Training in progress, step 682, 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 sha256:56d4950327e454eb546afcd24441040fcf134a356d000a4ee7f08f3e052c16f6
|
| 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 sha256:362f909bee5b34580b914d9065cdb21e5d0995cd8e786e547b2344f41615cc1e
|
| 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 sha256:4578a006fc558e462c5793425143656a8c4b614c7d3f96ea0cac6c8cb67c39ea
|
| 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:e7940a7e9a9ca38750caa51b3df3b13af904a9fadc3ec71ad157f8c2d907c0d1
|
| 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": 341,
|
| 6 |
-
"global_step":
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
@@ -2410,6 +2410,2401 @@
|
|
| 2410 |
"eval_samples_per_second": 27.624,
|
| 2411 |
"eval_steps_per_second": 13.812,
|
| 2412 |
"step": 341
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 2413 |
}
|
| 2414 |
],
|
| 2415 |
"logging_steps": 1,
|
|
@@ -2429,7 +4824,7 @@
|
|
| 2429 |
"attributes": {}
|
| 2430 |
}
|
| 2431 |
},
|
| 2432 |
-
"total_flos":
|
| 2433 |
"train_batch_size": 2,
|
| 2434 |
"trial_name": null,
|
| 2435 |
"trial_params": null
|
|
|
|
| 1 |
{
|
| 2 |
"best_metric": null,
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 0.1337386018237082,
|
| 5 |
"eval_steps": 341,
|
| 6 |
+
"global_step": 682,
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
|
|
| 2410 |
"eval_samples_per_second": 27.624,
|
| 2411 |
"eval_steps_per_second": 13.812,
|
| 2412 |
"step": 341
|
| 2413 |
+
},
|
| 2414 |
+
{
|
| 2415 |
+
"epoch": 0.0670653985684871,
|
| 2416 |
+
"grad_norm": 9.668785095214844,
|
| 2417 |
+
"learning_rate": 0.00017164979829596165,
|
| 2418 |
+
"loss": 3.3758,
|
| 2419 |
+
"step": 342
|
| 2420 |
+
},
|
| 2421 |
+
{
|
| 2422 |
+
"epoch": 0.06726149622512011,
|
| 2423 |
+
"grad_norm": 5.6129255294799805,
|
| 2424 |
+
"learning_rate": 0.00017148738851297256,
|
| 2425 |
+
"loss": 1.3369,
|
| 2426 |
+
"step": 343
|
| 2427 |
+
},
|
| 2428 |
+
{
|
| 2429 |
+
"epoch": 0.06745759388175311,
|
| 2430 |
+
"grad_norm": 8.042420387268066,
|
| 2431 |
+
"learning_rate": 0.0001713245921687629,
|
| 2432 |
+
"loss": 1.918,
|
| 2433 |
+
"step": 344
|
| 2434 |
+
},
|
| 2435 |
+
{
|
| 2436 |
+
"epoch": 0.06765369153838612,
|
| 2437 |
+
"grad_norm": 6.536214351654053,
|
| 2438 |
+
"learning_rate": 0.00017116141014363837,
|
| 2439 |
+
"loss": 0.97,
|
| 2440 |
+
"step": 345
|
| 2441 |
+
},
|
| 2442 |
+
{
|
| 2443 |
+
"epoch": 0.06784978919501912,
|
| 2444 |
+
"grad_norm": 10.576717376708984,
|
| 2445 |
+
"learning_rate": 0.0001709978433199901,
|
| 2446 |
+
"loss": 2.0315,
|
| 2447 |
+
"step": 346
|
| 2448 |
+
},
|
| 2449 |
+
{
|
| 2450 |
+
"epoch": 0.06804588685165212,
|
| 2451 |
+
"grad_norm": 13.645353317260742,
|
| 2452 |
+
"learning_rate": 0.00017083389258229013,
|
| 2453 |
+
"loss": 2.3226,
|
| 2454 |
+
"step": 347
|
| 2455 |
+
},
|
| 2456 |
+
{
|
| 2457 |
+
"epoch": 0.06824198450828513,
|
| 2458 |
+
"grad_norm": 10.135254859924316,
|
| 2459 |
+
"learning_rate": 0.00017066955881708636,
|
| 2460 |
+
"loss": 2.4686,
|
| 2461 |
+
"step": 348
|
| 2462 |
+
},
|
| 2463 |
+
{
|
| 2464 |
+
"epoch": 0.06843808216491813,
|
| 2465 |
+
"grad_norm": 5.7091569900512695,
|
| 2466 |
+
"learning_rate": 0.0001705048429129979,
|
| 2467 |
+
"loss": 1.633,
|
| 2468 |
+
"step": 349
|
| 2469 |
+
},
|
| 2470 |
+
{
|
| 2471 |
+
"epoch": 0.06863417982155114,
|
| 2472 |
+
"grad_norm": 7.3178935050964355,
|
| 2473 |
+
"learning_rate": 0.0001703397457607103,
|
| 2474 |
+
"loss": 1.5411,
|
| 2475 |
+
"step": 350
|
| 2476 |
+
},
|
| 2477 |
+
{
|
| 2478 |
+
"epoch": 0.06883027747818414,
|
| 2479 |
+
"grad_norm": 10.689863204956055,
|
| 2480 |
+
"learning_rate": 0.0001701742682529706,
|
| 2481 |
+
"loss": 2.8238,
|
| 2482 |
+
"step": 351
|
| 2483 |
+
},
|
| 2484 |
+
{
|
| 2485 |
+
"epoch": 0.06902637513481714,
|
| 2486 |
+
"grad_norm": 4.083006381988525,
|
| 2487 |
+
"learning_rate": 0.00017000841128458265,
|
| 2488 |
+
"loss": 1.7831,
|
| 2489 |
+
"step": 352
|
| 2490 |
+
},
|
| 2491 |
+
{
|
| 2492 |
+
"epoch": 0.06922247279145015,
|
| 2493 |
+
"grad_norm": 4.592381477355957,
|
| 2494 |
+
"learning_rate": 0.0001698421757524021,
|
| 2495 |
+
"loss": 1.1135,
|
| 2496 |
+
"step": 353
|
| 2497 |
+
},
|
| 2498 |
+
{
|
| 2499 |
+
"epoch": 0.06941857044808314,
|
| 2500 |
+
"grad_norm": 9.44848346710205,
|
| 2501 |
+
"learning_rate": 0.00016967556255533174,
|
| 2502 |
+
"loss": 1.6301,
|
| 2503 |
+
"step": 354
|
| 2504 |
+
},
|
| 2505 |
+
{
|
| 2506 |
+
"epoch": 0.06961466810471614,
|
| 2507 |
+
"grad_norm": 7.35884428024292,
|
| 2508 |
+
"learning_rate": 0.0001695085725943165,
|
| 2509 |
+
"loss": 1.6776,
|
| 2510 |
+
"step": 355
|
| 2511 |
+
},
|
| 2512 |
+
{
|
| 2513 |
+
"epoch": 0.06981076576134915,
|
| 2514 |
+
"grad_norm": 7.0129475593566895,
|
| 2515 |
+
"learning_rate": 0.00016934120677233863,
|
| 2516 |
+
"loss": 1.1665,
|
| 2517 |
+
"step": 356
|
| 2518 |
+
},
|
| 2519 |
+
{
|
| 2520 |
+
"epoch": 0.07000686341798215,
|
| 2521 |
+
"grad_norm": 11.612259864807129,
|
| 2522 |
+
"learning_rate": 0.00016917346599441276,
|
| 2523 |
+
"loss": 0.9044,
|
| 2524 |
+
"step": 357
|
| 2525 |
+
},
|
| 2526 |
+
{
|
| 2527 |
+
"epoch": 0.07020296107461516,
|
| 2528 |
+
"grad_norm": 5.861693382263184,
|
| 2529 |
+
"learning_rate": 0.00016900535116758125,
|
| 2530 |
+
"loss": 1.5738,
|
| 2531 |
+
"step": 358
|
| 2532 |
+
},
|
| 2533 |
+
{
|
| 2534 |
+
"epoch": 0.07039905873124816,
|
| 2535 |
+
"grad_norm": 4.579985618591309,
|
| 2536 |
+
"learning_rate": 0.00016883686320090884,
|
| 2537 |
+
"loss": 1.6305,
|
| 2538 |
+
"step": 359
|
| 2539 |
+
},
|
| 2540 |
+
{
|
| 2541 |
+
"epoch": 0.07059515638788116,
|
| 2542 |
+
"grad_norm": 6.142041206359863,
|
| 2543 |
+
"learning_rate": 0.00016866800300547813,
|
| 2544 |
+
"loss": 0.6206,
|
| 2545 |
+
"step": 360
|
| 2546 |
+
},
|
| 2547 |
+
{
|
| 2548 |
+
"epoch": 0.07079125404451417,
|
| 2549 |
+
"grad_norm": 3.1343085765838623,
|
| 2550 |
+
"learning_rate": 0.0001684987714943845,
|
| 2551 |
+
"loss": 0.7649,
|
| 2552 |
+
"step": 361
|
| 2553 |
+
},
|
| 2554 |
+
{
|
| 2555 |
+
"epoch": 0.07098735170114717,
|
| 2556 |
+
"grad_norm": 5.473539352416992,
|
| 2557 |
+
"learning_rate": 0.00016832916958273118,
|
| 2558 |
+
"loss": 1.4839,
|
| 2559 |
+
"step": 362
|
| 2560 |
+
},
|
| 2561 |
+
{
|
| 2562 |
+
"epoch": 0.07118344935778018,
|
| 2563 |
+
"grad_norm": 8.52957534790039,
|
| 2564 |
+
"learning_rate": 0.00016815919818762427,
|
| 2565 |
+
"loss": 2.5437,
|
| 2566 |
+
"step": 363
|
| 2567 |
+
},
|
| 2568 |
+
{
|
| 2569 |
+
"epoch": 0.07137954701441318,
|
| 2570 |
+
"grad_norm": 6.140387058258057,
|
| 2571 |
+
"learning_rate": 0.00016798885822816786,
|
| 2572 |
+
"loss": 1.2282,
|
| 2573 |
+
"step": 364
|
| 2574 |
+
},
|
| 2575 |
+
{
|
| 2576 |
+
"epoch": 0.07157564467104618,
|
| 2577 |
+
"grad_norm": 9.70067024230957,
|
| 2578 |
+
"learning_rate": 0.000167818150625459,
|
| 2579 |
+
"loss": 2.3286,
|
| 2580 |
+
"step": 365
|
| 2581 |
+
},
|
| 2582 |
+
{
|
| 2583 |
+
"epoch": 0.07177174232767919,
|
| 2584 |
+
"grad_norm": 6.14666223526001,
|
| 2585 |
+
"learning_rate": 0.0001676470763025827,
|
| 2586 |
+
"loss": 1.3825,
|
| 2587 |
+
"step": 366
|
| 2588 |
+
},
|
| 2589 |
+
{
|
| 2590 |
+
"epoch": 0.07196783998431219,
|
| 2591 |
+
"grad_norm": 10.377222061157227,
|
| 2592 |
+
"learning_rate": 0.0001674756361846071,
|
| 2593 |
+
"loss": 1.9278,
|
| 2594 |
+
"step": 367
|
| 2595 |
+
},
|
| 2596 |
+
{
|
| 2597 |
+
"epoch": 0.0721639376409452,
|
| 2598 |
+
"grad_norm": 7.09204626083374,
|
| 2599 |
+
"learning_rate": 0.00016730383119857817,
|
| 2600 |
+
"loss": 1.2413,
|
| 2601 |
+
"step": 368
|
| 2602 |
+
},
|
| 2603 |
+
{
|
| 2604 |
+
"epoch": 0.07236003529757819,
|
| 2605 |
+
"grad_norm": 12.448009490966797,
|
| 2606 |
+
"learning_rate": 0.00016713166227351497,
|
| 2607 |
+
"loss": 1.7415,
|
| 2608 |
+
"step": 369
|
| 2609 |
+
},
|
| 2610 |
+
{
|
| 2611 |
+
"epoch": 0.07255613295421119,
|
| 2612 |
+
"grad_norm": 4.5050811767578125,
|
| 2613 |
+
"learning_rate": 0.00016695913034040454,
|
| 2614 |
+
"loss": 1.7608,
|
| 2615 |
+
"step": 370
|
| 2616 |
+
},
|
| 2617 |
+
{
|
| 2618 |
+
"epoch": 0.0727522306108442,
|
| 2619 |
+
"grad_norm": 8.113517761230469,
|
| 2620 |
+
"learning_rate": 0.00016678623633219677,
|
| 2621 |
+
"loss": 1.8336,
|
| 2622 |
+
"step": 371
|
| 2623 |
+
},
|
| 2624 |
+
{
|
| 2625 |
+
"epoch": 0.0729483282674772,
|
| 2626 |
+
"grad_norm": 7.662284851074219,
|
| 2627 |
+
"learning_rate": 0.00016661298118379948,
|
| 2628 |
+
"loss": 2.5928,
|
| 2629 |
+
"step": 372
|
| 2630 |
+
},
|
| 2631 |
+
{
|
| 2632 |
+
"epoch": 0.0731444259241102,
|
| 2633 |
+
"grad_norm": 7.647708892822266,
|
| 2634 |
+
"learning_rate": 0.00016643936583207337,
|
| 2635 |
+
"loss": 3.006,
|
| 2636 |
+
"step": 373
|
| 2637 |
+
},
|
| 2638 |
+
{
|
| 2639 |
+
"epoch": 0.07334052358074321,
|
| 2640 |
+
"grad_norm": 7.049048900604248,
|
| 2641 |
+
"learning_rate": 0.00016626539121582685,
|
| 2642 |
+
"loss": 1.9476,
|
| 2643 |
+
"step": 374
|
| 2644 |
+
},
|
| 2645 |
+
{
|
| 2646 |
+
"epoch": 0.07353662123737621,
|
| 2647 |
+
"grad_norm": 4.551039218902588,
|
| 2648 |
+
"learning_rate": 0.000166091058275811,
|
| 2649 |
+
"loss": 2.2753,
|
| 2650 |
+
"step": 375
|
| 2651 |
+
},
|
| 2652 |
+
{
|
| 2653 |
+
"epoch": 0.07373271889400922,
|
| 2654 |
+
"grad_norm": 8.680680274963379,
|
| 2655 |
+
"learning_rate": 0.00016591636795471455,
|
| 2656 |
+
"loss": 2.0624,
|
| 2657 |
+
"step": 376
|
| 2658 |
+
},
|
| 2659 |
+
{
|
| 2660 |
+
"epoch": 0.07392881655064222,
|
| 2661 |
+
"grad_norm": 10.324525833129883,
|
| 2662 |
+
"learning_rate": 0.0001657413211971587,
|
| 2663 |
+
"loss": 1.6293,
|
| 2664 |
+
"step": 377
|
| 2665 |
+
},
|
| 2666 |
+
{
|
| 2667 |
+
"epoch": 0.07412491420727522,
|
| 2668 |
+
"grad_norm": 5.324687957763672,
|
| 2669 |
+
"learning_rate": 0.0001655659189496921,
|
| 2670 |
+
"loss": 1.4987,
|
| 2671 |
+
"step": 378
|
| 2672 |
+
},
|
| 2673 |
+
{
|
| 2674 |
+
"epoch": 0.07432101186390823,
|
| 2675 |
+
"grad_norm": 11.023552894592285,
|
| 2676 |
+
"learning_rate": 0.00016539016216078557,
|
| 2677 |
+
"loss": 1.9108,
|
| 2678 |
+
"step": 379
|
| 2679 |
+
},
|
| 2680 |
+
{
|
| 2681 |
+
"epoch": 0.07451710952054123,
|
| 2682 |
+
"grad_norm": 6.6586594581604,
|
| 2683 |
+
"learning_rate": 0.00016521405178082722,
|
| 2684 |
+
"loss": 2.4198,
|
| 2685 |
+
"step": 380
|
| 2686 |
+
},
|
| 2687 |
+
{
|
| 2688 |
+
"epoch": 0.07471320717717424,
|
| 2689 |
+
"grad_norm": 9.818254470825195,
|
| 2690 |
+
"learning_rate": 0.0001650375887621171,
|
| 2691 |
+
"loss": 2.162,
|
| 2692 |
+
"step": 381
|
| 2693 |
+
},
|
| 2694 |
+
{
|
| 2695 |
+
"epoch": 0.07490930483380724,
|
| 2696 |
+
"grad_norm": 6.547168254852295,
|
| 2697 |
+
"learning_rate": 0.000164860774058862,
|
| 2698 |
+
"loss": 1.8307,
|
| 2699 |
+
"step": 382
|
| 2700 |
+
},
|
| 2701 |
+
{
|
| 2702 |
+
"epoch": 0.07510540249044025,
|
| 2703 |
+
"grad_norm": 7.256173610687256,
|
| 2704 |
+
"learning_rate": 0.00016468360862717066,
|
| 2705 |
+
"loss": 0.9432,
|
| 2706 |
+
"step": 383
|
| 2707 |
+
},
|
| 2708 |
+
{
|
| 2709 |
+
"epoch": 0.07530150014707324,
|
| 2710 |
+
"grad_norm": 10.348817825317383,
|
| 2711 |
+
"learning_rate": 0.00016450609342504813,
|
| 2712 |
+
"loss": 1.7824,
|
| 2713 |
+
"step": 384
|
| 2714 |
+
},
|
| 2715 |
+
{
|
| 2716 |
+
"epoch": 0.07549759780370624,
|
| 2717 |
+
"grad_norm": 14.728826522827148,
|
| 2718 |
+
"learning_rate": 0.00016432822941239096,
|
| 2719 |
+
"loss": 3.1737,
|
| 2720 |
+
"step": 385
|
| 2721 |
+
},
|
| 2722 |
+
{
|
| 2723 |
+
"epoch": 0.07569369546033924,
|
| 2724 |
+
"grad_norm": 7.894331932067871,
|
| 2725 |
+
"learning_rate": 0.00016415001755098175,
|
| 2726 |
+
"loss": 2.0515,
|
| 2727 |
+
"step": 386
|
| 2728 |
+
},
|
| 2729 |
+
{
|
| 2730 |
+
"epoch": 0.07588979311697225,
|
| 2731 |
+
"grad_norm": 7.474819183349609,
|
| 2732 |
+
"learning_rate": 0.00016397145880448416,
|
| 2733 |
+
"loss": 3.1276,
|
| 2734 |
+
"step": 387
|
| 2735 |
+
},
|
| 2736 |
+
{
|
| 2737 |
+
"epoch": 0.07608589077360525,
|
| 2738 |
+
"grad_norm": 6.2442626953125,
|
| 2739 |
+
"learning_rate": 0.00016379255413843754,
|
| 2740 |
+
"loss": 0.9294,
|
| 2741 |
+
"step": 388
|
| 2742 |
+
},
|
| 2743 |
+
{
|
| 2744 |
+
"epoch": 0.07628198843023826,
|
| 2745 |
+
"grad_norm": 8.229086875915527,
|
| 2746 |
+
"learning_rate": 0.0001636133045202517,
|
| 2747 |
+
"loss": 2.122,
|
| 2748 |
+
"step": 389
|
| 2749 |
+
},
|
| 2750 |
+
{
|
| 2751 |
+
"epoch": 0.07647808608687126,
|
| 2752 |
+
"grad_norm": 5.9484100341796875,
|
| 2753 |
+
"learning_rate": 0.0001634337109192019,
|
| 2754 |
+
"loss": 1.3164,
|
| 2755 |
+
"step": 390
|
| 2756 |
+
},
|
| 2757 |
+
{
|
| 2758 |
+
"epoch": 0.07667418374350427,
|
| 2759 |
+
"grad_norm": 6.490477561950684,
|
| 2760 |
+
"learning_rate": 0.0001632537743064233,
|
| 2761 |
+
"loss": 1.1598,
|
| 2762 |
+
"step": 391
|
| 2763 |
+
},
|
| 2764 |
+
{
|
| 2765 |
+
"epoch": 0.07687028140013727,
|
| 2766 |
+
"grad_norm": 36.34891128540039,
|
| 2767 |
+
"learning_rate": 0.000163073495654906,
|
| 2768 |
+
"loss": 2.4203,
|
| 2769 |
+
"step": 392
|
| 2770 |
+
},
|
| 2771 |
+
{
|
| 2772 |
+
"epoch": 0.07706637905677027,
|
| 2773 |
+
"grad_norm": 5.949688911437988,
|
| 2774 |
+
"learning_rate": 0.00016289287593948952,
|
| 2775 |
+
"loss": 2.0893,
|
| 2776 |
+
"step": 393
|
| 2777 |
+
},
|
| 2778 |
+
{
|
| 2779 |
+
"epoch": 0.07726247671340328,
|
| 2780 |
+
"grad_norm": 5.93861198425293,
|
| 2781 |
+
"learning_rate": 0.00016271191613685776,
|
| 2782 |
+
"loss": 1.829,
|
| 2783 |
+
"step": 394
|
| 2784 |
+
},
|
| 2785 |
+
{
|
| 2786 |
+
"epoch": 0.07745857437003628,
|
| 2787 |
+
"grad_norm": 8.41703987121582,
|
| 2788 |
+
"learning_rate": 0.00016253061722553355,
|
| 2789 |
+
"loss": 2.1544,
|
| 2790 |
+
"step": 395
|
| 2791 |
+
},
|
| 2792 |
+
{
|
| 2793 |
+
"epoch": 0.07765467202666929,
|
| 2794 |
+
"grad_norm": 8.92647647857666,
|
| 2795 |
+
"learning_rate": 0.00016234898018587337,
|
| 2796 |
+
"loss": 1.3316,
|
| 2797 |
+
"step": 396
|
| 2798 |
+
},
|
| 2799 |
+
{
|
| 2800 |
+
"epoch": 0.07785076968330229,
|
| 2801 |
+
"grad_norm": 4.4784698486328125,
|
| 2802 |
+
"learning_rate": 0.0001621670060000622,
|
| 2803 |
+
"loss": 1.0305,
|
| 2804 |
+
"step": 397
|
| 2805 |
+
},
|
| 2806 |
+
{
|
| 2807 |
+
"epoch": 0.0780468673399353,
|
| 2808 |
+
"grad_norm": 10.080864906311035,
|
| 2809 |
+
"learning_rate": 0.00016198469565210805,
|
| 2810 |
+
"loss": 2.1012,
|
| 2811 |
+
"step": 398
|
| 2812 |
+
},
|
| 2813 |
+
{
|
| 2814 |
+
"epoch": 0.07824296499656828,
|
| 2815 |
+
"grad_norm": 8.505258560180664,
|
| 2816 |
+
"learning_rate": 0.0001618020501278367,
|
| 2817 |
+
"loss": 2.2348,
|
| 2818 |
+
"step": 399
|
| 2819 |
+
},
|
| 2820 |
+
{
|
| 2821 |
+
"epoch": 0.07843906265320129,
|
| 2822 |
+
"grad_norm": 6.3116559982299805,
|
| 2823 |
+
"learning_rate": 0.00016161907041488635,
|
| 2824 |
+
"loss": 1.3726,
|
| 2825 |
+
"step": 400
|
| 2826 |
+
},
|
| 2827 |
+
{
|
| 2828 |
+
"epoch": 0.07863516030983429,
|
| 2829 |
+
"grad_norm": 7.111060619354248,
|
| 2830 |
+
"learning_rate": 0.00016143575750270233,
|
| 2831 |
+
"loss": 1.5926,
|
| 2832 |
+
"step": 401
|
| 2833 |
+
},
|
| 2834 |
+
{
|
| 2835 |
+
"epoch": 0.0788312579664673,
|
| 2836 |
+
"grad_norm": 5.481679916381836,
|
| 2837 |
+
"learning_rate": 0.0001612521123825317,
|
| 2838 |
+
"loss": 0.8062,
|
| 2839 |
+
"step": 402
|
| 2840 |
+
},
|
| 2841 |
+
{
|
| 2842 |
+
"epoch": 0.0790273556231003,
|
| 2843 |
+
"grad_norm": 9.399847984313965,
|
| 2844 |
+
"learning_rate": 0.00016106813604741782,
|
| 2845 |
+
"loss": 1.7107,
|
| 2846 |
+
"step": 403
|
| 2847 |
+
},
|
| 2848 |
+
{
|
| 2849 |
+
"epoch": 0.0792234532797333,
|
| 2850 |
+
"grad_norm": 7.784383296966553,
|
| 2851 |
+
"learning_rate": 0.00016088382949219524,
|
| 2852 |
+
"loss": 1.4449,
|
| 2853 |
+
"step": 404
|
| 2854 |
+
},
|
| 2855 |
+
{
|
| 2856 |
+
"epoch": 0.07941955093636631,
|
| 2857 |
+
"grad_norm": 10.075730323791504,
|
| 2858 |
+
"learning_rate": 0.000160699193713484,
|
| 2859 |
+
"loss": 2.3277,
|
| 2860 |
+
"step": 405
|
| 2861 |
+
},
|
| 2862 |
+
{
|
| 2863 |
+
"epoch": 0.07961564859299931,
|
| 2864 |
+
"grad_norm": 4.376564979553223,
|
| 2865 |
+
"learning_rate": 0.00016051422970968438,
|
| 2866 |
+
"loss": 1.3489,
|
| 2867 |
+
"step": 406
|
| 2868 |
+
},
|
| 2869 |
+
{
|
| 2870 |
+
"epoch": 0.07981174624963232,
|
| 2871 |
+
"grad_norm": 28.959577560424805,
|
| 2872 |
+
"learning_rate": 0.00016032893848097165,
|
| 2873 |
+
"loss": 0.9703,
|
| 2874 |
+
"step": 407
|
| 2875 |
+
},
|
| 2876 |
+
{
|
| 2877 |
+
"epoch": 0.08000784390626532,
|
| 2878 |
+
"grad_norm": 9.604166030883789,
|
| 2879 |
+
"learning_rate": 0.00016014332102929027,
|
| 2880 |
+
"loss": 2.195,
|
| 2881 |
+
"step": 408
|
| 2882 |
+
},
|
| 2883 |
+
{
|
| 2884 |
+
"epoch": 0.08020394156289833,
|
| 2885 |
+
"grad_norm": 5.949897289276123,
|
| 2886 |
+
"learning_rate": 0.00015995737835834906,
|
| 2887 |
+
"loss": 0.8828,
|
| 2888 |
+
"step": 409
|
| 2889 |
+
},
|
| 2890 |
+
{
|
| 2891 |
+
"epoch": 0.08040003921953133,
|
| 2892 |
+
"grad_norm": 9.921558380126953,
|
| 2893 |
+
"learning_rate": 0.00015977111147361507,
|
| 2894 |
+
"loss": 2.1569,
|
| 2895 |
+
"step": 410
|
| 2896 |
+
},
|
| 2897 |
+
{
|
| 2898 |
+
"epoch": 0.08059613687616433,
|
| 2899 |
+
"grad_norm": 9.299111366271973,
|
| 2900 |
+
"learning_rate": 0.00015958452138230877,
|
| 2901 |
+
"loss": 1.4095,
|
| 2902 |
+
"step": 411
|
| 2903 |
+
},
|
| 2904 |
+
{
|
| 2905 |
+
"epoch": 0.08079223453279734,
|
| 2906 |
+
"grad_norm": 8.806063652038574,
|
| 2907 |
+
"learning_rate": 0.00015939760909339823,
|
| 2908 |
+
"loss": 1.7785,
|
| 2909 |
+
"step": 412
|
| 2910 |
+
},
|
| 2911 |
+
{
|
| 2912 |
+
"epoch": 0.08098833218943034,
|
| 2913 |
+
"grad_norm": 5.1091108322143555,
|
| 2914 |
+
"learning_rate": 0.00015921037561759377,
|
| 2915 |
+
"loss": 1.8415,
|
| 2916 |
+
"step": 413
|
| 2917 |
+
},
|
| 2918 |
+
{
|
| 2919 |
+
"epoch": 0.08118442984606333,
|
| 2920 |
+
"grad_norm": 12.836872100830078,
|
| 2921 |
+
"learning_rate": 0.0001590228219673425,
|
| 2922 |
+
"loss": 1.8644,
|
| 2923 |
+
"step": 414
|
| 2924 |
+
},
|
| 2925 |
+
{
|
| 2926 |
+
"epoch": 0.08138052750269634,
|
| 2927 |
+
"grad_norm": 6.567033767700195,
|
| 2928 |
+
"learning_rate": 0.00015883494915682289,
|
| 2929 |
+
"loss": 1.9557,
|
| 2930 |
+
"step": 415
|
| 2931 |
+
},
|
| 2932 |
+
{
|
| 2933 |
+
"epoch": 0.08157662515932934,
|
| 2934 |
+
"grad_norm": 8.996990203857422,
|
| 2935 |
+
"learning_rate": 0.00015864675820193922,
|
| 2936 |
+
"loss": 1.888,
|
| 2937 |
+
"step": 416
|
| 2938 |
+
},
|
| 2939 |
+
{
|
| 2940 |
+
"epoch": 0.08177272281596235,
|
| 2941 |
+
"grad_norm": 7.494692802429199,
|
| 2942 |
+
"learning_rate": 0.00015845825012031605,
|
| 2943 |
+
"loss": 2.4514,
|
| 2944 |
+
"step": 417
|
| 2945 |
+
},
|
| 2946 |
+
{
|
| 2947 |
+
"epoch": 0.08196882047259535,
|
| 2948 |
+
"grad_norm": 5.421525478363037,
|
| 2949 |
+
"learning_rate": 0.0001582694259312928,
|
| 2950 |
+
"loss": 1.0314,
|
| 2951 |
+
"step": 418
|
| 2952 |
+
},
|
| 2953 |
+
{
|
| 2954 |
+
"epoch": 0.08216491812922835,
|
| 2955 |
+
"grad_norm": 25.4990291595459,
|
| 2956 |
+
"learning_rate": 0.0001580802866559183,
|
| 2957 |
+
"loss": 0.9981,
|
| 2958 |
+
"step": 419
|
| 2959 |
+
},
|
| 2960 |
+
{
|
| 2961 |
+
"epoch": 0.08236101578586136,
|
| 2962 |
+
"grad_norm": 7.055636405944824,
|
| 2963 |
+
"learning_rate": 0.00015789083331694506,
|
| 2964 |
+
"loss": 1.718,
|
| 2965 |
+
"step": 420
|
| 2966 |
+
},
|
| 2967 |
+
{
|
| 2968 |
+
"epoch": 0.08255711344249436,
|
| 2969 |
+
"grad_norm": 8.986098289489746,
|
| 2970 |
+
"learning_rate": 0.00015770106693882387,
|
| 2971 |
+
"loss": 1.3962,
|
| 2972 |
+
"step": 421
|
| 2973 |
+
},
|
| 2974 |
+
{
|
| 2975 |
+
"epoch": 0.08275321109912737,
|
| 2976 |
+
"grad_norm": 4.7463788986206055,
|
| 2977 |
+
"learning_rate": 0.0001575109885476983,
|
| 2978 |
+
"loss": 1.4039,
|
| 2979 |
+
"step": 422
|
| 2980 |
+
},
|
| 2981 |
+
{
|
| 2982 |
+
"epoch": 0.08294930875576037,
|
| 2983 |
+
"grad_norm": 6.555906295776367,
|
| 2984 |
+
"learning_rate": 0.00015732059917139912,
|
| 2985 |
+
"loss": 1.0908,
|
| 2986 |
+
"step": 423
|
| 2987 |
+
},
|
| 2988 |
+
{
|
| 2989 |
+
"epoch": 0.08314540641239337,
|
| 2990 |
+
"grad_norm": 7.233346462249756,
|
| 2991 |
+
"learning_rate": 0.00015712989983943863,
|
| 2992 |
+
"loss": 1.9737,
|
| 2993 |
+
"step": 424
|
| 2994 |
+
},
|
| 2995 |
+
{
|
| 2996 |
+
"epoch": 0.08334150406902638,
|
| 2997 |
+
"grad_norm": 4.797744274139404,
|
| 2998 |
+
"learning_rate": 0.0001569388915830053,
|
| 2999 |
+
"loss": 1.2191,
|
| 3000 |
+
"step": 425
|
| 3001 |
+
},
|
| 3002 |
+
{
|
| 3003 |
+
"epoch": 0.08353760172565938,
|
| 3004 |
+
"grad_norm": 6.303066253662109,
|
| 3005 |
+
"learning_rate": 0.000156747575434958,
|
| 3006 |
+
"loss": 1.8937,
|
| 3007 |
+
"step": 426
|
| 3008 |
+
},
|
| 3009 |
+
{
|
| 3010 |
+
"epoch": 0.08373369938229239,
|
| 3011 |
+
"grad_norm": 12.17996597290039,
|
| 3012 |
+
"learning_rate": 0.00015655595242982048,
|
| 3013 |
+
"loss": 4.1888,
|
| 3014 |
+
"step": 427
|
| 3015 |
+
},
|
| 3016 |
+
{
|
| 3017 |
+
"epoch": 0.08392979703892539,
|
| 3018 |
+
"grad_norm": 7.345923900604248,
|
| 3019 |
+
"learning_rate": 0.00015636402360377587,
|
| 3020 |
+
"loss": 1.0776,
|
| 3021 |
+
"step": 428
|
| 3022 |
+
},
|
| 3023 |
+
{
|
| 3024 |
+
"epoch": 0.08412589469555838,
|
| 3025 |
+
"grad_norm": 7.688333034515381,
|
| 3026 |
+
"learning_rate": 0.0001561717899946609,
|
| 3027 |
+
"loss": 0.9205,
|
| 3028 |
+
"step": 429
|
| 3029 |
+
},
|
| 3030 |
+
{
|
| 3031 |
+
"epoch": 0.08432199235219139,
|
| 3032 |
+
"grad_norm": 5.487595558166504,
|
| 3033 |
+
"learning_rate": 0.00015597925264196049,
|
| 3034 |
+
"loss": 1.4256,
|
| 3035 |
+
"step": 430
|
| 3036 |
+
},
|
| 3037 |
+
{
|
| 3038 |
+
"epoch": 0.08451809000882439,
|
| 3039 |
+
"grad_norm": 7.871713161468506,
|
| 3040 |
+
"learning_rate": 0.0001557864125868019,
|
| 3041 |
+
"loss": 2.3536,
|
| 3042 |
+
"step": 431
|
| 3043 |
+
},
|
| 3044 |
+
{
|
| 3045 |
+
"epoch": 0.0847141876654574,
|
| 3046 |
+
"grad_norm": 5.0042033195495605,
|
| 3047 |
+
"learning_rate": 0.00015559327087194942,
|
| 3048 |
+
"loss": 3.0954,
|
| 3049 |
+
"step": 432
|
| 3050 |
+
},
|
| 3051 |
+
{
|
| 3052 |
+
"epoch": 0.0849102853220904,
|
| 3053 |
+
"grad_norm": 6.770071506500244,
|
| 3054 |
+
"learning_rate": 0.0001553998285417983,
|
| 3055 |
+
"loss": 1.5016,
|
| 3056 |
+
"step": 433
|
| 3057 |
+
},
|
| 3058 |
+
{
|
| 3059 |
+
"epoch": 0.0851063829787234,
|
| 3060 |
+
"grad_norm": 6.884315490722656,
|
| 3061 |
+
"learning_rate": 0.00015520608664236949,
|
| 3062 |
+
"loss": 0.9995,
|
| 3063 |
+
"step": 434
|
| 3064 |
+
},
|
| 3065 |
+
{
|
| 3066 |
+
"epoch": 0.0853024806353564,
|
| 3067 |
+
"grad_norm": 7.716054439544678,
|
| 3068 |
+
"learning_rate": 0.00015501204622130377,
|
| 3069 |
+
"loss": 1.525,
|
| 3070 |
+
"step": 435
|
| 3071 |
+
},
|
| 3072 |
+
{
|
| 3073 |
+
"epoch": 0.08549857829198941,
|
| 3074 |
+
"grad_norm": 9.22014045715332,
|
| 3075 |
+
"learning_rate": 0.0001548177083278562,
|
| 3076 |
+
"loss": 1.7368,
|
| 3077 |
+
"step": 436
|
| 3078 |
+
},
|
| 3079 |
+
{
|
| 3080 |
+
"epoch": 0.08569467594862241,
|
| 3081 |
+
"grad_norm": 11.081851959228516,
|
| 3082 |
+
"learning_rate": 0.0001546230740128904,
|
| 3083 |
+
"loss": 2.5292,
|
| 3084 |
+
"step": 437
|
| 3085 |
+
},
|
| 3086 |
+
{
|
| 3087 |
+
"epoch": 0.08589077360525542,
|
| 3088 |
+
"grad_norm": 3.8471319675445557,
|
| 3089 |
+
"learning_rate": 0.0001544281443288728,
|
| 3090 |
+
"loss": 1.8473,
|
| 3091 |
+
"step": 438
|
| 3092 |
+
},
|
| 3093 |
+
{
|
| 3094 |
+
"epoch": 0.08608687126188842,
|
| 3095 |
+
"grad_norm": 7.554915904998779,
|
| 3096 |
+
"learning_rate": 0.00015423292032986695,
|
| 3097 |
+
"loss": 1.4886,
|
| 3098 |
+
"step": 439
|
| 3099 |
+
},
|
| 3100 |
+
{
|
| 3101 |
+
"epoch": 0.08628296891852143,
|
| 3102 |
+
"grad_norm": 5.459715843200684,
|
| 3103 |
+
"learning_rate": 0.00015403740307152805,
|
| 3104 |
+
"loss": 2.7721,
|
| 3105 |
+
"step": 440
|
| 3106 |
+
},
|
| 3107 |
+
{
|
| 3108 |
+
"epoch": 0.08647906657515443,
|
| 3109 |
+
"grad_norm": 5.094114780426025,
|
| 3110 |
+
"learning_rate": 0.00015384159361109698,
|
| 3111 |
+
"loss": 0.6022,
|
| 3112 |
+
"step": 441
|
| 3113 |
+
},
|
| 3114 |
+
{
|
| 3115 |
+
"epoch": 0.08667516423178744,
|
| 3116 |
+
"grad_norm": 4.084317684173584,
|
| 3117 |
+
"learning_rate": 0.00015364549300739467,
|
| 3118 |
+
"loss": 0.8871,
|
| 3119 |
+
"step": 442
|
| 3120 |
+
},
|
| 3121 |
+
{
|
| 3122 |
+
"epoch": 0.08687126188842044,
|
| 3123 |
+
"grad_norm": 3.810145378112793,
|
| 3124 |
+
"learning_rate": 0.0001534491023208164,
|
| 3125 |
+
"loss": 2.2058,
|
| 3126 |
+
"step": 443
|
| 3127 |
+
},
|
| 3128 |
+
{
|
| 3129 |
+
"epoch": 0.08706735954505343,
|
| 3130 |
+
"grad_norm": 9.405007362365723,
|
| 3131 |
+
"learning_rate": 0.000153252422613326,
|
| 3132 |
+
"loss": 2.414,
|
| 3133 |
+
"step": 444
|
| 3134 |
+
},
|
| 3135 |
+
{
|
| 3136 |
+
"epoch": 0.08726345720168643,
|
| 3137 |
+
"grad_norm": 8.687947273254395,
|
| 3138 |
+
"learning_rate": 0.00015305545494845023,
|
| 3139 |
+
"loss": 0.8809,
|
| 3140 |
+
"step": 445
|
| 3141 |
+
},
|
| 3142 |
+
{
|
| 3143 |
+
"epoch": 0.08745955485831944,
|
| 3144 |
+
"grad_norm": 6.396414756774902,
|
| 3145 |
+
"learning_rate": 0.00015285820039127293,
|
| 3146 |
+
"loss": 1.488,
|
| 3147 |
+
"step": 446
|
| 3148 |
+
},
|
| 3149 |
+
{
|
| 3150 |
+
"epoch": 0.08765565251495244,
|
| 3151 |
+
"grad_norm": 4.447728157043457,
|
| 3152 |
+
"learning_rate": 0.00015266066000842917,
|
| 3153 |
+
"loss": 1.1921,
|
| 3154 |
+
"step": 447
|
| 3155 |
+
},
|
| 3156 |
+
{
|
| 3157 |
+
"epoch": 0.08785175017158545,
|
| 3158 |
+
"grad_norm": 9.093881607055664,
|
| 3159 |
+
"learning_rate": 0.00015246283486809977,
|
| 3160 |
+
"loss": 1.6693,
|
| 3161 |
+
"step": 448
|
| 3162 |
+
},
|
| 3163 |
+
{
|
| 3164 |
+
"epoch": 0.08804784782821845,
|
| 3165 |
+
"grad_norm": 8.160088539123535,
|
| 3166 |
+
"learning_rate": 0.0001522647260400053,
|
| 3167 |
+
"loss": 2.9597,
|
| 3168 |
+
"step": 449
|
| 3169 |
+
},
|
| 3170 |
+
{
|
| 3171 |
+
"epoch": 0.08824394548485145,
|
| 3172 |
+
"grad_norm": 4.759898662567139,
|
| 3173 |
+
"learning_rate": 0.00015206633459540023,
|
| 3174 |
+
"loss": 1.121,
|
| 3175 |
+
"step": 450
|
| 3176 |
+
},
|
| 3177 |
+
{
|
| 3178 |
+
"epoch": 0.08844004314148446,
|
| 3179 |
+
"grad_norm": 6.537527561187744,
|
| 3180 |
+
"learning_rate": 0.0001518676616070674,
|
| 3181 |
+
"loss": 1.1842,
|
| 3182 |
+
"step": 451
|
| 3183 |
+
},
|
| 3184 |
+
{
|
| 3185 |
+
"epoch": 0.08863614079811746,
|
| 3186 |
+
"grad_norm": 3.6689343452453613,
|
| 3187 |
+
"learning_rate": 0.0001516687081493119,
|
| 3188 |
+
"loss": 1.3464,
|
| 3189 |
+
"step": 452
|
| 3190 |
+
},
|
| 3191 |
+
{
|
| 3192 |
+
"epoch": 0.08883223845475047,
|
| 3193 |
+
"grad_norm": 13.292744636535645,
|
| 3194 |
+
"learning_rate": 0.00015146947529795567,
|
| 3195 |
+
"loss": 2.8284,
|
| 3196 |
+
"step": 453
|
| 3197 |
+
},
|
| 3198 |
+
{
|
| 3199 |
+
"epoch": 0.08902833611138347,
|
| 3200 |
+
"grad_norm": 3.7040257453918457,
|
| 3201 |
+
"learning_rate": 0.0001512699641303312,
|
| 3202 |
+
"loss": 1.0223,
|
| 3203 |
+
"step": 454
|
| 3204 |
+
},
|
| 3205 |
+
{
|
| 3206 |
+
"epoch": 0.08922443376801648,
|
| 3207 |
+
"grad_norm": 8.529545783996582,
|
| 3208 |
+
"learning_rate": 0.00015107017572527616,
|
| 3209 |
+
"loss": 1.9594,
|
| 3210 |
+
"step": 455
|
| 3211 |
+
},
|
| 3212 |
+
{
|
| 3213 |
+
"epoch": 0.08942053142464948,
|
| 3214 |
+
"grad_norm": 11.990189552307129,
|
| 3215 |
+
"learning_rate": 0.00015087011116312718,
|
| 3216 |
+
"loss": 1.2631,
|
| 3217 |
+
"step": 456
|
| 3218 |
+
},
|
| 3219 |
+
{
|
| 3220 |
+
"epoch": 0.08961662908128248,
|
| 3221 |
+
"grad_norm": 5.5952534675598145,
|
| 3222 |
+
"learning_rate": 0.00015066977152571428,
|
| 3223 |
+
"loss": 0.9935,
|
| 3224 |
+
"step": 457
|
| 3225 |
+
},
|
| 3226 |
+
{
|
| 3227 |
+
"epoch": 0.08981272673791549,
|
| 3228 |
+
"grad_norm": 6.168736457824707,
|
| 3229 |
+
"learning_rate": 0.0001504691578963549,
|
| 3230 |
+
"loss": 0.9699,
|
| 3231 |
+
"step": 458
|
| 3232 |
+
},
|
| 3233 |
+
{
|
| 3234 |
+
"epoch": 0.09000882439454848,
|
| 3235 |
+
"grad_norm": 6.674266338348389,
|
| 3236 |
+
"learning_rate": 0.0001502682713598481,
|
| 3237 |
+
"loss": 1.3238,
|
| 3238 |
+
"step": 459
|
| 3239 |
+
},
|
| 3240 |
+
{
|
| 3241 |
+
"epoch": 0.09020492205118148,
|
| 3242 |
+
"grad_norm": 7.915739059448242,
|
| 3243 |
+
"learning_rate": 0.00015006711300246853,
|
| 3244 |
+
"loss": 1.4752,
|
| 3245 |
+
"step": 460
|
| 3246 |
+
},
|
| 3247 |
+
{
|
| 3248 |
+
"epoch": 0.09040101970781449,
|
| 3249 |
+
"grad_norm": 9.858468055725098,
|
| 3250 |
+
"learning_rate": 0.00014986568391196092,
|
| 3251 |
+
"loss": 1.7595,
|
| 3252 |
+
"step": 461
|
| 3253 |
+
},
|
| 3254 |
+
{
|
| 3255 |
+
"epoch": 0.09059711736444749,
|
| 3256 |
+
"grad_norm": 7.993055820465088,
|
| 3257 |
+
"learning_rate": 0.0001496639851775337,
|
| 3258 |
+
"loss": 1.6935,
|
| 3259 |
+
"step": 462
|
| 3260 |
+
},
|
| 3261 |
+
{
|
| 3262 |
+
"epoch": 0.0907932150210805,
|
| 3263 |
+
"grad_norm": 8.133615493774414,
|
| 3264 |
+
"learning_rate": 0.00014946201788985358,
|
| 3265 |
+
"loss": 1.3022,
|
| 3266 |
+
"step": 463
|
| 3267 |
+
},
|
| 3268 |
+
{
|
| 3269 |
+
"epoch": 0.0909893126777135,
|
| 3270 |
+
"grad_norm": 23.21516990661621,
|
| 3271 |
+
"learning_rate": 0.0001492597831410393,
|
| 3272 |
+
"loss": 1.9148,
|
| 3273 |
+
"step": 464
|
| 3274 |
+
},
|
| 3275 |
+
{
|
| 3276 |
+
"epoch": 0.0911854103343465,
|
| 3277 |
+
"grad_norm": 5.873632431030273,
|
| 3278 |
+
"learning_rate": 0.00014905728202465595,
|
| 3279 |
+
"loss": 0.8053,
|
| 3280 |
+
"step": 465
|
| 3281 |
+
},
|
| 3282 |
+
{
|
| 3283 |
+
"epoch": 0.09138150799097951,
|
| 3284 |
+
"grad_norm": 9.700281143188477,
|
| 3285 |
+
"learning_rate": 0.000148854515635709,
|
| 3286 |
+
"loss": 1.8187,
|
| 3287 |
+
"step": 466
|
| 3288 |
+
},
|
| 3289 |
+
{
|
| 3290 |
+
"epoch": 0.09157760564761251,
|
| 3291 |
+
"grad_norm": 9.361727714538574,
|
| 3292 |
+
"learning_rate": 0.00014865148507063833,
|
| 3293 |
+
"loss": 1.5871,
|
| 3294 |
+
"step": 467
|
| 3295 |
+
},
|
| 3296 |
+
{
|
| 3297 |
+
"epoch": 0.09177370330424552,
|
| 3298 |
+
"grad_norm": 6.021685600280762,
|
| 3299 |
+
"learning_rate": 0.00014844819142731223,
|
| 3300 |
+
"loss": 0.6632,
|
| 3301 |
+
"step": 468
|
| 3302 |
+
},
|
| 3303 |
+
{
|
| 3304 |
+
"epoch": 0.09196980096087852,
|
| 3305 |
+
"grad_norm": 8.407185554504395,
|
| 3306 |
+
"learning_rate": 0.0001482446358050217,
|
| 3307 |
+
"loss": 2.1806,
|
| 3308 |
+
"step": 469
|
| 3309 |
+
},
|
| 3310 |
+
{
|
| 3311 |
+
"epoch": 0.09216589861751152,
|
| 3312 |
+
"grad_norm": 6.5181450843811035,
|
| 3313 |
+
"learning_rate": 0.00014804081930447433,
|
| 3314 |
+
"loss": 1.1511,
|
| 3315 |
+
"step": 470
|
| 3316 |
+
},
|
| 3317 |
+
{
|
| 3318 |
+
"epoch": 0.09236199627414453,
|
| 3319 |
+
"grad_norm": 6.047727584838867,
|
| 3320 |
+
"learning_rate": 0.00014783674302778832,
|
| 3321 |
+
"loss": 1.4668,
|
| 3322 |
+
"step": 471
|
| 3323 |
+
},
|
| 3324 |
+
{
|
| 3325 |
+
"epoch": 0.09255809393077753,
|
| 3326 |
+
"grad_norm": 5.636353969573975,
|
| 3327 |
+
"learning_rate": 0.00014763240807848666,
|
| 3328 |
+
"loss": 1.2796,
|
| 3329 |
+
"step": 472
|
| 3330 |
+
},
|
| 3331 |
+
{
|
| 3332 |
+
"epoch": 0.09275419158741054,
|
| 3333 |
+
"grad_norm": 10.02710247039795,
|
| 3334 |
+
"learning_rate": 0.00014742781556149103,
|
| 3335 |
+
"loss": 1.3822,
|
| 3336 |
+
"step": 473
|
| 3337 |
+
},
|
| 3338 |
+
{
|
| 3339 |
+
"epoch": 0.09295028924404353,
|
| 3340 |
+
"grad_norm": 5.736359596252441,
|
| 3341 |
+
"learning_rate": 0.00014722296658311595,
|
| 3342 |
+
"loss": 1.6941,
|
| 3343 |
+
"step": 474
|
| 3344 |
+
},
|
| 3345 |
+
{
|
| 3346 |
+
"epoch": 0.09314638690067653,
|
| 3347 |
+
"grad_norm": 4.80971622467041,
|
| 3348 |
+
"learning_rate": 0.0001470178622510627,
|
| 3349 |
+
"loss": 2.3505,
|
| 3350 |
+
"step": 475
|
| 3351 |
+
},
|
| 3352 |
+
{
|
| 3353 |
+
"epoch": 0.09334248455730954,
|
| 3354 |
+
"grad_norm": 6.357244491577148,
|
| 3355 |
+
"learning_rate": 0.00014681250367441328,
|
| 3356 |
+
"loss": 3.0606,
|
| 3357 |
+
"step": 476
|
| 3358 |
+
},
|
| 3359 |
+
{
|
| 3360 |
+
"epoch": 0.09353858221394254,
|
| 3361 |
+
"grad_norm": 6.115423202514648,
|
| 3362 |
+
"learning_rate": 0.00014660689196362462,
|
| 3363 |
+
"loss": 1.4211,
|
| 3364 |
+
"step": 477
|
| 3365 |
+
},
|
| 3366 |
+
{
|
| 3367 |
+
"epoch": 0.09373467987057554,
|
| 3368 |
+
"grad_norm": 10.913143157958984,
|
| 3369 |
+
"learning_rate": 0.0001464010282305224,
|
| 3370 |
+
"loss": 1.6531,
|
| 3371 |
+
"step": 478
|
| 3372 |
+
},
|
| 3373 |
+
{
|
| 3374 |
+
"epoch": 0.09393077752720855,
|
| 3375 |
+
"grad_norm": 7.621870517730713,
|
| 3376 |
+
"learning_rate": 0.000146194913588295,
|
| 3377 |
+
"loss": 1.964,
|
| 3378 |
+
"step": 479
|
| 3379 |
+
},
|
| 3380 |
+
{
|
| 3381 |
+
"epoch": 0.09412687518384155,
|
| 3382 |
+
"grad_norm": 11.227121353149414,
|
| 3383 |
+
"learning_rate": 0.0001459885491514878,
|
| 3384 |
+
"loss": 1.639,
|
| 3385 |
+
"step": 480
|
| 3386 |
+
},
|
| 3387 |
+
{
|
| 3388 |
+
"epoch": 0.09432297284047456,
|
| 3389 |
+
"grad_norm": 4.943283557891846,
|
| 3390 |
+
"learning_rate": 0.00014578193603599662,
|
| 3391 |
+
"loss": 0.5043,
|
| 3392 |
+
"step": 481
|
| 3393 |
+
},
|
| 3394 |
+
{
|
| 3395 |
+
"epoch": 0.09451907049710756,
|
| 3396 |
+
"grad_norm": 13.006143569946289,
|
| 3397 |
+
"learning_rate": 0.0001455750753590622,
|
| 3398 |
+
"loss": 3.0382,
|
| 3399 |
+
"step": 482
|
| 3400 |
+
},
|
| 3401 |
+
{
|
| 3402 |
+
"epoch": 0.09471516815374056,
|
| 3403 |
+
"grad_norm": 6.181257247924805,
|
| 3404 |
+
"learning_rate": 0.00014536796823926386,
|
| 3405 |
+
"loss": 1.3281,
|
| 3406 |
+
"step": 483
|
| 3407 |
+
},
|
| 3408 |
+
{
|
| 3409 |
+
"epoch": 0.09491126581037357,
|
| 3410 |
+
"grad_norm": 4.05950403213501,
|
| 3411 |
+
"learning_rate": 0.0001451606157965136,
|
| 3412 |
+
"loss": 1.5701,
|
| 3413 |
+
"step": 484
|
| 3414 |
+
},
|
| 3415 |
+
{
|
| 3416 |
+
"epoch": 0.09510736346700657,
|
| 3417 |
+
"grad_norm": 7.532302379608154,
|
| 3418 |
+
"learning_rate": 0.0001449530191520499,
|
| 3419 |
+
"loss": 1.811,
|
| 3420 |
+
"step": 485
|
| 3421 |
+
},
|
| 3422 |
+
{
|
| 3423 |
+
"epoch": 0.09530346112363958,
|
| 3424 |
+
"grad_norm": 7.403939247131348,
|
| 3425 |
+
"learning_rate": 0.00014474517942843175,
|
| 3426 |
+
"loss": 1.8269,
|
| 3427 |
+
"step": 486
|
| 3428 |
+
},
|
| 3429 |
+
{
|
| 3430 |
+
"epoch": 0.09549955878027258,
|
| 3431 |
+
"grad_norm": 18.15042495727539,
|
| 3432 |
+
"learning_rate": 0.0001445370977495326,
|
| 3433 |
+
"loss": 2.4586,
|
| 3434 |
+
"step": 487
|
| 3435 |
+
},
|
| 3436 |
+
{
|
| 3437 |
+
"epoch": 0.09569565643690559,
|
| 3438 |
+
"grad_norm": 13.836175918579102,
|
| 3439 |
+
"learning_rate": 0.00014432877524053427,
|
| 3440 |
+
"loss": 1.7828,
|
| 3441 |
+
"step": 488
|
| 3442 |
+
},
|
| 3443 |
+
{
|
| 3444 |
+
"epoch": 0.09589175409353858,
|
| 3445 |
+
"grad_norm": 5.9050822257995605,
|
| 3446 |
+
"learning_rate": 0.0001441202130279208,
|
| 3447 |
+
"loss": 2.7068,
|
| 3448 |
+
"step": 489
|
| 3449 |
+
},
|
| 3450 |
+
{
|
| 3451 |
+
"epoch": 0.09608785175017158,
|
| 3452 |
+
"grad_norm": 7.173941612243652,
|
| 3453 |
+
"learning_rate": 0.0001439114122394724,
|
| 3454 |
+
"loss": 2.7424,
|
| 3455 |
+
"step": 490
|
| 3456 |
+
},
|
| 3457 |
+
{
|
| 3458 |
+
"epoch": 0.09628394940680458,
|
| 3459 |
+
"grad_norm": 17.888187408447266,
|
| 3460 |
+
"learning_rate": 0.0001437023740042594,
|
| 3461 |
+
"loss": 2.0253,
|
| 3462 |
+
"step": 491
|
| 3463 |
+
},
|
| 3464 |
+
{
|
| 3465 |
+
"epoch": 0.09648004706343759,
|
| 3466 |
+
"grad_norm": 6.492856025695801,
|
| 3467 |
+
"learning_rate": 0.00014349309945263606,
|
| 3468 |
+
"loss": 2.1604,
|
| 3469 |
+
"step": 492
|
| 3470 |
+
},
|
| 3471 |
+
{
|
| 3472 |
+
"epoch": 0.09667614472007059,
|
| 3473 |
+
"grad_norm": 11.6004638671875,
|
| 3474 |
+
"learning_rate": 0.00014328358971623455,
|
| 3475 |
+
"loss": 2.1492,
|
| 3476 |
+
"step": 493
|
| 3477 |
+
},
|
| 3478 |
+
{
|
| 3479 |
+
"epoch": 0.0968722423767036,
|
| 3480 |
+
"grad_norm": 6.262972354888916,
|
| 3481 |
+
"learning_rate": 0.00014307384592795872,
|
| 3482 |
+
"loss": 2.3007,
|
| 3483 |
+
"step": 494
|
| 3484 |
+
},
|
| 3485 |
+
{
|
| 3486 |
+
"epoch": 0.0970683400333366,
|
| 3487 |
+
"grad_norm": 6.768704891204834,
|
| 3488 |
+
"learning_rate": 0.00014286386922197805,
|
| 3489 |
+
"loss": 0.6546,
|
| 3490 |
+
"step": 495
|
| 3491 |
+
},
|
| 3492 |
+
{
|
| 3493 |
+
"epoch": 0.0972644376899696,
|
| 3494 |
+
"grad_norm": 4.372698783874512,
|
| 3495 |
+
"learning_rate": 0.0001426536607337215,
|
| 3496 |
+
"loss": 2.3647,
|
| 3497 |
+
"step": 496
|
| 3498 |
+
},
|
| 3499 |
+
{
|
| 3500 |
+
"epoch": 0.09746053534660261,
|
| 3501 |
+
"grad_norm": 12.042864799499512,
|
| 3502 |
+
"learning_rate": 0.00014244322159987145,
|
| 3503 |
+
"loss": 2.1466,
|
| 3504 |
+
"step": 497
|
| 3505 |
+
},
|
| 3506 |
+
{
|
| 3507 |
+
"epoch": 0.09765663300323561,
|
| 3508 |
+
"grad_norm": 5.374109745025635,
|
| 3509 |
+
"learning_rate": 0.0001422325529583573,
|
| 3510 |
+
"loss": 0.8599,
|
| 3511 |
+
"step": 498
|
| 3512 |
+
},
|
| 3513 |
+
{
|
| 3514 |
+
"epoch": 0.09785273065986862,
|
| 3515 |
+
"grad_norm": 7.459904193878174,
|
| 3516 |
+
"learning_rate": 0.00014202165594834963,
|
| 3517 |
+
"loss": 1.7119,
|
| 3518 |
+
"step": 499
|
| 3519 |
+
},
|
| 3520 |
+
{
|
| 3521 |
+
"epoch": 0.09804882831650162,
|
| 3522 |
+
"grad_norm": 8.25338363647461,
|
| 3523 |
+
"learning_rate": 0.00014181053171025392,
|
| 3524 |
+
"loss": 1.3014,
|
| 3525 |
+
"step": 500
|
| 3526 |
+
},
|
| 3527 |
+
{
|
| 3528 |
+
"epoch": 0.09824492597313463,
|
| 3529 |
+
"grad_norm": 8.594246864318848,
|
| 3530 |
+
"learning_rate": 0.00014159918138570424,
|
| 3531 |
+
"loss": 1.9332,
|
| 3532 |
+
"step": 501
|
| 3533 |
+
},
|
| 3534 |
+
{
|
| 3535 |
+
"epoch": 0.09844102362976763,
|
| 3536 |
+
"grad_norm": 4.802338600158691,
|
| 3537 |
+
"learning_rate": 0.00014138760611755727,
|
| 3538 |
+
"loss": 1.6591,
|
| 3539 |
+
"step": 502
|
| 3540 |
+
},
|
| 3541 |
+
{
|
| 3542 |
+
"epoch": 0.09863712128640063,
|
| 3543 |
+
"grad_norm": 5.049933433532715,
|
| 3544 |
+
"learning_rate": 0.00014117580704988612,
|
| 3545 |
+
"loss": 1.3722,
|
| 3546 |
+
"step": 503
|
| 3547 |
+
},
|
| 3548 |
+
{
|
| 3549 |
+
"epoch": 0.09883321894303362,
|
| 3550 |
+
"grad_norm": 9.438610076904297,
|
| 3551 |
+
"learning_rate": 0.00014096378532797393,
|
| 3552 |
+
"loss": 2.1888,
|
| 3553 |
+
"step": 504
|
| 3554 |
+
},
|
| 3555 |
+
{
|
| 3556 |
+
"epoch": 0.09902931659966663,
|
| 3557 |
+
"grad_norm": 11.144261360168457,
|
| 3558 |
+
"learning_rate": 0.00014075154209830792,
|
| 3559 |
+
"loss": 1.2526,
|
| 3560 |
+
"step": 505
|
| 3561 |
+
},
|
| 3562 |
+
{
|
| 3563 |
+
"epoch": 0.09922541425629963,
|
| 3564 |
+
"grad_norm": 12.456216812133789,
|
| 3565 |
+
"learning_rate": 0.0001405390785085731,
|
| 3566 |
+
"loss": 2.8251,
|
| 3567 |
+
"step": 506
|
| 3568 |
+
},
|
| 3569 |
+
{
|
| 3570 |
+
"epoch": 0.09942151191293264,
|
| 3571 |
+
"grad_norm": 13.78912353515625,
|
| 3572 |
+
"learning_rate": 0.00014032639570764593,
|
| 3573 |
+
"loss": 2.3312,
|
| 3574 |
+
"step": 507
|
| 3575 |
+
},
|
| 3576 |
+
{
|
| 3577 |
+
"epoch": 0.09961760956956564,
|
| 3578 |
+
"grad_norm": 22.588031768798828,
|
| 3579 |
+
"learning_rate": 0.00014011349484558847,
|
| 3580 |
+
"loss": 1.8865,
|
| 3581 |
+
"step": 508
|
| 3582 |
+
},
|
| 3583 |
+
{
|
| 3584 |
+
"epoch": 0.09981370722619864,
|
| 3585 |
+
"grad_norm": 5.357258319854736,
|
| 3586 |
+
"learning_rate": 0.00013990037707364166,
|
| 3587 |
+
"loss": 2.1296,
|
| 3588 |
+
"step": 509
|
| 3589 |
+
},
|
| 3590 |
+
{
|
| 3591 |
+
"epoch": 0.10000980488283165,
|
| 3592 |
+
"grad_norm": 13.293752670288086,
|
| 3593 |
+
"learning_rate": 0.00013968704354421952,
|
| 3594 |
+
"loss": 2.0479,
|
| 3595 |
+
"step": 510
|
| 3596 |
+
},
|
| 3597 |
+
{
|
| 3598 |
+
"epoch": 0.10020590253946465,
|
| 3599 |
+
"grad_norm": 6.764191150665283,
|
| 3600 |
+
"learning_rate": 0.00013947349541090274,
|
| 3601 |
+
"loss": 1.3908,
|
| 3602 |
+
"step": 511
|
| 3603 |
+
},
|
| 3604 |
+
{
|
| 3605 |
+
"epoch": 0.10040200019609766,
|
| 3606 |
+
"grad_norm": 5.701320648193359,
|
| 3607 |
+
"learning_rate": 0.00013925973382843246,
|
| 3608 |
+
"loss": 1.1343,
|
| 3609 |
+
"step": 512
|
| 3610 |
+
},
|
| 3611 |
+
{
|
| 3612 |
+
"epoch": 0.10059809785273066,
|
| 3613 |
+
"grad_norm": 6.728299140930176,
|
| 3614 |
+
"learning_rate": 0.0001390457599527039,
|
| 3615 |
+
"loss": 1.3219,
|
| 3616 |
+
"step": 513
|
| 3617 |
+
},
|
| 3618 |
+
{
|
| 3619 |
+
"epoch": 0.10079419550936367,
|
| 3620 |
+
"grad_norm": 9.094908714294434,
|
| 3621 |
+
"learning_rate": 0.00013883157494076046,
|
| 3622 |
+
"loss": 1.1282,
|
| 3623 |
+
"step": 514
|
| 3624 |
+
},
|
| 3625 |
+
{
|
| 3626 |
+
"epoch": 0.10099029316599667,
|
| 3627 |
+
"grad_norm": 7.321430683135986,
|
| 3628 |
+
"learning_rate": 0.00013861717995078708,
|
| 3629 |
+
"loss": 1.0628,
|
| 3630 |
+
"step": 515
|
| 3631 |
+
},
|
| 3632 |
+
{
|
| 3633 |
+
"epoch": 0.10118639082262967,
|
| 3634 |
+
"grad_norm": 5.976141452789307,
|
| 3635 |
+
"learning_rate": 0.00013840257614210414,
|
| 3636 |
+
"loss": 1.5372,
|
| 3637 |
+
"step": 516
|
| 3638 |
+
},
|
| 3639 |
+
{
|
| 3640 |
+
"epoch": 0.10138248847926268,
|
| 3641 |
+
"grad_norm": 7.5356926918029785,
|
| 3642 |
+
"learning_rate": 0.00013818776467516125,
|
| 3643 |
+
"loss": 2.0415,
|
| 3644 |
+
"step": 517
|
| 3645 |
+
},
|
| 3646 |
+
{
|
| 3647 |
+
"epoch": 0.10157858613589568,
|
| 3648 |
+
"grad_norm": 5.656033992767334,
|
| 3649 |
+
"learning_rate": 0.00013797274671153092,
|
| 3650 |
+
"loss": 1.9171,
|
| 3651 |
+
"step": 518
|
| 3652 |
+
},
|
| 3653 |
+
{
|
| 3654 |
+
"epoch": 0.10177468379252867,
|
| 3655 |
+
"grad_norm": 4.822179317474365,
|
| 3656 |
+
"learning_rate": 0.0001377575234139022,
|
| 3657 |
+
"loss": 1.1379,
|
| 3658 |
+
"step": 519
|
| 3659 |
+
},
|
| 3660 |
+
{
|
| 3661 |
+
"epoch": 0.10197078144916168,
|
| 3662 |
+
"grad_norm": 2.5650954246520996,
|
| 3663 |
+
"learning_rate": 0.0001375420959460745,
|
| 3664 |
+
"loss": 0.3984,
|
| 3665 |
+
"step": 520
|
| 3666 |
+
},
|
| 3667 |
+
{
|
| 3668 |
+
"epoch": 0.10216687910579468,
|
| 3669 |
+
"grad_norm": 6.997866630554199,
|
| 3670 |
+
"learning_rate": 0.00013732646547295126,
|
| 3671 |
+
"loss": 1.2149,
|
| 3672 |
+
"step": 521
|
| 3673 |
+
},
|
| 3674 |
+
{
|
| 3675 |
+
"epoch": 0.10236297676242768,
|
| 3676 |
+
"grad_norm": 9.925884246826172,
|
| 3677 |
+
"learning_rate": 0.00013711063316053368,
|
| 3678 |
+
"loss": 1.9542,
|
| 3679 |
+
"step": 522
|
| 3680 |
+
},
|
| 3681 |
+
{
|
| 3682 |
+
"epoch": 0.10255907441906069,
|
| 3683 |
+
"grad_norm": 7.913006782531738,
|
| 3684 |
+
"learning_rate": 0.00013689460017591432,
|
| 3685 |
+
"loss": 1.3162,
|
| 3686 |
+
"step": 523
|
| 3687 |
+
},
|
| 3688 |
+
{
|
| 3689 |
+
"epoch": 0.1027551720756937,
|
| 3690 |
+
"grad_norm": 8.868382453918457,
|
| 3691 |
+
"learning_rate": 0.00013667836768727092,
|
| 3692 |
+
"loss": 2.8749,
|
| 3693 |
+
"step": 524
|
| 3694 |
+
},
|
| 3695 |
+
{
|
| 3696 |
+
"epoch": 0.1029512697323267,
|
| 3697 |
+
"grad_norm": 7.418759346008301,
|
| 3698 |
+
"learning_rate": 0.00013646193686386,
|
| 3699 |
+
"loss": 1.2697,
|
| 3700 |
+
"step": 525
|
| 3701 |
+
},
|
| 3702 |
+
{
|
| 3703 |
+
"epoch": 0.1031473673889597,
|
| 3704 |
+
"grad_norm": 9.501752853393555,
|
| 3705 |
+
"learning_rate": 0.00013624530887601055,
|
| 3706 |
+
"loss": 1.1473,
|
| 3707 |
+
"step": 526
|
| 3708 |
+
},
|
| 3709 |
+
{
|
| 3710 |
+
"epoch": 0.1033434650455927,
|
| 3711 |
+
"grad_norm": 6.225274562835693,
|
| 3712 |
+
"learning_rate": 0.0001360284848951177,
|
| 3713 |
+
"loss": 1.0762,
|
| 3714 |
+
"step": 527
|
| 3715 |
+
},
|
| 3716 |
+
{
|
| 3717 |
+
"epoch": 0.10353956270222571,
|
| 3718 |
+
"grad_norm": 6.502389907836914,
|
| 3719 |
+
"learning_rate": 0.0001358114660936364,
|
| 3720 |
+
"loss": 1.383,
|
| 3721 |
+
"step": 528
|
| 3722 |
+
},
|
| 3723 |
+
{
|
| 3724 |
+
"epoch": 0.10373566035885871,
|
| 3725 |
+
"grad_norm": 17.53512954711914,
|
| 3726 |
+
"learning_rate": 0.00013559425364507508,
|
| 3727 |
+
"loss": 1.7968,
|
| 3728 |
+
"step": 529
|
| 3729 |
+
},
|
| 3730 |
+
{
|
| 3731 |
+
"epoch": 0.10393175801549172,
|
| 3732 |
+
"grad_norm": 6.0248122215271,
|
| 3733 |
+
"learning_rate": 0.00013537684872398927,
|
| 3734 |
+
"loss": 1.2982,
|
| 3735 |
+
"step": 530
|
| 3736 |
+
},
|
| 3737 |
+
{
|
| 3738 |
+
"epoch": 0.10412785567212472,
|
| 3739 |
+
"grad_norm": 4.621517658233643,
|
| 3740 |
+
"learning_rate": 0.00013515925250597537,
|
| 3741 |
+
"loss": 1.5811,
|
| 3742 |
+
"step": 531
|
| 3743 |
+
},
|
| 3744 |
+
{
|
| 3745 |
+
"epoch": 0.10432395332875773,
|
| 3746 |
+
"grad_norm": 6.690089702606201,
|
| 3747 |
+
"learning_rate": 0.00013494146616766406,
|
| 3748 |
+
"loss": 1.3376,
|
| 3749 |
+
"step": 532
|
| 3750 |
+
},
|
| 3751 |
+
{
|
| 3752 |
+
"epoch": 0.10452005098539073,
|
| 3753 |
+
"grad_norm": 6.418732643127441,
|
| 3754 |
+
"learning_rate": 0.00013472349088671418,
|
| 3755 |
+
"loss": 1.507,
|
| 3756 |
+
"step": 533
|
| 3757 |
+
},
|
| 3758 |
+
{
|
| 3759 |
+
"epoch": 0.10471614864202373,
|
| 3760 |
+
"grad_norm": 12.271800994873047,
|
| 3761 |
+
"learning_rate": 0.0001345053278418062,
|
| 3762 |
+
"loss": 1.6329,
|
| 3763 |
+
"step": 534
|
| 3764 |
+
},
|
| 3765 |
+
{
|
| 3766 |
+
"epoch": 0.10491224629865673,
|
| 3767 |
+
"grad_norm": 10.394608497619629,
|
| 3768 |
+
"learning_rate": 0.000134286978212636,
|
| 3769 |
+
"loss": 1.8793,
|
| 3770 |
+
"step": 535
|
| 3771 |
+
},
|
| 3772 |
+
{
|
| 3773 |
+
"epoch": 0.10510834395528973,
|
| 3774 |
+
"grad_norm": 5.690524101257324,
|
| 3775 |
+
"learning_rate": 0.00013406844317990826,
|
| 3776 |
+
"loss": 2.2874,
|
| 3777 |
+
"step": 536
|
| 3778 |
+
},
|
| 3779 |
+
{
|
| 3780 |
+
"epoch": 0.10530444161192273,
|
| 3781 |
+
"grad_norm": 4.313036918640137,
|
| 3782 |
+
"learning_rate": 0.00013384972392533034,
|
| 3783 |
+
"loss": 2.2905,
|
| 3784 |
+
"step": 537
|
| 3785 |
+
},
|
| 3786 |
+
{
|
| 3787 |
+
"epoch": 0.10550053926855574,
|
| 3788 |
+
"grad_norm": 6.115211009979248,
|
| 3789 |
+
"learning_rate": 0.0001336308216316056,
|
| 3790 |
+
"loss": 2.3146,
|
| 3791 |
+
"step": 538
|
| 3792 |
+
},
|
| 3793 |
+
{
|
| 3794 |
+
"epoch": 0.10569663692518874,
|
| 3795 |
+
"grad_norm": 7.43934965133667,
|
| 3796 |
+
"learning_rate": 0.0001334117374824274,
|
| 3797 |
+
"loss": 2.5837,
|
| 3798 |
+
"step": 539
|
| 3799 |
+
},
|
| 3800 |
+
{
|
| 3801 |
+
"epoch": 0.10589273458182175,
|
| 3802 |
+
"grad_norm": 8.397297859191895,
|
| 3803 |
+
"learning_rate": 0.00013319247266247225,
|
| 3804 |
+
"loss": 2.1631,
|
| 3805 |
+
"step": 540
|
| 3806 |
+
},
|
| 3807 |
+
{
|
| 3808 |
+
"epoch": 0.10608883223845475,
|
| 3809 |
+
"grad_norm": 5.273848056793213,
|
| 3810 |
+
"learning_rate": 0.00013297302835739376,
|
| 3811 |
+
"loss": 1.261,
|
| 3812 |
+
"step": 541
|
| 3813 |
+
},
|
| 3814 |
+
{
|
| 3815 |
+
"epoch": 0.10628492989508775,
|
| 3816 |
+
"grad_norm": 5.789434909820557,
|
| 3817 |
+
"learning_rate": 0.00013275340575381598,
|
| 3818 |
+
"loss": 1.8465,
|
| 3819 |
+
"step": 542
|
| 3820 |
+
},
|
| 3821 |
+
{
|
| 3822 |
+
"epoch": 0.10648102755172076,
|
| 3823 |
+
"grad_norm": 8.769668579101562,
|
| 3824 |
+
"learning_rate": 0.0001325336060393272,
|
| 3825 |
+
"loss": 1.3243,
|
| 3826 |
+
"step": 543
|
| 3827 |
+
},
|
| 3828 |
+
{
|
| 3829 |
+
"epoch": 0.10667712520835376,
|
| 3830 |
+
"grad_norm": 7.573061943054199,
|
| 3831 |
+
"learning_rate": 0.00013231363040247334,
|
| 3832 |
+
"loss": 1.243,
|
| 3833 |
+
"step": 544
|
| 3834 |
+
},
|
| 3835 |
+
{
|
| 3836 |
+
"epoch": 0.10687322286498677,
|
| 3837 |
+
"grad_norm": 9.284170150756836,
|
| 3838 |
+
"learning_rate": 0.00013209348003275165,
|
| 3839 |
+
"loss": 1.6408,
|
| 3840 |
+
"step": 545
|
| 3841 |
+
},
|
| 3842 |
+
{
|
| 3843 |
+
"epoch": 0.10706932052161977,
|
| 3844 |
+
"grad_norm": 4.322136878967285,
|
| 3845 |
+
"learning_rate": 0.00013187315612060415,
|
| 3846 |
+
"loss": 3.2532,
|
| 3847 |
+
"step": 546
|
| 3848 |
+
},
|
| 3849 |
+
{
|
| 3850 |
+
"epoch": 0.10726541817825277,
|
| 3851 |
+
"grad_norm": 7.865033149719238,
|
| 3852 |
+
"learning_rate": 0.00013165265985741142,
|
| 3853 |
+
"loss": 3.5798,
|
| 3854 |
+
"step": 547
|
| 3855 |
+
},
|
| 3856 |
+
{
|
| 3857 |
+
"epoch": 0.10746151583488578,
|
| 3858 |
+
"grad_norm": 3.1747426986694336,
|
| 3859 |
+
"learning_rate": 0.00013143199243548587,
|
| 3860 |
+
"loss": 1.1817,
|
| 3861 |
+
"step": 548
|
| 3862 |
+
},
|
| 3863 |
+
{
|
| 3864 |
+
"epoch": 0.10765761349151878,
|
| 3865 |
+
"grad_norm": 6.890042781829834,
|
| 3866 |
+
"learning_rate": 0.00013121115504806553,
|
| 3867 |
+
"loss": 2.0318,
|
| 3868 |
+
"step": 549
|
| 3869 |
+
},
|
| 3870 |
+
{
|
| 3871 |
+
"epoch": 0.10785371114815177,
|
| 3872 |
+
"grad_norm": 7.632523059844971,
|
| 3873 |
+
"learning_rate": 0.00013099014888930748,
|
| 3874 |
+
"loss": 1.3925,
|
| 3875 |
+
"step": 550
|
| 3876 |
+
},
|
| 3877 |
+
{
|
| 3878 |
+
"epoch": 0.10804980880478478,
|
| 3879 |
+
"grad_norm": 4.939361095428467,
|
| 3880 |
+
"learning_rate": 0.00013076897515428132,
|
| 3881 |
+
"loss": 1.3661,
|
| 3882 |
+
"step": 551
|
| 3883 |
+
},
|
| 3884 |
+
{
|
| 3885 |
+
"epoch": 0.10824590646141778,
|
| 3886 |
+
"grad_norm": 6.433002948760986,
|
| 3887 |
+
"learning_rate": 0.00013054763503896294,
|
| 3888 |
+
"loss": 0.9917,
|
| 3889 |
+
"step": 552
|
| 3890 |
+
},
|
| 3891 |
+
{
|
| 3892 |
+
"epoch": 0.10844200411805079,
|
| 3893 |
+
"grad_norm": 4.507363319396973,
|
| 3894 |
+
"learning_rate": 0.00013032612974022784,
|
| 3895 |
+
"loss": 0.9992,
|
| 3896 |
+
"step": 553
|
| 3897 |
+
},
|
| 3898 |
+
{
|
| 3899 |
+
"epoch": 0.10863810177468379,
|
| 3900 |
+
"grad_norm": 9.477548599243164,
|
| 3901 |
+
"learning_rate": 0.00013010446045584479,
|
| 3902 |
+
"loss": 1.1149,
|
| 3903 |
+
"step": 554
|
| 3904 |
+
},
|
| 3905 |
+
{
|
| 3906 |
+
"epoch": 0.1088341994313168,
|
| 3907 |
+
"grad_norm": 5.909473419189453,
|
| 3908 |
+
"learning_rate": 0.00012988262838446922,
|
| 3909 |
+
"loss": 2.0005,
|
| 3910 |
+
"step": 555
|
| 3911 |
+
},
|
| 3912 |
+
{
|
| 3913 |
+
"epoch": 0.1090302970879498,
|
| 3914 |
+
"grad_norm": 8.154961585998535,
|
| 3915 |
+
"learning_rate": 0.00012966063472563685,
|
| 3916 |
+
"loss": 2.1173,
|
| 3917 |
+
"step": 556
|
| 3918 |
+
},
|
| 3919 |
+
{
|
| 3920 |
+
"epoch": 0.1092263947445828,
|
| 3921 |
+
"grad_norm": 4.8719868659973145,
|
| 3922 |
+
"learning_rate": 0.00012943848067975718,
|
| 3923 |
+
"loss": 2.925,
|
| 3924 |
+
"step": 557
|
| 3925 |
+
},
|
| 3926 |
+
{
|
| 3927 |
+
"epoch": 0.1094224924012158,
|
| 3928 |
+
"grad_norm": 5.895204544067383,
|
| 3929 |
+
"learning_rate": 0.00012921616744810698,
|
| 3930 |
+
"loss": 2.0375,
|
| 3931 |
+
"step": 558
|
| 3932 |
+
},
|
| 3933 |
+
{
|
| 3934 |
+
"epoch": 0.10961859005784881,
|
| 3935 |
+
"grad_norm": 5.050942897796631,
|
| 3936 |
+
"learning_rate": 0.0001289936962328238,
|
| 3937 |
+
"loss": 1.6125,
|
| 3938 |
+
"step": 559
|
| 3939 |
+
},
|
| 3940 |
+
{
|
| 3941 |
+
"epoch": 0.10981468771448182,
|
| 3942 |
+
"grad_norm": 5.890807151794434,
|
| 3943 |
+
"learning_rate": 0.0001287710682368995,
|
| 3944 |
+
"loss": 0.991,
|
| 3945 |
+
"step": 560
|
| 3946 |
+
},
|
| 3947 |
+
{
|
| 3948 |
+
"epoch": 0.11001078537111482,
|
| 3949 |
+
"grad_norm": 5.316340446472168,
|
| 3950 |
+
"learning_rate": 0.00012854828466417364,
|
| 3951 |
+
"loss": 1.8846,
|
| 3952 |
+
"step": 561
|
| 3953 |
+
},
|
| 3954 |
+
{
|
| 3955 |
+
"epoch": 0.11020688302774782,
|
| 3956 |
+
"grad_norm": 14.81187629699707,
|
| 3957 |
+
"learning_rate": 0.00012832534671932715,
|
| 3958 |
+
"loss": 2.141,
|
| 3959 |
+
"step": 562
|
| 3960 |
+
},
|
| 3961 |
+
{
|
| 3962 |
+
"epoch": 0.11040298068438083,
|
| 3963 |
+
"grad_norm": 11.379783630371094,
|
| 3964 |
+
"learning_rate": 0.0001281022556078756,
|
| 3965 |
+
"loss": 2.9432,
|
| 3966 |
+
"step": 563
|
| 3967 |
+
},
|
| 3968 |
+
{
|
| 3969 |
+
"epoch": 0.11059907834101383,
|
| 3970 |
+
"grad_norm": 17.28523826599121,
|
| 3971 |
+
"learning_rate": 0.0001278790125361629,
|
| 3972 |
+
"loss": 2.4028,
|
| 3973 |
+
"step": 564
|
| 3974 |
+
},
|
| 3975 |
+
{
|
| 3976 |
+
"epoch": 0.11079517599764682,
|
| 3977 |
+
"grad_norm": 6.626383304595947,
|
| 3978 |
+
"learning_rate": 0.0001276556187113546,
|
| 3979 |
+
"loss": 1.1534,
|
| 3980 |
+
"step": 565
|
| 3981 |
+
},
|
| 3982 |
+
{
|
| 3983 |
+
"epoch": 0.11099127365427983,
|
| 3984 |
+
"grad_norm": 7.6215128898620605,
|
| 3985 |
+
"learning_rate": 0.00012743207534143144,
|
| 3986 |
+
"loss": 1.6464,
|
| 3987 |
+
"step": 566
|
| 3988 |
+
},
|
| 3989 |
+
{
|
| 3990 |
+
"epoch": 0.11118737131091283,
|
| 3991 |
+
"grad_norm": 8.348710060119629,
|
| 3992 |
+
"learning_rate": 0.00012720838363518286,
|
| 3993 |
+
"loss": 1.1823,
|
| 3994 |
+
"step": 567
|
| 3995 |
+
},
|
| 3996 |
+
{
|
| 3997 |
+
"epoch": 0.11138346896754583,
|
| 3998 |
+
"grad_norm": 5.307929039001465,
|
| 3999 |
+
"learning_rate": 0.00012698454480220039,
|
| 4000 |
+
"loss": 2.3759,
|
| 4001 |
+
"step": 568
|
| 4002 |
+
},
|
| 4003 |
+
{
|
| 4004 |
+
"epoch": 0.11157956662417884,
|
| 4005 |
+
"grad_norm": 4.690614223480225,
|
| 4006 |
+
"learning_rate": 0.00012676056005287106,
|
| 4007 |
+
"loss": 1.3698,
|
| 4008 |
+
"step": 569
|
| 4009 |
+
},
|
| 4010 |
+
{
|
| 4011 |
+
"epoch": 0.11177566428081184,
|
| 4012 |
+
"grad_norm": 5.7534871101379395,
|
| 4013 |
+
"learning_rate": 0.00012653643059837107,
|
| 4014 |
+
"loss": 1.9587,
|
| 4015 |
+
"step": 570
|
| 4016 |
+
},
|
| 4017 |
+
{
|
| 4018 |
+
"epoch": 0.11197176193744485,
|
| 4019 |
+
"grad_norm": 5.088260650634766,
|
| 4020 |
+
"learning_rate": 0.00012631215765065897,
|
| 4021 |
+
"loss": 1.0195,
|
| 4022 |
+
"step": 571
|
| 4023 |
+
},
|
| 4024 |
+
{
|
| 4025 |
+
"epoch": 0.11216785959407785,
|
| 4026 |
+
"grad_norm": 5.987957000732422,
|
| 4027 |
+
"learning_rate": 0.0001260877424224693,
|
| 4028 |
+
"loss": 2.3612,
|
| 4029 |
+
"step": 572
|
| 4030 |
+
},
|
| 4031 |
+
{
|
| 4032 |
+
"epoch": 0.11236395725071086,
|
| 4033 |
+
"grad_norm": 5.895928859710693,
|
| 4034 |
+
"learning_rate": 0.000125863186127306,
|
| 4035 |
+
"loss": 2.0393,
|
| 4036 |
+
"step": 573
|
| 4037 |
+
},
|
| 4038 |
+
{
|
| 4039 |
+
"epoch": 0.11256005490734386,
|
| 4040 |
+
"grad_norm": 6.654543876647949,
|
| 4041 |
+
"learning_rate": 0.00012563848997943568,
|
| 4042 |
+
"loss": 1.7165,
|
| 4043 |
+
"step": 574
|
| 4044 |
+
},
|
| 4045 |
+
{
|
| 4046 |
+
"epoch": 0.11275615256397686,
|
| 4047 |
+
"grad_norm": 13.193924903869629,
|
| 4048 |
+
"learning_rate": 0.00012541365519388138,
|
| 4049 |
+
"loss": 1.3969,
|
| 4050 |
+
"step": 575
|
| 4051 |
+
},
|
| 4052 |
+
{
|
| 4053 |
+
"epoch": 0.11295225022060987,
|
| 4054 |
+
"grad_norm": 5.471423149108887,
|
| 4055 |
+
"learning_rate": 0.00012518868298641564,
|
| 4056 |
+
"loss": 1.091,
|
| 4057 |
+
"step": 576
|
| 4058 |
+
},
|
| 4059 |
+
{
|
| 4060 |
+
"epoch": 0.11314834787724287,
|
| 4061 |
+
"grad_norm": 9.188228607177734,
|
| 4062 |
+
"learning_rate": 0.00012496357457355422,
|
| 4063 |
+
"loss": 2.0812,
|
| 4064 |
+
"step": 577
|
| 4065 |
+
},
|
| 4066 |
+
{
|
| 4067 |
+
"epoch": 0.11334444553387588,
|
| 4068 |
+
"grad_norm": 11.622628211975098,
|
| 4069 |
+
"learning_rate": 0.00012473833117254931,
|
| 4070 |
+
"loss": 2.1117,
|
| 4071 |
+
"step": 578
|
| 4072 |
+
},
|
| 4073 |
+
{
|
| 4074 |
+
"epoch": 0.11354054319050888,
|
| 4075 |
+
"grad_norm": 4.972383975982666,
|
| 4076 |
+
"learning_rate": 0.00012451295400138314,
|
| 4077 |
+
"loss": 1.3126,
|
| 4078 |
+
"step": 579
|
| 4079 |
+
},
|
| 4080 |
+
{
|
| 4081 |
+
"epoch": 0.11373664084714187,
|
| 4082 |
+
"grad_norm": 7.87300968170166,
|
| 4083 |
+
"learning_rate": 0.0001242874442787611,
|
| 4084 |
+
"loss": 3.2112,
|
| 4085 |
+
"step": 580
|
| 4086 |
+
},
|
| 4087 |
+
{
|
| 4088 |
+
"epoch": 0.11393273850377487,
|
| 4089 |
+
"grad_norm": 5.145363807678223,
|
| 4090 |
+
"learning_rate": 0.0001240618032241056,
|
| 4091 |
+
"loss": 1.2819,
|
| 4092 |
+
"step": 581
|
| 4093 |
+
},
|
| 4094 |
+
{
|
| 4095 |
+
"epoch": 0.11412883616040788,
|
| 4096 |
+
"grad_norm": 12.539934158325195,
|
| 4097 |
+
"learning_rate": 0.00012383603205754894,
|
| 4098 |
+
"loss": 2.2262,
|
| 4099 |
+
"step": 582
|
| 4100 |
+
},
|
| 4101 |
+
{
|
| 4102 |
+
"epoch": 0.11432493381704088,
|
| 4103 |
+
"grad_norm": 5.802661895751953,
|
| 4104 |
+
"learning_rate": 0.00012361013199992726,
|
| 4105 |
+
"loss": 1.9663,
|
| 4106 |
+
"step": 583
|
| 4107 |
+
},
|
| 4108 |
+
{
|
| 4109 |
+
"epoch": 0.11452103147367389,
|
| 4110 |
+
"grad_norm": 16.773561477661133,
|
| 4111 |
+
"learning_rate": 0.0001233841042727734,
|
| 4112 |
+
"loss": 2.342,
|
| 4113 |
+
"step": 584
|
| 4114 |
+
},
|
| 4115 |
+
{
|
| 4116 |
+
"epoch": 0.11471712913030689,
|
| 4117 |
+
"grad_norm": 6.163604259490967,
|
| 4118 |
+
"learning_rate": 0.0001231579500983108,
|
| 4119 |
+
"loss": 1.3009,
|
| 4120 |
+
"step": 585
|
| 4121 |
+
},
|
| 4122 |
+
{
|
| 4123 |
+
"epoch": 0.1149132267869399,
|
| 4124 |
+
"grad_norm": 7.877258777618408,
|
| 4125 |
+
"learning_rate": 0.00012293167069944645,
|
| 4126 |
+
"loss": 1.9388,
|
| 4127 |
+
"step": 586
|
| 4128 |
+
},
|
| 4129 |
+
{
|
| 4130 |
+
"epoch": 0.1151093244435729,
|
| 4131 |
+
"grad_norm": 13.269306182861328,
|
| 4132 |
+
"learning_rate": 0.00012270526729976465,
|
| 4133 |
+
"loss": 2.3028,
|
| 4134 |
+
"step": 587
|
| 4135 |
+
},
|
| 4136 |
+
{
|
| 4137 |
+
"epoch": 0.1153054221002059,
|
| 4138 |
+
"grad_norm": 9.610801696777344,
|
| 4139 |
+
"learning_rate": 0.00012247874112352009,
|
| 4140 |
+
"loss": 1.0249,
|
| 4141 |
+
"step": 588
|
| 4142 |
+
},
|
| 4143 |
+
{
|
| 4144 |
+
"epoch": 0.11550151975683891,
|
| 4145 |
+
"grad_norm": 11.20408821105957,
|
| 4146 |
+
"learning_rate": 0.00012225209339563145,
|
| 4147 |
+
"loss": 2.2807,
|
| 4148 |
+
"step": 589
|
| 4149 |
+
},
|
| 4150 |
+
{
|
| 4151 |
+
"epoch": 0.11569761741347191,
|
| 4152 |
+
"grad_norm": 13.390026092529297,
|
| 4153 |
+
"learning_rate": 0.00012202532534167463,
|
| 4154 |
+
"loss": 1.7141,
|
| 4155 |
+
"step": 590
|
| 4156 |
+
},
|
| 4157 |
+
{
|
| 4158 |
+
"epoch": 0.11589371507010492,
|
| 4159 |
+
"grad_norm": 4.4530229568481445,
|
| 4160 |
+
"learning_rate": 0.00012179843818787624,
|
| 4161 |
+
"loss": 1.4108,
|
| 4162 |
+
"step": 591
|
| 4163 |
+
},
|
| 4164 |
+
{
|
| 4165 |
+
"epoch": 0.11608981272673792,
|
| 4166 |
+
"grad_norm": 12.671486854553223,
|
| 4167 |
+
"learning_rate": 0.00012157143316110684,
|
| 4168 |
+
"loss": 1.6154,
|
| 4169 |
+
"step": 592
|
| 4170 |
+
},
|
| 4171 |
+
{
|
| 4172 |
+
"epoch": 0.11628591038337092,
|
| 4173 |
+
"grad_norm": 6.195436477661133,
|
| 4174 |
+
"learning_rate": 0.00012134431148887442,
|
| 4175 |
+
"loss": 3.4713,
|
| 4176 |
+
"step": 593
|
| 4177 |
+
},
|
| 4178 |
+
{
|
| 4179 |
+
"epoch": 0.11648200804000393,
|
| 4180 |
+
"grad_norm": 5.860586643218994,
|
| 4181 |
+
"learning_rate": 0.00012111707439931776,
|
| 4182 |
+
"loss": 2.0938,
|
| 4183 |
+
"step": 594
|
| 4184 |
+
},
|
| 4185 |
+
{
|
| 4186 |
+
"epoch": 0.11667810569663692,
|
| 4187 |
+
"grad_norm": 7.4623260498046875,
|
| 4188 |
+
"learning_rate": 0.00012088972312119964,
|
| 4189 |
+
"loss": 1.9674,
|
| 4190 |
+
"step": 595
|
| 4191 |
+
},
|
| 4192 |
+
{
|
| 4193 |
+
"epoch": 0.11687420335326992,
|
| 4194 |
+
"grad_norm": 7.368410110473633,
|
| 4195 |
+
"learning_rate": 0.00012066225888390042,
|
| 4196 |
+
"loss": 1.3425,
|
| 4197 |
+
"step": 596
|
| 4198 |
+
},
|
| 4199 |
+
{
|
| 4200 |
+
"epoch": 0.11707030100990293,
|
| 4201 |
+
"grad_norm": 7.527665138244629,
|
| 4202 |
+
"learning_rate": 0.00012043468291741116,
|
| 4203 |
+
"loss": 1.6366,
|
| 4204 |
+
"step": 597
|
| 4205 |
+
},
|
| 4206 |
+
{
|
| 4207 |
+
"epoch": 0.11726639866653593,
|
| 4208 |
+
"grad_norm": 6.0890793800354,
|
| 4209 |
+
"learning_rate": 0.00012020699645232721,
|
| 4210 |
+
"loss": 1.9652,
|
| 4211 |
+
"step": 598
|
| 4212 |
+
},
|
| 4213 |
+
{
|
| 4214 |
+
"epoch": 0.11746249632316894,
|
| 4215 |
+
"grad_norm": 6.350461959838867,
|
| 4216 |
+
"learning_rate": 0.00011997920071984133,
|
| 4217 |
+
"loss": 2.2867,
|
| 4218 |
+
"step": 599
|
| 4219 |
+
},
|
| 4220 |
+
{
|
| 4221 |
+
"epoch": 0.11765859397980194,
|
| 4222 |
+
"grad_norm": 6.444486618041992,
|
| 4223 |
+
"learning_rate": 0.00011975129695173719,
|
| 4224 |
+
"loss": 0.7242,
|
| 4225 |
+
"step": 600
|
| 4226 |
+
},
|
| 4227 |
+
{
|
| 4228 |
+
"epoch": 0.11785469163643494,
|
| 4229 |
+
"grad_norm": 7.270365238189697,
|
| 4230 |
+
"learning_rate": 0.00011952328638038261,
|
| 4231 |
+
"loss": 2.2346,
|
| 4232 |
+
"step": 601
|
| 4233 |
+
},
|
| 4234 |
+
{
|
| 4235 |
+
"epoch": 0.11805078929306795,
|
| 4236 |
+
"grad_norm": 5.961056709289551,
|
| 4237 |
+
"learning_rate": 0.00011929517023872298,
|
| 4238 |
+
"loss": 2.4926,
|
| 4239 |
+
"step": 602
|
| 4240 |
+
},
|
| 4241 |
+
{
|
| 4242 |
+
"epoch": 0.11824688694970095,
|
| 4243 |
+
"grad_norm": 5.347055435180664,
|
| 4244 |
+
"learning_rate": 0.00011906694976027455,
|
| 4245 |
+
"loss": 1.3998,
|
| 4246 |
+
"step": 603
|
| 4247 |
+
},
|
| 4248 |
+
{
|
| 4249 |
+
"epoch": 0.11844298460633396,
|
| 4250 |
+
"grad_norm": 11.233309745788574,
|
| 4251 |
+
"learning_rate": 0.0001188386261791177,
|
| 4252 |
+
"loss": 2.0185,
|
| 4253 |
+
"step": 604
|
| 4254 |
+
},
|
| 4255 |
+
{
|
| 4256 |
+
"epoch": 0.11863908226296696,
|
| 4257 |
+
"grad_norm": 4.813647747039795,
|
| 4258 |
+
"learning_rate": 0.0001186102007298904,
|
| 4259 |
+
"loss": 1.3932,
|
| 4260 |
+
"step": 605
|
| 4261 |
+
},
|
| 4262 |
+
{
|
| 4263 |
+
"epoch": 0.11883517991959996,
|
| 4264 |
+
"grad_norm": 5.281280040740967,
|
| 4265 |
+
"learning_rate": 0.00011838167464778142,
|
| 4266 |
+
"loss": 1.7445,
|
| 4267 |
+
"step": 606
|
| 4268 |
+
},
|
| 4269 |
+
{
|
| 4270 |
+
"epoch": 0.11903127757623297,
|
| 4271 |
+
"grad_norm": 8.755961418151855,
|
| 4272 |
+
"learning_rate": 0.00011815304916852372,
|
| 4273 |
+
"loss": 1.656,
|
| 4274 |
+
"step": 607
|
| 4275 |
+
},
|
| 4276 |
+
{
|
| 4277 |
+
"epoch": 0.11922737523286597,
|
| 4278 |
+
"grad_norm": 15.501289367675781,
|
| 4279 |
+
"learning_rate": 0.00011792432552838772,
|
| 4280 |
+
"loss": 2.0894,
|
| 4281 |
+
"step": 608
|
| 4282 |
+
},
|
| 4283 |
+
{
|
| 4284 |
+
"epoch": 0.11942347288949898,
|
| 4285 |
+
"grad_norm": 6.1117753982543945,
|
| 4286 |
+
"learning_rate": 0.00011769550496417466,
|
| 4287 |
+
"loss": 1.0473,
|
| 4288 |
+
"step": 609
|
| 4289 |
+
},
|
| 4290 |
+
{
|
| 4291 |
+
"epoch": 0.11961957054613197,
|
| 4292 |
+
"grad_norm": 9.669829368591309,
|
| 4293 |
+
"learning_rate": 0.00011746658871320983,
|
| 4294 |
+
"loss": 0.7436,
|
| 4295 |
+
"step": 610
|
| 4296 |
+
},
|
| 4297 |
+
{
|
| 4298 |
+
"epoch": 0.11981566820276497,
|
| 4299 |
+
"grad_norm": 10.253652572631836,
|
| 4300 |
+
"learning_rate": 0.00011723757801333601,
|
| 4301 |
+
"loss": 2.7502,
|
| 4302 |
+
"step": 611
|
| 4303 |
+
},
|
| 4304 |
+
{
|
| 4305 |
+
"epoch": 0.12001176585939798,
|
| 4306 |
+
"grad_norm": 7.084132671356201,
|
| 4307 |
+
"learning_rate": 0.00011700847410290667,
|
| 4308 |
+
"loss": 2.5537,
|
| 4309 |
+
"step": 612
|
| 4310 |
+
},
|
| 4311 |
+
{
|
| 4312 |
+
"epoch": 0.12020786351603098,
|
| 4313 |
+
"grad_norm": 5.936439037322998,
|
| 4314 |
+
"learning_rate": 0.0001167792782207793,
|
| 4315 |
+
"loss": 1.1551,
|
| 4316 |
+
"step": 613
|
| 4317 |
+
},
|
| 4318 |
+
{
|
| 4319 |
+
"epoch": 0.12040396117266398,
|
| 4320 |
+
"grad_norm": 7.053248405456543,
|
| 4321 |
+
"learning_rate": 0.0001165499916063087,
|
| 4322 |
+
"loss": 1.2488,
|
| 4323 |
+
"step": 614
|
| 4324 |
+
},
|
| 4325 |
+
{
|
| 4326 |
+
"epoch": 0.12060005882929699,
|
| 4327 |
+
"grad_norm": 7.141607284545898,
|
| 4328 |
+
"learning_rate": 0.00011632061549934036,
|
| 4329 |
+
"loss": 1.4092,
|
| 4330 |
+
"step": 615
|
| 4331 |
+
},
|
| 4332 |
+
{
|
| 4333 |
+
"epoch": 0.12079615648592999,
|
| 4334 |
+
"grad_norm": 7.723854064941406,
|
| 4335 |
+
"learning_rate": 0.00011609115114020363,
|
| 4336 |
+
"loss": 1.7669,
|
| 4337 |
+
"step": 616
|
| 4338 |
+
},
|
| 4339 |
+
{
|
| 4340 |
+
"epoch": 0.120992254142563,
|
| 4341 |
+
"grad_norm": 7.095085144042969,
|
| 4342 |
+
"learning_rate": 0.00011586159976970514,
|
| 4343 |
+
"loss": 1.5018,
|
| 4344 |
+
"step": 617
|
| 4345 |
+
},
|
| 4346 |
+
{
|
| 4347 |
+
"epoch": 0.121188351799196,
|
| 4348 |
+
"grad_norm": 6.469241142272949,
|
| 4349 |
+
"learning_rate": 0.00011563196262912193,
|
| 4350 |
+
"loss": 2.3377,
|
| 4351 |
+
"step": 618
|
| 4352 |
+
},
|
| 4353 |
+
{
|
| 4354 |
+
"epoch": 0.121384449455829,
|
| 4355 |
+
"grad_norm": 6.16892671585083,
|
| 4356 |
+
"learning_rate": 0.00011540224096019494,
|
| 4357 |
+
"loss": 2.0011,
|
| 4358 |
+
"step": 619
|
| 4359 |
+
},
|
| 4360 |
+
{
|
| 4361 |
+
"epoch": 0.12158054711246201,
|
| 4362 |
+
"grad_norm": 4.764594554901123,
|
| 4363 |
+
"learning_rate": 0.00011517243600512209,
|
| 4364 |
+
"loss": 0.8995,
|
| 4365 |
+
"step": 620
|
| 4366 |
+
},
|
| 4367 |
+
{
|
| 4368 |
+
"epoch": 0.12177664476909501,
|
| 4369 |
+
"grad_norm": 8.256946563720703,
|
| 4370 |
+
"learning_rate": 0.00011494254900655175,
|
| 4371 |
+
"loss": 1.5867,
|
| 4372 |
+
"step": 621
|
| 4373 |
+
},
|
| 4374 |
+
{
|
| 4375 |
+
"epoch": 0.12197274242572802,
|
| 4376 |
+
"grad_norm": 9.220061302185059,
|
| 4377 |
+
"learning_rate": 0.00011471258120757586,
|
| 4378 |
+
"loss": 1.0146,
|
| 4379 |
+
"step": 622
|
| 4380 |
+
},
|
| 4381 |
+
{
|
| 4382 |
+
"epoch": 0.12216884008236102,
|
| 4383 |
+
"grad_norm": 6.284097194671631,
|
| 4384 |
+
"learning_rate": 0.00011448253385172335,
|
| 4385 |
+
"loss": 0.8305,
|
| 4386 |
+
"step": 623
|
| 4387 |
+
},
|
| 4388 |
+
{
|
| 4389 |
+
"epoch": 0.12236493773899403,
|
| 4390 |
+
"grad_norm": 6.183596134185791,
|
| 4391 |
+
"learning_rate": 0.00011425240818295326,
|
| 4392 |
+
"loss": 0.7749,
|
| 4393 |
+
"step": 624
|
| 4394 |
+
},
|
| 4395 |
+
{
|
| 4396 |
+
"epoch": 0.12256103539562702,
|
| 4397 |
+
"grad_norm": 7.550481796264648,
|
| 4398 |
+
"learning_rate": 0.00011402220544564819,
|
| 4399 |
+
"loss": 1.3157,
|
| 4400 |
+
"step": 625
|
| 4401 |
+
},
|
| 4402 |
+
{
|
| 4403 |
+
"epoch": 0.12275713305226002,
|
| 4404 |
+
"grad_norm": 4.810642719268799,
|
| 4405 |
+
"learning_rate": 0.0001137919268846074,
|
| 4406 |
+
"loss": 1.6442,
|
| 4407 |
+
"step": 626
|
| 4408 |
+
},
|
| 4409 |
+
{
|
| 4410 |
+
"epoch": 0.12295323070889302,
|
| 4411 |
+
"grad_norm": 10.603952407836914,
|
| 4412 |
+
"learning_rate": 0.00011356157374504027,
|
| 4413 |
+
"loss": 1.7275,
|
| 4414 |
+
"step": 627
|
| 4415 |
+
},
|
| 4416 |
+
{
|
| 4417 |
+
"epoch": 0.12314932836552603,
|
| 4418 |
+
"grad_norm": 13.009543418884277,
|
| 4419 |
+
"learning_rate": 0.00011333114727255929,
|
| 4420 |
+
"loss": 2.3718,
|
| 4421 |
+
"step": 628
|
| 4422 |
+
},
|
| 4423 |
+
{
|
| 4424 |
+
"epoch": 0.12334542602215903,
|
| 4425 |
+
"grad_norm": 8.453059196472168,
|
| 4426 |
+
"learning_rate": 0.00011310064871317366,
|
| 4427 |
+
"loss": 1.6558,
|
| 4428 |
+
"step": 629
|
| 4429 |
+
},
|
| 4430 |
+
{
|
| 4431 |
+
"epoch": 0.12354152367879204,
|
| 4432 |
+
"grad_norm": 7.005093574523926,
|
| 4433 |
+
"learning_rate": 0.00011287007931328226,
|
| 4434 |
+
"loss": 1.9709,
|
| 4435 |
+
"step": 630
|
| 4436 |
+
},
|
| 4437 |
+
{
|
| 4438 |
+
"epoch": 0.12373762133542504,
|
| 4439 |
+
"grad_norm": 5.710804462432861,
|
| 4440 |
+
"learning_rate": 0.00011263944031966714,
|
| 4441 |
+
"loss": 1.8595,
|
| 4442 |
+
"step": 631
|
| 4443 |
+
},
|
| 4444 |
+
{
|
| 4445 |
+
"epoch": 0.12393371899205805,
|
| 4446 |
+
"grad_norm": 3.608022689819336,
|
| 4447 |
+
"learning_rate": 0.00011240873297948653,
|
| 4448 |
+
"loss": 1.7165,
|
| 4449 |
+
"step": 632
|
| 4450 |
+
},
|
| 4451 |
+
{
|
| 4452 |
+
"epoch": 0.12412981664869105,
|
| 4453 |
+
"grad_norm": 8.973870277404785,
|
| 4454 |
+
"learning_rate": 0.0001121779585402684,
|
| 4455 |
+
"loss": 1.4924,
|
| 4456 |
+
"step": 633
|
| 4457 |
+
},
|
| 4458 |
+
{
|
| 4459 |
+
"epoch": 0.12432591430532405,
|
| 4460 |
+
"grad_norm": 4.735976219177246,
|
| 4461 |
+
"learning_rate": 0.00011194711824990344,
|
| 4462 |
+
"loss": 1.2721,
|
| 4463 |
+
"step": 634
|
| 4464 |
+
},
|
| 4465 |
+
{
|
| 4466 |
+
"epoch": 0.12452201196195706,
|
| 4467 |
+
"grad_norm": 4.169306755065918,
|
| 4468 |
+
"learning_rate": 0.00011171621335663844,
|
| 4469 |
+
"loss": 0.718,
|
| 4470 |
+
"step": 635
|
| 4471 |
+
},
|
| 4472 |
+
{
|
| 4473 |
+
"epoch": 0.12471810961859006,
|
| 4474 |
+
"grad_norm": 14.129206657409668,
|
| 4475 |
+
"learning_rate": 0.00011148524510906956,
|
| 4476 |
+
"loss": 1.1904,
|
| 4477 |
+
"step": 636
|
| 4478 |
+
},
|
| 4479 |
+
{
|
| 4480 |
+
"epoch": 0.12491420727522307,
|
| 4481 |
+
"grad_norm": 12.76352310180664,
|
| 4482 |
+
"learning_rate": 0.00011125421475613554,
|
| 4483 |
+
"loss": 2.4872,
|
| 4484 |
+
"step": 637
|
| 4485 |
+
},
|
| 4486 |
+
{
|
| 4487 |
+
"epoch": 0.12511030493185607,
|
| 4488 |
+
"grad_norm": 8.636000633239746,
|
| 4489 |
+
"learning_rate": 0.00011102312354711091,
|
| 4490 |
+
"loss": 0.7629,
|
| 4491 |
+
"step": 638
|
| 4492 |
+
},
|
| 4493 |
+
{
|
| 4494 |
+
"epoch": 0.12530640258848907,
|
| 4495 |
+
"grad_norm": 6.22203254699707,
|
| 4496 |
+
"learning_rate": 0.00011079197273159925,
|
| 4497 |
+
"loss": 1.7372,
|
| 4498 |
+
"step": 639
|
| 4499 |
+
},
|
| 4500 |
+
{
|
| 4501 |
+
"epoch": 0.12550250024512208,
|
| 4502 |
+
"grad_norm": 4.235958576202393,
|
| 4503 |
+
"learning_rate": 0.0001105607635595266,
|
| 4504 |
+
"loss": 1.8661,
|
| 4505 |
+
"step": 640
|
| 4506 |
+
},
|
| 4507 |
+
{
|
| 4508 |
+
"epoch": 0.12569859790175508,
|
| 4509 |
+
"grad_norm": 5.598705768585205,
|
| 4510 |
+
"learning_rate": 0.00011032949728113437,
|
| 4511 |
+
"loss": 1.7918,
|
| 4512 |
+
"step": 641
|
| 4513 |
+
},
|
| 4514 |
+
{
|
| 4515 |
+
"epoch": 0.1258946955583881,
|
| 4516 |
+
"grad_norm": 9.218809127807617,
|
| 4517 |
+
"learning_rate": 0.00011009817514697291,
|
| 4518 |
+
"loss": 1.2968,
|
| 4519 |
+
"step": 642
|
| 4520 |
+
},
|
| 4521 |
+
{
|
| 4522 |
+
"epoch": 0.1260907932150211,
|
| 4523 |
+
"grad_norm": 7.068932056427002,
|
| 4524 |
+
"learning_rate": 0.00010986679840789451,
|
| 4525 |
+
"loss": 1.8485,
|
| 4526 |
+
"step": 643
|
| 4527 |
+
},
|
| 4528 |
+
{
|
| 4529 |
+
"epoch": 0.1262868908716541,
|
| 4530 |
+
"grad_norm": 3.891483783721924,
|
| 4531 |
+
"learning_rate": 0.00010963536831504684,
|
| 4532 |
+
"loss": 0.3918,
|
| 4533 |
+
"step": 644
|
| 4534 |
+
},
|
| 4535 |
+
{
|
| 4536 |
+
"epoch": 0.1264829885282871,
|
| 4537 |
+
"grad_norm": 5.959576606750488,
|
| 4538 |
+
"learning_rate": 0.00010940388611986592,
|
| 4539 |
+
"loss": 0.7057,
|
| 4540 |
+
"step": 645
|
| 4541 |
+
},
|
| 4542 |
+
{
|
| 4543 |
+
"epoch": 0.1266790861849201,
|
| 4544 |
+
"grad_norm": 11.663713455200195,
|
| 4545 |
+
"learning_rate": 0.00010917235307406966,
|
| 4546 |
+
"loss": 1.5916,
|
| 4547 |
+
"step": 646
|
| 4548 |
+
},
|
| 4549 |
+
{
|
| 4550 |
+
"epoch": 0.12687518384155308,
|
| 4551 |
+
"grad_norm": 15.05775260925293,
|
| 4552 |
+
"learning_rate": 0.00010894077042965083,
|
| 4553 |
+
"loss": 2.979,
|
| 4554 |
+
"step": 647
|
| 4555 |
+
},
|
| 4556 |
+
{
|
| 4557 |
+
"epoch": 0.12707128149818608,
|
| 4558 |
+
"grad_norm": 7.2112650871276855,
|
| 4559 |
+
"learning_rate": 0.00010870913943887049,
|
| 4560 |
+
"loss": 1.6293,
|
| 4561 |
+
"step": 648
|
| 4562 |
+
},
|
| 4563 |
+
{
|
| 4564 |
+
"epoch": 0.1272673791548191,
|
| 4565 |
+
"grad_norm": 13.646367073059082,
|
| 4566 |
+
"learning_rate": 0.00010847746135425102,
|
| 4567 |
+
"loss": 2.5277,
|
| 4568 |
+
"step": 649
|
| 4569 |
+
},
|
| 4570 |
+
{
|
| 4571 |
+
"epoch": 0.1274634768114521,
|
| 4572 |
+
"grad_norm": 6.29921293258667,
|
| 4573 |
+
"learning_rate": 0.0001082457374285696,
|
| 4574 |
+
"loss": 2.4015,
|
| 4575 |
+
"step": 650
|
| 4576 |
+
},
|
| 4577 |
+
{
|
| 4578 |
+
"epoch": 0.1276595744680851,
|
| 4579 |
+
"grad_norm": 6.337920188903809,
|
| 4580 |
+
"learning_rate": 0.00010801396891485114,
|
| 4581 |
+
"loss": 0.9928,
|
| 4582 |
+
"step": 651
|
| 4583 |
+
},
|
| 4584 |
+
{
|
| 4585 |
+
"epoch": 0.1278556721247181,
|
| 4586 |
+
"grad_norm": 7.32103157043457,
|
| 4587 |
+
"learning_rate": 0.00010778215706636177,
|
| 4588 |
+
"loss": 2.4466,
|
| 4589 |
+
"step": 652
|
| 4590 |
+
},
|
| 4591 |
+
{
|
| 4592 |
+
"epoch": 0.1280517697813511,
|
| 4593 |
+
"grad_norm": 10.154195785522461,
|
| 4594 |
+
"learning_rate": 0.00010755030313660188,
|
| 4595 |
+
"loss": 2.4858,
|
| 4596 |
+
"step": 653
|
| 4597 |
+
},
|
| 4598 |
+
{
|
| 4599 |
+
"epoch": 0.1282478674379841,
|
| 4600 |
+
"grad_norm": 7.129855632781982,
|
| 4601 |
+
"learning_rate": 0.00010731840837929946,
|
| 4602 |
+
"loss": 1.2546,
|
| 4603 |
+
"step": 654
|
| 4604 |
+
},
|
| 4605 |
+
{
|
| 4606 |
+
"epoch": 0.1284439650946171,
|
| 4607 |
+
"grad_norm": 5.403641223907471,
|
| 4608 |
+
"learning_rate": 0.00010708647404840319,
|
| 4609 |
+
"loss": 1.1017,
|
| 4610 |
+
"step": 655
|
| 4611 |
+
},
|
| 4612 |
+
{
|
| 4613 |
+
"epoch": 0.12864006275125012,
|
| 4614 |
+
"grad_norm": 5.554705619812012,
|
| 4615 |
+
"learning_rate": 0.00010685450139807584,
|
| 4616 |
+
"loss": 2.0789,
|
| 4617 |
+
"step": 656
|
| 4618 |
+
},
|
| 4619 |
+
{
|
| 4620 |
+
"epoch": 0.12883616040788312,
|
| 4621 |
+
"grad_norm": 7.935024738311768,
|
| 4622 |
+
"learning_rate": 0.00010662249168268736,
|
| 4623 |
+
"loss": 1.9944,
|
| 4624 |
+
"step": 657
|
| 4625 |
+
},
|
| 4626 |
+
{
|
| 4627 |
+
"epoch": 0.12903225806451613,
|
| 4628 |
+
"grad_norm": 4.453607559204102,
|
| 4629 |
+
"learning_rate": 0.00010639044615680809,
|
| 4630 |
+
"loss": 0.8442,
|
| 4631 |
+
"step": 658
|
| 4632 |
+
},
|
| 4633 |
+
{
|
| 4634 |
+
"epoch": 0.12922835572114913,
|
| 4635 |
+
"grad_norm": 13.252726554870605,
|
| 4636 |
+
"learning_rate": 0.00010615836607520202,
|
| 4637 |
+
"loss": 1.1374,
|
| 4638 |
+
"step": 659
|
| 4639 |
+
},
|
| 4640 |
+
{
|
| 4641 |
+
"epoch": 0.12942445337778213,
|
| 4642 |
+
"grad_norm": 4.426390171051025,
|
| 4643 |
+
"learning_rate": 0.00010592625269282,
|
| 4644 |
+
"loss": 1.0554,
|
| 4645 |
+
"step": 660
|
| 4646 |
+
},
|
| 4647 |
+
{
|
| 4648 |
+
"epoch": 0.12962055103441514,
|
| 4649 |
+
"grad_norm": 6.692317008972168,
|
| 4650 |
+
"learning_rate": 0.000105694107264793,
|
| 4651 |
+
"loss": 0.6669,
|
| 4652 |
+
"step": 661
|
| 4653 |
+
},
|
| 4654 |
+
{
|
| 4655 |
+
"epoch": 0.12981664869104814,
|
| 4656 |
+
"grad_norm": 6.067706108093262,
|
| 4657 |
+
"learning_rate": 0.00010546193104642519,
|
| 4658 |
+
"loss": 1.488,
|
| 4659 |
+
"step": 662
|
| 4660 |
+
},
|
| 4661 |
+
{
|
| 4662 |
+
"epoch": 0.13001274634768115,
|
| 4663 |
+
"grad_norm": 7.749240875244141,
|
| 4664 |
+
"learning_rate": 0.00010522972529318733,
|
| 4665 |
+
"loss": 1.8632,
|
| 4666 |
+
"step": 663
|
| 4667 |
+
},
|
| 4668 |
+
{
|
| 4669 |
+
"epoch": 0.13020884400431415,
|
| 4670 |
+
"grad_norm": 4.964468955993652,
|
| 4671 |
+
"learning_rate": 0.00010499749126070979,
|
| 4672 |
+
"loss": 0.5845,
|
| 4673 |
+
"step": 664
|
| 4674 |
+
},
|
| 4675 |
+
{
|
| 4676 |
+
"epoch": 0.13040494166094715,
|
| 4677 |
+
"grad_norm": 6.432995319366455,
|
| 4678 |
+
"learning_rate": 0.00010476523020477593,
|
| 4679 |
+
"loss": 1.1617,
|
| 4680 |
+
"step": 665
|
| 4681 |
+
},
|
| 4682 |
+
{
|
| 4683 |
+
"epoch": 0.13060103931758016,
|
| 4684 |
+
"grad_norm": 6.4099249839782715,
|
| 4685 |
+
"learning_rate": 0.00010453294338131519,
|
| 4686 |
+
"loss": 1.2821,
|
| 4687 |
+
"step": 666
|
| 4688 |
+
},
|
| 4689 |
+
{
|
| 4690 |
+
"epoch": 0.13079713697421316,
|
| 4691 |
+
"grad_norm": 5.935438632965088,
|
| 4692 |
+
"learning_rate": 0.0001043006320463964,
|
| 4693 |
+
"loss": 1.4453,
|
| 4694 |
+
"step": 667
|
| 4695 |
+
},
|
| 4696 |
+
{
|
| 4697 |
+
"epoch": 0.13099323463084617,
|
| 4698 |
+
"grad_norm": 10.963349342346191,
|
| 4699 |
+
"learning_rate": 0.00010406829745622085,
|
| 4700 |
+
"loss": 1.6318,
|
| 4701 |
+
"step": 668
|
| 4702 |
+
},
|
| 4703 |
+
{
|
| 4704 |
+
"epoch": 0.13118933228747917,
|
| 4705 |
+
"grad_norm": 9.157288551330566,
|
| 4706 |
+
"learning_rate": 0.00010383594086711567,
|
| 4707 |
+
"loss": 1.6981,
|
| 4708 |
+
"step": 669
|
| 4709 |
+
},
|
| 4710 |
+
{
|
| 4711 |
+
"epoch": 0.13138542994411218,
|
| 4712 |
+
"grad_norm": 17.15442657470703,
|
| 4713 |
+
"learning_rate": 0.00010360356353552687,
|
| 4714 |
+
"loss": 1.498,
|
| 4715 |
+
"step": 670
|
| 4716 |
+
},
|
| 4717 |
+
{
|
| 4718 |
+
"epoch": 0.13158152760074518,
|
| 4719 |
+
"grad_norm": 3.763289451599121,
|
| 4720 |
+
"learning_rate": 0.00010337116671801272,
|
| 4721 |
+
"loss": 2.425,
|
| 4722 |
+
"step": 671
|
| 4723 |
+
},
|
| 4724 |
+
{
|
| 4725 |
+
"epoch": 0.13177762525737818,
|
| 4726 |
+
"grad_norm": 8.91812801361084,
|
| 4727 |
+
"learning_rate": 0.00010313875167123672,
|
| 4728 |
+
"loss": 2.3129,
|
| 4729 |
+
"step": 672
|
| 4730 |
+
},
|
| 4731 |
+
{
|
| 4732 |
+
"epoch": 0.1319737229140112,
|
| 4733 |
+
"grad_norm": 23.093202590942383,
|
| 4734 |
+
"learning_rate": 0.00010290631965196109,
|
| 4735 |
+
"loss": 2.3172,
|
| 4736 |
+
"step": 673
|
| 4737 |
+
},
|
| 4738 |
+
{
|
| 4739 |
+
"epoch": 0.1321698205706442,
|
| 4740 |
+
"grad_norm": 6.487079620361328,
|
| 4741 |
+
"learning_rate": 0.00010267387191703972,
|
| 4742 |
+
"loss": 1.6108,
|
| 4743 |
+
"step": 674
|
| 4744 |
+
},
|
| 4745 |
+
{
|
| 4746 |
+
"epoch": 0.1323659182272772,
|
| 4747 |
+
"grad_norm": 3.9093501567840576,
|
| 4748 |
+
"learning_rate": 0.00010244140972341155,
|
| 4749 |
+
"loss": 0.8954,
|
| 4750 |
+
"step": 675
|
| 4751 |
+
},
|
| 4752 |
+
{
|
| 4753 |
+
"epoch": 0.1325620158839102,
|
| 4754 |
+
"grad_norm": 11.078289985656738,
|
| 4755 |
+
"learning_rate": 0.00010220893432809365,
|
| 4756 |
+
"loss": 2.0604,
|
| 4757 |
+
"step": 676
|
| 4758 |
+
},
|
| 4759 |
+
{
|
| 4760 |
+
"epoch": 0.13275811354054318,
|
| 4761 |
+
"grad_norm": 7.805664539337158,
|
| 4762 |
+
"learning_rate": 0.00010197644698817446,
|
| 4763 |
+
"loss": 1.834,
|
| 4764 |
+
"step": 677
|
| 4765 |
+
},
|
| 4766 |
+
{
|
| 4767 |
+
"epoch": 0.13295421119717618,
|
| 4768 |
+
"grad_norm": 6.838125705718994,
|
| 4769 |
+
"learning_rate": 0.00010174394896080713,
|
| 4770 |
+
"loss": 1.4876,
|
| 4771 |
+
"step": 678
|
| 4772 |
+
},
|
| 4773 |
+
{
|
| 4774 |
+
"epoch": 0.13315030885380919,
|
| 4775 |
+
"grad_norm": 3.9814906120300293,
|
| 4776 |
+
"learning_rate": 0.00010151144150320247,
|
| 4777 |
+
"loss": 1.3479,
|
| 4778 |
+
"step": 679
|
| 4779 |
+
},
|
| 4780 |
+
{
|
| 4781 |
+
"epoch": 0.1333464065104422,
|
| 4782 |
+
"grad_norm": 6.396786212921143,
|
| 4783 |
+
"learning_rate": 0.00010127892587262233,
|
| 4784 |
+
"loss": 1.7642,
|
| 4785 |
+
"step": 680
|
| 4786 |
+
},
|
| 4787 |
+
{
|
| 4788 |
+
"epoch": 0.1335425041670752,
|
| 4789 |
+
"grad_norm": 10.133893013000488,
|
| 4790 |
+
"learning_rate": 0.00010104640332637278,
|
| 4791 |
+
"loss": 0.8957,
|
| 4792 |
+
"step": 681
|
| 4793 |
+
},
|
| 4794 |
+
{
|
| 4795 |
+
"epoch": 0.1337386018237082,
|
| 4796 |
+
"grad_norm": 6.3815412521362305,
|
| 4797 |
+
"learning_rate": 0.00010081387512179729,
|
| 4798 |
+
"loss": 1.4104,
|
| 4799 |
+
"step": 682
|
| 4800 |
+
},
|
| 4801 |
+
{
|
| 4802 |
+
"epoch": 0.1337386018237082,
|
| 4803 |
+
"eval_loss": 0.41039586067199707,
|
| 4804 |
+
"eval_runtime": 77.8442,
|
| 4805 |
+
"eval_samples_per_second": 27.594,
|
| 4806 |
+
"eval_steps_per_second": 13.797,
|
| 4807 |
+
"step": 682
|
| 4808 |
}
|
| 4809 |
],
|
| 4810 |
"logging_steps": 1,
|
|
|
|
| 4824 |
"attributes": {}
|
| 4825 |
}
|
| 4826 |
},
|
| 4827 |
+
"total_flos": 1.470491639540613e+17,
|
| 4828 |
"train_batch_size": 2,
|
| 4829 |
"trial_name": null,
|
| 4830 |
"trial_params": null
|