Training in progress, step 612, 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 50503544
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:0d21bb09eff34416685fbbd7e9c819db3b4a8bd18092777255987600c2b52cb3
|
| 3 |
size 50503544
|
last-checkpoint/optimizer.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 101184122
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:763b5a91612f432ad5cd4386cc9039d08b7dfceae177911b54c728b3d5924880
|
| 3 |
size 101184122
|
last-checkpoint/rng_state_0.pth
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 14512
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:95957e47c48f89f2d82005240408999cbda782b079d16061eb6dae69038672f8
|
| 3 |
size 14512
|
last-checkpoint/rng_state_1.pth
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 14512
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:2818223cc47cf96c694a9d3dd14069656697688b751dd9687bd461784a1f04a8
|
| 3 |
size 14512
|
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:e3954ee4b0cce08f11cf59acefd5fcef657cbc4f4e12f0d54a93447ea6f7f110
|
| 3 |
size 1064
|
last-checkpoint/trainer_state.json
CHANGED
|
@@ -1,9 +1,9 @@
|
|
| 1 |
{
|
| 2 |
"best_metric": null,
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
-
"epoch": 0.
|
| 5 |
"eval_steps": 204,
|
| 6 |
-
"global_step":
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
@@ -2887,6 +2887,1442 @@
|
|
| 2887 |
"eval_samples_per_second": 10.486,
|
| 2888 |
"eval_steps_per_second": 5.251,
|
| 2889 |
"step": 408
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 2890 |
}
|
| 2891 |
],
|
| 2892 |
"logging_steps": 1,
|
|
@@ -2906,7 +4342,7 @@
|
|
| 2906 |
"attributes": {}
|
| 2907 |
}
|
| 2908 |
},
|
| 2909 |
-
"total_flos":
|
| 2910 |
"train_batch_size": 1,
|
| 2911 |
"trial_name": null,
|
| 2912 |
"trial_params": null
|
|
|
|
| 1 |
{
|
| 2 |
"best_metric": null,
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 0.7523048555623848,
|
| 5 |
"eval_steps": 204,
|
| 6 |
+
"global_step": 612,
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
|
|
| 2887 |
"eval_samples_per_second": 10.486,
|
| 2888 |
"eval_steps_per_second": 5.251,
|
| 2889 |
"step": 408
|
| 2890 |
+
},
|
| 2891 |
+
{
|
| 2892 |
+
"epoch": 0.5027658266748617,
|
| 2893 |
+
"grad_norm": 0.31012120842933655,
|
| 2894 |
+
"learning_rate": 5.4625476241761196e-05,
|
| 2895 |
+
"loss": 1.1491,
|
| 2896 |
+
"step": 409
|
| 2897 |
+
},
|
| 2898 |
+
{
|
| 2899 |
+
"epoch": 0.5039950829748002,
|
| 2900 |
+
"grad_norm": 0.3182609975337982,
|
| 2901 |
+
"learning_rate": 5.442044935102375e-05,
|
| 2902 |
+
"loss": 1.0786,
|
| 2903 |
+
"step": 410
|
| 2904 |
+
},
|
| 2905 |
+
{
|
| 2906 |
+
"epoch": 0.5052243392747388,
|
| 2907 |
+
"grad_norm": 0.35413217544555664,
|
| 2908 |
+
"learning_rate": 5.421534751980556e-05,
|
| 2909 |
+
"loss": 1.2194,
|
| 2910 |
+
"step": 411
|
| 2911 |
+
},
|
| 2912 |
+
{
|
| 2913 |
+
"epoch": 0.5064535955746773,
|
| 2914 |
+
"grad_norm": 0.39505264163017273,
|
| 2915 |
+
"learning_rate": 5.401017422522594e-05,
|
| 2916 |
+
"loss": 1.0296,
|
| 2917 |
+
"step": 412
|
| 2918 |
+
},
|
| 2919 |
+
{
|
| 2920 |
+
"epoch": 0.5076828518746158,
|
| 2921 |
+
"grad_norm": 0.4267808496952057,
|
| 2922 |
+
"learning_rate": 5.380493294561573e-05,
|
| 2923 |
+
"loss": 1.0906,
|
| 2924 |
+
"step": 413
|
| 2925 |
+
},
|
| 2926 |
+
{
|
| 2927 |
+
"epoch": 0.5089121081745543,
|
| 2928 |
+
"grad_norm": 0.5495923161506653,
|
| 2929 |
+
"learning_rate": 5.359962716045835e-05,
|
| 2930 |
+
"loss": 1.2171,
|
| 2931 |
+
"step": 414
|
| 2932 |
+
},
|
| 2933 |
+
{
|
| 2934 |
+
"epoch": 0.510141364474493,
|
| 2935 |
+
"grad_norm": 0.6385165452957153,
|
| 2936 |
+
"learning_rate": 5.3394260350330796e-05,
|
| 2937 |
+
"loss": 0.8179,
|
| 2938 |
+
"step": 415
|
| 2939 |
+
},
|
| 2940 |
+
{
|
| 2941 |
+
"epoch": 0.5113706207744315,
|
| 2942 |
+
"grad_norm": 0.7211401462554932,
|
| 2943 |
+
"learning_rate": 5.318883599684456e-05,
|
| 2944 |
+
"loss": 0.7624,
|
| 2945 |
+
"step": 416
|
| 2946 |
+
},
|
| 2947 |
+
{
|
| 2948 |
+
"epoch": 0.51259987707437,
|
| 2949 |
+
"grad_norm": 0.800375759601593,
|
| 2950 |
+
"learning_rate": 5.298335758258678e-05,
|
| 2951 |
+
"loss": 0.9597,
|
| 2952 |
+
"step": 417
|
| 2953 |
+
},
|
| 2954 |
+
{
|
| 2955 |
+
"epoch": 0.5138291333743086,
|
| 2956 |
+
"grad_norm": 0.9896557927131653,
|
| 2957 |
+
"learning_rate": 5.2777828591060984e-05,
|
| 2958 |
+
"loss": 1.12,
|
| 2959 |
+
"step": 418
|
| 2960 |
+
},
|
| 2961 |
+
{
|
| 2962 |
+
"epoch": 0.5150583896742471,
|
| 2963 |
+
"grad_norm": 2.869112968444824,
|
| 2964 |
+
"learning_rate": 5.257225250662823e-05,
|
| 2965 |
+
"loss": 1.2992,
|
| 2966 |
+
"step": 419
|
| 2967 |
+
},
|
| 2968 |
+
{
|
| 2969 |
+
"epoch": 0.5162876459741856,
|
| 2970 |
+
"grad_norm": 3.394543409347534,
|
| 2971 |
+
"learning_rate": 5.236663281444791e-05,
|
| 2972 |
+
"loss": 1.298,
|
| 2973 |
+
"step": 420
|
| 2974 |
+
},
|
| 2975 |
+
{
|
| 2976 |
+
"epoch": 0.5175169022741242,
|
| 2977 |
+
"grad_norm": 3.3043065071105957,
|
| 2978 |
+
"learning_rate": 5.21609730004187e-05,
|
| 2979 |
+
"loss": 1.1924,
|
| 2980 |
+
"step": 421
|
| 2981 |
+
},
|
| 2982 |
+
{
|
| 2983 |
+
"epoch": 0.5187461585740627,
|
| 2984 |
+
"grad_norm": 2.603189706802368,
|
| 2985 |
+
"learning_rate": 5.1955276551119495e-05,
|
| 2986 |
+
"loss": 1.2155,
|
| 2987 |
+
"step": 422
|
| 2988 |
+
},
|
| 2989 |
+
{
|
| 2990 |
+
"epoch": 0.5199754148740012,
|
| 2991 |
+
"grad_norm": 3.018608808517456,
|
| 2992 |
+
"learning_rate": 5.174954695375023e-05,
|
| 2993 |
+
"loss": 1.2001,
|
| 2994 |
+
"step": 423
|
| 2995 |
+
},
|
| 2996 |
+
{
|
| 2997 |
+
"epoch": 0.5212046711739398,
|
| 2998 |
+
"grad_norm": 1.8929078578948975,
|
| 2999 |
+
"learning_rate": 5.154378769607286e-05,
|
| 3000 |
+
"loss": 0.7124,
|
| 3001 |
+
"step": 424
|
| 3002 |
+
},
|
| 3003 |
+
{
|
| 3004 |
+
"epoch": 0.5224339274738783,
|
| 3005 |
+
"grad_norm": 2.4811208248138428,
|
| 3006 |
+
"learning_rate": 5.1338002266352106e-05,
|
| 3007 |
+
"loss": 0.8491,
|
| 3008 |
+
"step": 425
|
| 3009 |
+
},
|
| 3010 |
+
{
|
| 3011 |
+
"epoch": 0.5236631837738168,
|
| 3012 |
+
"grad_norm": 0.22651559114456177,
|
| 3013 |
+
"learning_rate": 5.113219415329645e-05,
|
| 3014 |
+
"loss": 0.9983,
|
| 3015 |
+
"step": 426
|
| 3016 |
+
},
|
| 3017 |
+
{
|
| 3018 |
+
"epoch": 0.5248924400737554,
|
| 3019 |
+
"grad_norm": 0.2817627787590027,
|
| 3020 |
+
"learning_rate": 5.0926366845998904e-05,
|
| 3021 |
+
"loss": 1.1314,
|
| 3022 |
+
"step": 427
|
| 3023 |
+
},
|
| 3024 |
+
{
|
| 3025 |
+
"epoch": 0.5261216963736939,
|
| 3026 |
+
"grad_norm": 0.27301520109176636,
|
| 3027 |
+
"learning_rate": 5.072052383387786e-05,
|
| 3028 |
+
"loss": 1.2722,
|
| 3029 |
+
"step": 428
|
| 3030 |
+
},
|
| 3031 |
+
{
|
| 3032 |
+
"epoch": 0.5273509526736324,
|
| 3033 |
+
"grad_norm": 0.2770383059978485,
|
| 3034 |
+
"learning_rate": 5.0514668606618e-05,
|
| 3035 |
+
"loss": 1.1654,
|
| 3036 |
+
"step": 429
|
| 3037 |
+
},
|
| 3038 |
+
{
|
| 3039 |
+
"epoch": 0.5285802089735709,
|
| 3040 |
+
"grad_norm": 0.3147628903388977,
|
| 3041 |
+
"learning_rate": 5.0308804654111056e-05,
|
| 3042 |
+
"loss": 1.0896,
|
| 3043 |
+
"step": 430
|
| 3044 |
+
},
|
| 3045 |
+
{
|
| 3046 |
+
"epoch": 0.5298094652735095,
|
| 3047 |
+
"grad_norm": 0.305074542760849,
|
| 3048 |
+
"learning_rate": 5.01029354663967e-05,
|
| 3049 |
+
"loss": 1.1754,
|
| 3050 |
+
"step": 431
|
| 3051 |
+
},
|
| 3052 |
+
{
|
| 3053 |
+
"epoch": 0.5310387215734481,
|
| 3054 |
+
"grad_norm": 0.3240547776222229,
|
| 3055 |
+
"learning_rate": 4.9897064533603315e-05,
|
| 3056 |
+
"loss": 1.2413,
|
| 3057 |
+
"step": 432
|
| 3058 |
+
},
|
| 3059 |
+
{
|
| 3060 |
+
"epoch": 0.5322679778733866,
|
| 3061 |
+
"grad_norm": 0.34143832325935364,
|
| 3062 |
+
"learning_rate": 4.9691195345888956e-05,
|
| 3063 |
+
"loss": 1.1504,
|
| 3064 |
+
"step": 433
|
| 3065 |
+
},
|
| 3066 |
+
{
|
| 3067 |
+
"epoch": 0.5334972341733252,
|
| 3068 |
+
"grad_norm": 0.3945367634296417,
|
| 3069 |
+
"learning_rate": 4.948533139338202e-05,
|
| 3070 |
+
"loss": 1.2914,
|
| 3071 |
+
"step": 434
|
| 3072 |
+
},
|
| 3073 |
+
{
|
| 3074 |
+
"epoch": 0.5347264904732637,
|
| 3075 |
+
"grad_norm": 0.371367871761322,
|
| 3076 |
+
"learning_rate": 4.927947616612215e-05,
|
| 3077 |
+
"loss": 1.3269,
|
| 3078 |
+
"step": 435
|
| 3079 |
+
},
|
| 3080 |
+
{
|
| 3081 |
+
"epoch": 0.5359557467732022,
|
| 3082 |
+
"grad_norm": 0.4348413646221161,
|
| 3083 |
+
"learning_rate": 4.90736331540011e-05,
|
| 3084 |
+
"loss": 1.2061,
|
| 3085 |
+
"step": 436
|
| 3086 |
+
},
|
| 3087 |
+
{
|
| 3088 |
+
"epoch": 0.5371850030731408,
|
| 3089 |
+
"grad_norm": 0.5112064480781555,
|
| 3090 |
+
"learning_rate": 4.886780584670356e-05,
|
| 3091 |
+
"loss": 1.2283,
|
| 3092 |
+
"step": 437
|
| 3093 |
+
},
|
| 3094 |
+
{
|
| 3095 |
+
"epoch": 0.5384142593730793,
|
| 3096 |
+
"grad_norm": 0.6409791707992554,
|
| 3097 |
+
"learning_rate": 4.866199773364789e-05,
|
| 3098 |
+
"loss": 0.7663,
|
| 3099 |
+
"step": 438
|
| 3100 |
+
},
|
| 3101 |
+
{
|
| 3102 |
+
"epoch": 0.5396435156730178,
|
| 3103 |
+
"grad_norm": 0.7076222896575928,
|
| 3104 |
+
"learning_rate": 4.845621230392716e-05,
|
| 3105 |
+
"loss": 0.8073,
|
| 3106 |
+
"step": 439
|
| 3107 |
+
},
|
| 3108 |
+
{
|
| 3109 |
+
"epoch": 0.5408727719729564,
|
| 3110 |
+
"grad_norm": 0.6545782089233398,
|
| 3111 |
+
"learning_rate": 4.825045304624978e-05,
|
| 3112 |
+
"loss": 0.762,
|
| 3113 |
+
"step": 440
|
| 3114 |
+
},
|
| 3115 |
+
{
|
| 3116 |
+
"epoch": 0.5421020282728949,
|
| 3117 |
+
"grad_norm": 0.6853161454200745,
|
| 3118 |
+
"learning_rate": 4.804472344888052e-05,
|
| 3119 |
+
"loss": 0.9905,
|
| 3120 |
+
"step": 441
|
| 3121 |
+
},
|
| 3122 |
+
{
|
| 3123 |
+
"epoch": 0.5433312845728334,
|
| 3124 |
+
"grad_norm": 0.6758946776390076,
|
| 3125 |
+
"learning_rate": 4.7839026999581296e-05,
|
| 3126 |
+
"loss": 0.9353,
|
| 3127 |
+
"step": 442
|
| 3128 |
+
},
|
| 3129 |
+
{
|
| 3130 |
+
"epoch": 0.544560540872772,
|
| 3131 |
+
"grad_norm": 1.1472845077514648,
|
| 3132 |
+
"learning_rate": 4.7633367185552095e-05,
|
| 3133 |
+
"loss": 0.9796,
|
| 3134 |
+
"step": 443
|
| 3135 |
+
},
|
| 3136 |
+
{
|
| 3137 |
+
"epoch": 0.5457897971727105,
|
| 3138 |
+
"grad_norm": 2.543483018875122,
|
| 3139 |
+
"learning_rate": 4.742774749337179e-05,
|
| 3140 |
+
"loss": 1.5897,
|
| 3141 |
+
"step": 444
|
| 3142 |
+
},
|
| 3143 |
+
{
|
| 3144 |
+
"epoch": 0.547019053472649,
|
| 3145 |
+
"grad_norm": 3.029282331466675,
|
| 3146 |
+
"learning_rate": 4.7222171408939034e-05,
|
| 3147 |
+
"loss": 1.2527,
|
| 3148 |
+
"step": 445
|
| 3149 |
+
},
|
| 3150 |
+
{
|
| 3151 |
+
"epoch": 0.5482483097725875,
|
| 3152 |
+
"grad_norm": 3.1332318782806396,
|
| 3153 |
+
"learning_rate": 4.701664241741323e-05,
|
| 3154 |
+
"loss": 1.1093,
|
| 3155 |
+
"step": 446
|
| 3156 |
+
},
|
| 3157 |
+
{
|
| 3158 |
+
"epoch": 0.5494775660725261,
|
| 3159 |
+
"grad_norm": 2.407163381576538,
|
| 3160 |
+
"learning_rate": 4.681116400315544e-05,
|
| 3161 |
+
"loss": 0.75,
|
| 3162 |
+
"step": 447
|
| 3163 |
+
},
|
| 3164 |
+
{
|
| 3165 |
+
"epoch": 0.5507068223724647,
|
| 3166 |
+
"grad_norm": 2.622605085372925,
|
| 3167 |
+
"learning_rate": 4.6605739649669236e-05,
|
| 3168 |
+
"loss": 0.8924,
|
| 3169 |
+
"step": 448
|
| 3170 |
+
},
|
| 3171 |
+
{
|
| 3172 |
+
"epoch": 0.5519360786724032,
|
| 3173 |
+
"grad_norm": 2.5338971614837646,
|
| 3174 |
+
"learning_rate": 4.640037283954165e-05,
|
| 3175 |
+
"loss": 0.681,
|
| 3176 |
+
"step": 449
|
| 3177 |
+
},
|
| 3178 |
+
{
|
| 3179 |
+
"epoch": 0.5531653349723418,
|
| 3180 |
+
"grad_norm": 2.4635331630706787,
|
| 3181 |
+
"learning_rate": 4.619506705438428e-05,
|
| 3182 |
+
"loss": 0.7866,
|
| 3183 |
+
"step": 450
|
| 3184 |
+
},
|
| 3185 |
+
{
|
| 3186 |
+
"epoch": 0.5543945912722803,
|
| 3187 |
+
"grad_norm": 0.18258269131183624,
|
| 3188 |
+
"learning_rate": 4.598982577477408e-05,
|
| 3189 |
+
"loss": 0.8069,
|
| 3190 |
+
"step": 451
|
| 3191 |
+
},
|
| 3192 |
+
{
|
| 3193 |
+
"epoch": 0.5556238475722188,
|
| 3194 |
+
"grad_norm": 0.23453976213932037,
|
| 3195 |
+
"learning_rate": 4.578465248019445e-05,
|
| 3196 |
+
"loss": 1.1606,
|
| 3197 |
+
"step": 452
|
| 3198 |
+
},
|
| 3199 |
+
{
|
| 3200 |
+
"epoch": 0.5568531038721574,
|
| 3201 |
+
"grad_norm": 0.25334998965263367,
|
| 3202 |
+
"learning_rate": 4.557955064897626e-05,
|
| 3203 |
+
"loss": 1.2323,
|
| 3204 |
+
"step": 453
|
| 3205 |
+
},
|
| 3206 |
+
{
|
| 3207 |
+
"epoch": 0.5580823601720959,
|
| 3208 |
+
"grad_norm": 0.2757891118526459,
|
| 3209 |
+
"learning_rate": 4.537452375823881e-05,
|
| 3210 |
+
"loss": 1.4173,
|
| 3211 |
+
"step": 454
|
| 3212 |
+
},
|
| 3213 |
+
{
|
| 3214 |
+
"epoch": 0.5593116164720344,
|
| 3215 |
+
"grad_norm": 0.2755081355571747,
|
| 3216 |
+
"learning_rate": 4.5169575283830936e-05,
|
| 3217 |
+
"loss": 1.2255,
|
| 3218 |
+
"step": 455
|
| 3219 |
+
},
|
| 3220 |
+
{
|
| 3221 |
+
"epoch": 0.560540872771973,
|
| 3222 |
+
"grad_norm": 0.29173940420150757,
|
| 3223 |
+
"learning_rate": 4.496470870027209e-05,
|
| 3224 |
+
"loss": 1.4528,
|
| 3225 |
+
"step": 456
|
| 3226 |
+
},
|
| 3227 |
+
{
|
| 3228 |
+
"epoch": 0.5617701290719115,
|
| 3229 |
+
"grad_norm": 0.27483510971069336,
|
| 3230 |
+
"learning_rate": 4.475992748069339e-05,
|
| 3231 |
+
"loss": 0.7703,
|
| 3232 |
+
"step": 457
|
| 3233 |
+
},
|
| 3234 |
+
{
|
| 3235 |
+
"epoch": 0.56299938537185,
|
| 3236 |
+
"grad_norm": 0.3180837333202362,
|
| 3237 |
+
"learning_rate": 4.455523509677882e-05,
|
| 3238 |
+
"loss": 1.025,
|
| 3239 |
+
"step": 458
|
| 3240 |
+
},
|
| 3241 |
+
{
|
| 3242 |
+
"epoch": 0.5642286416717885,
|
| 3243 |
+
"grad_norm": 0.3103027939796448,
|
| 3244 |
+
"learning_rate": 4.435063501870622e-05,
|
| 3245 |
+
"loss": 1.0503,
|
| 3246 |
+
"step": 459
|
| 3247 |
+
},
|
| 3248 |
+
{
|
| 3249 |
+
"epoch": 0.5654578979717271,
|
| 3250 |
+
"grad_norm": 0.3448195457458496,
|
| 3251 |
+
"learning_rate": 4.4146130715088676e-05,
|
| 3252 |
+
"loss": 1.0366,
|
| 3253 |
+
"step": 460
|
| 3254 |
+
},
|
| 3255 |
+
{
|
| 3256 |
+
"epoch": 0.5666871542716656,
|
| 3257 |
+
"grad_norm": 0.3892152011394501,
|
| 3258 |
+
"learning_rate": 4.3941725652915494e-05,
|
| 3259 |
+
"loss": 1.1834,
|
| 3260 |
+
"step": 461
|
| 3261 |
+
},
|
| 3262 |
+
{
|
| 3263 |
+
"epoch": 0.5679164105716041,
|
| 3264 |
+
"grad_norm": 0.4770027697086334,
|
| 3265 |
+
"learning_rate": 4.373742329749362e-05,
|
| 3266 |
+
"loss": 1.4524,
|
| 3267 |
+
"step": 462
|
| 3268 |
+
},
|
| 3269 |
+
{
|
| 3270 |
+
"epoch": 0.5691456668715427,
|
| 3271 |
+
"grad_norm": 0.5856226682662964,
|
| 3272 |
+
"learning_rate": 4.3533227112388694e-05,
|
| 3273 |
+
"loss": 1.2538,
|
| 3274 |
+
"step": 463
|
| 3275 |
+
},
|
| 3276 |
+
{
|
| 3277 |
+
"epoch": 0.5703749231714813,
|
| 3278 |
+
"grad_norm": 0.6203077435493469,
|
| 3279 |
+
"learning_rate": 4.332914055936653e-05,
|
| 3280 |
+
"loss": 0.6276,
|
| 3281 |
+
"step": 464
|
| 3282 |
+
},
|
| 3283 |
+
{
|
| 3284 |
+
"epoch": 0.5716041794714198,
|
| 3285 |
+
"grad_norm": 0.6294959187507629,
|
| 3286 |
+
"learning_rate": 4.3125167098334286e-05,
|
| 3287 |
+
"loss": 0.642,
|
| 3288 |
+
"step": 465
|
| 3289 |
+
},
|
| 3290 |
+
{
|
| 3291 |
+
"epoch": 0.5728334357713584,
|
| 3292 |
+
"grad_norm": 0.7135624885559082,
|
| 3293 |
+
"learning_rate": 4.2921310187281864e-05,
|
| 3294 |
+
"loss": 1.1486,
|
| 3295 |
+
"step": 466
|
| 3296 |
+
},
|
| 3297 |
+
{
|
| 3298 |
+
"epoch": 0.5740626920712969,
|
| 3299 |
+
"grad_norm": 0.7505180835723877,
|
| 3300 |
+
"learning_rate": 4.27175732822233e-05,
|
| 3301 |
+
"loss": 0.9109,
|
| 3302 |
+
"step": 467
|
| 3303 |
+
},
|
| 3304 |
+
{
|
| 3305 |
+
"epoch": 0.5752919483712354,
|
| 3306 |
+
"grad_norm": 1.0719517469406128,
|
| 3307 |
+
"learning_rate": 4.251395983713809e-05,
|
| 3308 |
+
"loss": 1.2909,
|
| 3309 |
+
"step": 468
|
| 3310 |
+
},
|
| 3311 |
+
{
|
| 3312 |
+
"epoch": 0.576521204671174,
|
| 3313 |
+
"grad_norm": 2.5531132221221924,
|
| 3314 |
+
"learning_rate": 4.231047330391278e-05,
|
| 3315 |
+
"loss": 1.1723,
|
| 3316 |
+
"step": 469
|
| 3317 |
+
},
|
| 3318 |
+
{
|
| 3319 |
+
"epoch": 0.5777504609711125,
|
| 3320 |
+
"grad_norm": 3.104837656021118,
|
| 3321 |
+
"learning_rate": 4.21071171322823e-05,
|
| 3322 |
+
"loss": 1.3296,
|
| 3323 |
+
"step": 470
|
| 3324 |
+
},
|
| 3325 |
+
{
|
| 3326 |
+
"epoch": 0.578979717271051,
|
| 3327 |
+
"grad_norm": 3.239119052886963,
|
| 3328 |
+
"learning_rate": 4.190389476977156e-05,
|
| 3329 |
+
"loss": 0.8768,
|
| 3330 |
+
"step": 471
|
| 3331 |
+
},
|
| 3332 |
+
{
|
| 3333 |
+
"epoch": 0.5802089735709896,
|
| 3334 |
+
"grad_norm": 2.681823968887329,
|
| 3335 |
+
"learning_rate": 4.170080966163702e-05,
|
| 3336 |
+
"loss": 0.8809,
|
| 3337 |
+
"step": 472
|
| 3338 |
+
},
|
| 3339 |
+
{
|
| 3340 |
+
"epoch": 0.5814382298709281,
|
| 3341 |
+
"grad_norm": 2.557533025741577,
|
| 3342 |
+
"learning_rate": 4.149786525080819e-05,
|
| 3343 |
+
"loss": 0.8249,
|
| 3344 |
+
"step": 473
|
| 3345 |
+
},
|
| 3346 |
+
{
|
| 3347 |
+
"epoch": 0.5826674861708666,
|
| 3348 |
+
"grad_norm": 2.4591336250305176,
|
| 3349 |
+
"learning_rate": 4.1295064977829334e-05,
|
| 3350 |
+
"loss": 0.7384,
|
| 3351 |
+
"step": 474
|
| 3352 |
+
},
|
| 3353 |
+
{
|
| 3354 |
+
"epoch": 0.5838967424708051,
|
| 3355 |
+
"grad_norm": 2.414395809173584,
|
| 3356 |
+
"learning_rate": 4.109241228080115e-05,
|
| 3357 |
+
"loss": 0.7307,
|
| 3358 |
+
"step": 475
|
| 3359 |
+
},
|
| 3360 |
+
{
|
| 3361 |
+
"epoch": 0.5851259987707437,
|
| 3362 |
+
"grad_norm": 0.2034798115491867,
|
| 3363 |
+
"learning_rate": 4.088991059532248e-05,
|
| 3364 |
+
"loss": 0.8326,
|
| 3365 |
+
"step": 476
|
| 3366 |
+
},
|
| 3367 |
+
{
|
| 3368 |
+
"epoch": 0.5863552550706822,
|
| 3369 |
+
"grad_norm": 0.24682076275348663,
|
| 3370 |
+
"learning_rate": 4.0687563354431984e-05,
|
| 3371 |
+
"loss": 1.3382,
|
| 3372 |
+
"step": 477
|
| 3373 |
+
},
|
| 3374 |
+
{
|
| 3375 |
+
"epoch": 0.5875845113706207,
|
| 3376 |
+
"grad_norm": 0.2430254966020584,
|
| 3377 |
+
"learning_rate": 4.048537398855003e-05,
|
| 3378 |
+
"loss": 1.2718,
|
| 3379 |
+
"step": 478
|
| 3380 |
+
},
|
| 3381 |
+
{
|
| 3382 |
+
"epoch": 0.5888137676705593,
|
| 3383 |
+
"grad_norm": 0.27249017357826233,
|
| 3384 |
+
"learning_rate": 4.028334592542054e-05,
|
| 3385 |
+
"loss": 1.2328,
|
| 3386 |
+
"step": 479
|
| 3387 |
+
},
|
| 3388 |
+
{
|
| 3389 |
+
"epoch": 0.5900430239704979,
|
| 3390 |
+
"grad_norm": 0.28996631503105164,
|
| 3391 |
+
"learning_rate": 4.008148259005279e-05,
|
| 3392 |
+
"loss": 1.3963,
|
| 3393 |
+
"step": 480
|
| 3394 |
+
},
|
| 3395 |
+
{
|
| 3396 |
+
"epoch": 0.5912722802704364,
|
| 3397 |
+
"grad_norm": 0.28658372163772583,
|
| 3398 |
+
"learning_rate": 3.9879787404663474e-05,
|
| 3399 |
+
"loss": 1.2326,
|
| 3400 |
+
"step": 481
|
| 3401 |
+
},
|
| 3402 |
+
{
|
| 3403 |
+
"epoch": 0.592501536570375,
|
| 3404 |
+
"grad_norm": 0.3246002495288849,
|
| 3405 |
+
"learning_rate": 3.967826378861854e-05,
|
| 3406 |
+
"loss": 1.1071,
|
| 3407 |
+
"step": 482
|
| 3408 |
+
},
|
| 3409 |
+
{
|
| 3410 |
+
"epoch": 0.5937307928703135,
|
| 3411 |
+
"grad_norm": 0.33610400557518005,
|
| 3412 |
+
"learning_rate": 3.947691515837537e-05,
|
| 3413 |
+
"loss": 1.1798,
|
| 3414 |
+
"step": 483
|
| 3415 |
+
},
|
| 3416 |
+
{
|
| 3417 |
+
"epoch": 0.594960049170252,
|
| 3418 |
+
"grad_norm": 0.35427945852279663,
|
| 3419 |
+
"learning_rate": 3.927574492742473e-05,
|
| 3420 |
+
"loss": 1.3277,
|
| 3421 |
+
"step": 484
|
| 3422 |
+
},
|
| 3423 |
+
{
|
| 3424 |
+
"epoch": 0.5961893054701906,
|
| 3425 |
+
"grad_norm": 0.3942146599292755,
|
| 3426 |
+
"learning_rate": 3.907475650623299e-05,
|
| 3427 |
+
"loss": 1.0499,
|
| 3428 |
+
"step": 485
|
| 3429 |
+
},
|
| 3430 |
+
{
|
| 3431 |
+
"epoch": 0.5974185617701291,
|
| 3432 |
+
"grad_norm": 0.36504921317100525,
|
| 3433 |
+
"learning_rate": 3.887395330218429e-05,
|
| 3434 |
+
"loss": 1.1876,
|
| 3435 |
+
"step": 486
|
| 3436 |
+
},
|
| 3437 |
+
{
|
| 3438 |
+
"epoch": 0.5986478180700676,
|
| 3439 |
+
"grad_norm": 0.45135965943336487,
|
| 3440 |
+
"learning_rate": 3.867333871952269e-05,
|
| 3441 |
+
"loss": 1.1775,
|
| 3442 |
+
"step": 487
|
| 3443 |
+
},
|
| 3444 |
+
{
|
| 3445 |
+
"epoch": 0.5998770743700061,
|
| 3446 |
+
"grad_norm": 0.6297984719276428,
|
| 3447 |
+
"learning_rate": 3.84729161592946e-05,
|
| 3448 |
+
"loss": 0.9756,
|
| 3449 |
+
"step": 488
|
| 3450 |
+
},
|
| 3451 |
+
{
|
| 3452 |
+
"epoch": 0.6011063306699447,
|
| 3453 |
+
"grad_norm": 0.6099340915679932,
|
| 3454 |
+
"learning_rate": 3.827268901929102e-05,
|
| 3455 |
+
"loss": 0.7669,
|
| 3456 |
+
"step": 489
|
| 3457 |
+
},
|
| 3458 |
+
{
|
| 3459 |
+
"epoch": 0.6023355869698832,
|
| 3460 |
+
"grad_norm": 0.7176752686500549,
|
| 3461 |
+
"learning_rate": 3.8072660693989967e-05,
|
| 3462 |
+
"loss": 0.5387,
|
| 3463 |
+
"step": 490
|
| 3464 |
+
},
|
| 3465 |
+
{
|
| 3466 |
+
"epoch": 0.6035648432698217,
|
| 3467 |
+
"grad_norm": 0.6262489557266235,
|
| 3468 |
+
"learning_rate": 3.78728345744989e-05,
|
| 3469 |
+
"loss": 0.9306,
|
| 3470 |
+
"step": 491
|
| 3471 |
+
},
|
| 3472 |
+
{
|
| 3473 |
+
"epoch": 0.6047940995697603,
|
| 3474 |
+
"grad_norm": 1.0194954872131348,
|
| 3475 |
+
"learning_rate": 3.767321404849727e-05,
|
| 3476 |
+
"loss": 1.1677,
|
| 3477 |
+
"step": 492
|
| 3478 |
+
},
|
| 3479 |
+
{
|
| 3480 |
+
"epoch": 0.6060233558696988,
|
| 3481 |
+
"grad_norm": 1.6236090660095215,
|
| 3482 |
+
"learning_rate": 3.7473802500179114e-05,
|
| 3483 |
+
"loss": 1.0458,
|
| 3484 |
+
"step": 493
|
| 3485 |
+
},
|
| 3486 |
+
{
|
| 3487 |
+
"epoch": 0.6072526121696373,
|
| 3488 |
+
"grad_norm": 2.5714359283447266,
|
| 3489 |
+
"learning_rate": 3.727460331019559e-05,
|
| 3490 |
+
"loss": 1.2519,
|
| 3491 |
+
"step": 494
|
| 3492 |
+
},
|
| 3493 |
+
{
|
| 3494 |
+
"epoch": 0.6084818684695759,
|
| 3495 |
+
"grad_norm": 2.4615423679351807,
|
| 3496 |
+
"learning_rate": 3.7075619855597744e-05,
|
| 3497 |
+
"loss": 0.5842,
|
| 3498 |
+
"step": 495
|
| 3499 |
+
},
|
| 3500 |
+
{
|
| 3501 |
+
"epoch": 0.6097111247695144,
|
| 3502 |
+
"grad_norm": 3.4131109714508057,
|
| 3503 |
+
"learning_rate": 3.6876855509779206e-05,
|
| 3504 |
+
"loss": 1.3071,
|
| 3505 |
+
"step": 496
|
| 3506 |
+
},
|
| 3507 |
+
{
|
| 3508 |
+
"epoch": 0.610940381069453,
|
| 3509 |
+
"grad_norm": 3.461987257003784,
|
| 3510 |
+
"learning_rate": 3.667831364241904e-05,
|
| 3511 |
+
"loss": 1.1426,
|
| 3512 |
+
"step": 497
|
| 3513 |
+
},
|
| 3514 |
+
{
|
| 3515 |
+
"epoch": 0.6121696373693916,
|
| 3516 |
+
"grad_norm": 2.462989330291748,
|
| 3517 |
+
"learning_rate": 3.6479997619424605e-05,
|
| 3518 |
+
"loss": 0.8421,
|
| 3519 |
+
"step": 498
|
| 3520 |
+
},
|
| 3521 |
+
{
|
| 3522 |
+
"epoch": 0.6133988936693301,
|
| 3523 |
+
"grad_norm": 2.423410654067993,
|
| 3524 |
+
"learning_rate": 3.628191080287451e-05,
|
| 3525 |
+
"loss": 0.7845,
|
| 3526 |
+
"step": 499
|
| 3527 |
+
},
|
| 3528 |
+
{
|
| 3529 |
+
"epoch": 0.6146281499692686,
|
| 3530 |
+
"grad_norm": 2.5741212368011475,
|
| 3531 |
+
"learning_rate": 3.608405655096154e-05,
|
| 3532 |
+
"loss": 0.6916,
|
| 3533 |
+
"step": 500
|
| 3534 |
+
},
|
| 3535 |
+
{
|
| 3536 |
+
"epoch": 0.6158574062692072,
|
| 3537 |
+
"grad_norm": 0.2071269154548645,
|
| 3538 |
+
"learning_rate": 3.588643821793582e-05,
|
| 3539 |
+
"loss": 0.9648,
|
| 3540 |
+
"step": 501
|
| 3541 |
+
},
|
| 3542 |
+
{
|
| 3543 |
+
"epoch": 0.6170866625691457,
|
| 3544 |
+
"grad_norm": 0.23868902027606964,
|
| 3545 |
+
"learning_rate": 3.5689059154047915e-05,
|
| 3546 |
+
"loss": 1.261,
|
| 3547 |
+
"step": 502
|
| 3548 |
+
},
|
| 3549 |
+
{
|
| 3550 |
+
"epoch": 0.6183159188690842,
|
| 3551 |
+
"grad_norm": 0.2465473711490631,
|
| 3552 |
+
"learning_rate": 3.5491922705492e-05,
|
| 3553 |
+
"loss": 1.2999,
|
| 3554 |
+
"step": 503
|
| 3555 |
+
},
|
| 3556 |
+
{
|
| 3557 |
+
"epoch": 0.6195451751690227,
|
| 3558 |
+
"grad_norm": 0.25372591614723206,
|
| 3559 |
+
"learning_rate": 3.5295032214349196e-05,
|
| 3560 |
+
"loss": 1.3633,
|
| 3561 |
+
"step": 504
|
| 3562 |
+
},
|
| 3563 |
+
{
|
| 3564 |
+
"epoch": 0.6207744314689613,
|
| 3565 |
+
"grad_norm": 0.24923810362815857,
|
| 3566 |
+
"learning_rate": 3.5098391018530816e-05,
|
| 3567 |
+
"loss": 1.1408,
|
| 3568 |
+
"step": 505
|
| 3569 |
+
},
|
| 3570 |
+
{
|
| 3571 |
+
"epoch": 0.6220036877688998,
|
| 3572 |
+
"grad_norm": 0.2594156563282013,
|
| 3573 |
+
"learning_rate": 3.4902002451721916e-05,
|
| 3574 |
+
"loss": 1.158,
|
| 3575 |
+
"step": 506
|
| 3576 |
+
},
|
| 3577 |
+
{
|
| 3578 |
+
"epoch": 0.6232329440688383,
|
| 3579 |
+
"grad_norm": 0.28343746066093445,
|
| 3580 |
+
"learning_rate": 3.4705869843324614e-05,
|
| 3581 |
+
"loss": 1.1717,
|
| 3582 |
+
"step": 507
|
| 3583 |
+
},
|
| 3584 |
+
{
|
| 3585 |
+
"epoch": 0.6244622003687769,
|
| 3586 |
+
"grad_norm": 0.29203376173973083,
|
| 3587 |
+
"learning_rate": 3.450999651840179e-05,
|
| 3588 |
+
"loss": 1.0693,
|
| 3589 |
+
"step": 508
|
| 3590 |
+
},
|
| 3591 |
+
{
|
| 3592 |
+
"epoch": 0.6256914566687154,
|
| 3593 |
+
"grad_norm": 0.3236950933933258,
|
| 3594 |
+
"learning_rate": 3.431438579762066e-05,
|
| 3595 |
+
"loss": 1.0098,
|
| 3596 |
+
"step": 509
|
| 3597 |
+
},
|
| 3598 |
+
{
|
| 3599 |
+
"epoch": 0.6269207129686539,
|
| 3600 |
+
"grad_norm": 0.3463621139526367,
|
| 3601 |
+
"learning_rate": 3.411904099719642e-05,
|
| 3602 |
+
"loss": 1.1123,
|
| 3603 |
+
"step": 510
|
| 3604 |
+
},
|
| 3605 |
+
{
|
| 3606 |
+
"epoch": 0.6281499692685925,
|
| 3607 |
+
"grad_norm": 0.4081440269947052,
|
| 3608 |
+
"learning_rate": 3.3923965428836105e-05,
|
| 3609 |
+
"loss": 1.1303,
|
| 3610 |
+
"step": 511
|
| 3611 |
+
},
|
| 3612 |
+
{
|
| 3613 |
+
"epoch": 0.629379225568531,
|
| 3614 |
+
"grad_norm": 0.38460928201675415,
|
| 3615 |
+
"learning_rate": 3.3729162399682456e-05,
|
| 3616 |
+
"loss": 1.0858,
|
| 3617 |
+
"step": 512
|
| 3618 |
+
},
|
| 3619 |
+
{
|
| 3620 |
+
"epoch": 0.6306084818684696,
|
| 3621 |
+
"grad_norm": 0.474970281124115,
|
| 3622 |
+
"learning_rate": 3.35346352122578e-05,
|
| 3623 |
+
"loss": 1.1176,
|
| 3624 |
+
"step": 513
|
| 3625 |
+
},
|
| 3626 |
+
{
|
| 3627 |
+
"epoch": 0.6318377381684082,
|
| 3628 |
+
"grad_norm": 0.7059311270713806,
|
| 3629 |
+
"learning_rate": 3.3340387164408046e-05,
|
| 3630 |
+
"loss": 1.1786,
|
| 3631 |
+
"step": 514
|
| 3632 |
+
},
|
| 3633 |
+
{
|
| 3634 |
+
"epoch": 0.6330669944683467,
|
| 3635 |
+
"grad_norm": 0.7394426465034485,
|
| 3636 |
+
"learning_rate": 3.314642154924686e-05,
|
| 3637 |
+
"loss": 0.7836,
|
| 3638 |
+
"step": 515
|
| 3639 |
+
},
|
| 3640 |
+
{
|
| 3641 |
+
"epoch": 0.6342962507682852,
|
| 3642 |
+
"grad_norm": 0.766044020652771,
|
| 3643 |
+
"learning_rate": 3.295274165509979e-05,
|
| 3644 |
+
"loss": 0.8919,
|
| 3645 |
+
"step": 516
|
| 3646 |
+
},
|
| 3647 |
+
{
|
| 3648 |
+
"epoch": 0.6355255070682237,
|
| 3649 |
+
"grad_norm": 0.7920165061950684,
|
| 3650 |
+
"learning_rate": 3.275935076544845e-05,
|
| 3651 |
+
"loss": 1.0439,
|
| 3652 |
+
"step": 517
|
| 3653 |
+
},
|
| 3654 |
+
{
|
| 3655 |
+
"epoch": 0.6367547633681623,
|
| 3656 |
+
"grad_norm": 1.224188208580017,
|
| 3657 |
+
"learning_rate": 3.256625215887502e-05,
|
| 3658 |
+
"loss": 1.3103,
|
| 3659 |
+
"step": 518
|
| 3660 |
+
},
|
| 3661 |
+
{
|
| 3662 |
+
"epoch": 0.6379840196681008,
|
| 3663 |
+
"grad_norm": 2.0128793716430664,
|
| 3664 |
+
"learning_rate": 3.237344910900648e-05,
|
| 3665 |
+
"loss": 1.0804,
|
| 3666 |
+
"step": 519
|
| 3667 |
+
},
|
| 3668 |
+
{
|
| 3669 |
+
"epoch": 0.6392132759680393,
|
| 3670 |
+
"grad_norm": 3.370530366897583,
|
| 3671 |
+
"learning_rate": 3.218094488445923e-05,
|
| 3672 |
+
"loss": 0.5978,
|
| 3673 |
+
"step": 520
|
| 3674 |
+
},
|
| 3675 |
+
{
|
| 3676 |
+
"epoch": 0.6404425322679779,
|
| 3677 |
+
"grad_norm": 3.095421314239502,
|
| 3678 |
+
"learning_rate": 3.198874274878365e-05,
|
| 3679 |
+
"loss": 0.847,
|
| 3680 |
+
"step": 521
|
| 3681 |
+
},
|
| 3682 |
+
{
|
| 3683 |
+
"epoch": 0.6416717885679164,
|
| 3684 |
+
"grad_norm": 3.4100868701934814,
|
| 3685 |
+
"learning_rate": 3.179684596040878e-05,
|
| 3686 |
+
"loss": 0.8489,
|
| 3687 |
+
"step": 522
|
| 3688 |
+
},
|
| 3689 |
+
{
|
| 3690 |
+
"epoch": 0.6429010448678549,
|
| 3691 |
+
"grad_norm": 2.608668088912964,
|
| 3692 |
+
"learning_rate": 3.1605257772587035e-05,
|
| 3693 |
+
"loss": 0.8378,
|
| 3694 |
+
"step": 523
|
| 3695 |
+
},
|
| 3696 |
+
{
|
| 3697 |
+
"epoch": 0.6441303011677935,
|
| 3698 |
+
"grad_norm": 2.300096273422241,
|
| 3699 |
+
"learning_rate": 3.141398143333907e-05,
|
| 3700 |
+
"loss": 0.7825,
|
| 3701 |
+
"step": 524
|
| 3702 |
+
},
|
| 3703 |
+
{
|
| 3704 |
+
"epoch": 0.645359557467732,
|
| 3705 |
+
"grad_norm": 2.652704954147339,
|
| 3706 |
+
"learning_rate": 3.1223020185398765e-05,
|
| 3707 |
+
"loss": 0.6505,
|
| 3708 |
+
"step": 525
|
| 3709 |
+
},
|
| 3710 |
+
{
|
| 3711 |
+
"epoch": 0.6465888137676705,
|
| 3712 |
+
"grad_norm": 0.19123272597789764,
|
| 3713 |
+
"learning_rate": 3.103237726615822e-05,
|
| 3714 |
+
"loss": 0.8317,
|
| 3715 |
+
"step": 526
|
| 3716 |
+
},
|
| 3717 |
+
{
|
| 3718 |
+
"epoch": 0.647818070067609,
|
| 3719 |
+
"grad_norm": 0.22544066607952118,
|
| 3720 |
+
"learning_rate": 3.084205590761284e-05,
|
| 3721 |
+
"loss": 1.1174,
|
| 3722 |
+
"step": 527
|
| 3723 |
+
},
|
| 3724 |
+
{
|
| 3725 |
+
"epoch": 0.6490473263675476,
|
| 3726 |
+
"grad_norm": 0.25559771060943604,
|
| 3727 |
+
"learning_rate": 3.065205933630655e-05,
|
| 3728 |
+
"loss": 1.258,
|
| 3729 |
+
"step": 528
|
| 3730 |
+
},
|
| 3731 |
+
{
|
| 3732 |
+
"epoch": 0.6502765826674862,
|
| 3733 |
+
"grad_norm": 0.2453518509864807,
|
| 3734 |
+
"learning_rate": 3.0462390773277154e-05,
|
| 3735 |
+
"loss": 1.1197,
|
| 3736 |
+
"step": 529
|
| 3737 |
+
},
|
| 3738 |
+
{
|
| 3739 |
+
"epoch": 0.6515058389674248,
|
| 3740 |
+
"grad_norm": 0.26897767186164856,
|
| 3741 |
+
"learning_rate": 3.0273053434001662e-05,
|
| 3742 |
+
"loss": 1.2231,
|
| 3743 |
+
"step": 530
|
| 3744 |
+
},
|
| 3745 |
+
{
|
| 3746 |
+
"epoch": 0.6527350952673633,
|
| 3747 |
+
"grad_norm": 0.2956353425979614,
|
| 3748 |
+
"learning_rate": 3.0084050528341824e-05,
|
| 3749 |
+
"loss": 0.9809,
|
| 3750 |
+
"step": 531
|
| 3751 |
+
},
|
| 3752 |
+
{
|
| 3753 |
+
"epoch": 0.6539643515673018,
|
| 3754 |
+
"grad_norm": 0.3043264150619507,
|
| 3755 |
+
"learning_rate": 2.989538526048968e-05,
|
| 3756 |
+
"loss": 1.2625,
|
| 3757 |
+
"step": 532
|
| 3758 |
+
},
|
| 3759 |
+
{
|
| 3760 |
+
"epoch": 0.6551936078672403,
|
| 3761 |
+
"grad_norm": 0.3115137219429016,
|
| 3762 |
+
"learning_rate": 2.9707060828913225e-05,
|
| 3763 |
+
"loss": 1.1544,
|
| 3764 |
+
"step": 533
|
| 3765 |
+
},
|
| 3766 |
+
{
|
| 3767 |
+
"epoch": 0.6564228641671789,
|
| 3768 |
+
"grad_norm": 0.29051673412323,
|
| 3769 |
+
"learning_rate": 2.9519080426302238e-05,
|
| 3770 |
+
"loss": 0.8664,
|
| 3771 |
+
"step": 534
|
| 3772 |
+
},
|
| 3773 |
+
{
|
| 3774 |
+
"epoch": 0.6576521204671174,
|
| 3775 |
+
"grad_norm": 0.31704720854759216,
|
| 3776 |
+
"learning_rate": 2.933144723951414e-05,
|
| 3777 |
+
"loss": 0.8553,
|
| 3778 |
+
"step": 535
|
| 3779 |
+
},
|
| 3780 |
+
{
|
| 3781 |
+
"epoch": 0.6588813767670559,
|
| 3782 |
+
"grad_norm": 0.3343563675880432,
|
| 3783 |
+
"learning_rate": 2.9144164449519917e-05,
|
| 3784 |
+
"loss": 1.042,
|
| 3785 |
+
"step": 536
|
| 3786 |
+
},
|
| 3787 |
+
{
|
| 3788 |
+
"epoch": 0.6601106330669945,
|
| 3789 |
+
"grad_norm": 0.411081999540329,
|
| 3790 |
+
"learning_rate": 2.895723523135028e-05,
|
| 3791 |
+
"loss": 1.2587,
|
| 3792 |
+
"step": 537
|
| 3793 |
+
},
|
| 3794 |
+
{
|
| 3795 |
+
"epoch": 0.661339889366933,
|
| 3796 |
+
"grad_norm": 0.4190918505191803,
|
| 3797 |
+
"learning_rate": 2.877066275404172e-05,
|
| 3798 |
+
"loss": 1.1546,
|
| 3799 |
+
"step": 538
|
| 3800 |
+
},
|
| 3801 |
+
{
|
| 3802 |
+
"epoch": 0.6625691456668715,
|
| 3803 |
+
"grad_norm": 0.5606351494789124,
|
| 3804 |
+
"learning_rate": 2.8584450180582912e-05,
|
| 3805 |
+
"loss": 0.9889,
|
| 3806 |
+
"step": 539
|
| 3807 |
+
},
|
| 3808 |
+
{
|
| 3809 |
+
"epoch": 0.6637984019668101,
|
| 3810 |
+
"grad_norm": 0.6339558959007263,
|
| 3811 |
+
"learning_rate": 2.839860066786103e-05,
|
| 3812 |
+
"loss": 0.5278,
|
| 3813 |
+
"step": 540
|
| 3814 |
+
},
|
| 3815 |
+
{
|
| 3816 |
+
"epoch": 0.6650276582667486,
|
| 3817 |
+
"grad_norm": 0.6629421710968018,
|
| 3818 |
+
"learning_rate": 2.8213117366608188e-05,
|
| 3819 |
+
"loss": 0.752,
|
| 3820 |
+
"step": 541
|
| 3821 |
+
},
|
| 3822 |
+
{
|
| 3823 |
+
"epoch": 0.6662569145666871,
|
| 3824 |
+
"grad_norm": 0.7566842436790466,
|
| 3825 |
+
"learning_rate": 2.802800342134807e-05,
|
| 3826 |
+
"loss": 1.0395,
|
| 3827 |
+
"step": 542
|
| 3828 |
+
},
|
| 3829 |
+
{
|
| 3830 |
+
"epoch": 0.6674861708666256,
|
| 3831 |
+
"grad_norm": 1.169986605644226,
|
| 3832 |
+
"learning_rate": 2.784326197034266e-05,
|
| 3833 |
+
"loss": 1.1942,
|
| 3834 |
+
"step": 543
|
| 3835 |
+
},
|
| 3836 |
+
{
|
| 3837 |
+
"epoch": 0.6687154271665642,
|
| 3838 |
+
"grad_norm": 2.78249192237854,
|
| 3839 |
+
"learning_rate": 2.7658896145538983e-05,
|
| 3840 |
+
"loss": 1.1694,
|
| 3841 |
+
"step": 544
|
| 3842 |
+
},
|
| 3843 |
+
{
|
| 3844 |
+
"epoch": 0.6699446834665027,
|
| 3845 |
+
"grad_norm": 2.5933189392089844,
|
| 3846 |
+
"learning_rate": 2.7474909072515993e-05,
|
| 3847 |
+
"loss": 0.9325,
|
| 3848 |
+
"step": 545
|
| 3849 |
+
},
|
| 3850 |
+
{
|
| 3851 |
+
"epoch": 0.6711739397664414,
|
| 3852 |
+
"grad_norm": 2.37711238861084,
|
| 3853 |
+
"learning_rate": 2.7291303870431662e-05,
|
| 3854 |
+
"loss": 0.9246,
|
| 3855 |
+
"step": 546
|
| 3856 |
+
},
|
| 3857 |
+
{
|
| 3858 |
+
"epoch": 0.6724031960663799,
|
| 3859 |
+
"grad_norm": 3.3568241596221924,
|
| 3860 |
+
"learning_rate": 2.710808365197e-05,
|
| 3861 |
+
"loss": 0.9492,
|
| 3862 |
+
"step": 547
|
| 3863 |
+
},
|
| 3864 |
+
{
|
| 3865 |
+
"epoch": 0.6736324523663184,
|
| 3866 |
+
"grad_norm": 2.975985527038574,
|
| 3867 |
+
"learning_rate": 2.6925251523288346e-05,
|
| 3868 |
+
"loss": 0.8018,
|
| 3869 |
+
"step": 548
|
| 3870 |
+
},
|
| 3871 |
+
{
|
| 3872 |
+
"epoch": 0.6748617086662569,
|
| 3873 |
+
"grad_norm": 2.78823184967041,
|
| 3874 |
+
"learning_rate": 2.674281058396473e-05,
|
| 3875 |
+
"loss": 0.843,
|
| 3876 |
+
"step": 549
|
| 3877 |
+
},
|
| 3878 |
+
{
|
| 3879 |
+
"epoch": 0.6760909649661955,
|
| 3880 |
+
"grad_norm": 4.179988384246826,
|
| 3881 |
+
"learning_rate": 2.6560763926945275e-05,
|
| 3882 |
+
"loss": 1.0288,
|
| 3883 |
+
"step": 550
|
| 3884 |
+
},
|
| 3885 |
+
{
|
| 3886 |
+
"epoch": 0.677320221266134,
|
| 3887 |
+
"grad_norm": 0.21164661645889282,
|
| 3888 |
+
"learning_rate": 2.6379114638491807e-05,
|
| 3889 |
+
"loss": 0.9468,
|
| 3890 |
+
"step": 551
|
| 3891 |
+
},
|
| 3892 |
+
{
|
| 3893 |
+
"epoch": 0.6785494775660725,
|
| 3894 |
+
"grad_norm": 0.2241944521665573,
|
| 3895 |
+
"learning_rate": 2.6197865798129462e-05,
|
| 3896 |
+
"loss": 1.1578,
|
| 3897 |
+
"step": 552
|
| 3898 |
+
},
|
| 3899 |
+
{
|
| 3900 |
+
"epoch": 0.6797787338660111,
|
| 3901 |
+
"grad_norm": 0.25557973980903625,
|
| 3902 |
+
"learning_rate": 2.601702047859455e-05,
|
| 3903 |
+
"loss": 1.2432,
|
| 3904 |
+
"step": 553
|
| 3905 |
+
},
|
| 3906 |
+
{
|
| 3907 |
+
"epoch": 0.6810079901659496,
|
| 3908 |
+
"grad_norm": 0.2733787000179291,
|
| 3909 |
+
"learning_rate": 2.5836581745782475e-05,
|
| 3910 |
+
"loss": 1.2624,
|
| 3911 |
+
"step": 554
|
| 3912 |
+
},
|
| 3913 |
+
{
|
| 3914 |
+
"epoch": 0.6822372464658881,
|
| 3915 |
+
"grad_norm": 0.29645049571990967,
|
| 3916 |
+
"learning_rate": 2.5656552658695642e-05,
|
| 3917 |
+
"loss": 1.2544,
|
| 3918 |
+
"step": 555
|
| 3919 |
+
},
|
| 3920 |
+
{
|
| 3921 |
+
"epoch": 0.6834665027658267,
|
| 3922 |
+
"grad_norm": 0.29211127758026123,
|
| 3923 |
+
"learning_rate": 2.5476936269391726e-05,
|
| 3924 |
+
"loss": 1.0548,
|
| 3925 |
+
"step": 556
|
| 3926 |
+
},
|
| 3927 |
+
{
|
| 3928 |
+
"epoch": 0.6846957590657652,
|
| 3929 |
+
"grad_norm": 0.3013113737106323,
|
| 3930 |
+
"learning_rate": 2.5297735622931874e-05,
|
| 3931 |
+
"loss": 1.0941,
|
| 3932 |
+
"step": 557
|
| 3933 |
+
},
|
| 3934 |
+
{
|
| 3935 |
+
"epoch": 0.6859250153657037,
|
| 3936 |
+
"grad_norm": 0.33897635340690613,
|
| 3937 |
+
"learning_rate": 2.5118953757329088e-05,
|
| 3938 |
+
"loss": 1.0855,
|
| 3939 |
+
"step": 558
|
| 3940 |
+
},
|
| 3941 |
+
{
|
| 3942 |
+
"epoch": 0.6871542716656422,
|
| 3943 |
+
"grad_norm": 0.36230790615081787,
|
| 3944 |
+
"learning_rate": 2.494059370349673e-05,
|
| 3945 |
+
"loss": 1.0817,
|
| 3946 |
+
"step": 559
|
| 3947 |
+
},
|
| 3948 |
+
{
|
| 3949 |
+
"epoch": 0.6883835279655808,
|
| 3950 |
+
"grad_norm": 0.36360612511634827,
|
| 3951 |
+
"learning_rate": 2.4762658485197123e-05,
|
| 3952 |
+
"loss": 1.0868,
|
| 3953 |
+
"step": 560
|
| 3954 |
+
},
|
| 3955 |
+
{
|
| 3956 |
+
"epoch": 0.6896127842655193,
|
| 3957 |
+
"grad_norm": 0.45407038927078247,
|
| 3958 |
+
"learning_rate": 2.4585151118990286e-05,
|
| 3959 |
+
"loss": 1.1186,
|
| 3960 |
+
"step": 561
|
| 3961 |
+
},
|
| 3962 |
+
{
|
| 3963 |
+
"epoch": 0.690842040565458,
|
| 3964 |
+
"grad_norm": 0.5940976142883301,
|
| 3965 |
+
"learning_rate": 2.4408074614182773e-05,
|
| 3966 |
+
"loss": 1.1448,
|
| 3967 |
+
"step": 562
|
| 3968 |
+
},
|
| 3969 |
+
{
|
| 3970 |
+
"epoch": 0.6920712968653965,
|
| 3971 |
+
"grad_norm": 0.6550815105438232,
|
| 3972 |
+
"learning_rate": 2.4231431972776758e-05,
|
| 3973 |
+
"loss": 0.931,
|
| 3974 |
+
"step": 563
|
| 3975 |
+
},
|
| 3976 |
+
{
|
| 3977 |
+
"epoch": 0.693300553165335,
|
| 3978 |
+
"grad_norm": 0.6394757032394409,
|
| 3979 |
+
"learning_rate": 2.4055226189419018e-05,
|
| 3980 |
+
"loss": 0.8543,
|
| 3981 |
+
"step": 564
|
| 3982 |
+
},
|
| 3983 |
+
{
|
| 3984 |
+
"epoch": 0.6945298094652735,
|
| 3985 |
+
"grad_norm": 0.6903315186500549,
|
| 3986 |
+
"learning_rate": 2.3879460251350255e-05,
|
| 3987 |
+
"loss": 0.8152,
|
| 3988 |
+
"step": 565
|
| 3989 |
+
},
|
| 3990 |
+
{
|
| 3991 |
+
"epoch": 0.6957590657652121,
|
| 3992 |
+
"grad_norm": 0.9499403238296509,
|
| 3993 |
+
"learning_rate": 2.3704137138354355e-05,
|
| 3994 |
+
"loss": 1.2532,
|
| 3995 |
+
"step": 566
|
| 3996 |
+
},
|
| 3997 |
+
{
|
| 3998 |
+
"epoch": 0.6969883220651506,
|
| 3999 |
+
"grad_norm": 1.2395949363708496,
|
| 4000 |
+
"learning_rate": 2.3529259822708e-05,
|
| 4001 |
+
"loss": 1.3153,
|
| 4002 |
+
"step": 567
|
| 4003 |
+
},
|
| 4004 |
+
{
|
| 4005 |
+
"epoch": 0.6982175783650891,
|
| 4006 |
+
"grad_norm": 2.5209338665008545,
|
| 4007 |
+
"learning_rate": 2.3354831269130133e-05,
|
| 4008 |
+
"loss": 1.4484,
|
| 4009 |
+
"step": 568
|
| 4010 |
+
},
|
| 4011 |
+
{
|
| 4012 |
+
"epoch": 0.6994468346650277,
|
| 4013 |
+
"grad_norm": 3.072936773300171,
|
| 4014 |
+
"learning_rate": 2.318085443473185e-05,
|
| 4015 |
+
"loss": 0.9566,
|
| 4016 |
+
"step": 569
|
| 4017 |
+
},
|
| 4018 |
+
{
|
| 4019 |
+
"epoch": 0.7006760909649662,
|
| 4020 |
+
"grad_norm": 2.40030574798584,
|
| 4021 |
+
"learning_rate": 2.300733226896612e-05,
|
| 4022 |
+
"loss": 0.6958,
|
| 4023 |
+
"step": 570
|
| 4024 |
+
},
|
| 4025 |
+
{
|
| 4026 |
+
"epoch": 0.7019053472649047,
|
| 4027 |
+
"grad_norm": 3.1356310844421387,
|
| 4028 |
+
"learning_rate": 2.2834267713577904e-05,
|
| 4029 |
+
"loss": 1.261,
|
| 4030 |
+
"step": 571
|
| 4031 |
+
},
|
| 4032 |
+
{
|
| 4033 |
+
"epoch": 0.7031346035648433,
|
| 4034 |
+
"grad_norm": 2.5506911277770996,
|
| 4035 |
+
"learning_rate": 2.2661663702554208e-05,
|
| 4036 |
+
"loss": 0.7489,
|
| 4037 |
+
"step": 572
|
| 4038 |
+
},
|
| 4039 |
+
{
|
| 4040 |
+
"epoch": 0.7043638598647818,
|
| 4041 |
+
"grad_norm": 2.605987548828125,
|
| 4042 |
+
"learning_rate": 2.2489523162074393e-05,
|
| 4043 |
+
"loss": 0.8642,
|
| 4044 |
+
"step": 573
|
| 4045 |
+
},
|
| 4046 |
+
{
|
| 4047 |
+
"epoch": 0.7055931161647203,
|
| 4048 |
+
"grad_norm": 2.3878705501556396,
|
| 4049 |
+
"learning_rate": 2.2317849010460507e-05,
|
| 4050 |
+
"loss": 0.559,
|
| 4051 |
+
"step": 574
|
| 4052 |
+
},
|
| 4053 |
+
{
|
| 4054 |
+
"epoch": 0.7068223724646588,
|
| 4055 |
+
"grad_norm": 4.309600353240967,
|
| 4056 |
+
"learning_rate": 2.2146644158127827e-05,
|
| 4057 |
+
"loss": 1.3898,
|
| 4058 |
+
"step": 575
|
| 4059 |
+
},
|
| 4060 |
+
{
|
| 4061 |
+
"epoch": 0.7080516287645974,
|
| 4062 |
+
"grad_norm": 0.2340036779642105,
|
| 4063 |
+
"learning_rate": 2.197591150753559e-05,
|
| 4064 |
+
"loss": 1.1237,
|
| 4065 |
+
"step": 576
|
| 4066 |
+
},
|
| 4067 |
+
{
|
| 4068 |
+
"epoch": 0.7092808850645359,
|
| 4069 |
+
"grad_norm": 0.2404404878616333,
|
| 4070 |
+
"learning_rate": 2.1805653953137707e-05,
|
| 4071 |
+
"loss": 1.2235,
|
| 4072 |
+
"step": 577
|
| 4073 |
+
},
|
| 4074 |
+
{
|
| 4075 |
+
"epoch": 0.7105101413644745,
|
| 4076 |
+
"grad_norm": 0.2546350657939911,
|
| 4077 |
+
"learning_rate": 2.1635874381333714e-05,
|
| 4078 |
+
"loss": 1.1931,
|
| 4079 |
+
"step": 578
|
| 4080 |
+
},
|
| 4081 |
+
{
|
| 4082 |
+
"epoch": 0.7117393976644131,
|
| 4083 |
+
"grad_norm": 0.28081899881362915,
|
| 4084 |
+
"learning_rate": 2.1466575670419876e-05,
|
| 4085 |
+
"loss": 1.1851,
|
| 4086 |
+
"step": 579
|
| 4087 |
+
},
|
| 4088 |
+
{
|
| 4089 |
+
"epoch": 0.7129686539643516,
|
| 4090 |
+
"grad_norm": 0.27532532811164856,
|
| 4091 |
+
"learning_rate": 2.1297760690540302e-05,
|
| 4092 |
+
"loss": 1.0733,
|
| 4093 |
+
"step": 580
|
| 4094 |
+
},
|
| 4095 |
+
{
|
| 4096 |
+
"epoch": 0.7141979102642901,
|
| 4097 |
+
"grad_norm": 0.26407289505004883,
|
| 4098 |
+
"learning_rate": 2.1129432303638352e-05,
|
| 4099 |
+
"loss": 1.0863,
|
| 4100 |
+
"step": 581
|
| 4101 |
+
},
|
| 4102 |
+
{
|
| 4103 |
+
"epoch": 0.7154271665642287,
|
| 4104 |
+
"grad_norm": 0.31387853622436523,
|
| 4105 |
+
"learning_rate": 2.0961593363408156e-05,
|
| 4106 |
+
"loss": 1.1808,
|
| 4107 |
+
"step": 582
|
| 4108 |
+
},
|
| 4109 |
+
{
|
| 4110 |
+
"epoch": 0.7166564228641672,
|
| 4111 |
+
"grad_norm": 0.3309866487979889,
|
| 4112 |
+
"learning_rate": 2.079424671524616e-05,
|
| 4113 |
+
"loss": 1.0741,
|
| 4114 |
+
"step": 583
|
| 4115 |
+
},
|
| 4116 |
+
{
|
| 4117 |
+
"epoch": 0.7178856791641057,
|
| 4118 |
+
"grad_norm": 0.300689697265625,
|
| 4119 |
+
"learning_rate": 2.0627395196202898e-05,
|
| 4120 |
+
"loss": 0.9483,
|
| 4121 |
+
"step": 584
|
| 4122 |
+
},
|
| 4123 |
+
{
|
| 4124 |
+
"epoch": 0.7191149354640443,
|
| 4125 |
+
"grad_norm": 0.3478316366672516,
|
| 4126 |
+
"learning_rate": 2.046104163493493e-05,
|
| 4127 |
+
"loss": 1.144,
|
| 4128 |
+
"step": 585
|
| 4129 |
+
},
|
| 4130 |
+
{
|
| 4131 |
+
"epoch": 0.7203441917639828,
|
| 4132 |
+
"grad_norm": 0.44858890771865845,
|
| 4133 |
+
"learning_rate": 2.0295188851656892e-05,
|
| 4134 |
+
"loss": 1.1384,
|
| 4135 |
+
"step": 586
|
| 4136 |
+
},
|
| 4137 |
+
{
|
| 4138 |
+
"epoch": 0.7215734480639213,
|
| 4139 |
+
"grad_norm": 0.4991127550601959,
|
| 4140 |
+
"learning_rate": 2.0129839658093607e-05,
|
| 4141 |
+
"loss": 1.3527,
|
| 4142 |
+
"step": 587
|
| 4143 |
+
},
|
| 4144 |
+
{
|
| 4145 |
+
"epoch": 0.7228027043638598,
|
| 4146 |
+
"grad_norm": 0.5499400496482849,
|
| 4147 |
+
"learning_rate": 1.996499685743254e-05,
|
| 4148 |
+
"loss": 1.2116,
|
| 4149 |
+
"step": 588
|
| 4150 |
+
},
|
| 4151 |
+
{
|
| 4152 |
+
"epoch": 0.7240319606637984,
|
| 4153 |
+
"grad_norm": 0.6611432433128357,
|
| 4154 |
+
"learning_rate": 1.980066324427613e-05,
|
| 4155 |
+
"loss": 0.5055,
|
| 4156 |
+
"step": 589
|
| 4157 |
+
},
|
| 4158 |
+
{
|
| 4159 |
+
"epoch": 0.7252612169637369,
|
| 4160 |
+
"grad_norm": 0.7008240222930908,
|
| 4161 |
+
"learning_rate": 1.9636841604594557e-05,
|
| 4162 |
+
"loss": 0.7072,
|
| 4163 |
+
"step": 590
|
| 4164 |
+
},
|
| 4165 |
+
{
|
| 4166 |
+
"epoch": 0.7264904732636754,
|
| 4167 |
+
"grad_norm": 0.7243725061416626,
|
| 4168 |
+
"learning_rate": 1.9473534715678427e-05,
|
| 4169 |
+
"loss": 0.8031,
|
| 4170 |
+
"step": 591
|
| 4171 |
+
},
|
| 4172 |
+
{
|
| 4173 |
+
"epoch": 0.727719729563614,
|
| 4174 |
+
"grad_norm": 0.8577756285667419,
|
| 4175 |
+
"learning_rate": 1.9310745346091714e-05,
|
| 4176 |
+
"loss": 1.1383,
|
| 4177 |
+
"step": 592
|
| 4178 |
+
},
|
| 4179 |
+
{
|
| 4180 |
+
"epoch": 0.7289489858635525,
|
| 4181 |
+
"grad_norm": 1.2171465158462524,
|
| 4182 |
+
"learning_rate": 1.9148476255624764e-05,
|
| 4183 |
+
"loss": 0.9733,
|
| 4184 |
+
"step": 593
|
| 4185 |
+
},
|
| 4186 |
+
{
|
| 4187 |
+
"epoch": 0.7301782421634911,
|
| 4188 |
+
"grad_norm": 3.537208318710327,
|
| 4189 |
+
"learning_rate": 1.898673019524764e-05,
|
| 4190 |
+
"loss": 1.1197,
|
| 4191 |
+
"step": 594
|
| 4192 |
+
},
|
| 4193 |
+
{
|
| 4194 |
+
"epoch": 0.7314074984634297,
|
| 4195 |
+
"grad_norm": 2.420771360397339,
|
| 4196 |
+
"learning_rate": 1.8825509907063327e-05,
|
| 4197 |
+
"loss": 0.7744,
|
| 4198 |
+
"step": 595
|
| 4199 |
+
},
|
| 4200 |
+
{
|
| 4201 |
+
"epoch": 0.7326367547633682,
|
| 4202 |
+
"grad_norm": 3.1013336181640625,
|
| 4203 |
+
"learning_rate": 1.8664818124261374e-05,
|
| 4204 |
+
"loss": 0.7931,
|
| 4205 |
+
"step": 596
|
| 4206 |
+
},
|
| 4207 |
+
{
|
| 4208 |
+
"epoch": 0.7338660110633067,
|
| 4209 |
+
"grad_norm": 3.5393359661102295,
|
| 4210 |
+
"learning_rate": 1.8504657571071515e-05,
|
| 4211 |
+
"loss": 1.0527,
|
| 4212 |
+
"step": 597
|
| 4213 |
+
},
|
| 4214 |
+
{
|
| 4215 |
+
"epoch": 0.7350952673632453,
|
| 4216 |
+
"grad_norm": 2.782148838043213,
|
| 4217 |
+
"learning_rate": 1.8345030962717407e-05,
|
| 4218 |
+
"loss": 0.8034,
|
| 4219 |
+
"step": 598
|
| 4220 |
+
},
|
| 4221 |
+
{
|
| 4222 |
+
"epoch": 0.7363245236631838,
|
| 4223 |
+
"grad_norm": 2.12009596824646,
|
| 4224 |
+
"learning_rate": 1.8185941005370745e-05,
|
| 4225 |
+
"loss": 0.4618,
|
| 4226 |
+
"step": 599
|
| 4227 |
+
},
|
| 4228 |
+
{
|
| 4229 |
+
"epoch": 0.7375537799631223,
|
| 4230 |
+
"grad_norm": 3.6186935901641846,
|
| 4231 |
+
"learning_rate": 1.802739039610527e-05,
|
| 4232 |
+
"loss": 0.793,
|
| 4233 |
+
"step": 600
|
| 4234 |
+
},
|
| 4235 |
+
{
|
| 4236 |
+
"epoch": 0.7387830362630609,
|
| 4237 |
+
"grad_norm": 0.22188998758792877,
|
| 4238 |
+
"learning_rate": 1.786938182285107e-05,
|
| 4239 |
+
"loss": 1.1901,
|
| 4240 |
+
"step": 601
|
| 4241 |
+
},
|
| 4242 |
+
{
|
| 4243 |
+
"epoch": 0.7400122925629994,
|
| 4244 |
+
"grad_norm": 0.2288677990436554,
|
| 4245 |
+
"learning_rate": 1.7711917964349062e-05,
|
| 4246 |
+
"loss": 1.1755,
|
| 4247 |
+
"step": 602
|
| 4248 |
+
},
|
| 4249 |
+
{
|
| 4250 |
+
"epoch": 0.7412415488629379,
|
| 4251 |
+
"grad_norm": 0.25471025705337524,
|
| 4252 |
+
"learning_rate": 1.7555001490105488e-05,
|
| 4253 |
+
"loss": 1.2616,
|
| 4254 |
+
"step": 603
|
| 4255 |
+
},
|
| 4256 |
+
{
|
| 4257 |
+
"epoch": 0.7424708051628764,
|
| 4258 |
+
"grad_norm": 0.23840656876564026,
|
| 4259 |
+
"learning_rate": 1.7398635060346746e-05,
|
| 4260 |
+
"loss": 1.0882,
|
| 4261 |
+
"step": 604
|
| 4262 |
+
},
|
| 4263 |
+
{
|
| 4264 |
+
"epoch": 0.743700061462815,
|
| 4265 |
+
"grad_norm": 0.2534352242946625,
|
| 4266 |
+
"learning_rate": 1.7242821325974258e-05,
|
| 4267 |
+
"loss": 1.1328,
|
| 4268 |
+
"step": 605
|
| 4269 |
+
},
|
| 4270 |
+
{
|
| 4271 |
+
"epoch": 0.7449293177627535,
|
| 4272 |
+
"grad_norm": 0.279732346534729,
|
| 4273 |
+
"learning_rate": 1.7087562928519514e-05,
|
| 4274 |
+
"loss": 1.1703,
|
| 4275 |
+
"step": 606
|
| 4276 |
+
},
|
| 4277 |
+
{
|
| 4278 |
+
"epoch": 0.746158574062692,
|
| 4279 |
+
"grad_norm": 0.3084731698036194,
|
| 4280 |
+
"learning_rate": 1.69328625000993e-05,
|
| 4281 |
+
"loss": 1.1646,
|
| 4282 |
+
"step": 607
|
| 4283 |
+
},
|
| 4284 |
+
{
|
| 4285 |
+
"epoch": 0.7473878303626306,
|
| 4286 |
+
"grad_norm": 0.3052552044391632,
|
| 4287 |
+
"learning_rate": 1.6778722663371053e-05,
|
| 4288 |
+
"loss": 1.1481,
|
| 4289 |
+
"step": 608
|
| 4290 |
+
},
|
| 4291 |
+
{
|
| 4292 |
+
"epoch": 0.7486170866625691,
|
| 4293 |
+
"grad_norm": 0.3502384126186371,
|
| 4294 |
+
"learning_rate": 1.662514603148847e-05,
|
| 4295 |
+
"loss": 1.0333,
|
| 4296 |
+
"step": 609
|
| 4297 |
+
},
|
| 4298 |
+
{
|
| 4299 |
+
"epoch": 0.7498463429625076,
|
| 4300 |
+
"grad_norm": 0.3393804728984833,
|
| 4301 |
+
"learning_rate": 1.6472135208057126e-05,
|
| 4302 |
+
"loss": 1.0056,
|
| 4303 |
+
"step": 610
|
| 4304 |
+
},
|
| 4305 |
+
{
|
| 4306 |
+
"epoch": 0.7510755992624463,
|
| 4307 |
+
"grad_norm": 0.3970140814781189,
|
| 4308 |
+
"learning_rate": 1.631969278709041e-05,
|
| 4309 |
+
"loss": 1.1957,
|
| 4310 |
+
"step": 611
|
| 4311 |
+
},
|
| 4312 |
+
{
|
| 4313 |
+
"epoch": 0.7523048555623848,
|
| 4314 |
+
"grad_norm": 0.4345152974128723,
|
| 4315 |
+
"learning_rate": 1.616782135296544e-05,
|
| 4316 |
+
"loss": 1.0145,
|
| 4317 |
+
"step": 612
|
| 4318 |
+
},
|
| 4319 |
+
{
|
| 4320 |
+
"epoch": 0.7523048555623848,
|
| 4321 |
+
"eval_loss": 0.9307632446289062,
|
| 4322 |
+
"eval_runtime": 64.7426,
|
| 4323 |
+
"eval_samples_per_second": 10.58,
|
| 4324 |
+
"eval_steps_per_second": 5.298,
|
| 4325 |
+
"step": 612
|
| 4326 |
}
|
| 4327 |
],
|
| 4328 |
"logging_steps": 1,
|
|
|
|
| 4342 |
"attributes": {}
|
| 4343 |
}
|
| 4344 |
},
|
| 4345 |
+
"total_flos": 1.261491430967214e+17,
|
| 4346 |
"train_batch_size": 1,
|
| 4347 |
"trial_name": null,
|
| 4348 |
"trial_params": null
|