Hopefully fixed bug
Browse files
app.py
CHANGED
|
@@ -44,7 +44,7 @@ def speech_to_speech_translation(audio):
|
|
| 44 |
print(translated_text)
|
| 45 |
synthesised_speech = synthesise(translated_text)
|
| 46 |
synthesised_speech = (synthesised_speech.numpy() * 32767).astype(np.int16)
|
| 47 |
-
return 16000, synthesised_speech
|
| 48 |
|
| 49 |
|
| 50 |
title = "Cascaded STST"
|
|
@@ -60,7 +60,10 @@ demo = gr.Blocks()
|
|
| 60 |
mic_translate = gr.Interface(
|
| 61 |
fn=speech_to_speech_translation,
|
| 62 |
inputs=gr.Audio(source="microphone", type="filepath"),
|
| 63 |
-
outputs=
|
|
|
|
|
|
|
|
|
|
| 64 |
title=title,
|
| 65 |
description=description,
|
| 66 |
)
|
|
@@ -68,7 +71,10 @@ mic_translate = gr.Interface(
|
|
| 68 |
file_translate = gr.Interface(
|
| 69 |
fn=speech_to_speech_translation,
|
| 70 |
inputs=gr.Audio(source="upload", type="filepath"),
|
| 71 |
-
outputs=
|
|
|
|
|
|
|
|
|
|
| 72 |
examples=[["./example.wav"]],
|
| 73 |
title=title,
|
| 74 |
description=description,
|
|
|
|
| 44 |
print(translated_text)
|
| 45 |
synthesised_speech = synthesise(translated_text)
|
| 46 |
synthesised_speech = (synthesised_speech.numpy() * 32767).astype(np.int16)
|
| 47 |
+
return (16000, synthesised_speech.T), translated_text
|
| 48 |
|
| 49 |
|
| 50 |
title = "Cascaded STST"
|
|
|
|
| 60 |
mic_translate = gr.Interface(
|
| 61 |
fn=speech_to_speech_translation,
|
| 62 |
inputs=gr.Audio(source="microphone", type="filepath"),
|
| 63 |
+
outputs=[
|
| 64 |
+
gr.Audio(label="Generated Speech", type="numpy"),
|
| 65 |
+
gr.Text(label="Translation")
|
| 66 |
+
]
|
| 67 |
title=title,
|
| 68 |
description=description,
|
| 69 |
)
|
|
|
|
| 71 |
file_translate = gr.Interface(
|
| 72 |
fn=speech_to_speech_translation,
|
| 73 |
inputs=gr.Audio(source="upload", type="filepath"),
|
| 74 |
+
outputs=[
|
| 75 |
+
gr.Audio(label="Generated Speech", type="numpy"),
|
| 76 |
+
gr.Text(label="Translation")
|
| 77 |
+
]
|
| 78 |
examples=[["./example.wav"]],
|
| 79 |
title=title,
|
| 80 |
description=description,
|