Training in progress, step 400, 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 335604696
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:5cb855745e92129072755755d9fdaa0486ec6942c96133c5b09f950a7b7029bc
|
| 3 |
size 335604696
|
last-checkpoint/optimizer.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 90187222
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:d37d2ea5d99c8376477f778d2dcab535a6a9efb26d189d797cbde7bd3ee7a896
|
| 3 |
size 90187222
|
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:a34ad82c37dafd9525216378ae3698e89bb327fac4990ae643a10de8796b8c2e
|
| 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 2080
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:4940e2af5b9950a163fe737c94b79bcc3933d1cf31ad7ac593d5b072cbf9d00c
|
| 3 |
size 2080
|
last-checkpoint/trainer_state.json
CHANGED
|
@@ -1,9 +1,9 @@
|
|
| 1 |
{
|
| 2 |
"best_metric": 1.4815335273742676,
|
| 3 |
"best_model_checkpoint": "miner_id_24/checkpoint-200",
|
| 4 |
-
"epoch": 0.
|
| 5 |
"eval_steps": 200,
|
| 6 |
-
"global_step":
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
@@ -1423,6 +1423,1414 @@
|
|
| 1423 |
"eval_samples_per_second": 5.118,
|
| 1424 |
"eval_steps_per_second": 2.564,
|
| 1425 |
"step": 200
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1426 |
}
|
| 1427 |
],
|
| 1428 |
"logging_steps": 1,
|
|
@@ -1437,7 +2845,7 @@
|
|
| 1437 |
"early_stopping_threshold": 0.0
|
| 1438 |
},
|
| 1439 |
"attributes": {
|
| 1440 |
-
"early_stopping_patience_counter":
|
| 1441 |
}
|
| 1442 |
},
|
| 1443 |
"TrainerControl": {
|
|
@@ -1451,7 +2859,7 @@
|
|
| 1451 |
"attributes": {}
|
| 1452 |
}
|
| 1453 |
},
|
| 1454 |
-
"total_flos":
|
| 1455 |
"train_batch_size": 2,
|
| 1456 |
"trial_name": null,
|
| 1457 |
"trial_params": null
|
|
|
|
| 1 |
{
|
| 2 |
"best_metric": 1.4815335273742676,
|
| 3 |
"best_model_checkpoint": "miner_id_24/checkpoint-200",
|
| 4 |
+
"epoch": 0.031968670702711345,
|
| 5 |
"eval_steps": 200,
|
| 6 |
+
"global_step": 400,
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
|
|
| 1423 |
"eval_samples_per_second": 5.118,
|
| 1424 |
"eval_steps_per_second": 2.564,
|
| 1425 |
"step": 200
|
| 1426 |
+
},
|
| 1427 |
+
{
|
| 1428 |
+
"epoch": 0.01606425702811245,
|
| 1429 |
+
"grad_norm": 15.242115020751953,
|
| 1430 |
+
"learning_rate": 0.00019998973584733903,
|
| 1431 |
+
"loss": 6.5955,
|
| 1432 |
+
"step": 201
|
| 1433 |
+
},
|
| 1434 |
+
{
|
| 1435 |
+
"epoch": 0.016144178704869228,
|
| 1436 |
+
"grad_norm": 9.51711654663086,
|
| 1437 |
+
"learning_rate": 0.00019998960488010198,
|
| 1438 |
+
"loss": 5.5036,
|
| 1439 |
+
"step": 202
|
| 1440 |
+
},
|
| 1441 |
+
{
|
| 1442 |
+
"epoch": 0.016224100381626007,
|
| 1443 |
+
"grad_norm": 10.303105354309082,
|
| 1444 |
+
"learning_rate": 0.00019998948846478015,
|
| 1445 |
+
"loss": 6.4441,
|
| 1446 |
+
"step": 203
|
| 1447 |
+
},
|
| 1448 |
+
{
|
| 1449 |
+
"epoch": 0.016304022058382786,
|
| 1450 |
+
"grad_norm": 9.889719009399414,
|
| 1451 |
+
"learning_rate": 0.00019998937204945832,
|
| 1452 |
+
"loss": 4.8753,
|
| 1453 |
+
"step": 204
|
| 1454 |
+
},
|
| 1455 |
+
{
|
| 1456 |
+
"epoch": 0.016383943735139562,
|
| 1457 |
+
"grad_norm": 11.760963439941406,
|
| 1458 |
+
"learning_rate": 0.0001999892556341365,
|
| 1459 |
+
"loss": 6.3996,
|
| 1460 |
+
"step": 205
|
| 1461 |
+
},
|
| 1462 |
+
{
|
| 1463 |
+
"epoch": 0.01646386541189634,
|
| 1464 |
+
"grad_norm": 11.809788703918457,
|
| 1465 |
+
"learning_rate": 0.00019998912466689944,
|
| 1466 |
+
"loss": 5.3402,
|
| 1467 |
+
"step": 206
|
| 1468 |
+
},
|
| 1469 |
+
{
|
| 1470 |
+
"epoch": 0.01654378708865312,
|
| 1471 |
+
"grad_norm": 12.838322639465332,
|
| 1472 |
+
"learning_rate": 0.00019998900825157762,
|
| 1473 |
+
"loss": 5.6018,
|
| 1474 |
+
"step": 207
|
| 1475 |
+
},
|
| 1476 |
+
{
|
| 1477 |
+
"epoch": 0.0166237087654099,
|
| 1478 |
+
"grad_norm": 17.00547981262207,
|
| 1479 |
+
"learning_rate": 0.0001999888918362558,
|
| 1480 |
+
"loss": 5.6337,
|
| 1481 |
+
"step": 208
|
| 1482 |
+
},
|
| 1483 |
+
{
|
| 1484 |
+
"epoch": 0.016703630442166675,
|
| 1485 |
+
"grad_norm": 13.280611038208008,
|
| 1486 |
+
"learning_rate": 0.0001999887463171035,
|
| 1487 |
+
"loss": 5.0759,
|
| 1488 |
+
"step": 209
|
| 1489 |
+
},
|
| 1490 |
+
{
|
| 1491 |
+
"epoch": 0.016783552118923455,
|
| 1492 |
+
"grad_norm": 14.64804458618164,
|
| 1493 |
+
"learning_rate": 0.00019998862990178168,
|
| 1494 |
+
"loss": 6.039,
|
| 1495 |
+
"step": 210
|
| 1496 |
+
},
|
| 1497 |
+
{
|
| 1498 |
+
"epoch": 0.016863473795680234,
|
| 1499 |
+
"grad_norm": 14.111835479736328,
|
| 1500 |
+
"learning_rate": 0.00019998849893454462,
|
| 1501 |
+
"loss": 6.1179,
|
| 1502 |
+
"step": 211
|
| 1503 |
+
},
|
| 1504 |
+
{
|
| 1505 |
+
"epoch": 0.016943395472437013,
|
| 1506 |
+
"grad_norm": 13.955074310302734,
|
| 1507 |
+
"learning_rate": 0.00019998836796730757,
|
| 1508 |
+
"loss": 5.9847,
|
| 1509 |
+
"step": 212
|
| 1510 |
+
},
|
| 1511 |
+
{
|
| 1512 |
+
"epoch": 0.01702331714919379,
|
| 1513 |
+
"grad_norm": 16.10040855407715,
|
| 1514 |
+
"learning_rate": 0.00019998825155198574,
|
| 1515 |
+
"loss": 6.3676,
|
| 1516 |
+
"step": 213
|
| 1517 |
+
},
|
| 1518 |
+
{
|
| 1519 |
+
"epoch": 0.017103238825950568,
|
| 1520 |
+
"grad_norm": 14.447005271911621,
|
| 1521 |
+
"learning_rate": 0.00019998812058474869,
|
| 1522 |
+
"loss": 5.6564,
|
| 1523 |
+
"step": 214
|
| 1524 |
+
},
|
| 1525 |
+
{
|
| 1526 |
+
"epoch": 0.017183160502707347,
|
| 1527 |
+
"grad_norm": 10.266460418701172,
|
| 1528 |
+
"learning_rate": 0.00019998798961751163,
|
| 1529 |
+
"loss": 5.3369,
|
| 1530 |
+
"step": 215
|
| 1531 |
+
},
|
| 1532 |
+
{
|
| 1533 |
+
"epoch": 0.017263082179464127,
|
| 1534 |
+
"grad_norm": 17.443172454833984,
|
| 1535 |
+
"learning_rate": 0.00019998785865027457,
|
| 1536 |
+
"loss": 5.487,
|
| 1537 |
+
"step": 216
|
| 1538 |
+
},
|
| 1539 |
+
{
|
| 1540 |
+
"epoch": 0.017343003856220902,
|
| 1541 |
+
"grad_norm": 19.374847412109375,
|
| 1542 |
+
"learning_rate": 0.00019998772768303752,
|
| 1543 |
+
"loss": 5.2789,
|
| 1544 |
+
"step": 217
|
| 1545 |
+
},
|
| 1546 |
+
{
|
| 1547 |
+
"epoch": 0.01742292553297768,
|
| 1548 |
+
"grad_norm": 13.237730979919434,
|
| 1549 |
+
"learning_rate": 0.00019998759671580046,
|
| 1550 |
+
"loss": 5.4945,
|
| 1551 |
+
"step": 218
|
| 1552 |
+
},
|
| 1553 |
+
{
|
| 1554 |
+
"epoch": 0.01750284720973446,
|
| 1555 |
+
"grad_norm": 12.152576446533203,
|
| 1556 |
+
"learning_rate": 0.0001999874657485634,
|
| 1557 |
+
"loss": 5.4477,
|
| 1558 |
+
"step": 219
|
| 1559 |
+
},
|
| 1560 |
+
{
|
| 1561 |
+
"epoch": 0.01758276888649124,
|
| 1562 |
+
"grad_norm": 17.59059715270996,
|
| 1563 |
+
"learning_rate": 0.00019998733478132635,
|
| 1564 |
+
"loss": 6.2299,
|
| 1565 |
+
"step": 220
|
| 1566 |
+
},
|
| 1567 |
+
{
|
| 1568 |
+
"epoch": 0.017662690563248016,
|
| 1569 |
+
"grad_norm": 12.863672256469727,
|
| 1570 |
+
"learning_rate": 0.00019998718926217407,
|
| 1571 |
+
"loss": 6.4539,
|
| 1572 |
+
"step": 221
|
| 1573 |
+
},
|
| 1574 |
+
{
|
| 1575 |
+
"epoch": 0.017742612240004795,
|
| 1576 |
+
"grad_norm": 12.739554405212402,
|
| 1577 |
+
"learning_rate": 0.00019998705829493701,
|
| 1578 |
+
"loss": 4.7151,
|
| 1579 |
+
"step": 222
|
| 1580 |
+
},
|
| 1581 |
+
{
|
| 1582 |
+
"epoch": 0.017822533916761574,
|
| 1583 |
+
"grad_norm": 16.173961639404297,
|
| 1584 |
+
"learning_rate": 0.00019998692732769996,
|
| 1585 |
+
"loss": 7.2765,
|
| 1586 |
+
"step": 223
|
| 1587 |
+
},
|
| 1588 |
+
{
|
| 1589 |
+
"epoch": 0.017902455593518354,
|
| 1590 |
+
"grad_norm": 22.782018661499023,
|
| 1591 |
+
"learning_rate": 0.00019998678180854768,
|
| 1592 |
+
"loss": 6.582,
|
| 1593 |
+
"step": 224
|
| 1594 |
+
},
|
| 1595 |
+
{
|
| 1596 |
+
"epoch": 0.01798237727027513,
|
| 1597 |
+
"grad_norm": 16.2958927154541,
|
| 1598 |
+
"learning_rate": 0.00019998665084131062,
|
| 1599 |
+
"loss": 5.832,
|
| 1600 |
+
"step": 225
|
| 1601 |
+
},
|
| 1602 |
+
{
|
| 1603 |
+
"epoch": 0.01806229894703191,
|
| 1604 |
+
"grad_norm": 11.503108978271484,
|
| 1605 |
+
"learning_rate": 0.00019998651987407357,
|
| 1606 |
+
"loss": 5.2539,
|
| 1607 |
+
"step": 226
|
| 1608 |
+
},
|
| 1609 |
+
{
|
| 1610 |
+
"epoch": 0.018142220623788688,
|
| 1611 |
+
"grad_norm": 15.306110382080078,
|
| 1612 |
+
"learning_rate": 0.00019998637435492128,
|
| 1613 |
+
"loss": 5.9638,
|
| 1614 |
+
"step": 227
|
| 1615 |
+
},
|
| 1616 |
+
{
|
| 1617 |
+
"epoch": 0.018222142300545464,
|
| 1618 |
+
"grad_norm": 12.46424674987793,
|
| 1619 |
+
"learning_rate": 0.00019998624338768423,
|
| 1620 |
+
"loss": 5.5315,
|
| 1621 |
+
"step": 228
|
| 1622 |
+
},
|
| 1623 |
+
{
|
| 1624 |
+
"epoch": 0.018302063977302243,
|
| 1625 |
+
"grad_norm": 15.144165992736816,
|
| 1626 |
+
"learning_rate": 0.00019998611242044717,
|
| 1627 |
+
"loss": 6.3079,
|
| 1628 |
+
"step": 229
|
| 1629 |
+
},
|
| 1630 |
+
{
|
| 1631 |
+
"epoch": 0.018381985654059022,
|
| 1632 |
+
"grad_norm": 14.480883598327637,
|
| 1633 |
+
"learning_rate": 0.0001999859669012949,
|
| 1634 |
+
"loss": 5.4738,
|
| 1635 |
+
"step": 230
|
| 1636 |
+
},
|
| 1637 |
+
{
|
| 1638 |
+
"epoch": 0.0184619073308158,
|
| 1639 |
+
"grad_norm": 13.642375946044922,
|
| 1640 |
+
"learning_rate": 0.0001999858213821426,
|
| 1641 |
+
"loss": 5.5203,
|
| 1642 |
+
"step": 231
|
| 1643 |
+
},
|
| 1644 |
+
{
|
| 1645 |
+
"epoch": 0.018541829007572577,
|
| 1646 |
+
"grad_norm": 14.766783714294434,
|
| 1647 |
+
"learning_rate": 0.00019998567586299032,
|
| 1648 |
+
"loss": 6.6556,
|
| 1649 |
+
"step": 232
|
| 1650 |
+
},
|
| 1651 |
+
{
|
| 1652 |
+
"epoch": 0.018621750684329356,
|
| 1653 |
+
"grad_norm": 11.702398300170898,
|
| 1654 |
+
"learning_rate": 0.00019998553034383804,
|
| 1655 |
+
"loss": 6.1329,
|
| 1656 |
+
"step": 233
|
| 1657 |
+
},
|
| 1658 |
+
{
|
| 1659 |
+
"epoch": 0.018701672361086136,
|
| 1660 |
+
"grad_norm": 10.976285934448242,
|
| 1661 |
+
"learning_rate": 0.00019998538482468575,
|
| 1662 |
+
"loss": 5.2038,
|
| 1663 |
+
"step": 234
|
| 1664 |
+
},
|
| 1665 |
+
{
|
| 1666 |
+
"epoch": 0.018781594037842915,
|
| 1667 |
+
"grad_norm": 17.724811553955078,
|
| 1668 |
+
"learning_rate": 0.0001999852538574487,
|
| 1669 |
+
"loss": 6.0959,
|
| 1670 |
+
"step": 235
|
| 1671 |
+
},
|
| 1672 |
+
{
|
| 1673 |
+
"epoch": 0.01886151571459969,
|
| 1674 |
+
"grad_norm": 13.917747497558594,
|
| 1675 |
+
"learning_rate": 0.0001999851083382964,
|
| 1676 |
+
"loss": 5.6464,
|
| 1677 |
+
"step": 236
|
| 1678 |
+
},
|
| 1679 |
+
{
|
| 1680 |
+
"epoch": 0.01894143739135647,
|
| 1681 |
+
"grad_norm": 15.578882217407227,
|
| 1682 |
+
"learning_rate": 0.00019998496281914413,
|
| 1683 |
+
"loss": 6.1002,
|
| 1684 |
+
"step": 237
|
| 1685 |
+
},
|
| 1686 |
+
{
|
| 1687 |
+
"epoch": 0.01902135906811325,
|
| 1688 |
+
"grad_norm": 13.99124813079834,
|
| 1689 |
+
"learning_rate": 0.00019998481729999185,
|
| 1690 |
+
"loss": 6.1651,
|
| 1691 |
+
"step": 238
|
| 1692 |
+
},
|
| 1693 |
+
{
|
| 1694 |
+
"epoch": 0.01910128074487003,
|
| 1695 |
+
"grad_norm": 14.303272247314453,
|
| 1696 |
+
"learning_rate": 0.00019998467178083956,
|
| 1697 |
+
"loss": 6.2277,
|
| 1698 |
+
"step": 239
|
| 1699 |
+
},
|
| 1700 |
+
{
|
| 1701 |
+
"epoch": 0.019181202421626804,
|
| 1702 |
+
"grad_norm": 12.441352844238281,
|
| 1703 |
+
"learning_rate": 0.00019998452626168728,
|
| 1704 |
+
"loss": 5.3236,
|
| 1705 |
+
"step": 240
|
| 1706 |
+
},
|
| 1707 |
+
{
|
| 1708 |
+
"epoch": 0.019261124098383584,
|
| 1709 |
+
"grad_norm": 16.971145629882812,
|
| 1710 |
+
"learning_rate": 0.000199984380742535,
|
| 1711 |
+
"loss": 5.6737,
|
| 1712 |
+
"step": 241
|
| 1713 |
+
},
|
| 1714 |
+
{
|
| 1715 |
+
"epoch": 0.019341045775140363,
|
| 1716 |
+
"grad_norm": 17.1169490814209,
|
| 1717 |
+
"learning_rate": 0.00019998422067146748,
|
| 1718 |
+
"loss": 6.8689,
|
| 1719 |
+
"step": 242
|
| 1720 |
+
},
|
| 1721 |
+
{
|
| 1722 |
+
"epoch": 0.019420967451897142,
|
| 1723 |
+
"grad_norm": 13.830810546875,
|
| 1724 |
+
"learning_rate": 0.0001999840751523152,
|
| 1725 |
+
"loss": 4.9328,
|
| 1726 |
+
"step": 243
|
| 1727 |
+
},
|
| 1728 |
+
{
|
| 1729 |
+
"epoch": 0.019500889128653918,
|
| 1730 |
+
"grad_norm": 13.696432113647461,
|
| 1731 |
+
"learning_rate": 0.00019998392963316292,
|
| 1732 |
+
"loss": 5.6886,
|
| 1733 |
+
"step": 244
|
| 1734 |
+
},
|
| 1735 |
+
{
|
| 1736 |
+
"epoch": 0.019580810805410697,
|
| 1737 |
+
"grad_norm": 10.656487464904785,
|
| 1738 |
+
"learning_rate": 0.00019998378411401063,
|
| 1739 |
+
"loss": 6.0341,
|
| 1740 |
+
"step": 245
|
| 1741 |
+
},
|
| 1742 |
+
{
|
| 1743 |
+
"epoch": 0.019660732482167476,
|
| 1744 |
+
"grad_norm": 13.665307998657227,
|
| 1745 |
+
"learning_rate": 0.00019998362404294312,
|
| 1746 |
+
"loss": 5.4808,
|
| 1747 |
+
"step": 246
|
| 1748 |
+
},
|
| 1749 |
+
{
|
| 1750 |
+
"epoch": 0.019740654158924256,
|
| 1751 |
+
"grad_norm": 13.564827919006348,
|
| 1752 |
+
"learning_rate": 0.00019998347852379084,
|
| 1753 |
+
"loss": 5.8875,
|
| 1754 |
+
"step": 247
|
| 1755 |
+
},
|
| 1756 |
+
{
|
| 1757 |
+
"epoch": 0.01982057583568103,
|
| 1758 |
+
"grad_norm": 13.60628890991211,
|
| 1759 |
+
"learning_rate": 0.00019998333300463855,
|
| 1760 |
+
"loss": 5.9795,
|
| 1761 |
+
"step": 248
|
| 1762 |
+
},
|
| 1763 |
+
{
|
| 1764 |
+
"epoch": 0.01990049751243781,
|
| 1765 |
+
"grad_norm": 15.560327529907227,
|
| 1766 |
+
"learning_rate": 0.00019998317293357104,
|
| 1767 |
+
"loss": 5.5713,
|
| 1768 |
+
"step": 249
|
| 1769 |
+
},
|
| 1770 |
+
{
|
| 1771 |
+
"epoch": 0.01998041918919459,
|
| 1772 |
+
"grad_norm": 16.924474716186523,
|
| 1773 |
+
"learning_rate": 0.00019998301286250353,
|
| 1774 |
+
"loss": 6.835,
|
| 1775 |
+
"step": 250
|
| 1776 |
+
},
|
| 1777 |
+
{
|
| 1778 |
+
"epoch": 0.02006034086595137,
|
| 1779 |
+
"grad_norm": 13.334891319274902,
|
| 1780 |
+
"learning_rate": 0.00019998286734335124,
|
| 1781 |
+
"loss": 6.8343,
|
| 1782 |
+
"step": 251
|
| 1783 |
+
},
|
| 1784 |
+
{
|
| 1785 |
+
"epoch": 0.020140262542708145,
|
| 1786 |
+
"grad_norm": 14.925477027893066,
|
| 1787 |
+
"learning_rate": 0.00019998270727228373,
|
| 1788 |
+
"loss": 5.8623,
|
| 1789 |
+
"step": 252
|
| 1790 |
+
},
|
| 1791 |
+
{
|
| 1792 |
+
"epoch": 0.020220184219464924,
|
| 1793 |
+
"grad_norm": 10.16474437713623,
|
| 1794 |
+
"learning_rate": 0.00019998256175313145,
|
| 1795 |
+
"loss": 5.9831,
|
| 1796 |
+
"step": 253
|
| 1797 |
+
},
|
| 1798 |
+
{
|
| 1799 |
+
"epoch": 0.020300105896221703,
|
| 1800 |
+
"grad_norm": 15.137889862060547,
|
| 1801 |
+
"learning_rate": 0.0001999823871301487,
|
| 1802 |
+
"loss": 4.9667,
|
| 1803 |
+
"step": 254
|
| 1804 |
+
},
|
| 1805 |
+
{
|
| 1806 |
+
"epoch": 0.020380027572978483,
|
| 1807 |
+
"grad_norm": 18.675920486450195,
|
| 1808 |
+
"learning_rate": 0.0001999822270590812,
|
| 1809 |
+
"loss": 5.0038,
|
| 1810 |
+
"step": 255
|
| 1811 |
+
},
|
| 1812 |
+
{
|
| 1813 |
+
"epoch": 0.02045994924973526,
|
| 1814 |
+
"grad_norm": 18.873275756835938,
|
| 1815 |
+
"learning_rate": 0.0001999820815399289,
|
| 1816 |
+
"loss": 5.7058,
|
| 1817 |
+
"step": 256
|
| 1818 |
+
},
|
| 1819 |
+
{
|
| 1820 |
+
"epoch": 0.020539870926492038,
|
| 1821 |
+
"grad_norm": 15.434806823730469,
|
| 1822 |
+
"learning_rate": 0.0001999819214688614,
|
| 1823 |
+
"loss": 5.4421,
|
| 1824 |
+
"step": 257
|
| 1825 |
+
},
|
| 1826 |
+
{
|
| 1827 |
+
"epoch": 0.020619792603248817,
|
| 1828 |
+
"grad_norm": 15.471893310546875,
|
| 1829 |
+
"learning_rate": 0.0001999817613977939,
|
| 1830 |
+
"loss": 5.9829,
|
| 1831 |
+
"step": 258
|
| 1832 |
+
},
|
| 1833 |
+
{
|
| 1834 |
+
"epoch": 0.020699714280005596,
|
| 1835 |
+
"grad_norm": 12.350846290588379,
|
| 1836 |
+
"learning_rate": 0.00019998160132672638,
|
| 1837 |
+
"loss": 6.2408,
|
| 1838 |
+
"step": 259
|
| 1839 |
+
},
|
| 1840 |
+
{
|
| 1841 |
+
"epoch": 0.020779635956762372,
|
| 1842 |
+
"grad_norm": 14.518590927124023,
|
| 1843 |
+
"learning_rate": 0.00019998144125565886,
|
| 1844 |
+
"loss": 5.2333,
|
| 1845 |
+
"step": 260
|
| 1846 |
+
},
|
| 1847 |
+
{
|
| 1848 |
+
"epoch": 0.02085955763351915,
|
| 1849 |
+
"grad_norm": 14.197212219238281,
|
| 1850 |
+
"learning_rate": 0.00019998128118459135,
|
| 1851 |
+
"loss": 5.4315,
|
| 1852 |
+
"step": 261
|
| 1853 |
+
},
|
| 1854 |
+
{
|
| 1855 |
+
"epoch": 0.02093947931027593,
|
| 1856 |
+
"grad_norm": 11.51275634765625,
|
| 1857 |
+
"learning_rate": 0.0001999811065616086,
|
| 1858 |
+
"loss": 4.8215,
|
| 1859 |
+
"step": 262
|
| 1860 |
+
},
|
| 1861 |
+
{
|
| 1862 |
+
"epoch": 0.02101940098703271,
|
| 1863 |
+
"grad_norm": 13.645966529846191,
|
| 1864 |
+
"learning_rate": 0.0001999809464905411,
|
| 1865 |
+
"loss": 5.6501,
|
| 1866 |
+
"step": 263
|
| 1867 |
+
},
|
| 1868 |
+
{
|
| 1869 |
+
"epoch": 0.021099322663789485,
|
| 1870 |
+
"grad_norm": 12.8093843460083,
|
| 1871 |
+
"learning_rate": 0.0001999807864194736,
|
| 1872 |
+
"loss": 5.3092,
|
| 1873 |
+
"step": 264
|
| 1874 |
+
},
|
| 1875 |
+
{
|
| 1876 |
+
"epoch": 0.021179244340546265,
|
| 1877 |
+
"grad_norm": 12.535063743591309,
|
| 1878 |
+
"learning_rate": 0.00019998062634840608,
|
| 1879 |
+
"loss": 5.1969,
|
| 1880 |
+
"step": 265
|
| 1881 |
+
},
|
| 1882 |
+
{
|
| 1883 |
+
"epoch": 0.021259166017303044,
|
| 1884 |
+
"grad_norm": 18.322696685791016,
|
| 1885 |
+
"learning_rate": 0.00019998045172542334,
|
| 1886 |
+
"loss": 6.1928,
|
| 1887 |
+
"step": 266
|
| 1888 |
+
},
|
| 1889 |
+
{
|
| 1890 |
+
"epoch": 0.02133908769405982,
|
| 1891 |
+
"grad_norm": 12.753133773803711,
|
| 1892 |
+
"learning_rate": 0.00019998029165435582,
|
| 1893 |
+
"loss": 6.3926,
|
| 1894 |
+
"step": 267
|
| 1895 |
+
},
|
| 1896 |
+
{
|
| 1897 |
+
"epoch": 0.0214190093708166,
|
| 1898 |
+
"grad_norm": 10.863653182983398,
|
| 1899 |
+
"learning_rate": 0.00019998011703137308,
|
| 1900 |
+
"loss": 6.2499,
|
| 1901 |
+
"step": 268
|
| 1902 |
+
},
|
| 1903 |
+
{
|
| 1904 |
+
"epoch": 0.021498931047573378,
|
| 1905 |
+
"grad_norm": 15.866484642028809,
|
| 1906 |
+
"learning_rate": 0.00019997995696030557,
|
| 1907 |
+
"loss": 6.2885,
|
| 1908 |
+
"step": 269
|
| 1909 |
+
},
|
| 1910 |
+
{
|
| 1911 |
+
"epoch": 0.021578852724330157,
|
| 1912 |
+
"grad_norm": 18.528520584106445,
|
| 1913 |
+
"learning_rate": 0.00019997978233732283,
|
| 1914 |
+
"loss": 5.9365,
|
| 1915 |
+
"step": 270
|
| 1916 |
+
},
|
| 1917 |
+
{
|
| 1918 |
+
"epoch": 0.021658774401086933,
|
| 1919 |
+
"grad_norm": 17.615192413330078,
|
| 1920 |
+
"learning_rate": 0.00019997962226625532,
|
| 1921 |
+
"loss": 5.6754,
|
| 1922 |
+
"step": 271
|
| 1923 |
+
},
|
| 1924 |
+
{
|
| 1925 |
+
"epoch": 0.021738696077843712,
|
| 1926 |
+
"grad_norm": 16.032712936401367,
|
| 1927 |
+
"learning_rate": 0.0001999794621951878,
|
| 1928 |
+
"loss": 6.0404,
|
| 1929 |
+
"step": 272
|
| 1930 |
+
},
|
| 1931 |
+
{
|
| 1932 |
+
"epoch": 0.02181861775460049,
|
| 1933 |
+
"grad_norm": 13.623299598693848,
|
| 1934 |
+
"learning_rate": 0.00019997927302028984,
|
| 1935 |
+
"loss": 6.0127,
|
| 1936 |
+
"step": 273
|
| 1937 |
+
},
|
| 1938 |
+
{
|
| 1939 |
+
"epoch": 0.02189853943135727,
|
| 1940 |
+
"grad_norm": 12.06942081451416,
|
| 1941 |
+
"learning_rate": 0.00019997911294922233,
|
| 1942 |
+
"loss": 5.1208,
|
| 1943 |
+
"step": 274
|
| 1944 |
+
},
|
| 1945 |
+
{
|
| 1946 |
+
"epoch": 0.021978461108114047,
|
| 1947 |
+
"grad_norm": 13.500571250915527,
|
| 1948 |
+
"learning_rate": 0.00019997893832623959,
|
| 1949 |
+
"loss": 5.3658,
|
| 1950 |
+
"step": 275
|
| 1951 |
+
},
|
| 1952 |
+
{
|
| 1953 |
+
"epoch": 0.022058382784870826,
|
| 1954 |
+
"grad_norm": 14.099763870239258,
|
| 1955 |
+
"learning_rate": 0.00019997876370325685,
|
| 1956 |
+
"loss": 6.2572,
|
| 1957 |
+
"step": 276
|
| 1958 |
+
},
|
| 1959 |
+
{
|
| 1960 |
+
"epoch": 0.022138304461627605,
|
| 1961 |
+
"grad_norm": 15.197004318237305,
|
| 1962 |
+
"learning_rate": 0.00019997860363218933,
|
| 1963 |
+
"loss": 5.8418,
|
| 1964 |
+
"step": 277
|
| 1965 |
+
},
|
| 1966 |
+
{
|
| 1967 |
+
"epoch": 0.022218226138384384,
|
| 1968 |
+
"grad_norm": 16.935300827026367,
|
| 1969 |
+
"learning_rate": 0.00019997841445729136,
|
| 1970 |
+
"loss": 6.2711,
|
| 1971 |
+
"step": 278
|
| 1972 |
+
},
|
| 1973 |
+
{
|
| 1974 |
+
"epoch": 0.02229814781514116,
|
| 1975 |
+
"grad_norm": 15.207847595214844,
|
| 1976 |
+
"learning_rate": 0.00019997825438622385,
|
| 1977 |
+
"loss": 6.2159,
|
| 1978 |
+
"step": 279
|
| 1979 |
+
},
|
| 1980 |
+
{
|
| 1981 |
+
"epoch": 0.02237806949189794,
|
| 1982 |
+
"grad_norm": 15.968687057495117,
|
| 1983 |
+
"learning_rate": 0.00019997806521132588,
|
| 1984 |
+
"loss": 7.2275,
|
| 1985 |
+
"step": 280
|
| 1986 |
+
},
|
| 1987 |
+
{
|
| 1988 |
+
"epoch": 0.02245799116865472,
|
| 1989 |
+
"grad_norm": 12.241089820861816,
|
| 1990 |
+
"learning_rate": 0.00019997789058834314,
|
| 1991 |
+
"loss": 6.8563,
|
| 1992 |
+
"step": 281
|
| 1993 |
+
},
|
| 1994 |
+
{
|
| 1995 |
+
"epoch": 0.022537912845411498,
|
| 1996 |
+
"grad_norm": 12.12191390991211,
|
| 1997 |
+
"learning_rate": 0.0001999777159653604,
|
| 1998 |
+
"loss": 6.3478,
|
| 1999 |
+
"step": 282
|
| 2000 |
+
},
|
| 2001 |
+
{
|
| 2002 |
+
"epoch": 0.022617834522168274,
|
| 2003 |
+
"grad_norm": 15.668216705322266,
|
| 2004 |
+
"learning_rate": 0.00019997754134237766,
|
| 2005 |
+
"loss": 6.4098,
|
| 2006 |
+
"step": 283
|
| 2007 |
+
},
|
| 2008 |
+
{
|
| 2009 |
+
"epoch": 0.022697756198925053,
|
| 2010 |
+
"grad_norm": 13.400096893310547,
|
| 2011 |
+
"learning_rate": 0.00019997736671939492,
|
| 2012 |
+
"loss": 5.2505,
|
| 2013 |
+
"step": 284
|
| 2014 |
+
},
|
| 2015 |
+
{
|
| 2016 |
+
"epoch": 0.022777677875681832,
|
| 2017 |
+
"grad_norm": 11.265392303466797,
|
| 2018 |
+
"learning_rate": 0.00019997719209641218,
|
| 2019 |
+
"loss": 5.1597,
|
| 2020 |
+
"step": 285
|
| 2021 |
+
},
|
| 2022 |
+
{
|
| 2023 |
+
"epoch": 0.02285759955243861,
|
| 2024 |
+
"grad_norm": 19.430946350097656,
|
| 2025 |
+
"learning_rate": 0.0001999770029215142,
|
| 2026 |
+
"loss": 6.6648,
|
| 2027 |
+
"step": 286
|
| 2028 |
+
},
|
| 2029 |
+
{
|
| 2030 |
+
"epoch": 0.022937521229195387,
|
| 2031 |
+
"grad_norm": 13.3882474899292,
|
| 2032 |
+
"learning_rate": 0.00019997681374661624,
|
| 2033 |
+
"loss": 5.9156,
|
| 2034 |
+
"step": 287
|
| 2035 |
+
},
|
| 2036 |
+
{
|
| 2037 |
+
"epoch": 0.023017442905952167,
|
| 2038 |
+
"grad_norm": 12.684931755065918,
|
| 2039 |
+
"learning_rate": 0.0001999766391236335,
|
| 2040 |
+
"loss": 6.5754,
|
| 2041 |
+
"step": 288
|
| 2042 |
+
},
|
| 2043 |
+
{
|
| 2044 |
+
"epoch": 0.023097364582708946,
|
| 2045 |
+
"grad_norm": 15.125930786132812,
|
| 2046 |
+
"learning_rate": 0.00019997646450065076,
|
| 2047 |
+
"loss": 5.6977,
|
| 2048 |
+
"step": 289
|
| 2049 |
+
},
|
| 2050 |
+
{
|
| 2051 |
+
"epoch": 0.023177286259465725,
|
| 2052 |
+
"grad_norm": 14.20514965057373,
|
| 2053 |
+
"learning_rate": 0.00019997628987766802,
|
| 2054 |
+
"loss": 6.033,
|
| 2055 |
+
"step": 290
|
| 2056 |
+
},
|
| 2057 |
+
{
|
| 2058 |
+
"epoch": 0.0232572079362225,
|
| 2059 |
+
"grad_norm": 16.646081924438477,
|
| 2060 |
+
"learning_rate": 0.00019997610070277005,
|
| 2061 |
+
"loss": 6.4136,
|
| 2062 |
+
"step": 291
|
| 2063 |
+
},
|
| 2064 |
+
{
|
| 2065 |
+
"epoch": 0.02333712961297928,
|
| 2066 |
+
"grad_norm": 15.947591781616211,
|
| 2067 |
+
"learning_rate": 0.00019997591152787209,
|
| 2068 |
+
"loss": 6.8916,
|
| 2069 |
+
"step": 292
|
| 2070 |
+
},
|
| 2071 |
+
{
|
| 2072 |
+
"epoch": 0.02341705128973606,
|
| 2073 |
+
"grad_norm": 15.278575897216797,
|
| 2074 |
+
"learning_rate": 0.00019997573690488935,
|
| 2075 |
+
"loss": 6.1326,
|
| 2076 |
+
"step": 293
|
| 2077 |
+
},
|
| 2078 |
+
{
|
| 2079 |
+
"epoch": 0.02349697296649284,
|
| 2080 |
+
"grad_norm": 14.641289710998535,
|
| 2081 |
+
"learning_rate": 0.00019997554772999138,
|
| 2082 |
+
"loss": 6.288,
|
| 2083 |
+
"step": 294
|
| 2084 |
+
},
|
| 2085 |
+
{
|
| 2086 |
+
"epoch": 0.023576894643249614,
|
| 2087 |
+
"grad_norm": 13.565072059631348,
|
| 2088 |
+
"learning_rate": 0.00019997537310700864,
|
| 2089 |
+
"loss": 6.8969,
|
| 2090 |
+
"step": 295
|
| 2091 |
+
},
|
| 2092 |
+
{
|
| 2093 |
+
"epoch": 0.023656816320006394,
|
| 2094 |
+
"grad_norm": 17.934261322021484,
|
| 2095 |
+
"learning_rate": 0.00019997516938019544,
|
| 2096 |
+
"loss": 5.7942,
|
| 2097 |
+
"step": 296
|
| 2098 |
+
},
|
| 2099 |
+
{
|
| 2100 |
+
"epoch": 0.023736737996763173,
|
| 2101 |
+
"grad_norm": 19.171415328979492,
|
| 2102 |
+
"learning_rate": 0.00019997498020529747,
|
| 2103 |
+
"loss": 6.8676,
|
| 2104 |
+
"step": 297
|
| 2105 |
+
},
|
| 2106 |
+
{
|
| 2107 |
+
"epoch": 0.023816659673519952,
|
| 2108 |
+
"grad_norm": 13.919241905212402,
|
| 2109 |
+
"learning_rate": 0.0001999747910303995,
|
| 2110 |
+
"loss": 5.5413,
|
| 2111 |
+
"step": 298
|
| 2112 |
+
},
|
| 2113 |
+
{
|
| 2114 |
+
"epoch": 0.023896581350276728,
|
| 2115 |
+
"grad_norm": 18.490270614624023,
|
| 2116 |
+
"learning_rate": 0.00019997461640741676,
|
| 2117 |
+
"loss": 6.2958,
|
| 2118 |
+
"step": 299
|
| 2119 |
+
},
|
| 2120 |
+
{
|
| 2121 |
+
"epoch": 0.023976503027033507,
|
| 2122 |
+
"grad_norm": 16.66715431213379,
|
| 2123 |
+
"learning_rate": 0.0001999744272325188,
|
| 2124 |
+
"loss": 7.6855,
|
| 2125 |
+
"step": 300
|
| 2126 |
+
},
|
| 2127 |
+
{
|
| 2128 |
+
"epoch": 0.024056424703790286,
|
| 2129 |
+
"grad_norm": 16.368955612182617,
|
| 2130 |
+
"learning_rate": 0.00019997423805762082,
|
| 2131 |
+
"loss": 6.3887,
|
| 2132 |
+
"step": 301
|
| 2133 |
+
},
|
| 2134 |
+
{
|
| 2135 |
+
"epoch": 0.024136346380547066,
|
| 2136 |
+
"grad_norm": 21.846345901489258,
|
| 2137 |
+
"learning_rate": 0.00019997404888272285,
|
| 2138 |
+
"loss": 6.3832,
|
| 2139 |
+
"step": 302
|
| 2140 |
+
},
|
| 2141 |
+
{
|
| 2142 |
+
"epoch": 0.02421626805730384,
|
| 2143 |
+
"grad_norm": 12.367501258850098,
|
| 2144 |
+
"learning_rate": 0.00019997384515590966,
|
| 2145 |
+
"loss": 5.2022,
|
| 2146 |
+
"step": 303
|
| 2147 |
+
},
|
| 2148 |
+
{
|
| 2149 |
+
"epoch": 0.02429618973406062,
|
| 2150 |
+
"grad_norm": 18.50084686279297,
|
| 2151 |
+
"learning_rate": 0.0001999736559810117,
|
| 2152 |
+
"loss": 5.6139,
|
| 2153 |
+
"step": 304
|
| 2154 |
+
},
|
| 2155 |
+
{
|
| 2156 |
+
"epoch": 0.0243761114108174,
|
| 2157 |
+
"grad_norm": 15.343425750732422,
|
| 2158 |
+
"learning_rate": 0.00019997346680611372,
|
| 2159 |
+
"loss": 6.2095,
|
| 2160 |
+
"step": 305
|
| 2161 |
+
},
|
| 2162 |
+
{
|
| 2163 |
+
"epoch": 0.024456033087574176,
|
| 2164 |
+
"grad_norm": 12.259312629699707,
|
| 2165 |
+
"learning_rate": 0.00019997326307930052,
|
| 2166 |
+
"loss": 5.9394,
|
| 2167 |
+
"step": 306
|
| 2168 |
+
},
|
| 2169 |
+
{
|
| 2170 |
+
"epoch": 0.024535954764330955,
|
| 2171 |
+
"grad_norm": 15.81058406829834,
|
| 2172 |
+
"learning_rate": 0.00019997307390440255,
|
| 2173 |
+
"loss": 7.1533,
|
| 2174 |
+
"step": 307
|
| 2175 |
+
},
|
| 2176 |
+
{
|
| 2177 |
+
"epoch": 0.024615876441087734,
|
| 2178 |
+
"grad_norm": 22.787445068359375,
|
| 2179 |
+
"learning_rate": 0.00019997288472950459,
|
| 2180 |
+
"loss": 5.6197,
|
| 2181 |
+
"step": 308
|
| 2182 |
+
},
|
| 2183 |
+
{
|
| 2184 |
+
"epoch": 0.024695798117844513,
|
| 2185 |
+
"grad_norm": 16.70743751525879,
|
| 2186 |
+
"learning_rate": 0.0001999726810026914,
|
| 2187 |
+
"loss": 6.4227,
|
| 2188 |
+
"step": 309
|
| 2189 |
+
},
|
| 2190 |
+
{
|
| 2191 |
+
"epoch": 0.02477571979460129,
|
| 2192 |
+
"grad_norm": 16.33761215209961,
|
| 2193 |
+
"learning_rate": 0.00019997249182779342,
|
| 2194 |
+
"loss": 6.4147,
|
| 2195 |
+
"step": 310
|
| 2196 |
+
},
|
| 2197 |
+
{
|
| 2198 |
+
"epoch": 0.02485564147135807,
|
| 2199 |
+
"grad_norm": 12.479717254638672,
|
| 2200 |
+
"learning_rate": 0.00019997228810098022,
|
| 2201 |
+
"loss": 4.9899,
|
| 2202 |
+
"step": 311
|
| 2203 |
+
},
|
| 2204 |
+
{
|
| 2205 |
+
"epoch": 0.024935563148114848,
|
| 2206 |
+
"grad_norm": 13.511429786682129,
|
| 2207 |
+
"learning_rate": 0.00019997209892608225,
|
| 2208 |
+
"loss": 6.093,
|
| 2209 |
+
"step": 312
|
| 2210 |
+
},
|
| 2211 |
+
{
|
| 2212 |
+
"epoch": 0.025015484824871627,
|
| 2213 |
+
"grad_norm": 14.443562507629395,
|
| 2214 |
+
"learning_rate": 0.00019997189519926906,
|
| 2215 |
+
"loss": 6.141,
|
| 2216 |
+
"step": 313
|
| 2217 |
+
},
|
| 2218 |
+
{
|
| 2219 |
+
"epoch": 0.025095406501628403,
|
| 2220 |
+
"grad_norm": 15.19046401977539,
|
| 2221 |
+
"learning_rate": 0.0001999717060243711,
|
| 2222 |
+
"loss": 5.7764,
|
| 2223 |
+
"step": 314
|
| 2224 |
+
},
|
| 2225 |
+
{
|
| 2226 |
+
"epoch": 0.025175328178385182,
|
| 2227 |
+
"grad_norm": 17.57716178894043,
|
| 2228 |
+
"learning_rate": 0.0001999715022975579,
|
| 2229 |
+
"loss": 6.1878,
|
| 2230 |
+
"step": 315
|
| 2231 |
+
},
|
| 2232 |
+
{
|
| 2233 |
+
"epoch": 0.02525524985514196,
|
| 2234 |
+
"grad_norm": 12.924726486206055,
|
| 2235 |
+
"learning_rate": 0.0001999712985707447,
|
| 2236 |
+
"loss": 5.5529,
|
| 2237 |
+
"step": 316
|
| 2238 |
+
},
|
| 2239 |
+
{
|
| 2240 |
+
"epoch": 0.02533517153189874,
|
| 2241 |
+
"grad_norm": 17.26683235168457,
|
| 2242 |
+
"learning_rate": 0.0001999710948439315,
|
| 2243 |
+
"loss": 6.1933,
|
| 2244 |
+
"step": 317
|
| 2245 |
+
},
|
| 2246 |
+
{
|
| 2247 |
+
"epoch": 0.025415093208655516,
|
| 2248 |
+
"grad_norm": 11.450511932373047,
|
| 2249 |
+
"learning_rate": 0.0001999708911171183,
|
| 2250 |
+
"loss": 5.5608,
|
| 2251 |
+
"step": 318
|
| 2252 |
+
},
|
| 2253 |
+
{
|
| 2254 |
+
"epoch": 0.025495014885412295,
|
| 2255 |
+
"grad_norm": 16.322240829467773,
|
| 2256 |
+
"learning_rate": 0.0001999706873903051,
|
| 2257 |
+
"loss": 6.6088,
|
| 2258 |
+
"step": 319
|
| 2259 |
+
},
|
| 2260 |
+
{
|
| 2261 |
+
"epoch": 0.025574936562169075,
|
| 2262 |
+
"grad_norm": 14.587623596191406,
|
| 2263 |
+
"learning_rate": 0.0001999704836634919,
|
| 2264 |
+
"loss": 5.4771,
|
| 2265 |
+
"step": 320
|
| 2266 |
+
},
|
| 2267 |
+
{
|
| 2268 |
+
"epoch": 0.025654858238925854,
|
| 2269 |
+
"grad_norm": 13.803770065307617,
|
| 2270 |
+
"learning_rate": 0.00019997029448859394,
|
| 2271 |
+
"loss": 6.1518,
|
| 2272 |
+
"step": 321
|
| 2273 |
+
},
|
| 2274 |
+
{
|
| 2275 |
+
"epoch": 0.02573477991568263,
|
| 2276 |
+
"grad_norm": 17.529619216918945,
|
| 2277 |
+
"learning_rate": 0.00019997009076178074,
|
| 2278 |
+
"loss": 6.3047,
|
| 2279 |
+
"step": 322
|
| 2280 |
+
},
|
| 2281 |
+
{
|
| 2282 |
+
"epoch": 0.02581470159243941,
|
| 2283 |
+
"grad_norm": 18.123512268066406,
|
| 2284 |
+
"learning_rate": 0.00019996988703496754,
|
| 2285 |
+
"loss": 6.0607,
|
| 2286 |
+
"step": 323
|
| 2287 |
+
},
|
| 2288 |
+
{
|
| 2289 |
+
"epoch": 0.025894623269196188,
|
| 2290 |
+
"grad_norm": 15.56558895111084,
|
| 2291 |
+
"learning_rate": 0.00019996968330815434,
|
| 2292 |
+
"loss": 7.5001,
|
| 2293 |
+
"step": 324
|
| 2294 |
+
},
|
| 2295 |
+
{
|
| 2296 |
+
"epoch": 0.025974544945952967,
|
| 2297 |
+
"grad_norm": 19.938735961914062,
|
| 2298 |
+
"learning_rate": 0.00019996946502942592,
|
| 2299 |
+
"loss": 5.6603,
|
| 2300 |
+
"step": 325
|
| 2301 |
+
},
|
| 2302 |
+
{
|
| 2303 |
+
"epoch": 0.026054466622709743,
|
| 2304 |
+
"grad_norm": 12.854557991027832,
|
| 2305 |
+
"learning_rate": 0.00019996926130261272,
|
| 2306 |
+
"loss": 4.9226,
|
| 2307 |
+
"step": 326
|
| 2308 |
+
},
|
| 2309 |
+
{
|
| 2310 |
+
"epoch": 0.026134388299466522,
|
| 2311 |
+
"grad_norm": 13.187103271484375,
|
| 2312 |
+
"learning_rate": 0.0001999690430238843,
|
| 2313 |
+
"loss": 6.4182,
|
| 2314 |
+
"step": 327
|
| 2315 |
+
},
|
| 2316 |
+
{
|
| 2317 |
+
"epoch": 0.0262143099762233,
|
| 2318 |
+
"grad_norm": 16.57967758178711,
|
| 2319 |
+
"learning_rate": 0.0001999688392970711,
|
| 2320 |
+
"loss": 5.4498,
|
| 2321 |
+
"step": 328
|
| 2322 |
+
},
|
| 2323 |
+
{
|
| 2324 |
+
"epoch": 0.02629423165298008,
|
| 2325 |
+
"grad_norm": 13.449827194213867,
|
| 2326 |
+
"learning_rate": 0.0001999686355702579,
|
| 2327 |
+
"loss": 5.742,
|
| 2328 |
+
"step": 329
|
| 2329 |
+
},
|
| 2330 |
+
{
|
| 2331 |
+
"epoch": 0.026374153329736857,
|
| 2332 |
+
"grad_norm": 12.401226997375488,
|
| 2333 |
+
"learning_rate": 0.0001999684318434447,
|
| 2334 |
+
"loss": 7.1152,
|
| 2335 |
+
"step": 330
|
| 2336 |
+
},
|
| 2337 |
+
{
|
| 2338 |
+
"epoch": 0.026454075006493636,
|
| 2339 |
+
"grad_norm": 14.42686939239502,
|
| 2340 |
+
"learning_rate": 0.00019996821356471628,
|
| 2341 |
+
"loss": 5.869,
|
| 2342 |
+
"step": 331
|
| 2343 |
+
},
|
| 2344 |
+
{
|
| 2345 |
+
"epoch": 0.026533996683250415,
|
| 2346 |
+
"grad_norm": 15.052666664123535,
|
| 2347 |
+
"learning_rate": 0.00019996799528598785,
|
| 2348 |
+
"loss": 6.2315,
|
| 2349 |
+
"step": 332
|
| 2350 |
+
},
|
| 2351 |
+
{
|
| 2352 |
+
"epoch": 0.026613918360007194,
|
| 2353 |
+
"grad_norm": 17.430248260498047,
|
| 2354 |
+
"learning_rate": 0.00019996779155917466,
|
| 2355 |
+
"loss": 6.361,
|
| 2356 |
+
"step": 333
|
| 2357 |
+
},
|
| 2358 |
+
{
|
| 2359 |
+
"epoch": 0.02669384003676397,
|
| 2360 |
+
"grad_norm": 13.0830659866333,
|
| 2361 |
+
"learning_rate": 0.00019996758783236146,
|
| 2362 |
+
"loss": 5.9668,
|
| 2363 |
+
"step": 334
|
| 2364 |
+
},
|
| 2365 |
+
{
|
| 2366 |
+
"epoch": 0.02677376171352075,
|
| 2367 |
+
"grad_norm": 12.437070846557617,
|
| 2368 |
+
"learning_rate": 0.00019996736955363303,
|
| 2369 |
+
"loss": 5.1858,
|
| 2370 |
+
"step": 335
|
| 2371 |
+
},
|
| 2372 |
+
{
|
| 2373 |
+
"epoch": 0.02685368339027753,
|
| 2374 |
+
"grad_norm": 9.951759338378906,
|
| 2375 |
+
"learning_rate": 0.0001999671512749046,
|
| 2376 |
+
"loss": 6.8424,
|
| 2377 |
+
"step": 336
|
| 2378 |
+
},
|
| 2379 |
+
{
|
| 2380 |
+
"epoch": 0.026933605067034308,
|
| 2381 |
+
"grad_norm": 12.171687126159668,
|
| 2382 |
+
"learning_rate": 0.00019996693299617618,
|
| 2383 |
+
"loss": 6.4313,
|
| 2384 |
+
"step": 337
|
| 2385 |
+
},
|
| 2386 |
+
{
|
| 2387 |
+
"epoch": 0.027013526743791084,
|
| 2388 |
+
"grad_norm": 11.820364952087402,
|
| 2389 |
+
"learning_rate": 0.00019996672926936299,
|
| 2390 |
+
"loss": 4.9972,
|
| 2391 |
+
"step": 338
|
| 2392 |
+
},
|
| 2393 |
+
{
|
| 2394 |
+
"epoch": 0.027093448420547863,
|
| 2395 |
+
"grad_norm": 13.975296974182129,
|
| 2396 |
+
"learning_rate": 0.00019996649643871933,
|
| 2397 |
+
"loss": 7.0921,
|
| 2398 |
+
"step": 339
|
| 2399 |
+
},
|
| 2400 |
+
{
|
| 2401 |
+
"epoch": 0.027173370097304642,
|
| 2402 |
+
"grad_norm": 13.744132995605469,
|
| 2403 |
+
"learning_rate": 0.0001999662781599909,
|
| 2404 |
+
"loss": 6.0032,
|
| 2405 |
+
"step": 340
|
| 2406 |
+
},
|
| 2407 |
+
{
|
| 2408 |
+
"epoch": 0.02725329177406142,
|
| 2409 |
+
"grad_norm": 21.694072723388672,
|
| 2410 |
+
"learning_rate": 0.0001999660744331777,
|
| 2411 |
+
"loss": 5.7393,
|
| 2412 |
+
"step": 341
|
| 2413 |
+
},
|
| 2414 |
+
{
|
| 2415 |
+
"epoch": 0.027333213450818197,
|
| 2416 |
+
"grad_norm": 13.287113189697266,
|
| 2417 |
+
"learning_rate": 0.00019996585615444928,
|
| 2418 |
+
"loss": 6.1796,
|
| 2419 |
+
"step": 342
|
| 2420 |
+
},
|
| 2421 |
+
{
|
| 2422 |
+
"epoch": 0.027413135127574977,
|
| 2423 |
+
"grad_norm": 24.737030029296875,
|
| 2424 |
+
"learning_rate": 0.00019996562332380563,
|
| 2425 |
+
"loss": 7.4829,
|
| 2426 |
+
"step": 343
|
| 2427 |
+
},
|
| 2428 |
+
{
|
| 2429 |
+
"epoch": 0.027493056804331756,
|
| 2430 |
+
"grad_norm": 13.299663543701172,
|
| 2431 |
+
"learning_rate": 0.00019996541959699243,
|
| 2432 |
+
"loss": 5.4366,
|
| 2433 |
+
"step": 344
|
| 2434 |
+
},
|
| 2435 |
+
{
|
| 2436 |
+
"epoch": 0.02757297848108853,
|
| 2437 |
+
"grad_norm": 16.67266082763672,
|
| 2438 |
+
"learning_rate": 0.00019996518676634878,
|
| 2439 |
+
"loss": 6.7686,
|
| 2440 |
+
"step": 345
|
| 2441 |
+
},
|
| 2442 |
+
{
|
| 2443 |
+
"epoch": 0.02765290015784531,
|
| 2444 |
+
"grad_norm": 9.903456687927246,
|
| 2445 |
+
"learning_rate": 0.00019996496848762035,
|
| 2446 |
+
"loss": 5.6142,
|
| 2447 |
+
"step": 346
|
| 2448 |
+
},
|
| 2449 |
+
{
|
| 2450 |
+
"epoch": 0.02773282183460209,
|
| 2451 |
+
"grad_norm": 15.513213157653809,
|
| 2452 |
+
"learning_rate": 0.00019996475020889193,
|
| 2453 |
+
"loss": 5.1707,
|
| 2454 |
+
"step": 347
|
| 2455 |
+
},
|
| 2456 |
+
{
|
| 2457 |
+
"epoch": 0.02781274351135887,
|
| 2458 |
+
"grad_norm": 14.624857902526855,
|
| 2459 |
+
"learning_rate": 0.00019996451737824827,
|
| 2460 |
+
"loss": 6.4379,
|
| 2461 |
+
"step": 348
|
| 2462 |
+
},
|
| 2463 |
+
{
|
| 2464 |
+
"epoch": 0.027892665188115645,
|
| 2465 |
+
"grad_norm": 17.85902976989746,
|
| 2466 |
+
"learning_rate": 0.00019996429909951985,
|
| 2467 |
+
"loss": 6.0558,
|
| 2468 |
+
"step": 349
|
| 2469 |
+
},
|
| 2470 |
+
{
|
| 2471 |
+
"epoch": 0.027972586864872424,
|
| 2472 |
+
"grad_norm": 19.93682098388672,
|
| 2473 |
+
"learning_rate": 0.00019996408082079142,
|
| 2474 |
+
"loss": 6.3928,
|
| 2475 |
+
"step": 350
|
| 2476 |
+
},
|
| 2477 |
+
{
|
| 2478 |
+
"epoch": 0.028052508541629204,
|
| 2479 |
+
"grad_norm": 12.33932113647461,
|
| 2480 |
+
"learning_rate": 0.00019996384799014777,
|
| 2481 |
+
"loss": 5.8751,
|
| 2482 |
+
"step": 351
|
| 2483 |
+
},
|
| 2484 |
+
{
|
| 2485 |
+
"epoch": 0.028132430218385983,
|
| 2486 |
+
"grad_norm": 15.817342758178711,
|
| 2487 |
+
"learning_rate": 0.00019996362971141934,
|
| 2488 |
+
"loss": 6.8256,
|
| 2489 |
+
"step": 352
|
| 2490 |
+
},
|
| 2491 |
+
{
|
| 2492 |
+
"epoch": 0.02821235189514276,
|
| 2493 |
+
"grad_norm": 15.714959144592285,
|
| 2494 |
+
"learning_rate": 0.0001999633968807757,
|
| 2495 |
+
"loss": 7.1489,
|
| 2496 |
+
"step": 353
|
| 2497 |
+
},
|
| 2498 |
+
{
|
| 2499 |
+
"epoch": 0.028292273571899538,
|
| 2500 |
+
"grad_norm": 12.748688697814941,
|
| 2501 |
+
"learning_rate": 0.00019996316405013204,
|
| 2502 |
+
"loss": 5.0708,
|
| 2503 |
+
"step": 354
|
| 2504 |
+
},
|
| 2505 |
+
{
|
| 2506 |
+
"epoch": 0.028372195248656317,
|
| 2507 |
+
"grad_norm": 10.243889808654785,
|
| 2508 |
+
"learning_rate": 0.0001999629457714036,
|
| 2509 |
+
"loss": 4.5773,
|
| 2510 |
+
"step": 355
|
| 2511 |
+
},
|
| 2512 |
+
{
|
| 2513 |
+
"epoch": 0.028452116925413096,
|
| 2514 |
+
"grad_norm": 16.02983856201172,
|
| 2515 |
+
"learning_rate": 0.00019996271294075996,
|
| 2516 |
+
"loss": 6.7184,
|
| 2517 |
+
"step": 356
|
| 2518 |
+
},
|
| 2519 |
+
{
|
| 2520 |
+
"epoch": 0.028532038602169872,
|
| 2521 |
+
"grad_norm": 15.122602462768555,
|
| 2522 |
+
"learning_rate": 0.00019996249466203153,
|
| 2523 |
+
"loss": 5.815,
|
| 2524 |
+
"step": 357
|
| 2525 |
+
},
|
| 2526 |
+
{
|
| 2527 |
+
"epoch": 0.02861196027892665,
|
| 2528 |
+
"grad_norm": 16.68320083618164,
|
| 2529 |
+
"learning_rate": 0.00019996224727947265,
|
| 2530 |
+
"loss": 6.4434,
|
| 2531 |
+
"step": 358
|
| 2532 |
+
},
|
| 2533 |
+
{
|
| 2534 |
+
"epoch": 0.02869188195568343,
|
| 2535 |
+
"grad_norm": 15.334043502807617,
|
| 2536 |
+
"learning_rate": 0.00019996202900074422,
|
| 2537 |
+
"loss": 7.1437,
|
| 2538 |
+
"step": 359
|
| 2539 |
+
},
|
| 2540 |
+
{
|
| 2541 |
+
"epoch": 0.02877180363244021,
|
| 2542 |
+
"grad_norm": 10.798954963684082,
|
| 2543 |
+
"learning_rate": 0.00019996179617010057,
|
| 2544 |
+
"loss": 6.3215,
|
| 2545 |
+
"step": 360
|
| 2546 |
+
},
|
| 2547 |
+
{
|
| 2548 |
+
"epoch": 0.028851725309196986,
|
| 2549 |
+
"grad_norm": 12.362552642822266,
|
| 2550 |
+
"learning_rate": 0.00019996156333945692,
|
| 2551 |
+
"loss": 5.2076,
|
| 2552 |
+
"step": 361
|
| 2553 |
+
},
|
| 2554 |
+
{
|
| 2555 |
+
"epoch": 0.028931646985953765,
|
| 2556 |
+
"grad_norm": 19.123573303222656,
|
| 2557 |
+
"learning_rate": 0.00019996133050881326,
|
| 2558 |
+
"loss": 6.0658,
|
| 2559 |
+
"step": 362
|
| 2560 |
+
},
|
| 2561 |
+
{
|
| 2562 |
+
"epoch": 0.029011568662710544,
|
| 2563 |
+
"grad_norm": 14.757805824279785,
|
| 2564 |
+
"learning_rate": 0.0001999610976781696,
|
| 2565 |
+
"loss": 5.3874,
|
| 2566 |
+
"step": 363
|
| 2567 |
+
},
|
| 2568 |
+
{
|
| 2569 |
+
"epoch": 0.029091490339467323,
|
| 2570 |
+
"grad_norm": 17.045217514038086,
|
| 2571 |
+
"learning_rate": 0.00019996086484752595,
|
| 2572 |
+
"loss": 7.0085,
|
| 2573 |
+
"step": 364
|
| 2574 |
+
},
|
| 2575 |
+
{
|
| 2576 |
+
"epoch": 0.0291714120162241,
|
| 2577 |
+
"grad_norm": 16.436168670654297,
|
| 2578 |
+
"learning_rate": 0.0001999606320168823,
|
| 2579 |
+
"loss": 6.1691,
|
| 2580 |
+
"step": 365
|
| 2581 |
+
},
|
| 2582 |
+
{
|
| 2583 |
+
"epoch": 0.02925133369298088,
|
| 2584 |
+
"grad_norm": 17.719770431518555,
|
| 2585 |
+
"learning_rate": 0.00019996039918623865,
|
| 2586 |
+
"loss": 5.9578,
|
| 2587 |
+
"step": 366
|
| 2588 |
+
},
|
| 2589 |
+
{
|
| 2590 |
+
"epoch": 0.029331255369737658,
|
| 2591 |
+
"grad_norm": 15.637602806091309,
|
| 2592 |
+
"learning_rate": 0.00019996015180367976,
|
| 2593 |
+
"loss": 6.3728,
|
| 2594 |
+
"step": 367
|
| 2595 |
+
},
|
| 2596 |
+
{
|
| 2597 |
+
"epoch": 0.029411177046494437,
|
| 2598 |
+
"grad_norm": 16.410329818725586,
|
| 2599 |
+
"learning_rate": 0.0001999599189730361,
|
| 2600 |
+
"loss": 5.4258,
|
| 2601 |
+
"step": 368
|
| 2602 |
+
},
|
| 2603 |
+
{
|
| 2604 |
+
"epoch": 0.029491098723251213,
|
| 2605 |
+
"grad_norm": 10.284132957458496,
|
| 2606 |
+
"learning_rate": 0.00019995967159047723,
|
| 2607 |
+
"loss": 6.6986,
|
| 2608 |
+
"step": 369
|
| 2609 |
+
},
|
| 2610 |
+
{
|
| 2611 |
+
"epoch": 0.029571020400007992,
|
| 2612 |
+
"grad_norm": 14.195467948913574,
|
| 2613 |
+
"learning_rate": 0.00019995943875983357,
|
| 2614 |
+
"loss": 5.2009,
|
| 2615 |
+
"step": 370
|
| 2616 |
+
},
|
| 2617 |
+
{
|
| 2618 |
+
"epoch": 0.02965094207676477,
|
| 2619 |
+
"grad_norm": 11.96532917022705,
|
| 2620 |
+
"learning_rate": 0.00019995920592918992,
|
| 2621 |
+
"loss": 5.4514,
|
| 2622 |
+
"step": 371
|
| 2623 |
+
},
|
| 2624 |
+
{
|
| 2625 |
+
"epoch": 0.02973086375352155,
|
| 2626 |
+
"grad_norm": 11.906566619873047,
|
| 2627 |
+
"learning_rate": 0.00019995895854663104,
|
| 2628 |
+
"loss": 6.0655,
|
| 2629 |
+
"step": 372
|
| 2630 |
+
},
|
| 2631 |
+
{
|
| 2632 |
+
"epoch": 0.029810785430278326,
|
| 2633 |
+
"grad_norm": 13.493677139282227,
|
| 2634 |
+
"learning_rate": 0.00019995872571598738,
|
| 2635 |
+
"loss": 6.6973,
|
| 2636 |
+
"step": 373
|
| 2637 |
+
},
|
| 2638 |
+
{
|
| 2639 |
+
"epoch": 0.029890707107035105,
|
| 2640 |
+
"grad_norm": 17.246294021606445,
|
| 2641 |
+
"learning_rate": 0.00019995849288534373,
|
| 2642 |
+
"loss": 6.0919,
|
| 2643 |
+
"step": 374
|
| 2644 |
+
},
|
| 2645 |
+
{
|
| 2646 |
+
"epoch": 0.029970628783791885,
|
| 2647 |
+
"grad_norm": 16.358449935913086,
|
| 2648 |
+
"learning_rate": 0.00019995824550278485,
|
| 2649 |
+
"loss": 6.303,
|
| 2650 |
+
"step": 375
|
| 2651 |
+
},
|
| 2652 |
+
{
|
| 2653 |
+
"epoch": 0.030050550460548664,
|
| 2654 |
+
"grad_norm": 14.359763145446777,
|
| 2655 |
+
"learning_rate": 0.00019995799812022597,
|
| 2656 |
+
"loss": 5.8685,
|
| 2657 |
+
"step": 376
|
| 2658 |
+
},
|
| 2659 |
+
{
|
| 2660 |
+
"epoch": 0.03013047213730544,
|
| 2661 |
+
"grad_norm": 15.526153564453125,
|
| 2662 |
+
"learning_rate": 0.0001999577652895823,
|
| 2663 |
+
"loss": 6.4071,
|
| 2664 |
+
"step": 377
|
| 2665 |
+
},
|
| 2666 |
+
{
|
| 2667 |
+
"epoch": 0.03021039381406222,
|
| 2668 |
+
"grad_norm": 19.989709854125977,
|
| 2669 |
+
"learning_rate": 0.0001999575033551082,
|
| 2670 |
+
"loss": 5.7926,
|
| 2671 |
+
"step": 378
|
| 2672 |
+
},
|
| 2673 |
+
{
|
| 2674 |
+
"epoch": 0.030290315490818998,
|
| 2675 |
+
"grad_norm": 15.332075119018555,
|
| 2676 |
+
"learning_rate": 0.00019995727052446455,
|
| 2677 |
+
"loss": 6.5446,
|
| 2678 |
+
"step": 379
|
| 2679 |
+
},
|
| 2680 |
+
{
|
| 2681 |
+
"epoch": 0.030370237167575774,
|
| 2682 |
+
"grad_norm": 15.886347770690918,
|
| 2683 |
+
"learning_rate": 0.00019995702314190567,
|
| 2684 |
+
"loss": 5.7387,
|
| 2685 |
+
"step": 380
|
| 2686 |
+
},
|
| 2687 |
+
{
|
| 2688 |
+
"epoch": 0.030450158844332553,
|
| 2689 |
+
"grad_norm": 15.453957557678223,
|
| 2690 |
+
"learning_rate": 0.00019995677575934678,
|
| 2691 |
+
"loss": 5.9044,
|
| 2692 |
+
"step": 381
|
| 2693 |
+
},
|
| 2694 |
+
{
|
| 2695 |
+
"epoch": 0.030530080521089333,
|
| 2696 |
+
"grad_norm": 14.168845176696777,
|
| 2697 |
+
"learning_rate": 0.0001999565283767879,
|
| 2698 |
+
"loss": 5.3228,
|
| 2699 |
+
"step": 382
|
| 2700 |
+
},
|
| 2701 |
+
{
|
| 2702 |
+
"epoch": 0.030610002197846112,
|
| 2703 |
+
"grad_norm": 16.134654998779297,
|
| 2704 |
+
"learning_rate": 0.00019995628099422902,
|
| 2705 |
+
"loss": 6.2229,
|
| 2706 |
+
"step": 383
|
| 2707 |
+
},
|
| 2708 |
+
{
|
| 2709 |
+
"epoch": 0.030689923874602888,
|
| 2710 |
+
"grad_norm": 17.78655242919922,
|
| 2711 |
+
"learning_rate": 0.00019995603361167014,
|
| 2712 |
+
"loss": 7.0895,
|
| 2713 |
+
"step": 384
|
| 2714 |
+
},
|
| 2715 |
+
{
|
| 2716 |
+
"epoch": 0.030769845551359667,
|
| 2717 |
+
"grad_norm": 15.200431823730469,
|
| 2718 |
+
"learning_rate": 0.00019995578622911125,
|
| 2719 |
+
"loss": 6.4011,
|
| 2720 |
+
"step": 385
|
| 2721 |
+
},
|
| 2722 |
+
{
|
| 2723 |
+
"epoch": 0.030849767228116446,
|
| 2724 |
+
"grad_norm": 16.652729034423828,
|
| 2725 |
+
"learning_rate": 0.00019995553884655237,
|
| 2726 |
+
"loss": 5.9991,
|
| 2727 |
+
"step": 386
|
| 2728 |
+
},
|
| 2729 |
+
{
|
| 2730 |
+
"epoch": 0.030929688904873225,
|
| 2731 |
+
"grad_norm": 18.908794403076172,
|
| 2732 |
+
"learning_rate": 0.0001999552914639935,
|
| 2733 |
+
"loss": 5.4947,
|
| 2734 |
+
"step": 387
|
| 2735 |
+
},
|
| 2736 |
+
{
|
| 2737 |
+
"epoch": 0.03100961058163,
|
| 2738 |
+
"grad_norm": 22.241044998168945,
|
| 2739 |
+
"learning_rate": 0.0001999550440814346,
|
| 2740 |
+
"loss": 5.9547,
|
| 2741 |
+
"step": 388
|
| 2742 |
+
},
|
| 2743 |
+
{
|
| 2744 |
+
"epoch": 0.03108953225838678,
|
| 2745 |
+
"grad_norm": 15.065712928771973,
|
| 2746 |
+
"learning_rate": 0.00019995479669887573,
|
| 2747 |
+
"loss": 5.0218,
|
| 2748 |
+
"step": 389
|
| 2749 |
+
},
|
| 2750 |
+
{
|
| 2751 |
+
"epoch": 0.03116945393514356,
|
| 2752 |
+
"grad_norm": 16.05895233154297,
|
| 2753 |
+
"learning_rate": 0.00019995453476440161,
|
| 2754 |
+
"loss": 5.0685,
|
| 2755 |
+
"step": 390
|
| 2756 |
+
},
|
| 2757 |
+
{
|
| 2758 |
+
"epoch": 0.03124937561190034,
|
| 2759 |
+
"grad_norm": 16.954740524291992,
|
| 2760 |
+
"learning_rate": 0.0001999542728299275,
|
| 2761 |
+
"loss": 6.0334,
|
| 2762 |
+
"step": 391
|
| 2763 |
+
},
|
| 2764 |
+
{
|
| 2765 |
+
"epoch": 0.03132929728865712,
|
| 2766 |
+
"grad_norm": 14.711755752563477,
|
| 2767 |
+
"learning_rate": 0.00019995402544736862,
|
| 2768 |
+
"loss": 6.3458,
|
| 2769 |
+
"step": 392
|
| 2770 |
+
},
|
| 2771 |
+
{
|
| 2772 |
+
"epoch": 0.0314092189654139,
|
| 2773 |
+
"grad_norm": 20.665563583374023,
|
| 2774 |
+
"learning_rate": 0.00019995377806480974,
|
| 2775 |
+
"loss": 6.4642,
|
| 2776 |
+
"step": 393
|
| 2777 |
+
},
|
| 2778 |
+
{
|
| 2779 |
+
"epoch": 0.03148914064217067,
|
| 2780 |
+
"grad_norm": 15.423831939697266,
|
| 2781 |
+
"learning_rate": 0.00019995351613033563,
|
| 2782 |
+
"loss": 6.2702,
|
| 2783 |
+
"step": 394
|
| 2784 |
+
},
|
| 2785 |
+
{
|
| 2786 |
+
"epoch": 0.03156906231892745,
|
| 2787 |
+
"grad_norm": 36.09846878051758,
|
| 2788 |
+
"learning_rate": 0.00019995326874777675,
|
| 2789 |
+
"loss": 6.5314,
|
| 2790 |
+
"step": 395
|
| 2791 |
+
},
|
| 2792 |
+
{
|
| 2793 |
+
"epoch": 0.03164898399568423,
|
| 2794 |
+
"grad_norm": 15.841190338134766,
|
| 2795 |
+
"learning_rate": 0.00019995300681330264,
|
| 2796 |
+
"loss": 6.4815,
|
| 2797 |
+
"step": 396
|
| 2798 |
+
},
|
| 2799 |
+
{
|
| 2800 |
+
"epoch": 0.03172890567244101,
|
| 2801 |
+
"grad_norm": 15.846356391906738,
|
| 2802 |
+
"learning_rate": 0.00019995275943074375,
|
| 2803 |
+
"loss": 7.0392,
|
| 2804 |
+
"step": 397
|
| 2805 |
+
},
|
| 2806 |
+
{
|
| 2807 |
+
"epoch": 0.03180882734919779,
|
| 2808 |
+
"grad_norm": 12.423648834228516,
|
| 2809 |
+
"learning_rate": 0.00019995248294435441,
|
| 2810 |
+
"loss": 5.5649,
|
| 2811 |
+
"step": 398
|
| 2812 |
+
},
|
| 2813 |
+
{
|
| 2814 |
+
"epoch": 0.031888749025954566,
|
| 2815 |
+
"grad_norm": 14.614909172058105,
|
| 2816 |
+
"learning_rate": 0.0001999522210098803,
|
| 2817 |
+
"loss": 6.1323,
|
| 2818 |
+
"step": 399
|
| 2819 |
+
},
|
| 2820 |
+
{
|
| 2821 |
+
"epoch": 0.031968670702711345,
|
| 2822 |
+
"grad_norm": 14.6503267288208,
|
| 2823 |
+
"learning_rate": 0.00019995197362732142,
|
| 2824 |
+
"loss": 5.7497,
|
| 2825 |
+
"step": 400
|
| 2826 |
+
},
|
| 2827 |
+
{
|
| 2828 |
+
"epoch": 0.031968670702711345,
|
| 2829 |
+
"eval_loss": 1.4913572072982788,
|
| 2830 |
+
"eval_runtime": 98.2768,
|
| 2831 |
+
"eval_samples_per_second": 5.118,
|
| 2832 |
+
"eval_steps_per_second": 2.564,
|
| 2833 |
+
"step": 400
|
| 2834 |
}
|
| 2835 |
],
|
| 2836 |
"logging_steps": 1,
|
|
|
|
| 2845 |
"early_stopping_threshold": 0.0
|
| 2846 |
},
|
| 2847 |
"attributes": {
|
| 2848 |
+
"early_stopping_patience_counter": 1
|
| 2849 |
}
|
| 2850 |
},
|
| 2851 |
"TrainerControl": {
|
|
|
|
| 2859 |
"attributes": {}
|
| 2860 |
}
|
| 2861 |
},
|
| 2862 |
+
"total_flos": 7.673695808834765e+16,
|
| 2863 |
"train_batch_size": 2,
|
| 2864 |
"trial_name": null,
|
| 2865 |
"trial_params": null
|