
|
2886 |
+
"learning_rate": 0.000161454383118588,
|
2887 |
+
"loss": 0.9166,
|
2888 |
+
"step": 409
|
2889 |
+
},
|
2890 |
+
{
|
2891 |
+
"epoch": 0.08040003921953133,
|
2892 |
+
"grad_norm": 9.670071601867676,
|
2893 |
+
"learning_rate": 0.00016127450312957353,
|
2894 |
+
"loss": 2.2514,
|
2895 |
+
"step": 410
|
2896 |
+
},
|
2897 |
+
{
|
2898 |
+
"epoch": 0.08059613687616433,
|
2899 |
+
"grad_norm": 7.870420455932617,
|
2900 |
+
"learning_rate": 0.00016109430512320237,
|
2901 |
+
"loss": 1.417,
|
2902 |
+
"step": 411
|
2903 |
+
},
|
2904 |
+
{
|
2905 |
+
"epoch": 0.08079223453279734,
|
2906 |
+
"grad_norm": 8.569132804870605,
|
2907 |
+
"learning_rate": 0.00016091379003471007,
|
2908 |
+
"loss": 1.6958,
|
2909 |
+
"step": 412
|
2910 |
+
},
|
2911 |
+
{
|
2912 |
+
"epoch": 0.08098833218943034,
|
2913 |
+
"grad_norm": 4.8630900382995605,
|
2914 |
+
"learning_rate": 0.00016073295880097784,
|
2915 |
+
"loss": 1.8919,
|
2916 |
+
"step": 413
|
2917 |
+
},
|
2918 |
+
{
|
2919 |
+
"epoch": 0.08118442984606333,
|
2920 |
+
"grad_norm": 10.894848823547363,
|
2921 |
+
"learning_rate": 0.0001605518123605277,
|
2922 |
+
"loss": 1.7899,
|
2923 |
+
"step": 414
|
2924 |
+
},
|
2925 |
+
{
|
2926 |
+
"epoch": 0.08138052750269634,
|
2927 |
+
"grad_norm": 6.055443286895752,
|
2928 |
+
"learning_rate": 0.00016037035165351768,
|
2929 |
+
"loss": 1.9432,
|
2930 |
+
"step": 415
|
2931 |
+
},
|
2932 |
+
{
|
2933 |
+
"epoch": 0.08157662515932934,
|
2934 |
+
"grad_norm": 7.106583118438721,
|
2935 |
+
"learning_rate": 0.0001601885776217367,
|
2936 |
+
"loss": 1.736,
|
2937 |
+
"step": 416
|
2938 |
+
},
|
2939 |
+
{
|
2940 |
+
"epoch": 0.08177272281596235,
|
2941 |
+
"grad_norm": 7.7677764892578125,
|
2942 |
+
"learning_rate": 0.00016000649120860003,
|
2943 |
+
"loss": 2.4425,
|
2944 |
+
"step": 417
|
2945 |
+
},
|
2946 |
+
{
|
2947 |
+
"epoch": 0.08196882047259535,
|
2948 |
+
"grad_norm": 5.940224647521973,
|
2949 |
+
"learning_rate": 0.00015982409335914407,
|
2950 |
+
"loss": 1.0954,
|
2951 |
+
"step": 418
|
2952 |
+
},
|
2953 |
+
{
|
2954 |
+
"epoch": 0.08216491812922835,
|
2955 |
+
"grad_norm": 6.025548458099365,
|
2956 |
+
"learning_rate": 0.00015964138502002175,
|
2957 |
+
"loss": 0.93,
|
2958 |
+
"step": 419
|
2959 |
+
},
|
2960 |
+
{
|
2961 |
+
"epoch": 0.08236101578586136,
|
2962 |
+
"grad_norm": 7.528570652008057,
|
2963 |
+
"learning_rate": 0.00015945836713949726,
|
2964 |
+
"loss": 1.8381,
|
2965 |
+
"step": 420
|
2966 |
+
},
|
2967 |
+
{
|
2968 |
+
"epoch": 0.08255711344249436,
|
2969 |
+
"grad_norm": 15.305673599243164,
|
2970 |
+
"learning_rate": 0.00015927504066744148,
|
2971 |
+
"loss": 1.7413,
|
2972 |
+
"step": 421
|
2973 |
+
},
|
2974 |
+
{
|
2975 |
+
"epoch": 0.08275321109912737,
|
2976 |
+
"grad_norm": 4.707043647766113,
|
2977 |
+
"learning_rate": 0.0001590914065553268,
|
2978 |
+
"loss": 1.4186,
|
2979 |
+
"step": 422
|
2980 |
+
},
|
2981 |
+
{
|
2982 |
+
"epoch": 0.08294930875576037,
|
2983 |
+
"grad_norm": 5.808046817779541,
|
2984 |
+
"learning_rate": 0.00015890746575622231,
|
2985 |
+
"loss": 1.095,
|
2986 |
+
"step": 423
|
2987 |
+
},
|
2988 |
+
{
|
2989 |
+
"epoch": 0.08314540641239337,
|
2990 |
+
"grad_norm": 6.294839859008789,
|
2991 |
+
"learning_rate": 0.00015872321922478884,
|
2992 |
+
"loss": 1.918,
|
2993 |
+
"step": 424
|
2994 |
+
},
|
2995 |
+
{
|
2996 |
+
"epoch": 0.08334150406902638,
|
2997 |
+
"grad_norm": 4.934250354766846,
|
2998 |
+
"learning_rate": 0.00015853866791727396,
|
2999 |
+
"loss": 1.2578,
|
3000 |
+
"step": 425
|
3001 |
+
},
|
3002 |
+
{
|
3003 |
+
"epoch": 0.08353760172565938,
|
3004 |
+
"grad_norm": 8.470808982849121,
|
3005 |
+
"learning_rate": 0.00015835381279150705,
|
3006 |
+
"loss": 2.0231,
|
3007 |
+
"step": 426
|
3008 |
+
},
|
3009 |
+
{
|
3010 |
+
"epoch": 0.08373369938229239,
|
3011 |
+
"grad_norm": 18.526151657104492,
|
3012 |
+
"learning_rate": 0.00015816865480689426,
|
3013 |
+
"loss": 4.1691,
|
3014 |
+
"step": 427
|
3015 |
+
},
|
3016 |
+
{
|
3017 |
+
"epoch": 0.08392979703892539,
|
3018 |
+
"grad_norm": 7.958248615264893,
|
3019 |
+
"learning_rate": 0.0001579831949244137,
|
3020 |
+
"loss": 1.066,
|
3021 |
+
"step": 428
|
3022 |
+
},
|
3023 |
+
{
|
3024 |
+
"epoch": 0.08412589469555838,
|
3025 |
+
"grad_norm": 6.250977516174316,
|
3026 |
+
"learning_rate": 0.00015779743410661033,
|
3027 |
+
"loss": 0.8756,
|
3028 |
+
"step": 429
|
3029 |
+
},
|
3030 |
+
{
|
3031 |
+
"epoch": 0.08432199235219139,
|
3032 |
+
"grad_norm": 4.739164352416992,
|
3033 |
+
"learning_rate": 0.00015761137331759084,
|
3034 |
+
"loss": 1.3986,
|
3035 |
+
"step": 430
|
3036 |
+
},
|
3037 |
+
{
|
3038 |
+
"epoch": 0.08451809000882439,
|
3039 |
+
"grad_norm": 10.723671913146973,
|
3040 |
+
"learning_rate": 0.00015742501352301893,
|
3041 |
+
"loss": 2.3962,
|
3042 |
+
"step": 431
|
3043 |
+
},
|
3044 |
+
{
|
3045 |
+
"epoch": 0.0847141876654574,
|
3046 |
+
"grad_norm": 5.097165584564209,
|
3047 |
+
"learning_rate": 0.00015723835569011007,
|
3048 |
+
"loss": 3.1665,
|
3049 |
+
"step": 432
|
3050 |
+
},
|
3051 |
+
{
|
3052 |
+
"epoch": 0.0849102853220904,
|
3053 |
+
"grad_norm": 7.0065388679504395,
|
3054 |
+
"learning_rate": 0.00015705140078762665,
|
3055 |
+
"loss": 1.4634,
|
3056 |
+
"step": 433
|
3057 |
+
},
|
3058 |
+
{
|
3059 |
+
"epoch": 0.0851063829787234,
|
3060 |
+
"grad_norm": 6.814738750457764,
|
3061 |
+
"learning_rate": 0.00015686414978587277,
|
3062 |
+
"loss": 0.9439,
|
3063 |
+
"step": 434
|
3064 |
+
},
|
3065 |
+
{
|
3066 |
+
"epoch": 0.0853024806353564,
|
3067 |
+
"grad_norm": 8.381726264953613,
|
3068 |
+
"learning_rate": 0.0001566766036566893,
|
3069 |
+
"loss": 1.6046,
|
3070 |
+
"step": 435
|
3071 |
+
},
|
3072 |
+
{
|
3073 |
+
"epoch": 0.08549857829198941,
|
3074 |
+
"grad_norm": 8.249631881713867,
|
3075 |
+
"learning_rate": 0.00015648876337344896,
|
3076 |
+
"loss": 1.6886,
|
3077 |
+
"step": 436
|
3078 |
+
},
|
3079 |
+
{
|
3080 |
+
"epoch": 0.08569467594862241,
|
3081 |
+
"grad_norm": 9.020310401916504,
|
3082 |
+
"learning_rate": 0.00015630062991105098,
|
3083 |
+
"loss": 2.5881,
|
3084 |
+
"step": 437
|
3085 |
+
},
|
3086 |
+
{
|
3087 |
+
"epoch": 0.08589077360525542,
|
3088 |
+
"grad_norm": 3.999058723449707,
|
3089 |
+
"learning_rate": 0.0001561122042459163,
|
3090 |
+
"loss": 1.8363,
|
3091 |
+
"step": 438
|
3092 |
+
},
|
3093 |
+
{
|
3094 |
+
"epoch": 0.08608687126188842,
|
3095 |
+
"grad_norm": 7.487119674682617,
|
3096 |
+
"learning_rate": 0.00015592348735598237,
|
3097 |
+
"loss": 1.4798,
|
3098 |
+
"step": 439
|
3099 |
+
},
|
3100 |
+
{
|
3101 |
+
"epoch": 0.08628296891852143,
|
3102 |
+
"grad_norm": 4.982146739959717,
|
3103 |
+
"learning_rate": 0.00015573448022069815,
|
3104 |
+
"loss": 2.7132,
|
3105 |
+
"step": 440
|
3106 |
+
},
|
3107 |
+
{
|
3108 |
+
"epoch": 0.08647906657515443,
|
3109 |
+
"grad_norm": 3.5383474826812744,
|
3110 |
+
"learning_rate": 0.00015554518382101892,
|
3111 |
+
"loss": 0.6142,
|
3112 |
+
"step": 441
|
3113 |
+
},
|
3114 |
+
{
|
3115 |
+
"epoch": 0.08667516423178744,
|
3116 |
+
"grad_norm": 3.881410598754883,
|
3117 |
+
"learning_rate": 0.00015535559913940126,
|
3118 |
+
"loss": 0.8708,
|
3119 |
+
"step": 442
|
3120 |
+
},
|
3121 |
+
{
|
3122 |
+
"epoch": 0.08687126188842044,
|
3123 |
+
"grad_norm": 5.159698963165283,
|
3124 |
+
"learning_rate": 0.00015516572715979806,
|
3125 |
+
"loss": 2.32,
|
3126 |
+
"step": 443
|
3127 |
+
},
|
3128 |
+
{
|
3129 |
+
"epoch": 0.08706735954505343,
|
3130 |
+
"grad_norm": 9.29107666015625,
|
3131 |
+
"learning_rate": 0.00015497556886765316,
|
3132 |
+
"loss": 2.5744,
|
3133 |
+
"step": 444
|
3134 |
+
},
|
3135 |
+
{
|
3136 |
+
"epoch": 0.08726345720168643,
|
3137 |
+
"grad_norm": 8.70083999633789,
|
3138 |
+
"learning_rate": 0.00015478512524989645,
|
3139 |
+
"loss": 0.8412,
|
3140 |
+
"step": 445
|
3141 |
+
},
|
3142 |
+
{
|
3143 |
+
"epoch": 0.08745955485831944,
|
3144 |
+
"grad_norm": 6.164897918701172,
|
3145 |
+
"learning_rate": 0.00015459439729493865,
|
3146 |
+
"loss": 1.4857,
|
3147 |
+
"step": 446
|
3148 |
+
},
|
3149 |
+
{
|
3150 |
+
"epoch": 0.08765565251495244,
|
3151 |
+
"grad_norm": 4.213920593261719,
|
3152 |
+
"learning_rate": 0.00015440338599266622,
|
3153 |
+
"loss": 1.2077,
|
3154 |
+
"step": 447
|
3155 |
+
},
|
3156 |
+
{
|
3157 |
+
"epoch": 0.08785175017158545,
|
3158 |
+
"grad_norm": 14.897852897644043,
|
3159 |
+
"learning_rate": 0.00015421209233443617,
|
3160 |
+
"loss": 1.815,
|
3161 |
+
"step": 448
|
3162 |
+
},
|
3163 |
+
{
|
3164 |
+
"epoch": 0.08804784782821845,
|
3165 |
+
"grad_norm": 6.773970127105713,
|
3166 |
+
"learning_rate": 0.00015402051731307093,
|
3167 |
+
"loss": 2.9485,
|
3168 |
+
"step": 449
|
3169 |
+
},
|
3170 |
+
{
|
3171 |
+
"epoch": 0.08824394548485145,
|
3172 |
+
"grad_norm": 5.569252014160156,
|
3173 |
+
"learning_rate": 0.0001538286619228533,
|
3174 |
+
"loss": 1.1505,
|
3175 |
+
"step": 450
|
3176 |
+
},
|
3177 |
+
{
|
3178 |
+
"epoch": 0.08844004314148446,
|
3179 |
+
"grad_norm": 5.61276388168335,
|
3180 |
+
"learning_rate": 0.0001536365271595212,
|
3181 |
+
"loss": 1.2065,
|
3182 |
+
"step": 451
|
3183 |
+
},
|
3184 |
+
{
|
3185 |
+
"epoch": 0.08863614079811746,
|
3186 |
+
"grad_norm": 3.9499528408050537,
|
3187 |
+
"learning_rate": 0.00015344411402026245,
|
3188 |
+
"loss": 1.3547,
|
3189 |
+
"step": 452
|
3190 |
+
},
|
3191 |
+
{
|
3192 |
+
"epoch": 0.08883223845475047,
|
3193 |
+
"grad_norm": 11.833647727966309,
|
3194 |
+
"learning_rate": 0.00015325142350370967,
|
3195 |
+
"loss": 2.8476,
|
3196 |
+
"step": 453
|
3197 |
+
},
|
3198 |
+
{
|
3199 |
+
"epoch": 0.08902833611138347,
|
3200 |
+
"grad_norm": 6.281819820404053,
|
3201 |
+
"learning_rate": 0.00015305845660993503,
|
3202 |
+
"loss": 1.0563,
|
3203 |
+
"step": 454
|
3204 |
+
},
|
3205 |
+
{
|
3206 |
+
"epoch": 0.08922443376801648,
|
3207 |
+
"grad_norm": 9.338071823120117,
|
3208 |
+
"learning_rate": 0.00015286521434044526,
|
3209 |
+
"loss": 1.9847,
|
3210 |
+
"step": 455
|
3211 |
+
},
|
3212 |
+
{
|
3213 |
+
"epoch": 0.08942053142464948,
|
3214 |
+
"grad_norm": 12.811955451965332,
|
3215 |
+
"learning_rate": 0.0001526716976981761,
|
3216 |
+
"loss": 1.2632,
|
3217 |
+
"step": 456
|
3218 |
+
},
|
3219 |
+
{
|
3220 |
+
"epoch": 0.08961662908128248,
|
3221 |
+
"grad_norm": 5.077617168426514,
|
3222 |
+
"learning_rate": 0.0001524779076874875,
|
3223 |
+
"loss": 0.9728,
|
3224 |
+
"step": 457
|
3225 |
+
},
|
3226 |
+
{
|
3227 |
+
"epoch": 0.08981272673791549,
|
3228 |
+
"grad_norm": 5.802744388580322,
|
3229 |
+
"learning_rate": 0.0001522838453141581,
|
3230 |
+
"loss": 0.9826,
|
3231 |
+
"step": 458
|
3232 |
+
},
|
3233 |
+
{
|
3234 |
+
"epoch": 0.09000882439454848,
|
3235 |
+
"grad_norm": 6.643836498260498,
|
3236 |
+
"learning_rate": 0.00015208951158538004,
|
3237 |
+
"loss": 1.3868,
|
3238 |
+
"step": 459
|
3239 |
+
},
|
3240 |
+
{
|
3241 |
+
"epoch": 0.09020492205118148,
|
3242 |
+
"grad_norm": 8.100834846496582,
|
3243 |
+
"learning_rate": 0.000151894907509754,
|
3244 |
+
"loss": 1.5572,
|
3245 |
+
"step": 460
|
3246 |
+
},
|
3247 |
+
{
|
3248 |
+
"epoch": 0.09040101970781449,
|
3249 |
+
"grad_norm": 8.695015907287598,
|
3250 |
+
"learning_rate": 0.00015170003409728356,
|
3251 |
+
"loss": 1.7788,
|
3252 |
+
"step": 461
|
3253 |
+
},
|
3254 |
+
{
|
3255 |
+
"epoch": 0.09059711736444749,
|
3256 |
+
"grad_norm": 11.81883430480957,
|
3257 |
+
"learning_rate": 0.00015150489235937035,
|
3258 |
+
"loss": 1.7457,
|
3259 |
+
"step": 462
|
3260 |
+
},
|
3261 |
+
{
|
3262 |
+
"epoch": 0.0907932150210805,
|
3263 |
+
"grad_norm": 8.165855407714844,
|
3264 |
+
"learning_rate": 0.00015130948330880847,
|
3265 |
+
"loss": 1.3886,
|
3266 |
+
"step": 463
|
3267 |
+
},
|
3268 |
+
{
|
3269 |
+
"epoch": 0.0909893126777135,
|
3270 |
+
"grad_norm": 22.797456741333008,
|
3271 |
+
"learning_rate": 0.00015111380795977954,
|
3272 |
+
"loss": 1.8671,
|
3273 |
+
"step": 464
|
3274 |
+
},
|
3275 |
+
{
|
3276 |
+
"epoch": 0.0911854103343465,
|
3277 |
+
"grad_norm": 5.275067329406738,
|
3278 |
+
"learning_rate": 0.00015091786732784716,
|
3279 |
+
"loss": 0.7186,
|
3280 |
+
"step": 465
|
3281 |
+
},
|
3282 |
+
{
|
3283 |
+
"epoch": 0.09138150799097951,
|
3284 |
+
"grad_norm": 8.38714599609375,
|
3285 |
+
"learning_rate": 0.00015072166242995175,
|
3286 |
+
"loss": 1.7764,
|
3287 |
+
"step": 466
|
3288 |
+
},
|
3289 |
+
{
|
3290 |
+
"epoch": 0.09157760564761251,
|
3291 |
+
"grad_norm": 8.916413307189941,
|
3292 |
+
"learning_rate": 0.0001505251942844054,
|
3293 |
+
"loss": 1.5589,
|
3294 |
+
"step": 467
|
3295 |
+
},
|
3296 |
+
{
|
3297 |
+
"epoch": 0.09177370330424552,
|
3298 |
+
"grad_norm": 5.007913112640381,
|
3299 |
+
"learning_rate": 0.00015032846391088635,
|
3300 |
+
"loss": 0.6688,
|
3301 |
+
"step": 468
|
3302 |
+
},
|
3303 |
+
{
|
3304 |
+
"epoch": 0.09196980096087852,
|
3305 |
+
"grad_norm": 10.685876846313477,
|
3306 |
+
"learning_rate": 0.0001501314723304339,
|
3307 |
+
"loss": 2.2118,
|
3308 |
+
"step": 469
|
3309 |
+
},
|
3310 |
+
{
|
3311 |
+
"epoch": 0.09216589861751152,
|
3312 |
+
"grad_norm": 7.626431465148926,
|
3313 |
+
"learning_rate": 0.00014993422056544295,
|
3314 |
+
"loss": 1.1849,
|
3315 |
+
"step": 470
|
3316 |
+
},
|
3317 |
+
{
|
3318 |
+
"epoch": 0.09236199627414453,
|
3319 |
+
"grad_norm": 5.862109184265137,
|
3320 |
+
"learning_rate": 0.00014973670963965883,
|
3321 |
+
"loss": 1.4594,
|
3322 |
+
"step": 471
|
3323 |
+
},
|
3324 |
+
{
|
3325 |
+
"epoch": 0.09255809393077753,
|
3326 |
+
"grad_norm": 5.194091320037842,
|
3327 |
+
"learning_rate": 0.00014953894057817188,
|
3328 |
+
"loss": 1.2243,
|
3329 |
+
"step": 472
|
3330 |
+
},
|
3331 |
+
{
|
3332 |
+
"epoch": 0.09275419158741054,
|
3333 |
+
"grad_norm": 7.806075572967529,
|
3334 |
+
"learning_rate": 0.0001493409144074122,
|
3335 |
+
"loss": 1.2365,
|
3336 |
+
"step": 473
|
3337 |
+
},
|
3338 |
+
{
|
3339 |
+
"epoch": 0.09295028924404353,
|
3340 |
+
"grad_norm": 5.004404067993164,
|
3341 |
+
"learning_rate": 0.00014914263215514431,
|
3342 |
+
"loss": 1.6081,
|
3343 |
+
"step": 474
|
3344 |
+
},
|
3345 |
+
{
|
3346 |
+
"epoch": 0.09314638690067653,
|
3347 |
+
"grad_norm": 5.245405197143555,
|
3348 |
+
"learning_rate": 0.00014894409485046177,
|
3349 |
+
"loss": 2.2626,
|
3350 |
+
"step": 475
|
3351 |
+
},
|
3352 |
+
{
|
3353 |
+
"epoch": 0.09334248455730954,
|
3354 |
+
"grad_norm": 6.017452716827393,
|
3355 |
+
"learning_rate": 0.00014874530352378194,
|
3356 |
+
"loss": 3.0452,
|
3357 |
+
"step": 476
|
3358 |
+
},
|
3359 |
+
{
|
3360 |
+
"epoch": 0.09353858221394254,
|
3361 |
+
"grad_norm": 8.504718780517578,
|
3362 |
+
"learning_rate": 0.00014854625920684042,
|
3363 |
+
"loss": 1.4529,
|
3364 |
+
"step": 477
|
3365 |
+
},
|
3366 |
+
{
|
3367 |
+
"epoch": 0.09373467987057554,
|
3368 |
+
"grad_norm": 7.850302696228027,
|
3369 |
+
"learning_rate": 0.00014834696293268603,
|
3370 |
+
"loss": 1.5511,
|
3371 |
+
"step": 478
|
3372 |
+
},
|
3373 |
+
{
|
3374 |
+
"epoch": 0.09393077752720855,
|
3375 |
+
"grad_norm": 6.846382141113281,
|
3376 |
+
"learning_rate": 0.00014814741573567514,
|
3377 |
+
"loss": 2.0178,
|
3378 |
+
"step": 479
|
3379 |
+
},
|
3380 |
+
{
|
3381 |
+
"epoch": 0.09412687518384155,
|
3382 |
+
"grad_norm": 8.834015846252441,
|
3383 |
+
"learning_rate": 0.00014794761865146648,
|
3384 |
+
"loss": 1.6438,
|
3385 |
+
"step": 480
|
3386 |
+
},
|
3387 |
+
{
|
3388 |
+
"epoch": 0.09432297284047456,
|
3389 |
+
"grad_norm": 7.127365589141846,
|
3390 |
+
"learning_rate": 0.00014774757271701557,
|
3391 |
+
"loss": 0.5588,
|
3392 |
+
"step": 481
|
3393 |
+
},
|
3394 |
+
{
|
3395 |
+
"epoch": 0.09451907049710756,
|
3396 |
+
"grad_norm": 13.859881401062012,
|
3397 |
+
"learning_rate": 0.00014754727897056967,
|
3398 |
+
"loss": 2.9212,
|
3399 |
+
"step": 482
|
3400 |
+
},
|
3401 |
+
{
|
3402 |
+
"epoch": 0.09471516815374056,
|
3403 |
+
"grad_norm": 5.914462089538574,
|
3404 |
+
"learning_rate": 0.0001473467384516621,
|
3405 |
+
"loss": 1.334,
|
3406 |
+
"step": 483
|
3407 |
+
},
|
3408 |
+
{
|
3409 |
+
"epoch": 0.09491126581037357,
|
3410 |
+
"grad_norm": 3.947435140609741,
|
3411 |
+
"learning_rate": 0.0001471459522011069,
|
3412 |
+
"loss": 1.5284,
|
3413 |
+
"step": 484
|
3414 |
+
},
|
3415 |
+
{
|
3416 |
+
"epoch": 0.09510736346700657,
|
3417 |
+
"grad_norm": 7.171452522277832,
|
3418 |
+
"learning_rate": 0.00014694492126099353,
|
3419 |
+
"loss": 1.7676,
|
3420 |
+
"step": 485
|
3421 |
+
},
|
3422 |
+
{
|
3423 |
+
"epoch": 0.09530346112363958,
|
3424 |
+
"grad_norm": 7.364241600036621,
|
3425 |
+
"learning_rate": 0.0001467436466746814,
|
3426 |
+
"loss": 1.7551,
|
3427 |
+
"step": 486
|
3428 |
+
},
|
3429 |
+
{
|
3430 |
+
"epoch": 0.09549955878027258,
|
3431 |
+
"grad_norm": 12.173360824584961,
|
3432 |
+
"learning_rate": 0.0001465421294867944,
|
3433 |
+
"loss": 2.5703,
|
3434 |
+
"step": 487
|
3435 |
+
},
|
3436 |
+
{
|
3437 |
+
"epoch": 0.09569565643690559,
|
3438 |
+
"grad_norm": 14.59663200378418,
|
3439 |
+
"learning_rate": 0.00014634037074321557,
|
3440 |
+
"loss": 1.7594,
|
3441 |
+
"step": 488
|
3442 |
+
},
|
3443 |
+
{
|
3444 |
+
"epoch": 0.09589175409353858,
|
3445 |
+
"grad_norm": 5.692611217498779,
|
3446 |
+
"learning_rate": 0.00014613837149108163,
|
3447 |
+
"loss": 2.5973,
|
3448 |
+
"step": 489
|
3449 |
+
},
|
3450 |
+
{
|
3451 |
+
"epoch": 0.09608785175017158,
|
3452 |
+
"grad_norm": 6.9588470458984375,
|
3453 |
+
"learning_rate": 0.00014593613277877758,
|
3454 |
+
"loss": 2.6119,
|
3455 |
+
"step": 490
|
3456 |
+
},
|
3457 |
+
{
|
3458 |
+
"epoch": 0.09628394940680458,
|
3459 |
+
"grad_norm": 14.370820999145508,
|
3460 |
+
"learning_rate": 0.0001457336556559312,
|
3461 |
+
"loss": 1.9495,
|
3462 |
+
"step": 491
|
3463 |
+
},
|
3464 |
+
{
|
3465 |
+
"epoch": 0.09648004706343759,
|
3466 |
+
"grad_norm": 8.259851455688477,
|
3467 |
+
"learning_rate": 0.0001455309411734076,
|
3468 |
+
"loss": 2.2663,
|
3469 |
+
"step": 492
|
3470 |
+
},
|
3471 |
+
{
|
3472 |
+
"epoch": 0.09667614472007059,
|
3473 |
+
"grad_norm": 11.025699615478516,
|
3474 |
+
"learning_rate": 0.00014532799038330385,
|
3475 |
+
"loss": 2.2184,
|
3476 |
+
"step": 493
|
3477 |
+
},
|
3478 |
+
{
|
3479 |
+
"epoch": 0.0968722423767036,
|
3480 |
+
"grad_norm": 6.014841079711914,
|
3481 |
+
"learning_rate": 0.00014512480433894343,
|
3482 |
+
"loss": 2.3436,
|
3483 |
+
"step": 494
|
3484 |
+
},
|
3485 |
+
{
|
3486 |
+
"epoch": 0.0970683400333366,
|
3487 |
+
"grad_norm": 11.881712913513184,
|
3488 |
+
"learning_rate": 0.00014492138409487085,
|
3489 |
+
"loss": 0.7869,
|
3490 |
+
"step": 495
|
3491 |
+
},
|
3492 |
+
{
|
3493 |
+
"epoch": 0.0972644376899696,
|
3494 |
+
"grad_norm": 4.720561504364014,
|
3495 |
+
"learning_rate": 0.00014471773070684599,
|
3496 |
+
"loss": 2.417,
|
3497 |
+
"step": 496
|
3498 |
+
},
|
3499 |
+
{
|
3500 |
+
"epoch": 0.09746053534660261,
|
3501 |
+
"grad_norm": 11.112417221069336,
|
3502 |
+
"learning_rate": 0.00014451384523183903,
|
3503 |
+
"loss": 2.155,
|
3504 |
+
"step": 497
|
3505 |
+
},
|
3506 |
+
{
|
3507 |
+
"epoch": 0.09765663300323561,
|
3508 |
+
"grad_norm": 5.968791961669922,
|
3509 |
+
"learning_rate": 0.0001443097287280244,
|
3510 |
+
"loss": 0.8096,
|
3511 |
+
"step": 498
|
3512 |
+
},
|
3513 |
+
{
|
3514 |
+
"epoch": 0.09785273065986862,
|
3515 |
+
"grad_norm": 7.543048858642578,
|
3516 |
+
"learning_rate": 0.0001441053822547757,
|
3517 |
+
"loss": 1.7723,
|
3518 |
+
"step": 499
|
3519 |
+
},
|
3520 |
+
{
|
3521 |
+
"epoch": 0.09804882831650162,
|
3522 |
+
"grad_norm": 7.892668724060059,
|
3523 |
+
"learning_rate": 0.00014390080687266013,
|
3524 |
+
"loss": 1.4027,
|
3525 |
+
"step": 500
|
3526 |
+
},
|
3527 |
+
{
|
3528 |
+
"epoch": 0.09824492597313463,
|
3529 |
+
"grad_norm": 7.690077304840088,
|
3530 |
+
"learning_rate": 0.00014369600364343285,
|
3531 |
+
"loss": 1.901,
|
3532 |
+
"step": 501
|
3533 |
+
},
|
3534 |
+
{
|
3535 |
+
"epoch": 0.09844102362976763,
|
3536 |
+
"grad_norm": 4.714064598083496,
|
3537 |
+
"learning_rate": 0.00014349097363003163,
|
3538 |
+
"loss": 1.6198,
|
3539 |
+
"step": 502
|
3540 |
+
},
|
3541 |
+
{
|
3542 |
+
"epoch": 0.09863712128640063,
|
3543 |
+
"grad_norm": 4.764143466949463,
|
3544 |
+
"learning_rate": 0.0001432857178965712,
|
3545 |
+
"loss": 1.3231,
|
3546 |
+
"step": 503
|
3547 |
+
},
|
3548 |
+
{
|
3549 |
+
"epoch": 0.09883321894303362,
|
3550 |
+
"grad_norm": 9.042250633239746,
|
3551 |
+
"learning_rate": 0.00014308023750833783,
|
3552 |
+
"loss": 2.1337,
|
3553 |
+
"step": 504
|
3554 |
+
},
|
3555 |
+
{
|
3556 |
+
"epoch": 0.09902931659966663,
|
3557 |
+
"grad_norm": 7.619399547576904,
|
3558 |
+
"learning_rate": 0.00014287453353178372,
|
3559 |
+
"loss": 1.1525,
|
3560 |
+
"step": 505
|
3561 |
+
},
|
3562 |
+
{
|
3563 |
+
"epoch": 0.09922541425629963,
|
3564 |
+
"grad_norm": 10.186138153076172,
|
3565 |
+
"learning_rate": 0.00014266860703452156,
|
3566 |
+
"loss": 2.5917,
|
3567 |
+
"step": 506
|
3568 |
+
},
|
3569 |
+
{
|
3570 |
+
"epoch": 0.09942151191293264,
|
3571 |
+
"grad_norm": 6.784362316131592,
|
3572 |
+
"learning_rate": 0.00014246245908531882,
|
3573 |
+
"loss": 2.1194,
|
3574 |
+
"step": 507
|
3575 |
+
},
|
3576 |
+
{
|
3577 |
+
"epoch": 0.09961760956956564,
|
3578 |
+
"grad_norm": 46.69243621826172,
|
3579 |
+
"learning_rate": 0.0001422560907540925,
|
3580 |
+
"loss": 2.0196,
|
3581 |
+
"step": 508
|
3582 |
+
},
|
3583 |
+
{
|
3584 |
+
"epoch": 0.09981370722619864,
|
3585 |
+
"grad_norm": 5.98226261138916,
|
3586 |
+
"learning_rate": 0.00014204950311190318,
|
3587 |
+
"loss": 2.131,
|
3588 |
+
"step": 509
|
3589 |
+
},
|
3590 |
+
{
|
3591 |
+
"epoch": 0.10000980488283165,
|
3592 |
+
"grad_norm": 19.974212646484375,
|
3593 |
+
"learning_rate": 0.00014184269723094988,
|
3594 |
+
"loss": 1.9812,
|
3595 |
+
"step": 510
|
3596 |
+
},
|
3597 |
+
{
|
3598 |
+
"epoch": 0.10020590253946465,
|
3599 |
+
"grad_norm": 6.612239360809326,
|
3600 |
+
"learning_rate": 0.00014163567418456406,
|
3601 |
+
"loss": 1.2688,
|
3602 |
+
"step": 511
|
3603 |
+
},
|
3604 |
+
{
|
3605 |
+
"epoch": 0.10040200019609766,
|
3606 |
+
"grad_norm": 5.024367332458496,
|
3607 |
+
"learning_rate": 0.0001414284350472045,
|
3608 |
+
"loss": 1.0559,
|
3609 |
+
"step": 512
|
3610 |
+
},
|
3611 |
+
{
|
3612 |
+
"epoch": 0.10059809785273066,
|
3613 |
+
"grad_norm": 6.517631530761719,
|
3614 |
+
"learning_rate": 0.00014122098089445142,
|
3615 |
+
"loss": 1.2611,
|
3616 |
+
"step": 513
|
3617 |
+
},
|
3618 |
+
{
|
3619 |
+
"epoch": 0.10079419550936367,
|
3620 |
+
"grad_norm": 7.715580463409424,
|
3621 |
+
"learning_rate": 0.0001410133128030009,
|
3622 |
+
"loss": 1.0419,
|
3623 |
+
"step": 514
|
3624 |
+
},
|
3625 |
+
{
|
3626 |
+
"epoch": 0.10099029316599667,
|
3627 |
+
"grad_norm": 5.8916120529174805,
|
3628 |
+
"learning_rate": 0.00014080543185065943,
|
3629 |
+
"loss": 0.938,
|
3630 |
+
"step": 515
|
3631 |
+
},
|
3632 |
+
{
|
3633 |
+
"epoch": 0.10118639082262967,
|
3634 |
+
"grad_norm": 5.587378978729248,
|
3635 |
+
"learning_rate": 0.0001405973391163383,
|
3636 |
+
"loss": 1.4988,
|
3637 |
+
"step": 516
|
3638 |
+
},
|
3639 |
+
{
|
3640 |
+
"epoch": 0.10138248847926268,
|
3641 |
+
"grad_norm": 7.234095096588135,
|
3642 |
+
"learning_rate": 0.0001403890356800479,
|
3643 |
+
"loss": 2.0476,
|
3644 |
+
"step": 517
|
3645 |
+
},
|
3646 |
+
{
|
3647 |
+
"epoch": 0.10157858613589568,
|
3648 |
+
"grad_norm": 5.781263828277588,
|
3649 |
+
"learning_rate": 0.00014018052262289223,
|
3650 |
+
"loss": 1.9736,
|
3651 |
+
"step": 518
|
3652 |
+
},
|
3653 |
+
{
|
3654 |
+
"epoch": 0.10177468379252867,
|
3655 |
+
"grad_norm": 7.789620399475098,
|
3656 |
+
"learning_rate": 0.0001399718010270632,
|
3657 |
+
"loss": 1.2162,
|
3658 |
+
"step": 519
|
3659 |
+
},
|
3660 |
+
{
|
3661 |
+
"epoch": 0.10197078144916168,
|
3662 |
+
"grad_norm": 2.692659378051758,
|
3663 |
+
"learning_rate": 0.00013976287197583494,
|
3664 |
+
"loss": 0.391,
|
3665 |
+
"step": 520
|
3666 |
+
},
|
3667 |
+
{
|
3668 |
+
"epoch": 0.10216687910579468,
|
3669 |
+
"grad_norm": 7.2949676513671875,
|
3670 |
+
"learning_rate": 0.0001395537365535585,
|
3671 |
+
"loss": 1.2454,
|
3672 |
+
"step": 521
|
3673 |
+
},
|
3674 |
+
{
|
3675 |
+
"epoch": 0.10236297676242768,
|
3676 |
+
"grad_norm": 14.400397300720215,
|
3677 |
+
"learning_rate": 0.00013934439584565583,
|
3678 |
+
"loss": 1.844,
|
3679 |
+
"step": 522
|
3680 |
+
},
|
3681 |
+
{
|
3682 |
+
"epoch": 0.10255907441906069,
|
3683 |
+
"grad_norm": 5.972201347351074,
|
3684 |
+
"learning_rate": 0.0001391348509386144,
|
3685 |
+
"loss": 1.2853,
|
3686 |
+
"step": 523
|
3687 |
+
},
|
3688 |
+
{
|
3689 |
+
"epoch": 0.1027551720756937,
|
3690 |
+
"grad_norm": 8.810315132141113,
|
3691 |
+
"learning_rate": 0.00013892510291998146,
|
3692 |
+
"loss": 2.7965,
|
3693 |
+
"step": 524
|
3694 |
+
},
|
3695 |
+
{
|
3696 |
+
"epoch": 0.1029512697323267,
|
3697 |
+
"grad_norm": 8.922815322875977,
|
3698 |
+
"learning_rate": 0.00013871515287835839,
|
3699 |
+
"loss": 1.2606,
|
3700 |
+
"step": 525
|
3701 |
+
},
|
3702 |
+
{
|
3703 |
+
"epoch": 0.1031473673889597,
|
3704 |
+
"grad_norm": 7.5427021980285645,
|
3705 |
+
"learning_rate": 0.00013850500190339514,
|
3706 |
+
"loss": 1.0504,
|
3707 |
+
"step": 526
|
3708 |
+
},
|
3709 |
+
{
|
3710 |
+
"epoch": 0.1033434650455927,
|
3711 |
+
"grad_norm": 6.646986484527588,
|
3712 |
+
"learning_rate": 0.00013829465108578445,
|
3713 |
+
"loss": 1.1522,
|
3714 |
+
"step": 527
|
3715 |
+
},
|
3716 |
+
{
|
3717 |
+
"epoch": 0.10353956270222571,
|
3718 |
+
"grad_norm": 8.387070655822754,
|
3719 |
+
"learning_rate": 0.0001380841015172563,
|
3720 |
+
"loss": 1.4298,
|
3721 |
+
"step": 528
|
3722 |
+
},
|
3723 |
+
{
|
3724 |
+
"epoch": 0.10373566035885871,
|
3725 |
+
"grad_norm": 21.383251190185547,
|
3726 |
+
"learning_rate": 0.0001378733542905722,
|
3727 |
+
"loss": 1.9162,
|
3728 |
+
"step": 529
|
3729 |
+
},
|
3730 |
+
{
|
3731 |
+
"epoch": 0.10393175801549172,
|
3732 |
+
"grad_norm": 5.776891231536865,
|
3733 |
+
"learning_rate": 0.00013766241049951948,
|
3734 |
+
"loss": 1.2837,
|
3735 |
+
"step": 530
|
3736 |
+
},
|
3737 |
+
{
|
3738 |
+
"epoch": 0.10412785567212472,
|
3739 |
+
"grad_norm": 4.5066094398498535,
|
3740 |
+
"learning_rate": 0.00013745127123890565,
|
3741 |
+
"loss": 1.5433,
|
3742 |
+
"step": 531
|
3743 |
+
},
|
3744 |
+
{
|
3745 |
+
"epoch": 0.10432395332875773,
|
3746 |
+
"grad_norm": 6.012543201446533,
|
3747 |
+
"learning_rate": 0.00013723993760455272,
|
3748 |
+
"loss": 1.3534,
|
3749 |
+
"step": 532
|
3750 |
+
},
|
3751 |
+
{
|
3752 |
+
"epoch": 0.10452005098539073,
|
3753 |
+
"grad_norm": 8.182510375976562,
|
3754 |
+
"learning_rate": 0.0001370284106932915,
|
3755 |
+
"loss": 1.5422,
|
3756 |
+
"step": 533
|
3757 |
+
},
|
3758 |
+
{
|
3759 |
+
"epoch": 0.10471614864202373,
|
3760 |
+
"grad_norm": 12.530463218688965,
|
3761 |
+
"learning_rate": 0.00013681669160295597,
|
3762 |
+
"loss": 1.651,
|
3763 |
+
"step": 534
|
3764 |
+
},
|
3765 |
+
{
|
3766 |
+
"epoch": 0.10491224629865673,
|
3767 |
+
"grad_norm": 8.226496696472168,
|
3768 |
+
"learning_rate": 0.00013660478143237746,
|
3769 |
+
"loss": 1.8992,
|
3770 |
+
"step": 535
|
3771 |
+
},
|
3772 |
+
{
|
3773 |
+
"epoch": 0.10510834395528973,
|
3774 |
+
"grad_norm": 4.896340847015381,
|
3775 |
+
"learning_rate": 0.00013639268128137907,
|
3776 |
+
"loss": 2.2539,
|
3777 |
+
"step": 536
|
3778 |
+
},
|
3779 |
+
{
|
3780 |
+
"epoch": 0.10530444161192273,
|
3781 |
+
"grad_norm": 4.573820114135742,
|
3782 |
+
"learning_rate": 0.00013618039225076986,
|
3783 |
+
"loss": 2.4075,
|
3784 |
+
"step": 537
|
3785 |
+
},
|
3786 |
+
{
|
3787 |
+
"epoch": 0.10550053926855574,
|
3788 |
+
"grad_norm": 4.920849323272705,
|
3789 |
+
"learning_rate": 0.0001359679154423392,
|
3790 |
+
"loss": 2.2481,
|
3791 |
+
"step": 538
|
3792 |
+
},
|
3793 |
+
{
|
3794 |
+
"epoch": 0.10569663692518874,
|
3795 |
+
"grad_norm": 6.206282138824463,
|
3796 |
+
"learning_rate": 0.00013575525195885107,
|
3797 |
+
"loss": 2.5488,
|
3798 |
+
"step": 539
|
3799 |
+
},
|
3800 |
+
{
|
3801 |
+
"epoch": 0.10589273458182175,
|
3802 |
+
"grad_norm": 9.469644546508789,
|
3803 |
+
"learning_rate": 0.0001355424029040382,
|
3804 |
+
"loss": 2.0894,
|
3805 |
+
"step": 540
|
3806 |
+
},
|
3807 |
+
{
|
3808 |
+
"epoch": 0.10608883223845475,
|
3809 |
+
"grad_norm": 5.231925010681152,
|
3810 |
+
"learning_rate": 0.00013532936938259656,
|
3811 |
+
"loss": 1.2463,
|
3812 |
+
"step": 541
|
3813 |
+
},
|
3814 |
+
{
|
3815 |
+
"epoch": 0.10628492989508775,
|
3816 |
+
"grad_norm": 6.779435157775879,
|
3817 |
+
"learning_rate": 0.0001351161525001795,
|
3818 |
+
"loss": 1.8802,
|
3819 |
+
"step": 542
|
3820 |
+
},
|
3821 |
+
{
|
3822 |
+
"epoch": 0.10648102755172076,
|
3823 |
+
"grad_norm": 7.8035197257995605,
|
3824 |
+
"learning_rate": 0.00013490275336339188,
|
3825 |
+
"loss": 1.1796,
|
3826 |
+
"step": 543
|
3827 |
+
},
|
3828 |
+
{
|
3829 |
+
"epoch": 0.10667712520835376,
|
3830 |
+
"grad_norm": 5.93556022644043,
|
3831 |
+
"learning_rate": 0.00013468917307978467,
|
3832 |
+
"loss": 1.1082,
|
3833 |
+
"step": 544
|
3834 |
+
},
|
3835 |
+
{
|
3836 |
+
"epoch": 0.10687322286498677,
|
3837 |
+
"grad_norm": 8.9798002243042,
|
3838 |
+
"learning_rate": 0.00013447541275784887,
|
3839 |
+
"loss": 1.6289,
|
3840 |
+
"step": 545
|
3841 |
+
},
|
3842 |
+
{
|
3843 |
+
"epoch": 0.10706932052161977,
|
3844 |
+
"grad_norm": 4.270269393920898,
|
3845 |
+
"learning_rate": 0.00013426147350700996,
|
3846 |
+
"loss": 3.2414,
|
3847 |
+
"step": 546
|
3848 |
+
},
|
3849 |
+
{
|
3850 |
+
"epoch": 0.10726541817825277,
|
3851 |
+
"grad_norm": 7.755331516265869,
|
3852 |
+
"learning_rate": 0.00013404735643762192,
|
3853 |
+
"loss": 3.5263,
|
3854 |
+
"step": 547
|
3855 |
+
},
|
3856 |
+
{
|
3857 |
+
"epoch": 0.10746151583488578,
|
3858 |
+
"grad_norm": 3.364109754562378,
|
3859 |
+
"learning_rate": 0.0001338330626609618,
|
3860 |
+
"loss": 1.1847,
|
3861 |
+
"step": 548
|
3862 |
+
},
|
3863 |
+
{
|
3864 |
+
"epoch": 0.10765761349151878,
|
3865 |
+
"grad_norm": 7.73459529876709,
|
3866 |
+
"learning_rate": 0.0001336185932892237,
|
3867 |
+
"loss": 2.0363,
|
3868 |
+
"step": 549
|
3869 |
+
},
|
3870 |
+
{
|
3871 |
+
"epoch": 0.10785371114815177,
|
3872 |
+
"grad_norm": 7.035179138183594,
|
3873 |
+
"learning_rate": 0.000133403949435513,
|
3874 |
+
"loss": 1.3317,
|
3875 |
+
"step": 550
|
3876 |
+
},
|
3877 |
+
{
|
3878 |
+
"epoch": 0.10804980880478478,
|
3879 |
+
"grad_norm": 5.30495023727417,
|
3880 |
+
"learning_rate": 0.00013318913221384076,
|
3881 |
+
"loss": 1.3776,
|
3882 |
+
"step": 551
|
3883 |
+
},
|
3884 |
+
{
|
3885 |
+
"epoch": 0.10824590646141778,
|
3886 |
+
"grad_norm": 6.3127593994140625,
|
3887 |
+
"learning_rate": 0.00013297414273911784,
|
3888 |
+
"loss": 0.9931,
|
3889 |
+
"step": 552
|
3890 |
+
},
|
3891 |
+
{
|
3892 |
+
"epoch": 0.10844200411805079,
|
3893 |
+
"grad_norm": 3.763789653778076,
|
3894 |
+
"learning_rate": 0.00013275898212714889,
|
3895 |
+
"loss": 0.9685,
|
3896 |
+
"step": 553
|
3897 |
+
},
|
3898 |
+
{
|
3899 |
+
"epoch": 0.10863810177468379,
|
3900 |
+
"grad_norm": 9.935264587402344,
|
3901 |
+
"learning_rate": 0.00013254365149462699,
|
3902 |
+
"loss": 1.1951,
|
3903 |
+
"step": 554
|
3904 |
+
},
|
3905 |
+
{
|
3906 |
+
"epoch": 0.1088341994313168,
|
3907 |
+
"grad_norm": 6.151608943939209,
|
3908 |
+
"learning_rate": 0.00013232815195912754,
|
3909 |
+
"loss": 1.9978,
|
3910 |
+
"step": 555
|
3911 |
+
},
|
3912 |
+
{
|
3913 |
+
"epoch": 0.1090302970879498,
|
3914 |
+
"grad_norm": 7.497893333435059,
|
3915 |
+
"learning_rate": 0.00013211248463910262,
|
3916 |
+
"loss": 2.1158,
|
3917 |
+
"step": 556
|
3918 |
+
},
|
3919 |
+
{
|
3920 |
+
"epoch": 0.1092263947445828,
|
3921 |
+
"grad_norm": 4.703649520874023,
|
3922 |
+
"learning_rate": 0.00013189665065387507,
|
3923 |
+
"loss": 2.9424,
|
3924 |
+
"step": 557
|
3925 |
+
},
|
3926 |
+
{
|
3927 |
+
"epoch": 0.1094224924012158,
|
3928 |
+
"grad_norm": 6.602818489074707,
|
3929 |
+
"learning_rate": 0.00013168065112363264,
|
3930 |
+
"loss": 1.9994,
|
3931 |
+
"step": 558
|
3932 |
+
},
|
3933 |
+
{
|
3934 |
+
"epoch": 0.10961859005784881,
|
3935 |
+
"grad_norm": 5.672735691070557,
|
3936 |
+
"learning_rate": 0.00013146448716942245,
|
3937 |
+
"loss": 1.6569,
|
3938 |
+
"step": 559
|
3939 |
+
},
|
3940 |
+
{
|
3941 |
+
"epoch": 0.10981468771448182,
|
3942 |
+
"grad_norm": 6.252768039703369,
|
3943 |
+
"learning_rate": 0.0001312481599131449,
|
3944 |
+
"loss": 1.0263,
|
3945 |
+
"step": 560
|
3946 |
+
},
|
3947 |
+
{
|
3948 |
+
"epoch": 0.11001078537111482,
|
3949 |
+
"grad_norm": 5.062610626220703,
|
3950 |
+
"learning_rate": 0.00013103167047754784,
|
3951 |
+
"loss": 1.8557,
|
3952 |
+
"step": 561
|
3953 |
+
},
|
3954 |
+
{
|
3955 |
+
"epoch": 0.11020688302774782,
|
3956 |
+
"grad_norm": 17.327529907226562,
|
3957 |
+
"learning_rate": 0.000130815019986221,
|
3958 |
+
"loss": 2.2288,
|
3959 |
+
"step": 562
|
3960 |
+
},
|
3961 |
+
{
|
3962 |
+
"epoch": 0.11040298068438083,
|
3963 |
+
"grad_norm": 9.787447929382324,
|
3964 |
+
"learning_rate": 0.00013059820956358998,
|
3965 |
+
"loss": 2.8692,
|
3966 |
+
"step": 563
|
3967 |
+
},
|
3968 |
+
{
|
3969 |
+
"epoch": 0.11059907834101383,
|
3970 |
+
"grad_norm": 17.163454055786133,
|
3971 |
+
"learning_rate": 0.00013038124033491025,
|
3972 |
+
"loss": 2.3984,
|
3973 |
+
"step": 564
|
3974 |
+
},
|
3975 |
+
{
|
3976 |
+
"epoch": 0.11079517599764682,
|
3977 |
+
"grad_norm": 7.947646617889404,
|
3978 |
+
"learning_rate": 0.00013016411342626168,
|
3979 |
+
"loss": 1.2144,
|
3980 |
+
"step": 565
|
3981 |
+
},
|
3982 |
+
{
|
3983 |
+
"epoch": 0.11099127365427983,
|
3984 |
+
"grad_norm": 8.151859283447266,
|
3985 |
+
"learning_rate": 0.00012994682996454247,
|
3986 |
+
"loss": 1.7198,
|
3987 |
+
"step": 566
|
3988 |
+
},
|
3989 |
+
{
|
3990 |
+
"epoch": 0.11118737131091283,
|
3991 |
+
"grad_norm": 8.358308792114258,
|
3992 |
+
"learning_rate": 0.00012972939107746325,
|
3993 |
+
"loss": 1.1448,
|
3994 |
+
"step": 567
|
3995 |
+
},
|
3996 |
+
{
|
3997 |
+
"epoch": 0.11138346896754583,
|
3998 |
+
"grad_norm": 5.084352970123291,
|
3999 |
+
"learning_rate": 0.0001295117978935414,
|
4000 |
+
"loss": 2.3229,
|
4001 |
+
"step": 568
|
4002 |
+
},
|
4003 |
+
{
|
4004 |
+
"epoch": 0.11157956662417884,
|
4005 |
+
"grad_norm": 4.846659183502197,
|
4006 |
+
"learning_rate": 0.0001292940515420951,
|
4007 |
+
"loss": 1.3941,
|
4008 |
+
"step": 569
|
4009 |
+
},
|
4010 |
+
{
|
4011 |
+
"epoch": 0.11177566428081184,
|
4012 |
+
"grad_norm": 5.432335376739502,
|
4013 |
+
"learning_rate": 0.0001290761531532374,
|
4014 |
+
"loss": 1.9803,
|
4015 |
+
"step": 570
|
4016 |
+
},
|
4017 |
+
{
|
4018 |
+
"epoch": 0.11197176193744485,
|
4019 |
+
"grad_norm": 5.953497886657715,
|
4020 |
+
"learning_rate": 0.00012885810385787055,
|
4021 |
+
"loss": 1.0619,
|
4022 |
+
"step": 571
|
4023 |
+
},
|
4024 |
+
{
|
4025 |
+
"epoch": 0.11216785959407785,
|
4026 |
+
"grad_norm": 4.770197868347168,
|
4027 |
+
"learning_rate": 0.00012863990478767994,
|
4028 |
+
"loss": 2.2999,
|
4029 |
+
"step": 572
|
4030 |
+
},
|
4031 |
+
{
|
4032 |
+
"epoch": 0.11236395725071086,
|
4033 |
+
"grad_norm": 5.719841957092285,
|
4034 |
+
"learning_rate": 0.00012842155707512825,
|
4035 |
+
"loss": 2.0115,
|
4036 |
+
"step": 573
|
4037 |
+
},
|
4038 |
+
{
|
4039 |
+
"epoch": 0.11256005490734386,
|
4040 |
+
"grad_norm": 6.13926362991333,
|
4041 |
+
"learning_rate": 0.00012820306185344976,
|
4042 |
+
"loss": 1.7,
|
4043 |
+
"step": 574
|
4044 |
+
},
|
4045 |
+
{
|
4046 |
+
"epoch": 0.11275615256397686,
|
4047 |
+
"grad_norm": 14.388799667358398,
|
4048 |
+
"learning_rate": 0.0001279844202566442,
|
4049 |
+
"loss": 1.4477,
|
4050 |
+
"step": 575
|
4051 |
+
},
|
4052 |
+
{
|
4053 |
+
"epoch": 0.11295225022060987,
|
4054 |
+
"grad_norm": 5.510779857635498,
|
4055 |
+
"learning_rate": 0.00012776563341947104,
|
4056 |
+
"loss": 1.145,
|
4057 |
+
"step": 576
|
4058 |
+
},
|
4059 |
+
{
|
4060 |
+
"epoch": 0.11314834787724287,
|
4061 |
+
"grad_norm": 7.694248199462891,
|
4062 |
+
"learning_rate": 0.00012754670247744354,
|
4063 |
+
"loss": 2.0622,
|
4064 |
+
"step": 577
|
4065 |
+
},
|
4066 |
+
{
|
4067 |
+
"epoch": 0.11334444553387588,
|
4068 |
+
"grad_norm": 10.04172420501709,
|
4069 |
+
"learning_rate": 0.0001273276285668229,
|
4070 |
+
"loss": 2.1074,
|
4071 |
+
"step": 578
|
4072 |
+
},
|
4073 |
+
{
|
4074 |
+
"epoch": 0.11354054319050888,
|
4075 |
+
"grad_norm": 4.613293170928955,
|
4076 |
+
"learning_rate": 0.00012710841282461238,
|
4077 |
+
"loss": 1.3278,
|
4078 |
+
"step": 579
|
4079 |
+
},
|
4080 |
+
{
|
4081 |
+
"epoch": 0.11373664084714187,
|
4082 |
+
"grad_norm": 7.878372669219971,
|
4083 |
+
"learning_rate": 0.0001268890563885512,
|
4084 |
+
"loss": 3.3353,
|
4085 |
+
"step": 580
|
4086 |
+
},
|
4087 |
+
{
|
4088 |
+
"epoch": 0.11393273850377487,
|
4089 |
+
"grad_norm": 4.252342224121094,
|
4090 |
+
"learning_rate": 0.0001266695603971089,
|
4091 |
+
"loss": 1.2478,
|
4092 |
+
"step": 581
|
4093 |
+
},
|
4094 |
+
{
|
4095 |
+
"epoch": 0.11412883616040788,
|
4096 |
+
"grad_norm": 13.550718307495117,
|
4097 |
+
"learning_rate": 0.0001264499259894793,
|
4098 |
+
"loss": 2.3346,
|
4099 |
+
"step": 582
|
4100 |
+
},
|
4101 |
+
{
|
4102 |
+
"epoch": 0.11432493381704088,
|
4103 |
+
"grad_norm": 5.674976348876953,
|
4104 |
+
"learning_rate": 0.0001262301543055746,
|
4105 |
+
"loss": 1.9865,
|
4106 |
+
"step": 583
|
4107 |
+
},
|
4108 |
+
{
|
4109 |
+
"epoch": 0.11452103147367389,
|
4110 |
+
"grad_norm": 14.854180335998535,
|
4111 |
+
"learning_rate": 0.0001260102464860195,
|
4112 |
+
"loss": 2.1993,
|
4113 |
+
"step": 584
|
4114 |
+
},
|
4115 |
+
{
|
4116 |
+
"epoch": 0.11471712913030689,
|
4117 |
+
"grad_norm": 5.227717399597168,
|
4118 |
+
"learning_rate": 0.0001257902036721452,
|
4119 |
+
"loss": 1.3025,
|
4120 |
+
"step": 585
|
4121 |
+
},
|
4122 |
+
{
|
4123 |
+
"epoch": 0.1149132267869399,
|
4124 |
+
"grad_norm": 9.518735885620117,
|
4125 |
+
"learning_rate": 0.00012557002700598353,
|
4126 |
+
"loss": 1.9782,
|
4127 |
+
"step": 586
|
4128 |
+
},
|
4129 |
+
{
|
4130 |
+
"epoch": 0.1151093244435729,
|
4131 |
+
"grad_norm": 15.842093467712402,
|
4132 |
+
"learning_rate": 0.00012534971763026104,
|
4133 |
+
"loss": 2.2285,
|
4134 |
+
"step": 587
|
4135 |
+
},
|
4136 |
+
{
|
4137 |
+
"epoch": 0.1153054221002059,
|
4138 |
+
"grad_norm": 7.949627876281738,
|
4139 |
+
"learning_rate": 0.00012512927668839304,
|
4140 |
+
"loss": 1.0101,
|
4141 |
+
"step": 588
|
4142 |
+
},
|
4143 |
+
{
|
4144 |
+
"epoch": 0.11550151975683891,
|
4145 |
+
"grad_norm": 10.285954475402832,
|
4146 |
+
"learning_rate": 0.00012490870532447774,
|
4147 |
+
"loss": 2.2186,
|
4148 |
+
"step": 589
|
4149 |
+
},
|
4150 |
+
{
|
4151 |
+
"epoch": 0.11569761741347191,
|
4152 |
+
"grad_norm": 13.825008392333984,
|
4153 |
+
"learning_rate": 0.00012468800468329013,
|
4154 |
+
"loss": 1.9117,
|
4155 |
+
"step": 590
|
4156 |
+
},
|
4157 |
+
{
|
4158 |
+
"epoch": 0.11589371507010492,
|
4159 |
+
"grad_norm": 4.062519073486328,
|
4160 |
+
"learning_rate": 0.00012446717591027624,
|
4161 |
+
"loss": 1.3969,
|
4162 |
+
"step": 591
|
4163 |
+
},
|
4164 |
+
{
|
4165 |
+
"epoch": 0.11608981272673792,
|
4166 |
+
"grad_norm": 13.4169282913208,
|
4167 |
+
"learning_rate": 0.00012424622015154703,
|
4168 |
+
"loss": 1.6026,
|
4169 |
+
"step": 592
|
4170 |
+
},
|
4171 |
+
{
|
4172 |
+
"epoch": 0.11628591038337092,
|
4173 |
+
"grad_norm": 5.054627895355225,
|
4174 |
+
"learning_rate": 0.0001240251385538726,
|
4175 |
+
"loss": 3.3848,
|
4176 |
+
"step": 593
|
4177 |
+
},
|
4178 |
+
{
|
4179 |
+
"epoch": 0.11648200804000393,
|
4180 |
+
"grad_norm": 5.519535541534424,
|
4181 |
+
"learning_rate": 0.00012380393226467615,
|
4182 |
+
"loss": 2.0605,
|
4183 |
+
"step": 594
|
4184 |
+
},
|
4185 |
+
{
|
4186 |
+
"epoch": 0.11667810569663692,
|
4187 |
+
"grad_norm": 8.747673034667969,
|
4188 |
+
"learning_rate": 0.000123582602432028,
|
4189 |
+
"loss": 1.945,
|
4190 |
+
"step": 595
|
4191 |
+
},
|
4192 |
+
{
|
4193 |
+
"epoch": 0.11687420335326992,
|
4194 |
+
"grad_norm": 7.695986747741699,
|
4195 |
+
"learning_rate": 0.0001233611502046397,
|
4196 |
+
"loss": 1.2818,
|
4197 |
+
"step": 596
|
4198 |
+
},
|
4199 |
+
{
|
4200 |
+
"epoch": 0.11707030100990293,
|
4201 |
+
"grad_norm": 6.775882720947266,
|
4202 |
+
"learning_rate": 0.000123139576731858,
|
4203 |
+
"loss": 1.5774,
|
4204 |
+
"step": 597
|
4205 |
+
},
|
4206 |
+
{
|
4207 |
+
"epoch": 0.11726639866653593,
|
4208 |
+
"grad_norm": 6.141412258148193,
|
4209 |
+
"learning_rate": 0.00012291788316365888,
|
4210 |
+
"loss": 2.0431,
|
4211 |
+
"step": 598
|
4212 |
+
},
|
4213 |
+
{
|
4214 |
+
"epoch": 0.11746249632316894,
|
4215 |
+
"grad_norm": 6.908384323120117,
|
4216 |
+
"learning_rate": 0.00012269607065064177,
|
4217 |
+
"loss": 2.298,
|
4218 |
+
"step": 599
|
4219 |
+
},
|
4220 |
+
{
|
4221 |
+
"epoch": 0.11765859397980194,
|
4222 |
+
"grad_norm": 6.526587963104248,
|
4223 |
+
"learning_rate": 0.0001224741403440233,
|
4224 |
+
"loss": 0.7667,
|
4225 |
+
"step": 600
|
4226 |
+
},
|
4227 |
+
{
|
4228 |
+
"epoch": 0.11785469163643494,
|
4229 |
+
"grad_norm": 6.505580902099609,
|
4230 |
+
"learning_rate": 0.00012225209339563145,
|
4231 |
+
"loss": 2.2598,
|
4232 |
+
"step": 601
|
4233 |
+
},
|
4234 |
+
{
|
4235 |
+
"epoch": 0.11805078929306795,
|
4236 |
+
"grad_norm": 5.94541072845459,
|
4237 |
+
"learning_rate": 0.00012202993095789966,
|
4238 |
+
"loss": 2.5347,
|
4239 |
+
"step": 602
|
4240 |
+
},
|
4241 |
+
{
|
4242 |
+
"epoch": 0.11824688694970095,
|
4243 |
+
"grad_norm": 4.825118064880371,
|
4244 |
+
"learning_rate": 0.00012180765418386068,
|
4245 |
+
"loss": 1.3592,
|
4246 |
+
"step": 603
|
4247 |
+
},
|
4248 |
+
{
|
4249 |
+
"epoch": 0.11844298460633396,
|
4250 |
+
"grad_norm": 10.260912895202637,
|
4251 |
+
"learning_rate": 0.00012158526422714076,
|
4252 |
+
"loss": 1.9816,
|
4253 |
+
"step": 604
|
4254 |
+
},
|
4255 |
+
{
|
4256 |
+
"epoch": 0.11863908226296696,
|
4257 |
+
"grad_norm": 4.47437047958374,
|
4258 |
+
"learning_rate": 0.00012136276224195348,
|
4259 |
+
"loss": 1.3715,
|
4260 |
+
"step": 605
|
4261 |
+
},
|
4262 |
+
{
|
4263 |
+
"epoch": 0.11883517991959996,
|
4264 |
+
"grad_norm": 5.319314002990723,
|
4265 |
+
"learning_rate": 0.00012114014938309393,
|
4266 |
+
"loss": 1.7164,
|
4267 |
+
"step": 606
|
4268 |
+
},
|
4269 |
+
{
|
4270 |
+
"epoch": 0.11903127757623297,
|
4271 |
+
"grad_norm": 9.467094421386719,
|
4272 |
+
"learning_rate": 0.00012091742680593254,
|
4273 |
+
"loss": 1.711,
|
4274 |
+
"step": 607
|
4275 |
+
},
|
4276 |
+
{
|
4277 |
+
"epoch": 0.11922737523286597,
|
4278 |
+
"grad_norm": 15.36571979522705,
|
4279 |
+
"learning_rate": 0.0001206945956664093,
|
4280 |
+
"loss": 2.1494,
|
4281 |
+
"step": 608
|
4282 |
+
},
|
4283 |
+
{
|
4284 |
+
"epoch": 0.11942347288949898,
|
4285 |
+
"grad_norm": 5.132857322692871,
|
4286 |
+
"learning_rate": 0.00012047165712102759,
|
4287 |
+
"loss": 1.0723,
|
4288 |
+
"step": 609
|
4289 |
+
},
|
4290 |
+
{
|
4291 |
+
"epoch": 0.11961957054613197,
|
4292 |
+
"grad_norm": 7.701923847198486,
|
4293 |
+
"learning_rate": 0.00012024861232684823,
|
4294 |
+
"loss": 0.6081,
|
4295 |
+
"step": 610
|
4296 |
+
},
|
4297 |
+
{
|
4298 |
+
"epoch": 0.11981566820276497,
|
4299 |
+
"grad_norm": 8.608162879943848,
|
4300 |
+
"learning_rate": 0.00012002546244148345,
|
4301 |
+
"loss": 2.7694,
|
4302 |
+
"step": 611
|
4303 |
+
},
|
4304 |
+
{
|
4305 |
+
"epoch": 0.12001176585939798,
|
4306 |
+
"grad_norm": 6.971923351287842,
|
4307 |
+
"learning_rate": 0.00011980220862309097,
|
4308 |
+
"loss": 2.4804,
|
4309 |
+
"step": 612
|
4310 |
+
},
|
4311 |
+
{
|
4312 |
+
"epoch": 0.12020786351603098,
|
4313 |
+
"grad_norm": 6.421287536621094,
|
4314 |
+
"learning_rate": 0.00011957885203036785,
|
4315 |
+
"loss": 1.1601,
|
4316 |
+
"step": 613
|
4317 |
+
},
|
4318 |
+
{
|
4319 |
+
"epoch": 0.12040396117266398,
|
4320 |
+
"grad_norm": 5.901442527770996,
|
4321 |
+
"learning_rate": 0.00011935539382254459,
|
4322 |
+
"loss": 1.1592,
|
4323 |
+
"step": 614
|
4324 |
+
},
|
4325 |
+
{
|
4326 |
+
"epoch": 0.12060005882929699,
|
4327 |
+
"grad_norm": 8.312341690063477,
|
4328 |
+
"learning_rate": 0.00011913183515937916,
|
4329 |
+
"loss": 1.4096,
|
4330 |
+
"step": 615
|
4331 |
+
},
|
4332 |
+
{
|
4333 |
+
"epoch": 0.12079615648592999,
|
4334 |
+
"grad_norm": 9.224618911743164,
|
4335 |
+
"learning_rate": 0.00011890817720115075,
|
4336 |
+
"loss": 1.8728,
|
4337 |
+
"step": 616
|
4338 |
+
},
|
4339 |
+
{
|
4340 |
+
"epoch": 0.120992254142563,
|
4341 |
+
"grad_norm": 6.559041500091553,
|
4342 |
+
"learning_rate": 0.00011868442110865399,
|
4343 |
+
"loss": 1.4327,
|
4344 |
+
"step": 617
|
4345 |
+
},
|
4346 |
+
{
|
4347 |
+
"epoch": 0.121188351799196,
|
4348 |
+
"grad_norm": 6.139137268066406,
|
4349 |
+
"learning_rate": 0.0001184605680431928,
|
4350 |
+
"loss": 2.3143,
|
4351 |
+
"step": 618
|
4352 |
+
},
|
4353 |
+
{
|
4354 |
+
"epoch": 0.121384449455829,
|
4355 |
+
"grad_norm": 7.806881427764893,
|
4356 |
+
"learning_rate": 0.0001182366191665744,
|
4357 |
+
"loss": 2.0893,
|
4358 |
+
"step": 619
|
4359 |
+
},
|
4360 |
+
{
|
4361 |
+
"epoch": 0.12158054711246201,
|
4362 |
+
"grad_norm": 5.146418571472168,
|
4363 |
+
"learning_rate": 0.00011801257564110329,
|
4364 |
+
"loss": 0.9106,
|
4365 |
+
"step": 620
|
4366 |
+
},
|
4367 |
+
{
|
4368 |
+
"epoch": 0.12177664476909501,
|
4369 |
+
"grad_norm": 10.43996810913086,
|
4370 |
+
"learning_rate": 0.00011778843862957514,
|
4371 |
+
"loss": 1.5653,
|
4372 |
+
"step": 621
|
4373 |
+
},
|
4374 |
+
{
|
4375 |
+
"epoch": 0.12197274242572802,
|
4376 |
+
"grad_norm": 9.192870140075684,
|
4377 |
+
"learning_rate": 0.0001175642092952709,
|
4378 |
+
"loss": 1.097,
|
4379 |
+
"step": 622
|
4380 |
+
},
|
4381 |
+
{
|
4382 |
+
"epoch": 0.12216884008236102,
|
4383 |
+
"grad_norm": 12.34306526184082,
|
4384 |
+
"learning_rate": 0.00011733988880195068,
|
4385 |
+
"loss": 0.8703,
|
4386 |
+
"step": 623
|
4387 |
+
},
|
4388 |
+
{
|
4389 |
+
"epoch": 0.12236493773899403,
|
4390 |
+
"grad_norm": 6.297984600067139,
|
4391 |
+
"learning_rate": 0.00011711547831384761,
|
4392 |
+
"loss": 0.8209,
|
4393 |
+
"step": 624
|
4394 |
+
},
|
4395 |
+
{
|
4396 |
+
"epoch": 0.12256103539562702,
|
4397 |
+
"grad_norm": 6.844716548919678,
|
4398 |
+
"learning_rate": 0.00011689097899566198,
|
4399 |
+
"loss": 1.2469,
|
4400 |
+
"step": 625
|
4401 |
+
},
|
4402 |
+
{
|
4403 |
+
"epoch": 0.12275713305226002,
|
4404 |
+
"grad_norm": 4.610165119171143,
|
4405 |
+
"learning_rate": 0.00011666639201255506,
|
4406 |
+
"loss": 1.6045,
|
4407 |
+
"step": 626
|
4408 |
+
},
|
4409 |
+
{
|
4410 |
+
"epoch": 0.12295323070889302,
|
4411 |
+
"grad_norm": 8.704841613769531,
|
4412 |
+
"learning_rate": 0.00011644171853014319,
|
4413 |
+
"loss": 1.5989,
|
4414 |
+
"step": 627
|
4415 |
+
},
|
4416 |
+
{
|
4417 |
+
"epoch": 0.12314932836552603,
|
4418 |
+
"grad_norm": 14.035907745361328,
|
4419 |
+
"learning_rate": 0.00011621695971449154,
|
4420 |
+
"loss": 2.3428,
|
4421 |
+
"step": 628
|
4422 |
+
},
|
4423 |
+
{
|
4424 |
+
"epoch": 0.12334542602215903,
|
4425 |
+
"grad_norm": 8.880375862121582,
|
4426 |
+
"learning_rate": 0.00011599211673210826,
|
4427 |
+
"loss": 1.6628,
|
4428 |
+
"step": 629
|
4429 |
+
},
|
4430 |
+
{
|
4431 |
+
"epoch": 0.12354152367879204,
|
4432 |
+
"grad_norm": 5.904176235198975,
|
4433 |
+
"learning_rate": 0.00011576719074993827,
|
4434 |
+
"loss": 1.8985,
|
4435 |
+
"step": 630
|
4436 |
+
},
|
4437 |
+
{
|
4438 |
+
"epoch": 0.12373762133542504,
|
4439 |
+
"grad_norm": 5.038250923156738,
|
4440 |
+
"learning_rate": 0.00011554218293535725,
|
4441 |
+
"loss": 1.8648,
|
4442 |
+
"step": 631
|
4443 |
+
},
|
4444 |
+
{
|
4445 |
+
"epoch": 0.12393371899205805,
|
4446 |
+
"grad_norm": 4.206717014312744,
|
4447 |
+
"learning_rate": 0.00011531709445616563,
|
4448 |
+
"loss": 1.7799,
|
4449 |
+
"step": 632
|
4450 |
+
},
|
4451 |
+
{
|
4452 |
+
"epoch": 0.12412981664869105,
|
4453 |
+
"grad_norm": 9.580304145812988,
|
4454 |
+
"learning_rate": 0.00011509192648058249,
|
4455 |
+
"loss": 1.484,
|
4456 |
+
"step": 633
|
4457 |
+
},
|
4458 |
+
{
|
4459 |
+
"epoch": 0.12432591430532405,
|
4460 |
+
"grad_norm": 5.086032867431641,
|
4461 |
+
"learning_rate": 0.00011486668017723949,
|
4462 |
+
"loss": 1.2848,
|
4463 |
+
"step": 634
|
4464 |
+
},
|
4465 |
+
{
|
4466 |
+
"epoch": 0.12452201196195706,
|
4467 |
+
"grad_norm": 3.946556568145752,
|
4468 |
+
"learning_rate": 0.00011464135671517482,
|
4469 |
+
"loss": 0.6885,
|
4470 |
+
"step": 635
|
4471 |
+
},
|
4472 |
+
{
|
4473 |
+
"epoch": 0.12471810961859006,
|
4474 |
+
"grad_norm": 5.43235969543457,
|
4475 |
+
"learning_rate": 0.0001144159572638271,
|
4476 |
+
"loss": 1.1324,
|
4477 |
+
"step": 636
|
4478 |
+
},
|
4479 |
+
{
|
4480 |
+
"epoch": 0.12491420727522307,
|
4481 |
+
"grad_norm": 14.442760467529297,
|
4482 |
+
"learning_rate": 0.00011419048299302939,
|
4483 |
+
"loss": 2.5022,
|
4484 |
+
"step": 637
|
4485 |
+
},
|
4486 |
+
{
|
4487 |
+
"epoch": 0.12511030493185607,
|
4488 |
+
"grad_norm": 5.514584541320801,
|
4489 |
+
"learning_rate": 0.00011396493507300303,
|
4490 |
+
"loss": 0.714,
|
4491 |
+
"step": 638
|
4492 |
+
},
|
4493 |
+
{
|
4494 |
+
"epoch": 0.12530640258848907,
|
4495 |
+
"grad_norm": 9.007026672363281,
|
4496 |
+
"learning_rate": 0.00011373931467435159,
|
4497 |
+
"loss": 1.894,
|
4498 |
+
"step": 639
|
4499 |
+
},
|
4500 |
+
{
|
4501 |
+
"epoch": 0.12550250024512208,
|
4502 |
+
"grad_norm": 4.5296196937561035,
|
4503 |
+
"learning_rate": 0.00011351362296805485,
|
4504 |
+
"loss": 1.8764,
|
4505 |
+
"step": 640
|
4506 |
+
},
|
4507 |
+
{
|
4508 |
+
"epoch": 0.12569859790175508,
|
4509 |
+
"grad_norm": 6.608786106109619,
|
4510 |
+
"learning_rate": 0.00011328786112546269,
|
4511 |
+
"loss": 1.8515,
|
4512 |
+
"step": 641
|
4513 |
+
},
|
4514 |
+
{
|
4515 |
+
"epoch": 0.1258946955583881,
|
4516 |
+
"grad_norm": 5.0649261474609375,
|
4517 |
+
"learning_rate": 0.00011306203031828886,
|
4518 |
+
"loss": 1.1955,
|
4519 |
+
"step": 642
|
4520 |
+
},
|
4521 |
+
{
|
4522 |
+
"epoch": 0.1260907932150211,
|
4523 |
+
"grad_norm": 7.887217998504639,
|
4524 |
+
"learning_rate": 0.00011283613171860525,
|
4525 |
+
"loss": 1.9566,
|
4526 |
+
"step": 643
|
4527 |
+
},
|
4528 |
+
{
|
4529 |
+
"epoch": 0.1262868908716541,
|
4530 |
+
"grad_norm": 3.997429609298706,
|
4531 |
+
"learning_rate": 0.00011261016649883545,
|
4532 |
+
"loss": 0.4142,
|
4533 |
+
"step": 644
|
4534 |
+
},
|
4535 |
+
{
|
4536 |
+
"epoch": 0.1264829885282871,
|
4537 |
+
"grad_norm": 6.192634582519531,
|
4538 |
+
"learning_rate": 0.0001123841358317489,
|
4539 |
+
"loss": 0.7429,
|
4540 |
+
"step": 645
|
4541 |
+
},
|
4542 |
+
{
|
4543 |
+
"epoch": 0.1266790861849201,
|
4544 |
+
"grad_norm": 10.252041816711426,
|
4545 |
+
"learning_rate": 0.00011215804089045459,
|
4546 |
+
"loss": 1.5821,
|
4547 |
+
"step": 646
|
4548 |
+
},
|
4549 |
+
{
|
4550 |
+
"epoch": 0.12687518384155308,
|
4551 |
+
"grad_norm": 15.608290672302246,
|
4552 |
+
"learning_rate": 0.00011193188284839517,
|
4553 |
+
"loss": 2.9049,
|
4554 |
+
"step": 647
|
4555 |
+
},
|
4556 |
+
{
|
4557 |
+
"epoch": 0.12707128149818608,
|
4558 |
+
"grad_norm": 7.4965314865112305,
|
4559 |
+
"learning_rate": 0.00011170566287934088,
|
4560 |
+
"loss": 1.7288,
|
4561 |
+
"step": 648
|
4562 |
+
},
|
4563 |
+
{
|
4564 |
+
"epoch": 0.1272673791548191,
|
4565 |
+
"grad_norm": 13.268543243408203,
|
4566 |
+
"learning_rate": 0.00011147938215738323,
|
4567 |
+
"loss": 2.6662,
|
4568 |
+
"step": 649
|
4569 |
+
},
|
4570 |
+
{
|
4571 |
+
"epoch": 0.1274634768114521,
|
4572 |
+
"grad_norm": 6.431972503662109,
|
4573 |
+
"learning_rate": 0.00011125304185692907,
|
4574 |
+
"loss": 2.416,
|
4575 |
+
"step": 650
|
4576 |
+
},
|
4577 |
+
{
|
4578 |
+
"epoch": 0.1276595744680851,
|
4579 |
+
"grad_norm": 5.82267951965332,
|
4580 |
+
"learning_rate": 0.0001110266431526945,
|
4581 |
+
"loss": 1.0295,
|
4582 |
+
"step": 651
|
4583 |
+
},
|
4584 |
+
{
|
4585 |
+
"epoch": 0.1278556721247181,
|
4586 |
+
"grad_norm": 8.612664222717285,
|
4587 |
+
"learning_rate": 0.00011080018721969871,
|
4588 |
+
"loss": 2.5151,
|
4589 |
+
"step": 652
|
4590 |
+
},
|
4591 |
+
{
|
4592 |
+
"epoch": 0.1280517697813511,
|
4593 |
+
"grad_norm": 10.103599548339844,
|
4594 |
+
"learning_rate": 0.00011057367523325792,
|
4595 |
+
"loss": 2.4646,
|
4596 |
+
"step": 653
|
4597 |
+
},
|
4598 |
+
{
|
4599 |
+
"epoch": 0.1282478674379841,
|
4600 |
+
"grad_norm": 6.955473899841309,
|
4601 |
+
"learning_rate": 0.00011034710836897921,
|
4602 |
+
"loss": 1.3256,
|
4603 |
+
"step": 654
|
4604 |
+
},
|
4605 |
+
{
|
4606 |
+
"epoch": 0.1284439650946171,
|
4607 |
+
"grad_norm": 6.604331016540527,
|
4608 |
+
"learning_rate": 0.00011012048780275463,
|
4609 |
+
"loss": 1.1124,
|
4610 |
+
"step": 655
|
4611 |
+
},
|
4612 |
+
{
|
4613 |
+
"epoch": 0.12864006275125012,
|
4614 |
+
"grad_norm": 7.2269511222839355,
|
4615 |
+
"learning_rate": 0.00010989381471075482,
|
4616 |
+
"loss": 2.1239,
|
4617 |
+
"step": 656
|
4618 |
+
},
|
4619 |
+
{
|
4620 |
+
"epoch": 0.12883616040788312,
|
4621 |
+
"grad_norm": 7.591269016265869,
|
4622 |
+
"learning_rate": 0.00010966709026942303,
|
4623 |
+
"loss": 1.9995,
|
4624 |
+
"step": 657
|
4625 |
+
},
|
4626 |
+
{
|
4627 |
+
"epoch": 0.12903225806451613,
|
4628 |
+
"grad_norm": 5.342310428619385,
|
4629 |
+
"learning_rate": 0.00010944031565546906,
|
4630 |
+
"loss": 0.8718,
|
4631 |
+
"step": 658
|
4632 |
+
},
|
4633 |
+
{
|
4634 |
+
"epoch": 0.12922835572114913,
|
4635 |
+
"grad_norm": 9.045146942138672,
|
4636 |
+
"learning_rate": 0.0001092134920458631,
|
4637 |
+
"loss": 0.9157,
|
4638 |
+
"step": 659
|
4639 |
+
},
|
4640 |
+
{
|
4641 |
+
"epoch": 0.12942445337778213,
|
4642 |
+
"grad_norm": 4.300910472869873,
|
4643 |
+
"learning_rate": 0.00010898662061782965,
|
4644 |
+
"loss": 1.0317,
|
4645 |
+
"step": 660
|
4646 |
+
},
|
4647 |
+
{
|
4648 |
+
"epoch": 0.12962055103441514,
|
4649 |
+
"grad_norm": 11.28966999053955,
|
4650 |
+
"learning_rate": 0.0001087597025488413,
|
4651 |
+
"loss": 0.7102,
|
4652 |
+
"step": 661
|
4653 |
+
},
|
4654 |
+
{
|
4655 |
+
"epoch": 0.12981664869104814,
|
4656 |
+
"grad_norm": 7.143383502960205,
|
4657 |
+
"learning_rate": 0.00010853273901661285,
|
4658 |
+
"loss": 1.4524,
|
4659 |
+
"step": 662
|
4660 |
+
},
|
4661 |
+
{
|
4662 |
+
"epoch": 0.13001274634768115,
|
4663 |
+
"grad_norm": 8.20251750946045,
|
4664 |
+
"learning_rate": 0.00010830573119909493,
|
4665 |
+
"loss": 1.825,
|
4666 |
+
"step": 663
|
4667 |
+
},
|
4668 |
+
{
|
4669 |
+
"epoch": 0.13020884400431415,
|
4670 |
+
"grad_norm": 8.748811721801758,
|
4671 |
+
"learning_rate": 0.00010807868027446808,
|
4672 |
+
"loss": 0.6214,
|
4673 |
+
"step": 664
|
4674 |
+
},
|
4675 |
+
{
|
4676 |
+
"epoch": 0.13040494166094715,
|
4677 |
+
"grad_norm": 6.7635040283203125,
|
4678 |
+
"learning_rate": 0.00010785158742113655,
|
4679 |
+
"loss": 1.158,
|
4680 |
+
"step": 665
|
4681 |
+
},
|
4682 |
+
{
|
4683 |
+
"epoch": 0.13060103931758016,
|
4684 |
+
"grad_norm": 5.804599761962891,
|
4685 |
+
"learning_rate": 0.00010762445381772217,
|
4686 |
+
"loss": 1.2651,
|
4687 |
+
"step": 666
|
4688 |
+
},
|
4689 |
+
{
|
4690 |
+
"epoch": 0.13079713697421316,
|
4691 |
+
"grad_norm": 6.046361923217773,
|
4692 |
+
"learning_rate": 0.00010739728064305834,
|
4693 |
+
"loss": 1.5022,
|
4694 |
+
"step": 667
|
4695 |
+
},
|
4696 |
+
{
|
4697 |
+
"epoch": 0.13099323463084617,
|
4698 |
+
"grad_norm": 10.987192153930664,
|
4699 |
+
"learning_rate": 0.00010717006907618377,
|
4700 |
+
"loss": 1.5359,
|
4701 |
+
"step": 668
|
4702 |
+
},
|
4703 |
+
{
|
4704 |
+
"epoch": 0.13118933228747917,
|
4705 |
+
"grad_norm": 8.694954872131348,
|
4706 |
+
"learning_rate": 0.00010694282029633647,
|
4707 |
+
"loss": 1.682,
|
4708 |
+
"step": 669
|
4709 |
+
},
|
4710 |
+
{
|
4711 |
+
"epoch": 0.13138542994411218,
|
4712 |
+
"grad_norm": 15.458917617797852,
|
4713 |
+
"learning_rate": 0.00010671553548294753,
|
4714 |
+
"loss": 1.3832,
|
4715 |
+
"step": 670
|
4716 |
+
},
|
4717 |
+
{
|
4718 |
+
"epoch": 0.13158152760074518,
|
4719 |
+
"grad_norm": 3.7156457901000977,
|
4720 |
+
"learning_rate": 0.00010648821581563513,
|
4721 |
+
"loss": 2.4507,
|
4722 |
+
"step": 671
|
4723 |
+
},
|
4724 |
+
{
|
4725 |
+
"epoch": 0.13177762525737818,
|
4726 |
+
"grad_norm": 9.7117338180542,
|
4727 |
+
"learning_rate": 0.00010626086247419826,
|
4728 |
+
"loss": 2.3141,
|
4729 |
+
"step": 672
|
4730 |
+
},
|
4731 |
+
{
|
4732 |
+
"epoch": 0.1319737229140112,
|
4733 |
+
"grad_norm": 7.740649700164795,
|
4734 |
+
"learning_rate": 0.00010603347663861079,
|
4735 |
+
"loss": 2.1845,
|
4736 |
+
"step": 673
|
4737 |
+
},
|
4738 |
+
{
|
4739 |
+
"epoch": 0.1321698205706442,
|
4740 |
+
"grad_norm": 7.079351425170898,
|
4741 |
+
"learning_rate": 0.00010580605948901514,
|
4742 |
+
"loss": 1.5912,
|
4743 |
+
"step": 674
|
4744 |
+
},
|
4745 |
+
{
|
4746 |
+
"epoch": 0.1323659182272772,
|
4747 |
+
"grad_norm": 3.896484613418579,
|
4748 |
+
"learning_rate": 0.00010557861220571625,
|
4749 |
+
"loss": 0.8481,
|
4750 |
+
"step": 675
|
4751 |
+
},
|
4752 |
+
{
|
4753 |
+
"epoch": 0.1325620158839102,
|
4754 |
+
"grad_norm": 12.182394981384277,
|
4755 |
+
"learning_rate": 0.00010535113596917556,
|
4756 |
+
"loss": 2.1278,
|
4757 |
+
"step": 676
|
4758 |
+
},
|
4759 |
+
{
|
4760 |
+
"epoch": 0.13275811354054318,
|
4761 |
+
"grad_norm": 7.2080230712890625,
|
4762 |
+
"learning_rate": 0.00010512363196000465,
|
4763 |
+
"loss": 1.8256,
|
4764 |
+
"step": 677
|
4765 |
+
},
|
4766 |
+
{
|
4767 |
+
"epoch": 0.13295421119717618,
|
4768 |
+
"grad_norm": 7.081282615661621,
|
4769 |
+
"learning_rate": 0.00010489610135895933,
|
4770 |
+
"loss": 1.5351,
|
4771 |
+
"step": 678
|
4772 |
+
},
|
4773 |
+
{
|
4774 |
+
"epoch": 0.13315030885380919,
|
4775 |
+
"grad_norm": 4.274497985839844,
|
4776 |
+
"learning_rate": 0.00010466854534693335,
|
4777 |
+
"loss": 1.3264,
|
4778 |
+
"step": 679
|
4779 |
+
},
|
4780 |
+
{
|
4781 |
+
"epoch": 0.1333464065104422,
|
4782 |
+
"grad_norm": 6.364016532897949,
|
4783 |
+
"learning_rate": 0.00010444096510495243,
|
4784 |
+
"loss": 1.751,
|
4785 |
+
"step": 680
|
4786 |
+
},
|
4787 |
+
{
|
4788 |
+
"epoch": 0.1335425041670752,
|
4789 |
+
"grad_norm": 6.791689872741699,
|
4790 |
+
"learning_rate": 0.00010421336181416797,
|
4791 |
+
"loss": 0.7944,
|
4792 |
+
"step": 681
|
4793 |
+
},
|
4794 |
+
{
|
4795 |
+
"epoch": 0.1337386018237082,
|
4796 |
+
"grad_norm": 9.867424964904785,
|
4797 |
+
"learning_rate": 0.00010398573665585105,
|
4798 |
+
"loss": 1.6282,
|
4799 |
+
"step": 682
|
4800 |
+
},
|
4801 |
+
{
|
4802 |
+
"epoch": 0.1339346994803412,
|
4803 |
+
"grad_norm": 12.289506912231445,
|
4804 |
+
"learning_rate": 0.0001037580908113862,
|
4805 |
+
"loss": 2.1015,
|
4806 |
+
"step": 683
|
4807 |
+
},
|
4808 |
+
{
|
4809 |
+
"epoch": 0.1341307971369742,
|
4810 |
+
"grad_norm": 5.21271276473999,
|
4811 |
+
"learning_rate": 0.00010353042546226537,
|
4812 |
+
"loss": 1.0861,
|
4813 |
+
"step": 684
|
4814 |
+
},
|
4815 |
+
{
|
4816 |
+
"epoch": 0.1343268947936072,
|
4817 |
+
"grad_norm": 6.344359874725342,
|
4818 |
+
"learning_rate": 0.00010330274179008161,
|
4819 |
+
"loss": 1.295,
|
4820 |
+
"step": 685
|
4821 |
+
},
|
4822 |
+
{
|
4823 |
+
"epoch": 0.13452299245024021,
|
4824 |
+
"grad_norm": 4.4853196144104,
|
4825 |
+
"learning_rate": 0.00010307504097652323,
|
4826 |
+
"loss": 1.5746,
|
4827 |
+
"step": 686
|
4828 |
+
},
|
4829 |
+
{
|
4830 |
+
"epoch": 0.13471909010687322,
|
4831 |
+
"grad_norm": 4.868480205535889,
|
4832 |
+
"learning_rate": 0.0001028473242033674,
|
4833 |
+
"loss": 0.7534,
|
4834 |
+
"step": 687
|
4835 |
+
},
|
4836 |
+
{
|
4837 |
+
"epoch": 0.13491518776350622,
|
4838 |
+
"grad_norm": 33.27044677734375,
|
4839 |
+
"learning_rate": 0.00010261959265247419,
|
4840 |
+
"loss": 3.1026,
|
4841 |
+
"step": 688
|
4842 |
+
},
|
4843 |
+
{
|
4844 |
+
"epoch": 0.13511128542013923,
|
4845 |
+
"grad_norm": 5.2553253173828125,
|
4846 |
+
"learning_rate": 0.0001023918475057803,
|
4847 |
+
"loss": 1.7087,
|
4848 |
+
"step": 689
|
4849 |
+
},
|
4850 |
+
{
|
4851 |
+
"epoch": 0.13530738307677223,
|
4852 |
+
"grad_norm": 9.47342586517334,
|
4853 |
+
"learning_rate": 0.00010216408994529303,
|
4854 |
+
"loss": 1.2846,
|
4855 |
+
"step": 690
|
4856 |
+
},
|
4857 |
+
{
|
4858 |
+
"epoch": 0.13550348073340523,
|
4859 |
+
"grad_norm": 5.372720241546631,
|
4860 |
+
"learning_rate": 0.00010193632115308411,
|
4861 |
+
"loss": 2.6163,
|
4862 |
+
"step": 691
|
4863 |
+
},
|
4864 |
+
{
|
4865 |
+
"epoch": 0.13569957839003824,
|
4866 |
+
"grad_norm": 9.89535903930664,
|
4867 |
+
"learning_rate": 0.00010170854231128352,
|
4868 |
+
"loss": 2.3616,
|
4869 |
+
"step": 692
|
4870 |
+
},
|
4871 |
+
{
|
4872 |
+
"epoch": 0.13589567604667124,
|
4873 |
+
"grad_norm": 8.19672966003418,
|
4874 |
+
"learning_rate": 0.00010148075460207347,
|
4875 |
+
"loss": 1.8404,
|
4876 |
+
"step": 693
|
4877 |
+
},
|
4878 |
+
{
|
4879 |
+
"epoch": 0.13609177370330425,
|
4880 |
+
"grad_norm": 5.280313491821289,
|
4881 |
+
"learning_rate": 0.0001012529592076821,
|
4882 |
+
"loss": 2.2078,
|
4883 |
+
"step": 694
|
4884 |
+
},
|
4885 |
+
{
|
4886 |
+
"epoch": 0.13628787135993725,
|
4887 |
+
"grad_norm": 7.142581462860107,
|
4888 |
+
"learning_rate": 0.00010102515731037758,
|
4889 |
+
"loss": 0.9421,
|
4890 |
+
"step": 695
|
4891 |
+
},
|
4892 |
+
{
|
4893 |
+
"epoch": 0.13648396901657026,
|
4894 |
+
"grad_norm": 8.142248153686523,
|
4895 |
+
"learning_rate": 0.00010079735009246167,
|
4896 |
+
"loss": 0.83,
|
4897 |
+
"step": 696
|
4898 |
+
},
|
4899 |
+
{
|
4900 |
+
"epoch": 0.13648396901657026,
|
4901 |
+
"eval_loss": 0.40602046251296997,
|
4902 |
+
"eval_runtime": 78.837,
|
4903 |
+
"eval_samples_per_second": 27.246,
|
4904 |
+
"eval_steps_per_second": 13.623,
|
4905 |
+
"step": 696
|
4906 |
}
|
4907 |
],
|
4908 |
"logging_steps": 1,
|
|
|
4922 |
"attributes": {}
|
4923 |
}
|
4924 |
},
|
4925 |
+
"total_flos": 1.498972556277842e+17,
|
4926 |
"train_batch_size": 2,
|
4927 |
"trial_name": null,
|
4928 |
"trial_params": null
|