Update README.md
Browse files
README.md
CHANGED
@@ -28,9 +28,37 @@ model-index:
|
|
28 |
|
29 |
# Wav2Vec2-Large-XLSR-53-Chinese-zh-CN-aishell1
|
30 |
|
31 |
-
Fine-tuned [facebook/wav2vec2-large-xlsr-53](https://huggingface.co/facebook/wav2vec2-large-xlsr-53) on Chinese using the [AISHELL-1](https://github.com/kaldi-asr/kaldi/tree/master/egs/aishell).
|
32 |
When using this model, make sure that your speech input is sampled at 16kHz.
|
33 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
34 |
## Evaluation
|
35 |
|
36 |
```python
|
|
|
28 |
|
29 |
# Wav2Vec2-Large-XLSR-53-Chinese-zh-CN-aishell1
|
30 |
|
31 |
+
Fine-tuned [facebook/wav2vec2-large-xlsr-53](https://huggingface.co/facebook/wav2vec2-large-xlsr-53) on Chinese using the [AISHELL-1](https://github.com/kaldi-asr/kaldi/tree/master/egs/aishell) dataset.
|
32 |
When using this model, make sure that your speech input is sampled at 16kHz.
|
33 |
|
34 |
+
## Usage
|
35 |
+
|
36 |
+
The model can be used directly (without a language model) as follows:
|
37 |
+
|
38 |
+
```python
|
39 |
+
import torch
|
40 |
+
import librosa
|
41 |
+
from transformers import Wav2Vec2ForCTC, Wav2Vec2Processor
|
42 |
+
|
43 |
+
device = "cuda:0" if torch.cuda.is_available() else "cpu"
|
44 |
+
|
45 |
+
processor = Wav2Vec2Processor.from_pretrained(
|
46 |
+
'qinyue/wav2vec2-large-xlsr-53-chinese-zn-cn-aishell1')
|
47 |
+
model = Wav2Vec2ForCTC.from_pretrained(
|
48 |
+
'qinyue/wav2vec2-large-xlsr-53-chinese-zn-cn-aishell1').to(device)
|
49 |
+
|
50 |
+
filepath = 'test.wav'
|
51 |
+
audio, sr = librosa.load(filepath, sr=16000, mono=True)
|
52 |
+
inputs = processor(audio, sample_rate=16000, return_tensors="pt").to(device)
|
53 |
+
with torch.no_grad():
|
54 |
+
logits = model(inputs.input_values,
|
55 |
+
attention_mask=inputs.attention_mask).logits
|
56 |
+
predicted_ids = torch.argmax(logits, dim=-1)
|
57 |
+
pred_str = processor.decode(predicted_ids[0])
|
58 |
+
|
59 |
+
print(pred_str)
|
60 |
+
```
|
61 |
+
|
62 |
## Evaluation
|
63 |
|
64 |
```python
|