Update app.py
Browse files
app.py
CHANGED
|
@@ -5,13 +5,14 @@ from transformers import pipeline
|
|
| 5 |
|
| 6 |
# Load models
|
| 7 |
asr = pipeline("automatic-speech-recognition", model="openai/whisper-small")
|
| 8 |
-
tts = pipeline("text-to-speech", model="facebook/mms-tts-
|
| 9 |
|
| 10 |
def speech_to_speech(audio):
|
| 11 |
# Speech → Text
|
| 12 |
-
|
|
|
|
| 13 |
|
| 14 |
-
# Text → Speech
|
| 15 |
speech = tts(text)
|
| 16 |
|
| 17 |
# Save audio
|
|
@@ -24,10 +25,10 @@ demo = gr.Interface(
|
|
| 24 |
inputs=gr.Audio(type="filepath", label="Speak here"),
|
| 25 |
outputs=[
|
| 26 |
gr.Textbox(label="Recognized Text"),
|
| 27 |
-
gr.Audio(label="
|
| 28 |
],
|
| 29 |
-
title="Speech to Speech AI",
|
| 30 |
-
description="Speak into the mic, AI listens and
|
| 31 |
)
|
| 32 |
|
| 33 |
demo.launch()
|
|
|
|
| 5 |
|
| 6 |
# Load models
|
| 7 |
asr = pipeline("automatic-speech-recognition", model="openai/whisper-small")
|
| 8 |
+
tts = pipeline("text-to-speech", model="facebook/mms-tts-hin")
|
| 9 |
|
| 10 |
def speech_to_speech(audio):
|
| 11 |
# Speech → Text
|
| 12 |
+
result = asr(audio)
|
| 13 |
+
text = result["text"]
|
| 14 |
|
| 15 |
+
# Text → Speech (Hindi voice)
|
| 16 |
speech = tts(text)
|
| 17 |
|
| 18 |
# Save audio
|
|
|
|
| 25 |
inputs=gr.Audio(type="filepath", label="Speak here"),
|
| 26 |
outputs=[
|
| 27 |
gr.Textbox(label="Recognized Text"),
|
| 28 |
+
gr.Audio(label="Hindi Speech Output")
|
| 29 |
],
|
| 30 |
+
title="Speech to Speech AI (Hindi)",
|
| 31 |
+
description="Speak into the mic, AI listens and replies in Hindi voice"
|
| 32 |
)
|
| 33 |
|
| 34 |
demo.launch()
|