Training in progress, step 711, 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 125918320
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:8a3407d30d0c0cfbf15d4561a6c3c2e60e51f387c6006b81ec98129a09360217
|
| 3 |
size 125918320
|
last-checkpoint/optimizer.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 64684244
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:216e3dcd0e4a1866d078f676c9fe160b4ad75410883f6b6c2a01e5cdb87eb944
|
| 3 |
size 64684244
|
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:afccc137a363b0dd47f30df2ed42c833cac16b3f284a231abe45baa1320ad8ae
|
| 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 sha256:422b4e12aed6e68bb7e6ff0e852bb990c7f0d4ac2c13311a762eee9cc8a9875f
|
| 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":
|
| 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,
|
|
@@ -3745,6 +3745,1253 @@
|
|
| 3745 |
"learning_rate": 1.4723411326807807e-05,
|
| 3746 |
"loss": 0.5797,
|
| 3747 |
"step": 534
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 3748 |
}
|
| 3749 |
],
|
| 3750 |
"logging_steps": 1,
|
|
@@ -3759,12 +5006,12 @@
|
|
| 3759 |
"should_evaluate": false,
|
| 3760 |
"should_log": false,
|
| 3761 |
"should_save": true,
|
| 3762 |
-
"should_training_stop":
|
| 3763 |
},
|
| 3764 |
"attributes": {}
|
| 3765 |
}
|
| 3766 |
},
|
| 3767 |
-
"total_flos":
|
| 3768 |
"train_batch_size": 4,
|
| 3769 |
"trial_name": null,
|
| 3770 |
"trial_params": null
|
|
|
|
| 1 |
{
|
| 2 |
"best_metric": null,
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 1.0010559662090812,
|
| 5 |
"eval_steps": 500,
|
| 6 |
+
"global_step": 711,
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
|
|
| 3745 |
"learning_rate": 1.4723411326807807e-05,
|
| 3746 |
"loss": 0.5797,
|
| 3747 |
"step": 534
|
| 3748 |
+
},
|
| 3749 |
+
{
|
| 3750 |
+
"epoch": 0.753255895811334,
|
| 3751 |
+
"grad_norm": 7.255380630493164,
|
| 3752 |
+
"learning_rate": 1.4566085573529874e-05,
|
| 3753 |
+
"loss": 0.7906,
|
| 3754 |
+
"step": 535
|
| 3755 |
+
},
|
| 3756 |
+
{
|
| 3757 |
+
"epoch": 0.7546638507567758,
|
| 3758 |
+
"grad_norm": 6.509925842285156,
|
| 3759 |
+
"learning_rate": 1.4409461451235068e-05,
|
| 3760 |
+
"loss": 0.796,
|
| 3761 |
+
"step": 536
|
| 3762 |
+
},
|
| 3763 |
+
{
|
| 3764 |
+
"epoch": 0.7560718057022175,
|
| 3765 |
+
"grad_norm": 5.435834884643555,
|
| 3766 |
+
"learning_rate": 1.4253542061255515e-05,
|
| 3767 |
+
"loss": 0.4876,
|
| 3768 |
+
"step": 537
|
| 3769 |
+
},
|
| 3770 |
+
{
|
| 3771 |
+
"epoch": 0.7574797606476593,
|
| 3772 |
+
"grad_norm": 4.782072067260742,
|
| 3773 |
+
"learning_rate": 1.409833049096882e-05,
|
| 3774 |
+
"loss": 0.2748,
|
| 3775 |
+
"step": 538
|
| 3776 |
+
},
|
| 3777 |
+
{
|
| 3778 |
+
"epoch": 0.758887715593101,
|
| 3779 |
+
"grad_norm": 5.33141565322876,
|
| 3780 |
+
"learning_rate": 1.3943829813737008e-05,
|
| 3781 |
+
"loss": 0.6006,
|
| 3782 |
+
"step": 539
|
| 3783 |
+
},
|
| 3784 |
+
{
|
| 3785 |
+
"epoch": 0.7602956705385427,
|
| 3786 |
+
"grad_norm": 5.939730167388916,
|
| 3787 |
+
"learning_rate": 1.3790043088845644e-05,
|
| 3788 |
+
"loss": 0.8508,
|
| 3789 |
+
"step": 540
|
| 3790 |
+
},
|
| 3791 |
+
{
|
| 3792 |
+
"epoch": 0.7617036254839845,
|
| 3793 |
+
"grad_norm": 8.518722534179688,
|
| 3794 |
+
"learning_rate": 1.3636973361443183e-05,
|
| 3795 |
+
"loss": 0.8655,
|
| 3796 |
+
"step": 541
|
| 3797 |
+
},
|
| 3798 |
+
{
|
| 3799 |
+
"epoch": 0.7631115804294263,
|
| 3800 |
+
"grad_norm": 13.114689826965332,
|
| 3801 |
+
"learning_rate": 1.3484623662480778e-05,
|
| 3802 |
+
"loss": 1.176,
|
| 3803 |
+
"step": 542
|
| 3804 |
+
},
|
| 3805 |
+
{
|
| 3806 |
+
"epoch": 0.764519535374868,
|
| 3807 |
+
"grad_norm": 6.1082844734191895,
|
| 3808 |
+
"learning_rate": 1.3332997008652204e-05,
|
| 3809 |
+
"loss": 0.6226,
|
| 3810 |
+
"step": 543
|
| 3811 |
+
},
|
| 3812 |
+
{
|
| 3813 |
+
"epoch": 0.7659274903203097,
|
| 3814 |
+
"grad_norm": 3.8953821659088135,
|
| 3815 |
+
"learning_rate": 1.3182096402334154e-05,
|
| 3816 |
+
"loss": 0.3033,
|
| 3817 |
+
"step": 544
|
| 3818 |
+
},
|
| 3819 |
+
{
|
| 3820 |
+
"epoch": 0.7673354452657515,
|
| 3821 |
+
"grad_norm": 9.015904426574707,
|
| 3822 |
+
"learning_rate": 1.3031924831526737e-05,
|
| 3823 |
+
"loss": 0.9219,
|
| 3824 |
+
"step": 545
|
| 3825 |
+
},
|
| 3826 |
+
{
|
| 3827 |
+
"epoch": 0.7687434002111933,
|
| 3828 |
+
"grad_norm": 4.156109809875488,
|
| 3829 |
+
"learning_rate": 1.2882485269794337e-05,
|
| 3830 |
+
"loss": 0.4274,
|
| 3831 |
+
"step": 546
|
| 3832 |
+
},
|
| 3833 |
+
{
|
| 3834 |
+
"epoch": 0.770151355156635,
|
| 3835 |
+
"grad_norm": 4.576930522918701,
|
| 3836 |
+
"learning_rate": 1.2733780676206824e-05,
|
| 3837 |
+
"loss": 0.5323,
|
| 3838 |
+
"step": 547
|
| 3839 |
+
},
|
| 3840 |
+
{
|
| 3841 |
+
"epoch": 0.7715593101020767,
|
| 3842 |
+
"grad_norm": 11.672577857971191,
|
| 3843 |
+
"learning_rate": 1.2585813995280782e-05,
|
| 3844 |
+
"loss": 0.9249,
|
| 3845 |
+
"step": 548
|
| 3846 |
+
},
|
| 3847 |
+
{
|
| 3848 |
+
"epoch": 0.7729672650475184,
|
| 3849 |
+
"grad_norm": 6.854877948760986,
|
| 3850 |
+
"learning_rate": 1.2438588156921377e-05,
|
| 3851 |
+
"loss": 0.7525,
|
| 3852 |
+
"step": 549
|
| 3853 |
+
},
|
| 3854 |
+
{
|
| 3855 |
+
"epoch": 0.7743752199929602,
|
| 3856 |
+
"grad_norm": 13.515569686889648,
|
| 3857 |
+
"learning_rate": 1.2292106076364201e-05,
|
| 3858 |
+
"loss": 1.2491,
|
| 3859 |
+
"step": 550
|
| 3860 |
+
},
|
| 3861 |
+
{
|
| 3862 |
+
"epoch": 0.775783174938402,
|
| 3863 |
+
"grad_norm": 2.5775089263916016,
|
| 3864 |
+
"learning_rate": 1.2146370654117672e-05,
|
| 3865 |
+
"loss": 0.0877,
|
| 3866 |
+
"step": 551
|
| 3867 |
+
},
|
| 3868 |
+
{
|
| 3869 |
+
"epoch": 0.7771911298838438,
|
| 3870 |
+
"grad_norm": 4.166385650634766,
|
| 3871 |
+
"learning_rate": 1.2001384775905517e-05,
|
| 3872 |
+
"loss": 0.34,
|
| 3873 |
+
"step": 552
|
| 3874 |
+
},
|
| 3875 |
+
{
|
| 3876 |
+
"epoch": 0.7785990848292854,
|
| 3877 |
+
"grad_norm": 8.945438385009766,
|
| 3878 |
+
"learning_rate": 1.1857151312609622e-05,
|
| 3879 |
+
"loss": 0.6992,
|
| 3880 |
+
"step": 553
|
| 3881 |
+
},
|
| 3882 |
+
{
|
| 3883 |
+
"epoch": 0.7800070397747272,
|
| 3884 |
+
"grad_norm": 3.7914435863494873,
|
| 3885 |
+
"learning_rate": 1.1713673120213308e-05,
|
| 3886 |
+
"loss": 0.4275,
|
| 3887 |
+
"step": 554
|
| 3888 |
+
},
|
| 3889 |
+
{
|
| 3890 |
+
"epoch": 0.781414994720169,
|
| 3891 |
+
"grad_norm": 4.656171798706055,
|
| 3892 |
+
"learning_rate": 1.1570953039744591e-05,
|
| 3893 |
+
"loss": 0.5339,
|
| 3894 |
+
"step": 555
|
| 3895 |
+
},
|
| 3896 |
+
{
|
| 3897 |
+
"epoch": 0.7828229496656107,
|
| 3898 |
+
"grad_norm": 7.4065446853637695,
|
| 3899 |
+
"learning_rate": 1.1428993897220103e-05,
|
| 3900 |
+
"loss": 0.8365,
|
| 3901 |
+
"step": 556
|
| 3902 |
+
},
|
| 3903 |
+
{
|
| 3904 |
+
"epoch": 0.7842309046110525,
|
| 3905 |
+
"grad_norm": 3.4411070346832275,
|
| 3906 |
+
"learning_rate": 1.1287798503588992e-05,
|
| 3907 |
+
"loss": 0.4935,
|
| 3908 |
+
"step": 557
|
| 3909 |
+
},
|
| 3910 |
+
{
|
| 3911 |
+
"epoch": 0.7856388595564942,
|
| 3912 |
+
"grad_norm": 1.0402673482894897,
|
| 3913 |
+
"learning_rate": 1.1147369654677375e-05,
|
| 3914 |
+
"loss": 0.0769,
|
| 3915 |
+
"step": 558
|
| 3916 |
+
},
|
| 3917 |
+
{
|
| 3918 |
+
"epoch": 0.7870468145019359,
|
| 3919 |
+
"grad_norm": 5.989752292633057,
|
| 3920 |
+
"learning_rate": 1.1007710131132932e-05,
|
| 3921 |
+
"loss": 0.5214,
|
| 3922 |
+
"step": 559
|
| 3923 |
+
},
|
| 3924 |
+
{
|
| 3925 |
+
"epoch": 0.7884547694473777,
|
| 3926 |
+
"grad_norm": 12.650126457214355,
|
| 3927 |
+
"learning_rate": 1.0868822698369785e-05,
|
| 3928 |
+
"loss": 0.7226,
|
| 3929 |
+
"step": 560
|
| 3930 |
+
},
|
| 3931 |
+
{
|
| 3932 |
+
"epoch": 0.7898627243928195,
|
| 3933 |
+
"grad_norm": 10.798233985900879,
|
| 3934 |
+
"learning_rate": 1.0730710106513886e-05,
|
| 3935 |
+
"loss": 0.8267,
|
| 3936 |
+
"step": 561
|
| 3937 |
+
},
|
| 3938 |
+
{
|
| 3939 |
+
"epoch": 0.7912706793382612,
|
| 3940 |
+
"grad_norm": 17.764671325683594,
|
| 3941 |
+
"learning_rate": 1.0593375090348396e-05,
|
| 3942 |
+
"loss": 1.073,
|
| 3943 |
+
"step": 562
|
| 3944 |
+
},
|
| 3945 |
+
{
|
| 3946 |
+
"epoch": 0.7926786342837029,
|
| 3947 |
+
"grad_norm": 8.220035552978516,
|
| 3948 |
+
"learning_rate": 1.0456820369259667e-05,
|
| 3949 |
+
"loss": 0.8961,
|
| 3950 |
+
"step": 563
|
| 3951 |
+
},
|
| 3952 |
+
{
|
| 3953 |
+
"epoch": 0.7940865892291447,
|
| 3954 |
+
"grad_norm": 7.768228054046631,
|
| 3955 |
+
"learning_rate": 1.0321048647183278e-05,
|
| 3956 |
+
"loss": 0.7928,
|
| 3957 |
+
"step": 564
|
| 3958 |
+
},
|
| 3959 |
+
{
|
| 3960 |
+
"epoch": 0.7954945441745864,
|
| 3961 |
+
"grad_norm": 5.817502498626709,
|
| 3962 |
+
"learning_rate": 1.0186062612550617e-05,
|
| 3963 |
+
"loss": 0.6983,
|
| 3964 |
+
"step": 565
|
| 3965 |
+
},
|
| 3966 |
+
{
|
| 3967 |
+
"epoch": 0.7969024991200282,
|
| 3968 |
+
"grad_norm": 4.239863395690918,
|
| 3969 |
+
"learning_rate": 1.0051864938235572e-05,
|
| 3970 |
+
"loss": 0.536,
|
| 3971 |
+
"step": 566
|
| 3972 |
+
},
|
| 3973 |
+
{
|
| 3974 |
+
"epoch": 0.7983104540654699,
|
| 3975 |
+
"grad_norm": 4.070102691650391,
|
| 3976 |
+
"learning_rate": 9.91845828150157e-06,
|
| 3977 |
+
"loss": 0.5088,
|
| 3978 |
+
"step": 567
|
| 3979 |
+
},
|
| 3980 |
+
{
|
| 3981 |
+
"epoch": 0.7997184090109116,
|
| 3982 |
+
"grad_norm": 3.5287749767303467,
|
| 3983 |
+
"learning_rate": 9.78584528394907e-06,
|
| 3984 |
+
"loss": 0.4654,
|
| 3985 |
+
"step": 568
|
| 3986 |
+
},
|
| 3987 |
+
{
|
| 3988 |
+
"epoch": 0.8011263639563534,
|
| 3989 |
+
"grad_norm": 3.3061466217041016,
|
| 3990 |
+
"learning_rate": 9.65402857146317e-06,
|
| 3991 |
+
"loss": 0.4742,
|
| 3992 |
+
"step": 569
|
| 3993 |
+
},
|
| 3994 |
+
{
|
| 3995 |
+
"epoch": 0.8025343189017952,
|
| 3996 |
+
"grad_norm": 3.7039663791656494,
|
| 3997 |
+
"learning_rate": 9.523010754161621e-06,
|
| 3998 |
+
"loss": 0.5101,
|
| 3999 |
+
"step": 570
|
| 4000 |
+
},
|
| 4001 |
+
{
|
| 4002 |
+
"epoch": 0.8039422738472369,
|
| 4003 |
+
"grad_norm": 3.39595627784729,
|
| 4004 |
+
"learning_rate": 9.392794426343188e-06,
|
| 4005 |
+
"loss": 0.487,
|
| 4006 |
+
"step": 571
|
| 4007 |
+
},
|
| 4008 |
+
{
|
| 4009 |
+
"epoch": 0.8053502287926786,
|
| 4010 |
+
"grad_norm": 2.967733383178711,
|
| 4011 |
+
"learning_rate": 9.263382166436224e-06,
|
| 4012 |
+
"loss": 0.4949,
|
| 4013 |
+
"step": 572
|
| 4014 |
+
},
|
| 4015 |
+
{
|
| 4016 |
+
"epoch": 0.8067581837381204,
|
| 4017 |
+
"grad_norm": 3.1134448051452637,
|
| 4018 |
+
"learning_rate": 9.134776536947687e-06,
|
| 4019 |
+
"loss": 0.458,
|
| 4020 |
+
"step": 573
|
| 4021 |
+
},
|
| 4022 |
+
{
|
| 4023 |
+
"epoch": 0.8081661386835621,
|
| 4024 |
+
"grad_norm": 4.439833164215088,
|
| 4025 |
+
"learning_rate": 9.00698008441228e-06,
|
| 4026 |
+
"loss": 0.5411,
|
| 4027 |
+
"step": 574
|
| 4028 |
+
},
|
| 4029 |
+
{
|
| 4030 |
+
"epoch": 0.8095740936290039,
|
| 4031 |
+
"grad_norm": 4.359611511230469,
|
| 4032 |
+
"learning_rate": 8.879995339342167e-06,
|
| 4033 |
+
"loss": 0.6781,
|
| 4034 |
+
"step": 575
|
| 4035 |
+
},
|
| 4036 |
+
{
|
| 4037 |
+
"epoch": 0.8109820485744457,
|
| 4038 |
+
"grad_norm": 4.5389556884765625,
|
| 4039 |
+
"learning_rate": 8.753824816176787e-06,
|
| 4040 |
+
"loss": 0.5552,
|
| 4041 |
+
"step": 576
|
| 4042 |
+
},
|
| 4043 |
+
{
|
| 4044 |
+
"epoch": 0.8123900035198873,
|
| 4045 |
+
"grad_norm": 5.031420707702637,
|
| 4046 |
+
"learning_rate": 8.628471013233036e-06,
|
| 4047 |
+
"loss": 0.5982,
|
| 4048 |
+
"step": 577
|
| 4049 |
+
},
|
| 4050 |
+
{
|
| 4051 |
+
"epoch": 0.8137979584653291,
|
| 4052 |
+
"grad_norm": 4.413736820220947,
|
| 4053 |
+
"learning_rate": 8.5039364126559e-06,
|
| 4054 |
+
"loss": 0.5859,
|
| 4055 |
+
"step": 578
|
| 4056 |
+
},
|
| 4057 |
+
{
|
| 4058 |
+
"epoch": 0.8152059134107709,
|
| 4059 |
+
"grad_norm": 2.110487461090088,
|
| 4060 |
+
"learning_rate": 8.380223480369182e-06,
|
| 4061 |
+
"loss": 0.3305,
|
| 4062 |
+
"step": 579
|
| 4063 |
+
},
|
| 4064 |
+
{
|
| 4065 |
+
"epoch": 0.8166138683562126,
|
| 4066 |
+
"grad_norm": 2.8831546306610107,
|
| 4067 |
+
"learning_rate": 8.257334666026817e-06,
|
| 4068 |
+
"loss": 0.2105,
|
| 4069 |
+
"step": 580
|
| 4070 |
+
},
|
| 4071 |
+
{
|
| 4072 |
+
"epoch": 0.8180218233016543,
|
| 4073 |
+
"grad_norm": 2.9017462730407715,
|
| 4074 |
+
"learning_rate": 8.135272402964212e-06,
|
| 4075 |
+
"loss": 0.2621,
|
| 4076 |
+
"step": 581
|
| 4077 |
+
},
|
| 4078 |
+
{
|
| 4079 |
+
"epoch": 0.8194297782470961,
|
| 4080 |
+
"grad_norm": 3.5329201221466064,
|
| 4081 |
+
"learning_rate": 8.014039108150168e-06,
|
| 4082 |
+
"loss": 0.3293,
|
| 4083 |
+
"step": 582
|
| 4084 |
+
},
|
| 4085 |
+
{
|
| 4086 |
+
"epoch": 0.8208377331925378,
|
| 4087 |
+
"grad_norm": 4.189939498901367,
|
| 4088 |
+
"learning_rate": 7.893637182139002e-06,
|
| 4089 |
+
"loss": 0.6052,
|
| 4090 |
+
"step": 583
|
| 4091 |
+
},
|
| 4092 |
+
{
|
| 4093 |
+
"epoch": 0.8222456881379796,
|
| 4094 |
+
"grad_norm": 6.137105464935303,
|
| 4095 |
+
"learning_rate": 7.774069009022982e-06,
|
| 4096 |
+
"loss": 0.5093,
|
| 4097 |
+
"step": 584
|
| 4098 |
+
},
|
| 4099 |
+
{
|
| 4100 |
+
"epoch": 0.8236536430834214,
|
| 4101 |
+
"grad_norm": 5.461522579193115,
|
| 4102 |
+
"learning_rate": 7.655336956385156e-06,
|
| 4103 |
+
"loss": 0.5406,
|
| 4104 |
+
"step": 585
|
| 4105 |
+
},
|
| 4106 |
+
{
|
| 4107 |
+
"epoch": 0.825061598028863,
|
| 4108 |
+
"grad_norm": 3.9772427082061768,
|
| 4109 |
+
"learning_rate": 7.537443375252423e-06,
|
| 4110 |
+
"loss": 0.3464,
|
| 4111 |
+
"step": 586
|
| 4112 |
+
},
|
| 4113 |
+
{
|
| 4114 |
+
"epoch": 0.8264695529743048,
|
| 4115 |
+
"grad_norm": 4.120124340057373,
|
| 4116 |
+
"learning_rate": 7.420390600049076e-06,
|
| 4117 |
+
"loss": 0.3298,
|
| 4118 |
+
"step": 587
|
| 4119 |
+
},
|
| 4120 |
+
{
|
| 4121 |
+
"epoch": 0.8278775079197466,
|
| 4122 |
+
"grad_norm": 4.581118583679199,
|
| 4123 |
+
"learning_rate": 7.304180948550454e-06,
|
| 4124 |
+
"loss": 0.523,
|
| 4125 |
+
"step": 588
|
| 4126 |
+
},
|
| 4127 |
+
{
|
| 4128 |
+
"epoch": 0.8292854628651883,
|
| 4129 |
+
"grad_norm": 6.187935829162598,
|
| 4130 |
+
"learning_rate": 7.188816721837116e-06,
|
| 4131 |
+
"loss": 0.6198,
|
| 4132 |
+
"step": 589
|
| 4133 |
+
},
|
| 4134 |
+
{
|
| 4135 |
+
"epoch": 0.8306934178106301,
|
| 4136 |
+
"grad_norm": 9.697196006774902,
|
| 4137 |
+
"learning_rate": 7.074300204249296e-06,
|
| 4138 |
+
"loss": 0.8589,
|
| 4139 |
+
"step": 590
|
| 4140 |
+
},
|
| 4141 |
+
{
|
| 4142 |
+
"epoch": 0.8321013727560718,
|
| 4143 |
+
"grad_norm": 7.243709087371826,
|
| 4144 |
+
"learning_rate": 6.960633663341615e-06,
|
| 4145 |
+
"loss": 0.6008,
|
| 4146 |
+
"step": 591
|
| 4147 |
+
},
|
| 4148 |
+
{
|
| 4149 |
+
"epoch": 0.8335093277015135,
|
| 4150 |
+
"grad_norm": 11.357503890991211,
|
| 4151 |
+
"learning_rate": 6.847819349838247e-06,
|
| 4152 |
+
"loss": 0.999,
|
| 4153 |
+
"step": 592
|
| 4154 |
+
},
|
| 4155 |
+
{
|
| 4156 |
+
"epoch": 0.8349172826469553,
|
| 4157 |
+
"grad_norm": 5.490504741668701,
|
| 4158 |
+
"learning_rate": 6.7358594975882535e-06,
|
| 4159 |
+
"loss": 0.6455,
|
| 4160 |
+
"step": 593
|
| 4161 |
+
},
|
| 4162 |
+
{
|
| 4163 |
+
"epoch": 0.8363252375923971,
|
| 4164 |
+
"grad_norm": 5.9745965003967285,
|
| 4165 |
+
"learning_rate": 6.624756323521459e-06,
|
| 4166 |
+
"loss": 0.5208,
|
| 4167 |
+
"step": 594
|
| 4168 |
+
},
|
| 4169 |
+
{
|
| 4170 |
+
"epoch": 0.8377331925378388,
|
| 4171 |
+
"grad_norm": 4.54645299911499,
|
| 4172 |
+
"learning_rate": 6.514512027604508e-06,
|
| 4173 |
+
"loss": 0.6638,
|
| 4174 |
+
"step": 595
|
| 4175 |
+
},
|
| 4176 |
+
{
|
| 4177 |
+
"epoch": 0.8391411474832805,
|
| 4178 |
+
"grad_norm": 4.623635292053223,
|
| 4179 |
+
"learning_rate": 6.405128792797255e-06,
|
| 4180 |
+
"loss": 0.5462,
|
| 4181 |
+
"step": 596
|
| 4182 |
+
},
|
| 4183 |
+
{
|
| 4184 |
+
"epoch": 0.8405491024287223,
|
| 4185 |
+
"grad_norm": 3.521913528442383,
|
| 4186 |
+
"learning_rate": 6.296608785009633e-06,
|
| 4187 |
+
"loss": 0.5212,
|
| 4188 |
+
"step": 597
|
| 4189 |
+
},
|
| 4190 |
+
{
|
| 4191 |
+
"epoch": 0.841957057374164,
|
| 4192 |
+
"grad_norm": 8.050673484802246,
|
| 4193 |
+
"learning_rate": 6.1889541530586955e-06,
|
| 4194 |
+
"loss": 0.7836,
|
| 4195 |
+
"step": 598
|
| 4196 |
+
},
|
| 4197 |
+
{
|
| 4198 |
+
"epoch": 0.8433650123196058,
|
| 4199 |
+
"grad_norm": 6.162160873413086,
|
| 4200 |
+
"learning_rate": 6.0821670286261016e-06,
|
| 4201 |
+
"loss": 0.5981,
|
| 4202 |
+
"step": 599
|
| 4203 |
+
},
|
| 4204 |
+
{
|
| 4205 |
+
"epoch": 0.8447729672650475,
|
| 4206 |
+
"grad_norm": 14.778404235839844,
|
| 4207 |
+
"learning_rate": 5.9762495262158764e-06,
|
| 4208 |
+
"loss": 1.2597,
|
| 4209 |
+
"step": 600
|
| 4210 |
+
},
|
| 4211 |
+
{
|
| 4212 |
+
"epoch": 0.8461809222104892,
|
| 4213 |
+
"grad_norm": 2.480350971221924,
|
| 4214 |
+
"learning_rate": 5.87120374311258e-06,
|
| 4215 |
+
"loss": 0.1038,
|
| 4216 |
+
"step": 601
|
| 4217 |
+
},
|
| 4218 |
+
{
|
| 4219 |
+
"epoch": 0.847588877155931,
|
| 4220 |
+
"grad_norm": 4.1250481605529785,
|
| 4221 |
+
"learning_rate": 5.767031759339758e-06,
|
| 4222 |
+
"loss": 0.3867,
|
| 4223 |
+
"step": 602
|
| 4224 |
+
},
|
| 4225 |
+
{
|
| 4226 |
+
"epoch": 0.8489968321013728,
|
| 4227 |
+
"grad_norm": 9.889833450317383,
|
| 4228 |
+
"learning_rate": 5.663735637618728e-06,
|
| 4229 |
+
"loss": 0.6262,
|
| 4230 |
+
"step": 603
|
| 4231 |
+
},
|
| 4232 |
+
{
|
| 4233 |
+
"epoch": 0.8504047870468145,
|
| 4234 |
+
"grad_norm": 3.7069804668426514,
|
| 4235 |
+
"learning_rate": 5.5613174233278e-06,
|
| 4236 |
+
"loss": 0.4572,
|
| 4237 |
+
"step": 604
|
| 4238 |
+
},
|
| 4239 |
+
{
|
| 4240 |
+
"epoch": 0.8518127419922562,
|
| 4241 |
+
"grad_norm": 3.344693422317505,
|
| 4242 |
+
"learning_rate": 5.459779144461713e-06,
|
| 4243 |
+
"loss": 0.5041,
|
| 4244 |
+
"step": 605
|
| 4245 |
+
},
|
| 4246 |
+
{
|
| 4247 |
+
"epoch": 0.853220696937698,
|
| 4248 |
+
"grad_norm": 6.324820041656494,
|
| 4249 |
+
"learning_rate": 5.359122811591533e-06,
|
| 4250 |
+
"loss": 0.481,
|
| 4251 |
+
"step": 606
|
| 4252 |
+
},
|
| 4253 |
+
{
|
| 4254 |
+
"epoch": 0.8546286518831397,
|
| 4255 |
+
"grad_norm": 6.167655944824219,
|
| 4256 |
+
"learning_rate": 5.259350417824776e-06,
|
| 4257 |
+
"loss": 0.5111,
|
| 4258 |
+
"step": 607
|
| 4259 |
+
},
|
| 4260 |
+
{
|
| 4261 |
+
"epoch": 0.8560366068285815,
|
| 4262 |
+
"grad_norm": 7.6900811195373535,
|
| 4263 |
+
"learning_rate": 5.160463938765986e-06,
|
| 4264 |
+
"loss": 0.6785,
|
| 4265 |
+
"step": 608
|
| 4266 |
+
},
|
| 4267 |
+
{
|
| 4268 |
+
"epoch": 0.8574445617740233,
|
| 4269 |
+
"grad_norm": 4.303120136260986,
|
| 4270 |
+
"learning_rate": 5.062465332477651e-06,
|
| 4271 |
+
"loss": 0.4381,
|
| 4272 |
+
"step": 609
|
| 4273 |
+
},
|
| 4274 |
+
{
|
| 4275 |
+
"epoch": 0.8588525167194649,
|
| 4276 |
+
"grad_norm": 2.0334081649780273,
|
| 4277 |
+
"learning_rate": 4.9653565394413295e-06,
|
| 4278 |
+
"loss": 0.184,
|
| 4279 |
+
"step": 610
|
| 4280 |
+
},
|
| 4281 |
+
{
|
| 4282 |
+
"epoch": 0.8602604716649067,
|
| 4283 |
+
"grad_norm": 4.938886642456055,
|
| 4284 |
+
"learning_rate": 4.869139482519325e-06,
|
| 4285 |
+
"loss": 0.5076,
|
| 4286 |
+
"step": 611
|
| 4287 |
+
},
|
| 4288 |
+
{
|
| 4289 |
+
"epoch": 0.8616684266103485,
|
| 4290 |
+
"grad_norm": 7.609766960144043,
|
| 4291 |
+
"learning_rate": 4.77381606691657e-06,
|
| 4292 |
+
"loss": 0.7531,
|
| 4293 |
+
"step": 612
|
| 4294 |
+
},
|
| 4295 |
+
{
|
| 4296 |
+
"epoch": 0.8630763815557903,
|
| 4297 |
+
"grad_norm": 3.9302780628204346,
|
| 4298 |
+
"learning_rate": 4.6793881801428796e-06,
|
| 4299 |
+
"loss": 0.4682,
|
| 4300 |
+
"step": 613
|
| 4301 |
+
},
|
| 4302 |
+
{
|
| 4303 |
+
"epoch": 0.8644843365012319,
|
| 4304 |
+
"grad_norm": 6.8657684326171875,
|
| 4305 |
+
"learning_rate": 4.585857691975626e-06,
|
| 4306 |
+
"loss": 0.8234,
|
| 4307 |
+
"step": 614
|
| 4308 |
+
},
|
| 4309 |
+
{
|
| 4310 |
+
"epoch": 0.8658922914466737,
|
| 4311 |
+
"grad_norm": 6.67000150680542,
|
| 4312 |
+
"learning_rate": 4.493226454422661e-06,
|
| 4313 |
+
"loss": 0.8253,
|
| 4314 |
+
"step": 615
|
| 4315 |
+
},
|
| 4316 |
+
{
|
| 4317 |
+
"epoch": 0.8673002463921154,
|
| 4318 |
+
"grad_norm": 6.91273307800293,
|
| 4319 |
+
"learning_rate": 4.40149630168572e-06,
|
| 4320 |
+
"loss": 0.83,
|
| 4321 |
+
"step": 616
|
| 4322 |
+
},
|
| 4323 |
+
{
|
| 4324 |
+
"epoch": 0.8687082013375572,
|
| 4325 |
+
"grad_norm": 7.153740882873535,
|
| 4326 |
+
"learning_rate": 4.310669050123994e-06,
|
| 4327 |
+
"loss": 0.8146,
|
| 4328 |
+
"step": 617
|
| 4329 |
+
},
|
| 4330 |
+
{
|
| 4331 |
+
"epoch": 0.870116156282999,
|
| 4332 |
+
"grad_norm": 3.3541862964630127,
|
| 4333 |
+
"learning_rate": 4.220746498218286e-06,
|
| 4334 |
+
"loss": 0.4726,
|
| 4335 |
+
"step": 618
|
| 4336 |
+
},
|
| 4337 |
+
{
|
| 4338 |
+
"epoch": 0.8715241112284406,
|
| 4339 |
+
"grad_norm": 3.593432664871216,
|
| 4340 |
+
"learning_rate": 4.131730426535285e-06,
|
| 4341 |
+
"loss": 0.5584,
|
| 4342 |
+
"step": 619
|
| 4343 |
+
},
|
| 4344 |
+
{
|
| 4345 |
+
"epoch": 0.8729320661738824,
|
| 4346 |
+
"grad_norm": 2.7953832149505615,
|
| 4347 |
+
"learning_rate": 4.043622597692387e-06,
|
| 4348 |
+
"loss": 0.3818,
|
| 4349 |
+
"step": 620
|
| 4350 |
+
},
|
| 4351 |
+
{
|
| 4352 |
+
"epoch": 0.8743400211193242,
|
| 4353 |
+
"grad_norm": 3.4306161403656006,
|
| 4354 |
+
"learning_rate": 3.956424756322785e-06,
|
| 4355 |
+
"loss": 0.4769,
|
| 4356 |
+
"step": 621
|
| 4357 |
+
},
|
| 4358 |
+
{
|
| 4359 |
+
"epoch": 0.875747976064766,
|
| 4360 |
+
"grad_norm": 2.890202760696411,
|
| 4361 |
+
"learning_rate": 3.870138629040848e-06,
|
| 4362 |
+
"loss": 0.4188,
|
| 4363 |
+
"step": 622
|
| 4364 |
+
},
|
| 4365 |
+
{
|
| 4366 |
+
"epoch": 0.8771559310102077,
|
| 4367 |
+
"grad_norm": 2.4199206829071045,
|
| 4368 |
+
"learning_rate": 3.7847659244080736e-06,
|
| 4369 |
+
"loss": 0.414,
|
| 4370 |
+
"step": 623
|
| 4371 |
+
},
|
| 4372 |
+
{
|
| 4373 |
+
"epoch": 0.8785638859556494,
|
| 4374 |
+
"grad_norm": 2.4400713443756104,
|
| 4375 |
+
"learning_rate": 3.7003083328990993e-06,
|
| 4376 |
+
"loss": 0.4695,
|
| 4377 |
+
"step": 624
|
| 4378 |
+
},
|
| 4379 |
+
{
|
| 4380 |
+
"epoch": 0.8799718409010912,
|
| 4381 |
+
"grad_norm": 3.7911550998687744,
|
| 4382 |
+
"learning_rate": 3.616767526868353e-06,
|
| 4383 |
+
"loss": 0.5002,
|
| 4384 |
+
"step": 625
|
| 4385 |
+
},
|
| 4386 |
+
{
|
| 4387 |
+
"epoch": 0.8813797958465329,
|
| 4388 |
+
"grad_norm": 2.611201286315918,
|
| 4389 |
+
"learning_rate": 3.534145160516844e-06,
|
| 4390 |
+
"loss": 0.345,
|
| 4391 |
+
"step": 626
|
| 4392 |
+
},
|
| 4393 |
+
{
|
| 4394 |
+
"epoch": 0.8827877507919747,
|
| 4395 |
+
"grad_norm": 2.117814064025879,
|
| 4396 |
+
"learning_rate": 3.45244286985949e-06,
|
| 4397 |
+
"loss": 0.3955,
|
| 4398 |
+
"step": 627
|
| 4399 |
+
},
|
| 4400 |
+
{
|
| 4401 |
+
"epoch": 0.8841957057374163,
|
| 4402 |
+
"grad_norm": 1.9709141254425049,
|
| 4403 |
+
"learning_rate": 3.371662272692677e-06,
|
| 4404 |
+
"loss": 0.239,
|
| 4405 |
+
"step": 628
|
| 4406 |
+
},
|
| 4407 |
+
{
|
| 4408 |
+
"epoch": 0.8856036606828581,
|
| 4409 |
+
"grad_norm": 2.1946871280670166,
|
| 4410 |
+
"learning_rate": 3.2918049685622175e-06,
|
| 4411 |
+
"loss": 0.3965,
|
| 4412 |
+
"step": 629
|
| 4413 |
+
},
|
| 4414 |
+
{
|
| 4415 |
+
"epoch": 0.8870116156282999,
|
| 4416 |
+
"grad_norm": 2.6868765354156494,
|
| 4417 |
+
"learning_rate": 3.212872538731704e-06,
|
| 4418 |
+
"loss": 0.2324,
|
| 4419 |
+
"step": 630
|
| 4420 |
+
},
|
| 4421 |
+
{
|
| 4422 |
+
"epoch": 0.8884195705737417,
|
| 4423 |
+
"grad_norm": 2.5215296745300293,
|
| 4424 |
+
"learning_rate": 3.1348665461511994e-06,
|
| 4425 |
+
"loss": 0.2174,
|
| 4426 |
+
"step": 631
|
| 4427 |
+
},
|
| 4428 |
+
{
|
| 4429 |
+
"epoch": 0.8898275255191834,
|
| 4430 |
+
"grad_norm": 2.135326862335205,
|
| 4431 |
+
"learning_rate": 3.0577885354262534e-06,
|
| 4432 |
+
"loss": 0.2029,
|
| 4433 |
+
"step": 632
|
| 4434 |
+
},
|
| 4435 |
+
{
|
| 4436 |
+
"epoch": 0.8912354804646251,
|
| 4437 |
+
"grad_norm": 7.6683220863342285,
|
| 4438 |
+
"learning_rate": 2.981640032787347e-06,
|
| 4439 |
+
"loss": 0.6038,
|
| 4440 |
+
"step": 633
|
| 4441 |
+
},
|
| 4442 |
+
{
|
| 4443 |
+
"epoch": 0.8926434354100669,
|
| 4444 |
+
"grad_norm": 9.006458282470703,
|
| 4445 |
+
"learning_rate": 2.9064225460596837e-06,
|
| 4446 |
+
"loss": 0.586,
|
| 4447 |
+
"step": 634
|
| 4448 |
+
},
|
| 4449 |
+
{
|
| 4450 |
+
"epoch": 0.8940513903555086,
|
| 4451 |
+
"grad_norm": 8.218254089355469,
|
| 4452 |
+
"learning_rate": 2.8321375646333027e-06,
|
| 4453 |
+
"loss": 0.4863,
|
| 4454 |
+
"step": 635
|
| 4455 |
+
},
|
| 4456 |
+
{
|
| 4457 |
+
"epoch": 0.8954593453009504,
|
| 4458 |
+
"grad_norm": 4.616533279418945,
|
| 4459 |
+
"learning_rate": 2.758786559433585e-06,
|
| 4460 |
+
"loss": 0.4851,
|
| 4461 |
+
"step": 636
|
| 4462 |
+
},
|
| 4463 |
+
{
|
| 4464 |
+
"epoch": 0.8968673002463922,
|
| 4465 |
+
"grad_norm": 4.0166239738464355,
|
| 4466 |
+
"learning_rate": 2.6863709828921614e-06,
|
| 4467 |
+
"loss": 0.2963,
|
| 4468 |
+
"step": 637
|
| 4469 |
+
},
|
| 4470 |
+
{
|
| 4471 |
+
"epoch": 0.8982752551918338,
|
| 4472 |
+
"grad_norm": 4.9515275955200195,
|
| 4473 |
+
"learning_rate": 2.6148922689181397e-06,
|
| 4474 |
+
"loss": 0.2624,
|
| 4475 |
+
"step": 638
|
| 4476 |
+
},
|
| 4477 |
+
{
|
| 4478 |
+
"epoch": 0.8996832101372756,
|
| 4479 |
+
"grad_norm": 4.517815589904785,
|
| 4480 |
+
"learning_rate": 2.544351832869668e-06,
|
| 4481 |
+
"loss": 0.3128,
|
| 4482 |
+
"step": 639
|
| 4483 |
+
},
|
| 4484 |
+
{
|
| 4485 |
+
"epoch": 0.9010911650827174,
|
| 4486 |
+
"grad_norm": 4.481226444244385,
|
| 4487 |
+
"learning_rate": 2.474751071525988e-06,
|
| 4488 |
+
"loss": 0.5452,
|
| 4489 |
+
"step": 640
|
| 4490 |
+
},
|
| 4491 |
+
{
|
| 4492 |
+
"epoch": 0.9024991200281591,
|
| 4493 |
+
"grad_norm": 6.772841453552246,
|
| 4494 |
+
"learning_rate": 2.4060913630596983e-06,
|
| 4495 |
+
"loss": 0.7146,
|
| 4496 |
+
"step": 641
|
| 4497 |
+
},
|
| 4498 |
+
{
|
| 4499 |
+
"epoch": 0.9039070749736009,
|
| 4500 |
+
"grad_norm": 5.757582664489746,
|
| 4501 |
+
"learning_rate": 2.3383740670095524e-06,
|
| 4502 |
+
"loss": 0.7659,
|
| 4503 |
+
"step": 642
|
| 4504 |
+
},
|
| 4505 |
+
{
|
| 4506 |
+
"epoch": 0.9053150299190426,
|
| 4507 |
+
"grad_norm": 6.684267997741699,
|
| 4508 |
+
"learning_rate": 2.2716005242534257e-06,
|
| 4509 |
+
"loss": 0.7003,
|
| 4510 |
+
"step": 643
|
| 4511 |
+
},
|
| 4512 |
+
{
|
| 4513 |
+
"epoch": 0.9067229848644843,
|
| 4514 |
+
"grad_norm": 4.235751152038574,
|
| 4515 |
+
"learning_rate": 2.205772056981875e-06,
|
| 4516 |
+
"loss": 0.2978,
|
| 4517 |
+
"step": 644
|
| 4518 |
+
},
|
| 4519 |
+
{
|
| 4520 |
+
"epoch": 0.9081309398099261,
|
| 4521 |
+
"grad_norm": 5.094956874847412,
|
| 4522 |
+
"learning_rate": 2.1408899686718997e-06,
|
| 4523 |
+
"loss": 0.5,
|
| 4524 |
+
"step": 645
|
| 4525 |
+
},
|
| 4526 |
+
{
|
| 4527 |
+
"epoch": 0.9095388947553679,
|
| 4528 |
+
"grad_norm": 4.281081676483154,
|
| 4529 |
+
"learning_rate": 2.0769555440611154e-06,
|
| 4530 |
+
"loss": 0.4886,
|
| 4531 |
+
"step": 646
|
| 4532 |
+
},
|
| 4533 |
+
{
|
| 4534 |
+
"epoch": 0.9109468497008095,
|
| 4535 |
+
"grad_norm": 3.601447343826294,
|
| 4536 |
+
"learning_rate": 2.0139700491223766e-06,
|
| 4537 |
+
"loss": 0.4544,
|
| 4538 |
+
"step": 647
|
| 4539 |
+
},
|
| 4540 |
+
{
|
| 4541 |
+
"epoch": 0.9123548046462513,
|
| 4542 |
+
"grad_norm": 4.476624965667725,
|
| 4543 |
+
"learning_rate": 1.951934731038635e-06,
|
| 4544 |
+
"loss": 0.521,
|
| 4545 |
+
"step": 648
|
| 4546 |
+
},
|
| 4547 |
+
{
|
| 4548 |
+
"epoch": 0.9137627595916931,
|
| 4549 |
+
"grad_norm": 7.66337251663208,
|
| 4550 |
+
"learning_rate": 1.8908508181783247e-06,
|
| 4551 |
+
"loss": 0.8661,
|
| 4552 |
+
"step": 649
|
| 4553 |
+
},
|
| 4554 |
+
{
|
| 4555 |
+
"epoch": 0.9151707145371348,
|
| 4556 |
+
"grad_norm": 9.684638977050781,
|
| 4557 |
+
"learning_rate": 1.8307195200709727e-06,
|
| 4558 |
+
"loss": 0.9013,
|
| 4559 |
+
"step": 650
|
| 4560 |
+
},
|
| 4561 |
+
{
|
| 4562 |
+
"epoch": 0.9165786694825766,
|
| 4563 |
+
"grad_norm": 2.1627516746520996,
|
| 4564 |
+
"learning_rate": 1.771542027383255e-06,
|
| 4565 |
+
"loss": 0.124,
|
| 4566 |
+
"step": 651
|
| 4567 |
+
},
|
| 4568 |
+
{
|
| 4569 |
+
"epoch": 0.9179866244280183,
|
| 4570 |
+
"grad_norm": 9.195147514343262,
|
| 4571 |
+
"learning_rate": 1.7133195118954838e-06,
|
| 4572 |
+
"loss": 0.5165,
|
| 4573 |
+
"step": 652
|
| 4574 |
+
},
|
| 4575 |
+
{
|
| 4576 |
+
"epoch": 0.91939457937346,
|
| 4577 |
+
"grad_norm": 5.70101261138916,
|
| 4578 |
+
"learning_rate": 1.656053126478313e-06,
|
| 4579 |
+
"loss": 0.4579,
|
| 4580 |
+
"step": 653
|
| 4581 |
+
},
|
| 4582 |
+
{
|
| 4583 |
+
"epoch": 0.9208025343189018,
|
| 4584 |
+
"grad_norm": 3.837534189224243,
|
| 4585 |
+
"learning_rate": 1.5997440050699863e-06,
|
| 4586 |
+
"loss": 0.5213,
|
| 4587 |
+
"step": 654
|
| 4588 |
+
},
|
| 4589 |
+
{
|
| 4590 |
+
"epoch": 0.9222104892643436,
|
| 4591 |
+
"grad_norm": 5.631805896759033,
|
| 4592 |
+
"learning_rate": 1.5443932626538316e-06,
|
| 4593 |
+
"loss": 0.4975,
|
| 4594 |
+
"step": 655
|
| 4595 |
+
},
|
| 4596 |
+
{
|
| 4597 |
+
"epoch": 0.9236184442097853,
|
| 4598 |
+
"grad_norm": 5.074072360992432,
|
| 4599 |
+
"learning_rate": 1.4900019952362232e-06,
|
| 4600 |
+
"loss": 0.541,
|
| 4601 |
+
"step": 656
|
| 4602 |
+
},
|
| 4603 |
+
{
|
| 4604 |
+
"epoch": 0.925026399155227,
|
| 4605 |
+
"grad_norm": 5.553448677062988,
|
| 4606 |
+
"learning_rate": 1.436571279824861e-06,
|
| 4607 |
+
"loss": 0.6306,
|
| 4608 |
+
"step": 657
|
| 4609 |
+
},
|
| 4610 |
+
{
|
| 4611 |
+
"epoch": 0.9264343541006688,
|
| 4612 |
+
"grad_norm": 4.038352966308594,
|
| 4613 |
+
"learning_rate": 1.3841021744074257e-06,
|
| 4614 |
+
"loss": 0.3506,
|
| 4615 |
+
"step": 658
|
| 4616 |
+
},
|
| 4617 |
+
{
|
| 4618 |
+
"epoch": 0.9278423090461105,
|
| 4619 |
+
"grad_norm": 2.5761544704437256,
|
| 4620 |
+
"learning_rate": 1.3325957179306626e-06,
|
| 4621 |
+
"loss": 0.2508,
|
| 4622 |
+
"step": 659
|
| 4623 |
+
},
|
| 4624 |
+
{
|
| 4625 |
+
"epoch": 0.9292502639915523,
|
| 4626 |
+
"grad_norm": 3.6852400302886963,
|
| 4627 |
+
"learning_rate": 1.2820529302798035e-06,
|
| 4628 |
+
"loss": 0.4874,
|
| 4629 |
+
"step": 660
|
| 4630 |
+
},
|
| 4631 |
+
{
|
| 4632 |
+
"epoch": 0.930658218936994,
|
| 4633 |
+
"grad_norm": 3.062720537185669,
|
| 4634 |
+
"learning_rate": 1.232474812258344e-06,
|
| 4635 |
+
"loss": 0.3289,
|
| 4636 |
+
"step": 661
|
| 4637 |
+
},
|
| 4638 |
+
{
|
| 4639 |
+
"epoch": 0.9320661738824357,
|
| 4640 |
+
"grad_norm": 5.507415771484375,
|
| 4641 |
+
"learning_rate": 1.1838623455682652e-06,
|
| 4642 |
+
"loss": 0.5935,
|
| 4643 |
+
"step": 662
|
| 4644 |
+
},
|
| 4645 |
+
{
|
| 4646 |
+
"epoch": 0.9334741288278775,
|
| 4647 |
+
"grad_norm": 6.665245532989502,
|
| 4648 |
+
"learning_rate": 1.1362164927905595e-06,
|
| 4649 |
+
"loss": 0.7645,
|
| 4650 |
+
"step": 663
|
| 4651 |
+
},
|
| 4652 |
+
{
|
| 4653 |
+
"epoch": 0.9348820837733193,
|
| 4654 |
+
"grad_norm": 6.005698204040527,
|
| 4655 |
+
"learning_rate": 1.0895381973662023e-06,
|
| 4656 |
+
"loss": 0.6456,
|
| 4657 |
+
"step": 664
|
| 4658 |
+
},
|
| 4659 |
+
{
|
| 4660 |
+
"epoch": 0.936290038718761,
|
| 4661 |
+
"grad_norm": 5.974649429321289,
|
| 4662 |
+
"learning_rate": 1.0438283835774387e-06,
|
| 4663 |
+
"loss": 0.7067,
|
| 4664 |
+
"step": 665
|
| 4665 |
+
},
|
| 4666 |
+
{
|
| 4667 |
+
"epoch": 0.9376979936642027,
|
| 4668 |
+
"grad_norm": 3.828660011291504,
|
| 4669 |
+
"learning_rate": 9.990879565295153e-07,
|
| 4670 |
+
"loss": 0.5732,
|
| 4671 |
+
"step": 666
|
| 4672 |
+
},
|
| 4673 |
+
{
|
| 4674 |
+
"epoch": 0.9391059486096445,
|
| 4675 |
+
"grad_norm": 4.172199249267578,
|
| 4676 |
+
"learning_rate": 9.553178021327326e-07,
|
| 4677 |
+
"loss": 0.4813,
|
| 4678 |
+
"step": 667
|
| 4679 |
+
},
|
| 4680 |
+
{
|
| 4681 |
+
"epoch": 0.9405139035550862,
|
| 4682 |
+
"grad_norm": 3.0832149982452393,
|
| 4683 |
+
"learning_rate": 9.12518787084915e-07,
|
| 4684 |
+
"loss": 0.3741,
|
| 4685 |
+
"step": 668
|
| 4686 |
+
},
|
| 4687 |
+
{
|
| 4688 |
+
"epoch": 0.941921858500528,
|
| 4689 |
+
"grad_norm": 2.6431021690368652,
|
| 4690 |
+
"learning_rate": 8.706917588542418e-07,
|
| 4691 |
+
"loss": 0.4311,
|
| 4692 |
+
"step": 669
|
| 4693 |
+
},
|
| 4694 |
+
{
|
| 4695 |
+
"epoch": 0.9433298134459698,
|
| 4696 |
+
"grad_norm": 3.608599901199341,
|
| 4697 |
+
"learning_rate": 8.298375456624652e-07,
|
| 4698 |
+
"loss": 0.4787,
|
| 4699 |
+
"step": 670
|
| 4700 |
+
},
|
| 4701 |
+
{
|
| 4702 |
+
"epoch": 0.9447377683914114,
|
| 4703 |
+
"grad_norm": 2.496652364730835,
|
| 4704 |
+
"learning_rate": 7.899569564685294e-07,
|
| 4705 |
+
"loss": 0.392,
|
| 4706 |
+
"step": 671
|
| 4707 |
+
},
|
| 4708 |
+
{
|
| 4709 |
+
"epoch": 0.9461457233368532,
|
| 4710 |
+
"grad_norm": 2.865622043609619,
|
| 4711 |
+
"learning_rate": 7.510507809525335e-07,
|
| 4712 |
+
"loss": 0.4334,
|
| 4713 |
+
"step": 672
|
| 4714 |
+
},
|
| 4715 |
+
{
|
| 4716 |
+
"epoch": 0.947553678282295,
|
| 4717 |
+
"grad_norm": 2.384401798248291,
|
| 4718 |
+
"learning_rate": 7.131197895000941e-07,
|
| 4719 |
+
"loss": 0.3749,
|
| 4720 |
+
"step": 673
|
| 4721 |
+
},
|
| 4722 |
+
{
|
| 4723 |
+
"epoch": 0.9489616332277367,
|
| 4724 |
+
"grad_norm": 3.3137624263763428,
|
| 4725 |
+
"learning_rate": 6.761647331871013e-07,
|
| 4726 |
+
"loss": 0.3988,
|
| 4727 |
+
"step": 674
|
| 4728 |
+
},
|
| 4729 |
+
{
|
| 4730 |
+
"epoch": 0.9503695881731784,
|
| 4731 |
+
"grad_norm": 3.8170225620269775,
|
| 4732 |
+
"learning_rate": 6.401863437648481e-07,
|
| 4733 |
+
"loss": 0.4322,
|
| 4734 |
+
"step": 675
|
| 4735 |
+
},
|
| 4736 |
+
{
|
| 4737 |
+
"epoch": 0.9517775431186202,
|
| 4738 |
+
"grad_norm": 3.7484734058380127,
|
| 4739 |
+
"learning_rate": 6.051853336455249e-07,
|
| 4740 |
+
"loss": 0.3821,
|
| 4741 |
+
"step": 676
|
| 4742 |
+
},
|
| 4743 |
+
{
|
| 4744 |
+
"epoch": 0.9531854980640619,
|
| 4745 |
+
"grad_norm": 1.4913954734802246,
|
| 4746 |
+
"learning_rate": 5.711623958881195e-07,
|
| 4747 |
+
"loss": 0.2762,
|
| 4748 |
+
"step": 677
|
| 4749 |
+
},
|
| 4750 |
+
{
|
| 4751 |
+
"epoch": 0.9545934530095037,
|
| 4752 |
+
"grad_norm": 1.4391337633132935,
|
| 4753 |
+
"learning_rate": 5.381182041847232e-07,
|
| 4754 |
+
"loss": 0.2613,
|
| 4755 |
+
"step": 678
|
| 4756 |
+
},
|
| 4757 |
+
{
|
| 4758 |
+
"epoch": 0.9560014079549455,
|
| 4759 |
+
"grad_norm": 1.9773917198181152,
|
| 4760 |
+
"learning_rate": 5.06053412847135e-07,
|
| 4761 |
+
"loss": 0.2667,
|
| 4762 |
+
"step": 679
|
| 4763 |
+
},
|
| 4764 |
+
{
|
| 4765 |
+
"epoch": 0.9574093629003871,
|
| 4766 |
+
"grad_norm": 2.6283700466156006,
|
| 4767 |
+
"learning_rate": 4.749686567939504e-07,
|
| 4768 |
+
"loss": 0.286,
|
| 4769 |
+
"step": 680
|
| 4770 |
+
},
|
| 4771 |
+
{
|
| 4772 |
+
"epoch": 0.9588173178458289,
|
| 4773 |
+
"grad_norm": 2.4537582397460938,
|
| 4774 |
+
"learning_rate": 4.448645515379657e-07,
|
| 4775 |
+
"loss": 0.2895,
|
| 4776 |
+
"step": 681
|
| 4777 |
+
},
|
| 4778 |
+
{
|
| 4779 |
+
"epoch": 0.9602252727912707,
|
| 4780 |
+
"grad_norm": 1.875128149986267,
|
| 4781 |
+
"learning_rate": 4.157416931740099e-07,
|
| 4782 |
+
"loss": 0.1732,
|
| 4783 |
+
"step": 682
|
| 4784 |
+
},
|
| 4785 |
+
{
|
| 4786 |
+
"epoch": 0.9616332277367124,
|
| 4787 |
+
"grad_norm": 1.9811691045761108,
|
| 4788 |
+
"learning_rate": 3.8760065836712653e-07,
|
| 4789 |
+
"loss": 0.2129,
|
| 4790 |
+
"step": 683
|
| 4791 |
+
},
|
| 4792 |
+
{
|
| 4793 |
+
"epoch": 0.9630411826821542,
|
| 4794 |
+
"grad_norm": 5.41964864730835,
|
| 4795 |
+
"learning_rate": 3.604420043411494e-07,
|
| 4796 |
+
"loss": 0.451,
|
| 4797 |
+
"step": 684
|
| 4798 |
+
},
|
| 4799 |
+
{
|
| 4800 |
+
"epoch": 0.9644491376275959,
|
| 4801 |
+
"grad_norm": 5.268420219421387,
|
| 4802 |
+
"learning_rate": 3.342662688676945e-07,
|
| 4803 |
+
"loss": 0.5786,
|
| 4804 |
+
"step": 685
|
| 4805 |
+
},
|
| 4806 |
+
{
|
| 4807 |
+
"epoch": 0.9658570925730376,
|
| 4808 |
+
"grad_norm": 6.128338813781738,
|
| 4809 |
+
"learning_rate": 3.090739702554801e-07,
|
| 4810 |
+
"loss": 0.4144,
|
| 4811 |
+
"step": 686
|
| 4812 |
+
},
|
| 4813 |
+
{
|
| 4814 |
+
"epoch": 0.9672650475184794,
|
| 4815 |
+
"grad_norm": 3.8407442569732666,
|
| 4816 |
+
"learning_rate": 2.8486560734009573e-07,
|
| 4817 |
+
"loss": 0.5015,
|
| 4818 |
+
"step": 687
|
| 4819 |
+
},
|
| 4820 |
+
{
|
| 4821 |
+
"epoch": 0.9686730024639212,
|
| 4822 |
+
"grad_norm": 3.2146735191345215,
|
| 4823 |
+
"learning_rate": 2.6164165947408226e-07,
|
| 4824 |
+
"loss": 0.2868,
|
| 4825 |
+
"step": 688
|
| 4826 |
+
},
|
| 4827 |
+
{
|
| 4828 |
+
"epoch": 0.970080957409363,
|
| 4829 |
+
"grad_norm": 2.745905876159668,
|
| 4830 |
+
"learning_rate": 2.3940258651749536e-07,
|
| 4831 |
+
"loss": 0.2684,
|
| 4832 |
+
"step": 689
|
| 4833 |
+
},
|
| 4834 |
+
{
|
| 4835 |
+
"epoch": 0.9714889123548046,
|
| 4836 |
+
"grad_norm": 5.39619255065918,
|
| 4837 |
+
"learning_rate": 2.1814882882875698e-07,
|
| 4838 |
+
"loss": 0.6061,
|
| 4839 |
+
"step": 690
|
| 4840 |
+
},
|
| 4841 |
+
{
|
| 4842 |
+
"epoch": 0.9728968673002464,
|
| 4843 |
+
"grad_norm": 6.267408847808838,
|
| 4844 |
+
"learning_rate": 1.9788080725593462e-07,
|
| 4845 |
+
"loss": 0.7417,
|
| 4846 |
+
"step": 691
|
| 4847 |
+
},
|
| 4848 |
+
{
|
| 4849 |
+
"epoch": 0.9743048222456882,
|
| 4850 |
+
"grad_norm": 9.174079895019531,
|
| 4851 |
+
"learning_rate": 1.7859892312844795e-07,
|
| 4852 |
+
"loss": 0.5494,
|
| 4853 |
+
"step": 692
|
| 4854 |
+
},
|
| 4855 |
+
{
|
| 4856 |
+
"epoch": 0.9757127771911299,
|
| 4857 |
+
"grad_norm": 8.090996742248535,
|
| 4858 |
+
"learning_rate": 1.6030355824908639e-07,
|
| 4859 |
+
"loss": 0.8298,
|
| 4860 |
+
"step": 693
|
| 4861 |
+
},
|
| 4862 |
+
{
|
| 4863 |
+
"epoch": 0.9771207321365716,
|
| 4864 |
+
"grad_norm": 4.43288516998291,
|
| 4865 |
+
"learning_rate": 1.4299507488644838e-07,
|
| 4866 |
+
"loss": 0.4159,
|
| 4867 |
+
"step": 694
|
| 4868 |
+
},
|
| 4869 |
+
{
|
| 4870 |
+
"epoch": 0.9785286870820133,
|
| 4871 |
+
"grad_norm": 2.211864948272705,
|
| 4872 |
+
"learning_rate": 1.2667381576779712e-07,
|
| 4873 |
+
"loss": 0.274,
|
| 4874 |
+
"step": 695
|
| 4875 |
+
},
|
| 4876 |
+
{
|
| 4877 |
+
"epoch": 0.9799366420274551,
|
| 4878 |
+
"grad_norm": 4.036583423614502,
|
| 4879 |
+
"learning_rate": 1.1134010407223838e-07,
|
| 4880 |
+
"loss": 0.573,
|
| 4881 |
+
"step": 696
|
| 4882 |
+
},
|
| 4883 |
+
{
|
| 4884 |
+
"epoch": 0.9813445969728969,
|
| 4885 |
+
"grad_norm": 3.924792766571045,
|
| 4886 |
+
"learning_rate": 9.69942434243587e-08,
|
| 4887 |
+
"loss": 0.4368,
|
| 4888 |
+
"step": 697
|
| 4889 |
+
},
|
| 4890 |
+
{
|
| 4891 |
+
"epoch": 0.9827525519183387,
|
| 4892 |
+
"grad_norm": 6.395772457122803,
|
| 4893 |
+
"learning_rate": 8.363651788816929e-08,
|
| 4894 |
+
"loss": 0.7059,
|
| 4895 |
+
"step": 698
|
| 4896 |
+
},
|
| 4897 |
+
{
|
| 4898 |
+
"epoch": 0.9841605068637803,
|
| 4899 |
+
"grad_norm": 6.783596992492676,
|
| 4900 |
+
"learning_rate": 7.126719196152154e-08,
|
| 4901 |
+
"loss": 0.8242,
|
| 4902 |
+
"step": 699
|
| 4903 |
+
},
|
| 4904 |
+
{
|
| 4905 |
+
"epoch": 0.9855684618092221,
|
| 4906 |
+
"grad_norm": 9.767477035522461,
|
| 4907 |
+
"learning_rate": 5.98865105708557e-08,
|
| 4908 |
+
"loss": 0.9057,
|
| 4909 |
+
"step": 700
|
| 4910 |
+
},
|
| 4911 |
+
{
|
| 4912 |
+
"epoch": 0.9869764167546639,
|
| 4913 |
+
"grad_norm": 1.9957053661346436,
|
| 4914 |
+
"learning_rate": 4.949469906633808e-08,
|
| 4915 |
+
"loss": 0.3597,
|
| 4916 |
+
"step": 701
|
| 4917 |
+
},
|
| 4918 |
+
{
|
| 4919 |
+
"epoch": 0.9883843717001056,
|
| 4920 |
+
"grad_norm": 1.9467065334320068,
|
| 4921 |
+
"learning_rate": 4.0091963217420147e-08,
|
| 4922 |
+
"loss": 0.3514,
|
| 4923 |
+
"step": 702
|
| 4924 |
+
},
|
| 4925 |
+
{
|
| 4926 |
+
"epoch": 0.9897923266455474,
|
| 4927 |
+
"grad_norm": 5.04243803024292,
|
| 4928 |
+
"learning_rate": 3.167848920873628e-08,
|
| 4929 |
+
"loss": 0.6304,
|
| 4930 |
+
"step": 703
|
| 4931 |
+
},
|
| 4932 |
+
{
|
| 4933 |
+
"epoch": 0.991200281590989,
|
| 4934 |
+
"grad_norm": 2.3117592334747314,
|
| 4935 |
+
"learning_rate": 2.425444363645113e-08,
|
| 4936 |
+
"loss": 0.3404,
|
| 4937 |
+
"step": 704
|
| 4938 |
+
},
|
| 4939 |
+
{
|
| 4940 |
+
"epoch": 0.9926082365364308,
|
| 4941 |
+
"grad_norm": 1.944746732711792,
|
| 4942 |
+
"learning_rate": 1.7819973504940025e-08,
|
| 4943 |
+
"loss": 0.3422,
|
| 4944 |
+
"step": 705
|
| 4945 |
+
},
|
| 4946 |
+
{
|
| 4947 |
+
"epoch": 0.9940161914818726,
|
| 4948 |
+
"grad_norm": 1.3621710538864136,
|
| 4949 |
+
"learning_rate": 1.2375206223874669e-08,
|
| 4950 |
+
"loss": 0.2572,
|
| 4951 |
+
"step": 706
|
| 4952 |
+
},
|
| 4953 |
+
{
|
| 4954 |
+
"epoch": 0.9954241464273144,
|
| 4955 |
+
"grad_norm": 3.6393015384674072,
|
| 4956 |
+
"learning_rate": 7.920249605719576e-09,
|
| 4957 |
+
"loss": 0.4432,
|
| 4958 |
+
"step": 707
|
| 4959 |
+
},
|
| 4960 |
+
{
|
| 4961 |
+
"epoch": 0.996832101372756,
|
| 4962 |
+
"grad_norm": 3.2820541858673096,
|
| 4963 |
+
"learning_rate": 4.455191863578234e-09,
|
| 4964 |
+
"loss": 0.5374,
|
| 4965 |
+
"step": 708
|
| 4966 |
+
},
|
| 4967 |
+
{
|
| 4968 |
+
"epoch": 0.9982400563181978,
|
| 4969 |
+
"grad_norm": 2.710822820663452,
|
| 4970 |
+
"learning_rate": 1.980101609455609e-09,
|
| 4971 |
+
"loss": 0.4186,
|
| 4972 |
+
"step": 709
|
| 4973 |
+
},
|
| 4974 |
+
{
|
| 4975 |
+
"epoch": 0.9996480112636396,
|
| 4976 |
+
"grad_norm": 4.686365604400635,
|
| 4977 |
+
"learning_rate": 4.9502785288702e-10,
|
| 4978 |
+
"loss": 0.7881,
|
| 4979 |
+
"step": 710
|
| 4980 |
+
},
|
| 4981 |
+
{
|
| 4982 |
+
"epoch": 0.9996480112636396,
|
| 4983 |
+
"eval_loss": 0.11581157892942429,
|
| 4984 |
+
"eval_runtime": 44.8917,
|
| 4985 |
+
"eval_samples_per_second": 13.321,
|
| 4986 |
+
"eval_steps_per_second": 3.341,
|
| 4987 |
+
"step": 710
|
| 4988 |
+
},
|
| 4989 |
+
{
|
| 4990 |
+
"epoch": 1.0010559662090812,
|
| 4991 |
+
"grad_norm": 5.4512128829956055,
|
| 4992 |
+
"learning_rate": 0.0,
|
| 4993 |
+
"loss": 0.2848,
|
| 4994 |
+
"step": 711
|
| 4995 |
}
|
| 4996 |
],
|
| 4997 |
"logging_steps": 1,
|
|
|
|
| 5006 |
"should_evaluate": false,
|
| 5007 |
"should_log": false,
|
| 5008 |
"should_save": true,
|
| 5009 |
+
"should_training_stop": true
|
| 5010 |
},
|
| 5011 |
"attributes": {}
|
| 5012 |
}
|
| 5013 |
},
|
| 5014 |
+
"total_flos": 7.429122375674757e+17,
|
| 5015 |
"train_batch_size": 4,
|
| 5016 |
"trial_name": null,
|
| 5017 |
"trial_params": null
|