import gradio as gr
from asr import transcribe, ASR_EXAMPLES, ASR_LANGUAGES, ASR_NOTE
mms_transcribe = gr.Interface(
fn=transcribe,
inputs=[
gr.Audio(),
gr.Dropdown(
[f"{k} ({v})" for k, v in ASR_LANGUAGES.items()],
label="Language",
value="eng English",
),
# gr.Checkbox(label="Use Language Model (if available)", default=True),
],
outputs="text",
examples=ASR_EXAMPLES,
title="Speech-to-text",
description=(
"Transcribe audio from a microphone or input file in your desired language."
),
article=ASR_NOTE,
allow_flagging="never",
)
with gr.Blocks() as demo:
gr.HTML(
"""
Language Transcriber powered by UNESCO, Meta and Hugging Face
"""
)
gr.Markdown(
"MMS: Scaling Speech Technology to 1000+ languages demo. See our blog post and paper.
"
)
gr.HTML(
"""You can also finetune MMS models on your data using the recipes provides here."""
)
gr.HTML(
"""
for more control and no queue."""
)
mms_transcribe.render()
gr.HTML(
"""
"""
)
if __name__ == "__main__":
demo.queue()
demo.launch()