Training in progress, step 1800, 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 22573704
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:5b2771309b54cad22a6d12b43f2a88fbb5afc22c03f682444f479a7fce98561d
|
| 3 |
size 22573704
|
last-checkpoint/optimizer.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 17437818
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:a61685d18f810f6a4fed1647d54ba4b3b16ad699872c167b11c2edaa87c2eb01
|
| 3 |
size 17437818
|
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:ae96998c01a7ee338eee30377bf1e5f04258e6479e2af6c6075624f0f7738fe3
|
| 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:d8e1e886ba269f608335b2a2a84983b4f136968a26a27413a91ebbba4d3699d3
|
| 3 |
size 1064
|
last-checkpoint/trainer_state.json
CHANGED
|
@@ -1,9 +1,9 @@
|
|
| 1 |
{
|
| 2 |
"best_metric": 0.32298070192337036,
|
| 3 |
"best_model_checkpoint": "miner_id_24/checkpoint-1600",
|
| 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,
|
|
@@ -11279,6 +11279,1414 @@
|
|
| 11279 |
"eval_samples_per_second": 8.651,
|
| 11280 |
"eval_steps_per_second": 4.343,
|
| 11281 |
"step": 1600
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 11282 |
}
|
| 11283 |
],
|
| 11284 |
"logging_steps": 1,
|
|
@@ -11293,7 +12701,7 @@
|
|
| 11293 |
"early_stopping_threshold": 0.0
|
| 11294 |
},
|
| 11295 |
"attributes": {
|
| 11296 |
-
"early_stopping_patience_counter":
|
| 11297 |
}
|
| 11298 |
},
|
| 11299 |
"TrainerControl": {
|
|
@@ -11307,7 +12715,7 @@
|
|
| 11307 |
"attributes": {}
|
| 11308 |
}
|
| 11309 |
},
|
| 11310 |
-
"total_flos": 1.
|
| 11311 |
"train_batch_size": 2,
|
| 11312 |
"trial_name": null,
|
| 11313 |
"trial_params": null
|
|
|
|
| 1 |
{
|
| 2 |
"best_metric": 0.32298070192337036,
|
| 3 |
"best_model_checkpoint": "miner_id_24/checkpoint-1600",
|
| 4 |
+
"epoch": 0.05790713946773687,
|
| 5 |
"eval_steps": 200,
|
| 6 |
+
"global_step": 1800,
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
|
|
| 11279 |
"eval_samples_per_second": 8.651,
|
| 11280 |
"eval_steps_per_second": 4.343,
|
| 11281 |
"step": 1600
|
| 11282 |
+
},
|
| 11283 |
+
{
|
| 11284 |
+
"epoch": 0.05150518349324819,
|
| 11285 |
+
"grad_norm": 1.293373703956604,
|
| 11286 |
+
"learning_rate": 0.0002,
|
| 11287 |
+
"loss": 0.2867,
|
| 11288 |
+
"step": 1601
|
| 11289 |
+
},
|
| 11290 |
+
{
|
| 11291 |
+
"epoch": 0.05153735412628582,
|
| 11292 |
+
"grad_norm": 1.954478144645691,
|
| 11293 |
+
"learning_rate": 0.0002,
|
| 11294 |
+
"loss": 0.2713,
|
| 11295 |
+
"step": 1602
|
| 11296 |
+
},
|
| 11297 |
+
{
|
| 11298 |
+
"epoch": 0.05156952475932345,
|
| 11299 |
+
"grad_norm": 1.4226385354995728,
|
| 11300 |
+
"learning_rate": 0.0002,
|
| 11301 |
+
"loss": 0.3493,
|
| 11302 |
+
"step": 1603
|
| 11303 |
+
},
|
| 11304 |
+
{
|
| 11305 |
+
"epoch": 0.051601695392361084,
|
| 11306 |
+
"grad_norm": 1.4402284622192383,
|
| 11307 |
+
"learning_rate": 0.0002,
|
| 11308 |
+
"loss": 0.3488,
|
| 11309 |
+
"step": 1604
|
| 11310 |
+
},
|
| 11311 |
+
{
|
| 11312 |
+
"epoch": 0.05163386602539871,
|
| 11313 |
+
"grad_norm": 1.2787245512008667,
|
| 11314 |
+
"learning_rate": 0.0002,
|
| 11315 |
+
"loss": 0.3158,
|
| 11316 |
+
"step": 1605
|
| 11317 |
+
},
|
| 11318 |
+
{
|
| 11319 |
+
"epoch": 0.051666036658436346,
|
| 11320 |
+
"grad_norm": 1.3623970746994019,
|
| 11321 |
+
"learning_rate": 0.0002,
|
| 11322 |
+
"loss": 0.326,
|
| 11323 |
+
"step": 1606
|
| 11324 |
+
},
|
| 11325 |
+
{
|
| 11326 |
+
"epoch": 0.05169820729147398,
|
| 11327 |
+
"grad_norm": 1.5353353023529053,
|
| 11328 |
+
"learning_rate": 0.0002,
|
| 11329 |
+
"loss": 0.3247,
|
| 11330 |
+
"step": 1607
|
| 11331 |
+
},
|
| 11332 |
+
{
|
| 11333 |
+
"epoch": 0.05173037792451161,
|
| 11334 |
+
"grad_norm": 1.0746046304702759,
|
| 11335 |
+
"learning_rate": 0.0002,
|
| 11336 |
+
"loss": 0.276,
|
| 11337 |
+
"step": 1608
|
| 11338 |
+
},
|
| 11339 |
+
{
|
| 11340 |
+
"epoch": 0.05176254855754924,
|
| 11341 |
+
"grad_norm": 1.6327682733535767,
|
| 11342 |
+
"learning_rate": 0.0002,
|
| 11343 |
+
"loss": 0.4567,
|
| 11344 |
+
"step": 1609
|
| 11345 |
+
},
|
| 11346 |
+
{
|
| 11347 |
+
"epoch": 0.051794719190586876,
|
| 11348 |
+
"grad_norm": 1.284000039100647,
|
| 11349 |
+
"learning_rate": 0.0002,
|
| 11350 |
+
"loss": 0.2964,
|
| 11351 |
+
"step": 1610
|
| 11352 |
+
},
|
| 11353 |
+
{
|
| 11354 |
+
"epoch": 0.0518268898236245,
|
| 11355 |
+
"grad_norm": 1.4247043132781982,
|
| 11356 |
+
"learning_rate": 0.0002,
|
| 11357 |
+
"loss": 0.4059,
|
| 11358 |
+
"step": 1611
|
| 11359 |
+
},
|
| 11360 |
+
{
|
| 11361 |
+
"epoch": 0.05185906045666214,
|
| 11362 |
+
"grad_norm": 1.3963816165924072,
|
| 11363 |
+
"learning_rate": 0.0002,
|
| 11364 |
+
"loss": 0.2186,
|
| 11365 |
+
"step": 1612
|
| 11366 |
+
},
|
| 11367 |
+
{
|
| 11368 |
+
"epoch": 0.051891231089699764,
|
| 11369 |
+
"grad_norm": 1.5223742723464966,
|
| 11370 |
+
"learning_rate": 0.0002,
|
| 11371 |
+
"loss": 0.4245,
|
| 11372 |
+
"step": 1613
|
| 11373 |
+
},
|
| 11374 |
+
{
|
| 11375 |
+
"epoch": 0.0519234017227374,
|
| 11376 |
+
"grad_norm": 1.524560570716858,
|
| 11377 |
+
"learning_rate": 0.0002,
|
| 11378 |
+
"loss": 0.3594,
|
| 11379 |
+
"step": 1614
|
| 11380 |
+
},
|
| 11381 |
+
{
|
| 11382 |
+
"epoch": 0.05195557235577503,
|
| 11383 |
+
"grad_norm": 1.2883237600326538,
|
| 11384 |
+
"learning_rate": 0.0002,
|
| 11385 |
+
"loss": 0.3207,
|
| 11386 |
+
"step": 1615
|
| 11387 |
+
},
|
| 11388 |
+
{
|
| 11389 |
+
"epoch": 0.05198774298881266,
|
| 11390 |
+
"grad_norm": 1.3281550407409668,
|
| 11391 |
+
"learning_rate": 0.0002,
|
| 11392 |
+
"loss": 0.2525,
|
| 11393 |
+
"step": 1616
|
| 11394 |
+
},
|
| 11395 |
+
{
|
| 11396 |
+
"epoch": 0.052019913621850294,
|
| 11397 |
+
"grad_norm": 1.4164742231369019,
|
| 11398 |
+
"learning_rate": 0.0002,
|
| 11399 |
+
"loss": 0.3592,
|
| 11400 |
+
"step": 1617
|
| 11401 |
+
},
|
| 11402 |
+
{
|
| 11403 |
+
"epoch": 0.05205208425488793,
|
| 11404 |
+
"grad_norm": 1.3430583477020264,
|
| 11405 |
+
"learning_rate": 0.0002,
|
| 11406 |
+
"loss": 0.3874,
|
| 11407 |
+
"step": 1618
|
| 11408 |
+
},
|
| 11409 |
+
{
|
| 11410 |
+
"epoch": 0.052084254887925556,
|
| 11411 |
+
"grad_norm": 1.2102442979812622,
|
| 11412 |
+
"learning_rate": 0.0002,
|
| 11413 |
+
"loss": 0.3332,
|
| 11414 |
+
"step": 1619
|
| 11415 |
+
},
|
| 11416 |
+
{
|
| 11417 |
+
"epoch": 0.05211642552096319,
|
| 11418 |
+
"grad_norm": 1.2358921766281128,
|
| 11419 |
+
"learning_rate": 0.0002,
|
| 11420 |
+
"loss": 0.3808,
|
| 11421 |
+
"step": 1620
|
| 11422 |
+
},
|
| 11423 |
+
{
|
| 11424 |
+
"epoch": 0.05214859615400082,
|
| 11425 |
+
"grad_norm": 2.1207656860351562,
|
| 11426 |
+
"learning_rate": 0.0002,
|
| 11427 |
+
"loss": 0.3696,
|
| 11428 |
+
"step": 1621
|
| 11429 |
+
},
|
| 11430 |
+
{
|
| 11431 |
+
"epoch": 0.05218076678703845,
|
| 11432 |
+
"grad_norm": 1.2918682098388672,
|
| 11433 |
+
"learning_rate": 0.0002,
|
| 11434 |
+
"loss": 0.3035,
|
| 11435 |
+
"step": 1622
|
| 11436 |
+
},
|
| 11437 |
+
{
|
| 11438 |
+
"epoch": 0.052212937420076086,
|
| 11439 |
+
"grad_norm": 1.468856930732727,
|
| 11440 |
+
"learning_rate": 0.0002,
|
| 11441 |
+
"loss": 0.4139,
|
| 11442 |
+
"step": 1623
|
| 11443 |
+
},
|
| 11444 |
+
{
|
| 11445 |
+
"epoch": 0.05224510805311371,
|
| 11446 |
+
"grad_norm": 1.2353788614273071,
|
| 11447 |
+
"learning_rate": 0.0002,
|
| 11448 |
+
"loss": 0.3485,
|
| 11449 |
+
"step": 1624
|
| 11450 |
+
},
|
| 11451 |
+
{
|
| 11452 |
+
"epoch": 0.05227727868615135,
|
| 11453 |
+
"grad_norm": 1.2309379577636719,
|
| 11454 |
+
"learning_rate": 0.0002,
|
| 11455 |
+
"loss": 0.3515,
|
| 11456 |
+
"step": 1625
|
| 11457 |
+
},
|
| 11458 |
+
{
|
| 11459 |
+
"epoch": 0.05230944931918898,
|
| 11460 |
+
"grad_norm": 1.4282273054122925,
|
| 11461 |
+
"learning_rate": 0.0002,
|
| 11462 |
+
"loss": 0.2622,
|
| 11463 |
+
"step": 1626
|
| 11464 |
+
},
|
| 11465 |
+
{
|
| 11466 |
+
"epoch": 0.05234161995222661,
|
| 11467 |
+
"grad_norm": 1.4643419981002808,
|
| 11468 |
+
"learning_rate": 0.0002,
|
| 11469 |
+
"loss": 0.3148,
|
| 11470 |
+
"step": 1627
|
| 11471 |
+
},
|
| 11472 |
+
{
|
| 11473 |
+
"epoch": 0.05237379058526424,
|
| 11474 |
+
"grad_norm": 1.2791427373886108,
|
| 11475 |
+
"learning_rate": 0.0002,
|
| 11476 |
+
"loss": 0.2369,
|
| 11477 |
+
"step": 1628
|
| 11478 |
+
},
|
| 11479 |
+
{
|
| 11480 |
+
"epoch": 0.05240596121830187,
|
| 11481 |
+
"grad_norm": 1.4609307050704956,
|
| 11482 |
+
"learning_rate": 0.0002,
|
| 11483 |
+
"loss": 0.3716,
|
| 11484 |
+
"step": 1629
|
| 11485 |
+
},
|
| 11486 |
+
{
|
| 11487 |
+
"epoch": 0.052438131851339505,
|
| 11488 |
+
"grad_norm": 1.3073168992996216,
|
| 11489 |
+
"learning_rate": 0.0002,
|
| 11490 |
+
"loss": 0.3514,
|
| 11491 |
+
"step": 1630
|
| 11492 |
+
},
|
| 11493 |
+
{
|
| 11494 |
+
"epoch": 0.05247030248437714,
|
| 11495 |
+
"grad_norm": 1.682662844657898,
|
| 11496 |
+
"learning_rate": 0.0002,
|
| 11497 |
+
"loss": 0.4786,
|
| 11498 |
+
"step": 1631
|
| 11499 |
+
},
|
| 11500 |
+
{
|
| 11501 |
+
"epoch": 0.052502473117414766,
|
| 11502 |
+
"grad_norm": 1.6667532920837402,
|
| 11503 |
+
"learning_rate": 0.0002,
|
| 11504 |
+
"loss": 0.3054,
|
| 11505 |
+
"step": 1632
|
| 11506 |
+
},
|
| 11507 |
+
{
|
| 11508 |
+
"epoch": 0.0525346437504524,
|
| 11509 |
+
"grad_norm": 1.7224571704864502,
|
| 11510 |
+
"learning_rate": 0.0002,
|
| 11511 |
+
"loss": 0.3226,
|
| 11512 |
+
"step": 1633
|
| 11513 |
+
},
|
| 11514 |
+
{
|
| 11515 |
+
"epoch": 0.05256681438349003,
|
| 11516 |
+
"grad_norm": 1.6179438829421997,
|
| 11517 |
+
"learning_rate": 0.0002,
|
| 11518 |
+
"loss": 0.3544,
|
| 11519 |
+
"step": 1634
|
| 11520 |
+
},
|
| 11521 |
+
{
|
| 11522 |
+
"epoch": 0.05259898501652766,
|
| 11523 |
+
"grad_norm": 1.5665251016616821,
|
| 11524 |
+
"learning_rate": 0.0002,
|
| 11525 |
+
"loss": 0.2653,
|
| 11526 |
+
"step": 1635
|
| 11527 |
+
},
|
| 11528 |
+
{
|
| 11529 |
+
"epoch": 0.052631155649565296,
|
| 11530 |
+
"grad_norm": 1.6938865184783936,
|
| 11531 |
+
"learning_rate": 0.0002,
|
| 11532 |
+
"loss": 0.3111,
|
| 11533 |
+
"step": 1636
|
| 11534 |
+
},
|
| 11535 |
+
{
|
| 11536 |
+
"epoch": 0.052663326282602924,
|
| 11537 |
+
"grad_norm": 1.654742956161499,
|
| 11538 |
+
"learning_rate": 0.0002,
|
| 11539 |
+
"loss": 0.296,
|
| 11540 |
+
"step": 1637
|
| 11541 |
+
},
|
| 11542 |
+
{
|
| 11543 |
+
"epoch": 0.05269549691564056,
|
| 11544 |
+
"grad_norm": 1.2744419574737549,
|
| 11545 |
+
"learning_rate": 0.0002,
|
| 11546 |
+
"loss": 0.3968,
|
| 11547 |
+
"step": 1638
|
| 11548 |
+
},
|
| 11549 |
+
{
|
| 11550 |
+
"epoch": 0.05272766754867819,
|
| 11551 |
+
"grad_norm": 1.5279433727264404,
|
| 11552 |
+
"learning_rate": 0.0002,
|
| 11553 |
+
"loss": 0.3807,
|
| 11554 |
+
"step": 1639
|
| 11555 |
+
},
|
| 11556 |
+
{
|
| 11557 |
+
"epoch": 0.05275983818171582,
|
| 11558 |
+
"grad_norm": 1.4322067499160767,
|
| 11559 |
+
"learning_rate": 0.0002,
|
| 11560 |
+
"loss": 0.3944,
|
| 11561 |
+
"step": 1640
|
| 11562 |
+
},
|
| 11563 |
+
{
|
| 11564 |
+
"epoch": 0.052792008814753454,
|
| 11565 |
+
"grad_norm": 1.7231345176696777,
|
| 11566 |
+
"learning_rate": 0.0002,
|
| 11567 |
+
"loss": 0.3679,
|
| 11568 |
+
"step": 1641
|
| 11569 |
+
},
|
| 11570 |
+
{
|
| 11571 |
+
"epoch": 0.05282417944779108,
|
| 11572 |
+
"grad_norm": 1.3890626430511475,
|
| 11573 |
+
"learning_rate": 0.0002,
|
| 11574 |
+
"loss": 0.3432,
|
| 11575 |
+
"step": 1642
|
| 11576 |
+
},
|
| 11577 |
+
{
|
| 11578 |
+
"epoch": 0.052856350080828715,
|
| 11579 |
+
"grad_norm": 1.8426936864852905,
|
| 11580 |
+
"learning_rate": 0.0002,
|
| 11581 |
+
"loss": 0.3827,
|
| 11582 |
+
"step": 1643
|
| 11583 |
+
},
|
| 11584 |
+
{
|
| 11585 |
+
"epoch": 0.05288852071386635,
|
| 11586 |
+
"grad_norm": 1.3114937543869019,
|
| 11587 |
+
"learning_rate": 0.0002,
|
| 11588 |
+
"loss": 0.3065,
|
| 11589 |
+
"step": 1644
|
| 11590 |
+
},
|
| 11591 |
+
{
|
| 11592 |
+
"epoch": 0.05292069134690398,
|
| 11593 |
+
"grad_norm": 1.37386155128479,
|
| 11594 |
+
"learning_rate": 0.0002,
|
| 11595 |
+
"loss": 0.3879,
|
| 11596 |
+
"step": 1645
|
| 11597 |
+
},
|
| 11598 |
+
{
|
| 11599 |
+
"epoch": 0.05295286197994161,
|
| 11600 |
+
"grad_norm": 1.399117112159729,
|
| 11601 |
+
"learning_rate": 0.0002,
|
| 11602 |
+
"loss": 0.3109,
|
| 11603 |
+
"step": 1646
|
| 11604 |
+
},
|
| 11605 |
+
{
|
| 11606 |
+
"epoch": 0.052985032612979245,
|
| 11607 |
+
"grad_norm": 1.4952045679092407,
|
| 11608 |
+
"learning_rate": 0.0002,
|
| 11609 |
+
"loss": 0.3416,
|
| 11610 |
+
"step": 1647
|
| 11611 |
+
},
|
| 11612 |
+
{
|
| 11613 |
+
"epoch": 0.05301720324601687,
|
| 11614 |
+
"grad_norm": 1.2361903190612793,
|
| 11615 |
+
"learning_rate": 0.0002,
|
| 11616 |
+
"loss": 0.3465,
|
| 11617 |
+
"step": 1648
|
| 11618 |
+
},
|
| 11619 |
+
{
|
| 11620 |
+
"epoch": 0.05304937387905451,
|
| 11621 |
+
"grad_norm": 1.3692654371261597,
|
| 11622 |
+
"learning_rate": 0.0002,
|
| 11623 |
+
"loss": 0.309,
|
| 11624 |
+
"step": 1649
|
| 11625 |
+
},
|
| 11626 |
+
{
|
| 11627 |
+
"epoch": 0.053081544512092134,
|
| 11628 |
+
"grad_norm": 1.401586651802063,
|
| 11629 |
+
"learning_rate": 0.0002,
|
| 11630 |
+
"loss": 0.3781,
|
| 11631 |
+
"step": 1650
|
| 11632 |
+
},
|
| 11633 |
+
{
|
| 11634 |
+
"epoch": 0.05311371514512977,
|
| 11635 |
+
"grad_norm": 1.240088939666748,
|
| 11636 |
+
"learning_rate": 0.0002,
|
| 11637 |
+
"loss": 0.2762,
|
| 11638 |
+
"step": 1651
|
| 11639 |
+
},
|
| 11640 |
+
{
|
| 11641 |
+
"epoch": 0.0531458857781674,
|
| 11642 |
+
"grad_norm": 1.792125940322876,
|
| 11643 |
+
"learning_rate": 0.0002,
|
| 11644 |
+
"loss": 0.4918,
|
| 11645 |
+
"step": 1652
|
| 11646 |
+
},
|
| 11647 |
+
{
|
| 11648 |
+
"epoch": 0.05317805641120503,
|
| 11649 |
+
"grad_norm": 1.360266923904419,
|
| 11650 |
+
"learning_rate": 0.0002,
|
| 11651 |
+
"loss": 0.3424,
|
| 11652 |
+
"step": 1653
|
| 11653 |
+
},
|
| 11654 |
+
{
|
| 11655 |
+
"epoch": 0.053210227044242664,
|
| 11656 |
+
"grad_norm": 1.3013075590133667,
|
| 11657 |
+
"learning_rate": 0.0002,
|
| 11658 |
+
"loss": 0.3393,
|
| 11659 |
+
"step": 1654
|
| 11660 |
+
},
|
| 11661 |
+
{
|
| 11662 |
+
"epoch": 0.0532423976772803,
|
| 11663 |
+
"grad_norm": 1.6117784976959229,
|
| 11664 |
+
"learning_rate": 0.0002,
|
| 11665 |
+
"loss": 0.2893,
|
| 11666 |
+
"step": 1655
|
| 11667 |
+
},
|
| 11668 |
+
{
|
| 11669 |
+
"epoch": 0.053274568310317925,
|
| 11670 |
+
"grad_norm": 1.3652844429016113,
|
| 11671 |
+
"learning_rate": 0.0002,
|
| 11672 |
+
"loss": 0.3569,
|
| 11673 |
+
"step": 1656
|
| 11674 |
+
},
|
| 11675 |
+
{
|
| 11676 |
+
"epoch": 0.05330673894335556,
|
| 11677 |
+
"grad_norm": 1.434241533279419,
|
| 11678 |
+
"learning_rate": 0.0002,
|
| 11679 |
+
"loss": 0.3126,
|
| 11680 |
+
"step": 1657
|
| 11681 |
+
},
|
| 11682 |
+
{
|
| 11683 |
+
"epoch": 0.05333890957639319,
|
| 11684 |
+
"grad_norm": 1.442833662033081,
|
| 11685 |
+
"learning_rate": 0.0002,
|
| 11686 |
+
"loss": 0.2712,
|
| 11687 |
+
"step": 1658
|
| 11688 |
+
},
|
| 11689 |
+
{
|
| 11690 |
+
"epoch": 0.05337108020943082,
|
| 11691 |
+
"grad_norm": 1.1605110168457031,
|
| 11692 |
+
"learning_rate": 0.0002,
|
| 11693 |
+
"loss": 0.2821,
|
| 11694 |
+
"step": 1659
|
| 11695 |
+
},
|
| 11696 |
+
{
|
| 11697 |
+
"epoch": 0.053403250842468455,
|
| 11698 |
+
"grad_norm": 1.3498077392578125,
|
| 11699 |
+
"learning_rate": 0.0002,
|
| 11700 |
+
"loss": 0.4021,
|
| 11701 |
+
"step": 1660
|
| 11702 |
+
},
|
| 11703 |
+
{
|
| 11704 |
+
"epoch": 0.05343542147550608,
|
| 11705 |
+
"grad_norm": 1.5127185583114624,
|
| 11706 |
+
"learning_rate": 0.0002,
|
| 11707 |
+
"loss": 0.4155,
|
| 11708 |
+
"step": 1661
|
| 11709 |
+
},
|
| 11710 |
+
{
|
| 11711 |
+
"epoch": 0.05346759210854372,
|
| 11712 |
+
"grad_norm": 1.2818982601165771,
|
| 11713 |
+
"learning_rate": 0.0002,
|
| 11714 |
+
"loss": 0.3557,
|
| 11715 |
+
"step": 1662
|
| 11716 |
+
},
|
| 11717 |
+
{
|
| 11718 |
+
"epoch": 0.053499762741581344,
|
| 11719 |
+
"grad_norm": 1.0614053010940552,
|
| 11720 |
+
"learning_rate": 0.0002,
|
| 11721 |
+
"loss": 0.2353,
|
| 11722 |
+
"step": 1663
|
| 11723 |
+
},
|
| 11724 |
+
{
|
| 11725 |
+
"epoch": 0.05353193337461898,
|
| 11726 |
+
"grad_norm": 1.554489016532898,
|
| 11727 |
+
"learning_rate": 0.0002,
|
| 11728 |
+
"loss": 0.3121,
|
| 11729 |
+
"step": 1664
|
| 11730 |
+
},
|
| 11731 |
+
{
|
| 11732 |
+
"epoch": 0.05356410400765661,
|
| 11733 |
+
"grad_norm": 1.8659179210662842,
|
| 11734 |
+
"learning_rate": 0.0002,
|
| 11735 |
+
"loss": 0.3481,
|
| 11736 |
+
"step": 1665
|
| 11737 |
+
},
|
| 11738 |
+
{
|
| 11739 |
+
"epoch": 0.05359627464069424,
|
| 11740 |
+
"grad_norm": 1.556249737739563,
|
| 11741 |
+
"learning_rate": 0.0002,
|
| 11742 |
+
"loss": 0.3768,
|
| 11743 |
+
"step": 1666
|
| 11744 |
+
},
|
| 11745 |
+
{
|
| 11746 |
+
"epoch": 0.053628445273731874,
|
| 11747 |
+
"grad_norm": 1.8973748683929443,
|
| 11748 |
+
"learning_rate": 0.0002,
|
| 11749 |
+
"loss": 0.4356,
|
| 11750 |
+
"step": 1667
|
| 11751 |
+
},
|
| 11752 |
+
{
|
| 11753 |
+
"epoch": 0.05366061590676951,
|
| 11754 |
+
"grad_norm": 1.4099444150924683,
|
| 11755 |
+
"learning_rate": 0.0002,
|
| 11756 |
+
"loss": 0.3529,
|
| 11757 |
+
"step": 1668
|
| 11758 |
+
},
|
| 11759 |
+
{
|
| 11760 |
+
"epoch": 0.053692786539807136,
|
| 11761 |
+
"grad_norm": 1.4847983121871948,
|
| 11762 |
+
"learning_rate": 0.0002,
|
| 11763 |
+
"loss": 0.228,
|
| 11764 |
+
"step": 1669
|
| 11765 |
+
},
|
| 11766 |
+
{
|
| 11767 |
+
"epoch": 0.05372495717284477,
|
| 11768 |
+
"grad_norm": 1.6904289722442627,
|
| 11769 |
+
"learning_rate": 0.0002,
|
| 11770 |
+
"loss": 0.4133,
|
| 11771 |
+
"step": 1670
|
| 11772 |
+
},
|
| 11773 |
+
{
|
| 11774 |
+
"epoch": 0.0537571278058824,
|
| 11775 |
+
"grad_norm": 1.9371005296707153,
|
| 11776 |
+
"learning_rate": 0.0002,
|
| 11777 |
+
"loss": 0.4057,
|
| 11778 |
+
"step": 1671
|
| 11779 |
+
},
|
| 11780 |
+
{
|
| 11781 |
+
"epoch": 0.05378929843892003,
|
| 11782 |
+
"grad_norm": 1.69685697555542,
|
| 11783 |
+
"learning_rate": 0.0002,
|
| 11784 |
+
"loss": 0.351,
|
| 11785 |
+
"step": 1672
|
| 11786 |
+
},
|
| 11787 |
+
{
|
| 11788 |
+
"epoch": 0.053821469071957666,
|
| 11789 |
+
"grad_norm": 1.3266903162002563,
|
| 11790 |
+
"learning_rate": 0.0002,
|
| 11791 |
+
"loss": 0.3825,
|
| 11792 |
+
"step": 1673
|
| 11793 |
+
},
|
| 11794 |
+
{
|
| 11795 |
+
"epoch": 0.05385363970499529,
|
| 11796 |
+
"grad_norm": 1.290021538734436,
|
| 11797 |
+
"learning_rate": 0.0002,
|
| 11798 |
+
"loss": 0.3304,
|
| 11799 |
+
"step": 1674
|
| 11800 |
+
},
|
| 11801 |
+
{
|
| 11802 |
+
"epoch": 0.05388581033803293,
|
| 11803 |
+
"grad_norm": 1.6238058805465698,
|
| 11804 |
+
"learning_rate": 0.0002,
|
| 11805 |
+
"loss": 0.3901,
|
| 11806 |
+
"step": 1675
|
| 11807 |
+
},
|
| 11808 |
+
{
|
| 11809 |
+
"epoch": 0.05391798097107056,
|
| 11810 |
+
"grad_norm": 1.4297298192977905,
|
| 11811 |
+
"learning_rate": 0.0002,
|
| 11812 |
+
"loss": 0.3819,
|
| 11813 |
+
"step": 1676
|
| 11814 |
+
},
|
| 11815 |
+
{
|
| 11816 |
+
"epoch": 0.05395015160410819,
|
| 11817 |
+
"grad_norm": 1.24575936794281,
|
| 11818 |
+
"learning_rate": 0.0002,
|
| 11819 |
+
"loss": 0.2963,
|
| 11820 |
+
"step": 1677
|
| 11821 |
+
},
|
| 11822 |
+
{
|
| 11823 |
+
"epoch": 0.05398232223714582,
|
| 11824 |
+
"grad_norm": 1.966524362564087,
|
| 11825 |
+
"learning_rate": 0.0002,
|
| 11826 |
+
"loss": 0.4083,
|
| 11827 |
+
"step": 1678
|
| 11828 |
+
},
|
| 11829 |
+
{
|
| 11830 |
+
"epoch": 0.05401449287018345,
|
| 11831 |
+
"grad_norm": 1.5879671573638916,
|
| 11832 |
+
"learning_rate": 0.0002,
|
| 11833 |
+
"loss": 0.2824,
|
| 11834 |
+
"step": 1679
|
| 11835 |
+
},
|
| 11836 |
+
{
|
| 11837 |
+
"epoch": 0.054046663503221085,
|
| 11838 |
+
"grad_norm": 1.4699482917785645,
|
| 11839 |
+
"learning_rate": 0.0002,
|
| 11840 |
+
"loss": 0.3661,
|
| 11841 |
+
"step": 1680
|
| 11842 |
+
},
|
| 11843 |
+
{
|
| 11844 |
+
"epoch": 0.05407883413625872,
|
| 11845 |
+
"grad_norm": 1.4641050100326538,
|
| 11846 |
+
"learning_rate": 0.0002,
|
| 11847 |
+
"loss": 0.4356,
|
| 11848 |
+
"step": 1681
|
| 11849 |
+
},
|
| 11850 |
+
{
|
| 11851 |
+
"epoch": 0.054111004769296346,
|
| 11852 |
+
"grad_norm": 1.235504388809204,
|
| 11853 |
+
"learning_rate": 0.0002,
|
| 11854 |
+
"loss": 0.3851,
|
| 11855 |
+
"step": 1682
|
| 11856 |
+
},
|
| 11857 |
+
{
|
| 11858 |
+
"epoch": 0.05414317540233398,
|
| 11859 |
+
"grad_norm": 1.5182708501815796,
|
| 11860 |
+
"learning_rate": 0.0002,
|
| 11861 |
+
"loss": 0.4544,
|
| 11862 |
+
"step": 1683
|
| 11863 |
+
},
|
| 11864 |
+
{
|
| 11865 |
+
"epoch": 0.05417534603537161,
|
| 11866 |
+
"grad_norm": 1.5309251546859741,
|
| 11867 |
+
"learning_rate": 0.0002,
|
| 11868 |
+
"loss": 0.4457,
|
| 11869 |
+
"step": 1684
|
| 11870 |
+
},
|
| 11871 |
+
{
|
| 11872 |
+
"epoch": 0.05420751666840924,
|
| 11873 |
+
"grad_norm": 1.1779826879501343,
|
| 11874 |
+
"learning_rate": 0.0002,
|
| 11875 |
+
"loss": 0.3409,
|
| 11876 |
+
"step": 1685
|
| 11877 |
+
},
|
| 11878 |
+
{
|
| 11879 |
+
"epoch": 0.054239687301446876,
|
| 11880 |
+
"grad_norm": 1.5744260549545288,
|
| 11881 |
+
"learning_rate": 0.0002,
|
| 11882 |
+
"loss": 0.4168,
|
| 11883 |
+
"step": 1686
|
| 11884 |
+
},
|
| 11885 |
+
{
|
| 11886 |
+
"epoch": 0.0542718579344845,
|
| 11887 |
+
"grad_norm": 1.139620304107666,
|
| 11888 |
+
"learning_rate": 0.0002,
|
| 11889 |
+
"loss": 0.2686,
|
| 11890 |
+
"step": 1687
|
| 11891 |
+
},
|
| 11892 |
+
{
|
| 11893 |
+
"epoch": 0.05430402856752214,
|
| 11894 |
+
"grad_norm": 2.2010507583618164,
|
| 11895 |
+
"learning_rate": 0.0002,
|
| 11896 |
+
"loss": 0.4929,
|
| 11897 |
+
"step": 1688
|
| 11898 |
+
},
|
| 11899 |
+
{
|
| 11900 |
+
"epoch": 0.05433619920055977,
|
| 11901 |
+
"grad_norm": 1.3266719579696655,
|
| 11902 |
+
"learning_rate": 0.0002,
|
| 11903 |
+
"loss": 0.4041,
|
| 11904 |
+
"step": 1689
|
| 11905 |
+
},
|
| 11906 |
+
{
|
| 11907 |
+
"epoch": 0.0543683698335974,
|
| 11908 |
+
"grad_norm": 1.3262860774993896,
|
| 11909 |
+
"learning_rate": 0.0002,
|
| 11910 |
+
"loss": 0.3981,
|
| 11911 |
+
"step": 1690
|
| 11912 |
+
},
|
| 11913 |
+
{
|
| 11914 |
+
"epoch": 0.05440054046663503,
|
| 11915 |
+
"grad_norm": 1.2893167734146118,
|
| 11916 |
+
"learning_rate": 0.0002,
|
| 11917 |
+
"loss": 0.4294,
|
| 11918 |
+
"step": 1691
|
| 11919 |
+
},
|
| 11920 |
+
{
|
| 11921 |
+
"epoch": 0.05443271109967266,
|
| 11922 |
+
"grad_norm": 1.4548931121826172,
|
| 11923 |
+
"learning_rate": 0.0002,
|
| 11924 |
+
"loss": 0.362,
|
| 11925 |
+
"step": 1692
|
| 11926 |
+
},
|
| 11927 |
+
{
|
| 11928 |
+
"epoch": 0.054464881732710295,
|
| 11929 |
+
"grad_norm": 1.2379785776138306,
|
| 11930 |
+
"learning_rate": 0.0002,
|
| 11931 |
+
"loss": 0.3363,
|
| 11932 |
+
"step": 1693
|
| 11933 |
+
},
|
| 11934 |
+
{
|
| 11935 |
+
"epoch": 0.05449705236574793,
|
| 11936 |
+
"grad_norm": 1.282060980796814,
|
| 11937 |
+
"learning_rate": 0.0002,
|
| 11938 |
+
"loss": 0.3301,
|
| 11939 |
+
"step": 1694
|
| 11940 |
+
},
|
| 11941 |
+
{
|
| 11942 |
+
"epoch": 0.054529222998785556,
|
| 11943 |
+
"grad_norm": 1.3144274950027466,
|
| 11944 |
+
"learning_rate": 0.0002,
|
| 11945 |
+
"loss": 0.3098,
|
| 11946 |
+
"step": 1695
|
| 11947 |
+
},
|
| 11948 |
+
{
|
| 11949 |
+
"epoch": 0.05456139363182319,
|
| 11950 |
+
"grad_norm": 1.3063935041427612,
|
| 11951 |
+
"learning_rate": 0.0002,
|
| 11952 |
+
"loss": 0.4182,
|
| 11953 |
+
"step": 1696
|
| 11954 |
+
},
|
| 11955 |
+
{
|
| 11956 |
+
"epoch": 0.054593564264860825,
|
| 11957 |
+
"grad_norm": 1.367294430732727,
|
| 11958 |
+
"learning_rate": 0.0002,
|
| 11959 |
+
"loss": 0.3198,
|
| 11960 |
+
"step": 1697
|
| 11961 |
+
},
|
| 11962 |
+
{
|
| 11963 |
+
"epoch": 0.05462573489789845,
|
| 11964 |
+
"grad_norm": 1.199182152748108,
|
| 11965 |
+
"learning_rate": 0.0002,
|
| 11966 |
+
"loss": 0.2503,
|
| 11967 |
+
"step": 1698
|
| 11968 |
+
},
|
| 11969 |
+
{
|
| 11970 |
+
"epoch": 0.054657905530936086,
|
| 11971 |
+
"grad_norm": 1.6057204008102417,
|
| 11972 |
+
"learning_rate": 0.0002,
|
| 11973 |
+
"loss": 0.4208,
|
| 11974 |
+
"step": 1699
|
| 11975 |
+
},
|
| 11976 |
+
{
|
| 11977 |
+
"epoch": 0.054690076163973714,
|
| 11978 |
+
"grad_norm": 1.375246286392212,
|
| 11979 |
+
"learning_rate": 0.0002,
|
| 11980 |
+
"loss": 0.369,
|
| 11981 |
+
"step": 1700
|
| 11982 |
+
},
|
| 11983 |
+
{
|
| 11984 |
+
"epoch": 0.05472224679701135,
|
| 11985 |
+
"grad_norm": 1.4277704954147339,
|
| 11986 |
+
"learning_rate": 0.0002,
|
| 11987 |
+
"loss": 0.4014,
|
| 11988 |
+
"step": 1701
|
| 11989 |
+
},
|
| 11990 |
+
{
|
| 11991 |
+
"epoch": 0.05475441743004898,
|
| 11992 |
+
"grad_norm": 1.760572075843811,
|
| 11993 |
+
"learning_rate": 0.0002,
|
| 11994 |
+
"loss": 0.4971,
|
| 11995 |
+
"step": 1702
|
| 11996 |
+
},
|
| 11997 |
+
{
|
| 11998 |
+
"epoch": 0.05478658806308661,
|
| 11999 |
+
"grad_norm": 1.1767444610595703,
|
| 12000 |
+
"learning_rate": 0.0002,
|
| 12001 |
+
"loss": 0.3887,
|
| 12002 |
+
"step": 1703
|
| 12003 |
+
},
|
| 12004 |
+
{
|
| 12005 |
+
"epoch": 0.054818758696124244,
|
| 12006 |
+
"grad_norm": 1.5637484788894653,
|
| 12007 |
+
"learning_rate": 0.0002,
|
| 12008 |
+
"loss": 0.423,
|
| 12009 |
+
"step": 1704
|
| 12010 |
+
},
|
| 12011 |
+
{
|
| 12012 |
+
"epoch": 0.05485092932916188,
|
| 12013 |
+
"grad_norm": 1.5630580186843872,
|
| 12014 |
+
"learning_rate": 0.0002,
|
| 12015 |
+
"loss": 0.3064,
|
| 12016 |
+
"step": 1705
|
| 12017 |
+
},
|
| 12018 |
+
{
|
| 12019 |
+
"epoch": 0.054883099962199505,
|
| 12020 |
+
"grad_norm": 1.4884583950042725,
|
| 12021 |
+
"learning_rate": 0.0002,
|
| 12022 |
+
"loss": 0.4479,
|
| 12023 |
+
"step": 1706
|
| 12024 |
+
},
|
| 12025 |
+
{
|
| 12026 |
+
"epoch": 0.05491527059523714,
|
| 12027 |
+
"grad_norm": 1.4628565311431885,
|
| 12028 |
+
"learning_rate": 0.0002,
|
| 12029 |
+
"loss": 0.3498,
|
| 12030 |
+
"step": 1707
|
| 12031 |
+
},
|
| 12032 |
+
{
|
| 12033 |
+
"epoch": 0.05494744122827477,
|
| 12034 |
+
"grad_norm": 1.5383150577545166,
|
| 12035 |
+
"learning_rate": 0.0002,
|
| 12036 |
+
"loss": 0.3374,
|
| 12037 |
+
"step": 1708
|
| 12038 |
+
},
|
| 12039 |
+
{
|
| 12040 |
+
"epoch": 0.0549796118613124,
|
| 12041 |
+
"grad_norm": 1.2963407039642334,
|
| 12042 |
+
"learning_rate": 0.0002,
|
| 12043 |
+
"loss": 0.3092,
|
| 12044 |
+
"step": 1709
|
| 12045 |
+
},
|
| 12046 |
+
{
|
| 12047 |
+
"epoch": 0.055011782494350035,
|
| 12048 |
+
"grad_norm": 1.195543646812439,
|
| 12049 |
+
"learning_rate": 0.0002,
|
| 12050 |
+
"loss": 0.2963,
|
| 12051 |
+
"step": 1710
|
| 12052 |
+
},
|
| 12053 |
+
{
|
| 12054 |
+
"epoch": 0.05504395312738766,
|
| 12055 |
+
"grad_norm": 1.6504194736480713,
|
| 12056 |
+
"learning_rate": 0.0002,
|
| 12057 |
+
"loss": 0.4263,
|
| 12058 |
+
"step": 1711
|
| 12059 |
+
},
|
| 12060 |
+
{
|
| 12061 |
+
"epoch": 0.0550761237604253,
|
| 12062 |
+
"grad_norm": 1.2097197771072388,
|
| 12063 |
+
"learning_rate": 0.0002,
|
| 12064 |
+
"loss": 0.3503,
|
| 12065 |
+
"step": 1712
|
| 12066 |
+
},
|
| 12067 |
+
{
|
| 12068 |
+
"epoch": 0.055108294393462924,
|
| 12069 |
+
"grad_norm": 1.5624823570251465,
|
| 12070 |
+
"learning_rate": 0.0002,
|
| 12071 |
+
"loss": 0.4007,
|
| 12072 |
+
"step": 1713
|
| 12073 |
+
},
|
| 12074 |
+
{
|
| 12075 |
+
"epoch": 0.05514046502650056,
|
| 12076 |
+
"grad_norm": 1.2153539657592773,
|
| 12077 |
+
"learning_rate": 0.0002,
|
| 12078 |
+
"loss": 0.3014,
|
| 12079 |
+
"step": 1714
|
| 12080 |
+
},
|
| 12081 |
+
{
|
| 12082 |
+
"epoch": 0.05517263565953819,
|
| 12083 |
+
"grad_norm": 1.3091020584106445,
|
| 12084 |
+
"learning_rate": 0.0002,
|
| 12085 |
+
"loss": 0.3353,
|
| 12086 |
+
"step": 1715
|
| 12087 |
+
},
|
| 12088 |
+
{
|
| 12089 |
+
"epoch": 0.05520480629257582,
|
| 12090 |
+
"grad_norm": 1.2249510288238525,
|
| 12091 |
+
"learning_rate": 0.0002,
|
| 12092 |
+
"loss": 0.3109,
|
| 12093 |
+
"step": 1716
|
| 12094 |
+
},
|
| 12095 |
+
{
|
| 12096 |
+
"epoch": 0.055236976925613454,
|
| 12097 |
+
"grad_norm": 1.6217988729476929,
|
| 12098 |
+
"learning_rate": 0.0002,
|
| 12099 |
+
"loss": 0.4435,
|
| 12100 |
+
"step": 1717
|
| 12101 |
+
},
|
| 12102 |
+
{
|
| 12103 |
+
"epoch": 0.05526914755865109,
|
| 12104 |
+
"grad_norm": 1.774591088294983,
|
| 12105 |
+
"learning_rate": 0.0002,
|
| 12106 |
+
"loss": 0.4297,
|
| 12107 |
+
"step": 1718
|
| 12108 |
+
},
|
| 12109 |
+
{
|
| 12110 |
+
"epoch": 0.055301318191688716,
|
| 12111 |
+
"grad_norm": 1.2918236255645752,
|
| 12112 |
+
"learning_rate": 0.0002,
|
| 12113 |
+
"loss": 0.404,
|
| 12114 |
+
"step": 1719
|
| 12115 |
+
},
|
| 12116 |
+
{
|
| 12117 |
+
"epoch": 0.05533348882472635,
|
| 12118 |
+
"grad_norm": 1.9299715757369995,
|
| 12119 |
+
"learning_rate": 0.0002,
|
| 12120 |
+
"loss": 0.3822,
|
| 12121 |
+
"step": 1720
|
| 12122 |
+
},
|
| 12123 |
+
{
|
| 12124 |
+
"epoch": 0.05536565945776398,
|
| 12125 |
+
"grad_norm": 1.424896478652954,
|
| 12126 |
+
"learning_rate": 0.0002,
|
| 12127 |
+
"loss": 0.3738,
|
| 12128 |
+
"step": 1721
|
| 12129 |
+
},
|
| 12130 |
+
{
|
| 12131 |
+
"epoch": 0.05539783009080161,
|
| 12132 |
+
"grad_norm": 1.4504587650299072,
|
| 12133 |
+
"learning_rate": 0.0002,
|
| 12134 |
+
"loss": 0.4341,
|
| 12135 |
+
"step": 1722
|
| 12136 |
+
},
|
| 12137 |
+
{
|
| 12138 |
+
"epoch": 0.055430000723839246,
|
| 12139 |
+
"grad_norm": 1.7236900329589844,
|
| 12140 |
+
"learning_rate": 0.0002,
|
| 12141 |
+
"loss": 0.3607,
|
| 12142 |
+
"step": 1723
|
| 12143 |
+
},
|
| 12144 |
+
{
|
| 12145 |
+
"epoch": 0.05546217135687687,
|
| 12146 |
+
"grad_norm": 1.5733604431152344,
|
| 12147 |
+
"learning_rate": 0.0002,
|
| 12148 |
+
"loss": 0.3498,
|
| 12149 |
+
"step": 1724
|
| 12150 |
+
},
|
| 12151 |
+
{
|
| 12152 |
+
"epoch": 0.05549434198991451,
|
| 12153 |
+
"grad_norm": 1.577148199081421,
|
| 12154 |
+
"learning_rate": 0.0002,
|
| 12155 |
+
"loss": 0.3519,
|
| 12156 |
+
"step": 1725
|
| 12157 |
+
},
|
| 12158 |
+
{
|
| 12159 |
+
"epoch": 0.05552651262295214,
|
| 12160 |
+
"grad_norm": 1.3532757759094238,
|
| 12161 |
+
"learning_rate": 0.0002,
|
| 12162 |
+
"loss": 0.2584,
|
| 12163 |
+
"step": 1726
|
| 12164 |
+
},
|
| 12165 |
+
{
|
| 12166 |
+
"epoch": 0.05555868325598977,
|
| 12167 |
+
"grad_norm": 1.3213073015213013,
|
| 12168 |
+
"learning_rate": 0.0002,
|
| 12169 |
+
"loss": 0.2993,
|
| 12170 |
+
"step": 1727
|
| 12171 |
+
},
|
| 12172 |
+
{
|
| 12173 |
+
"epoch": 0.0555908538890274,
|
| 12174 |
+
"grad_norm": 1.2618013620376587,
|
| 12175 |
+
"learning_rate": 0.0002,
|
| 12176 |
+
"loss": 0.2884,
|
| 12177 |
+
"step": 1728
|
| 12178 |
+
},
|
| 12179 |
+
{
|
| 12180 |
+
"epoch": 0.05562302452206503,
|
| 12181 |
+
"grad_norm": 1.2675014734268188,
|
| 12182 |
+
"learning_rate": 0.0002,
|
| 12183 |
+
"loss": 0.3005,
|
| 12184 |
+
"step": 1729
|
| 12185 |
+
},
|
| 12186 |
+
{
|
| 12187 |
+
"epoch": 0.055655195155102664,
|
| 12188 |
+
"grad_norm": 1.7479132413864136,
|
| 12189 |
+
"learning_rate": 0.0002,
|
| 12190 |
+
"loss": 0.4917,
|
| 12191 |
+
"step": 1730
|
| 12192 |
+
},
|
| 12193 |
+
{
|
| 12194 |
+
"epoch": 0.0556873657881403,
|
| 12195 |
+
"grad_norm": 1.501369595527649,
|
| 12196 |
+
"learning_rate": 0.0002,
|
| 12197 |
+
"loss": 0.3491,
|
| 12198 |
+
"step": 1731
|
| 12199 |
+
},
|
| 12200 |
+
{
|
| 12201 |
+
"epoch": 0.055719536421177926,
|
| 12202 |
+
"grad_norm": 1.3929688930511475,
|
| 12203 |
+
"learning_rate": 0.0002,
|
| 12204 |
+
"loss": 0.3628,
|
| 12205 |
+
"step": 1732
|
| 12206 |
+
},
|
| 12207 |
+
{
|
| 12208 |
+
"epoch": 0.05575170705421556,
|
| 12209 |
+
"grad_norm": 1.3868218660354614,
|
| 12210 |
+
"learning_rate": 0.0002,
|
| 12211 |
+
"loss": 0.4068,
|
| 12212 |
+
"step": 1733
|
| 12213 |
+
},
|
| 12214 |
+
{
|
| 12215 |
+
"epoch": 0.05578387768725319,
|
| 12216 |
+
"grad_norm": 1.8729850053787231,
|
| 12217 |
+
"learning_rate": 0.0002,
|
| 12218 |
+
"loss": 0.4231,
|
| 12219 |
+
"step": 1734
|
| 12220 |
+
},
|
| 12221 |
+
{
|
| 12222 |
+
"epoch": 0.05581604832029082,
|
| 12223 |
+
"grad_norm": 1.530115008354187,
|
| 12224 |
+
"learning_rate": 0.0002,
|
| 12225 |
+
"loss": 0.2897,
|
| 12226 |
+
"step": 1735
|
| 12227 |
+
},
|
| 12228 |
+
{
|
| 12229 |
+
"epoch": 0.055848218953328456,
|
| 12230 |
+
"grad_norm": 1.0476127862930298,
|
| 12231 |
+
"learning_rate": 0.0002,
|
| 12232 |
+
"loss": 0.2669,
|
| 12233 |
+
"step": 1736
|
| 12234 |
+
},
|
| 12235 |
+
{
|
| 12236 |
+
"epoch": 0.05588038958636608,
|
| 12237 |
+
"grad_norm": 1.67778742313385,
|
| 12238 |
+
"learning_rate": 0.0002,
|
| 12239 |
+
"loss": 0.309,
|
| 12240 |
+
"step": 1737
|
| 12241 |
+
},
|
| 12242 |
+
{
|
| 12243 |
+
"epoch": 0.05591256021940372,
|
| 12244 |
+
"grad_norm": 1.740838646888733,
|
| 12245 |
+
"learning_rate": 0.0002,
|
| 12246 |
+
"loss": 0.3692,
|
| 12247 |
+
"step": 1738
|
| 12248 |
+
},
|
| 12249 |
+
{
|
| 12250 |
+
"epoch": 0.05594473085244135,
|
| 12251 |
+
"grad_norm": 1.2610220909118652,
|
| 12252 |
+
"learning_rate": 0.0002,
|
| 12253 |
+
"loss": 0.3833,
|
| 12254 |
+
"step": 1739
|
| 12255 |
+
},
|
| 12256 |
+
{
|
| 12257 |
+
"epoch": 0.05597690148547898,
|
| 12258 |
+
"grad_norm": 1.3762145042419434,
|
| 12259 |
+
"learning_rate": 0.0002,
|
| 12260 |
+
"loss": 0.3073,
|
| 12261 |
+
"step": 1740
|
| 12262 |
+
},
|
| 12263 |
+
{
|
| 12264 |
+
"epoch": 0.05600907211851661,
|
| 12265 |
+
"grad_norm": 1.4256093502044678,
|
| 12266 |
+
"learning_rate": 0.0002,
|
| 12267 |
+
"loss": 0.2849,
|
| 12268 |
+
"step": 1741
|
| 12269 |
+
},
|
| 12270 |
+
{
|
| 12271 |
+
"epoch": 0.05604124275155424,
|
| 12272 |
+
"grad_norm": 1.1765705347061157,
|
| 12273 |
+
"learning_rate": 0.0002,
|
| 12274 |
+
"loss": 0.3274,
|
| 12275 |
+
"step": 1742
|
| 12276 |
+
},
|
| 12277 |
+
{
|
| 12278 |
+
"epoch": 0.056073413384591875,
|
| 12279 |
+
"grad_norm": 1.3820139169692993,
|
| 12280 |
+
"learning_rate": 0.0002,
|
| 12281 |
+
"loss": 0.2519,
|
| 12282 |
+
"step": 1743
|
| 12283 |
+
},
|
| 12284 |
+
{
|
| 12285 |
+
"epoch": 0.05610558401762951,
|
| 12286 |
+
"grad_norm": 1.3904192447662354,
|
| 12287 |
+
"learning_rate": 0.0002,
|
| 12288 |
+
"loss": 0.3481,
|
| 12289 |
+
"step": 1744
|
| 12290 |
+
},
|
| 12291 |
+
{
|
| 12292 |
+
"epoch": 0.056137754650667136,
|
| 12293 |
+
"grad_norm": 1.578075647354126,
|
| 12294 |
+
"learning_rate": 0.0002,
|
| 12295 |
+
"loss": 0.3838,
|
| 12296 |
+
"step": 1745
|
| 12297 |
+
},
|
| 12298 |
+
{
|
| 12299 |
+
"epoch": 0.05616992528370477,
|
| 12300 |
+
"grad_norm": 2.1633901596069336,
|
| 12301 |
+
"learning_rate": 0.0002,
|
| 12302 |
+
"loss": 0.5132,
|
| 12303 |
+
"step": 1746
|
| 12304 |
+
},
|
| 12305 |
+
{
|
| 12306 |
+
"epoch": 0.056202095916742405,
|
| 12307 |
+
"grad_norm": 1.5895938873291016,
|
| 12308 |
+
"learning_rate": 0.0002,
|
| 12309 |
+
"loss": 0.3685,
|
| 12310 |
+
"step": 1747
|
| 12311 |
+
},
|
| 12312 |
+
{
|
| 12313 |
+
"epoch": 0.05623426654978003,
|
| 12314 |
+
"grad_norm": 1.3263061046600342,
|
| 12315 |
+
"learning_rate": 0.0002,
|
| 12316 |
+
"loss": 0.3321,
|
| 12317 |
+
"step": 1748
|
| 12318 |
+
},
|
| 12319 |
+
{
|
| 12320 |
+
"epoch": 0.056266437182817666,
|
| 12321 |
+
"grad_norm": 1.7387034893035889,
|
| 12322 |
+
"learning_rate": 0.0002,
|
| 12323 |
+
"loss": 0.4441,
|
| 12324 |
+
"step": 1749
|
| 12325 |
+
},
|
| 12326 |
+
{
|
| 12327 |
+
"epoch": 0.056298607815855294,
|
| 12328 |
+
"grad_norm": 1.5756828784942627,
|
| 12329 |
+
"learning_rate": 0.0002,
|
| 12330 |
+
"loss": 0.4804,
|
| 12331 |
+
"step": 1750
|
| 12332 |
+
},
|
| 12333 |
+
{
|
| 12334 |
+
"epoch": 0.05633077844889293,
|
| 12335 |
+
"grad_norm": 1.5541918277740479,
|
| 12336 |
+
"learning_rate": 0.0002,
|
| 12337 |
+
"loss": 0.4562,
|
| 12338 |
+
"step": 1751
|
| 12339 |
+
},
|
| 12340 |
+
{
|
| 12341 |
+
"epoch": 0.05636294908193056,
|
| 12342 |
+
"grad_norm": 1.2700080871582031,
|
| 12343 |
+
"learning_rate": 0.0002,
|
| 12344 |
+
"loss": 0.3591,
|
| 12345 |
+
"step": 1752
|
| 12346 |
+
},
|
| 12347 |
+
{
|
| 12348 |
+
"epoch": 0.05639511971496819,
|
| 12349 |
+
"grad_norm": 1.5825947523117065,
|
| 12350 |
+
"learning_rate": 0.0002,
|
| 12351 |
+
"loss": 0.4042,
|
| 12352 |
+
"step": 1753
|
| 12353 |
+
},
|
| 12354 |
+
{
|
| 12355 |
+
"epoch": 0.056427290348005824,
|
| 12356 |
+
"grad_norm": 1.5457230806350708,
|
| 12357 |
+
"learning_rate": 0.0002,
|
| 12358 |
+
"loss": 0.3447,
|
| 12359 |
+
"step": 1754
|
| 12360 |
+
},
|
| 12361 |
+
{
|
| 12362 |
+
"epoch": 0.05645946098104346,
|
| 12363 |
+
"grad_norm": 1.6879940032958984,
|
| 12364 |
+
"learning_rate": 0.0002,
|
| 12365 |
+
"loss": 0.4371,
|
| 12366 |
+
"step": 1755
|
| 12367 |
+
},
|
| 12368 |
+
{
|
| 12369 |
+
"epoch": 0.056491631614081085,
|
| 12370 |
+
"grad_norm": 1.3551236391067505,
|
| 12371 |
+
"learning_rate": 0.0002,
|
| 12372 |
+
"loss": 0.2969,
|
| 12373 |
+
"step": 1756
|
| 12374 |
+
},
|
| 12375 |
+
{
|
| 12376 |
+
"epoch": 0.05652380224711872,
|
| 12377 |
+
"grad_norm": 1.4704535007476807,
|
| 12378 |
+
"learning_rate": 0.0002,
|
| 12379 |
+
"loss": 0.3708,
|
| 12380 |
+
"step": 1757
|
| 12381 |
+
},
|
| 12382 |
+
{
|
| 12383 |
+
"epoch": 0.05655597288015635,
|
| 12384 |
+
"grad_norm": 1.2458895444869995,
|
| 12385 |
+
"learning_rate": 0.0002,
|
| 12386 |
+
"loss": 0.2914,
|
| 12387 |
+
"step": 1758
|
| 12388 |
+
},
|
| 12389 |
+
{
|
| 12390 |
+
"epoch": 0.05658814351319398,
|
| 12391 |
+
"grad_norm": 1.4119230508804321,
|
| 12392 |
+
"learning_rate": 0.0002,
|
| 12393 |
+
"loss": 0.4314,
|
| 12394 |
+
"step": 1759
|
| 12395 |
+
},
|
| 12396 |
+
{
|
| 12397 |
+
"epoch": 0.056620314146231615,
|
| 12398 |
+
"grad_norm": 1.4540210962295532,
|
| 12399 |
+
"learning_rate": 0.0002,
|
| 12400 |
+
"loss": 0.3175,
|
| 12401 |
+
"step": 1760
|
| 12402 |
+
},
|
| 12403 |
+
{
|
| 12404 |
+
"epoch": 0.05665248477926924,
|
| 12405 |
+
"grad_norm": 1.4042888879776,
|
| 12406 |
+
"learning_rate": 0.0002,
|
| 12407 |
+
"loss": 0.3327,
|
| 12408 |
+
"step": 1761
|
| 12409 |
+
},
|
| 12410 |
+
{
|
| 12411 |
+
"epoch": 0.05668465541230688,
|
| 12412 |
+
"grad_norm": 1.3778269290924072,
|
| 12413 |
+
"learning_rate": 0.0002,
|
| 12414 |
+
"loss": 0.3051,
|
| 12415 |
+
"step": 1762
|
| 12416 |
+
},
|
| 12417 |
+
{
|
| 12418 |
+
"epoch": 0.056716826045344504,
|
| 12419 |
+
"grad_norm": 1.4808207750320435,
|
| 12420 |
+
"learning_rate": 0.0002,
|
| 12421 |
+
"loss": 0.2844,
|
| 12422 |
+
"step": 1763
|
| 12423 |
+
},
|
| 12424 |
+
{
|
| 12425 |
+
"epoch": 0.05674899667838214,
|
| 12426 |
+
"grad_norm": 1.471560001373291,
|
| 12427 |
+
"learning_rate": 0.0002,
|
| 12428 |
+
"loss": 0.2853,
|
| 12429 |
+
"step": 1764
|
| 12430 |
+
},
|
| 12431 |
+
{
|
| 12432 |
+
"epoch": 0.05678116731141977,
|
| 12433 |
+
"grad_norm": 1.3863550424575806,
|
| 12434 |
+
"learning_rate": 0.0002,
|
| 12435 |
+
"loss": 0.3715,
|
| 12436 |
+
"step": 1765
|
| 12437 |
+
},
|
| 12438 |
+
{
|
| 12439 |
+
"epoch": 0.0568133379444574,
|
| 12440 |
+
"grad_norm": 1.444294810295105,
|
| 12441 |
+
"learning_rate": 0.0002,
|
| 12442 |
+
"loss": 0.3452,
|
| 12443 |
+
"step": 1766
|
| 12444 |
+
},
|
| 12445 |
+
{
|
| 12446 |
+
"epoch": 0.056845508577495034,
|
| 12447 |
+
"grad_norm": 1.2170288562774658,
|
| 12448 |
+
"learning_rate": 0.0002,
|
| 12449 |
+
"loss": 0.3307,
|
| 12450 |
+
"step": 1767
|
| 12451 |
+
},
|
| 12452 |
+
{
|
| 12453 |
+
"epoch": 0.05687767921053267,
|
| 12454 |
+
"grad_norm": 1.3601940870285034,
|
| 12455 |
+
"learning_rate": 0.0002,
|
| 12456 |
+
"loss": 0.3935,
|
| 12457 |
+
"step": 1768
|
| 12458 |
+
},
|
| 12459 |
+
{
|
| 12460 |
+
"epoch": 0.056909849843570295,
|
| 12461 |
+
"grad_norm": 1.5399080514907837,
|
| 12462 |
+
"learning_rate": 0.0002,
|
| 12463 |
+
"loss": 0.4176,
|
| 12464 |
+
"step": 1769
|
| 12465 |
+
},
|
| 12466 |
+
{
|
| 12467 |
+
"epoch": 0.05694202047660793,
|
| 12468 |
+
"grad_norm": 1.3395329713821411,
|
| 12469 |
+
"learning_rate": 0.0002,
|
| 12470 |
+
"loss": 0.3628,
|
| 12471 |
+
"step": 1770
|
| 12472 |
+
},
|
| 12473 |
+
{
|
| 12474 |
+
"epoch": 0.05697419110964556,
|
| 12475 |
+
"grad_norm": 1.9464349746704102,
|
| 12476 |
+
"learning_rate": 0.0002,
|
| 12477 |
+
"loss": 0.3275,
|
| 12478 |
+
"step": 1771
|
| 12479 |
+
},
|
| 12480 |
+
{
|
| 12481 |
+
"epoch": 0.05700636174268319,
|
| 12482 |
+
"grad_norm": 1.6518911123275757,
|
| 12483 |
+
"learning_rate": 0.0002,
|
| 12484 |
+
"loss": 0.3581,
|
| 12485 |
+
"step": 1772
|
| 12486 |
+
},
|
| 12487 |
+
{
|
| 12488 |
+
"epoch": 0.057038532375720825,
|
| 12489 |
+
"grad_norm": 1.468802809715271,
|
| 12490 |
+
"learning_rate": 0.0002,
|
| 12491 |
+
"loss": 0.4129,
|
| 12492 |
+
"step": 1773
|
| 12493 |
+
},
|
| 12494 |
+
{
|
| 12495 |
+
"epoch": 0.05707070300875845,
|
| 12496 |
+
"grad_norm": 1.3149415254592896,
|
| 12497 |
+
"learning_rate": 0.0002,
|
| 12498 |
+
"loss": 0.3162,
|
| 12499 |
+
"step": 1774
|
| 12500 |
+
},
|
| 12501 |
+
{
|
| 12502 |
+
"epoch": 0.05710287364179609,
|
| 12503 |
+
"grad_norm": 1.7053868770599365,
|
| 12504 |
+
"learning_rate": 0.0002,
|
| 12505 |
+
"loss": 0.3398,
|
| 12506 |
+
"step": 1775
|
| 12507 |
+
},
|
| 12508 |
+
{
|
| 12509 |
+
"epoch": 0.05713504427483372,
|
| 12510 |
+
"grad_norm": 1.2897893190383911,
|
| 12511 |
+
"learning_rate": 0.0002,
|
| 12512 |
+
"loss": 0.2758,
|
| 12513 |
+
"step": 1776
|
| 12514 |
+
},
|
| 12515 |
+
{
|
| 12516 |
+
"epoch": 0.05716721490787135,
|
| 12517 |
+
"grad_norm": 1.4252231121063232,
|
| 12518 |
+
"learning_rate": 0.0002,
|
| 12519 |
+
"loss": 0.2262,
|
| 12520 |
+
"step": 1777
|
| 12521 |
+
},
|
| 12522 |
+
{
|
| 12523 |
+
"epoch": 0.05719938554090898,
|
| 12524 |
+
"grad_norm": 1.3744853734970093,
|
| 12525 |
+
"learning_rate": 0.0002,
|
| 12526 |
+
"loss": 0.2802,
|
| 12527 |
+
"step": 1778
|
| 12528 |
+
},
|
| 12529 |
+
{
|
| 12530 |
+
"epoch": 0.05723155617394661,
|
| 12531 |
+
"grad_norm": 1.2926393747329712,
|
| 12532 |
+
"learning_rate": 0.0002,
|
| 12533 |
+
"loss": 0.3477,
|
| 12534 |
+
"step": 1779
|
| 12535 |
+
},
|
| 12536 |
+
{
|
| 12537 |
+
"epoch": 0.057263726806984244,
|
| 12538 |
+
"grad_norm": 1.5595306158065796,
|
| 12539 |
+
"learning_rate": 0.0002,
|
| 12540 |
+
"loss": 0.3823,
|
| 12541 |
+
"step": 1780
|
| 12542 |
+
},
|
| 12543 |
+
{
|
| 12544 |
+
"epoch": 0.05729589744002188,
|
| 12545 |
+
"grad_norm": 1.5525274276733398,
|
| 12546 |
+
"learning_rate": 0.0002,
|
| 12547 |
+
"loss": 0.3788,
|
| 12548 |
+
"step": 1781
|
| 12549 |
+
},
|
| 12550 |
+
{
|
| 12551 |
+
"epoch": 0.057328068073059506,
|
| 12552 |
+
"grad_norm": 1.4625494480133057,
|
| 12553 |
+
"learning_rate": 0.0002,
|
| 12554 |
+
"loss": 0.338,
|
| 12555 |
+
"step": 1782
|
| 12556 |
+
},
|
| 12557 |
+
{
|
| 12558 |
+
"epoch": 0.05736023870609714,
|
| 12559 |
+
"grad_norm": 1.4190728664398193,
|
| 12560 |
+
"learning_rate": 0.0002,
|
| 12561 |
+
"loss": 0.3539,
|
| 12562 |
+
"step": 1783
|
| 12563 |
+
},
|
| 12564 |
+
{
|
| 12565 |
+
"epoch": 0.057392409339134774,
|
| 12566 |
+
"grad_norm": 1.702358365058899,
|
| 12567 |
+
"learning_rate": 0.0002,
|
| 12568 |
+
"loss": 0.2488,
|
| 12569 |
+
"step": 1784
|
| 12570 |
+
},
|
| 12571 |
+
{
|
| 12572 |
+
"epoch": 0.0574245799721724,
|
| 12573 |
+
"grad_norm": 1.401620626449585,
|
| 12574 |
+
"learning_rate": 0.0002,
|
| 12575 |
+
"loss": 0.3911,
|
| 12576 |
+
"step": 1785
|
| 12577 |
+
},
|
| 12578 |
+
{
|
| 12579 |
+
"epoch": 0.057456750605210036,
|
| 12580 |
+
"grad_norm": 1.429564118385315,
|
| 12581 |
+
"learning_rate": 0.0002,
|
| 12582 |
+
"loss": 0.3282,
|
| 12583 |
+
"step": 1786
|
| 12584 |
+
},
|
| 12585 |
+
{
|
| 12586 |
+
"epoch": 0.05748892123824766,
|
| 12587 |
+
"grad_norm": 1.3767783641815186,
|
| 12588 |
+
"learning_rate": 0.0002,
|
| 12589 |
+
"loss": 0.3155,
|
| 12590 |
+
"step": 1787
|
| 12591 |
+
},
|
| 12592 |
+
{
|
| 12593 |
+
"epoch": 0.0575210918712853,
|
| 12594 |
+
"grad_norm": 1.6560885906219482,
|
| 12595 |
+
"learning_rate": 0.0002,
|
| 12596 |
+
"loss": 0.4185,
|
| 12597 |
+
"step": 1788
|
| 12598 |
+
},
|
| 12599 |
+
{
|
| 12600 |
+
"epoch": 0.05755326250432293,
|
| 12601 |
+
"grad_norm": 1.4894121885299683,
|
| 12602 |
+
"learning_rate": 0.0002,
|
| 12603 |
+
"loss": 0.3895,
|
| 12604 |
+
"step": 1789
|
| 12605 |
+
},
|
| 12606 |
+
{
|
| 12607 |
+
"epoch": 0.05758543313736056,
|
| 12608 |
+
"grad_norm": 1.457728624343872,
|
| 12609 |
+
"learning_rate": 0.0002,
|
| 12610 |
+
"loss": 0.3405,
|
| 12611 |
+
"step": 1790
|
| 12612 |
+
},
|
| 12613 |
+
{
|
| 12614 |
+
"epoch": 0.05761760377039819,
|
| 12615 |
+
"grad_norm": 1.3831778764724731,
|
| 12616 |
+
"learning_rate": 0.0002,
|
| 12617 |
+
"loss": 0.2808,
|
| 12618 |
+
"step": 1791
|
| 12619 |
+
},
|
| 12620 |
+
{
|
| 12621 |
+
"epoch": 0.05764977440343582,
|
| 12622 |
+
"grad_norm": 1.5047637224197388,
|
| 12623 |
+
"learning_rate": 0.0002,
|
| 12624 |
+
"loss": 0.3211,
|
| 12625 |
+
"step": 1792
|
| 12626 |
+
},
|
| 12627 |
+
{
|
| 12628 |
+
"epoch": 0.057681945036473455,
|
| 12629 |
+
"grad_norm": 1.3648213148117065,
|
| 12630 |
+
"learning_rate": 0.0002,
|
| 12631 |
+
"loss": 0.3884,
|
| 12632 |
+
"step": 1793
|
| 12633 |
+
},
|
| 12634 |
+
{
|
| 12635 |
+
"epoch": 0.05771411566951109,
|
| 12636 |
+
"grad_norm": 1.5376144647598267,
|
| 12637 |
+
"learning_rate": 0.0002,
|
| 12638 |
+
"loss": 0.3959,
|
| 12639 |
+
"step": 1794
|
| 12640 |
+
},
|
| 12641 |
+
{
|
| 12642 |
+
"epoch": 0.057746286302548716,
|
| 12643 |
+
"grad_norm": 1.6056605577468872,
|
| 12644 |
+
"learning_rate": 0.0002,
|
| 12645 |
+
"loss": 0.4887,
|
| 12646 |
+
"step": 1795
|
| 12647 |
+
},
|
| 12648 |
+
{
|
| 12649 |
+
"epoch": 0.05777845693558635,
|
| 12650 |
+
"grad_norm": 1.2603166103363037,
|
| 12651 |
+
"learning_rate": 0.0002,
|
| 12652 |
+
"loss": 0.2566,
|
| 12653 |
+
"step": 1796
|
| 12654 |
+
},
|
| 12655 |
+
{
|
| 12656 |
+
"epoch": 0.057810627568623985,
|
| 12657 |
+
"grad_norm": 1.444352626800537,
|
| 12658 |
+
"learning_rate": 0.0002,
|
| 12659 |
+
"loss": 0.3696,
|
| 12660 |
+
"step": 1797
|
| 12661 |
+
},
|
| 12662 |
+
{
|
| 12663 |
+
"epoch": 0.05784279820166161,
|
| 12664 |
+
"grad_norm": 1.2577990293502808,
|
| 12665 |
+
"learning_rate": 0.0002,
|
| 12666 |
+
"loss": 0.3349,
|
| 12667 |
+
"step": 1798
|
| 12668 |
+
},
|
| 12669 |
+
{
|
| 12670 |
+
"epoch": 0.057874968834699246,
|
| 12671 |
+
"grad_norm": 1.9209790229797363,
|
| 12672 |
+
"learning_rate": 0.0002,
|
| 12673 |
+
"loss": 0.4819,
|
| 12674 |
+
"step": 1799
|
| 12675 |
+
},
|
| 12676 |
+
{
|
| 12677 |
+
"epoch": 0.05790713946773687,
|
| 12678 |
+
"grad_norm": 1.4944988489151,
|
| 12679 |
+
"learning_rate": 0.0002,
|
| 12680 |
+
"loss": 0.2463,
|
| 12681 |
+
"step": 1800
|
| 12682 |
+
},
|
| 12683 |
+
{
|
| 12684 |
+
"epoch": 0.05790713946773687,
|
| 12685 |
+
"eval_loss": 0.32331234216690063,
|
| 12686 |
+
"eval_runtime": 28.8437,
|
| 12687 |
+
"eval_samples_per_second": 8.633,
|
| 12688 |
+
"eval_steps_per_second": 4.334,
|
| 12689 |
+
"step": 1800
|
| 12690 |
}
|
| 12691 |
],
|
| 12692 |
"logging_steps": 1,
|
|
|
|
| 12701 |
"early_stopping_threshold": 0.0
|
| 12702 |
},
|
| 12703 |
"attributes": {
|
| 12704 |
+
"early_stopping_patience_counter": 1
|
| 12705 |
}
|
| 12706 |
},
|
| 12707 |
"TrainerControl": {
|
|
|
|
| 12715 |
"attributes": {}
|
| 12716 |
}
|
| 12717 |
},
|
| 12718 |
+
"total_flos": 1.1852880273211392e+17,
|
| 12719 |
"train_batch_size": 2,
|
| 12720 |
"trial_name": null,
|
| 12721 |
"trial_params": null
|