Spaces:
Runtime error
Runtime error
Update app.py
Browse files
app.py
CHANGED
|
@@ -24,8 +24,6 @@ speaker_embeddings = torch.tensor(embeddings_dataset[7306]["xvector"]).unsqueeze
|
|
| 24 |
|
| 25 |
|
| 26 |
def translate(audio):
|
| 27 |
-
print(audio)
|
| 28 |
-
audio = librosa.resample(audio, orig_sr=22050, target_sr=16000)
|
| 29 |
outputs = asr_pipe(audio, max_new_tokens=256, generate_kwargs={"language": "nl","task": "transcribe"})
|
| 30 |
return outputs["text"]
|
| 31 |
|
|
@@ -37,7 +35,10 @@ def synthesise(text):
|
|
| 37 |
|
| 38 |
|
| 39 |
def speech_to_speech_translation(audio):
|
| 40 |
-
|
|
|
|
|
|
|
|
|
|
| 41 |
synthesised_speech = synthesise(translated_text)
|
| 42 |
synthesised_speech = (synthesised_speech.numpy() * 32767).astype(np.int16)
|
| 43 |
return 16000, synthesised_speech
|
|
|
|
| 24 |
|
| 25 |
|
| 26 |
def translate(audio):
|
|
|
|
|
|
|
| 27 |
outputs = asr_pipe(audio, max_new_tokens=256, generate_kwargs={"language": "nl","task": "transcribe"})
|
| 28 |
return outputs["text"]
|
| 29 |
|
|
|
|
| 35 |
|
| 36 |
|
| 37 |
def speech_to_speech_translation(audio):
|
| 38 |
+
sampling_rate = 16000
|
| 39 |
+
data_array,samplerate = librosa.load(audio)
|
| 40 |
+
data_16 = librosa.resample(data_array, orig_sr=samplerate, target_sr=sampling_rate)
|
| 41 |
+
translated_text = translate(data_16)
|
| 42 |
synthesised_speech = synthesise(translated_text)
|
| 43 |
synthesised_speech = (synthesised_speech.numpy() * 32767).astype(np.int16)
|
| 44 |
return 16000, synthesised_speech
|