Training in progress, step 600, 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 295488936
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:7ebf9523ced4cd841390ae63e532fa706d54acc659271afe17daf4fe46a0fd81
|
| 3 |
size 295488936
|
last-checkpoint/optimizer.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 150487412
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:32fef6dc43e6ccba98378b739fc38b21a34b8c1abf8063665b01be9c8534559b
|
| 3 |
size 150487412
|
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 sha256:6135eda0e32d9cea2b960aff951c19589d3553b1f9b09a87975dc58dcbdb2c43
|
| 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:c4b7d5ebd1d82824f6c080341b10910a5587a67df516cd4ded83038671436c74
|
| 3 |
size 1064
|
last-checkpoint/trainer_state.json
CHANGED
|
@@ -1,9 +1,9 @@
|
|
| 1 |
{
|
| 2 |
-
"best_metric": 0.
|
| 3 |
-
"best_model_checkpoint": "miner_id_24/checkpoint-
|
| 4 |
-
"epoch": 0.
|
| 5 |
"eval_steps": 150,
|
| 6 |
-
"global_step":
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
@@ -2131,6 +2131,2122 @@
|
|
| 2131 |
"eval_samples_per_second": 46.421,
|
| 2132 |
"eval_steps_per_second": 11.609,
|
| 2133 |
"step": 300
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 2134 |
}
|
| 2135 |
],
|
| 2136 |
"logging_steps": 1,
|
|
@@ -2154,12 +4270,12 @@
|
|
| 2154 |
"should_evaluate": false,
|
| 2155 |
"should_log": false,
|
| 2156 |
"should_save": true,
|
| 2157 |
-
"should_training_stop":
|
| 2158 |
},
|
| 2159 |
"attributes": {}
|
| 2160 |
}
|
| 2161 |
},
|
| 2162 |
-
"total_flos":
|
| 2163 |
"train_batch_size": 16,
|
| 2164 |
"trial_name": null,
|
| 2165 |
"trial_params": null
|
|
|
|
| 1 |
{
|
| 2 |
+
"best_metric": 0.7933821678161621,
|
| 3 |
+
"best_model_checkpoint": "miner_id_24/checkpoint-600",
|
| 4 |
+
"epoch": 0.6037735849056604,
|
| 5 |
"eval_steps": 150,
|
| 6 |
+
"global_step": 600,
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
|
|
| 2131 |
"eval_samples_per_second": 46.421,
|
| 2132 |
"eval_steps_per_second": 11.609,
|
| 2133 |
"step": 300
|
| 2134 |
+
},
|
| 2135 |
+
{
|
| 2136 |
+
"epoch": 0.30289308176100627,
|
| 2137 |
+
"grad_norm": 0.2989448010921478,
|
| 2138 |
+
"learning_rate": 8.999753262144806e-05,
|
| 2139 |
+
"loss": 0.555,
|
| 2140 |
+
"step": 301
|
| 2141 |
+
},
|
| 2142 |
+
{
|
| 2143 |
+
"epoch": 0.3038993710691824,
|
| 2144 |
+
"grad_norm": 0.33250898122787476,
|
| 2145 |
+
"learning_rate": 8.999013075636805e-05,
|
| 2146 |
+
"loss": 0.6292,
|
| 2147 |
+
"step": 302
|
| 2148 |
+
},
|
| 2149 |
+
{
|
| 2150 |
+
"epoch": 0.3049056603773585,
|
| 2151 |
+
"grad_norm": 0.33274582028388977,
|
| 2152 |
+
"learning_rate": 8.997779521645793e-05,
|
| 2153 |
+
"loss": 0.6798,
|
| 2154 |
+
"step": 303
|
| 2155 |
+
},
|
| 2156 |
+
{
|
| 2157 |
+
"epoch": 0.3059119496855346,
|
| 2158 |
+
"grad_norm": 0.26289069652557373,
|
| 2159 |
+
"learning_rate": 8.996052735444863e-05,
|
| 2160 |
+
"loss": 0.6443,
|
| 2161 |
+
"step": 304
|
| 2162 |
+
},
|
| 2163 |
+
{
|
| 2164 |
+
"epoch": 0.3069182389937107,
|
| 2165 |
+
"grad_norm": 0.24797621369361877,
|
| 2166 |
+
"learning_rate": 8.993832906395582e-05,
|
| 2167 |
+
"loss": 0.6181,
|
| 2168 |
+
"step": 305
|
| 2169 |
+
},
|
| 2170 |
+
{
|
| 2171 |
+
"epoch": 0.30792452830188677,
|
| 2172 |
+
"grad_norm": 0.2333655059337616,
|
| 2173 |
+
"learning_rate": 8.991120277927223e-05,
|
| 2174 |
+
"loss": 0.6594,
|
| 2175 |
+
"step": 306
|
| 2176 |
+
},
|
| 2177 |
+
{
|
| 2178 |
+
"epoch": 0.3089308176100629,
|
| 2179 |
+
"grad_norm": 0.23653578758239746,
|
| 2180 |
+
"learning_rate": 8.987915147510061e-05,
|
| 2181 |
+
"loss": 0.5947,
|
| 2182 |
+
"step": 307
|
| 2183 |
+
},
|
| 2184 |
+
{
|
| 2185 |
+
"epoch": 0.309937106918239,
|
| 2186 |
+
"grad_norm": 0.2812041938304901,
|
| 2187 |
+
"learning_rate": 8.98421786662277e-05,
|
| 2188 |
+
"loss": 0.7297,
|
| 2189 |
+
"step": 308
|
| 2190 |
+
},
|
| 2191 |
+
{
|
| 2192 |
+
"epoch": 0.3109433962264151,
|
| 2193 |
+
"grad_norm": 0.28146234154701233,
|
| 2194 |
+
"learning_rate": 8.98002884071386e-05,
|
| 2195 |
+
"loss": 0.6798,
|
| 2196 |
+
"step": 309
|
| 2197 |
+
},
|
| 2198 |
+
{
|
| 2199 |
+
"epoch": 0.3119496855345912,
|
| 2200 |
+
"grad_norm": 0.2789071798324585,
|
| 2201 |
+
"learning_rate": 8.97534852915723e-05,
|
| 2202 |
+
"loss": 0.7207,
|
| 2203 |
+
"step": 310
|
| 2204 |
+
},
|
| 2205 |
+
{
|
| 2206 |
+
"epoch": 0.31295597484276727,
|
| 2207 |
+
"grad_norm": 0.27146562933921814,
|
| 2208 |
+
"learning_rate": 8.970177445201784e-05,
|
| 2209 |
+
"loss": 0.7418,
|
| 2210 |
+
"step": 311
|
| 2211 |
+
},
|
| 2212 |
+
{
|
| 2213 |
+
"epoch": 0.3139622641509434,
|
| 2214 |
+
"grad_norm": 0.27337393164634705,
|
| 2215 |
+
"learning_rate": 8.964516155915151e-05,
|
| 2216 |
+
"loss": 0.7257,
|
| 2217 |
+
"step": 312
|
| 2218 |
+
},
|
| 2219 |
+
{
|
| 2220 |
+
"epoch": 0.3149685534591195,
|
| 2221 |
+
"grad_norm": 0.29097968339920044,
|
| 2222 |
+
"learning_rate": 8.958365282121497e-05,
|
| 2223 |
+
"loss": 0.8054,
|
| 2224 |
+
"step": 313
|
| 2225 |
+
},
|
| 2226 |
+
{
|
| 2227 |
+
"epoch": 0.3159748427672956,
|
| 2228 |
+
"grad_norm": 0.29273757338523865,
|
| 2229 |
+
"learning_rate": 8.951725498333449e-05,
|
| 2230 |
+
"loss": 0.7513,
|
| 2231 |
+
"step": 314
|
| 2232 |
+
},
|
| 2233 |
+
{
|
| 2234 |
+
"epoch": 0.3169811320754717,
|
| 2235 |
+
"grad_norm": 0.27156925201416016,
|
| 2236 |
+
"learning_rate": 8.94459753267812e-05,
|
| 2237 |
+
"loss": 0.7676,
|
| 2238 |
+
"step": 315
|
| 2239 |
+
},
|
| 2240 |
+
{
|
| 2241 |
+
"epoch": 0.3179874213836478,
|
| 2242 |
+
"grad_norm": 0.2903954088687897,
|
| 2243 |
+
"learning_rate": 8.936982166817273e-05,
|
| 2244 |
+
"loss": 0.8394,
|
| 2245 |
+
"step": 316
|
| 2246 |
+
},
|
| 2247 |
+
{
|
| 2248 |
+
"epoch": 0.3189937106918239,
|
| 2249 |
+
"grad_norm": 0.30967020988464355,
|
| 2250 |
+
"learning_rate": 8.928880235861588e-05,
|
| 2251 |
+
"loss": 0.8251,
|
| 2252 |
+
"step": 317
|
| 2253 |
+
},
|
| 2254 |
+
{
|
| 2255 |
+
"epoch": 0.32,
|
| 2256 |
+
"grad_norm": 0.302234411239624,
|
| 2257 |
+
"learning_rate": 8.9202926282791e-05,
|
| 2258 |
+
"loss": 0.789,
|
| 2259 |
+
"step": 318
|
| 2260 |
+
},
|
| 2261 |
+
{
|
| 2262 |
+
"epoch": 0.3210062893081761,
|
| 2263 |
+
"grad_norm": 0.3056583106517792,
|
| 2264 |
+
"learning_rate": 8.911220285797748e-05,
|
| 2265 |
+
"loss": 0.815,
|
| 2266 |
+
"step": 319
|
| 2267 |
+
},
|
| 2268 |
+
{
|
| 2269 |
+
"epoch": 0.3220125786163522,
|
| 2270 |
+
"grad_norm": 0.3035435974597931,
|
| 2271 |
+
"learning_rate": 8.901664203302126e-05,
|
| 2272 |
+
"loss": 0.819,
|
| 2273 |
+
"step": 320
|
| 2274 |
+
},
|
| 2275 |
+
{
|
| 2276 |
+
"epoch": 0.3230188679245283,
|
| 2277 |
+
"grad_norm": 0.3102681636810303,
|
| 2278 |
+
"learning_rate": 8.891625428724365e-05,
|
| 2279 |
+
"loss": 0.8169,
|
| 2280 |
+
"step": 321
|
| 2281 |
+
},
|
| 2282 |
+
{
|
| 2283 |
+
"epoch": 0.3240251572327044,
|
| 2284 |
+
"grad_norm": 0.2978799045085907,
|
| 2285 |
+
"learning_rate": 8.881105062929222e-05,
|
| 2286 |
+
"loss": 0.7624,
|
| 2287 |
+
"step": 322
|
| 2288 |
+
},
|
| 2289 |
+
{
|
| 2290 |
+
"epoch": 0.3250314465408805,
|
| 2291 |
+
"grad_norm": 0.32888859510421753,
|
| 2292 |
+
"learning_rate": 8.870104259593363e-05,
|
| 2293 |
+
"loss": 0.8504,
|
| 2294 |
+
"step": 323
|
| 2295 |
+
},
|
| 2296 |
+
{
|
| 2297 |
+
"epoch": 0.3260377358490566,
|
| 2298 |
+
"grad_norm": 0.32155942916870117,
|
| 2299 |
+
"learning_rate": 8.858624225078841e-05,
|
| 2300 |
+
"loss": 0.7412,
|
| 2301 |
+
"step": 324
|
| 2302 |
+
},
|
| 2303 |
+
{
|
| 2304 |
+
"epoch": 0.3270440251572327,
|
| 2305 |
+
"grad_norm": 0.312429279088974,
|
| 2306 |
+
"learning_rate": 8.846666218300807e-05,
|
| 2307 |
+
"loss": 0.8747,
|
| 2308 |
+
"step": 325
|
| 2309 |
+
},
|
| 2310 |
+
{
|
| 2311 |
+
"epoch": 0.32805031446540883,
|
| 2312 |
+
"grad_norm": 0.3254411518573761,
|
| 2313 |
+
"learning_rate": 8.834231550589462e-05,
|
| 2314 |
+
"loss": 0.8664,
|
| 2315 |
+
"step": 326
|
| 2316 |
+
},
|
| 2317 |
+
{
|
| 2318 |
+
"epoch": 0.3290566037735849,
|
| 2319 |
+
"grad_norm": 0.354501336812973,
|
| 2320 |
+
"learning_rate": 8.821321585546244e-05,
|
| 2321 |
+
"loss": 0.8942,
|
| 2322 |
+
"step": 327
|
| 2323 |
+
},
|
| 2324 |
+
{
|
| 2325 |
+
"epoch": 0.330062893081761,
|
| 2326 |
+
"grad_norm": 0.3486211597919464,
|
| 2327 |
+
"learning_rate": 8.807937738894302e-05,
|
| 2328 |
+
"loss": 0.8619,
|
| 2329 |
+
"step": 328
|
| 2330 |
+
},
|
| 2331 |
+
{
|
| 2332 |
+
"epoch": 0.3310691823899371,
|
| 2333 |
+
"grad_norm": 0.34248775243759155,
|
| 2334 |
+
"learning_rate": 8.794081478323246e-05,
|
| 2335 |
+
"loss": 0.8448,
|
| 2336 |
+
"step": 329
|
| 2337 |
+
},
|
| 2338 |
+
{
|
| 2339 |
+
"epoch": 0.3320754716981132,
|
| 2340 |
+
"grad_norm": 0.34060388803482056,
|
| 2341 |
+
"learning_rate": 8.779754323328193e-05,
|
| 2342 |
+
"loss": 0.7516,
|
| 2343 |
+
"step": 330
|
| 2344 |
+
},
|
| 2345 |
+
{
|
| 2346 |
+
"epoch": 0.33308176100628933,
|
| 2347 |
+
"grad_norm": 0.3704814016819,
|
| 2348 |
+
"learning_rate": 8.764957845043137e-05,
|
| 2349 |
+
"loss": 0.8106,
|
| 2350 |
+
"step": 331
|
| 2351 |
+
},
|
| 2352 |
+
{
|
| 2353 |
+
"epoch": 0.3340880503144654,
|
| 2354 |
+
"grad_norm": 1.31474769115448,
|
| 2355 |
+
"learning_rate": 8.749693666068665e-05,
|
| 2356 |
+
"loss": 0.8051,
|
| 2357 |
+
"step": 332
|
| 2358 |
+
},
|
| 2359 |
+
{
|
| 2360 |
+
"epoch": 0.3350943396226415,
|
| 2361 |
+
"grad_norm": 0.3882513642311096,
|
| 2362 |
+
"learning_rate": 8.733963460294016e-05,
|
| 2363 |
+
"loss": 0.8016,
|
| 2364 |
+
"step": 333
|
| 2365 |
+
},
|
| 2366 |
+
{
|
| 2367 |
+
"epoch": 0.3361006289308176,
|
| 2368 |
+
"grad_norm": 0.438913494348526,
|
| 2369 |
+
"learning_rate": 8.717768952713513e-05,
|
| 2370 |
+
"loss": 0.8598,
|
| 2371 |
+
"step": 334
|
| 2372 |
+
},
|
| 2373 |
+
{
|
| 2374 |
+
"epoch": 0.3371069182389937,
|
| 2375 |
+
"grad_norm": 0.4014555513858795,
|
| 2376 |
+
"learning_rate": 8.701111919237408e-05,
|
| 2377 |
+
"loss": 0.8749,
|
| 2378 |
+
"step": 335
|
| 2379 |
+
},
|
| 2380 |
+
{
|
| 2381 |
+
"epoch": 0.33811320754716984,
|
| 2382 |
+
"grad_norm": 0.41690000891685486,
|
| 2383 |
+
"learning_rate": 8.683994186497131e-05,
|
| 2384 |
+
"loss": 0.8127,
|
| 2385 |
+
"step": 336
|
| 2386 |
+
},
|
| 2387 |
+
{
|
| 2388 |
+
"epoch": 0.3391194968553459,
|
| 2389 |
+
"grad_norm": 0.456018328666687,
|
| 2390 |
+
"learning_rate": 8.666417631644977e-05,
|
| 2391 |
+
"loss": 0.8541,
|
| 2392 |
+
"step": 337
|
| 2393 |
+
},
|
| 2394 |
+
{
|
| 2395 |
+
"epoch": 0.340125786163522,
|
| 2396 |
+
"grad_norm": 0.4461841881275177,
|
| 2397 |
+
"learning_rate": 8.648384182148252e-05,
|
| 2398 |
+
"loss": 0.7593,
|
| 2399 |
+
"step": 338
|
| 2400 |
+
},
|
| 2401 |
+
{
|
| 2402 |
+
"epoch": 0.3411320754716981,
|
| 2403 |
+
"grad_norm": 0.5013781189918518,
|
| 2404 |
+
"learning_rate": 8.629895815577916e-05,
|
| 2405 |
+
"loss": 0.8718,
|
| 2406 |
+
"step": 339
|
| 2407 |
+
},
|
| 2408 |
+
{
|
| 2409 |
+
"epoch": 0.3421383647798742,
|
| 2410 |
+
"grad_norm": 0.5359609723091125,
|
| 2411 |
+
"learning_rate": 8.610954559391704e-05,
|
| 2412 |
+
"loss": 0.809,
|
| 2413 |
+
"step": 340
|
| 2414 |
+
},
|
| 2415 |
+
{
|
| 2416 |
+
"epoch": 0.34314465408805034,
|
| 2417 |
+
"grad_norm": 0.5478375554084778,
|
| 2418 |
+
"learning_rate": 8.59156249071181e-05,
|
| 2419 |
+
"loss": 0.9652,
|
| 2420 |
+
"step": 341
|
| 2421 |
+
},
|
| 2422 |
+
{
|
| 2423 |
+
"epoch": 0.3441509433962264,
|
| 2424 |
+
"grad_norm": 0.6687819361686707,
|
| 2425 |
+
"learning_rate": 8.571721736097089e-05,
|
| 2426 |
+
"loss": 0.9975,
|
| 2427 |
+
"step": 342
|
| 2428 |
+
},
|
| 2429 |
+
{
|
| 2430 |
+
"epoch": 0.3451572327044025,
|
| 2431 |
+
"grad_norm": 0.6184760332107544,
|
| 2432 |
+
"learning_rate": 8.551434471309872e-05,
|
| 2433 |
+
"loss": 1.0351,
|
| 2434 |
+
"step": 343
|
| 2435 |
+
},
|
| 2436 |
+
{
|
| 2437 |
+
"epoch": 0.3461635220125786,
|
| 2438 |
+
"grad_norm": 0.6045582294464111,
|
| 2439 |
+
"learning_rate": 8.530702921077359e-05,
|
| 2440 |
+
"loss": 0.8445,
|
| 2441 |
+
"step": 344
|
| 2442 |
+
},
|
| 2443 |
+
{
|
| 2444 |
+
"epoch": 0.3471698113207547,
|
| 2445 |
+
"grad_norm": 0.6508563160896301,
|
| 2446 |
+
"learning_rate": 8.509529358847655e-05,
|
| 2447 |
+
"loss": 0.9425,
|
| 2448 |
+
"step": 345
|
| 2449 |
+
},
|
| 2450 |
+
{
|
| 2451 |
+
"epoch": 0.34817610062893084,
|
| 2452 |
+
"grad_norm": 0.6897462606430054,
|
| 2453 |
+
"learning_rate": 8.487916106540466e-05,
|
| 2454 |
+
"loss": 0.945,
|
| 2455 |
+
"step": 346
|
| 2456 |
+
},
|
| 2457 |
+
{
|
| 2458 |
+
"epoch": 0.34918238993710693,
|
| 2459 |
+
"grad_norm": 0.7266098856925964,
|
| 2460 |
+
"learning_rate": 8.465865534292465e-05,
|
| 2461 |
+
"loss": 0.9754,
|
| 2462 |
+
"step": 347
|
| 2463 |
+
},
|
| 2464 |
+
{
|
| 2465 |
+
"epoch": 0.350188679245283,
|
| 2466 |
+
"grad_norm": 0.796158492565155,
|
| 2467 |
+
"learning_rate": 8.443380060197386e-05,
|
| 2468 |
+
"loss": 0.9235,
|
| 2469 |
+
"step": 348
|
| 2470 |
+
},
|
| 2471 |
+
{
|
| 2472 |
+
"epoch": 0.3511949685534591,
|
| 2473 |
+
"grad_norm": 1.028557300567627,
|
| 2474 |
+
"learning_rate": 8.420462150040853e-05,
|
| 2475 |
+
"loss": 1.0911,
|
| 2476 |
+
"step": 349
|
| 2477 |
+
},
|
| 2478 |
+
{
|
| 2479 |
+
"epoch": 0.3522012578616352,
|
| 2480 |
+
"grad_norm": 1.4935519695281982,
|
| 2481 |
+
"learning_rate": 8.397114317029975e-05,
|
| 2482 |
+
"loss": 1.034,
|
| 2483 |
+
"step": 350
|
| 2484 |
+
},
|
| 2485 |
+
{
|
| 2486 |
+
"epoch": 0.35320754716981134,
|
| 2487 |
+
"grad_norm": 0.3518058955669403,
|
| 2488 |
+
"learning_rate": 8.373339121517747e-05,
|
| 2489 |
+
"loss": 0.6323,
|
| 2490 |
+
"step": 351
|
| 2491 |
+
},
|
| 2492 |
+
{
|
| 2493 |
+
"epoch": 0.35421383647798743,
|
| 2494 |
+
"grad_norm": 0.3483763039112091,
|
| 2495 |
+
"learning_rate": 8.34913917072228e-05,
|
| 2496 |
+
"loss": 0.559,
|
| 2497 |
+
"step": 352
|
| 2498 |
+
},
|
| 2499 |
+
{
|
| 2500 |
+
"epoch": 0.3552201257861635,
|
| 2501 |
+
"grad_norm": 0.33547666668891907,
|
| 2502 |
+
"learning_rate": 8.324517118440889e-05,
|
| 2503 |
+
"loss": 0.5886,
|
| 2504 |
+
"step": 353
|
| 2505 |
+
},
|
| 2506 |
+
{
|
| 2507 |
+
"epoch": 0.3562264150943396,
|
| 2508 |
+
"grad_norm": 0.2576180696487427,
|
| 2509 |
+
"learning_rate": 8.299475664759068e-05,
|
| 2510 |
+
"loss": 0.674,
|
| 2511 |
+
"step": 354
|
| 2512 |
+
},
|
| 2513 |
+
{
|
| 2514 |
+
"epoch": 0.3572327044025157,
|
| 2515 |
+
"grad_norm": 0.21154595911502838,
|
| 2516 |
+
"learning_rate": 8.274017555754409e-05,
|
| 2517 |
+
"loss": 0.6139,
|
| 2518 |
+
"step": 355
|
| 2519 |
+
},
|
| 2520 |
+
{
|
| 2521 |
+
"epoch": 0.35823899371069184,
|
| 2522 |
+
"grad_norm": 0.23689445853233337,
|
| 2523 |
+
"learning_rate": 8.248145583195448e-05,
|
| 2524 |
+
"loss": 0.6871,
|
| 2525 |
+
"step": 356
|
| 2526 |
+
},
|
| 2527 |
+
{
|
| 2528 |
+
"epoch": 0.35924528301886793,
|
| 2529 |
+
"grad_norm": 0.2390032261610031,
|
| 2530 |
+
"learning_rate": 8.221862584235528e-05,
|
| 2531 |
+
"loss": 0.6175,
|
| 2532 |
+
"step": 357
|
| 2533 |
+
},
|
| 2534 |
+
{
|
| 2535 |
+
"epoch": 0.360251572327044,
|
| 2536 |
+
"grad_norm": 0.26824384927749634,
|
| 2537 |
+
"learning_rate": 8.195171441101669e-05,
|
| 2538 |
+
"loss": 0.7,
|
| 2539 |
+
"step": 358
|
| 2540 |
+
},
|
| 2541 |
+
{
|
| 2542 |
+
"epoch": 0.3612578616352201,
|
| 2543 |
+
"grad_norm": 0.2665193974971771,
|
| 2544 |
+
"learning_rate": 8.168075080778494e-05,
|
| 2545 |
+
"loss": 0.671,
|
| 2546 |
+
"step": 359
|
| 2547 |
+
},
|
| 2548 |
+
{
|
| 2549 |
+
"epoch": 0.3622641509433962,
|
| 2550 |
+
"grad_norm": 0.2973785698413849,
|
| 2551 |
+
"learning_rate": 8.140576474687264e-05,
|
| 2552 |
+
"loss": 0.7634,
|
| 2553 |
+
"step": 360
|
| 2554 |
+
},
|
| 2555 |
+
{
|
| 2556 |
+
"epoch": 0.36327044025157235,
|
| 2557 |
+
"grad_norm": 0.2718108296394348,
|
| 2558 |
+
"learning_rate": 8.112678638360016e-05,
|
| 2559 |
+
"loss": 0.6838,
|
| 2560 |
+
"step": 361
|
| 2561 |
+
},
|
| 2562 |
+
{
|
| 2563 |
+
"epoch": 0.36427672955974844,
|
| 2564 |
+
"grad_norm": 0.27566713094711304,
|
| 2565 |
+
"learning_rate": 8.084384631108883e-05,
|
| 2566 |
+
"loss": 0.6967,
|
| 2567 |
+
"step": 362
|
| 2568 |
+
},
|
| 2569 |
+
{
|
| 2570 |
+
"epoch": 0.3652830188679245,
|
| 2571 |
+
"grad_norm": 0.24386361241340637,
|
| 2572 |
+
"learning_rate": 8.055697555690607e-05,
|
| 2573 |
+
"loss": 0.6934,
|
| 2574 |
+
"step": 363
|
| 2575 |
+
},
|
| 2576 |
+
{
|
| 2577 |
+
"epoch": 0.3662893081761006,
|
| 2578 |
+
"grad_norm": 0.28102463483810425,
|
| 2579 |
+
"learning_rate": 8.02662055796628e-05,
|
| 2580 |
+
"loss": 0.7469,
|
| 2581 |
+
"step": 364
|
| 2582 |
+
},
|
| 2583 |
+
{
|
| 2584 |
+
"epoch": 0.3672955974842767,
|
| 2585 |
+
"grad_norm": 0.27854305505752563,
|
| 2586 |
+
"learning_rate": 7.99715682655637e-05,
|
| 2587 |
+
"loss": 0.76,
|
| 2588 |
+
"step": 365
|
| 2589 |
+
},
|
| 2590 |
+
{
|
| 2591 |
+
"epoch": 0.36830188679245285,
|
| 2592 |
+
"grad_norm": 0.32097896933555603,
|
| 2593 |
+
"learning_rate": 7.967309592491052e-05,
|
| 2594 |
+
"loss": 0.8277,
|
| 2595 |
+
"step": 366
|
| 2596 |
+
},
|
| 2597 |
+
{
|
| 2598 |
+
"epoch": 0.36930817610062894,
|
| 2599 |
+
"grad_norm": 0.338775098323822,
|
| 2600 |
+
"learning_rate": 7.937082128855891e-05,
|
| 2601 |
+
"loss": 0.8518,
|
| 2602 |
+
"step": 367
|
| 2603 |
+
},
|
| 2604 |
+
{
|
| 2605 |
+
"epoch": 0.370314465408805,
|
| 2606 |
+
"grad_norm": 0.3111404478549957,
|
| 2607 |
+
"learning_rate": 7.906477750432904e-05,
|
| 2608 |
+
"loss": 0.8281,
|
| 2609 |
+
"step": 368
|
| 2610 |
+
},
|
| 2611 |
+
{
|
| 2612 |
+
"epoch": 0.3713207547169811,
|
| 2613 |
+
"grad_norm": 0.29169967770576477,
|
| 2614 |
+
"learning_rate": 7.875499813337069e-05,
|
| 2615 |
+
"loss": 0.7992,
|
| 2616 |
+
"step": 369
|
| 2617 |
+
},
|
| 2618 |
+
{
|
| 2619 |
+
"epoch": 0.3723270440251572,
|
| 2620 |
+
"grad_norm": 0.2766212522983551,
|
| 2621 |
+
"learning_rate": 7.844151714648274e-05,
|
| 2622 |
+
"loss": 0.7662,
|
| 2623 |
+
"step": 370
|
| 2624 |
+
},
|
| 2625 |
+
{
|
| 2626 |
+
"epoch": 0.37333333333333335,
|
| 2627 |
+
"grad_norm": 0.2962402403354645,
|
| 2628 |
+
"learning_rate": 7.812436892038805e-05,
|
| 2629 |
+
"loss": 0.8313,
|
| 2630 |
+
"step": 371
|
| 2631 |
+
},
|
| 2632 |
+
{
|
| 2633 |
+
"epoch": 0.37433962264150944,
|
| 2634 |
+
"grad_norm": 0.292148619890213,
|
| 2635 |
+
"learning_rate": 7.780358823396352e-05,
|
| 2636 |
+
"loss": 0.8729,
|
| 2637 |
+
"step": 372
|
| 2638 |
+
},
|
| 2639 |
+
{
|
| 2640 |
+
"epoch": 0.37534591194968553,
|
| 2641 |
+
"grad_norm": 0.3062426447868347,
|
| 2642 |
+
"learning_rate": 7.747921026442631e-05,
|
| 2643 |
+
"loss": 0.8486,
|
| 2644 |
+
"step": 373
|
| 2645 |
+
},
|
| 2646 |
+
{
|
| 2647 |
+
"epoch": 0.3763522012578616,
|
| 2648 |
+
"grad_norm": 0.3405818045139313,
|
| 2649 |
+
"learning_rate": 7.715127058347615e-05,
|
| 2650 |
+
"loss": 0.8567,
|
| 2651 |
+
"step": 374
|
| 2652 |
+
},
|
| 2653 |
+
{
|
| 2654 |
+
"epoch": 0.37735849056603776,
|
| 2655 |
+
"grad_norm": 0.3373412489891052,
|
| 2656 |
+
"learning_rate": 7.681980515339464e-05,
|
| 2657 |
+
"loss": 0.8583,
|
| 2658 |
+
"step": 375
|
| 2659 |
+
},
|
| 2660 |
+
{
|
| 2661 |
+
"epoch": 0.37836477987421385,
|
| 2662 |
+
"grad_norm": 0.33258765935897827,
|
| 2663 |
+
"learning_rate": 7.648485032310145e-05,
|
| 2664 |
+
"loss": 0.8829,
|
| 2665 |
+
"step": 376
|
| 2666 |
+
},
|
| 2667 |
+
{
|
| 2668 |
+
"epoch": 0.37937106918238994,
|
| 2669 |
+
"grad_norm": 0.3391346037387848,
|
| 2670 |
+
"learning_rate": 7.614644282416831e-05,
|
| 2671 |
+
"loss": 0.871,
|
| 2672 |
+
"step": 377
|
| 2673 |
+
},
|
| 2674 |
+
{
|
| 2675 |
+
"epoch": 0.38037735849056603,
|
| 2676 |
+
"grad_norm": 0.35854992270469666,
|
| 2677 |
+
"learning_rate": 7.5804619766791e-05,
|
| 2678 |
+
"loss": 0.8654,
|
| 2679 |
+
"step": 378
|
| 2680 |
+
},
|
| 2681 |
+
{
|
| 2682 |
+
"epoch": 0.3813836477987421,
|
| 2683 |
+
"grad_norm": 0.3469408452510834,
|
| 2684 |
+
"learning_rate": 7.545941863571974e-05,
|
| 2685 |
+
"loss": 0.8359,
|
| 2686 |
+
"step": 379
|
| 2687 |
+
},
|
| 2688 |
+
{
|
| 2689 |
+
"epoch": 0.38238993710691827,
|
| 2690 |
+
"grad_norm": 0.3572901785373688,
|
| 2691 |
+
"learning_rate": 7.511087728614862e-05,
|
| 2692 |
+
"loss": 0.8464,
|
| 2693 |
+
"step": 380
|
| 2694 |
+
},
|
| 2695 |
+
{
|
| 2696 |
+
"epoch": 0.38339622641509435,
|
| 2697 |
+
"grad_norm": 0.35687223076820374,
|
| 2698 |
+
"learning_rate": 7.475903393956434e-05,
|
| 2699 |
+
"loss": 0.7826,
|
| 2700 |
+
"step": 381
|
| 2701 |
+
},
|
| 2702 |
+
{
|
| 2703 |
+
"epoch": 0.38440251572327044,
|
| 2704 |
+
"grad_norm": 0.3538658022880554,
|
| 2705 |
+
"learning_rate": 7.440392717955476e-05,
|
| 2706 |
+
"loss": 0.8013,
|
| 2707 |
+
"step": 382
|
| 2708 |
+
},
|
| 2709 |
+
{
|
| 2710 |
+
"epoch": 0.38540880503144653,
|
| 2711 |
+
"grad_norm": 0.39203977584838867,
|
| 2712 |
+
"learning_rate": 7.404559594757779e-05,
|
| 2713 |
+
"loss": 0.8344,
|
| 2714 |
+
"step": 383
|
| 2715 |
+
},
|
| 2716 |
+
{
|
| 2717 |
+
"epoch": 0.3864150943396226,
|
| 2718 |
+
"grad_norm": 0.4001169502735138,
|
| 2719 |
+
"learning_rate": 7.368407953869104e-05,
|
| 2720 |
+
"loss": 0.7899,
|
| 2721 |
+
"step": 384
|
| 2722 |
+
},
|
| 2723 |
+
{
|
| 2724 |
+
"epoch": 0.38742138364779877,
|
| 2725 |
+
"grad_norm": 0.41093799471855164,
|
| 2726 |
+
"learning_rate": 7.33194175972427e-05,
|
| 2727 |
+
"loss": 0.9041,
|
| 2728 |
+
"step": 385
|
| 2729 |
+
},
|
| 2730 |
+
{
|
| 2731 |
+
"epoch": 0.38842767295597486,
|
| 2732 |
+
"grad_norm": 0.4247496724128723,
|
| 2733 |
+
"learning_rate": 7.295165011252397e-05,
|
| 2734 |
+
"loss": 0.8063,
|
| 2735 |
+
"step": 386
|
| 2736 |
+
},
|
| 2737 |
+
{
|
| 2738 |
+
"epoch": 0.38943396226415095,
|
| 2739 |
+
"grad_norm": 0.4747501313686371,
|
| 2740 |
+
"learning_rate": 7.258081741438396e-05,
|
| 2741 |
+
"loss": 0.8265,
|
| 2742 |
+
"step": 387
|
| 2743 |
+
},
|
| 2744 |
+
{
|
| 2745 |
+
"epoch": 0.39044025157232704,
|
| 2746 |
+
"grad_norm": 0.4680624008178711,
|
| 2747 |
+
"learning_rate": 7.220696016880688e-05,
|
| 2748 |
+
"loss": 0.8525,
|
| 2749 |
+
"step": 388
|
| 2750 |
+
},
|
| 2751 |
+
{
|
| 2752 |
+
"epoch": 0.3914465408805031,
|
| 2753 |
+
"grad_norm": 0.45939162373542786,
|
| 2754 |
+
"learning_rate": 7.183011937345271e-05,
|
| 2755 |
+
"loss": 0.8456,
|
| 2756 |
+
"step": 389
|
| 2757 |
+
},
|
| 2758 |
+
{
|
| 2759 |
+
"epoch": 0.39245283018867927,
|
| 2760 |
+
"grad_norm": 0.5063957571983337,
|
| 2761 |
+
"learning_rate": 7.14503363531613e-05,
|
| 2762 |
+
"loss": 0.89,
|
| 2763 |
+
"step": 390
|
| 2764 |
+
},
|
| 2765 |
+
{
|
| 2766 |
+
"epoch": 0.39345911949685536,
|
| 2767 |
+
"grad_norm": 0.5353115797042847,
|
| 2768 |
+
"learning_rate": 7.106765275542055e-05,
|
| 2769 |
+
"loss": 0.8865,
|
| 2770 |
+
"step": 391
|
| 2771 |
+
},
|
| 2772 |
+
{
|
| 2773 |
+
"epoch": 0.39446540880503145,
|
| 2774 |
+
"grad_norm": 0.6021161079406738,
|
| 2775 |
+
"learning_rate": 7.068211054579944e-05,
|
| 2776 |
+
"loss": 0.9897,
|
| 2777 |
+
"step": 392
|
| 2778 |
+
},
|
| 2779 |
+
{
|
| 2780 |
+
"epoch": 0.39547169811320754,
|
| 2781 |
+
"grad_norm": 0.6731549501419067,
|
| 2782 |
+
"learning_rate": 7.029375200334588e-05,
|
| 2783 |
+
"loss": 0.9746,
|
| 2784 |
+
"step": 393
|
| 2785 |
+
},
|
| 2786 |
+
{
|
| 2787 |
+
"epoch": 0.3964779874213836,
|
| 2788 |
+
"grad_norm": 0.6880476474761963,
|
| 2789 |
+
"learning_rate": 6.99026197159505e-05,
|
| 2790 |
+
"loss": 1.0113,
|
| 2791 |
+
"step": 394
|
| 2792 |
+
},
|
| 2793 |
+
{
|
| 2794 |
+
"epoch": 0.39748427672955977,
|
| 2795 |
+
"grad_norm": 0.5883579254150391,
|
| 2796 |
+
"learning_rate": 6.950875657567623e-05,
|
| 2797 |
+
"loss": 0.9356,
|
| 2798 |
+
"step": 395
|
| 2799 |
+
},
|
| 2800 |
+
{
|
| 2801 |
+
"epoch": 0.39849056603773586,
|
| 2802 |
+
"grad_norm": 0.7087289094924927,
|
| 2803 |
+
"learning_rate": 6.911220577405484e-05,
|
| 2804 |
+
"loss": 0.9998,
|
| 2805 |
+
"step": 396
|
| 2806 |
+
},
|
| 2807 |
+
{
|
| 2808 |
+
"epoch": 0.39949685534591195,
|
| 2809 |
+
"grad_norm": 0.7184171676635742,
|
| 2810 |
+
"learning_rate": 6.87130107973505e-05,
|
| 2811 |
+
"loss": 0.9281,
|
| 2812 |
+
"step": 397
|
| 2813 |
+
},
|
| 2814 |
+
{
|
| 2815 |
+
"epoch": 0.40050314465408804,
|
| 2816 |
+
"grad_norm": 0.7750122547149658,
|
| 2817 |
+
"learning_rate": 6.831121542179087e-05,
|
| 2818 |
+
"loss": 0.9402,
|
| 2819 |
+
"step": 398
|
| 2820 |
+
},
|
| 2821 |
+
{
|
| 2822 |
+
"epoch": 0.40150943396226413,
|
| 2823 |
+
"grad_norm": 0.8327257633209229,
|
| 2824 |
+
"learning_rate": 6.790686370876671e-05,
|
| 2825 |
+
"loss": 0.9312,
|
| 2826 |
+
"step": 399
|
| 2827 |
+
},
|
| 2828 |
+
{
|
| 2829 |
+
"epoch": 0.4025157232704403,
|
| 2830 |
+
"grad_norm": 1.3496246337890625,
|
| 2831 |
+
"learning_rate": 6.75e-05,
|
| 2832 |
+
"loss": 0.8687,
|
| 2833 |
+
"step": 400
|
| 2834 |
+
},
|
| 2835 |
+
{
|
| 2836 |
+
"epoch": 0.40352201257861636,
|
| 2837 |
+
"grad_norm": 0.22164778411388397,
|
| 2838 |
+
"learning_rate": 6.709066891268135e-05,
|
| 2839 |
+
"loss": 0.5538,
|
| 2840 |
+
"step": 401
|
| 2841 |
+
},
|
| 2842 |
+
{
|
| 2843 |
+
"epoch": 0.40452830188679245,
|
| 2844 |
+
"grad_norm": 0.2639901340007782,
|
| 2845 |
+
"learning_rate": 6.667891533457719e-05,
|
| 2846 |
+
"loss": 0.6009,
|
| 2847 |
+
"step": 402
|
| 2848 |
+
},
|
| 2849 |
+
{
|
| 2850 |
+
"epoch": 0.40553459119496854,
|
| 2851 |
+
"grad_norm": 0.2394772469997406,
|
| 2852 |
+
"learning_rate": 6.626478441910744e-05,
|
| 2853 |
+
"loss": 0.5637,
|
| 2854 |
+
"step": 403
|
| 2855 |
+
},
|
| 2856 |
+
{
|
| 2857 |
+
"epoch": 0.40654088050314463,
|
| 2858 |
+
"grad_norm": 0.2308882623910904,
|
| 2859 |
+
"learning_rate": 6.584832158039378e-05,
|
| 2860 |
+
"loss": 0.6271,
|
| 2861 |
+
"step": 404
|
| 2862 |
+
},
|
| 2863 |
+
{
|
| 2864 |
+
"epoch": 0.4075471698113208,
|
| 2865 |
+
"grad_norm": 0.22672177851200104,
|
| 2866 |
+
"learning_rate": 6.542957248827961e-05,
|
| 2867 |
+
"loss": 0.6491,
|
| 2868 |
+
"step": 405
|
| 2869 |
+
},
|
| 2870 |
+
{
|
| 2871 |
+
"epoch": 0.40855345911949686,
|
| 2872 |
+
"grad_norm": 0.21780024468898773,
|
| 2873 |
+
"learning_rate": 6.500858306332174e-05,
|
| 2874 |
+
"loss": 0.5769,
|
| 2875 |
+
"step": 406
|
| 2876 |
+
},
|
| 2877 |
+
{
|
| 2878 |
+
"epoch": 0.40955974842767295,
|
| 2879 |
+
"grad_norm": 0.2447304129600525,
|
| 2880 |
+
"learning_rate": 6.458539947175475e-05,
|
| 2881 |
+
"loss": 0.6887,
|
| 2882 |
+
"step": 407
|
| 2883 |
+
},
|
| 2884 |
+
{
|
| 2885 |
+
"epoch": 0.41056603773584904,
|
| 2886 |
+
"grad_norm": 0.2344241440296173,
|
| 2887 |
+
"learning_rate": 6.416006812042828e-05,
|
| 2888 |
+
"loss": 0.7118,
|
| 2889 |
+
"step": 408
|
| 2890 |
+
},
|
| 2891 |
+
{
|
| 2892 |
+
"epoch": 0.41157232704402513,
|
| 2893 |
+
"grad_norm": 0.2348852902650833,
|
| 2894 |
+
"learning_rate": 6.373263565171806e-05,
|
| 2895 |
+
"loss": 0.7149,
|
| 2896 |
+
"step": 409
|
| 2897 |
+
},
|
| 2898 |
+
{
|
| 2899 |
+
"epoch": 0.4125786163522013,
|
| 2900 |
+
"grad_norm": 0.24856218695640564,
|
| 2901 |
+
"learning_rate": 6.330314893841102e-05,
|
| 2902 |
+
"loss": 0.7624,
|
| 2903 |
+
"step": 410
|
| 2904 |
+
},
|
| 2905 |
+
{
|
| 2906 |
+
"epoch": 0.41358490566037737,
|
| 2907 |
+
"grad_norm": 0.2626706063747406,
|
| 2908 |
+
"learning_rate": 6.287165507856512e-05,
|
| 2909 |
+
"loss": 0.77,
|
| 2910 |
+
"step": 411
|
| 2911 |
+
},
|
| 2912 |
+
{
|
| 2913 |
+
"epoch": 0.41459119496855346,
|
| 2914 |
+
"grad_norm": 0.2993844449520111,
|
| 2915 |
+
"learning_rate": 6.243820139034464e-05,
|
| 2916 |
+
"loss": 0.737,
|
| 2917 |
+
"step": 412
|
| 2918 |
+
},
|
| 2919 |
+
{
|
| 2920 |
+
"epoch": 0.41559748427672955,
|
| 2921 |
+
"grad_norm": 0.245309978723526,
|
| 2922 |
+
"learning_rate": 6.200283540683103e-05,
|
| 2923 |
+
"loss": 0.6663,
|
| 2924 |
+
"step": 413
|
| 2925 |
+
},
|
| 2926 |
+
{
|
| 2927 |
+
"epoch": 0.41660377358490563,
|
| 2928 |
+
"grad_norm": 0.303198903799057,
|
| 2929 |
+
"learning_rate": 6.156560487081051e-05,
|
| 2930 |
+
"loss": 0.887,
|
| 2931 |
+
"step": 414
|
| 2932 |
+
},
|
| 2933 |
+
{
|
| 2934 |
+
"epoch": 0.4176100628930818,
|
| 2935 |
+
"grad_norm": 0.26473936438560486,
|
| 2936 |
+
"learning_rate": 6.112655772953851e-05,
|
| 2937 |
+
"loss": 0.75,
|
| 2938 |
+
"step": 415
|
| 2939 |
+
},
|
| 2940 |
+
{
|
| 2941 |
+
"epoch": 0.41861635220125787,
|
| 2942 |
+
"grad_norm": 0.2525455057621002,
|
| 2943 |
+
"learning_rate": 6.068574212948169e-05,
|
| 2944 |
+
"loss": 0.7307,
|
| 2945 |
+
"step": 416
|
| 2946 |
+
},
|
| 2947 |
+
{
|
| 2948 |
+
"epoch": 0.41962264150943396,
|
| 2949 |
+
"grad_norm": 0.2966090738773346,
|
| 2950 |
+
"learning_rate": 6.024320641103812e-05,
|
| 2951 |
+
"loss": 0.8675,
|
| 2952 |
+
"step": 417
|
| 2953 |
+
},
|
| 2954 |
+
{
|
| 2955 |
+
"epoch": 0.42062893081761005,
|
| 2956 |
+
"grad_norm": 0.29838553071022034,
|
| 2957 |
+
"learning_rate": 5.979899910323625e-05,
|
| 2958 |
+
"loss": 0.8116,
|
| 2959 |
+
"step": 418
|
| 2960 |
+
},
|
| 2961 |
+
{
|
| 2962 |
+
"epoch": 0.42163522012578614,
|
| 2963 |
+
"grad_norm": 0.2876933813095093,
|
| 2964 |
+
"learning_rate": 5.935316891841316e-05,
|
| 2965 |
+
"loss": 0.8379,
|
| 2966 |
+
"step": 419
|
| 2967 |
+
},
|
| 2968 |
+
{
|
| 2969 |
+
"epoch": 0.4226415094339623,
|
| 2970 |
+
"grad_norm": 0.31865549087524414,
|
| 2971 |
+
"learning_rate": 5.890576474687264e-05,
|
| 2972 |
+
"loss": 0.8156,
|
| 2973 |
+
"step": 420
|
| 2974 |
+
},
|
| 2975 |
+
{
|
| 2976 |
+
"epoch": 0.42364779874213837,
|
| 2977 |
+
"grad_norm": 0.3124519884586334,
|
| 2978 |
+
"learning_rate": 5.845683565152391e-05,
|
| 2979 |
+
"loss": 0.8597,
|
| 2980 |
+
"step": 421
|
| 2981 |
+
},
|
| 2982 |
+
{
|
| 2983 |
+
"epoch": 0.42465408805031446,
|
| 2984 |
+
"grad_norm": 0.30120527744293213,
|
| 2985 |
+
"learning_rate": 5.800643086250122e-05,
|
| 2986 |
+
"loss": 0.8902,
|
| 2987 |
+
"step": 422
|
| 2988 |
+
},
|
| 2989 |
+
{
|
| 2990 |
+
"epoch": 0.42566037735849055,
|
| 2991 |
+
"grad_norm": 0.2954394817352295,
|
| 2992 |
+
"learning_rate": 5.7554599771765325e-05,
|
| 2993 |
+
"loss": 0.8416,
|
| 2994 |
+
"step": 423
|
| 2995 |
+
},
|
| 2996 |
+
{
|
| 2997 |
+
"epoch": 0.4266666666666667,
|
| 2998 |
+
"grad_norm": 0.3026507496833801,
|
| 2999 |
+
"learning_rate": 5.710139192768696e-05,
|
| 3000 |
+
"loss": 0.872,
|
| 3001 |
+
"step": 424
|
| 3002 |
+
},
|
| 3003 |
+
{
|
| 3004 |
+
"epoch": 0.4276729559748428,
|
| 3005 |
+
"grad_norm": 0.32046762108802795,
|
| 3006 |
+
"learning_rate": 5.6646857029613434e-05,
|
| 3007 |
+
"loss": 0.7951,
|
| 3008 |
+
"step": 425
|
| 3009 |
+
},
|
| 3010 |
+
{
|
| 3011 |
+
"epoch": 0.4286792452830189,
|
| 3012 |
+
"grad_norm": 0.3316304087638855,
|
| 3013 |
+
"learning_rate": 5.6191044922418485e-05,
|
| 3014 |
+
"loss": 0.8383,
|
| 3015 |
+
"step": 426
|
| 3016 |
+
},
|
| 3017 |
+
{
|
| 3018 |
+
"epoch": 0.42968553459119496,
|
| 3019 |
+
"grad_norm": 0.33417731523513794,
|
| 3020 |
+
"learning_rate": 5.5734005591036144e-05,
|
| 3021 |
+
"loss": 0.8762,
|
| 3022 |
+
"step": 427
|
| 3023 |
+
},
|
| 3024 |
+
{
|
| 3025 |
+
"epoch": 0.43069182389937105,
|
| 3026 |
+
"grad_norm": 0.34470945596694946,
|
| 3027 |
+
"learning_rate": 5.527578915497952e-05,
|
| 3028 |
+
"loss": 0.9325,
|
| 3029 |
+
"step": 428
|
| 3030 |
+
},
|
| 3031 |
+
{
|
| 3032 |
+
"epoch": 0.4316981132075472,
|
| 3033 |
+
"grad_norm": 0.34728768467903137,
|
| 3034 |
+
"learning_rate": 5.4816445862844426e-05,
|
| 3035 |
+
"loss": 0.8731,
|
| 3036 |
+
"step": 429
|
| 3037 |
+
},
|
| 3038 |
+
{
|
| 3039 |
+
"epoch": 0.4327044025157233,
|
| 3040 |
+
"grad_norm": 0.36172977089881897,
|
| 3041 |
+
"learning_rate": 5.435602608679916e-05,
|
| 3042 |
+
"loss": 0.7652,
|
| 3043 |
+
"step": 430
|
| 3044 |
+
},
|
| 3045 |
+
{
|
| 3046 |
+
"epoch": 0.4337106918238994,
|
| 3047 |
+
"grad_norm": 0.3733101487159729,
|
| 3048 |
+
"learning_rate": 5.3894580317060684e-05,
|
| 3049 |
+
"loss": 0.7679,
|
| 3050 |
+
"step": 431
|
| 3051 |
+
},
|
| 3052 |
+
{
|
| 3053 |
+
"epoch": 0.43471698113207546,
|
| 3054 |
+
"grad_norm": 0.3756862282752991,
|
| 3055 |
+
"learning_rate": 5.343215915635762e-05,
|
| 3056 |
+
"loss": 0.8794,
|
| 3057 |
+
"step": 432
|
| 3058 |
+
},
|
| 3059 |
+
{
|
| 3060 |
+
"epoch": 0.43572327044025155,
|
| 3061 |
+
"grad_norm": 0.4108087122440338,
|
| 3062 |
+
"learning_rate": 5.2968813314381255e-05,
|
| 3063 |
+
"loss": 0.8592,
|
| 3064 |
+
"step": 433
|
| 3065 |
+
},
|
| 3066 |
+
{
|
| 3067 |
+
"epoch": 0.4367295597484277,
|
| 3068 |
+
"grad_norm": 0.41381314396858215,
|
| 3069 |
+
"learning_rate": 5.250459360222461e-05,
|
| 3070 |
+
"loss": 0.9277,
|
| 3071 |
+
"step": 434
|
| 3072 |
+
},
|
| 3073 |
+
{
|
| 3074 |
+
"epoch": 0.4377358490566038,
|
| 3075 |
+
"grad_norm": 0.39701637625694275,
|
| 3076 |
+
"learning_rate": 5.20395509268104e-05,
|
| 3077 |
+
"loss": 0.779,
|
| 3078 |
+
"step": 435
|
| 3079 |
+
},
|
| 3080 |
+
{
|
| 3081 |
+
"epoch": 0.4387421383647799,
|
| 3082 |
+
"grad_norm": 0.39671677350997925,
|
| 3083 |
+
"learning_rate": 5.157373628530853e-05,
|
| 3084 |
+
"loss": 0.7655,
|
| 3085 |
+
"step": 436
|
| 3086 |
+
},
|
| 3087 |
+
{
|
| 3088 |
+
"epoch": 0.43974842767295597,
|
| 3089 |
+
"grad_norm": 0.4550817906856537,
|
| 3090 |
+
"learning_rate": 5.1107200759543704e-05,
|
| 3091 |
+
"loss": 0.8008,
|
| 3092 |
+
"step": 437
|
| 3093 |
+
},
|
| 3094 |
+
{
|
| 3095 |
+
"epoch": 0.44075471698113206,
|
| 3096 |
+
"grad_norm": 0.4520381987094879,
|
| 3097 |
+
"learning_rate": 5.06399955103937e-05,
|
| 3098 |
+
"loss": 0.9125,
|
| 3099 |
+
"step": 438
|
| 3100 |
+
},
|
| 3101 |
+
{
|
| 3102 |
+
"epoch": 0.4417610062893082,
|
| 3103 |
+
"grad_norm": 0.49547475576400757,
|
| 3104 |
+
"learning_rate": 5.017217177217901e-05,
|
| 3105 |
+
"loss": 0.9307,
|
| 3106 |
+
"step": 439
|
| 3107 |
+
},
|
| 3108 |
+
{
|
| 3109 |
+
"epoch": 0.4427672955974843,
|
| 3110 |
+
"grad_norm": 0.5199328660964966,
|
| 3111 |
+
"learning_rate": 4.9703780847044415e-05,
|
| 3112 |
+
"loss": 0.8968,
|
| 3113 |
+
"step": 440
|
| 3114 |
+
},
|
| 3115 |
+
{
|
| 3116 |
+
"epoch": 0.4437735849056604,
|
| 3117 |
+
"grad_norm": 0.4823533594608307,
|
| 3118 |
+
"learning_rate": 4.923487409933316e-05,
|
| 3119 |
+
"loss": 0.836,
|
| 3120 |
+
"step": 441
|
| 3121 |
+
},
|
| 3122 |
+
{
|
| 3123 |
+
"epoch": 0.44477987421383647,
|
| 3124 |
+
"grad_norm": 0.573940634727478,
|
| 3125 |
+
"learning_rate": 4.876550294995421e-05,
|
| 3126 |
+
"loss": 0.9542,
|
| 3127 |
+
"step": 442
|
| 3128 |
+
},
|
| 3129 |
+
{
|
| 3130 |
+
"epoch": 0.44578616352201256,
|
| 3131 |
+
"grad_norm": 0.6316177248954773,
|
| 3132 |
+
"learning_rate": 4.829571887074343e-05,
|
| 3133 |
+
"loss": 1.0012,
|
| 3134 |
+
"step": 443
|
| 3135 |
+
},
|
| 3136 |
+
{
|
| 3137 |
+
"epoch": 0.4467924528301887,
|
| 3138 |
+
"grad_norm": 0.6314241886138916,
|
| 3139 |
+
"learning_rate": 4.782557337881911e-05,
|
| 3140 |
+
"loss": 0.9749,
|
| 3141 |
+
"step": 444
|
| 3142 |
+
},
|
| 3143 |
+
{
|
| 3144 |
+
"epoch": 0.4477987421383648,
|
| 3145 |
+
"grad_norm": 0.5996174216270447,
|
| 3146 |
+
"learning_rate": 4.7355118030932484e-05,
|
| 3147 |
+
"loss": 0.7982,
|
| 3148 |
+
"step": 445
|
| 3149 |
+
},
|
| 3150 |
+
{
|
| 3151 |
+
"epoch": 0.4488050314465409,
|
| 3152 |
+
"grad_norm": 0.6762143969535828,
|
| 3153 |
+
"learning_rate": 4.688440441781398e-05,
|
| 3154 |
+
"loss": 1.026,
|
| 3155 |
+
"step": 446
|
| 3156 |
+
},
|
| 3157 |
+
{
|
| 3158 |
+
"epoch": 0.44981132075471697,
|
| 3159 |
+
"grad_norm": 0.7565874457359314,
|
| 3160 |
+
"learning_rate": 4.6413484158515774e-05,
|
| 3161 |
+
"loss": 0.877,
|
| 3162 |
+
"step": 447
|
| 3163 |
+
},
|
| 3164 |
+
{
|
| 3165 |
+
"epoch": 0.45081761006289306,
|
| 3166 |
+
"grad_norm": 0.6773599982261658,
|
| 3167 |
+
"learning_rate": 4.594240889475107e-05,
|
| 3168 |
+
"loss": 0.8928,
|
| 3169 |
+
"step": 448
|
| 3170 |
+
},
|
| 3171 |
+
{
|
| 3172 |
+
"epoch": 0.4518238993710692,
|
| 3173 |
+
"grad_norm": 0.9175851345062256,
|
| 3174 |
+
"learning_rate": 4.547123028523106e-05,
|
| 3175 |
+
"loss": 0.9389,
|
| 3176 |
+
"step": 449
|
| 3177 |
+
},
|
| 3178 |
+
{
|
| 3179 |
+
"epoch": 0.4528301886792453,
|
| 3180 |
+
"grad_norm": 1.3272944688796997,
|
| 3181 |
+
"learning_rate": 4.5e-05,
|
| 3182 |
+
"loss": 0.9653,
|
| 3183 |
+
"step": 450
|
| 3184 |
+
},
|
| 3185 |
+
{
|
| 3186 |
+
"epoch": 0.4528301886792453,
|
| 3187 |
+
"eval_loss": 0.8040257692337036,
|
| 3188 |
+
"eval_runtime": 72.3709,
|
| 3189 |
+
"eval_samples_per_second": 46.248,
|
| 3190 |
+
"eval_steps_per_second": 11.565,
|
| 3191 |
+
"step": 450
|
| 3192 |
+
},
|
| 3193 |
+
{
|
| 3194 |
+
"epoch": 0.4538364779874214,
|
| 3195 |
+
"grad_norm": 0.20471110939979553,
|
| 3196 |
+
"learning_rate": 4.452876971476896e-05,
|
| 3197 |
+
"loss": 0.4702,
|
| 3198 |
+
"step": 451
|
| 3199 |
+
},
|
| 3200 |
+
{
|
| 3201 |
+
"epoch": 0.45484276729559747,
|
| 3202 |
+
"grad_norm": 0.25521793961524963,
|
| 3203 |
+
"learning_rate": 4.4057591105248945e-05,
|
| 3204 |
+
"loss": 0.6283,
|
| 3205 |
+
"step": 452
|
| 3206 |
+
},
|
| 3207 |
+
{
|
| 3208 |
+
"epoch": 0.45584905660377356,
|
| 3209 |
+
"grad_norm": 0.1867275834083557,
|
| 3210 |
+
"learning_rate": 4.358651584148423e-05,
|
| 3211 |
+
"loss": 0.5012,
|
| 3212 |
+
"step": 453
|
| 3213 |
+
},
|
| 3214 |
+
{
|
| 3215 |
+
"epoch": 0.4568553459119497,
|
| 3216 |
+
"grad_norm": 0.20557887852191925,
|
| 3217 |
+
"learning_rate": 4.311559558218603e-05,
|
| 3218 |
+
"loss": 0.5983,
|
| 3219 |
+
"step": 454
|
| 3220 |
+
},
|
| 3221 |
+
{
|
| 3222 |
+
"epoch": 0.4578616352201258,
|
| 3223 |
+
"grad_norm": 0.2151852697134018,
|
| 3224 |
+
"learning_rate": 4.264488196906753e-05,
|
| 3225 |
+
"loss": 0.6148,
|
| 3226 |
+
"step": 455
|
| 3227 |
+
},
|
| 3228 |
+
{
|
| 3229 |
+
"epoch": 0.4588679245283019,
|
| 3230 |
+
"grad_norm": 0.20975513756275177,
|
| 3231 |
+
"learning_rate": 4.21744266211809e-05,
|
| 3232 |
+
"loss": 0.5828,
|
| 3233 |
+
"step": 456
|
| 3234 |
+
},
|
| 3235 |
+
{
|
| 3236 |
+
"epoch": 0.459874213836478,
|
| 3237 |
+
"grad_norm": 0.2262539118528366,
|
| 3238 |
+
"learning_rate": 4.1704281129256585e-05,
|
| 3239 |
+
"loss": 0.6302,
|
| 3240 |
+
"step": 457
|
| 3241 |
+
},
|
| 3242 |
+
{
|
| 3243 |
+
"epoch": 0.46088050314465406,
|
| 3244 |
+
"grad_norm": 0.24988995492458344,
|
| 3245 |
+
"learning_rate": 4.1234497050045815e-05,
|
| 3246 |
+
"loss": 0.7265,
|
| 3247 |
+
"step": 458
|
| 3248 |
+
},
|
| 3249 |
+
{
|
| 3250 |
+
"epoch": 0.4618867924528302,
|
| 3251 |
+
"grad_norm": 0.259901762008667,
|
| 3252 |
+
"learning_rate": 4.076512590066686e-05,
|
| 3253 |
+
"loss": 0.7655,
|
| 3254 |
+
"step": 459
|
| 3255 |
+
},
|
| 3256 |
+
{
|
| 3257 |
+
"epoch": 0.4628930817610063,
|
| 3258 |
+
"grad_norm": 0.2539641559123993,
|
| 3259 |
+
"learning_rate": 4.0296219152955604e-05,
|
| 3260 |
+
"loss": 0.6835,
|
| 3261 |
+
"step": 460
|
| 3262 |
+
},
|
| 3263 |
+
{
|
| 3264 |
+
"epoch": 0.4638993710691824,
|
| 3265 |
+
"grad_norm": 0.2466258704662323,
|
| 3266 |
+
"learning_rate": 3.982782822782101e-05,
|
| 3267 |
+
"loss": 0.6605,
|
| 3268 |
+
"step": 461
|
| 3269 |
+
},
|
| 3270 |
+
{
|
| 3271 |
+
"epoch": 0.4649056603773585,
|
| 3272 |
+
"grad_norm": 0.2474479377269745,
|
| 3273 |
+
"learning_rate": 3.936000448960631e-05,
|
| 3274 |
+
"loss": 0.7396,
|
| 3275 |
+
"step": 462
|
| 3276 |
+
},
|
| 3277 |
+
{
|
| 3278 |
+
"epoch": 0.46591194968553457,
|
| 3279 |
+
"grad_norm": 0.2521885335445404,
|
| 3280 |
+
"learning_rate": 3.889279924045631e-05,
|
| 3281 |
+
"loss": 0.7581,
|
| 3282 |
+
"step": 463
|
| 3283 |
+
},
|
| 3284 |
+
{
|
| 3285 |
+
"epoch": 0.4669182389937107,
|
| 3286 |
+
"grad_norm": 0.2826155722141266,
|
| 3287 |
+
"learning_rate": 3.842626371469148e-05,
|
| 3288 |
+
"loss": 0.8262,
|
| 3289 |
+
"step": 464
|
| 3290 |
+
},
|
| 3291 |
+
{
|
| 3292 |
+
"epoch": 0.4679245283018868,
|
| 3293 |
+
"grad_norm": 0.2689575254917145,
|
| 3294 |
+
"learning_rate": 3.796044907318961e-05,
|
| 3295 |
+
"loss": 0.7999,
|
| 3296 |
+
"step": 465
|
| 3297 |
+
},
|
| 3298 |
+
{
|
| 3299 |
+
"epoch": 0.4689308176100629,
|
| 3300 |
+
"grad_norm": 0.26612672209739685,
|
| 3301 |
+
"learning_rate": 3.74954063977754e-05,
|
| 3302 |
+
"loss": 0.7519,
|
| 3303 |
+
"step": 466
|
| 3304 |
+
},
|
| 3305 |
+
{
|
| 3306 |
+
"epoch": 0.469937106918239,
|
| 3307 |
+
"grad_norm": 0.26734238862991333,
|
| 3308 |
+
"learning_rate": 3.703118668561876e-05,
|
| 3309 |
+
"loss": 0.7494,
|
| 3310 |
+
"step": 467
|
| 3311 |
+
},
|
| 3312 |
+
{
|
| 3313 |
+
"epoch": 0.47094339622641507,
|
| 3314 |
+
"grad_norm": 0.27751103043556213,
|
| 3315 |
+
"learning_rate": 3.6567840843642385e-05,
|
| 3316 |
+
"loss": 0.7699,
|
| 3317 |
+
"step": 468
|
| 3318 |
+
},
|
| 3319 |
+
{
|
| 3320 |
+
"epoch": 0.4719496855345912,
|
| 3321 |
+
"grad_norm": 0.2924068570137024,
|
| 3322 |
+
"learning_rate": 3.610541968293932e-05,
|
| 3323 |
+
"loss": 0.8906,
|
| 3324 |
+
"step": 469
|
| 3325 |
+
},
|
| 3326 |
+
{
|
| 3327 |
+
"epoch": 0.4729559748427673,
|
| 3328 |
+
"grad_norm": 0.3333476185798645,
|
| 3329 |
+
"learning_rate": 3.564397391320084e-05,
|
| 3330 |
+
"loss": 0.8296,
|
| 3331 |
+
"step": 470
|
| 3332 |
+
},
|
| 3333 |
+
{
|
| 3334 |
+
"epoch": 0.4739622641509434,
|
| 3335 |
+
"grad_norm": 0.2821838855743408,
|
| 3336 |
+
"learning_rate": 3.51835541371556e-05,
|
| 3337 |
+
"loss": 0.7995,
|
| 3338 |
+
"step": 471
|
| 3339 |
+
},
|
| 3340 |
+
{
|
| 3341 |
+
"epoch": 0.4749685534591195,
|
| 3342 |
+
"grad_norm": 0.29105469584465027,
|
| 3343 |
+
"learning_rate": 3.472421084502049e-05,
|
| 3344 |
+
"loss": 0.8302,
|
| 3345 |
+
"step": 472
|
| 3346 |
+
},
|
| 3347 |
+
{
|
| 3348 |
+
"epoch": 0.4759748427672956,
|
| 3349 |
+
"grad_norm": 0.3232196867465973,
|
| 3350 |
+
"learning_rate": 3.426599440896387e-05,
|
| 3351 |
+
"loss": 0.8682,
|
| 3352 |
+
"step": 473
|
| 3353 |
+
},
|
| 3354 |
+
{
|
| 3355 |
+
"epoch": 0.4769811320754717,
|
| 3356 |
+
"grad_norm": 0.3052535355091095,
|
| 3357 |
+
"learning_rate": 3.380895507758154e-05,
|
| 3358 |
+
"loss": 0.8917,
|
| 3359 |
+
"step": 474
|
| 3360 |
+
},
|
| 3361 |
+
{
|
| 3362 |
+
"epoch": 0.4779874213836478,
|
| 3363 |
+
"grad_norm": 0.31709784269332886,
|
| 3364 |
+
"learning_rate": 3.3353142970386565e-05,
|
| 3365 |
+
"loss": 0.9024,
|
| 3366 |
+
"step": 475
|
| 3367 |
+
},
|
| 3368 |
+
{
|
| 3369 |
+
"epoch": 0.4789937106918239,
|
| 3370 |
+
"grad_norm": 0.30279234051704407,
|
| 3371 |
+
"learning_rate": 3.2898608072313045e-05,
|
| 3372 |
+
"loss": 0.8268,
|
| 3373 |
+
"step": 476
|
| 3374 |
+
},
|
| 3375 |
+
{
|
| 3376 |
+
"epoch": 0.48,
|
| 3377 |
+
"grad_norm": 0.31982421875,
|
| 3378 |
+
"learning_rate": 3.244540022823469e-05,
|
| 3379 |
+
"loss": 0.8574,
|
| 3380 |
+
"step": 477
|
| 3381 |
+
},
|
| 3382 |
+
{
|
| 3383 |
+
"epoch": 0.4810062893081761,
|
| 3384 |
+
"grad_norm": 0.34274452924728394,
|
| 3385 |
+
"learning_rate": 3.199356913749877e-05,
|
| 3386 |
+
"loss": 0.8376,
|
| 3387 |
+
"step": 478
|
| 3388 |
+
},
|
| 3389 |
+
{
|
| 3390 |
+
"epoch": 0.4820125786163522,
|
| 3391 |
+
"grad_norm": 0.3448265790939331,
|
| 3392 |
+
"learning_rate": 3.1543164348476105e-05,
|
| 3393 |
+
"loss": 0.8395,
|
| 3394 |
+
"step": 479
|
| 3395 |
+
},
|
| 3396 |
+
{
|
| 3397 |
+
"epoch": 0.4830188679245283,
|
| 3398 |
+
"grad_norm": 0.3254213333129883,
|
| 3399 |
+
"learning_rate": 3.1094235253127374e-05,
|
| 3400 |
+
"loss": 0.7239,
|
| 3401 |
+
"step": 480
|
| 3402 |
+
},
|
| 3403 |
+
{
|
| 3404 |
+
"epoch": 0.4840251572327044,
|
| 3405 |
+
"grad_norm": 0.3562363386154175,
|
| 3406 |
+
"learning_rate": 3.064683108158685e-05,
|
| 3407 |
+
"loss": 0.8632,
|
| 3408 |
+
"step": 481
|
| 3409 |
+
},
|
| 3410 |
+
{
|
| 3411 |
+
"epoch": 0.4850314465408805,
|
| 3412 |
+
"grad_norm": 0.36696115136146545,
|
| 3413 |
+
"learning_rate": 3.0201000896763757e-05,
|
| 3414 |
+
"loss": 0.8333,
|
| 3415 |
+
"step": 482
|
| 3416 |
+
},
|
| 3417 |
+
{
|
| 3418 |
+
"epoch": 0.48603773584905663,
|
| 3419 |
+
"grad_norm": 0.39277511835098267,
|
| 3420 |
+
"learning_rate": 2.975679358896189e-05,
|
| 3421 |
+
"loss": 0.799,
|
| 3422 |
+
"step": 483
|
| 3423 |
+
},
|
| 3424 |
+
{
|
| 3425 |
+
"epoch": 0.4870440251572327,
|
| 3426 |
+
"grad_norm": 0.3960510790348053,
|
| 3427 |
+
"learning_rate": 2.9314257870518325e-05,
|
| 3428 |
+
"loss": 0.8517,
|
| 3429 |
+
"step": 484
|
| 3430 |
+
},
|
| 3431 |
+
{
|
| 3432 |
+
"epoch": 0.4880503144654088,
|
| 3433 |
+
"grad_norm": 0.40880805253982544,
|
| 3434 |
+
"learning_rate": 2.887344227046149e-05,
|
| 3435 |
+
"loss": 0.768,
|
| 3436 |
+
"step": 485
|
| 3437 |
+
},
|
| 3438 |
+
{
|
| 3439 |
+
"epoch": 0.4890566037735849,
|
| 3440 |
+
"grad_norm": 0.4367224872112274,
|
| 3441 |
+
"learning_rate": 2.8434395129189495e-05,
|
| 3442 |
+
"loss": 0.8494,
|
| 3443 |
+
"step": 486
|
| 3444 |
+
},
|
| 3445 |
+
{
|
| 3446 |
+
"epoch": 0.490062893081761,
|
| 3447 |
+
"grad_norm": 0.41261956095695496,
|
| 3448 |
+
"learning_rate": 2.7997164593168986e-05,
|
| 3449 |
+
"loss": 0.772,
|
| 3450 |
+
"step": 487
|
| 3451 |
+
},
|
| 3452 |
+
{
|
| 3453 |
+
"epoch": 0.49106918238993713,
|
| 3454 |
+
"grad_norm": 0.4327767491340637,
|
| 3455 |
+
"learning_rate": 2.756179860965537e-05,
|
| 3456 |
+
"loss": 0.8299,
|
| 3457 |
+
"step": 488
|
| 3458 |
+
},
|
| 3459 |
+
{
|
| 3460 |
+
"epoch": 0.4920754716981132,
|
| 3461 |
+
"grad_norm": 0.5124357342720032,
|
| 3462 |
+
"learning_rate": 2.7128344921434877e-05,
|
| 3463 |
+
"loss": 0.9382,
|
| 3464 |
+
"step": 489
|
| 3465 |
+
},
|
| 3466 |
+
{
|
| 3467 |
+
"epoch": 0.4930817610062893,
|
| 3468 |
+
"grad_norm": 0.47919076681137085,
|
| 3469 |
+
"learning_rate": 2.6696851061589e-05,
|
| 3470 |
+
"loss": 0.7916,
|
| 3471 |
+
"step": 490
|
| 3472 |
+
},
|
| 3473 |
+
{
|
| 3474 |
+
"epoch": 0.4940880503144654,
|
| 3475 |
+
"grad_norm": 0.5297518968582153,
|
| 3476 |
+
"learning_rate": 2.6267364348281954e-05,
|
| 3477 |
+
"loss": 0.8588,
|
| 3478 |
+
"step": 491
|
| 3479 |
+
},
|
| 3480 |
+
{
|
| 3481 |
+
"epoch": 0.4950943396226415,
|
| 3482 |
+
"grad_norm": 0.5296423435211182,
|
| 3483 |
+
"learning_rate": 2.5839931879571733e-05,
|
| 3484 |
+
"loss": 0.851,
|
| 3485 |
+
"step": 492
|
| 3486 |
+
},
|
| 3487 |
+
{
|
| 3488 |
+
"epoch": 0.49610062893081763,
|
| 3489 |
+
"grad_norm": 0.5914133191108704,
|
| 3490 |
+
"learning_rate": 2.541460052824527e-05,
|
| 3491 |
+
"loss": 0.9017,
|
| 3492 |
+
"step": 493
|
| 3493 |
+
},
|
| 3494 |
+
{
|
| 3495 |
+
"epoch": 0.4971069182389937,
|
| 3496 |
+
"grad_norm": 0.6070476174354553,
|
| 3497 |
+
"learning_rate": 2.4991416936678276e-05,
|
| 3498 |
+
"loss": 0.8965,
|
| 3499 |
+
"step": 494
|
| 3500 |
+
},
|
| 3501 |
+
{
|
| 3502 |
+
"epoch": 0.4981132075471698,
|
| 3503 |
+
"grad_norm": 0.6474441289901733,
|
| 3504 |
+
"learning_rate": 2.4570427511720398e-05,
|
| 3505 |
+
"loss": 0.9863,
|
| 3506 |
+
"step": 495
|
| 3507 |
+
},
|
| 3508 |
+
{
|
| 3509 |
+
"epoch": 0.4991194968553459,
|
| 3510 |
+
"grad_norm": 0.6955735683441162,
|
| 3511 |
+
"learning_rate": 2.4151678419606235e-05,
|
| 3512 |
+
"loss": 0.9854,
|
| 3513 |
+
"step": 496
|
| 3514 |
+
},
|
| 3515 |
+
{
|
| 3516 |
+
"epoch": 0.500125786163522,
|
| 3517 |
+
"grad_norm": 0.6878324747085571,
|
| 3518 |
+
"learning_rate": 2.3735215580892577e-05,
|
| 3519 |
+
"loss": 0.8784,
|
| 3520 |
+
"step": 497
|
| 3521 |
+
},
|
| 3522 |
+
{
|
| 3523 |
+
"epoch": 0.5011320754716981,
|
| 3524 |
+
"grad_norm": 0.7594687938690186,
|
| 3525 |
+
"learning_rate": 2.3321084665422807e-05,
|
| 3526 |
+
"loss": 1.0182,
|
| 3527 |
+
"step": 498
|
| 3528 |
+
},
|
| 3529 |
+
{
|
| 3530 |
+
"epoch": 0.5021383647798742,
|
| 3531 |
+
"grad_norm": 0.902470588684082,
|
| 3532 |
+
"learning_rate": 2.2909331087318664e-05,
|
| 3533 |
+
"loss": 0.8469,
|
| 3534 |
+
"step": 499
|
| 3535 |
+
},
|
| 3536 |
+
{
|
| 3537 |
+
"epoch": 0.5031446540880503,
|
| 3538 |
+
"grad_norm": 1.4358625411987305,
|
| 3539 |
+
"learning_rate": 2.250000000000001e-05,
|
| 3540 |
+
"loss": 0.8747,
|
| 3541 |
+
"step": 500
|
| 3542 |
+
},
|
| 3543 |
+
{
|
| 3544 |
+
"epoch": 0.5041509433962265,
|
| 3545 |
+
"grad_norm": 0.23904839158058167,
|
| 3546 |
+
"learning_rate": 2.209313629123329e-05,
|
| 3547 |
+
"loss": 0.5938,
|
| 3548 |
+
"step": 501
|
| 3549 |
+
},
|
| 3550 |
+
{
|
| 3551 |
+
"epoch": 0.5051572327044025,
|
| 3552 |
+
"grad_norm": 0.20867227017879486,
|
| 3553 |
+
"learning_rate": 2.168878457820915e-05,
|
| 3554 |
+
"loss": 0.5036,
|
| 3555 |
+
"step": 502
|
| 3556 |
+
},
|
| 3557 |
+
{
|
| 3558 |
+
"epoch": 0.5061635220125786,
|
| 3559 |
+
"grad_norm": 0.2144133299589157,
|
| 3560 |
+
"learning_rate": 2.128698920264951e-05,
|
| 3561 |
+
"loss": 0.6249,
|
| 3562 |
+
"step": 503
|
| 3563 |
+
},
|
| 3564 |
+
{
|
| 3565 |
+
"epoch": 0.5071698113207547,
|
| 3566 |
+
"grad_norm": 0.20212629437446594,
|
| 3567 |
+
"learning_rate": 2.088779422594514e-05,
|
| 3568 |
+
"loss": 0.6889,
|
| 3569 |
+
"step": 504
|
| 3570 |
+
},
|
| 3571 |
+
{
|
| 3572 |
+
"epoch": 0.5081761006289308,
|
| 3573 |
+
"grad_norm": 0.18371202051639557,
|
| 3574 |
+
"learning_rate": 2.0491243424323783e-05,
|
| 3575 |
+
"loss": 0.6188,
|
| 3576 |
+
"step": 505
|
| 3577 |
+
},
|
| 3578 |
+
{
|
| 3579 |
+
"epoch": 0.509182389937107,
|
| 3580 |
+
"grad_norm": 0.2056654989719391,
|
| 3581 |
+
"learning_rate": 2.009738028404952e-05,
|
| 3582 |
+
"loss": 0.6117,
|
| 3583 |
+
"step": 506
|
| 3584 |
+
},
|
| 3585 |
+
{
|
| 3586 |
+
"epoch": 0.510188679245283,
|
| 3587 |
+
"grad_norm": 0.21647100150585175,
|
| 3588 |
+
"learning_rate": 1.9706247996654134e-05,
|
| 3589 |
+
"loss": 0.6288,
|
| 3590 |
+
"step": 507
|
| 3591 |
+
},
|
| 3592 |
+
{
|
| 3593 |
+
"epoch": 0.5111949685534591,
|
| 3594 |
+
"grad_norm": 0.2186782956123352,
|
| 3595 |
+
"learning_rate": 1.9317889454200578e-05,
|
| 3596 |
+
"loss": 0.6399,
|
| 3597 |
+
"step": 508
|
| 3598 |
+
},
|
| 3599 |
+
{
|
| 3600 |
+
"epoch": 0.5122012578616352,
|
| 3601 |
+
"grad_norm": 0.23389363288879395,
|
| 3602 |
+
"learning_rate": 1.8932347244579463e-05,
|
| 3603 |
+
"loss": 0.7042,
|
| 3604 |
+
"step": 509
|
| 3605 |
+
},
|
| 3606 |
+
{
|
| 3607 |
+
"epoch": 0.5132075471698113,
|
| 3608 |
+
"grad_norm": 0.24194186925888062,
|
| 3609 |
+
"learning_rate": 1.8549663646838714e-05,
|
| 3610 |
+
"loss": 0.7191,
|
| 3611 |
+
"step": 510
|
| 3612 |
+
},
|
| 3613 |
+
{
|
| 3614 |
+
"epoch": 0.5142138364779875,
|
| 3615 |
+
"grad_norm": 0.24342100322246552,
|
| 3616 |
+
"learning_rate": 1.8169880626547285e-05,
|
| 3617 |
+
"loss": 0.7147,
|
| 3618 |
+
"step": 511
|
| 3619 |
+
},
|
| 3620 |
+
{
|
| 3621 |
+
"epoch": 0.5152201257861635,
|
| 3622 |
+
"grad_norm": 0.2450750470161438,
|
| 3623 |
+
"learning_rate": 1.7793039831193134e-05,
|
| 3624 |
+
"loss": 0.7187,
|
| 3625 |
+
"step": 512
|
| 3626 |
+
},
|
| 3627 |
+
{
|
| 3628 |
+
"epoch": 0.5162264150943396,
|
| 3629 |
+
"grad_norm": 0.26462146639823914,
|
| 3630 |
+
"learning_rate": 1.741918258561607e-05,
|
| 3631 |
+
"loss": 0.8313,
|
| 3632 |
+
"step": 513
|
| 3633 |
+
},
|
| 3634 |
+
{
|
| 3635 |
+
"epoch": 0.5172327044025157,
|
| 3636 |
+
"grad_norm": 0.26424697041511536,
|
| 3637 |
+
"learning_rate": 1.7048349887476038e-05,
|
| 3638 |
+
"loss": 0.7624,
|
| 3639 |
+
"step": 514
|
| 3640 |
+
},
|
| 3641 |
+
{
|
| 3642 |
+
"epoch": 0.5182389937106918,
|
| 3643 |
+
"grad_norm": 0.2656416594982147,
|
| 3644 |
+
"learning_rate": 1.6680582402757324e-05,
|
| 3645 |
+
"loss": 0.7959,
|
| 3646 |
+
"step": 515
|
| 3647 |
+
},
|
| 3648 |
+
{
|
| 3649 |
+
"epoch": 0.519245283018868,
|
| 3650 |
+
"grad_norm": 0.27823954820632935,
|
| 3651 |
+
"learning_rate": 1.631592046130896e-05,
|
| 3652 |
+
"loss": 0.7825,
|
| 3653 |
+
"step": 516
|
| 3654 |
+
},
|
| 3655 |
+
{
|
| 3656 |
+
"epoch": 0.520251572327044,
|
| 3657 |
+
"grad_norm": 0.26520705223083496,
|
| 3658 |
+
"learning_rate": 1.5954404052422217e-05,
|
| 3659 |
+
"loss": 0.7833,
|
| 3660 |
+
"step": 517
|
| 3661 |
+
},
|
| 3662 |
+
{
|
| 3663 |
+
"epoch": 0.5212578616352201,
|
| 3664 |
+
"grad_norm": 0.3015387952327728,
|
| 3665 |
+
"learning_rate": 1.5596072820445255e-05,
|
| 3666 |
+
"loss": 0.8261,
|
| 3667 |
+
"step": 518
|
| 3668 |
+
},
|
| 3669 |
+
{
|
| 3670 |
+
"epoch": 0.5222641509433962,
|
| 3671 |
+
"grad_norm": 0.28979843854904175,
|
| 3672 |
+
"learning_rate": 1.5240966060435674e-05,
|
| 3673 |
+
"loss": 0.785,
|
| 3674 |
+
"step": 519
|
| 3675 |
+
},
|
| 3676 |
+
{
|
| 3677 |
+
"epoch": 0.5232704402515723,
|
| 3678 |
+
"grad_norm": 0.2958517372608185,
|
| 3679 |
+
"learning_rate": 1.4889122713851395e-05,
|
| 3680 |
+
"loss": 0.7631,
|
| 3681 |
+
"step": 520
|
| 3682 |
+
},
|
| 3683 |
+
{
|
| 3684 |
+
"epoch": 0.5242767295597485,
|
| 3685 |
+
"grad_norm": 0.2933093011379242,
|
| 3686 |
+
"learning_rate": 1.4540581364280274e-05,
|
| 3687 |
+
"loss": 0.8831,
|
| 3688 |
+
"step": 521
|
| 3689 |
+
},
|
| 3690 |
+
{
|
| 3691 |
+
"epoch": 0.5252830188679245,
|
| 3692 |
+
"grad_norm": 0.27697688341140747,
|
| 3693 |
+
"learning_rate": 1.4195380233209009e-05,
|
| 3694 |
+
"loss": 0.7368,
|
| 3695 |
+
"step": 522
|
| 3696 |
+
},
|
| 3697 |
+
{
|
| 3698 |
+
"epoch": 0.5262893081761006,
|
| 3699 |
+
"grad_norm": 0.3134765326976776,
|
| 3700 |
+
"learning_rate": 1.38535571758317e-05,
|
| 3701 |
+
"loss": 0.8383,
|
| 3702 |
+
"step": 523
|
| 3703 |
+
},
|
| 3704 |
+
{
|
| 3705 |
+
"epoch": 0.5272955974842767,
|
| 3706 |
+
"grad_norm": 0.28508156538009644,
|
| 3707 |
+
"learning_rate": 1.3515149676898551e-05,
|
| 3708 |
+
"loss": 0.778,
|
| 3709 |
+
"step": 524
|
| 3710 |
+
},
|
| 3711 |
+
{
|
| 3712 |
+
"epoch": 0.5283018867924528,
|
| 3713 |
+
"grad_norm": 0.31699931621551514,
|
| 3714 |
+
"learning_rate": 1.3180194846605365e-05,
|
| 3715 |
+
"loss": 0.8089,
|
| 3716 |
+
"step": 525
|
| 3717 |
+
},
|
| 3718 |
+
{
|
| 3719 |
+
"epoch": 0.529308176100629,
|
| 3720 |
+
"grad_norm": 0.3305254876613617,
|
| 3721 |
+
"learning_rate": 1.284872941652386e-05,
|
| 3722 |
+
"loss": 0.831,
|
| 3723 |
+
"step": 526
|
| 3724 |
+
},
|
| 3725 |
+
{
|
| 3726 |
+
"epoch": 0.530314465408805,
|
| 3727 |
+
"grad_norm": 0.31761541962623596,
|
| 3728 |
+
"learning_rate": 1.2520789735573703e-05,
|
| 3729 |
+
"loss": 0.8284,
|
| 3730 |
+
"step": 527
|
| 3731 |
+
},
|
| 3732 |
+
{
|
| 3733 |
+
"epoch": 0.5313207547169811,
|
| 3734 |
+
"grad_norm": 0.31628575921058655,
|
| 3735 |
+
"learning_rate": 1.2196411766036491e-05,
|
| 3736 |
+
"loss": 0.7306,
|
| 3737 |
+
"step": 528
|
| 3738 |
+
},
|
| 3739 |
+
{
|
| 3740 |
+
"epoch": 0.5323270440251572,
|
| 3741 |
+
"grad_norm": 0.352816104888916,
|
| 3742 |
+
"learning_rate": 1.1875631079611956e-05,
|
| 3743 |
+
"loss": 0.8164,
|
| 3744 |
+
"step": 529
|
| 3745 |
+
},
|
| 3746 |
+
{
|
| 3747 |
+
"epoch": 0.5333333333333333,
|
| 3748 |
+
"grad_norm": 0.35125187039375305,
|
| 3749 |
+
"learning_rate": 1.1558482853517254e-05,
|
| 3750 |
+
"loss": 0.8364,
|
| 3751 |
+
"step": 530
|
| 3752 |
+
},
|
| 3753 |
+
{
|
| 3754 |
+
"epoch": 0.5343396226415095,
|
| 3755 |
+
"grad_norm": 0.3974522352218628,
|
| 3756 |
+
"learning_rate": 1.124500186662932e-05,
|
| 3757 |
+
"loss": 0.8929,
|
| 3758 |
+
"step": 531
|
| 3759 |
+
},
|
| 3760 |
+
{
|
| 3761 |
+
"epoch": 0.5353459119496855,
|
| 3762 |
+
"grad_norm": 0.3924785256385803,
|
| 3763 |
+
"learning_rate": 1.0935222495670969e-05,
|
| 3764 |
+
"loss": 0.8117,
|
| 3765 |
+
"step": 532
|
| 3766 |
+
},
|
| 3767 |
+
{
|
| 3768 |
+
"epoch": 0.5363522012578616,
|
| 3769 |
+
"grad_norm": 0.37499430775642395,
|
| 3770 |
+
"learning_rate": 1.0629178711441115e-05,
|
| 3771 |
+
"loss": 0.8571,
|
| 3772 |
+
"step": 533
|
| 3773 |
+
},
|
| 3774 |
+
{
|
| 3775 |
+
"epoch": 0.5373584905660377,
|
| 3776 |
+
"grad_norm": 0.4152454137802124,
|
| 3777 |
+
"learning_rate": 1.032690407508949e-05,
|
| 3778 |
+
"loss": 0.8176,
|
| 3779 |
+
"step": 534
|
| 3780 |
+
},
|
| 3781 |
+
{
|
| 3782 |
+
"epoch": 0.5383647798742138,
|
| 3783 |
+
"grad_norm": 0.4613153040409088,
|
| 3784 |
+
"learning_rate": 1.002843173443631e-05,
|
| 3785 |
+
"loss": 0.8914,
|
| 3786 |
+
"step": 535
|
| 3787 |
+
},
|
| 3788 |
+
{
|
| 3789 |
+
"epoch": 0.53937106918239,
|
| 3790 |
+
"grad_norm": 0.4119192659854889,
|
| 3791 |
+
"learning_rate": 9.733794420337214e-06,
|
| 3792 |
+
"loss": 0.7557,
|
| 3793 |
+
"step": 536
|
| 3794 |
+
},
|
| 3795 |
+
{
|
| 3796 |
+
"epoch": 0.540377358490566,
|
| 3797 |
+
"grad_norm": 0.47877928614616394,
|
| 3798 |
+
"learning_rate": 9.443024443093932e-06,
|
| 3799 |
+
"loss": 0.7729,
|
| 3800 |
+
"step": 537
|
| 3801 |
+
},
|
| 3802 |
+
{
|
| 3803 |
+
"epoch": 0.5413836477987422,
|
| 3804 |
+
"grad_norm": 0.5126065611839294,
|
| 3805 |
+
"learning_rate": 9.15615368891117e-06,
|
| 3806 |
+
"loss": 0.8311,
|
| 3807 |
+
"step": 538
|
| 3808 |
+
},
|
| 3809 |
+
{
|
| 3810 |
+
"epoch": 0.5423899371069182,
|
| 3811 |
+
"grad_norm": 0.48336324095726013,
|
| 3812 |
+
"learning_rate": 8.873213616399854e-06,
|
| 3813 |
+
"loss": 0.8465,
|
| 3814 |
+
"step": 539
|
| 3815 |
+
},
|
| 3816 |
+
{
|
| 3817 |
+
"epoch": 0.5433962264150943,
|
| 3818 |
+
"grad_norm": 0.5260478258132935,
|
| 3819 |
+
"learning_rate": 8.59423525312737e-06,
|
| 3820 |
+
"loss": 0.8989,
|
| 3821 |
+
"step": 540
|
| 3822 |
+
},
|
| 3823 |
+
{
|
| 3824 |
+
"epoch": 0.5444025157232705,
|
| 3825 |
+
"grad_norm": 0.6092353463172913,
|
| 3826 |
+
"learning_rate": 8.319249192215056e-06,
|
| 3827 |
+
"loss": 0.92,
|
| 3828 |
+
"step": 541
|
| 3829 |
+
},
|
| 3830 |
+
{
|
| 3831 |
+
"epoch": 0.5454088050314465,
|
| 3832 |
+
"grad_norm": 0.5647766590118408,
|
| 3833 |
+
"learning_rate": 8.04828558898332e-06,
|
| 3834 |
+
"loss": 0.9122,
|
| 3835 |
+
"step": 542
|
| 3836 |
+
},
|
| 3837 |
+
{
|
| 3838 |
+
"epoch": 0.5464150943396227,
|
| 3839 |
+
"grad_norm": 0.6063531041145325,
|
| 3840 |
+
"learning_rate": 7.781374157644714e-06,
|
| 3841 |
+
"loss": 1.0103,
|
| 3842 |
+
"step": 543
|
| 3843 |
+
},
|
| 3844 |
+
{
|
| 3845 |
+
"epoch": 0.5474213836477987,
|
| 3846 |
+
"grad_norm": 1.2413952350616455,
|
| 3847 |
+
"learning_rate": 7.518544168045526e-06,
|
| 3848 |
+
"loss": 0.7649,
|
| 3849 |
+
"step": 544
|
| 3850 |
+
},
|
| 3851 |
+
{
|
| 3852 |
+
"epoch": 0.5484276729559748,
|
| 3853 |
+
"grad_norm": 0.634516716003418,
|
| 3854 |
+
"learning_rate": 7.259824442455923e-06,
|
| 3855 |
+
"loss": 1.0046,
|
| 3856 |
+
"step": 545
|
| 3857 |
+
},
|
| 3858 |
+
{
|
| 3859 |
+
"epoch": 0.549433962264151,
|
| 3860 |
+
"grad_norm": 0.7817690968513489,
|
| 3861 |
+
"learning_rate": 7.005243352409332e-06,
|
| 3862 |
+
"loss": 1.0795,
|
| 3863 |
+
"step": 546
|
| 3864 |
+
},
|
| 3865 |
+
{
|
| 3866 |
+
"epoch": 0.550440251572327,
|
| 3867 |
+
"grad_norm": 0.7421931624412537,
|
| 3868 |
+
"learning_rate": 6.754828815591131e-06,
|
| 3869 |
+
"loss": 0.9726,
|
| 3870 |
+
"step": 547
|
| 3871 |
+
},
|
| 3872 |
+
{
|
| 3873 |
+
"epoch": 0.5514465408805032,
|
| 3874 |
+
"grad_norm": 0.8387755155563354,
|
| 3875 |
+
"learning_rate": 6.508608292777203e-06,
|
| 3876 |
+
"loss": 0.9694,
|
| 3877 |
+
"step": 548
|
| 3878 |
+
},
|
| 3879 |
+
{
|
| 3880 |
+
"epoch": 0.5524528301886793,
|
| 3881 |
+
"grad_norm": 0.8762068748474121,
|
| 3882 |
+
"learning_rate": 6.266608784822542e-06,
|
| 3883 |
+
"loss": 0.9576,
|
| 3884 |
+
"step": 549
|
| 3885 |
+
},
|
| 3886 |
+
{
|
| 3887 |
+
"epoch": 0.5534591194968553,
|
| 3888 |
+
"grad_norm": 1.0989902019500732,
|
| 3889 |
+
"learning_rate": 6.028856829700258e-06,
|
| 3890 |
+
"loss": 0.9774,
|
| 3891 |
+
"step": 550
|
| 3892 |
+
},
|
| 3893 |
+
{
|
| 3894 |
+
"epoch": 0.5544654088050315,
|
| 3895 |
+
"grad_norm": 0.19267253577709198,
|
| 3896 |
+
"learning_rate": 5.795378499591479e-06,
|
| 3897 |
+
"loss": 0.5068,
|
| 3898 |
+
"step": 551
|
| 3899 |
+
},
|
| 3900 |
+
{
|
| 3901 |
+
"epoch": 0.5554716981132075,
|
| 3902 |
+
"grad_norm": 0.20570369064807892,
|
| 3903 |
+
"learning_rate": 5.566199398026149e-06,
|
| 3904 |
+
"loss": 0.4821,
|
| 3905 |
+
"step": 552
|
| 3906 |
+
},
|
| 3907 |
+
{
|
| 3908 |
+
"epoch": 0.5564779874213837,
|
| 3909 |
+
"grad_norm": 0.2014710158109665,
|
| 3910 |
+
"learning_rate": 5.341344657075353e-06,
|
| 3911 |
+
"loss": 0.5268,
|
| 3912 |
+
"step": 553
|
| 3913 |
+
},
|
| 3914 |
+
{
|
| 3915 |
+
"epoch": 0.5574842767295598,
|
| 3916 |
+
"grad_norm": 0.17889411747455597,
|
| 3917 |
+
"learning_rate": 5.120838934595337e-06,
|
| 3918 |
+
"loss": 0.5612,
|
| 3919 |
+
"step": 554
|
| 3920 |
+
},
|
| 3921 |
+
{
|
| 3922 |
+
"epoch": 0.5584905660377358,
|
| 3923 |
+
"grad_norm": 0.1814723163843155,
|
| 3924 |
+
"learning_rate": 4.90470641152345e-06,
|
| 3925 |
+
"loss": 0.6024,
|
| 3926 |
+
"step": 555
|
| 3927 |
+
},
|
| 3928 |
+
{
|
| 3929 |
+
"epoch": 0.559496855345912,
|
| 3930 |
+
"grad_norm": 0.19090139865875244,
|
| 3931 |
+
"learning_rate": 4.69297078922642e-06,
|
| 3932 |
+
"loss": 0.6081,
|
| 3933 |
+
"step": 556
|
| 3934 |
+
},
|
| 3935 |
+
{
|
| 3936 |
+
"epoch": 0.560503144654088,
|
| 3937 |
+
"grad_norm": 0.21327486634254456,
|
| 3938 |
+
"learning_rate": 4.485655286901292e-06,
|
| 3939 |
+
"loss": 0.6522,
|
| 3940 |
+
"step": 557
|
| 3941 |
+
},
|
| 3942 |
+
{
|
| 3943 |
+
"epoch": 0.5615094339622642,
|
| 3944 |
+
"grad_norm": 0.20898205041885376,
|
| 3945 |
+
"learning_rate": 4.28278263902913e-06,
|
| 3946 |
+
"loss": 0.6344,
|
| 3947 |
+
"step": 558
|
| 3948 |
+
},
|
| 3949 |
+
{
|
| 3950 |
+
"epoch": 0.5625157232704403,
|
| 3951 |
+
"grad_norm": 0.22488388419151306,
|
| 3952 |
+
"learning_rate": 4.084375092881916e-06,
|
| 3953 |
+
"loss": 0.6422,
|
| 3954 |
+
"step": 559
|
| 3955 |
+
},
|
| 3956 |
+
{
|
| 3957 |
+
"epoch": 0.5635220125786163,
|
| 3958 |
+
"grad_norm": 0.2517217695713043,
|
| 3959 |
+
"learning_rate": 3.890454406082956e-06,
|
| 3960 |
+
"loss": 0.7872,
|
| 3961 |
+
"step": 560
|
| 3962 |
+
},
|
| 3963 |
+
{
|
| 3964 |
+
"epoch": 0.5645283018867925,
|
| 3965 |
+
"grad_norm": 0.23601284623146057,
|
| 3966 |
+
"learning_rate": 3.701041844220849e-06,
|
| 3967 |
+
"loss": 0.677,
|
| 3968 |
+
"step": 561
|
| 3969 |
+
},
|
| 3970 |
+
{
|
| 3971 |
+
"epoch": 0.5655345911949685,
|
| 3972 |
+
"grad_norm": 0.2342206984758377,
|
| 3973 |
+
"learning_rate": 3.516158178517482e-06,
|
| 3974 |
+
"loss": 0.6674,
|
| 3975 |
+
"step": 562
|
| 3976 |
+
},
|
| 3977 |
+
{
|
| 3978 |
+
"epoch": 0.5665408805031447,
|
| 3979 |
+
"grad_norm": 0.2706093192100525,
|
| 3980 |
+
"learning_rate": 3.335823683550237e-06,
|
| 3981 |
+
"loss": 0.7669,
|
| 3982 |
+
"step": 563
|
| 3983 |
+
},
|
| 3984 |
+
{
|
| 3985 |
+
"epoch": 0.5675471698113208,
|
| 3986 |
+
"grad_norm": 0.25333014130592346,
|
| 3987 |
+
"learning_rate": 3.1600581350286897e-06,
|
| 3988 |
+
"loss": 0.7338,
|
| 3989 |
+
"step": 564
|
| 3990 |
+
},
|
| 3991 |
+
{
|
| 3992 |
+
"epoch": 0.5685534591194968,
|
| 3993 |
+
"grad_norm": 0.2653476595878601,
|
| 3994 |
+
"learning_rate": 2.9888808076259267e-06,
|
| 3995 |
+
"loss": 0.782,
|
| 3996 |
+
"step": 565
|
| 3997 |
+
},
|
| 3998 |
+
{
|
| 3999 |
+
"epoch": 0.569559748427673,
|
| 4000 |
+
"grad_norm": 0.27972614765167236,
|
| 4001 |
+
"learning_rate": 2.822310472864885e-06,
|
| 4002 |
+
"loss": 0.7923,
|
| 4003 |
+
"step": 566
|
| 4004 |
+
},
|
| 4005 |
+
{
|
| 4006 |
+
"epoch": 0.570566037735849,
|
| 4007 |
+
"grad_norm": 0.26939818263053894,
|
| 4008 |
+
"learning_rate": 2.660365397059855e-06,
|
| 4009 |
+
"loss": 0.7194,
|
| 4010 |
+
"step": 567
|
| 4011 |
+
},
|
| 4012 |
+
{
|
| 4013 |
+
"epoch": 0.5715723270440252,
|
| 4014 |
+
"grad_norm": 0.28501391410827637,
|
| 4015 |
+
"learning_rate": 2.503063339313355e-06,
|
| 4016 |
+
"loss": 0.824,
|
| 4017 |
+
"step": 568
|
| 4018 |
+
},
|
| 4019 |
+
{
|
| 4020 |
+
"epoch": 0.5725786163522013,
|
| 4021 |
+
"grad_norm": 0.2938201427459717,
|
| 4022 |
+
"learning_rate": 2.3504215495686498e-06,
|
| 4023 |
+
"loss": 0.8437,
|
| 4024 |
+
"step": 569
|
| 4025 |
+
},
|
| 4026 |
+
{
|
| 4027 |
+
"epoch": 0.5735849056603773,
|
| 4028 |
+
"grad_norm": 0.28671878576278687,
|
| 4029 |
+
"learning_rate": 2.2024567667180914e-06,
|
| 4030 |
+
"loss": 0.8355,
|
| 4031 |
+
"step": 570
|
| 4032 |
+
},
|
| 4033 |
+
{
|
| 4034 |
+
"epoch": 0.5745911949685535,
|
| 4035 |
+
"grad_norm": 0.29585692286491394,
|
| 4036 |
+
"learning_rate": 2.059185216767543e-06,
|
| 4037 |
+
"loss": 0.8325,
|
| 4038 |
+
"step": 571
|
| 4039 |
+
},
|
| 4040 |
+
{
|
| 4041 |
+
"epoch": 0.5755974842767295,
|
| 4042 |
+
"grad_norm": 0.3155832886695862,
|
| 4043 |
+
"learning_rate": 1.9206226110569742e-06,
|
| 4044 |
+
"loss": 0.7434,
|
| 4045 |
+
"step": 572
|
| 4046 |
+
},
|
| 4047 |
+
{
|
| 4048 |
+
"epoch": 0.5766037735849057,
|
| 4049 |
+
"grad_norm": 0.29947665333747864,
|
| 4050 |
+
"learning_rate": 1.7867841445375621e-06,
|
| 4051 |
+
"loss": 0.8107,
|
| 4052 |
+
"step": 573
|
| 4053 |
+
},
|
| 4054 |
+
{
|
| 4055 |
+
"epoch": 0.5776100628930818,
|
| 4056 |
+
"grad_norm": 0.3051373362541199,
|
| 4057 |
+
"learning_rate": 1.6576844941053854e-06,
|
| 4058 |
+
"loss": 0.789,
|
| 4059 |
+
"step": 574
|
| 4060 |
+
},
|
| 4061 |
+
{
|
| 4062 |
+
"epoch": 0.5786163522012578,
|
| 4063 |
+
"grad_norm": 0.31009525060653687,
|
| 4064 |
+
"learning_rate": 1.533337816991931e-06,
|
| 4065 |
+
"loss": 0.8255,
|
| 4066 |
+
"step": 575
|
| 4067 |
+
},
|
| 4068 |
+
{
|
| 4069 |
+
"epoch": 0.579622641509434,
|
| 4070 |
+
"grad_norm": 0.3172381818294525,
|
| 4071 |
+
"learning_rate": 1.4137577492116016e-06,
|
| 4072 |
+
"loss": 0.8534,
|
| 4073 |
+
"step": 576
|
| 4074 |
+
},
|
| 4075 |
+
{
|
| 4076 |
+
"epoch": 0.58062893081761,
|
| 4077 |
+
"grad_norm": 0.3247847557067871,
|
| 4078 |
+
"learning_rate": 1.2989574040663816e-06,
|
| 4079 |
+
"loss": 0.7678,
|
| 4080 |
+
"step": 577
|
| 4081 |
+
},
|
| 4082 |
+
{
|
| 4083 |
+
"epoch": 0.5816352201257862,
|
| 4084 |
+
"grad_norm": 0.3526797592639923,
|
| 4085 |
+
"learning_rate": 1.188949370707787e-06,
|
| 4086 |
+
"loss": 0.8184,
|
| 4087 |
+
"step": 578
|
| 4088 |
+
},
|
| 4089 |
+
{
|
| 4090 |
+
"epoch": 0.5826415094339623,
|
| 4091 |
+
"grad_norm": 0.3355388641357422,
|
| 4092 |
+
"learning_rate": 1.0837457127563656e-06,
|
| 4093 |
+
"loss": 0.7369,
|
| 4094 |
+
"step": 579
|
| 4095 |
+
},
|
| 4096 |
+
{
|
| 4097 |
+
"epoch": 0.5836477987421383,
|
| 4098 |
+
"grad_norm": 0.35080787539482117,
|
| 4099 |
+
"learning_rate": 9.83357966978744e-07,
|
| 4100 |
+
"loss": 0.7977,
|
| 4101 |
+
"step": 580
|
| 4102 |
+
},
|
| 4103 |
+
{
|
| 4104 |
+
"epoch": 0.5846540880503145,
|
| 4105 |
+
"grad_norm": 0.3781813681125641,
|
| 4106 |
+
"learning_rate": 8.877971420225212e-07,
|
| 4107 |
+
"loss": 0.8562,
|
| 4108 |
+
"step": 581
|
| 4109 |
+
},
|
| 4110 |
+
{
|
| 4111 |
+
"epoch": 0.5856603773584905,
|
| 4112 |
+
"grad_norm": 0.3460877537727356,
|
| 4113 |
+
"learning_rate": 7.970737172090126e-07,
|
| 4114 |
+
"loss": 0.7582,
|
| 4115 |
+
"step": 582
|
| 4116 |
+
},
|
| 4117 |
+
{
|
| 4118 |
+
"epoch": 0.5866666666666667,
|
| 4119 |
+
"grad_norm": 0.40400418639183044,
|
| 4120 |
+
"learning_rate": 7.111976413841153e-07,
|
| 4121 |
+
"loss": 0.8586,
|
| 4122 |
+
"step": 583
|
| 4123 |
+
},
|
| 4124 |
+
{
|
| 4125 |
+
"epoch": 0.5876729559748428,
|
| 4126 |
+
"grad_norm": 0.3923444449901581,
|
| 4127 |
+
"learning_rate": 6.301783318272809e-07,
|
| 4128 |
+
"loss": 0.769,
|
| 4129 |
+
"step": 584
|
| 4130 |
+
},
|
| 4131 |
+
{
|
| 4132 |
+
"epoch": 0.5886792452830188,
|
| 4133 |
+
"grad_norm": 0.45618340373039246,
|
| 4134 |
+
"learning_rate": 5.540246732188054e-07,
|
| 4135 |
+
"loss": 0.8698,
|
| 4136 |
+
"step": 585
|
| 4137 |
+
},
|
| 4138 |
+
{
|
| 4139 |
+
"epoch": 0.589685534591195,
|
| 4140 |
+
"grad_norm": 0.4493141174316406,
|
| 4141 |
+
"learning_rate": 4.827450166655251e-07,
|
| 4142 |
+
"loss": 0.774,
|
| 4143 |
+
"step": 586
|
| 4144 |
+
},
|
| 4145 |
+
{
|
| 4146 |
+
"epoch": 0.590691823899371,
|
| 4147 |
+
"grad_norm": 0.4435656666755676,
|
| 4148 |
+
"learning_rate": 4.1634717878503816e-07,
|
| 4149 |
+
"loss": 0.7854,
|
| 4150 |
+
"step": 587
|
| 4151 |
+
},
|
| 4152 |
+
{
|
| 4153 |
+
"epoch": 0.5916981132075472,
|
| 4154 |
+
"grad_norm": 0.4567503035068512,
|
| 4155 |
+
"learning_rate": 3.548384408485006e-07,
|
| 4156 |
+
"loss": 0.8518,
|
| 4157 |
+
"step": 588
|
| 4158 |
+
},
|
| 4159 |
+
{
|
| 4160 |
+
"epoch": 0.5927044025157233,
|
| 4161 |
+
"grad_norm": 0.4994966387748718,
|
| 4162 |
+
"learning_rate": 2.9822554798215994e-07,
|
| 4163 |
+
"loss": 0.8974,
|
| 4164 |
+
"step": 589
|
| 4165 |
+
},
|
| 4166 |
+
{
|
| 4167 |
+
"epoch": 0.5937106918238994,
|
| 4168 |
+
"grad_norm": 0.5203002095222473,
|
| 4169 |
+
"learning_rate": 2.4651470842770196e-07,
|
| 4170 |
+
"loss": 0.8934,
|
| 4171 |
+
"step": 590
|
| 4172 |
+
},
|
| 4173 |
+
{
|
| 4174 |
+
"epoch": 0.5947169811320755,
|
| 4175 |
+
"grad_norm": 0.5493276119232178,
|
| 4176 |
+
"learning_rate": 1.9971159286140017e-07,
|
| 4177 |
+
"loss": 0.9349,
|
| 4178 |
+
"step": 591
|
| 4179 |
+
},
|
| 4180 |
+
{
|
| 4181 |
+
"epoch": 0.5957232704402515,
|
| 4182 |
+
"grad_norm": 0.611061692237854,
|
| 4183 |
+
"learning_rate": 1.5782133377230334e-07,
|
| 4184 |
+
"loss": 0.909,
|
| 4185 |
+
"step": 592
|
| 4186 |
+
},
|
| 4187 |
+
{
|
| 4188 |
+
"epoch": 0.5967295597484277,
|
| 4189 |
+
"grad_norm": 0.5774553418159485,
|
| 4190 |
+
"learning_rate": 1.208485248993857e-07,
|
| 4191 |
+
"loss": 0.9196,
|
| 4192 |
+
"step": 593
|
| 4193 |
+
},
|
| 4194 |
+
{
|
| 4195 |
+
"epoch": 0.5977358490566038,
|
| 4196 |
+
"grad_norm": 0.6194422841072083,
|
| 4197 |
+
"learning_rate": 8.879722072777986e-08,
|
| 4198 |
+
"loss": 0.9283,
|
| 4199 |
+
"step": 594
|
| 4200 |
+
},
|
| 4201 |
+
{
|
| 4202 |
+
"epoch": 0.5987421383647799,
|
| 4203 |
+
"grad_norm": 0.7021675705909729,
|
| 4204 |
+
"learning_rate": 6.167093604417751e-08,
|
| 4205 |
+
"loss": 1.0583,
|
| 4206 |
+
"step": 595
|
| 4207 |
+
},
|
| 4208 |
+
{
|
| 4209 |
+
"epoch": 0.599748427672956,
|
| 4210 |
+
"grad_norm": 0.6871606707572937,
|
| 4211 |
+
"learning_rate": 3.9472645551372757e-08,
|
| 4212 |
+
"loss": 0.9733,
|
| 4213 |
+
"step": 596
|
| 4214 |
+
},
|
| 4215 |
+
{
|
| 4216 |
+
"epoch": 0.600754716981132,
|
| 4217 |
+
"grad_norm": 0.7465012669563293,
|
| 4218 |
+
"learning_rate": 2.2204783542078e-08,
|
| 4219 |
+
"loss": 0.9695,
|
| 4220 |
+
"step": 597
|
| 4221 |
+
},
|
| 4222 |
+
{
|
| 4223 |
+
"epoch": 0.6017610062893082,
|
| 4224 |
+
"grad_norm": 0.8198745250701904,
|
| 4225 |
+
"learning_rate": 9.869243631952518e-09,
|
| 4226 |
+
"loss": 0.9072,
|
| 4227 |
+
"step": 598
|
| 4228 |
+
},
|
| 4229 |
+
{
|
| 4230 |
+
"epoch": 0.6027672955974843,
|
| 4231 |
+
"grad_norm": 0.9207664132118225,
|
| 4232 |
+
"learning_rate": 2.467378551953559e-09,
|
| 4233 |
+
"loss": 0.9067,
|
| 4234 |
+
"step": 599
|
| 4235 |
+
},
|
| 4236 |
+
{
|
| 4237 |
+
"epoch": 0.6037735849056604,
|
| 4238 |
+
"grad_norm": 1.3537120819091797,
|
| 4239 |
+
"learning_rate": 0.0,
|
| 4240 |
+
"loss": 0.8588,
|
| 4241 |
+
"step": 600
|
| 4242 |
+
},
|
| 4243 |
+
{
|
| 4244 |
+
"epoch": 0.6037735849056604,
|
| 4245 |
+
"eval_loss": 0.7933821678161621,
|
| 4246 |
+
"eval_runtime": 72.1908,
|
| 4247 |
+
"eval_samples_per_second": 46.363,
|
| 4248 |
+
"eval_steps_per_second": 11.594,
|
| 4249 |
+
"step": 600
|
| 4250 |
}
|
| 4251 |
],
|
| 4252 |
"logging_steps": 1,
|
|
|
|
| 4270 |
"should_evaluate": false,
|
| 4271 |
"should_log": false,
|
| 4272 |
"should_save": true,
|
| 4273 |
+
"should_training_stop": true
|
| 4274 |
},
|
| 4275 |
"attributes": {}
|
| 4276 |
}
|
| 4277 |
},
|
| 4278 |
+
"total_flos": 3.430384330289971e+17,
|
| 4279 |
"train_batch_size": 16,
|
| 4280 |
"trial_name": null,
|
| 4281 |
"trial_params": null
|