Commit
·
b584d69
1
Parent(s):
748ad62
Update app.py
Browse files
app.py
CHANGED
@@ -33,11 +33,11 @@ def speech_to_speech_translation(audio):
|
|
33 |
|
34 |
|
35 |
title = "Cascaded STST"
|
36 |
-
|
37 |
-
|
38 |
-
|
39 |
-
|
40 |
-
|
41 |
|
42 |
demo = gr.Blocks()
|
43 |
|
@@ -45,8 +45,9 @@ mic_translate = gr.Interface(
|
|
45 |
fn=speech_to_speech_translation,
|
46 |
inputs=gr.Audio(source="microphone", type="filepath"),
|
47 |
outputs=gr.Audio(label="Generated Speech", type="numpy"),
|
|
|
48 |
title=title,
|
49 |
-
|
50 |
)
|
51 |
|
52 |
file_translate = gr.Interface(
|
@@ -55,7 +56,7 @@ file_translate = gr.Interface(
|
|
55 |
outputs=gr.Audio(label="Generated Speech", type="numpy"),
|
56 |
examples=[["./example.wav"]],
|
57 |
title=title,
|
58 |
-
|
59 |
)
|
60 |
|
61 |
with demo:
|
|
|
33 |
|
34 |
|
35 |
title = "Cascaded STST"
|
36 |
+
description = """
|
37 |
+
Demo for cascaded speech-to-speech translation (STST), mapping from source speech in any language to target speech in French. Demo uses OpenAI's [Whisper Base](https://huggingface.co/openai/whisper-base) model for speech translation, and Facebook's
|
38 |
+
[MMS TTS](https://huggingface.co/facebook/mms-tts) model for text-to-speech:
|
39 |
+

|
40 |
+
"""
|
41 |
|
42 |
demo = gr.Blocks()
|
43 |
|
|
|
45 |
fn=speech_to_speech_translation,
|
46 |
inputs=gr.Audio(source="microphone", type="filepath"),
|
47 |
outputs=gr.Audio(label="Generated Speech", type="numpy"),
|
48 |
+
examples=[["./example.wav"]],
|
49 |
title=title,
|
50 |
+
description=description,
|
51 |
)
|
52 |
|
53 |
file_translate = gr.Interface(
|
|
|
56 |
outputs=gr.Audio(label="Generated Speech", type="numpy"),
|
57 |
examples=[["./example.wav"]],
|
58 |
title=title,
|
59 |
+
description=description,
|
60 |
)
|
61 |
|
62 |
with demo:
|