app.py

import gradio as gr
from asr import transcribe
# from tts import synthesize


mms_transcribe = gr.Interface(
    fn=transcribe,
    inputs=[
        gr.Audio()
    ],
    outputs="text",
    #examples=ASR_EXAMPLES,
    title="Speech-to-text",
    description=(
        "Transcribe audio from a microphone or input file."
    ),
    #article=ASR_NOTE,
    allow_flagging="never",
)

# mms_synthesize = gr.Interface(
#     fn=synthesize,
#     inputs=[
#         gr.Text(label="Input text"),
#     ],
#     outputs=[
#         gr.Audio(label="Generated Audio", type="numpy"),
#         gr.Text(label="Filtered text after removing OOVs"),
#     ],
#     #examples=TTS_EXAMPLES,
#     title="Text-to-speech",
#     description=("Generate audio from input text."),
#     allow_flagging="never",
# )

tabbed_interface = gr.TabbedInterface(
    [mms_transcribe],
    ["Speech-to-text"],
)

# tabbed_interface = gr.TabbedInterface(
#     [mms_transcribe, mms_synthesize],
#     ["Speech-to-text", "Text-to-speech"],
# )

with gr.Blocks() as demo:
    tabbed_interface.render()

if __name__ == "__main__":
    demo.queue()
    demo.launch()