Training in progress, step 710, 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:c697e2ed1b482ec16a63e74e225c1f1f09d12d31e958814ccc91ba74a24608c8
|
| 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 sha256:154077384bf3c6df248077d6b6383ead0b226736aad66daa3149d3f0b8d1a54a
|
| 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 sha256:4cd027f5ae87b577ed5a4fa3457707907f6254fa67a2de6f1a39e26db0b2b2ab
|
| 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 sha256:73af69cd6b5c51f4c2d1771021579a9a5f370051abeeccd5aec8430ea77c8c75
|
| 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": 355,
|
| 6 |
-
"global_step":
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
@@ -2508,6 +2508,2499 @@
|
|
| 2508 |
"eval_samples_per_second": 17.319,
|
| 2509 |
"eval_steps_per_second": 8.668,
|
| 2510 |
"step": 355
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 2511 |
}
|
| 2512 |
],
|
| 2513 |
"logging_steps": 1,
|
|
@@ -2527,7 +5020,7 @@
|
|
| 2527 |
"attributes": {}
|
| 2528 |
}
|
| 2529 |
},
|
| 2530 |
-
"total_flos":
|
| 2531 |
"train_batch_size": 2,
|
| 2532 |
"trial_name": null,
|
| 2533 |
"trial_params": null
|
|
|
|
| 1 |
{
|
| 2 |
"best_metric": null,
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 0.28035538005923,
|
| 5 |
"eval_steps": 355,
|
| 6 |
+
"global_step": 710,
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
|
|
| 2508 |
"eval_samples_per_second": 17.319,
|
| 2509 |
"eval_steps_per_second": 8.668,
|
| 2510 |
"step": 355
|
| 2511 |
+
},
|
| 2512 |
+
{
|
| 2513 |
+
"epoch": 0.1405725567620928,
|
| 2514 |
+
"grad_norm": 0.5210056304931641,
|
| 2515 |
+
"learning_rate": 0.000171727295664559,
|
| 2516 |
+
"loss": 1.1635,
|
| 2517 |
+
"step": 356
|
| 2518 |
+
},
|
| 2519 |
+
{
|
| 2520 |
+
"epoch": 0.1409674234945706,
|
| 2521 |
+
"grad_norm": 0.5472628474235535,
|
| 2522 |
+
"learning_rate": 0.00017157186686278766,
|
| 2523 |
+
"loss": 1.2106,
|
| 2524 |
+
"step": 357
|
| 2525 |
+
},
|
| 2526 |
+
{
|
| 2527 |
+
"epoch": 0.14136229022704838,
|
| 2528 |
+
"grad_norm": 0.459087073802948,
|
| 2529 |
+
"learning_rate": 0.00017141608275408006,
|
| 2530 |
+
"loss": 1.0337,
|
| 2531 |
+
"step": 358
|
| 2532 |
+
},
|
| 2533 |
+
{
|
| 2534 |
+
"epoch": 0.14175715695952615,
|
| 2535 |
+
"grad_norm": 0.41874152421951294,
|
| 2536 |
+
"learning_rate": 0.00017125994411180124,
|
| 2537 |
+
"loss": 0.8032,
|
| 2538 |
+
"step": 359
|
| 2539 |
+
},
|
| 2540 |
+
{
|
| 2541 |
+
"epoch": 0.14215202369200394,
|
| 2542 |
+
"grad_norm": 0.4521096348762512,
|
| 2543 |
+
"learning_rate": 0.0001711034517110761,
|
| 2544 |
+
"loss": 0.9549,
|
| 2545 |
+
"step": 360
|
| 2546 |
+
},
|
| 2547 |
+
{
|
| 2548 |
+
"epoch": 0.14254689042448174,
|
| 2549 |
+
"grad_norm": 0.48767751455307007,
|
| 2550 |
+
"learning_rate": 0.00017094660632878582,
|
| 2551 |
+
"loss": 0.9779,
|
| 2552 |
+
"step": 361
|
| 2553 |
+
},
|
| 2554 |
+
{
|
| 2555 |
+
"epoch": 0.14294175715695953,
|
| 2556 |
+
"grad_norm": 0.4864053428173065,
|
| 2557 |
+
"learning_rate": 0.00017078940874356392,
|
| 2558 |
+
"loss": 0.7642,
|
| 2559 |
+
"step": 362
|
| 2560 |
+
},
|
| 2561 |
+
{
|
| 2562 |
+
"epoch": 0.14333662388943733,
|
| 2563 |
+
"grad_norm": 0.46765899658203125,
|
| 2564 |
+
"learning_rate": 0.00017063185973579232,
|
| 2565 |
+
"loss": 1.0457,
|
| 2566 |
+
"step": 363
|
| 2567 |
+
},
|
| 2568 |
+
{
|
| 2569 |
+
"epoch": 0.1437314906219151,
|
| 2570 |
+
"grad_norm": 0.4682892858982086,
|
| 2571 |
+
"learning_rate": 0.00017047396008759754,
|
| 2572 |
+
"loss": 0.974,
|
| 2573 |
+
"step": 364
|
| 2574 |
+
},
|
| 2575 |
+
{
|
| 2576 |
+
"epoch": 0.1441263573543929,
|
| 2577 |
+
"grad_norm": 0.4890439212322235,
|
| 2578 |
+
"learning_rate": 0.00017031571058284678,
|
| 2579 |
+
"loss": 0.9047,
|
| 2580 |
+
"step": 365
|
| 2581 |
+
},
|
| 2582 |
+
{
|
| 2583 |
+
"epoch": 0.14452122408687068,
|
| 2584 |
+
"grad_norm": 0.4934488832950592,
|
| 2585 |
+
"learning_rate": 0.00017015711200714414,
|
| 2586 |
+
"loss": 1.1,
|
| 2587 |
+
"step": 366
|
| 2588 |
+
},
|
| 2589 |
+
{
|
| 2590 |
+
"epoch": 0.14491609081934848,
|
| 2591 |
+
"grad_norm": 0.5664051175117493,
|
| 2592 |
+
"learning_rate": 0.00016999816514782647,
|
| 2593 |
+
"loss": 1.0985,
|
| 2594 |
+
"step": 367
|
| 2595 |
+
},
|
| 2596 |
+
{
|
| 2597 |
+
"epoch": 0.14531095755182627,
|
| 2598 |
+
"grad_norm": 0.5750765800476074,
|
| 2599 |
+
"learning_rate": 0.00016983887079395974,
|
| 2600 |
+
"loss": 0.936,
|
| 2601 |
+
"step": 368
|
| 2602 |
+
},
|
| 2603 |
+
{
|
| 2604 |
+
"epoch": 0.14570582428430404,
|
| 2605 |
+
"grad_norm": 0.49568259716033936,
|
| 2606 |
+
"learning_rate": 0.00016967922973633494,
|
| 2607 |
+
"loss": 1.1489,
|
| 2608 |
+
"step": 369
|
| 2609 |
+
},
|
| 2610 |
+
{
|
| 2611 |
+
"epoch": 0.14610069101678183,
|
| 2612 |
+
"grad_norm": 0.5348165035247803,
|
| 2613 |
+
"learning_rate": 0.00016951924276746425,
|
| 2614 |
+
"loss": 1.1904,
|
| 2615 |
+
"step": 370
|
| 2616 |
+
},
|
| 2617 |
+
{
|
| 2618 |
+
"epoch": 0.14649555774925963,
|
| 2619 |
+
"grad_norm": 0.4326549470424652,
|
| 2620 |
+
"learning_rate": 0.00016935891068157704,
|
| 2621 |
+
"loss": 0.8516,
|
| 2622 |
+
"step": 371
|
| 2623 |
+
},
|
| 2624 |
+
{
|
| 2625 |
+
"epoch": 0.14689042448173742,
|
| 2626 |
+
"grad_norm": 0.5540488362312317,
|
| 2627 |
+
"learning_rate": 0.000169198234274616,
|
| 2628 |
+
"loss": 1.0405,
|
| 2629 |
+
"step": 372
|
| 2630 |
+
},
|
| 2631 |
+
{
|
| 2632 |
+
"epoch": 0.1472852912142152,
|
| 2633 |
+
"grad_norm": 0.5236465930938721,
|
| 2634 |
+
"learning_rate": 0.00016903721434423306,
|
| 2635 |
+
"loss": 0.9151,
|
| 2636 |
+
"step": 373
|
| 2637 |
+
},
|
| 2638 |
+
{
|
| 2639 |
+
"epoch": 0.14768015794669298,
|
| 2640 |
+
"grad_norm": 0.5312307476997375,
|
| 2641 |
+
"learning_rate": 0.00016887585168978562,
|
| 2642 |
+
"loss": 1.1763,
|
| 2643 |
+
"step": 374
|
| 2644 |
+
},
|
| 2645 |
+
{
|
| 2646 |
+
"epoch": 0.14807502467917077,
|
| 2647 |
+
"grad_norm": 0.4888836145401001,
|
| 2648 |
+
"learning_rate": 0.0001687141471123324,
|
| 2649 |
+
"loss": 1.0494,
|
| 2650 |
+
"step": 375
|
| 2651 |
+
},
|
| 2652 |
+
{
|
| 2653 |
+
"epoch": 0.14846989141164857,
|
| 2654 |
+
"grad_norm": 0.49629417061805725,
|
| 2655 |
+
"learning_rate": 0.00016855210141462963,
|
| 2656 |
+
"loss": 0.697,
|
| 2657 |
+
"step": 376
|
| 2658 |
+
},
|
| 2659 |
+
{
|
| 2660 |
+
"epoch": 0.14886475814412636,
|
| 2661 |
+
"grad_norm": 0.5097388029098511,
|
| 2662 |
+
"learning_rate": 0.0001683897154011269,
|
| 2663 |
+
"loss": 1.0308,
|
| 2664 |
+
"step": 377
|
| 2665 |
+
},
|
| 2666 |
+
{
|
| 2667 |
+
"epoch": 0.14925962487660416,
|
| 2668 |
+
"grad_norm": 0.4330967664718628,
|
| 2669 |
+
"learning_rate": 0.0001682269898779632,
|
| 2670 |
+
"loss": 0.8466,
|
| 2671 |
+
"step": 378
|
| 2672 |
+
},
|
| 2673 |
+
{
|
| 2674 |
+
"epoch": 0.14965449160908192,
|
| 2675 |
+
"grad_norm": 0.4622458219528198,
|
| 2676 |
+
"learning_rate": 0.00016806392565296311,
|
| 2677 |
+
"loss": 0.849,
|
| 2678 |
+
"step": 379
|
| 2679 |
+
},
|
| 2680 |
+
{
|
| 2681 |
+
"epoch": 0.15004935834155972,
|
| 2682 |
+
"grad_norm": 0.4904235303401947,
|
| 2683 |
+
"learning_rate": 0.00016790052353563253,
|
| 2684 |
+
"loss": 1.0324,
|
| 2685 |
+
"step": 380
|
| 2686 |
+
},
|
| 2687 |
+
{
|
| 2688 |
+
"epoch": 0.1504442250740375,
|
| 2689 |
+
"grad_norm": 0.5324286222457886,
|
| 2690 |
+
"learning_rate": 0.00016773678433715475,
|
| 2691 |
+
"loss": 0.905,
|
| 2692 |
+
"step": 381
|
| 2693 |
+
},
|
| 2694 |
+
{
|
| 2695 |
+
"epoch": 0.1508390918065153,
|
| 2696 |
+
"grad_norm": 0.5701109766960144,
|
| 2697 |
+
"learning_rate": 0.00016757270887038654,
|
| 2698 |
+
"loss": 1.1105,
|
| 2699 |
+
"step": 382
|
| 2700 |
+
},
|
| 2701 |
+
{
|
| 2702 |
+
"epoch": 0.1512339585389931,
|
| 2703 |
+
"grad_norm": 0.45347732305526733,
|
| 2704 |
+
"learning_rate": 0.00016740829794985394,
|
| 2705 |
+
"loss": 0.875,
|
| 2706 |
+
"step": 383
|
| 2707 |
+
},
|
| 2708 |
+
{
|
| 2709 |
+
"epoch": 0.15162882527147087,
|
| 2710 |
+
"grad_norm": 0.4763396084308624,
|
| 2711 |
+
"learning_rate": 0.00016724355239174833,
|
| 2712 |
+
"loss": 0.9732,
|
| 2713 |
+
"step": 384
|
| 2714 |
+
},
|
| 2715 |
+
{
|
| 2716 |
+
"epoch": 0.15202369200394866,
|
| 2717 |
+
"grad_norm": 0.4246227443218231,
|
| 2718 |
+
"learning_rate": 0.00016707847301392236,
|
| 2719 |
+
"loss": 0.9235,
|
| 2720 |
+
"step": 385
|
| 2721 |
+
},
|
| 2722 |
+
{
|
| 2723 |
+
"epoch": 0.15241855873642646,
|
| 2724 |
+
"grad_norm": 0.6034351587295532,
|
| 2725 |
+
"learning_rate": 0.00016691306063588583,
|
| 2726 |
+
"loss": 0.8553,
|
| 2727 |
+
"step": 386
|
| 2728 |
+
},
|
| 2729 |
+
{
|
| 2730 |
+
"epoch": 0.15281342546890425,
|
| 2731 |
+
"grad_norm": 0.48924902081489563,
|
| 2732 |
+
"learning_rate": 0.0001667473160788017,
|
| 2733 |
+
"loss": 0.967,
|
| 2734 |
+
"step": 387
|
| 2735 |
+
},
|
| 2736 |
+
{
|
| 2737 |
+
"epoch": 0.15320829220138205,
|
| 2738 |
+
"grad_norm": 0.5340859889984131,
|
| 2739 |
+
"learning_rate": 0.00016658124016548197,
|
| 2740 |
+
"loss": 1.1052,
|
| 2741 |
+
"step": 388
|
| 2742 |
+
},
|
| 2743 |
+
{
|
| 2744 |
+
"epoch": 0.1536031589338598,
|
| 2745 |
+
"grad_norm": 0.5492063760757446,
|
| 2746 |
+
"learning_rate": 0.0001664148337203836,
|
| 2747 |
+
"loss": 1.1952,
|
| 2748 |
+
"step": 389
|
| 2749 |
+
},
|
| 2750 |
+
{
|
| 2751 |
+
"epoch": 0.1539980256663376,
|
| 2752 |
+
"grad_norm": 0.4856433868408203,
|
| 2753 |
+
"learning_rate": 0.00016624809756960444,
|
| 2754 |
+
"loss": 0.9994,
|
| 2755 |
+
"step": 390
|
| 2756 |
+
},
|
| 2757 |
+
{
|
| 2758 |
+
"epoch": 0.1543928923988154,
|
| 2759 |
+
"grad_norm": 0.4147414565086365,
|
| 2760 |
+
"learning_rate": 0.00016608103254087906,
|
| 2761 |
+
"loss": 0.9976,
|
| 2762 |
+
"step": 391
|
| 2763 |
+
},
|
| 2764 |
+
{
|
| 2765 |
+
"epoch": 0.1547877591312932,
|
| 2766 |
+
"grad_norm": 0.5175687074661255,
|
| 2767 |
+
"learning_rate": 0.00016591363946357474,
|
| 2768 |
+
"loss": 1.0245,
|
| 2769 |
+
"step": 392
|
| 2770 |
+
},
|
| 2771 |
+
{
|
| 2772 |
+
"epoch": 0.155182625863771,
|
| 2773 |
+
"grad_norm": 0.620985209941864,
|
| 2774 |
+
"learning_rate": 0.00016574591916868728,
|
| 2775 |
+
"loss": 1.1981,
|
| 2776 |
+
"step": 393
|
| 2777 |
+
},
|
| 2778 |
+
{
|
| 2779 |
+
"epoch": 0.15557749259624876,
|
| 2780 |
+
"grad_norm": 0.42090892791748047,
|
| 2781 |
+
"learning_rate": 0.00016557787248883696,
|
| 2782 |
+
"loss": 0.9171,
|
| 2783 |
+
"step": 394
|
| 2784 |
+
},
|
| 2785 |
+
{
|
| 2786 |
+
"epoch": 0.15597235932872655,
|
| 2787 |
+
"grad_norm": 0.6142613291740417,
|
| 2788 |
+
"learning_rate": 0.00016540950025826422,
|
| 2789 |
+
"loss": 1.0901,
|
| 2790 |
+
"step": 395
|
| 2791 |
+
},
|
| 2792 |
+
{
|
| 2793 |
+
"epoch": 0.15636722606120435,
|
| 2794 |
+
"grad_norm": 0.6569430232048035,
|
| 2795 |
+
"learning_rate": 0.00016524080331282577,
|
| 2796 |
+
"loss": 1.0362,
|
| 2797 |
+
"step": 396
|
| 2798 |
+
},
|
| 2799 |
+
{
|
| 2800 |
+
"epoch": 0.15676209279368214,
|
| 2801 |
+
"grad_norm": 0.5111309885978699,
|
| 2802 |
+
"learning_rate": 0.00016507178248999024,
|
| 2803 |
+
"loss": 0.9666,
|
| 2804 |
+
"step": 397
|
| 2805 |
+
},
|
| 2806 |
+
{
|
| 2807 |
+
"epoch": 0.15715695952615993,
|
| 2808 |
+
"grad_norm": 0.5645838975906372,
|
| 2809 |
+
"learning_rate": 0.00016490243862883413,
|
| 2810 |
+
"loss": 0.9295,
|
| 2811 |
+
"step": 398
|
| 2812 |
+
},
|
| 2813 |
+
{
|
| 2814 |
+
"epoch": 0.1575518262586377,
|
| 2815 |
+
"grad_norm": 0.46207594871520996,
|
| 2816 |
+
"learning_rate": 0.00016473277257003757,
|
| 2817 |
+
"loss": 1.0462,
|
| 2818 |
+
"step": 399
|
| 2819 |
+
},
|
| 2820 |
+
{
|
| 2821 |
+
"epoch": 0.1579466929911155,
|
| 2822 |
+
"grad_norm": 0.5260053873062134,
|
| 2823 |
+
"learning_rate": 0.00016456278515588024,
|
| 2824 |
+
"loss": 0.9745,
|
| 2825 |
+
"step": 400
|
| 2826 |
+
},
|
| 2827 |
+
{
|
| 2828 |
+
"epoch": 0.1583415597235933,
|
| 2829 |
+
"grad_norm": 0.5953394174575806,
|
| 2830 |
+
"learning_rate": 0.00016439247723023712,
|
| 2831 |
+
"loss": 1.0208,
|
| 2832 |
+
"step": 401
|
| 2833 |
+
},
|
| 2834 |
+
{
|
| 2835 |
+
"epoch": 0.15873642645607108,
|
| 2836 |
+
"grad_norm": 0.5679177641868591,
|
| 2837 |
+
"learning_rate": 0.00016422184963857432,
|
| 2838 |
+
"loss": 0.8679,
|
| 2839 |
+
"step": 402
|
| 2840 |
+
},
|
| 2841 |
+
{
|
| 2842 |
+
"epoch": 0.15913129318854888,
|
| 2843 |
+
"grad_norm": 0.47382932901382446,
|
| 2844 |
+
"learning_rate": 0.00016405090322794483,
|
| 2845 |
+
"loss": 0.9579,
|
| 2846 |
+
"step": 403
|
| 2847 |
+
},
|
| 2848 |
+
{
|
| 2849 |
+
"epoch": 0.15952615992102664,
|
| 2850 |
+
"grad_norm": 0.49775707721710205,
|
| 2851 |
+
"learning_rate": 0.00016387963884698448,
|
| 2852 |
+
"loss": 1.2399,
|
| 2853 |
+
"step": 404
|
| 2854 |
+
},
|
| 2855 |
+
{
|
| 2856 |
+
"epoch": 0.15992102665350444,
|
| 2857 |
+
"grad_norm": 0.49022993445396423,
|
| 2858 |
+
"learning_rate": 0.00016370805734590747,
|
| 2859 |
+
"loss": 1.0984,
|
| 2860 |
+
"step": 405
|
| 2861 |
+
},
|
| 2862 |
+
{
|
| 2863 |
+
"epoch": 0.16031589338598223,
|
| 2864 |
+
"grad_norm": 0.5563966035842896,
|
| 2865 |
+
"learning_rate": 0.00016353615957650236,
|
| 2866 |
+
"loss": 0.9777,
|
| 2867 |
+
"step": 406
|
| 2868 |
+
},
|
| 2869 |
+
{
|
| 2870 |
+
"epoch": 0.16071076011846003,
|
| 2871 |
+
"grad_norm": 0.46516212821006775,
|
| 2872 |
+
"learning_rate": 0.00016336394639212783,
|
| 2873 |
+
"loss": 0.8274,
|
| 2874 |
+
"step": 407
|
| 2875 |
+
},
|
| 2876 |
+
{
|
| 2877 |
+
"epoch": 0.16110562685093782,
|
| 2878 |
+
"grad_norm": 0.4698953926563263,
|
| 2879 |
+
"learning_rate": 0.00016319141864770827,
|
| 2880 |
+
"loss": 0.9419,
|
| 2881 |
+
"step": 408
|
| 2882 |
+
},
|
| 2883 |
+
{
|
| 2884 |
+
"epoch": 0.1615004935834156,
|
| 2885 |
+
"grad_norm": 0.4373069405555725,
|
| 2886 |
+
"learning_rate": 0.00016301857719972976,
|
| 2887 |
+
"loss": 0.8235,
|
| 2888 |
+
"step": 409
|
| 2889 |
+
},
|
| 2890 |
+
{
|
| 2891 |
+
"epoch": 0.16189536031589338,
|
| 2892 |
+
"grad_norm": 0.6190903186798096,
|
| 2893 |
+
"learning_rate": 0.00016284542290623567,
|
| 2894 |
+
"loss": 0.9092,
|
| 2895 |
+
"step": 410
|
| 2896 |
+
},
|
| 2897 |
+
{
|
| 2898 |
+
"epoch": 0.16229022704837118,
|
| 2899 |
+
"grad_norm": 0.5729885101318359,
|
| 2900 |
+
"learning_rate": 0.0001626719566268224,
|
| 2901 |
+
"loss": 0.9563,
|
| 2902 |
+
"step": 411
|
| 2903 |
+
},
|
| 2904 |
+
{
|
| 2905 |
+
"epoch": 0.16268509378084897,
|
| 2906 |
+
"grad_norm": 0.5004550814628601,
|
| 2907 |
+
"learning_rate": 0.00016249817922263517,
|
| 2908 |
+
"loss": 1.1011,
|
| 2909 |
+
"step": 412
|
| 2910 |
+
},
|
| 2911 |
+
{
|
| 2912 |
+
"epoch": 0.16307996051332677,
|
| 2913 |
+
"grad_norm": 0.4848381578922272,
|
| 2914 |
+
"learning_rate": 0.0001623240915563638,
|
| 2915 |
+
"loss": 1.0606,
|
| 2916 |
+
"step": 413
|
| 2917 |
+
},
|
| 2918 |
+
{
|
| 2919 |
+
"epoch": 0.16347482724580453,
|
| 2920 |
+
"grad_norm": 0.5314562916755676,
|
| 2921 |
+
"learning_rate": 0.00016214969449223824,
|
| 2922 |
+
"loss": 0.7504,
|
| 2923 |
+
"step": 414
|
| 2924 |
+
},
|
| 2925 |
+
{
|
| 2926 |
+
"epoch": 0.16386969397828233,
|
| 2927 |
+
"grad_norm": 0.5033890604972839,
|
| 2928 |
+
"learning_rate": 0.00016197498889602448,
|
| 2929 |
+
"loss": 1.0749,
|
| 2930 |
+
"step": 415
|
| 2931 |
+
},
|
| 2932 |
+
{
|
| 2933 |
+
"epoch": 0.16426456071076012,
|
| 2934 |
+
"grad_norm": 0.5094935894012451,
|
| 2935 |
+
"learning_rate": 0.0001617999756350202,
|
| 2936 |
+
"loss": 0.8851,
|
| 2937 |
+
"step": 416
|
| 2938 |
+
},
|
| 2939 |
+
{
|
| 2940 |
+
"epoch": 0.16465942744323792,
|
| 2941 |
+
"grad_norm": 0.4637800455093384,
|
| 2942 |
+
"learning_rate": 0.00016162465557805034,
|
| 2943 |
+
"loss": 0.8553,
|
| 2944 |
+
"step": 417
|
| 2945 |
+
},
|
| 2946 |
+
{
|
| 2947 |
+
"epoch": 0.16505429417571568,
|
| 2948 |
+
"grad_norm": 0.43399858474731445,
|
| 2949 |
+
"learning_rate": 0.00016144902959546286,
|
| 2950 |
+
"loss": 0.9576,
|
| 2951 |
+
"step": 418
|
| 2952 |
+
},
|
| 2953 |
+
{
|
| 2954 |
+
"epoch": 0.16544916090819348,
|
| 2955 |
+
"grad_norm": 0.5109582543373108,
|
| 2956 |
+
"learning_rate": 0.00016127309855912457,
|
| 2957 |
+
"loss": 1.1576,
|
| 2958 |
+
"step": 419
|
| 2959 |
+
},
|
| 2960 |
+
{
|
| 2961 |
+
"epoch": 0.16584402764067127,
|
| 2962 |
+
"grad_norm": 0.5534142851829529,
|
| 2963 |
+
"learning_rate": 0.00016109686334241655,
|
| 2964 |
+
"loss": 0.7701,
|
| 2965 |
+
"step": 420
|
| 2966 |
+
},
|
| 2967 |
+
{
|
| 2968 |
+
"epoch": 0.16623889437314907,
|
| 2969 |
+
"grad_norm": 0.46211493015289307,
|
| 2970 |
+
"learning_rate": 0.00016092032482023,
|
| 2971 |
+
"loss": 0.9263,
|
| 2972 |
+
"step": 421
|
| 2973 |
+
},
|
| 2974 |
+
{
|
| 2975 |
+
"epoch": 0.16663376110562686,
|
| 2976 |
+
"grad_norm": 0.5489344596862793,
|
| 2977 |
+
"learning_rate": 0.00016074348386896177,
|
| 2978 |
+
"loss": 1.0165,
|
| 2979 |
+
"step": 422
|
| 2980 |
+
},
|
| 2981 |
+
{
|
| 2982 |
+
"epoch": 0.16702862783810463,
|
| 2983 |
+
"grad_norm": 0.4893771708011627,
|
| 2984 |
+
"learning_rate": 0.0001605663413665102,
|
| 2985 |
+
"loss": 1.0284,
|
| 2986 |
+
"step": 423
|
| 2987 |
+
},
|
| 2988 |
+
{
|
| 2989 |
+
"epoch": 0.16742349457058242,
|
| 2990 |
+
"grad_norm": 0.46742960810661316,
|
| 2991 |
+
"learning_rate": 0.00016038889819227045,
|
| 2992 |
+
"loss": 1.0603,
|
| 2993 |
+
"step": 424
|
| 2994 |
+
},
|
| 2995 |
+
{
|
| 2996 |
+
"epoch": 0.16781836130306022,
|
| 2997 |
+
"grad_norm": 0.6020311713218689,
|
| 2998 |
+
"learning_rate": 0.00016021115522713047,
|
| 2999 |
+
"loss": 0.8771,
|
| 3000 |
+
"step": 425
|
| 3001 |
+
},
|
| 3002 |
+
{
|
| 3003 |
+
"epoch": 0.168213228035538,
|
| 3004 |
+
"grad_norm": 0.5273780822753906,
|
| 3005 |
+
"learning_rate": 0.00016003311335346636,
|
| 3006 |
+
"loss": 0.9648,
|
| 3007 |
+
"step": 426
|
| 3008 |
+
},
|
| 3009 |
+
{
|
| 3010 |
+
"epoch": 0.1686080947680158,
|
| 3011 |
+
"grad_norm": 0.5141377449035645,
|
| 3012 |
+
"learning_rate": 0.00015985477345513817,
|
| 3013 |
+
"loss": 0.8773,
|
| 3014 |
+
"step": 427
|
| 3015 |
+
},
|
| 3016 |
+
{
|
| 3017 |
+
"epoch": 0.16900296150049357,
|
| 3018 |
+
"grad_norm": 0.4852812588214874,
|
| 3019 |
+
"learning_rate": 0.00015967613641748542,
|
| 3020 |
+
"loss": 0.7764,
|
| 3021 |
+
"step": 428
|
| 3022 |
+
},
|
| 3023 |
+
{
|
| 3024 |
+
"epoch": 0.16939782823297136,
|
| 3025 |
+
"grad_norm": 0.43280699849128723,
|
| 3026 |
+
"learning_rate": 0.0001594972031273228,
|
| 3027 |
+
"loss": 0.7407,
|
| 3028 |
+
"step": 429
|
| 3029 |
+
},
|
| 3030 |
+
{
|
| 3031 |
+
"epoch": 0.16979269496544916,
|
| 3032 |
+
"grad_norm": 0.5069910883903503,
|
| 3033 |
+
"learning_rate": 0.00015931797447293552,
|
| 3034 |
+
"loss": 0.9313,
|
| 3035 |
+
"step": 430
|
| 3036 |
+
},
|
| 3037 |
+
{
|
| 3038 |
+
"epoch": 0.17018756169792695,
|
| 3039 |
+
"grad_norm": 0.5123517513275146,
|
| 3040 |
+
"learning_rate": 0.00015913845134407533,
|
| 3041 |
+
"loss": 1.0705,
|
| 3042 |
+
"step": 431
|
| 3043 |
+
},
|
| 3044 |
+
{
|
| 3045 |
+
"epoch": 0.17058242843040475,
|
| 3046 |
+
"grad_norm": 0.724429190158844,
|
| 3047 |
+
"learning_rate": 0.00015895863463195558,
|
| 3048 |
+
"loss": 0.9353,
|
| 3049 |
+
"step": 432
|
| 3050 |
+
},
|
| 3051 |
+
{
|
| 3052 |
+
"epoch": 0.17097729516288251,
|
| 3053 |
+
"grad_norm": 0.5428094267845154,
|
| 3054 |
+
"learning_rate": 0.00015877852522924732,
|
| 3055 |
+
"loss": 1.1227,
|
| 3056 |
+
"step": 433
|
| 3057 |
+
},
|
| 3058 |
+
{
|
| 3059 |
+
"epoch": 0.1713721618953603,
|
| 3060 |
+
"grad_norm": 0.6102519631385803,
|
| 3061 |
+
"learning_rate": 0.00015859812403007443,
|
| 3062 |
+
"loss": 0.9353,
|
| 3063 |
+
"step": 434
|
| 3064 |
+
},
|
| 3065 |
+
{
|
| 3066 |
+
"epoch": 0.1717670286278381,
|
| 3067 |
+
"grad_norm": 0.5151787400245667,
|
| 3068 |
+
"learning_rate": 0.00015841743193000944,
|
| 3069 |
+
"loss": 0.9646,
|
| 3070 |
+
"step": 435
|
| 3071 |
+
},
|
| 3072 |
+
{
|
| 3073 |
+
"epoch": 0.1721618953603159,
|
| 3074 |
+
"grad_norm": 0.6272695064544678,
|
| 3075 |
+
"learning_rate": 0.00015823644982606905,
|
| 3076 |
+
"loss": 0.8384,
|
| 3077 |
+
"step": 436
|
| 3078 |
+
},
|
| 3079 |
+
{
|
| 3080 |
+
"epoch": 0.1725567620927937,
|
| 3081 |
+
"grad_norm": 0.49354809522628784,
|
| 3082 |
+
"learning_rate": 0.00015805517861670952,
|
| 3083 |
+
"loss": 0.7855,
|
| 3084 |
+
"step": 437
|
| 3085 |
+
},
|
| 3086 |
+
{
|
| 3087 |
+
"epoch": 0.17295162882527146,
|
| 3088 |
+
"grad_norm": 0.47559288144111633,
|
| 3089 |
+
"learning_rate": 0.0001578736192018224,
|
| 3090 |
+
"loss": 0.9591,
|
| 3091 |
+
"step": 438
|
| 3092 |
+
},
|
| 3093 |
+
{
|
| 3094 |
+
"epoch": 0.17334649555774925,
|
| 3095 |
+
"grad_norm": 0.5615376830101013,
|
| 3096 |
+
"learning_rate": 0.00015769177248273008,
|
| 3097 |
+
"loss": 1.1537,
|
| 3098 |
+
"step": 439
|
| 3099 |
+
},
|
| 3100 |
+
{
|
| 3101 |
+
"epoch": 0.17374136229022705,
|
| 3102 |
+
"grad_norm": 0.5301774144172668,
|
| 3103 |
+
"learning_rate": 0.00015750963936218105,
|
| 3104 |
+
"loss": 0.7773,
|
| 3105 |
+
"step": 440
|
| 3106 |
+
},
|
| 3107 |
+
{
|
| 3108 |
+
"epoch": 0.17413622902270484,
|
| 3109 |
+
"grad_norm": 0.5083664059638977,
|
| 3110 |
+
"learning_rate": 0.0001573272207443457,
|
| 3111 |
+
"loss": 0.9705,
|
| 3112 |
+
"step": 441
|
| 3113 |
+
},
|
| 3114 |
+
{
|
| 3115 |
+
"epoch": 0.17453109575518264,
|
| 3116 |
+
"grad_norm": 0.5633112788200378,
|
| 3117 |
+
"learning_rate": 0.00015714451753481168,
|
| 3118 |
+
"loss": 1.0109,
|
| 3119 |
+
"step": 442
|
| 3120 |
+
},
|
| 3121 |
+
{
|
| 3122 |
+
"epoch": 0.1749259624876604,
|
| 3123 |
+
"grad_norm": 0.5243581533432007,
|
| 3124 |
+
"learning_rate": 0.00015696153064057947,
|
| 3125 |
+
"loss": 1.0258,
|
| 3126 |
+
"step": 443
|
| 3127 |
+
},
|
| 3128 |
+
{
|
| 3129 |
+
"epoch": 0.1753208292201382,
|
| 3130 |
+
"grad_norm": 0.6054911613464355,
|
| 3131 |
+
"learning_rate": 0.0001567782609700579,
|
| 3132 |
+
"loss": 1.0102,
|
| 3133 |
+
"step": 444
|
| 3134 |
+
},
|
| 3135 |
+
{
|
| 3136 |
+
"epoch": 0.175715695952616,
|
| 3137 |
+
"grad_norm": 0.5889274477958679,
|
| 3138 |
+
"learning_rate": 0.00015659470943305955,
|
| 3139 |
+
"loss": 1.1549,
|
| 3140 |
+
"step": 445
|
| 3141 |
+
},
|
| 3142 |
+
{
|
| 3143 |
+
"epoch": 0.17611056268509379,
|
| 3144 |
+
"grad_norm": 0.5765202045440674,
|
| 3145 |
+
"learning_rate": 0.0001564108769407962,
|
| 3146 |
+
"loss": 0.7791,
|
| 3147 |
+
"step": 446
|
| 3148 |
+
},
|
| 3149 |
+
{
|
| 3150 |
+
"epoch": 0.17650542941757158,
|
| 3151 |
+
"grad_norm": 0.5080841779708862,
|
| 3152 |
+
"learning_rate": 0.0001562267644058746,
|
| 3153 |
+
"loss": 0.9962,
|
| 3154 |
+
"step": 447
|
| 3155 |
+
},
|
| 3156 |
+
{
|
| 3157 |
+
"epoch": 0.17690029615004935,
|
| 3158 |
+
"grad_norm": 0.5185093879699707,
|
| 3159 |
+
"learning_rate": 0.00015604237274229147,
|
| 3160 |
+
"loss": 1.1927,
|
| 3161 |
+
"step": 448
|
| 3162 |
+
},
|
| 3163 |
+
{
|
| 3164 |
+
"epoch": 0.17729516288252714,
|
| 3165 |
+
"grad_norm": 0.5385391116142273,
|
| 3166 |
+
"learning_rate": 0.00015585770286542945,
|
| 3167 |
+
"loss": 1.0555,
|
| 3168 |
+
"step": 449
|
| 3169 |
+
},
|
| 3170 |
+
{
|
| 3171 |
+
"epoch": 0.17769002961500494,
|
| 3172 |
+
"grad_norm": 0.6289413571357727,
|
| 3173 |
+
"learning_rate": 0.00015567275569205218,
|
| 3174 |
+
"loss": 1.0431,
|
| 3175 |
+
"step": 450
|
| 3176 |
+
},
|
| 3177 |
+
{
|
| 3178 |
+
"epoch": 0.17808489634748273,
|
| 3179 |
+
"grad_norm": 0.6271052956581116,
|
| 3180 |
+
"learning_rate": 0.0001554875321402999,
|
| 3181 |
+
"loss": 1.0078,
|
| 3182 |
+
"step": 451
|
| 3183 |
+
},
|
| 3184 |
+
{
|
| 3185 |
+
"epoch": 0.17847976307996052,
|
| 3186 |
+
"grad_norm": 0.6165266633033752,
|
| 3187 |
+
"learning_rate": 0.00015530203312968502,
|
| 3188 |
+
"loss": 0.9761,
|
| 3189 |
+
"step": 452
|
| 3190 |
+
},
|
| 3191 |
+
{
|
| 3192 |
+
"epoch": 0.1788746298124383,
|
| 3193 |
+
"grad_norm": 0.4909960627555847,
|
| 3194 |
+
"learning_rate": 0.00015511625958108719,
|
| 3195 |
+
"loss": 1.061,
|
| 3196 |
+
"step": 453
|
| 3197 |
+
},
|
| 3198 |
+
{
|
| 3199 |
+
"epoch": 0.17926949654491608,
|
| 3200 |
+
"grad_norm": 0.5358797311782837,
|
| 3201 |
+
"learning_rate": 0.00015493021241674918,
|
| 3202 |
+
"loss": 1.0878,
|
| 3203 |
+
"step": 454
|
| 3204 |
+
},
|
| 3205 |
+
{
|
| 3206 |
+
"epoch": 0.17966436327739388,
|
| 3207 |
+
"grad_norm": 0.6802231073379517,
|
| 3208 |
+
"learning_rate": 0.000154743892560272,
|
| 3209 |
+
"loss": 1.0068,
|
| 3210 |
+
"step": 455
|
| 3211 |
+
},
|
| 3212 |
+
{
|
| 3213 |
+
"epoch": 0.18005923000987167,
|
| 3214 |
+
"grad_norm": 0.6705336570739746,
|
| 3215 |
+
"learning_rate": 0.00015455730093661034,
|
| 3216 |
+
"loss": 1.0845,
|
| 3217 |
+
"step": 456
|
| 3218 |
+
},
|
| 3219 |
+
{
|
| 3220 |
+
"epoch": 0.18045409674234947,
|
| 3221 |
+
"grad_norm": 0.43329593539237976,
|
| 3222 |
+
"learning_rate": 0.0001543704384720681,
|
| 3223 |
+
"loss": 0.8636,
|
| 3224 |
+
"step": 457
|
| 3225 |
+
},
|
| 3226 |
+
{
|
| 3227 |
+
"epoch": 0.18084896347482723,
|
| 3228 |
+
"grad_norm": 0.5168460011482239,
|
| 3229 |
+
"learning_rate": 0.0001541833060942937,
|
| 3230 |
+
"loss": 0.7497,
|
| 3231 |
+
"step": 458
|
| 3232 |
+
},
|
| 3233 |
+
{
|
| 3234 |
+
"epoch": 0.18124383020730503,
|
| 3235 |
+
"grad_norm": 0.4275985062122345,
|
| 3236 |
+
"learning_rate": 0.0001539959047322755,
|
| 3237 |
+
"loss": 0.9042,
|
| 3238 |
+
"step": 459
|
| 3239 |
+
},
|
| 3240 |
+
{
|
| 3241 |
+
"epoch": 0.18163869693978282,
|
| 3242 |
+
"grad_norm": 0.44853127002716064,
|
| 3243 |
+
"learning_rate": 0.00015380823531633729,
|
| 3244 |
+
"loss": 0.9091,
|
| 3245 |
+
"step": 460
|
| 3246 |
+
},
|
| 3247 |
+
{
|
| 3248 |
+
"epoch": 0.18203356367226062,
|
| 3249 |
+
"grad_norm": 0.4683418571949005,
|
| 3250 |
+
"learning_rate": 0.00015362029877813332,
|
| 3251 |
+
"loss": 0.8174,
|
| 3252 |
+
"step": 461
|
| 3253 |
+
},
|
| 3254 |
+
{
|
| 3255 |
+
"epoch": 0.1824284304047384,
|
| 3256 |
+
"grad_norm": 0.4813165068626404,
|
| 3257 |
+
"learning_rate": 0.00015343209605064422,
|
| 3258 |
+
"loss": 0.7648,
|
| 3259 |
+
"step": 462
|
| 3260 |
+
},
|
| 3261 |
+
{
|
| 3262 |
+
"epoch": 0.18282329713721618,
|
| 3263 |
+
"grad_norm": 0.47031188011169434,
|
| 3264 |
+
"learning_rate": 0.00015324362806817186,
|
| 3265 |
+
"loss": 0.8667,
|
| 3266 |
+
"step": 463
|
| 3267 |
+
},
|
| 3268 |
+
{
|
| 3269 |
+
"epoch": 0.18321816386969397,
|
| 3270 |
+
"grad_norm": 0.5744631886482239,
|
| 3271 |
+
"learning_rate": 0.00015305489576633504,
|
| 3272 |
+
"loss": 0.8219,
|
| 3273 |
+
"step": 464
|
| 3274 |
+
},
|
| 3275 |
+
{
|
| 3276 |
+
"epoch": 0.18361303060217177,
|
| 3277 |
+
"grad_norm": 0.48958727717399597,
|
| 3278 |
+
"learning_rate": 0.00015286590008206465,
|
| 3279 |
+
"loss": 1.018,
|
| 3280 |
+
"step": 465
|
| 3281 |
+
},
|
| 3282 |
+
{
|
| 3283 |
+
"epoch": 0.18400789733464956,
|
| 3284 |
+
"grad_norm": 0.5247324109077454,
|
| 3285 |
+
"learning_rate": 0.00015267664195359917,
|
| 3286 |
+
"loss": 0.8933,
|
| 3287 |
+
"step": 466
|
| 3288 |
+
},
|
| 3289 |
+
{
|
| 3290 |
+
"epoch": 0.18440276406712736,
|
| 3291 |
+
"grad_norm": 0.545970618724823,
|
| 3292 |
+
"learning_rate": 0.00015248712232047992,
|
| 3293 |
+
"loss": 1.0508,
|
| 3294 |
+
"step": 467
|
| 3295 |
+
},
|
| 3296 |
+
{
|
| 3297 |
+
"epoch": 0.18479763079960512,
|
| 3298 |
+
"grad_norm": 0.5140126347541809,
|
| 3299 |
+
"learning_rate": 0.0001522973421235464,
|
| 3300 |
+
"loss": 0.7581,
|
| 3301 |
+
"step": 468
|
| 3302 |
+
},
|
| 3303 |
+
{
|
| 3304 |
+
"epoch": 0.18519249753208292,
|
| 3305 |
+
"grad_norm": 0.4570896625518799,
|
| 3306 |
+
"learning_rate": 0.00015210730230493162,
|
| 3307 |
+
"loss": 1.0665,
|
| 3308 |
+
"step": 469
|
| 3309 |
+
},
|
| 3310 |
+
{
|
| 3311 |
+
"epoch": 0.1855873642645607,
|
| 3312 |
+
"grad_norm": 0.43096470832824707,
|
| 3313 |
+
"learning_rate": 0.00015191700380805752,
|
| 3314 |
+
"loss": 0.8313,
|
| 3315 |
+
"step": 470
|
| 3316 |
+
},
|
| 3317 |
+
{
|
| 3318 |
+
"epoch": 0.1859822309970385,
|
| 3319 |
+
"grad_norm": 0.460275799036026,
|
| 3320 |
+
"learning_rate": 0.00015172644757763015,
|
| 3321 |
+
"loss": 0.9575,
|
| 3322 |
+
"step": 471
|
| 3323 |
+
},
|
| 3324 |
+
{
|
| 3325 |
+
"epoch": 0.1863770977295163,
|
| 3326 |
+
"grad_norm": 0.4163447618484497,
|
| 3327 |
+
"learning_rate": 0.00015153563455963499,
|
| 3328 |
+
"loss": 0.8838,
|
| 3329 |
+
"step": 472
|
| 3330 |
+
},
|
| 3331 |
+
{
|
| 3332 |
+
"epoch": 0.18677196446199407,
|
| 3333 |
+
"grad_norm": 0.542241096496582,
|
| 3334 |
+
"learning_rate": 0.0001513445657013324,
|
| 3335 |
+
"loss": 0.8143,
|
| 3336 |
+
"step": 473
|
| 3337 |
+
},
|
| 3338 |
+
{
|
| 3339 |
+
"epoch": 0.18716683119447186,
|
| 3340 |
+
"grad_norm": 0.6614589691162109,
|
| 3341 |
+
"learning_rate": 0.00015115324195125274,
|
| 3342 |
+
"loss": 0.9645,
|
| 3343 |
+
"step": 474
|
| 3344 |
+
},
|
| 3345 |
+
{
|
| 3346 |
+
"epoch": 0.18756169792694966,
|
| 3347 |
+
"grad_norm": 0.4719527065753937,
|
| 3348 |
+
"learning_rate": 0.00015096166425919175,
|
| 3349 |
+
"loss": 0.9894,
|
| 3350 |
+
"step": 475
|
| 3351 |
+
},
|
| 3352 |
+
{
|
| 3353 |
+
"epoch": 0.18795656465942745,
|
| 3354 |
+
"grad_norm": 0.4972122013568878,
|
| 3355 |
+
"learning_rate": 0.0001507698335762059,
|
| 3356 |
+
"loss": 0.8865,
|
| 3357 |
+
"step": 476
|
| 3358 |
+
},
|
| 3359 |
+
{
|
| 3360 |
+
"epoch": 0.18835143139190524,
|
| 3361 |
+
"grad_norm": 0.5407273769378662,
|
| 3362 |
+
"learning_rate": 0.00015057775085460749,
|
| 3363 |
+
"loss": 0.8714,
|
| 3364 |
+
"step": 477
|
| 3365 |
+
},
|
| 3366 |
+
{
|
| 3367 |
+
"epoch": 0.188746298124383,
|
| 3368 |
+
"grad_norm": 0.4692353308200836,
|
| 3369 |
+
"learning_rate": 0.00015038541704796003,
|
| 3370 |
+
"loss": 0.9357,
|
| 3371 |
+
"step": 478
|
| 3372 |
+
},
|
| 3373 |
+
{
|
| 3374 |
+
"epoch": 0.1891411648568608,
|
| 3375 |
+
"grad_norm": 0.4748722314834595,
|
| 3376 |
+
"learning_rate": 0.00015019283311107367,
|
| 3377 |
+
"loss": 1.0376,
|
| 3378 |
+
"step": 479
|
| 3379 |
+
},
|
| 3380 |
+
{
|
| 3381 |
+
"epoch": 0.1895360315893386,
|
| 3382 |
+
"grad_norm": 0.4381348490715027,
|
| 3383 |
+
"learning_rate": 0.00015000000000000001,
|
| 3384 |
+
"loss": 1.0059,
|
| 3385 |
+
"step": 480
|
| 3386 |
+
},
|
| 3387 |
+
{
|
| 3388 |
+
"epoch": 0.1899308983218164,
|
| 3389 |
+
"grad_norm": 0.48987191915512085,
|
| 3390 |
+
"learning_rate": 0.0001498069186720279,
|
| 3391 |
+
"loss": 0.8901,
|
| 3392 |
+
"step": 481
|
| 3393 |
+
},
|
| 3394 |
+
{
|
| 3395 |
+
"epoch": 0.1903257650542942,
|
| 3396 |
+
"grad_norm": 0.5566233396530151,
|
| 3397 |
+
"learning_rate": 0.0001496135900856782,
|
| 3398 |
+
"loss": 1.1464,
|
| 3399 |
+
"step": 482
|
| 3400 |
+
},
|
| 3401 |
+
{
|
| 3402 |
+
"epoch": 0.19072063178677195,
|
| 3403 |
+
"grad_norm": 0.5021251440048218,
|
| 3404 |
+
"learning_rate": 0.00014942001520069947,
|
| 3405 |
+
"loss": 1.0947,
|
| 3406 |
+
"step": 483
|
| 3407 |
+
},
|
| 3408 |
+
{
|
| 3409 |
+
"epoch": 0.19111549851924975,
|
| 3410 |
+
"grad_norm": 0.5681881904602051,
|
| 3411 |
+
"learning_rate": 0.00014922619497806277,
|
| 3412 |
+
"loss": 1.0981,
|
| 3413 |
+
"step": 484
|
| 3414 |
+
},
|
| 3415 |
+
{
|
| 3416 |
+
"epoch": 0.19151036525172754,
|
| 3417 |
+
"grad_norm": 0.5182890892028809,
|
| 3418 |
+
"learning_rate": 0.00014903213037995724,
|
| 3419 |
+
"loss": 1.1017,
|
| 3420 |
+
"step": 485
|
| 3421 |
+
},
|
| 3422 |
+
{
|
| 3423 |
+
"epoch": 0.19190523198420534,
|
| 3424 |
+
"grad_norm": 0.4682919979095459,
|
| 3425 |
+
"learning_rate": 0.0001488378223697851,
|
| 3426 |
+
"loss": 0.6508,
|
| 3427 |
+
"step": 486
|
| 3428 |
+
},
|
| 3429 |
+
{
|
| 3430 |
+
"epoch": 0.19230009871668313,
|
| 3431 |
+
"grad_norm": 0.4784727096557617,
|
| 3432 |
+
"learning_rate": 0.00014864327191215702,
|
| 3433 |
+
"loss": 0.874,
|
| 3434 |
+
"step": 487
|
| 3435 |
+
},
|
| 3436 |
+
{
|
| 3437 |
+
"epoch": 0.1926949654491609,
|
| 3438 |
+
"grad_norm": 0.5247599482536316,
|
| 3439 |
+
"learning_rate": 0.00014844847997288717,
|
| 3440 |
+
"loss": 1.1797,
|
| 3441 |
+
"step": 488
|
| 3442 |
+
},
|
| 3443 |
+
{
|
| 3444 |
+
"epoch": 0.1930898321816387,
|
| 3445 |
+
"grad_norm": 0.48195531964302063,
|
| 3446 |
+
"learning_rate": 0.00014825344751898863,
|
| 3447 |
+
"loss": 1.0463,
|
| 3448 |
+
"step": 489
|
| 3449 |
+
},
|
| 3450 |
+
{
|
| 3451 |
+
"epoch": 0.1934846989141165,
|
| 3452 |
+
"grad_norm": 0.5093549489974976,
|
| 3453 |
+
"learning_rate": 0.00014805817551866838,
|
| 3454 |
+
"loss": 1.0409,
|
| 3455 |
+
"step": 490
|
| 3456 |
+
},
|
| 3457 |
+
{
|
| 3458 |
+
"epoch": 0.19387956564659428,
|
| 3459 |
+
"grad_norm": 0.5454416275024414,
|
| 3460 |
+
"learning_rate": 0.00014786266494132267,
|
| 3461 |
+
"loss": 1.035,
|
| 3462 |
+
"step": 491
|
| 3463 |
+
},
|
| 3464 |
+
{
|
| 3465 |
+
"epoch": 0.19427443237907208,
|
| 3466 |
+
"grad_norm": 0.49301639199256897,
|
| 3467 |
+
"learning_rate": 0.00014766691675753202,
|
| 3468 |
+
"loss": 1.1046,
|
| 3469 |
+
"step": 492
|
| 3470 |
+
},
|
| 3471 |
+
{
|
| 3472 |
+
"epoch": 0.19466929911154984,
|
| 3473 |
+
"grad_norm": 0.4534429907798767,
|
| 3474 |
+
"learning_rate": 0.00014747093193905657,
|
| 3475 |
+
"loss": 0.9061,
|
| 3476 |
+
"step": 493
|
| 3477 |
+
},
|
| 3478 |
+
{
|
| 3479 |
+
"epoch": 0.19506416584402764,
|
| 3480 |
+
"grad_norm": 0.599082887172699,
|
| 3481 |
+
"learning_rate": 0.00014727471145883127,
|
| 3482 |
+
"loss": 1.0882,
|
| 3483 |
+
"step": 494
|
| 3484 |
+
},
|
| 3485 |
+
{
|
| 3486 |
+
"epoch": 0.19545903257650543,
|
| 3487 |
+
"grad_norm": 0.5382128953933716,
|
| 3488 |
+
"learning_rate": 0.00014707825629096084,
|
| 3489 |
+
"loss": 1.0369,
|
| 3490 |
+
"step": 495
|
| 3491 |
+
},
|
| 3492 |
+
{
|
| 3493 |
+
"epoch": 0.19585389930898323,
|
| 3494 |
+
"grad_norm": 0.5784068703651428,
|
| 3495 |
+
"learning_rate": 0.00014688156741071514,
|
| 3496 |
+
"loss": 0.9614,
|
| 3497 |
+
"step": 496
|
| 3498 |
+
},
|
| 3499 |
+
{
|
| 3500 |
+
"epoch": 0.19624876604146102,
|
| 3501 |
+
"grad_norm": 0.5423576235771179,
|
| 3502 |
+
"learning_rate": 0.00014668464579452425,
|
| 3503 |
+
"loss": 0.9217,
|
| 3504 |
+
"step": 497
|
| 3505 |
+
},
|
| 3506 |
+
{
|
| 3507 |
+
"epoch": 0.1966436327739388,
|
| 3508 |
+
"grad_norm": 0.47836440801620483,
|
| 3509 |
+
"learning_rate": 0.00014648749241997363,
|
| 3510 |
+
"loss": 0.918,
|
| 3511 |
+
"step": 498
|
| 3512 |
+
},
|
| 3513 |
+
{
|
| 3514 |
+
"epoch": 0.19703849950641658,
|
| 3515 |
+
"grad_norm": 0.5091026425361633,
|
| 3516 |
+
"learning_rate": 0.00014629010826579928,
|
| 3517 |
+
"loss": 1.0415,
|
| 3518 |
+
"step": 499
|
| 3519 |
+
},
|
| 3520 |
+
{
|
| 3521 |
+
"epoch": 0.19743336623889438,
|
| 3522 |
+
"grad_norm": 0.553871214389801,
|
| 3523 |
+
"learning_rate": 0.00014609249431188278,
|
| 3524 |
+
"loss": 0.8315,
|
| 3525 |
+
"step": 500
|
| 3526 |
+
},
|
| 3527 |
+
{
|
| 3528 |
+
"epoch": 0.19782823297137217,
|
| 3529 |
+
"grad_norm": 0.4480036199092865,
|
| 3530 |
+
"learning_rate": 0.00014589465153924672,
|
| 3531 |
+
"loss": 0.9431,
|
| 3532 |
+
"step": 501
|
| 3533 |
+
},
|
| 3534 |
+
{
|
| 3535 |
+
"epoch": 0.19822309970384996,
|
| 3536 |
+
"grad_norm": 0.5340292453765869,
|
| 3537 |
+
"learning_rate": 0.00014569658093004935,
|
| 3538 |
+
"loss": 0.8736,
|
| 3539 |
+
"step": 502
|
| 3540 |
+
},
|
| 3541 |
+
{
|
| 3542 |
+
"epoch": 0.19861796643632773,
|
| 3543 |
+
"grad_norm": 0.5622652173042297,
|
| 3544 |
+
"learning_rate": 0.0001454982834675802,
|
| 3545 |
+
"loss": 1.1759,
|
| 3546 |
+
"step": 503
|
| 3547 |
+
},
|
| 3548 |
+
{
|
| 3549 |
+
"epoch": 0.19901283316880553,
|
| 3550 |
+
"grad_norm": 0.5641468167304993,
|
| 3551 |
+
"learning_rate": 0.00014529976013625482,
|
| 3552 |
+
"loss": 0.9721,
|
| 3553 |
+
"step": 504
|
| 3554 |
+
},
|
| 3555 |
+
{
|
| 3556 |
+
"epoch": 0.19940769990128332,
|
| 3557 |
+
"grad_norm": 0.49418380856513977,
|
| 3558 |
+
"learning_rate": 0.00014510101192161018,
|
| 3559 |
+
"loss": 0.8389,
|
| 3560 |
+
"step": 505
|
| 3561 |
+
},
|
| 3562 |
+
{
|
| 3563 |
+
"epoch": 0.19980256663376111,
|
| 3564 |
+
"grad_norm": 0.47910451889038086,
|
| 3565 |
+
"learning_rate": 0.0001449020398102996,
|
| 3566 |
+
"loss": 0.9339,
|
| 3567 |
+
"step": 506
|
| 3568 |
+
},
|
| 3569 |
+
{
|
| 3570 |
+
"epoch": 0.20019743336623888,
|
| 3571 |
+
"grad_norm": 0.5156650543212891,
|
| 3572 |
+
"learning_rate": 0.00014470284479008782,
|
| 3573 |
+
"loss": 0.9458,
|
| 3574 |
+
"step": 507
|
| 3575 |
+
},
|
| 3576 |
+
{
|
| 3577 |
+
"epoch": 0.20059230009871667,
|
| 3578 |
+
"grad_norm": 0.4681549072265625,
|
| 3579 |
+
"learning_rate": 0.00014450342784984633,
|
| 3580 |
+
"loss": 0.8954,
|
| 3581 |
+
"step": 508
|
| 3582 |
+
},
|
| 3583 |
+
{
|
| 3584 |
+
"epoch": 0.20098716683119447,
|
| 3585 |
+
"grad_norm": 0.5173560380935669,
|
| 3586 |
+
"learning_rate": 0.00014430378997954817,
|
| 3587 |
+
"loss": 1.0272,
|
| 3588 |
+
"step": 509
|
| 3589 |
+
},
|
| 3590 |
+
{
|
| 3591 |
+
"epoch": 0.20138203356367226,
|
| 3592 |
+
"grad_norm": 0.5966143012046814,
|
| 3593 |
+
"learning_rate": 0.00014410393217026318,
|
| 3594 |
+
"loss": 0.8915,
|
| 3595 |
+
"step": 510
|
| 3596 |
+
},
|
| 3597 |
+
{
|
| 3598 |
+
"epoch": 0.20177690029615006,
|
| 3599 |
+
"grad_norm": 0.5026108026504517,
|
| 3600 |
+
"learning_rate": 0.00014390385541415308,
|
| 3601 |
+
"loss": 0.9169,
|
| 3602 |
+
"step": 511
|
| 3603 |
+
},
|
| 3604 |
+
{
|
| 3605 |
+
"epoch": 0.20217176702862782,
|
| 3606 |
+
"grad_norm": 0.4867290258407593,
|
| 3607 |
+
"learning_rate": 0.00014370356070446654,
|
| 3608 |
+
"loss": 1.133,
|
| 3609 |
+
"step": 512
|
| 3610 |
+
},
|
| 3611 |
+
{
|
| 3612 |
+
"epoch": 0.20256663376110562,
|
| 3613 |
+
"grad_norm": 0.4962225556373596,
|
| 3614 |
+
"learning_rate": 0.00014350304903553416,
|
| 3615 |
+
"loss": 0.9498,
|
| 3616 |
+
"step": 513
|
| 3617 |
+
},
|
| 3618 |
+
{
|
| 3619 |
+
"epoch": 0.2029615004935834,
|
| 3620 |
+
"grad_norm": 0.4522517919540405,
|
| 3621 |
+
"learning_rate": 0.00014330232140276366,
|
| 3622 |
+
"loss": 0.8796,
|
| 3623 |
+
"step": 514
|
| 3624 |
+
},
|
| 3625 |
+
{
|
| 3626 |
+
"epoch": 0.2033563672260612,
|
| 3627 |
+
"grad_norm": 0.4583631455898285,
|
| 3628 |
+
"learning_rate": 0.00014310137880263482,
|
| 3629 |
+
"loss": 0.9822,
|
| 3630 |
+
"step": 515
|
| 3631 |
+
},
|
| 3632 |
+
{
|
| 3633 |
+
"epoch": 0.203751233958539,
|
| 3634 |
+
"grad_norm": 0.5668156147003174,
|
| 3635 |
+
"learning_rate": 0.00014290022223269463,
|
| 3636 |
+
"loss": 0.8197,
|
| 3637 |
+
"step": 516
|
| 3638 |
+
},
|
| 3639 |
+
{
|
| 3640 |
+
"epoch": 0.20414610069101677,
|
| 3641 |
+
"grad_norm": 0.47949913144111633,
|
| 3642 |
+
"learning_rate": 0.0001426988526915523,
|
| 3643 |
+
"loss": 0.9435,
|
| 3644 |
+
"step": 517
|
| 3645 |
+
},
|
| 3646 |
+
{
|
| 3647 |
+
"epoch": 0.20454096742349456,
|
| 3648 |
+
"grad_norm": 0.497577965259552,
|
| 3649 |
+
"learning_rate": 0.00014249727117887425,
|
| 3650 |
+
"loss": 0.8928,
|
| 3651 |
+
"step": 518
|
| 3652 |
+
},
|
| 3653 |
+
{
|
| 3654 |
+
"epoch": 0.20493583415597236,
|
| 3655 |
+
"grad_norm": 0.4772360622882843,
|
| 3656 |
+
"learning_rate": 0.0001422954786953793,
|
| 3657 |
+
"loss": 1.0171,
|
| 3658 |
+
"step": 519
|
| 3659 |
+
},
|
| 3660 |
+
{
|
| 3661 |
+
"epoch": 0.20533070088845015,
|
| 3662 |
+
"grad_norm": 0.5616466999053955,
|
| 3663 |
+
"learning_rate": 0.0001420934762428335,
|
| 3664 |
+
"loss": 0.9387,
|
| 3665 |
+
"step": 520
|
| 3666 |
+
},
|
| 3667 |
+
{
|
| 3668 |
+
"epoch": 0.20572556762092795,
|
| 3669 |
+
"grad_norm": 0.6362606287002563,
|
| 3670 |
+
"learning_rate": 0.00014189126482404532,
|
| 3671 |
+
"loss": 0.7291,
|
| 3672 |
+
"step": 521
|
| 3673 |
+
},
|
| 3674 |
+
{
|
| 3675 |
+
"epoch": 0.2061204343534057,
|
| 3676 |
+
"grad_norm": 0.4801354706287384,
|
| 3677 |
+
"learning_rate": 0.00014168884544286053,
|
| 3678 |
+
"loss": 0.7701,
|
| 3679 |
+
"step": 522
|
| 3680 |
+
},
|
| 3681 |
+
{
|
| 3682 |
+
"epoch": 0.2065153010858835,
|
| 3683 |
+
"grad_norm": 0.45586976408958435,
|
| 3684 |
+
"learning_rate": 0.0001414862191041574,
|
| 3685 |
+
"loss": 0.9861,
|
| 3686 |
+
"step": 523
|
| 3687 |
+
},
|
| 3688 |
+
{
|
| 3689 |
+
"epoch": 0.2069101678183613,
|
| 3690 |
+
"grad_norm": 0.5662283301353455,
|
| 3691 |
+
"learning_rate": 0.00014128338681384153,
|
| 3692 |
+
"loss": 0.9726,
|
| 3693 |
+
"step": 524
|
| 3694 |
+
},
|
| 3695 |
+
{
|
| 3696 |
+
"epoch": 0.2073050345508391,
|
| 3697 |
+
"grad_norm": 0.5532451868057251,
|
| 3698 |
+
"learning_rate": 0.00014108034957884094,
|
| 3699 |
+
"loss": 1.2107,
|
| 3700 |
+
"step": 525
|
| 3701 |
+
},
|
| 3702 |
+
{
|
| 3703 |
+
"epoch": 0.2076999012833169,
|
| 3704 |
+
"grad_norm": 0.5659233331680298,
|
| 3705 |
+
"learning_rate": 0.0001408771084071012,
|
| 3706 |
+
"loss": 0.9739,
|
| 3707 |
+
"step": 526
|
| 3708 |
+
},
|
| 3709 |
+
{
|
| 3710 |
+
"epoch": 0.20809476801579466,
|
| 3711 |
+
"grad_norm": 0.5252379775047302,
|
| 3712 |
+
"learning_rate": 0.00014067366430758004,
|
| 3713 |
+
"loss": 0.9744,
|
| 3714 |
+
"step": 527
|
| 3715 |
+
},
|
| 3716 |
+
{
|
| 3717 |
+
"epoch": 0.20848963474827245,
|
| 3718 |
+
"grad_norm": 0.6081037521362305,
|
| 3719 |
+
"learning_rate": 0.0001404700182902428,
|
| 3720 |
+
"loss": 1.0524,
|
| 3721 |
+
"step": 528
|
| 3722 |
+
},
|
| 3723 |
+
{
|
| 3724 |
+
"epoch": 0.20888450148075025,
|
| 3725 |
+
"grad_norm": 0.6748234033584595,
|
| 3726 |
+
"learning_rate": 0.0001402661713660571,
|
| 3727 |
+
"loss": 0.9468,
|
| 3728 |
+
"step": 529
|
| 3729 |
+
},
|
| 3730 |
+
{
|
| 3731 |
+
"epoch": 0.20927936821322804,
|
| 3732 |
+
"grad_norm": 0.4868841767311096,
|
| 3733 |
+
"learning_rate": 0.00014006212454698797,
|
| 3734 |
+
"loss": 1.0739,
|
| 3735 |
+
"step": 530
|
| 3736 |
+
},
|
| 3737 |
+
{
|
| 3738 |
+
"epoch": 0.20967423494570583,
|
| 3739 |
+
"grad_norm": 0.5020860433578491,
|
| 3740 |
+
"learning_rate": 0.00013985787884599282,
|
| 3741 |
+
"loss": 0.9586,
|
| 3742 |
+
"step": 531
|
| 3743 |
+
},
|
| 3744 |
+
{
|
| 3745 |
+
"epoch": 0.2100691016781836,
|
| 3746 |
+
"grad_norm": 0.579229474067688,
|
| 3747 |
+
"learning_rate": 0.00013965343527701628,
|
| 3748 |
+
"loss": 0.8937,
|
| 3749 |
+
"step": 532
|
| 3750 |
+
},
|
| 3751 |
+
{
|
| 3752 |
+
"epoch": 0.2104639684106614,
|
| 3753 |
+
"grad_norm": 0.48879000544548035,
|
| 3754 |
+
"learning_rate": 0.00013944879485498538,
|
| 3755 |
+
"loss": 0.956,
|
| 3756 |
+
"step": 533
|
| 3757 |
+
},
|
| 3758 |
+
{
|
| 3759 |
+
"epoch": 0.2108588351431392,
|
| 3760 |
+
"grad_norm": 0.5132958292961121,
|
| 3761 |
+
"learning_rate": 0.00013924395859580432,
|
| 3762 |
+
"loss": 0.8762,
|
| 3763 |
+
"step": 534
|
| 3764 |
+
},
|
| 3765 |
+
{
|
| 3766 |
+
"epoch": 0.21125370187561698,
|
| 3767 |
+
"grad_norm": 0.5426986217498779,
|
| 3768 |
+
"learning_rate": 0.00013903892751634947,
|
| 3769 |
+
"loss": 1.0018,
|
| 3770 |
+
"step": 535
|
| 3771 |
+
},
|
| 3772 |
+
{
|
| 3773 |
+
"epoch": 0.21164856860809478,
|
| 3774 |
+
"grad_norm": 0.5325112342834473,
|
| 3775 |
+
"learning_rate": 0.0001388337026344645,
|
| 3776 |
+
"loss": 0.9227,
|
| 3777 |
+
"step": 536
|
| 3778 |
+
},
|
| 3779 |
+
{
|
| 3780 |
+
"epoch": 0.21204343534057254,
|
| 3781 |
+
"grad_norm": 0.5148372650146484,
|
| 3782 |
+
"learning_rate": 0.000138628284968955,
|
| 3783 |
+
"loss": 1.0617,
|
| 3784 |
+
"step": 537
|
| 3785 |
+
},
|
| 3786 |
+
{
|
| 3787 |
+
"epoch": 0.21243830207305034,
|
| 3788 |
+
"grad_norm": 0.4726095199584961,
|
| 3789 |
+
"learning_rate": 0.00013842267553958371,
|
| 3790 |
+
"loss": 0.9038,
|
| 3791 |
+
"step": 538
|
| 3792 |
+
},
|
| 3793 |
+
{
|
| 3794 |
+
"epoch": 0.21283316880552813,
|
| 3795 |
+
"grad_norm": 0.504017174243927,
|
| 3796 |
+
"learning_rate": 0.00013821687536706533,
|
| 3797 |
+
"loss": 1.0946,
|
| 3798 |
+
"step": 539
|
| 3799 |
+
},
|
| 3800 |
+
{
|
| 3801 |
+
"epoch": 0.21322803553800593,
|
| 3802 |
+
"grad_norm": 0.483732670545578,
|
| 3803 |
+
"learning_rate": 0.00013801088547306148,
|
| 3804 |
+
"loss": 0.7506,
|
| 3805 |
+
"step": 540
|
| 3806 |
+
},
|
| 3807 |
+
{
|
| 3808 |
+
"epoch": 0.21362290227048372,
|
| 3809 |
+
"grad_norm": 0.49207037687301636,
|
| 3810 |
+
"learning_rate": 0.00013780470688017562,
|
| 3811 |
+
"loss": 0.8905,
|
| 3812 |
+
"step": 541
|
| 3813 |
+
},
|
| 3814 |
+
{
|
| 3815 |
+
"epoch": 0.2140177690029615,
|
| 3816 |
+
"grad_norm": 0.5545893311500549,
|
| 3817 |
+
"learning_rate": 0.00013759834061194794,
|
| 3818 |
+
"loss": 0.808,
|
| 3819 |
+
"step": 542
|
| 3820 |
+
},
|
| 3821 |
+
{
|
| 3822 |
+
"epoch": 0.21441263573543928,
|
| 3823 |
+
"grad_norm": 0.657805323600769,
|
| 3824 |
+
"learning_rate": 0.00013739178769285032,
|
| 3825 |
+
"loss": 0.8566,
|
| 3826 |
+
"step": 543
|
| 3827 |
+
},
|
| 3828 |
+
{
|
| 3829 |
+
"epoch": 0.21480750246791708,
|
| 3830 |
+
"grad_norm": 0.5868344902992249,
|
| 3831 |
+
"learning_rate": 0.00013718504914828135,
|
| 3832 |
+
"loss": 0.9001,
|
| 3833 |
+
"step": 544
|
| 3834 |
+
},
|
| 3835 |
+
{
|
| 3836 |
+
"epoch": 0.21520236920039487,
|
| 3837 |
+
"grad_norm": 0.4816092550754547,
|
| 3838 |
+
"learning_rate": 0.00013697812600456093,
|
| 3839 |
+
"loss": 0.993,
|
| 3840 |
+
"step": 545
|
| 3841 |
+
},
|
| 3842 |
+
{
|
| 3843 |
+
"epoch": 0.21559723593287267,
|
| 3844 |
+
"grad_norm": 0.5755246877670288,
|
| 3845 |
+
"learning_rate": 0.00013677101928892554,
|
| 3846 |
+
"loss": 1.1376,
|
| 3847 |
+
"step": 546
|
| 3848 |
+
},
|
| 3849 |
+
{
|
| 3850 |
+
"epoch": 0.21599210266535043,
|
| 3851 |
+
"grad_norm": 0.5235198736190796,
|
| 3852 |
+
"learning_rate": 0.0001365637300295229,
|
| 3853 |
+
"loss": 1.036,
|
| 3854 |
+
"step": 547
|
| 3855 |
+
},
|
| 3856 |
+
{
|
| 3857 |
+
"epoch": 0.21638696939782823,
|
| 3858 |
+
"grad_norm": 0.4904315173625946,
|
| 3859 |
+
"learning_rate": 0.00013635625925540696,
|
| 3860 |
+
"loss": 1.0033,
|
| 3861 |
+
"step": 548
|
| 3862 |
+
},
|
| 3863 |
+
{
|
| 3864 |
+
"epoch": 0.21678183613030602,
|
| 3865 |
+
"grad_norm": 0.49426376819610596,
|
| 3866 |
+
"learning_rate": 0.00013614860799653276,
|
| 3867 |
+
"loss": 1.0455,
|
| 3868 |
+
"step": 549
|
| 3869 |
+
},
|
| 3870 |
+
{
|
| 3871 |
+
"epoch": 0.21717670286278382,
|
| 3872 |
+
"grad_norm": 0.5230404734611511,
|
| 3873 |
+
"learning_rate": 0.00013594077728375128,
|
| 3874 |
+
"loss": 0.8619,
|
| 3875 |
+
"step": 550
|
| 3876 |
+
},
|
| 3877 |
+
{
|
| 3878 |
+
"epoch": 0.2175715695952616,
|
| 3879 |
+
"grad_norm": 0.4252125918865204,
|
| 3880 |
+
"learning_rate": 0.0001357327681488045,
|
| 3881 |
+
"loss": 0.552,
|
| 3882 |
+
"step": 551
|
| 3883 |
+
},
|
| 3884 |
+
{
|
| 3885 |
+
"epoch": 0.21796643632773938,
|
| 3886 |
+
"grad_norm": 0.6176772117614746,
|
| 3887 |
+
"learning_rate": 0.00013552458162432003,
|
| 3888 |
+
"loss": 0.9374,
|
| 3889 |
+
"step": 552
|
| 3890 |
+
},
|
| 3891 |
+
{
|
| 3892 |
+
"epoch": 0.21836130306021717,
|
| 3893 |
+
"grad_norm": 0.476182222366333,
|
| 3894 |
+
"learning_rate": 0.00013531621874380613,
|
| 3895 |
+
"loss": 0.9189,
|
| 3896 |
+
"step": 553
|
| 3897 |
+
},
|
| 3898 |
+
{
|
| 3899 |
+
"epoch": 0.21875616979269497,
|
| 3900 |
+
"grad_norm": 0.5183379054069519,
|
| 3901 |
+
"learning_rate": 0.00013510768054164653,
|
| 3902 |
+
"loss": 0.9177,
|
| 3903 |
+
"step": 554
|
| 3904 |
+
},
|
| 3905 |
+
{
|
| 3906 |
+
"epoch": 0.21915103652517276,
|
| 3907 |
+
"grad_norm": 0.5251573920249939,
|
| 3908 |
+
"learning_rate": 0.00013489896805309542,
|
| 3909 |
+
"loss": 0.8619,
|
| 3910 |
+
"step": 555
|
| 3911 |
+
},
|
| 3912 |
+
{
|
| 3913 |
+
"epoch": 0.21954590325765055,
|
| 3914 |
+
"grad_norm": 0.508030116558075,
|
| 3915 |
+
"learning_rate": 0.00013469008231427207,
|
| 3916 |
+
"loss": 1.032,
|
| 3917 |
+
"step": 556
|
| 3918 |
+
},
|
| 3919 |
+
{
|
| 3920 |
+
"epoch": 0.21994076999012832,
|
| 3921 |
+
"grad_norm": 0.5486171245574951,
|
| 3922 |
+
"learning_rate": 0.00013448102436215592,
|
| 3923 |
+
"loss": 0.8481,
|
| 3924 |
+
"step": 557
|
| 3925 |
+
},
|
| 3926 |
+
{
|
| 3927 |
+
"epoch": 0.22033563672260612,
|
| 3928 |
+
"grad_norm": 0.5298491716384888,
|
| 3929 |
+
"learning_rate": 0.00013427179523458127,
|
| 3930 |
+
"loss": 0.7748,
|
| 3931 |
+
"step": 558
|
| 3932 |
+
},
|
| 3933 |
+
{
|
| 3934 |
+
"epoch": 0.2207305034550839,
|
| 3935 |
+
"grad_norm": 0.6214480996131897,
|
| 3936 |
+
"learning_rate": 0.00013406239597023225,
|
| 3937 |
+
"loss": 1.0314,
|
| 3938 |
+
"step": 559
|
| 3939 |
+
},
|
| 3940 |
+
{
|
| 3941 |
+
"epoch": 0.2211253701875617,
|
| 3942 |
+
"grad_norm": 0.5228508710861206,
|
| 3943 |
+
"learning_rate": 0.00013385282760863758,
|
| 3944 |
+
"loss": 0.9916,
|
| 3945 |
+
"step": 560
|
| 3946 |
+
},
|
| 3947 |
+
{
|
| 3948 |
+
"epoch": 0.2215202369200395,
|
| 3949 |
+
"grad_norm": 0.5693901777267456,
|
| 3950 |
+
"learning_rate": 0.00013364309119016538,
|
| 3951 |
+
"loss": 1.0026,
|
| 3952 |
+
"step": 561
|
| 3953 |
+
},
|
| 3954 |
+
{
|
| 3955 |
+
"epoch": 0.22191510365251726,
|
| 3956 |
+
"grad_norm": 0.4827022850513458,
|
| 3957 |
+
"learning_rate": 0.0001334331877560182,
|
| 3958 |
+
"loss": 0.8068,
|
| 3959 |
+
"step": 562
|
| 3960 |
+
},
|
| 3961 |
+
{
|
| 3962 |
+
"epoch": 0.22230997038499506,
|
| 3963 |
+
"grad_norm": 0.4529504179954529,
|
| 3964 |
+
"learning_rate": 0.00013322311834822756,
|
| 3965 |
+
"loss": 0.8318,
|
| 3966 |
+
"step": 563
|
| 3967 |
+
},
|
| 3968 |
+
{
|
| 3969 |
+
"epoch": 0.22270483711747285,
|
| 3970 |
+
"grad_norm": 0.5850614309310913,
|
| 3971 |
+
"learning_rate": 0.00013301288400964902,
|
| 3972 |
+
"loss": 0.8946,
|
| 3973 |
+
"step": 564
|
| 3974 |
+
},
|
| 3975 |
+
{
|
| 3976 |
+
"epoch": 0.22309970384995065,
|
| 3977 |
+
"grad_norm": 0.5779694318771362,
|
| 3978 |
+
"learning_rate": 0.0001328024857839569,
|
| 3979 |
+
"loss": 0.7861,
|
| 3980 |
+
"step": 565
|
| 3981 |
+
},
|
| 3982 |
+
{
|
| 3983 |
+
"epoch": 0.22349457058242844,
|
| 3984 |
+
"grad_norm": 0.5600647926330566,
|
| 3985 |
+
"learning_rate": 0.00013259192471563912,
|
| 3986 |
+
"loss": 0.7921,
|
| 3987 |
+
"step": 566
|
| 3988 |
+
},
|
| 3989 |
+
{
|
| 3990 |
+
"epoch": 0.2238894373149062,
|
| 3991 |
+
"grad_norm": 0.5468648076057434,
|
| 3992 |
+
"learning_rate": 0.00013238120184999195,
|
| 3993 |
+
"loss": 0.7732,
|
| 3994 |
+
"step": 567
|
| 3995 |
+
},
|
| 3996 |
+
{
|
| 3997 |
+
"epoch": 0.224284304047384,
|
| 3998 |
+
"grad_norm": 0.619243323802948,
|
| 3999 |
+
"learning_rate": 0.00013217031823311488,
|
| 4000 |
+
"loss": 1.2012,
|
| 4001 |
+
"step": 568
|
| 4002 |
+
},
|
| 4003 |
+
{
|
| 4004 |
+
"epoch": 0.2246791707798618,
|
| 4005 |
+
"grad_norm": 0.5478827953338623,
|
| 4006 |
+
"learning_rate": 0.00013195927491190554,
|
| 4007 |
+
"loss": 1.0279,
|
| 4008 |
+
"step": 569
|
| 4009 |
+
},
|
| 4010 |
+
{
|
| 4011 |
+
"epoch": 0.2250740375123396,
|
| 4012 |
+
"grad_norm": 0.520038366317749,
|
| 4013 |
+
"learning_rate": 0.00013174807293405428,
|
| 4014 |
+
"loss": 0.8806,
|
| 4015 |
+
"step": 570
|
| 4016 |
+
},
|
| 4017 |
+
{
|
| 4018 |
+
"epoch": 0.2254689042448174,
|
| 4019 |
+
"grad_norm": 0.46674588322639465,
|
| 4020 |
+
"learning_rate": 0.00013153671334803905,
|
| 4021 |
+
"loss": 0.7596,
|
| 4022 |
+
"step": 571
|
| 4023 |
+
},
|
| 4024 |
+
{
|
| 4025 |
+
"epoch": 0.22586377097729515,
|
| 4026 |
+
"grad_norm": 0.510772705078125,
|
| 4027 |
+
"learning_rate": 0.0001313251972031203,
|
| 4028 |
+
"loss": 0.873,
|
| 4029 |
+
"step": 572
|
| 4030 |
+
},
|
| 4031 |
+
{
|
| 4032 |
+
"epoch": 0.22625863770977295,
|
| 4033 |
+
"grad_norm": 0.4360644221305847,
|
| 4034 |
+
"learning_rate": 0.00013111352554933563,
|
| 4035 |
+
"loss": 0.9622,
|
| 4036 |
+
"step": 573
|
| 4037 |
+
},
|
| 4038 |
+
{
|
| 4039 |
+
"epoch": 0.22665350444225074,
|
| 4040 |
+
"grad_norm": 0.4816182255744934,
|
| 4041 |
+
"learning_rate": 0.00013090169943749476,
|
| 4042 |
+
"loss": 1.0487,
|
| 4043 |
+
"step": 574
|
| 4044 |
+
},
|
| 4045 |
+
{
|
| 4046 |
+
"epoch": 0.22704837117472854,
|
| 4047 |
+
"grad_norm": 0.48583704233169556,
|
| 4048 |
+
"learning_rate": 0.000130689719919174,
|
| 4049 |
+
"loss": 0.8041,
|
| 4050 |
+
"step": 575
|
| 4051 |
+
},
|
| 4052 |
+
{
|
| 4053 |
+
"epoch": 0.22744323790720633,
|
| 4054 |
+
"grad_norm": 0.5958275198936462,
|
| 4055 |
+
"learning_rate": 0.00013047758804671136,
|
| 4056 |
+
"loss": 0.9466,
|
| 4057 |
+
"step": 576
|
| 4058 |
+
},
|
| 4059 |
+
{
|
| 4060 |
+
"epoch": 0.2278381046396841,
|
| 4061 |
+
"grad_norm": 0.4736819863319397,
|
| 4062 |
+
"learning_rate": 0.00013026530487320113,
|
| 4063 |
+
"loss": 0.9483,
|
| 4064 |
+
"step": 577
|
| 4065 |
+
},
|
| 4066 |
+
{
|
| 4067 |
+
"epoch": 0.2282329713721619,
|
| 4068 |
+
"grad_norm": 0.5173781514167786,
|
| 4069 |
+
"learning_rate": 0.00013005287145248878,
|
| 4070 |
+
"loss": 0.8428,
|
| 4071 |
+
"step": 578
|
| 4072 |
+
},
|
| 4073 |
+
{
|
| 4074 |
+
"epoch": 0.22862783810463969,
|
| 4075 |
+
"grad_norm": 0.44686052203178406,
|
| 4076 |
+
"learning_rate": 0.00012984028883916552,
|
| 4077 |
+
"loss": 1.009,
|
| 4078 |
+
"step": 579
|
| 4079 |
+
},
|
| 4080 |
+
{
|
| 4081 |
+
"epoch": 0.22902270483711748,
|
| 4082 |
+
"grad_norm": 0.5016387701034546,
|
| 4083 |
+
"learning_rate": 0.00012962755808856342,
|
| 4084 |
+
"loss": 1.0226,
|
| 4085 |
+
"step": 580
|
| 4086 |
+
},
|
| 4087 |
+
{
|
| 4088 |
+
"epoch": 0.22941757156959527,
|
| 4089 |
+
"grad_norm": 0.6307184100151062,
|
| 4090 |
+
"learning_rate": 0.0001294146802567497,
|
| 4091 |
+
"loss": 0.9179,
|
| 4092 |
+
"step": 581
|
| 4093 |
+
},
|
| 4094 |
+
{
|
| 4095 |
+
"epoch": 0.22981243830207304,
|
| 4096 |
+
"grad_norm": 0.43451613187789917,
|
| 4097 |
+
"learning_rate": 0.0001292016564005219,
|
| 4098 |
+
"loss": 0.9468,
|
| 4099 |
+
"step": 582
|
| 4100 |
+
},
|
| 4101 |
+
{
|
| 4102 |
+
"epoch": 0.23020730503455084,
|
| 4103 |
+
"grad_norm": 0.5214070081710815,
|
| 4104 |
+
"learning_rate": 0.00012898848757740246,
|
| 4105 |
+
"loss": 0.9226,
|
| 4106 |
+
"step": 583
|
| 4107 |
+
},
|
| 4108 |
+
{
|
| 4109 |
+
"epoch": 0.23060217176702863,
|
| 4110 |
+
"grad_norm": 0.6036335825920105,
|
| 4111 |
+
"learning_rate": 0.00012877517484563344,
|
| 4112 |
+
"loss": 0.9585,
|
| 4113 |
+
"step": 584
|
| 4114 |
+
},
|
| 4115 |
+
{
|
| 4116 |
+
"epoch": 0.23099703849950642,
|
| 4117 |
+
"grad_norm": 0.5829451084136963,
|
| 4118 |
+
"learning_rate": 0.00012856171926417133,
|
| 4119 |
+
"loss": 1.2637,
|
| 4120 |
+
"step": 585
|
| 4121 |
+
},
|
| 4122 |
+
{
|
| 4123 |
+
"epoch": 0.23139190523198422,
|
| 4124 |
+
"grad_norm": 0.5726488828659058,
|
| 4125 |
+
"learning_rate": 0.0001283481218926818,
|
| 4126 |
+
"loss": 0.9223,
|
| 4127 |
+
"step": 586
|
| 4128 |
+
},
|
| 4129 |
+
{
|
| 4130 |
+
"epoch": 0.23178677196446199,
|
| 4131 |
+
"grad_norm": 0.501338005065918,
|
| 4132 |
+
"learning_rate": 0.0001281343837915344,
|
| 4133 |
+
"loss": 0.8892,
|
| 4134 |
+
"step": 587
|
| 4135 |
+
},
|
| 4136 |
+
{
|
| 4137 |
+
"epoch": 0.23218163869693978,
|
| 4138 |
+
"grad_norm": 0.5073863863945007,
|
| 4139 |
+
"learning_rate": 0.00012792050602179725,
|
| 4140 |
+
"loss": 0.8541,
|
| 4141 |
+
"step": 588
|
| 4142 |
+
},
|
| 4143 |
+
{
|
| 4144 |
+
"epoch": 0.23257650542941757,
|
| 4145 |
+
"grad_norm": 0.6474023461341858,
|
| 4146 |
+
"learning_rate": 0.00012770648964523194,
|
| 4147 |
+
"loss": 1.1276,
|
| 4148 |
+
"step": 589
|
| 4149 |
+
},
|
| 4150 |
+
{
|
| 4151 |
+
"epoch": 0.23297137216189537,
|
| 4152 |
+
"grad_norm": 0.48500892519950867,
|
| 4153 |
+
"learning_rate": 0.00012749233572428804,
|
| 4154 |
+
"loss": 1.0324,
|
| 4155 |
+
"step": 590
|
| 4156 |
+
},
|
| 4157 |
+
{
|
| 4158 |
+
"epoch": 0.23336623889437316,
|
| 4159 |
+
"grad_norm": 0.5808454155921936,
|
| 4160 |
+
"learning_rate": 0.00012727804532209803,
|
| 4161 |
+
"loss": 1.0817,
|
| 4162 |
+
"step": 591
|
| 4163 |
+
},
|
| 4164 |
+
{
|
| 4165 |
+
"epoch": 0.23376110562685093,
|
| 4166 |
+
"grad_norm": 0.49606916308403015,
|
| 4167 |
+
"learning_rate": 0.0001270636195024719,
|
| 4168 |
+
"loss": 0.9889,
|
| 4169 |
+
"step": 592
|
| 4170 |
+
},
|
| 4171 |
+
{
|
| 4172 |
+
"epoch": 0.23415597235932872,
|
| 4173 |
+
"grad_norm": 0.6760202050209045,
|
| 4174 |
+
"learning_rate": 0.00012684905932989186,
|
| 4175 |
+
"loss": 1.1171,
|
| 4176 |
+
"step": 593
|
| 4177 |
+
},
|
| 4178 |
+
{
|
| 4179 |
+
"epoch": 0.23455083909180652,
|
| 4180 |
+
"grad_norm": 0.44945040345191956,
|
| 4181 |
+
"learning_rate": 0.00012663436586950714,
|
| 4182 |
+
"loss": 0.9308,
|
| 4183 |
+
"step": 594
|
| 4184 |
+
},
|
| 4185 |
+
{
|
| 4186 |
+
"epoch": 0.2349457058242843,
|
| 4187 |
+
"grad_norm": 0.6141867637634277,
|
| 4188 |
+
"learning_rate": 0.00012641954018712863,
|
| 4189 |
+
"loss": 1.1256,
|
| 4190 |
+
"step": 595
|
| 4191 |
+
},
|
| 4192 |
+
{
|
| 4193 |
+
"epoch": 0.2353405725567621,
|
| 4194 |
+
"grad_norm": 0.5368825197219849,
|
| 4195 |
+
"learning_rate": 0.0001262045833492236,
|
| 4196 |
+
"loss": 1.0181,
|
| 4197 |
+
"step": 596
|
| 4198 |
+
},
|
| 4199 |
+
{
|
| 4200 |
+
"epoch": 0.23573543928923987,
|
| 4201 |
+
"grad_norm": 0.5402265787124634,
|
| 4202 |
+
"learning_rate": 0.00012598949642291047,
|
| 4203 |
+
"loss": 1.2135,
|
| 4204 |
+
"step": 597
|
| 4205 |
+
},
|
| 4206 |
+
{
|
| 4207 |
+
"epoch": 0.23613030602171767,
|
| 4208 |
+
"grad_norm": 0.5296156406402588,
|
| 4209 |
+
"learning_rate": 0.00012577428047595344,
|
| 4210 |
+
"loss": 0.8361,
|
| 4211 |
+
"step": 598
|
| 4212 |
+
},
|
| 4213 |
+
{
|
| 4214 |
+
"epoch": 0.23652517275419546,
|
| 4215 |
+
"grad_norm": 0.4741549491882324,
|
| 4216 |
+
"learning_rate": 0.00012555893657675718,
|
| 4217 |
+
"loss": 0.9022,
|
| 4218 |
+
"step": 599
|
| 4219 |
+
},
|
| 4220 |
+
{
|
| 4221 |
+
"epoch": 0.23692003948667326,
|
| 4222 |
+
"grad_norm": 0.4711611866950989,
|
| 4223 |
+
"learning_rate": 0.0001253434657943616,
|
| 4224 |
+
"loss": 1.0165,
|
| 4225 |
+
"step": 600
|
| 4226 |
+
},
|
| 4227 |
+
{
|
| 4228 |
+
"epoch": 0.23731490621915102,
|
| 4229 |
+
"grad_norm": 0.4527783989906311,
|
| 4230 |
+
"learning_rate": 0.00012512786919843648,
|
| 4231 |
+
"loss": 0.8954,
|
| 4232 |
+
"step": 601
|
| 4233 |
+
},
|
| 4234 |
+
{
|
| 4235 |
+
"epoch": 0.23770977295162882,
|
| 4236 |
+
"grad_norm": 0.5278099775314331,
|
| 4237 |
+
"learning_rate": 0.0001249121478592762,
|
| 4238 |
+
"loss": 1.0807,
|
| 4239 |
+
"step": 602
|
| 4240 |
+
},
|
| 4241 |
+
{
|
| 4242 |
+
"epoch": 0.2381046396841066,
|
| 4243 |
+
"grad_norm": 0.45427364110946655,
|
| 4244 |
+
"learning_rate": 0.00012469630284779438,
|
| 4245 |
+
"loss": 0.9407,
|
| 4246 |
+
"step": 603
|
| 4247 |
+
},
|
| 4248 |
+
{
|
| 4249 |
+
"epoch": 0.2384995064165844,
|
| 4250 |
+
"grad_norm": 0.5772978663444519,
|
| 4251 |
+
"learning_rate": 0.00012448033523551865,
|
| 4252 |
+
"loss": 1.2734,
|
| 4253 |
+
"step": 604
|
| 4254 |
+
},
|
| 4255 |
+
{
|
| 4256 |
+
"epoch": 0.2388943731490622,
|
| 4257 |
+
"grad_norm": 0.5643810033798218,
|
| 4258 |
+
"learning_rate": 0.00012426424609458518,
|
| 4259 |
+
"loss": 1.0125,
|
| 4260 |
+
"step": 605
|
| 4261 |
+
},
|
| 4262 |
+
{
|
| 4263 |
+
"epoch": 0.23928923988153997,
|
| 4264 |
+
"grad_norm": 0.5393880009651184,
|
| 4265 |
+
"learning_rate": 0.0001240480364977335,
|
| 4266 |
+
"loss": 0.8613,
|
| 4267 |
+
"step": 606
|
| 4268 |
+
},
|
| 4269 |
+
{
|
| 4270 |
+
"epoch": 0.23968410661401776,
|
| 4271 |
+
"grad_norm": 0.49924036860466003,
|
| 4272 |
+
"learning_rate": 0.0001238317075183011,
|
| 4273 |
+
"loss": 0.9862,
|
| 4274 |
+
"step": 607
|
| 4275 |
+
},
|
| 4276 |
+
{
|
| 4277 |
+
"epoch": 0.24007897334649556,
|
| 4278 |
+
"grad_norm": 0.4382825493812561,
|
| 4279 |
+
"learning_rate": 0.00012361526023021822,
|
| 4280 |
+
"loss": 0.8692,
|
| 4281 |
+
"step": 608
|
| 4282 |
+
},
|
| 4283 |
+
{
|
| 4284 |
+
"epoch": 0.24047384007897335,
|
| 4285 |
+
"grad_norm": 0.5655290484428406,
|
| 4286 |
+
"learning_rate": 0.00012339869570800232,
|
| 4287 |
+
"loss": 0.9122,
|
| 4288 |
+
"step": 609
|
| 4289 |
+
},
|
| 4290 |
+
{
|
| 4291 |
+
"epoch": 0.24086870681145114,
|
| 4292 |
+
"grad_norm": 0.5245158076286316,
|
| 4293 |
+
"learning_rate": 0.00012318201502675285,
|
| 4294 |
+
"loss": 0.8347,
|
| 4295 |
+
"step": 610
|
| 4296 |
+
},
|
| 4297 |
+
{
|
| 4298 |
+
"epoch": 0.2412635735439289,
|
| 4299 |
+
"grad_norm": 0.5116402506828308,
|
| 4300 |
+
"learning_rate": 0.00012296521926214596,
|
| 4301 |
+
"loss": 1.0317,
|
| 4302 |
+
"step": 611
|
| 4303 |
+
},
|
| 4304 |
+
{
|
| 4305 |
+
"epoch": 0.2416584402764067,
|
| 4306 |
+
"grad_norm": 0.4878910779953003,
|
| 4307 |
+
"learning_rate": 0.00012274830949042908,
|
| 4308 |
+
"loss": 0.8947,
|
| 4309 |
+
"step": 612
|
| 4310 |
+
},
|
| 4311 |
+
{
|
| 4312 |
+
"epoch": 0.2420533070088845,
|
| 4313 |
+
"grad_norm": 0.5890231132507324,
|
| 4314 |
+
"learning_rate": 0.00012253128678841568,
|
| 4315 |
+
"loss": 1.0683,
|
| 4316 |
+
"step": 613
|
| 4317 |
+
},
|
| 4318 |
+
{
|
| 4319 |
+
"epoch": 0.2424481737413623,
|
| 4320 |
+
"grad_norm": 0.4757651388645172,
|
| 4321 |
+
"learning_rate": 0.00012231415223347972,
|
| 4322 |
+
"loss": 0.9816,
|
| 4323 |
+
"step": 614
|
| 4324 |
+
},
|
| 4325 |
+
{
|
| 4326 |
+
"epoch": 0.2428430404738401,
|
| 4327 |
+
"grad_norm": 0.5415259599685669,
|
| 4328 |
+
"learning_rate": 0.0001220969069035506,
|
| 4329 |
+
"loss": 0.9583,
|
| 4330 |
+
"step": 615
|
| 4331 |
+
},
|
| 4332 |
+
{
|
| 4333 |
+
"epoch": 0.24323790720631785,
|
| 4334 |
+
"grad_norm": 0.5389688014984131,
|
| 4335 |
+
"learning_rate": 0.0001218795518771075,
|
| 4336 |
+
"loss": 0.7091,
|
| 4337 |
+
"step": 616
|
| 4338 |
+
},
|
| 4339 |
+
{
|
| 4340 |
+
"epoch": 0.24363277393879565,
|
| 4341 |
+
"grad_norm": 0.8125389814376831,
|
| 4342 |
+
"learning_rate": 0.00012166208823317427,
|
| 4343 |
+
"loss": 1.0042,
|
| 4344 |
+
"step": 617
|
| 4345 |
+
},
|
| 4346 |
+
{
|
| 4347 |
+
"epoch": 0.24402764067127344,
|
| 4348 |
+
"grad_norm": 0.5463893413543701,
|
| 4349 |
+
"learning_rate": 0.0001214445170513139,
|
| 4350 |
+
"loss": 1.0685,
|
| 4351 |
+
"step": 618
|
| 4352 |
+
},
|
| 4353 |
+
{
|
| 4354 |
+
"epoch": 0.24442250740375124,
|
| 4355 |
+
"grad_norm": 0.4813181161880493,
|
| 4356 |
+
"learning_rate": 0.0001212268394116233,
|
| 4357 |
+
"loss": 0.8353,
|
| 4358 |
+
"step": 619
|
| 4359 |
+
},
|
| 4360 |
+
{
|
| 4361 |
+
"epoch": 0.24481737413622903,
|
| 4362 |
+
"grad_norm": 0.4448351263999939,
|
| 4363 |
+
"learning_rate": 0.00012100905639472779,
|
| 4364 |
+
"loss": 1.0261,
|
| 4365 |
+
"step": 620
|
| 4366 |
+
},
|
| 4367 |
+
{
|
| 4368 |
+
"epoch": 0.2452122408687068,
|
| 4369 |
+
"grad_norm": 0.509598433971405,
|
| 4370 |
+
"learning_rate": 0.00012079116908177593,
|
| 4371 |
+
"loss": 1.0274,
|
| 4372 |
+
"step": 621
|
| 4373 |
+
},
|
| 4374 |
+
{
|
| 4375 |
+
"epoch": 0.2456071076011846,
|
| 4376 |
+
"grad_norm": 0.5108718276023865,
|
| 4377 |
+
"learning_rate": 0.00012057317855443395,
|
| 4378 |
+
"loss": 0.8464,
|
| 4379 |
+
"step": 622
|
| 4380 |
+
},
|
| 4381 |
+
{
|
| 4382 |
+
"epoch": 0.2460019743336624,
|
| 4383 |
+
"grad_norm": 0.4307985007762909,
|
| 4384 |
+
"learning_rate": 0.00012035508589488053,
|
| 4385 |
+
"loss": 0.9281,
|
| 4386 |
+
"step": 623
|
| 4387 |
+
},
|
| 4388 |
+
{
|
| 4389 |
+
"epoch": 0.24639684106614018,
|
| 4390 |
+
"grad_norm": 0.5565152764320374,
|
| 4391 |
+
"learning_rate": 0.00012013689218580132,
|
| 4392 |
+
"loss": 0.9916,
|
| 4393 |
+
"step": 624
|
| 4394 |
+
},
|
| 4395 |
+
{
|
| 4396 |
+
"epoch": 0.24679170779861798,
|
| 4397 |
+
"grad_norm": 0.7218130826950073,
|
| 4398 |
+
"learning_rate": 0.0001199185985103836,
|
| 4399 |
+
"loss": 1.0378,
|
| 4400 |
+
"step": 625
|
| 4401 |
+
},
|
| 4402 |
+
{
|
| 4403 |
+
"epoch": 0.24718657453109574,
|
| 4404 |
+
"grad_norm": 0.5790725350379944,
|
| 4405 |
+
"learning_rate": 0.00011970020595231101,
|
| 4406 |
+
"loss": 1.0886,
|
| 4407 |
+
"step": 626
|
| 4408 |
+
},
|
| 4409 |
+
{
|
| 4410 |
+
"epoch": 0.24758144126357354,
|
| 4411 |
+
"grad_norm": 0.5239992141723633,
|
| 4412 |
+
"learning_rate": 0.000119481715595758,
|
| 4413 |
+
"loss": 1.0763,
|
| 4414 |
+
"step": 627
|
| 4415 |
+
},
|
| 4416 |
+
{
|
| 4417 |
+
"epoch": 0.24797630799605133,
|
| 4418 |
+
"grad_norm": 0.5147425532341003,
|
| 4419 |
+
"learning_rate": 0.00011926312852538455,
|
| 4420 |
+
"loss": 1.1421,
|
| 4421 |
+
"step": 628
|
| 4422 |
+
},
|
| 4423 |
+
{
|
| 4424 |
+
"epoch": 0.24837117472852913,
|
| 4425 |
+
"grad_norm": 0.6474115252494812,
|
| 4426 |
+
"learning_rate": 0.0001190444458263307,
|
| 4427 |
+
"loss": 0.8813,
|
| 4428 |
+
"step": 629
|
| 4429 |
+
},
|
| 4430 |
+
{
|
| 4431 |
+
"epoch": 0.24876604146100692,
|
| 4432 |
+
"grad_norm": 0.46953749656677246,
|
| 4433 |
+
"learning_rate": 0.00011882566858421135,
|
| 4434 |
+
"loss": 0.6636,
|
| 4435 |
+
"step": 630
|
| 4436 |
+
},
|
| 4437 |
+
{
|
| 4438 |
+
"epoch": 0.2491609081934847,
|
| 4439 |
+
"grad_norm": 0.6197345852851868,
|
| 4440 |
+
"learning_rate": 0.00011860679788511064,
|
| 4441 |
+
"loss": 0.8935,
|
| 4442 |
+
"step": 631
|
| 4443 |
+
},
|
| 4444 |
+
{
|
| 4445 |
+
"epoch": 0.24955577492596248,
|
| 4446 |
+
"grad_norm": 0.5545244812965393,
|
| 4447 |
+
"learning_rate": 0.00011838783481557664,
|
| 4448 |
+
"loss": 0.7358,
|
| 4449 |
+
"step": 632
|
| 4450 |
+
},
|
| 4451 |
+
{
|
| 4452 |
+
"epoch": 0.24995064165844028,
|
| 4453 |
+
"grad_norm": 0.5150088667869568,
|
| 4454 |
+
"learning_rate": 0.00011816878046261615,
|
| 4455 |
+
"loss": 0.8935,
|
| 4456 |
+
"step": 633
|
| 4457 |
+
},
|
| 4458 |
+
{
|
| 4459 |
+
"epoch": 0.25034550839091807,
|
| 4460 |
+
"grad_norm": 0.5777970552444458,
|
| 4461 |
+
"learning_rate": 0.00011794963591368893,
|
| 4462 |
+
"loss": 0.9984,
|
| 4463 |
+
"step": 634
|
| 4464 |
+
},
|
| 4465 |
+
{
|
| 4466 |
+
"epoch": 0.25074037512339586,
|
| 4467 |
+
"grad_norm": 0.6271523237228394,
|
| 4468 |
+
"learning_rate": 0.00011773040225670256,
|
| 4469 |
+
"loss": 1.0425,
|
| 4470 |
+
"step": 635
|
| 4471 |
+
},
|
| 4472 |
+
{
|
| 4473 |
+
"epoch": 0.25113524185587366,
|
| 4474 |
+
"grad_norm": 0.5507854223251343,
|
| 4475 |
+
"learning_rate": 0.00011751108058000706,
|
| 4476 |
+
"loss": 1.017,
|
| 4477 |
+
"step": 636
|
| 4478 |
+
},
|
| 4479 |
+
{
|
| 4480 |
+
"epoch": 0.25153010858835145,
|
| 4481 |
+
"grad_norm": 0.5012635588645935,
|
| 4482 |
+
"learning_rate": 0.00011729167197238935,
|
| 4483 |
+
"loss": 1.0421,
|
| 4484 |
+
"step": 637
|
| 4485 |
+
},
|
| 4486 |
+
{
|
| 4487 |
+
"epoch": 0.25192497532082925,
|
| 4488 |
+
"grad_norm": 0.5927073955535889,
|
| 4489 |
+
"learning_rate": 0.0001170721775230679,
|
| 4490 |
+
"loss": 0.8634,
|
| 4491 |
+
"step": 638
|
| 4492 |
+
},
|
| 4493 |
+
{
|
| 4494 |
+
"epoch": 0.252319842053307,
|
| 4495 |
+
"grad_norm": 0.5467817783355713,
|
| 4496 |
+
"learning_rate": 0.0001168525983216873,
|
| 4497 |
+
"loss": 0.828,
|
| 4498 |
+
"step": 639
|
| 4499 |
+
},
|
| 4500 |
+
{
|
| 4501 |
+
"epoch": 0.2527147087857848,
|
| 4502 |
+
"grad_norm": 0.5477814078330994,
|
| 4503 |
+
"learning_rate": 0.00011663293545831302,
|
| 4504 |
+
"loss": 0.8878,
|
| 4505 |
+
"step": 640
|
| 4506 |
+
},
|
| 4507 |
+
{
|
| 4508 |
+
"epoch": 0.2531095755182626,
|
| 4509 |
+
"grad_norm": 0.5831199884414673,
|
| 4510 |
+
"learning_rate": 0.00011641319002342568,
|
| 4511 |
+
"loss": 0.9146,
|
| 4512 |
+
"step": 641
|
| 4513 |
+
},
|
| 4514 |
+
{
|
| 4515 |
+
"epoch": 0.25350444225074037,
|
| 4516 |
+
"grad_norm": 0.4861442446708679,
|
| 4517 |
+
"learning_rate": 0.00011619336310791586,
|
| 4518 |
+
"loss": 0.9527,
|
| 4519 |
+
"step": 642
|
| 4520 |
+
},
|
| 4521 |
+
{
|
| 4522 |
+
"epoch": 0.25389930898321816,
|
| 4523 |
+
"grad_norm": 0.4937233030796051,
|
| 4524 |
+
"learning_rate": 0.00011597345580307875,
|
| 4525 |
+
"loss": 0.8631,
|
| 4526 |
+
"step": 643
|
| 4527 |
+
},
|
| 4528 |
+
{
|
| 4529 |
+
"epoch": 0.25429417571569596,
|
| 4530 |
+
"grad_norm": 0.5312657952308655,
|
| 4531 |
+
"learning_rate": 0.00011575346920060846,
|
| 4532 |
+
"loss": 1.0702,
|
| 4533 |
+
"step": 644
|
| 4534 |
+
},
|
| 4535 |
+
{
|
| 4536 |
+
"epoch": 0.25468904244817375,
|
| 4537 |
+
"grad_norm": 0.5473558306694031,
|
| 4538 |
+
"learning_rate": 0.00011553340439259286,
|
| 4539 |
+
"loss": 0.8867,
|
| 4540 |
+
"step": 645
|
| 4541 |
+
},
|
| 4542 |
+
{
|
| 4543 |
+
"epoch": 0.25508390918065155,
|
| 4544 |
+
"grad_norm": 0.5094785094261169,
|
| 4545 |
+
"learning_rate": 0.00011531326247150803,
|
| 4546 |
+
"loss": 0.9356,
|
| 4547 |
+
"step": 646
|
| 4548 |
+
},
|
| 4549 |
+
{
|
| 4550 |
+
"epoch": 0.25547877591312934,
|
| 4551 |
+
"grad_norm": 0.4709779620170593,
|
| 4552 |
+
"learning_rate": 0.00011509304453021288,
|
| 4553 |
+
"loss": 0.8547,
|
| 4554 |
+
"step": 647
|
| 4555 |
+
},
|
| 4556 |
+
{
|
| 4557 |
+
"epoch": 0.2558736426456071,
|
| 4558 |
+
"grad_norm": 0.5478555560112,
|
| 4559 |
+
"learning_rate": 0.00011487275166194367,
|
| 4560 |
+
"loss": 1.013,
|
| 4561 |
+
"step": 648
|
| 4562 |
+
},
|
| 4563 |
+
{
|
| 4564 |
+
"epoch": 0.2562685093780849,
|
| 4565 |
+
"grad_norm": 0.5569909811019897,
|
| 4566 |
+
"learning_rate": 0.00011465238496030868,
|
| 4567 |
+
"loss": 0.9655,
|
| 4568 |
+
"step": 649
|
| 4569 |
+
},
|
| 4570 |
+
{
|
| 4571 |
+
"epoch": 0.25666337611056267,
|
| 4572 |
+
"grad_norm": 0.6182945370674133,
|
| 4573 |
+
"learning_rate": 0.00011443194551928266,
|
| 4574 |
+
"loss": 1.0553,
|
| 4575 |
+
"step": 650
|
| 4576 |
+
},
|
| 4577 |
+
{
|
| 4578 |
+
"epoch": 0.25705824284304046,
|
| 4579 |
+
"grad_norm": 0.49789944291114807,
|
| 4580 |
+
"learning_rate": 0.00011421143443320155,
|
| 4581 |
+
"loss": 1.0104,
|
| 4582 |
+
"step": 651
|
| 4583 |
+
},
|
| 4584 |
+
{
|
| 4585 |
+
"epoch": 0.25745310957551826,
|
| 4586 |
+
"grad_norm": 0.5329926013946533,
|
| 4587 |
+
"learning_rate": 0.00011399085279675687,
|
| 4588 |
+
"loss": 1.1935,
|
| 4589 |
+
"step": 652
|
| 4590 |
+
},
|
| 4591 |
+
{
|
| 4592 |
+
"epoch": 0.25784797630799605,
|
| 4593 |
+
"grad_norm": 0.6216697692871094,
|
| 4594 |
+
"learning_rate": 0.0001137702017049904,
|
| 4595 |
+
"loss": 0.9502,
|
| 4596 |
+
"step": 653
|
| 4597 |
+
},
|
| 4598 |
+
{
|
| 4599 |
+
"epoch": 0.25824284304047385,
|
| 4600 |
+
"grad_norm": 0.46455976366996765,
|
| 4601 |
+
"learning_rate": 0.00011354948225328877,
|
| 4602 |
+
"loss": 1.0217,
|
| 4603 |
+
"step": 654
|
| 4604 |
+
},
|
| 4605 |
+
{
|
| 4606 |
+
"epoch": 0.25863770977295164,
|
| 4607 |
+
"grad_norm": 0.5468802452087402,
|
| 4608 |
+
"learning_rate": 0.0001133286955373779,
|
| 4609 |
+
"loss": 0.9669,
|
| 4610 |
+
"step": 655
|
| 4611 |
+
},
|
| 4612 |
+
{
|
| 4613 |
+
"epoch": 0.25903257650542943,
|
| 4614 |
+
"grad_norm": 0.6737267971038818,
|
| 4615 |
+
"learning_rate": 0.00011310784265331769,
|
| 4616 |
+
"loss": 0.828,
|
| 4617 |
+
"step": 656
|
| 4618 |
+
},
|
| 4619 |
+
{
|
| 4620 |
+
"epoch": 0.25942744323790723,
|
| 4621 |
+
"grad_norm": 0.5070087909698486,
|
| 4622 |
+
"learning_rate": 0.00011288692469749649,
|
| 4623 |
+
"loss": 0.9463,
|
| 4624 |
+
"step": 657
|
| 4625 |
+
},
|
| 4626 |
+
{
|
| 4627 |
+
"epoch": 0.25982230997038497,
|
| 4628 |
+
"grad_norm": 0.5430976152420044,
|
| 4629 |
+
"learning_rate": 0.0001126659427666257,
|
| 4630 |
+
"loss": 0.9986,
|
| 4631 |
+
"step": 658
|
| 4632 |
+
},
|
| 4633 |
+
{
|
| 4634 |
+
"epoch": 0.26021717670286276,
|
| 4635 |
+
"grad_norm": 0.4886017143726349,
|
| 4636 |
+
"learning_rate": 0.00011244489795773432,
|
| 4637 |
+
"loss": 1.0702,
|
| 4638 |
+
"step": 659
|
| 4639 |
+
},
|
| 4640 |
+
{
|
| 4641 |
+
"epoch": 0.26061204343534056,
|
| 4642 |
+
"grad_norm": 0.4727013111114502,
|
| 4643 |
+
"learning_rate": 0.00011222379136816345,
|
| 4644 |
+
"loss": 0.9179,
|
| 4645 |
+
"step": 660
|
| 4646 |
+
},
|
| 4647 |
+
{
|
| 4648 |
+
"epoch": 0.26100691016781835,
|
| 4649 |
+
"grad_norm": 0.48751479387283325,
|
| 4650 |
+
"learning_rate": 0.00011200262409556097,
|
| 4651 |
+
"loss": 0.9176,
|
| 4652 |
+
"step": 661
|
| 4653 |
+
},
|
| 4654 |
+
{
|
| 4655 |
+
"epoch": 0.26140177690029615,
|
| 4656 |
+
"grad_norm": 0.511191189289093,
|
| 4657 |
+
"learning_rate": 0.00011178139723787597,
|
| 4658 |
+
"loss": 1.0286,
|
| 4659 |
+
"step": 662
|
| 4660 |
+
},
|
| 4661 |
+
{
|
| 4662 |
+
"epoch": 0.26179664363277394,
|
| 4663 |
+
"grad_norm": 0.48295167088508606,
|
| 4664 |
+
"learning_rate": 0.00011156011189335332,
|
| 4665 |
+
"loss": 0.9306,
|
| 4666 |
+
"step": 663
|
| 4667 |
+
},
|
| 4668 |
+
{
|
| 4669 |
+
"epoch": 0.26219151036525173,
|
| 4670 |
+
"grad_norm": 0.4702792167663574,
|
| 4671 |
+
"learning_rate": 0.00011133876916052821,
|
| 4672 |
+
"loss": 0.9666,
|
| 4673 |
+
"step": 664
|
| 4674 |
+
},
|
| 4675 |
+
{
|
| 4676 |
+
"epoch": 0.26258637709772953,
|
| 4677 |
+
"grad_norm": 0.5391889810562134,
|
| 4678 |
+
"learning_rate": 0.00011111737013822088,
|
| 4679 |
+
"loss": 0.9043,
|
| 4680 |
+
"step": 665
|
| 4681 |
+
},
|
| 4682 |
+
{
|
| 4683 |
+
"epoch": 0.2629812438302073,
|
| 4684 |
+
"grad_norm": 0.6235263347625732,
|
| 4685 |
+
"learning_rate": 0.00011089591592553082,
|
| 4686 |
+
"loss": 1.0091,
|
| 4687 |
+
"step": 666
|
| 4688 |
+
},
|
| 4689 |
+
{
|
| 4690 |
+
"epoch": 0.2633761105626851,
|
| 4691 |
+
"grad_norm": 0.5476643443107605,
|
| 4692 |
+
"learning_rate": 0.00011067440762183164,
|
| 4693 |
+
"loss": 0.9399,
|
| 4694 |
+
"step": 667
|
| 4695 |
+
},
|
| 4696 |
+
{
|
| 4697 |
+
"epoch": 0.26377097729516286,
|
| 4698 |
+
"grad_norm": 0.42856502532958984,
|
| 4699 |
+
"learning_rate": 0.00011045284632676536,
|
| 4700 |
+
"loss": 0.8856,
|
| 4701 |
+
"step": 668
|
| 4702 |
+
},
|
| 4703 |
+
{
|
| 4704 |
+
"epoch": 0.26416584402764065,
|
| 4705 |
+
"grad_norm": 0.6507248878479004,
|
| 4706 |
+
"learning_rate": 0.00011023123314023717,
|
| 4707 |
+
"loss": 1.0047,
|
| 4708 |
+
"step": 669
|
| 4709 |
+
},
|
| 4710 |
+
{
|
| 4711 |
+
"epoch": 0.26456071076011844,
|
| 4712 |
+
"grad_norm": 0.5612763166427612,
|
| 4713 |
+
"learning_rate": 0.00011000956916240985,
|
| 4714 |
+
"loss": 0.9552,
|
| 4715 |
+
"step": 670
|
| 4716 |
+
},
|
| 4717 |
+
{
|
| 4718 |
+
"epoch": 0.26495557749259624,
|
| 4719 |
+
"grad_norm": 0.6156942844390869,
|
| 4720 |
+
"learning_rate": 0.00010978785549369823,
|
| 4721 |
+
"loss": 0.7834,
|
| 4722 |
+
"step": 671
|
| 4723 |
+
},
|
| 4724 |
+
{
|
| 4725 |
+
"epoch": 0.26535044422507403,
|
| 4726 |
+
"grad_norm": 0.5807977914810181,
|
| 4727 |
+
"learning_rate": 0.00010956609323476399,
|
| 4728 |
+
"loss": 1.0304,
|
| 4729 |
+
"step": 672
|
| 4730 |
+
},
|
| 4731 |
+
{
|
| 4732 |
+
"epoch": 0.26574531095755183,
|
| 4733 |
+
"grad_norm": 0.4886478781700134,
|
| 4734 |
+
"learning_rate": 0.00010934428348650986,
|
| 4735 |
+
"loss": 1.0725,
|
| 4736 |
+
"step": 673
|
| 4737 |
+
},
|
| 4738 |
+
{
|
| 4739 |
+
"epoch": 0.2661401776900296,
|
| 4740 |
+
"grad_norm": 0.6015396118164062,
|
| 4741 |
+
"learning_rate": 0.00010912242735007441,
|
| 4742 |
+
"loss": 0.9459,
|
| 4743 |
+
"step": 674
|
| 4744 |
+
},
|
| 4745 |
+
{
|
| 4746 |
+
"epoch": 0.2665350444225074,
|
| 4747 |
+
"grad_norm": 0.5361436605453491,
|
| 4748 |
+
"learning_rate": 0.0001089005259268265,
|
| 4749 |
+
"loss": 1.0034,
|
| 4750 |
+
"step": 675
|
| 4751 |
+
},
|
| 4752 |
+
{
|
| 4753 |
+
"epoch": 0.2669299111549852,
|
| 4754 |
+
"grad_norm": 0.5702396035194397,
|
| 4755 |
+
"learning_rate": 0.00010867858031835975,
|
| 4756 |
+
"loss": 1.1566,
|
| 4757 |
+
"step": 676
|
| 4758 |
+
},
|
| 4759 |
+
{
|
| 4760 |
+
"epoch": 0.267324777887463,
|
| 4761 |
+
"grad_norm": 0.5034629702568054,
|
| 4762 |
+
"learning_rate": 0.00010845659162648723,
|
| 4763 |
+
"loss": 0.9551,
|
| 4764 |
+
"step": 677
|
| 4765 |
+
},
|
| 4766 |
+
{
|
| 4767 |
+
"epoch": 0.26771964461994074,
|
| 4768 |
+
"grad_norm": 0.49666622281074524,
|
| 4769 |
+
"learning_rate": 0.00010823456095323579,
|
| 4770 |
+
"loss": 0.9059,
|
| 4771 |
+
"step": 678
|
| 4772 |
+
},
|
| 4773 |
+
{
|
| 4774 |
+
"epoch": 0.26811451135241854,
|
| 4775 |
+
"grad_norm": 0.5842538475990295,
|
| 4776 |
+
"learning_rate": 0.00010801248940084074,
|
| 4777 |
+
"loss": 0.8893,
|
| 4778 |
+
"step": 679
|
| 4779 |
+
},
|
| 4780 |
+
{
|
| 4781 |
+
"epoch": 0.26850937808489633,
|
| 4782 |
+
"grad_norm": 0.5735609531402588,
|
| 4783 |
+
"learning_rate": 0.00010779037807174033,
|
| 4784 |
+
"loss": 0.983,
|
| 4785 |
+
"step": 680
|
| 4786 |
+
},
|
| 4787 |
+
{
|
| 4788 |
+
"epoch": 0.2689042448173741,
|
| 4789 |
+
"grad_norm": 0.5265182256698608,
|
| 4790 |
+
"learning_rate": 0.00010756822806857028,
|
| 4791 |
+
"loss": 0.8681,
|
| 4792 |
+
"step": 681
|
| 4793 |
+
},
|
| 4794 |
+
{
|
| 4795 |
+
"epoch": 0.2692991115498519,
|
| 4796 |
+
"grad_norm": 0.5432460308074951,
|
| 4797 |
+
"learning_rate": 0.00010734604049415822,
|
| 4798 |
+
"loss": 0.7613,
|
| 4799 |
+
"step": 682
|
| 4800 |
+
},
|
| 4801 |
+
{
|
| 4802 |
+
"epoch": 0.2696939782823297,
|
| 4803 |
+
"grad_norm": 0.5309749245643616,
|
| 4804 |
+
"learning_rate": 0.00010712381645151844,
|
| 4805 |
+
"loss": 1.1094,
|
| 4806 |
+
"step": 683
|
| 4807 |
+
},
|
| 4808 |
+
{
|
| 4809 |
+
"epoch": 0.2700888450148075,
|
| 4810 |
+
"grad_norm": 0.588448703289032,
|
| 4811 |
+
"learning_rate": 0.00010690155704384615,
|
| 4812 |
+
"loss": 1.082,
|
| 4813 |
+
"step": 684
|
| 4814 |
+
},
|
| 4815 |
+
{
|
| 4816 |
+
"epoch": 0.2704837117472853,
|
| 4817 |
+
"grad_norm": 0.4991750717163086,
|
| 4818 |
+
"learning_rate": 0.00010667926337451217,
|
| 4819 |
+
"loss": 0.8744,
|
| 4820 |
+
"step": 685
|
| 4821 |
+
},
|
| 4822 |
+
{
|
| 4823 |
+
"epoch": 0.2708785784797631,
|
| 4824 |
+
"grad_norm": 0.4775620698928833,
|
| 4825 |
+
"learning_rate": 0.0001064569365470574,
|
| 4826 |
+
"loss": 1.025,
|
| 4827 |
+
"step": 686
|
| 4828 |
+
},
|
| 4829 |
+
{
|
| 4830 |
+
"epoch": 0.2712734452122409,
|
| 4831 |
+
"grad_norm": 0.5142999291419983,
|
| 4832 |
+
"learning_rate": 0.00010623457766518736,
|
| 4833 |
+
"loss": 0.9704,
|
| 4834 |
+
"step": 687
|
| 4835 |
+
},
|
| 4836 |
+
{
|
| 4837 |
+
"epoch": 0.27166831194471863,
|
| 4838 |
+
"grad_norm": 0.48825737833976746,
|
| 4839 |
+
"learning_rate": 0.00010601218783276672,
|
| 4840 |
+
"loss": 1.122,
|
| 4841 |
+
"step": 688
|
| 4842 |
+
},
|
| 4843 |
+
{
|
| 4844 |
+
"epoch": 0.2720631786771964,
|
| 4845 |
+
"grad_norm": 0.5517759919166565,
|
| 4846 |
+
"learning_rate": 0.00010578976815381372,
|
| 4847 |
+
"loss": 0.8859,
|
| 4848 |
+
"step": 689
|
| 4849 |
+
},
|
| 4850 |
+
{
|
| 4851 |
+
"epoch": 0.2724580454096742,
|
| 4852 |
+
"grad_norm": 0.5661524534225464,
|
| 4853 |
+
"learning_rate": 0.00010556731973249485,
|
| 4854 |
+
"loss": 0.89,
|
| 4855 |
+
"step": 690
|
| 4856 |
+
},
|
| 4857 |
+
{
|
| 4858 |
+
"epoch": 0.272852912142152,
|
| 4859 |
+
"grad_norm": 0.5802194476127625,
|
| 4860 |
+
"learning_rate": 0.00010534484367311923,
|
| 4861 |
+
"loss": 0.8048,
|
| 4862 |
+
"step": 691
|
| 4863 |
+
},
|
| 4864 |
+
{
|
| 4865 |
+
"epoch": 0.2732477788746298,
|
| 4866 |
+
"grad_norm": 0.5603414177894592,
|
| 4867 |
+
"learning_rate": 0.00010512234108013319,
|
| 4868 |
+
"loss": 1.0869,
|
| 4869 |
+
"step": 692
|
| 4870 |
+
},
|
| 4871 |
+
{
|
| 4872 |
+
"epoch": 0.2736426456071076,
|
| 4873 |
+
"grad_norm": 0.5277503728866577,
|
| 4874 |
+
"learning_rate": 0.00010489981305811487,
|
| 4875 |
+
"loss": 0.7287,
|
| 4876 |
+
"step": 693
|
| 4877 |
+
},
|
| 4878 |
+
{
|
| 4879 |
+
"epoch": 0.2740375123395854,
|
| 4880 |
+
"grad_norm": 0.6045374870300293,
|
| 4881 |
+
"learning_rate": 0.00010467726071176853,
|
| 4882 |
+
"loss": 0.9417,
|
| 4883 |
+
"step": 694
|
| 4884 |
+
},
|
| 4885 |
+
{
|
| 4886 |
+
"epoch": 0.2744323790720632,
|
| 4887 |
+
"grad_norm": 0.6021410822868347,
|
| 4888 |
+
"learning_rate": 0.00010445468514591925,
|
| 4889 |
+
"loss": 1.1399,
|
| 4890 |
+
"step": 695
|
| 4891 |
+
},
|
| 4892 |
+
{
|
| 4893 |
+
"epoch": 0.274827245804541,
|
| 4894 |
+
"grad_norm": 0.5544920563697815,
|
| 4895 |
+
"learning_rate": 0.00010423208746550732,
|
| 4896 |
+
"loss": 0.7279,
|
| 4897 |
+
"step": 696
|
| 4898 |
+
},
|
| 4899 |
+
{
|
| 4900 |
+
"epoch": 0.2752221125370188,
|
| 4901 |
+
"grad_norm": 0.5612072348594666,
|
| 4902 |
+
"learning_rate": 0.00010400946877558293,
|
| 4903 |
+
"loss": 0.8715,
|
| 4904 |
+
"step": 697
|
| 4905 |
+
},
|
| 4906 |
+
{
|
| 4907 |
+
"epoch": 0.2756169792694965,
|
| 4908 |
+
"grad_norm": 0.5820972323417664,
|
| 4909 |
+
"learning_rate": 0.00010378683018130047,
|
| 4910 |
+
"loss": 0.9456,
|
| 4911 |
+
"step": 698
|
| 4912 |
+
},
|
| 4913 |
+
{
|
| 4914 |
+
"epoch": 0.2760118460019743,
|
| 4915 |
+
"grad_norm": 0.43969717621803284,
|
| 4916 |
+
"learning_rate": 0.0001035641727879131,
|
| 4917 |
+
"loss": 0.8192,
|
| 4918 |
+
"step": 699
|
| 4919 |
+
},
|
| 4920 |
+
{
|
| 4921 |
+
"epoch": 0.2764067127344521,
|
| 4922 |
+
"grad_norm": 0.6528908610343933,
|
| 4923 |
+
"learning_rate": 0.00010334149770076747,
|
| 4924 |
+
"loss": 0.9742,
|
| 4925 |
+
"step": 700
|
| 4926 |
+
},
|
| 4927 |
+
{
|
| 4928 |
+
"epoch": 0.2768015794669299,
|
| 4929 |
+
"grad_norm": 0.4668010473251343,
|
| 4930 |
+
"learning_rate": 0.00010311880602529794,
|
| 4931 |
+
"loss": 0.9471,
|
| 4932 |
+
"step": 701
|
| 4933 |
+
},
|
| 4934 |
+
{
|
| 4935 |
+
"epoch": 0.2771964461994077,
|
| 4936 |
+
"grad_norm": 0.5474753379821777,
|
| 4937 |
+
"learning_rate": 0.0001028960988670212,
|
| 4938 |
+
"loss": 0.9224,
|
| 4939 |
+
"step": 702
|
| 4940 |
+
},
|
| 4941 |
+
{
|
| 4942 |
+
"epoch": 0.2775913129318855,
|
| 4943 |
+
"grad_norm": 0.5672122240066528,
|
| 4944 |
+
"learning_rate": 0.00010267337733153089,
|
| 4945 |
+
"loss": 0.8441,
|
| 4946 |
+
"step": 703
|
| 4947 |
+
},
|
| 4948 |
+
{
|
| 4949 |
+
"epoch": 0.2779861796643633,
|
| 4950 |
+
"grad_norm": 0.5570011138916016,
|
| 4951 |
+
"learning_rate": 0.00010245064252449201,
|
| 4952 |
+
"loss": 0.9897,
|
| 4953 |
+
"step": 704
|
| 4954 |
+
},
|
| 4955 |
+
{
|
| 4956 |
+
"epoch": 0.2783810463968411,
|
| 4957 |
+
"grad_norm": 0.5113812685012817,
|
| 4958 |
+
"learning_rate": 0.0001022278955516354,
|
| 4959 |
+
"loss": 1.0066,
|
| 4960 |
+
"step": 705
|
| 4961 |
+
},
|
| 4962 |
+
{
|
| 4963 |
+
"epoch": 0.2787759131293189,
|
| 4964 |
+
"grad_norm": 0.48175865411758423,
|
| 4965 |
+
"learning_rate": 0.00010200513751875227,
|
| 4966 |
+
"loss": 0.9212,
|
| 4967 |
+
"step": 706
|
| 4968 |
+
},
|
| 4969 |
+
{
|
| 4970 |
+
"epoch": 0.27917077986179667,
|
| 4971 |
+
"grad_norm": 0.5643198490142822,
|
| 4972 |
+
"learning_rate": 0.00010178236953168885,
|
| 4973 |
+
"loss": 1.2371,
|
| 4974 |
+
"step": 707
|
| 4975 |
+
},
|
| 4976 |
+
{
|
| 4977 |
+
"epoch": 0.2795656465942744,
|
| 4978 |
+
"grad_norm": 0.5115451216697693,
|
| 4979 |
+
"learning_rate": 0.00010155959269634068,
|
| 4980 |
+
"loss": 0.9701,
|
| 4981 |
+
"step": 708
|
| 4982 |
+
},
|
| 4983 |
+
{
|
| 4984 |
+
"epoch": 0.2799605133267522,
|
| 4985 |
+
"grad_norm": 0.5253522396087646,
|
| 4986 |
+
"learning_rate": 0.00010133680811864727,
|
| 4987 |
+
"loss": 1.0685,
|
| 4988 |
+
"step": 709
|
| 4989 |
+
},
|
| 4990 |
+
{
|
| 4991 |
+
"epoch": 0.28035538005923,
|
| 4992 |
+
"grad_norm": 0.7160750031471252,
|
| 4993 |
+
"learning_rate": 0.00010111401690458654,
|
| 4994 |
+
"loss": 1.0837,
|
| 4995 |
+
"step": 710
|
| 4996 |
+
},
|
| 4997 |
+
{
|
| 4998 |
+
"epoch": 0.28035538005923,
|
| 4999 |
+
"eval_loss": 0.9457208514213562,
|
| 5000 |
+
"eval_runtime": 61.5761,
|
| 5001 |
+
"eval_samples_per_second": 17.328,
|
| 5002 |
+
"eval_steps_per_second": 8.672,
|
| 5003 |
+
"step": 710
|
| 5004 |
}
|
| 5005 |
],
|
| 5006 |
"logging_steps": 1,
|
|
|
|
| 5020 |
"attributes": {}
|
| 5021 |
}
|
| 5022 |
},
|
| 5023 |
+
"total_flos": 2.602799288203346e+17,
|
| 5024 |
"train_batch_size": 2,
|
| 5025 |
"trial_name": null,
|
| 5026 |
"trial_params": null
|