Training in progress, step 400, 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 639691872
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:9961be420a9f5b1ece0bf8352ac72f0825cae7d693fa0e49096ffb7c6fd9d324
|
| 3 |
size 639691872
|
last-checkpoint/optimizer.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 1279647314
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:e27e9ec9ac3798574a12ad5881eb724672736596feeec845aa325cf61868ceae
|
| 3 |
size 1279647314
|
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 sha256:13af9b7c642e074997a9bdbae88ba56e792139bc8360f4873ff633fa70291205
|
| 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 sha256:6dd3cfeea20c315ac503882fc53b8b9eae9cdcaf6dd61efec7bfa78209d384fd
|
| 3 |
size 1064
|
last-checkpoint/trainer_state.json
CHANGED
|
@@ -1,9 +1,9 @@
|
|
| 1 |
{
|
| 2 |
-
"best_metric": 1.
|
| 3 |
-
"best_model_checkpoint": "miner_id_24/checkpoint-
|
| 4 |
-
"epoch": 0.
|
| 5 |
"eval_steps": 200,
|
| 6 |
-
"global_step":
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
@@ -1423,6 +1423,1414 @@
|
|
| 1423 |
"eval_samples_per_second": 2.12,
|
| 1424 |
"eval_steps_per_second": 2.12,
|
| 1425 |
"step": 200
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1426 |
}
|
| 1427 |
],
|
| 1428 |
"logging_steps": 1,
|
|
@@ -1451,7 +2859,7 @@
|
|
| 1451 |
"attributes": {}
|
| 1452 |
}
|
| 1453 |
},
|
| 1454 |
-
"total_flos":
|
| 1455 |
"train_batch_size": 1,
|
| 1456 |
"trial_name": null,
|
| 1457 |
"trial_params": null
|
|
|
|
| 1 |
{
|
| 2 |
+
"best_metric": 1.6627388000488281,
|
| 3 |
+
"best_model_checkpoint": "miner_id_24/checkpoint-400",
|
| 4 |
+
"epoch": 0.2157715518694582,
|
| 5 |
"eval_steps": 200,
|
| 6 |
+
"global_step": 400,
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
|
|
| 1423 |
"eval_samples_per_second": 2.12,
|
| 1424 |
"eval_steps_per_second": 2.12,
|
| 1425 |
"step": 200
|
| 1426 |
+
},
|
| 1427 |
+
{
|
| 1428 |
+
"epoch": 0.10842520481440275,
|
| 1429 |
+
"grad_norm": 23.590002059936523,
|
| 1430 |
+
"learning_rate": 0.0001998311255158628,
|
| 1431 |
+
"loss": 53.9458,
|
| 1432 |
+
"step": 201
|
| 1433 |
+
},
|
| 1434 |
+
{
|
| 1435 |
+
"epoch": 0.1089646336940764,
|
| 1436 |
+
"grad_norm": 39.737159729003906,
|
| 1437 |
+
"learning_rate": 0.00019982776587698666,
|
| 1438 |
+
"loss": 85.7514,
|
| 1439 |
+
"step": 202
|
| 1440 |
+
},
|
| 1441 |
+
{
|
| 1442 |
+
"epoch": 0.10950406257375005,
|
| 1443 |
+
"grad_norm": 35.41561508178711,
|
| 1444 |
+
"learning_rate": 0.00019982437317643217,
|
| 1445 |
+
"loss": 84.9662,
|
| 1446 |
+
"step": 203
|
| 1447 |
+
},
|
| 1448 |
+
{
|
| 1449 |
+
"epoch": 0.11004349145342368,
|
| 1450 |
+
"grad_norm": 31.39605140686035,
|
| 1451 |
+
"learning_rate": 0.0001998209474153229,
|
| 1452 |
+
"loss": 110.0561,
|
| 1453 |
+
"step": 204
|
| 1454 |
+
},
|
| 1455 |
+
{
|
| 1456 |
+
"epoch": 0.11058292033309733,
|
| 1457 |
+
"grad_norm": 30.160261154174805,
|
| 1458 |
+
"learning_rate": 0.00019981748859479348,
|
| 1459 |
+
"loss": 101.1574,
|
| 1460 |
+
"step": 205
|
| 1461 |
+
},
|
| 1462 |
+
{
|
| 1463 |
+
"epoch": 0.11112234921277098,
|
| 1464 |
+
"grad_norm": 33.4417724609375,
|
| 1465 |
+
"learning_rate": 0.00019981399671598939,
|
| 1466 |
+
"loss": 116.0456,
|
| 1467 |
+
"step": 206
|
| 1468 |
+
},
|
| 1469 |
+
{
|
| 1470 |
+
"epoch": 0.11166177809244462,
|
| 1471 |
+
"grad_norm": 34.16884994506836,
|
| 1472 |
+
"learning_rate": 0.0001998104717800671,
|
| 1473 |
+
"loss": 103.0287,
|
| 1474 |
+
"step": 207
|
| 1475 |
+
},
|
| 1476 |
+
{
|
| 1477 |
+
"epoch": 0.11220120697211827,
|
| 1478 |
+
"grad_norm": 33.58393859863281,
|
| 1479 |
+
"learning_rate": 0.00019980691378819406,
|
| 1480 |
+
"loss": 95.5024,
|
| 1481 |
+
"step": 208
|
| 1482 |
+
},
|
| 1483 |
+
{
|
| 1484 |
+
"epoch": 0.11274063585179192,
|
| 1485 |
+
"grad_norm": 29.785871505737305,
|
| 1486 |
+
"learning_rate": 0.00019980332274154857,
|
| 1487 |
+
"loss": 91.5854,
|
| 1488 |
+
"step": 209
|
| 1489 |
+
},
|
| 1490 |
+
{
|
| 1491 |
+
"epoch": 0.11328006473146555,
|
| 1492 |
+
"grad_norm": 29.184667587280273,
|
| 1493 |
+
"learning_rate": 0.00019979969864131997,
|
| 1494 |
+
"loss": 86.9138,
|
| 1495 |
+
"step": 210
|
| 1496 |
+
},
|
| 1497 |
+
{
|
| 1498 |
+
"epoch": 0.1138194936111392,
|
| 1499 |
+
"grad_norm": 25.164024353027344,
|
| 1500 |
+
"learning_rate": 0.00019979604148870854,
|
| 1501 |
+
"loss": 72.7827,
|
| 1502 |
+
"step": 211
|
| 1503 |
+
},
|
| 1504 |
+
{
|
| 1505 |
+
"epoch": 0.11435892249081285,
|
| 1506 |
+
"grad_norm": 18.179292678833008,
|
| 1507 |
+
"learning_rate": 0.00019979235128492545,
|
| 1508 |
+
"loss": 67.364,
|
| 1509 |
+
"step": 212
|
| 1510 |
+
},
|
| 1511 |
+
{
|
| 1512 |
+
"epoch": 0.1148983513704865,
|
| 1513 |
+
"grad_norm": 20.353260040283203,
|
| 1514 |
+
"learning_rate": 0.00019978862803119284,
|
| 1515 |
+
"loss": 60.0141,
|
| 1516 |
+
"step": 213
|
| 1517 |
+
},
|
| 1518 |
+
{
|
| 1519 |
+
"epoch": 0.11543778025016015,
|
| 1520 |
+
"grad_norm": 27.25603485107422,
|
| 1521 |
+
"learning_rate": 0.00019978487172874382,
|
| 1522 |
+
"loss": 61.8063,
|
| 1523 |
+
"step": 214
|
| 1524 |
+
},
|
| 1525 |
+
{
|
| 1526 |
+
"epoch": 0.1159772091298338,
|
| 1527 |
+
"grad_norm": 40.56468963623047,
|
| 1528 |
+
"learning_rate": 0.00019978108237882244,
|
| 1529 |
+
"loss": 51.2483,
|
| 1530 |
+
"step": 215
|
| 1531 |
+
},
|
| 1532 |
+
{
|
| 1533 |
+
"epoch": 0.11651663800950743,
|
| 1534 |
+
"grad_norm": 64.65696716308594,
|
| 1535 |
+
"learning_rate": 0.00019977725998268365,
|
| 1536 |
+
"loss": 37.8312,
|
| 1537 |
+
"step": 216
|
| 1538 |
+
},
|
| 1539 |
+
{
|
| 1540 |
+
"epoch": 0.11705606688918108,
|
| 1541 |
+
"grad_norm": 80.94468688964844,
|
| 1542 |
+
"learning_rate": 0.00019977340454159343,
|
| 1543 |
+
"loss": 55.2775,
|
| 1544 |
+
"step": 217
|
| 1545 |
+
},
|
| 1546 |
+
{
|
| 1547 |
+
"epoch": 0.11759549576885472,
|
| 1548 |
+
"grad_norm": 100.61930084228516,
|
| 1549 |
+
"learning_rate": 0.00019976951605682862,
|
| 1550 |
+
"loss": 65.5767,
|
| 1551 |
+
"step": 218
|
| 1552 |
+
},
|
| 1553 |
+
{
|
| 1554 |
+
"epoch": 0.11813492464852837,
|
| 1555 |
+
"grad_norm": 71.5768051147461,
|
| 1556 |
+
"learning_rate": 0.00019976559452967703,
|
| 1557 |
+
"loss": 57.5296,
|
| 1558 |
+
"step": 219
|
| 1559 |
+
},
|
| 1560 |
+
{
|
| 1561 |
+
"epoch": 0.11867435352820202,
|
| 1562 |
+
"grad_norm": 37.10725021362305,
|
| 1563 |
+
"learning_rate": 0.00019976163996143745,
|
| 1564 |
+
"loss": 48.8497,
|
| 1565 |
+
"step": 220
|
| 1566 |
+
},
|
| 1567 |
+
{
|
| 1568 |
+
"epoch": 0.11921378240787567,
|
| 1569 |
+
"grad_norm": 40.85627746582031,
|
| 1570 |
+
"learning_rate": 0.00019975765235341955,
|
| 1571 |
+
"loss": 47.6466,
|
| 1572 |
+
"step": 221
|
| 1573 |
+
},
|
| 1574 |
+
{
|
| 1575 |
+
"epoch": 0.1197532112875493,
|
| 1576 |
+
"grad_norm": 55.1395263671875,
|
| 1577 |
+
"learning_rate": 0.000199753631706944,
|
| 1578 |
+
"loss": 60.2519,
|
| 1579 |
+
"step": 222
|
| 1580 |
+
},
|
| 1581 |
+
{
|
| 1582 |
+
"epoch": 0.12029264016722295,
|
| 1583 |
+
"grad_norm": 42.060585021972656,
|
| 1584 |
+
"learning_rate": 0.00019974957802334234,
|
| 1585 |
+
"loss": 48.1031,
|
| 1586 |
+
"step": 223
|
| 1587 |
+
},
|
| 1588 |
+
{
|
| 1589 |
+
"epoch": 0.1208320690468966,
|
| 1590 |
+
"grad_norm": 36.57340621948242,
|
| 1591 |
+
"learning_rate": 0.00019974549130395713,
|
| 1592 |
+
"loss": 43.3995,
|
| 1593 |
+
"step": 224
|
| 1594 |
+
},
|
| 1595 |
+
{
|
| 1596 |
+
"epoch": 0.12137149792657025,
|
| 1597 |
+
"grad_norm": 31.497970581054688,
|
| 1598 |
+
"learning_rate": 0.0001997413715501419,
|
| 1599 |
+
"loss": 41.1591,
|
| 1600 |
+
"step": 225
|
| 1601 |
+
},
|
| 1602 |
+
{
|
| 1603 |
+
"epoch": 0.1219109268062439,
|
| 1604 |
+
"grad_norm": 30.481502532958984,
|
| 1605 |
+
"learning_rate": 0.00019973721876326094,
|
| 1606 |
+
"loss": 38.0712,
|
| 1607 |
+
"step": 226
|
| 1608 |
+
},
|
| 1609 |
+
{
|
| 1610 |
+
"epoch": 0.12245035568591753,
|
| 1611 |
+
"grad_norm": 38.2381477355957,
|
| 1612 |
+
"learning_rate": 0.00019973303294468968,
|
| 1613 |
+
"loss": 46.3861,
|
| 1614 |
+
"step": 227
|
| 1615 |
+
},
|
| 1616 |
+
{
|
| 1617 |
+
"epoch": 0.12298978456559118,
|
| 1618 |
+
"grad_norm": 37.4508171081543,
|
| 1619 |
+
"learning_rate": 0.0001997288140958144,
|
| 1620 |
+
"loss": 49.3107,
|
| 1621 |
+
"step": 228
|
| 1622 |
+
},
|
| 1623 |
+
{
|
| 1624 |
+
"epoch": 0.12352921344526482,
|
| 1625 |
+
"grad_norm": 37.3139533996582,
|
| 1626 |
+
"learning_rate": 0.0001997245622180323,
|
| 1627 |
+
"loss": 43.1914,
|
| 1628 |
+
"step": 229
|
| 1629 |
+
},
|
| 1630 |
+
{
|
| 1631 |
+
"epoch": 0.12406864232493847,
|
| 1632 |
+
"grad_norm": 35.13384246826172,
|
| 1633 |
+
"learning_rate": 0.0001997202773127516,
|
| 1634 |
+
"loss": 45.7228,
|
| 1635 |
+
"step": 230
|
| 1636 |
+
},
|
| 1637 |
+
{
|
| 1638 |
+
"epoch": 0.12460807120461212,
|
| 1639 |
+
"grad_norm": 37.45779037475586,
|
| 1640 |
+
"learning_rate": 0.00019971595938139135,
|
| 1641 |
+
"loss": 45.0848,
|
| 1642 |
+
"step": 231
|
| 1643 |
+
},
|
| 1644 |
+
{
|
| 1645 |
+
"epoch": 0.12514750008428577,
|
| 1646 |
+
"grad_norm": 37.03962707519531,
|
| 1647 |
+
"learning_rate": 0.00019971160842538162,
|
| 1648 |
+
"loss": 46.3705,
|
| 1649 |
+
"step": 232
|
| 1650 |
+
},
|
| 1651 |
+
{
|
| 1652 |
+
"epoch": 0.12568692896395942,
|
| 1653 |
+
"grad_norm": 30.98250389099121,
|
| 1654 |
+
"learning_rate": 0.0001997072244461634,
|
| 1655 |
+
"loss": 41.1065,
|
| 1656 |
+
"step": 233
|
| 1657 |
+
},
|
| 1658 |
+
{
|
| 1659 |
+
"epoch": 0.12622635784363306,
|
| 1660 |
+
"grad_norm": 33.62482833862305,
|
| 1661 |
+
"learning_rate": 0.00019970280744518854,
|
| 1662 |
+
"loss": 41.8594,
|
| 1663 |
+
"step": 234
|
| 1664 |
+
},
|
| 1665 |
+
{
|
| 1666 |
+
"epoch": 0.1267657867233067,
|
| 1667 |
+
"grad_norm": 45.488739013671875,
|
| 1668 |
+
"learning_rate": 0.00019969835742392,
|
| 1669 |
+
"loss": 38.6525,
|
| 1670 |
+
"step": 235
|
| 1671 |
+
},
|
| 1672 |
+
{
|
| 1673 |
+
"epoch": 0.12730521560298033,
|
| 1674 |
+
"grad_norm": 43.84321594238281,
|
| 1675 |
+
"learning_rate": 0.0001996938743838315,
|
| 1676 |
+
"loss": 53.2114,
|
| 1677 |
+
"step": 236
|
| 1678 |
+
},
|
| 1679 |
+
{
|
| 1680 |
+
"epoch": 0.12784464448265398,
|
| 1681 |
+
"grad_norm": 40.51958084106445,
|
| 1682 |
+
"learning_rate": 0.00019968935832640782,
|
| 1683 |
+
"loss": 50.4725,
|
| 1684 |
+
"step": 237
|
| 1685 |
+
},
|
| 1686 |
+
{
|
| 1687 |
+
"epoch": 0.12838407336232763,
|
| 1688 |
+
"grad_norm": 35.1596794128418,
|
| 1689 |
+
"learning_rate": 0.00019968480925314458,
|
| 1690 |
+
"loss": 45.1618,
|
| 1691 |
+
"step": 238
|
| 1692 |
+
},
|
| 1693 |
+
{
|
| 1694 |
+
"epoch": 0.12892350224200128,
|
| 1695 |
+
"grad_norm": 32.27614974975586,
|
| 1696 |
+
"learning_rate": 0.00019968022716554832,
|
| 1697 |
+
"loss": 38.2164,
|
| 1698 |
+
"step": 239
|
| 1699 |
+
},
|
| 1700 |
+
{
|
| 1701 |
+
"epoch": 0.12946293112167492,
|
| 1702 |
+
"grad_norm": 33.67794418334961,
|
| 1703 |
+
"learning_rate": 0.00019967561206513668,
|
| 1704 |
+
"loss": 43.3203,
|
| 1705 |
+
"step": 240
|
| 1706 |
+
},
|
| 1707 |
+
{
|
| 1708 |
+
"epoch": 0.13000236000134857,
|
| 1709 |
+
"grad_norm": 26.34979820251465,
|
| 1710 |
+
"learning_rate": 0.00019967096395343806,
|
| 1711 |
+
"loss": 32.1165,
|
| 1712 |
+
"step": 241
|
| 1713 |
+
},
|
| 1714 |
+
{
|
| 1715 |
+
"epoch": 0.13054178888102222,
|
| 1716 |
+
"grad_norm": 33.10830307006836,
|
| 1717 |
+
"learning_rate": 0.00019966628283199186,
|
| 1718 |
+
"loss": 45.5207,
|
| 1719 |
+
"step": 242
|
| 1720 |
+
},
|
| 1721 |
+
{
|
| 1722 |
+
"epoch": 0.13108121776069587,
|
| 1723 |
+
"grad_norm": 47.04872131347656,
|
| 1724 |
+
"learning_rate": 0.00019966156870234844,
|
| 1725 |
+
"loss": 44.7497,
|
| 1726 |
+
"step": 243
|
| 1727 |
+
},
|
| 1728 |
+
{
|
| 1729 |
+
"epoch": 0.13162064664036952,
|
| 1730 |
+
"grad_norm": 38.99346160888672,
|
| 1731 |
+
"learning_rate": 0.000199656821566069,
|
| 1732 |
+
"loss": 43.9255,
|
| 1733 |
+
"step": 244
|
| 1734 |
+
},
|
| 1735 |
+
{
|
| 1736 |
+
"epoch": 0.13216007552004316,
|
| 1737 |
+
"grad_norm": 29.892854690551758,
|
| 1738 |
+
"learning_rate": 0.00019965204142472574,
|
| 1739 |
+
"loss": 48.4896,
|
| 1740 |
+
"step": 245
|
| 1741 |
+
},
|
| 1742 |
+
{
|
| 1743 |
+
"epoch": 0.1326995043997168,
|
| 1744 |
+
"grad_norm": 37.65726089477539,
|
| 1745 |
+
"learning_rate": 0.00019964722827990185,
|
| 1746 |
+
"loss": 37.7987,
|
| 1747 |
+
"step": 246
|
| 1748 |
+
},
|
| 1749 |
+
{
|
| 1750 |
+
"epoch": 0.13323893327939046,
|
| 1751 |
+
"grad_norm": 41.673274993896484,
|
| 1752 |
+
"learning_rate": 0.00019964238213319134,
|
| 1753 |
+
"loss": 48.4095,
|
| 1754 |
+
"step": 247
|
| 1755 |
+
},
|
| 1756 |
+
{
|
| 1757 |
+
"epoch": 0.13377836215906408,
|
| 1758 |
+
"grad_norm": 37.152793884277344,
|
| 1759 |
+
"learning_rate": 0.00019963750298619917,
|
| 1760 |
+
"loss": 33.8212,
|
| 1761 |
+
"step": 248
|
| 1762 |
+
},
|
| 1763 |
+
{
|
| 1764 |
+
"epoch": 0.13431779103873773,
|
| 1765 |
+
"grad_norm": 43.92071533203125,
|
| 1766 |
+
"learning_rate": 0.00019963259084054128,
|
| 1767 |
+
"loss": 35.554,
|
| 1768 |
+
"step": 249
|
| 1769 |
+
},
|
| 1770 |
+
{
|
| 1771 |
+
"epoch": 0.13485721991841138,
|
| 1772 |
+
"grad_norm": 39.161903381347656,
|
| 1773 |
+
"learning_rate": 0.0001996276456978445,
|
| 1774 |
+
"loss": 33.8096,
|
| 1775 |
+
"step": 250
|
| 1776 |
+
},
|
| 1777 |
+
{
|
| 1778 |
+
"epoch": 0.13539664879808502,
|
| 1779 |
+
"grad_norm": 24.633363723754883,
|
| 1780 |
+
"learning_rate": 0.00019962266755974657,
|
| 1781 |
+
"loss": 46.0338,
|
| 1782 |
+
"step": 251
|
| 1783 |
+
},
|
| 1784 |
+
{
|
| 1785 |
+
"epoch": 0.13593607767775867,
|
| 1786 |
+
"grad_norm": 54.83051300048828,
|
| 1787 |
+
"learning_rate": 0.00019961765642789625,
|
| 1788 |
+
"loss": 80.4599,
|
| 1789 |
+
"step": 252
|
| 1790 |
+
},
|
| 1791 |
+
{
|
| 1792 |
+
"epoch": 0.13647550655743232,
|
| 1793 |
+
"grad_norm": 43.1768684387207,
|
| 1794 |
+
"learning_rate": 0.0001996126123039531,
|
| 1795 |
+
"loss": 84.3379,
|
| 1796 |
+
"step": 253
|
| 1797 |
+
},
|
| 1798 |
+
{
|
| 1799 |
+
"epoch": 0.13701493543710597,
|
| 1800 |
+
"grad_norm": 24.49346160888672,
|
| 1801 |
+
"learning_rate": 0.00019960753518958772,
|
| 1802 |
+
"loss": 100.9898,
|
| 1803 |
+
"step": 254
|
| 1804 |
+
},
|
| 1805 |
+
{
|
| 1806 |
+
"epoch": 0.13755436431677961,
|
| 1807 |
+
"grad_norm": 38.09309768676758,
|
| 1808 |
+
"learning_rate": 0.00019960242508648154,
|
| 1809 |
+
"loss": 101.0717,
|
| 1810 |
+
"step": 255
|
| 1811 |
+
},
|
| 1812 |
+
{
|
| 1813 |
+
"epoch": 0.13809379319645326,
|
| 1814 |
+
"grad_norm": 40.072296142578125,
|
| 1815 |
+
"learning_rate": 0.00019959728199632699,
|
| 1816 |
+
"loss": 108.2131,
|
| 1817 |
+
"step": 256
|
| 1818 |
+
},
|
| 1819 |
+
{
|
| 1820 |
+
"epoch": 0.1386332220761269,
|
| 1821 |
+
"grad_norm": 43.77210235595703,
|
| 1822 |
+
"learning_rate": 0.0001995921059208274,
|
| 1823 |
+
"loss": 111.636,
|
| 1824 |
+
"step": 257
|
| 1825 |
+
},
|
| 1826 |
+
{
|
| 1827 |
+
"epoch": 0.13917265095580056,
|
| 1828 |
+
"grad_norm": 42.023155212402344,
|
| 1829 |
+
"learning_rate": 0.00019958689686169697,
|
| 1830 |
+
"loss": 90.4911,
|
| 1831 |
+
"step": 258
|
| 1832 |
+
},
|
| 1833 |
+
{
|
| 1834 |
+
"epoch": 0.13971207983547418,
|
| 1835 |
+
"grad_norm": 27.917343139648438,
|
| 1836 |
+
"learning_rate": 0.00019958165482066094,
|
| 1837 |
+
"loss": 92.3676,
|
| 1838 |
+
"step": 259
|
| 1839 |
+
},
|
| 1840 |
+
{
|
| 1841 |
+
"epoch": 0.14025150871514783,
|
| 1842 |
+
"grad_norm": 19.174135208129883,
|
| 1843 |
+
"learning_rate": 0.00019957637979945537,
|
| 1844 |
+
"loss": 88.4276,
|
| 1845 |
+
"step": 260
|
| 1846 |
+
},
|
| 1847 |
+
{
|
| 1848 |
+
"epoch": 0.14079093759482147,
|
| 1849 |
+
"grad_norm": 22.779672622680664,
|
| 1850 |
+
"learning_rate": 0.0001995710717998273,
|
| 1851 |
+
"loss": 88.3991,
|
| 1852 |
+
"step": 261
|
| 1853 |
+
},
|
| 1854 |
+
{
|
| 1855 |
+
"epoch": 0.14133036647449512,
|
| 1856 |
+
"grad_norm": 17.607568740844727,
|
| 1857 |
+
"learning_rate": 0.00019956573082353463,
|
| 1858 |
+
"loss": 77.4426,
|
| 1859 |
+
"step": 262
|
| 1860 |
+
},
|
| 1861 |
+
{
|
| 1862 |
+
"epoch": 0.14186979535416877,
|
| 1863 |
+
"grad_norm": 22.228328704833984,
|
| 1864 |
+
"learning_rate": 0.00019956035687234626,
|
| 1865 |
+
"loss": 68.3415,
|
| 1866 |
+
"step": 263
|
| 1867 |
+
},
|
| 1868 |
+
{
|
| 1869 |
+
"epoch": 0.14240922423384242,
|
| 1870 |
+
"grad_norm": 21.00279998779297,
|
| 1871 |
+
"learning_rate": 0.00019955494994804198,
|
| 1872 |
+
"loss": 70.7203,
|
| 1873 |
+
"step": 264
|
| 1874 |
+
},
|
| 1875 |
+
{
|
| 1876 |
+
"epoch": 0.14294865311351607,
|
| 1877 |
+
"grad_norm": 27.789443969726562,
|
| 1878 |
+
"learning_rate": 0.00019954951005241248,
|
| 1879 |
+
"loss": 62.4471,
|
| 1880 |
+
"step": 265
|
| 1881 |
+
},
|
| 1882 |
+
{
|
| 1883 |
+
"epoch": 0.14348808199318971,
|
| 1884 |
+
"grad_norm": 21.813310623168945,
|
| 1885 |
+
"learning_rate": 0.0001995440371872594,
|
| 1886 |
+
"loss": 65.5364,
|
| 1887 |
+
"step": 266
|
| 1888 |
+
},
|
| 1889 |
+
{
|
| 1890 |
+
"epoch": 0.14402751087286336,
|
| 1891 |
+
"grad_norm": 22.338788986206055,
|
| 1892 |
+
"learning_rate": 0.00019953853135439522,
|
| 1893 |
+
"loss": 53.7872,
|
| 1894 |
+
"step": 267
|
| 1895 |
+
},
|
| 1896 |
+
{
|
| 1897 |
+
"epoch": 0.144566939752537,
|
| 1898 |
+
"grad_norm": 17.053470611572266,
|
| 1899 |
+
"learning_rate": 0.00019953299255564346,
|
| 1900 |
+
"loss": 46.6823,
|
| 1901 |
+
"step": 268
|
| 1902 |
+
},
|
| 1903 |
+
{
|
| 1904 |
+
"epoch": 0.14510636863221066,
|
| 1905 |
+
"grad_norm": 34.75794219970703,
|
| 1906 |
+
"learning_rate": 0.0001995274207928385,
|
| 1907 |
+
"loss": 32.208,
|
| 1908 |
+
"step": 269
|
| 1909 |
+
},
|
| 1910 |
+
{
|
| 1911 |
+
"epoch": 0.1456457975118843,
|
| 1912 |
+
"grad_norm": 76.52667236328125,
|
| 1913 |
+
"learning_rate": 0.00019952181606782565,
|
| 1914 |
+
"loss": 52.4054,
|
| 1915 |
+
"step": 270
|
| 1916 |
+
},
|
| 1917 |
+
{
|
| 1918 |
+
"epoch": 0.14618522639155793,
|
| 1919 |
+
"grad_norm": 71.48796844482422,
|
| 1920 |
+
"learning_rate": 0.00019951617838246107,
|
| 1921 |
+
"loss": 48.9668,
|
| 1922 |
+
"step": 271
|
| 1923 |
+
},
|
| 1924 |
+
{
|
| 1925 |
+
"epoch": 0.14672465527123157,
|
| 1926 |
+
"grad_norm": 79.96577453613281,
|
| 1927 |
+
"learning_rate": 0.00019951050773861192,
|
| 1928 |
+
"loss": 61.6082,
|
| 1929 |
+
"step": 272
|
| 1930 |
+
},
|
| 1931 |
+
{
|
| 1932 |
+
"epoch": 0.14726408415090522,
|
| 1933 |
+
"grad_norm": 42.05474090576172,
|
| 1934 |
+
"learning_rate": 0.0001995048041381562,
|
| 1935 |
+
"loss": 50.8627,
|
| 1936 |
+
"step": 273
|
| 1937 |
+
},
|
| 1938 |
+
{
|
| 1939 |
+
"epoch": 0.14780351303057887,
|
| 1940 |
+
"grad_norm": 43.19125747680664,
|
| 1941 |
+
"learning_rate": 0.00019949906758298295,
|
| 1942 |
+
"loss": 45.519,
|
| 1943 |
+
"step": 274
|
| 1944 |
+
},
|
| 1945 |
+
{
|
| 1946 |
+
"epoch": 0.14834294191025252,
|
| 1947 |
+
"grad_norm": 47.39426040649414,
|
| 1948 |
+
"learning_rate": 0.00019949329807499198,
|
| 1949 |
+
"loss": 51.654,
|
| 1950 |
+
"step": 275
|
| 1951 |
+
},
|
| 1952 |
+
{
|
| 1953 |
+
"epoch": 0.14888237078992617,
|
| 1954 |
+
"grad_norm": 36.0722770690918,
|
| 1955 |
+
"learning_rate": 0.00019948749561609415,
|
| 1956 |
+
"loss": 46.8854,
|
| 1957 |
+
"step": 276
|
| 1958 |
+
},
|
| 1959 |
+
{
|
| 1960 |
+
"epoch": 0.14942179966959981,
|
| 1961 |
+
"grad_norm": 33.252742767333984,
|
| 1962 |
+
"learning_rate": 0.00019948166020821107,
|
| 1963 |
+
"loss": 46.7532,
|
| 1964 |
+
"step": 277
|
| 1965 |
+
},
|
| 1966 |
+
{
|
| 1967 |
+
"epoch": 0.14996122854927346,
|
| 1968 |
+
"grad_norm": 33.89019012451172,
|
| 1969 |
+
"learning_rate": 0.0001994757918532754,
|
| 1970 |
+
"loss": 49.6403,
|
| 1971 |
+
"step": 278
|
| 1972 |
+
},
|
| 1973 |
+
{
|
| 1974 |
+
"epoch": 0.1505006574289471,
|
| 1975 |
+
"grad_norm": 37.914676666259766,
|
| 1976 |
+
"learning_rate": 0.00019946989055323066,
|
| 1977 |
+
"loss": 54.5018,
|
| 1978 |
+
"step": 279
|
| 1979 |
+
},
|
| 1980 |
+
{
|
| 1981 |
+
"epoch": 0.15104008630862076,
|
| 1982 |
+
"grad_norm": 37.611061096191406,
|
| 1983 |
+
"learning_rate": 0.00019946395631003128,
|
| 1984 |
+
"loss": 50.6423,
|
| 1985 |
+
"step": 280
|
| 1986 |
+
},
|
| 1987 |
+
{
|
| 1988 |
+
"epoch": 0.1515795151882944,
|
| 1989 |
+
"grad_norm": 36.489723205566406,
|
| 1990 |
+
"learning_rate": 0.00019945798912564264,
|
| 1991 |
+
"loss": 45.9299,
|
| 1992 |
+
"step": 281
|
| 1993 |
+
},
|
| 1994 |
+
{
|
| 1995 |
+
"epoch": 0.15211894406796803,
|
| 1996 |
+
"grad_norm": 31.33220100402832,
|
| 1997 |
+
"learning_rate": 0.00019945198900204095,
|
| 1998 |
+
"loss": 47.4519,
|
| 1999 |
+
"step": 282
|
| 2000 |
+
},
|
| 2001 |
+
{
|
| 2002 |
+
"epoch": 0.15265837294764167,
|
| 2003 |
+
"grad_norm": 32.4266242980957,
|
| 2004 |
+
"learning_rate": 0.00019944595594121337,
|
| 2005 |
+
"loss": 40.0806,
|
| 2006 |
+
"step": 283
|
| 2007 |
+
},
|
| 2008 |
+
{
|
| 2009 |
+
"epoch": 0.15319780182731532,
|
| 2010 |
+
"grad_norm": 38.17313003540039,
|
| 2011 |
+
"learning_rate": 0.00019943988994515797,
|
| 2012 |
+
"loss": 39.9765,
|
| 2013 |
+
"step": 284
|
| 2014 |
+
},
|
| 2015 |
+
{
|
| 2016 |
+
"epoch": 0.15373723070698897,
|
| 2017 |
+
"grad_norm": 40.299354553222656,
|
| 2018 |
+
"learning_rate": 0.00019943379101588376,
|
| 2019 |
+
"loss": 40.7812,
|
| 2020 |
+
"step": 285
|
| 2021 |
+
},
|
| 2022 |
+
{
|
| 2023 |
+
"epoch": 0.15427665958666262,
|
| 2024 |
+
"grad_norm": 42.34661102294922,
|
| 2025 |
+
"learning_rate": 0.00019942765915541063,
|
| 2026 |
+
"loss": 31.2513,
|
| 2027 |
+
"step": 286
|
| 2028 |
+
},
|
| 2029 |
+
{
|
| 2030 |
+
"epoch": 0.15481608846633627,
|
| 2031 |
+
"grad_norm": 46.61203384399414,
|
| 2032 |
+
"learning_rate": 0.00019942149436576938,
|
| 2033 |
+
"loss": 41.5619,
|
| 2034 |
+
"step": 287
|
| 2035 |
+
},
|
| 2036 |
+
{
|
| 2037 |
+
"epoch": 0.1553555173460099,
|
| 2038 |
+
"grad_norm": 39.79526901245117,
|
| 2039 |
+
"learning_rate": 0.00019941529664900168,
|
| 2040 |
+
"loss": 38.13,
|
| 2041 |
+
"step": 288
|
| 2042 |
+
},
|
| 2043 |
+
{
|
| 2044 |
+
"epoch": 0.15589494622568356,
|
| 2045 |
+
"grad_norm": 42.995567321777344,
|
| 2046 |
+
"learning_rate": 0.0001994090660071601,
|
| 2047 |
+
"loss": 41.3515,
|
| 2048 |
+
"step": 289
|
| 2049 |
+
},
|
| 2050 |
+
{
|
| 2051 |
+
"epoch": 0.1564343751053572,
|
| 2052 |
+
"grad_norm": 34.27892303466797,
|
| 2053 |
+
"learning_rate": 0.00019940280244230824,
|
| 2054 |
+
"loss": 41.1277,
|
| 2055 |
+
"step": 290
|
| 2056 |
+
},
|
| 2057 |
+
{
|
| 2058 |
+
"epoch": 0.15697380398503086,
|
| 2059 |
+
"grad_norm": 29.622488021850586,
|
| 2060 |
+
"learning_rate": 0.00019939650595652045,
|
| 2061 |
+
"loss": 49.2284,
|
| 2062 |
+
"step": 291
|
| 2063 |
+
},
|
| 2064 |
+
{
|
| 2065 |
+
"epoch": 0.1575132328647045,
|
| 2066 |
+
"grad_norm": 36.693119049072266,
|
| 2067 |
+
"learning_rate": 0.00019939017655188206,
|
| 2068 |
+
"loss": 35.5444,
|
| 2069 |
+
"step": 292
|
| 2070 |
+
},
|
| 2071 |
+
{
|
| 2072 |
+
"epoch": 0.15805266174437815,
|
| 2073 |
+
"grad_norm": 30.75679588317871,
|
| 2074 |
+
"learning_rate": 0.00019938381423048932,
|
| 2075 |
+
"loss": 34.9666,
|
| 2076 |
+
"step": 293
|
| 2077 |
+
},
|
| 2078 |
+
{
|
| 2079 |
+
"epoch": 0.15859209062405177,
|
| 2080 |
+
"grad_norm": 35.84019088745117,
|
| 2081 |
+
"learning_rate": 0.00019937741899444928,
|
| 2082 |
+
"loss": 39.4625,
|
| 2083 |
+
"step": 294
|
| 2084 |
+
},
|
| 2085 |
+
{
|
| 2086 |
+
"epoch": 0.15913151950372542,
|
| 2087 |
+
"grad_norm": 35.854496002197266,
|
| 2088 |
+
"learning_rate": 0.00019937099084588002,
|
| 2089 |
+
"loss": 37.2887,
|
| 2090 |
+
"step": 295
|
| 2091 |
+
},
|
| 2092 |
+
{
|
| 2093 |
+
"epoch": 0.15967094838339907,
|
| 2094 |
+
"grad_norm": 33.07613754272461,
|
| 2095 |
+
"learning_rate": 0.00019936452978691044,
|
| 2096 |
+
"loss": 34.5375,
|
| 2097 |
+
"step": 296
|
| 2098 |
+
},
|
| 2099 |
+
{
|
| 2100 |
+
"epoch": 0.16021037726307272,
|
| 2101 |
+
"grad_norm": 43.46371078491211,
|
| 2102 |
+
"learning_rate": 0.00019935803581968035,
|
| 2103 |
+
"loss": 30.3173,
|
| 2104 |
+
"step": 297
|
| 2105 |
+
},
|
| 2106 |
+
{
|
| 2107 |
+
"epoch": 0.16074980614274637,
|
| 2108 |
+
"grad_norm": 52.03241729736328,
|
| 2109 |
+
"learning_rate": 0.00019935150894634046,
|
| 2110 |
+
"loss": 42.4725,
|
| 2111 |
+
"step": 298
|
| 2112 |
+
},
|
| 2113 |
+
{
|
| 2114 |
+
"epoch": 0.16128923502242,
|
| 2115 |
+
"grad_norm": 50.36249542236328,
|
| 2116 |
+
"learning_rate": 0.00019934494916905245,
|
| 2117 |
+
"loss": 37.3647,
|
| 2118 |
+
"step": 299
|
| 2119 |
+
},
|
| 2120 |
+
{
|
| 2121 |
+
"epoch": 0.16182866390209366,
|
| 2122 |
+
"grad_norm": 41.50126647949219,
|
| 2123 |
+
"learning_rate": 0.00019933835648998875,
|
| 2124 |
+
"loss": 24.2931,
|
| 2125 |
+
"step": 300
|
| 2126 |
+
},
|
| 2127 |
+
{
|
| 2128 |
+
"epoch": 0.1623680927817673,
|
| 2129 |
+
"grad_norm": 31.253141403198242,
|
| 2130 |
+
"learning_rate": 0.00019933173091133286,
|
| 2131 |
+
"loss": 44.7853,
|
| 2132 |
+
"step": 301
|
| 2133 |
+
},
|
| 2134 |
+
{
|
| 2135 |
+
"epoch": 0.16290752166144096,
|
| 2136 |
+
"grad_norm": 96.83972930908203,
|
| 2137 |
+
"learning_rate": 0.000199325072435279,
|
| 2138 |
+
"loss": 84.9808,
|
| 2139 |
+
"step": 302
|
| 2140 |
+
},
|
| 2141 |
+
{
|
| 2142 |
+
"epoch": 0.1634469505411146,
|
| 2143 |
+
"grad_norm": 91.9966049194336,
|
| 2144 |
+
"learning_rate": 0.0001993183810640324,
|
| 2145 |
+
"loss": 99.5531,
|
| 2146 |
+
"step": 303
|
| 2147 |
+
},
|
| 2148 |
+
{
|
| 2149 |
+
"epoch": 0.16398637942078825,
|
| 2150 |
+
"grad_norm": 66.43877410888672,
|
| 2151 |
+
"learning_rate": 0.00019931165679980918,
|
| 2152 |
+
"loss": 105.7665,
|
| 2153 |
+
"step": 304
|
| 2154 |
+
},
|
| 2155 |
+
{
|
| 2156 |
+
"epoch": 0.16452580830046187,
|
| 2157 |
+
"grad_norm": 35.26411056518555,
|
| 2158 |
+
"learning_rate": 0.00019930489964483633,
|
| 2159 |
+
"loss": 109.6819,
|
| 2160 |
+
"step": 305
|
| 2161 |
+
},
|
| 2162 |
+
{
|
| 2163 |
+
"epoch": 0.16506523718013552,
|
| 2164 |
+
"grad_norm": 47.18457794189453,
|
| 2165 |
+
"learning_rate": 0.00019929810960135172,
|
| 2166 |
+
"loss": 113.4221,
|
| 2167 |
+
"step": 306
|
| 2168 |
+
},
|
| 2169 |
+
{
|
| 2170 |
+
"epoch": 0.16560466605980917,
|
| 2171 |
+
"grad_norm": 49.24475860595703,
|
| 2172 |
+
"learning_rate": 0.00019929128667160408,
|
| 2173 |
+
"loss": 108.0158,
|
| 2174 |
+
"step": 307
|
| 2175 |
+
},
|
| 2176 |
+
{
|
| 2177 |
+
"epoch": 0.16614409493948282,
|
| 2178 |
+
"grad_norm": 45.63924026489258,
|
| 2179 |
+
"learning_rate": 0.00019928443085785318,
|
| 2180 |
+
"loss": 94.1414,
|
| 2181 |
+
"step": 308
|
| 2182 |
+
},
|
| 2183 |
+
{
|
| 2184 |
+
"epoch": 0.16668352381915646,
|
| 2185 |
+
"grad_norm": 46.688350677490234,
|
| 2186 |
+
"learning_rate": 0.00019927754216236948,
|
| 2187 |
+
"loss": 87.8688,
|
| 2188 |
+
"step": 309
|
| 2189 |
+
},
|
| 2190 |
+
{
|
| 2191 |
+
"epoch": 0.1672229526988301,
|
| 2192 |
+
"grad_norm": 39.54045486450195,
|
| 2193 |
+
"learning_rate": 0.00019927062058743448,
|
| 2194 |
+
"loss": 92.6019,
|
| 2195 |
+
"step": 310
|
| 2196 |
+
},
|
| 2197 |
+
{
|
| 2198 |
+
"epoch": 0.16776238157850376,
|
| 2199 |
+
"grad_norm": 29.866121292114258,
|
| 2200 |
+
"learning_rate": 0.0001992636661353405,
|
| 2201 |
+
"loss": 81.9024,
|
| 2202 |
+
"step": 311
|
| 2203 |
+
},
|
| 2204 |
+
{
|
| 2205 |
+
"epoch": 0.1683018104581774,
|
| 2206 |
+
"grad_norm": 22.350112915039062,
|
| 2207 |
+
"learning_rate": 0.0001992566788083908,
|
| 2208 |
+
"loss": 68.4321,
|
| 2209 |
+
"step": 312
|
| 2210 |
+
},
|
| 2211 |
+
{
|
| 2212 |
+
"epoch": 0.16884123933785106,
|
| 2213 |
+
"grad_norm": 21.657258987426758,
|
| 2214 |
+
"learning_rate": 0.00019924965860889944,
|
| 2215 |
+
"loss": 65.7434,
|
| 2216 |
+
"step": 313
|
| 2217 |
+
},
|
| 2218 |
+
{
|
| 2219 |
+
"epoch": 0.1693806682175247,
|
| 2220 |
+
"grad_norm": 18.347572326660156,
|
| 2221 |
+
"learning_rate": 0.00019924260553919146,
|
| 2222 |
+
"loss": 62.485,
|
| 2223 |
+
"step": 314
|
| 2224 |
+
},
|
| 2225 |
+
{
|
| 2226 |
+
"epoch": 0.16992009709719835,
|
| 2227 |
+
"grad_norm": 28.368114471435547,
|
| 2228 |
+
"learning_rate": 0.00019923551960160268,
|
| 2229 |
+
"loss": 53.7759,
|
| 2230 |
+
"step": 315
|
| 2231 |
+
},
|
| 2232 |
+
{
|
| 2233 |
+
"epoch": 0.170459525976872,
|
| 2234 |
+
"grad_norm": 35.214988708496094,
|
| 2235 |
+
"learning_rate": 0.00019922840079848,
|
| 2236 |
+
"loss": 45.4414,
|
| 2237 |
+
"step": 316
|
| 2238 |
+
},
|
| 2239 |
+
{
|
| 2240 |
+
"epoch": 0.17099895485654562,
|
| 2241 |
+
"grad_norm": 38.698760986328125,
|
| 2242 |
+
"learning_rate": 0.00019922124913218094,
|
| 2243 |
+
"loss": 37.665,
|
| 2244 |
+
"step": 317
|
| 2245 |
+
},
|
| 2246 |
+
{
|
| 2247 |
+
"epoch": 0.17153838373621927,
|
| 2248 |
+
"grad_norm": 43.39471435546875,
|
| 2249 |
+
"learning_rate": 0.0001992140646050741,
|
| 2250 |
+
"loss": 51.4899,
|
| 2251 |
+
"step": 318
|
| 2252 |
+
},
|
| 2253 |
+
{
|
| 2254 |
+
"epoch": 0.17207781261589292,
|
| 2255 |
+
"grad_norm": 43.52251434326172,
|
| 2256 |
+
"learning_rate": 0.00019920684721953894,
|
| 2257 |
+
"loss": 48.5712,
|
| 2258 |
+
"step": 319
|
| 2259 |
+
},
|
| 2260 |
+
{
|
| 2261 |
+
"epoch": 0.17261724149556656,
|
| 2262 |
+
"grad_norm": 60.897579193115234,
|
| 2263 |
+
"learning_rate": 0.00019919959697796568,
|
| 2264 |
+
"loss": 59.9231,
|
| 2265 |
+
"step": 320
|
| 2266 |
+
},
|
| 2267 |
+
{
|
| 2268 |
+
"epoch": 0.1731566703752402,
|
| 2269 |
+
"grad_norm": 37.93972396850586,
|
| 2270 |
+
"learning_rate": 0.0001991923138827556,
|
| 2271 |
+
"loss": 47.906,
|
| 2272 |
+
"step": 321
|
| 2273 |
+
},
|
| 2274 |
+
{
|
| 2275 |
+
"epoch": 0.17369609925491386,
|
| 2276 |
+
"grad_norm": 44.32222366333008,
|
| 2277 |
+
"learning_rate": 0.0001991849979363207,
|
| 2278 |
+
"loss": 54.5404,
|
| 2279 |
+
"step": 322
|
| 2280 |
+
},
|
| 2281 |
+
{
|
| 2282 |
+
"epoch": 0.1742355281345875,
|
| 2283 |
+
"grad_norm": 37.367671966552734,
|
| 2284 |
+
"learning_rate": 0.00019917764914108394,
|
| 2285 |
+
"loss": 49.3113,
|
| 2286 |
+
"step": 323
|
| 2287 |
+
},
|
| 2288 |
+
{
|
| 2289 |
+
"epoch": 0.17477495701426116,
|
| 2290 |
+
"grad_norm": 43.20479965209961,
|
| 2291 |
+
"learning_rate": 0.00019917026749947917,
|
| 2292 |
+
"loss": 41.9015,
|
| 2293 |
+
"step": 324
|
| 2294 |
+
},
|
| 2295 |
+
{
|
| 2296 |
+
"epoch": 0.1753143858939348,
|
| 2297 |
+
"grad_norm": 36.7598991394043,
|
| 2298 |
+
"learning_rate": 0.0001991628530139511,
|
| 2299 |
+
"loss": 43.7222,
|
| 2300 |
+
"step": 325
|
| 2301 |
+
},
|
| 2302 |
+
{
|
| 2303 |
+
"epoch": 0.17585381477360845,
|
| 2304 |
+
"grad_norm": 33.30655288696289,
|
| 2305 |
+
"learning_rate": 0.0001991554056869553,
|
| 2306 |
+
"loss": 48.4387,
|
| 2307 |
+
"step": 326
|
| 2308 |
+
},
|
| 2309 |
+
{
|
| 2310 |
+
"epoch": 0.1763932436532821,
|
| 2311 |
+
"grad_norm": 32.89339828491211,
|
| 2312 |
+
"learning_rate": 0.00019914792552095818,
|
| 2313 |
+
"loss": 51.108,
|
| 2314 |
+
"step": 327
|
| 2315 |
+
},
|
| 2316 |
+
{
|
| 2317 |
+
"epoch": 0.17693267253295572,
|
| 2318 |
+
"grad_norm": 31.422489166259766,
|
| 2319 |
+
"learning_rate": 0.00019914041251843716,
|
| 2320 |
+
"loss": 42.9287,
|
| 2321 |
+
"step": 328
|
| 2322 |
+
},
|
| 2323 |
+
{
|
| 2324 |
+
"epoch": 0.17747210141262937,
|
| 2325 |
+
"grad_norm": 33.38264465332031,
|
| 2326 |
+
"learning_rate": 0.00019913286668188037,
|
| 2327 |
+
"loss": 47.0867,
|
| 2328 |
+
"step": 329
|
| 2329 |
+
},
|
| 2330 |
+
{
|
| 2331 |
+
"epoch": 0.17801153029230302,
|
| 2332 |
+
"grad_norm": 37.976837158203125,
|
| 2333 |
+
"learning_rate": 0.00019912528801378698,
|
| 2334 |
+
"loss": 38.2593,
|
| 2335 |
+
"step": 330
|
| 2336 |
+
},
|
| 2337 |
+
{
|
| 2338 |
+
"epoch": 0.17855095917197666,
|
| 2339 |
+
"grad_norm": 35.707054138183594,
|
| 2340 |
+
"learning_rate": 0.0001991176765166669,
|
| 2341 |
+
"loss": 44.5348,
|
| 2342 |
+
"step": 331
|
| 2343 |
+
},
|
| 2344 |
+
{
|
| 2345 |
+
"epoch": 0.1790903880516503,
|
| 2346 |
+
"grad_norm": 43.86237335205078,
|
| 2347 |
+
"learning_rate": 0.00019911003219304094,
|
| 2348 |
+
"loss": 40.4868,
|
| 2349 |
+
"step": 332
|
| 2350 |
+
},
|
| 2351 |
+
{
|
| 2352 |
+
"epoch": 0.17962981693132396,
|
| 2353 |
+
"grad_norm": 54.88194274902344,
|
| 2354 |
+
"learning_rate": 0.00019910235504544082,
|
| 2355 |
+
"loss": 38.935,
|
| 2356 |
+
"step": 333
|
| 2357 |
+
},
|
| 2358 |
+
{
|
| 2359 |
+
"epoch": 0.1801692458109976,
|
| 2360 |
+
"grad_norm": 43.87349319458008,
|
| 2361 |
+
"learning_rate": 0.00019909464507640915,
|
| 2362 |
+
"loss": 43.0978,
|
| 2363 |
+
"step": 334
|
| 2364 |
+
},
|
| 2365 |
+
{
|
| 2366 |
+
"epoch": 0.18070867469067126,
|
| 2367 |
+
"grad_norm": 43.421932220458984,
|
| 2368 |
+
"learning_rate": 0.0001990869022884993,
|
| 2369 |
+
"loss": 39.2888,
|
| 2370 |
+
"step": 335
|
| 2371 |
+
},
|
| 2372 |
+
{
|
| 2373 |
+
"epoch": 0.1812481035703449,
|
| 2374 |
+
"grad_norm": 41.14269256591797,
|
| 2375 |
+
"learning_rate": 0.00019907912668427566,
|
| 2376 |
+
"loss": 42.6139,
|
| 2377 |
+
"step": 336
|
| 2378 |
+
},
|
| 2379 |
+
{
|
| 2380 |
+
"epoch": 0.18178753245001855,
|
| 2381 |
+
"grad_norm": 38.619380950927734,
|
| 2382 |
+
"learning_rate": 0.00019907131826631336,
|
| 2383 |
+
"loss": 40.0248,
|
| 2384 |
+
"step": 337
|
| 2385 |
+
},
|
| 2386 |
+
{
|
| 2387 |
+
"epoch": 0.1823269613296922,
|
| 2388 |
+
"grad_norm": 33.65724563598633,
|
| 2389 |
+
"learning_rate": 0.00019906347703719845,
|
| 2390 |
+
"loss": 38.7406,
|
| 2391 |
+
"step": 338
|
| 2392 |
+
},
|
| 2393 |
+
{
|
| 2394 |
+
"epoch": 0.18286639020936585,
|
| 2395 |
+
"grad_norm": 35.25956344604492,
|
| 2396 |
+
"learning_rate": 0.0001990556029995279,
|
| 2397 |
+
"loss": 39.2734,
|
| 2398 |
+
"step": 339
|
| 2399 |
+
},
|
| 2400 |
+
{
|
| 2401 |
+
"epoch": 0.18340581908903947,
|
| 2402 |
+
"grad_norm": 36.87468719482422,
|
| 2403 |
+
"learning_rate": 0.00019904769615590942,
|
| 2404 |
+
"loss": 40.6619,
|
| 2405 |
+
"step": 340
|
| 2406 |
+
},
|
| 2407 |
+
{
|
| 2408 |
+
"epoch": 0.18394524796871312,
|
| 2409 |
+
"grad_norm": 32.0380973815918,
|
| 2410 |
+
"learning_rate": 0.00019903975650896168,
|
| 2411 |
+
"loss": 39.8376,
|
| 2412 |
+
"step": 341
|
| 2413 |
+
},
|
| 2414 |
+
{
|
| 2415 |
+
"epoch": 0.18448467684838676,
|
| 2416 |
+
"grad_norm": 33.44660949707031,
|
| 2417 |
+
"learning_rate": 0.0001990317840613142,
|
| 2418 |
+
"loss": 33.338,
|
| 2419 |
+
"step": 342
|
| 2420 |
+
},
|
| 2421 |
+
{
|
| 2422 |
+
"epoch": 0.1850241057280604,
|
| 2423 |
+
"grad_norm": 36.242523193359375,
|
| 2424 |
+
"learning_rate": 0.00019902377881560735,
|
| 2425 |
+
"loss": 35.0493,
|
| 2426 |
+
"step": 343
|
| 2427 |
+
},
|
| 2428 |
+
{
|
| 2429 |
+
"epoch": 0.18556353460773406,
|
| 2430 |
+
"grad_norm": 37.39813232421875,
|
| 2431 |
+
"learning_rate": 0.00019901574077449232,
|
| 2432 |
+
"loss": 26.9563,
|
| 2433 |
+
"step": 344
|
| 2434 |
+
},
|
| 2435 |
+
{
|
| 2436 |
+
"epoch": 0.1861029634874077,
|
| 2437 |
+
"grad_norm": 35.84196472167969,
|
| 2438 |
+
"learning_rate": 0.0001990076699406313,
|
| 2439 |
+
"loss": 33.7825,
|
| 2440 |
+
"step": 345
|
| 2441 |
+
},
|
| 2442 |
+
{
|
| 2443 |
+
"epoch": 0.18664239236708136,
|
| 2444 |
+
"grad_norm": 38.69563293457031,
|
| 2445 |
+
"learning_rate": 0.00019899956631669717,
|
| 2446 |
+
"loss": 29.9582,
|
| 2447 |
+
"step": 346
|
| 2448 |
+
},
|
| 2449 |
+
{
|
| 2450 |
+
"epoch": 0.187181821246755,
|
| 2451 |
+
"grad_norm": 47.82805633544922,
|
| 2452 |
+
"learning_rate": 0.00019899142990537376,
|
| 2453 |
+
"loss": 33.9471,
|
| 2454 |
+
"step": 347
|
| 2455 |
+
},
|
| 2456 |
+
{
|
| 2457 |
+
"epoch": 0.18772125012642865,
|
| 2458 |
+
"grad_norm": 36.29233169555664,
|
| 2459 |
+
"learning_rate": 0.00019898326070935579,
|
| 2460 |
+
"loss": 28.1711,
|
| 2461 |
+
"step": 348
|
| 2462 |
+
},
|
| 2463 |
+
{
|
| 2464 |
+
"epoch": 0.1882606790061023,
|
| 2465 |
+
"grad_norm": 45.26416015625,
|
| 2466 |
+
"learning_rate": 0.00019897505873134872,
|
| 2467 |
+
"loss": 33.76,
|
| 2468 |
+
"step": 349
|
| 2469 |
+
},
|
| 2470 |
+
{
|
| 2471 |
+
"epoch": 0.18880010788577595,
|
| 2472 |
+
"grad_norm": 39.766441345214844,
|
| 2473 |
+
"learning_rate": 0.000198966823974069,
|
| 2474 |
+
"loss": 25.629,
|
| 2475 |
+
"step": 350
|
| 2476 |
+
},
|
| 2477 |
+
{
|
| 2478 |
+
"epoch": 0.18933953676544957,
|
| 2479 |
+
"grad_norm": 30.092906951904297,
|
| 2480 |
+
"learning_rate": 0.00019895855644024387,
|
| 2481 |
+
"loss": 45.1687,
|
| 2482 |
+
"step": 351
|
| 2483 |
+
},
|
| 2484 |
+
{
|
| 2485 |
+
"epoch": 0.18987896564512322,
|
| 2486 |
+
"grad_norm": 61.02379608154297,
|
| 2487 |
+
"learning_rate": 0.00019895025613261136,
|
| 2488 |
+
"loss": 77.4727,
|
| 2489 |
+
"step": 352
|
| 2490 |
+
},
|
| 2491 |
+
{
|
| 2492 |
+
"epoch": 0.19041839452479686,
|
| 2493 |
+
"grad_norm": 51.788063049316406,
|
| 2494 |
+
"learning_rate": 0.00019894192305392055,
|
| 2495 |
+
"loss": 82.3816,
|
| 2496 |
+
"step": 353
|
| 2497 |
+
},
|
| 2498 |
+
{
|
| 2499 |
+
"epoch": 0.1909578234044705,
|
| 2500 |
+
"grad_norm": 72.1239242553711,
|
| 2501 |
+
"learning_rate": 0.0001989335572069311,
|
| 2502 |
+
"loss": 103.2545,
|
| 2503 |
+
"step": 354
|
| 2504 |
+
},
|
| 2505 |
+
{
|
| 2506 |
+
"epoch": 0.19149725228414416,
|
| 2507 |
+
"grad_norm": 29.279748916625977,
|
| 2508 |
+
"learning_rate": 0.00019892515859441383,
|
| 2509 |
+
"loss": 113.7908,
|
| 2510 |
+
"step": 355
|
| 2511 |
+
},
|
| 2512 |
+
{
|
| 2513 |
+
"epoch": 0.1920366811638178,
|
| 2514 |
+
"grad_norm": 43.08776092529297,
|
| 2515 |
+
"learning_rate": 0.00019891672721915015,
|
| 2516 |
+
"loss": 107.6541,
|
| 2517 |
+
"step": 356
|
| 2518 |
+
},
|
| 2519 |
+
{
|
| 2520 |
+
"epoch": 0.19257611004349146,
|
| 2521 |
+
"grad_norm": 54.121192932128906,
|
| 2522 |
+
"learning_rate": 0.00019890826308393243,
|
| 2523 |
+
"loss": 102.3774,
|
| 2524 |
+
"step": 357
|
| 2525 |
+
},
|
| 2526 |
+
{
|
| 2527 |
+
"epoch": 0.1931155389231651,
|
| 2528 |
+
"grad_norm": 52.771793365478516,
|
| 2529 |
+
"learning_rate": 0.0001988997661915639,
|
| 2530 |
+
"loss": 87.3872,
|
| 2531 |
+
"step": 358
|
| 2532 |
+
},
|
| 2533 |
+
{
|
| 2534 |
+
"epoch": 0.19365496780283875,
|
| 2535 |
+
"grad_norm": 58.10847854614258,
|
| 2536 |
+
"learning_rate": 0.00019889123654485866,
|
| 2537 |
+
"loss": 97.106,
|
| 2538 |
+
"step": 359
|
| 2539 |
+
},
|
| 2540 |
+
{
|
| 2541 |
+
"epoch": 0.1941943966825124,
|
| 2542 |
+
"grad_norm": 52.38351058959961,
|
| 2543 |
+
"learning_rate": 0.00019888267414664156,
|
| 2544 |
+
"loss": 91.256,
|
| 2545 |
+
"step": 360
|
| 2546 |
+
},
|
| 2547 |
+
{
|
| 2548 |
+
"epoch": 0.19473382556218605,
|
| 2549 |
+
"grad_norm": 48.153804779052734,
|
| 2550 |
+
"learning_rate": 0.0001988740789997484,
|
| 2551 |
+
"loss": 81.894,
|
| 2552 |
+
"step": 361
|
| 2553 |
+
},
|
| 2554 |
+
{
|
| 2555 |
+
"epoch": 0.19527325444185967,
|
| 2556 |
+
"grad_norm": 25.811304092407227,
|
| 2557 |
+
"learning_rate": 0.00019886545110702576,
|
| 2558 |
+
"loss": 69.6325,
|
| 2559 |
+
"step": 362
|
| 2560 |
+
},
|
| 2561 |
+
{
|
| 2562 |
+
"epoch": 0.19581268332153332,
|
| 2563 |
+
"grad_norm": 22.911964416503906,
|
| 2564 |
+
"learning_rate": 0.00019885679047133107,
|
| 2565 |
+
"loss": 65.5302,
|
| 2566 |
+
"step": 363
|
| 2567 |
+
},
|
| 2568 |
+
{
|
| 2569 |
+
"epoch": 0.19635211220120696,
|
| 2570 |
+
"grad_norm": 37.54278564453125,
|
| 2571 |
+
"learning_rate": 0.00019884809709553265,
|
| 2572 |
+
"loss": 60.65,
|
| 2573 |
+
"step": 364
|
| 2574 |
+
},
|
| 2575 |
+
{
|
| 2576 |
+
"epoch": 0.1968915410808806,
|
| 2577 |
+
"grad_norm": 20.303857803344727,
|
| 2578 |
+
"learning_rate": 0.00019883937098250963,
|
| 2579 |
+
"loss": 44.1299,
|
| 2580 |
+
"step": 365
|
| 2581 |
+
},
|
| 2582 |
+
{
|
| 2583 |
+
"epoch": 0.19743096996055426,
|
| 2584 |
+
"grad_norm": 31.87704849243164,
|
| 2585 |
+
"learning_rate": 0.00019883061213515197,
|
| 2586 |
+
"loss": 34.1489,
|
| 2587 |
+
"step": 366
|
| 2588 |
+
},
|
| 2589 |
+
{
|
| 2590 |
+
"epoch": 0.1979703988402279,
|
| 2591 |
+
"grad_norm": 39.10615539550781,
|
| 2592 |
+
"learning_rate": 0.00019882182055636053,
|
| 2593 |
+
"loss": 37.5989,
|
| 2594 |
+
"step": 367
|
| 2595 |
+
},
|
| 2596 |
+
{
|
| 2597 |
+
"epoch": 0.19850982771990155,
|
| 2598 |
+
"grad_norm": 41.10018539428711,
|
| 2599 |
+
"learning_rate": 0.00019881299624904692,
|
| 2600 |
+
"loss": 48.6169,
|
| 2601 |
+
"step": 368
|
| 2602 |
+
},
|
| 2603 |
+
{
|
| 2604 |
+
"epoch": 0.1990492565995752,
|
| 2605 |
+
"grad_norm": 34.8628044128418,
|
| 2606 |
+
"learning_rate": 0.00019880413921613367,
|
| 2607 |
+
"loss": 51.3889,
|
| 2608 |
+
"step": 369
|
| 2609 |
+
},
|
| 2610 |
+
{
|
| 2611 |
+
"epoch": 0.19958868547924885,
|
| 2612 |
+
"grad_norm": 41.81850051879883,
|
| 2613 |
+
"learning_rate": 0.0001987952494605541,
|
| 2614 |
+
"loss": 46.2857,
|
| 2615 |
+
"step": 370
|
| 2616 |
+
},
|
| 2617 |
+
{
|
| 2618 |
+
"epoch": 0.2001281143589225,
|
| 2619 |
+
"grad_norm": 46.00803756713867,
|
| 2620 |
+
"learning_rate": 0.00019878632698525238,
|
| 2621 |
+
"loss": 42.1201,
|
| 2622 |
+
"step": 371
|
| 2623 |
+
},
|
| 2624 |
+
{
|
| 2625 |
+
"epoch": 0.20066754323859615,
|
| 2626 |
+
"grad_norm": 37.3172492980957,
|
| 2627 |
+
"learning_rate": 0.00019877737179318353,
|
| 2628 |
+
"loss": 44.8517,
|
| 2629 |
+
"step": 372
|
| 2630 |
+
},
|
| 2631 |
+
{
|
| 2632 |
+
"epoch": 0.2012069721182698,
|
| 2633 |
+
"grad_norm": 30.38181495666504,
|
| 2634 |
+
"learning_rate": 0.0001987683838873134,
|
| 2635 |
+
"loss": 30.3321,
|
| 2636 |
+
"step": 373
|
| 2637 |
+
},
|
| 2638 |
+
{
|
| 2639 |
+
"epoch": 0.20174640099794341,
|
| 2640 |
+
"grad_norm": 36.00757598876953,
|
| 2641 |
+
"learning_rate": 0.00019875936327061865,
|
| 2642 |
+
"loss": 41.3805,
|
| 2643 |
+
"step": 374
|
| 2644 |
+
},
|
| 2645 |
+
{
|
| 2646 |
+
"epoch": 0.20228582987761706,
|
| 2647 |
+
"grad_norm": 36.742733001708984,
|
| 2648 |
+
"learning_rate": 0.00019875030994608684,
|
| 2649 |
+
"loss": 48.6651,
|
| 2650 |
+
"step": 375
|
| 2651 |
+
},
|
| 2652 |
+
{
|
| 2653 |
+
"epoch": 0.2028252587572907,
|
| 2654 |
+
"grad_norm": 42.53518295288086,
|
| 2655 |
+
"learning_rate": 0.00019874122391671622,
|
| 2656 |
+
"loss": 32.5649,
|
| 2657 |
+
"step": 376
|
| 2658 |
+
},
|
| 2659 |
+
{
|
| 2660 |
+
"epoch": 0.20336468763696436,
|
| 2661 |
+
"grad_norm": 35.77900314331055,
|
| 2662 |
+
"learning_rate": 0.00019873210518551608,
|
| 2663 |
+
"loss": 46.6955,
|
| 2664 |
+
"step": 377
|
| 2665 |
+
},
|
| 2666 |
+
{
|
| 2667 |
+
"epoch": 0.203904116516638,
|
| 2668 |
+
"grad_norm": 44.95616149902344,
|
| 2669 |
+
"learning_rate": 0.00019872295375550635,
|
| 2670 |
+
"loss": 41.271,
|
| 2671 |
+
"step": 378
|
| 2672 |
+
},
|
| 2673 |
+
{
|
| 2674 |
+
"epoch": 0.20444354539631165,
|
| 2675 |
+
"grad_norm": 34.28546142578125,
|
| 2676 |
+
"learning_rate": 0.00019871376962971789,
|
| 2677 |
+
"loss": 41.4059,
|
| 2678 |
+
"step": 379
|
| 2679 |
+
},
|
| 2680 |
+
{
|
| 2681 |
+
"epoch": 0.2049829742759853,
|
| 2682 |
+
"grad_norm": 35.807682037353516,
|
| 2683 |
+
"learning_rate": 0.00019870455281119237,
|
| 2684 |
+
"loss": 45.8892,
|
| 2685 |
+
"step": 380
|
| 2686 |
+
},
|
| 2687 |
+
{
|
| 2688 |
+
"epoch": 0.20552240315565895,
|
| 2689 |
+
"grad_norm": 30.27015495300293,
|
| 2690 |
+
"learning_rate": 0.00019869530330298227,
|
| 2691 |
+
"loss": 34.013,
|
| 2692 |
+
"step": 381
|
| 2693 |
+
},
|
| 2694 |
+
{
|
| 2695 |
+
"epoch": 0.2060618320353326,
|
| 2696 |
+
"grad_norm": 38.26789093017578,
|
| 2697 |
+
"learning_rate": 0.00019868602110815093,
|
| 2698 |
+
"loss": 42.6953,
|
| 2699 |
+
"step": 382
|
| 2700 |
+
},
|
| 2701 |
+
{
|
| 2702 |
+
"epoch": 0.20660126091500625,
|
| 2703 |
+
"grad_norm": 39.61716079711914,
|
| 2704 |
+
"learning_rate": 0.00019867670622977248,
|
| 2705 |
+
"loss": 40.4979,
|
| 2706 |
+
"step": 383
|
| 2707 |
+
},
|
| 2708 |
+
{
|
| 2709 |
+
"epoch": 0.2071406897946799,
|
| 2710 |
+
"grad_norm": 35.717227935791016,
|
| 2711 |
+
"learning_rate": 0.00019866735867093188,
|
| 2712 |
+
"loss": 31.5146,
|
| 2713 |
+
"step": 384
|
| 2714 |
+
},
|
| 2715 |
+
{
|
| 2716 |
+
"epoch": 0.20768011867435351,
|
| 2717 |
+
"grad_norm": 43.41541290283203,
|
| 2718 |
+
"learning_rate": 0.0001986579784347249,
|
| 2719 |
+
"loss": 37.5416,
|
| 2720 |
+
"step": 385
|
| 2721 |
+
},
|
| 2722 |
+
{
|
| 2723 |
+
"epoch": 0.20821954755402716,
|
| 2724 |
+
"grad_norm": 40.18928146362305,
|
| 2725 |
+
"learning_rate": 0.0001986485655242582,
|
| 2726 |
+
"loss": 39.0367,
|
| 2727 |
+
"step": 386
|
| 2728 |
+
},
|
| 2729 |
+
{
|
| 2730 |
+
"epoch": 0.2087589764337008,
|
| 2731 |
+
"grad_norm": 35.295291900634766,
|
| 2732 |
+
"learning_rate": 0.00019863911994264926,
|
| 2733 |
+
"loss": 36.8243,
|
| 2734 |
+
"step": 387
|
| 2735 |
+
},
|
| 2736 |
+
{
|
| 2737 |
+
"epoch": 0.20929840531337446,
|
| 2738 |
+
"grad_norm": 52.24161148071289,
|
| 2739 |
+
"learning_rate": 0.00019862964169302621,
|
| 2740 |
+
"loss": 41.7241,
|
| 2741 |
+
"step": 388
|
| 2742 |
+
},
|
| 2743 |
+
{
|
| 2744 |
+
"epoch": 0.2098378341930481,
|
| 2745 |
+
"grad_norm": 53.32133483886719,
|
| 2746 |
+
"learning_rate": 0.00019862013077852822,
|
| 2747 |
+
"loss": 38.7999,
|
| 2748 |
+
"step": 389
|
| 2749 |
+
},
|
| 2750 |
+
{
|
| 2751 |
+
"epoch": 0.21037726307272175,
|
| 2752 |
+
"grad_norm": 42.945804595947266,
|
| 2753 |
+
"learning_rate": 0.00019861058720230514,
|
| 2754 |
+
"loss": 34.0199,
|
| 2755 |
+
"step": 390
|
| 2756 |
+
},
|
| 2757 |
+
{
|
| 2758 |
+
"epoch": 0.2109166919523954,
|
| 2759 |
+
"grad_norm": 38.77582931518555,
|
| 2760 |
+
"learning_rate": 0.00019860101096751768,
|
| 2761 |
+
"loss": 33.4203,
|
| 2762 |
+
"step": 391
|
| 2763 |
+
},
|
| 2764 |
+
{
|
| 2765 |
+
"epoch": 0.21145612083206905,
|
| 2766 |
+
"grad_norm": 30.80617332458496,
|
| 2767 |
+
"learning_rate": 0.0001985914020773374,
|
| 2768 |
+
"loss": 27.0483,
|
| 2769 |
+
"step": 392
|
| 2770 |
+
},
|
| 2771 |
+
{
|
| 2772 |
+
"epoch": 0.2119955497117427,
|
| 2773 |
+
"grad_norm": 43.676090240478516,
|
| 2774 |
+
"learning_rate": 0.00019858176053494663,
|
| 2775 |
+
"loss": 33.954,
|
| 2776 |
+
"step": 393
|
| 2777 |
+
},
|
| 2778 |
+
{
|
| 2779 |
+
"epoch": 0.21253497859141635,
|
| 2780 |
+
"grad_norm": 38.32650375366211,
|
| 2781 |
+
"learning_rate": 0.00019857208634353852,
|
| 2782 |
+
"loss": 29.378,
|
| 2783 |
+
"step": 394
|
| 2784 |
+
},
|
| 2785 |
+
{
|
| 2786 |
+
"epoch": 0.21307440747109,
|
| 2787 |
+
"grad_norm": 39.12830352783203,
|
| 2788 |
+
"learning_rate": 0.000198562379506317,
|
| 2789 |
+
"loss": 27.9634,
|
| 2790 |
+
"step": 395
|
| 2791 |
+
},
|
| 2792 |
+
{
|
| 2793 |
+
"epoch": 0.21361383635076364,
|
| 2794 |
+
"grad_norm": 47.39609909057617,
|
| 2795 |
+
"learning_rate": 0.00019855264002649692,
|
| 2796 |
+
"loss": 34.1847,
|
| 2797 |
+
"step": 396
|
| 2798 |
+
},
|
| 2799 |
+
{
|
| 2800 |
+
"epoch": 0.21415326523043726,
|
| 2801 |
+
"grad_norm": 38.62258529663086,
|
| 2802 |
+
"learning_rate": 0.00019854286790730384,
|
| 2803 |
+
"loss": 26.0765,
|
| 2804 |
+
"step": 397
|
| 2805 |
+
},
|
| 2806 |
+
{
|
| 2807 |
+
"epoch": 0.2146926941101109,
|
| 2808 |
+
"grad_norm": 42.81424331665039,
|
| 2809 |
+
"learning_rate": 0.00019853306315197413,
|
| 2810 |
+
"loss": 34.1509,
|
| 2811 |
+
"step": 398
|
| 2812 |
+
},
|
| 2813 |
+
{
|
| 2814 |
+
"epoch": 0.21523212298978456,
|
| 2815 |
+
"grad_norm": 45.57196807861328,
|
| 2816 |
+
"learning_rate": 0.00019852322576375503,
|
| 2817 |
+
"loss": 32.0371,
|
| 2818 |
+
"step": 399
|
| 2819 |
+
},
|
| 2820 |
+
{
|
| 2821 |
+
"epoch": 0.2157715518694582,
|
| 2822 |
+
"grad_norm": 35.20758819580078,
|
| 2823 |
+
"learning_rate": 0.0001985133557459046,
|
| 2824 |
+
"loss": 20.3634,
|
| 2825 |
+
"step": 400
|
| 2826 |
+
},
|
| 2827 |
+
{
|
| 2828 |
+
"epoch": 0.2157715518694582,
|
| 2829 |
+
"eval_loss": 1.6627388000488281,
|
| 2830 |
+
"eval_runtime": 141.0153,
|
| 2831 |
+
"eval_samples_per_second": 2.12,
|
| 2832 |
+
"eval_steps_per_second": 2.12,
|
| 2833 |
+
"step": 400
|
| 2834 |
}
|
| 2835 |
],
|
| 2836 |
"logging_steps": 1,
|
|
|
|
| 2859 |
"attributes": {}
|
| 2860 |
}
|
| 2861 |
},
|
| 2862 |
+
"total_flos": 7.146078851574006e+17,
|
| 2863 |
"train_batch_size": 1,
|
| 2864 |
"trial_name": null,
|
| 2865 |
"trial_params": null
|