Vu Minh Chien
commited on
Commit
•
6182f4f
1
Parent(s):
05014f2
update
Browse files
README.md
CHANGED
|
@@ -117,9 +117,10 @@ def evaluate(batch):
|
|
| 117 |
result = test_dataset.map(evaluate, batched=True, batch_size=8)
|
| 118 |
print("WER: {:2f}".format(100 * wer.compute(predictions=result["pred_strings"], references=result["sentence"])))
|
| 119 |
```
|
| 120 |
-
|
| 121 |
**WER:** 30.84%
|
| 122 |
**CER:** 17.85
|
|
|
|
| 123 |
## Training
|
| 124 |
The Common Voice `train`, `validation` datasets and Japanese speech corpus `basic5000` datasets were used for training.
|
| 125 |
|
|
|
|
| 117 |
result = test_dataset.map(evaluate, batched=True, batch_size=8)
|
| 118 |
print("WER: {:2f}".format(100 * wer.compute(predictions=result["pred_strings"], references=result["sentence"])))
|
| 119 |
```
|
| 120 |
+
## Test Result
|
| 121 |
**WER:** 30.84%
|
| 122 |
**CER:** 17.85
|
| 123 |
+
|
| 124 |
## Training
|
| 125 |
The Common Voice `train`, `validation` datasets and Japanese speech corpus `basic5000` datasets were used for training.
|
| 126 |
|