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 201361312
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:23e7a6d5b7aad440e78deec56146ff348af3d4e8f8beae15cf733c5f24d05919
|
| 3 |
size 201361312
|
last-checkpoint/optimizer.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
-
size
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:51ada906bb55415a9b47ce9b7358d4c6ec8104de8825a3ecb6529237e55d1768
|
| 3 |
+
size 102537812
|
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 sha256:6754fab244ca44d88a841a60271b5dc8c5d8d98d90006bf01abc71d28f65b63d
|
| 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 sha256:23d544576b5cbfe5ea315089265e38649a716f01fec6390e5015f8bf833d3574
|
| 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": 8.738,
|
| 1424 |
"eval_steps_per_second": 4.374,
|
| 1425 |
"step": 200
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1426 |
}
|
| 1427 |
],
|
| 1428 |
"logging_steps": 1,
|
|
@@ -1451,7 +2859,7 @@
|
|
| 1451 |
"attributes": {}
|
| 1452 |
}
|
| 1453 |
},
|
| 1454 |
-
"total_flos":
|
| 1455 |
"train_batch_size": 2,
|
| 1456 |
"trial_name": null,
|
| 1457 |
"trial_params": null
|
|
|
|
| 1 |
{
|
| 2 |
+
"best_metric": 1.665988564491272,
|
| 3 |
+
"best_model_checkpoint": "miner_id_24/checkpoint-400",
|
| 4 |
+
"epoch": 0.019090572836501174,
|
| 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": 8.738,
|
| 1424 |
"eval_steps_per_second": 4.374,
|
| 1425 |
"step": 200
|
| 1426 |
+
},
|
| 1427 |
+
{
|
| 1428 |
+
"epoch": 0.00959301285034184,
|
| 1429 |
+
"grad_norm": 3.706386089324951,
|
| 1430 |
+
"learning_rate": 0.00019999634421619673,
|
| 1431 |
+
"loss": 5.2003,
|
| 1432 |
+
"step": 201
|
| 1433 |
+
},
|
| 1434 |
+
{
|
| 1435 |
+
"epoch": 0.009640739282433093,
|
| 1436 |
+
"grad_norm": 4.2931294441223145,
|
| 1437 |
+
"learning_rate": 0.00019999630133373325,
|
| 1438 |
+
"loss": 5.9312,
|
| 1439 |
+
"step": 202
|
| 1440 |
+
},
|
| 1441 |
+
{
|
| 1442 |
+
"epoch": 0.009688465714524346,
|
| 1443 |
+
"grad_norm": 4.851586818695068,
|
| 1444 |
+
"learning_rate": 0.00019999625820123245,
|
| 1445 |
+
"loss": 6.7144,
|
| 1446 |
+
"step": 203
|
| 1447 |
+
},
|
| 1448 |
+
{
|
| 1449 |
+
"epoch": 0.0097361921466156,
|
| 1450 |
+
"grad_norm": 5.349869728088379,
|
| 1451 |
+
"learning_rate": 0.0001999962148186944,
|
| 1452 |
+
"loss": 6.6918,
|
| 1453 |
+
"step": 204
|
| 1454 |
+
},
|
| 1455 |
+
{
|
| 1456 |
+
"epoch": 0.009783918578706852,
|
| 1457 |
+
"grad_norm": 5.367208003997803,
|
| 1458 |
+
"learning_rate": 0.00019999617118611924,
|
| 1459 |
+
"loss": 7.5381,
|
| 1460 |
+
"step": 205
|
| 1461 |
+
},
|
| 1462 |
+
{
|
| 1463 |
+
"epoch": 0.009831645010798105,
|
| 1464 |
+
"grad_norm": 4.719504356384277,
|
| 1465 |
+
"learning_rate": 0.00019999612730350707,
|
| 1466 |
+
"loss": 6.669,
|
| 1467 |
+
"step": 206
|
| 1468 |
+
},
|
| 1469 |
+
{
|
| 1470 |
+
"epoch": 0.009879371442889359,
|
| 1471 |
+
"grad_norm": 5.995404243469238,
|
| 1472 |
+
"learning_rate": 0.000199996083170858,
|
| 1473 |
+
"loss": 7.8411,
|
| 1474 |
+
"step": 207
|
| 1475 |
+
},
|
| 1476 |
+
{
|
| 1477 |
+
"epoch": 0.009927097874980612,
|
| 1478 |
+
"grad_norm": 5.326706409454346,
|
| 1479 |
+
"learning_rate": 0.0001999960387881721,
|
| 1480 |
+
"loss": 6.2895,
|
| 1481 |
+
"step": 208
|
| 1482 |
+
},
|
| 1483 |
+
{
|
| 1484 |
+
"epoch": 0.009974824307071865,
|
| 1485 |
+
"grad_norm": 4.975574016571045,
|
| 1486 |
+
"learning_rate": 0.00019999599415544957,
|
| 1487 |
+
"loss": 5.5179,
|
| 1488 |
+
"step": 209
|
| 1489 |
+
},
|
| 1490 |
+
{
|
| 1491 |
+
"epoch": 0.010022550739163118,
|
| 1492 |
+
"grad_norm": 5.673884391784668,
|
| 1493 |
+
"learning_rate": 0.00019999594927269047,
|
| 1494 |
+
"loss": 6.7112,
|
| 1495 |
+
"step": 210
|
| 1496 |
+
},
|
| 1497 |
+
{
|
| 1498 |
+
"epoch": 0.01007027717125437,
|
| 1499 |
+
"grad_norm": 5.526071071624756,
|
| 1500 |
+
"learning_rate": 0.0001999959041398949,
|
| 1501 |
+
"loss": 6.4561,
|
| 1502 |
+
"step": 211
|
| 1503 |
+
},
|
| 1504 |
+
{
|
| 1505 |
+
"epoch": 0.010118003603345622,
|
| 1506 |
+
"grad_norm": 5.906366348266602,
|
| 1507 |
+
"learning_rate": 0.000199995858757063,
|
| 1508 |
+
"loss": 6.8685,
|
| 1509 |
+
"step": 212
|
| 1510 |
+
},
|
| 1511 |
+
{
|
| 1512 |
+
"epoch": 0.010165730035436875,
|
| 1513 |
+
"grad_norm": 5.4884257316589355,
|
| 1514 |
+
"learning_rate": 0.0001999958131241949,
|
| 1515 |
+
"loss": 8.4239,
|
| 1516 |
+
"step": 213
|
| 1517 |
+
},
|
| 1518 |
+
{
|
| 1519 |
+
"epoch": 0.010213456467528128,
|
| 1520 |
+
"grad_norm": 3.8919754028320312,
|
| 1521 |
+
"learning_rate": 0.0001999957672412906,
|
| 1522 |
+
"loss": 6.4004,
|
| 1523 |
+
"step": 214
|
| 1524 |
+
},
|
| 1525 |
+
{
|
| 1526 |
+
"epoch": 0.010261182899619381,
|
| 1527 |
+
"grad_norm": 5.335618495941162,
|
| 1528 |
+
"learning_rate": 0.0001999957211083504,
|
| 1529 |
+
"loss": 5.9267,
|
| 1530 |
+
"step": 215
|
| 1531 |
+
},
|
| 1532 |
+
{
|
| 1533 |
+
"epoch": 0.010308909331710634,
|
| 1534 |
+
"grad_norm": 5.428693771362305,
|
| 1535 |
+
"learning_rate": 0.00019999567472537424,
|
| 1536 |
+
"loss": 7.1497,
|
| 1537 |
+
"step": 216
|
| 1538 |
+
},
|
| 1539 |
+
{
|
| 1540 |
+
"epoch": 0.010356635763801888,
|
| 1541 |
+
"grad_norm": 3.8885908126831055,
|
| 1542 |
+
"learning_rate": 0.0001999956280923623,
|
| 1543 |
+
"loss": 5.886,
|
| 1544 |
+
"step": 217
|
| 1545 |
+
},
|
| 1546 |
+
{
|
| 1547 |
+
"epoch": 0.01040436219589314,
|
| 1548 |
+
"grad_norm": 4.225867748260498,
|
| 1549 |
+
"learning_rate": 0.0001999955812093148,
|
| 1550 |
+
"loss": 7.0125,
|
| 1551 |
+
"step": 218
|
| 1552 |
+
},
|
| 1553 |
+
{
|
| 1554 |
+
"epoch": 0.010452088627984394,
|
| 1555 |
+
"grad_norm": 4.775414943695068,
|
| 1556 |
+
"learning_rate": 0.0001999955340762317,
|
| 1557 |
+
"loss": 6.0621,
|
| 1558 |
+
"step": 219
|
| 1559 |
+
},
|
| 1560 |
+
{
|
| 1561 |
+
"epoch": 0.010499815060075647,
|
| 1562 |
+
"grad_norm": 4.572904586791992,
|
| 1563 |
+
"learning_rate": 0.00019999548669311318,
|
| 1564 |
+
"loss": 5.5764,
|
| 1565 |
+
"step": 220
|
| 1566 |
+
},
|
| 1567 |
+
{
|
| 1568 |
+
"epoch": 0.0105475414921669,
|
| 1569 |
+
"grad_norm": 5.60729455947876,
|
| 1570 |
+
"learning_rate": 0.00019999543905995938,
|
| 1571 |
+
"loss": 8.5128,
|
| 1572 |
+
"step": 221
|
| 1573 |
+
},
|
| 1574 |
+
{
|
| 1575 |
+
"epoch": 0.010595267924258153,
|
| 1576 |
+
"grad_norm": 5.341472148895264,
|
| 1577 |
+
"learning_rate": 0.0001999953911767704,
|
| 1578 |
+
"loss": 7.4026,
|
| 1579 |
+
"step": 222
|
| 1580 |
+
},
|
| 1581 |
+
{
|
| 1582 |
+
"epoch": 0.010642994356349406,
|
| 1583 |
+
"grad_norm": 3.933389663696289,
|
| 1584 |
+
"learning_rate": 0.00019999534304354635,
|
| 1585 |
+
"loss": 4.8309,
|
| 1586 |
+
"step": 223
|
| 1587 |
+
},
|
| 1588 |
+
{
|
| 1589 |
+
"epoch": 0.010690720788440657,
|
| 1590 |
+
"grad_norm": 5.747697830200195,
|
| 1591 |
+
"learning_rate": 0.00019999529466028737,
|
| 1592 |
+
"loss": 7.3435,
|
| 1593 |
+
"step": 224
|
| 1594 |
+
},
|
| 1595 |
+
{
|
| 1596 |
+
"epoch": 0.01073844722053191,
|
| 1597 |
+
"grad_norm": 4.927333831787109,
|
| 1598 |
+
"learning_rate": 0.00019999524602699358,
|
| 1599 |
+
"loss": 6.4677,
|
| 1600 |
+
"step": 225
|
| 1601 |
+
},
|
| 1602 |
+
{
|
| 1603 |
+
"epoch": 0.010786173652623163,
|
| 1604 |
+
"grad_norm": 3.5867760181427,
|
| 1605 |
+
"learning_rate": 0.0001999951971436651,
|
| 1606 |
+
"loss": 5.4795,
|
| 1607 |
+
"step": 226
|
| 1608 |
+
},
|
| 1609 |
+
{
|
| 1610 |
+
"epoch": 0.010833900084714417,
|
| 1611 |
+
"grad_norm": 4.812954425811768,
|
| 1612 |
+
"learning_rate": 0.000199995148010302,
|
| 1613 |
+
"loss": 6.1549,
|
| 1614 |
+
"step": 227
|
| 1615 |
+
},
|
| 1616 |
+
{
|
| 1617 |
+
"epoch": 0.01088162651680567,
|
| 1618 |
+
"grad_norm": 6.369991302490234,
|
| 1619 |
+
"learning_rate": 0.00019999509862690448,
|
| 1620 |
+
"loss": 7.4918,
|
| 1621 |
+
"step": 228
|
| 1622 |
+
},
|
| 1623 |
+
{
|
| 1624 |
+
"epoch": 0.010929352948896923,
|
| 1625 |
+
"grad_norm": 5.212950229644775,
|
| 1626 |
+
"learning_rate": 0.00019999504899347262,
|
| 1627 |
+
"loss": 5.5875,
|
| 1628 |
+
"step": 229
|
| 1629 |
+
},
|
| 1630 |
+
{
|
| 1631 |
+
"epoch": 0.010977079380988176,
|
| 1632 |
+
"grad_norm": 4.195418357849121,
|
| 1633 |
+
"learning_rate": 0.00019999499911000656,
|
| 1634 |
+
"loss": 5.2198,
|
| 1635 |
+
"step": 230
|
| 1636 |
+
},
|
| 1637 |
+
{
|
| 1638 |
+
"epoch": 0.011024805813079429,
|
| 1639 |
+
"grad_norm": 4.676207065582275,
|
| 1640 |
+
"learning_rate": 0.00019999494897650645,
|
| 1641 |
+
"loss": 6.0417,
|
| 1642 |
+
"step": 231
|
| 1643 |
+
},
|
| 1644 |
+
{
|
| 1645 |
+
"epoch": 0.011072532245170682,
|
| 1646 |
+
"grad_norm": 4.369086265563965,
|
| 1647 |
+
"learning_rate": 0.00019999489859297235,
|
| 1648 |
+
"loss": 4.7328,
|
| 1649 |
+
"step": 232
|
| 1650 |
+
},
|
| 1651 |
+
{
|
| 1652 |
+
"epoch": 0.011120258677261935,
|
| 1653 |
+
"grad_norm": 5.925018310546875,
|
| 1654 |
+
"learning_rate": 0.00019999484795940443,
|
| 1655 |
+
"loss": 6.0987,
|
| 1656 |
+
"step": 233
|
| 1657 |
+
},
|
| 1658 |
+
{
|
| 1659 |
+
"epoch": 0.011167985109353188,
|
| 1660 |
+
"grad_norm": 4.76014518737793,
|
| 1661 |
+
"learning_rate": 0.00019999479707580282,
|
| 1662 |
+
"loss": 6.571,
|
| 1663 |
+
"step": 234
|
| 1664 |
+
},
|
| 1665 |
+
{
|
| 1666 |
+
"epoch": 0.011215711541444441,
|
| 1667 |
+
"grad_norm": 5.120473861694336,
|
| 1668 |
+
"learning_rate": 0.00019999474594216762,
|
| 1669 |
+
"loss": 7.5342,
|
| 1670 |
+
"step": 235
|
| 1671 |
+
},
|
| 1672 |
+
{
|
| 1673 |
+
"epoch": 0.011263437973535694,
|
| 1674 |
+
"grad_norm": 5.6909027099609375,
|
| 1675 |
+
"learning_rate": 0.000199994694558499,
|
| 1676 |
+
"loss": 8.6062,
|
| 1677 |
+
"step": 236
|
| 1678 |
+
},
|
| 1679 |
+
{
|
| 1680 |
+
"epoch": 0.011311164405626946,
|
| 1681 |
+
"grad_norm": 8.290360450744629,
|
| 1682 |
+
"learning_rate": 0.00019999464292479703,
|
| 1683 |
+
"loss": 7.7553,
|
| 1684 |
+
"step": 237
|
| 1685 |
+
},
|
| 1686 |
+
{
|
| 1687 |
+
"epoch": 0.011358890837718199,
|
| 1688 |
+
"grad_norm": 5.976494312286377,
|
| 1689 |
+
"learning_rate": 0.00019999459104106187,
|
| 1690 |
+
"loss": 8.6744,
|
| 1691 |
+
"step": 238
|
| 1692 |
+
},
|
| 1693 |
+
{
|
| 1694 |
+
"epoch": 0.011406617269809452,
|
| 1695 |
+
"grad_norm": 10.440987586975098,
|
| 1696 |
+
"learning_rate": 0.0001999945389072937,
|
| 1697 |
+
"loss": 8.9354,
|
| 1698 |
+
"step": 239
|
| 1699 |
+
},
|
| 1700 |
+
{
|
| 1701 |
+
"epoch": 0.011454343701900705,
|
| 1702 |
+
"grad_norm": 4.592039585113525,
|
| 1703 |
+
"learning_rate": 0.00019999448652349258,
|
| 1704 |
+
"loss": 7.377,
|
| 1705 |
+
"step": 240
|
| 1706 |
+
},
|
| 1707 |
+
{
|
| 1708 |
+
"epoch": 0.011502070133991958,
|
| 1709 |
+
"grad_norm": 4.670874118804932,
|
| 1710 |
+
"learning_rate": 0.00019999443388965863,
|
| 1711 |
+
"loss": 7.5659,
|
| 1712 |
+
"step": 241
|
| 1713 |
+
},
|
| 1714 |
+
{
|
| 1715 |
+
"epoch": 0.011549796566083211,
|
| 1716 |
+
"grad_norm": 4.679845809936523,
|
| 1717 |
+
"learning_rate": 0.00019999438100579204,
|
| 1718 |
+
"loss": 6.6256,
|
| 1719 |
+
"step": 242
|
| 1720 |
+
},
|
| 1721 |
+
{
|
| 1722 |
+
"epoch": 0.011597522998174464,
|
| 1723 |
+
"grad_norm": 4.715898036956787,
|
| 1724 |
+
"learning_rate": 0.0001999943278718929,
|
| 1725 |
+
"loss": 6.702,
|
| 1726 |
+
"step": 243
|
| 1727 |
+
},
|
| 1728 |
+
{
|
| 1729 |
+
"epoch": 0.011645249430265717,
|
| 1730 |
+
"grad_norm": 7.395649433135986,
|
| 1731 |
+
"learning_rate": 0.0001999942744879614,
|
| 1732 |
+
"loss": 7.1643,
|
| 1733 |
+
"step": 244
|
| 1734 |
+
},
|
| 1735 |
+
{
|
| 1736 |
+
"epoch": 0.01169297586235697,
|
| 1737 |
+
"grad_norm": 4.151764392852783,
|
| 1738 |
+
"learning_rate": 0.0001999942208539976,
|
| 1739 |
+
"loss": 5.9845,
|
| 1740 |
+
"step": 245
|
| 1741 |
+
},
|
| 1742 |
+
{
|
| 1743 |
+
"epoch": 0.011740702294448223,
|
| 1744 |
+
"grad_norm": 5.3397345542907715,
|
| 1745 |
+
"learning_rate": 0.00019999416697000165,
|
| 1746 |
+
"loss": 8.978,
|
| 1747 |
+
"step": 246
|
| 1748 |
+
},
|
| 1749 |
+
{
|
| 1750 |
+
"epoch": 0.011788428726539476,
|
| 1751 |
+
"grad_norm": 5.556086540222168,
|
| 1752 |
+
"learning_rate": 0.00019999411283597374,
|
| 1753 |
+
"loss": 7.4888,
|
| 1754 |
+
"step": 247
|
| 1755 |
+
},
|
| 1756 |
+
{
|
| 1757 |
+
"epoch": 0.01183615515863073,
|
| 1758 |
+
"grad_norm": 4.385659217834473,
|
| 1759 |
+
"learning_rate": 0.00019999405845191393,
|
| 1760 |
+
"loss": 6.167,
|
| 1761 |
+
"step": 248
|
| 1762 |
+
},
|
| 1763 |
+
{
|
| 1764 |
+
"epoch": 0.011883881590721982,
|
| 1765 |
+
"grad_norm": 4.701001167297363,
|
| 1766 |
+
"learning_rate": 0.00019999400381782244,
|
| 1767 |
+
"loss": 6.2042,
|
| 1768 |
+
"step": 249
|
| 1769 |
+
},
|
| 1770 |
+
{
|
| 1771 |
+
"epoch": 0.011931608022813234,
|
| 1772 |
+
"grad_norm": 4.775174140930176,
|
| 1773 |
+
"learning_rate": 0.0001999939489336993,
|
| 1774 |
+
"loss": 7.4549,
|
| 1775 |
+
"step": 250
|
| 1776 |
+
},
|
| 1777 |
+
{
|
| 1778 |
+
"epoch": 0.011979334454904487,
|
| 1779 |
+
"grad_norm": 5.019827365875244,
|
| 1780 |
+
"learning_rate": 0.00019999389379954477,
|
| 1781 |
+
"loss": 5.8486,
|
| 1782 |
+
"step": 251
|
| 1783 |
+
},
|
| 1784 |
+
{
|
| 1785 |
+
"epoch": 0.01202706088699574,
|
| 1786 |
+
"grad_norm": 5.453334331512451,
|
| 1787 |
+
"learning_rate": 0.00019999383841535888,
|
| 1788 |
+
"loss": 7.1586,
|
| 1789 |
+
"step": 252
|
| 1790 |
+
},
|
| 1791 |
+
{
|
| 1792 |
+
"epoch": 0.012074787319086993,
|
| 1793 |
+
"grad_norm": 4.455638408660889,
|
| 1794 |
+
"learning_rate": 0.00019999378278114183,
|
| 1795 |
+
"loss": 5.3958,
|
| 1796 |
+
"step": 253
|
| 1797 |
+
},
|
| 1798 |
+
{
|
| 1799 |
+
"epoch": 0.012122513751178246,
|
| 1800 |
+
"grad_norm": 4.937852382659912,
|
| 1801 |
+
"learning_rate": 0.00019999372689689376,
|
| 1802 |
+
"loss": 7.0083,
|
| 1803 |
+
"step": 254
|
| 1804 |
+
},
|
| 1805 |
+
{
|
| 1806 |
+
"epoch": 0.012170240183269499,
|
| 1807 |
+
"grad_norm": 4.422460079193115,
|
| 1808 |
+
"learning_rate": 0.00019999367076261476,
|
| 1809 |
+
"loss": 6.4996,
|
| 1810 |
+
"step": 255
|
| 1811 |
+
},
|
| 1812 |
+
{
|
| 1813 |
+
"epoch": 0.012217966615360752,
|
| 1814 |
+
"grad_norm": 3.780395269393921,
|
| 1815 |
+
"learning_rate": 0.000199993614378305,
|
| 1816 |
+
"loss": 6.0372,
|
| 1817 |
+
"step": 256
|
| 1818 |
+
},
|
| 1819 |
+
{
|
| 1820 |
+
"epoch": 0.012265693047452005,
|
| 1821 |
+
"grad_norm": 5.636499404907227,
|
| 1822 |
+
"learning_rate": 0.00019999355774396465,
|
| 1823 |
+
"loss": 6.3515,
|
| 1824 |
+
"step": 257
|
| 1825 |
+
},
|
| 1826 |
+
{
|
| 1827 |
+
"epoch": 0.012313419479543258,
|
| 1828 |
+
"grad_norm": 5.77448844909668,
|
| 1829 |
+
"learning_rate": 0.0001999935008595938,
|
| 1830 |
+
"loss": 8.0423,
|
| 1831 |
+
"step": 258
|
| 1832 |
+
},
|
| 1833 |
+
{
|
| 1834 |
+
"epoch": 0.012361145911634511,
|
| 1835 |
+
"grad_norm": 4.834639549255371,
|
| 1836 |
+
"learning_rate": 0.00019999344372519264,
|
| 1837 |
+
"loss": 5.8949,
|
| 1838 |
+
"step": 259
|
| 1839 |
+
},
|
| 1840 |
+
{
|
| 1841 |
+
"epoch": 0.012408872343725764,
|
| 1842 |
+
"grad_norm": 6.200581073760986,
|
| 1843 |
+
"learning_rate": 0.00019999338634076126,
|
| 1844 |
+
"loss": 8.6784,
|
| 1845 |
+
"step": 260
|
| 1846 |
+
},
|
| 1847 |
+
{
|
| 1848 |
+
"epoch": 0.012456598775817018,
|
| 1849 |
+
"grad_norm": 5.017780303955078,
|
| 1850 |
+
"learning_rate": 0.00019999332870629987,
|
| 1851 |
+
"loss": 5.4993,
|
| 1852 |
+
"step": 261
|
| 1853 |
+
},
|
| 1854 |
+
{
|
| 1855 |
+
"epoch": 0.01250432520790827,
|
| 1856 |
+
"grad_norm": 4.737110137939453,
|
| 1857 |
+
"learning_rate": 0.00019999327082180854,
|
| 1858 |
+
"loss": 6.2233,
|
| 1859 |
+
"step": 262
|
| 1860 |
+
},
|
| 1861 |
+
{
|
| 1862 |
+
"epoch": 0.012552051639999522,
|
| 1863 |
+
"grad_norm": 4.9615278244018555,
|
| 1864 |
+
"learning_rate": 0.00019999321268728747,
|
| 1865 |
+
"loss": 6.5653,
|
| 1866 |
+
"step": 263
|
| 1867 |
+
},
|
| 1868 |
+
{
|
| 1869 |
+
"epoch": 0.012599778072090775,
|
| 1870 |
+
"grad_norm": 4.462545871734619,
|
| 1871 |
+
"learning_rate": 0.00019999315430273683,
|
| 1872 |
+
"loss": 6.815,
|
| 1873 |
+
"step": 264
|
| 1874 |
+
},
|
| 1875 |
+
{
|
| 1876 |
+
"epoch": 0.012647504504182028,
|
| 1877 |
+
"grad_norm": 5.788268566131592,
|
| 1878 |
+
"learning_rate": 0.00019999309566815665,
|
| 1879 |
+
"loss": 5.2953,
|
| 1880 |
+
"step": 265
|
| 1881 |
+
},
|
| 1882 |
+
{
|
| 1883 |
+
"epoch": 0.012695230936273281,
|
| 1884 |
+
"grad_norm": 5.634494304656982,
|
| 1885 |
+
"learning_rate": 0.0001999930367835472,
|
| 1886 |
+
"loss": 6.1009,
|
| 1887 |
+
"step": 266
|
| 1888 |
+
},
|
| 1889 |
+
{
|
| 1890 |
+
"epoch": 0.012742957368364534,
|
| 1891 |
+
"grad_norm": 6.640275955200195,
|
| 1892 |
+
"learning_rate": 0.00019999297764890854,
|
| 1893 |
+
"loss": 6.454,
|
| 1894 |
+
"step": 267
|
| 1895 |
+
},
|
| 1896 |
+
{
|
| 1897 |
+
"epoch": 0.012790683800455787,
|
| 1898 |
+
"grad_norm": 5.759767055511475,
|
| 1899 |
+
"learning_rate": 0.0001999929182642409,
|
| 1900 |
+
"loss": 8.0766,
|
| 1901 |
+
"step": 268
|
| 1902 |
+
},
|
| 1903 |
+
{
|
| 1904 |
+
"epoch": 0.01283841023254704,
|
| 1905 |
+
"grad_norm": 4.624085426330566,
|
| 1906 |
+
"learning_rate": 0.00019999285862954436,
|
| 1907 |
+
"loss": 6.848,
|
| 1908 |
+
"step": 269
|
| 1909 |
+
},
|
| 1910 |
+
{
|
| 1911 |
+
"epoch": 0.012886136664638293,
|
| 1912 |
+
"grad_norm": 3.952364206314087,
|
| 1913 |
+
"learning_rate": 0.00019999279874481908,
|
| 1914 |
+
"loss": 5.7864,
|
| 1915 |
+
"step": 270
|
| 1916 |
+
},
|
| 1917 |
+
{
|
| 1918 |
+
"epoch": 0.012933863096729547,
|
| 1919 |
+
"grad_norm": 5.424389839172363,
|
| 1920 |
+
"learning_rate": 0.00019999273861006525,
|
| 1921 |
+
"loss": 6.1687,
|
| 1922 |
+
"step": 271
|
| 1923 |
+
},
|
| 1924 |
+
{
|
| 1925 |
+
"epoch": 0.0129815895288208,
|
| 1926 |
+
"grad_norm": 4.6196088790893555,
|
| 1927 |
+
"learning_rate": 0.00019999267822528297,
|
| 1928 |
+
"loss": 6.6814,
|
| 1929 |
+
"step": 272
|
| 1930 |
+
},
|
| 1931 |
+
{
|
| 1932 |
+
"epoch": 0.013029315960912053,
|
| 1933 |
+
"grad_norm": 4.935333251953125,
|
| 1934 |
+
"learning_rate": 0.00019999261759047243,
|
| 1935 |
+
"loss": 5.5486,
|
| 1936 |
+
"step": 273
|
| 1937 |
+
},
|
| 1938 |
+
{
|
| 1939 |
+
"epoch": 0.013077042393003306,
|
| 1940 |
+
"grad_norm": 5.395534038543701,
|
| 1941 |
+
"learning_rate": 0.00019999255670563376,
|
| 1942 |
+
"loss": 6.8173,
|
| 1943 |
+
"step": 274
|
| 1944 |
+
},
|
| 1945 |
+
{
|
| 1946 |
+
"epoch": 0.013124768825094559,
|
| 1947 |
+
"grad_norm": 6.942729949951172,
|
| 1948 |
+
"learning_rate": 0.0001999924955707671,
|
| 1949 |
+
"loss": 8.094,
|
| 1950 |
+
"step": 275
|
| 1951 |
+
},
|
| 1952 |
+
{
|
| 1953 |
+
"epoch": 0.01317249525718581,
|
| 1954 |
+
"grad_norm": 5.219112396240234,
|
| 1955 |
+
"learning_rate": 0.00019999243418587266,
|
| 1956 |
+
"loss": 6.5134,
|
| 1957 |
+
"step": 276
|
| 1958 |
+
},
|
| 1959 |
+
{
|
| 1960 |
+
"epoch": 0.013220221689277063,
|
| 1961 |
+
"grad_norm": 4.332059383392334,
|
| 1962 |
+
"learning_rate": 0.00019999237255095053,
|
| 1963 |
+
"loss": 6.2888,
|
| 1964 |
+
"step": 277
|
| 1965 |
+
},
|
| 1966 |
+
{
|
| 1967 |
+
"epoch": 0.013267948121368316,
|
| 1968 |
+
"grad_norm": 5.0153679847717285,
|
| 1969 |
+
"learning_rate": 0.0001999923106660009,
|
| 1970 |
+
"loss": 6.3642,
|
| 1971 |
+
"step": 278
|
| 1972 |
+
},
|
| 1973 |
+
{
|
| 1974 |
+
"epoch": 0.01331567455345957,
|
| 1975 |
+
"grad_norm": 3.461423873901367,
|
| 1976 |
+
"learning_rate": 0.0001999922485310239,
|
| 1977 |
+
"loss": 5.3825,
|
| 1978 |
+
"step": 279
|
| 1979 |
+
},
|
| 1980 |
+
{
|
| 1981 |
+
"epoch": 0.013363400985550822,
|
| 1982 |
+
"grad_norm": 4.702268600463867,
|
| 1983 |
+
"learning_rate": 0.0001999921861460197,
|
| 1984 |
+
"loss": 5.971,
|
| 1985 |
+
"step": 280
|
| 1986 |
+
},
|
| 1987 |
+
{
|
| 1988 |
+
"epoch": 0.013411127417642076,
|
| 1989 |
+
"grad_norm": 4.084397792816162,
|
| 1990 |
+
"learning_rate": 0.00019999212351098846,
|
| 1991 |
+
"loss": 6.3295,
|
| 1992 |
+
"step": 281
|
| 1993 |
+
},
|
| 1994 |
+
{
|
| 1995 |
+
"epoch": 0.013458853849733329,
|
| 1996 |
+
"grad_norm": 6.114694118499756,
|
| 1997 |
+
"learning_rate": 0.00019999206062593032,
|
| 1998 |
+
"loss": 7.4146,
|
| 1999 |
+
"step": 282
|
| 2000 |
+
},
|
| 2001 |
+
{
|
| 2002 |
+
"epoch": 0.013506580281824582,
|
| 2003 |
+
"grad_norm": 6.846054553985596,
|
| 2004 |
+
"learning_rate": 0.00019999199749084546,
|
| 2005 |
+
"loss": 6.3659,
|
| 2006 |
+
"step": 283
|
| 2007 |
+
},
|
| 2008 |
+
{
|
| 2009 |
+
"epoch": 0.013554306713915835,
|
| 2010 |
+
"grad_norm": 4.534778594970703,
|
| 2011 |
+
"learning_rate": 0.00019999193410573404,
|
| 2012 |
+
"loss": 6.4923,
|
| 2013 |
+
"step": 284
|
| 2014 |
+
},
|
| 2015 |
+
{
|
| 2016 |
+
"epoch": 0.013602033146007088,
|
| 2017 |
+
"grad_norm": 8.201200485229492,
|
| 2018 |
+
"learning_rate": 0.0001999918704705962,
|
| 2019 |
+
"loss": 8.3697,
|
| 2020 |
+
"step": 285
|
| 2021 |
+
},
|
| 2022 |
+
{
|
| 2023 |
+
"epoch": 0.013649759578098341,
|
| 2024 |
+
"grad_norm": 4.86665153503418,
|
| 2025 |
+
"learning_rate": 0.00019999180658543207,
|
| 2026 |
+
"loss": 6.4496,
|
| 2027 |
+
"step": 286
|
| 2028 |
+
},
|
| 2029 |
+
{
|
| 2030 |
+
"epoch": 0.013697486010189594,
|
| 2031 |
+
"grad_norm": 5.482726573944092,
|
| 2032 |
+
"learning_rate": 0.00019999174245024186,
|
| 2033 |
+
"loss": 6.7978,
|
| 2034 |
+
"step": 287
|
| 2035 |
+
},
|
| 2036 |
+
{
|
| 2037 |
+
"epoch": 0.013745212442280845,
|
| 2038 |
+
"grad_norm": 5.493237018585205,
|
| 2039 |
+
"learning_rate": 0.00019999167806502573,
|
| 2040 |
+
"loss": 7.1984,
|
| 2041 |
+
"step": 288
|
| 2042 |
+
},
|
| 2043 |
+
{
|
| 2044 |
+
"epoch": 0.013792938874372098,
|
| 2045 |
+
"grad_norm": 5.953492641448975,
|
| 2046 |
+
"learning_rate": 0.0001999916134297838,
|
| 2047 |
+
"loss": 7.8598,
|
| 2048 |
+
"step": 289
|
| 2049 |
+
},
|
| 2050 |
+
{
|
| 2051 |
+
"epoch": 0.013840665306463351,
|
| 2052 |
+
"grad_norm": 6.032693386077881,
|
| 2053 |
+
"learning_rate": 0.0001999915485445163,
|
| 2054 |
+
"loss": 9.7618,
|
| 2055 |
+
"step": 290
|
| 2056 |
+
},
|
| 2057 |
+
{
|
| 2058 |
+
"epoch": 0.013888391738554605,
|
| 2059 |
+
"grad_norm": 5.2243733406066895,
|
| 2060 |
+
"learning_rate": 0.00019999148340922333,
|
| 2061 |
+
"loss": 8.777,
|
| 2062 |
+
"step": 291
|
| 2063 |
+
},
|
| 2064 |
+
{
|
| 2065 |
+
"epoch": 0.013936118170645858,
|
| 2066 |
+
"grad_norm": 4.763432025909424,
|
| 2067 |
+
"learning_rate": 0.00019999141802390505,
|
| 2068 |
+
"loss": 8.0905,
|
| 2069 |
+
"step": 292
|
| 2070 |
+
},
|
| 2071 |
+
{
|
| 2072 |
+
"epoch": 0.01398384460273711,
|
| 2073 |
+
"grad_norm": 4.205863952636719,
|
| 2074 |
+
"learning_rate": 0.00019999135238856164,
|
| 2075 |
+
"loss": 4.598,
|
| 2076 |
+
"step": 293
|
| 2077 |
+
},
|
| 2078 |
+
{
|
| 2079 |
+
"epoch": 0.014031571034828364,
|
| 2080 |
+
"grad_norm": 4.153672218322754,
|
| 2081 |
+
"learning_rate": 0.0001999912865031933,
|
| 2082 |
+
"loss": 7.0555,
|
| 2083 |
+
"step": 294
|
| 2084 |
+
},
|
| 2085 |
+
{
|
| 2086 |
+
"epoch": 0.014079297466919617,
|
| 2087 |
+
"grad_norm": 4.621336460113525,
|
| 2088 |
+
"learning_rate": 0.00019999122036780013,
|
| 2089 |
+
"loss": 5.8933,
|
| 2090 |
+
"step": 295
|
| 2091 |
+
},
|
| 2092 |
+
{
|
| 2093 |
+
"epoch": 0.01412702389901087,
|
| 2094 |
+
"grad_norm": 4.004661560058594,
|
| 2095 |
+
"learning_rate": 0.00019999115398238235,
|
| 2096 |
+
"loss": 7.7509,
|
| 2097 |
+
"step": 296
|
| 2098 |
+
},
|
| 2099 |
+
{
|
| 2100 |
+
"epoch": 0.014174750331102123,
|
| 2101 |
+
"grad_norm": 4.914793014526367,
|
| 2102 |
+
"learning_rate": 0.00019999108734694012,
|
| 2103 |
+
"loss": 7.7317,
|
| 2104 |
+
"step": 297
|
| 2105 |
+
},
|
| 2106 |
+
{
|
| 2107 |
+
"epoch": 0.014222476763193376,
|
| 2108 |
+
"grad_norm": 5.124854564666748,
|
| 2109 |
+
"learning_rate": 0.00019999102046147358,
|
| 2110 |
+
"loss": 7.1103,
|
| 2111 |
+
"step": 298
|
| 2112 |
+
},
|
| 2113 |
+
{
|
| 2114 |
+
"epoch": 0.01427020319528463,
|
| 2115 |
+
"grad_norm": 4.222902297973633,
|
| 2116 |
+
"learning_rate": 0.0001999909533259829,
|
| 2117 |
+
"loss": 6.7856,
|
| 2118 |
+
"step": 299
|
| 2119 |
+
},
|
| 2120 |
+
{
|
| 2121 |
+
"epoch": 0.014317929627375882,
|
| 2122 |
+
"grad_norm": 3.855794668197632,
|
| 2123 |
+
"learning_rate": 0.00019999088594046827,
|
| 2124 |
+
"loss": 7.6669,
|
| 2125 |
+
"step": 300
|
| 2126 |
+
},
|
| 2127 |
+
{
|
| 2128 |
+
"epoch": 0.014365656059467134,
|
| 2129 |
+
"grad_norm": 4.996542453765869,
|
| 2130 |
+
"learning_rate": 0.00019999081830492983,
|
| 2131 |
+
"loss": 7.5863,
|
| 2132 |
+
"step": 301
|
| 2133 |
+
},
|
| 2134 |
+
{
|
| 2135 |
+
"epoch": 0.014413382491558387,
|
| 2136 |
+
"grad_norm": 3.9295811653137207,
|
| 2137 |
+
"learning_rate": 0.0001999907504193678,
|
| 2138 |
+
"loss": 6.4964,
|
| 2139 |
+
"step": 302
|
| 2140 |
+
},
|
| 2141 |
+
{
|
| 2142 |
+
"epoch": 0.01446110892364964,
|
| 2143 |
+
"grad_norm": 5.005653381347656,
|
| 2144 |
+
"learning_rate": 0.00019999068228378225,
|
| 2145 |
+
"loss": 7.0248,
|
| 2146 |
+
"step": 303
|
| 2147 |
+
},
|
| 2148 |
+
{
|
| 2149 |
+
"epoch": 0.014508835355740893,
|
| 2150 |
+
"grad_norm": 5.204959392547607,
|
| 2151 |
+
"learning_rate": 0.00019999061389817347,
|
| 2152 |
+
"loss": 6.8771,
|
| 2153 |
+
"step": 304
|
| 2154 |
+
},
|
| 2155 |
+
{
|
| 2156 |
+
"epoch": 0.014556561787832146,
|
| 2157 |
+
"grad_norm": 5.9217681884765625,
|
| 2158 |
+
"learning_rate": 0.00019999054526254154,
|
| 2159 |
+
"loss": 6.7342,
|
| 2160 |
+
"step": 305
|
| 2161 |
+
},
|
| 2162 |
+
{
|
| 2163 |
+
"epoch": 0.014604288219923399,
|
| 2164 |
+
"grad_norm": 6.10482931137085,
|
| 2165 |
+
"learning_rate": 0.0001999904763768867,
|
| 2166 |
+
"loss": 7.8614,
|
| 2167 |
+
"step": 306
|
| 2168 |
+
},
|
| 2169 |
+
{
|
| 2170 |
+
"epoch": 0.014652014652014652,
|
| 2171 |
+
"grad_norm": 4.919510841369629,
|
| 2172 |
+
"learning_rate": 0.00019999040724120909,
|
| 2173 |
+
"loss": 6.5849,
|
| 2174 |
+
"step": 307
|
| 2175 |
+
},
|
| 2176 |
+
{
|
| 2177 |
+
"epoch": 0.014699741084105905,
|
| 2178 |
+
"grad_norm": 3.9289333820343018,
|
| 2179 |
+
"learning_rate": 0.00019999033785550886,
|
| 2180 |
+
"loss": 4.5725,
|
| 2181 |
+
"step": 308
|
| 2182 |
+
},
|
| 2183 |
+
{
|
| 2184 |
+
"epoch": 0.014747467516197158,
|
| 2185 |
+
"grad_norm": 5.739213943481445,
|
| 2186 |
+
"learning_rate": 0.0001999902682197862,
|
| 2187 |
+
"loss": 7.8427,
|
| 2188 |
+
"step": 309
|
| 2189 |
+
},
|
| 2190 |
+
{
|
| 2191 |
+
"epoch": 0.014795193948288411,
|
| 2192 |
+
"grad_norm": 3.7051210403442383,
|
| 2193 |
+
"learning_rate": 0.00019999019833404132,
|
| 2194 |
+
"loss": 6.2502,
|
| 2195 |
+
"step": 310
|
| 2196 |
+
},
|
| 2197 |
+
{
|
| 2198 |
+
"epoch": 0.014842920380379664,
|
| 2199 |
+
"grad_norm": 4.647250175476074,
|
| 2200 |
+
"learning_rate": 0.00019999012819827433,
|
| 2201 |
+
"loss": 6.5101,
|
| 2202 |
+
"step": 311
|
| 2203 |
+
},
|
| 2204 |
+
{
|
| 2205 |
+
"epoch": 0.014890646812470917,
|
| 2206 |
+
"grad_norm": 3.8817758560180664,
|
| 2207 |
+
"learning_rate": 0.0001999900578124855,
|
| 2208 |
+
"loss": 4.8714,
|
| 2209 |
+
"step": 312
|
| 2210 |
+
},
|
| 2211 |
+
{
|
| 2212 |
+
"epoch": 0.01493837324456217,
|
| 2213 |
+
"grad_norm": 4.8853607177734375,
|
| 2214 |
+
"learning_rate": 0.0001999899871766749,
|
| 2215 |
+
"loss": 7.3876,
|
| 2216 |
+
"step": 313
|
| 2217 |
+
},
|
| 2218 |
+
{
|
| 2219 |
+
"epoch": 0.014986099676653422,
|
| 2220 |
+
"grad_norm": 5.011137962341309,
|
| 2221 |
+
"learning_rate": 0.00019998991629084276,
|
| 2222 |
+
"loss": 5.4221,
|
| 2223 |
+
"step": 314
|
| 2224 |
+
},
|
| 2225 |
+
{
|
| 2226 |
+
"epoch": 0.015033826108744675,
|
| 2227 |
+
"grad_norm": 4.79543399810791,
|
| 2228 |
+
"learning_rate": 0.00019998984515498926,
|
| 2229 |
+
"loss": 5.3682,
|
| 2230 |
+
"step": 315
|
| 2231 |
+
},
|
| 2232 |
+
{
|
| 2233 |
+
"epoch": 0.015081552540835928,
|
| 2234 |
+
"grad_norm": 4.967861652374268,
|
| 2235 |
+
"learning_rate": 0.00019998977376911454,
|
| 2236 |
+
"loss": 6.0652,
|
| 2237 |
+
"step": 316
|
| 2238 |
+
},
|
| 2239 |
+
{
|
| 2240 |
+
"epoch": 0.015129278972927181,
|
| 2241 |
+
"grad_norm": 5.12961483001709,
|
| 2242 |
+
"learning_rate": 0.00019998970213321883,
|
| 2243 |
+
"loss": 6.722,
|
| 2244 |
+
"step": 317
|
| 2245 |
+
},
|
| 2246 |
+
{
|
| 2247 |
+
"epoch": 0.015177005405018434,
|
| 2248 |
+
"grad_norm": 5.37574577331543,
|
| 2249 |
+
"learning_rate": 0.0001999896302473023,
|
| 2250 |
+
"loss": 6.4294,
|
| 2251 |
+
"step": 318
|
| 2252 |
+
},
|
| 2253 |
+
{
|
| 2254 |
+
"epoch": 0.015224731837109687,
|
| 2255 |
+
"grad_norm": 4.319894790649414,
|
| 2256 |
+
"learning_rate": 0.0001999895581113651,
|
| 2257 |
+
"loss": 6.6173,
|
| 2258 |
+
"step": 319
|
| 2259 |
+
},
|
| 2260 |
+
{
|
| 2261 |
+
"epoch": 0.01527245826920094,
|
| 2262 |
+
"grad_norm": 4.902976989746094,
|
| 2263 |
+
"learning_rate": 0.0001999894857254074,
|
| 2264 |
+
"loss": 6.5298,
|
| 2265 |
+
"step": 320
|
| 2266 |
+
},
|
| 2267 |
+
{
|
| 2268 |
+
"epoch": 0.015320184701292193,
|
| 2269 |
+
"grad_norm": 5.665980815887451,
|
| 2270 |
+
"learning_rate": 0.00019998941308942944,
|
| 2271 |
+
"loss": 6.9341,
|
| 2272 |
+
"step": 321
|
| 2273 |
+
},
|
| 2274 |
+
{
|
| 2275 |
+
"epoch": 0.015367911133383446,
|
| 2276 |
+
"grad_norm": 4.651304721832275,
|
| 2277 |
+
"learning_rate": 0.00019998934020343137,
|
| 2278 |
+
"loss": 5.3411,
|
| 2279 |
+
"step": 322
|
| 2280 |
+
},
|
| 2281 |
+
{
|
| 2282 |
+
"epoch": 0.0154156375654747,
|
| 2283 |
+
"grad_norm": 5.079586982727051,
|
| 2284 |
+
"learning_rate": 0.00019998926706741335,
|
| 2285 |
+
"loss": 5.7383,
|
| 2286 |
+
"step": 323
|
| 2287 |
+
},
|
| 2288 |
+
{
|
| 2289 |
+
"epoch": 0.015463363997565953,
|
| 2290 |
+
"grad_norm": 5.262585163116455,
|
| 2291 |
+
"learning_rate": 0.0001999891936813756,
|
| 2292 |
+
"loss": 7.7724,
|
| 2293 |
+
"step": 324
|
| 2294 |
+
},
|
| 2295 |
+
{
|
| 2296 |
+
"epoch": 0.015511090429657206,
|
| 2297 |
+
"grad_norm": 4.025634288787842,
|
| 2298 |
+
"learning_rate": 0.00019998912004531828,
|
| 2299 |
+
"loss": 4.6858,
|
| 2300 |
+
"step": 325
|
| 2301 |
+
},
|
| 2302 |
+
{
|
| 2303 |
+
"epoch": 0.015558816861748459,
|
| 2304 |
+
"grad_norm": 4.3046393394470215,
|
| 2305 |
+
"learning_rate": 0.00019998904615924158,
|
| 2306 |
+
"loss": 5.6154,
|
| 2307 |
+
"step": 326
|
| 2308 |
+
},
|
| 2309 |
+
{
|
| 2310 |
+
"epoch": 0.01560654329383971,
|
| 2311 |
+
"grad_norm": 5.861588478088379,
|
| 2312 |
+
"learning_rate": 0.00019998897202314569,
|
| 2313 |
+
"loss": 7.1517,
|
| 2314 |
+
"step": 327
|
| 2315 |
+
},
|
| 2316 |
+
{
|
| 2317 |
+
"epoch": 0.015654269725930963,
|
| 2318 |
+
"grad_norm": 4.287512302398682,
|
| 2319 |
+
"learning_rate": 0.00019998889763703077,
|
| 2320 |
+
"loss": 5.763,
|
| 2321 |
+
"step": 328
|
| 2322 |
+
},
|
| 2323 |
+
{
|
| 2324 |
+
"epoch": 0.015701996158022216,
|
| 2325 |
+
"grad_norm": 5.883302211761475,
|
| 2326 |
+
"learning_rate": 0.00019998882300089704,
|
| 2327 |
+
"loss": 7.2864,
|
| 2328 |
+
"step": 329
|
| 2329 |
+
},
|
| 2330 |
+
{
|
| 2331 |
+
"epoch": 0.01574972259011347,
|
| 2332 |
+
"grad_norm": 5.099184989929199,
|
| 2333 |
+
"learning_rate": 0.0001999887481147447,
|
| 2334 |
+
"loss": 7.4756,
|
| 2335 |
+
"step": 330
|
| 2336 |
+
},
|
| 2337 |
+
{
|
| 2338 |
+
"epoch": 0.015797449022204722,
|
| 2339 |
+
"grad_norm": 4.795839309692383,
|
| 2340 |
+
"learning_rate": 0.00019998867297857387,
|
| 2341 |
+
"loss": 7.5285,
|
| 2342 |
+
"step": 331
|
| 2343 |
+
},
|
| 2344 |
+
{
|
| 2345 |
+
"epoch": 0.015845175454295975,
|
| 2346 |
+
"grad_norm": 4.008845329284668,
|
| 2347 |
+
"learning_rate": 0.0001999885975923848,
|
| 2348 |
+
"loss": 5.1818,
|
| 2349 |
+
"step": 332
|
| 2350 |
+
},
|
| 2351 |
+
{
|
| 2352 |
+
"epoch": 0.01589290188638723,
|
| 2353 |
+
"grad_norm": 5.107858180999756,
|
| 2354 |
+
"learning_rate": 0.00019998852195617767,
|
| 2355 |
+
"loss": 7.3373,
|
| 2356 |
+
"step": 333
|
| 2357 |
+
},
|
| 2358 |
+
{
|
| 2359 |
+
"epoch": 0.01594062831847848,
|
| 2360 |
+
"grad_norm": 4.369069576263428,
|
| 2361 |
+
"learning_rate": 0.00019998844606995263,
|
| 2362 |
+
"loss": 6.1056,
|
| 2363 |
+
"step": 334
|
| 2364 |
+
},
|
| 2365 |
+
{
|
| 2366 |
+
"epoch": 0.015988354750569735,
|
| 2367 |
+
"grad_norm": 4.465891361236572,
|
| 2368 |
+
"learning_rate": 0.0001999883699337099,
|
| 2369 |
+
"loss": 6.0833,
|
| 2370 |
+
"step": 335
|
| 2371 |
+
},
|
| 2372 |
+
{
|
| 2373 |
+
"epoch": 0.016036081182660988,
|
| 2374 |
+
"grad_norm": 4.340141773223877,
|
| 2375 |
+
"learning_rate": 0.0001999882935474497,
|
| 2376 |
+
"loss": 7.0717,
|
| 2377 |
+
"step": 336
|
| 2378 |
+
},
|
| 2379 |
+
{
|
| 2380 |
+
"epoch": 0.01608380761475224,
|
| 2381 |
+
"grad_norm": 4.589889049530029,
|
| 2382 |
+
"learning_rate": 0.00019998821691117217,
|
| 2383 |
+
"loss": 5.5123,
|
| 2384 |
+
"step": 337
|
| 2385 |
+
},
|
| 2386 |
+
{
|
| 2387 |
+
"epoch": 0.016131534046843494,
|
| 2388 |
+
"grad_norm": 4.784783840179443,
|
| 2389 |
+
"learning_rate": 0.0001999881400248775,
|
| 2390 |
+
"loss": 7.857,
|
| 2391 |
+
"step": 338
|
| 2392 |
+
},
|
| 2393 |
+
{
|
| 2394 |
+
"epoch": 0.016179260478934747,
|
| 2395 |
+
"grad_norm": 5.159041881561279,
|
| 2396 |
+
"learning_rate": 0.00019998806288856592,
|
| 2397 |
+
"loss": 6.3091,
|
| 2398 |
+
"step": 339
|
| 2399 |
+
},
|
| 2400 |
+
{
|
| 2401 |
+
"epoch": 0.016226986911026,
|
| 2402 |
+
"grad_norm": 4.404861927032471,
|
| 2403 |
+
"learning_rate": 0.00019998798550223762,
|
| 2404 |
+
"loss": 7.1938,
|
| 2405 |
+
"step": 340
|
| 2406 |
+
},
|
| 2407 |
+
{
|
| 2408 |
+
"epoch": 0.016274713343117253,
|
| 2409 |
+
"grad_norm": 4.286370277404785,
|
| 2410 |
+
"learning_rate": 0.00019998790786589275,
|
| 2411 |
+
"loss": 7.7826,
|
| 2412 |
+
"step": 341
|
| 2413 |
+
},
|
| 2414 |
+
{
|
| 2415 |
+
"epoch": 0.016322439775208506,
|
| 2416 |
+
"grad_norm": 4.845062732696533,
|
| 2417 |
+
"learning_rate": 0.00019998782997953155,
|
| 2418 |
+
"loss": 7.5879,
|
| 2419 |
+
"step": 342
|
| 2420 |
+
},
|
| 2421 |
+
{
|
| 2422 |
+
"epoch": 0.01637016620729976,
|
| 2423 |
+
"grad_norm": 4.884746074676514,
|
| 2424 |
+
"learning_rate": 0.0001999877518431542,
|
| 2425 |
+
"loss": 6.0634,
|
| 2426 |
+
"step": 343
|
| 2427 |
+
},
|
| 2428 |
+
{
|
| 2429 |
+
"epoch": 0.016417892639391012,
|
| 2430 |
+
"grad_norm": 5.123991012573242,
|
| 2431 |
+
"learning_rate": 0.00019998767345676087,
|
| 2432 |
+
"loss": 7.1257,
|
| 2433 |
+
"step": 344
|
| 2434 |
+
},
|
| 2435 |
+
{
|
| 2436 |
+
"epoch": 0.016465619071482265,
|
| 2437 |
+
"grad_norm": 5.908779144287109,
|
| 2438 |
+
"learning_rate": 0.0001999875948203518,
|
| 2439 |
+
"loss": 9.0678,
|
| 2440 |
+
"step": 345
|
| 2441 |
+
},
|
| 2442 |
+
{
|
| 2443 |
+
"epoch": 0.016513345503573515,
|
| 2444 |
+
"grad_norm": 4.132058143615723,
|
| 2445 |
+
"learning_rate": 0.00019998751593392714,
|
| 2446 |
+
"loss": 5.7171,
|
| 2447 |
+
"step": 346
|
| 2448 |
+
},
|
| 2449 |
+
{
|
| 2450 |
+
"epoch": 0.016561071935664768,
|
| 2451 |
+
"grad_norm": 3.8147695064544678,
|
| 2452 |
+
"learning_rate": 0.00019998743679748712,
|
| 2453 |
+
"loss": 5.4728,
|
| 2454 |
+
"step": 347
|
| 2455 |
+
},
|
| 2456 |
+
{
|
| 2457 |
+
"epoch": 0.01660879836775602,
|
| 2458 |
+
"grad_norm": 3.7608211040496826,
|
| 2459 |
+
"learning_rate": 0.00019998735741103194,
|
| 2460 |
+
"loss": 5.5926,
|
| 2461 |
+
"step": 348
|
| 2462 |
+
},
|
| 2463 |
+
{
|
| 2464 |
+
"epoch": 0.016656524799847274,
|
| 2465 |
+
"grad_norm": 5.518613815307617,
|
| 2466 |
+
"learning_rate": 0.00019998727777456178,
|
| 2467 |
+
"loss": 7.8168,
|
| 2468 |
+
"step": 349
|
| 2469 |
+
},
|
| 2470 |
+
{
|
| 2471 |
+
"epoch": 0.016704251231938527,
|
| 2472 |
+
"grad_norm": 5.009797096252441,
|
| 2473 |
+
"learning_rate": 0.00019998719788807688,
|
| 2474 |
+
"loss": 7.069,
|
| 2475 |
+
"step": 350
|
| 2476 |
+
},
|
| 2477 |
+
{
|
| 2478 |
+
"epoch": 0.01675197766402978,
|
| 2479 |
+
"grad_norm": 4.233580589294434,
|
| 2480 |
+
"learning_rate": 0.00019998711775157734,
|
| 2481 |
+
"loss": 6.5305,
|
| 2482 |
+
"step": 351
|
| 2483 |
+
},
|
| 2484 |
+
{
|
| 2485 |
+
"epoch": 0.016799704096121033,
|
| 2486 |
+
"grad_norm": 3.6126842498779297,
|
| 2487 |
+
"learning_rate": 0.0001999870373650635,
|
| 2488 |
+
"loss": 4.6525,
|
| 2489 |
+
"step": 352
|
| 2490 |
+
},
|
| 2491 |
+
{
|
| 2492 |
+
"epoch": 0.016847430528212286,
|
| 2493 |
+
"grad_norm": 5.284352779388428,
|
| 2494 |
+
"learning_rate": 0.00019998695672853546,
|
| 2495 |
+
"loss": 7.4143,
|
| 2496 |
+
"step": 353
|
| 2497 |
+
},
|
| 2498 |
+
{
|
| 2499 |
+
"epoch": 0.01689515696030354,
|
| 2500 |
+
"grad_norm": 5.172549724578857,
|
| 2501 |
+
"learning_rate": 0.00019998687584199342,
|
| 2502 |
+
"loss": 7.0653,
|
| 2503 |
+
"step": 354
|
| 2504 |
+
},
|
| 2505 |
+
{
|
| 2506 |
+
"epoch": 0.016942883392394793,
|
| 2507 |
+
"grad_norm": 4.566983699798584,
|
| 2508 |
+
"learning_rate": 0.00019998679470543764,
|
| 2509 |
+
"loss": 7.1398,
|
| 2510 |
+
"step": 355
|
| 2511 |
+
},
|
| 2512 |
+
{
|
| 2513 |
+
"epoch": 0.016990609824486046,
|
| 2514 |
+
"grad_norm": 6.174628734588623,
|
| 2515 |
+
"learning_rate": 0.0001999867133188683,
|
| 2516 |
+
"loss": 6.8418,
|
| 2517 |
+
"step": 356
|
| 2518 |
+
},
|
| 2519 |
+
{
|
| 2520 |
+
"epoch": 0.0170383362565773,
|
| 2521 |
+
"grad_norm": 6.573761463165283,
|
| 2522 |
+
"learning_rate": 0.0001999866316822856,
|
| 2523 |
+
"loss": 7.3079,
|
| 2524 |
+
"step": 357
|
| 2525 |
+
},
|
| 2526 |
+
{
|
| 2527 |
+
"epoch": 0.017086062688668552,
|
| 2528 |
+
"grad_norm": 4.254095554351807,
|
| 2529 |
+
"learning_rate": 0.00019998654979568975,
|
| 2530 |
+
"loss": 5.5805,
|
| 2531 |
+
"step": 358
|
| 2532 |
+
},
|
| 2533 |
+
{
|
| 2534 |
+
"epoch": 0.017133789120759805,
|
| 2535 |
+
"grad_norm": 5.071292400360107,
|
| 2536 |
+
"learning_rate": 0.00019998646765908093,
|
| 2537 |
+
"loss": 7.3112,
|
| 2538 |
+
"step": 359
|
| 2539 |
+
},
|
| 2540 |
+
{
|
| 2541 |
+
"epoch": 0.017181515552851058,
|
| 2542 |
+
"grad_norm": 6.035824298858643,
|
| 2543 |
+
"learning_rate": 0.00019998638527245937,
|
| 2544 |
+
"loss": 7.6335,
|
| 2545 |
+
"step": 360
|
| 2546 |
+
},
|
| 2547 |
+
{
|
| 2548 |
+
"epoch": 0.01722924198494231,
|
| 2549 |
+
"grad_norm": 4.634344577789307,
|
| 2550 |
+
"learning_rate": 0.00019998630263582526,
|
| 2551 |
+
"loss": 6.9934,
|
| 2552 |
+
"step": 361
|
| 2553 |
+
},
|
| 2554 |
+
{
|
| 2555 |
+
"epoch": 0.017276968417033564,
|
| 2556 |
+
"grad_norm": 4.159761428833008,
|
| 2557 |
+
"learning_rate": 0.00019998621974917885,
|
| 2558 |
+
"loss": 5.933,
|
| 2559 |
+
"step": 362
|
| 2560 |
+
},
|
| 2561 |
+
{
|
| 2562 |
+
"epoch": 0.017324694849124817,
|
| 2563 |
+
"grad_norm": 5.788492679595947,
|
| 2564 |
+
"learning_rate": 0.0001999861366125203,
|
| 2565 |
+
"loss": 9.5469,
|
| 2566 |
+
"step": 363
|
| 2567 |
+
},
|
| 2568 |
+
{
|
| 2569 |
+
"epoch": 0.01737242128121607,
|
| 2570 |
+
"grad_norm": 5.0350542068481445,
|
| 2571 |
+
"learning_rate": 0.0001999860532258498,
|
| 2572 |
+
"loss": 7.0258,
|
| 2573 |
+
"step": 364
|
| 2574 |
+
},
|
| 2575 |
+
{
|
| 2576 |
+
"epoch": 0.017420147713307323,
|
| 2577 |
+
"grad_norm": 3.9429879188537598,
|
| 2578 |
+
"learning_rate": 0.00019998596958916763,
|
| 2579 |
+
"loss": 6.1842,
|
| 2580 |
+
"step": 365
|
| 2581 |
+
},
|
| 2582 |
+
{
|
| 2583 |
+
"epoch": 0.017467874145398576,
|
| 2584 |
+
"grad_norm": 4.744482517242432,
|
| 2585 |
+
"learning_rate": 0.00019998588570247395,
|
| 2586 |
+
"loss": 6.4803,
|
| 2587 |
+
"step": 366
|
| 2588 |
+
},
|
| 2589 |
+
{
|
| 2590 |
+
"epoch": 0.01751560057748983,
|
| 2591 |
+
"grad_norm": 3.6584131717681885,
|
| 2592 |
+
"learning_rate": 0.00019998580156576894,
|
| 2593 |
+
"loss": 4.6629,
|
| 2594 |
+
"step": 367
|
| 2595 |
+
},
|
| 2596 |
+
{
|
| 2597 |
+
"epoch": 0.017563327009581083,
|
| 2598 |
+
"grad_norm": 4.291102409362793,
|
| 2599 |
+
"learning_rate": 0.00019998571717905287,
|
| 2600 |
+
"loss": 6.2968,
|
| 2601 |
+
"step": 368
|
| 2602 |
+
},
|
| 2603 |
+
{
|
| 2604 |
+
"epoch": 0.017611053441672336,
|
| 2605 |
+
"grad_norm": 4.7433247566223145,
|
| 2606 |
+
"learning_rate": 0.00019998563254232594,
|
| 2607 |
+
"loss": 5.3619,
|
| 2608 |
+
"step": 369
|
| 2609 |
+
},
|
| 2610 |
+
{
|
| 2611 |
+
"epoch": 0.01765877987376359,
|
| 2612 |
+
"grad_norm": 7.7083048820495605,
|
| 2613 |
+
"learning_rate": 0.00019998554765558835,
|
| 2614 |
+
"loss": 7.1767,
|
| 2615 |
+
"step": 370
|
| 2616 |
+
},
|
| 2617 |
+
{
|
| 2618 |
+
"epoch": 0.017706506305854842,
|
| 2619 |
+
"grad_norm": 6.283130168914795,
|
| 2620 |
+
"learning_rate": 0.00019998546251884033,
|
| 2621 |
+
"loss": 5.8467,
|
| 2622 |
+
"step": 371
|
| 2623 |
+
},
|
| 2624 |
+
{
|
| 2625 |
+
"epoch": 0.01775423273794609,
|
| 2626 |
+
"grad_norm": 5.176546573638916,
|
| 2627 |
+
"learning_rate": 0.00019998537713208206,
|
| 2628 |
+
"loss": 7.7189,
|
| 2629 |
+
"step": 372
|
| 2630 |
+
},
|
| 2631 |
+
{
|
| 2632 |
+
"epoch": 0.017801959170037344,
|
| 2633 |
+
"grad_norm": 4.718236446380615,
|
| 2634 |
+
"learning_rate": 0.00019998529149531374,
|
| 2635 |
+
"loss": 7.1753,
|
| 2636 |
+
"step": 373
|
| 2637 |
+
},
|
| 2638 |
+
{
|
| 2639 |
+
"epoch": 0.017849685602128598,
|
| 2640 |
+
"grad_norm": 4.217880725860596,
|
| 2641 |
+
"learning_rate": 0.00019998520560853567,
|
| 2642 |
+
"loss": 6.6313,
|
| 2643 |
+
"step": 374
|
| 2644 |
+
},
|
| 2645 |
+
{
|
| 2646 |
+
"epoch": 0.01789741203421985,
|
| 2647 |
+
"grad_norm": 4.156486511230469,
|
| 2648 |
+
"learning_rate": 0.00019998511947174797,
|
| 2649 |
+
"loss": 6.8616,
|
| 2650 |
+
"step": 375
|
| 2651 |
+
},
|
| 2652 |
+
{
|
| 2653 |
+
"epoch": 0.017945138466311104,
|
| 2654 |
+
"grad_norm": 5.970060348510742,
|
| 2655 |
+
"learning_rate": 0.00019998503308495092,
|
| 2656 |
+
"loss": 8.0937,
|
| 2657 |
+
"step": 376
|
| 2658 |
+
},
|
| 2659 |
+
{
|
| 2660 |
+
"epoch": 0.017992864898402357,
|
| 2661 |
+
"grad_norm": 5.400389671325684,
|
| 2662 |
+
"learning_rate": 0.00019998494644814468,
|
| 2663 |
+
"loss": 7.7833,
|
| 2664 |
+
"step": 377
|
| 2665 |
+
},
|
| 2666 |
+
{
|
| 2667 |
+
"epoch": 0.01804059133049361,
|
| 2668 |
+
"grad_norm": 6.085134983062744,
|
| 2669 |
+
"learning_rate": 0.00019998485956132953,
|
| 2670 |
+
"loss": 8.5722,
|
| 2671 |
+
"step": 378
|
| 2672 |
+
},
|
| 2673 |
+
{
|
| 2674 |
+
"epoch": 0.018088317762584863,
|
| 2675 |
+
"grad_norm": 4.581125736236572,
|
| 2676 |
+
"learning_rate": 0.00019998477242450565,
|
| 2677 |
+
"loss": 4.9362,
|
| 2678 |
+
"step": 379
|
| 2679 |
+
},
|
| 2680 |
+
{
|
| 2681 |
+
"epoch": 0.018136044194676116,
|
| 2682 |
+
"grad_norm": 4.854187965393066,
|
| 2683 |
+
"learning_rate": 0.00019998468503767324,
|
| 2684 |
+
"loss": 5.7646,
|
| 2685 |
+
"step": 380
|
| 2686 |
+
},
|
| 2687 |
+
{
|
| 2688 |
+
"epoch": 0.01818377062676737,
|
| 2689 |
+
"grad_norm": 5.279097080230713,
|
| 2690 |
+
"learning_rate": 0.00019998459740083255,
|
| 2691 |
+
"loss": 5.9699,
|
| 2692 |
+
"step": 381
|
| 2693 |
+
},
|
| 2694 |
+
{
|
| 2695 |
+
"epoch": 0.018231497058858622,
|
| 2696 |
+
"grad_norm": 6.0616655349731445,
|
| 2697 |
+
"learning_rate": 0.0001999845095139838,
|
| 2698 |
+
"loss": 6.7272,
|
| 2699 |
+
"step": 382
|
| 2700 |
+
},
|
| 2701 |
+
{
|
| 2702 |
+
"epoch": 0.018279223490949875,
|
| 2703 |
+
"grad_norm": 4.58502197265625,
|
| 2704 |
+
"learning_rate": 0.00019998442137712718,
|
| 2705 |
+
"loss": 5.8863,
|
| 2706 |
+
"step": 383
|
| 2707 |
+
},
|
| 2708 |
+
{
|
| 2709 |
+
"epoch": 0.01832694992304113,
|
| 2710 |
+
"grad_norm": 5.454066753387451,
|
| 2711 |
+
"learning_rate": 0.00019998433299026297,
|
| 2712 |
+
"loss": 7.9997,
|
| 2713 |
+
"step": 384
|
| 2714 |
+
},
|
| 2715 |
+
{
|
| 2716 |
+
"epoch": 0.01837467635513238,
|
| 2717 |
+
"grad_norm": 5.5297017097473145,
|
| 2718 |
+
"learning_rate": 0.00019998424435339128,
|
| 2719 |
+
"loss": 9.2768,
|
| 2720 |
+
"step": 385
|
| 2721 |
+
},
|
| 2722 |
+
{
|
| 2723 |
+
"epoch": 0.018422402787223634,
|
| 2724 |
+
"grad_norm": 6.259234428405762,
|
| 2725 |
+
"learning_rate": 0.00019998415546651247,
|
| 2726 |
+
"loss": 7.4493,
|
| 2727 |
+
"step": 386
|
| 2728 |
+
},
|
| 2729 |
+
{
|
| 2730 |
+
"epoch": 0.018470129219314887,
|
| 2731 |
+
"grad_norm": 7.039416790008545,
|
| 2732 |
+
"learning_rate": 0.00019998406632962666,
|
| 2733 |
+
"loss": 9.3582,
|
| 2734 |
+
"step": 387
|
| 2735 |
+
},
|
| 2736 |
+
{
|
| 2737 |
+
"epoch": 0.01851785565140614,
|
| 2738 |
+
"grad_norm": 5.677418231964111,
|
| 2739 |
+
"learning_rate": 0.0001999839769427341,
|
| 2740 |
+
"loss": 6.5286,
|
| 2741 |
+
"step": 388
|
| 2742 |
+
},
|
| 2743 |
+
{
|
| 2744 |
+
"epoch": 0.018565582083497394,
|
| 2745 |
+
"grad_norm": 5.97587776184082,
|
| 2746 |
+
"learning_rate": 0.00019998388730583505,
|
| 2747 |
+
"loss": 7.9887,
|
| 2748 |
+
"step": 389
|
| 2749 |
+
},
|
| 2750 |
+
{
|
| 2751 |
+
"epoch": 0.018613308515588647,
|
| 2752 |
+
"grad_norm": 20.873300552368164,
|
| 2753 |
+
"learning_rate": 0.00019998379741892973,
|
| 2754 |
+
"loss": 8.7424,
|
| 2755 |
+
"step": 390
|
| 2756 |
+
},
|
| 2757 |
+
{
|
| 2758 |
+
"epoch": 0.0186610349476799,
|
| 2759 |
+
"grad_norm": 5.740731239318848,
|
| 2760 |
+
"learning_rate": 0.00019998370728201828,
|
| 2761 |
+
"loss": 8.7783,
|
| 2762 |
+
"step": 391
|
| 2763 |
+
},
|
| 2764 |
+
{
|
| 2765 |
+
"epoch": 0.018708761379771153,
|
| 2766 |
+
"grad_norm": 6.921543121337891,
|
| 2767 |
+
"learning_rate": 0.00019998361689510104,
|
| 2768 |
+
"loss": 7.8717,
|
| 2769 |
+
"step": 392
|
| 2770 |
+
},
|
| 2771 |
+
{
|
| 2772 |
+
"epoch": 0.018756487811862406,
|
| 2773 |
+
"grad_norm": 4.221953868865967,
|
| 2774 |
+
"learning_rate": 0.00019998352625817813,
|
| 2775 |
+
"loss": 6.5247,
|
| 2776 |
+
"step": 393
|
| 2777 |
+
},
|
| 2778 |
+
{
|
| 2779 |
+
"epoch": 0.01880421424395366,
|
| 2780 |
+
"grad_norm": 4.8449249267578125,
|
| 2781 |
+
"learning_rate": 0.00019998343537124986,
|
| 2782 |
+
"loss": 6.8002,
|
| 2783 |
+
"step": 394
|
| 2784 |
+
},
|
| 2785 |
+
{
|
| 2786 |
+
"epoch": 0.018851940676044912,
|
| 2787 |
+
"grad_norm": 4.915500164031982,
|
| 2788 |
+
"learning_rate": 0.0001999833442343164,
|
| 2789 |
+
"loss": 6.2396,
|
| 2790 |
+
"step": 395
|
| 2791 |
+
},
|
| 2792 |
+
{
|
| 2793 |
+
"epoch": 0.018899667108136165,
|
| 2794 |
+
"grad_norm": 3.3595244884490967,
|
| 2795 |
+
"learning_rate": 0.00019998325284737806,
|
| 2796 |
+
"loss": 4.481,
|
| 2797 |
+
"step": 396
|
| 2798 |
+
},
|
| 2799 |
+
{
|
| 2800 |
+
"epoch": 0.018947393540227415,
|
| 2801 |
+
"grad_norm": 4.997402667999268,
|
| 2802 |
+
"learning_rate": 0.00019998316121043493,
|
| 2803 |
+
"loss": 5.8485,
|
| 2804 |
+
"step": 397
|
| 2805 |
+
},
|
| 2806 |
+
{
|
| 2807 |
+
"epoch": 0.018995119972318668,
|
| 2808 |
+
"grad_norm": 5.331072807312012,
|
| 2809 |
+
"learning_rate": 0.0001999830693234874,
|
| 2810 |
+
"loss": 6.5482,
|
| 2811 |
+
"step": 398
|
| 2812 |
+
},
|
| 2813 |
+
{
|
| 2814 |
+
"epoch": 0.01904284640440992,
|
| 2815 |
+
"grad_norm": 5.49896764755249,
|
| 2816 |
+
"learning_rate": 0.00019998297718653557,
|
| 2817 |
+
"loss": 7.7411,
|
| 2818 |
+
"step": 399
|
| 2819 |
+
},
|
| 2820 |
+
{
|
| 2821 |
+
"epoch": 0.019090572836501174,
|
| 2822 |
+
"grad_norm": 5.086883068084717,
|
| 2823 |
+
"learning_rate": 0.0001999828847995797,
|
| 2824 |
+
"loss": 6.3793,
|
| 2825 |
+
"step": 400
|
| 2826 |
+
},
|
| 2827 |
+
{
|
| 2828 |
+
"epoch": 0.019090572836501174,
|
| 2829 |
+
"eval_loss": 1.665988564491272,
|
| 2830 |
+
"eval_runtime": 96.4558,
|
| 2831 |
+
"eval_samples_per_second": 8.74,
|
| 2832 |
+
"eval_steps_per_second": 4.375,
|
| 2833 |
+
"step": 400
|
| 2834 |
}
|
| 2835 |
],
|
| 2836 |
"logging_steps": 1,
|
|
|
|
| 2859 |
"attributes": {}
|
| 2860 |
}
|
| 2861 |
},
|
| 2862 |
+
"total_flos": 3.70892202835968e+16,
|
| 2863 |
"train_batch_size": 2,
|
| 2864 |
"trial_name": null,
|
| 2865 |
"trial_params": null
|