Update README.md
Browse files
README.md
CHANGED
|
@@ -38,15 +38,16 @@ $ pip install -r requirements.txt
|
|
| 38 |
```
|
| 39 |
|
| 40 |
## Download Audio Data
|
|
|
|
| 41 |
```
|
| 42 |
import datasets
|
| 43 |
cv4en = datasets.load_dataset(
|
| 44 |
-
"mozilla-foundation/common_voice_4_0",
|
| 45 |
-
"en",
|
| 46 |
-
cache_dir='path/to/cv4/download',
|
| 47 |
)
|
| 48 |
```
|
| 49 |
## Evaluation
|
|
|
|
|
|
|
| 50 |
```bash
|
| 51 |
$ python speechqe/score_speechqe.py \
|
| 52 |
--speechqe_model=h-j-han/SpeechQE-TowerInstruct-7B-en2de \
|
|
@@ -57,8 +58,6 @@ $ python speechqe/score_speechqe.py \
|
|
| 57 |
```
|
| 58 |
|
| 59 |
|
| 60 |
-
|
| 61 |
-
|
| 62 |
## Reference
|
| 63 |
Please find details in this paper :
|
| 64 |
```
|
|
|
|
| 38 |
```
|
| 39 |
|
| 40 |
## Download Audio Data
|
| 41 |
+
Download the audio data from Common Voice. Here, we use mozilla-foundation/common_voice_4_0.
|
| 42 |
```
|
| 43 |
import datasets
|
| 44 |
cv4en = datasets.load_dataset(
|
| 45 |
+
"mozilla-foundation/common_voice_4_0", "en", cache_dir='path/to/cv4/download',
|
|
|
|
|
|
|
| 46 |
)
|
| 47 |
```
|
| 48 |
## Evaluation
|
| 49 |
+
We provide SpeechQE benchmark: [h-j-han/SpeechQE-CoVoST2](https://huggingface.co/datasets/h-j-han/SpeechQE-CoVoST2).
|
| 50 |
+
BASE_AUDIO_PATH is the path of downloaded Common Voice dataset.
|
| 51 |
```bash
|
| 52 |
$ python speechqe/score_speechqe.py \
|
| 53 |
--speechqe_model=h-j-han/SpeechQE-TowerInstruct-7B-en2de \
|
|
|
|
| 58 |
```
|
| 59 |
|
| 60 |
|
|
|
|
|
|
|
| 61 |
## Reference
|
| 62 |
Please find details in this paper :
|
| 63 |
```
|