Training in progress, step 699, 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 14293800
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:0b098c267df4058757e013da2353ed0e41ef6d403ed82445ed658dcf91b1ba7d
|
| 3 |
size 14293800
|
last-checkpoint/optimizer.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 7580068
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:e04bd14d7c7ad63b28c7159a22b510d6fc239bb5d11f324e22a1faef3024d006
|
| 3 |
size 7580068
|
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 sha256:17a9390150a36562e39e016d5aad5d61ae4c54518262bb0bae25d0421c3afecd
|
| 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 sha256:876db3afeec1a47b9cd23534bb10590b976ccab59f5486942f107d4e7111af82
|
| 3 |
size 1064
|
last-checkpoint/trainer_state.json
CHANGED
|
@@ -1,9 +1,9 @@
|
|
| 1 |
{
|
| 2 |
"best_metric": null,
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
-
"epoch": 0.
|
| 5 |
"eval_steps": 233,
|
| 6 |
-
"global_step":
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
@@ -3293,6 +3293,1645 @@
|
|
| 3293 |
"eval_samples_per_second": 70.571,
|
| 3294 |
"eval_steps_per_second": 35.286,
|
| 3295 |
"step": 466
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 3296 |
}
|
| 3297 |
],
|
| 3298 |
"logging_steps": 1,
|
|
@@ -3312,7 +4951,7 @@
|
|
| 3312 |
"attributes": {}
|
| 3313 |
}
|
| 3314 |
},
|
| 3315 |
-
"total_flos":
|
| 3316 |
"train_batch_size": 2,
|
| 3317 |
"trial_name": null,
|
| 3318 |
"trial_params": null
|
|
|
|
| 1 |
{
|
| 2 |
"best_metric": null,
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 0.7514109110454179,
|
| 5 |
"eval_steps": 233,
|
| 6 |
+
"global_step": 699,
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
|
|
| 3293 |
"eval_samples_per_second": 70.571,
|
| 3294 |
"eval_steps_per_second": 35.286,
|
| 3295 |
"step": 466
|
| 3296 |
+
},
|
| 3297 |
+
{
|
| 3298 |
+
"epoch": 0.5020155872077399,
|
| 3299 |
+
"grad_norm": 14.241241455078125,
|
| 3300 |
+
"learning_rate": 0.0001011938450687699,
|
| 3301 |
+
"loss": 2.6351,
|
| 3302 |
+
"step": 467
|
| 3303 |
+
},
|
| 3304 |
+
{
|
| 3305 |
+
"epoch": 0.5030905670518678,
|
| 3306 |
+
"grad_norm": 15.222787857055664,
|
| 3307 |
+
"learning_rate": 0.00010085275639984904,
|
| 3308 |
+
"loss": 3.6475,
|
| 3309 |
+
"step": 468
|
| 3310 |
+
},
|
| 3311 |
+
{
|
| 3312 |
+
"epoch": 0.5041655468959957,
|
| 3313 |
+
"grad_norm": 12.280945777893066,
|
| 3314 |
+
"learning_rate": 0.00010051165780879504,
|
| 3315 |
+
"loss": 2.5128,
|
| 3316 |
+
"step": 469
|
| 3317 |
+
},
|
| 3318 |
+
{
|
| 3319 |
+
"epoch": 0.5052405267401237,
|
| 3320 |
+
"grad_norm": 14.948744773864746,
|
| 3321 |
+
"learning_rate": 0.00010017055326441494,
|
| 3322 |
+
"loss": 2.9114,
|
| 3323 |
+
"step": 470
|
| 3324 |
+
},
|
| 3325 |
+
{
|
| 3326 |
+
"epoch": 0.5063155065842515,
|
| 3327 |
+
"grad_norm": 17.295438766479492,
|
| 3328 |
+
"learning_rate": 9.982944673558508e-05,
|
| 3329 |
+
"loss": 3.4193,
|
| 3330 |
+
"step": 471
|
| 3331 |
+
},
|
| 3332 |
+
{
|
| 3333 |
+
"epoch": 0.5073904864283795,
|
| 3334 |
+
"grad_norm": 18.69068145751953,
|
| 3335 |
+
"learning_rate": 9.9488342191205e-05,
|
| 3336 |
+
"loss": 4.0637,
|
| 3337 |
+
"step": 472
|
| 3338 |
+
},
|
| 3339 |
+
{
|
| 3340 |
+
"epoch": 0.5084654662725074,
|
| 3341 |
+
"grad_norm": 17.060588836669922,
|
| 3342 |
+
"learning_rate": 9.914724360015099e-05,
|
| 3343 |
+
"loss": 3.1746,
|
| 3344 |
+
"step": 473
|
| 3345 |
+
},
|
| 3346 |
+
{
|
| 3347 |
+
"epoch": 0.5095404461166353,
|
| 3348 |
+
"grad_norm": 13.83548641204834,
|
| 3349 |
+
"learning_rate": 9.880615493123012e-05,
|
| 3350 |
+
"loss": 2.757,
|
| 3351 |
+
"step": 474
|
| 3352 |
+
},
|
| 3353 |
+
{
|
| 3354 |
+
"epoch": 0.5106154259607633,
|
| 3355 |
+
"grad_norm": 12.663851737976074,
|
| 3356 |
+
"learning_rate": 9.846508015313408e-05,
|
| 3357 |
+
"loss": 2.7217,
|
| 3358 |
+
"step": 475
|
| 3359 |
+
},
|
| 3360 |
+
{
|
| 3361 |
+
"epoch": 0.5116904058048911,
|
| 3362 |
+
"grad_norm": 14.959012985229492,
|
| 3363 |
+
"learning_rate": 9.812402323439284e-05,
|
| 3364 |
+
"loss": 3.3345,
|
| 3365 |
+
"step": 476
|
| 3366 |
+
},
|
| 3367 |
+
{
|
| 3368 |
+
"epoch": 0.5127653856490191,
|
| 3369 |
+
"grad_norm": 14.708747863769531,
|
| 3370 |
+
"learning_rate": 9.778298814332863e-05,
|
| 3371 |
+
"loss": 2.8671,
|
| 3372 |
+
"step": 477
|
| 3373 |
+
},
|
| 3374 |
+
{
|
| 3375 |
+
"epoch": 0.513840365493147,
|
| 3376 |
+
"grad_norm": 10.635260581970215,
|
| 3377 |
+
"learning_rate": 9.744197884800969e-05,
|
| 3378 |
+
"loss": 2.6943,
|
| 3379 |
+
"step": 478
|
| 3380 |
+
},
|
| 3381 |
+
{
|
| 3382 |
+
"epoch": 0.5149153453372749,
|
| 3383 |
+
"grad_norm": 14.44356632232666,
|
| 3384 |
+
"learning_rate": 9.710099931620408e-05,
|
| 3385 |
+
"loss": 2.1394,
|
| 3386 |
+
"step": 479
|
| 3387 |
+
},
|
| 3388 |
+
{
|
| 3389 |
+
"epoch": 0.5159903251814029,
|
| 3390 |
+
"grad_norm": 9.918120384216309,
|
| 3391 |
+
"learning_rate": 9.676005351533366e-05,
|
| 3392 |
+
"loss": 2.2011,
|
| 3393 |
+
"step": 480
|
| 3394 |
+
},
|
| 3395 |
+
{
|
| 3396 |
+
"epoch": 0.5170653050255307,
|
| 3397 |
+
"grad_norm": 17.1123046875,
|
| 3398 |
+
"learning_rate": 9.64191454124277e-05,
|
| 3399 |
+
"loss": 3.2862,
|
| 3400 |
+
"step": 481
|
| 3401 |
+
},
|
| 3402 |
+
{
|
| 3403 |
+
"epoch": 0.5181402848696587,
|
| 3404 |
+
"grad_norm": 13.601349830627441,
|
| 3405 |
+
"learning_rate": 9.60782789740769e-05,
|
| 3406 |
+
"loss": 2.3652,
|
| 3407 |
+
"step": 482
|
| 3408 |
+
},
|
| 3409 |
+
{
|
| 3410 |
+
"epoch": 0.5192152647137867,
|
| 3411 |
+
"grad_norm": 16.888429641723633,
|
| 3412 |
+
"learning_rate": 9.573745816638716e-05,
|
| 3413 |
+
"loss": 2.7698,
|
| 3414 |
+
"step": 483
|
| 3415 |
+
},
|
| 3416 |
+
{
|
| 3417 |
+
"epoch": 0.5202902445579145,
|
| 3418 |
+
"grad_norm": 16.15688705444336,
|
| 3419 |
+
"learning_rate": 9.539668695493344e-05,
|
| 3420 |
+
"loss": 2.6367,
|
| 3421 |
+
"step": 484
|
| 3422 |
+
},
|
| 3423 |
+
{
|
| 3424 |
+
"epoch": 0.5213652244020425,
|
| 3425 |
+
"grad_norm": 16.414520263671875,
|
| 3426 |
+
"learning_rate": 9.505596930471367e-05,
|
| 3427 |
+
"loss": 3.9243,
|
| 3428 |
+
"step": 485
|
| 3429 |
+
},
|
| 3430 |
+
{
|
| 3431 |
+
"epoch": 0.5224402042461704,
|
| 3432 |
+
"grad_norm": 13.967657089233398,
|
| 3433 |
+
"learning_rate": 9.471530918010253e-05,
|
| 3434 |
+
"loss": 3.1243,
|
| 3435 |
+
"step": 486
|
| 3436 |
+
},
|
| 3437 |
+
{
|
| 3438 |
+
"epoch": 0.5235151840902983,
|
| 3439 |
+
"grad_norm": 15.140684127807617,
|
| 3440 |
+
"learning_rate": 9.43747105448054e-05,
|
| 3441 |
+
"loss": 2.659,
|
| 3442 |
+
"step": 487
|
| 3443 |
+
},
|
| 3444 |
+
{
|
| 3445 |
+
"epoch": 0.5245901639344263,
|
| 3446 |
+
"grad_norm": 13.074856758117676,
|
| 3447 |
+
"learning_rate": 9.40341773618122e-05,
|
| 3448 |
+
"loss": 3.3224,
|
| 3449 |
+
"step": 488
|
| 3450 |
+
},
|
| 3451 |
+
{
|
| 3452 |
+
"epoch": 0.5256651437785541,
|
| 3453 |
+
"grad_norm": 15.123608589172363,
|
| 3454 |
+
"learning_rate": 9.369371359335128e-05,
|
| 3455 |
+
"loss": 3.385,
|
| 3456 |
+
"step": 489
|
| 3457 |
+
},
|
| 3458 |
+
{
|
| 3459 |
+
"epoch": 0.5267401236226821,
|
| 3460 |
+
"grad_norm": 15.648529052734375,
|
| 3461 |
+
"learning_rate": 9.335332320084331e-05,
|
| 3462 |
+
"loss": 2.8329,
|
| 3463 |
+
"step": 490
|
| 3464 |
+
},
|
| 3465 |
+
{
|
| 3466 |
+
"epoch": 0.52781510346681,
|
| 3467 |
+
"grad_norm": 15.20040225982666,
|
| 3468 |
+
"learning_rate": 9.301301014485528e-05,
|
| 3469 |
+
"loss": 3.5456,
|
| 3470 |
+
"step": 491
|
| 3471 |
+
},
|
| 3472 |
+
{
|
| 3473 |
+
"epoch": 0.5288900833109379,
|
| 3474 |
+
"grad_norm": 23.638113021850586,
|
| 3475 |
+
"learning_rate": 9.267277838505423e-05,
|
| 3476 |
+
"loss": 4.8434,
|
| 3477 |
+
"step": 492
|
| 3478 |
+
},
|
| 3479 |
+
{
|
| 3480 |
+
"epoch": 0.5299650631550659,
|
| 3481 |
+
"grad_norm": 11.857388496398926,
|
| 3482 |
+
"learning_rate": 9.233263188016138e-05,
|
| 3483 |
+
"loss": 2.2761,
|
| 3484 |
+
"step": 493
|
| 3485 |
+
},
|
| 3486 |
+
{
|
| 3487 |
+
"epoch": 0.5310400429991937,
|
| 3488 |
+
"grad_norm": 12.123178482055664,
|
| 3489 |
+
"learning_rate": 9.199257458790591e-05,
|
| 3490 |
+
"loss": 3.0025,
|
| 3491 |
+
"step": 494
|
| 3492 |
+
},
|
| 3493 |
+
{
|
| 3494 |
+
"epoch": 0.5321150228433217,
|
| 3495 |
+
"grad_norm": 11.024534225463867,
|
| 3496 |
+
"learning_rate": 9.165261046497907e-05,
|
| 3497 |
+
"loss": 2.265,
|
| 3498 |
+
"step": 495
|
| 3499 |
+
},
|
| 3500 |
+
{
|
| 3501 |
+
"epoch": 0.5331900026874496,
|
| 3502 |
+
"grad_norm": 16.32103157043457,
|
| 3503 |
+
"learning_rate": 9.131274346698796e-05,
|
| 3504 |
+
"loss": 3.5393,
|
| 3505 |
+
"step": 496
|
| 3506 |
+
},
|
| 3507 |
+
{
|
| 3508 |
+
"epoch": 0.5342649825315775,
|
| 3509 |
+
"grad_norm": 25.771560668945312,
|
| 3510 |
+
"learning_rate": 9.097297754840962e-05,
|
| 3511 |
+
"loss": 3.8375,
|
| 3512 |
+
"step": 497
|
| 3513 |
+
},
|
| 3514 |
+
{
|
| 3515 |
+
"epoch": 0.5353399623757055,
|
| 3516 |
+
"grad_norm": 12.820847511291504,
|
| 3517 |
+
"learning_rate": 9.063331666254503e-05,
|
| 3518 |
+
"loss": 2.6361,
|
| 3519 |
+
"step": 498
|
| 3520 |
+
},
|
| 3521 |
+
{
|
| 3522 |
+
"epoch": 0.5364149422198333,
|
| 3523 |
+
"grad_norm": 12.816265106201172,
|
| 3524 |
+
"learning_rate": 9.029376476147302e-05,
|
| 3525 |
+
"loss": 2.2486,
|
| 3526 |
+
"step": 499
|
| 3527 |
+
},
|
| 3528 |
+
{
|
| 3529 |
+
"epoch": 0.5374899220639613,
|
| 3530 |
+
"grad_norm": 9.368247032165527,
|
| 3531 |
+
"learning_rate": 8.995432579600439e-05,
|
| 3532 |
+
"loss": 2.5467,
|
| 3533 |
+
"step": 500
|
| 3534 |
+
},
|
| 3535 |
+
{
|
| 3536 |
+
"epoch": 0.5385649019080893,
|
| 3537 |
+
"grad_norm": 16.314271926879883,
|
| 3538 |
+
"learning_rate": 8.961500371563585e-05,
|
| 3539 |
+
"loss": 3.1917,
|
| 3540 |
+
"step": 501
|
| 3541 |
+
},
|
| 3542 |
+
{
|
| 3543 |
+
"epoch": 0.5396398817522171,
|
| 3544 |
+
"grad_norm": 17.877838134765625,
|
| 3545 |
+
"learning_rate": 8.927580246850418e-05,
|
| 3546 |
+
"loss": 3.636,
|
| 3547 |
+
"step": 502
|
| 3548 |
+
},
|
| 3549 |
+
{
|
| 3550 |
+
"epoch": 0.5407148615963451,
|
| 3551 |
+
"grad_norm": 15.371697425842285,
|
| 3552 |
+
"learning_rate": 8.893672600134013e-05,
|
| 3553 |
+
"loss": 4.3843,
|
| 3554 |
+
"step": 503
|
| 3555 |
+
},
|
| 3556 |
+
{
|
| 3557 |
+
"epoch": 0.541789841440473,
|
| 3558 |
+
"grad_norm": 14.944757461547852,
|
| 3559 |
+
"learning_rate": 8.859777825942267e-05,
|
| 3560 |
+
"loss": 2.2189,
|
| 3561 |
+
"step": 504
|
| 3562 |
+
},
|
| 3563 |
+
{
|
| 3564 |
+
"epoch": 0.5428648212846009,
|
| 3565 |
+
"grad_norm": 17.238252639770508,
|
| 3566 |
+
"learning_rate": 8.825896318653293e-05,
|
| 3567 |
+
"loss": 2.525,
|
| 3568 |
+
"step": 505
|
| 3569 |
+
},
|
| 3570 |
+
{
|
| 3571 |
+
"epoch": 0.5439398011287289,
|
| 3572 |
+
"grad_norm": 14.758814811706543,
|
| 3573 |
+
"learning_rate": 8.792028472490844e-05,
|
| 3574 |
+
"loss": 3.1758,
|
| 3575 |
+
"step": 506
|
| 3576 |
+
},
|
| 3577 |
+
{
|
| 3578 |
+
"epoch": 0.5450147809728567,
|
| 3579 |
+
"grad_norm": 9.887633323669434,
|
| 3580 |
+
"learning_rate": 8.758174681519721e-05,
|
| 3581 |
+
"loss": 2.2908,
|
| 3582 |
+
"step": 507
|
| 3583 |
+
},
|
| 3584 |
+
{
|
| 3585 |
+
"epoch": 0.5460897608169847,
|
| 3586 |
+
"grad_norm": 13.68622875213623,
|
| 3587 |
+
"learning_rate": 8.724335339641184e-05,
|
| 3588 |
+
"loss": 2.105,
|
| 3589 |
+
"step": 508
|
| 3590 |
+
},
|
| 3591 |
+
{
|
| 3592 |
+
"epoch": 0.5471647406611126,
|
| 3593 |
+
"grad_norm": 12.679695129394531,
|
| 3594 |
+
"learning_rate": 8.690510840588373e-05,
|
| 3595 |
+
"loss": 2.1756,
|
| 3596 |
+
"step": 509
|
| 3597 |
+
},
|
| 3598 |
+
{
|
| 3599 |
+
"epoch": 0.5482397205052405,
|
| 3600 |
+
"grad_norm": 14.024535179138184,
|
| 3601 |
+
"learning_rate": 8.656701577921732e-05,
|
| 3602 |
+
"loss": 3.0431,
|
| 3603 |
+
"step": 510
|
| 3604 |
+
},
|
| 3605 |
+
{
|
| 3606 |
+
"epoch": 0.5493147003493685,
|
| 3607 |
+
"grad_norm": 12.965935707092285,
|
| 3608 |
+
"learning_rate": 8.622907945024417e-05,
|
| 3609 |
+
"loss": 2.1099,
|
| 3610 |
+
"step": 511
|
| 3611 |
+
},
|
| 3612 |
+
{
|
| 3613 |
+
"epoch": 0.5503896801934963,
|
| 3614 |
+
"grad_norm": 19.419710159301758,
|
| 3615 |
+
"learning_rate": 8.589130335097732e-05,
|
| 3616 |
+
"loss": 3.3639,
|
| 3617 |
+
"step": 512
|
| 3618 |
+
},
|
| 3619 |
+
{
|
| 3620 |
+
"epoch": 0.5514646600376243,
|
| 3621 |
+
"grad_norm": 18.27731704711914,
|
| 3622 |
+
"learning_rate": 8.55536914115654e-05,
|
| 3623 |
+
"loss": 3.5416,
|
| 3624 |
+
"step": 513
|
| 3625 |
+
},
|
| 3626 |
+
{
|
| 3627 |
+
"epoch": 0.5525396398817523,
|
| 3628 |
+
"grad_norm": 16.81820297241211,
|
| 3629 |
+
"learning_rate": 8.521624756024705e-05,
|
| 3630 |
+
"loss": 3.8419,
|
| 3631 |
+
"step": 514
|
| 3632 |
+
},
|
| 3633 |
+
{
|
| 3634 |
+
"epoch": 0.5536146197258801,
|
| 3635 |
+
"grad_norm": 14.465489387512207,
|
| 3636 |
+
"learning_rate": 8.487897572330513e-05,
|
| 3637 |
+
"loss": 2.3487,
|
| 3638 |
+
"step": 515
|
| 3639 |
+
},
|
| 3640 |
+
{
|
| 3641 |
+
"epoch": 0.5546895995700081,
|
| 3642 |
+
"grad_norm": 11.499032974243164,
|
| 3643 |
+
"learning_rate": 8.454187982502101e-05,
|
| 3644 |
+
"loss": 2.6283,
|
| 3645 |
+
"step": 516
|
| 3646 |
+
},
|
| 3647 |
+
{
|
| 3648 |
+
"epoch": 0.555764579414136,
|
| 3649 |
+
"grad_norm": 21.029111862182617,
|
| 3650 |
+
"learning_rate": 8.4204963787629e-05,
|
| 3651 |
+
"loss": 4.7078,
|
| 3652 |
+
"step": 517
|
| 3653 |
+
},
|
| 3654 |
+
{
|
| 3655 |
+
"epoch": 0.5568395592582639,
|
| 3656 |
+
"grad_norm": 18.131210327148438,
|
| 3657 |
+
"learning_rate": 8.386823153127064e-05,
|
| 3658 |
+
"loss": 3.6223,
|
| 3659 |
+
"step": 518
|
| 3660 |
+
},
|
| 3661 |
+
{
|
| 3662 |
+
"epoch": 0.5579145391023919,
|
| 3663 |
+
"grad_norm": 15.802128791809082,
|
| 3664 |
+
"learning_rate": 8.353168697394913e-05,
|
| 3665 |
+
"loss": 2.6126,
|
| 3666 |
+
"step": 519
|
| 3667 |
+
},
|
| 3668 |
+
{
|
| 3669 |
+
"epoch": 0.5589895189465197,
|
| 3670 |
+
"grad_norm": 16.3378849029541,
|
| 3671 |
+
"learning_rate": 8.319533403148367e-05,
|
| 3672 |
+
"loss": 2.8075,
|
| 3673 |
+
"step": 520
|
| 3674 |
+
},
|
| 3675 |
+
{
|
| 3676 |
+
"epoch": 0.5600644987906477,
|
| 3677 |
+
"grad_norm": 13.895936012268066,
|
| 3678 |
+
"learning_rate": 8.285917661746401e-05,
|
| 3679 |
+
"loss": 2.7503,
|
| 3680 |
+
"step": 521
|
| 3681 |
+
},
|
| 3682 |
+
{
|
| 3683 |
+
"epoch": 0.5611394786347756,
|
| 3684 |
+
"grad_norm": 13.693531036376953,
|
| 3685 |
+
"learning_rate": 8.25232186432048e-05,
|
| 3686 |
+
"loss": 2.9142,
|
| 3687 |
+
"step": 522
|
| 3688 |
+
},
|
| 3689 |
+
{
|
| 3690 |
+
"epoch": 0.5622144584789035,
|
| 3691 |
+
"grad_norm": 14.982376098632812,
|
| 3692 |
+
"learning_rate": 8.218746401770022e-05,
|
| 3693 |
+
"loss": 3.0101,
|
| 3694 |
+
"step": 523
|
| 3695 |
+
},
|
| 3696 |
+
{
|
| 3697 |
+
"epoch": 0.5632894383230315,
|
| 3698 |
+
"grad_norm": 18.6697998046875,
|
| 3699 |
+
"learning_rate": 8.185191664757828e-05,
|
| 3700 |
+
"loss": 3.6426,
|
| 3701 |
+
"step": 524
|
| 3702 |
+
},
|
| 3703 |
+
{
|
| 3704 |
+
"epoch": 0.5643644181671593,
|
| 3705 |
+
"grad_norm": 12.057175636291504,
|
| 3706 |
+
"learning_rate": 8.151658043705565e-05,
|
| 3707 |
+
"loss": 2.9482,
|
| 3708 |
+
"step": 525
|
| 3709 |
+
},
|
| 3710 |
+
{
|
| 3711 |
+
"epoch": 0.5654393980112873,
|
| 3712 |
+
"grad_norm": 14.079970359802246,
|
| 3713 |
+
"learning_rate": 8.118145928789199e-05,
|
| 3714 |
+
"loss": 3.1769,
|
| 3715 |
+
"step": 526
|
| 3716 |
+
},
|
| 3717 |
+
{
|
| 3718 |
+
"epoch": 0.5665143778554153,
|
| 3719 |
+
"grad_norm": 12.814187049865723,
|
| 3720 |
+
"learning_rate": 8.084655709934462e-05,
|
| 3721 |
+
"loss": 2.472,
|
| 3722 |
+
"step": 527
|
| 3723 |
+
},
|
| 3724 |
+
{
|
| 3725 |
+
"epoch": 0.5675893576995431,
|
| 3726 |
+
"grad_norm": 10.802642822265625,
|
| 3727 |
+
"learning_rate": 8.051187776812326e-05,
|
| 3728 |
+
"loss": 2.0466,
|
| 3729 |
+
"step": 528
|
| 3730 |
+
},
|
| 3731 |
+
{
|
| 3732 |
+
"epoch": 0.5686643375436711,
|
| 3733 |
+
"grad_norm": 12.31850528717041,
|
| 3734 |
+
"learning_rate": 8.017742518834454e-05,
|
| 3735 |
+
"loss": 2.4457,
|
| 3736 |
+
"step": 529
|
| 3737 |
+
},
|
| 3738 |
+
{
|
| 3739 |
+
"epoch": 0.5697393173877989,
|
| 3740 |
+
"grad_norm": 13.78878116607666,
|
| 3741 |
+
"learning_rate": 7.984320325148675e-05,
|
| 3742 |
+
"loss": 2.6326,
|
| 3743 |
+
"step": 530
|
| 3744 |
+
},
|
| 3745 |
+
{
|
| 3746 |
+
"epoch": 0.5708142972319269,
|
| 3747 |
+
"grad_norm": 9.785225868225098,
|
| 3748 |
+
"learning_rate": 7.950921584634461e-05,
|
| 3749 |
+
"loss": 2.8243,
|
| 3750 |
+
"step": 531
|
| 3751 |
+
},
|
| 3752 |
+
{
|
| 3753 |
+
"epoch": 0.5718892770760549,
|
| 3754 |
+
"grad_norm": 16.127605438232422,
|
| 3755 |
+
"learning_rate": 7.917546685898391e-05,
|
| 3756 |
+
"loss": 3.5011,
|
| 3757 |
+
"step": 532
|
| 3758 |
+
},
|
| 3759 |
+
{
|
| 3760 |
+
"epoch": 0.5729642569201827,
|
| 3761 |
+
"grad_norm": 20.46214485168457,
|
| 3762 |
+
"learning_rate": 7.884196017269648e-05,
|
| 3763 |
+
"loss": 2.5311,
|
| 3764 |
+
"step": 533
|
| 3765 |
+
},
|
| 3766 |
+
{
|
| 3767 |
+
"epoch": 0.5740392367643107,
|
| 3768 |
+
"grad_norm": 13.586955070495605,
|
| 3769 |
+
"learning_rate": 7.850869966795476e-05,
|
| 3770 |
+
"loss": 2.8393,
|
| 3771 |
+
"step": 534
|
| 3772 |
+
},
|
| 3773 |
+
{
|
| 3774 |
+
"epoch": 0.5751142166084386,
|
| 3775 |
+
"grad_norm": 16.564584732055664,
|
| 3776 |
+
"learning_rate": 7.817568922236682e-05,
|
| 3777 |
+
"loss": 2.3696,
|
| 3778 |
+
"step": 535
|
| 3779 |
+
},
|
| 3780 |
+
{
|
| 3781 |
+
"epoch": 0.5761891964525665,
|
| 3782 |
+
"grad_norm": 21.446279525756836,
|
| 3783 |
+
"learning_rate": 7.784293271063124e-05,
|
| 3784 |
+
"loss": 4.4285,
|
| 3785 |
+
"step": 536
|
| 3786 |
+
},
|
| 3787 |
+
{
|
| 3788 |
+
"epoch": 0.5772641762966945,
|
| 3789 |
+
"grad_norm": 18.923242568969727,
|
| 3790 |
+
"learning_rate": 7.751043400449197e-05,
|
| 3791 |
+
"loss": 3.2939,
|
| 3792 |
+
"step": 537
|
| 3793 |
+
},
|
| 3794 |
+
{
|
| 3795 |
+
"epoch": 0.5783391561408223,
|
| 3796 |
+
"grad_norm": 16.000579833984375,
|
| 3797 |
+
"learning_rate": 7.717819697269321e-05,
|
| 3798 |
+
"loss": 3.8915,
|
| 3799 |
+
"step": 538
|
| 3800 |
+
},
|
| 3801 |
+
{
|
| 3802 |
+
"epoch": 0.5794141359849503,
|
| 3803 |
+
"grad_norm": 11.695368766784668,
|
| 3804 |
+
"learning_rate": 7.684622548093461e-05,
|
| 3805 |
+
"loss": 2.5856,
|
| 3806 |
+
"step": 539
|
| 3807 |
+
},
|
| 3808 |
+
{
|
| 3809 |
+
"epoch": 0.5804891158290783,
|
| 3810 |
+
"grad_norm": 15.072840690612793,
|
| 3811 |
+
"learning_rate": 7.651452339182613e-05,
|
| 3812 |
+
"loss": 2.8462,
|
| 3813 |
+
"step": 540
|
| 3814 |
+
},
|
| 3815 |
+
{
|
| 3816 |
+
"epoch": 0.5815640956732061,
|
| 3817 |
+
"grad_norm": 18.407136917114258,
|
| 3818 |
+
"learning_rate": 7.618309456484308e-05,
|
| 3819 |
+
"loss": 2.5811,
|
| 3820 |
+
"step": 541
|
| 3821 |
+
},
|
| 3822 |
+
{
|
| 3823 |
+
"epoch": 0.5826390755173341,
|
| 3824 |
+
"grad_norm": 17.274293899536133,
|
| 3825 |
+
"learning_rate": 7.58519428562813e-05,
|
| 3826 |
+
"loss": 3.2455,
|
| 3827 |
+
"step": 542
|
| 3828 |
+
},
|
| 3829 |
+
{
|
| 3830 |
+
"epoch": 0.5837140553614619,
|
| 3831 |
+
"grad_norm": 15.445805549621582,
|
| 3832 |
+
"learning_rate": 7.552107211921229e-05,
|
| 3833 |
+
"loss": 3.3812,
|
| 3834 |
+
"step": 543
|
| 3835 |
+
},
|
| 3836 |
+
{
|
| 3837 |
+
"epoch": 0.5847890352055899,
|
| 3838 |
+
"grad_norm": 16.473222732543945,
|
| 3839 |
+
"learning_rate": 7.519048620343825e-05,
|
| 3840 |
+
"loss": 2.9544,
|
| 3841 |
+
"step": 544
|
| 3842 |
+
},
|
| 3843 |
+
{
|
| 3844 |
+
"epoch": 0.5858640150497179,
|
| 3845 |
+
"grad_norm": 16.593351364135742,
|
| 3846 |
+
"learning_rate": 7.486018895544748e-05,
|
| 3847 |
+
"loss": 3.8982,
|
| 3848 |
+
"step": 545
|
| 3849 |
+
},
|
| 3850 |
+
{
|
| 3851 |
+
"epoch": 0.5869389948938457,
|
| 3852 |
+
"grad_norm": 18.829208374023438,
|
| 3853 |
+
"learning_rate": 7.453018421836946e-05,
|
| 3854 |
+
"loss": 2.933,
|
| 3855 |
+
"step": 546
|
| 3856 |
+
},
|
| 3857 |
+
{
|
| 3858 |
+
"epoch": 0.5880139747379737,
|
| 3859 |
+
"grad_norm": 21.860137939453125,
|
| 3860 |
+
"learning_rate": 7.420047583193019e-05,
|
| 3861 |
+
"loss": 3.5987,
|
| 3862 |
+
"step": 547
|
| 3863 |
+
},
|
| 3864 |
+
{
|
| 3865 |
+
"epoch": 0.5890889545821015,
|
| 3866 |
+
"grad_norm": 15.280534744262695,
|
| 3867 |
+
"learning_rate": 7.387106763240763e-05,
|
| 3868 |
+
"loss": 2.8143,
|
| 3869 |
+
"step": 548
|
| 3870 |
+
},
|
| 3871 |
+
{
|
| 3872 |
+
"epoch": 0.5901639344262295,
|
| 3873 |
+
"grad_norm": 15.49770736694336,
|
| 3874 |
+
"learning_rate": 7.354196345258683e-05,
|
| 3875 |
+
"loss": 2.4214,
|
| 3876 |
+
"step": 549
|
| 3877 |
+
},
|
| 3878 |
+
{
|
| 3879 |
+
"epoch": 0.5912389142703575,
|
| 3880 |
+
"grad_norm": 15.602472305297852,
|
| 3881 |
+
"learning_rate": 7.32131671217155e-05,
|
| 3882 |
+
"loss": 3.4367,
|
| 3883 |
+
"step": 550
|
| 3884 |
+
},
|
| 3885 |
+
{
|
| 3886 |
+
"epoch": 0.5923138941144853,
|
| 3887 |
+
"grad_norm": 14.388925552368164,
|
| 3888 |
+
"learning_rate": 7.288468246545946e-05,
|
| 3889 |
+
"loss": 3.4037,
|
| 3890 |
+
"step": 551
|
| 3891 |
+
},
|
| 3892 |
+
{
|
| 3893 |
+
"epoch": 0.5933888739586133,
|
| 3894 |
+
"grad_norm": 19.788185119628906,
|
| 3895 |
+
"learning_rate": 7.255651330585797e-05,
|
| 3896 |
+
"loss": 3.3473,
|
| 3897 |
+
"step": 552
|
| 3898 |
+
},
|
| 3899 |
+
{
|
| 3900 |
+
"epoch": 0.5944638538027412,
|
| 3901 |
+
"grad_norm": 13.820460319519043,
|
| 3902 |
+
"learning_rate": 7.222866346127953e-05,
|
| 3903 |
+
"loss": 2.5223,
|
| 3904 |
+
"step": 553
|
| 3905 |
+
},
|
| 3906 |
+
{
|
| 3907 |
+
"epoch": 0.5955388336468691,
|
| 3908 |
+
"grad_norm": 16.7007999420166,
|
| 3909 |
+
"learning_rate": 7.190113674637714e-05,
|
| 3910 |
+
"loss": 2.8172,
|
| 3911 |
+
"step": 554
|
| 3912 |
+
},
|
| 3913 |
+
{
|
| 3914 |
+
"epoch": 0.5966138134909971,
|
| 3915 |
+
"grad_norm": 16.35044288635254,
|
| 3916 |
+
"learning_rate": 7.157393697204416e-05,
|
| 3917 |
+
"loss": 2.9871,
|
| 3918 |
+
"step": 555
|
| 3919 |
+
},
|
| 3920 |
+
{
|
| 3921 |
+
"epoch": 0.5976887933351249,
|
| 3922 |
+
"grad_norm": 13.284764289855957,
|
| 3923 |
+
"learning_rate": 7.124706794536983e-05,
|
| 3924 |
+
"loss": 3.6496,
|
| 3925 |
+
"step": 556
|
| 3926 |
+
},
|
| 3927 |
+
{
|
| 3928 |
+
"epoch": 0.5987637731792529,
|
| 3929 |
+
"grad_norm": 12.027867317199707,
|
| 3930 |
+
"learning_rate": 7.09205334695951e-05,
|
| 3931 |
+
"loss": 2.2565,
|
| 3932 |
+
"step": 557
|
| 3933 |
+
},
|
| 3934 |
+
{
|
| 3935 |
+
"epoch": 0.5998387530233809,
|
| 3936 |
+
"grad_norm": 17.299827575683594,
|
| 3937 |
+
"learning_rate": 7.059433734406818e-05,
|
| 3938 |
+
"loss": 2.7168,
|
| 3939 |
+
"step": 558
|
| 3940 |
+
},
|
| 3941 |
+
{
|
| 3942 |
+
"epoch": 0.6009137328675087,
|
| 3943 |
+
"grad_norm": 17.865158081054688,
|
| 3944 |
+
"learning_rate": 7.026848336420054e-05,
|
| 3945 |
+
"loss": 3.6538,
|
| 3946 |
+
"step": 559
|
| 3947 |
+
},
|
| 3948 |
+
{
|
| 3949 |
+
"epoch": 0.6019887127116367,
|
| 3950 |
+
"grad_norm": 12.184990882873535,
|
| 3951 |
+
"learning_rate": 6.99429753214226e-05,
|
| 3952 |
+
"loss": 1.9531,
|
| 3953 |
+
"step": 560
|
| 3954 |
+
},
|
| 3955 |
+
{
|
| 3956 |
+
"epoch": 0.6030636925557645,
|
| 3957 |
+
"grad_norm": 15.416332244873047,
|
| 3958 |
+
"learning_rate": 6.961781700313972e-05,
|
| 3959 |
+
"loss": 2.8138,
|
| 3960 |
+
"step": 561
|
| 3961 |
+
},
|
| 3962 |
+
{
|
| 3963 |
+
"epoch": 0.6041386723998925,
|
| 3964 |
+
"grad_norm": 16.23126220703125,
|
| 3965 |
+
"learning_rate": 6.929301219268805e-05,
|
| 3966 |
+
"loss": 2.6759,
|
| 3967 |
+
"step": 562
|
| 3968 |
+
},
|
| 3969 |
+
{
|
| 3970 |
+
"epoch": 0.6052136522440205,
|
| 3971 |
+
"grad_norm": 21.727291107177734,
|
| 3972 |
+
"learning_rate": 6.896856466929062e-05,
|
| 3973 |
+
"loss": 3.2578,
|
| 3974 |
+
"step": 563
|
| 3975 |
+
},
|
| 3976 |
+
{
|
| 3977 |
+
"epoch": 0.6062886320881483,
|
| 3978 |
+
"grad_norm": 9.553668975830078,
|
| 3979 |
+
"learning_rate": 6.86444782080132e-05,
|
| 3980 |
+
"loss": 2.1244,
|
| 3981 |
+
"step": 564
|
| 3982 |
+
},
|
| 3983 |
+
{
|
| 3984 |
+
"epoch": 0.6073636119322763,
|
| 3985 |
+
"grad_norm": 21.076509475708008,
|
| 3986 |
+
"learning_rate": 6.832075657972054e-05,
|
| 3987 |
+
"loss": 3.2957,
|
| 3988 |
+
"step": 565
|
| 3989 |
+
},
|
| 3990 |
+
{
|
| 3991 |
+
"epoch": 0.6084385917764042,
|
| 3992 |
+
"grad_norm": 11.683152198791504,
|
| 3993 |
+
"learning_rate": 6.799740355103239e-05,
|
| 3994 |
+
"loss": 2.4247,
|
| 3995 |
+
"step": 566
|
| 3996 |
+
},
|
| 3997 |
+
{
|
| 3998 |
+
"epoch": 0.6095135716205321,
|
| 3999 |
+
"grad_norm": 13.204034805297852,
|
| 4000 |
+
"learning_rate": 6.76744228842797e-05,
|
| 4001 |
+
"loss": 2.9528,
|
| 4002 |
+
"step": 567
|
| 4003 |
+
},
|
| 4004 |
+
{
|
| 4005 |
+
"epoch": 0.6105885514646601,
|
| 4006 |
+
"grad_norm": 14.521734237670898,
|
| 4007 |
+
"learning_rate": 6.735181833746086e-05,
|
| 4008 |
+
"loss": 2.5978,
|
| 4009 |
+
"step": 568
|
| 4010 |
+
},
|
| 4011 |
+
{
|
| 4012 |
+
"epoch": 0.6116635313087879,
|
| 4013 |
+
"grad_norm": 17.30651092529297,
|
| 4014 |
+
"learning_rate": 6.702959366419801e-05,
|
| 4015 |
+
"loss": 3.5166,
|
| 4016 |
+
"step": 569
|
| 4017 |
+
},
|
| 4018 |
+
{
|
| 4019 |
+
"epoch": 0.6127385111529159,
|
| 4020 |
+
"grad_norm": 15.7838716506958,
|
| 4021 |
+
"learning_rate": 6.670775261369325e-05,
|
| 4022 |
+
"loss": 2.6126,
|
| 4023 |
+
"step": 570
|
| 4024 |
+
},
|
| 4025 |
+
{
|
| 4026 |
+
"epoch": 0.6138134909970439,
|
| 4027 |
+
"grad_norm": 13.338095664978027,
|
| 4028 |
+
"learning_rate": 6.638629893068515e-05,
|
| 4029 |
+
"loss": 3.1597,
|
| 4030 |
+
"step": 571
|
| 4031 |
+
},
|
| 4032 |
+
{
|
| 4033 |
+
"epoch": 0.6148884708411717,
|
| 4034 |
+
"grad_norm": 16.482463836669922,
|
| 4035 |
+
"learning_rate": 6.60652363554051e-05,
|
| 4036 |
+
"loss": 2.4673,
|
| 4037 |
+
"step": 572
|
| 4038 |
+
},
|
| 4039 |
+
{
|
| 4040 |
+
"epoch": 0.6159634506852997,
|
| 4041 |
+
"grad_norm": 13.602096557617188,
|
| 4042 |
+
"learning_rate": 6.574456862353377e-05,
|
| 4043 |
+
"loss": 2.7441,
|
| 4044 |
+
"step": 573
|
| 4045 |
+
},
|
| 4046 |
+
{
|
| 4047 |
+
"epoch": 0.6170384305294275,
|
| 4048 |
+
"grad_norm": 15.007840156555176,
|
| 4049 |
+
"learning_rate": 6.542429946615774e-05,
|
| 4050 |
+
"loss": 2.7128,
|
| 4051 |
+
"step": 574
|
| 4052 |
+
},
|
| 4053 |
+
{
|
| 4054 |
+
"epoch": 0.6181134103735555,
|
| 4055 |
+
"grad_norm": 11.352252006530762,
|
| 4056 |
+
"learning_rate": 6.510443260972599e-05,
|
| 4057 |
+
"loss": 3.2629,
|
| 4058 |
+
"step": 575
|
| 4059 |
+
},
|
| 4060 |
+
{
|
| 4061 |
+
"epoch": 0.6191883902176835,
|
| 4062 |
+
"grad_norm": 14.53456974029541,
|
| 4063 |
+
"learning_rate": 6.47849717760066e-05,
|
| 4064 |
+
"loss": 1.8599,
|
| 4065 |
+
"step": 576
|
| 4066 |
+
},
|
| 4067 |
+
{
|
| 4068 |
+
"epoch": 0.6202633700618113,
|
| 4069 |
+
"grad_norm": 12.216171264648438,
|
| 4070 |
+
"learning_rate": 6.446592068204341e-05,
|
| 4071 |
+
"loss": 2.2911,
|
| 4072 |
+
"step": 577
|
| 4073 |
+
},
|
| 4074 |
+
{
|
| 4075 |
+
"epoch": 0.6213383499059393,
|
| 4076 |
+
"grad_norm": 15.579245567321777,
|
| 4077 |
+
"learning_rate": 6.41472830401128e-05,
|
| 4078 |
+
"loss": 2.4855,
|
| 4079 |
+
"step": 578
|
| 4080 |
+
},
|
| 4081 |
+
{
|
| 4082 |
+
"epoch": 0.6224133297500671,
|
| 4083 |
+
"grad_norm": 19.280820846557617,
|
| 4084 |
+
"learning_rate": 6.382906255768051e-05,
|
| 4085 |
+
"loss": 4.8336,
|
| 4086 |
+
"step": 579
|
| 4087 |
+
},
|
| 4088 |
+
{
|
| 4089 |
+
"epoch": 0.6234883095941951,
|
| 4090 |
+
"grad_norm": 19.063934326171875,
|
| 4091 |
+
"learning_rate": 6.351126293735843e-05,
|
| 4092 |
+
"loss": 2.7687,
|
| 4093 |
+
"step": 580
|
| 4094 |
+
},
|
| 4095 |
+
{
|
| 4096 |
+
"epoch": 0.6245632894383231,
|
| 4097 |
+
"grad_norm": 13.213850975036621,
|
| 4098 |
+
"learning_rate": 6.319388787686158e-05,
|
| 4099 |
+
"loss": 3.1479,
|
| 4100 |
+
"step": 581
|
| 4101 |
+
},
|
| 4102 |
+
{
|
| 4103 |
+
"epoch": 0.6256382692824509,
|
| 4104 |
+
"grad_norm": 19.5723934173584,
|
| 4105 |
+
"learning_rate": 6.287694106896509e-05,
|
| 4106 |
+
"loss": 4.7255,
|
| 4107 |
+
"step": 582
|
| 4108 |
+
},
|
| 4109 |
+
{
|
| 4110 |
+
"epoch": 0.6267132491265789,
|
| 4111 |
+
"grad_norm": 12.059754371643066,
|
| 4112 |
+
"learning_rate": 6.256042620146119e-05,
|
| 4113 |
+
"loss": 2.5616,
|
| 4114 |
+
"step": 583
|
| 4115 |
+
},
|
| 4116 |
+
{
|
| 4117 |
+
"epoch": 0.6277882289707069,
|
| 4118 |
+
"grad_norm": 10.33979320526123,
|
| 4119 |
+
"learning_rate": 6.224434695711631e-05,
|
| 4120 |
+
"loss": 1.6791,
|
| 4121 |
+
"step": 584
|
| 4122 |
+
},
|
| 4123 |
+
{
|
| 4124 |
+
"epoch": 0.6288632088148347,
|
| 4125 |
+
"grad_norm": 12.472517967224121,
|
| 4126 |
+
"learning_rate": 6.19287070136283e-05,
|
| 4127 |
+
"loss": 2.8273,
|
| 4128 |
+
"step": 585
|
| 4129 |
+
},
|
| 4130 |
+
{
|
| 4131 |
+
"epoch": 0.6299381886589627,
|
| 4132 |
+
"grad_norm": 13.145999908447266,
|
| 4133 |
+
"learning_rate": 6.16135100435836e-05,
|
| 4134 |
+
"loss": 2.4934,
|
| 4135 |
+
"step": 586
|
| 4136 |
+
},
|
| 4137 |
+
{
|
| 4138 |
+
"epoch": 0.6310131685030905,
|
| 4139 |
+
"grad_norm": 12.09467887878418,
|
| 4140 |
+
"learning_rate": 6.129875971441434e-05,
|
| 4141 |
+
"loss": 2.5874,
|
| 4142 |
+
"step": 587
|
| 4143 |
+
},
|
| 4144 |
+
{
|
| 4145 |
+
"epoch": 0.6320881483472185,
|
| 4146 |
+
"grad_norm": 12.754231452941895,
|
| 4147 |
+
"learning_rate": 6.0984459688356e-05,
|
| 4148 |
+
"loss": 3.2328,
|
| 4149 |
+
"step": 588
|
| 4150 |
+
},
|
| 4151 |
+
{
|
| 4152 |
+
"epoch": 0.6331631281913465,
|
| 4153 |
+
"grad_norm": 20.645830154418945,
|
| 4154 |
+
"learning_rate": 6.0670613622404496e-05,
|
| 4155 |
+
"loss": 3.202,
|
| 4156 |
+
"step": 589
|
| 4157 |
+
},
|
| 4158 |
+
{
|
| 4159 |
+
"epoch": 0.6342381080354743,
|
| 4160 |
+
"grad_norm": 13.244977951049805,
|
| 4161 |
+
"learning_rate": 6.035722516827382e-05,
|
| 4162 |
+
"loss": 2.1665,
|
| 4163 |
+
"step": 590
|
| 4164 |
+
},
|
| 4165 |
+
{
|
| 4166 |
+
"epoch": 0.6353130878796023,
|
| 4167 |
+
"grad_norm": 17.669931411743164,
|
| 4168 |
+
"learning_rate": 6.004429797235349e-05,
|
| 4169 |
+
"loss": 2.6613,
|
| 4170 |
+
"step": 591
|
| 4171 |
+
},
|
| 4172 |
+
{
|
| 4173 |
+
"epoch": 0.6363880677237301,
|
| 4174 |
+
"grad_norm": 10.736635208129883,
|
| 4175 |
+
"learning_rate": 5.973183567566605e-05,
|
| 4176 |
+
"loss": 2.4063,
|
| 4177 |
+
"step": 592
|
| 4178 |
+
},
|
| 4179 |
+
{
|
| 4180 |
+
"epoch": 0.6374630475678581,
|
| 4181 |
+
"grad_norm": 18.680782318115234,
|
| 4182 |
+
"learning_rate": 5.9419841913824824e-05,
|
| 4183 |
+
"loss": 2.6796,
|
| 4184 |
+
"step": 593
|
| 4185 |
+
},
|
| 4186 |
+
{
|
| 4187 |
+
"epoch": 0.6385380274119861,
|
| 4188 |
+
"grad_norm": 19.364147186279297,
|
| 4189 |
+
"learning_rate": 5.9108320316991536e-05,
|
| 4190 |
+
"loss": 3.8844,
|
| 4191 |
+
"step": 594
|
| 4192 |
+
},
|
| 4193 |
+
{
|
| 4194 |
+
"epoch": 0.6396130072561139,
|
| 4195 |
+
"grad_norm": 13.238618850708008,
|
| 4196 |
+
"learning_rate": 5.879727450983412e-05,
|
| 4197 |
+
"loss": 3.3821,
|
| 4198 |
+
"step": 595
|
| 4199 |
+
},
|
| 4200 |
+
{
|
| 4201 |
+
"epoch": 0.6406879871002419,
|
| 4202 |
+
"grad_norm": 21.127750396728516,
|
| 4203 |
+
"learning_rate": 5.848670811148451e-05,
|
| 4204 |
+
"loss": 3.8302,
|
| 4205 |
+
"step": 596
|
| 4206 |
+
},
|
| 4207 |
+
{
|
| 4208 |
+
"epoch": 0.6417629669443697,
|
| 4209 |
+
"grad_norm": 19.388427734375,
|
| 4210 |
+
"learning_rate": 5.817662473549651e-05,
|
| 4211 |
+
"loss": 2.4551,
|
| 4212 |
+
"step": 597
|
| 4213 |
+
},
|
| 4214 |
+
{
|
| 4215 |
+
"epoch": 0.6428379467884977,
|
| 4216 |
+
"grad_norm": 13.296785354614258,
|
| 4217 |
+
"learning_rate": 5.786702798980388e-05,
|
| 4218 |
+
"loss": 2.7313,
|
| 4219 |
+
"step": 598
|
| 4220 |
+
},
|
| 4221 |
+
{
|
| 4222 |
+
"epoch": 0.6439129266326257,
|
| 4223 |
+
"grad_norm": 13.282078742980957,
|
| 4224 |
+
"learning_rate": 5.755792147667811e-05,
|
| 4225 |
+
"loss": 2.6865,
|
| 4226 |
+
"step": 599
|
| 4227 |
+
},
|
| 4228 |
+
{
|
| 4229 |
+
"epoch": 0.6449879064767535,
|
| 4230 |
+
"grad_norm": 16.131546020507812,
|
| 4231 |
+
"learning_rate": 5.7249308792686815e-05,
|
| 4232 |
+
"loss": 2.9787,
|
| 4233 |
+
"step": 600
|
| 4234 |
+
},
|
| 4235 |
+
{
|
| 4236 |
+
"epoch": 0.6460628863208815,
|
| 4237 |
+
"grad_norm": 13.292814254760742,
|
| 4238 |
+
"learning_rate": 5.6941193528651596e-05,
|
| 4239 |
+
"loss": 2.7872,
|
| 4240 |
+
"step": 601
|
| 4241 |
+
},
|
| 4242 |
+
{
|
| 4243 |
+
"epoch": 0.6471378661650095,
|
| 4244 |
+
"grad_norm": 13.973624229431152,
|
| 4245 |
+
"learning_rate": 5.663357926960644e-05,
|
| 4246 |
+
"loss": 2.6566,
|
| 4247 |
+
"step": 602
|
| 4248 |
+
},
|
| 4249 |
+
{
|
| 4250 |
+
"epoch": 0.6482128460091373,
|
| 4251 |
+
"grad_norm": 14.040596961975098,
|
| 4252 |
+
"learning_rate": 5.6326469594756034e-05,
|
| 4253 |
+
"loss": 3.0928,
|
| 4254 |
+
"step": 603
|
| 4255 |
+
},
|
| 4256 |
+
{
|
| 4257 |
+
"epoch": 0.6492878258532653,
|
| 4258 |
+
"grad_norm": 17.149208068847656,
|
| 4259 |
+
"learning_rate": 5.6019868077433876e-05,
|
| 4260 |
+
"loss": 3.2098,
|
| 4261 |
+
"step": 604
|
| 4262 |
+
},
|
| 4263 |
+
{
|
| 4264 |
+
"epoch": 0.6503628056973931,
|
| 4265 |
+
"grad_norm": 14.205854415893555,
|
| 4266 |
+
"learning_rate": 5.5713778285061046e-05,
|
| 4267 |
+
"loss": 2.1934,
|
| 4268 |
+
"step": 605
|
| 4269 |
+
},
|
| 4270 |
+
{
|
| 4271 |
+
"epoch": 0.6514377855415211,
|
| 4272 |
+
"grad_norm": 20.856319427490234,
|
| 4273 |
+
"learning_rate": 5.540820377910435e-05,
|
| 4274 |
+
"loss": 4.2625,
|
| 4275 |
+
"step": 606
|
| 4276 |
+
},
|
| 4277 |
+
{
|
| 4278 |
+
"epoch": 0.6525127653856491,
|
| 4279 |
+
"grad_norm": 19.353553771972656,
|
| 4280 |
+
"learning_rate": 5.5103148115035195e-05,
|
| 4281 |
+
"loss": 3.3242,
|
| 4282 |
+
"step": 607
|
| 4283 |
+
},
|
| 4284 |
+
{
|
| 4285 |
+
"epoch": 0.6535877452297769,
|
| 4286 |
+
"grad_norm": 13.190366744995117,
|
| 4287 |
+
"learning_rate": 5.479861484228794e-05,
|
| 4288 |
+
"loss": 2.3837,
|
| 4289 |
+
"step": 608
|
| 4290 |
+
},
|
| 4291 |
+
{
|
| 4292 |
+
"epoch": 0.6546627250739049,
|
| 4293 |
+
"grad_norm": 16.302879333496094,
|
| 4294 |
+
"learning_rate": 5.449460750421883e-05,
|
| 4295 |
+
"loss": 3.2505,
|
| 4296 |
+
"step": 609
|
| 4297 |
+
},
|
| 4298 |
+
{
|
| 4299 |
+
"epoch": 0.6557377049180327,
|
| 4300 |
+
"grad_norm": 13.280342102050781,
|
| 4301 |
+
"learning_rate": 5.419112963806468e-05,
|
| 4302 |
+
"loss": 1.8674,
|
| 4303 |
+
"step": 610
|
| 4304 |
+
},
|
| 4305 |
+
{
|
| 4306 |
+
"epoch": 0.6568126847621607,
|
| 4307 |
+
"grad_norm": 21.006906509399414,
|
| 4308 |
+
"learning_rate": 5.388818477490154e-05,
|
| 4309 |
+
"loss": 3.6557,
|
| 4310 |
+
"step": 611
|
| 4311 |
+
},
|
| 4312 |
+
{
|
| 4313 |
+
"epoch": 0.6578876646062887,
|
| 4314 |
+
"grad_norm": 19.307729721069336,
|
| 4315 |
+
"learning_rate": 5.358577643960403e-05,
|
| 4316 |
+
"loss": 2.2382,
|
| 4317 |
+
"step": 612
|
| 4318 |
+
},
|
| 4319 |
+
{
|
| 4320 |
+
"epoch": 0.6589626444504165,
|
| 4321 |
+
"grad_norm": 11.849048614501953,
|
| 4322 |
+
"learning_rate": 5.328390815080381e-05,
|
| 4323 |
+
"loss": 2.0229,
|
| 4324 |
+
"step": 613
|
| 4325 |
+
},
|
| 4326 |
+
{
|
| 4327 |
+
"epoch": 0.6600376242945445,
|
| 4328 |
+
"grad_norm": 11.787457466125488,
|
| 4329 |
+
"learning_rate": 5.2982583420849116e-05,
|
| 4330 |
+
"loss": 2.6637,
|
| 4331 |
+
"step": 614
|
| 4332 |
+
},
|
| 4333 |
+
{
|
| 4334 |
+
"epoch": 0.6611126041386725,
|
| 4335 |
+
"grad_norm": 15.777995109558105,
|
| 4336 |
+
"learning_rate": 5.268180575576352e-05,
|
| 4337 |
+
"loss": 2.893,
|
| 4338 |
+
"step": 615
|
| 4339 |
+
},
|
| 4340 |
+
{
|
| 4341 |
+
"epoch": 0.6621875839828003,
|
| 4342 |
+
"grad_norm": 19.96381378173828,
|
| 4343 |
+
"learning_rate": 5.238157865520539e-05,
|
| 4344 |
+
"loss": 3.2706,
|
| 4345 |
+
"step": 616
|
| 4346 |
+
},
|
| 4347 |
+
{
|
| 4348 |
+
"epoch": 0.6632625638269283,
|
| 4349 |
+
"grad_norm": 20.699359893798828,
|
| 4350 |
+
"learning_rate": 5.208190561242708e-05,
|
| 4351 |
+
"loss": 2.6676,
|
| 4352 |
+
"step": 617
|
| 4353 |
+
},
|
| 4354 |
+
{
|
| 4355 |
+
"epoch": 0.6643375436710561,
|
| 4356 |
+
"grad_norm": 14.418145179748535,
|
| 4357 |
+
"learning_rate": 5.178279011423417e-05,
|
| 4358 |
+
"loss": 2.6929,
|
| 4359 |
+
"step": 618
|
| 4360 |
+
},
|
| 4361 |
+
{
|
| 4362 |
+
"epoch": 0.6654125235151841,
|
| 4363 |
+
"grad_norm": 12.969032287597656,
|
| 4364 |
+
"learning_rate": 5.148423564094517e-05,
|
| 4365 |
+
"loss": 2.5543,
|
| 4366 |
+
"step": 619
|
| 4367 |
+
},
|
| 4368 |
+
{
|
| 4369 |
+
"epoch": 0.6664875033593121,
|
| 4370 |
+
"grad_norm": 25.956907272338867,
|
| 4371 |
+
"learning_rate": 5.118624566635066e-05,
|
| 4372 |
+
"loss": 3.7616,
|
| 4373 |
+
"step": 620
|
| 4374 |
+
},
|
| 4375 |
+
{
|
| 4376 |
+
"epoch": 0.6675624832034399,
|
| 4377 |
+
"grad_norm": 16.259069442749023,
|
| 4378 |
+
"learning_rate": 5.0888823657673266e-05,
|
| 4379 |
+
"loss": 2.6596,
|
| 4380 |
+
"step": 621
|
| 4381 |
+
},
|
| 4382 |
+
{
|
| 4383 |
+
"epoch": 0.6686374630475679,
|
| 4384 |
+
"grad_norm": 15.737533569335938,
|
| 4385 |
+
"learning_rate": 5.059197307552698e-05,
|
| 4386 |
+
"loss": 3.4037,
|
| 4387 |
+
"step": 622
|
| 4388 |
+
},
|
| 4389 |
+
{
|
| 4390 |
+
"epoch": 0.6697124428916957,
|
| 4391 |
+
"grad_norm": 15.57932186126709,
|
| 4392 |
+
"learning_rate": 5.0295697373877096e-05,
|
| 4393 |
+
"loss": 3.3037,
|
| 4394 |
+
"step": 623
|
| 4395 |
+
},
|
| 4396 |
+
{
|
| 4397 |
+
"epoch": 0.6707874227358237,
|
| 4398 |
+
"grad_norm": 16.854917526245117,
|
| 4399 |
+
"learning_rate": 5.000000000000002e-05,
|
| 4400 |
+
"loss": 2.7998,
|
| 4401 |
+
"step": 624
|
| 4402 |
+
},
|
| 4403 |
+
{
|
| 4404 |
+
"epoch": 0.6718624025799517,
|
| 4405 |
+
"grad_norm": 20.23788070678711,
|
| 4406 |
+
"learning_rate": 4.9704884394442964e-05,
|
| 4407 |
+
"loss": 2.6882,
|
| 4408 |
+
"step": 625
|
| 4409 |
+
},
|
| 4410 |
+
{
|
| 4411 |
+
"epoch": 0.6729373824240795,
|
| 4412 |
+
"grad_norm": 13.206673622131348,
|
| 4413 |
+
"learning_rate": 4.941035399098418e-05,
|
| 4414 |
+
"loss": 2.8392,
|
| 4415 |
+
"step": 626
|
| 4416 |
+
},
|
| 4417 |
+
{
|
| 4418 |
+
"epoch": 0.6740123622682075,
|
| 4419 |
+
"grad_norm": 11.313912391662598,
|
| 4420 |
+
"learning_rate": 4.911641221659279e-05,
|
| 4421 |
+
"loss": 2.6687,
|
| 4422 |
+
"step": 627
|
| 4423 |
+
},
|
| 4424 |
+
{
|
| 4425 |
+
"epoch": 0.6750873421123353,
|
| 4426 |
+
"grad_norm": 17.33076286315918,
|
| 4427 |
+
"learning_rate": 4.8823062491389094e-05,
|
| 4428 |
+
"loss": 2.5485,
|
| 4429 |
+
"step": 628
|
| 4430 |
+
},
|
| 4431 |
+
{
|
| 4432 |
+
"epoch": 0.6761623219564633,
|
| 4433 |
+
"grad_norm": 17.54669952392578,
|
| 4434 |
+
"learning_rate": 4.853030822860455e-05,
|
| 4435 |
+
"loss": 3.4131,
|
| 4436 |
+
"step": 629
|
| 4437 |
+
},
|
| 4438 |
+
{
|
| 4439 |
+
"epoch": 0.6772373018005913,
|
| 4440 |
+
"grad_norm": 14.568754196166992,
|
| 4441 |
+
"learning_rate": 4.823815283454235e-05,
|
| 4442 |
+
"loss": 3.2074,
|
| 4443 |
+
"step": 630
|
| 4444 |
+
},
|
| 4445 |
+
{
|
| 4446 |
+
"epoch": 0.6783122816447191,
|
| 4447 |
+
"grad_norm": 14.021873474121094,
|
| 4448 |
+
"learning_rate": 4.794659970853749e-05,
|
| 4449 |
+
"loss": 2.3931,
|
| 4450 |
+
"step": 631
|
| 4451 |
+
},
|
| 4452 |
+
{
|
| 4453 |
+
"epoch": 0.6793872614888471,
|
| 4454 |
+
"grad_norm": 15.752306938171387,
|
| 4455 |
+
"learning_rate": 4.765565224291743e-05,
|
| 4456 |
+
"loss": 2.4038,
|
| 4457 |
+
"step": 632
|
| 4458 |
+
},
|
| 4459 |
+
{
|
| 4460 |
+
"epoch": 0.6804622413329751,
|
| 4461 |
+
"grad_norm": 17.143024444580078,
|
| 4462 |
+
"learning_rate": 4.7365313822962576e-05,
|
| 4463 |
+
"loss": 2.8179,
|
| 4464 |
+
"step": 633
|
| 4465 |
+
},
|
| 4466 |
+
{
|
| 4467 |
+
"epoch": 0.6815372211771029,
|
| 4468 |
+
"grad_norm": 18.36078643798828,
|
| 4469 |
+
"learning_rate": 4.707558782686677e-05,
|
| 4470 |
+
"loss": 2.4392,
|
| 4471 |
+
"step": 634
|
| 4472 |
+
},
|
| 4473 |
+
{
|
| 4474 |
+
"epoch": 0.6826122010212309,
|
| 4475 |
+
"grad_norm": 17.252010345458984,
|
| 4476 |
+
"learning_rate": 4.67864776256982e-05,
|
| 4477 |
+
"loss": 2.6834,
|
| 4478 |
+
"step": 635
|
| 4479 |
+
},
|
| 4480 |
+
{
|
| 4481 |
+
"epoch": 0.6836871808653587,
|
| 4482 |
+
"grad_norm": 25.259824752807617,
|
| 4483 |
+
"learning_rate": 4.64979865833599e-05,
|
| 4484 |
+
"loss": 3.477,
|
| 4485 |
+
"step": 636
|
| 4486 |
+
},
|
| 4487 |
+
{
|
| 4488 |
+
"epoch": 0.6847621607094867,
|
| 4489 |
+
"grad_norm": 14.85805606842041,
|
| 4490 |
+
"learning_rate": 4.621011805655093e-05,
|
| 4491 |
+
"loss": 2.3259,
|
| 4492 |
+
"step": 637
|
| 4493 |
+
},
|
| 4494 |
+
{
|
| 4495 |
+
"epoch": 0.6858371405536147,
|
| 4496 |
+
"grad_norm": 14.022159576416016,
|
| 4497 |
+
"learning_rate": 4.592287539472701e-05,
|
| 4498 |
+
"loss": 2.6373,
|
| 4499 |
+
"step": 638
|
| 4500 |
+
},
|
| 4501 |
+
{
|
| 4502 |
+
"epoch": 0.6869121203977425,
|
| 4503 |
+
"grad_norm": 13.847736358642578,
|
| 4504 |
+
"learning_rate": 4.563626194006178e-05,
|
| 4505 |
+
"loss": 2.5128,
|
| 4506 |
+
"step": 639
|
| 4507 |
+
},
|
| 4508 |
+
{
|
| 4509 |
+
"epoch": 0.6879871002418705,
|
| 4510 |
+
"grad_norm": 12.617048263549805,
|
| 4511 |
+
"learning_rate": 4.535028102740785e-05,
|
| 4512 |
+
"loss": 2.6718,
|
| 4513 |
+
"step": 640
|
| 4514 |
+
},
|
| 4515 |
+
{
|
| 4516 |
+
"epoch": 0.6890620800859983,
|
| 4517 |
+
"grad_norm": 15.006109237670898,
|
| 4518 |
+
"learning_rate": 4.5064935984257826e-05,
|
| 4519 |
+
"loss": 3.0097,
|
| 4520 |
+
"step": 641
|
| 4521 |
+
},
|
| 4522 |
+
{
|
| 4523 |
+
"epoch": 0.6901370599301263,
|
| 4524 |
+
"grad_norm": 18.60867691040039,
|
| 4525 |
+
"learning_rate": 4.478023013070595e-05,
|
| 4526 |
+
"loss": 3.4391,
|
| 4527 |
+
"step": 642
|
| 4528 |
+
},
|
| 4529 |
+
{
|
| 4530 |
+
"epoch": 0.6912120397742543,
|
| 4531 |
+
"grad_norm": 18.005126953125,
|
| 4532 |
+
"learning_rate": 4.449616677940903e-05,
|
| 4533 |
+
"loss": 2.8679,
|
| 4534 |
+
"step": 643
|
| 4535 |
+
},
|
| 4536 |
+
{
|
| 4537 |
+
"epoch": 0.6922870196183821,
|
| 4538 |
+
"grad_norm": 15.206576347351074,
|
| 4539 |
+
"learning_rate": 4.421274923554835e-05,
|
| 4540 |
+
"loss": 2.5725,
|
| 4541 |
+
"step": 644
|
| 4542 |
+
},
|
| 4543 |
+
{
|
| 4544 |
+
"epoch": 0.6933619994625101,
|
| 4545 |
+
"grad_norm": 10.55386734008789,
|
| 4546 |
+
"learning_rate": 4.392998079679076e-05,
|
| 4547 |
+
"loss": 2.5634,
|
| 4548 |
+
"step": 645
|
| 4549 |
+
},
|
| 4550 |
+
{
|
| 4551 |
+
"epoch": 0.694436979306638,
|
| 4552 |
+
"grad_norm": 17.997459411621094,
|
| 4553 |
+
"learning_rate": 4.364786475325072e-05,
|
| 4554 |
+
"loss": 4.0269,
|
| 4555 |
+
"step": 646
|
| 4556 |
+
},
|
| 4557 |
+
{
|
| 4558 |
+
"epoch": 0.6955119591507659,
|
| 4559 |
+
"grad_norm": 17.598691940307617,
|
| 4560 |
+
"learning_rate": 4.33664043874518e-05,
|
| 4561 |
+
"loss": 3.0651,
|
| 4562 |
+
"step": 647
|
| 4563 |
+
},
|
| 4564 |
+
{
|
| 4565 |
+
"epoch": 0.6965869389948939,
|
| 4566 |
+
"grad_norm": 19.84808349609375,
|
| 4567 |
+
"learning_rate": 4.30856029742884e-05,
|
| 4568 |
+
"loss": 3.4776,
|
| 4569 |
+
"step": 648
|
| 4570 |
+
},
|
| 4571 |
+
{
|
| 4572 |
+
"epoch": 0.6976619188390217,
|
| 4573 |
+
"grad_norm": 19.29751968383789,
|
| 4574 |
+
"learning_rate": 4.280546378098792e-05,
|
| 4575 |
+
"loss": 3.4422,
|
| 4576 |
+
"step": 649
|
| 4577 |
+
},
|
| 4578 |
+
{
|
| 4579 |
+
"epoch": 0.6987368986831497,
|
| 4580 |
+
"grad_norm": 19.02079963684082,
|
| 4581 |
+
"learning_rate": 4.252599006707245e-05,
|
| 4582 |
+
"loss": 2.9825,
|
| 4583 |
+
"step": 650
|
| 4584 |
+
},
|
| 4585 |
+
{
|
| 4586 |
+
"epoch": 0.6998118785272777,
|
| 4587 |
+
"grad_norm": 17.91227912902832,
|
| 4588 |
+
"learning_rate": 4.224718508432113e-05,
|
| 4589 |
+
"loss": 2.7378,
|
| 4590 |
+
"step": 651
|
| 4591 |
+
},
|
| 4592 |
+
{
|
| 4593 |
+
"epoch": 0.7008868583714055,
|
| 4594 |
+
"grad_norm": 20.459753036499023,
|
| 4595 |
+
"learning_rate": 4.196905207673201e-05,
|
| 4596 |
+
"loss": 3.7669,
|
| 4597 |
+
"step": 652
|
| 4598 |
+
},
|
| 4599 |
+
{
|
| 4600 |
+
"epoch": 0.7019618382155335,
|
| 4601 |
+
"grad_norm": 12.434891700744629,
|
| 4602 |
+
"learning_rate": 4.16915942804846e-05,
|
| 4603 |
+
"loss": 2.6119,
|
| 4604 |
+
"step": 653
|
| 4605 |
+
},
|
| 4606 |
+
{
|
| 4607 |
+
"epoch": 0.7030368180596613,
|
| 4608 |
+
"grad_norm": 17.53223419189453,
|
| 4609 |
+
"learning_rate": 4.141481492390197e-05,
|
| 4610 |
+
"loss": 2.7561,
|
| 4611 |
+
"step": 654
|
| 4612 |
+
},
|
| 4613 |
+
{
|
| 4614 |
+
"epoch": 0.7041117979037893,
|
| 4615 |
+
"grad_norm": 15.005716323852539,
|
| 4616 |
+
"learning_rate": 4.113871722741337e-05,
|
| 4617 |
+
"loss": 2.5018,
|
| 4618 |
+
"step": 655
|
| 4619 |
+
},
|
| 4620 |
+
{
|
| 4621 |
+
"epoch": 0.7051867777479173,
|
| 4622 |
+
"grad_norm": 19.277578353881836,
|
| 4623 |
+
"learning_rate": 4.08633044035167e-05,
|
| 4624 |
+
"loss": 3.2063,
|
| 4625 |
+
"step": 656
|
| 4626 |
+
},
|
| 4627 |
+
{
|
| 4628 |
+
"epoch": 0.7062617575920451,
|
| 4629 |
+
"grad_norm": 18.722848892211914,
|
| 4630 |
+
"learning_rate": 4.058857965674101e-05,
|
| 4631 |
+
"loss": 2.4138,
|
| 4632 |
+
"step": 657
|
| 4633 |
+
},
|
| 4634 |
+
{
|
| 4635 |
+
"epoch": 0.7073367374361731,
|
| 4636 |
+
"grad_norm": 12.36088752746582,
|
| 4637 |
+
"learning_rate": 4.031454618360945e-05,
|
| 4638 |
+
"loss": 2.4916,
|
| 4639 |
+
"step": 658
|
| 4640 |
+
},
|
| 4641 |
+
{
|
| 4642 |
+
"epoch": 0.708411717280301,
|
| 4643 |
+
"grad_norm": 18.981199264526367,
|
| 4644 |
+
"learning_rate": 4.0041207172601826e-05,
|
| 4645 |
+
"loss": 3.578,
|
| 4646 |
+
"step": 659
|
| 4647 |
+
},
|
| 4648 |
+
{
|
| 4649 |
+
"epoch": 0.7094866971244289,
|
| 4650 |
+
"grad_norm": 15.9765625,
|
| 4651 |
+
"learning_rate": 3.976856580411774e-05,
|
| 4652 |
+
"loss": 2.896,
|
| 4653 |
+
"step": 660
|
| 4654 |
+
},
|
| 4655 |
+
{
|
| 4656 |
+
"epoch": 0.7105616769685569,
|
| 4657 |
+
"grad_norm": 10.438156127929688,
|
| 4658 |
+
"learning_rate": 3.9496625250439344e-05,
|
| 4659 |
+
"loss": 2.5757,
|
| 4660 |
+
"step": 661
|
| 4661 |
+
},
|
| 4662 |
+
{
|
| 4663 |
+
"epoch": 0.7116366568126847,
|
| 4664 |
+
"grad_norm": 15.564522743225098,
|
| 4665 |
+
"learning_rate": 3.922538867569466e-05,
|
| 4666 |
+
"loss": 2.7255,
|
| 4667 |
+
"step": 662
|
| 4668 |
+
},
|
| 4669 |
+
{
|
| 4670 |
+
"epoch": 0.7127116366568127,
|
| 4671 |
+
"grad_norm": 10.465269088745117,
|
| 4672 |
+
"learning_rate": 3.8954859235820664e-05,
|
| 4673 |
+
"loss": 1.888,
|
| 4674 |
+
"step": 663
|
| 4675 |
+
},
|
| 4676 |
+
{
|
| 4677 |
+
"epoch": 0.7137866165009407,
|
| 4678 |
+
"grad_norm": 10.99638843536377,
|
| 4679 |
+
"learning_rate": 3.8685040078526415e-05,
|
| 4680 |
+
"loss": 2.6292,
|
| 4681 |
+
"step": 664
|
| 4682 |
+
},
|
| 4683 |
+
{
|
| 4684 |
+
"epoch": 0.7148615963450685,
|
| 4685 |
+
"grad_norm": 14.457728385925293,
|
| 4686 |
+
"learning_rate": 3.841593434325675e-05,
|
| 4687 |
+
"loss": 2.3995,
|
| 4688 |
+
"step": 665
|
| 4689 |
+
},
|
| 4690 |
+
{
|
| 4691 |
+
"epoch": 0.7159365761891965,
|
| 4692 |
+
"grad_norm": 17.016172409057617,
|
| 4693 |
+
"learning_rate": 3.814754516115544e-05,
|
| 4694 |
+
"loss": 2.8181,
|
| 4695 |
+
"step": 666
|
| 4696 |
+
},
|
| 4697 |
+
{
|
| 4698 |
+
"epoch": 0.7170115560333243,
|
| 4699 |
+
"grad_norm": 18.437854766845703,
|
| 4700 |
+
"learning_rate": 3.787987565502902e-05,
|
| 4701 |
+
"loss": 2.1409,
|
| 4702 |
+
"step": 667
|
| 4703 |
+
},
|
| 4704 |
+
{
|
| 4705 |
+
"epoch": 0.7180865358774523,
|
| 4706 |
+
"grad_norm": 19.192140579223633,
|
| 4707 |
+
"learning_rate": 3.761292893931019e-05,
|
| 4708 |
+
"loss": 3.0362,
|
| 4709 |
+
"step": 668
|
| 4710 |
+
},
|
| 4711 |
+
{
|
| 4712 |
+
"epoch": 0.7191615157215803,
|
| 4713 |
+
"grad_norm": 18.337635040283203,
|
| 4714 |
+
"learning_rate": 3.734670812002183e-05,
|
| 4715 |
+
"loss": 3.1711,
|
| 4716 |
+
"step": 669
|
| 4717 |
+
},
|
| 4718 |
+
{
|
| 4719 |
+
"epoch": 0.7202364955657081,
|
| 4720 |
+
"grad_norm": 12.289979934692383,
|
| 4721 |
+
"learning_rate": 3.708121629474077e-05,
|
| 4722 |
+
"loss": 2.4776,
|
| 4723 |
+
"step": 670
|
| 4724 |
+
},
|
| 4725 |
+
{
|
| 4726 |
+
"epoch": 0.7213114754098361,
|
| 4727 |
+
"grad_norm": 16.519603729248047,
|
| 4728 |
+
"learning_rate": 3.681645655256159e-05,
|
| 4729 |
+
"loss": 2.5979,
|
| 4730 |
+
"step": 671
|
| 4731 |
+
},
|
| 4732 |
+
{
|
| 4733 |
+
"epoch": 0.7223864552539639,
|
| 4734 |
+
"grad_norm": 16.15471076965332,
|
| 4735 |
+
"learning_rate": 3.655243197406097e-05,
|
| 4736 |
+
"loss": 3.1705,
|
| 4737 |
+
"step": 672
|
| 4738 |
+
},
|
| 4739 |
+
{
|
| 4740 |
+
"epoch": 0.7234614350980919,
|
| 4741 |
+
"grad_norm": 14.030595779418945,
|
| 4742 |
+
"learning_rate": 3.628914563126156e-05,
|
| 4743 |
+
"loss": 2.5392,
|
| 4744 |
+
"step": 673
|
| 4745 |
+
},
|
| 4746 |
+
{
|
| 4747 |
+
"epoch": 0.7245364149422199,
|
| 4748 |
+
"grad_norm": 14.324786186218262,
|
| 4749 |
+
"learning_rate": 3.6026600587596484e-05,
|
| 4750 |
+
"loss": 3.4405,
|
| 4751 |
+
"step": 674
|
| 4752 |
+
},
|
| 4753 |
+
{
|
| 4754 |
+
"epoch": 0.7256113947863477,
|
| 4755 |
+
"grad_norm": 14.356512069702148,
|
| 4756 |
+
"learning_rate": 3.576479989787345e-05,
|
| 4757 |
+
"loss": 2.7351,
|
| 4758 |
+
"step": 675
|
| 4759 |
+
},
|
| 4760 |
+
{
|
| 4761 |
+
"epoch": 0.7266863746304757,
|
| 4762 |
+
"grad_norm": 22.63331413269043,
|
| 4763 |
+
"learning_rate": 3.550374660823949e-05,
|
| 4764 |
+
"loss": 4.4579,
|
| 4765 |
+
"step": 676
|
| 4766 |
+
},
|
| 4767 |
+
{
|
| 4768 |
+
"epoch": 0.7277613544746036,
|
| 4769 |
+
"grad_norm": 16.112449645996094,
|
| 4770 |
+
"learning_rate": 3.52434437561452e-05,
|
| 4771 |
+
"loss": 3.2883,
|
| 4772 |
+
"step": 677
|
| 4773 |
+
},
|
| 4774 |
+
{
|
| 4775 |
+
"epoch": 0.7288363343187315,
|
| 4776 |
+
"grad_norm": 15.933297157287598,
|
| 4777 |
+
"learning_rate": 3.4983894370309665e-05,
|
| 4778 |
+
"loss": 3.1271,
|
| 4779 |
+
"step": 678
|
| 4780 |
+
},
|
| 4781 |
+
{
|
| 4782 |
+
"epoch": 0.7299113141628595,
|
| 4783 |
+
"grad_norm": 18.30646514892578,
|
| 4784 |
+
"learning_rate": 3.472510147068515e-05,
|
| 4785 |
+
"loss": 3.1462,
|
| 4786 |
+
"step": 679
|
| 4787 |
+
},
|
| 4788 |
+
{
|
| 4789 |
+
"epoch": 0.7309862940069873,
|
| 4790 |
+
"grad_norm": 14.52500057220459,
|
| 4791 |
+
"learning_rate": 3.446706806842177e-05,
|
| 4792 |
+
"loss": 3.5975,
|
| 4793 |
+
"step": 680
|
| 4794 |
+
},
|
| 4795 |
+
{
|
| 4796 |
+
"epoch": 0.7320612738511153,
|
| 4797 |
+
"grad_norm": 19.05596160888672,
|
| 4798 |
+
"learning_rate": 3.420979716583279e-05,
|
| 4799 |
+
"loss": 3.3156,
|
| 4800 |
+
"step": 681
|
| 4801 |
+
},
|
| 4802 |
+
{
|
| 4803 |
+
"epoch": 0.7331362536952433,
|
| 4804 |
+
"grad_norm": 17.002887725830078,
|
| 4805 |
+
"learning_rate": 3.395329175635935e-05,
|
| 4806 |
+
"loss": 2.8059,
|
| 4807 |
+
"step": 682
|
| 4808 |
+
},
|
| 4809 |
+
{
|
| 4810 |
+
"epoch": 0.7342112335393711,
|
| 4811 |
+
"grad_norm": 13.947765350341797,
|
| 4812 |
+
"learning_rate": 3.369755482453594e-05,
|
| 4813 |
+
"loss": 2.3958,
|
| 4814 |
+
"step": 683
|
| 4815 |
+
},
|
| 4816 |
+
{
|
| 4817 |
+
"epoch": 0.7352862133834991,
|
| 4818 |
+
"grad_norm": 14.15211009979248,
|
| 4819 |
+
"learning_rate": 3.344258934595539e-05,
|
| 4820 |
+
"loss": 2.7893,
|
| 4821 |
+
"step": 684
|
| 4822 |
+
},
|
| 4823 |
+
{
|
| 4824 |
+
"epoch": 0.7363611932276269,
|
| 4825 |
+
"grad_norm": 16.130531311035156,
|
| 4826 |
+
"learning_rate": 3.31883982872345e-05,
|
| 4827 |
+
"loss": 2.4694,
|
| 4828 |
+
"step": 685
|
| 4829 |
+
},
|
| 4830 |
+
{
|
| 4831 |
+
"epoch": 0.7374361730717549,
|
| 4832 |
+
"grad_norm": 19.640533447265625,
|
| 4833 |
+
"learning_rate": 3.2934984605979424e-05,
|
| 4834 |
+
"loss": 3.6272,
|
| 4835 |
+
"step": 686
|
| 4836 |
+
},
|
| 4837 |
+
{
|
| 4838 |
+
"epoch": 0.7385111529158829,
|
| 4839 |
+
"grad_norm": 16.92746925354004,
|
| 4840 |
+
"learning_rate": 3.268235125075111e-05,
|
| 4841 |
+
"loss": 3.1489,
|
| 4842 |
+
"step": 687
|
| 4843 |
+
},
|
| 4844 |
+
{
|
| 4845 |
+
"epoch": 0.7395861327600107,
|
| 4846 |
+
"grad_norm": 13.478626251220703,
|
| 4847 |
+
"learning_rate": 3.243050116103128e-05,
|
| 4848 |
+
"loss": 4.0042,
|
| 4849 |
+
"step": 688
|
| 4850 |
+
},
|
| 4851 |
+
{
|
| 4852 |
+
"epoch": 0.7406611126041387,
|
| 4853 |
+
"grad_norm": 16.014270782470703,
|
| 4854 |
+
"learning_rate": 3.217943726718795e-05,
|
| 4855 |
+
"loss": 2.8463,
|
| 4856 |
+
"step": 689
|
| 4857 |
+
},
|
| 4858 |
+
{
|
| 4859 |
+
"epoch": 0.7417360924482665,
|
| 4860 |
+
"grad_norm": 10.819585800170898,
|
| 4861 |
+
"learning_rate": 3.1929162490441565e-05,
|
| 4862 |
+
"loss": 2.3828,
|
| 4863 |
+
"step": 690
|
| 4864 |
+
},
|
| 4865 |
+
{
|
| 4866 |
+
"epoch": 0.7428110722923945,
|
| 4867 |
+
"grad_norm": 13.190820693969727,
|
| 4868 |
+
"learning_rate": 3.16796797428308e-05,
|
| 4869 |
+
"loss": 2.5904,
|
| 4870 |
+
"step": 691
|
| 4871 |
+
},
|
| 4872 |
+
{
|
| 4873 |
+
"epoch": 0.7438860521365225,
|
| 4874 |
+
"grad_norm": 16.281742095947266,
|
| 4875 |
+
"learning_rate": 3.1430991927178866e-05,
|
| 4876 |
+
"loss": 2.2573,
|
| 4877 |
+
"step": 692
|
| 4878 |
+
},
|
| 4879 |
+
{
|
| 4880 |
+
"epoch": 0.7449610319806503,
|
| 4881 |
+
"grad_norm": 17.6481990814209,
|
| 4882 |
+
"learning_rate": 3.1183101937059647e-05,
|
| 4883 |
+
"loss": 2.8613,
|
| 4884 |
+
"step": 693
|
| 4885 |
+
},
|
| 4886 |
+
{
|
| 4887 |
+
"epoch": 0.7460360118247783,
|
| 4888 |
+
"grad_norm": 15.284998893737793,
|
| 4889 |
+
"learning_rate": 3.093601265676393e-05,
|
| 4890 |
+
"loss": 3.2382,
|
| 4891 |
+
"step": 694
|
| 4892 |
+
},
|
| 4893 |
+
{
|
| 4894 |
+
"epoch": 0.7471109916689062,
|
| 4895 |
+
"grad_norm": 15.303094863891602,
|
| 4896 |
+
"learning_rate": 3.068972696126611e-05,
|
| 4897 |
+
"loss": 2.2713,
|
| 4898 |
+
"step": 695
|
| 4899 |
+
},
|
| 4900 |
+
{
|
| 4901 |
+
"epoch": 0.7481859715130341,
|
| 4902 |
+
"grad_norm": 17.59407615661621,
|
| 4903 |
+
"learning_rate": 3.044424771619041e-05,
|
| 4904 |
+
"loss": 2.976,
|
| 4905 |
+
"step": 696
|
| 4906 |
+
},
|
| 4907 |
+
{
|
| 4908 |
+
"epoch": 0.7492609513571621,
|
| 4909 |
+
"grad_norm": 15.894051551818848,
|
| 4910 |
+
"learning_rate": 3.0199577777777875e-05,
|
| 4911 |
+
"loss": 3.3442,
|
| 4912 |
+
"step": 697
|
| 4913 |
+
},
|
| 4914 |
+
{
|
| 4915 |
+
"epoch": 0.7503359312012899,
|
| 4916 |
+
"grad_norm": 13.673702239990234,
|
| 4917 |
+
"learning_rate": 2.9955719992852804e-05,
|
| 4918 |
+
"loss": 2.5122,
|
| 4919 |
+
"step": 698
|
| 4920 |
+
},
|
| 4921 |
+
{
|
| 4922 |
+
"epoch": 0.7514109110454179,
|
| 4923 |
+
"grad_norm": 16.84321403503418,
|
| 4924 |
+
"learning_rate": 2.9712677198789916e-05,
|
| 4925 |
+
"loss": 2.9834,
|
| 4926 |
+
"step": 699
|
| 4927 |
+
},
|
| 4928 |
+
{
|
| 4929 |
+
"epoch": 0.7514109110454179,
|
| 4930 |
+
"eval_loss": 0.6793206930160522,
|
| 4931 |
+
"eval_runtime": 5.5979,
|
| 4932 |
+
"eval_samples_per_second": 70.026,
|
| 4933 |
+
"eval_steps_per_second": 35.013,
|
| 4934 |
+
"step": 699
|
| 4935 |
}
|
| 4936 |
],
|
| 4937 |
"logging_steps": 1,
|
|
|
|
| 4951 |
"attributes": {}
|
| 4952 |
}
|
| 4953 |
},
|
| 4954 |
+
"total_flos": 5272488565014528.0,
|
| 4955 |
"train_batch_size": 2,
|
| 4956 |
"trial_name": null,
|
| 4957 |
"trial_params": null
|