End of training
Browse files- README.md +11 -7
- all_results.json +16 -0
- eval_results.json +10 -0
- train_results.json +9 -0
- trainer_state.json +2181 -0
README.md
CHANGED
|
@@ -1,8 +1,12 @@
|
|
| 1 |
---
|
| 2 |
library_name: transformers
|
|
|
|
|
|
|
| 3 |
license: apache-2.0
|
| 4 |
base_model: rinna/japanese-hubert-base
|
| 5 |
tags:
|
|
|
|
|
|
|
| 6 |
- generated_from_trainer
|
| 7 |
datasets:
|
| 8 |
- common_voice_13_0
|
|
@@ -15,15 +19,15 @@ model-index:
|
|
| 15 |
name: Automatic Speech Recognition
|
| 16 |
type: automatic-speech-recognition
|
| 17 |
dataset:
|
| 18 |
-
name:
|
| 19 |
type: common_voice_13_0
|
| 20 |
config: ja
|
| 21 |
split: test
|
| 22 |
-
args: ja
|
| 23 |
metrics:
|
| 24 |
- name: Wer
|
| 25 |
type: wer
|
| 26 |
-
value: 0.
|
| 27 |
---
|
| 28 |
|
| 29 |
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
|
|
@@ -31,11 +35,11 @@ should probably proofread and complete it, then remove this comment. -->
|
|
| 31 |
|
| 32 |
# Hubert-common_voice-ja-demo-roma-cosine-3e-4
|
| 33 |
|
| 34 |
-
This model is a fine-tuned version of [rinna/japanese-hubert-base](https://huggingface.co/rinna/japanese-hubert-base) on the
|
| 35 |
It achieves the following results on the evaluation set:
|
| 36 |
-
- Loss: 0.
|
| 37 |
-
- Wer: 0.
|
| 38 |
-
- Cer: 0.
|
| 39 |
|
| 40 |
## Model description
|
| 41 |
|
|
|
|
| 1 |
---
|
| 2 |
library_name: transformers
|
| 3 |
+
language:
|
| 4 |
+
- ja
|
| 5 |
license: apache-2.0
|
| 6 |
base_model: rinna/japanese-hubert-base
|
| 7 |
tags:
|
| 8 |
+
- automatic-speech-recognition
|
| 9 |
+
- mozilla-foundation/common_voice_13_0
|
| 10 |
- generated_from_trainer
|
| 11 |
datasets:
|
| 12 |
- common_voice_13_0
|
|
|
|
| 19 |
name: Automatic Speech Recognition
|
| 20 |
type: automatic-speech-recognition
|
| 21 |
dataset:
|
| 22 |
+
name: MOZILLA-FOUNDATION/COMMON_VOICE_13_0 - JA
|
| 23 |
type: common_voice_13_0
|
| 24 |
config: ja
|
| 25 |
split: test
|
| 26 |
+
args: 'Config: ja, Training split: train+validation, Eval split: test'
|
| 27 |
metrics:
|
| 28 |
- name: Wer
|
| 29 |
type: wer
|
| 30 |
+
value: 0.9987900786448881
|
| 31 |
---
|
| 32 |
|
| 33 |
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
|
|
|
|
| 35 |
|
| 36 |
# Hubert-common_voice-ja-demo-roma-cosine-3e-4
|
| 37 |
|
| 38 |
+
This model is a fine-tuned version of [rinna/japanese-hubert-base](https://huggingface.co/rinna/japanese-hubert-base) on the MOZILLA-FOUNDATION/COMMON_VOICE_13_0 - JA dataset.
|
| 39 |
It achieves the following results on the evaluation set:
|
| 40 |
+
- Loss: 0.3032
|
| 41 |
+
- Wer: 0.9988
|
| 42 |
+
- Cer: 0.1667
|
| 43 |
|
| 44 |
## Model description
|
| 45 |
|
all_results.json
ADDED
|
@@ -0,0 +1,16 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 50.0,
|
| 3 |
+
"eval_cer": 0.16668555816562242,
|
| 4 |
+
"eval_loss": 0.30323880910873413,
|
| 5 |
+
"eval_runtime": 252.3866,
|
| 6 |
+
"eval_samples": 4959,
|
| 7 |
+
"eval_samples_per_second": 19.648,
|
| 8 |
+
"eval_steps_per_second": 2.457,
|
| 9 |
+
"eval_wer": 0.9987900786448881,
|
| 10 |
+
"total_flos": 2.6273093616415412e+19,
|
| 11 |
+
"train_loss": 0.4882018888757584,
|
| 12 |
+
"train_runtime": 107059.6152,
|
| 13 |
+
"train_samples": 12031,
|
| 14 |
+
"train_samples_per_second": 5.619,
|
| 15 |
+
"train_steps_per_second": 0.176
|
| 16 |
+
}
|
eval_results.json
ADDED
|
@@ -0,0 +1,10 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 50.0,
|
| 3 |
+
"eval_cer": 0.16668555816562242,
|
| 4 |
+
"eval_loss": 0.30323880910873413,
|
| 5 |
+
"eval_runtime": 252.3866,
|
| 6 |
+
"eval_samples": 4959,
|
| 7 |
+
"eval_samples_per_second": 19.648,
|
| 8 |
+
"eval_steps_per_second": 2.457,
|
| 9 |
+
"eval_wer": 0.9987900786448881
|
| 10 |
+
}
|
train_results.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 50.0,
|
| 3 |
+
"total_flos": 2.6273093616415412e+19,
|
| 4 |
+
"train_loss": 0.4882018888757584,
|
| 5 |
+
"train_runtime": 107059.6152,
|
| 6 |
+
"train_samples": 12031,
|
| 7 |
+
"train_samples_per_second": 5.619,
|
| 8 |
+
"train_steps_per_second": 0.176
|
| 9 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,2181 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 0.3024933338165283,
|
| 3 |
+
"best_model_checkpoint": "./Hubert-common_voice-ja-demo-roma-cosine-3e-4/checkpoint-4400",
|
| 4 |
+
"epoch": 50.0,
|
| 5 |
+
"eval_steps": 100,
|
| 6 |
+
"global_step": 18800,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.26595744680851063,
|
| 13 |
+
"eval_cer": 1.6163371077245385,
|
| 14 |
+
"eval_loss": 15.823559761047363,
|
| 15 |
+
"eval_runtime": 210.3767,
|
| 16 |
+
"eval_samples_per_second": 23.572,
|
| 17 |
+
"eval_steps_per_second": 2.947,
|
| 18 |
+
"eval_wer": 1.792498487598306,
|
| 19 |
+
"step": 100
|
| 20 |
+
},
|
| 21 |
+
{
|
| 22 |
+
"epoch": 0.5319148936170213,
|
| 23 |
+
"eval_cer": 0.9276012616915579,
|
| 24 |
+
"eval_loss": 6.4206719398498535,
|
| 25 |
+
"eval_runtime": 221.1891,
|
| 26 |
+
"eval_samples_per_second": 22.42,
|
| 27 |
+
"eval_steps_per_second": 2.803,
|
| 28 |
+
"eval_wer": 1.0,
|
| 29 |
+
"step": 200
|
| 30 |
+
},
|
| 31 |
+
{
|
| 32 |
+
"epoch": 0.7978723404255319,
|
| 33 |
+
"eval_cer": 0.9276012616915579,
|
| 34 |
+
"eval_loss": 5.419256210327148,
|
| 35 |
+
"eval_runtime": 241.9799,
|
| 36 |
+
"eval_samples_per_second": 20.493,
|
| 37 |
+
"eval_steps_per_second": 2.562,
|
| 38 |
+
"eval_wer": 1.0,
|
| 39 |
+
"step": 300
|
| 40 |
+
},
|
| 41 |
+
{
|
| 42 |
+
"epoch": 1.0638297872340425,
|
| 43 |
+
"eval_cer": 0.9276012616915579,
|
| 44 |
+
"eval_loss": 4.907326698303223,
|
| 45 |
+
"eval_runtime": 261.1648,
|
| 46 |
+
"eval_samples_per_second": 18.988,
|
| 47 |
+
"eval_steps_per_second": 2.374,
|
| 48 |
+
"eval_wer": 1.0,
|
| 49 |
+
"step": 400
|
| 50 |
+
},
|
| 51 |
+
{
|
| 52 |
+
"epoch": 1.3297872340425532,
|
| 53 |
+
"grad_norm": 20.339691162109375,
|
| 54 |
+
"learning_rate": 1.1903999999999998e-05,
|
| 55 |
+
"loss": 7.456,
|
| 56 |
+
"step": 500
|
| 57 |
+
},
|
| 58 |
+
{
|
| 59 |
+
"epoch": 1.3297872340425532,
|
| 60 |
+
"eval_cer": 0.9276012616915579,
|
| 61 |
+
"eval_loss": 4.358646869659424,
|
| 62 |
+
"eval_runtime": 256.9257,
|
| 63 |
+
"eval_samples_per_second": 19.301,
|
| 64 |
+
"eval_steps_per_second": 2.413,
|
| 65 |
+
"eval_wer": 1.0,
|
| 66 |
+
"step": 500
|
| 67 |
+
},
|
| 68 |
+
{
|
| 69 |
+
"epoch": 1.5957446808510638,
|
| 70 |
+
"eval_cer": 0.9276012616915579,
|
| 71 |
+
"eval_loss": 3.8341214656829834,
|
| 72 |
+
"eval_runtime": 284.7473,
|
| 73 |
+
"eval_samples_per_second": 17.415,
|
| 74 |
+
"eval_steps_per_second": 2.177,
|
| 75 |
+
"eval_wer": 1.0,
|
| 76 |
+
"step": 600
|
| 77 |
+
},
|
| 78 |
+
{
|
| 79 |
+
"epoch": 1.8617021276595744,
|
| 80 |
+
"eval_cer": 0.9276012616915579,
|
| 81 |
+
"eval_loss": 3.4148387908935547,
|
| 82 |
+
"eval_runtime": 301.2939,
|
| 83 |
+
"eval_samples_per_second": 16.459,
|
| 84 |
+
"eval_steps_per_second": 2.058,
|
| 85 |
+
"eval_wer": 1.0,
|
| 86 |
+
"step": 700
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 2.127659574468085,
|
| 90 |
+
"eval_cer": 0.9276012616915579,
|
| 91 |
+
"eval_loss": 3.104691982269287,
|
| 92 |
+
"eval_runtime": 236.1726,
|
| 93 |
+
"eval_samples_per_second": 20.997,
|
| 94 |
+
"eval_steps_per_second": 2.625,
|
| 95 |
+
"eval_wer": 1.0,
|
| 96 |
+
"step": 800
|
| 97 |
+
},
|
| 98 |
+
{
|
| 99 |
+
"epoch": 2.393617021276596,
|
| 100 |
+
"eval_cer": 0.9276012616915579,
|
| 101 |
+
"eval_loss": 2.961205244064331,
|
| 102 |
+
"eval_runtime": 231.5216,
|
| 103 |
+
"eval_samples_per_second": 21.419,
|
| 104 |
+
"eval_steps_per_second": 2.678,
|
| 105 |
+
"eval_wer": 1.0,
|
| 106 |
+
"step": 900
|
| 107 |
+
},
|
| 108 |
+
{
|
| 109 |
+
"epoch": 2.6595744680851063,
|
| 110 |
+
"grad_norm": 1.9079935550689697,
|
| 111 |
+
"learning_rate": 2.3903999999999997e-05,
|
| 112 |
+
"loss": 3.2563,
|
| 113 |
+
"step": 1000
|
| 114 |
+
},
|
| 115 |
+
{
|
| 116 |
+
"epoch": 2.6595744680851063,
|
| 117 |
+
"eval_cer": 0.9276012616915579,
|
| 118 |
+
"eval_loss": 2.902658700942993,
|
| 119 |
+
"eval_runtime": 265.314,
|
| 120 |
+
"eval_samples_per_second": 18.691,
|
| 121 |
+
"eval_steps_per_second": 2.337,
|
| 122 |
+
"eval_wer": 1.0,
|
| 123 |
+
"step": 1000
|
| 124 |
+
},
|
| 125 |
+
{
|
| 126 |
+
"epoch": 2.925531914893617,
|
| 127 |
+
"eval_cer": 0.9276012616915579,
|
| 128 |
+
"eval_loss": 2.876744031906128,
|
| 129 |
+
"eval_runtime": 268.5469,
|
| 130 |
+
"eval_samples_per_second": 18.466,
|
| 131 |
+
"eval_steps_per_second": 2.309,
|
| 132 |
+
"eval_wer": 1.0,
|
| 133 |
+
"step": 1100
|
| 134 |
+
},
|
| 135 |
+
{
|
| 136 |
+
"epoch": 3.1914893617021276,
|
| 137 |
+
"eval_cer": 0.9276012616915579,
|
| 138 |
+
"eval_loss": 2.8588435649871826,
|
| 139 |
+
"eval_runtime": 250.7138,
|
| 140 |
+
"eval_samples_per_second": 19.78,
|
| 141 |
+
"eval_steps_per_second": 2.473,
|
| 142 |
+
"eval_wer": 1.0,
|
| 143 |
+
"step": 1200
|
| 144 |
+
},
|
| 145 |
+
{
|
| 146 |
+
"epoch": 3.4574468085106385,
|
| 147 |
+
"eval_cer": 0.9276012616915579,
|
| 148 |
+
"eval_loss": 2.803792953491211,
|
| 149 |
+
"eval_runtime": 253.4886,
|
| 150 |
+
"eval_samples_per_second": 19.563,
|
| 151 |
+
"eval_steps_per_second": 2.446,
|
| 152 |
+
"eval_wer": 1.0,
|
| 153 |
+
"step": 1300
|
| 154 |
+
},
|
| 155 |
+
{
|
| 156 |
+
"epoch": 3.723404255319149,
|
| 157 |
+
"eval_cer": 0.8359292858622536,
|
| 158 |
+
"eval_loss": 1.9932550191879272,
|
| 159 |
+
"eval_runtime": 266.0365,
|
| 160 |
+
"eval_samples_per_second": 18.64,
|
| 161 |
+
"eval_steps_per_second": 2.331,
|
| 162 |
+
"eval_wer": 1.0,
|
| 163 |
+
"step": 1400
|
| 164 |
+
},
|
| 165 |
+
{
|
| 166 |
+
"epoch": 3.9893617021276597,
|
| 167 |
+
"grad_norm": 4.3663716316223145,
|
| 168 |
+
"learning_rate": 3.5903999999999994e-05,
|
| 169 |
+
"loss": 2.5032,
|
| 170 |
+
"step": 1500
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 3.9893617021276597,
|
| 174 |
+
"eval_cer": 0.38112231138141156,
|
| 175 |
+
"eval_loss": 1.1380858421325684,
|
| 176 |
+
"eval_runtime": 256.0779,
|
| 177 |
+
"eval_samples_per_second": 19.365,
|
| 178 |
+
"eval_steps_per_second": 2.421,
|
| 179 |
+
"eval_wer": 0.9997983464408147,
|
| 180 |
+
"step": 1500
|
| 181 |
+
},
|
| 182 |
+
{
|
| 183 |
+
"epoch": 4.25531914893617,
|
| 184 |
+
"eval_cer": 0.2746446118187823,
|
| 185 |
+
"eval_loss": 0.8252492547035217,
|
| 186 |
+
"eval_runtime": 240.6394,
|
| 187 |
+
"eval_samples_per_second": 20.608,
|
| 188 |
+
"eval_steps_per_second": 2.576,
|
| 189 |
+
"eval_wer": 0.9983867715265174,
|
| 190 |
+
"step": 1600
|
| 191 |
+
},
|
| 192 |
+
{
|
| 193 |
+
"epoch": 4.5212765957446805,
|
| 194 |
+
"eval_cer": 0.2507318827267859,
|
| 195 |
+
"eval_loss": 0.6762787103652954,
|
| 196 |
+
"eval_runtime": 212.2043,
|
| 197 |
+
"eval_samples_per_second": 23.369,
|
| 198 |
+
"eval_steps_per_second": 2.922,
|
| 199 |
+
"eval_wer": 0.9973785037305909,
|
| 200 |
+
"step": 1700
|
| 201 |
+
},
|
| 202 |
+
{
|
| 203 |
+
"epoch": 4.787234042553192,
|
| 204 |
+
"eval_cer": 0.24174213494783992,
|
| 205 |
+
"eval_loss": 0.6010992527008057,
|
| 206 |
+
"eval_runtime": 212.3652,
|
| 207 |
+
"eval_samples_per_second": 23.351,
|
| 208 |
+
"eval_steps_per_second": 2.919,
|
| 209 |
+
"eval_wer": 0.9979834644081468,
|
| 210 |
+
"step": 1800
|
| 211 |
+
},
|
| 212 |
+
{
|
| 213 |
+
"epoch": 5.053191489361702,
|
| 214 |
+
"eval_cer": 0.23079418558748002,
|
| 215 |
+
"eval_loss": 0.5514463782310486,
|
| 216 |
+
"eval_runtime": 209.5514,
|
| 217 |
+
"eval_samples_per_second": 23.665,
|
| 218 |
+
"eval_steps_per_second": 2.959,
|
| 219 |
+
"eval_wer": 0.9975801572897761,
|
| 220 |
+
"step": 1900
|
| 221 |
+
},
|
| 222 |
+
{
|
| 223 |
+
"epoch": 5.319148936170213,
|
| 224 |
+
"grad_norm": 2.237888813018799,
|
| 225 |
+
"learning_rate": 4.7903999999999994e-05,
|
| 226 |
+
"loss": 0.678,
|
| 227 |
+
"step": 2000
|
| 228 |
+
},
|
| 229 |
+
{
|
| 230 |
+
"epoch": 5.319148936170213,
|
| 231 |
+
"eval_cer": 0.22485313488139397,
|
| 232 |
+
"eval_loss": 0.5128617882728577,
|
| 233 |
+
"eval_runtime": 210.7223,
|
| 234 |
+
"eval_samples_per_second": 23.533,
|
| 235 |
+
"eval_steps_per_second": 2.942,
|
| 236 |
+
"eval_wer": 0.9985884250857028,
|
| 237 |
+
"step": 2000
|
| 238 |
+
},
|
| 239 |
+
{
|
| 240 |
+
"epoch": 5.585106382978723,
|
| 241 |
+
"eval_cer": 0.23253350635340653,
|
| 242 |
+
"eval_loss": 0.4957906901836395,
|
| 243 |
+
"eval_runtime": 209.1903,
|
| 244 |
+
"eval_samples_per_second": 23.706,
|
| 245 |
+
"eval_steps_per_second": 2.964,
|
| 246 |
+
"eval_wer": 0.994353700342811,
|
| 247 |
+
"step": 2100
|
| 248 |
+
},
|
| 249 |
+
{
|
| 250 |
+
"epoch": 5.851063829787234,
|
| 251 |
+
"eval_cer": 0.21945146904204466,
|
| 252 |
+
"eval_loss": 0.47169229388237,
|
| 253 |
+
"eval_runtime": 208.7444,
|
| 254 |
+
"eval_samples_per_second": 23.756,
|
| 255 |
+
"eval_steps_per_second": 2.97,
|
| 256 |
+
"eval_wer": 0.9977818108489614,
|
| 257 |
+
"step": 2200
|
| 258 |
+
},
|
| 259 |
+
{
|
| 260 |
+
"epoch": 6.117021276595745,
|
| 261 |
+
"eval_cer": 0.217309563919061,
|
| 262 |
+
"eval_loss": 0.4651758670806885,
|
| 263 |
+
"eval_runtime": 239.8374,
|
| 264 |
+
"eval_samples_per_second": 20.677,
|
| 265 |
+
"eval_steps_per_second": 2.585,
|
| 266 |
+
"eval_wer": 0.9991933857632587,
|
| 267 |
+
"step": 2300
|
| 268 |
+
},
|
| 269 |
+
{
|
| 270 |
+
"epoch": 6.382978723404255,
|
| 271 |
+
"eval_cer": 0.21218149909907094,
|
| 272 |
+
"eval_loss": 0.45553550124168396,
|
| 273 |
+
"eval_runtime": 209.0918,
|
| 274 |
+
"eval_samples_per_second": 23.717,
|
| 275 |
+
"eval_steps_per_second": 2.965,
|
| 276 |
+
"eval_wer": 0.9973785037305909,
|
| 277 |
+
"step": 2400
|
| 278 |
+
},
|
| 279 |
+
{
|
| 280 |
+
"epoch": 6.648936170212766,
|
| 281 |
+
"grad_norm": 3.0910284519195557,
|
| 282 |
+
"learning_rate": 5.9903999999999994e-05,
|
| 283 |
+
"loss": 0.4238,
|
| 284 |
+
"step": 2500
|
| 285 |
+
},
|
| 286 |
+
{
|
| 287 |
+
"epoch": 6.648936170212766,
|
| 288 |
+
"eval_cer": 0.21164211423233417,
|
| 289 |
+
"eval_loss": 0.4360628128051758,
|
| 290 |
+
"eval_runtime": 211.1218,
|
| 291 |
+
"eval_samples_per_second": 23.489,
|
| 292 |
+
"eval_steps_per_second": 2.937,
|
| 293 |
+
"eval_wer": 0.9987900786448881,
|
| 294 |
+
"step": 2500
|
| 295 |
+
},
|
| 296 |
+
{
|
| 297 |
+
"epoch": 6.914893617021277,
|
| 298 |
+
"eval_cer": 0.20628344283888417,
|
| 299 |
+
"eval_loss": 0.4180741608142853,
|
| 300 |
+
"eval_runtime": 210.3304,
|
| 301 |
+
"eval_samples_per_second": 23.577,
|
| 302 |
+
"eval_steps_per_second": 2.948,
|
| 303 |
+
"eval_wer": 0.9989917322040734,
|
| 304 |
+
"step": 2600
|
| 305 |
+
},
|
| 306 |
+
{
|
| 307 |
+
"epoch": 7.180851063829787,
|
| 308 |
+
"eval_cer": 0.2011983724647934,
|
| 309 |
+
"eval_loss": 0.4185584485530853,
|
| 310 |
+
"eval_runtime": 209.9592,
|
| 311 |
+
"eval_samples_per_second": 23.619,
|
| 312 |
+
"eval_steps_per_second": 2.953,
|
| 313 |
+
"eval_wer": 0.9957652752571083,
|
| 314 |
+
"step": 2700
|
| 315 |
+
},
|
| 316 |
+
{
|
| 317 |
+
"epoch": 7.446808510638298,
|
| 318 |
+
"eval_cer": 0.20350052961340176,
|
| 319 |
+
"eval_loss": 0.4253988265991211,
|
| 320 |
+
"eval_runtime": 212.3471,
|
| 321 |
+
"eval_samples_per_second": 23.353,
|
| 322 |
+
"eval_steps_per_second": 2.92,
|
| 323 |
+
"eval_wer": 0.999395039322444,
|
| 324 |
+
"step": 2800
|
| 325 |
+
},
|
| 326 |
+
{
|
| 327 |
+
"epoch": 7.712765957446808,
|
| 328 |
+
"eval_cer": 0.1980519607421623,
|
| 329 |
+
"eval_loss": 0.4013788104057312,
|
| 330 |
+
"eval_runtime": 213.4065,
|
| 331 |
+
"eval_samples_per_second": 23.237,
|
| 332 |
+
"eval_steps_per_second": 2.905,
|
| 333 |
+
"eval_wer": 0.9983867715265174,
|
| 334 |
+
"step": 2900
|
| 335 |
+
},
|
| 336 |
+
{
|
| 337 |
+
"epoch": 7.9787234042553195,
|
| 338 |
+
"grad_norm": 2.3959779739379883,
|
| 339 |
+
"learning_rate": 7.1904e-05,
|
| 340 |
+
"loss": 0.3375,
|
| 341 |
+
"step": 3000
|
| 342 |
+
},
|
| 343 |
+
{
|
| 344 |
+
"epoch": 7.9787234042553195,
|
| 345 |
+
"eval_cer": 0.19799333195229962,
|
| 346 |
+
"eval_loss": 0.38772091269493103,
|
| 347 |
+
"eval_runtime": 212.9922,
|
| 348 |
+
"eval_samples_per_second": 23.283,
|
| 349 |
+
"eval_steps_per_second": 2.911,
|
| 350 |
+
"eval_wer": 0.9975801572897761,
|
| 351 |
+
"step": 3000
|
| 352 |
+
},
|
| 353 |
+
{
|
| 354 |
+
"epoch": 8.24468085106383,
|
| 355 |
+
"eval_cer": 0.19256821459700524,
|
| 356 |
+
"eval_loss": 0.3868231475353241,
|
| 357 |
+
"eval_runtime": 211.4945,
|
| 358 |
+
"eval_samples_per_second": 23.447,
|
| 359 |
+
"eval_steps_per_second": 2.932,
|
| 360 |
+
"eval_wer": 0.9981851179673321,
|
| 361 |
+
"step": 3100
|
| 362 |
+
},
|
| 363 |
+
{
|
| 364 |
+
"epoch": 8.51063829787234,
|
| 365 |
+
"eval_cer": 0.19027778320636943,
|
| 366 |
+
"eval_loss": 0.374016672372818,
|
| 367 |
+
"eval_runtime": 207.4949,
|
| 368 |
+
"eval_samples_per_second": 23.899,
|
| 369 |
+
"eval_steps_per_second": 2.988,
|
| 370 |
+
"eval_wer": 0.9977818108489614,
|
| 371 |
+
"step": 3200
|
| 372 |
+
},
|
| 373 |
+
{
|
| 374 |
+
"epoch": 8.77659574468085,
|
| 375 |
+
"eval_cer": 0.18545067950767452,
|
| 376 |
+
"eval_loss": 0.36447063088417053,
|
| 377 |
+
"eval_runtime": 208.0584,
|
| 378 |
+
"eval_samples_per_second": 23.835,
|
| 379 |
+
"eval_steps_per_second": 2.98,
|
| 380 |
+
"eval_wer": 0.9981851179673321,
|
| 381 |
+
"step": 3300
|
| 382 |
+
},
|
| 383 |
+
{
|
| 384 |
+
"epoch": 9.042553191489361,
|
| 385 |
+
"eval_cer": 0.18235117081693356,
|
| 386 |
+
"eval_loss": 0.35864943265914917,
|
| 387 |
+
"eval_runtime": 233.9479,
|
| 388 |
+
"eval_samples_per_second": 21.197,
|
| 389 |
+
"eval_steps_per_second": 2.65,
|
| 390 |
+
"eval_wer": 0.9991933857632587,
|
| 391 |
+
"step": 3400
|
| 392 |
+
},
|
| 393 |
+
{
|
| 394 |
+
"epoch": 9.308510638297872,
|
| 395 |
+
"grad_norm": 1.6243752241134644,
|
| 396 |
+
"learning_rate": 8.390399999999999e-05,
|
| 397 |
+
"loss": 0.2553,
|
| 398 |
+
"step": 3500
|
| 399 |
+
},
|
| 400 |
+
{
|
| 401 |
+
"epoch": 9.308510638297872,
|
| 402 |
+
"eval_cer": 0.1774537125704034,
|
| 403 |
+
"eval_loss": 0.3328181505203247,
|
| 404 |
+
"eval_runtime": 208.1328,
|
| 405 |
+
"eval_samples_per_second": 23.826,
|
| 406 |
+
"eval_steps_per_second": 2.979,
|
| 407 |
+
"eval_wer": 0.9979834644081468,
|
| 408 |
+
"step": 3500
|
| 409 |
+
},
|
| 410 |
+
{
|
| 411 |
+
"epoch": 9.574468085106384,
|
| 412 |
+
"eval_cer": 0.174365929637635,
|
| 413 |
+
"eval_loss": 0.34012970328330994,
|
| 414 |
+
"eval_runtime": 208.9838,
|
| 415 |
+
"eval_samples_per_second": 23.729,
|
| 416 |
+
"eval_steps_per_second": 2.967,
|
| 417 |
+
"eval_wer": 0.9977818108489614,
|
| 418 |
+
"step": 3600
|
| 419 |
+
},
|
| 420 |
+
{
|
| 421 |
+
"epoch": 9.840425531914894,
|
| 422 |
+
"eval_cer": 0.1727282321074705,
|
| 423 |
+
"eval_loss": 0.31242451071739197,
|
| 424 |
+
"eval_runtime": 209.0076,
|
| 425 |
+
"eval_samples_per_second": 23.726,
|
| 426 |
+
"eval_steps_per_second": 2.966,
|
| 427 |
+
"eval_wer": 0.9975801572897761,
|
| 428 |
+
"step": 3700
|
| 429 |
+
},
|
| 430 |
+
{
|
| 431 |
+
"epoch": 10.106382978723405,
|
| 432 |
+
"eval_cer": 0.1708716537618186,
|
| 433 |
+
"eval_loss": 0.322468101978302,
|
| 434 |
+
"eval_runtime": 207.4311,
|
| 435 |
+
"eval_samples_per_second": 23.907,
|
| 436 |
+
"eval_steps_per_second": 2.989,
|
| 437 |
+
"eval_wer": 0.9987900786448881,
|
| 438 |
+
"step": 3800
|
| 439 |
+
},
|
| 440 |
+
{
|
| 441 |
+
"epoch": 10.372340425531915,
|
| 442 |
+
"eval_cer": 0.17599971858180866,
|
| 443 |
+
"eval_loss": 0.3310896158218384,
|
| 444 |
+
"eval_runtime": 206.3784,
|
| 445 |
+
"eval_samples_per_second": 24.029,
|
| 446 |
+
"eval_steps_per_second": 3.004,
|
| 447 |
+
"eval_wer": 0.9973785037305909,
|
| 448 |
+
"step": 3900
|
| 449 |
+
},
|
| 450 |
+
{
|
| 451 |
+
"epoch": 10.638297872340425,
|
| 452 |
+
"grad_norm": 2.6917850971221924,
|
| 453 |
+
"learning_rate": 9.5904e-05,
|
| 454 |
+
"loss": 0.2035,
|
| 455 |
+
"step": 4000
|
| 456 |
+
},
|
| 457 |
+
{
|
| 458 |
+
"epoch": 10.638297872340425,
|
| 459 |
+
"eval_cer": 0.17048861233471568,
|
| 460 |
+
"eval_loss": 0.3098428249359131,
|
| 461 |
+
"eval_runtime": 207.2374,
|
| 462 |
+
"eval_samples_per_second": 23.929,
|
| 463 |
+
"eval_steps_per_second": 2.992,
|
| 464 |
+
"eval_wer": 0.9979834644081468,
|
| 465 |
+
"step": 4000
|
| 466 |
+
},
|
| 467 |
+
{
|
| 468 |
+
"epoch": 10.904255319148936,
|
| 469 |
+
"eval_cer": 0.1713641355966652,
|
| 470 |
+
"eval_loss": 0.32444971799850464,
|
| 471 |
+
"eval_runtime": 207.2878,
|
| 472 |
+
"eval_samples_per_second": 23.923,
|
| 473 |
+
"eval_steps_per_second": 2.991,
|
| 474 |
+
"eval_wer": 0.9979834644081468,
|
| 475 |
+
"step": 4100
|
| 476 |
+
},
|
| 477 |
+
{
|
| 478 |
+
"epoch": 11.170212765957446,
|
| 479 |
+
"eval_cer": 0.16859294812915532,
|
| 480 |
+
"eval_loss": 0.3280084431171417,
|
| 481 |
+
"eval_runtime": 206.6779,
|
| 482 |
+
"eval_samples_per_second": 23.994,
|
| 483 |
+
"eval_steps_per_second": 3.0,
|
| 484 |
+
"eval_wer": 0.9925388183101431,
|
| 485 |
+
"step": 4200
|
| 486 |
+
},
|
| 487 |
+
{
|
| 488 |
+
"epoch": 11.436170212765958,
|
| 489 |
+
"eval_cer": 0.17051206385066076,
|
| 490 |
+
"eval_loss": 0.31342190504074097,
|
| 491 |
+
"eval_runtime": 214.4373,
|
| 492 |
+
"eval_samples_per_second": 23.126,
|
| 493 |
+
"eval_steps_per_second": 2.891,
|
| 494 |
+
"eval_wer": 0.9983867715265174,
|
| 495 |
+
"step": 4300
|
| 496 |
+
},
|
| 497 |
+
{
|
| 498 |
+
"epoch": 11.702127659574469,
|
| 499 |
+
"eval_cer": 0.16671291826755835,
|
| 500 |
+
"eval_loss": 0.3024933338165283,
|
| 501 |
+
"eval_runtime": 208.7095,
|
| 502 |
+
"eval_samples_per_second": 23.76,
|
| 503 |
+
"eval_steps_per_second": 2.971,
|
| 504 |
+
"eval_wer": 0.9987900786448881,
|
| 505 |
+
"step": 4400
|
| 506 |
+
},
|
| 507 |
+
{
|
| 508 |
+
"epoch": 11.96808510638298,
|
| 509 |
+
"grad_norm": 1.652050256729126,
|
| 510 |
+
"learning_rate": 0.00010787999999999998,
|
| 511 |
+
"loss": 0.1772,
|
| 512 |
+
"step": 4500
|
| 513 |
+
},
|
| 514 |
+
{
|
| 515 |
+
"epoch": 11.96808510638298,
|
| 516 |
+
"eval_cer": 0.16904634410409347,
|
| 517 |
+
"eval_loss": 0.31564900279045105,
|
| 518 |
+
"eval_runtime": 207.6579,
|
| 519 |
+
"eval_samples_per_second": 23.881,
|
| 520 |
+
"eval_steps_per_second": 2.986,
|
| 521 |
+
"eval_wer": 0.9979834644081468,
|
| 522 |
+
"step": 4500
|
| 523 |
+
},
|
| 524 |
+
{
|
| 525 |
+
"epoch": 12.23404255319149,
|
| 526 |
+
"eval_cer": 0.16571232025390176,
|
| 527 |
+
"eval_loss": 0.3212853670120239,
|
| 528 |
+
"eval_runtime": 208.1891,
|
| 529 |
+
"eval_samples_per_second": 23.82,
|
| 530 |
+
"eval_steps_per_second": 2.978,
|
| 531 |
+
"eval_wer": 0.9967735430530349,
|
| 532 |
+
"step": 4600
|
| 533 |
+
},
|
| 534 |
+
{
|
| 535 |
+
"epoch": 12.5,
|
| 536 |
+
"eval_cer": 0.17018374262742966,
|
| 537 |
+
"eval_loss": 0.31842583417892456,
|
| 538 |
+
"eval_runtime": 207.4188,
|
| 539 |
+
"eval_samples_per_second": 23.908,
|
| 540 |
+
"eval_steps_per_second": 2.989,
|
| 541 |
+
"eval_wer": 0.9975801572897761,
|
| 542 |
+
"step": 4700
|
| 543 |
+
},
|
| 544 |
+
{
|
| 545 |
+
"epoch": 12.76595744680851,
|
| 546 |
+
"eval_cer": 0.16590774955344406,
|
| 547 |
+
"eval_loss": 0.3348098397254944,
|
| 548 |
+
"eval_runtime": 206.5664,
|
| 549 |
+
"eval_samples_per_second": 24.007,
|
| 550 |
+
"eval_steps_per_second": 3.001,
|
| 551 |
+
"eval_wer": 0.9989917322040734,
|
| 552 |
+
"step": 4800
|
| 553 |
+
},
|
| 554 |
+
{
|
| 555 |
+
"epoch": 13.03191489361702,
|
| 556 |
+
"eval_cer": 0.16545044499251504,
|
| 557 |
+
"eval_loss": 0.31754830479621887,
|
| 558 |
+
"eval_runtime": 204.6212,
|
| 559 |
+
"eval_samples_per_second": 24.235,
|
| 560 |
+
"eval_steps_per_second": 3.03,
|
| 561 |
+
"eval_wer": 0.9977818108489614,
|
| 562 |
+
"step": 4900
|
| 563 |
+
},
|
| 564 |
+
{
|
| 565 |
+
"epoch": 13.297872340425531,
|
| 566 |
+
"grad_norm": 1.4975948333740234,
|
| 567 |
+
"learning_rate": 0.00011988,
|
| 568 |
+
"loss": 0.1542,
|
| 569 |
+
"step": 5000
|
| 570 |
+
},
|
| 571 |
+
{
|
| 572 |
+
"epoch": 13.297872340425531,
|
| 573 |
+
"eval_cer": 0.16795975719863826,
|
| 574 |
+
"eval_loss": 0.3414091467857361,
|
| 575 |
+
"eval_runtime": 212.8965,
|
| 576 |
+
"eval_samples_per_second": 23.293,
|
| 577 |
+
"eval_steps_per_second": 2.912,
|
| 578 |
+
"eval_wer": 0.9997983464408147,
|
| 579 |
+
"step": 5000
|
| 580 |
+
},
|
| 581 |
+
{
|
| 582 |
+
"epoch": 13.563829787234042,
|
| 583 |
+
"eval_cer": 0.17008993656364937,
|
| 584 |
+
"eval_loss": 0.3142850399017334,
|
| 585 |
+
"eval_runtime": 206.8293,
|
| 586 |
+
"eval_samples_per_second": 23.976,
|
| 587 |
+
"eval_steps_per_second": 2.998,
|
| 588 |
+
"eval_wer": 0.999395039322444,
|
| 589 |
+
"step": 5100
|
| 590 |
+
},
|
| 591 |
+
{
|
| 592 |
+
"epoch": 13.829787234042554,
|
| 593 |
+
"eval_cer": 0.16879228601468846,
|
| 594 |
+
"eval_loss": 0.3204251229763031,
|
| 595 |
+
"eval_runtime": 205.7182,
|
| 596 |
+
"eval_samples_per_second": 24.106,
|
| 597 |
+
"eval_steps_per_second": 3.014,
|
| 598 |
+
"eval_wer": 0.9985884250857028,
|
| 599 |
+
"step": 5200
|
| 600 |
+
},
|
| 601 |
+
{
|
| 602 |
+
"epoch": 14.095744680851064,
|
| 603 |
+
"eval_cer": 0.16618525915879412,
|
| 604 |
+
"eval_loss": 0.35486334562301636,
|
| 605 |
+
"eval_runtime": 206.3465,
|
| 606 |
+
"eval_samples_per_second": 24.032,
|
| 607 |
+
"eval_steps_per_second": 3.005,
|
| 608 |
+
"eval_wer": 0.9989917322040734,
|
| 609 |
+
"step": 5300
|
| 610 |
+
},
|
| 611 |
+
{
|
| 612 |
+
"epoch": 14.361702127659575,
|
| 613 |
+
"eval_cer": 0.1666073864458055,
|
| 614 |
+
"eval_loss": 0.40914931893348694,
|
| 615 |
+
"eval_runtime": 206.0512,
|
| 616 |
+
"eval_samples_per_second": 24.067,
|
| 617 |
+
"eval_steps_per_second": 3.009,
|
| 618 |
+
"eval_wer": 0.9973785037305909,
|
| 619 |
+
"step": 5400
|
| 620 |
+
},
|
| 621 |
+
{
|
| 622 |
+
"epoch": 14.627659574468085,
|
| 623 |
+
"grad_norm": 2.959421157836914,
|
| 624 |
+
"learning_rate": 0.000131856,
|
| 625 |
+
"loss": 0.1449,
|
| 626 |
+
"step": 5500
|
| 627 |
+
},
|
| 628 |
+
{
|
| 629 |
+
"epoch": 14.627659574468085,
|
| 630 |
+
"eval_cer": 0.16760407587347126,
|
| 631 |
+
"eval_loss": 0.3907996714115143,
|
| 632 |
+
"eval_runtime": 205.2139,
|
| 633 |
+
"eval_samples_per_second": 24.165,
|
| 634 |
+
"eval_steps_per_second": 3.021,
|
| 635 |
+
"eval_wer": 0.9985884250857028,
|
| 636 |
+
"step": 5500
|
| 637 |
+
},
|
| 638 |
+
{
|
| 639 |
+
"epoch": 14.893617021276595,
|
| 640 |
+
"eval_cer": 0.1662243450187026,
|
| 641 |
+
"eval_loss": 0.3705514073371887,
|
| 642 |
+
"eval_runtime": 205.3677,
|
| 643 |
+
"eval_samples_per_second": 24.147,
|
| 644 |
+
"eval_steps_per_second": 3.019,
|
| 645 |
+
"eval_wer": 0.9983867715265174,
|
| 646 |
+
"step": 5600
|
| 647 |
+
},
|
| 648 |
+
{
|
| 649 |
+
"epoch": 15.159574468085106,
|
| 650 |
+
"eval_cer": 0.164105891411664,
|
| 651 |
+
"eval_loss": 0.3971535861492157,
|
| 652 |
+
"eval_runtime": 205.6896,
|
| 653 |
+
"eval_samples_per_second": 24.109,
|
| 654 |
+
"eval_steps_per_second": 3.014,
|
| 655 |
+
"eval_wer": 0.9971768501714056,
|
| 656 |
+
"step": 5700
|
| 657 |
+
},
|
| 658 |
+
{
|
| 659 |
+
"epoch": 15.425531914893616,
|
| 660 |
+
"eval_cer": 0.1653097358968446,
|
| 661 |
+
"eval_loss": 0.3461838662624359,
|
| 662 |
+
"eval_runtime": 206.3509,
|
| 663 |
+
"eval_samples_per_second": 24.032,
|
| 664 |
+
"eval_steps_per_second": 3.005,
|
| 665 |
+
"eval_wer": 0.9983867715265174,
|
| 666 |
+
"step": 5800
|
| 667 |
+
},
|
| 668 |
+
{
|
| 669 |
+
"epoch": 15.691489361702128,
|
| 670 |
+
"eval_cer": 0.16992186736604298,
|
| 671 |
+
"eval_loss": 0.35435542464256287,
|
| 672 |
+
"eval_runtime": 205.3227,
|
| 673 |
+
"eval_samples_per_second": 24.152,
|
| 674 |
+
"eval_steps_per_second": 3.02,
|
| 675 |
+
"eval_wer": 0.9983867715265174,
|
| 676 |
+
"step": 5900
|
| 677 |
+
},
|
| 678 |
+
{
|
| 679 |
+
"epoch": 15.957446808510639,
|
| 680 |
+
"grad_norm": 1.915339469909668,
|
| 681 |
+
"learning_rate": 0.00014385599999999999,
|
| 682 |
+
"loss": 0.1396,
|
| 683 |
+
"step": 6000
|
| 684 |
+
},
|
| 685 |
+
{
|
| 686 |
+
"epoch": 15.957446808510639,
|
| 687 |
+
"eval_cer": 0.1681590950841714,
|
| 688 |
+
"eval_loss": 0.33965829014778137,
|
| 689 |
+
"eval_runtime": 205.3964,
|
| 690 |
+
"eval_samples_per_second": 24.144,
|
| 691 |
+
"eval_steps_per_second": 3.019,
|
| 692 |
+
"eval_wer": 0.9987900786448881,
|
| 693 |
+
"step": 6000
|
| 694 |
+
},
|
| 695 |
+
{
|
| 696 |
+
"epoch": 16.22340425531915,
|
| 697 |
+
"eval_cer": 0.16804574609043688,
|
| 698 |
+
"eval_loss": 0.3451753556728363,
|
| 699 |
+
"eval_runtime": 205.8971,
|
| 700 |
+
"eval_samples_per_second": 24.085,
|
| 701 |
+
"eval_steps_per_second": 3.011,
|
| 702 |
+
"eval_wer": 0.9983867715265174,
|
| 703 |
+
"step": 6100
|
| 704 |
+
},
|
| 705 |
+
{
|
| 706 |
+
"epoch": 16.48936170212766,
|
| 707 |
+
"eval_cer": 0.16652921472598858,
|
| 708 |
+
"eval_loss": 0.3533567488193512,
|
| 709 |
+
"eval_runtime": 205.4803,
|
| 710 |
+
"eval_samples_per_second": 24.134,
|
| 711 |
+
"eval_steps_per_second": 3.017,
|
| 712 |
+
"eval_wer": 0.9981851179673321,
|
| 713 |
+
"step": 6200
|
| 714 |
+
},
|
| 715 |
+
{
|
| 716 |
+
"epoch": 16.75531914893617,
|
| 717 |
+
"eval_cer": 0.17028145727720082,
|
| 718 |
+
"eval_loss": 0.3502165377140045,
|
| 719 |
+
"eval_runtime": 205.7767,
|
| 720 |
+
"eval_samples_per_second": 24.099,
|
| 721 |
+
"eval_steps_per_second": 3.013,
|
| 722 |
+
"eval_wer": 0.9985884250857028,
|
| 723 |
+
"step": 6300
|
| 724 |
+
},
|
| 725 |
+
{
|
| 726 |
+
"epoch": 17.02127659574468,
|
| 727 |
+
"eval_cer": 0.17007430221968597,
|
| 728 |
+
"eval_loss": 0.34754782915115356,
|
| 729 |
+
"eval_runtime": 205.7613,
|
| 730 |
+
"eval_samples_per_second": 24.101,
|
| 731 |
+
"eval_steps_per_second": 3.013,
|
| 732 |
+
"eval_wer": 0.9977818108489614,
|
| 733 |
+
"step": 6400
|
| 734 |
+
},
|
| 735 |
+
{
|
| 736 |
+
"epoch": 17.28723404255319,
|
| 737 |
+
"grad_norm": 1.2938112020492554,
|
| 738 |
+
"learning_rate": 0.00015585599999999998,
|
| 739 |
+
"loss": 0.1293,
|
| 740 |
+
"step": 6500
|
| 741 |
+
},
|
| 742 |
+
{
|
| 743 |
+
"epoch": 17.28723404255319,
|
| 744 |
+
"eval_cer": 0.1681239178102538,
|
| 745 |
+
"eval_loss": 0.3350105881690979,
|
| 746 |
+
"eval_runtime": 206.0245,
|
| 747 |
+
"eval_samples_per_second": 24.07,
|
| 748 |
+
"eval_steps_per_second": 3.009,
|
| 749 |
+
"eval_wer": 0.9987900786448881,
|
| 750 |
+
"step": 6500
|
| 751 |
+
},
|
| 752 |
+
{
|
| 753 |
+
"epoch": 17.5531914893617,
|
| 754 |
+
"eval_cer": 0.16944501987515975,
|
| 755 |
+
"eval_loss": 0.3442031443119049,
|
| 756 |
+
"eval_runtime": 205.3927,
|
| 757 |
+
"eval_samples_per_second": 24.144,
|
| 758 |
+
"eval_steps_per_second": 3.019,
|
| 759 |
+
"eval_wer": 0.9977818108489614,
|
| 760 |
+
"step": 6600
|
| 761 |
+
},
|
| 762 |
+
{
|
| 763 |
+
"epoch": 17.819148936170212,
|
| 764 |
+
"eval_cer": 0.16870629712288984,
|
| 765 |
+
"eval_loss": 0.33415013551712036,
|
| 766 |
+
"eval_runtime": 205.9912,
|
| 767 |
+
"eval_samples_per_second": 24.074,
|
| 768 |
+
"eval_steps_per_second": 3.01,
|
| 769 |
+
"eval_wer": 0.9987900786448881,
|
| 770 |
+
"step": 6700
|
| 771 |
+
},
|
| 772 |
+
{
|
| 773 |
+
"epoch": 18.085106382978722,
|
| 774 |
+
"eval_cer": 0.1696013633147936,
|
| 775 |
+
"eval_loss": 0.3668520450592041,
|
| 776 |
+
"eval_runtime": 205.5898,
|
| 777 |
+
"eval_samples_per_second": 24.121,
|
| 778 |
+
"eval_steps_per_second": 3.016,
|
| 779 |
+
"eval_wer": 0.9985884250857028,
|
| 780 |
+
"step": 6800
|
| 781 |
+
},
|
| 782 |
+
{
|
| 783 |
+
"epoch": 18.351063829787233,
|
| 784 |
+
"eval_cer": 0.16906197844805684,
|
| 785 |
+
"eval_loss": 0.34041330218315125,
|
| 786 |
+
"eval_runtime": 205.1961,
|
| 787 |
+
"eval_samples_per_second": 24.167,
|
| 788 |
+
"eval_steps_per_second": 3.022,
|
| 789 |
+
"eval_wer": 0.9969751966122202,
|
| 790 |
+
"step": 6900
|
| 791 |
+
},
|
| 792 |
+
{
|
| 793 |
+
"epoch": 18.617021276595743,
|
| 794 |
+
"grad_norm": 2.553234815597534,
|
| 795 |
+
"learning_rate": 0.000167856,
|
| 796 |
+
"loss": 0.1276,
|
| 797 |
+
"step": 7000
|
| 798 |
+
},
|
| 799 |
+
{
|
| 800 |
+
"epoch": 18.617021276595743,
|
| 801 |
+
"eval_cer": 0.16789331123679385,
|
| 802 |
+
"eval_loss": 0.3463904857635498,
|
| 803 |
+
"eval_runtime": 204.7479,
|
| 804 |
+
"eval_samples_per_second": 24.22,
|
| 805 |
+
"eval_steps_per_second": 3.028,
|
| 806 |
+
"eval_wer": 0.9989917322040734,
|
| 807 |
+
"step": 7000
|
| 808 |
+
},
|
| 809 |
+
{
|
| 810 |
+
"epoch": 18.882978723404257,
|
| 811 |
+
"eval_cer": 0.16948801432105906,
|
| 812 |
+
"eval_loss": 0.3496261537075043,
|
| 813 |
+
"eval_runtime": 205.3143,
|
| 814 |
+
"eval_samples_per_second": 24.153,
|
| 815 |
+
"eval_steps_per_second": 3.02,
|
| 816 |
+
"eval_wer": 0.9983867715265174,
|
| 817 |
+
"step": 7100
|
| 818 |
+
},
|
| 819 |
+
{
|
| 820 |
+
"epoch": 19.148936170212767,
|
| 821 |
+
"eval_cer": 0.16978897544235422,
|
| 822 |
+
"eval_loss": 0.3436301350593567,
|
| 823 |
+
"eval_runtime": 205.1734,
|
| 824 |
+
"eval_samples_per_second": 24.17,
|
| 825 |
+
"eval_steps_per_second": 3.022,
|
| 826 |
+
"eval_wer": 0.9967735430530349,
|
| 827 |
+
"step": 7200
|
| 828 |
+
},
|
| 829 |
+
{
|
| 830 |
+
"epoch": 19.414893617021278,
|
| 831 |
+
"eval_cer": 0.16896035521229485,
|
| 832 |
+
"eval_loss": 0.36047497391700745,
|
| 833 |
+
"eval_runtime": 205.4789,
|
| 834 |
+
"eval_samples_per_second": 24.134,
|
| 835 |
+
"eval_steps_per_second": 3.017,
|
| 836 |
+
"eval_wer": 0.9953619681387377,
|
| 837 |
+
"step": 7300
|
| 838 |
+
},
|
| 839 |
+
{
|
| 840 |
+
"epoch": 19.680851063829788,
|
| 841 |
+
"eval_cer": 0.16866330267699053,
|
| 842 |
+
"eval_loss": 0.3582172989845276,
|
| 843 |
+
"eval_runtime": 205.1102,
|
| 844 |
+
"eval_samples_per_second": 24.177,
|
| 845 |
+
"eval_steps_per_second": 3.023,
|
| 846 |
+
"eval_wer": 0.9973785037305909,
|
| 847 |
+
"step": 7400
|
| 848 |
+
},
|
| 849 |
+
{
|
| 850 |
+
"epoch": 19.9468085106383,
|
| 851 |
+
"grad_norm": 1.7617504596710205,
|
| 852 |
+
"learning_rate": 0.000179856,
|
| 853 |
+
"loss": 0.1264,
|
| 854 |
+
"step": 7500
|
| 855 |
+
},
|
| 856 |
+
{
|
| 857 |
+
"epoch": 19.9468085106383,
|
| 858 |
+
"eval_cer": 0.17402197407044054,
|
| 859 |
+
"eval_loss": 0.3575660288333893,
|
| 860 |
+
"eval_runtime": 204.7059,
|
| 861 |
+
"eval_samples_per_second": 24.225,
|
| 862 |
+
"eval_steps_per_second": 3.029,
|
| 863 |
+
"eval_wer": 0.9981851179673321,
|
| 864 |
+
"step": 7500
|
| 865 |
+
},
|
| 866 |
+
{
|
| 867 |
+
"epoch": 20.21276595744681,
|
| 868 |
+
"eval_cer": 0.17258752301180003,
|
| 869 |
+
"eval_loss": 0.3668830990791321,
|
| 870 |
+
"eval_runtime": 205.0195,
|
| 871 |
+
"eval_samples_per_second": 24.188,
|
| 872 |
+
"eval_steps_per_second": 3.024,
|
| 873 |
+
"eval_wer": 0.9985884250857028,
|
| 874 |
+
"step": 7600
|
| 875 |
+
},
|
| 876 |
+
{
|
| 877 |
+
"epoch": 20.47872340425532,
|
| 878 |
+
"eval_cer": 0.17064495577434952,
|
| 879 |
+
"eval_loss": 0.361765593290329,
|
| 880 |
+
"eval_runtime": 204.8268,
|
| 881 |
+
"eval_samples_per_second": 24.211,
|
| 882 |
+
"eval_steps_per_second": 3.027,
|
| 883 |
+
"eval_wer": 0.9979834644081468,
|
| 884 |
+
"step": 7700
|
| 885 |
+
},
|
| 886 |
+
{
|
| 887 |
+
"epoch": 20.74468085106383,
|
| 888 |
+
"eval_cer": 0.17463562207100336,
|
| 889 |
+
"eval_loss": 0.34747639298439026,
|
| 890 |
+
"eval_runtime": 204.8046,
|
| 891 |
+
"eval_samples_per_second": 24.213,
|
| 892 |
+
"eval_steps_per_second": 3.027,
|
| 893 |
+
"eval_wer": 0.9989917322040734,
|
| 894 |
+
"step": 7800
|
| 895 |
+
},
|
| 896 |
+
{
|
| 897 |
+
"epoch": 21.01063829787234,
|
| 898 |
+
"eval_cer": 0.1715165704503082,
|
| 899 |
+
"eval_loss": 0.34254932403564453,
|
| 900 |
+
"eval_runtime": 204.6469,
|
| 901 |
+
"eval_samples_per_second": 24.232,
|
| 902 |
+
"eval_steps_per_second": 3.03,
|
| 903 |
+
"eval_wer": 0.9971768501714056,
|
| 904 |
+
"step": 7900
|
| 905 |
+
},
|
| 906 |
+
{
|
| 907 |
+
"epoch": 21.27659574468085,
|
| 908 |
+
"grad_norm": 1.5520191192626953,
|
| 909 |
+
"learning_rate": 0.000191856,
|
| 910 |
+
"loss": 0.1219,
|
| 911 |
+
"step": 8000
|
| 912 |
+
},
|
| 913 |
+
{
|
| 914 |
+
"epoch": 21.27659574468085,
|
| 915 |
+
"eval_cer": 0.1716377366160244,
|
| 916 |
+
"eval_loss": 0.3685350716114044,
|
| 917 |
+
"eval_runtime": 204.5671,
|
| 918 |
+
"eval_samples_per_second": 24.241,
|
| 919 |
+
"eval_steps_per_second": 3.031,
|
| 920 |
+
"eval_wer": 0.9983867715265174,
|
| 921 |
+
"step": 8000
|
| 922 |
+
},
|
| 923 |
+
{
|
| 924 |
+
"epoch": 21.54255319148936,
|
| 925 |
+
"eval_cer": 0.1755893170527698,
|
| 926 |
+
"eval_loss": 0.38032132387161255,
|
| 927 |
+
"eval_runtime": 204.7288,
|
| 928 |
+
"eval_samples_per_second": 24.222,
|
| 929 |
+
"eval_steps_per_second": 3.028,
|
| 930 |
+
"eval_wer": 0.9989917322040734,
|
| 931 |
+
"step": 8100
|
| 932 |
+
},
|
| 933 |
+
{
|
| 934 |
+
"epoch": 21.80851063829787,
|
| 935 |
+
"eval_cer": 0.17974805254703005,
|
| 936 |
+
"eval_loss": 0.36628487706184387,
|
| 937 |
+
"eval_runtime": 205.2564,
|
| 938 |
+
"eval_samples_per_second": 24.16,
|
| 939 |
+
"eval_steps_per_second": 3.021,
|
| 940 |
+
"eval_wer": 0.9983867715265174,
|
| 941 |
+
"step": 8200
|
| 942 |
+
},
|
| 943 |
+
{
|
| 944 |
+
"epoch": 22.074468085106382,
|
| 945 |
+
"eval_cer": 0.1709576426536172,
|
| 946 |
+
"eval_loss": 0.3641684055328369,
|
| 947 |
+
"eval_runtime": 205.4639,
|
| 948 |
+
"eval_samples_per_second": 24.136,
|
| 949 |
+
"eval_steps_per_second": 3.018,
|
| 950 |
+
"eval_wer": 0.9977818108489614,
|
| 951 |
+
"step": 8300
|
| 952 |
+
},
|
| 953 |
+
{
|
| 954 |
+
"epoch": 22.340425531914892,
|
| 955 |
+
"eval_cer": 0.17367801850324607,
|
| 956 |
+
"eval_loss": 0.38052114844322205,
|
| 957 |
+
"eval_runtime": 205.4218,
|
| 958 |
+
"eval_samples_per_second": 24.141,
|
| 959 |
+
"eval_steps_per_second": 3.018,
|
| 960 |
+
"eval_wer": 0.9987900786448881,
|
| 961 |
+
"step": 8400
|
| 962 |
+
},
|
| 963 |
+
{
|
| 964 |
+
"epoch": 22.606382978723403,
|
| 965 |
+
"grad_norm": 2.203418493270874,
|
| 966 |
+
"learning_rate": 0.00020385599999999998,
|
| 967 |
+
"loss": 0.1177,
|
| 968 |
+
"step": 8500
|
| 969 |
+
},
|
| 970 |
+
{
|
| 971 |
+
"epoch": 22.606382978723403,
|
| 972 |
+
"eval_cer": 0.17468252510289353,
|
| 973 |
+
"eval_loss": 0.36295458674430847,
|
| 974 |
+
"eval_runtime": 205.462,
|
| 975 |
+
"eval_samples_per_second": 24.136,
|
| 976 |
+
"eval_steps_per_second": 3.018,
|
| 977 |
+
"eval_wer": 0.9985884250857028,
|
| 978 |
+
"step": 8500
|
| 979 |
+
},
|
| 980 |
+
{
|
| 981 |
+
"epoch": 22.872340425531917,
|
| 982 |
+
"eval_cer": 0.1753078988614289,
|
| 983 |
+
"eval_loss": 0.4001345634460449,
|
| 984 |
+
"eval_runtime": 205.3238,
|
| 985 |
+
"eval_samples_per_second": 24.152,
|
| 986 |
+
"eval_steps_per_second": 3.02,
|
| 987 |
+
"eval_wer": 0.9973785037305909,
|
| 988 |
+
"step": 8600
|
| 989 |
+
},
|
| 990 |
+
{
|
| 991 |
+
"epoch": 23.138297872340427,
|
| 992 |
+
"eval_cer": 0.17578865493830298,
|
| 993 |
+
"eval_loss": 0.375848650932312,
|
| 994 |
+
"eval_runtime": 205.2532,
|
| 995 |
+
"eval_samples_per_second": 24.16,
|
| 996 |
+
"eval_steps_per_second": 3.021,
|
| 997 |
+
"eval_wer": 0.9977818108489614,
|
| 998 |
+
"step": 8700
|
| 999 |
+
},
|
| 1000 |
+
{
|
| 1001 |
+
"epoch": 23.404255319148938,
|
| 1002 |
+
"eval_cer": 0.17470988520482944,
|
| 1003 |
+
"eval_loss": 0.37709999084472656,
|
| 1004 |
+
"eval_runtime": 205.004,
|
| 1005 |
+
"eval_samples_per_second": 24.19,
|
| 1006 |
+
"eval_steps_per_second": 3.024,
|
| 1007 |
+
"eval_wer": 0.9983867715265174,
|
| 1008 |
+
"step": 8800
|
| 1009 |
+
},
|
| 1010 |
+
{
|
| 1011 |
+
"epoch": 23.670212765957448,
|
| 1012 |
+
"eval_cer": 0.17936501111992714,
|
| 1013 |
+
"eval_loss": 0.4001403748989105,
|
| 1014 |
+
"eval_runtime": 204.8778,
|
| 1015 |
+
"eval_samples_per_second": 24.205,
|
| 1016 |
+
"eval_steps_per_second": 3.026,
|
| 1017 |
+
"eval_wer": 0.9983867715265174,
|
| 1018 |
+
"step": 8900
|
| 1019 |
+
},
|
| 1020 |
+
{
|
| 1021 |
+
"epoch": 23.93617021276596,
|
| 1022 |
+
"grad_norm": 1.3760440349578857,
|
| 1023 |
+
"learning_rate": 0.000215856,
|
| 1024 |
+
"loss": 0.1241,
|
| 1025 |
+
"step": 9000
|
| 1026 |
+
},
|
| 1027 |
+
{
|
| 1028 |
+
"epoch": 23.93617021276596,
|
| 1029 |
+
"eval_cer": 0.17689087618772156,
|
| 1030 |
+
"eval_loss": 0.3929387629032135,
|
| 1031 |
+
"eval_runtime": 205.6986,
|
| 1032 |
+
"eval_samples_per_second": 24.108,
|
| 1033 |
+
"eval_steps_per_second": 3.014,
|
| 1034 |
+
"eval_wer": 0.9997983464408147,
|
| 1035 |
+
"step": 9000
|
| 1036 |
+
},
|
| 1037 |
+
{
|
| 1038 |
+
"epoch": 24.20212765957447,
|
| 1039 |
+
"eval_cer": 0.17517500693774013,
|
| 1040 |
+
"eval_loss": 0.3732432425022125,
|
| 1041 |
+
"eval_runtime": 205.6021,
|
| 1042 |
+
"eval_samples_per_second": 24.119,
|
| 1043 |
+
"eval_steps_per_second": 3.016,
|
| 1044 |
+
"eval_wer": 0.9991933857632587,
|
| 1045 |
+
"step": 9100
|
| 1046 |
+
},
|
| 1047 |
+
{
|
| 1048 |
+
"epoch": 24.46808510638298,
|
| 1049 |
+
"eval_cer": 0.17383045335688907,
|
| 1050 |
+
"eval_loss": 0.3813112676143646,
|
| 1051 |
+
"eval_runtime": 205.3982,
|
| 1052 |
+
"eval_samples_per_second": 24.143,
|
| 1053 |
+
"eval_steps_per_second": 3.019,
|
| 1054 |
+
"eval_wer": 0.9983867715265174,
|
| 1055 |
+
"step": 9200
|
| 1056 |
+
},
|
| 1057 |
+
{
|
| 1058 |
+
"epoch": 24.73404255319149,
|
| 1059 |
+
"eval_cer": 0.17939627980785391,
|
| 1060 |
+
"eval_loss": 0.41280829906463623,
|
| 1061 |
+
"eval_runtime": 219.7977,
|
| 1062 |
+
"eval_samples_per_second": 22.562,
|
| 1063 |
+
"eval_steps_per_second": 2.821,
|
| 1064 |
+
"eval_wer": 0.9989917322040734,
|
| 1065 |
+
"step": 9300
|
| 1066 |
+
},
|
| 1067 |
+
{
|
| 1068 |
+
"epoch": 25.0,
|
| 1069 |
+
"eval_cer": 0.17513592107783169,
|
| 1070 |
+
"eval_loss": 0.3756047785282135,
|
| 1071 |
+
"eval_runtime": 219.9345,
|
| 1072 |
+
"eval_samples_per_second": 22.548,
|
| 1073 |
+
"eval_steps_per_second": 2.819,
|
| 1074 |
+
"eval_wer": 0.9989917322040734,
|
| 1075 |
+
"step": 9400
|
| 1076 |
+
},
|
| 1077 |
+
{
|
| 1078 |
+
"epoch": 25.26595744680851,
|
| 1079 |
+
"grad_norm": 9.107454299926758,
|
| 1080 |
+
"learning_rate": 0.00022783199999999997,
|
| 1081 |
+
"loss": 0.121,
|
| 1082 |
+
"step": 9500
|
| 1083 |
+
},
|
| 1084 |
+
{
|
| 1085 |
+
"epoch": 25.26595744680851,
|
| 1086 |
+
"eval_cer": 0.17971678385910328,
|
| 1087 |
+
"eval_loss": 0.39162904024124146,
|
| 1088 |
+
"eval_runtime": 219.1957,
|
| 1089 |
+
"eval_samples_per_second": 22.624,
|
| 1090 |
+
"eval_steps_per_second": 2.829,
|
| 1091 |
+
"eval_wer": 0.9989917322040734,
|
| 1092 |
+
"step": 9500
|
| 1093 |
+
},
|
| 1094 |
+
{
|
| 1095 |
+
"epoch": 25.53191489361702,
|
| 1096 |
+
"eval_cer": 0.18236680516089696,
|
| 1097 |
+
"eval_loss": 0.38820743560791016,
|
| 1098 |
+
"eval_runtime": 219.4833,
|
| 1099 |
+
"eval_samples_per_second": 22.594,
|
| 1100 |
+
"eval_steps_per_second": 2.825,
|
| 1101 |
+
"eval_wer": 0.9983867715265174,
|
| 1102 |
+
"step": 9600
|
| 1103 |
+
},
|
| 1104 |
+
{
|
| 1105 |
+
"epoch": 25.79787234042553,
|
| 1106 |
+
"eval_cer": 0.18378562187557407,
|
| 1107 |
+
"eval_loss": 0.39170223474502563,
|
| 1108 |
+
"eval_runtime": 219.5415,
|
| 1109 |
+
"eval_samples_per_second": 22.588,
|
| 1110 |
+
"eval_steps_per_second": 2.824,
|
| 1111 |
+
"eval_wer": 0.9975801572897761,
|
| 1112 |
+
"step": 9700
|
| 1113 |
+
},
|
| 1114 |
+
{
|
| 1115 |
+
"epoch": 26.06382978723404,
|
| 1116 |
+
"eval_cer": 0.17702376811141032,
|
| 1117 |
+
"eval_loss": 0.3927652835845947,
|
| 1118 |
+
"eval_runtime": 219.475,
|
| 1119 |
+
"eval_samples_per_second": 22.595,
|
| 1120 |
+
"eval_steps_per_second": 2.825,
|
| 1121 |
+
"eval_wer": 0.9983867715265174,
|
| 1122 |
+
"step": 9800
|
| 1123 |
+
},
|
| 1124 |
+
{
|
| 1125 |
+
"epoch": 26.329787234042552,
|
| 1126 |
+
"eval_cer": 0.18011155104417875,
|
| 1127 |
+
"eval_loss": 0.3928898870944977,
|
| 1128 |
+
"eval_runtime": 219.0682,
|
| 1129 |
+
"eval_samples_per_second": 22.637,
|
| 1130 |
+
"eval_steps_per_second": 2.83,
|
| 1131 |
+
"eval_wer": 0.9995966928816293,
|
| 1132 |
+
"step": 9900
|
| 1133 |
+
},
|
| 1134 |
+
{
|
| 1135 |
+
"epoch": 26.595744680851062,
|
| 1136 |
+
"grad_norm": 1.6058417558670044,
|
| 1137 |
+
"learning_rate": 0.000239832,
|
| 1138 |
+
"loss": 0.1206,
|
| 1139 |
+
"step": 10000
|
| 1140 |
+
},
|
| 1141 |
+
{
|
| 1142 |
+
"epoch": 26.595744680851062,
|
| 1143 |
+
"eval_cer": 0.17812598936082893,
|
| 1144 |
+
"eval_loss": 0.39845964312553406,
|
| 1145 |
+
"eval_runtime": 219.0245,
|
| 1146 |
+
"eval_samples_per_second": 22.641,
|
| 1147 |
+
"eval_steps_per_second": 2.831,
|
| 1148 |
+
"eval_wer": 0.9987900786448881,
|
| 1149 |
+
"step": 10000
|
| 1150 |
+
},
|
| 1151 |
+
{
|
| 1152 |
+
"epoch": 26.861702127659573,
|
| 1153 |
+
"eval_cer": 0.1770628539713188,
|
| 1154 |
+
"eval_loss": 0.3799264132976532,
|
| 1155 |
+
"eval_runtime": 219.8342,
|
| 1156 |
+
"eval_samples_per_second": 22.558,
|
| 1157 |
+
"eval_steps_per_second": 2.82,
|
| 1158 |
+
"eval_wer": 0.999395039322444,
|
| 1159 |
+
"step": 10100
|
| 1160 |
+
},
|
| 1161 |
+
{
|
| 1162 |
+
"epoch": 27.127659574468087,
|
| 1163 |
+
"eval_cer": 0.1785715681637854,
|
| 1164 |
+
"eval_loss": 0.4022609293460846,
|
| 1165 |
+
"eval_runtime": 221.0183,
|
| 1166 |
+
"eval_samples_per_second": 22.437,
|
| 1167 |
+
"eval_steps_per_second": 2.805,
|
| 1168 |
+
"eval_wer": 0.999395039322444,
|
| 1169 |
+
"step": 10200
|
| 1170 |
+
},
|
| 1171 |
+
{
|
| 1172 |
+
"epoch": 27.393617021276597,
|
| 1173 |
+
"eval_cer": 0.17838395603622478,
|
| 1174 |
+
"eval_loss": 0.4000260531902313,
|
| 1175 |
+
"eval_runtime": 218.5093,
|
| 1176 |
+
"eval_samples_per_second": 22.695,
|
| 1177 |
+
"eval_steps_per_second": 2.837,
|
| 1178 |
+
"eval_wer": 0.9991933857632587,
|
| 1179 |
+
"step": 10300
|
| 1180 |
+
},
|
| 1181 |
+
{
|
| 1182 |
+
"epoch": 27.659574468085108,
|
| 1183 |
+
"eval_cer": 0.18251142284255825,
|
| 1184 |
+
"eval_loss": 0.37564513087272644,
|
| 1185 |
+
"eval_runtime": 218.7597,
|
| 1186 |
+
"eval_samples_per_second": 22.669,
|
| 1187 |
+
"eval_steps_per_second": 2.834,
|
| 1188 |
+
"eval_wer": 0.9975801572897761,
|
| 1189 |
+
"step": 10400
|
| 1190 |
+
},
|
| 1191 |
+
{
|
| 1192 |
+
"epoch": 27.925531914893618,
|
| 1193 |
+
"grad_norm": 5.426018238067627,
|
| 1194 |
+
"learning_rate": 0.00025183199999999996,
|
| 1195 |
+
"loss": 0.124,
|
| 1196 |
+
"step": 10500
|
| 1197 |
+
},
|
| 1198 |
+
{
|
| 1199 |
+
"epoch": 27.925531914893618,
|
| 1200 |
+
"eval_cer": 0.17786411409944225,
|
| 1201 |
+
"eval_loss": 0.39711853861808777,
|
| 1202 |
+
"eval_runtime": 219.1126,
|
| 1203 |
+
"eval_samples_per_second": 22.632,
|
| 1204 |
+
"eval_steps_per_second": 2.83,
|
| 1205 |
+
"eval_wer": 0.9985884250857028,
|
| 1206 |
+
"step": 10500
|
| 1207 |
+
},
|
| 1208 |
+
{
|
| 1209 |
+
"epoch": 28.19148936170213,
|
| 1210 |
+
"eval_cer": 0.178911615144989,
|
| 1211 |
+
"eval_loss": 0.42398887872695923,
|
| 1212 |
+
"eval_runtime": 216.8565,
|
| 1213 |
+
"eval_samples_per_second": 22.868,
|
| 1214 |
+
"eval_steps_per_second": 2.859,
|
| 1215 |
+
"eval_wer": 0.9995966928816293,
|
| 1216 |
+
"step": 10600
|
| 1217 |
+
},
|
| 1218 |
+
{
|
| 1219 |
+
"epoch": 28.45744680851064,
|
| 1220 |
+
"eval_cer": 0.17917349040637567,
|
| 1221 |
+
"eval_loss": 0.37179791927337646,
|
| 1222 |
+
"eval_runtime": 219.2208,
|
| 1223 |
+
"eval_samples_per_second": 22.621,
|
| 1224 |
+
"eval_steps_per_second": 2.828,
|
| 1225 |
+
"eval_wer": 0.9979834644081468,
|
| 1226 |
+
"step": 10700
|
| 1227 |
+
},
|
| 1228 |
+
{
|
| 1229 |
+
"epoch": 28.72340425531915,
|
| 1230 |
+
"eval_cer": 0.17998256770648083,
|
| 1231 |
+
"eval_loss": 0.41135987639427185,
|
| 1232 |
+
"eval_runtime": 220.5483,
|
| 1233 |
+
"eval_samples_per_second": 22.485,
|
| 1234 |
+
"eval_steps_per_second": 2.811,
|
| 1235 |
+
"eval_wer": 0.9985884250857028,
|
| 1236 |
+
"step": 10800
|
| 1237 |
+
},
|
| 1238 |
+
{
|
| 1239 |
+
"epoch": 28.98936170212766,
|
| 1240 |
+
"eval_cer": 0.18003337932436184,
|
| 1241 |
+
"eval_loss": 0.41742414236068726,
|
| 1242 |
+
"eval_runtime": 218.3591,
|
| 1243 |
+
"eval_samples_per_second": 22.71,
|
| 1244 |
+
"eval_steps_per_second": 2.839,
|
| 1245 |
+
"eval_wer": 0.9977818108489614,
|
| 1246 |
+
"step": 10900
|
| 1247 |
+
},
|
| 1248 |
+
{
|
| 1249 |
+
"epoch": 29.25531914893617,
|
| 1250 |
+
"grad_norm": 2.8614120483398438,
|
| 1251 |
+
"learning_rate": 0.000263832,
|
| 1252 |
+
"loss": 0.122,
|
| 1253 |
+
"step": 11000
|
| 1254 |
+
},
|
| 1255 |
+
{
|
| 1256 |
+
"epoch": 29.25531914893617,
|
| 1257 |
+
"eval_cer": 0.18530997041200406,
|
| 1258 |
+
"eval_loss": 0.4062168598175049,
|
| 1259 |
+
"eval_runtime": 220.2511,
|
| 1260 |
+
"eval_samples_per_second": 22.515,
|
| 1261 |
+
"eval_steps_per_second": 2.815,
|
| 1262 |
+
"eval_wer": 0.9987900786448881,
|
| 1263 |
+
"step": 11000
|
| 1264 |
+
},
|
| 1265 |
+
{
|
| 1266 |
+
"epoch": 29.52127659574468,
|
| 1267 |
+
"eval_cer": 0.1861112305401275,
|
| 1268 |
+
"eval_loss": 0.4202740788459778,
|
| 1269 |
+
"eval_runtime": 217.0728,
|
| 1270 |
+
"eval_samples_per_second": 22.845,
|
| 1271 |
+
"eval_steps_per_second": 2.856,
|
| 1272 |
+
"eval_wer": 0.9977818108489614,
|
| 1273 |
+
"step": 11100
|
| 1274 |
+
},
|
| 1275 |
+
{
|
| 1276 |
+
"epoch": 29.78723404255319,
|
| 1277 |
+
"eval_cer": 0.18606432750823734,
|
| 1278 |
+
"eval_loss": 0.43755000829696655,
|
| 1279 |
+
"eval_runtime": 219.2588,
|
| 1280 |
+
"eval_samples_per_second": 22.617,
|
| 1281 |
+
"eval_steps_per_second": 2.828,
|
| 1282 |
+
"eval_wer": 0.9985884250857028,
|
| 1283 |
+
"step": 11200
|
| 1284 |
+
},
|
| 1285 |
+
{
|
| 1286 |
+
"epoch": 30.0531914893617,
|
| 1287 |
+
"eval_cer": 0.18122940663756074,
|
| 1288 |
+
"eval_loss": 0.409412682056427,
|
| 1289 |
+
"eval_runtime": 223.5339,
|
| 1290 |
+
"eval_samples_per_second": 22.185,
|
| 1291 |
+
"eval_steps_per_second": 2.774,
|
| 1292 |
+
"eval_wer": 0.9991933857632587,
|
| 1293 |
+
"step": 11300
|
| 1294 |
+
},
|
| 1295 |
+
{
|
| 1296 |
+
"epoch": 30.319148936170212,
|
| 1297 |
+
"eval_cer": 0.18185868898208696,
|
| 1298 |
+
"eval_loss": 0.40995094180107117,
|
| 1299 |
+
"eval_runtime": 225.4254,
|
| 1300 |
+
"eval_samples_per_second": 21.998,
|
| 1301 |
+
"eval_steps_per_second": 2.75,
|
| 1302 |
+
"eval_wer": 0.9987900786448881,
|
| 1303 |
+
"step": 11400
|
| 1304 |
+
},
|
| 1305 |
+
{
|
| 1306 |
+
"epoch": 30.585106382978722,
|
| 1307 |
+
"grad_norm": 1.4252536296844482,
|
| 1308 |
+
"learning_rate": 0.000275832,
|
| 1309 |
+
"loss": 0.125,
|
| 1310 |
+
"step": 11500
|
| 1311 |
+
},
|
| 1312 |
+
{
|
| 1313 |
+
"epoch": 30.585106382978722,
|
| 1314 |
+
"eval_cer": 0.18692812501221434,
|
| 1315 |
+
"eval_loss": 0.39966416358947754,
|
| 1316 |
+
"eval_runtime": 219.2997,
|
| 1317 |
+
"eval_samples_per_second": 22.613,
|
| 1318 |
+
"eval_steps_per_second": 2.827,
|
| 1319 |
+
"eval_wer": 0.9981851179673321,
|
| 1320 |
+
"step": 11500
|
| 1321 |
+
},
|
| 1322 |
+
{
|
| 1323 |
+
"epoch": 30.851063829787233,
|
| 1324 |
+
"eval_cer": 0.18203848393766586,
|
| 1325 |
+
"eval_loss": 0.4437381625175476,
|
| 1326 |
+
"eval_runtime": 219.4853,
|
| 1327 |
+
"eval_samples_per_second": 22.594,
|
| 1328 |
+
"eval_steps_per_second": 2.825,
|
| 1329 |
+
"eval_wer": 0.9989917322040734,
|
| 1330 |
+
"step": 11600
|
| 1331 |
+
},
|
| 1332 |
+
{
|
| 1333 |
+
"epoch": 31.117021276595743,
|
| 1334 |
+
"eval_cer": 0.18580245224685066,
|
| 1335 |
+
"eval_loss": 0.44233861565589905,
|
| 1336 |
+
"eval_runtime": 216.9437,
|
| 1337 |
+
"eval_samples_per_second": 22.858,
|
| 1338 |
+
"eval_steps_per_second": 2.858,
|
| 1339 |
+
"eval_wer": 0.9989917322040734,
|
| 1340 |
+
"step": 11700
|
| 1341 |
+
},
|
| 1342 |
+
{
|
| 1343 |
+
"epoch": 31.382978723404257,
|
| 1344 |
+
"eval_cer": 0.18947652307824597,
|
| 1345 |
+
"eval_loss": 0.4216567575931549,
|
| 1346 |
+
"eval_runtime": 217.1103,
|
| 1347 |
+
"eval_samples_per_second": 22.841,
|
| 1348 |
+
"eval_steps_per_second": 2.856,
|
| 1349 |
+
"eval_wer": 0.9987900786448881,
|
| 1350 |
+
"step": 11800
|
| 1351 |
+
},
|
| 1352 |
+
{
|
| 1353 |
+
"epoch": 31.648936170212767,
|
| 1354 |
+
"eval_cer": 0.1966175096835218,
|
| 1355 |
+
"eval_loss": 0.4611862301826477,
|
| 1356 |
+
"eval_runtime": 226.5232,
|
| 1357 |
+
"eval_samples_per_second": 21.892,
|
| 1358 |
+
"eval_steps_per_second": 2.737,
|
| 1359 |
+
"eval_wer": 0.9991933857632587,
|
| 1360 |
+
"step": 11900
|
| 1361 |
+
},
|
| 1362 |
+
{
|
| 1363 |
+
"epoch": 31.914893617021278,
|
| 1364 |
+
"grad_norm": 5.578627586364746,
|
| 1365 |
+
"learning_rate": 0.00028783199999999996,
|
| 1366 |
+
"loss": 0.1294,
|
| 1367 |
+
"step": 12000
|
| 1368 |
+
},
|
| 1369 |
+
{
|
| 1370 |
+
"epoch": 31.914893617021278,
|
| 1371 |
+
"eval_cer": 0.18622067094787118,
|
| 1372 |
+
"eval_loss": 0.4385882318019867,
|
| 1373 |
+
"eval_runtime": 216.3818,
|
| 1374 |
+
"eval_samples_per_second": 22.918,
|
| 1375 |
+
"eval_steps_per_second": 2.865,
|
| 1376 |
+
"eval_wer": 0.9973785037305909,
|
| 1377 |
+
"step": 12000
|
| 1378 |
+
},
|
| 1379 |
+
{
|
| 1380 |
+
"epoch": 32.180851063829785,
|
| 1381 |
+
"eval_cer": 0.1892146478168593,
|
| 1382 |
+
"eval_loss": 0.42784062027931213,
|
| 1383 |
+
"eval_runtime": 236.4807,
|
| 1384 |
+
"eval_samples_per_second": 20.97,
|
| 1385 |
+
"eval_steps_per_second": 2.622,
|
| 1386 |
+
"eval_wer": 0.9983867715265174,
|
| 1387 |
+
"step": 12100
|
| 1388 |
+
},
|
| 1389 |
+
{
|
| 1390 |
+
"epoch": 32.4468085106383,
|
| 1391 |
+
"eval_cer": 0.1856109315332992,
|
| 1392 |
+
"eval_loss": 0.4187084436416626,
|
| 1393 |
+
"eval_runtime": 222.7058,
|
| 1394 |
+
"eval_samples_per_second": 22.267,
|
| 1395 |
+
"eval_steps_per_second": 2.784,
|
| 1396 |
+
"eval_wer": 0.9983867715265174,
|
| 1397 |
+
"step": 12200
|
| 1398 |
+
},
|
| 1399 |
+
{
|
| 1400 |
+
"epoch": 32.712765957446805,
|
| 1401 |
+
"eval_cer": 0.18286710416772525,
|
| 1402 |
+
"eval_loss": 0.4047478437423706,
|
| 1403 |
+
"eval_runtime": 222.5261,
|
| 1404 |
+
"eval_samples_per_second": 22.285,
|
| 1405 |
+
"eval_steps_per_second": 2.786,
|
| 1406 |
+
"eval_wer": 0.9985884250857028,
|
| 1407 |
+
"step": 12300
|
| 1408 |
+
},
|
| 1409 |
+
{
|
| 1410 |
+
"epoch": 32.97872340425532,
|
| 1411 |
+
"eval_cer": 0.18515753555836106,
|
| 1412 |
+
"eval_loss": 0.4230673015117645,
|
| 1413 |
+
"eval_runtime": 219.6428,
|
| 1414 |
+
"eval_samples_per_second": 22.578,
|
| 1415 |
+
"eval_steps_per_second": 2.823,
|
| 1416 |
+
"eval_wer": 0.9979834644081468,
|
| 1417 |
+
"step": 12400
|
| 1418 |
+
},
|
| 1419 |
+
{
|
| 1420 |
+
"epoch": 33.244680851063826,
|
| 1421 |
+
"grad_norm": 3.530928134918213,
|
| 1422 |
+
"learning_rate": 0.000299832,
|
| 1423 |
+
"loss": 0.1275,
|
| 1424 |
+
"step": 12500
|
| 1425 |
+
},
|
| 1426 |
+
{
|
| 1427 |
+
"epoch": 33.244680851063826,
|
| 1428 |
+
"eval_cer": 0.1842781037104207,
|
| 1429 |
+
"eval_loss": 0.41239944100379944,
|
| 1430 |
+
"eval_runtime": 212.9629,
|
| 1431 |
+
"eval_samples_per_second": 23.286,
|
| 1432 |
+
"eval_steps_per_second": 2.911,
|
| 1433 |
+
"eval_wer": 0.999395039322444,
|
| 1434 |
+
"step": 12500
|
| 1435 |
+
},
|
| 1436 |
+
{
|
| 1437 |
+
"epoch": 33.51063829787234,
|
| 1438 |
+
"eval_cer": 0.18698284521608619,
|
| 1439 |
+
"eval_loss": 0.41910141706466675,
|
| 1440 |
+
"eval_runtime": 209.3165,
|
| 1441 |
+
"eval_samples_per_second": 23.691,
|
| 1442 |
+
"eval_steps_per_second": 2.962,
|
| 1443 |
+
"eval_wer": 0.999395039322444,
|
| 1444 |
+
"step": 12600
|
| 1445 |
+
},
|
| 1446 |
+
{
|
| 1447 |
+
"epoch": 33.776595744680854,
|
| 1448 |
+
"eval_cer": 0.19266983783276723,
|
| 1449 |
+
"eval_loss": 0.4845859110355377,
|
| 1450 |
+
"eval_runtime": 209.6184,
|
| 1451 |
+
"eval_samples_per_second": 23.657,
|
| 1452 |
+
"eval_steps_per_second": 2.958,
|
| 1453 |
+
"eval_wer": 0.9979834644081468,
|
| 1454 |
+
"step": 12700
|
| 1455 |
+
},
|
| 1456 |
+
{
|
| 1457 |
+
"epoch": 34.04255319148936,
|
| 1458 |
+
"eval_cer": 0.18445399008000876,
|
| 1459 |
+
"eval_loss": 0.421185702085495,
|
| 1460 |
+
"eval_runtime": 209.5402,
|
| 1461 |
+
"eval_samples_per_second": 23.666,
|
| 1462 |
+
"eval_steps_per_second": 2.959,
|
| 1463 |
+
"eval_wer": 0.9983867715265174,
|
| 1464 |
+
"step": 12800
|
| 1465 |
+
},
|
| 1466 |
+
{
|
| 1467 |
+
"epoch": 34.308510638297875,
|
| 1468 |
+
"eval_cer": 0.183660547123867,
|
| 1469 |
+
"eval_loss": 0.43256160616874695,
|
| 1470 |
+
"eval_runtime": 219.5902,
|
| 1471 |
+
"eval_samples_per_second": 22.583,
|
| 1472 |
+
"eval_steps_per_second": 2.823,
|
| 1473 |
+
"eval_wer": 0.9983867715265174,
|
| 1474 |
+
"step": 12900
|
| 1475 |
+
},
|
| 1476 |
+
{
|
| 1477 |
+
"epoch": 34.57446808510638,
|
| 1478 |
+
"grad_norm": 1.3231768608093262,
|
| 1479 |
+
"learning_rate": 0.00029548990895071305,
|
| 1480 |
+
"loss": 0.134,
|
| 1481 |
+
"step": 13000
|
| 1482 |
+
},
|
| 1483 |
+
{
|
| 1484 |
+
"epoch": 34.57446808510638,
|
| 1485 |
+
"eval_cer": 0.18797171747177024,
|
| 1486 |
+
"eval_loss": 0.4103536009788513,
|
| 1487 |
+
"eval_runtime": 213.3906,
|
| 1488 |
+
"eval_samples_per_second": 23.239,
|
| 1489 |
+
"eval_steps_per_second": 2.905,
|
| 1490 |
+
"eval_wer": 0.9991933857632587,
|
| 1491 |
+
"step": 13000
|
| 1492 |
+
},
|
| 1493 |
+
{
|
| 1494 |
+
"epoch": 34.840425531914896,
|
| 1495 |
+
"eval_cer": 0.18773329372632863,
|
| 1496 |
+
"eval_loss": 0.3964749276638031,
|
| 1497 |
+
"eval_runtime": 219.4036,
|
| 1498 |
+
"eval_samples_per_second": 22.602,
|
| 1499 |
+
"eval_steps_per_second": 2.826,
|
| 1500 |
+
"eval_wer": 0.9977818108489614,
|
| 1501 |
+
"step": 13100
|
| 1502 |
+
},
|
| 1503 |
+
{
|
| 1504 |
+
"epoch": 35.1063829787234,
|
| 1505 |
+
"eval_cer": 0.18439145270415522,
|
| 1506 |
+
"eval_loss": 0.4146926701068878,
|
| 1507 |
+
"eval_runtime": 211.7608,
|
| 1508 |
+
"eval_samples_per_second": 23.418,
|
| 1509 |
+
"eval_steps_per_second": 2.928,
|
| 1510 |
+
"eval_wer": 0.999395039322444,
|
| 1511 |
+
"step": 13200
|
| 1512 |
+
},
|
| 1513 |
+
{
|
| 1514 |
+
"epoch": 35.37234042553192,
|
| 1515 |
+
"eval_cer": 0.18589625831063097,
|
| 1516 |
+
"eval_loss": 0.4251463711261749,
|
| 1517 |
+
"eval_runtime": 208.6917,
|
| 1518 |
+
"eval_samples_per_second": 23.762,
|
| 1519 |
+
"eval_steps_per_second": 2.971,
|
| 1520 |
+
"eval_wer": 0.9977818108489614,
|
| 1521 |
+
"step": 13300
|
| 1522 |
+
},
|
| 1523 |
+
{
|
| 1524 |
+
"epoch": 35.638297872340424,
|
| 1525 |
+
"eval_cer": 0.19017225138461658,
|
| 1526 |
+
"eval_loss": 0.4457683861255646,
|
| 1527 |
+
"eval_runtime": 211.8043,
|
| 1528 |
+
"eval_samples_per_second": 23.413,
|
| 1529 |
+
"eval_steps_per_second": 2.927,
|
| 1530 |
+
"eval_wer": 0.999395039322444,
|
| 1531 |
+
"step": 13400
|
| 1532 |
+
},
|
| 1533 |
+
{
|
| 1534 |
+
"epoch": 35.90425531914894,
|
| 1535 |
+
"grad_norm": 3.084670066833496,
|
| 1536 |
+
"learning_rate": 0.00028198284613067677,
|
| 1537 |
+
"loss": 0.1293,
|
| 1538 |
+
"step": 13500
|
| 1539 |
+
},
|
| 1540 |
+
{
|
| 1541 |
+
"epoch": 35.90425531914894,
|
| 1542 |
+
"eval_cer": 0.19420591212716976,
|
| 1543 |
+
"eval_loss": 0.43544086813926697,
|
| 1544 |
+
"eval_runtime": 209.3943,
|
| 1545 |
+
"eval_samples_per_second": 23.683,
|
| 1546 |
+
"eval_steps_per_second": 2.961,
|
| 1547 |
+
"eval_wer": 0.9991933857632587,
|
| 1548 |
+
"step": 13500
|
| 1549 |
+
},
|
| 1550 |
+
{
|
| 1551 |
+
"epoch": 36.170212765957444,
|
| 1552 |
+
"eval_cer": 0.1863222941836332,
|
| 1553 |
+
"eval_loss": 0.4197835326194763,
|
| 1554 |
+
"eval_runtime": 210.9951,
|
| 1555 |
+
"eval_samples_per_second": 23.503,
|
| 1556 |
+
"eval_steps_per_second": 2.938,
|
| 1557 |
+
"eval_wer": 0.9995966928816293,
|
| 1558 |
+
"step": 13600
|
| 1559 |
+
},
|
| 1560 |
+
{
|
| 1561 |
+
"epoch": 36.43617021276596,
|
| 1562 |
+
"eval_cer": 0.18905830437722546,
|
| 1563 |
+
"eval_loss": 0.4278818666934967,
|
| 1564 |
+
"eval_runtime": 214.736,
|
| 1565 |
+
"eval_samples_per_second": 23.093,
|
| 1566 |
+
"eval_steps_per_second": 2.887,
|
| 1567 |
+
"eval_wer": 0.9985884250857028,
|
| 1568 |
+
"step": 13700
|
| 1569 |
+
},
|
| 1570 |
+
{
|
| 1571 |
+
"epoch": 36.702127659574465,
|
| 1572 |
+
"eval_cer": 0.18459079058968836,
|
| 1573 |
+
"eval_loss": 0.4114536941051483,
|
| 1574 |
+
"eval_runtime": 212.7503,
|
| 1575 |
+
"eval_samples_per_second": 23.309,
|
| 1576 |
+
"eval_steps_per_second": 2.914,
|
| 1577 |
+
"eval_wer": 0.9975801572897761,
|
| 1578 |
+
"step": 13800
|
| 1579 |
+
},
|
| 1580 |
+
{
|
| 1581 |
+
"epoch": 36.96808510638298,
|
| 1582 |
+
"eval_cer": 0.18683822753442486,
|
| 1583 |
+
"eval_loss": 0.43593207001686096,
|
| 1584 |
+
"eval_runtime": 213.278,
|
| 1585 |
+
"eval_samples_per_second": 23.251,
|
| 1586 |
+
"eval_steps_per_second": 2.907,
|
| 1587 |
+
"eval_wer": 0.9985884250857028,
|
| 1588 |
+
"step": 13900
|
| 1589 |
+
},
|
| 1590 |
+
{
|
| 1591 |
+
"epoch": 37.234042553191486,
|
| 1592 |
+
"grad_norm": 2.2776243686676025,
|
| 1593 |
+
"learning_rate": 0.00026031324776349936,
|
| 1594 |
+
"loss": 0.1193,
|
| 1595 |
+
"step": 14000
|
| 1596 |
+
},
|
| 1597 |
+
{
|
| 1598 |
+
"epoch": 37.234042553191486,
|
| 1599 |
+
"eval_cer": 0.19045757816194836,
|
| 1600 |
+
"eval_loss": 0.43160808086395264,
|
| 1601 |
+
"eval_runtime": 209.1934,
|
| 1602 |
+
"eval_samples_per_second": 23.705,
|
| 1603 |
+
"eval_steps_per_second": 2.964,
|
| 1604 |
+
"eval_wer": 0.9991933857632587,
|
| 1605 |
+
"step": 14000
|
| 1606 |
+
},
|
| 1607 |
+
{
|
| 1608 |
+
"epoch": 37.5,
|
| 1609 |
+
"eval_cer": 0.18992991905318413,
|
| 1610 |
+
"eval_loss": 0.4389212727546692,
|
| 1611 |
+
"eval_runtime": 215.9667,
|
| 1612 |
+
"eval_samples_per_second": 22.962,
|
| 1613 |
+
"eval_steps_per_second": 2.871,
|
| 1614 |
+
"eval_wer": 0.999395039322444,
|
| 1615 |
+
"step": 14100
|
| 1616 |
+
},
|
| 1617 |
+
{
|
| 1618 |
+
"epoch": 37.765957446808514,
|
| 1619 |
+
"eval_cer": 0.1825231486005308,
|
| 1620 |
+
"eval_loss": 0.42151185870170593,
|
| 1621 |
+
"eval_runtime": 208.7788,
|
| 1622 |
+
"eval_samples_per_second": 23.752,
|
| 1623 |
+
"eval_steps_per_second": 2.97,
|
| 1624 |
+
"eval_wer": 0.9991933857632587,
|
| 1625 |
+
"step": 14200
|
| 1626 |
+
},
|
| 1627 |
+
{
|
| 1628 |
+
"epoch": 38.03191489361702,
|
| 1629 |
+
"eval_cer": 0.18895668114146344,
|
| 1630 |
+
"eval_loss": 0.47925981879234314,
|
| 1631 |
+
"eval_runtime": 211.6717,
|
| 1632 |
+
"eval_samples_per_second": 23.428,
|
| 1633 |
+
"eval_steps_per_second": 2.929,
|
| 1634 |
+
"eval_wer": 0.9991933857632587,
|
| 1635 |
+
"step": 14300
|
| 1636 |
+
},
|
| 1637 |
+
{
|
| 1638 |
+
"epoch": 38.297872340425535,
|
| 1639 |
+
"eval_cer": 0.1853920507178118,
|
| 1640 |
+
"eval_loss": 0.4381818473339081,
|
| 1641 |
+
"eval_runtime": 209.7616,
|
| 1642 |
+
"eval_samples_per_second": 23.641,
|
| 1643 |
+
"eval_steps_per_second": 2.956,
|
| 1644 |
+
"eval_wer": 0.9983867715265174,
|
| 1645 |
+
"step": 14400
|
| 1646 |
+
},
|
| 1647 |
+
{
|
| 1648 |
+
"epoch": 38.56382978723404,
|
| 1649 |
+
"grad_norm": 1.5227301120758057,
|
| 1650 |
+
"learning_rate": 0.00023182127942485896,
|
| 1651 |
+
"loss": 0.1132,
|
| 1652 |
+
"step": 14500
|
| 1653 |
+
},
|
| 1654 |
+
{
|
| 1655 |
+
"epoch": 38.56382978723404,
|
| 1656 |
+
"eval_cer": 0.182441068294723,
|
| 1657 |
+
"eval_loss": 0.40106824040412903,
|
| 1658 |
+
"eval_runtime": 209.0948,
|
| 1659 |
+
"eval_samples_per_second": 23.717,
|
| 1660 |
+
"eval_steps_per_second": 2.965,
|
| 1661 |
+
"eval_wer": 0.9975801572897761,
|
| 1662 |
+
"step": 14500
|
| 1663 |
+
},
|
| 1664 |
+
{
|
| 1665 |
+
"epoch": 38.829787234042556,
|
| 1666 |
+
"eval_cer": 0.1797675954769843,
|
| 1667 |
+
"eval_loss": 0.4283459186553955,
|
| 1668 |
+
"eval_runtime": 209.0149,
|
| 1669 |
+
"eval_samples_per_second": 23.726,
|
| 1670 |
+
"eval_steps_per_second": 2.966,
|
| 1671 |
+
"eval_wer": 0.994958661020367,
|
| 1672 |
+
"step": 14600
|
| 1673 |
+
},
|
| 1674 |
+
{
|
| 1675 |
+
"epoch": 39.09574468085106,
|
| 1676 |
+
"eval_cer": 0.18034997478962037,
|
| 1677 |
+
"eval_loss": 0.4303878843784332,
|
| 1678 |
+
"eval_runtime": 211.9384,
|
| 1679 |
+
"eval_samples_per_second": 23.398,
|
| 1680 |
+
"eval_steps_per_second": 2.925,
|
| 1681 |
+
"eval_wer": 0.9979834644081468,
|
| 1682 |
+
"step": 14700
|
| 1683 |
+
},
|
| 1684 |
+
{
|
| 1685 |
+
"epoch": 39.361702127659576,
|
| 1686 |
+
"eval_cer": 0.18114732633175296,
|
| 1687 |
+
"eval_loss": 0.40491461753845215,
|
| 1688 |
+
"eval_runtime": 210.1586,
|
| 1689 |
+
"eval_samples_per_second": 23.596,
|
| 1690 |
+
"eval_steps_per_second": 2.95,
|
| 1691 |
+
"eval_wer": 0.9983867715265174,
|
| 1692 |
+
"step": 14800
|
| 1693 |
+
},
|
| 1694 |
+
{
|
| 1695 |
+
"epoch": 39.62765957446808,
|
| 1696 |
+
"eval_cer": 0.1785129393739227,
|
| 1697 |
+
"eval_loss": 0.41461843252182007,
|
| 1698 |
+
"eval_runtime": 209.4294,
|
| 1699 |
+
"eval_samples_per_second": 23.679,
|
| 1700 |
+
"eval_steps_per_second": 2.96,
|
| 1701 |
+
"eval_wer": 0.9987900786448881,
|
| 1702 |
+
"step": 14900
|
| 1703 |
+
},
|
| 1704 |
+
{
|
| 1705 |
+
"epoch": 39.8936170212766,
|
| 1706 |
+
"grad_norm": 3.955990791320801,
|
| 1707 |
+
"learning_rate": 0.00019826903907816743,
|
| 1708 |
+
"loss": 0.0949,
|
| 1709 |
+
"step": 15000
|
| 1710 |
+
},
|
| 1711 |
+
{
|
| 1712 |
+
"epoch": 39.8936170212766,
|
| 1713 |
+
"eval_cer": 0.18166325968254465,
|
| 1714 |
+
"eval_loss": 0.44991689920425415,
|
| 1715 |
+
"eval_runtime": 208.9423,
|
| 1716 |
+
"eval_samples_per_second": 23.734,
|
| 1717 |
+
"eval_steps_per_second": 2.967,
|
| 1718 |
+
"eval_wer": 0.9997983464408147,
|
| 1719 |
+
"step": 15000
|
| 1720 |
+
},
|
| 1721 |
+
{
|
| 1722 |
+
"epoch": 40.159574468085104,
|
| 1723 |
+
"eval_cer": 0.17907968434259539,
|
| 1724 |
+
"eval_loss": 0.420516699552536,
|
| 1725 |
+
"eval_runtime": 210.185,
|
| 1726 |
+
"eval_samples_per_second": 23.593,
|
| 1727 |
+
"eval_steps_per_second": 2.95,
|
| 1728 |
+
"eval_wer": 0.9977818108489614,
|
| 1729 |
+
"step": 15100
|
| 1730 |
+
},
|
| 1731 |
+
{
|
| 1732 |
+
"epoch": 40.42553191489362,
|
| 1733 |
+
"eval_cer": 0.1807017475287965,
|
| 1734 |
+
"eval_loss": 0.4418542683124542,
|
| 1735 |
+
"eval_runtime": 210.2936,
|
| 1736 |
+
"eval_samples_per_second": 23.581,
|
| 1737 |
+
"eval_steps_per_second": 2.948,
|
| 1738 |
+
"eval_wer": 0.9985884250857028,
|
| 1739 |
+
"step": 15200
|
| 1740 |
+
},
|
| 1741 |
+
{
|
| 1742 |
+
"epoch": 40.691489361702125,
|
| 1743 |
+
"eval_cer": 0.18009591670021535,
|
| 1744 |
+
"eval_loss": 0.4282812476158142,
|
| 1745 |
+
"eval_runtime": 209.7755,
|
| 1746 |
+
"eval_samples_per_second": 23.64,
|
| 1747 |
+
"eval_steps_per_second": 2.956,
|
| 1748 |
+
"eval_wer": 0.9981851179673321,
|
| 1749 |
+
"step": 15300
|
| 1750 |
+
},
|
| 1751 |
+
{
|
| 1752 |
+
"epoch": 40.95744680851064,
|
| 1753 |
+
"eval_cer": 0.17690651053168496,
|
| 1754 |
+
"eval_loss": 0.4327284097671509,
|
| 1755 |
+
"eval_runtime": 210.8473,
|
| 1756 |
+
"eval_samples_per_second": 23.519,
|
| 1757 |
+
"eval_steps_per_second": 2.941,
|
| 1758 |
+
"eval_wer": 0.9995966928816293,
|
| 1759 |
+
"step": 15400
|
| 1760 |
+
},
|
| 1761 |
+
{
|
| 1762 |
+
"epoch": 41.223404255319146,
|
| 1763 |
+
"grad_norm": 1.6714894771575928,
|
| 1764 |
+
"learning_rate": 0.0001617315793726526,
|
| 1765 |
+
"loss": 0.0876,
|
| 1766 |
+
"step": 15500
|
| 1767 |
+
},
|
| 1768 |
+
{
|
| 1769 |
+
"epoch": 41.223404255319146,
|
| 1770 |
+
"eval_cer": 0.17806736057096625,
|
| 1771 |
+
"eval_loss": 0.44882500171661377,
|
| 1772 |
+
"eval_runtime": 210.3046,
|
| 1773 |
+
"eval_samples_per_second": 23.58,
|
| 1774 |
+
"eval_steps_per_second": 2.948,
|
| 1775 |
+
"eval_wer": 0.9995966928816293,
|
| 1776 |
+
"step": 15500
|
| 1777 |
+
},
|
| 1778 |
+
{
|
| 1779 |
+
"epoch": 41.48936170212766,
|
| 1780 |
+
"eval_cer": 0.17360766395541086,
|
| 1781 |
+
"eval_loss": 0.41944339871406555,
|
| 1782 |
+
"eval_runtime": 210.2221,
|
| 1783 |
+
"eval_samples_per_second": 23.589,
|
| 1784 |
+
"eval_steps_per_second": 2.949,
|
| 1785 |
+
"eval_wer": 0.9989917322040734,
|
| 1786 |
+
"step": 15600
|
| 1787 |
+
},
|
| 1788 |
+
{
|
| 1789 |
+
"epoch": 41.755319148936174,
|
| 1790 |
+
"eval_cer": 0.17544079078511768,
|
| 1791 |
+
"eval_loss": 0.4319663345813751,
|
| 1792 |
+
"eval_runtime": 209.0871,
|
| 1793 |
+
"eval_samples_per_second": 23.717,
|
| 1794 |
+
"eval_steps_per_second": 2.965,
|
| 1795 |
+
"eval_wer": 0.9991933857632587,
|
| 1796 |
+
"step": 15700
|
| 1797 |
+
},
|
| 1798 |
+
{
|
| 1799 |
+
"epoch": 42.02127659574468,
|
| 1800 |
+
"eval_cer": 0.1729080270630494,
|
| 1801 |
+
"eval_loss": 0.4347435235977173,
|
| 1802 |
+
"eval_runtime": 209.3479,
|
| 1803 |
+
"eval_samples_per_second": 23.688,
|
| 1804 |
+
"eval_steps_per_second": 2.962,
|
| 1805 |
+
"eval_wer": 0.9989917322040734,
|
| 1806 |
+
"step": 15800
|
| 1807 |
+
},
|
| 1808 |
+
{
|
| 1809 |
+
"epoch": 42.287234042553195,
|
| 1810 |
+
"eval_cer": 0.17443628418547022,
|
| 1811 |
+
"eval_loss": 0.48187553882598877,
|
| 1812 |
+
"eval_runtime": 208.6066,
|
| 1813 |
+
"eval_samples_per_second": 23.772,
|
| 1814 |
+
"eval_steps_per_second": 2.972,
|
| 1815 |
+
"eval_wer": 0.999395039322444,
|
| 1816 |
+
"step": 15900
|
| 1817 |
+
},
|
| 1818 |
+
{
|
| 1819 |
+
"epoch": 42.5531914893617,
|
| 1820 |
+
"grad_norm": 0.7495588660240173,
|
| 1821 |
+
"learning_rate": 0.00012446857512509515,
|
| 1822 |
+
"loss": 0.0725,
|
| 1823 |
+
"step": 16000
|
| 1824 |
+
},
|
| 1825 |
+
{
|
| 1826 |
+
"epoch": 42.5531914893617,
|
| 1827 |
+
"eval_cer": 0.17520236703967607,
|
| 1828 |
+
"eval_loss": 0.44911447167396545,
|
| 1829 |
+
"eval_runtime": 208.5357,
|
| 1830 |
+
"eval_samples_per_second": 23.78,
|
| 1831 |
+
"eval_steps_per_second": 2.973,
|
| 1832 |
+
"eval_wer": 0.9989917322040734,
|
| 1833 |
+
"step": 16000
|
| 1834 |
+
},
|
| 1835 |
+
{
|
| 1836 |
+
"epoch": 42.819148936170215,
|
| 1837 |
+
"eval_cer": 0.17413923165016593,
|
| 1838 |
+
"eval_loss": 0.45369845628738403,
|
| 1839 |
+
"eval_runtime": 210.1991,
|
| 1840 |
+
"eval_samples_per_second": 23.592,
|
| 1841 |
+
"eval_steps_per_second": 2.95,
|
| 1842 |
+
"eval_wer": 0.9985884250857028,
|
| 1843 |
+
"step": 16100
|
| 1844 |
+
},
|
| 1845 |
+
{
|
| 1846 |
+
"epoch": 43.08510638297872,
|
| 1847 |
+
"eval_cer": 0.17170809116385966,
|
| 1848 |
+
"eval_loss": 0.4587579071521759,
|
| 1849 |
+
"eval_runtime": 209.6677,
|
| 1850 |
+
"eval_samples_per_second": 23.652,
|
| 1851 |
+
"eval_steps_per_second": 2.957,
|
| 1852 |
+
"eval_wer": 0.9983867715265174,
|
| 1853 |
+
"step": 16200
|
| 1854 |
+
},
|
| 1855 |
+
{
|
| 1856 |
+
"epoch": 43.351063829787236,
|
| 1857 |
+
"eval_cer": 0.17147357600440888,
|
| 1858 |
+
"eval_loss": 0.4416778087615967,
|
| 1859 |
+
"eval_runtime": 209.0934,
|
| 1860 |
+
"eval_samples_per_second": 23.717,
|
| 1861 |
+
"eval_steps_per_second": 2.965,
|
| 1862 |
+
"eval_wer": 0.9981851179673321,
|
| 1863 |
+
"step": 16300
|
| 1864 |
+
},
|
| 1865 |
+
{
|
| 1866 |
+
"epoch": 43.61702127659574,
|
| 1867 |
+
"eval_cer": 0.17290411847705855,
|
| 1868 |
+
"eval_loss": 0.45538103580474854,
|
| 1869 |
+
"eval_runtime": 210.428,
|
| 1870 |
+
"eval_samples_per_second": 23.566,
|
| 1871 |
+
"eval_steps_per_second": 2.946,
|
| 1872 |
+
"eval_wer": 0.9983867715265174,
|
| 1873 |
+
"step": 16400
|
| 1874 |
+
},
|
| 1875 |
+
{
|
| 1876 |
+
"epoch": 43.88297872340426,
|
| 1877 |
+
"grad_norm": 1.7286139726638794,
|
| 1878 |
+
"learning_rate": 8.878457276467852e-05,
|
| 1879 |
+
"loss": 0.0615,
|
| 1880 |
+
"step": 16500
|
| 1881 |
+
},
|
| 1882 |
+
{
|
| 1883 |
+
"epoch": 43.88297872340426,
|
| 1884 |
+
"eval_cer": 0.17189570329142026,
|
| 1885 |
+
"eval_loss": 0.44644975662231445,
|
| 1886 |
+
"eval_runtime": 208.6857,
|
| 1887 |
+
"eval_samples_per_second": 23.763,
|
| 1888 |
+
"eval_steps_per_second": 2.971,
|
| 1889 |
+
"eval_wer": 0.9995966928816293,
|
| 1890 |
+
"step": 16500
|
| 1891 |
+
},
|
| 1892 |
+
{
|
| 1893 |
+
"epoch": 44.148936170212764,
|
| 1894 |
+
"eval_cer": 0.17184489167353925,
|
| 1895 |
+
"eval_loss": 0.47256070375442505,
|
| 1896 |
+
"eval_runtime": 214.9431,
|
| 1897 |
+
"eval_samples_per_second": 23.071,
|
| 1898 |
+
"eval_steps_per_second": 2.884,
|
| 1899 |
+
"eval_wer": 0.9981851179673321,
|
| 1900 |
+
"step": 16600
|
| 1901 |
+
},
|
| 1902 |
+
{
|
| 1903 |
+
"epoch": 44.41489361702128,
|
| 1904 |
+
"eval_cer": 0.1704495264748072,
|
| 1905 |
+
"eval_loss": 0.4455854296684265,
|
| 1906 |
+
"eval_runtime": 212.2795,
|
| 1907 |
+
"eval_samples_per_second": 23.361,
|
| 1908 |
+
"eval_steps_per_second": 2.921,
|
| 1909 |
+
"eval_wer": 0.9979834644081468,
|
| 1910 |
+
"step": 16700
|
| 1911 |
+
},
|
| 1912 |
+
{
|
| 1913 |
+
"epoch": 44.680851063829785,
|
| 1914 |
+
"eval_cer": 0.16927304209156255,
|
| 1915 |
+
"eval_loss": 0.42471006512641907,
|
| 1916 |
+
"eval_runtime": 208.6276,
|
| 1917 |
+
"eval_samples_per_second": 23.77,
|
| 1918 |
+
"eval_steps_per_second": 2.972,
|
| 1919 |
+
"eval_wer": 0.9979834644081468,
|
| 1920 |
+
"step": 16800
|
| 1921 |
+
},
|
| 1922 |
+
{
|
| 1923 |
+
"epoch": 44.9468085106383,
|
| 1924 |
+
"eval_cer": 0.1683897016576313,
|
| 1925 |
+
"eval_loss": 0.4499261975288391,
|
| 1926 |
+
"eval_runtime": 208.8575,
|
| 1927 |
+
"eval_samples_per_second": 23.743,
|
| 1928 |
+
"eval_steps_per_second": 2.969,
|
| 1929 |
+
"eval_wer": 0.9985884250857028,
|
| 1930 |
+
"step": 16900
|
| 1931 |
+
},
|
| 1932 |
+
{
|
| 1933 |
+
"epoch": 45.212765957446805,
|
| 1934 |
+
"grad_norm": 1.9694031476974487,
|
| 1935 |
+
"learning_rate": 5.6886464673787096e-05,
|
| 1936 |
+
"loss": 0.0524,
|
| 1937 |
+
"step": 17000
|
| 1938 |
+
},
|
| 1939 |
+
{
|
| 1940 |
+
"epoch": 45.212765957446805,
|
| 1941 |
+
"eval_cer": 0.16670119250958582,
|
| 1942 |
+
"eval_loss": 0.46098417043685913,
|
| 1943 |
+
"eval_runtime": 209.0801,
|
| 1944 |
+
"eval_samples_per_second": 23.718,
|
| 1945 |
+
"eval_steps_per_second": 2.965,
|
| 1946 |
+
"eval_wer": 0.9987900786448881,
|
| 1947 |
+
"step": 17000
|
| 1948 |
+
},
|
| 1949 |
+
{
|
| 1950 |
+
"epoch": 45.47872340425532,
|
| 1951 |
+
"eval_cer": 0.16754153849761771,
|
| 1952 |
+
"eval_loss": 0.42516496777534485,
|
| 1953 |
+
"eval_runtime": 215.2938,
|
| 1954 |
+
"eval_samples_per_second": 23.034,
|
| 1955 |
+
"eval_steps_per_second": 2.88,
|
| 1956 |
+
"eval_wer": 0.9981851179673321,
|
| 1957 |
+
"step": 17100
|
| 1958 |
+
},
|
| 1959 |
+
{
|
| 1960 |
+
"epoch": 45.744680851063826,
|
| 1961 |
+
"eval_cer": 0.16701778797484435,
|
| 1962 |
+
"eval_loss": 0.41850271821022034,
|
| 1963 |
+
"eval_runtime": 208.7292,
|
| 1964 |
+
"eval_samples_per_second": 23.758,
|
| 1965 |
+
"eval_steps_per_second": 2.97,
|
| 1966 |
+
"eval_wer": 0.9979834644081468,
|
| 1967 |
+
"step": 17200
|
| 1968 |
+
},
|
| 1969 |
+
{
|
| 1970 |
+
"epoch": 46.01063829787234,
|
| 1971 |
+
"eval_cer": 0.16651748896801605,
|
| 1972 |
+
"eval_loss": 0.4377266764640808,
|
| 1973 |
+
"eval_runtime": 208.7268,
|
| 1974 |
+
"eval_samples_per_second": 23.758,
|
| 1975 |
+
"eval_steps_per_second": 2.97,
|
| 1976 |
+
"eval_wer": 0.9979834644081468,
|
| 1977 |
+
"step": 17300
|
| 1978 |
+
},
|
| 1979 |
+
{
|
| 1980 |
+
"epoch": 46.276595744680854,
|
| 1981 |
+
"eval_cer": 0.16658784351585126,
|
| 1982 |
+
"eval_loss": 0.4386991262435913,
|
| 1983 |
+
"eval_runtime": 208.9381,
|
| 1984 |
+
"eval_samples_per_second": 23.734,
|
| 1985 |
+
"eval_steps_per_second": 2.967,
|
| 1986 |
+
"eval_wer": 0.9991933857632587,
|
| 1987 |
+
"step": 17400
|
| 1988 |
+
},
|
| 1989 |
+
{
|
| 1990 |
+
"epoch": 46.54255319148936,
|
| 1991 |
+
"grad_norm": 0.5704839825630188,
|
| 1992 |
+
"learning_rate": 3.074700298507786e-05,
|
| 1993 |
+
"loss": 0.0466,
|
| 1994 |
+
"step": 17500
|
| 1995 |
+
},
|
| 1996 |
+
{
|
| 1997 |
+
"epoch": 46.54255319148936,
|
| 1998 |
+
"eval_cer": 0.16593510965537997,
|
| 1999 |
+
"eval_loss": 0.4388461112976074,
|
| 2000 |
+
"eval_runtime": 208.6064,
|
| 2001 |
+
"eval_samples_per_second": 23.772,
|
| 2002 |
+
"eval_steps_per_second": 2.972,
|
| 2003 |
+
"eval_wer": 0.9985884250857028,
|
| 2004 |
+
"step": 17500
|
| 2005 |
+
},
|
| 2006 |
+
{
|
| 2007 |
+
"epoch": 46.808510638297875,
|
| 2008 |
+
"eval_cer": 0.16539963337463406,
|
| 2009 |
+
"eval_loss": 0.4407631754875183,
|
| 2010 |
+
"eval_runtime": 209.5029,
|
| 2011 |
+
"eval_samples_per_second": 23.67,
|
| 2012 |
+
"eval_steps_per_second": 2.959,
|
| 2013 |
+
"eval_wer": 0.9985884250857028,
|
| 2014 |
+
"step": 17600
|
| 2015 |
+
},
|
| 2016 |
+
{
|
| 2017 |
+
"epoch": 47.07446808510638,
|
| 2018 |
+
"eval_cer": 0.16512603235527484,
|
| 2019 |
+
"eval_loss": 0.4277360141277313,
|
| 2020 |
+
"eval_runtime": 208.9388,
|
| 2021 |
+
"eval_samples_per_second": 23.734,
|
| 2022 |
+
"eval_steps_per_second": 2.967,
|
| 2023 |
+
"eval_wer": 0.9983867715265174,
|
| 2024 |
+
"step": 17700
|
| 2025 |
+
},
|
| 2026 |
+
{
|
| 2027 |
+
"epoch": 47.340425531914896,
|
| 2028 |
+
"eval_cer": 0.16503222629149453,
|
| 2029 |
+
"eval_loss": 0.4244243800640106,
|
| 2030 |
+
"eval_runtime": 208.9653,
|
| 2031 |
+
"eval_samples_per_second": 23.731,
|
| 2032 |
+
"eval_steps_per_second": 2.967,
|
| 2033 |
+
"eval_wer": 0.9985884250857028,
|
| 2034 |
+
"step": 17800
|
| 2035 |
+
},
|
| 2036 |
+
{
|
| 2037 |
+
"epoch": 47.6063829787234,
|
| 2038 |
+
"eval_cer": 0.16441857829093168,
|
| 2039 |
+
"eval_loss": 0.42959603667259216,
|
| 2040 |
+
"eval_runtime": 209.1668,
|
| 2041 |
+
"eval_samples_per_second": 23.708,
|
| 2042 |
+
"eval_steps_per_second": 2.964,
|
| 2043 |
+
"eval_wer": 0.9977818108489614,
|
| 2044 |
+
"step": 17900
|
| 2045 |
+
},
|
| 2046 |
+
{
|
| 2047 |
+
"epoch": 47.87234042553192,
|
| 2048 |
+
"grad_norm": 1.8080158233642578,
|
| 2049 |
+
"learning_rate": 1.1982793882402725e-05,
|
| 2050 |
+
"loss": 0.0393,
|
| 2051 |
+
"step": 18000
|
| 2052 |
+
},
|
| 2053 |
+
{
|
| 2054 |
+
"epoch": 47.87234042553192,
|
| 2055 |
+
"eval_cer": 0.1648250712339797,
|
| 2056 |
+
"eval_loss": 0.43411195278167725,
|
| 2057 |
+
"eval_runtime": 215.2833,
|
| 2058 |
+
"eval_samples_per_second": 23.035,
|
| 2059 |
+
"eval_steps_per_second": 2.88,
|
| 2060 |
+
"eval_wer": 0.9983867715265174,
|
| 2061 |
+
"step": 18000
|
| 2062 |
+
},
|
| 2063 |
+
{
|
| 2064 |
+
"epoch": 48.138297872340424,
|
| 2065 |
+
"eval_cer": 0.16460228183250145,
|
| 2066 |
+
"eval_loss": 0.43372273445129395,
|
| 2067 |
+
"eval_runtime": 259.5957,
|
| 2068 |
+
"eval_samples_per_second": 19.103,
|
| 2069 |
+
"eval_steps_per_second": 2.388,
|
| 2070 |
+
"eval_wer": 0.9981851179673321,
|
| 2071 |
+
"step": 18100
|
| 2072 |
+
},
|
| 2073 |
+
{
|
| 2074 |
+
"epoch": 48.40425531914894,
|
| 2075 |
+
"eval_cer": 0.16421142323341684,
|
| 2076 |
+
"eval_loss": 0.43311166763305664,
|
| 2077 |
+
"eval_runtime": 209.0523,
|
| 2078 |
+
"eval_samples_per_second": 23.721,
|
| 2079 |
+
"eval_steps_per_second": 2.966,
|
| 2080 |
+
"eval_wer": 0.9977818108489614,
|
| 2081 |
+
"step": 18200
|
| 2082 |
+
},
|
| 2083 |
+
{
|
| 2084 |
+
"epoch": 48.670212765957444,
|
| 2085 |
+
"eval_cer": 0.16413716009959078,
|
| 2086 |
+
"eval_loss": 0.4280834197998047,
|
| 2087 |
+
"eval_runtime": 210.4305,
|
| 2088 |
+
"eval_samples_per_second": 23.566,
|
| 2089 |
+
"eval_steps_per_second": 2.946,
|
| 2090 |
+
"eval_wer": 0.9979834644081468,
|
| 2091 |
+
"step": 18300
|
| 2092 |
+
},
|
| 2093 |
+
{
|
| 2094 |
+
"epoch": 48.93617021276596,
|
| 2095 |
+
"eval_cer": 0.1640824398957189,
|
| 2096 |
+
"eval_loss": 0.42681965231895447,
|
| 2097 |
+
"eval_runtime": 261.2993,
|
| 2098 |
+
"eval_samples_per_second": 18.978,
|
| 2099 |
+
"eval_steps_per_second": 2.373,
|
| 2100 |
+
"eval_wer": 0.9981851179673321,
|
| 2101 |
+
"step": 18400
|
| 2102 |
+
},
|
| 2103 |
+
{
|
| 2104 |
+
"epoch": 49.202127659574465,
|
| 2105 |
+
"grad_norm": 0.8930312395095825,
|
| 2106 |
+
"learning_rate": 1.7543179002012464e-06,
|
| 2107 |
+
"loss": 0.0373,
|
| 2108 |
+
"step": 18500
|
| 2109 |
+
},
|
| 2110 |
+
{
|
| 2111 |
+
"epoch": 49.202127659574465,
|
| 2112 |
+
"eval_cer": 0.16407071413774638,
|
| 2113 |
+
"eval_loss": 0.42752090096473694,
|
| 2114 |
+
"eval_runtime": 209.2156,
|
| 2115 |
+
"eval_samples_per_second": 23.703,
|
| 2116 |
+
"eval_steps_per_second": 2.963,
|
| 2117 |
+
"eval_wer": 0.9981851179673321,
|
| 2118 |
+
"step": 18500
|
| 2119 |
+
},
|
| 2120 |
+
{
|
| 2121 |
+
"epoch": 49.46808510638298,
|
| 2122 |
+
"eval_cer": 0.16406680555175554,
|
| 2123 |
+
"eval_loss": 0.4268626868724823,
|
| 2124 |
+
"eval_runtime": 212.2555,
|
| 2125 |
+
"eval_samples_per_second": 23.363,
|
| 2126 |
+
"eval_steps_per_second": 2.921,
|
| 2127 |
+
"eval_wer": 0.9981851179673321,
|
| 2128 |
+
"step": 18600
|
| 2129 |
+
},
|
| 2130 |
+
{
|
| 2131 |
+
"epoch": 49.734042553191486,
|
| 2132 |
+
"eval_cer": 0.16399254241792946,
|
| 2133 |
+
"eval_loss": 0.4268469214439392,
|
| 2134 |
+
"eval_runtime": 208.7487,
|
| 2135 |
+
"eval_samples_per_second": 23.756,
|
| 2136 |
+
"eval_steps_per_second": 2.97,
|
| 2137 |
+
"eval_wer": 0.9977818108489614,
|
| 2138 |
+
"step": 18700
|
| 2139 |
+
},
|
| 2140 |
+
{
|
| 2141 |
+
"epoch": 50.0,
|
| 2142 |
+
"eval_cer": 0.16415279444355416,
|
| 2143 |
+
"eval_loss": 0.42705675959587097,
|
| 2144 |
+
"eval_runtime": 208.3387,
|
| 2145 |
+
"eval_samples_per_second": 23.803,
|
| 2146 |
+
"eval_steps_per_second": 2.976,
|
| 2147 |
+
"eval_wer": 0.9977818108489614,
|
| 2148 |
+
"step": 18800
|
| 2149 |
+
},
|
| 2150 |
+
{
|
| 2151 |
+
"epoch": 50.0,
|
| 2152 |
+
"step": 18800,
|
| 2153 |
+
"total_flos": 2.6273093616415412e+19,
|
| 2154 |
+
"train_loss": 0.4882018888757584,
|
| 2155 |
+
"train_runtime": 107059.6152,
|
| 2156 |
+
"train_samples_per_second": 5.619,
|
| 2157 |
+
"train_steps_per_second": 0.176
|
| 2158 |
+
}
|
| 2159 |
+
],
|
| 2160 |
+
"logging_steps": 500,
|
| 2161 |
+
"max_steps": 18800,
|
| 2162 |
+
"num_input_tokens_seen": 0,
|
| 2163 |
+
"num_train_epochs": 50,
|
| 2164 |
+
"save_steps": 400,
|
| 2165 |
+
"stateful_callbacks": {
|
| 2166 |
+
"TrainerControl": {
|
| 2167 |
+
"args": {
|
| 2168 |
+
"should_epoch_stop": false,
|
| 2169 |
+
"should_evaluate": false,
|
| 2170 |
+
"should_log": false,
|
| 2171 |
+
"should_save": true,
|
| 2172 |
+
"should_training_stop": true
|
| 2173 |
+
},
|
| 2174 |
+
"attributes": {}
|
| 2175 |
+
}
|
| 2176 |
+
},
|
| 2177 |
+
"total_flos": 2.6273093616415412e+19,
|
| 2178 |
+
"train_batch_size": 16,
|
| 2179 |
+
"trial_name": null,
|
| 2180 |
+
"trial_params": null
|
| 2181 |
+
}
|