Training in progress, step 600, checkpoint
Browse files
last-checkpoint/adapter_model.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 239536272
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:380a9f8bd1452df8063d91dc4f5608799c0eed96ed4cdc18292e7cb70a2346a6
|
| 3 |
size 239536272
|
last-checkpoint/optimizer.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 183010548
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:ce3b58265d130261bec295db93a3ae710f09eb686a51c216c625ecac4a9bbd35
|
| 3 |
size 183010548
|
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 sha256:84a73f7114eacaa8f3d0a34708cd3c380dc843298650a38ecf02f4e2db17ae9f
|
| 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:d4dd75a8923222c498d00e1167f4c1be1383dd919c1ebf448c63527342992a3c
|
| 3 |
size 1064
|
last-checkpoint/trainer_state.json
CHANGED
|
@@ -1,9 +1,9 @@
|
|
| 1 |
{
|
| 2 |
-
"best_metric": 0.
|
| 3 |
-
"best_model_checkpoint": "miner_id_24/checkpoint-
|
| 4 |
-
"epoch": 0.
|
| 5 |
"eval_steps": 100,
|
| 6 |
-
"global_step":
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
@@ -2847,6 +2847,1422 @@
|
|
| 2847 |
"eval_samples_per_second": 3.45,
|
| 2848 |
"eval_steps_per_second": 3.45,
|
| 2849 |
"step": 400
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 2850 |
}
|
| 2851 |
],
|
| 2852 |
"logging_steps": 1,
|
|
@@ -2875,7 +4291,7 @@
|
|
| 2875 |
"attributes": {}
|
| 2876 |
}
|
| 2877 |
},
|
| 2878 |
-
"total_flos":
|
| 2879 |
"train_batch_size": 1,
|
| 2880 |
"trial_name": null,
|
| 2881 |
"trial_params": null
|
|
|
|
| 1 |
{
|
| 2 |
+
"best_metric": 0.7834404110908508,
|
| 3 |
+
"best_model_checkpoint": "miner_id_24/checkpoint-600",
|
| 4 |
+
"epoch": 0.09845144087785868,
|
| 5 |
"eval_steps": 100,
|
| 6 |
+
"global_step": 600,
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
|
|
| 2847 |
"eval_samples_per_second": 3.45,
|
| 2848 |
"eval_steps_per_second": 3.45,
|
| 2849 |
"step": 400
|
| 2850 |
+
},
|
| 2851 |
+
{
|
| 2852 |
+
"epoch": 0.06579837965336889,
|
| 2853 |
+
"grad_norm": 0.15986493229866028,
|
| 2854 |
+
"learning_rate": 0.0001997741149322696,
|
| 2855 |
+
"loss": 0.5572,
|
| 2856 |
+
"step": 401
|
| 2857 |
+
},
|
| 2858 |
+
{
|
| 2859 |
+
"epoch": 0.06596246538816532,
|
| 2860 |
+
"grad_norm": 0.2333621084690094,
|
| 2861 |
+
"learning_rate": 0.00019977295847037837,
|
| 2862 |
+
"loss": 0.7397,
|
| 2863 |
+
"step": 402
|
| 2864 |
+
},
|
| 2865 |
+
{
|
| 2866 |
+
"epoch": 0.06612655112296174,
|
| 2867 |
+
"grad_norm": 0.2305152714252472,
|
| 2868 |
+
"learning_rate": 0.0001997717990590413,
|
| 2869 |
+
"loss": 0.7124,
|
| 2870 |
+
"step": 403
|
| 2871 |
+
},
|
| 2872 |
+
{
|
| 2873 |
+
"epoch": 0.06629063685775818,
|
| 2874 |
+
"grad_norm": 0.22249913215637207,
|
| 2875 |
+
"learning_rate": 0.00019977063669829271,
|
| 2876 |
+
"loss": 0.672,
|
| 2877 |
+
"step": 404
|
| 2878 |
+
},
|
| 2879 |
+
{
|
| 2880 |
+
"epoch": 0.06645472259255461,
|
| 2881 |
+
"grad_norm": 0.23822623491287231,
|
| 2882 |
+
"learning_rate": 0.00019976947138816695,
|
| 2883 |
+
"loss": 0.7609,
|
| 2884 |
+
"step": 405
|
| 2885 |
+
},
|
| 2886 |
+
{
|
| 2887 |
+
"epoch": 0.06661880832735104,
|
| 2888 |
+
"grad_norm": 0.4435589611530304,
|
| 2889 |
+
"learning_rate": 0.00019976830312869848,
|
| 2890 |
+
"loss": 0.7179,
|
| 2891 |
+
"step": 406
|
| 2892 |
+
},
|
| 2893 |
+
{
|
| 2894 |
+
"epoch": 0.06678289406214748,
|
| 2895 |
+
"grad_norm": 8.647324562072754,
|
| 2896 |
+
"learning_rate": 0.0001997671319199218,
|
| 2897 |
+
"loss": 0.8875,
|
| 2898 |
+
"step": 407
|
| 2899 |
+
},
|
| 2900 |
+
{
|
| 2901 |
+
"epoch": 0.0669469797969439,
|
| 2902 |
+
"grad_norm": 0.37018343806266785,
|
| 2903 |
+
"learning_rate": 0.00019976595776187154,
|
| 2904 |
+
"loss": 0.8438,
|
| 2905 |
+
"step": 408
|
| 2906 |
+
},
|
| 2907 |
+
{
|
| 2908 |
+
"epoch": 0.06711106553174033,
|
| 2909 |
+
"grad_norm": 0.2479041963815689,
|
| 2910 |
+
"learning_rate": 0.00019976478065458243,
|
| 2911 |
+
"loss": 0.686,
|
| 2912 |
+
"step": 409
|
| 2913 |
+
},
|
| 2914 |
+
{
|
| 2915 |
+
"epoch": 0.06727515126653677,
|
| 2916 |
+
"grad_norm": 0.2105472832918167,
|
| 2917 |
+
"learning_rate": 0.00019976360059808927,
|
| 2918 |
+
"loss": 0.7365,
|
| 2919 |
+
"step": 410
|
| 2920 |
+
},
|
| 2921 |
+
{
|
| 2922 |
+
"epoch": 0.0674392370013332,
|
| 2923 |
+
"grad_norm": 0.30589988827705383,
|
| 2924 |
+
"learning_rate": 0.00019976241759242692,
|
| 2925 |
+
"loss": 0.8328,
|
| 2926 |
+
"step": 411
|
| 2927 |
+
},
|
| 2928 |
+
{
|
| 2929 |
+
"epoch": 0.06760332273612962,
|
| 2930 |
+
"grad_norm": 0.24568642675876617,
|
| 2931 |
+
"learning_rate": 0.0001997612316376304,
|
| 2932 |
+
"loss": 0.8076,
|
| 2933 |
+
"step": 412
|
| 2934 |
+
},
|
| 2935 |
+
{
|
| 2936 |
+
"epoch": 0.06776740847092606,
|
| 2937 |
+
"grad_norm": 0.22725167870521545,
|
| 2938 |
+
"learning_rate": 0.00019976004273373468,
|
| 2939 |
+
"loss": 0.6269,
|
| 2940 |
+
"step": 413
|
| 2941 |
+
},
|
| 2942 |
+
{
|
| 2943 |
+
"epoch": 0.06793149420572249,
|
| 2944 |
+
"grad_norm": 0.30800437927246094,
|
| 2945 |
+
"learning_rate": 0.00019975885088077499,
|
| 2946 |
+
"loss": 0.685,
|
| 2947 |
+
"step": 414
|
| 2948 |
+
},
|
| 2949 |
+
{
|
| 2950 |
+
"epoch": 0.06809557994051892,
|
| 2951 |
+
"grad_norm": 0.27556002140045166,
|
| 2952 |
+
"learning_rate": 0.00019975765607878655,
|
| 2953 |
+
"loss": 0.5931,
|
| 2954 |
+
"step": 415
|
| 2955 |
+
},
|
| 2956 |
+
{
|
| 2957 |
+
"epoch": 0.06825966567531536,
|
| 2958 |
+
"grad_norm": 0.46520912647247314,
|
| 2959 |
+
"learning_rate": 0.0001997564583278046,
|
| 2960 |
+
"loss": 0.7079,
|
| 2961 |
+
"step": 416
|
| 2962 |
+
},
|
| 2963 |
+
{
|
| 2964 |
+
"epoch": 0.06842375141011178,
|
| 2965 |
+
"grad_norm": 0.5119454264640808,
|
| 2966 |
+
"learning_rate": 0.00019975525762786468,
|
| 2967 |
+
"loss": 0.9523,
|
| 2968 |
+
"step": 417
|
| 2969 |
+
},
|
| 2970 |
+
{
|
| 2971 |
+
"epoch": 0.06858783714490821,
|
| 2972 |
+
"grad_norm": 0.23642492294311523,
|
| 2973 |
+
"learning_rate": 0.0001997540539790022,
|
| 2974 |
+
"loss": 0.723,
|
| 2975 |
+
"step": 418
|
| 2976 |
+
},
|
| 2977 |
+
{
|
| 2978 |
+
"epoch": 0.06875192287970465,
|
| 2979 |
+
"grad_norm": 0.3586544096469879,
|
| 2980 |
+
"learning_rate": 0.0001997528473812527,
|
| 2981 |
+
"loss": 0.864,
|
| 2982 |
+
"step": 419
|
| 2983 |
+
},
|
| 2984 |
+
{
|
| 2985 |
+
"epoch": 0.06891600861450108,
|
| 2986 |
+
"grad_norm": 0.30147257447242737,
|
| 2987 |
+
"learning_rate": 0.00019975163783465195,
|
| 2988 |
+
"loss": 0.8027,
|
| 2989 |
+
"step": 420
|
| 2990 |
+
},
|
| 2991 |
+
{
|
| 2992 |
+
"epoch": 0.0690800943492975,
|
| 2993 |
+
"grad_norm": 0.2807484269142151,
|
| 2994 |
+
"learning_rate": 0.00019975042533923568,
|
| 2995 |
+
"loss": 0.7956,
|
| 2996 |
+
"step": 421
|
| 2997 |
+
},
|
| 2998 |
+
{
|
| 2999 |
+
"epoch": 0.06924418008409394,
|
| 3000 |
+
"grad_norm": 0.3472043573856354,
|
| 3001 |
+
"learning_rate": 0.00019974920989503968,
|
| 3002 |
+
"loss": 0.8275,
|
| 3003 |
+
"step": 422
|
| 3004 |
+
},
|
| 3005 |
+
{
|
| 3006 |
+
"epoch": 0.06940826581889037,
|
| 3007 |
+
"grad_norm": 0.2816649377346039,
|
| 3008 |
+
"learning_rate": 0.00019974799150209992,
|
| 3009 |
+
"loss": 0.7084,
|
| 3010 |
+
"step": 423
|
| 3011 |
+
},
|
| 3012 |
+
{
|
| 3013 |
+
"epoch": 0.0695723515536868,
|
| 3014 |
+
"grad_norm": 0.35236573219299316,
|
| 3015 |
+
"learning_rate": 0.00019974677016045244,
|
| 3016 |
+
"loss": 0.8452,
|
| 3017 |
+
"step": 424
|
| 3018 |
+
},
|
| 3019 |
+
{
|
| 3020 |
+
"epoch": 0.06973643728848324,
|
| 3021 |
+
"grad_norm": 0.2761305868625641,
|
| 3022 |
+
"learning_rate": 0.0001997455458701333,
|
| 3023 |
+
"loss": 0.8538,
|
| 3024 |
+
"step": 425
|
| 3025 |
+
},
|
| 3026 |
+
{
|
| 3027 |
+
"epoch": 0.06990052302327966,
|
| 3028 |
+
"grad_norm": 0.25361061096191406,
|
| 3029 |
+
"learning_rate": 0.0001997443186311787,
|
| 3030 |
+
"loss": 0.5883,
|
| 3031 |
+
"step": 426
|
| 3032 |
+
},
|
| 3033 |
+
{
|
| 3034 |
+
"epoch": 0.07006460875807609,
|
| 3035 |
+
"grad_norm": 0.3191013038158417,
|
| 3036 |
+
"learning_rate": 0.00019974308844362496,
|
| 3037 |
+
"loss": 0.8699,
|
| 3038 |
+
"step": 427
|
| 3039 |
+
},
|
| 3040 |
+
{
|
| 3041 |
+
"epoch": 0.07022869449287253,
|
| 3042 |
+
"grad_norm": 0.28123271465301514,
|
| 3043 |
+
"learning_rate": 0.00019974185530750838,
|
| 3044 |
+
"loss": 0.8439,
|
| 3045 |
+
"step": 428
|
| 3046 |
+
},
|
| 3047 |
+
{
|
| 3048 |
+
"epoch": 0.07039278022766896,
|
| 3049 |
+
"grad_norm": 0.5743336081504822,
|
| 3050 |
+
"learning_rate": 0.00019974061922286546,
|
| 3051 |
+
"loss": 0.9119,
|
| 3052 |
+
"step": 429
|
| 3053 |
+
},
|
| 3054 |
+
{
|
| 3055 |
+
"epoch": 0.07055686596246538,
|
| 3056 |
+
"grad_norm": 0.3127213418483734,
|
| 3057 |
+
"learning_rate": 0.00019973938018973274,
|
| 3058 |
+
"loss": 0.9264,
|
| 3059 |
+
"step": 430
|
| 3060 |
+
},
|
| 3061 |
+
{
|
| 3062 |
+
"epoch": 0.07072095169726182,
|
| 3063 |
+
"grad_norm": 0.34507542848587036,
|
| 3064 |
+
"learning_rate": 0.0001997381382081468,
|
| 3065 |
+
"loss": 0.7802,
|
| 3066 |
+
"step": 431
|
| 3067 |
+
},
|
| 3068 |
+
{
|
| 3069 |
+
"epoch": 0.07088503743205825,
|
| 3070 |
+
"grad_norm": 0.28120848536491394,
|
| 3071 |
+
"learning_rate": 0.00019973689327814443,
|
| 3072 |
+
"loss": 0.7031,
|
| 3073 |
+
"step": 432
|
| 3074 |
+
},
|
| 3075 |
+
{
|
| 3076 |
+
"epoch": 0.07104912316685467,
|
| 3077 |
+
"grad_norm": 0.4257034659385681,
|
| 3078 |
+
"learning_rate": 0.0001997356453997624,
|
| 3079 |
+
"loss": 0.8775,
|
| 3080 |
+
"step": 433
|
| 3081 |
+
},
|
| 3082 |
+
{
|
| 3083 |
+
"epoch": 0.07121320890165111,
|
| 3084 |
+
"grad_norm": 0.3616534173488617,
|
| 3085 |
+
"learning_rate": 0.00019973439457303757,
|
| 3086 |
+
"loss": 0.8013,
|
| 3087 |
+
"step": 434
|
| 3088 |
+
},
|
| 3089 |
+
{
|
| 3090 |
+
"epoch": 0.07137729463644754,
|
| 3091 |
+
"grad_norm": 0.38141894340515137,
|
| 3092 |
+
"learning_rate": 0.00019973314079800697,
|
| 3093 |
+
"loss": 0.8423,
|
| 3094 |
+
"step": 435
|
| 3095 |
+
},
|
| 3096 |
+
{
|
| 3097 |
+
"epoch": 0.07154138037124398,
|
| 3098 |
+
"grad_norm": 0.31338822841644287,
|
| 3099 |
+
"learning_rate": 0.0001997318840747076,
|
| 3100 |
+
"loss": 0.8506,
|
| 3101 |
+
"step": 436
|
| 3102 |
+
},
|
| 3103 |
+
{
|
| 3104 |
+
"epoch": 0.07170546610604041,
|
| 3105 |
+
"grad_norm": 0.3113965690135956,
|
| 3106 |
+
"learning_rate": 0.00019973062440317664,
|
| 3107 |
+
"loss": 0.8534,
|
| 3108 |
+
"step": 437
|
| 3109 |
+
},
|
| 3110 |
+
{
|
| 3111 |
+
"epoch": 0.07186955184083683,
|
| 3112 |
+
"grad_norm": 0.39697837829589844,
|
| 3113 |
+
"learning_rate": 0.00019972936178345134,
|
| 3114 |
+
"loss": 0.8224,
|
| 3115 |
+
"step": 438
|
| 3116 |
+
},
|
| 3117 |
+
{
|
| 3118 |
+
"epoch": 0.07203363757563327,
|
| 3119 |
+
"grad_norm": 0.30266883969306946,
|
| 3120 |
+
"learning_rate": 0.00019972809621556902,
|
| 3121 |
+
"loss": 0.7729,
|
| 3122 |
+
"step": 439
|
| 3123 |
+
},
|
| 3124 |
+
{
|
| 3125 |
+
"epoch": 0.0721977233104297,
|
| 3126 |
+
"grad_norm": 0.3997204601764679,
|
| 3127 |
+
"learning_rate": 0.0001997268276995671,
|
| 3128 |
+
"loss": 0.7422,
|
| 3129 |
+
"step": 440
|
| 3130 |
+
},
|
| 3131 |
+
{
|
| 3132 |
+
"epoch": 0.07236180904522613,
|
| 3133 |
+
"grad_norm": 0.4488981366157532,
|
| 3134 |
+
"learning_rate": 0.00019972555623548306,
|
| 3135 |
+
"loss": 0.898,
|
| 3136 |
+
"step": 441
|
| 3137 |
+
},
|
| 3138 |
+
{
|
| 3139 |
+
"epoch": 0.07252589478002257,
|
| 3140 |
+
"grad_norm": 0.3362194299697876,
|
| 3141 |
+
"learning_rate": 0.0001997242818233545,
|
| 3142 |
+
"loss": 0.8179,
|
| 3143 |
+
"step": 442
|
| 3144 |
+
},
|
| 3145 |
+
{
|
| 3146 |
+
"epoch": 0.072689980514819,
|
| 3147 |
+
"grad_norm": 0.4512924551963806,
|
| 3148 |
+
"learning_rate": 0.00019972300446321909,
|
| 3149 |
+
"loss": 0.8867,
|
| 3150 |
+
"step": 443
|
| 3151 |
+
},
|
| 3152 |
+
{
|
| 3153 |
+
"epoch": 0.07285406624961542,
|
| 3154 |
+
"grad_norm": 0.5159667134284973,
|
| 3155 |
+
"learning_rate": 0.00019972172415511457,
|
| 3156 |
+
"loss": 0.7261,
|
| 3157 |
+
"step": 444
|
| 3158 |
+
},
|
| 3159 |
+
{
|
| 3160 |
+
"epoch": 0.07301815198441186,
|
| 3161 |
+
"grad_norm": 0.46027064323425293,
|
| 3162 |
+
"learning_rate": 0.00019972044089907882,
|
| 3163 |
+
"loss": 1.0376,
|
| 3164 |
+
"step": 445
|
| 3165 |
+
},
|
| 3166 |
+
{
|
| 3167 |
+
"epoch": 0.07318223771920829,
|
| 3168 |
+
"grad_norm": 0.4172539710998535,
|
| 3169 |
+
"learning_rate": 0.00019971915469514973,
|
| 3170 |
+
"loss": 0.7398,
|
| 3171 |
+
"step": 446
|
| 3172 |
+
},
|
| 3173 |
+
{
|
| 3174 |
+
"epoch": 0.07334632345400471,
|
| 3175 |
+
"grad_norm": 0.3354978561401367,
|
| 3176 |
+
"learning_rate": 0.0001997178655433654,
|
| 3177 |
+
"loss": 0.9241,
|
| 3178 |
+
"step": 447
|
| 3179 |
+
},
|
| 3180 |
+
{
|
| 3181 |
+
"epoch": 0.07351040918880115,
|
| 3182 |
+
"grad_norm": 0.48640313744544983,
|
| 3183 |
+
"learning_rate": 0.0001997165734437639,
|
| 3184 |
+
"loss": 0.9889,
|
| 3185 |
+
"step": 448
|
| 3186 |
+
},
|
| 3187 |
+
{
|
| 3188 |
+
"epoch": 0.07367449492359758,
|
| 3189 |
+
"grad_norm": 0.36919528245925903,
|
| 3190 |
+
"learning_rate": 0.0001997152783963834,
|
| 3191 |
+
"loss": 0.9512,
|
| 3192 |
+
"step": 449
|
| 3193 |
+
},
|
| 3194 |
+
{
|
| 3195 |
+
"epoch": 0.073838580658394,
|
| 3196 |
+
"grad_norm": 0.410256028175354,
|
| 3197 |
+
"learning_rate": 0.0001997139804012622,
|
| 3198 |
+
"loss": 0.9793,
|
| 3199 |
+
"step": 450
|
| 3200 |
+
},
|
| 3201 |
+
{
|
| 3202 |
+
"epoch": 0.07400266639319045,
|
| 3203 |
+
"grad_norm": 0.25387945771217346,
|
| 3204 |
+
"learning_rate": 0.00019971267945843865,
|
| 3205 |
+
"loss": 0.4638,
|
| 3206 |
+
"step": 451
|
| 3207 |
+
},
|
| 3208 |
+
{
|
| 3209 |
+
"epoch": 0.07416675212798687,
|
| 3210 |
+
"grad_norm": 0.2956909239292145,
|
| 3211 |
+
"learning_rate": 0.0001997113755679513,
|
| 3212 |
+
"loss": 0.6796,
|
| 3213 |
+
"step": 452
|
| 3214 |
+
},
|
| 3215 |
+
{
|
| 3216 |
+
"epoch": 0.0743308378627833,
|
| 3217 |
+
"grad_norm": 0.26147258281707764,
|
| 3218 |
+
"learning_rate": 0.00019971006872983858,
|
| 3219 |
+
"loss": 0.6444,
|
| 3220 |
+
"step": 453
|
| 3221 |
+
},
|
| 3222 |
+
{
|
| 3223 |
+
"epoch": 0.07449492359757974,
|
| 3224 |
+
"grad_norm": 0.29692378640174866,
|
| 3225 |
+
"learning_rate": 0.00019970875894413916,
|
| 3226 |
+
"loss": 0.7089,
|
| 3227 |
+
"step": 454
|
| 3228 |
+
},
|
| 3229 |
+
{
|
| 3230 |
+
"epoch": 0.07465900933237617,
|
| 3231 |
+
"grad_norm": 0.3056322932243347,
|
| 3232 |
+
"learning_rate": 0.00019970744621089178,
|
| 3233 |
+
"loss": 0.6574,
|
| 3234 |
+
"step": 455
|
| 3235 |
+
},
|
| 3236 |
+
{
|
| 3237 |
+
"epoch": 0.0748230950671726,
|
| 3238 |
+
"grad_norm": 0.290412962436676,
|
| 3239 |
+
"learning_rate": 0.00019970613053013527,
|
| 3240 |
+
"loss": 0.7592,
|
| 3241 |
+
"step": 456
|
| 3242 |
+
},
|
| 3243 |
+
{
|
| 3244 |
+
"epoch": 0.07498718080196903,
|
| 3245 |
+
"grad_norm": 0.2817339301109314,
|
| 3246 |
+
"learning_rate": 0.00019970481190190843,
|
| 3247 |
+
"loss": 0.8211,
|
| 3248 |
+
"step": 457
|
| 3249 |
+
},
|
| 3250 |
+
{
|
| 3251 |
+
"epoch": 0.07515126653676546,
|
| 3252 |
+
"grad_norm": 0.2601938545703888,
|
| 3253 |
+
"learning_rate": 0.00019970349032625035,
|
| 3254 |
+
"loss": 0.7817,
|
| 3255 |
+
"step": 458
|
| 3256 |
+
},
|
| 3257 |
+
{
|
| 3258 |
+
"epoch": 0.07531535227156189,
|
| 3259 |
+
"grad_norm": 0.30698585510253906,
|
| 3260 |
+
"learning_rate": 0.00019970216580320003,
|
| 3261 |
+
"loss": 0.6768,
|
| 3262 |
+
"step": 459
|
| 3263 |
+
},
|
| 3264 |
+
{
|
| 3265 |
+
"epoch": 0.07547943800635833,
|
| 3266 |
+
"grad_norm": 0.27613088488578796,
|
| 3267 |
+
"learning_rate": 0.00019970083833279666,
|
| 3268 |
+
"loss": 0.7026,
|
| 3269 |
+
"step": 460
|
| 3270 |
+
},
|
| 3271 |
+
{
|
| 3272 |
+
"epoch": 0.07564352374115475,
|
| 3273 |
+
"grad_norm": 0.2695368826389313,
|
| 3274 |
+
"learning_rate": 0.00019969950791507942,
|
| 3275 |
+
"loss": 0.7914,
|
| 3276 |
+
"step": 461
|
| 3277 |
+
},
|
| 3278 |
+
{
|
| 3279 |
+
"epoch": 0.07580760947595118,
|
| 3280 |
+
"grad_norm": 0.2841225266456604,
|
| 3281 |
+
"learning_rate": 0.0001996981745500877,
|
| 3282 |
+
"loss": 0.7972,
|
| 3283 |
+
"step": 462
|
| 3284 |
+
},
|
| 3285 |
+
{
|
| 3286 |
+
"epoch": 0.07597169521074762,
|
| 3287 |
+
"grad_norm": 0.25151097774505615,
|
| 3288 |
+
"learning_rate": 0.00019969683823786093,
|
| 3289 |
+
"loss": 0.6287,
|
| 3290 |
+
"step": 463
|
| 3291 |
+
},
|
| 3292 |
+
{
|
| 3293 |
+
"epoch": 0.07613578094554405,
|
| 3294 |
+
"grad_norm": 0.2680940330028534,
|
| 3295 |
+
"learning_rate": 0.00019969549897843857,
|
| 3296 |
+
"loss": 0.6865,
|
| 3297 |
+
"step": 464
|
| 3298 |
+
},
|
| 3299 |
+
{
|
| 3300 |
+
"epoch": 0.07629986668034047,
|
| 3301 |
+
"grad_norm": 0.2942931354045868,
|
| 3302 |
+
"learning_rate": 0.0001996941567718602,
|
| 3303 |
+
"loss": 0.7175,
|
| 3304 |
+
"step": 465
|
| 3305 |
+
},
|
| 3306 |
+
{
|
| 3307 |
+
"epoch": 0.07646395241513691,
|
| 3308 |
+
"grad_norm": 0.33051207661628723,
|
| 3309 |
+
"learning_rate": 0.00019969281161816556,
|
| 3310 |
+
"loss": 0.8518,
|
| 3311 |
+
"step": 466
|
| 3312 |
+
},
|
| 3313 |
+
{
|
| 3314 |
+
"epoch": 0.07662803814993334,
|
| 3315 |
+
"grad_norm": 0.28793030977249146,
|
| 3316 |
+
"learning_rate": 0.00019969146351739436,
|
| 3317 |
+
"loss": 0.806,
|
| 3318 |
+
"step": 467
|
| 3319 |
+
},
|
| 3320 |
+
{
|
| 3321 |
+
"epoch": 0.07679212388472977,
|
| 3322 |
+
"grad_norm": 0.27401405572891235,
|
| 3323 |
+
"learning_rate": 0.00019969011246958647,
|
| 3324 |
+
"loss": 0.7563,
|
| 3325 |
+
"step": 468
|
| 3326 |
+
},
|
| 3327 |
+
{
|
| 3328 |
+
"epoch": 0.0769562096195262,
|
| 3329 |
+
"grad_norm": 0.28435268998146057,
|
| 3330 |
+
"learning_rate": 0.00019968875847478184,
|
| 3331 |
+
"loss": 0.8319,
|
| 3332 |
+
"step": 469
|
| 3333 |
+
},
|
| 3334 |
+
{
|
| 3335 |
+
"epoch": 0.07712029535432263,
|
| 3336 |
+
"grad_norm": 0.37528687715530396,
|
| 3337 |
+
"learning_rate": 0.00019968740153302047,
|
| 3338 |
+
"loss": 0.7616,
|
| 3339 |
+
"step": 470
|
| 3340 |
+
},
|
| 3341 |
+
{
|
| 3342 |
+
"epoch": 0.07728438108911906,
|
| 3343 |
+
"grad_norm": 0.3924008309841156,
|
| 3344 |
+
"learning_rate": 0.00019968604164434246,
|
| 3345 |
+
"loss": 0.8162,
|
| 3346 |
+
"step": 471
|
| 3347 |
+
},
|
| 3348 |
+
{
|
| 3349 |
+
"epoch": 0.0774484668239155,
|
| 3350 |
+
"grad_norm": 0.3189226984977722,
|
| 3351 |
+
"learning_rate": 0.0001996846788087881,
|
| 3352 |
+
"loss": 0.718,
|
| 3353 |
+
"step": 472
|
| 3354 |
+
},
|
| 3355 |
+
{
|
| 3356 |
+
"epoch": 0.07761255255871193,
|
| 3357 |
+
"grad_norm": 0.2787996828556061,
|
| 3358 |
+
"learning_rate": 0.0001996833130263976,
|
| 3359 |
+
"loss": 0.7551,
|
| 3360 |
+
"step": 473
|
| 3361 |
+
},
|
| 3362 |
+
{
|
| 3363 |
+
"epoch": 0.07777663829350835,
|
| 3364 |
+
"grad_norm": 0.353397935628891,
|
| 3365 |
+
"learning_rate": 0.0001996819442972113,
|
| 3366 |
+
"loss": 0.6889,
|
| 3367 |
+
"step": 474
|
| 3368 |
+
},
|
| 3369 |
+
{
|
| 3370 |
+
"epoch": 0.07794072402830479,
|
| 3371 |
+
"grad_norm": 0.28389084339141846,
|
| 3372 |
+
"learning_rate": 0.00019968057262126972,
|
| 3373 |
+
"loss": 0.8333,
|
| 3374 |
+
"step": 475
|
| 3375 |
+
},
|
| 3376 |
+
{
|
| 3377 |
+
"epoch": 0.07810480976310122,
|
| 3378 |
+
"grad_norm": 0.2652452886104584,
|
| 3379 |
+
"learning_rate": 0.00019967919799861346,
|
| 3380 |
+
"loss": 0.7011,
|
| 3381 |
+
"step": 476
|
| 3382 |
+
},
|
| 3383 |
+
{
|
| 3384 |
+
"epoch": 0.07826889549789764,
|
| 3385 |
+
"grad_norm": 0.2802514135837555,
|
| 3386 |
+
"learning_rate": 0.00019967782042928307,
|
| 3387 |
+
"loss": 0.7594,
|
| 3388 |
+
"step": 477
|
| 3389 |
+
},
|
| 3390 |
+
{
|
| 3391 |
+
"epoch": 0.07843298123269409,
|
| 3392 |
+
"grad_norm": 0.4574797451496124,
|
| 3393 |
+
"learning_rate": 0.0001996764399133193,
|
| 3394 |
+
"loss": 0.7706,
|
| 3395 |
+
"step": 478
|
| 3396 |
+
},
|
| 3397 |
+
{
|
| 3398 |
+
"epoch": 0.07859706696749051,
|
| 3399 |
+
"grad_norm": 0.3003863990306854,
|
| 3400 |
+
"learning_rate": 0.00019967505645076294,
|
| 3401 |
+
"loss": 0.7685,
|
| 3402 |
+
"step": 479
|
| 3403 |
+
},
|
| 3404 |
+
{
|
| 3405 |
+
"epoch": 0.07876115270228695,
|
| 3406 |
+
"grad_norm": 0.3658941388130188,
|
| 3407 |
+
"learning_rate": 0.0001996736700416549,
|
| 3408 |
+
"loss": 0.7536,
|
| 3409 |
+
"step": 480
|
| 3410 |
+
},
|
| 3411 |
+
{
|
| 3412 |
+
"epoch": 0.07892523843708338,
|
| 3413 |
+
"grad_norm": 0.3780055344104767,
|
| 3414 |
+
"learning_rate": 0.00019967228068603623,
|
| 3415 |
+
"loss": 0.7793,
|
| 3416 |
+
"step": 481
|
| 3417 |
+
},
|
| 3418 |
+
{
|
| 3419 |
+
"epoch": 0.0790893241718798,
|
| 3420 |
+
"grad_norm": 0.3696160316467285,
|
| 3421 |
+
"learning_rate": 0.00019967088838394789,
|
| 3422 |
+
"loss": 0.8083,
|
| 3423 |
+
"step": 482
|
| 3424 |
+
},
|
| 3425 |
+
{
|
| 3426 |
+
"epoch": 0.07925340990667624,
|
| 3427 |
+
"grad_norm": 0.3387402296066284,
|
| 3428 |
+
"learning_rate": 0.00019966949313543112,
|
| 3429 |
+
"loss": 0.8436,
|
| 3430 |
+
"step": 483
|
| 3431 |
+
},
|
| 3432 |
+
{
|
| 3433 |
+
"epoch": 0.07941749564147267,
|
| 3434 |
+
"grad_norm": 0.3094363212585449,
|
| 3435 |
+
"learning_rate": 0.00019966809494052713,
|
| 3436 |
+
"loss": 0.7768,
|
| 3437 |
+
"step": 484
|
| 3438 |
+
},
|
| 3439 |
+
{
|
| 3440 |
+
"epoch": 0.0795815813762691,
|
| 3441 |
+
"grad_norm": 0.3164433240890503,
|
| 3442 |
+
"learning_rate": 0.00019966669379927726,
|
| 3443 |
+
"loss": 0.8517,
|
| 3444 |
+
"step": 485
|
| 3445 |
+
},
|
| 3446 |
+
{
|
| 3447 |
+
"epoch": 0.07974566711106554,
|
| 3448 |
+
"grad_norm": 0.3255369961261749,
|
| 3449 |
+
"learning_rate": 0.00019966528971172295,
|
| 3450 |
+
"loss": 0.8573,
|
| 3451 |
+
"step": 486
|
| 3452 |
+
},
|
| 3453 |
+
{
|
| 3454 |
+
"epoch": 0.07990975284586196,
|
| 3455 |
+
"grad_norm": 0.31875723600387573,
|
| 3456 |
+
"learning_rate": 0.00019966388267790566,
|
| 3457 |
+
"loss": 0.8253,
|
| 3458 |
+
"step": 487
|
| 3459 |
+
},
|
| 3460 |
+
{
|
| 3461 |
+
"epoch": 0.08007383858065839,
|
| 3462 |
+
"grad_norm": 0.32112181186676025,
|
| 3463 |
+
"learning_rate": 0.00019966247269786701,
|
| 3464 |
+
"loss": 0.9794,
|
| 3465 |
+
"step": 488
|
| 3466 |
+
},
|
| 3467 |
+
{
|
| 3468 |
+
"epoch": 0.08023792431545483,
|
| 3469 |
+
"grad_norm": 0.34695595502853394,
|
| 3470 |
+
"learning_rate": 0.00019966105977164872,
|
| 3471 |
+
"loss": 0.7167,
|
| 3472 |
+
"step": 489
|
| 3473 |
+
},
|
| 3474 |
+
{
|
| 3475 |
+
"epoch": 0.08040201005025126,
|
| 3476 |
+
"grad_norm": 0.33684611320495605,
|
| 3477 |
+
"learning_rate": 0.0001996596438992925,
|
| 3478 |
+
"loss": 0.8618,
|
| 3479 |
+
"step": 490
|
| 3480 |
+
},
|
| 3481 |
+
{
|
| 3482 |
+
"epoch": 0.08056609578504768,
|
| 3483 |
+
"grad_norm": 0.3515377342700958,
|
| 3484 |
+
"learning_rate": 0.00019965822508084022,
|
| 3485 |
+
"loss": 0.7417,
|
| 3486 |
+
"step": 491
|
| 3487 |
+
},
|
| 3488 |
+
{
|
| 3489 |
+
"epoch": 0.08073018151984412,
|
| 3490 |
+
"grad_norm": 0.45770692825317383,
|
| 3491 |
+
"learning_rate": 0.00019965680331633382,
|
| 3492 |
+
"loss": 0.7696,
|
| 3493 |
+
"step": 492
|
| 3494 |
+
},
|
| 3495 |
+
{
|
| 3496 |
+
"epoch": 0.08089426725464055,
|
| 3497 |
+
"grad_norm": 0.35616007447242737,
|
| 3498 |
+
"learning_rate": 0.00019965537860581537,
|
| 3499 |
+
"loss": 0.9678,
|
| 3500 |
+
"step": 493
|
| 3501 |
+
},
|
| 3502 |
+
{
|
| 3503 |
+
"epoch": 0.08105835298943698,
|
| 3504 |
+
"grad_norm": 0.5278313159942627,
|
| 3505 |
+
"learning_rate": 0.00019965395094932693,
|
| 3506 |
+
"loss": 0.8417,
|
| 3507 |
+
"step": 494
|
| 3508 |
+
},
|
| 3509 |
+
{
|
| 3510 |
+
"epoch": 0.08122243872423342,
|
| 3511 |
+
"grad_norm": 0.40865620970726013,
|
| 3512 |
+
"learning_rate": 0.00019965252034691075,
|
| 3513 |
+
"loss": 0.7845,
|
| 3514 |
+
"step": 495
|
| 3515 |
+
},
|
| 3516 |
+
{
|
| 3517 |
+
"epoch": 0.08138652445902984,
|
| 3518 |
+
"grad_norm": 0.37308233976364136,
|
| 3519 |
+
"learning_rate": 0.0001996510867986091,
|
| 3520 |
+
"loss": 0.8607,
|
| 3521 |
+
"step": 496
|
| 3522 |
+
},
|
| 3523 |
+
{
|
| 3524 |
+
"epoch": 0.08155061019382627,
|
| 3525 |
+
"grad_norm": 0.5070668458938599,
|
| 3526 |
+
"learning_rate": 0.00019964965030446434,
|
| 3527 |
+
"loss": 0.8668,
|
| 3528 |
+
"step": 497
|
| 3529 |
+
},
|
| 3530 |
+
{
|
| 3531 |
+
"epoch": 0.08171469592862271,
|
| 3532 |
+
"grad_norm": 0.40923750400543213,
|
| 3533 |
+
"learning_rate": 0.00019964821086451896,
|
| 3534 |
+
"loss": 0.9865,
|
| 3535 |
+
"step": 498
|
| 3536 |
+
},
|
| 3537 |
+
{
|
| 3538 |
+
"epoch": 0.08187878166341914,
|
| 3539 |
+
"grad_norm": 0.42455339431762695,
|
| 3540 |
+
"learning_rate": 0.00019964676847881551,
|
| 3541 |
+
"loss": 1.0331,
|
| 3542 |
+
"step": 499
|
| 3543 |
+
},
|
| 3544 |
+
{
|
| 3545 |
+
"epoch": 0.08204286739821556,
|
| 3546 |
+
"grad_norm": 0.5710748434066772,
|
| 3547 |
+
"learning_rate": 0.00019964532314739662,
|
| 3548 |
+
"loss": 0.9817,
|
| 3549 |
+
"step": 500
|
| 3550 |
+
},
|
| 3551 |
+
{
|
| 3552 |
+
"epoch": 0.08204286739821556,
|
| 3553 |
+
"eval_loss": 0.7902427911758423,
|
| 3554 |
+
"eval_runtime": 70.7583,
|
| 3555 |
+
"eval_samples_per_second": 3.462,
|
| 3556 |
+
"eval_steps_per_second": 3.462,
|
| 3557 |
+
"step": 500
|
| 3558 |
+
},
|
| 3559 |
+
{
|
| 3560 |
+
"epoch": 0.082206953133012,
|
| 3561 |
+
"grad_norm": 0.2542221248149872,
|
| 3562 |
+
"learning_rate": 0.00019964387487030503,
|
| 3563 |
+
"loss": 0.544,
|
| 3564 |
+
"step": 501
|
| 3565 |
+
},
|
| 3566 |
+
{
|
| 3567 |
+
"epoch": 0.08237103886780843,
|
| 3568 |
+
"grad_norm": 0.2645345628261566,
|
| 3569 |
+
"learning_rate": 0.00019964242364758355,
|
| 3570 |
+
"loss": 0.5941,
|
| 3571 |
+
"step": 502
|
| 3572 |
+
},
|
| 3573 |
+
{
|
| 3574 |
+
"epoch": 0.08253512460260486,
|
| 3575 |
+
"grad_norm": 0.24813871085643768,
|
| 3576 |
+
"learning_rate": 0.00019964096947927508,
|
| 3577 |
+
"loss": 0.6608,
|
| 3578 |
+
"step": 503
|
| 3579 |
+
},
|
| 3580 |
+
{
|
| 3581 |
+
"epoch": 0.0826992103374013,
|
| 3582 |
+
"grad_norm": 0.24130459129810333,
|
| 3583 |
+
"learning_rate": 0.00019963951236542257,
|
| 3584 |
+
"loss": 0.637,
|
| 3585 |
+
"step": 504
|
| 3586 |
+
},
|
| 3587 |
+
{
|
| 3588 |
+
"epoch": 0.08286329607219772,
|
| 3589 |
+
"grad_norm": 0.3400072157382965,
|
| 3590 |
+
"learning_rate": 0.00019963805230606915,
|
| 3591 |
+
"loss": 0.638,
|
| 3592 |
+
"step": 505
|
| 3593 |
+
},
|
| 3594 |
+
{
|
| 3595 |
+
"epoch": 0.08302738180699415,
|
| 3596 |
+
"grad_norm": 0.2924500107765198,
|
| 3597 |
+
"learning_rate": 0.00019963658930125794,
|
| 3598 |
+
"loss": 0.7683,
|
| 3599 |
+
"step": 506
|
| 3600 |
+
},
|
| 3601 |
+
{
|
| 3602 |
+
"epoch": 0.08319146754179059,
|
| 3603 |
+
"grad_norm": 0.47059324383735657,
|
| 3604 |
+
"learning_rate": 0.00019963512335103222,
|
| 3605 |
+
"loss": 0.7232,
|
| 3606 |
+
"step": 507
|
| 3607 |
+
},
|
| 3608 |
+
{
|
| 3609 |
+
"epoch": 0.08335555327658702,
|
| 3610 |
+
"grad_norm": 0.29273051023483276,
|
| 3611 |
+
"learning_rate": 0.00019963365445543532,
|
| 3612 |
+
"loss": 0.7313,
|
| 3613 |
+
"step": 508
|
| 3614 |
+
},
|
| 3615 |
+
{
|
| 3616 |
+
"epoch": 0.08351963901138344,
|
| 3617 |
+
"grad_norm": 0.30427420139312744,
|
| 3618 |
+
"learning_rate": 0.00019963218261451066,
|
| 3619 |
+
"loss": 0.7675,
|
| 3620 |
+
"step": 509
|
| 3621 |
+
},
|
| 3622 |
+
{
|
| 3623 |
+
"epoch": 0.08368372474617988,
|
| 3624 |
+
"grad_norm": 0.27146315574645996,
|
| 3625 |
+
"learning_rate": 0.00019963070782830173,
|
| 3626 |
+
"loss": 0.6338,
|
| 3627 |
+
"step": 510
|
| 3628 |
+
},
|
| 3629 |
+
{
|
| 3630 |
+
"epoch": 0.08384781048097631,
|
| 3631 |
+
"grad_norm": 0.26816409826278687,
|
| 3632 |
+
"learning_rate": 0.00019962923009685216,
|
| 3633 |
+
"loss": 0.692,
|
| 3634 |
+
"step": 511
|
| 3635 |
+
},
|
| 3636 |
+
{
|
| 3637 |
+
"epoch": 0.08401189621577274,
|
| 3638 |
+
"grad_norm": 0.2671113610267639,
|
| 3639 |
+
"learning_rate": 0.0001996277494202056,
|
| 3640 |
+
"loss": 0.7029,
|
| 3641 |
+
"step": 512
|
| 3642 |
+
},
|
| 3643 |
+
{
|
| 3644 |
+
"epoch": 0.08417598195056918,
|
| 3645 |
+
"grad_norm": 0.26612916588783264,
|
| 3646 |
+
"learning_rate": 0.00019962626579840583,
|
| 3647 |
+
"loss": 0.8666,
|
| 3648 |
+
"step": 513
|
| 3649 |
+
},
|
| 3650 |
+
{
|
| 3651 |
+
"epoch": 0.0843400676853656,
|
| 3652 |
+
"grad_norm": 0.2911565601825714,
|
| 3653 |
+
"learning_rate": 0.00019962477923149674,
|
| 3654 |
+
"loss": 0.6847,
|
| 3655 |
+
"step": 514
|
| 3656 |
+
},
|
| 3657 |
+
{
|
| 3658 |
+
"epoch": 0.08450415342016203,
|
| 3659 |
+
"grad_norm": 0.2518673241138458,
|
| 3660 |
+
"learning_rate": 0.00019962328971952225,
|
| 3661 |
+
"loss": 0.6946,
|
| 3662 |
+
"step": 515
|
| 3663 |
+
},
|
| 3664 |
+
{
|
| 3665 |
+
"epoch": 0.08466823915495847,
|
| 3666 |
+
"grad_norm": 0.32353517413139343,
|
| 3667 |
+
"learning_rate": 0.0001996217972625264,
|
| 3668 |
+
"loss": 0.7648,
|
| 3669 |
+
"step": 516
|
| 3670 |
+
},
|
| 3671 |
+
{
|
| 3672 |
+
"epoch": 0.0848323248897549,
|
| 3673 |
+
"grad_norm": 0.33887815475463867,
|
| 3674 |
+
"learning_rate": 0.00019962030186055328,
|
| 3675 |
+
"loss": 0.74,
|
| 3676 |
+
"step": 517
|
| 3677 |
+
},
|
| 3678 |
+
{
|
| 3679 |
+
"epoch": 0.08499641062455132,
|
| 3680 |
+
"grad_norm": 0.30510279536247253,
|
| 3681 |
+
"learning_rate": 0.00019961880351364712,
|
| 3682 |
+
"loss": 0.8442,
|
| 3683 |
+
"step": 518
|
| 3684 |
+
},
|
| 3685 |
+
{
|
| 3686 |
+
"epoch": 0.08516049635934776,
|
| 3687 |
+
"grad_norm": 0.4953973889350891,
|
| 3688 |
+
"learning_rate": 0.00019961730222185225,
|
| 3689 |
+
"loss": 0.7548,
|
| 3690 |
+
"step": 519
|
| 3691 |
+
},
|
| 3692 |
+
{
|
| 3693 |
+
"epoch": 0.08532458209414419,
|
| 3694 |
+
"grad_norm": 0.3338676393032074,
|
| 3695 |
+
"learning_rate": 0.00019961579798521297,
|
| 3696 |
+
"loss": 0.7259,
|
| 3697 |
+
"step": 520
|
| 3698 |
+
},
|
| 3699 |
+
{
|
| 3700 |
+
"epoch": 0.08548866782894061,
|
| 3701 |
+
"grad_norm": 0.29042428731918335,
|
| 3702 |
+
"learning_rate": 0.0001996142908037738,
|
| 3703 |
+
"loss": 0.8281,
|
| 3704 |
+
"step": 521
|
| 3705 |
+
},
|
| 3706 |
+
{
|
| 3707 |
+
"epoch": 0.08565275356373706,
|
| 3708 |
+
"grad_norm": 0.30394816398620605,
|
| 3709 |
+
"learning_rate": 0.0001996127806775793,
|
| 3710 |
+
"loss": 0.6939,
|
| 3711 |
+
"step": 522
|
| 3712 |
+
},
|
| 3713 |
+
{
|
| 3714 |
+
"epoch": 0.08581683929853348,
|
| 3715 |
+
"grad_norm": 0.2836878001689911,
|
| 3716 |
+
"learning_rate": 0.0001996112676066741,
|
| 3717 |
+
"loss": 0.6618,
|
| 3718 |
+
"step": 523
|
| 3719 |
+
},
|
| 3720 |
+
{
|
| 3721 |
+
"epoch": 0.08598092503332991,
|
| 3722 |
+
"grad_norm": 0.33087509870529175,
|
| 3723 |
+
"learning_rate": 0.00019960975159110295,
|
| 3724 |
+
"loss": 0.7448,
|
| 3725 |
+
"step": 524
|
| 3726 |
+
},
|
| 3727 |
+
{
|
| 3728 |
+
"epoch": 0.08614501076812635,
|
| 3729 |
+
"grad_norm": 0.26727747917175293,
|
| 3730 |
+
"learning_rate": 0.0001996082326309106,
|
| 3731 |
+
"loss": 0.693,
|
| 3732 |
+
"step": 525
|
| 3733 |
+
},
|
| 3734 |
+
{
|
| 3735 |
+
"epoch": 0.08630909650292277,
|
| 3736 |
+
"grad_norm": 0.2853117287158966,
|
| 3737 |
+
"learning_rate": 0.000199606710726142,
|
| 3738 |
+
"loss": 0.7681,
|
| 3739 |
+
"step": 526
|
| 3740 |
+
},
|
| 3741 |
+
{
|
| 3742 |
+
"epoch": 0.08647318223771921,
|
| 3743 |
+
"grad_norm": 0.5314649343490601,
|
| 3744 |
+
"learning_rate": 0.00019960518587684213,
|
| 3745 |
+
"loss": 0.6646,
|
| 3746 |
+
"step": 527
|
| 3747 |
+
},
|
| 3748 |
+
{
|
| 3749 |
+
"epoch": 0.08663726797251564,
|
| 3750 |
+
"grad_norm": 0.33825939893722534,
|
| 3751 |
+
"learning_rate": 0.00019960365808305609,
|
| 3752 |
+
"loss": 0.7447,
|
| 3753 |
+
"step": 528
|
| 3754 |
+
},
|
| 3755 |
+
{
|
| 3756 |
+
"epoch": 0.08680135370731207,
|
| 3757 |
+
"grad_norm": 0.3486720323562622,
|
| 3758 |
+
"learning_rate": 0.00019960212734482902,
|
| 3759 |
+
"loss": 0.7324,
|
| 3760 |
+
"step": 529
|
| 3761 |
+
},
|
| 3762 |
+
{
|
| 3763 |
+
"epoch": 0.08696543944210851,
|
| 3764 |
+
"grad_norm": 0.44422647356987,
|
| 3765 |
+
"learning_rate": 0.00019960059366220617,
|
| 3766 |
+
"loss": 0.798,
|
| 3767 |
+
"step": 530
|
| 3768 |
+
},
|
| 3769 |
+
{
|
| 3770 |
+
"epoch": 0.08712952517690493,
|
| 3771 |
+
"grad_norm": 0.2957271635532379,
|
| 3772 |
+
"learning_rate": 0.00019959905703523288,
|
| 3773 |
+
"loss": 0.7875,
|
| 3774 |
+
"step": 531
|
| 3775 |
+
},
|
| 3776 |
+
{
|
| 3777 |
+
"epoch": 0.08729361091170136,
|
| 3778 |
+
"grad_norm": 0.32711222767829895,
|
| 3779 |
+
"learning_rate": 0.00019959751746395461,
|
| 3780 |
+
"loss": 0.8314,
|
| 3781 |
+
"step": 532
|
| 3782 |
+
},
|
| 3783 |
+
{
|
| 3784 |
+
"epoch": 0.0874576966464978,
|
| 3785 |
+
"grad_norm": 0.3585834205150604,
|
| 3786 |
+
"learning_rate": 0.00019959597494841681,
|
| 3787 |
+
"loss": 0.81,
|
| 3788 |
+
"step": 533
|
| 3789 |
+
},
|
| 3790 |
+
{
|
| 3791 |
+
"epoch": 0.08762178238129423,
|
| 3792 |
+
"grad_norm": 0.4582952857017517,
|
| 3793 |
+
"learning_rate": 0.00019959442948866513,
|
| 3794 |
+
"loss": 0.8199,
|
| 3795 |
+
"step": 534
|
| 3796 |
+
},
|
| 3797 |
+
{
|
| 3798 |
+
"epoch": 0.08778586811609065,
|
| 3799 |
+
"grad_norm": 0.3768620491027832,
|
| 3800 |
+
"learning_rate": 0.00019959288108474527,
|
| 3801 |
+
"loss": 0.7852,
|
| 3802 |
+
"step": 535
|
| 3803 |
+
},
|
| 3804 |
+
{
|
| 3805 |
+
"epoch": 0.0879499538508871,
|
| 3806 |
+
"grad_norm": 0.37209224700927734,
|
| 3807 |
+
"learning_rate": 0.00019959132973670292,
|
| 3808 |
+
"loss": 0.8902,
|
| 3809 |
+
"step": 536
|
| 3810 |
+
},
|
| 3811 |
+
{
|
| 3812 |
+
"epoch": 0.08811403958568352,
|
| 3813 |
+
"grad_norm": 0.3185194730758667,
|
| 3814 |
+
"learning_rate": 0.00019958977544458402,
|
| 3815 |
+
"loss": 0.7754,
|
| 3816 |
+
"step": 537
|
| 3817 |
+
},
|
| 3818 |
+
{
|
| 3819 |
+
"epoch": 0.08827812532047995,
|
| 3820 |
+
"grad_norm": 0.32522639632225037,
|
| 3821 |
+
"learning_rate": 0.00019958821820843448,
|
| 3822 |
+
"loss": 0.8702,
|
| 3823 |
+
"step": 538
|
| 3824 |
+
},
|
| 3825 |
+
{
|
| 3826 |
+
"epoch": 0.08844221105527639,
|
| 3827 |
+
"grad_norm": 0.33792081475257874,
|
| 3828 |
+
"learning_rate": 0.00019958665802830036,
|
| 3829 |
+
"loss": 0.7749,
|
| 3830 |
+
"step": 539
|
| 3831 |
+
},
|
| 3832 |
+
{
|
| 3833 |
+
"epoch": 0.08860629679007281,
|
| 3834 |
+
"grad_norm": 0.3337690830230713,
|
| 3835 |
+
"learning_rate": 0.0001995850949042277,
|
| 3836 |
+
"loss": 0.8281,
|
| 3837 |
+
"step": 540
|
| 3838 |
+
},
|
| 3839 |
+
{
|
| 3840 |
+
"epoch": 0.08877038252486924,
|
| 3841 |
+
"grad_norm": 0.3441987931728363,
|
| 3842 |
+
"learning_rate": 0.00019958352883626284,
|
| 3843 |
+
"loss": 0.7142,
|
| 3844 |
+
"step": 541
|
| 3845 |
+
},
|
| 3846 |
+
{
|
| 3847 |
+
"epoch": 0.08893446825966568,
|
| 3848 |
+
"grad_norm": 0.42077013850212097,
|
| 3849 |
+
"learning_rate": 0.00019958195982445199,
|
| 3850 |
+
"loss": 0.7891,
|
| 3851 |
+
"step": 542
|
| 3852 |
+
},
|
| 3853 |
+
{
|
| 3854 |
+
"epoch": 0.0890985539944621,
|
| 3855 |
+
"grad_norm": 0.4405931830406189,
|
| 3856 |
+
"learning_rate": 0.00019958038786884155,
|
| 3857 |
+
"loss": 0.9158,
|
| 3858 |
+
"step": 543
|
| 3859 |
+
},
|
| 3860 |
+
{
|
| 3861 |
+
"epoch": 0.08926263972925853,
|
| 3862 |
+
"grad_norm": 0.3358551859855652,
|
| 3863 |
+
"learning_rate": 0.00019957881296947798,
|
| 3864 |
+
"loss": 0.6941,
|
| 3865 |
+
"step": 544
|
| 3866 |
+
},
|
| 3867 |
+
{
|
| 3868 |
+
"epoch": 0.08942672546405497,
|
| 3869 |
+
"grad_norm": 0.39642512798309326,
|
| 3870 |
+
"learning_rate": 0.00019957723512640784,
|
| 3871 |
+
"loss": 0.7531,
|
| 3872 |
+
"step": 545
|
| 3873 |
+
},
|
| 3874 |
+
{
|
| 3875 |
+
"epoch": 0.0895908111988514,
|
| 3876 |
+
"grad_norm": 0.428423672914505,
|
| 3877 |
+
"learning_rate": 0.0001995756543396778,
|
| 3878 |
+
"loss": 0.8928,
|
| 3879 |
+
"step": 546
|
| 3880 |
+
},
|
| 3881 |
+
{
|
| 3882 |
+
"epoch": 0.08975489693364783,
|
| 3883 |
+
"grad_norm": 0.49737995862960815,
|
| 3884 |
+
"learning_rate": 0.00019957407060933457,
|
| 3885 |
+
"loss": 0.8893,
|
| 3886 |
+
"step": 547
|
| 3887 |
+
},
|
| 3888 |
+
{
|
| 3889 |
+
"epoch": 0.08991898266844427,
|
| 3890 |
+
"grad_norm": 0.45987364649772644,
|
| 3891 |
+
"learning_rate": 0.00019957248393542498,
|
| 3892 |
+
"loss": 0.6813,
|
| 3893 |
+
"step": 548
|
| 3894 |
+
},
|
| 3895 |
+
{
|
| 3896 |
+
"epoch": 0.0900830684032407,
|
| 3897 |
+
"grad_norm": 0.5005274415016174,
|
| 3898 |
+
"learning_rate": 0.0001995708943179959,
|
| 3899 |
+
"loss": 0.7574,
|
| 3900 |
+
"step": 549
|
| 3901 |
+
},
|
| 3902 |
+
{
|
| 3903 |
+
"epoch": 0.09024715413803712,
|
| 3904 |
+
"grad_norm": 0.6577406525611877,
|
| 3905 |
+
"learning_rate": 0.00019956930175709436,
|
| 3906 |
+
"loss": 0.9468,
|
| 3907 |
+
"step": 550
|
| 3908 |
+
},
|
| 3909 |
+
{
|
| 3910 |
+
"epoch": 0.09041123987283356,
|
| 3911 |
+
"grad_norm": 0.30720874667167664,
|
| 3912 |
+
"learning_rate": 0.0001995677062527674,
|
| 3913 |
+
"loss": 0.6136,
|
| 3914 |
+
"step": 551
|
| 3915 |
+
},
|
| 3916 |
+
{
|
| 3917 |
+
"epoch": 0.09057532560762999,
|
| 3918 |
+
"grad_norm": 0.24331681430339813,
|
| 3919 |
+
"learning_rate": 0.00019956610780506222,
|
| 3920 |
+
"loss": 0.6703,
|
| 3921 |
+
"step": 552
|
| 3922 |
+
},
|
| 3923 |
+
{
|
| 3924 |
+
"epoch": 0.09073941134242641,
|
| 3925 |
+
"grad_norm": 0.24991828203201294,
|
| 3926 |
+
"learning_rate": 0.00019956450641402609,
|
| 3927 |
+
"loss": 0.6179,
|
| 3928 |
+
"step": 553
|
| 3929 |
+
},
|
| 3930 |
+
{
|
| 3931 |
+
"epoch": 0.09090349707722285,
|
| 3932 |
+
"grad_norm": 0.26630428433418274,
|
| 3933 |
+
"learning_rate": 0.0001995629020797063,
|
| 3934 |
+
"loss": 0.9014,
|
| 3935 |
+
"step": 554
|
| 3936 |
+
},
|
| 3937 |
+
{
|
| 3938 |
+
"epoch": 0.09106758281201928,
|
| 3939 |
+
"grad_norm": 0.2855633795261383,
|
| 3940 |
+
"learning_rate": 0.00019956129480215026,
|
| 3941 |
+
"loss": 0.6828,
|
| 3942 |
+
"step": 555
|
| 3943 |
+
},
|
| 3944 |
+
{
|
| 3945 |
+
"epoch": 0.0912316685468157,
|
| 3946 |
+
"grad_norm": 0.292644739151001,
|
| 3947 |
+
"learning_rate": 0.00019955968458140557,
|
| 3948 |
+
"loss": 0.6975,
|
| 3949 |
+
"step": 556
|
| 3950 |
+
},
|
| 3951 |
+
{
|
| 3952 |
+
"epoch": 0.09139575428161215,
|
| 3953 |
+
"grad_norm": 0.3429616689682007,
|
| 3954 |
+
"learning_rate": 0.00019955807141751975,
|
| 3955 |
+
"loss": 0.7558,
|
| 3956 |
+
"step": 557
|
| 3957 |
+
},
|
| 3958 |
+
{
|
| 3959 |
+
"epoch": 0.09155984001640857,
|
| 3960 |
+
"grad_norm": 0.3144456744194031,
|
| 3961 |
+
"learning_rate": 0.00019955645531054056,
|
| 3962 |
+
"loss": 0.6571,
|
| 3963 |
+
"step": 558
|
| 3964 |
+
},
|
| 3965 |
+
{
|
| 3966 |
+
"epoch": 0.091723925751205,
|
| 3967 |
+
"grad_norm": 0.2671104669570923,
|
| 3968 |
+
"learning_rate": 0.0001995548362605157,
|
| 3969 |
+
"loss": 0.7479,
|
| 3970 |
+
"step": 559
|
| 3971 |
+
},
|
| 3972 |
+
{
|
| 3973 |
+
"epoch": 0.09188801148600144,
|
| 3974 |
+
"grad_norm": 0.31418734788894653,
|
| 3975 |
+
"learning_rate": 0.0001995532142674931,
|
| 3976 |
+
"loss": 0.6673,
|
| 3977 |
+
"step": 560
|
| 3978 |
+
},
|
| 3979 |
+
{
|
| 3980 |
+
"epoch": 0.09205209722079787,
|
| 3981 |
+
"grad_norm": 0.3422031104564667,
|
| 3982 |
+
"learning_rate": 0.00019955158933152064,
|
| 3983 |
+
"loss": 0.7965,
|
| 3984 |
+
"step": 561
|
| 3985 |
+
},
|
| 3986 |
+
{
|
| 3987 |
+
"epoch": 0.09221618295559429,
|
| 3988 |
+
"grad_norm": 0.2794967591762543,
|
| 3989 |
+
"learning_rate": 0.00019954996145264643,
|
| 3990 |
+
"loss": 0.716,
|
| 3991 |
+
"step": 562
|
| 3992 |
+
},
|
| 3993 |
+
{
|
| 3994 |
+
"epoch": 0.09238026869039073,
|
| 3995 |
+
"grad_norm": 0.39105239510536194,
|
| 3996 |
+
"learning_rate": 0.00019954833063091853,
|
| 3997 |
+
"loss": 0.6283,
|
| 3998 |
+
"step": 563
|
| 3999 |
+
},
|
| 4000 |
+
{
|
| 4001 |
+
"epoch": 0.09254435442518716,
|
| 4002 |
+
"grad_norm": 0.3515319526195526,
|
| 4003 |
+
"learning_rate": 0.00019954669686638514,
|
| 4004 |
+
"loss": 0.7045,
|
| 4005 |
+
"step": 564
|
| 4006 |
+
},
|
| 4007 |
+
{
|
| 4008 |
+
"epoch": 0.09270844015998359,
|
| 4009 |
+
"grad_norm": 0.32754653692245483,
|
| 4010 |
+
"learning_rate": 0.00019954506015909464,
|
| 4011 |
+
"loss": 0.7455,
|
| 4012 |
+
"step": 565
|
| 4013 |
+
},
|
| 4014 |
+
{
|
| 4015 |
+
"epoch": 0.09287252589478003,
|
| 4016 |
+
"grad_norm": 0.38944128155708313,
|
| 4017 |
+
"learning_rate": 0.00019954342050909534,
|
| 4018 |
+
"loss": 0.7544,
|
| 4019 |
+
"step": 566
|
| 4020 |
+
},
|
| 4021 |
+
{
|
| 4022 |
+
"epoch": 0.09303661162957645,
|
| 4023 |
+
"grad_norm": 0.32821983098983765,
|
| 4024 |
+
"learning_rate": 0.00019954177791643574,
|
| 4025 |
+
"loss": 0.6731,
|
| 4026 |
+
"step": 567
|
| 4027 |
+
},
|
| 4028 |
+
{
|
| 4029 |
+
"epoch": 0.09320069736437288,
|
| 4030 |
+
"grad_norm": 0.34350642561912537,
|
| 4031 |
+
"learning_rate": 0.00019954013238116438,
|
| 4032 |
+
"loss": 0.8399,
|
| 4033 |
+
"step": 568
|
| 4034 |
+
},
|
| 4035 |
+
{
|
| 4036 |
+
"epoch": 0.09336478309916932,
|
| 4037 |
+
"grad_norm": 0.33525383472442627,
|
| 4038 |
+
"learning_rate": 0.00019953848390332992,
|
| 4039 |
+
"loss": 0.8189,
|
| 4040 |
+
"step": 569
|
| 4041 |
+
},
|
| 4042 |
+
{
|
| 4043 |
+
"epoch": 0.09352886883396574,
|
| 4044 |
+
"grad_norm": 0.35323095321655273,
|
| 4045 |
+
"learning_rate": 0.00019953683248298106,
|
| 4046 |
+
"loss": 0.8978,
|
| 4047 |
+
"step": 570
|
| 4048 |
+
},
|
| 4049 |
+
{
|
| 4050 |
+
"epoch": 0.09369295456876219,
|
| 4051 |
+
"grad_norm": 0.34625014662742615,
|
| 4052 |
+
"learning_rate": 0.0001995351781201667,
|
| 4053 |
+
"loss": 0.9358,
|
| 4054 |
+
"step": 571
|
| 4055 |
+
},
|
| 4056 |
+
{
|
| 4057 |
+
"epoch": 0.09385704030355861,
|
| 4058 |
+
"grad_norm": 0.34705492854118347,
|
| 4059 |
+
"learning_rate": 0.00019953352081493567,
|
| 4060 |
+
"loss": 0.6916,
|
| 4061 |
+
"step": 572
|
| 4062 |
+
},
|
| 4063 |
+
{
|
| 4064 |
+
"epoch": 0.09402112603835504,
|
| 4065 |
+
"grad_norm": 0.3524523675441742,
|
| 4066 |
+
"learning_rate": 0.00019953186056733698,
|
| 4067 |
+
"loss": 0.8914,
|
| 4068 |
+
"step": 573
|
| 4069 |
+
},
|
| 4070 |
+
{
|
| 4071 |
+
"epoch": 0.09418521177315148,
|
| 4072 |
+
"grad_norm": 0.3258850574493408,
|
| 4073 |
+
"learning_rate": 0.00019953019737741973,
|
| 4074 |
+
"loss": 0.798,
|
| 4075 |
+
"step": 574
|
| 4076 |
+
},
|
| 4077 |
+
{
|
| 4078 |
+
"epoch": 0.0943492975079479,
|
| 4079 |
+
"grad_norm": 0.34496602416038513,
|
| 4080 |
+
"learning_rate": 0.00019952853124523306,
|
| 4081 |
+
"loss": 0.7979,
|
| 4082 |
+
"step": 575
|
| 4083 |
+
},
|
| 4084 |
+
{
|
| 4085 |
+
"epoch": 0.09451338324274433,
|
| 4086 |
+
"grad_norm": 0.3083381652832031,
|
| 4087 |
+
"learning_rate": 0.00019952686217082621,
|
| 4088 |
+
"loss": 0.8979,
|
| 4089 |
+
"step": 576
|
| 4090 |
+
},
|
| 4091 |
+
{
|
| 4092 |
+
"epoch": 0.09467746897754077,
|
| 4093 |
+
"grad_norm": 0.3396521508693695,
|
| 4094 |
+
"learning_rate": 0.0001995251901542486,
|
| 4095 |
+
"loss": 0.9223,
|
| 4096 |
+
"step": 577
|
| 4097 |
+
},
|
| 4098 |
+
{
|
| 4099 |
+
"epoch": 0.0948415547123372,
|
| 4100 |
+
"grad_norm": 0.30544814467430115,
|
| 4101 |
+
"learning_rate": 0.00019952351519554956,
|
| 4102 |
+
"loss": 0.9028,
|
| 4103 |
+
"step": 578
|
| 4104 |
+
},
|
| 4105 |
+
{
|
| 4106 |
+
"epoch": 0.09500564044713362,
|
| 4107 |
+
"grad_norm": 0.4109005033969879,
|
| 4108 |
+
"learning_rate": 0.00019952183729477865,
|
| 4109 |
+
"loss": 0.7769,
|
| 4110 |
+
"step": 579
|
| 4111 |
+
},
|
| 4112 |
+
{
|
| 4113 |
+
"epoch": 0.09516972618193006,
|
| 4114 |
+
"grad_norm": 0.40897905826568604,
|
| 4115 |
+
"learning_rate": 0.00019952015645198547,
|
| 4116 |
+
"loss": 0.6695,
|
| 4117 |
+
"step": 580
|
| 4118 |
+
},
|
| 4119 |
+
{
|
| 4120 |
+
"epoch": 0.09533381191672649,
|
| 4121 |
+
"grad_norm": 0.4280332624912262,
|
| 4122 |
+
"learning_rate": 0.0001995184726672197,
|
| 4123 |
+
"loss": 0.8043,
|
| 4124 |
+
"step": 581
|
| 4125 |
+
},
|
| 4126 |
+
{
|
| 4127 |
+
"epoch": 0.09549789765152292,
|
| 4128 |
+
"grad_norm": 0.38212618231773376,
|
| 4129 |
+
"learning_rate": 0.00019951678594053114,
|
| 4130 |
+
"loss": 0.8747,
|
| 4131 |
+
"step": 582
|
| 4132 |
+
},
|
| 4133 |
+
{
|
| 4134 |
+
"epoch": 0.09566198338631936,
|
| 4135 |
+
"grad_norm": 0.3472382128238678,
|
| 4136 |
+
"learning_rate": 0.0001995150962719696,
|
| 4137 |
+
"loss": 0.8115,
|
| 4138 |
+
"step": 583
|
| 4139 |
+
},
|
| 4140 |
+
{
|
| 4141 |
+
"epoch": 0.09582606912111578,
|
| 4142 |
+
"grad_norm": 0.43354499340057373,
|
| 4143 |
+
"learning_rate": 0.0001995134036615851,
|
| 4144 |
+
"loss": 0.7358,
|
| 4145 |
+
"step": 584
|
| 4146 |
+
},
|
| 4147 |
+
{
|
| 4148 |
+
"epoch": 0.09599015485591221,
|
| 4149 |
+
"grad_norm": 0.40974995493888855,
|
| 4150 |
+
"learning_rate": 0.00019951170810942763,
|
| 4151 |
+
"loss": 0.8083,
|
| 4152 |
+
"step": 585
|
| 4153 |
+
},
|
| 4154 |
+
{
|
| 4155 |
+
"epoch": 0.09615424059070865,
|
| 4156 |
+
"grad_norm": 0.33038821816444397,
|
| 4157 |
+
"learning_rate": 0.00019951000961554728,
|
| 4158 |
+
"loss": 0.8238,
|
| 4159 |
+
"step": 586
|
| 4160 |
+
},
|
| 4161 |
+
{
|
| 4162 |
+
"epoch": 0.09631832632550508,
|
| 4163 |
+
"grad_norm": 0.32848745584487915,
|
| 4164 |
+
"learning_rate": 0.00019950830817999434,
|
| 4165 |
+
"loss": 0.7763,
|
| 4166 |
+
"step": 587
|
| 4167 |
+
},
|
| 4168 |
+
{
|
| 4169 |
+
"epoch": 0.0964824120603015,
|
| 4170 |
+
"grad_norm": 0.434444397687912,
|
| 4171 |
+
"learning_rate": 0.0001995066038028191,
|
| 4172 |
+
"loss": 0.7627,
|
| 4173 |
+
"step": 588
|
| 4174 |
+
},
|
| 4175 |
+
{
|
| 4176 |
+
"epoch": 0.09664649779509794,
|
| 4177 |
+
"grad_norm": 0.5674214959144592,
|
| 4178 |
+
"learning_rate": 0.00019950489648407188,
|
| 4179 |
+
"loss": 0.8407,
|
| 4180 |
+
"step": 589
|
| 4181 |
+
},
|
| 4182 |
+
{
|
| 4183 |
+
"epoch": 0.09681058352989437,
|
| 4184 |
+
"grad_norm": 0.49122345447540283,
|
| 4185 |
+
"learning_rate": 0.00019950318622380316,
|
| 4186 |
+
"loss": 0.8482,
|
| 4187 |
+
"step": 590
|
| 4188 |
+
},
|
| 4189 |
+
{
|
| 4190 |
+
"epoch": 0.0969746692646908,
|
| 4191 |
+
"grad_norm": 0.40209364891052246,
|
| 4192 |
+
"learning_rate": 0.00019950147302206353,
|
| 4193 |
+
"loss": 0.8458,
|
| 4194 |
+
"step": 591
|
| 4195 |
+
},
|
| 4196 |
+
{
|
| 4197 |
+
"epoch": 0.09713875499948724,
|
| 4198 |
+
"grad_norm": 0.4982354938983917,
|
| 4199 |
+
"learning_rate": 0.00019949975687890366,
|
| 4200 |
+
"loss": 0.9152,
|
| 4201 |
+
"step": 592
|
| 4202 |
+
},
|
| 4203 |
+
{
|
| 4204 |
+
"epoch": 0.09730284073428366,
|
| 4205 |
+
"grad_norm": 0.4886113703250885,
|
| 4206 |
+
"learning_rate": 0.00019949803779437426,
|
| 4207 |
+
"loss": 0.8261,
|
| 4208 |
+
"step": 593
|
| 4209 |
+
},
|
| 4210 |
+
{
|
| 4211 |
+
"epoch": 0.09746692646908009,
|
| 4212 |
+
"grad_norm": 0.3699701130390167,
|
| 4213 |
+
"learning_rate": 0.0001994963157685261,
|
| 4214 |
+
"loss": 0.8339,
|
| 4215 |
+
"step": 594
|
| 4216 |
+
},
|
| 4217 |
+
{
|
| 4218 |
+
"epoch": 0.09763101220387653,
|
| 4219 |
+
"grad_norm": 0.3579908013343811,
|
| 4220 |
+
"learning_rate": 0.00019949459080141014,
|
| 4221 |
+
"loss": 0.7867,
|
| 4222 |
+
"step": 595
|
| 4223 |
+
},
|
| 4224 |
+
{
|
| 4225 |
+
"epoch": 0.09779509793867296,
|
| 4226 |
+
"grad_norm": 0.3716687262058258,
|
| 4227 |
+
"learning_rate": 0.00019949286289307737,
|
| 4228 |
+
"loss": 0.7762,
|
| 4229 |
+
"step": 596
|
| 4230 |
+
},
|
| 4231 |
+
{
|
| 4232 |
+
"epoch": 0.09795918367346938,
|
| 4233 |
+
"grad_norm": 0.5843008160591125,
|
| 4234 |
+
"learning_rate": 0.00019949113204357883,
|
| 4235 |
+
"loss": 0.9435,
|
| 4236 |
+
"step": 597
|
| 4237 |
+
},
|
| 4238 |
+
{
|
| 4239 |
+
"epoch": 0.09812326940826582,
|
| 4240 |
+
"grad_norm": 0.5935463309288025,
|
| 4241 |
+
"learning_rate": 0.00019948939825296572,
|
| 4242 |
+
"loss": 0.7525,
|
| 4243 |
+
"step": 598
|
| 4244 |
+
},
|
| 4245 |
+
{
|
| 4246 |
+
"epoch": 0.09828735514306225,
|
| 4247 |
+
"grad_norm": 0.6296449303627014,
|
| 4248 |
+
"learning_rate": 0.00019948766152128928,
|
| 4249 |
+
"loss": 0.8801,
|
| 4250 |
+
"step": 599
|
| 4251 |
+
},
|
| 4252 |
+
{
|
| 4253 |
+
"epoch": 0.09845144087785868,
|
| 4254 |
+
"grad_norm": 0.7150892615318298,
|
| 4255 |
+
"learning_rate": 0.00019948592184860088,
|
| 4256 |
+
"loss": 0.7897,
|
| 4257 |
+
"step": 600
|
| 4258 |
+
},
|
| 4259 |
+
{
|
| 4260 |
+
"epoch": 0.09845144087785868,
|
| 4261 |
+
"eval_loss": 0.7834404110908508,
|
| 4262 |
+
"eval_runtime": 70.8563,
|
| 4263 |
+
"eval_samples_per_second": 3.458,
|
| 4264 |
+
"eval_steps_per_second": 3.458,
|
| 4265 |
+
"step": 600
|
| 4266 |
}
|
| 4267 |
],
|
| 4268 |
"logging_steps": 1,
|
|
|
|
| 4291 |
"attributes": {}
|
| 4292 |
}
|
| 4293 |
},
|
| 4294 |
+
"total_flos": 8.361437762578022e+16,
|
| 4295 |
"train_batch_size": 1,
|
| 4296 |
"trial_name": null,
|
| 4297 |
"trial_params": null
|