Commit
•
88ac64a
1
Parent(s):
411cb77
Training in progress, epoch 5, 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 111475752
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:a44e3ee4a48b92e9fd6889e4bce40aeb655f5bb47d4d2f6ef7ac3dcf3857dbe4
|
| 3 |
size 111475752
|
last-checkpoint/optimizer.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 223212738
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:7f51ab0fd9137a0e08c0ff296c77c846e80363f3e180149de7ec71d491048b6c
|
| 3 |
size 223212738
|
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:efd50c45dad39ec69546c509423085891e75595e9e34cdb0627953a1082923b5
|
| 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 sha256:624a32e37d981af4a144a16cbf7efaf716d31484043b1f98eb31df69fa78222b
|
| 3 |
size 1064
|
last-checkpoint/trainer_state.json
CHANGED
|
@@ -1,9 +1,9 @@
|
|
| 1 |
{
|
| 2 |
-
"best_metric": 32.
|
| 3 |
-
"best_model_checkpoint": "./exp/whisper-large-v3-turbo-augmented/checkpoint-
|
| 4 |
-
"epoch":
|
| 5 |
"eval_steps": 500,
|
| 6 |
-
"global_step":
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
@@ -5371,6 +5371,1347 @@
|
|
| 5371 |
"eval_steps_per_second": 0.865,
|
| 5372 |
"eval_wer": 32.66918584678805,
|
| 5373 |
"step": 760
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 5374 |
}
|
| 5375 |
],
|
| 5376 |
"logging_steps": 1,
|
|
@@ -5385,12 +6726,12 @@
|
|
| 5385 |
"should_evaluate": false,
|
| 5386 |
"should_log": false,
|
| 5387 |
"should_save": true,
|
| 5388 |
-
"should_training_stop":
|
| 5389 |
},
|
| 5390 |
"attributes": {}
|
| 5391 |
}
|
| 5392 |
},
|
| 5393 |
-
"total_flos": 2.
|
| 5394 |
"train_batch_size": 16,
|
| 5395 |
"trial_name": null,
|
| 5396 |
"trial_params": null
|
|
|
|
| 1 |
{
|
| 2 |
+
"best_metric": 32.36001374098248,
|
| 3 |
+
"best_model_checkpoint": "./exp/whisper-large-v3-turbo-augmented/checkpoint-950",
|
| 4 |
+
"epoch": 5.0,
|
| 5 |
"eval_steps": 500,
|
| 6 |
+
"global_step": 950,
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
|
|
| 5371 |
"eval_steps_per_second": 0.865,
|
| 5372 |
"eval_wer": 32.66918584678805,
|
| 5373 |
"step": 760
|
| 5374 |
+
},
|
| 5375 |
+
{
|
| 5376 |
+
"epoch": 4.005263157894737,
|
| 5377 |
+
"grad_norm": 0.036575622856616974,
|
| 5378 |
+
"learning_rate": 9.947368421052632e-05,
|
| 5379 |
+
"loss": 0.0011,
|
| 5380 |
+
"step": 761
|
| 5381 |
+
},
|
| 5382 |
+
{
|
| 5383 |
+
"epoch": 4.010526315789473,
|
| 5384 |
+
"grad_norm": 0.008182702586054802,
|
| 5385 |
+
"learning_rate": 9.894736842105263e-05,
|
| 5386 |
+
"loss": 0.0005,
|
| 5387 |
+
"step": 762
|
| 5388 |
+
},
|
| 5389 |
+
{
|
| 5390 |
+
"epoch": 4.015789473684211,
|
| 5391 |
+
"grad_norm": 0.2233087420463562,
|
| 5392 |
+
"learning_rate": 9.842105263157896e-05,
|
| 5393 |
+
"loss": 0.0123,
|
| 5394 |
+
"step": 763
|
| 5395 |
+
},
|
| 5396 |
+
{
|
| 5397 |
+
"epoch": 4.021052631578947,
|
| 5398 |
+
"grad_norm": 0.018053319305181503,
|
| 5399 |
+
"learning_rate": 9.789473684210526e-05,
|
| 5400 |
+
"loss": 0.0011,
|
| 5401 |
+
"step": 764
|
| 5402 |
+
},
|
| 5403 |
+
{
|
| 5404 |
+
"epoch": 4.026315789473684,
|
| 5405 |
+
"grad_norm": 0.04312720522284508,
|
| 5406 |
+
"learning_rate": 9.736842105263158e-05,
|
| 5407 |
+
"loss": 0.0012,
|
| 5408 |
+
"step": 765
|
| 5409 |
+
},
|
| 5410 |
+
{
|
| 5411 |
+
"epoch": 4.031578947368421,
|
| 5412 |
+
"grad_norm": 0.008370849303901196,
|
| 5413 |
+
"learning_rate": 9.68421052631579e-05,
|
| 5414 |
+
"loss": 0.0006,
|
| 5415 |
+
"step": 766
|
| 5416 |
+
},
|
| 5417 |
+
{
|
| 5418 |
+
"epoch": 4.036842105263158,
|
| 5419 |
+
"grad_norm": 0.039640624076128006,
|
| 5420 |
+
"learning_rate": 9.631578947368422e-05,
|
| 5421 |
+
"loss": 0.0016,
|
| 5422 |
+
"step": 767
|
| 5423 |
+
},
|
| 5424 |
+
{
|
| 5425 |
+
"epoch": 4.042105263157895,
|
| 5426 |
+
"grad_norm": 0.014259401708841324,
|
| 5427 |
+
"learning_rate": 9.578947368421052e-05,
|
| 5428 |
+
"loss": 0.0008,
|
| 5429 |
+
"step": 768
|
| 5430 |
+
},
|
| 5431 |
+
{
|
| 5432 |
+
"epoch": 4.0473684210526315,
|
| 5433 |
+
"grad_norm": 0.12600314617156982,
|
| 5434 |
+
"learning_rate": 9.526315789473685e-05,
|
| 5435 |
+
"loss": 0.0048,
|
| 5436 |
+
"step": 769
|
| 5437 |
+
},
|
| 5438 |
+
{
|
| 5439 |
+
"epoch": 4.052631578947368,
|
| 5440 |
+
"grad_norm": 0.04803254455327988,
|
| 5441 |
+
"learning_rate": 9.473684210526316e-05,
|
| 5442 |
+
"loss": 0.0011,
|
| 5443 |
+
"step": 770
|
| 5444 |
+
},
|
| 5445 |
+
{
|
| 5446 |
+
"epoch": 4.057894736842106,
|
| 5447 |
+
"grad_norm": 0.04734767973423004,
|
| 5448 |
+
"learning_rate": 9.421052631578947e-05,
|
| 5449 |
+
"loss": 0.0016,
|
| 5450 |
+
"step": 771
|
| 5451 |
+
},
|
| 5452 |
+
{
|
| 5453 |
+
"epoch": 4.063157894736842,
|
| 5454 |
+
"grad_norm": 0.014112943783402443,
|
| 5455 |
+
"learning_rate": 9.368421052631579e-05,
|
| 5456 |
+
"loss": 0.0007,
|
| 5457 |
+
"step": 772
|
| 5458 |
+
},
|
| 5459 |
+
{
|
| 5460 |
+
"epoch": 4.068421052631579,
|
| 5461 |
+
"grad_norm": 0.01292451936751604,
|
| 5462 |
+
"learning_rate": 9.315789473684211e-05,
|
| 5463 |
+
"loss": 0.0007,
|
| 5464 |
+
"step": 773
|
| 5465 |
+
},
|
| 5466 |
+
{
|
| 5467 |
+
"epoch": 4.073684210526316,
|
| 5468 |
+
"grad_norm": 0.013298957608640194,
|
| 5469 |
+
"learning_rate": 9.263157894736843e-05,
|
| 5470 |
+
"loss": 0.0008,
|
| 5471 |
+
"step": 774
|
| 5472 |
+
},
|
| 5473 |
+
{
|
| 5474 |
+
"epoch": 4.078947368421052,
|
| 5475 |
+
"grad_norm": 0.01097803097218275,
|
| 5476 |
+
"learning_rate": 9.210526315789474e-05,
|
| 5477 |
+
"loss": 0.001,
|
| 5478 |
+
"step": 775
|
| 5479 |
+
},
|
| 5480 |
+
{
|
| 5481 |
+
"epoch": 4.08421052631579,
|
| 5482 |
+
"grad_norm": 0.007426980417221785,
|
| 5483 |
+
"learning_rate": 9.157894736842105e-05,
|
| 5484 |
+
"loss": 0.0005,
|
| 5485 |
+
"step": 776
|
| 5486 |
+
},
|
| 5487 |
+
{
|
| 5488 |
+
"epoch": 4.089473684210526,
|
| 5489 |
+
"grad_norm": 0.632396936416626,
|
| 5490 |
+
"learning_rate": 9.105263157894738e-05,
|
| 5491 |
+
"loss": 0.0093,
|
| 5492 |
+
"step": 777
|
| 5493 |
+
},
|
| 5494 |
+
{
|
| 5495 |
+
"epoch": 4.094736842105263,
|
| 5496 |
+
"grad_norm": 0.01156445313245058,
|
| 5497 |
+
"learning_rate": 9.052631578947369e-05,
|
| 5498 |
+
"loss": 0.0006,
|
| 5499 |
+
"step": 778
|
| 5500 |
+
},
|
| 5501 |
+
{
|
| 5502 |
+
"epoch": 4.1,
|
| 5503 |
+
"grad_norm": 0.07284267246723175,
|
| 5504 |
+
"learning_rate": 8.999999999999999e-05,
|
| 5505 |
+
"loss": 0.0018,
|
| 5506 |
+
"step": 779
|
| 5507 |
+
},
|
| 5508 |
+
{
|
| 5509 |
+
"epoch": 4.105263157894737,
|
| 5510 |
+
"grad_norm": 0.01863129995763302,
|
| 5511 |
+
"learning_rate": 8.947368421052632e-05,
|
| 5512 |
+
"loss": 0.0008,
|
| 5513 |
+
"step": 780
|
| 5514 |
+
},
|
| 5515 |
+
{
|
| 5516 |
+
"epoch": 4.110526315789474,
|
| 5517 |
+
"grad_norm": 0.02850412204861641,
|
| 5518 |
+
"learning_rate": 8.894736842105263e-05,
|
| 5519 |
+
"loss": 0.0016,
|
| 5520 |
+
"step": 781
|
| 5521 |
+
},
|
| 5522 |
+
{
|
| 5523 |
+
"epoch": 4.11578947368421,
|
| 5524 |
+
"grad_norm": 0.02035311982035637,
|
| 5525 |
+
"learning_rate": 8.842105263157894e-05,
|
| 5526 |
+
"loss": 0.0008,
|
| 5527 |
+
"step": 782
|
| 5528 |
+
},
|
| 5529 |
+
{
|
| 5530 |
+
"epoch": 4.121052631578947,
|
| 5531 |
+
"grad_norm": 0.017460035160183907,
|
| 5532 |
+
"learning_rate": 8.789473684210526e-05,
|
| 5533 |
+
"loss": 0.0012,
|
| 5534 |
+
"step": 783
|
| 5535 |
+
},
|
| 5536 |
+
{
|
| 5537 |
+
"epoch": 4.126315789473685,
|
| 5538 |
+
"grad_norm": 0.024095656350255013,
|
| 5539 |
+
"learning_rate": 8.736842105263158e-05,
|
| 5540 |
+
"loss": 0.0009,
|
| 5541 |
+
"step": 784
|
| 5542 |
+
},
|
| 5543 |
+
{
|
| 5544 |
+
"epoch": 4.131578947368421,
|
| 5545 |
+
"grad_norm": 0.009611714631319046,
|
| 5546 |
+
"learning_rate": 8.68421052631579e-05,
|
| 5547 |
+
"loss": 0.0005,
|
| 5548 |
+
"step": 785
|
| 5549 |
+
},
|
| 5550 |
+
{
|
| 5551 |
+
"epoch": 4.136842105263158,
|
| 5552 |
+
"grad_norm": 0.023250168189406395,
|
| 5553 |
+
"learning_rate": 8.631578947368421e-05,
|
| 5554 |
+
"loss": 0.001,
|
| 5555 |
+
"step": 786
|
| 5556 |
+
},
|
| 5557 |
+
{
|
| 5558 |
+
"epoch": 4.1421052631578945,
|
| 5559 |
+
"grad_norm": 0.09312010556459427,
|
| 5560 |
+
"learning_rate": 8.578947368421052e-05,
|
| 5561 |
+
"loss": 0.0032,
|
| 5562 |
+
"step": 787
|
| 5563 |
+
},
|
| 5564 |
+
{
|
| 5565 |
+
"epoch": 4.147368421052631,
|
| 5566 |
+
"grad_norm": 0.3276028633117676,
|
| 5567 |
+
"learning_rate": 8.526315789473685e-05,
|
| 5568 |
+
"loss": 0.0145,
|
| 5569 |
+
"step": 788
|
| 5570 |
+
},
|
| 5571 |
+
{
|
| 5572 |
+
"epoch": 4.152631578947369,
|
| 5573 |
+
"grad_norm": 0.013059087097644806,
|
| 5574 |
+
"learning_rate": 8.473684210526315e-05,
|
| 5575 |
+
"loss": 0.0009,
|
| 5576 |
+
"step": 789
|
| 5577 |
+
},
|
| 5578 |
+
{
|
| 5579 |
+
"epoch": 4.157894736842105,
|
| 5580 |
+
"grad_norm": 0.01524921040982008,
|
| 5581 |
+
"learning_rate": 8.421052631578948e-05,
|
| 5582 |
+
"loss": 0.0013,
|
| 5583 |
+
"step": 790
|
| 5584 |
+
},
|
| 5585 |
+
{
|
| 5586 |
+
"epoch": 4.163157894736842,
|
| 5587 |
+
"grad_norm": 0.013316730968654156,
|
| 5588 |
+
"learning_rate": 8.368421052631579e-05,
|
| 5589 |
+
"loss": 0.0007,
|
| 5590 |
+
"step": 791
|
| 5591 |
+
},
|
| 5592 |
+
{
|
| 5593 |
+
"epoch": 4.168421052631579,
|
| 5594 |
+
"grad_norm": 0.04916347563266754,
|
| 5595 |
+
"learning_rate": 8.315789473684212e-05,
|
| 5596 |
+
"loss": 0.0008,
|
| 5597 |
+
"step": 792
|
| 5598 |
+
},
|
| 5599 |
+
{
|
| 5600 |
+
"epoch": 4.173684210526316,
|
| 5601 |
+
"grad_norm": 0.007756952662020922,
|
| 5602 |
+
"learning_rate": 8.263157894736841e-05,
|
| 5603 |
+
"loss": 0.0005,
|
| 5604 |
+
"step": 793
|
| 5605 |
+
},
|
| 5606 |
+
{
|
| 5607 |
+
"epoch": 4.178947368421053,
|
| 5608 |
+
"grad_norm": 0.02531874179840088,
|
| 5609 |
+
"learning_rate": 8.210526315789474e-05,
|
| 5610 |
+
"loss": 0.0009,
|
| 5611 |
+
"step": 794
|
| 5612 |
+
},
|
| 5613 |
+
{
|
| 5614 |
+
"epoch": 4.184210526315789,
|
| 5615 |
+
"grad_norm": 0.08359131217002869,
|
| 5616 |
+
"learning_rate": 8.157894736842105e-05,
|
| 5617 |
+
"loss": 0.0012,
|
| 5618 |
+
"step": 795
|
| 5619 |
+
},
|
| 5620 |
+
{
|
| 5621 |
+
"epoch": 4.189473684210526,
|
| 5622 |
+
"grad_norm": 0.016808507964015007,
|
| 5623 |
+
"learning_rate": 8.105263157894737e-05,
|
| 5624 |
+
"loss": 0.0006,
|
| 5625 |
+
"step": 796
|
| 5626 |
+
},
|
| 5627 |
+
{
|
| 5628 |
+
"epoch": 4.1947368421052635,
|
| 5629 |
+
"grad_norm": 0.13686731457710266,
|
| 5630 |
+
"learning_rate": 8.052631578947368e-05,
|
| 5631 |
+
"loss": 0.0099,
|
| 5632 |
+
"step": 797
|
| 5633 |
+
},
|
| 5634 |
+
{
|
| 5635 |
+
"epoch": 4.2,
|
| 5636 |
+
"grad_norm": 0.009160361252725124,
|
| 5637 |
+
"learning_rate": 8e-05,
|
| 5638 |
+
"loss": 0.0007,
|
| 5639 |
+
"step": 798
|
| 5640 |
+
},
|
| 5641 |
+
{
|
| 5642 |
+
"epoch": 4.205263157894737,
|
| 5643 |
+
"grad_norm": 0.11400540173053741,
|
| 5644 |
+
"learning_rate": 7.947368421052632e-05,
|
| 5645 |
+
"loss": 0.0042,
|
| 5646 |
+
"step": 799
|
| 5647 |
+
},
|
| 5648 |
+
{
|
| 5649 |
+
"epoch": 4.2105263157894735,
|
| 5650 |
+
"grad_norm": 0.016534345224499702,
|
| 5651 |
+
"learning_rate": 7.894736842105263e-05,
|
| 5652 |
+
"loss": 0.0005,
|
| 5653 |
+
"step": 800
|
| 5654 |
+
},
|
| 5655 |
+
{
|
| 5656 |
+
"epoch": 4.215789473684211,
|
| 5657 |
+
"grad_norm": 0.06011210009455681,
|
| 5658 |
+
"learning_rate": 7.842105263157895e-05,
|
| 5659 |
+
"loss": 0.0016,
|
| 5660 |
+
"step": 801
|
| 5661 |
+
},
|
| 5662 |
+
{
|
| 5663 |
+
"epoch": 4.221052631578948,
|
| 5664 |
+
"grad_norm": 0.02660367079079151,
|
| 5665 |
+
"learning_rate": 7.789473684210527e-05,
|
| 5666 |
+
"loss": 0.0011,
|
| 5667 |
+
"step": 802
|
| 5668 |
+
},
|
| 5669 |
+
{
|
| 5670 |
+
"epoch": 4.226315789473684,
|
| 5671 |
+
"grad_norm": 0.010432078503072262,
|
| 5672 |
+
"learning_rate": 7.736842105263157e-05,
|
| 5673 |
+
"loss": 0.0006,
|
| 5674 |
+
"step": 803
|
| 5675 |
+
},
|
| 5676 |
+
{
|
| 5677 |
+
"epoch": 4.231578947368421,
|
| 5678 |
+
"grad_norm": 0.05256740376353264,
|
| 5679 |
+
"learning_rate": 7.68421052631579e-05,
|
| 5680 |
+
"loss": 0.0017,
|
| 5681 |
+
"step": 804
|
| 5682 |
+
},
|
| 5683 |
+
{
|
| 5684 |
+
"epoch": 4.2368421052631575,
|
| 5685 |
+
"grad_norm": 0.07292202860116959,
|
| 5686 |
+
"learning_rate": 7.631578947368421e-05,
|
| 5687 |
+
"loss": 0.0011,
|
| 5688 |
+
"step": 805
|
| 5689 |
+
},
|
| 5690 |
+
{
|
| 5691 |
+
"epoch": 4.242105263157895,
|
| 5692 |
+
"grad_norm": 0.03265571966767311,
|
| 5693 |
+
"learning_rate": 7.578947368421054e-05,
|
| 5694 |
+
"loss": 0.001,
|
| 5695 |
+
"step": 806
|
| 5696 |
+
},
|
| 5697 |
+
{
|
| 5698 |
+
"epoch": 4.247368421052632,
|
| 5699 |
+
"grad_norm": 0.008702615275979042,
|
| 5700 |
+
"learning_rate": 7.526315789473684e-05,
|
| 5701 |
+
"loss": 0.0006,
|
| 5702 |
+
"step": 807
|
| 5703 |
+
},
|
| 5704 |
+
{
|
| 5705 |
+
"epoch": 4.252631578947368,
|
| 5706 |
+
"grad_norm": 0.01023191586136818,
|
| 5707 |
+
"learning_rate": 7.473684210526316e-05,
|
| 5708 |
+
"loss": 0.0005,
|
| 5709 |
+
"step": 808
|
| 5710 |
+
},
|
| 5711 |
+
{
|
| 5712 |
+
"epoch": 4.257894736842105,
|
| 5713 |
+
"grad_norm": 0.009560521692037582,
|
| 5714 |
+
"learning_rate": 7.421052631578948e-05,
|
| 5715 |
+
"loss": 0.0005,
|
| 5716 |
+
"step": 809
|
| 5717 |
+
},
|
| 5718 |
+
{
|
| 5719 |
+
"epoch": 4.2631578947368425,
|
| 5720 |
+
"grad_norm": 0.008541513234376907,
|
| 5721 |
+
"learning_rate": 7.368421052631579e-05,
|
| 5722 |
+
"loss": 0.0006,
|
| 5723 |
+
"step": 810
|
| 5724 |
+
},
|
| 5725 |
+
{
|
| 5726 |
+
"epoch": 4.268421052631579,
|
| 5727 |
+
"grad_norm": 0.01207665540277958,
|
| 5728 |
+
"learning_rate": 7.31578947368421e-05,
|
| 5729 |
+
"loss": 0.0008,
|
| 5730 |
+
"step": 811
|
| 5731 |
+
},
|
| 5732 |
+
{
|
| 5733 |
+
"epoch": 4.273684210526316,
|
| 5734 |
+
"grad_norm": 0.08192358165979385,
|
| 5735 |
+
"learning_rate": 7.263157894736843e-05,
|
| 5736 |
+
"loss": 0.0016,
|
| 5737 |
+
"step": 812
|
| 5738 |
+
},
|
| 5739 |
+
{
|
| 5740 |
+
"epoch": 4.278947368421052,
|
| 5741 |
+
"grad_norm": 0.02339164726436138,
|
| 5742 |
+
"learning_rate": 7.210526315789474e-05,
|
| 5743 |
+
"loss": 0.0008,
|
| 5744 |
+
"step": 813
|
| 5745 |
+
},
|
| 5746 |
+
{
|
| 5747 |
+
"epoch": 4.284210526315789,
|
| 5748 |
+
"grad_norm": 0.20728661119937897,
|
| 5749 |
+
"learning_rate": 7.157894736842105e-05,
|
| 5750 |
+
"loss": 0.012,
|
| 5751 |
+
"step": 814
|
| 5752 |
+
},
|
| 5753 |
+
{
|
| 5754 |
+
"epoch": 4.2894736842105265,
|
| 5755 |
+
"grad_norm": 0.3511962294578552,
|
| 5756 |
+
"learning_rate": 7.105263157894737e-05,
|
| 5757 |
+
"loss": 0.0077,
|
| 5758 |
+
"step": 815
|
| 5759 |
+
},
|
| 5760 |
+
{
|
| 5761 |
+
"epoch": 4.294736842105263,
|
| 5762 |
+
"grad_norm": 0.010202188976109028,
|
| 5763 |
+
"learning_rate": 7.05263157894737e-05,
|
| 5764 |
+
"loss": 0.0005,
|
| 5765 |
+
"step": 816
|
| 5766 |
+
},
|
| 5767 |
+
{
|
| 5768 |
+
"epoch": 4.3,
|
| 5769 |
+
"grad_norm": 0.015126821584999561,
|
| 5770 |
+
"learning_rate": 7.000000000000001e-05,
|
| 5771 |
+
"loss": 0.0008,
|
| 5772 |
+
"step": 817
|
| 5773 |
+
},
|
| 5774 |
+
{
|
| 5775 |
+
"epoch": 4.3052631578947365,
|
| 5776 |
+
"grad_norm": 0.006620699539780617,
|
| 5777 |
+
"learning_rate": 6.947368421052632e-05,
|
| 5778 |
+
"loss": 0.0006,
|
| 5779 |
+
"step": 818
|
| 5780 |
+
},
|
| 5781 |
+
{
|
| 5782 |
+
"epoch": 4.310526315789474,
|
| 5783 |
+
"grad_norm": 0.26067501306533813,
|
| 5784 |
+
"learning_rate": 6.894736842105263e-05,
|
| 5785 |
+
"loss": 0.0325,
|
| 5786 |
+
"step": 819
|
| 5787 |
+
},
|
| 5788 |
+
{
|
| 5789 |
+
"epoch": 4.315789473684211,
|
| 5790 |
+
"grad_norm": 0.004113540519028902,
|
| 5791 |
+
"learning_rate": 6.842105263157896e-05,
|
| 5792 |
+
"loss": 0.0003,
|
| 5793 |
+
"step": 820
|
| 5794 |
+
},
|
| 5795 |
+
{
|
| 5796 |
+
"epoch": 4.321052631578947,
|
| 5797 |
+
"grad_norm": 0.008812862448394299,
|
| 5798 |
+
"learning_rate": 6.789473684210526e-05,
|
| 5799 |
+
"loss": 0.0006,
|
| 5800 |
+
"step": 821
|
| 5801 |
+
},
|
| 5802 |
+
{
|
| 5803 |
+
"epoch": 4.326315789473684,
|
| 5804 |
+
"grad_norm": 0.028753219172358513,
|
| 5805 |
+
"learning_rate": 6.736842105263157e-05,
|
| 5806 |
+
"loss": 0.0009,
|
| 5807 |
+
"step": 822
|
| 5808 |
+
},
|
| 5809 |
+
{
|
| 5810 |
+
"epoch": 4.331578947368421,
|
| 5811 |
+
"grad_norm": 0.009436219930648804,
|
| 5812 |
+
"learning_rate": 6.68421052631579e-05,
|
| 5813 |
+
"loss": 0.0006,
|
| 5814 |
+
"step": 823
|
| 5815 |
+
},
|
| 5816 |
+
{
|
| 5817 |
+
"epoch": 4.336842105263158,
|
| 5818 |
+
"grad_norm": 0.010652395896613598,
|
| 5819 |
+
"learning_rate": 6.631578947368421e-05,
|
| 5820 |
+
"loss": 0.0005,
|
| 5821 |
+
"step": 824
|
| 5822 |
+
},
|
| 5823 |
+
{
|
| 5824 |
+
"epoch": 4.342105263157895,
|
| 5825 |
+
"grad_norm": 0.012191063724458218,
|
| 5826 |
+
"learning_rate": 6.578947368421052e-05,
|
| 5827 |
+
"loss": 0.0006,
|
| 5828 |
+
"step": 825
|
| 5829 |
+
},
|
| 5830 |
+
{
|
| 5831 |
+
"epoch": 4.347368421052631,
|
| 5832 |
+
"grad_norm": 0.014493409544229507,
|
| 5833 |
+
"learning_rate": 6.526315789473684e-05,
|
| 5834 |
+
"loss": 0.0007,
|
| 5835 |
+
"step": 826
|
| 5836 |
+
},
|
| 5837 |
+
{
|
| 5838 |
+
"epoch": 4.352631578947369,
|
| 5839 |
+
"grad_norm": 0.09270607680082321,
|
| 5840 |
+
"learning_rate": 6.473684210526316e-05,
|
| 5841 |
+
"loss": 0.0018,
|
| 5842 |
+
"step": 827
|
| 5843 |
+
},
|
| 5844 |
+
{
|
| 5845 |
+
"epoch": 4.3578947368421055,
|
| 5846 |
+
"grad_norm": 0.04721800237894058,
|
| 5847 |
+
"learning_rate": 6.421052631578946e-05,
|
| 5848 |
+
"loss": 0.0014,
|
| 5849 |
+
"step": 828
|
| 5850 |
+
},
|
| 5851 |
+
{
|
| 5852 |
+
"epoch": 4.363157894736842,
|
| 5853 |
+
"grad_norm": 0.028600886464118958,
|
| 5854 |
+
"learning_rate": 6.368421052631579e-05,
|
| 5855 |
+
"loss": 0.001,
|
| 5856 |
+
"step": 829
|
| 5857 |
+
},
|
| 5858 |
+
{
|
| 5859 |
+
"epoch": 4.368421052631579,
|
| 5860 |
+
"grad_norm": 0.010532131418585777,
|
| 5861 |
+
"learning_rate": 6.31578947368421e-05,
|
| 5862 |
+
"loss": 0.0007,
|
| 5863 |
+
"step": 830
|
| 5864 |
+
},
|
| 5865 |
+
{
|
| 5866 |
+
"epoch": 4.373684210526315,
|
| 5867 |
+
"grad_norm": 0.06029532477259636,
|
| 5868 |
+
"learning_rate": 6.263157894736843e-05,
|
| 5869 |
+
"loss": 0.0014,
|
| 5870 |
+
"step": 831
|
| 5871 |
+
},
|
| 5872 |
+
{
|
| 5873 |
+
"epoch": 4.378947368421053,
|
| 5874 |
+
"grad_norm": 0.012682969681918621,
|
| 5875 |
+
"learning_rate": 6.210526315789474e-05,
|
| 5876 |
+
"loss": 0.0006,
|
| 5877 |
+
"step": 832
|
| 5878 |
+
},
|
| 5879 |
+
{
|
| 5880 |
+
"epoch": 4.38421052631579,
|
| 5881 |
+
"grad_norm": 0.01644139178097248,
|
| 5882 |
+
"learning_rate": 6.157894736842106e-05,
|
| 5883 |
+
"loss": 0.0007,
|
| 5884 |
+
"step": 833
|
| 5885 |
+
},
|
| 5886 |
+
{
|
| 5887 |
+
"epoch": 4.389473684210526,
|
| 5888 |
+
"grad_norm": 0.3973478376865387,
|
| 5889 |
+
"learning_rate": 6.105263157894737e-05,
|
| 5890 |
+
"loss": 0.0164,
|
| 5891 |
+
"step": 834
|
| 5892 |
+
},
|
| 5893 |
+
{
|
| 5894 |
+
"epoch": 4.394736842105263,
|
| 5895 |
+
"grad_norm": 0.009873030707240105,
|
| 5896 |
+
"learning_rate": 6.052631578947369e-05,
|
| 5897 |
+
"loss": 0.0006,
|
| 5898 |
+
"step": 835
|
| 5899 |
+
},
|
| 5900 |
+
{
|
| 5901 |
+
"epoch": 4.4,
|
| 5902 |
+
"grad_norm": 0.007480965461581945,
|
| 5903 |
+
"learning_rate": 6e-05,
|
| 5904 |
+
"loss": 0.0005,
|
| 5905 |
+
"step": 836
|
| 5906 |
+
},
|
| 5907 |
+
{
|
| 5908 |
+
"epoch": 4.405263157894737,
|
| 5909 |
+
"grad_norm": 0.062482405453920364,
|
| 5910 |
+
"learning_rate": 5.947368421052632e-05,
|
| 5911 |
+
"loss": 0.0015,
|
| 5912 |
+
"step": 837
|
| 5913 |
+
},
|
| 5914 |
+
{
|
| 5915 |
+
"epoch": 4.410526315789474,
|
| 5916 |
+
"grad_norm": 0.12598900496959686,
|
| 5917 |
+
"learning_rate": 5.8947368421052634e-05,
|
| 5918 |
+
"loss": 0.0076,
|
| 5919 |
+
"step": 838
|
| 5920 |
+
},
|
| 5921 |
+
{
|
| 5922 |
+
"epoch": 4.41578947368421,
|
| 5923 |
+
"grad_norm": 0.006310358177870512,
|
| 5924 |
+
"learning_rate": 5.8421052631578954e-05,
|
| 5925 |
+
"loss": 0.0006,
|
| 5926 |
+
"step": 839
|
| 5927 |
+
},
|
| 5928 |
+
{
|
| 5929 |
+
"epoch": 4.421052631578947,
|
| 5930 |
+
"grad_norm": 0.22700683772563934,
|
| 5931 |
+
"learning_rate": 5.789473684210527e-05,
|
| 5932 |
+
"loss": 0.0018,
|
| 5933 |
+
"step": 840
|
| 5934 |
+
},
|
| 5935 |
+
{
|
| 5936 |
+
"epoch": 4.426315789473684,
|
| 5937 |
+
"grad_norm": 0.027526654303073883,
|
| 5938 |
+
"learning_rate": 5.736842105263158e-05,
|
| 5939 |
+
"loss": 0.0012,
|
| 5940 |
+
"step": 841
|
| 5941 |
+
},
|
| 5942 |
+
{
|
| 5943 |
+
"epoch": 4.431578947368421,
|
| 5944 |
+
"grad_norm": 0.009056963957846165,
|
| 5945 |
+
"learning_rate": 5.68421052631579e-05,
|
| 5946 |
+
"loss": 0.0006,
|
| 5947 |
+
"step": 842
|
| 5948 |
+
},
|
| 5949 |
+
{
|
| 5950 |
+
"epoch": 4.436842105263158,
|
| 5951 |
+
"grad_norm": 0.06567683815956116,
|
| 5952 |
+
"learning_rate": 5.6315789473684206e-05,
|
| 5953 |
+
"loss": 0.002,
|
| 5954 |
+
"step": 843
|
| 5955 |
+
},
|
| 5956 |
+
{
|
| 5957 |
+
"epoch": 4.442105263157894,
|
| 5958 |
+
"grad_norm": 0.02286943607032299,
|
| 5959 |
+
"learning_rate": 5.5789473684210526e-05,
|
| 5960 |
+
"loss": 0.0009,
|
| 5961 |
+
"step": 844
|
| 5962 |
+
},
|
| 5963 |
+
{
|
| 5964 |
+
"epoch": 4.447368421052632,
|
| 5965 |
+
"grad_norm": 0.022757651284337044,
|
| 5966 |
+
"learning_rate": 5.526315789473684e-05,
|
| 5967 |
+
"loss": 0.0007,
|
| 5968 |
+
"step": 845
|
| 5969 |
+
},
|
| 5970 |
+
{
|
| 5971 |
+
"epoch": 4.4526315789473685,
|
| 5972 |
+
"grad_norm": 0.0074144755490124226,
|
| 5973 |
+
"learning_rate": 5.473684210526316e-05,
|
| 5974 |
+
"loss": 0.0005,
|
| 5975 |
+
"step": 846
|
| 5976 |
+
},
|
| 5977 |
+
{
|
| 5978 |
+
"epoch": 4.457894736842105,
|
| 5979 |
+
"grad_norm": 0.013096978887915611,
|
| 5980 |
+
"learning_rate": 5.421052631578947e-05,
|
| 5981 |
+
"loss": 0.0007,
|
| 5982 |
+
"step": 847
|
| 5983 |
+
},
|
| 5984 |
+
{
|
| 5985 |
+
"epoch": 4.463157894736842,
|
| 5986 |
+
"grad_norm": 0.017680592834949493,
|
| 5987 |
+
"learning_rate": 5.368421052631579e-05,
|
| 5988 |
+
"loss": 0.0009,
|
| 5989 |
+
"step": 848
|
| 5990 |
+
},
|
| 5991 |
+
{
|
| 5992 |
+
"epoch": 4.468421052631579,
|
| 5993 |
+
"grad_norm": 0.010339250788092613,
|
| 5994 |
+
"learning_rate": 5.3157894736842104e-05,
|
| 5995 |
+
"loss": 0.0005,
|
| 5996 |
+
"step": 849
|
| 5997 |
+
},
|
| 5998 |
+
{
|
| 5999 |
+
"epoch": 4.473684210526316,
|
| 6000 |
+
"grad_norm": 0.01019520778208971,
|
| 6001 |
+
"learning_rate": 5.263157894736842e-05,
|
| 6002 |
+
"loss": 0.0008,
|
| 6003 |
+
"step": 850
|
| 6004 |
+
},
|
| 6005 |
+
{
|
| 6006 |
+
"epoch": 4.478947368421053,
|
| 6007 |
+
"grad_norm": 0.018972480669617653,
|
| 6008 |
+
"learning_rate": 5.210526315789474e-05,
|
| 6009 |
+
"loss": 0.0009,
|
| 6010 |
+
"step": 851
|
| 6011 |
+
},
|
| 6012 |
+
{
|
| 6013 |
+
"epoch": 4.484210526315789,
|
| 6014 |
+
"grad_norm": 0.028711460530757904,
|
| 6015 |
+
"learning_rate": 5.157894736842105e-05,
|
| 6016 |
+
"loss": 0.0016,
|
| 6017 |
+
"step": 852
|
| 6018 |
+
},
|
| 6019 |
+
{
|
| 6020 |
+
"epoch": 4.489473684210527,
|
| 6021 |
+
"grad_norm": 0.010612289421260357,
|
| 6022 |
+
"learning_rate": 5.105263157894737e-05,
|
| 6023 |
+
"loss": 0.0005,
|
| 6024 |
+
"step": 853
|
| 6025 |
+
},
|
| 6026 |
+
{
|
| 6027 |
+
"epoch": 4.494736842105263,
|
| 6028 |
+
"grad_norm": 0.6425463557243347,
|
| 6029 |
+
"learning_rate": 5.052631578947368e-05,
|
| 6030 |
+
"loss": 0.0031,
|
| 6031 |
+
"step": 854
|
| 6032 |
+
},
|
| 6033 |
+
{
|
| 6034 |
+
"epoch": 4.5,
|
| 6035 |
+
"grad_norm": 0.008190907537937164,
|
| 6036 |
+
"learning_rate": 5e-05,
|
| 6037 |
+
"loss": 0.0005,
|
| 6038 |
+
"step": 855
|
| 6039 |
+
},
|
| 6040 |
+
{
|
| 6041 |
+
"epoch": 4.505263157894737,
|
| 6042 |
+
"grad_norm": 0.018427042290568352,
|
| 6043 |
+
"learning_rate": 4.9473684210526315e-05,
|
| 6044 |
+
"loss": 0.0009,
|
| 6045 |
+
"step": 856
|
| 6046 |
+
},
|
| 6047 |
+
{
|
| 6048 |
+
"epoch": 4.510526315789473,
|
| 6049 |
+
"grad_norm": 0.00848427601158619,
|
| 6050 |
+
"learning_rate": 4.894736842105263e-05,
|
| 6051 |
+
"loss": 0.0005,
|
| 6052 |
+
"step": 857
|
| 6053 |
+
},
|
| 6054 |
+
{
|
| 6055 |
+
"epoch": 4.515789473684211,
|
| 6056 |
+
"grad_norm": 0.019971899688243866,
|
| 6057 |
+
"learning_rate": 4.842105263157895e-05,
|
| 6058 |
+
"loss": 0.0007,
|
| 6059 |
+
"step": 858
|
| 6060 |
+
},
|
| 6061 |
+
{
|
| 6062 |
+
"epoch": 4.521052631578947,
|
| 6063 |
+
"grad_norm": 0.07693297415971756,
|
| 6064 |
+
"learning_rate": 4.789473684210526e-05,
|
| 6065 |
+
"loss": 0.0011,
|
| 6066 |
+
"step": 859
|
| 6067 |
+
},
|
| 6068 |
+
{
|
| 6069 |
+
"epoch": 4.526315789473684,
|
| 6070 |
+
"grad_norm": 0.006678177043795586,
|
| 6071 |
+
"learning_rate": 4.736842105263158e-05,
|
| 6072 |
+
"loss": 0.0005,
|
| 6073 |
+
"step": 860
|
| 6074 |
+
},
|
| 6075 |
+
{
|
| 6076 |
+
"epoch": 4.531578947368421,
|
| 6077 |
+
"grad_norm": 0.006219150498509407,
|
| 6078 |
+
"learning_rate": 4.6842105263157894e-05,
|
| 6079 |
+
"loss": 0.0004,
|
| 6080 |
+
"step": 861
|
| 6081 |
+
},
|
| 6082 |
+
{
|
| 6083 |
+
"epoch": 4.536842105263158,
|
| 6084 |
+
"grad_norm": 0.048225317150354385,
|
| 6085 |
+
"learning_rate": 4.6315789473684214e-05,
|
| 6086 |
+
"loss": 0.0013,
|
| 6087 |
+
"step": 862
|
| 6088 |
+
},
|
| 6089 |
+
{
|
| 6090 |
+
"epoch": 4.542105263157895,
|
| 6091 |
+
"grad_norm": 0.009762817993760109,
|
| 6092 |
+
"learning_rate": 4.5789473684210527e-05,
|
| 6093 |
+
"loss": 0.0006,
|
| 6094 |
+
"step": 863
|
| 6095 |
+
},
|
| 6096 |
+
{
|
| 6097 |
+
"epoch": 4.5473684210526315,
|
| 6098 |
+
"grad_norm": 0.01545332744717598,
|
| 6099 |
+
"learning_rate": 4.5263157894736846e-05,
|
| 6100 |
+
"loss": 0.0006,
|
| 6101 |
+
"step": 864
|
| 6102 |
+
},
|
| 6103 |
+
{
|
| 6104 |
+
"epoch": 4.552631578947368,
|
| 6105 |
+
"grad_norm": 0.03185436129570007,
|
| 6106 |
+
"learning_rate": 4.473684210526316e-05,
|
| 6107 |
+
"loss": 0.0015,
|
| 6108 |
+
"step": 865
|
| 6109 |
+
},
|
| 6110 |
+
{
|
| 6111 |
+
"epoch": 4.557894736842105,
|
| 6112 |
+
"grad_norm": 0.021765299141407013,
|
| 6113 |
+
"learning_rate": 4.421052631578947e-05,
|
| 6114 |
+
"loss": 0.0009,
|
| 6115 |
+
"step": 866
|
| 6116 |
+
},
|
| 6117 |
+
{
|
| 6118 |
+
"epoch": 4.563157894736842,
|
| 6119 |
+
"grad_norm": 0.02456175908446312,
|
| 6120 |
+
"learning_rate": 4.368421052631579e-05,
|
| 6121 |
+
"loss": 0.001,
|
| 6122 |
+
"step": 867
|
| 6123 |
+
},
|
| 6124 |
+
{
|
| 6125 |
+
"epoch": 4.568421052631579,
|
| 6126 |
+
"grad_norm": 0.013163010589778423,
|
| 6127 |
+
"learning_rate": 4.3157894736842105e-05,
|
| 6128 |
+
"loss": 0.0007,
|
| 6129 |
+
"step": 868
|
| 6130 |
+
},
|
| 6131 |
+
{
|
| 6132 |
+
"epoch": 4.573684210526316,
|
| 6133 |
+
"grad_norm": 0.007158617954701185,
|
| 6134 |
+
"learning_rate": 4.2631578947368425e-05,
|
| 6135 |
+
"loss": 0.0004,
|
| 6136 |
+
"step": 869
|
| 6137 |
+
},
|
| 6138 |
+
{
|
| 6139 |
+
"epoch": 4.578947368421053,
|
| 6140 |
+
"grad_norm": 0.012921416200697422,
|
| 6141 |
+
"learning_rate": 4.210526315789474e-05,
|
| 6142 |
+
"loss": 0.0008,
|
| 6143 |
+
"step": 870
|
| 6144 |
+
},
|
| 6145 |
+
{
|
| 6146 |
+
"epoch": 4.58421052631579,
|
| 6147 |
+
"grad_norm": 0.02498648315668106,
|
| 6148 |
+
"learning_rate": 4.157894736842106e-05,
|
| 6149 |
+
"loss": 0.0012,
|
| 6150 |
+
"step": 871
|
| 6151 |
+
},
|
| 6152 |
+
{
|
| 6153 |
+
"epoch": 4.589473684210526,
|
| 6154 |
+
"grad_norm": 0.009193144738674164,
|
| 6155 |
+
"learning_rate": 4.105263157894737e-05,
|
| 6156 |
+
"loss": 0.0006,
|
| 6157 |
+
"step": 872
|
| 6158 |
+
},
|
| 6159 |
+
{
|
| 6160 |
+
"epoch": 4.594736842105263,
|
| 6161 |
+
"grad_norm": 0.07746303826570511,
|
| 6162 |
+
"learning_rate": 4.0526315789473684e-05,
|
| 6163 |
+
"loss": 0.0027,
|
| 6164 |
+
"step": 873
|
| 6165 |
+
},
|
| 6166 |
+
{
|
| 6167 |
+
"epoch": 4.6,
|
| 6168 |
+
"grad_norm": 0.12564599514007568,
|
| 6169 |
+
"learning_rate": 4e-05,
|
| 6170 |
+
"loss": 0.0042,
|
| 6171 |
+
"step": 874
|
| 6172 |
+
},
|
| 6173 |
+
{
|
| 6174 |
+
"epoch": 4.605263157894737,
|
| 6175 |
+
"grad_norm": 0.013173549436032772,
|
| 6176 |
+
"learning_rate": 3.9473684210526316e-05,
|
| 6177 |
+
"loss": 0.0006,
|
| 6178 |
+
"step": 875
|
| 6179 |
+
},
|
| 6180 |
+
{
|
| 6181 |
+
"epoch": 4.610526315789474,
|
| 6182 |
+
"grad_norm": 0.014632917940616608,
|
| 6183 |
+
"learning_rate": 3.8947368421052636e-05,
|
| 6184 |
+
"loss": 0.0009,
|
| 6185 |
+
"step": 876
|
| 6186 |
+
},
|
| 6187 |
+
{
|
| 6188 |
+
"epoch": 4.61578947368421,
|
| 6189 |
+
"grad_norm": 0.004674790892750025,
|
| 6190 |
+
"learning_rate": 3.842105263157895e-05,
|
| 6191 |
+
"loss": 0.0003,
|
| 6192 |
+
"step": 877
|
| 6193 |
+
},
|
| 6194 |
+
{
|
| 6195 |
+
"epoch": 4.621052631578947,
|
| 6196 |
+
"grad_norm": 0.01062940526753664,
|
| 6197 |
+
"learning_rate": 3.789473684210527e-05,
|
| 6198 |
+
"loss": 0.0006,
|
| 6199 |
+
"step": 878
|
| 6200 |
+
},
|
| 6201 |
+
{
|
| 6202 |
+
"epoch": 4.626315789473685,
|
| 6203 |
+
"grad_norm": 0.03135712072253227,
|
| 6204 |
+
"learning_rate": 3.736842105263158e-05,
|
| 6205 |
+
"loss": 0.0009,
|
| 6206 |
+
"step": 879
|
| 6207 |
+
},
|
| 6208 |
+
{
|
| 6209 |
+
"epoch": 4.631578947368421,
|
| 6210 |
+
"grad_norm": 0.01560395397245884,
|
| 6211 |
+
"learning_rate": 3.6842105263157895e-05,
|
| 6212 |
+
"loss": 0.0007,
|
| 6213 |
+
"step": 880
|
| 6214 |
+
},
|
| 6215 |
+
{
|
| 6216 |
+
"epoch": 4.636842105263158,
|
| 6217 |
+
"grad_norm": 0.06484178453683853,
|
| 6218 |
+
"learning_rate": 3.6315789473684214e-05,
|
| 6219 |
+
"loss": 0.0022,
|
| 6220 |
+
"step": 881
|
| 6221 |
+
},
|
| 6222 |
+
{
|
| 6223 |
+
"epoch": 4.6421052631578945,
|
| 6224 |
+
"grad_norm": 0.007599903270602226,
|
| 6225 |
+
"learning_rate": 3.578947368421053e-05,
|
| 6226 |
+
"loss": 0.0004,
|
| 6227 |
+
"step": 882
|
| 6228 |
+
},
|
| 6229 |
+
{
|
| 6230 |
+
"epoch": 4.647368421052631,
|
| 6231 |
+
"grad_norm": 0.07640688121318817,
|
| 6232 |
+
"learning_rate": 3.526315789473685e-05,
|
| 6233 |
+
"loss": 0.002,
|
| 6234 |
+
"step": 883
|
| 6235 |
+
},
|
| 6236 |
+
{
|
| 6237 |
+
"epoch": 4.652631578947369,
|
| 6238 |
+
"grad_norm": 0.018931038677692413,
|
| 6239 |
+
"learning_rate": 3.473684210526316e-05,
|
| 6240 |
+
"loss": 0.0006,
|
| 6241 |
+
"step": 884
|
| 6242 |
+
},
|
| 6243 |
+
{
|
| 6244 |
+
"epoch": 4.657894736842105,
|
| 6245 |
+
"grad_norm": 0.03390808403491974,
|
| 6246 |
+
"learning_rate": 3.421052631578948e-05,
|
| 6247 |
+
"loss": 0.0012,
|
| 6248 |
+
"step": 885
|
| 6249 |
+
},
|
| 6250 |
+
{
|
| 6251 |
+
"epoch": 4.663157894736842,
|
| 6252 |
+
"grad_norm": 0.011840026825666428,
|
| 6253 |
+
"learning_rate": 3.3684210526315786e-05,
|
| 6254 |
+
"loss": 0.0008,
|
| 6255 |
+
"step": 886
|
| 6256 |
+
},
|
| 6257 |
+
{
|
| 6258 |
+
"epoch": 4.668421052631579,
|
| 6259 |
+
"grad_norm": 0.011166613548994064,
|
| 6260 |
+
"learning_rate": 3.3157894736842106e-05,
|
| 6261 |
+
"loss": 0.0006,
|
| 6262 |
+
"step": 887
|
| 6263 |
+
},
|
| 6264 |
+
{
|
| 6265 |
+
"epoch": 4.673684210526316,
|
| 6266 |
+
"grad_norm": 0.09616900235414505,
|
| 6267 |
+
"learning_rate": 3.263157894736842e-05,
|
| 6268 |
+
"loss": 0.0013,
|
| 6269 |
+
"step": 888
|
| 6270 |
+
},
|
| 6271 |
+
{
|
| 6272 |
+
"epoch": 4.678947368421053,
|
| 6273 |
+
"grad_norm": 0.04197359457612038,
|
| 6274 |
+
"learning_rate": 3.210526315789473e-05,
|
| 6275 |
+
"loss": 0.0014,
|
| 6276 |
+
"step": 889
|
| 6277 |
+
},
|
| 6278 |
+
{
|
| 6279 |
+
"epoch": 4.684210526315789,
|
| 6280 |
+
"grad_norm": 0.025599906221032143,
|
| 6281 |
+
"learning_rate": 3.157894736842105e-05,
|
| 6282 |
+
"loss": 0.0008,
|
| 6283 |
+
"step": 890
|
| 6284 |
+
},
|
| 6285 |
+
{
|
| 6286 |
+
"epoch": 4.689473684210526,
|
| 6287 |
+
"grad_norm": 0.010700304992496967,
|
| 6288 |
+
"learning_rate": 3.105263157894737e-05,
|
| 6289 |
+
"loss": 0.0007,
|
| 6290 |
+
"step": 891
|
| 6291 |
+
},
|
| 6292 |
+
{
|
| 6293 |
+
"epoch": 4.6947368421052635,
|
| 6294 |
+
"grad_norm": 0.004940670914947987,
|
| 6295 |
+
"learning_rate": 3.0526315789473684e-05,
|
| 6296 |
+
"loss": 0.0003,
|
| 6297 |
+
"step": 892
|
| 6298 |
+
},
|
| 6299 |
+
{
|
| 6300 |
+
"epoch": 4.7,
|
| 6301 |
+
"grad_norm": 0.023425649851560593,
|
| 6302 |
+
"learning_rate": 3e-05,
|
| 6303 |
+
"loss": 0.0006,
|
| 6304 |
+
"step": 893
|
| 6305 |
+
},
|
| 6306 |
+
{
|
| 6307 |
+
"epoch": 4.705263157894737,
|
| 6308 |
+
"grad_norm": 0.009202427230775356,
|
| 6309 |
+
"learning_rate": 2.9473684210526317e-05,
|
| 6310 |
+
"loss": 0.0006,
|
| 6311 |
+
"step": 894
|
| 6312 |
+
},
|
| 6313 |
+
{
|
| 6314 |
+
"epoch": 4.7105263157894735,
|
| 6315 |
+
"grad_norm": 0.012214015237987041,
|
| 6316 |
+
"learning_rate": 2.8947368421052634e-05,
|
| 6317 |
+
"loss": 0.0007,
|
| 6318 |
+
"step": 895
|
| 6319 |
+
},
|
| 6320 |
+
{
|
| 6321 |
+
"epoch": 4.715789473684211,
|
| 6322 |
+
"grad_norm": 0.02451181598007679,
|
| 6323 |
+
"learning_rate": 2.842105263157895e-05,
|
| 6324 |
+
"loss": 0.0011,
|
| 6325 |
+
"step": 896
|
| 6326 |
+
},
|
| 6327 |
+
{
|
| 6328 |
+
"epoch": 4.721052631578948,
|
| 6329 |
+
"grad_norm": 0.09701679646968842,
|
| 6330 |
+
"learning_rate": 2.7894736842105263e-05,
|
| 6331 |
+
"loss": 0.0029,
|
| 6332 |
+
"step": 897
|
| 6333 |
+
},
|
| 6334 |
+
{
|
| 6335 |
+
"epoch": 4.726315789473684,
|
| 6336 |
+
"grad_norm": 0.00955820269882679,
|
| 6337 |
+
"learning_rate": 2.736842105263158e-05,
|
| 6338 |
+
"loss": 0.0006,
|
| 6339 |
+
"step": 898
|
| 6340 |
+
},
|
| 6341 |
+
{
|
| 6342 |
+
"epoch": 4.731578947368421,
|
| 6343 |
+
"grad_norm": 0.037156615406274796,
|
| 6344 |
+
"learning_rate": 2.6842105263157896e-05,
|
| 6345 |
+
"loss": 0.0012,
|
| 6346 |
+
"step": 899
|
| 6347 |
+
},
|
| 6348 |
+
{
|
| 6349 |
+
"epoch": 4.7368421052631575,
|
| 6350 |
+
"grad_norm": 0.008509020321071148,
|
| 6351 |
+
"learning_rate": 2.631578947368421e-05,
|
| 6352 |
+
"loss": 0.0006,
|
| 6353 |
+
"step": 900
|
| 6354 |
+
},
|
| 6355 |
+
{
|
| 6356 |
+
"epoch": 4.742105263157895,
|
| 6357 |
+
"grad_norm": 0.004725357051938772,
|
| 6358 |
+
"learning_rate": 2.5789473684210525e-05,
|
| 6359 |
+
"loss": 0.0004,
|
| 6360 |
+
"step": 901
|
| 6361 |
+
},
|
| 6362 |
+
{
|
| 6363 |
+
"epoch": 4.747368421052632,
|
| 6364 |
+
"grad_norm": 0.04897700622677803,
|
| 6365 |
+
"learning_rate": 2.526315789473684e-05,
|
| 6366 |
+
"loss": 0.0012,
|
| 6367 |
+
"step": 902
|
| 6368 |
+
},
|
| 6369 |
+
{
|
| 6370 |
+
"epoch": 4.752631578947368,
|
| 6371 |
+
"grad_norm": 0.006657022517174482,
|
| 6372 |
+
"learning_rate": 2.4736842105263158e-05,
|
| 6373 |
+
"loss": 0.0004,
|
| 6374 |
+
"step": 903
|
| 6375 |
+
},
|
| 6376 |
+
{
|
| 6377 |
+
"epoch": 4.757894736842105,
|
| 6378 |
+
"grad_norm": 0.01627969741821289,
|
| 6379 |
+
"learning_rate": 2.4210526315789474e-05,
|
| 6380 |
+
"loss": 0.0006,
|
| 6381 |
+
"step": 904
|
| 6382 |
+
},
|
| 6383 |
+
{
|
| 6384 |
+
"epoch": 4.7631578947368425,
|
| 6385 |
+
"grad_norm": 0.0063805836252868176,
|
| 6386 |
+
"learning_rate": 2.368421052631579e-05,
|
| 6387 |
+
"loss": 0.0006,
|
| 6388 |
+
"step": 905
|
| 6389 |
+
},
|
| 6390 |
+
{
|
| 6391 |
+
"epoch": 4.768421052631579,
|
| 6392 |
+
"grad_norm": 0.01253101509064436,
|
| 6393 |
+
"learning_rate": 2.3157894736842107e-05,
|
| 6394 |
+
"loss": 0.0005,
|
| 6395 |
+
"step": 906
|
| 6396 |
+
},
|
| 6397 |
+
{
|
| 6398 |
+
"epoch": 4.773684210526316,
|
| 6399 |
+
"grad_norm": 0.01995311863720417,
|
| 6400 |
+
"learning_rate": 2.2631578947368423e-05,
|
| 6401 |
+
"loss": 0.0007,
|
| 6402 |
+
"step": 907
|
| 6403 |
+
},
|
| 6404 |
+
{
|
| 6405 |
+
"epoch": 4.778947368421052,
|
| 6406 |
+
"grad_norm": 0.019288456067442894,
|
| 6407 |
+
"learning_rate": 2.2105263157894736e-05,
|
| 6408 |
+
"loss": 0.0008,
|
| 6409 |
+
"step": 908
|
| 6410 |
+
},
|
| 6411 |
+
{
|
| 6412 |
+
"epoch": 4.784210526315789,
|
| 6413 |
+
"grad_norm": 0.014500236138701439,
|
| 6414 |
+
"learning_rate": 2.1578947368421053e-05,
|
| 6415 |
+
"loss": 0.0007,
|
| 6416 |
+
"step": 909
|
| 6417 |
+
},
|
| 6418 |
+
{
|
| 6419 |
+
"epoch": 4.7894736842105265,
|
| 6420 |
+
"grad_norm": 0.011677373200654984,
|
| 6421 |
+
"learning_rate": 2.105263157894737e-05,
|
| 6422 |
+
"loss": 0.0005,
|
| 6423 |
+
"step": 910
|
| 6424 |
+
},
|
| 6425 |
+
{
|
| 6426 |
+
"epoch": 4.794736842105263,
|
| 6427 |
+
"grad_norm": 0.03549593687057495,
|
| 6428 |
+
"learning_rate": 2.0526315789473685e-05,
|
| 6429 |
+
"loss": 0.0009,
|
| 6430 |
+
"step": 911
|
| 6431 |
+
},
|
| 6432 |
+
{
|
| 6433 |
+
"epoch": 4.8,
|
| 6434 |
+
"grad_norm": 0.007476325146853924,
|
| 6435 |
+
"learning_rate": 2e-05,
|
| 6436 |
+
"loss": 0.0004,
|
| 6437 |
+
"step": 912
|
| 6438 |
+
},
|
| 6439 |
+
{
|
| 6440 |
+
"epoch": 4.8052631578947365,
|
| 6441 |
+
"grad_norm": 0.02231527678668499,
|
| 6442 |
+
"learning_rate": 1.9473684210526318e-05,
|
| 6443 |
+
"loss": 0.0008,
|
| 6444 |
+
"step": 913
|
| 6445 |
+
},
|
| 6446 |
+
{
|
| 6447 |
+
"epoch": 4.810526315789474,
|
| 6448 |
+
"grad_norm": 0.31320345401763916,
|
| 6449 |
+
"learning_rate": 1.8947368421052634e-05,
|
| 6450 |
+
"loss": 0.0164,
|
| 6451 |
+
"step": 914
|
| 6452 |
+
},
|
| 6453 |
+
{
|
| 6454 |
+
"epoch": 4.815789473684211,
|
| 6455 |
+
"grad_norm": 0.031381040811538696,
|
| 6456 |
+
"learning_rate": 1.8421052631578947e-05,
|
| 6457 |
+
"loss": 0.0009,
|
| 6458 |
+
"step": 915
|
| 6459 |
+
},
|
| 6460 |
+
{
|
| 6461 |
+
"epoch": 4.821052631578947,
|
| 6462 |
+
"grad_norm": 0.13466821610927582,
|
| 6463 |
+
"learning_rate": 1.7894736842105264e-05,
|
| 6464 |
+
"loss": 0.0024,
|
| 6465 |
+
"step": 916
|
| 6466 |
+
},
|
| 6467 |
+
{
|
| 6468 |
+
"epoch": 4.826315789473684,
|
| 6469 |
+
"grad_norm": 0.02074410393834114,
|
| 6470 |
+
"learning_rate": 1.736842105263158e-05,
|
| 6471 |
+
"loss": 0.0008,
|
| 6472 |
+
"step": 917
|
| 6473 |
+
},
|
| 6474 |
+
{
|
| 6475 |
+
"epoch": 4.831578947368421,
|
| 6476 |
+
"grad_norm": 0.013965011574327946,
|
| 6477 |
+
"learning_rate": 1.6842105263157893e-05,
|
| 6478 |
+
"loss": 0.0007,
|
| 6479 |
+
"step": 918
|
| 6480 |
+
},
|
| 6481 |
+
{
|
| 6482 |
+
"epoch": 4.836842105263158,
|
| 6483 |
+
"grad_norm": 0.012210886925458908,
|
| 6484 |
+
"learning_rate": 1.631578947368421e-05,
|
| 6485 |
+
"loss": 0.0005,
|
| 6486 |
+
"step": 919
|
| 6487 |
+
},
|
| 6488 |
+
{
|
| 6489 |
+
"epoch": 4.842105263157895,
|
| 6490 |
+
"grad_norm": 0.009945962578058243,
|
| 6491 |
+
"learning_rate": 1.5789473684210526e-05,
|
| 6492 |
+
"loss": 0.0005,
|
| 6493 |
+
"step": 920
|
| 6494 |
+
},
|
| 6495 |
+
{
|
| 6496 |
+
"epoch": 4.847368421052631,
|
| 6497 |
+
"grad_norm": 0.008283141069114208,
|
| 6498 |
+
"learning_rate": 1.5263157894736842e-05,
|
| 6499 |
+
"loss": 0.0005,
|
| 6500 |
+
"step": 921
|
| 6501 |
+
},
|
| 6502 |
+
{
|
| 6503 |
+
"epoch": 4.852631578947369,
|
| 6504 |
+
"grad_norm": 0.025940481573343277,
|
| 6505 |
+
"learning_rate": 1.4736842105263159e-05,
|
| 6506 |
+
"loss": 0.0009,
|
| 6507 |
+
"step": 922
|
| 6508 |
+
},
|
| 6509 |
+
{
|
| 6510 |
+
"epoch": 4.8578947368421055,
|
| 6511 |
+
"grad_norm": 0.35328882932662964,
|
| 6512 |
+
"learning_rate": 1.4210526315789475e-05,
|
| 6513 |
+
"loss": 0.0238,
|
| 6514 |
+
"step": 923
|
| 6515 |
+
},
|
| 6516 |
+
{
|
| 6517 |
+
"epoch": 4.863157894736842,
|
| 6518 |
+
"grad_norm": 0.007774712517857552,
|
| 6519 |
+
"learning_rate": 1.368421052631579e-05,
|
| 6520 |
+
"loss": 0.0006,
|
| 6521 |
+
"step": 924
|
| 6522 |
+
},
|
| 6523 |
+
{
|
| 6524 |
+
"epoch": 4.868421052631579,
|
| 6525 |
+
"grad_norm": 0.016255423426628113,
|
| 6526 |
+
"learning_rate": 1.3157894736842104e-05,
|
| 6527 |
+
"loss": 0.0008,
|
| 6528 |
+
"step": 925
|
| 6529 |
+
},
|
| 6530 |
+
{
|
| 6531 |
+
"epoch": 4.873684210526315,
|
| 6532 |
+
"grad_norm": 0.005064685363322496,
|
| 6533 |
+
"learning_rate": 1.263157894736842e-05,
|
| 6534 |
+
"loss": 0.0003,
|
| 6535 |
+
"step": 926
|
| 6536 |
+
},
|
| 6537 |
+
{
|
| 6538 |
+
"epoch": 4.878947368421053,
|
| 6539 |
+
"grad_norm": 0.016532612964510918,
|
| 6540 |
+
"learning_rate": 1.2105263157894737e-05,
|
| 6541 |
+
"loss": 0.0008,
|
| 6542 |
+
"step": 927
|
| 6543 |
+
},
|
| 6544 |
+
{
|
| 6545 |
+
"epoch": 4.88421052631579,
|
| 6546 |
+
"grad_norm": 0.007061973214149475,
|
| 6547 |
+
"learning_rate": 1.1578947368421053e-05,
|
| 6548 |
+
"loss": 0.0005,
|
| 6549 |
+
"step": 928
|
| 6550 |
+
},
|
| 6551 |
+
{
|
| 6552 |
+
"epoch": 4.889473684210526,
|
| 6553 |
+
"grad_norm": 0.02170943096280098,
|
| 6554 |
+
"learning_rate": 1.1052631578947368e-05,
|
| 6555 |
+
"loss": 0.0007,
|
| 6556 |
+
"step": 929
|
| 6557 |
+
},
|
| 6558 |
+
{
|
| 6559 |
+
"epoch": 4.894736842105263,
|
| 6560 |
+
"grad_norm": 0.07484550029039383,
|
| 6561 |
+
"learning_rate": 1.0526315789473684e-05,
|
| 6562 |
+
"loss": 0.0008,
|
| 6563 |
+
"step": 930
|
| 6564 |
+
},
|
| 6565 |
+
{
|
| 6566 |
+
"epoch": 4.9,
|
| 6567 |
+
"grad_norm": 0.005272566806524992,
|
| 6568 |
+
"learning_rate": 1e-05,
|
| 6569 |
+
"loss": 0.0003,
|
| 6570 |
+
"step": 931
|
| 6571 |
+
},
|
| 6572 |
+
{
|
| 6573 |
+
"epoch": 4.905263157894737,
|
| 6574 |
+
"grad_norm": 0.08144894987344742,
|
| 6575 |
+
"learning_rate": 9.473684210526317e-06,
|
| 6576 |
+
"loss": 0.0018,
|
| 6577 |
+
"step": 932
|
| 6578 |
+
},
|
| 6579 |
+
{
|
| 6580 |
+
"epoch": 4.910526315789474,
|
| 6581 |
+
"grad_norm": 0.016598021611571312,
|
| 6582 |
+
"learning_rate": 8.947368421052632e-06,
|
| 6583 |
+
"loss": 0.0006,
|
| 6584 |
+
"step": 933
|
| 6585 |
+
},
|
| 6586 |
+
{
|
| 6587 |
+
"epoch": 4.91578947368421,
|
| 6588 |
+
"grad_norm": 0.0372316800057888,
|
| 6589 |
+
"learning_rate": 8.421052631578947e-06,
|
| 6590 |
+
"loss": 0.0008,
|
| 6591 |
+
"step": 934
|
| 6592 |
+
},
|
| 6593 |
+
{
|
| 6594 |
+
"epoch": 4.921052631578947,
|
| 6595 |
+
"grad_norm": 0.050694022327661514,
|
| 6596 |
+
"learning_rate": 7.894736842105263e-06,
|
| 6597 |
+
"loss": 0.0013,
|
| 6598 |
+
"step": 935
|
| 6599 |
+
},
|
| 6600 |
+
{
|
| 6601 |
+
"epoch": 4.926315789473684,
|
| 6602 |
+
"grad_norm": 0.009993389248847961,
|
| 6603 |
+
"learning_rate": 7.368421052631579e-06,
|
| 6604 |
+
"loss": 0.0006,
|
| 6605 |
+
"step": 936
|
| 6606 |
+
},
|
| 6607 |
+
{
|
| 6608 |
+
"epoch": 4.931578947368421,
|
| 6609 |
+
"grad_norm": 0.009024999104440212,
|
| 6610 |
+
"learning_rate": 6.842105263157895e-06,
|
| 6611 |
+
"loss": 0.0005,
|
| 6612 |
+
"step": 937
|
| 6613 |
+
},
|
| 6614 |
+
{
|
| 6615 |
+
"epoch": 4.936842105263158,
|
| 6616 |
+
"grad_norm": 0.008197016082704067,
|
| 6617 |
+
"learning_rate": 6.31578947368421e-06,
|
| 6618 |
+
"loss": 0.0005,
|
| 6619 |
+
"step": 938
|
| 6620 |
+
},
|
| 6621 |
+
{
|
| 6622 |
+
"epoch": 4.942105263157895,
|
| 6623 |
+
"grad_norm": 0.011617634445428848,
|
| 6624 |
+
"learning_rate": 5.789473684210527e-06,
|
| 6625 |
+
"loss": 0.0006,
|
| 6626 |
+
"step": 939
|
| 6627 |
+
},
|
| 6628 |
+
{
|
| 6629 |
+
"epoch": 4.947368421052632,
|
| 6630 |
+
"grad_norm": 0.013150883838534355,
|
| 6631 |
+
"learning_rate": 5.263157894736842e-06,
|
| 6632 |
+
"loss": 0.0006,
|
| 6633 |
+
"step": 940
|
| 6634 |
+
},
|
| 6635 |
+
{
|
| 6636 |
+
"epoch": 4.9526315789473685,
|
| 6637 |
+
"grad_norm": 0.01023543905466795,
|
| 6638 |
+
"learning_rate": 4.736842105263159e-06,
|
| 6639 |
+
"loss": 0.0006,
|
| 6640 |
+
"step": 941
|
| 6641 |
+
},
|
| 6642 |
+
{
|
| 6643 |
+
"epoch": 4.957894736842105,
|
| 6644 |
+
"grad_norm": 0.11015679687261581,
|
| 6645 |
+
"learning_rate": 4.210526315789473e-06,
|
| 6646 |
+
"loss": 0.0035,
|
| 6647 |
+
"step": 942
|
| 6648 |
+
},
|
| 6649 |
+
{
|
| 6650 |
+
"epoch": 4.963157894736842,
|
| 6651 |
+
"grad_norm": 0.04862911254167557,
|
| 6652 |
+
"learning_rate": 3.6842105263157896e-06,
|
| 6653 |
+
"loss": 0.0013,
|
| 6654 |
+
"step": 943
|
| 6655 |
+
},
|
| 6656 |
+
{
|
| 6657 |
+
"epoch": 4.968421052631579,
|
| 6658 |
+
"grad_norm": 0.006454069633036852,
|
| 6659 |
+
"learning_rate": 3.157894736842105e-06,
|
| 6660 |
+
"loss": 0.0004,
|
| 6661 |
+
"step": 944
|
| 6662 |
+
},
|
| 6663 |
+
{
|
| 6664 |
+
"epoch": 4.973684210526316,
|
| 6665 |
+
"grad_norm": 0.13466902077198029,
|
| 6666 |
+
"learning_rate": 2.631578947368421e-06,
|
| 6667 |
+
"loss": 0.0018,
|
| 6668 |
+
"step": 945
|
| 6669 |
+
},
|
| 6670 |
+
{
|
| 6671 |
+
"epoch": 4.978947368421053,
|
| 6672 |
+
"grad_norm": 0.01431818027049303,
|
| 6673 |
+
"learning_rate": 2.1052631578947366e-06,
|
| 6674 |
+
"loss": 0.0009,
|
| 6675 |
+
"step": 946
|
| 6676 |
+
},
|
| 6677 |
+
{
|
| 6678 |
+
"epoch": 4.984210526315789,
|
| 6679 |
+
"grad_norm": 0.010833046399056911,
|
| 6680 |
+
"learning_rate": 1.5789473684210526e-06,
|
| 6681 |
+
"loss": 0.0006,
|
| 6682 |
+
"step": 947
|
| 6683 |
+
},
|
| 6684 |
+
{
|
| 6685 |
+
"epoch": 4.989473684210527,
|
| 6686 |
+
"grad_norm": 0.010601919144392014,
|
| 6687 |
+
"learning_rate": 1.0526315789473683e-06,
|
| 6688 |
+
"loss": 0.0006,
|
| 6689 |
+
"step": 948
|
| 6690 |
+
},
|
| 6691 |
+
{
|
| 6692 |
+
"epoch": 4.994736842105263,
|
| 6693 |
+
"grad_norm": 0.010958509519696236,
|
| 6694 |
+
"learning_rate": 5.263157894736842e-07,
|
| 6695 |
+
"loss": 0.0005,
|
| 6696 |
+
"step": 949
|
| 6697 |
+
},
|
| 6698 |
+
{
|
| 6699 |
+
"epoch": 5.0,
|
| 6700 |
+
"grad_norm": 0.005408676341176033,
|
| 6701 |
+
"learning_rate": 0.0,
|
| 6702 |
+
"loss": 0.0002,
|
| 6703 |
+
"step": 950
|
| 6704 |
+
},
|
| 6705 |
+
{
|
| 6706 |
+
"epoch": 5.0,
|
| 6707 |
+
"eval_cer": 18.474692202462382,
|
| 6708 |
+
"eval_loss": 1.3566256761550903,
|
| 6709 |
+
"eval_pred": "| i | Label | Prediction |\n| --- | --- | --- |\n| 0 | I think that in the people in the river. They because I\u2026 | I ... that in ... people in the river they they ... I\u2026 |\n| 1 | This is a restaurant because there has | I ... ... restaurant because there\u2026\u2026 |\n| 2 | food. | It\u2026 |\n| 3 | The people in the picture are playing soccer. I\u2019ve played soccer twice before in physical education class and I liked it. Well, mostly because I have really strong muscles in my legs from running, so I have a lot of advantages in soccer. If I was a parent, I would I would agree for my kid to play soccer. Mostly because playing a sport helps you stay healthy and fit and that\u2019s what ??? society thinks you should do. Stay fit and healthy. | The people in the picture are playing soccer. I play\ufffdve played soccer twice before in physical education class and I liked it well Well, mostly because I have really strong muscles in my legs from running so so I have a lot of advantages in soccer. If I was a parent, I ... ... would agree for my kid to play soccer mostly Mostly because playing a sport helps you stay healthy and fit and that\u2019s what society society thinks you should do, Stay fit and healthy |\n| 4 | And it\u2019s also good for your health too. You can have lower a lower risk of getting any diseases from body fat. Also, the people in this picture mostly are wearing jerseys and shorts. Some of them are wearing knee-high socks. And all of them are wearing sneakers. And the details in here. There are a lot of trees, which I really like. And really beautiful grass. And there are also two buildings in the background. | And it\u2019s also good for your help too, You could have lower ... lower risk of getting any diseases from body fat. Also, the people in this picture mostly are wearing jerseys and shorts. Some of them are wearing ne highhigh socks and And all of them are wearing sneakers. And the details in here, There are a lot of trees which which I really like. And the beautiful grass. And there are also two buildings in the background |\n| 5 | Also a bridge. There is a silver car silver car and there are multiple soccer balls which um means that they are probably probably practicing and not playing against each other. The people um Also there are benches in the background which also indicates that they ??? might might be in a park instead of a soccer field. And there are | umso, bridge, There is a silver car ... car and there are multiple soccer balls which um means that they are probably probably practicing and not playing against each other. The people um also there are benches in the background which also indicates that they might might be be in the park instead of a soccer field and And there are |\n| 6 | some\u2026 | And\u2026 |\n| 7 | um I think people in the picture is playing soccer, if I\u2019m not wrong. Yes, eN they are playing soccer. And um did I? I did. When I was in elementary school, we we had we had a class we had a PE class and um the teacher taught us how to play soccer before. But, honestly, I\u2019m very poor at at um at sports, so I\u2019m not really enjoy it. But I did see some people, like my some of my classmate, really do know how to play soccer. | um I think people in the picture is playing soccer if if I\u2019m not wrong, Yes, they they they are playing soccer and And um did I ... I did uh And I was in my school, we ... ... a had a class, ... a PE class and the the teacher taught us how to play soccer before but But honestly honestly, I am\ufffdm very poor at ... ... at sport so so I\u2019m not really enjoy it. But I did say some people like like my ... of my classmatesmate really really do know how to play soccer |\n| 8 | I I was like, oh my god, this is a very good, very, very um phenomenal cause um it's like it\u2019s a it\u2019s very hard to see some for some the students in Taiwan to play soccer, so as I think it\u2019s quite quite cool. And if I am parents, well, um um because I\u2019m not really interested in this, so, um if they want to, of course, I would encourage them, but um if they doesn\u2019t like that, I I won\u2019t force them to do it cause I think it\u2019s not really um, it\u2019s alright. It doesn\u2019t like it\u2019s not | And ... was like, oh my god, this is a very good ... very ... very um phenomenal because um it\ufffd like,'s\ufffds a ...'s\ufffds very hard to see some for some the students in Taiwan to play soccer so so it I think it's\ufffds quite ... cool and And if I am parents well well um um because because I am\ufffdm not really interested in this so so um um if they want to, of course, I will encourage them, but um if they doesn\u2019t like that, I ... won't\ufffdt force them to do it, I think it's\ufffds not really um it it\u2019s alright, It doesn\u2019t like it is\ufffds not |\n| 9 | necessary and just like if they want, I will. And um if I have time, um so people and it\u2019s um only boys. Why? ish In the picture, they should have girls, but um in the picture, um the only boys in the pictures and uN they separate divided into two groups, is it? And all they\u2019re wearing this long socks is quite cool and um it\u2019s quite a beautiful place. It\u2019s a really beautiful place and um I think they enjoyed very | necessary and just like if they want to I will and And um if I have time um um so people in ...'s\ufffds that only boys why Why? In that in the picture they they should have girls, but um in the picture um um the only boys in the pictures, uh they they separate it into two groups is is it? And oh they're\ufffdre wearing this long socks is quite cool and um it\u2019s quite a beautiful place, It\u2019s a really beautiful place and um I think they enjoyed very |\n| 10 | much. It's quite\u2026 | um much It It\ufffd a |\n| 11 | Um I think the picture is taken um at a park and it\u2019s a very bright sunny day. And um there are some people are in the park and they are painting. And and there is a woman uh on the right of the picture. She\u2019s sitting on the chair and she is she has short hair and wearing some, wearing dress. And she | um I think the picture is taken um at a park and um\u2019s a very bright, day and And um there are some people are in the park and they are pant. And um um is a woman uh on the right of the picture, She is\ufffds sitting on the chair and she is ... has short hair and wearing some ... wearing dress and And she un un |\n| 12 | is painting some trees and I really like the picture. It\u2019s beautiful. And there are a bags beside the woman. I think there\u2019s uh the wore the pants or some color in the bags that she wants to to draw. And there are also um lots of people near nearby the the girl the woman | She painting some trees and I really like the picture it It\u2019s beautiful and And there are a bags uh a woman. I think there's\ufffds uh the ... the ... or some color in the bags that she wants to ... draw. And there are also um lots of people near ... the ... girl ... woman un un un un un un un un un un un un un un |\n| 13 | are painting painting and there are two people which has a bags and others are besides her and they are discussing um how to draw the picture. And I think uh the advantage to join a park is that you can really near the the picture you want to draw and it\u2019s in | un painting the and ... are two people which has a bags and others are besides her and I are discussing um how to draw the picture and And eh think uh the advantage to draw a park is that you can really near the ... picture you want to draw and it's\ufffds in un un un un un un un un un |\n| 14 | nature scenery\u2026 | um nature scenery un un un un un un un un un un un un un un un |\n| 15 | Well, I see at least nine people in the picture, and I can see that all they are all young men, and they are probably professional soccer players, or a soccer team at school, since they are all wearing sports wears that look quite professional. And, the man at the back of the picture is has a funky look, while he has spiky haircut. | Well, I see at least nine people in the picture and and I can see that all ... are all young man, and they are probably professional soccer players or or a soccer team at school, since they're all wearing sports wear that look quite professional, And the the man at the back of the picture is ... a funky look, well he has spunkyy haircut |\n| 16 | And he\u2019s wearing a white short sleeved t-shirt and like he\u2019s wearing blue shirts and long socks which soccer players usually wear, and he\u2019s also wearing blue sneakers. He\u2019s trying to chase the yellow soccer ball. um I see many people wearing long soccer socks, which really impressed me, and they\u2019re all wearing like um red | And he\u2019s wearing a white short-ved t-shirt and like he\u2019s wearing a shirts and long socks which soccer players usually wear and and he\u2019s also wearing blue sneakers he He\u2019s trying to chase the yellow soccer ball um Um I see many people wearing long soccer socks which which really impressed me and and they're\ufffdre all wearing like um red |\n| 17 | shorts or eh like green shirts with numbers on it. So they might be quite professional. And the weather looks good. At the back of the picture, I could see many trees too. And I could even see the MRT. eh Also, I see two buildings. And I\u2019m not really good at soccer, but I like watching soccer games. So I hope maybe | Shorts or I I green shirts with numbers on it so So they might be quite professional and And the weather looks good. At the back of the picture, I could see many trees too. And I could even see the MRT not Na not I I see two buildings. And I\u2019m not really good at soccer, but I like watching soccer games. So I hope maybe |\n| 18 | I can\u2026 | e can\u2026 |\n| 19 | I think this might be a room up a up in a up in the top building in the city because the windows out outside the windows there\u2019s a lot of colorful buil buildings and it\u2019s also really high up on the ground. The woman in the middle is playing her violin to the to the guests and and lots of people are taking pictures of her. I think this is a good place to have a celebration because it is really it is really | I think this might be a room up ... ... in the ... in the top building in the city because the windows ... ... the windows there\u2019s a lot of colorful buildings buildings buildings and is is\ufffds also really hype up the the ground the The woman in the middle is playing her violin to the ... the guests and ... lots of people are taking pictures of her I I think this is a good place to have a celebration because it is really is really |\n| 20 | um it looks very comfortable and the food and the food must taste really good. I think um the woman on the the woman on the left is wearing a red skirt red dress and is wearing black heels. She\u2019s looking happily at the woman playing the violin while filming her on the phone. And then there is a waiter on the on the top right corner serving food to the guest that that lives beside him. There is also a man | um It looks really comfortable and the food ... the food must taste really good I I think um the woman on the ... woman on the that is wearing a red skirt ... dress and is wearing black heel. She is\ufffds looking happily at the woman playing the violin while filming her on the phone and And then there is a waiter on ... ... the top right corner serving food to the guest that ... leads it him there There is also um man |\n| 21 | with a with a white T sh with a white shirt smiling happily. And behind the man, there is there is one guy on his iPad while another while another guy looks at him. There is four pers There's four people in total that\u2019s looking at the violin looking at the woman playing the violin, and they seem very satisfied with it. Um On the table, there are wines and different drinks for them. I don\u2019t see any | And ... ... a white ...shhh a white shirt smelling happily andAnd behind the men there there ... ... is one guy on his iPad, another ... another guy looks at him. There is four person there's four people in total that's\ufffds looking at the violin ... at the woman playing the violin and and the seem very satisfied with it. On on the table there there are wines and different drinks for them. I don\u2019t see any |\n| 22 | food yet, so maybe\u2026 | If, so so maybe |\n| 23 | In the picture, I can see peoples having dinner in a marvelous restaurant. And the woman woman standing in front of the picture are playing the violin. And the people around her are taking photos of of her. They are all smiling and happy with it. I think this | In the picture, I can see people having dinner in a marvelous restaurant and And the woman ... standing in front of the picture are ... the violin, And the people around her are taking photos of ... her. They are all smiling and happy with it. think |\n| 24 | place is good for people having marriage or having birthday party cause this place can can be full of people. Many people can go to the restaurant at once the same time. And it\u2019s fun | This is ... for people having marriage or having a party because this place can ... be full of people many Many people can go to the restaurant at once at same time and And it\u2019s fun |\n| 25 | to have so many people celebrate your birthday or marriage. mm And I don\u2019t think there are any children in the picture. I think it might be the all-adults party. And the waiters are | to have so many people celebrate your birthday or marriage and And I eh don\u2019t think there are any children in the picture I I think it might be the all adults adultsults party and And lo waiters are |\n| 26 | busy. | ohy |\n| 27 | Here is a river, and the river must in the country because you wouldn't you it's impossible to see this appearance in the city. And if you want to play uh in the river or beside the river, you have to be rea careful that you might get drowned. You have to find some river who has many lifeguard, and you can\u2019t go it uh by yourself alone uh because it's rea really dangerous that if you drown, | Here is a river and and the river must in the country because you ...\ufffd ... ... is impossible to see this appearance in the city and And if you want to play uh in the river or beside the river, you have to uh careful careful careful that you might get drawn, You have to find some river who has many lifeguard and and you can go\ufffdt go it by by yourself alone uh because it\ufffd re re real dangerous that if you drown |\n| 28 | no one can rescue you. And I think the place in the picture it's summer because everybody wear t-shirt and shirts and on only summer you will go to river to play because it\u2019s really uh comfortable and cool. And uh in the picture I think there are a big family. There are grandpa, grandma uh s sitting on on the rock in the river and there are many child | No one can rescue you andAnd eh think the play ... the picture is is summer because everybody wear t-shirt and shirts and only ... summer you will go to the to play because it's\ufffds really comfortable comfortable and cool and And uh in the picture I think there are a big family there There are grandpa grandma grandma uh sitting sitting on ... the rock in the river and there are a child |\n| 29 | many child play in the river. uh the river seems really cool and they have a great time to play with uh in it. uh because it is summer, so they everybody want to uh get into the cool river water and play. And there is a old man uh his hand holding a camera or vi camera to take picture. uh he might want | many child play in the river uh Uh The river thinks very cool and they have a great time to play with uh in it uh uh because it is summer so so they everybody want to uh get into the cool river water and play and And there is a old man uh he hand holding a camera or the a to take picture uh uh he might want |\n| 30 | to.... | To ... |\n| 31 | I think the people in the picture are playing soccer. And I ever did the such activities before. And I really like it because playing soccer can make me make a lot of friend. And I when I play soccer, I feel very happy. And if I were a parent, I would encourage my children to do this activities | I think the people in the picture are playing soccer and and I ever did the such activities before and And I really like it because playing soccer can make me make a lot of friends. And I ... I play soccer, I feel very happy and And if I were parents parent, I would encourage my children to do this activities |\n| 32 | because I hope they can make they can develop the habit of re doing regular exercise. And I also hope them can know what is the teamwork. And in the picture, I see a man with green clothes with 48 number. He is more | Because I hope they can ... ... can develop the habit of reg doing regular exercise and And eh also hope them can know what is the teamwork and And in the picture I I see a man with green clothes, forty s, He is more |\n| 33 | black than others people. And I think a yellow ball floats on the air and a man with white clothes and blue pants is ready for running to get the ball, I think. And in the according to the picture, I think the place is in the | more ... other people and And I think a yellow ball floats on the air and a man with white clothes and blue pants is ready for running to get the ball, I think. And in the ... to the picture, I think the place is in |\n| 34 | park. | uh |\n| 35 | I think this picture is in the park because there are a lot of tree and a lot of people in there. The girl who sit in in the right of the picture may might look might drawing eh some trees. eh I think through this activity you can enjoy some outside activity and you can smell some | I think this picture is in the park because there are a lot of tree and a lot of people in there the The girl who sit in ... the right of the picture may ... look might drawing a some trees uh Uh I think through this activity, can enjoy some outside activity and you can smell some |\n| 36 | fresh airs in the park. And I in this picture, I look some people drawing and looking their pictures. And s there are still another still other people running in front of them. And I think it is a good place to walk and do other activities. | umresh airs in the park and And I ... this picture I I look some people drawing and looking their pictures, And still there are still another ... other people running in front of them. I I think this is a good place to work and to other activities |\n| 37 | There is a girl who still she drawing her pictures that is very beautiful and a lot of trees on it. It is very fresh and very good and very beautiful. I think this is a good activity that you should go to there. If I | There is a girl who still ... drawing her pictures that is very beautiful and a lot of trees on it it It is very fresh and very good and very beautiful I I think this is a good activity that you should go to there if I |\n| 38 | have a opportunity, I would do that maybe be\u2026 | There a ..., I would do that maybe |\n| 39 | This place might be someone is getting married because they look some look someone look really normal and it\u2019s a restaurant. But and the the picture and the middle of the picture, the lady is playing a violin. And a lot of people are taking pictures of him is taking pictures of her. And I don\u2019t think this place is really great for to hold this | This place might be someone is getting married because they look someone look someone look really normal and eh\u2019s a restaurant but But ... ... ... picture in the middle of the picture, the ... is playing a violin and And a lot of people are taking pictures of him ... taking pictures of her and And I don\u2019t think this place is really great for ... hold this |\n| 40 | kind of activity because there are other people in the restaurants too. So I think if they can change to a another place that don\u2019t have too much people. This this will be better. And there are some there are some people in the in the labs taking photos and pictures of the lady playing violin. And the lady | eh of activity because eh are other people in the restaurants too so So I think if they can change to a ... place and don\u2019t have too much people, This ... will be better and And eh are some ... are some people in the ... the labs taking photos and pictures of the lady playing violin and And the lady |\n| 41 | playing violin is wearing high heels and there are some waiters waiters behind to serve the meals. And there are other customers too. And and the pink and a lady on the left is wearing a pink dress. And others are wearing normal except except of the waiter and the lady playing violin. | l a wearing high heels and there are some waiters ...ers behind to serve the meals andAnd there are other customers too andAnd ... a pink ... the lady on the left is wearing a pink dress and And others are wearing normal is of of the waiter and ... lady playing violin |\n| 42 | It may be a restaurant. Equally some food on the table. The lady stand in the center is playing music. And a man in left is ... | It's be a restaurant who Sheally some food on the table, The lady stand in the center is playing music and And a man in ... is\u2026 |\n| 43 | It\u2019s cameras. | It's\ufffds cameras\u2026 |\n| 44 | mm This is a restaurant and and she there and she is playing violin. And maybe there | um This is a restaurant and ... ... ... ... she is playing violin and And maybe there |\n| 45 | And maybe there can | And maybe there can\u2026 |\n| 46 | be\u2026 | um\u2026 |\n| 47 | I think thi this there is this is wedding conference because we see many people sit around sit around the table. And I think the mi, the wo | I think the I ... is ... is wedding conference because we see many people sit around ... around the table andAnd eh think the ... the the ... |\n| 48 | the woman stand in front of everybody is playing violin. And I think this is good to good to hold the campaign in here. | um ... sent in front of everybody is playing violin and And eh think this is good to ... to hold the pom in ... |\n| 49 | And and and some people are taking pictures the woman playing violin. | And ... ... ... people are taking pictures of woman playing violin |\n| 50 | And the waiters... | And ... waiters\u2026 |\n| 51 | It may be a restaurant because there are so many temple tem table and some ?? chairs. I think the lady in stand in the middle is playing the pia piano, the violin yes. I think it\u2019s suitable of holding a important and | It may be a restaurant because there are so many temple ... table and some natural they I I think the lady in ... ... the media is playing the ...iano ... ... the violin yes I I think is's\ufffds suitable of holding a important\u2026 |\n| 52 | celebrate event because it\u2019s it's their turn enough to show their show their respect. And they all wear a ties dress or dress up with a nature dre dress. And they have a romance romance light and ??? at | um Iate event because it ...\ufffds ...'s there true enough to show their ... their ... and And they old's a ties ... ... dress are with a lecture dress dress and And they are a ... ... rights and ... ... |\n| 53 | head. They are all ??? to to show out in this place. And they are all They have a ??? clothes and they all hold a everyone\u2019s hands are all | The'ss ... ... friends they ... show out in this place and And eh are all ... have a special they and they all hold a everyone's\ufffds hands are all |\n| 54 | a\u2026 | oh ... |\n| 55 | This is a rest room room because there have a table, round round uh round table and chairs. And this is have this uh the woman is playing uh music instrument and uh this this place uh uh to to this place is | This is the restaurant of uh uh uh have a table wrong wrong ... uh round table and the and And eh ... ... a uh the woman is uh uh music's and eh this ... place uh ... to ... this place is |\n| 56 | very good to a red wedding place eh because of uh the there have carpet, there have a high high glass and plate. And everyone uh have a suit and uh uh have a good good manner. This is a brighter and brighter and good a brighter and good place have a shine shine | It good to uh ... waiting place uh because of eh the ... have a and they have a high ... ... and the and And uh uh have a suit and um have have a ... ... manner and And is uh ... and the and good ... brighter and good place, a uh shine |\n| 57 | light and uh very good uh view. eh Everyone is very happy and talk uh enjoy the music and uh have a good uh have a good life. And everyone uh is talking each other heN. I think there is a very good place. And the woman | un and uh very good uh view and And everyone is very happy and uh uh enjoy the music and uh uh a good uh have a good life and And eh uh is talking each other eh I I I think uh is a very good place and And the woman |\n| 58 | play... | um |\n| 59 | I think here is Wulai because this there are there is a beautiful s river there and everyone are playing water happily and we should no take the raincoat down because | I think we is a allai because this ... ... ... is a filter river river there and eh are playing water happily and we should no take the eh prod done because |\n| 60 | if rain pour and the river will full of water and it might be a flood. Un The perfect season to come here was is summer. Summer is very hot hot so everyone will be play will be | If a broke and the river will\u2026 of water and ... might be a flood umum The perfect season to come here whereas it eh ... Summer is very hard pot so everyone will be play ... be |\n| 61 | happy to come to come here. And we come to Wulai because this last Sunday was my grandma\u2019s birthday and every everyone in my family planned a surprise to to celebrate at | eh to come ... ... here and And eh come to all likeai because last ... Sunday was my grandma\u2019s birthday and every ... in my family plan a surprise to ... celebrate |\n| 62 | my\u2026 | M \u2026 |\n| 63 | I think they are in the riverbank because it has water and stones and and have a tree. And you need to and they need to notice that if there if rains, it may it might be a lot of water and may and they might be be caught in they. And I | I think they are in the riverbank because it has to and stones and ... have a tree and And you need to ... they need to notice that if ... ... rains, it may ... might be a lot of water and ... ... I might be be caught in day and And I |\n| 64 | think this place and I think this is is summer because the water is cold and you can enjoy the water and and with your family to spend the afternoon and and the old guy is | I this place ... I think this is ... ... because uh water is cold and you can enjoy the water and ... with your family to spend the uh and ... the old guy is |\n| 65 | taking a picture and they seem to have fun in the riverbank and and there are 11 people in there. There're might be a family a family trip. | taking a picture and they seem to have fun in the riverbank and ... uh are ... people in there ... There are might be a family ... family creep |\n| 66 | eh\u2026 | um And |\n| 67 | This place might be a restaurant and and it\u2019s because it\u2019s having a wedding wedding ceremonies. The middle the middle woman is uh woman standing in the middle is playing violins. The restaurant is the best place to having this ceremony because everybody can have great great | This place might speed at restaurant and ... is is\ufffds because it's\ufffds having the wedding ... ceremony the The ... ... middle woman is ... woman standing in the middle is playing violins. The restaurant is the best place to having this ceremony because it can have great ... |\n| 68 | because everybody can have have to share their happiness to everyone. The middle woman standing in the middle is wearing black black dress and black shoes. The woman stand in the left side is wearing a red red dress. The man beside the red dress woman is wearing a | Because everybody can have ... to ... their happiness to everyone. The middle ... standing in the middle is wearing black ... dress and black shoes. The woman stand ... ... left side is wearing a red ... dress ... The man beside the red dress woman is wearing |\n| 69 | blue jeans blue jeans and brown color brown pants. And he's heh he is holding his cellphone to take a picture. | uh jeans jeans ... jeans and the color ... pants and and he is ... he is holding his cellphone to take a picture |\n| 70 | The people in the picture is playing the soccer. I used to playing the soccer in my junior high school exercise class. And I like to playing ball because I prefer to playing outside or outdoor rather than uh staying at home indoor. | The people in the picture is playing the soccer I I used to playing the soccer in my un high school exercise class and And I like to playing ball because I prefer to playing outside or outdoor rather than uh stay at home, |\n| 71 | If I a prac If I a parent, I would I would push my child child to engage this sports or act activity because the sport is healthy and have some pleasure and can | uh I ... ... ... I a present, I ... ... would push my child to to engage this sports or activity activity because the sport is healthy and have some pleasure and can |\n| 72 | can promote your health. The figure in the pictures is wearing a soccer and a soccer shirt and and foot. They wearing the blue | eh ... your health The The figure in the pictures is wearing the soccer and a soccer shirt and ... foot they They wearing the blog |\n| 73 | clothes and\u2026 | Clothes is\u2026 |\n| 74 | This may be the river eh near the river. And because there are a lot of waters in there, and rocks, they can like, bank in there. And this this this place, you need to notice that the the the mountains upstairs that up up there, you above there, you | um may be the uh ... near the river and And eh there are a lot of waters in there and and the they they can like bank bank in there. And this ... ... place, you need to notice that the ... ... mountains upstairs that up ... there you you above there |\n| 75 | may have the rains and the down the river may have lots of lots of flood uh flooded there. So you may notice that for for the summer. And for or you'll be dead or something. eh You may you can come there in the summer or | You have the rains and the down the river may have lots of ... of floods ... flood there's So you may notice that it ... the uh and and you ... you will be dead or something uh Uh you may ... can come there in the eh or |\n| 76 | spring. Because this quite cold you can there you can be cold there and you don\u2019t really need the\u2026 You can be be ca calm down there and it\u2019s very reliev relieves and just just calm down calm down and and Winter come there is too cold | uh because Because it quite too you can there ... can be cold there's ... ...\u2019t really need the ... You can be ... calm calm down there and it was\ufffds very reallyivesieves and just ... calm down ... down and ... uh come there it too cold |\n| 77 | and they all\u2026 | And the\u2026 |\n",
|
| 6710 |
+
"eval_runtime": 11.714,
|
| 6711 |
+
"eval_samples_per_second": 6.659,
|
| 6712 |
+
"eval_steps_per_second": 0.854,
|
| 6713 |
+
"eval_wer": 32.36001374098248,
|
| 6714 |
+
"step": 950
|
| 6715 |
}
|
| 6716 |
],
|
| 6717 |
"logging_steps": 1,
|
|
|
|
| 6726 |
"should_evaluate": false,
|
| 6727 |
"should_log": false,
|
| 6728 |
"should_save": true,
|
| 6729 |
+
"should_training_stop": true
|
| 6730 |
},
|
| 6731 |
"attributes": {}
|
| 6732 |
}
|
| 6733 |
},
|
| 6734 |
+
"total_flos": 2.6802228658176e+19,
|
| 6735 |
"train_batch_size": 16,
|
| 6736 |
"trial_name": null,
|
| 6737 |
"trial_params": null
|