Training in progress, step 678, 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 59933632
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:ba19194e4a5b9086e60652ceb255afdf27cf1615639b7ec033a6ef1c49e3025b
|
| 3 |
size 59933632
|
last-checkpoint/optimizer.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 31823460
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:5cdf9379ee6afa80350d79bcd5263969bfa1069cfd31b296b13a8324eb8102e9
|
| 3 |
size 31823460
|
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 sha256:8d80092e98990ad16bd04c6bdc13ab21f8968fa024d4ffb0e439fb35359d90a4
|
| 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:e58a1f189fa49844e0a8b0fc017ed772705566337859dc459c55af78a8975c99
|
| 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": 500,
|
| 6 |
-
"global_step":
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
@@ -2380,6 +2380,2379 @@
|
|
| 2380 |
"learning_rate": 8.560166594400878e-05,
|
| 2381 |
"loss": 1.4472,
|
| 2382 |
"step": 339
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 2383 |
}
|
| 2384 |
],
|
| 2385 |
"logging_steps": 1,
|
|
@@ -2399,7 +4772,7 @@
|
|
| 2399 |
"attributes": {}
|
| 2400 |
}
|
| 2401 |
},
|
| 2402 |
-
"total_flos":
|
| 2403 |
"train_batch_size": 4,
|
| 2404 |
"trial_name": null,
|
| 2405 |
"trial_params": null
|
|
|
|
| 1 |
{
|
| 2 |
"best_metric": null,
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 0.0321007528052649,
|
| 5 |
"eval_steps": 500,
|
| 6 |
+
"global_step": 678,
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
|
|
| 2380 |
"learning_rate": 8.560166594400878e-05,
|
| 2381 |
"loss": 1.4472,
|
| 2382 |
"step": 339
|
| 2383 |
+
},
|
| 2384 |
+
{
|
| 2385 |
+
"epoch": 0.016097722645708062,
|
| 2386 |
+
"grad_norm": 1.2615294456481934,
|
| 2387 |
+
"learning_rate": 8.551974975930194e-05,
|
| 2388 |
+
"loss": 1.7209,
|
| 2389 |
+
"step": 340
|
| 2390 |
+
},
|
| 2391 |
+
{
|
| 2392 |
+
"epoch": 0.016145068888783674,
|
| 2393 |
+
"grad_norm": 1.0856550931930542,
|
| 2394 |
+
"learning_rate": 8.543764064908295e-05,
|
| 2395 |
+
"loss": 1.406,
|
| 2396 |
+
"step": 341
|
| 2397 |
+
},
|
| 2398 |
+
{
|
| 2399 |
+
"epoch": 0.016192415131859286,
|
| 2400 |
+
"grad_norm": 1.315007209777832,
|
| 2401 |
+
"learning_rate": 8.535533905932738e-05,
|
| 2402 |
+
"loss": 1.6584,
|
| 2403 |
+
"step": 342
|
| 2404 |
+
},
|
| 2405 |
+
{
|
| 2406 |
+
"epoch": 0.016239761374934897,
|
| 2407 |
+
"grad_norm": 1.180970311164856,
|
| 2408 |
+
"learning_rate": 8.527284543705632e-05,
|
| 2409 |
+
"loss": 1.3438,
|
| 2410 |
+
"step": 343
|
| 2411 |
+
},
|
| 2412 |
+
{
|
| 2413 |
+
"epoch": 0.01628710761801051,
|
| 2414 |
+
"grad_norm": 1.186798095703125,
|
| 2415 |
+
"learning_rate": 8.519016023033383e-05,
|
| 2416 |
+
"loss": 1.3434,
|
| 2417 |
+
"step": 344
|
| 2418 |
+
},
|
| 2419 |
+
{
|
| 2420 |
+
"epoch": 0.01633445386108612,
|
| 2421 |
+
"grad_norm": 1.1144177913665771,
|
| 2422 |
+
"learning_rate": 8.510728388826463e-05,
|
| 2423 |
+
"loss": 1.5506,
|
| 2424 |
+
"step": 345
|
| 2425 |
+
},
|
| 2426 |
+
{
|
| 2427 |
+
"epoch": 0.016381800104161733,
|
| 2428 |
+
"grad_norm": 1.2593318223953247,
|
| 2429 |
+
"learning_rate": 8.502421686099153e-05,
|
| 2430 |
+
"loss": 1.6704,
|
| 2431 |
+
"step": 346
|
| 2432 |
+
},
|
| 2433 |
+
{
|
| 2434 |
+
"epoch": 0.016429146347237345,
|
| 2435 |
+
"grad_norm": 1.1712514162063599,
|
| 2436 |
+
"learning_rate": 8.494095959969308e-05,
|
| 2437 |
+
"loss": 1.2738,
|
| 2438 |
+
"step": 347
|
| 2439 |
+
},
|
| 2440 |
+
{
|
| 2441 |
+
"epoch": 0.016476492590312957,
|
| 2442 |
+
"grad_norm": 1.7519192695617676,
|
| 2443 |
+
"learning_rate": 8.485751255658106e-05,
|
| 2444 |
+
"loss": 1.4788,
|
| 2445 |
+
"step": 348
|
| 2446 |
+
},
|
| 2447 |
+
{
|
| 2448 |
+
"epoch": 0.01652383883338857,
|
| 2449 |
+
"grad_norm": 1.3057172298431396,
|
| 2450 |
+
"learning_rate": 8.477387618489807e-05,
|
| 2451 |
+
"loss": 1.1369,
|
| 2452 |
+
"step": 349
|
| 2453 |
+
},
|
| 2454 |
+
{
|
| 2455 |
+
"epoch": 0.01657118507646418,
|
| 2456 |
+
"grad_norm": 1.533386468887329,
|
| 2457 |
+
"learning_rate": 8.469005093891505e-05,
|
| 2458 |
+
"loss": 1.1792,
|
| 2459 |
+
"step": 350
|
| 2460 |
+
},
|
| 2461 |
+
{
|
| 2462 |
+
"epoch": 0.016618531319539793,
|
| 2463 |
+
"grad_norm": 1.0570282936096191,
|
| 2464 |
+
"learning_rate": 8.460603727392877e-05,
|
| 2465 |
+
"loss": 1.6635,
|
| 2466 |
+
"step": 351
|
| 2467 |
+
},
|
| 2468 |
+
{
|
| 2469 |
+
"epoch": 0.016665877562615405,
|
| 2470 |
+
"grad_norm": 0.9694927930831909,
|
| 2471 |
+
"learning_rate": 8.452183564625948e-05,
|
| 2472 |
+
"loss": 1.5728,
|
| 2473 |
+
"step": 352
|
| 2474 |
+
},
|
| 2475 |
+
{
|
| 2476 |
+
"epoch": 0.01671322380569102,
|
| 2477 |
+
"grad_norm": 1.0785024166107178,
|
| 2478 |
+
"learning_rate": 8.443744651324827e-05,
|
| 2479 |
+
"loss": 1.8904,
|
| 2480 |
+
"step": 353
|
| 2481 |
+
},
|
| 2482 |
+
{
|
| 2483 |
+
"epoch": 0.016760570048766632,
|
| 2484 |
+
"grad_norm": 1.146831750869751,
|
| 2485 |
+
"learning_rate": 8.435287033325472e-05,
|
| 2486 |
+
"loss": 1.4965,
|
| 2487 |
+
"step": 354
|
| 2488 |
+
},
|
| 2489 |
+
{
|
| 2490 |
+
"epoch": 0.016807916291842244,
|
| 2491 |
+
"grad_norm": 0.9391350150108337,
|
| 2492 |
+
"learning_rate": 8.426810756565429e-05,
|
| 2493 |
+
"loss": 1.4763,
|
| 2494 |
+
"step": 355
|
| 2495 |
+
},
|
| 2496 |
+
{
|
| 2497 |
+
"epoch": 0.016855262534917856,
|
| 2498 |
+
"grad_norm": 1.0124608278274536,
|
| 2499 |
+
"learning_rate": 8.418315867083595e-05,
|
| 2500 |
+
"loss": 1.3741,
|
| 2501 |
+
"step": 356
|
| 2502 |
+
},
|
| 2503 |
+
{
|
| 2504 |
+
"epoch": 0.016902608777993468,
|
| 2505 |
+
"grad_norm": 1.037788987159729,
|
| 2506 |
+
"learning_rate": 8.409802411019963e-05,
|
| 2507 |
+
"loss": 1.2912,
|
| 2508 |
+
"step": 357
|
| 2509 |
+
},
|
| 2510 |
+
{
|
| 2511 |
+
"epoch": 0.01694995502106908,
|
| 2512 |
+
"grad_norm": 1.1288650035858154,
|
| 2513 |
+
"learning_rate": 8.401270434615365e-05,
|
| 2514 |
+
"loss": 1.3734,
|
| 2515 |
+
"step": 358
|
| 2516 |
+
},
|
| 2517 |
+
{
|
| 2518 |
+
"epoch": 0.01699730126414469,
|
| 2519 |
+
"grad_norm": 0.9866798520088196,
|
| 2520 |
+
"learning_rate": 8.392719984211228e-05,
|
| 2521 |
+
"loss": 1.2654,
|
| 2522 |
+
"step": 359
|
| 2523 |
+
},
|
| 2524 |
+
{
|
| 2525 |
+
"epoch": 0.017044647507220304,
|
| 2526 |
+
"grad_norm": 1.1551830768585205,
|
| 2527 |
+
"learning_rate": 8.384151106249323e-05,
|
| 2528 |
+
"loss": 1.7818,
|
| 2529 |
+
"step": 360
|
| 2530 |
+
},
|
| 2531 |
+
{
|
| 2532 |
+
"epoch": 0.017091993750295915,
|
| 2533 |
+
"grad_norm": 0.9992442727088928,
|
| 2534 |
+
"learning_rate": 8.375563847271506e-05,
|
| 2535 |
+
"loss": 1.4939,
|
| 2536 |
+
"step": 361
|
| 2537 |
+
},
|
| 2538 |
+
{
|
| 2539 |
+
"epoch": 0.017139339993371527,
|
| 2540 |
+
"grad_norm": 1.0358177423477173,
|
| 2541 |
+
"learning_rate": 8.366958253919475e-05,
|
| 2542 |
+
"loss": 1.3252,
|
| 2543 |
+
"step": 362
|
| 2544 |
+
},
|
| 2545 |
+
{
|
| 2546 |
+
"epoch": 0.01718668623644714,
|
| 2547 |
+
"grad_norm": 1.1302683353424072,
|
| 2548 |
+
"learning_rate": 8.358334372934509e-05,
|
| 2549 |
+
"loss": 1.5499,
|
| 2550 |
+
"step": 363
|
| 2551 |
+
},
|
| 2552 |
+
{
|
| 2553 |
+
"epoch": 0.01723403247952275,
|
| 2554 |
+
"grad_norm": 1.0650432109832764,
|
| 2555 |
+
"learning_rate": 8.349692251157211e-05,
|
| 2556 |
+
"loss": 1.6008,
|
| 2557 |
+
"step": 364
|
| 2558 |
+
},
|
| 2559 |
+
{
|
| 2560 |
+
"epoch": 0.017281378722598363,
|
| 2561 |
+
"grad_norm": 1.2165616750717163,
|
| 2562 |
+
"learning_rate": 8.341031935527267e-05,
|
| 2563 |
+
"loss": 1.8869,
|
| 2564 |
+
"step": 365
|
| 2565 |
+
},
|
| 2566 |
+
{
|
| 2567 |
+
"epoch": 0.017328724965673975,
|
| 2568 |
+
"grad_norm": 1.2423419952392578,
|
| 2569 |
+
"learning_rate": 8.332353473083179e-05,
|
| 2570 |
+
"loss": 1.5879,
|
| 2571 |
+
"step": 366
|
| 2572 |
+
},
|
| 2573 |
+
{
|
| 2574 |
+
"epoch": 0.017376071208749587,
|
| 2575 |
+
"grad_norm": 1.2204476594924927,
|
| 2576 |
+
"learning_rate": 8.323656910962011e-05,
|
| 2577 |
+
"loss": 1.5686,
|
| 2578 |
+
"step": 367
|
| 2579 |
+
},
|
| 2580 |
+
{
|
| 2581 |
+
"epoch": 0.0174234174518252,
|
| 2582 |
+
"grad_norm": 1.0241512060165405,
|
| 2583 |
+
"learning_rate": 8.314942296399141e-05,
|
| 2584 |
+
"loss": 1.5046,
|
| 2585 |
+
"step": 368
|
| 2586 |
+
},
|
| 2587 |
+
{
|
| 2588 |
+
"epoch": 0.01747076369490081,
|
| 2589 |
+
"grad_norm": 1.2489361763000488,
|
| 2590 |
+
"learning_rate": 8.306209676727994e-05,
|
| 2591 |
+
"loss": 1.6071,
|
| 2592 |
+
"step": 369
|
| 2593 |
+
},
|
| 2594 |
+
{
|
| 2595 |
+
"epoch": 0.017518109937976423,
|
| 2596 |
+
"grad_norm": 1.0107746124267578,
|
| 2597 |
+
"learning_rate": 8.297459099379794e-05,
|
| 2598 |
+
"loss": 1.3928,
|
| 2599 |
+
"step": 370
|
| 2600 |
+
},
|
| 2601 |
+
{
|
| 2602 |
+
"epoch": 0.017565456181052035,
|
| 2603 |
+
"grad_norm": 1.192285180091858,
|
| 2604 |
+
"learning_rate": 8.288690611883295e-05,
|
| 2605 |
+
"loss": 1.4469,
|
| 2606 |
+
"step": 371
|
| 2607 |
+
},
|
| 2608 |
+
{
|
| 2609 |
+
"epoch": 0.017612802424127647,
|
| 2610 |
+
"grad_norm": 1.2728925943374634,
|
| 2611 |
+
"learning_rate": 8.279904261864539e-05,
|
| 2612 |
+
"loss": 1.5513,
|
| 2613 |
+
"step": 372
|
| 2614 |
+
},
|
| 2615 |
+
{
|
| 2616 |
+
"epoch": 0.01766014866720326,
|
| 2617 |
+
"grad_norm": 1.3038040399551392,
|
| 2618 |
+
"learning_rate": 8.271100097046584e-05,
|
| 2619 |
+
"loss": 1.3809,
|
| 2620 |
+
"step": 373
|
| 2621 |
+
},
|
| 2622 |
+
{
|
| 2623 |
+
"epoch": 0.01770749491027887,
|
| 2624 |
+
"grad_norm": 1.1194548606872559,
|
| 2625 |
+
"learning_rate": 8.262278165249247e-05,
|
| 2626 |
+
"loss": 1.4155,
|
| 2627 |
+
"step": 374
|
| 2628 |
+
},
|
| 2629 |
+
{
|
| 2630 |
+
"epoch": 0.017754841153354482,
|
| 2631 |
+
"grad_norm": 1.4333120584487915,
|
| 2632 |
+
"learning_rate": 8.25343851438885e-05,
|
| 2633 |
+
"loss": 1.8168,
|
| 2634 |
+
"step": 375
|
| 2635 |
+
},
|
| 2636 |
+
{
|
| 2637 |
+
"epoch": 0.017802187396430094,
|
| 2638 |
+
"grad_norm": 1.1379622220993042,
|
| 2639 |
+
"learning_rate": 8.244581192477953e-05,
|
| 2640 |
+
"loss": 1.7745,
|
| 2641 |
+
"step": 376
|
| 2642 |
+
},
|
| 2643 |
+
{
|
| 2644 |
+
"epoch": 0.017849533639505706,
|
| 2645 |
+
"grad_norm": 1.1049047708511353,
|
| 2646 |
+
"learning_rate": 8.235706247625098e-05,
|
| 2647 |
+
"loss": 1.5711,
|
| 2648 |
+
"step": 377
|
| 2649 |
+
},
|
| 2650 |
+
{
|
| 2651 |
+
"epoch": 0.017896879882581318,
|
| 2652 |
+
"grad_norm": 1.0843724012374878,
|
| 2653 |
+
"learning_rate": 8.226813728034546e-05,
|
| 2654 |
+
"loss": 1.5343,
|
| 2655 |
+
"step": 378
|
| 2656 |
+
},
|
| 2657 |
+
{
|
| 2658 |
+
"epoch": 0.01794422612565693,
|
| 2659 |
+
"grad_norm": 0.9932480454444885,
|
| 2660 |
+
"learning_rate": 8.217903682006017e-05,
|
| 2661 |
+
"loss": 1.5807,
|
| 2662 |
+
"step": 379
|
| 2663 |
+
},
|
| 2664 |
+
{
|
| 2665 |
+
"epoch": 0.017991572368732542,
|
| 2666 |
+
"grad_norm": 1.000180721282959,
|
| 2667 |
+
"learning_rate": 8.208976157934422e-05,
|
| 2668 |
+
"loss": 1.517,
|
| 2669 |
+
"step": 380
|
| 2670 |
+
},
|
| 2671 |
+
{
|
| 2672 |
+
"epoch": 0.018038918611808154,
|
| 2673 |
+
"grad_norm": 1.5340940952301025,
|
| 2674 |
+
"learning_rate": 8.200031204309603e-05,
|
| 2675 |
+
"loss": 1.5985,
|
| 2676 |
+
"step": 381
|
| 2677 |
+
},
|
| 2678 |
+
{
|
| 2679 |
+
"epoch": 0.018086264854883766,
|
| 2680 |
+
"grad_norm": 1.100239634513855,
|
| 2681 |
+
"learning_rate": 8.19106886971608e-05,
|
| 2682 |
+
"loss": 1.7124,
|
| 2683 |
+
"step": 382
|
| 2684 |
+
},
|
| 2685 |
+
{
|
| 2686 |
+
"epoch": 0.018133611097959378,
|
| 2687 |
+
"grad_norm": 1.123328685760498,
|
| 2688 |
+
"learning_rate": 8.182089202832767e-05,
|
| 2689 |
+
"loss": 1.6103,
|
| 2690 |
+
"step": 383
|
| 2691 |
+
},
|
| 2692 |
+
{
|
| 2693 |
+
"epoch": 0.01818095734103499,
|
| 2694 |
+
"grad_norm": 1.1449183225631714,
|
| 2695 |
+
"learning_rate": 8.173092252432721e-05,
|
| 2696 |
+
"loss": 1.4786,
|
| 2697 |
+
"step": 384
|
| 2698 |
+
},
|
| 2699 |
+
{
|
| 2700 |
+
"epoch": 0.0182283035841106,
|
| 2701 |
+
"grad_norm": 1.0747967958450317,
|
| 2702 |
+
"learning_rate": 8.164078067382882e-05,
|
| 2703 |
+
"loss": 1.6038,
|
| 2704 |
+
"step": 385
|
| 2705 |
+
},
|
| 2706 |
+
{
|
| 2707 |
+
"epoch": 0.018275649827186213,
|
| 2708 |
+
"grad_norm": 0.9646826982498169,
|
| 2709 |
+
"learning_rate": 8.155046696643787e-05,
|
| 2710 |
+
"loss": 1.2486,
|
| 2711 |
+
"step": 386
|
| 2712 |
+
},
|
| 2713 |
+
{
|
| 2714 |
+
"epoch": 0.018322996070261825,
|
| 2715 |
+
"grad_norm": 1.163071632385254,
|
| 2716 |
+
"learning_rate": 8.145998189269326e-05,
|
| 2717 |
+
"loss": 1.3863,
|
| 2718 |
+
"step": 387
|
| 2719 |
+
},
|
| 2720 |
+
{
|
| 2721 |
+
"epoch": 0.018370342313337437,
|
| 2722 |
+
"grad_norm": 1.092169165611267,
|
| 2723 |
+
"learning_rate": 8.136932594406468e-05,
|
| 2724 |
+
"loss": 1.4198,
|
| 2725 |
+
"step": 388
|
| 2726 |
+
},
|
| 2727 |
+
{
|
| 2728 |
+
"epoch": 0.01841768855641305,
|
| 2729 |
+
"grad_norm": 1.1086021661758423,
|
| 2730 |
+
"learning_rate": 8.127849961294984e-05,
|
| 2731 |
+
"loss": 1.4193,
|
| 2732 |
+
"step": 389
|
| 2733 |
+
},
|
| 2734 |
+
{
|
| 2735 |
+
"epoch": 0.01846503479948866,
|
| 2736 |
+
"grad_norm": 1.2102797031402588,
|
| 2737 |
+
"learning_rate": 8.118750339267198e-05,
|
| 2738 |
+
"loss": 1.6943,
|
| 2739 |
+
"step": 390
|
| 2740 |
+
},
|
| 2741 |
+
{
|
| 2742 |
+
"epoch": 0.018512381042564273,
|
| 2743 |
+
"grad_norm": 1.0669124126434326,
|
| 2744 |
+
"learning_rate": 8.109633777747703e-05,
|
| 2745 |
+
"loss": 1.5252,
|
| 2746 |
+
"step": 391
|
| 2747 |
+
},
|
| 2748 |
+
{
|
| 2749 |
+
"epoch": 0.018559727285639885,
|
| 2750 |
+
"grad_norm": 1.1494920253753662,
|
| 2751 |
+
"learning_rate": 8.100500326253101e-05,
|
| 2752 |
+
"loss": 1.5225,
|
| 2753 |
+
"step": 392
|
| 2754 |
+
},
|
| 2755 |
+
{
|
| 2756 |
+
"epoch": 0.018607073528715497,
|
| 2757 |
+
"grad_norm": 1.2124680280685425,
|
| 2758 |
+
"learning_rate": 8.091350034391732e-05,
|
| 2759 |
+
"loss": 1.7474,
|
| 2760 |
+
"step": 393
|
| 2761 |
+
},
|
| 2762 |
+
{
|
| 2763 |
+
"epoch": 0.01865441977179111,
|
| 2764 |
+
"grad_norm": 1.1262034177780151,
|
| 2765 |
+
"learning_rate": 8.082182951863402e-05,
|
| 2766 |
+
"loss": 1.2812,
|
| 2767 |
+
"step": 394
|
| 2768 |
+
},
|
| 2769 |
+
{
|
| 2770 |
+
"epoch": 0.01870176601486672,
|
| 2771 |
+
"grad_norm": 1.3504730463027954,
|
| 2772 |
+
"learning_rate": 8.072999128459119e-05,
|
| 2773 |
+
"loss": 1.4917,
|
| 2774 |
+
"step": 395
|
| 2775 |
+
},
|
| 2776 |
+
{
|
| 2777 |
+
"epoch": 0.018749112257942332,
|
| 2778 |
+
"grad_norm": 1.1630171537399292,
|
| 2779 |
+
"learning_rate": 8.063798614060819e-05,
|
| 2780 |
+
"loss": 1.4038,
|
| 2781 |
+
"step": 396
|
| 2782 |
+
},
|
| 2783 |
+
{
|
| 2784 |
+
"epoch": 0.018796458501017944,
|
| 2785 |
+
"grad_norm": 1.2764859199523926,
|
| 2786 |
+
"learning_rate": 8.05458145864109e-05,
|
| 2787 |
+
"loss": 1.489,
|
| 2788 |
+
"step": 397
|
| 2789 |
+
},
|
| 2790 |
+
{
|
| 2791 |
+
"epoch": 0.018843804744093556,
|
| 2792 |
+
"grad_norm": 1.1614313125610352,
|
| 2793 |
+
"learning_rate": 8.045347712262912e-05,
|
| 2794 |
+
"loss": 1.2777,
|
| 2795 |
+
"step": 398
|
| 2796 |
+
},
|
| 2797 |
+
{
|
| 2798 |
+
"epoch": 0.018891150987169168,
|
| 2799 |
+
"grad_norm": 1.272247314453125,
|
| 2800 |
+
"learning_rate": 8.036097425079378e-05,
|
| 2801 |
+
"loss": 1.3734,
|
| 2802 |
+
"step": 399
|
| 2803 |
+
},
|
| 2804 |
+
{
|
| 2805 |
+
"epoch": 0.01893849723024478,
|
| 2806 |
+
"grad_norm": 1.4819856882095337,
|
| 2807 |
+
"learning_rate": 8.026830647333416e-05,
|
| 2808 |
+
"loss": 1.4693,
|
| 2809 |
+
"step": 400
|
| 2810 |
+
},
|
| 2811 |
+
{
|
| 2812 |
+
"epoch": 0.018985843473320392,
|
| 2813 |
+
"grad_norm": 1.1174564361572266,
|
| 2814 |
+
"learning_rate": 8.017547429357532e-05,
|
| 2815 |
+
"loss": 1.7608,
|
| 2816 |
+
"step": 401
|
| 2817 |
+
},
|
| 2818 |
+
{
|
| 2819 |
+
"epoch": 0.019033189716396004,
|
| 2820 |
+
"grad_norm": 0.9090219140052795,
|
| 2821 |
+
"learning_rate": 8.008247821573518e-05,
|
| 2822 |
+
"loss": 1.2873,
|
| 2823 |
+
"step": 402
|
| 2824 |
+
},
|
| 2825 |
+
{
|
| 2826 |
+
"epoch": 0.019080535959471616,
|
| 2827 |
+
"grad_norm": 1.025678038597107,
|
| 2828 |
+
"learning_rate": 7.998931874492193e-05,
|
| 2829 |
+
"loss": 1.4831,
|
| 2830 |
+
"step": 403
|
| 2831 |
+
},
|
| 2832 |
+
{
|
| 2833 |
+
"epoch": 0.019127882202547228,
|
| 2834 |
+
"grad_norm": 1.1207436323165894,
|
| 2835 |
+
"learning_rate": 7.989599638713122e-05,
|
| 2836 |
+
"loss": 1.6693,
|
| 2837 |
+
"step": 404
|
| 2838 |
+
},
|
| 2839 |
+
{
|
| 2840 |
+
"epoch": 0.01917522844562284,
|
| 2841 |
+
"grad_norm": 1.031245231628418,
|
| 2842 |
+
"learning_rate": 7.98025116492434e-05,
|
| 2843 |
+
"loss": 1.4521,
|
| 2844 |
+
"step": 405
|
| 2845 |
+
},
|
| 2846 |
+
{
|
| 2847 |
+
"epoch": 0.01922257468869845,
|
| 2848 |
+
"grad_norm": 1.0481898784637451,
|
| 2849 |
+
"learning_rate": 7.970886503902082e-05,
|
| 2850 |
+
"loss": 1.5691,
|
| 2851 |
+
"step": 406
|
| 2852 |
+
},
|
| 2853 |
+
{
|
| 2854 |
+
"epoch": 0.019269920931774064,
|
| 2855 |
+
"grad_norm": 0.929305911064148,
|
| 2856 |
+
"learning_rate": 7.961505706510499e-05,
|
| 2857 |
+
"loss": 1.2938,
|
| 2858 |
+
"step": 407
|
| 2859 |
+
},
|
| 2860 |
+
{
|
| 2861 |
+
"epoch": 0.019317267174849675,
|
| 2862 |
+
"grad_norm": 1.1896798610687256,
|
| 2863 |
+
"learning_rate": 7.952108823701393e-05,
|
| 2864 |
+
"loss": 1.2882,
|
| 2865 |
+
"step": 408
|
| 2866 |
+
},
|
| 2867 |
+
{
|
| 2868 |
+
"epoch": 0.019364613417925287,
|
| 2869 |
+
"grad_norm": 1.116339087486267,
|
| 2870 |
+
"learning_rate": 7.94269590651393e-05,
|
| 2871 |
+
"loss": 1.5144,
|
| 2872 |
+
"step": 409
|
| 2873 |
+
},
|
| 2874 |
+
{
|
| 2875 |
+
"epoch": 0.0194119596610009,
|
| 2876 |
+
"grad_norm": 1.0536943674087524,
|
| 2877 |
+
"learning_rate": 7.933267006074366e-05,
|
| 2878 |
+
"loss": 1.4844,
|
| 2879 |
+
"step": 410
|
| 2880 |
+
},
|
| 2881 |
+
{
|
| 2882 |
+
"epoch": 0.01945930590407651,
|
| 2883 |
+
"grad_norm": 1.1302292346954346,
|
| 2884 |
+
"learning_rate": 7.923822173595772e-05,
|
| 2885 |
+
"loss": 1.5484,
|
| 2886 |
+
"step": 411
|
| 2887 |
+
},
|
| 2888 |
+
{
|
| 2889 |
+
"epoch": 0.019506652147152123,
|
| 2890 |
+
"grad_norm": 1.0949054956436157,
|
| 2891 |
+
"learning_rate": 7.914361460377755e-05,
|
| 2892 |
+
"loss": 1.5731,
|
| 2893 |
+
"step": 412
|
| 2894 |
+
},
|
| 2895 |
+
{
|
| 2896 |
+
"epoch": 0.019553998390227735,
|
| 2897 |
+
"grad_norm": 1.1493390798568726,
|
| 2898 |
+
"learning_rate": 7.904884917806174e-05,
|
| 2899 |
+
"loss": 1.5251,
|
| 2900 |
+
"step": 413
|
| 2901 |
+
},
|
| 2902 |
+
{
|
| 2903 |
+
"epoch": 0.019601344633303347,
|
| 2904 |
+
"grad_norm": 1.150673270225525,
|
| 2905 |
+
"learning_rate": 7.895392597352869e-05,
|
| 2906 |
+
"loss": 1.4642,
|
| 2907 |
+
"step": 414
|
| 2908 |
+
},
|
| 2909 |
+
{
|
| 2910 |
+
"epoch": 0.01964869087637896,
|
| 2911 |
+
"grad_norm": 1.1738483905792236,
|
| 2912 |
+
"learning_rate": 7.885884550575376e-05,
|
| 2913 |
+
"loss": 1.6378,
|
| 2914 |
+
"step": 415
|
| 2915 |
+
},
|
| 2916 |
+
{
|
| 2917 |
+
"epoch": 0.01969603711945457,
|
| 2918 |
+
"grad_norm": 1.1144845485687256,
|
| 2919 |
+
"learning_rate": 7.876360829116647e-05,
|
| 2920 |
+
"loss": 1.4802,
|
| 2921 |
+
"step": 416
|
| 2922 |
+
},
|
| 2923 |
+
{
|
| 2924 |
+
"epoch": 0.019743383362530183,
|
| 2925 |
+
"grad_norm": 1.1597155332565308,
|
| 2926 |
+
"learning_rate": 7.866821484704776e-05,
|
| 2927 |
+
"loss": 1.4787,
|
| 2928 |
+
"step": 417
|
| 2929 |
+
},
|
| 2930 |
+
{
|
| 2931 |
+
"epoch": 0.019790729605605795,
|
| 2932 |
+
"grad_norm": 1.2034460306167603,
|
| 2933 |
+
"learning_rate": 7.85726656915271e-05,
|
| 2934 |
+
"loss": 1.3434,
|
| 2935 |
+
"step": 418
|
| 2936 |
+
},
|
| 2937 |
+
{
|
| 2938 |
+
"epoch": 0.019838075848681407,
|
| 2939 |
+
"grad_norm": 1.213287115097046,
|
| 2940 |
+
"learning_rate": 7.847696134357966e-05,
|
| 2941 |
+
"loss": 1.5816,
|
| 2942 |
+
"step": 419
|
| 2943 |
+
},
|
| 2944 |
+
{
|
| 2945 |
+
"epoch": 0.01988542209175702,
|
| 2946 |
+
"grad_norm": 1.4430410861968994,
|
| 2947 |
+
"learning_rate": 7.838110232302362e-05,
|
| 2948 |
+
"loss": 1.5156,
|
| 2949 |
+
"step": 420
|
| 2950 |
+
},
|
| 2951 |
+
{
|
| 2952 |
+
"epoch": 0.01993276833483263,
|
| 2953 |
+
"grad_norm": 1.4468246698379517,
|
| 2954 |
+
"learning_rate": 7.828508915051724e-05,
|
| 2955 |
+
"loss": 1.6498,
|
| 2956 |
+
"step": 421
|
| 2957 |
+
},
|
| 2958 |
+
{
|
| 2959 |
+
"epoch": 0.019980114577908242,
|
| 2960 |
+
"grad_norm": 1.3097630739212036,
|
| 2961 |
+
"learning_rate": 7.818892234755599e-05,
|
| 2962 |
+
"loss": 1.7392,
|
| 2963 |
+
"step": 422
|
| 2964 |
+
},
|
| 2965 |
+
{
|
| 2966 |
+
"epoch": 0.020027460820983854,
|
| 2967 |
+
"grad_norm": 1.0946062803268433,
|
| 2968 |
+
"learning_rate": 7.80926024364699e-05,
|
| 2969 |
+
"loss": 1.5731,
|
| 2970 |
+
"step": 423
|
| 2971 |
+
},
|
| 2972 |
+
{
|
| 2973 |
+
"epoch": 0.020074807064059466,
|
| 2974 |
+
"grad_norm": 1.2427587509155273,
|
| 2975 |
+
"learning_rate": 7.799612994042049e-05,
|
| 2976 |
+
"loss": 1.7339,
|
| 2977 |
+
"step": 424
|
| 2978 |
+
},
|
| 2979 |
+
{
|
| 2980 |
+
"epoch": 0.020122153307135078,
|
| 2981 |
+
"grad_norm": 1.1726187467575073,
|
| 2982 |
+
"learning_rate": 7.789950538339812e-05,
|
| 2983 |
+
"loss": 1.6152,
|
| 2984 |
+
"step": 425
|
| 2985 |
+
},
|
| 2986 |
+
{
|
| 2987 |
+
"epoch": 0.02016949955021069,
|
| 2988 |
+
"grad_norm": 1.4452691078186035,
|
| 2989 |
+
"learning_rate": 7.780272929021906e-05,
|
| 2990 |
+
"loss": 1.9271,
|
| 2991 |
+
"step": 426
|
| 2992 |
+
},
|
| 2993 |
+
{
|
| 2994 |
+
"epoch": 0.020216845793286302,
|
| 2995 |
+
"grad_norm": 1.174172282218933,
|
| 2996 |
+
"learning_rate": 7.770580218652261e-05,
|
| 2997 |
+
"loss": 1.5024,
|
| 2998 |
+
"step": 427
|
| 2999 |
+
},
|
| 3000 |
+
{
|
| 3001 |
+
"epoch": 0.020264192036361914,
|
| 3002 |
+
"grad_norm": 1.14840567111969,
|
| 3003 |
+
"learning_rate": 7.760872459876834e-05,
|
| 3004 |
+
"loss": 1.467,
|
| 3005 |
+
"step": 428
|
| 3006 |
+
},
|
| 3007 |
+
{
|
| 3008 |
+
"epoch": 0.020311538279437526,
|
| 3009 |
+
"grad_norm": 1.1591323614120483,
|
| 3010 |
+
"learning_rate": 7.751149705423312e-05,
|
| 3011 |
+
"loss": 1.5206,
|
| 3012 |
+
"step": 429
|
| 3013 |
+
},
|
| 3014 |
+
{
|
| 3015 |
+
"epoch": 0.020358884522513138,
|
| 3016 |
+
"grad_norm": 1.0279210805892944,
|
| 3017 |
+
"learning_rate": 7.741412008100834e-05,
|
| 3018 |
+
"loss": 1.6793,
|
| 3019 |
+
"step": 430
|
| 3020 |
+
},
|
| 3021 |
+
{
|
| 3022 |
+
"epoch": 0.02040623076558875,
|
| 3023 |
+
"grad_norm": 1.0675665140151978,
|
| 3024 |
+
"learning_rate": 7.731659420799703e-05,
|
| 3025 |
+
"loss": 1.2579,
|
| 3026 |
+
"step": 431
|
| 3027 |
+
},
|
| 3028 |
+
{
|
| 3029 |
+
"epoch": 0.02045357700866436,
|
| 3030 |
+
"grad_norm": 1.1949347257614136,
|
| 3031 |
+
"learning_rate": 7.721891996491091e-05,
|
| 3032 |
+
"loss": 1.7004,
|
| 3033 |
+
"step": 432
|
| 3034 |
+
},
|
| 3035 |
+
{
|
| 3036 |
+
"epoch": 0.020500923251739973,
|
| 3037 |
+
"grad_norm": 1.1350997686386108,
|
| 3038 |
+
"learning_rate": 7.712109788226762e-05,
|
| 3039 |
+
"loss": 1.5066,
|
| 3040 |
+
"step": 433
|
| 3041 |
+
},
|
| 3042 |
+
{
|
| 3043 |
+
"epoch": 0.020548269494815585,
|
| 3044 |
+
"grad_norm": 1.492813229560852,
|
| 3045 |
+
"learning_rate": 7.70231284913878e-05,
|
| 3046 |
+
"loss": 1.5321,
|
| 3047 |
+
"step": 434
|
| 3048 |
+
},
|
| 3049 |
+
{
|
| 3050 |
+
"epoch": 0.020595615737891197,
|
| 3051 |
+
"grad_norm": 1.1236470937728882,
|
| 3052 |
+
"learning_rate": 7.692501232439214e-05,
|
| 3053 |
+
"loss": 1.5698,
|
| 3054 |
+
"step": 435
|
| 3055 |
+
},
|
| 3056 |
+
{
|
| 3057 |
+
"epoch": 0.02064296198096681,
|
| 3058 |
+
"grad_norm": 1.1587574481964111,
|
| 3059 |
+
"learning_rate": 7.682674991419861e-05,
|
| 3060 |
+
"loss": 1.5557,
|
| 3061 |
+
"step": 436
|
| 3062 |
+
},
|
| 3063 |
+
{
|
| 3064 |
+
"epoch": 0.02069030822404242,
|
| 3065 |
+
"grad_norm": 1.0571719408035278,
|
| 3066 |
+
"learning_rate": 7.672834179451942e-05,
|
| 3067 |
+
"loss": 1.3085,
|
| 3068 |
+
"step": 437
|
| 3069 |
+
},
|
| 3070 |
+
{
|
| 3071 |
+
"epoch": 0.020737654467118033,
|
| 3072 |
+
"grad_norm": 1.024495005607605,
|
| 3073 |
+
"learning_rate": 7.662978849985829e-05,
|
| 3074 |
+
"loss": 1.4495,
|
| 3075 |
+
"step": 438
|
| 3076 |
+
},
|
| 3077 |
+
{
|
| 3078 |
+
"epoch": 0.020785000710193645,
|
| 3079 |
+
"grad_norm": 1.062393307685852,
|
| 3080 |
+
"learning_rate": 7.653109056550741e-05,
|
| 3081 |
+
"loss": 1.16,
|
| 3082 |
+
"step": 439
|
| 3083 |
+
},
|
| 3084 |
+
{
|
| 3085 |
+
"epoch": 0.020832346953269257,
|
| 3086 |
+
"grad_norm": 1.0015718936920166,
|
| 3087 |
+
"learning_rate": 7.643224852754458e-05,
|
| 3088 |
+
"loss": 1.419,
|
| 3089 |
+
"step": 440
|
| 3090 |
+
},
|
| 3091 |
+
{
|
| 3092 |
+
"epoch": 0.02087969319634487,
|
| 3093 |
+
"grad_norm": 1.134616494178772,
|
| 3094 |
+
"learning_rate": 7.633326292283028e-05,
|
| 3095 |
+
"loss": 1.3081,
|
| 3096 |
+
"step": 441
|
| 3097 |
+
},
|
| 3098 |
+
{
|
| 3099 |
+
"epoch": 0.02092703943942048,
|
| 3100 |
+
"grad_norm": 1.2353401184082031,
|
| 3101 |
+
"learning_rate": 7.623413428900485e-05,
|
| 3102 |
+
"loss": 1.3393,
|
| 3103 |
+
"step": 442
|
| 3104 |
+
},
|
| 3105 |
+
{
|
| 3106 |
+
"epoch": 0.020974385682496093,
|
| 3107 |
+
"grad_norm": 1.1951806545257568,
|
| 3108 |
+
"learning_rate": 7.613486316448539e-05,
|
| 3109 |
+
"loss": 1.4668,
|
| 3110 |
+
"step": 443
|
| 3111 |
+
},
|
| 3112 |
+
{
|
| 3113 |
+
"epoch": 0.021021731925571704,
|
| 3114 |
+
"grad_norm": 1.1466597318649292,
|
| 3115 |
+
"learning_rate": 7.603545008846303e-05,
|
| 3116 |
+
"loss": 1.4044,
|
| 3117 |
+
"step": 444
|
| 3118 |
+
},
|
| 3119 |
+
{
|
| 3120 |
+
"epoch": 0.021069078168647316,
|
| 3121 |
+
"grad_norm": 1.498052716255188,
|
| 3122 |
+
"learning_rate": 7.593589560089985e-05,
|
| 3123 |
+
"loss": 1.4792,
|
| 3124 |
+
"step": 445
|
| 3125 |
+
},
|
| 3126 |
+
{
|
| 3127 |
+
"epoch": 0.02111642441172293,
|
| 3128 |
+
"grad_norm": 1.357369065284729,
|
| 3129 |
+
"learning_rate": 7.5836200242526e-05,
|
| 3130 |
+
"loss": 1.6335,
|
| 3131 |
+
"step": 446
|
| 3132 |
+
},
|
| 3133 |
+
{
|
| 3134 |
+
"epoch": 0.02116377065479854,
|
| 3135 |
+
"grad_norm": 1.073910117149353,
|
| 3136 |
+
"learning_rate": 7.573636455483683e-05,
|
| 3137 |
+
"loss": 1.2684,
|
| 3138 |
+
"step": 447
|
| 3139 |
+
},
|
| 3140 |
+
{
|
| 3141 |
+
"epoch": 0.021211116897874152,
|
| 3142 |
+
"grad_norm": 1.3562556505203247,
|
| 3143 |
+
"learning_rate": 7.563638908008985e-05,
|
| 3144 |
+
"loss": 1.4344,
|
| 3145 |
+
"step": 448
|
| 3146 |
+
},
|
| 3147 |
+
{
|
| 3148 |
+
"epoch": 0.021258463140949764,
|
| 3149 |
+
"grad_norm": 1.2978427410125732,
|
| 3150 |
+
"learning_rate": 7.553627436130183e-05,
|
| 3151 |
+
"loss": 1.3625,
|
| 3152 |
+
"step": 449
|
| 3153 |
+
},
|
| 3154 |
+
{
|
| 3155 |
+
"epoch": 0.021305809384025376,
|
| 3156 |
+
"grad_norm": 1.2368243932724,
|
| 3157 |
+
"learning_rate": 7.543602094224582e-05,
|
| 3158 |
+
"loss": 1.0972,
|
| 3159 |
+
"step": 450
|
| 3160 |
+
},
|
| 3161 |
+
{
|
| 3162 |
+
"epoch": 0.021353155627100988,
|
| 3163 |
+
"grad_norm": 0.983315110206604,
|
| 3164 |
+
"learning_rate": 7.533562936744825e-05,
|
| 3165 |
+
"loss": 1.6439,
|
| 3166 |
+
"step": 451
|
| 3167 |
+
},
|
| 3168 |
+
{
|
| 3169 |
+
"epoch": 0.0214005018701766,
|
| 3170 |
+
"grad_norm": 1.003436803817749,
|
| 3171 |
+
"learning_rate": 7.523510018218593e-05,
|
| 3172 |
+
"loss": 1.6331,
|
| 3173 |
+
"step": 452
|
| 3174 |
+
},
|
| 3175 |
+
{
|
| 3176 |
+
"epoch": 0.021447848113252215,
|
| 3177 |
+
"grad_norm": 0.9159024357795715,
|
| 3178 |
+
"learning_rate": 7.513443393248312e-05,
|
| 3179 |
+
"loss": 1.4203,
|
| 3180 |
+
"step": 453
|
| 3181 |
+
},
|
| 3182 |
+
{
|
| 3183 |
+
"epoch": 0.021495194356327827,
|
| 3184 |
+
"grad_norm": 1.0519555807113647,
|
| 3185 |
+
"learning_rate": 7.503363116510851e-05,
|
| 3186 |
+
"loss": 1.4515,
|
| 3187 |
+
"step": 454
|
| 3188 |
+
},
|
| 3189 |
+
{
|
| 3190 |
+
"epoch": 0.02154254059940344,
|
| 3191 |
+
"grad_norm": 1.0524705648422241,
|
| 3192 |
+
"learning_rate": 7.493269242757233e-05,
|
| 3193 |
+
"loss": 1.5003,
|
| 3194 |
+
"step": 455
|
| 3195 |
+
},
|
| 3196 |
+
{
|
| 3197 |
+
"epoch": 0.02158988684247905,
|
| 3198 |
+
"grad_norm": 1.0478808879852295,
|
| 3199 |
+
"learning_rate": 7.483161826812328e-05,
|
| 3200 |
+
"loss": 1.4334,
|
| 3201 |
+
"step": 456
|
| 3202 |
+
},
|
| 3203 |
+
{
|
| 3204 |
+
"epoch": 0.021637233085554663,
|
| 3205 |
+
"grad_norm": 1.092322826385498,
|
| 3206 |
+
"learning_rate": 7.473040923574567e-05,
|
| 3207 |
+
"loss": 1.4022,
|
| 3208 |
+
"step": 457
|
| 3209 |
+
},
|
| 3210 |
+
{
|
| 3211 |
+
"epoch": 0.021684579328630275,
|
| 3212 |
+
"grad_norm": 1.1703879833221436,
|
| 3213 |
+
"learning_rate": 7.46290658801563e-05,
|
| 3214 |
+
"loss": 1.7451,
|
| 3215 |
+
"step": 458
|
| 3216 |
+
},
|
| 3217 |
+
{
|
| 3218 |
+
"epoch": 0.021731925571705887,
|
| 3219 |
+
"grad_norm": 1.0476280450820923,
|
| 3220 |
+
"learning_rate": 7.45275887518016e-05,
|
| 3221 |
+
"loss": 1.5566,
|
| 3222 |
+
"step": 459
|
| 3223 |
+
},
|
| 3224 |
+
{
|
| 3225 |
+
"epoch": 0.0217792718147815,
|
| 3226 |
+
"grad_norm": 1.0837757587432861,
|
| 3227 |
+
"learning_rate": 7.442597840185456e-05,
|
| 3228 |
+
"loss": 1.6164,
|
| 3229 |
+
"step": 460
|
| 3230 |
+
},
|
| 3231 |
+
{
|
| 3232 |
+
"epoch": 0.02182661805785711,
|
| 3233 |
+
"grad_norm": 1.053000569343567,
|
| 3234 |
+
"learning_rate": 7.432423538221178e-05,
|
| 3235 |
+
"loss": 1.3632,
|
| 3236 |
+
"step": 461
|
| 3237 |
+
},
|
| 3238 |
+
{
|
| 3239 |
+
"epoch": 0.021873964300932722,
|
| 3240 |
+
"grad_norm": 1.14542555809021,
|
| 3241 |
+
"learning_rate": 7.422236024549043e-05,
|
| 3242 |
+
"loss": 1.3115,
|
| 3243 |
+
"step": 462
|
| 3244 |
+
},
|
| 3245 |
+
{
|
| 3246 |
+
"epoch": 0.021921310544008334,
|
| 3247 |
+
"grad_norm": 1.0363187789916992,
|
| 3248 |
+
"learning_rate": 7.412035354502532e-05,
|
| 3249 |
+
"loss": 1.1298,
|
| 3250 |
+
"step": 463
|
| 3251 |
+
},
|
| 3252 |
+
{
|
| 3253 |
+
"epoch": 0.021968656787083946,
|
| 3254 |
+
"grad_norm": 1.152509093284607,
|
| 3255 |
+
"learning_rate": 7.401821583486576e-05,
|
| 3256 |
+
"loss": 1.6052,
|
| 3257 |
+
"step": 464
|
| 3258 |
+
},
|
| 3259 |
+
{
|
| 3260 |
+
"epoch": 0.022016003030159558,
|
| 3261 |
+
"grad_norm": 1.1238806247711182,
|
| 3262 |
+
"learning_rate": 7.391594766977277e-05,
|
| 3263 |
+
"loss": 1.3573,
|
| 3264 |
+
"step": 465
|
| 3265 |
+
},
|
| 3266 |
+
{
|
| 3267 |
+
"epoch": 0.02206334927323517,
|
| 3268 |
+
"grad_norm": 1.072085976600647,
|
| 3269 |
+
"learning_rate": 7.381354960521582e-05,
|
| 3270 |
+
"loss": 1.4479,
|
| 3271 |
+
"step": 466
|
| 3272 |
+
},
|
| 3273 |
+
{
|
| 3274 |
+
"epoch": 0.022110695516310782,
|
| 3275 |
+
"grad_norm": 1.342141032218933,
|
| 3276 |
+
"learning_rate": 7.371102219737e-05,
|
| 3277 |
+
"loss": 1.6094,
|
| 3278 |
+
"step": 467
|
| 3279 |
+
},
|
| 3280 |
+
{
|
| 3281 |
+
"epoch": 0.022158041759386394,
|
| 3282 |
+
"grad_norm": 1.165441870689392,
|
| 3283 |
+
"learning_rate": 7.360836600311286e-05,
|
| 3284 |
+
"loss": 1.589,
|
| 3285 |
+
"step": 468
|
| 3286 |
+
},
|
| 3287 |
+
{
|
| 3288 |
+
"epoch": 0.022205388002462006,
|
| 3289 |
+
"grad_norm": 1.105671763420105,
|
| 3290 |
+
"learning_rate": 7.350558158002154e-05,
|
| 3291 |
+
"loss": 1.5632,
|
| 3292 |
+
"step": 469
|
| 3293 |
+
},
|
| 3294 |
+
{
|
| 3295 |
+
"epoch": 0.022252734245537618,
|
| 3296 |
+
"grad_norm": 1.151079535484314,
|
| 3297 |
+
"learning_rate": 7.340266948636957e-05,
|
| 3298 |
+
"loss": 1.5647,
|
| 3299 |
+
"step": 470
|
| 3300 |
+
},
|
| 3301 |
+
{
|
| 3302 |
+
"epoch": 0.02230008048861323,
|
| 3303 |
+
"grad_norm": 1.3369741439819336,
|
| 3304 |
+
"learning_rate": 7.329963028112398e-05,
|
| 3305 |
+
"loss": 1.7085,
|
| 3306 |
+
"step": 471
|
| 3307 |
+
},
|
| 3308 |
+
{
|
| 3309 |
+
"epoch": 0.02234742673168884,
|
| 3310 |
+
"grad_norm": 1.3554760217666626,
|
| 3311 |
+
"learning_rate": 7.31964645239422e-05,
|
| 3312 |
+
"loss": 1.7893,
|
| 3313 |
+
"step": 472
|
| 3314 |
+
},
|
| 3315 |
+
{
|
| 3316 |
+
"epoch": 0.022394772974764453,
|
| 3317 |
+
"grad_norm": 1.2745559215545654,
|
| 3318 |
+
"learning_rate": 7.3093172775169e-05,
|
| 3319 |
+
"loss": 1.5171,
|
| 3320 |
+
"step": 473
|
| 3321 |
+
},
|
| 3322 |
+
{
|
| 3323 |
+
"epoch": 0.022442119217840065,
|
| 3324 |
+
"grad_norm": 1.1908695697784424,
|
| 3325 |
+
"learning_rate": 7.298975559583348e-05,
|
| 3326 |
+
"loss": 1.6895,
|
| 3327 |
+
"step": 474
|
| 3328 |
+
},
|
| 3329 |
+
{
|
| 3330 |
+
"epoch": 0.022489465460915677,
|
| 3331 |
+
"grad_norm": 1.2285094261169434,
|
| 3332 |
+
"learning_rate": 7.288621354764604e-05,
|
| 3333 |
+
"loss": 1.7527,
|
| 3334 |
+
"step": 475
|
| 3335 |
+
},
|
| 3336 |
+
{
|
| 3337 |
+
"epoch": 0.02253681170399129,
|
| 3338 |
+
"grad_norm": 2.27028489112854,
|
| 3339 |
+
"learning_rate": 7.278254719299528e-05,
|
| 3340 |
+
"loss": 1.671,
|
| 3341 |
+
"step": 476
|
| 3342 |
+
},
|
| 3343 |
+
{
|
| 3344 |
+
"epoch": 0.0225841579470669,
|
| 3345 |
+
"grad_norm": 1.16270911693573,
|
| 3346 |
+
"learning_rate": 7.267875709494499e-05,
|
| 3347 |
+
"loss": 1.5574,
|
| 3348 |
+
"step": 477
|
| 3349 |
+
},
|
| 3350 |
+
{
|
| 3351 |
+
"epoch": 0.022631504190142513,
|
| 3352 |
+
"grad_norm": 1.2359905242919922,
|
| 3353 |
+
"learning_rate": 7.257484381723107e-05,
|
| 3354 |
+
"loss": 1.8739,
|
| 3355 |
+
"step": 478
|
| 3356 |
+
},
|
| 3357 |
+
{
|
| 3358 |
+
"epoch": 0.022678850433218125,
|
| 3359 |
+
"grad_norm": 1.1558997631072998,
|
| 3360 |
+
"learning_rate": 7.247080792425844e-05,
|
| 3361 |
+
"loss": 1.7564,
|
| 3362 |
+
"step": 479
|
| 3363 |
+
},
|
| 3364 |
+
{
|
| 3365 |
+
"epoch": 0.022726196676293737,
|
| 3366 |
+
"grad_norm": 1.1397159099578857,
|
| 3367 |
+
"learning_rate": 7.236664998109803e-05,
|
| 3368 |
+
"loss": 1.5661,
|
| 3369 |
+
"step": 480
|
| 3370 |
+
},
|
| 3371 |
+
{
|
| 3372 |
+
"epoch": 0.02277354291936935,
|
| 3373 |
+
"grad_norm": 1.050812840461731,
|
| 3374 |
+
"learning_rate": 7.22623705534837e-05,
|
| 3375 |
+
"loss": 1.4053,
|
| 3376 |
+
"step": 481
|
| 3377 |
+
},
|
| 3378 |
+
{
|
| 3379 |
+
"epoch": 0.02282088916244496,
|
| 3380 |
+
"grad_norm": 1.070478916168213,
|
| 3381 |
+
"learning_rate": 7.21579702078091e-05,
|
| 3382 |
+
"loss": 1.5102,
|
| 3383 |
+
"step": 482
|
| 3384 |
+
},
|
| 3385 |
+
{
|
| 3386 |
+
"epoch": 0.022868235405520573,
|
| 3387 |
+
"grad_norm": 1.0024526119232178,
|
| 3388 |
+
"learning_rate": 7.205344951112474e-05,
|
| 3389 |
+
"loss": 1.631,
|
| 3390 |
+
"step": 483
|
| 3391 |
+
},
|
| 3392 |
+
{
|
| 3393 |
+
"epoch": 0.022915581648596185,
|
| 3394 |
+
"grad_norm": 1.031777024269104,
|
| 3395 |
+
"learning_rate": 7.194880903113471e-05,
|
| 3396 |
+
"loss": 1.5878,
|
| 3397 |
+
"step": 484
|
| 3398 |
+
},
|
| 3399 |
+
{
|
| 3400 |
+
"epoch": 0.022962927891671796,
|
| 3401 |
+
"grad_norm": 1.2618625164031982,
|
| 3402 |
+
"learning_rate": 7.184404933619377e-05,
|
| 3403 |
+
"loss": 1.6548,
|
| 3404 |
+
"step": 485
|
| 3405 |
+
},
|
| 3406 |
+
{
|
| 3407 |
+
"epoch": 0.02301027413474741,
|
| 3408 |
+
"grad_norm": 1.1872750520706177,
|
| 3409 |
+
"learning_rate": 7.173917099530418e-05,
|
| 3410 |
+
"loss": 1.7397,
|
| 3411 |
+
"step": 486
|
| 3412 |
+
},
|
| 3413 |
+
{
|
| 3414 |
+
"epoch": 0.02305762037782302,
|
| 3415 |
+
"grad_norm": 1.1944125890731812,
|
| 3416 |
+
"learning_rate": 7.163417457811261e-05,
|
| 3417 |
+
"loss": 1.6646,
|
| 3418 |
+
"step": 487
|
| 3419 |
+
},
|
| 3420 |
+
{
|
| 3421 |
+
"epoch": 0.023104966620898632,
|
| 3422 |
+
"grad_norm": 1.050458312034607,
|
| 3423 |
+
"learning_rate": 7.152906065490708e-05,
|
| 3424 |
+
"loss": 1.3215,
|
| 3425 |
+
"step": 488
|
| 3426 |
+
},
|
| 3427 |
+
{
|
| 3428 |
+
"epoch": 0.023152312863974244,
|
| 3429 |
+
"grad_norm": 0.999083936214447,
|
| 3430 |
+
"learning_rate": 7.142382979661386e-05,
|
| 3431 |
+
"loss": 1.3683,
|
| 3432 |
+
"step": 489
|
| 3433 |
+
},
|
| 3434 |
+
{
|
| 3435 |
+
"epoch": 0.023199659107049856,
|
| 3436 |
+
"grad_norm": 1.1480274200439453,
|
| 3437 |
+
"learning_rate": 7.13184825747943e-05,
|
| 3438 |
+
"loss": 1.4883,
|
| 3439 |
+
"step": 490
|
| 3440 |
+
},
|
| 3441 |
+
{
|
| 3442 |
+
"epoch": 0.023247005350125468,
|
| 3443 |
+
"grad_norm": 1.0551904439926147,
|
| 3444 |
+
"learning_rate": 7.121301956164184e-05,
|
| 3445 |
+
"loss": 1.462,
|
| 3446 |
+
"step": 491
|
| 3447 |
+
},
|
| 3448 |
+
{
|
| 3449 |
+
"epoch": 0.02329435159320108,
|
| 3450 |
+
"grad_norm": 1.1817728281021118,
|
| 3451 |
+
"learning_rate": 7.110744132997877e-05,
|
| 3452 |
+
"loss": 1.4887,
|
| 3453 |
+
"step": 492
|
| 3454 |
+
},
|
| 3455 |
+
{
|
| 3456 |
+
"epoch": 0.023341697836276692,
|
| 3457 |
+
"grad_norm": 1.1094098091125488,
|
| 3458 |
+
"learning_rate": 7.100174845325327e-05,
|
| 3459 |
+
"loss": 1.4593,
|
| 3460 |
+
"step": 493
|
| 3461 |
+
},
|
| 3462 |
+
{
|
| 3463 |
+
"epoch": 0.023389044079352304,
|
| 3464 |
+
"grad_norm": 1.0678788423538208,
|
| 3465 |
+
"learning_rate": 7.089594150553615e-05,
|
| 3466 |
+
"loss": 1.3145,
|
| 3467 |
+
"step": 494
|
| 3468 |
+
},
|
| 3469 |
+
{
|
| 3470 |
+
"epoch": 0.023436390322427916,
|
| 3471 |
+
"grad_norm": 1.2436941862106323,
|
| 3472 |
+
"learning_rate": 7.079002106151782e-05,
|
| 3473 |
+
"loss": 1.3822,
|
| 3474 |
+
"step": 495
|
| 3475 |
+
},
|
| 3476 |
+
{
|
| 3477 |
+
"epoch": 0.023483736565503528,
|
| 3478 |
+
"grad_norm": 1.3193278312683105,
|
| 3479 |
+
"learning_rate": 7.068398769650516e-05,
|
| 3480 |
+
"loss": 1.5251,
|
| 3481 |
+
"step": 496
|
| 3482 |
+
},
|
| 3483 |
+
{
|
| 3484 |
+
"epoch": 0.02353108280857914,
|
| 3485 |
+
"grad_norm": 1.4025723934173584,
|
| 3486 |
+
"learning_rate": 7.057784198641834e-05,
|
| 3487 |
+
"loss": 1.3896,
|
| 3488 |
+
"step": 497
|
| 3489 |
+
},
|
| 3490 |
+
{
|
| 3491 |
+
"epoch": 0.02357842905165475,
|
| 3492 |
+
"grad_norm": 1.537299394607544,
|
| 3493 |
+
"learning_rate": 7.047158450778782e-05,
|
| 3494 |
+
"loss": 1.6854,
|
| 3495 |
+
"step": 498
|
| 3496 |
+
},
|
| 3497 |
+
{
|
| 3498 |
+
"epoch": 0.023625775294730363,
|
| 3499 |
+
"grad_norm": 1.4015412330627441,
|
| 3500 |
+
"learning_rate": 7.036521583775099e-05,
|
| 3501 |
+
"loss": 1.6335,
|
| 3502 |
+
"step": 499
|
| 3503 |
+
},
|
| 3504 |
+
{
|
| 3505 |
+
"epoch": 0.023673121537805975,
|
| 3506 |
+
"grad_norm": 1.3199678659439087,
|
| 3507 |
+
"learning_rate": 7.025873655404931e-05,
|
| 3508 |
+
"loss": 1.0554,
|
| 3509 |
+
"step": 500
|
| 3510 |
+
},
|
| 3511 |
+
{
|
| 3512 |
+
"epoch": 0.023720467780881587,
|
| 3513 |
+
"grad_norm": 1.0300636291503906,
|
| 3514 |
+
"learning_rate": 7.015214723502496e-05,
|
| 3515 |
+
"loss": 1.854,
|
| 3516 |
+
"step": 501
|
| 3517 |
+
},
|
| 3518 |
+
{
|
| 3519 |
+
"epoch": 0.0237678140239572,
|
| 3520 |
+
"grad_norm": 1.021327018737793,
|
| 3521 |
+
"learning_rate": 7.00454484596178e-05,
|
| 3522 |
+
"loss": 1.5568,
|
| 3523 |
+
"step": 502
|
| 3524 |
+
},
|
| 3525 |
+
{
|
| 3526 |
+
"epoch": 0.02381516026703281,
|
| 3527 |
+
"grad_norm": 1.1301056146621704,
|
| 3528 |
+
"learning_rate": 6.99386408073622e-05,
|
| 3529 |
+
"loss": 1.703,
|
| 3530 |
+
"step": 503
|
| 3531 |
+
},
|
| 3532 |
+
{
|
| 3533 |
+
"epoch": 0.023862506510108423,
|
| 3534 |
+
"grad_norm": 1.0863914489746094,
|
| 3535 |
+
"learning_rate": 6.983172485838394e-05,
|
| 3536 |
+
"loss": 1.4904,
|
| 3537 |
+
"step": 504
|
| 3538 |
+
},
|
| 3539 |
+
{
|
| 3540 |
+
"epoch": 0.023909852753184035,
|
| 3541 |
+
"grad_norm": 1.0221482515335083,
|
| 3542 |
+
"learning_rate": 6.972470119339691e-05,
|
| 3543 |
+
"loss": 1.5025,
|
| 3544 |
+
"step": 505
|
| 3545 |
+
},
|
| 3546 |
+
{
|
| 3547 |
+
"epoch": 0.023957198996259647,
|
| 3548 |
+
"grad_norm": 0.9677168726921082,
|
| 3549 |
+
"learning_rate": 6.961757039370016e-05,
|
| 3550 |
+
"loss": 1.5255,
|
| 3551 |
+
"step": 506
|
| 3552 |
+
},
|
| 3553 |
+
{
|
| 3554 |
+
"epoch": 0.02400454523933526,
|
| 3555 |
+
"grad_norm": 1.0375566482543945,
|
| 3556 |
+
"learning_rate": 6.951033304117458e-05,
|
| 3557 |
+
"loss": 1.4896,
|
| 3558 |
+
"step": 507
|
| 3559 |
+
},
|
| 3560 |
+
{
|
| 3561 |
+
"epoch": 0.02405189148241087,
|
| 3562 |
+
"grad_norm": 1.0411720275878906,
|
| 3563 |
+
"learning_rate": 6.940298971827985e-05,
|
| 3564 |
+
"loss": 1.4138,
|
| 3565 |
+
"step": 508
|
| 3566 |
+
},
|
| 3567 |
+
{
|
| 3568 |
+
"epoch": 0.024099237725486482,
|
| 3569 |
+
"grad_norm": 0.9705979824066162,
|
| 3570 |
+
"learning_rate": 6.929554100805118e-05,
|
| 3571 |
+
"loss": 1.3712,
|
| 3572 |
+
"step": 509
|
| 3573 |
+
},
|
| 3574 |
+
{
|
| 3575 |
+
"epoch": 0.024146583968562094,
|
| 3576 |
+
"grad_norm": 1.141237735748291,
|
| 3577 |
+
"learning_rate": 6.91879874940962e-05,
|
| 3578 |
+
"loss": 1.4621,
|
| 3579 |
+
"step": 510
|
| 3580 |
+
},
|
| 3581 |
+
{
|
| 3582 |
+
"epoch": 0.024193930211637706,
|
| 3583 |
+
"grad_norm": 1.0020304918289185,
|
| 3584 |
+
"learning_rate": 6.908032976059183e-05,
|
| 3585 |
+
"loss": 1.3593,
|
| 3586 |
+
"step": 511
|
| 3587 |
+
},
|
| 3588 |
+
{
|
| 3589 |
+
"epoch": 0.024241276454713318,
|
| 3590 |
+
"grad_norm": 1.116170048713684,
|
| 3591 |
+
"learning_rate": 6.897256839228101e-05,
|
| 3592 |
+
"loss": 1.3145,
|
| 3593 |
+
"step": 512
|
| 3594 |
+
},
|
| 3595 |
+
{
|
| 3596 |
+
"epoch": 0.02428862269778893,
|
| 3597 |
+
"grad_norm": 1.0256245136260986,
|
| 3598 |
+
"learning_rate": 6.886470397446958e-05,
|
| 3599 |
+
"loss": 1.195,
|
| 3600 |
+
"step": 513
|
| 3601 |
+
},
|
| 3602 |
+
{
|
| 3603 |
+
"epoch": 0.024335968940864542,
|
| 3604 |
+
"grad_norm": 1.0521621704101562,
|
| 3605 |
+
"learning_rate": 6.875673709302311e-05,
|
| 3606 |
+
"loss": 1.4941,
|
| 3607 |
+
"step": 514
|
| 3608 |
+
},
|
| 3609 |
+
{
|
| 3610 |
+
"epoch": 0.024383315183940154,
|
| 3611 |
+
"grad_norm": 1.030476450920105,
|
| 3612 |
+
"learning_rate": 6.864866833436368e-05,
|
| 3613 |
+
"loss": 1.2982,
|
| 3614 |
+
"step": 515
|
| 3615 |
+
},
|
| 3616 |
+
{
|
| 3617 |
+
"epoch": 0.024430661427015766,
|
| 3618 |
+
"grad_norm": 1.0907087326049805,
|
| 3619 |
+
"learning_rate": 6.854049828546674e-05,
|
| 3620 |
+
"loss": 1.4964,
|
| 3621 |
+
"step": 516
|
| 3622 |
+
},
|
| 3623 |
+
{
|
| 3624 |
+
"epoch": 0.024478007670091378,
|
| 3625 |
+
"grad_norm": 1.2865772247314453,
|
| 3626 |
+
"learning_rate": 6.843222753385786e-05,
|
| 3627 |
+
"loss": 1.638,
|
| 3628 |
+
"step": 517
|
| 3629 |
+
},
|
| 3630 |
+
{
|
| 3631 |
+
"epoch": 0.02452535391316699,
|
| 3632 |
+
"grad_norm": 1.044796109199524,
|
| 3633 |
+
"learning_rate": 6.832385666760962e-05,
|
| 3634 |
+
"loss": 1.4634,
|
| 3635 |
+
"step": 518
|
| 3636 |
+
},
|
| 3637 |
+
{
|
| 3638 |
+
"epoch": 0.0245727001562426,
|
| 3639 |
+
"grad_norm": 1.0215106010437012,
|
| 3640 |
+
"learning_rate": 6.821538627533833e-05,
|
| 3641 |
+
"loss": 1.2617,
|
| 3642 |
+
"step": 519
|
| 3643 |
+
},
|
| 3644 |
+
{
|
| 3645 |
+
"epoch": 0.024620046399318214,
|
| 3646 |
+
"grad_norm": 1.2929142713546753,
|
| 3647 |
+
"learning_rate": 6.810681694620093e-05,
|
| 3648 |
+
"loss": 1.5473,
|
| 3649 |
+
"step": 520
|
| 3650 |
+
},
|
| 3651 |
+
{
|
| 3652 |
+
"epoch": 0.024667392642393825,
|
| 3653 |
+
"grad_norm": 1.5338375568389893,
|
| 3654 |
+
"learning_rate": 6.79981492698917e-05,
|
| 3655 |
+
"loss": 1.8097,
|
| 3656 |
+
"step": 521
|
| 3657 |
+
},
|
| 3658 |
+
{
|
| 3659 |
+
"epoch": 0.024714738885469437,
|
| 3660 |
+
"grad_norm": 1.2600690126419067,
|
| 3661 |
+
"learning_rate": 6.78893838366391e-05,
|
| 3662 |
+
"loss": 1.4513,
|
| 3663 |
+
"step": 522
|
| 3664 |
+
},
|
| 3665 |
+
{
|
| 3666 |
+
"epoch": 0.02476208512854505,
|
| 3667 |
+
"grad_norm": 1.225058674812317,
|
| 3668 |
+
"learning_rate": 6.778052123720252e-05,
|
| 3669 |
+
"loss": 1.6699,
|
| 3670 |
+
"step": 523
|
| 3671 |
+
},
|
| 3672 |
+
{
|
| 3673 |
+
"epoch": 0.02480943137162066,
|
| 3674 |
+
"grad_norm": 1.2140684127807617,
|
| 3675 |
+
"learning_rate": 6.767156206286917e-05,
|
| 3676 |
+
"loss": 1.6982,
|
| 3677 |
+
"step": 524
|
| 3678 |
+
},
|
| 3679 |
+
{
|
| 3680 |
+
"epoch": 0.024856777614696273,
|
| 3681 |
+
"grad_norm": 1.104151964187622,
|
| 3682 |
+
"learning_rate": 6.756250690545079e-05,
|
| 3683 |
+
"loss": 1.6381,
|
| 3684 |
+
"step": 525
|
| 3685 |
+
},
|
| 3686 |
+
{
|
| 3687 |
+
"epoch": 0.024904123857771885,
|
| 3688 |
+
"grad_norm": 1.1987253427505493,
|
| 3689 |
+
"learning_rate": 6.745335635728041e-05,
|
| 3690 |
+
"loss": 1.7105,
|
| 3691 |
+
"step": 526
|
| 3692 |
+
},
|
| 3693 |
+
{
|
| 3694 |
+
"epoch": 0.024951470100847497,
|
| 3695 |
+
"grad_norm": 1.0873751640319824,
|
| 3696 |
+
"learning_rate": 6.734411101120924e-05,
|
| 3697 |
+
"loss": 1.6234,
|
| 3698 |
+
"step": 527
|
| 3699 |
+
},
|
| 3700 |
+
{
|
| 3701 |
+
"epoch": 0.02499881634392311,
|
| 3702 |
+
"grad_norm": 1.1724483966827393,
|
| 3703 |
+
"learning_rate": 6.723477146060333e-05,
|
| 3704 |
+
"loss": 1.4267,
|
| 3705 |
+
"step": 528
|
| 3706 |
+
},
|
| 3707 |
+
{
|
| 3708 |
+
"epoch": 0.02504616258699872,
|
| 3709 |
+
"grad_norm": 1.009129285812378,
|
| 3710 |
+
"learning_rate": 6.712533829934042e-05,
|
| 3711 |
+
"loss": 1.347,
|
| 3712 |
+
"step": 529
|
| 3713 |
+
},
|
| 3714 |
+
{
|
| 3715 |
+
"epoch": 0.025093508830074333,
|
| 3716 |
+
"grad_norm": 1.1746597290039062,
|
| 3717 |
+
"learning_rate": 6.701581212180671e-05,
|
| 3718 |
+
"loss": 1.6072,
|
| 3719 |
+
"step": 530
|
| 3720 |
+
},
|
| 3721 |
+
{
|
| 3722 |
+
"epoch": 0.025140855073149945,
|
| 3723 |
+
"grad_norm": 1.082383394241333,
|
| 3724 |
+
"learning_rate": 6.690619352289359e-05,
|
| 3725 |
+
"loss": 1.5734,
|
| 3726 |
+
"step": 531
|
| 3727 |
+
},
|
| 3728 |
+
{
|
| 3729 |
+
"epoch": 0.025188201316225556,
|
| 3730 |
+
"grad_norm": 1.2403680086135864,
|
| 3731 |
+
"learning_rate": 6.679648309799447e-05,
|
| 3732 |
+
"loss": 1.8346,
|
| 3733 |
+
"step": 532
|
| 3734 |
+
},
|
| 3735 |
+
{
|
| 3736 |
+
"epoch": 0.02523554755930117,
|
| 3737 |
+
"grad_norm": 1.135392427444458,
|
| 3738 |
+
"learning_rate": 6.668668144300149e-05,
|
| 3739 |
+
"loss": 1.4422,
|
| 3740 |
+
"step": 533
|
| 3741 |
+
},
|
| 3742 |
+
{
|
| 3743 |
+
"epoch": 0.02528289380237678,
|
| 3744 |
+
"grad_norm": 1.0954493284225464,
|
| 3745 |
+
"learning_rate": 6.65767891543023e-05,
|
| 3746 |
+
"loss": 1.5561,
|
| 3747 |
+
"step": 534
|
| 3748 |
+
},
|
| 3749 |
+
{
|
| 3750 |
+
"epoch": 0.025330240045452392,
|
| 3751 |
+
"grad_norm": 1.1450875997543335,
|
| 3752 |
+
"learning_rate": 6.646680682877687e-05,
|
| 3753 |
+
"loss": 1.4436,
|
| 3754 |
+
"step": 535
|
| 3755 |
+
},
|
| 3756 |
+
{
|
| 3757 |
+
"epoch": 0.025377586288528004,
|
| 3758 |
+
"grad_norm": 1.028868317604065,
|
| 3759 |
+
"learning_rate": 6.635673506379415e-05,
|
| 3760 |
+
"loss": 1.4269,
|
| 3761 |
+
"step": 536
|
| 3762 |
+
},
|
| 3763 |
+
{
|
| 3764 |
+
"epoch": 0.025424932531603616,
|
| 3765 |
+
"grad_norm": 1.1591053009033203,
|
| 3766 |
+
"learning_rate": 6.62465744572089e-05,
|
| 3767 |
+
"loss": 1.4876,
|
| 3768 |
+
"step": 537
|
| 3769 |
+
},
|
| 3770 |
+
{
|
| 3771 |
+
"epoch": 0.025472278774679228,
|
| 3772 |
+
"grad_norm": 1.1316996812820435,
|
| 3773 |
+
"learning_rate": 6.613632560735845e-05,
|
| 3774 |
+
"loss": 1.593,
|
| 3775 |
+
"step": 538
|
| 3776 |
+
},
|
| 3777 |
+
{
|
| 3778 |
+
"epoch": 0.02551962501775484,
|
| 3779 |
+
"grad_norm": 1.2574400901794434,
|
| 3780 |
+
"learning_rate": 6.602598911305938e-05,
|
| 3781 |
+
"loss": 1.5639,
|
| 3782 |
+
"step": 539
|
| 3783 |
+
},
|
| 3784 |
+
{
|
| 3785 |
+
"epoch": 0.025566971260830452,
|
| 3786 |
+
"grad_norm": 1.1778851747512817,
|
| 3787 |
+
"learning_rate": 6.591556557360436e-05,
|
| 3788 |
+
"loss": 1.5958,
|
| 3789 |
+
"step": 540
|
| 3790 |
+
},
|
| 3791 |
+
{
|
| 3792 |
+
"epoch": 0.025614317503906064,
|
| 3793 |
+
"grad_norm": 1.0932775735855103,
|
| 3794 |
+
"learning_rate": 6.580505558875877e-05,
|
| 3795 |
+
"loss": 1.3915,
|
| 3796 |
+
"step": 541
|
| 3797 |
+
},
|
| 3798 |
+
{
|
| 3799 |
+
"epoch": 0.025661663746981676,
|
| 3800 |
+
"grad_norm": 0.9905120134353638,
|
| 3801 |
+
"learning_rate": 6.569445975875763e-05,
|
| 3802 |
+
"loss": 1.2847,
|
| 3803 |
+
"step": 542
|
| 3804 |
+
},
|
| 3805 |
+
{
|
| 3806 |
+
"epoch": 0.025709009990057288,
|
| 3807 |
+
"grad_norm": 1.07041597366333,
|
| 3808 |
+
"learning_rate": 6.558377868430211e-05,
|
| 3809 |
+
"loss": 1.4003,
|
| 3810 |
+
"step": 543
|
| 3811 |
+
},
|
| 3812 |
+
{
|
| 3813 |
+
"epoch": 0.0257563562331329,
|
| 3814 |
+
"grad_norm": 1.1834315061569214,
|
| 3815 |
+
"learning_rate": 6.547301296655647e-05,
|
| 3816 |
+
"loss": 1.4869,
|
| 3817 |
+
"step": 544
|
| 3818 |
+
},
|
| 3819 |
+
{
|
| 3820 |
+
"epoch": 0.02580370247620851,
|
| 3821 |
+
"grad_norm": 1.2603247165679932,
|
| 3822 |
+
"learning_rate": 6.536216320714466e-05,
|
| 3823 |
+
"loss": 1.2922,
|
| 3824 |
+
"step": 545
|
| 3825 |
+
},
|
| 3826 |
+
{
|
| 3827 |
+
"epoch": 0.025851048719284123,
|
| 3828 |
+
"grad_norm": 1.1863409280776978,
|
| 3829 |
+
"learning_rate": 6.525123000814712e-05,
|
| 3830 |
+
"loss": 1.2475,
|
| 3831 |
+
"step": 546
|
| 3832 |
+
},
|
| 3833 |
+
{
|
| 3834 |
+
"epoch": 0.025898394962359735,
|
| 3835 |
+
"grad_norm": 1.3121163845062256,
|
| 3836 |
+
"learning_rate": 6.51402139720975e-05,
|
| 3837 |
+
"loss": 1.3717,
|
| 3838 |
+
"step": 547
|
| 3839 |
+
},
|
| 3840 |
+
{
|
| 3841 |
+
"epoch": 0.025945741205435347,
|
| 3842 |
+
"grad_norm": 1.8999866247177124,
|
| 3843 |
+
"learning_rate": 6.502911570197939e-05,
|
| 3844 |
+
"loss": 1.3112,
|
| 3845 |
+
"step": 548
|
| 3846 |
+
},
|
| 3847 |
+
{
|
| 3848 |
+
"epoch": 0.02599308744851096,
|
| 3849 |
+
"grad_norm": 1.1127601861953735,
|
| 3850 |
+
"learning_rate": 6.491793580122301e-05,
|
| 3851 |
+
"loss": 1.334,
|
| 3852 |
+
"step": 549
|
| 3853 |
+
},
|
| 3854 |
+
{
|
| 3855 |
+
"epoch": 0.02604043369158657,
|
| 3856 |
+
"grad_norm": 1.446142554283142,
|
| 3857 |
+
"learning_rate": 6.480667487370193e-05,
|
| 3858 |
+
"loss": 1.1253,
|
| 3859 |
+
"step": 550
|
| 3860 |
+
},
|
| 3861 |
+
{
|
| 3862 |
+
"epoch": 0.026087779934662183,
|
| 3863 |
+
"grad_norm": 1.0348849296569824,
|
| 3864 |
+
"learning_rate": 6.469533352372989e-05,
|
| 3865 |
+
"loss": 1.7429,
|
| 3866 |
+
"step": 551
|
| 3867 |
+
},
|
| 3868 |
+
{
|
| 3869 |
+
"epoch": 0.026135126177737795,
|
| 3870 |
+
"grad_norm": 1.0058255195617676,
|
| 3871 |
+
"learning_rate": 6.458391235605739e-05,
|
| 3872 |
+
"loss": 1.5335,
|
| 3873 |
+
"step": 552
|
| 3874 |
+
},
|
| 3875 |
+
{
|
| 3876 |
+
"epoch": 0.026182472420813407,
|
| 3877 |
+
"grad_norm": 1.1435742378234863,
|
| 3878 |
+
"learning_rate": 6.447241197586847e-05,
|
| 3879 |
+
"loss": 1.5263,
|
| 3880 |
+
"step": 553
|
| 3881 |
+
},
|
| 3882 |
+
{
|
| 3883 |
+
"epoch": 0.026229818663889022,
|
| 3884 |
+
"grad_norm": 1.0566860437393188,
|
| 3885 |
+
"learning_rate": 6.436083298877744e-05,
|
| 3886 |
+
"loss": 1.3886,
|
| 3887 |
+
"step": 554
|
| 3888 |
+
},
|
| 3889 |
+
{
|
| 3890 |
+
"epoch": 0.026277164906964634,
|
| 3891 |
+
"grad_norm": 1.139036774635315,
|
| 3892 |
+
"learning_rate": 6.424917600082551e-05,
|
| 3893 |
+
"loss": 1.3902,
|
| 3894 |
+
"step": 555
|
| 3895 |
+
},
|
| 3896 |
+
{
|
| 3897 |
+
"epoch": 0.026324511150040246,
|
| 3898 |
+
"grad_norm": 1.220698356628418,
|
| 3899 |
+
"learning_rate": 6.413744161847761e-05,
|
| 3900 |
+
"loss": 1.4474,
|
| 3901 |
+
"step": 556
|
| 3902 |
+
},
|
| 3903 |
+
{
|
| 3904 |
+
"epoch": 0.026371857393115858,
|
| 3905 |
+
"grad_norm": 1.146551489830017,
|
| 3906 |
+
"learning_rate": 6.4025630448619e-05,
|
| 3907 |
+
"loss": 1.5885,
|
| 3908 |
+
"step": 557
|
| 3909 |
+
},
|
| 3910 |
+
{
|
| 3911 |
+
"epoch": 0.02641920363619147,
|
| 3912 |
+
"grad_norm": 1.2537118196487427,
|
| 3913 |
+
"learning_rate": 6.391374309855201e-05,
|
| 3914 |
+
"loss": 1.6038,
|
| 3915 |
+
"step": 558
|
| 3916 |
+
},
|
| 3917 |
+
{
|
| 3918 |
+
"epoch": 0.02646654987926708,
|
| 3919 |
+
"grad_norm": 1.116015076637268,
|
| 3920 |
+
"learning_rate": 6.380178017599276e-05,
|
| 3921 |
+
"loss": 1.4602,
|
| 3922 |
+
"step": 559
|
| 3923 |
+
},
|
| 3924 |
+
{
|
| 3925 |
+
"epoch": 0.026513896122342694,
|
| 3926 |
+
"grad_norm": 1.0770164728164673,
|
| 3927 |
+
"learning_rate": 6.368974228906785e-05,
|
| 3928 |
+
"loss": 1.5029,
|
| 3929 |
+
"step": 560
|
| 3930 |
+
},
|
| 3931 |
+
{
|
| 3932 |
+
"epoch": 0.026561242365418306,
|
| 3933 |
+
"grad_norm": 1.1019155979156494,
|
| 3934 |
+
"learning_rate": 6.357763004631104e-05,
|
| 3935 |
+
"loss": 1.5474,
|
| 3936 |
+
"step": 561
|
| 3937 |
+
},
|
| 3938 |
+
{
|
| 3939 |
+
"epoch": 0.026608588608493917,
|
| 3940 |
+
"grad_norm": 1.127123475074768,
|
| 3941 |
+
"learning_rate": 6.346544405665992e-05,
|
| 3942 |
+
"loss": 1.6436,
|
| 3943 |
+
"step": 562
|
| 3944 |
+
},
|
| 3945 |
+
{
|
| 3946 |
+
"epoch": 0.02665593485156953,
|
| 3947 |
+
"grad_norm": 1.0497205257415771,
|
| 3948 |
+
"learning_rate": 6.335318492945271e-05,
|
| 3949 |
+
"loss": 1.3378,
|
| 3950 |
+
"step": 563
|
| 3951 |
+
},
|
| 3952 |
+
{
|
| 3953 |
+
"epoch": 0.02670328109464514,
|
| 3954 |
+
"grad_norm": 1.2064182758331299,
|
| 3955 |
+
"learning_rate": 6.324085327442477e-05,
|
| 3956 |
+
"loss": 1.3007,
|
| 3957 |
+
"step": 564
|
| 3958 |
+
},
|
| 3959 |
+
{
|
| 3960 |
+
"epoch": 0.026750627337720753,
|
| 3961 |
+
"grad_norm": 1.085641860961914,
|
| 3962 |
+
"learning_rate": 6.31284497017055e-05,
|
| 3963 |
+
"loss": 1.3083,
|
| 3964 |
+
"step": 565
|
| 3965 |
+
},
|
| 3966 |
+
{
|
| 3967 |
+
"epoch": 0.026797973580796365,
|
| 3968 |
+
"grad_norm": 1.1102137565612793,
|
| 3969 |
+
"learning_rate": 6.30159748218149e-05,
|
| 3970 |
+
"loss": 1.3017,
|
| 3971 |
+
"step": 566
|
| 3972 |
+
},
|
| 3973 |
+
{
|
| 3974 |
+
"epoch": 0.026845319823871977,
|
| 3975 |
+
"grad_norm": 1.2144014835357666,
|
| 3976 |
+
"learning_rate": 6.290342924566019e-05,
|
| 3977 |
+
"loss": 1.7009,
|
| 3978 |
+
"step": 567
|
| 3979 |
+
},
|
| 3980 |
+
{
|
| 3981 |
+
"epoch": 0.02689266606694759,
|
| 3982 |
+
"grad_norm": 1.1694127321243286,
|
| 3983 |
+
"learning_rate": 6.279081358453268e-05,
|
| 3984 |
+
"loss": 1.3694,
|
| 3985 |
+
"step": 568
|
| 3986 |
+
},
|
| 3987 |
+
{
|
| 3988 |
+
"epoch": 0.0269400123100232,
|
| 3989 |
+
"grad_norm": 1.145224690437317,
|
| 3990 |
+
"learning_rate": 6.26781284501043e-05,
|
| 3991 |
+
"loss": 1.5255,
|
| 3992 |
+
"step": 569
|
| 3993 |
+
},
|
| 3994 |
+
{
|
| 3995 |
+
"epoch": 0.026987358553098813,
|
| 3996 |
+
"grad_norm": 1.0720521211624146,
|
| 3997 |
+
"learning_rate": 6.256537445442432e-05,
|
| 3998 |
+
"loss": 1.3011,
|
| 3999 |
+
"step": 570
|
| 4000 |
+
},
|
| 4001 |
+
{
|
| 4002 |
+
"epoch": 0.027034704796174425,
|
| 4003 |
+
"grad_norm": 1.1498228311538696,
|
| 4004 |
+
"learning_rate": 6.245255220991606e-05,
|
| 4005 |
+
"loss": 1.6562,
|
| 4006 |
+
"step": 571
|
| 4007 |
+
},
|
| 4008 |
+
{
|
| 4009 |
+
"epoch": 0.027082051039250037,
|
| 4010 |
+
"grad_norm": 1.1388801336288452,
|
| 4011 |
+
"learning_rate": 6.233966232937349e-05,
|
| 4012 |
+
"loss": 1.4365,
|
| 4013 |
+
"step": 572
|
| 4014 |
+
},
|
| 4015 |
+
{
|
| 4016 |
+
"epoch": 0.02712939728232565,
|
| 4017 |
+
"grad_norm": 1.1685620546340942,
|
| 4018 |
+
"learning_rate": 6.2226705425958e-05,
|
| 4019 |
+
"loss": 1.4001,
|
| 4020 |
+
"step": 573
|
| 4021 |
+
},
|
| 4022 |
+
{
|
| 4023 |
+
"epoch": 0.02717674352540126,
|
| 4024 |
+
"grad_norm": 1.1403052806854248,
|
| 4025 |
+
"learning_rate": 6.211368211319495e-05,
|
| 4026 |
+
"loss": 1.584,
|
| 4027 |
+
"step": 574
|
| 4028 |
+
},
|
| 4029 |
+
{
|
| 4030 |
+
"epoch": 0.027224089768476872,
|
| 4031 |
+
"grad_norm": 1.2580426931381226,
|
| 4032 |
+
"learning_rate": 6.200059300497045e-05,
|
| 4033 |
+
"loss": 1.6061,
|
| 4034 |
+
"step": 575
|
| 4035 |
+
},
|
| 4036 |
+
{
|
| 4037 |
+
"epoch": 0.027271436011552484,
|
| 4038 |
+
"grad_norm": 1.106146216392517,
|
| 4039 |
+
"learning_rate": 6.188743871552796e-05,
|
| 4040 |
+
"loss": 1.6067,
|
| 4041 |
+
"step": 576
|
| 4042 |
+
},
|
| 4043 |
+
{
|
| 4044 |
+
"epoch": 0.027318782254628096,
|
| 4045 |
+
"grad_norm": 1.1391223669052124,
|
| 4046 |
+
"learning_rate": 6.177421985946499e-05,
|
| 4047 |
+
"loss": 1.7163,
|
| 4048 |
+
"step": 577
|
| 4049 |
+
},
|
| 4050 |
+
{
|
| 4051 |
+
"epoch": 0.027366128497703708,
|
| 4052 |
+
"grad_norm": 1.0916639566421509,
|
| 4053 |
+
"learning_rate": 6.166093705172971e-05,
|
| 4054 |
+
"loss": 1.4578,
|
| 4055 |
+
"step": 578
|
| 4056 |
+
},
|
| 4057 |
+
{
|
| 4058 |
+
"epoch": 0.02741347474077932,
|
| 4059 |
+
"grad_norm": 1.0076122283935547,
|
| 4060 |
+
"learning_rate": 6.154759090761768e-05,
|
| 4061 |
+
"loss": 1.4665,
|
| 4062 |
+
"step": 579
|
| 4063 |
+
},
|
| 4064 |
+
{
|
| 4065 |
+
"epoch": 0.027460820983854932,
|
| 4066 |
+
"grad_norm": 1.1860275268554688,
|
| 4067 |
+
"learning_rate": 6.143418204276843e-05,
|
| 4068 |
+
"loss": 1.5437,
|
| 4069 |
+
"step": 580
|
| 4070 |
+
},
|
| 4071 |
+
{
|
| 4072 |
+
"epoch": 0.027508167226930544,
|
| 4073 |
+
"grad_norm": 1.1275054216384888,
|
| 4074 |
+
"learning_rate": 6.13207110731622e-05,
|
| 4075 |
+
"loss": 1.6615,
|
| 4076 |
+
"step": 581
|
| 4077 |
+
},
|
| 4078 |
+
{
|
| 4079 |
+
"epoch": 0.027555513470006156,
|
| 4080 |
+
"grad_norm": 1.229531168937683,
|
| 4081 |
+
"learning_rate": 6.120717861511655e-05,
|
| 4082 |
+
"loss": 1.6994,
|
| 4083 |
+
"step": 582
|
| 4084 |
+
},
|
| 4085 |
+
{
|
| 4086 |
+
"epoch": 0.027602859713081768,
|
| 4087 |
+
"grad_norm": 1.1395683288574219,
|
| 4088 |
+
"learning_rate": 6.109358528528296e-05,
|
| 4089 |
+
"loss": 1.6029,
|
| 4090 |
+
"step": 583
|
| 4091 |
+
},
|
| 4092 |
+
{
|
| 4093 |
+
"epoch": 0.02765020595615738,
|
| 4094 |
+
"grad_norm": 1.202475905418396,
|
| 4095 |
+
"learning_rate": 6.09799317006436e-05,
|
| 4096 |
+
"loss": 1.4413,
|
| 4097 |
+
"step": 584
|
| 4098 |
+
},
|
| 4099 |
+
{
|
| 4100 |
+
"epoch": 0.02769755219923299,
|
| 4101 |
+
"grad_norm": 1.087605595588684,
|
| 4102 |
+
"learning_rate": 6.086621847850788e-05,
|
| 4103 |
+
"loss": 1.4957,
|
| 4104 |
+
"step": 585
|
| 4105 |
+
},
|
| 4106 |
+
{
|
| 4107 |
+
"epoch": 0.027744898442308603,
|
| 4108 |
+
"grad_norm": 1.20675790309906,
|
| 4109 |
+
"learning_rate": 6.075244623650913e-05,
|
| 4110 |
+
"loss": 1.487,
|
| 4111 |
+
"step": 586
|
| 4112 |
+
},
|
| 4113 |
+
{
|
| 4114 |
+
"epoch": 0.027792244685384215,
|
| 4115 |
+
"grad_norm": 1.214881181716919,
|
| 4116 |
+
"learning_rate": 6.0638615592601265e-05,
|
| 4117 |
+
"loss": 1.5525,
|
| 4118 |
+
"step": 587
|
| 4119 |
+
},
|
| 4120 |
+
{
|
| 4121 |
+
"epoch": 0.027839590928459827,
|
| 4122 |
+
"grad_norm": 1.6232247352600098,
|
| 4123 |
+
"learning_rate": 6.0524727165055395e-05,
|
| 4124 |
+
"loss": 1.5852,
|
| 4125 |
+
"step": 588
|
| 4126 |
+
},
|
| 4127 |
+
{
|
| 4128 |
+
"epoch": 0.02788693717153544,
|
| 4129 |
+
"grad_norm": 1.1975793838500977,
|
| 4130 |
+
"learning_rate": 6.0410781572456486e-05,
|
| 4131 |
+
"loss": 1.3506,
|
| 4132 |
+
"step": 589
|
| 4133 |
+
},
|
| 4134 |
+
{
|
| 4135 |
+
"epoch": 0.02793428341461105,
|
| 4136 |
+
"grad_norm": 1.2239601612091064,
|
| 4137 |
+
"learning_rate": 6.029677943370001e-05,
|
| 4138 |
+
"loss": 1.3936,
|
| 4139 |
+
"step": 590
|
| 4140 |
+
},
|
| 4141 |
+
{
|
| 4142 |
+
"epoch": 0.027981629657686663,
|
| 4143 |
+
"grad_norm": 1.2145081758499146,
|
| 4144 |
+
"learning_rate": 6.018272136798854e-05,
|
| 4145 |
+
"loss": 1.3082,
|
| 4146 |
+
"step": 591
|
| 4147 |
+
},
|
| 4148 |
+
{
|
| 4149 |
+
"epoch": 0.028028975900762275,
|
| 4150 |
+
"grad_norm": 1.1785577535629272,
|
| 4151 |
+
"learning_rate": 6.006860799482844e-05,
|
| 4152 |
+
"loss": 1.2617,
|
| 4153 |
+
"step": 592
|
| 4154 |
+
},
|
| 4155 |
+
{
|
| 4156 |
+
"epoch": 0.028076322143837887,
|
| 4157 |
+
"grad_norm": 1.4417402744293213,
|
| 4158 |
+
"learning_rate": 5.995443993402647e-05,
|
| 4159 |
+
"loss": 1.3944,
|
| 4160 |
+
"step": 593
|
| 4161 |
+
},
|
| 4162 |
+
{
|
| 4163 |
+
"epoch": 0.0281236683869135,
|
| 4164 |
+
"grad_norm": 1.0939041376113892,
|
| 4165 |
+
"learning_rate": 5.9840217805686426e-05,
|
| 4166 |
+
"loss": 1.322,
|
| 4167 |
+
"step": 594
|
| 4168 |
+
},
|
| 4169 |
+
{
|
| 4170 |
+
"epoch": 0.02817101462998911,
|
| 4171 |
+
"grad_norm": 1.2456567287445068,
|
| 4172 |
+
"learning_rate": 5.972594223020576e-05,
|
| 4173 |
+
"loss": 1.3403,
|
| 4174 |
+
"step": 595
|
| 4175 |
+
},
|
| 4176 |
+
{
|
| 4177 |
+
"epoch": 0.028218360873064723,
|
| 4178 |
+
"grad_norm": 1.145764708518982,
|
| 4179 |
+
"learning_rate": 5.961161382827223e-05,
|
| 4180 |
+
"loss": 1.1613,
|
| 4181 |
+
"step": 596
|
| 4182 |
+
},
|
| 4183 |
+
{
|
| 4184 |
+
"epoch": 0.028265707116140334,
|
| 4185 |
+
"grad_norm": 1.1802617311477661,
|
| 4186 |
+
"learning_rate": 5.949723322086053e-05,
|
| 4187 |
+
"loss": 1.1702,
|
| 4188 |
+
"step": 597
|
| 4189 |
+
},
|
| 4190 |
+
{
|
| 4191 |
+
"epoch": 0.028313053359215946,
|
| 4192 |
+
"grad_norm": 1.3801965713500977,
|
| 4193 |
+
"learning_rate": 5.9382801029228904e-05,
|
| 4194 |
+
"loss": 1.599,
|
| 4195 |
+
"step": 598
|
| 4196 |
+
},
|
| 4197 |
+
{
|
| 4198 |
+
"epoch": 0.02836039960229156,
|
| 4199 |
+
"grad_norm": 1.3491289615631104,
|
| 4200 |
+
"learning_rate": 5.926831787491577e-05,
|
| 4201 |
+
"loss": 1.3966,
|
| 4202 |
+
"step": 599
|
| 4203 |
+
},
|
| 4204 |
+
{
|
| 4205 |
+
"epoch": 0.02840774584536717,
|
| 4206 |
+
"grad_norm": 1.4206387996673584,
|
| 4207 |
+
"learning_rate": 5.9153784379736355e-05,
|
| 4208 |
+
"loss": 1.2928,
|
| 4209 |
+
"step": 600
|
| 4210 |
+
},
|
| 4211 |
+
{
|
| 4212 |
+
"epoch": 0.028455092088442782,
|
| 4213 |
+
"grad_norm": 1.088247299194336,
|
| 4214 |
+
"learning_rate": 5.9039201165779315e-05,
|
| 4215 |
+
"loss": 1.7898,
|
| 4216 |
+
"step": 601
|
| 4217 |
+
},
|
| 4218 |
+
{
|
| 4219 |
+
"epoch": 0.028502438331518394,
|
| 4220 |
+
"grad_norm": 1.12989342212677,
|
| 4221 |
+
"learning_rate": 5.892456885540333e-05,
|
| 4222 |
+
"loss": 1.6844,
|
| 4223 |
+
"step": 602
|
| 4224 |
+
},
|
| 4225 |
+
{
|
| 4226 |
+
"epoch": 0.028549784574594006,
|
| 4227 |
+
"grad_norm": 1.0441310405731201,
|
| 4228 |
+
"learning_rate": 5.8809888071233785e-05,
|
| 4229 |
+
"loss": 1.392,
|
| 4230 |
+
"step": 603
|
| 4231 |
+
},
|
| 4232 |
+
{
|
| 4233 |
+
"epoch": 0.028597130817669618,
|
| 4234 |
+
"grad_norm": 0.9600141048431396,
|
| 4235 |
+
"learning_rate": 5.869515943615932e-05,
|
| 4236 |
+
"loss": 1.5506,
|
| 4237 |
+
"step": 604
|
| 4238 |
+
},
|
| 4239 |
+
{
|
| 4240 |
+
"epoch": 0.02864447706074523,
|
| 4241 |
+
"grad_norm": 1.1483334302902222,
|
| 4242 |
+
"learning_rate": 5.85803835733285e-05,
|
| 4243 |
+
"loss": 1.5322,
|
| 4244 |
+
"step": 605
|
| 4245 |
+
},
|
| 4246 |
+
{
|
| 4247 |
+
"epoch": 0.02869182330382084,
|
| 4248 |
+
"grad_norm": 1.0461188554763794,
|
| 4249 |
+
"learning_rate": 5.8465561106146386e-05,
|
| 4250 |
+
"loss": 1.404,
|
| 4251 |
+
"step": 606
|
| 4252 |
+
},
|
| 4253 |
+
{
|
| 4254 |
+
"epoch": 0.028739169546896454,
|
| 4255 |
+
"grad_norm": 1.073137640953064,
|
| 4256 |
+
"learning_rate": 5.835069265827119e-05,
|
| 4257 |
+
"loss": 1.3054,
|
| 4258 |
+
"step": 607
|
| 4259 |
+
},
|
| 4260 |
+
{
|
| 4261 |
+
"epoch": 0.028786515789972066,
|
| 4262 |
+
"grad_norm": 1.1550793647766113,
|
| 4263 |
+
"learning_rate": 5.8235778853610845e-05,
|
| 4264 |
+
"loss": 1.6782,
|
| 4265 |
+
"step": 608
|
| 4266 |
+
},
|
| 4267 |
+
{
|
| 4268 |
+
"epoch": 0.028833862033047677,
|
| 4269 |
+
"grad_norm": 1.1815844774246216,
|
| 4270 |
+
"learning_rate": 5.812082031631966e-05,
|
| 4271 |
+
"loss": 1.7661,
|
| 4272 |
+
"step": 609
|
| 4273 |
+
},
|
| 4274 |
+
{
|
| 4275 |
+
"epoch": 0.02888120827612329,
|
| 4276 |
+
"grad_norm": 1.0953624248504639,
|
| 4277 |
+
"learning_rate": 5.80058176707949e-05,
|
| 4278 |
+
"loss": 1.3478,
|
| 4279 |
+
"step": 610
|
| 4280 |
+
},
|
| 4281 |
+
{
|
| 4282 |
+
"epoch": 0.0289285545191989,
|
| 4283 |
+
"grad_norm": 0.9961740970611572,
|
| 4284 |
+
"learning_rate": 5.7890771541673415e-05,
|
| 4285 |
+
"loss": 1.1861,
|
| 4286 |
+
"step": 611
|
| 4287 |
+
},
|
| 4288 |
+
{
|
| 4289 |
+
"epoch": 0.028975900762274513,
|
| 4290 |
+
"grad_norm": 1.1952297687530518,
|
| 4291 |
+
"learning_rate": 5.7775682553828224e-05,
|
| 4292 |
+
"loss": 1.5975,
|
| 4293 |
+
"step": 612
|
| 4294 |
+
},
|
| 4295 |
+
{
|
| 4296 |
+
"epoch": 0.029023247005350125,
|
| 4297 |
+
"grad_norm": 1.2598936557769775,
|
| 4298 |
+
"learning_rate": 5.766055133236513e-05,
|
| 4299 |
+
"loss": 1.5,
|
| 4300 |
+
"step": 613
|
| 4301 |
+
},
|
| 4302 |
+
{
|
| 4303 |
+
"epoch": 0.029070593248425737,
|
| 4304 |
+
"grad_norm": 1.236576795578003,
|
| 4305 |
+
"learning_rate": 5.754537850261934e-05,
|
| 4306 |
+
"loss": 1.5108,
|
| 4307 |
+
"step": 614
|
| 4308 |
+
},
|
| 4309 |
+
{
|
| 4310 |
+
"epoch": 0.02911793949150135,
|
| 4311 |
+
"grad_norm": 1.1170661449432373,
|
| 4312 |
+
"learning_rate": 5.743016469015204e-05,
|
| 4313 |
+
"loss": 1.3738,
|
| 4314 |
+
"step": 615
|
| 4315 |
+
},
|
| 4316 |
+
{
|
| 4317 |
+
"epoch": 0.02916528573457696,
|
| 4318 |
+
"grad_norm": 1.0549659729003906,
|
| 4319 |
+
"learning_rate": 5.731491052074704e-05,
|
| 4320 |
+
"loss": 1.3512,
|
| 4321 |
+
"step": 616
|
| 4322 |
+
},
|
| 4323 |
+
{
|
| 4324 |
+
"epoch": 0.029212631977652573,
|
| 4325 |
+
"grad_norm": 1.2418321371078491,
|
| 4326 |
+
"learning_rate": 5.719961662040733e-05,
|
| 4327 |
+
"loss": 1.6309,
|
| 4328 |
+
"step": 617
|
| 4329 |
+
},
|
| 4330 |
+
{
|
| 4331 |
+
"epoch": 0.029259978220728185,
|
| 4332 |
+
"grad_norm": 1.1811386346817017,
|
| 4333 |
+
"learning_rate": 5.7084283615351694e-05,
|
| 4334 |
+
"loss": 1.5017,
|
| 4335 |
+
"step": 618
|
| 4336 |
+
},
|
| 4337 |
+
{
|
| 4338 |
+
"epoch": 0.029307324463803797,
|
| 4339 |
+
"grad_norm": 1.3130184412002563,
|
| 4340 |
+
"learning_rate": 5.6968912132011334e-05,
|
| 4341 |
+
"loss": 1.5756,
|
| 4342 |
+
"step": 619
|
| 4343 |
+
},
|
| 4344 |
+
{
|
| 4345 |
+
"epoch": 0.02935467070687941,
|
| 4346 |
+
"grad_norm": 1.171229362487793,
|
| 4347 |
+
"learning_rate": 5.685350279702641e-05,
|
| 4348 |
+
"loss": 1.3932,
|
| 4349 |
+
"step": 620
|
| 4350 |
+
},
|
| 4351 |
+
{
|
| 4352 |
+
"epoch": 0.02940201694995502,
|
| 4353 |
+
"grad_norm": 1.0718367099761963,
|
| 4354 |
+
"learning_rate": 5.673805623724272e-05,
|
| 4355 |
+
"loss": 1.2716,
|
| 4356 |
+
"step": 621
|
| 4357 |
+
},
|
| 4358 |
+
{
|
| 4359 |
+
"epoch": 0.029449363193030632,
|
| 4360 |
+
"grad_norm": 1.3737154006958008,
|
| 4361 |
+
"learning_rate": 5.662257307970822e-05,
|
| 4362 |
+
"loss": 1.6383,
|
| 4363 |
+
"step": 622
|
| 4364 |
+
},
|
| 4365 |
+
{
|
| 4366 |
+
"epoch": 0.029496709436106244,
|
| 4367 |
+
"grad_norm": 1.2305822372436523,
|
| 4368 |
+
"learning_rate": 5.650705395166964e-05,
|
| 4369 |
+
"loss": 1.6972,
|
| 4370 |
+
"step": 623
|
| 4371 |
+
},
|
| 4372 |
+
{
|
| 4373 |
+
"epoch": 0.029544055679181856,
|
| 4374 |
+
"grad_norm": 1.3053735494613647,
|
| 4375 |
+
"learning_rate": 5.639149948056911e-05,
|
| 4376 |
+
"loss": 1.5476,
|
| 4377 |
+
"step": 624
|
| 4378 |
+
},
|
| 4379 |
+
{
|
| 4380 |
+
"epoch": 0.029591401922257468,
|
| 4381 |
+
"grad_norm": 1.2132328748703003,
|
| 4382 |
+
"learning_rate": 5.627591029404071e-05,
|
| 4383 |
+
"loss": 1.6751,
|
| 4384 |
+
"step": 625
|
| 4385 |
+
},
|
| 4386 |
+
{
|
| 4387 |
+
"epoch": 0.02963874816533308,
|
| 4388 |
+
"grad_norm": 1.1572853326797485,
|
| 4389 |
+
"learning_rate": 5.616028701990708e-05,
|
| 4390 |
+
"loss": 1.7647,
|
| 4391 |
+
"step": 626
|
| 4392 |
+
},
|
| 4393 |
+
{
|
| 4394 |
+
"epoch": 0.029686094408408692,
|
| 4395 |
+
"grad_norm": 1.0607131719589233,
|
| 4396 |
+
"learning_rate": 5.604463028617598e-05,
|
| 4397 |
+
"loss": 1.3218,
|
| 4398 |
+
"step": 627
|
| 4399 |
+
},
|
| 4400 |
+
{
|
| 4401 |
+
"epoch": 0.029733440651484304,
|
| 4402 |
+
"grad_norm": 1.1433614492416382,
|
| 4403 |
+
"learning_rate": 5.592894072103697e-05,
|
| 4404 |
+
"loss": 1.4143,
|
| 4405 |
+
"step": 628
|
| 4406 |
+
},
|
| 4407 |
+
{
|
| 4408 |
+
"epoch": 0.029780786894559916,
|
| 4409 |
+
"grad_norm": 1.2053921222686768,
|
| 4410 |
+
"learning_rate": 5.581321895285787e-05,
|
| 4411 |
+
"loss": 1.5408,
|
| 4412 |
+
"step": 629
|
| 4413 |
+
},
|
| 4414 |
+
{
|
| 4415 |
+
"epoch": 0.029828133137635528,
|
| 4416 |
+
"grad_norm": 1.139135479927063,
|
| 4417 |
+
"learning_rate": 5.5697465610181445e-05,
|
| 4418 |
+
"loss": 1.5326,
|
| 4419 |
+
"step": 630
|
| 4420 |
+
},
|
| 4421 |
+
{
|
| 4422 |
+
"epoch": 0.02987547938071114,
|
| 4423 |
+
"grad_norm": 1.0945171117782593,
|
| 4424 |
+
"learning_rate": 5.558168132172194e-05,
|
| 4425 |
+
"loss": 1.5827,
|
| 4426 |
+
"step": 631
|
| 4427 |
+
},
|
| 4428 |
+
{
|
| 4429 |
+
"epoch": 0.02992282562378675,
|
| 4430 |
+
"grad_norm": 1.1602097749710083,
|
| 4431 |
+
"learning_rate": 5.5465866716361704e-05,
|
| 4432 |
+
"loss": 1.5157,
|
| 4433 |
+
"step": 632
|
| 4434 |
+
},
|
| 4435 |
+
{
|
| 4436 |
+
"epoch": 0.029970171866862363,
|
| 4437 |
+
"grad_norm": 1.279569387435913,
|
| 4438 |
+
"learning_rate": 5.535002242314772e-05,
|
| 4439 |
+
"loss": 1.4067,
|
| 4440 |
+
"step": 633
|
| 4441 |
+
},
|
| 4442 |
+
{
|
| 4443 |
+
"epoch": 0.030017518109937975,
|
| 4444 |
+
"grad_norm": 1.4093965291976929,
|
| 4445 |
+
"learning_rate": 5.5234149071288245e-05,
|
| 4446 |
+
"loss": 1.6564,
|
| 4447 |
+
"step": 634
|
| 4448 |
+
},
|
| 4449 |
+
{
|
| 4450 |
+
"epoch": 0.030064864353013587,
|
| 4451 |
+
"grad_norm": 1.4158297777175903,
|
| 4452 |
+
"learning_rate": 5.511824729014936e-05,
|
| 4453 |
+
"loss": 1.5333,
|
| 4454 |
+
"step": 635
|
| 4455 |
+
},
|
| 4456 |
+
{
|
| 4457 |
+
"epoch": 0.0301122105960892,
|
| 4458 |
+
"grad_norm": 1.002573847770691,
|
| 4459 |
+
"learning_rate": 5.500231770925157e-05,
|
| 4460 |
+
"loss": 1.1366,
|
| 4461 |
+
"step": 636
|
| 4462 |
+
},
|
| 4463 |
+
{
|
| 4464 |
+
"epoch": 0.03015955683916481,
|
| 4465 |
+
"grad_norm": 1.1367785930633545,
|
| 4466 |
+
"learning_rate": 5.488636095826636e-05,
|
| 4467 |
+
"loss": 1.2451,
|
| 4468 |
+
"step": 637
|
| 4469 |
+
},
|
| 4470 |
+
{
|
| 4471 |
+
"epoch": 0.030206903082240423,
|
| 4472 |
+
"grad_norm": 1.1791261434555054,
|
| 4473 |
+
"learning_rate": 5.477037766701278e-05,
|
| 4474 |
+
"loss": 1.3949,
|
| 4475 |
+
"step": 638
|
| 4476 |
+
},
|
| 4477 |
+
{
|
| 4478 |
+
"epoch": 0.030254249325316035,
|
| 4479 |
+
"grad_norm": 1.093300700187683,
|
| 4480 |
+
"learning_rate": 5.465436846545407e-05,
|
| 4481 |
+
"loss": 1.3358,
|
| 4482 |
+
"step": 639
|
| 4483 |
+
},
|
| 4484 |
+
{
|
| 4485 |
+
"epoch": 0.030301595568391647,
|
| 4486 |
+
"grad_norm": 1.2502402067184448,
|
| 4487 |
+
"learning_rate": 5.453833398369419e-05,
|
| 4488 |
+
"loss": 1.535,
|
| 4489 |
+
"step": 640
|
| 4490 |
+
},
|
| 4491 |
+
{
|
| 4492 |
+
"epoch": 0.03034894181146726,
|
| 4493 |
+
"grad_norm": 1.2500826120376587,
|
| 4494 |
+
"learning_rate": 5.442227485197435e-05,
|
| 4495 |
+
"loss": 1.5574,
|
| 4496 |
+
"step": 641
|
| 4497 |
+
},
|
| 4498 |
+
{
|
| 4499 |
+
"epoch": 0.03039628805454287,
|
| 4500 |
+
"grad_norm": 1.234018325805664,
|
| 4501 |
+
"learning_rate": 5.430619170066975e-05,
|
| 4502 |
+
"loss": 1.2698,
|
| 4503 |
+
"step": 642
|
| 4504 |
+
},
|
| 4505 |
+
{
|
| 4506 |
+
"epoch": 0.030443634297618483,
|
| 4507 |
+
"grad_norm": 1.372062087059021,
|
| 4508 |
+
"learning_rate": 5.419008516028596e-05,
|
| 4509 |
+
"loss": 1.3608,
|
| 4510 |
+
"step": 643
|
| 4511 |
+
},
|
| 4512 |
+
{
|
| 4513 |
+
"epoch": 0.030490980540694095,
|
| 4514 |
+
"grad_norm": 1.2977724075317383,
|
| 4515 |
+
"learning_rate": 5.407395586145566e-05,
|
| 4516 |
+
"loss": 1.1454,
|
| 4517 |
+
"step": 644
|
| 4518 |
+
},
|
| 4519 |
+
{
|
| 4520 |
+
"epoch": 0.030538326783769706,
|
| 4521 |
+
"grad_norm": 1.3347288370132446,
|
| 4522 |
+
"learning_rate": 5.395780443493508e-05,
|
| 4523 |
+
"loss": 1.6372,
|
| 4524 |
+
"step": 645
|
| 4525 |
+
},
|
| 4526 |
+
{
|
| 4527 |
+
"epoch": 0.03058567302684532,
|
| 4528 |
+
"grad_norm": 1.4768805503845215,
|
| 4529 |
+
"learning_rate": 5.384163151160068e-05,
|
| 4530 |
+
"loss": 1.5887,
|
| 4531 |
+
"step": 646
|
| 4532 |
+
},
|
| 4533 |
+
{
|
| 4534 |
+
"epoch": 0.03063301926992093,
|
| 4535 |
+
"grad_norm": 1.2934019565582275,
|
| 4536 |
+
"learning_rate": 5.3725437722445656e-05,
|
| 4537 |
+
"loss": 1.3144,
|
| 4538 |
+
"step": 647
|
| 4539 |
+
},
|
| 4540 |
+
{
|
| 4541 |
+
"epoch": 0.030680365512996542,
|
| 4542 |
+
"grad_norm": 1.254329800605774,
|
| 4543 |
+
"learning_rate": 5.360922369857655e-05,
|
| 4544 |
+
"loss": 1.2691,
|
| 4545 |
+
"step": 648
|
| 4546 |
+
},
|
| 4547 |
+
{
|
| 4548 |
+
"epoch": 0.030727711756072154,
|
| 4549 |
+
"grad_norm": 1.3698008060455322,
|
| 4550 |
+
"learning_rate": 5.3492990071209806e-05,
|
| 4551 |
+
"loss": 1.2337,
|
| 4552 |
+
"step": 649
|
| 4553 |
+
},
|
| 4554 |
+
{
|
| 4555 |
+
"epoch": 0.030775057999147766,
|
| 4556 |
+
"grad_norm": 1.4372755289077759,
|
| 4557 |
+
"learning_rate": 5.337673747166834e-05,
|
| 4558 |
+
"loss": 1.1736,
|
| 4559 |
+
"step": 650
|
| 4560 |
+
},
|
| 4561 |
+
{
|
| 4562 |
+
"epoch": 0.030822404242223378,
|
| 4563 |
+
"grad_norm": 1.0708941221237183,
|
| 4564 |
+
"learning_rate": 5.3260466531378106e-05,
|
| 4565 |
+
"loss": 1.6468,
|
| 4566 |
+
"step": 651
|
| 4567 |
+
},
|
| 4568 |
+
{
|
| 4569 |
+
"epoch": 0.03086975048529899,
|
| 4570 |
+
"grad_norm": 1.1300220489501953,
|
| 4571 |
+
"learning_rate": 5.314417788186471e-05,
|
| 4572 |
+
"loss": 1.6485,
|
| 4573 |
+
"step": 652
|
| 4574 |
+
},
|
| 4575 |
+
{
|
| 4576 |
+
"epoch": 0.030917096728374602,
|
| 4577 |
+
"grad_norm": 1.0736169815063477,
|
| 4578 |
+
"learning_rate": 5.3027872154749915e-05,
|
| 4579 |
+
"loss": 1.3527,
|
| 4580 |
+
"step": 653
|
| 4581 |
+
},
|
| 4582 |
+
{
|
| 4583 |
+
"epoch": 0.030964442971450217,
|
| 4584 |
+
"grad_norm": 1.1438918113708496,
|
| 4585 |
+
"learning_rate": 5.291154998174824e-05,
|
| 4586 |
+
"loss": 1.4718,
|
| 4587 |
+
"step": 654
|
| 4588 |
+
},
|
| 4589 |
+
{
|
| 4590 |
+
"epoch": 0.03101178921452583,
|
| 4591 |
+
"grad_norm": 1.1480180025100708,
|
| 4592 |
+
"learning_rate": 5.279521199466356e-05,
|
| 4593 |
+
"loss": 1.6733,
|
| 4594 |
+
"step": 655
|
| 4595 |
+
},
|
| 4596 |
+
{
|
| 4597 |
+
"epoch": 0.03105913545760144,
|
| 4598 |
+
"grad_norm": 1.0300452709197998,
|
| 4599 |
+
"learning_rate": 5.2678858825385604e-05,
|
| 4600 |
+
"loss": 1.1472,
|
| 4601 |
+
"step": 656
|
| 4602 |
+
},
|
| 4603 |
+
{
|
| 4604 |
+
"epoch": 0.031106481700677053,
|
| 4605 |
+
"grad_norm": 1.1770920753479004,
|
| 4606 |
+
"learning_rate": 5.256249110588659e-05,
|
| 4607 |
+
"loss": 1.5572,
|
| 4608 |
+
"step": 657
|
| 4609 |
+
},
|
| 4610 |
+
{
|
| 4611 |
+
"epoch": 0.031153827943752665,
|
| 4612 |
+
"grad_norm": 1.1315271854400635,
|
| 4613 |
+
"learning_rate": 5.244610946821776e-05,
|
| 4614 |
+
"loss": 1.4018,
|
| 4615 |
+
"step": 658
|
| 4616 |
+
},
|
| 4617 |
+
{
|
| 4618 |
+
"epoch": 0.031201174186828277,
|
| 4619 |
+
"grad_norm": 1.072265386581421,
|
| 4620 |
+
"learning_rate": 5.232971454450595e-05,
|
| 4621 |
+
"loss": 1.3648,
|
| 4622 |
+
"step": 659
|
| 4623 |
+
},
|
| 4624 |
+
{
|
| 4625 |
+
"epoch": 0.03124852042990389,
|
| 4626 |
+
"grad_norm": 1.0530691146850586,
|
| 4627 |
+
"learning_rate": 5.2213306966950146e-05,
|
| 4628 |
+
"loss": 1.3735,
|
| 4629 |
+
"step": 660
|
| 4630 |
+
},
|
| 4631 |
+
{
|
| 4632 |
+
"epoch": 0.0312958666729795,
|
| 4633 |
+
"grad_norm": 1.23915433883667,
|
| 4634 |
+
"learning_rate": 5.2096887367818105e-05,
|
| 4635 |
+
"loss": 1.6796,
|
| 4636 |
+
"step": 661
|
| 4637 |
+
},
|
| 4638 |
+
{
|
| 4639 |
+
"epoch": 0.03134321291605511,
|
| 4640 |
+
"grad_norm": 1.2694331407546997,
|
| 4641 |
+
"learning_rate": 5.198045637944283e-05,
|
| 4642 |
+
"loss": 1.4603,
|
| 4643 |
+
"step": 662
|
| 4644 |
+
},
|
| 4645 |
+
{
|
| 4646 |
+
"epoch": 0.03139055915913072,
|
| 4647 |
+
"grad_norm": 1.1831574440002441,
|
| 4648 |
+
"learning_rate": 5.1864014634219214e-05,
|
| 4649 |
+
"loss": 1.0448,
|
| 4650 |
+
"step": 663
|
| 4651 |
+
},
|
| 4652 |
+
{
|
| 4653 |
+
"epoch": 0.03143790540220633,
|
| 4654 |
+
"grad_norm": 1.2854231595993042,
|
| 4655 |
+
"learning_rate": 5.174756276460056e-05,
|
| 4656 |
+
"loss": 1.4597,
|
| 4657 |
+
"step": 664
|
| 4658 |
+
},
|
| 4659 |
+
{
|
| 4660 |
+
"epoch": 0.031485251645281945,
|
| 4661 |
+
"grad_norm": 1.0730315446853638,
|
| 4662 |
+
"learning_rate": 5.1631101403095184e-05,
|
| 4663 |
+
"loss": 1.3382,
|
| 4664 |
+
"step": 665
|
| 4665 |
+
},
|
| 4666 |
+
{
|
| 4667 |
+
"epoch": 0.03153259788835756,
|
| 4668 |
+
"grad_norm": 1.163493275642395,
|
| 4669 |
+
"learning_rate": 5.151463118226294e-05,
|
| 4670 |
+
"loss": 1.3837,
|
| 4671 |
+
"step": 666
|
| 4672 |
+
},
|
| 4673 |
+
{
|
| 4674 |
+
"epoch": 0.03157994413143317,
|
| 4675 |
+
"grad_norm": 1.2980921268463135,
|
| 4676 |
+
"learning_rate": 5.139815273471177e-05,
|
| 4677 |
+
"loss": 1.5276,
|
| 4678 |
+
"step": 667
|
| 4679 |
+
},
|
| 4680 |
+
{
|
| 4681 |
+
"epoch": 0.03162729037450878,
|
| 4682 |
+
"grad_norm": 1.4798357486724854,
|
| 4683 |
+
"learning_rate": 5.128166669309438e-05,
|
| 4684 |
+
"loss": 1.8276,
|
| 4685 |
+
"step": 668
|
| 4686 |
+
},
|
| 4687 |
+
{
|
| 4688 |
+
"epoch": 0.03167463661758439,
|
| 4689 |
+
"grad_norm": 1.2314108610153198,
|
| 4690 |
+
"learning_rate": 5.116517369010466e-05,
|
| 4691 |
+
"loss": 1.3072,
|
| 4692 |
+
"step": 669
|
| 4693 |
+
},
|
| 4694 |
+
{
|
| 4695 |
+
"epoch": 0.031721982860660004,
|
| 4696 |
+
"grad_norm": 1.2994263172149658,
|
| 4697 |
+
"learning_rate": 5.1048674358474325e-05,
|
| 4698 |
+
"loss": 1.5165,
|
| 4699 |
+
"step": 670
|
| 4700 |
+
},
|
| 4701 |
+
{
|
| 4702 |
+
"epoch": 0.031769329103735616,
|
| 4703 |
+
"grad_norm": 1.2647054195404053,
|
| 4704 |
+
"learning_rate": 5.093216933096946e-05,
|
| 4705 |
+
"loss": 1.5609,
|
| 4706 |
+
"step": 671
|
| 4707 |
+
},
|
| 4708 |
+
{
|
| 4709 |
+
"epoch": 0.03181667534681123,
|
| 4710 |
+
"grad_norm": 1.350826382637024,
|
| 4711 |
+
"learning_rate": 5.081565924038711e-05,
|
| 4712 |
+
"loss": 1.6365,
|
| 4713 |
+
"step": 672
|
| 4714 |
+
},
|
| 4715 |
+
{
|
| 4716 |
+
"epoch": 0.03186402158988684,
|
| 4717 |
+
"grad_norm": 1.1930540800094604,
|
| 4718 |
+
"learning_rate": 5.069914471955178e-05,
|
| 4719 |
+
"loss": 1.5473,
|
| 4720 |
+
"step": 673
|
| 4721 |
+
},
|
| 4722 |
+
{
|
| 4723 |
+
"epoch": 0.03191136783296245,
|
| 4724 |
+
"grad_norm": 1.2137759923934937,
|
| 4725 |
+
"learning_rate": 5.058262640131209e-05,
|
| 4726 |
+
"loss": 1.6863,
|
| 4727 |
+
"step": 674
|
| 4728 |
+
},
|
| 4729 |
+
{
|
| 4730 |
+
"epoch": 0.031958714076038064,
|
| 4731 |
+
"grad_norm": 1.1367132663726807,
|
| 4732 |
+
"learning_rate": 5.046610491853724e-05,
|
| 4733 |
+
"loss": 1.5055,
|
| 4734 |
+
"step": 675
|
| 4735 |
+
},
|
| 4736 |
+
{
|
| 4737 |
+
"epoch": 0.032006060319113676,
|
| 4738 |
+
"grad_norm": 1.1324983835220337,
|
| 4739 |
+
"learning_rate": 5.034958090411364e-05,
|
| 4740 |
+
"loss": 1.5342,
|
| 4741 |
+
"step": 676
|
| 4742 |
+
},
|
| 4743 |
+
{
|
| 4744 |
+
"epoch": 0.03205340656218929,
|
| 4745 |
+
"grad_norm": 1.3015636205673218,
|
| 4746 |
+
"learning_rate": 5.023305499094144e-05,
|
| 4747 |
+
"loss": 1.8074,
|
| 4748 |
+
"step": 677
|
| 4749 |
+
},
|
| 4750 |
+
{
|
| 4751 |
+
"epoch": 0.0321007528052649,
|
| 4752 |
+
"grad_norm": 1.1993989944458008,
|
| 4753 |
+
"learning_rate": 5.011652781193111e-05,
|
| 4754 |
+
"loss": 1.7739,
|
| 4755 |
+
"step": 678
|
| 4756 |
}
|
| 4757 |
],
|
| 4758 |
"logging_steps": 1,
|
|
|
|
| 4772 |
"attributes": {}
|
| 4773 |
}
|
| 4774 |
},
|
| 4775 |
+
"total_flos": 1.8593651970854093e+17,
|
| 4776 |
"train_batch_size": 4,
|
| 4777 |
"trial_name": null,
|
| 4778 |
"trial_params": null
|