app.py

from transformers import pipeline
import gradio as gr
import time
import unicodedata
p = pipeline("automatic-speech-recognition",model="kingabzpro/wav2vec2-large-xls-r-300m-Urdu")

def transcribe(audio, state=""):
    time.sleep(2)
    text = p(audio)["text"]
    state += unicodedata.normalize("NFC",text) + " "
    return state, state

################### Gradio Web APP ################################

title = "Real-Time Urdu ASR"

description = """
<p>
<center>
This model is a fine-tuned version of facebook/wav2vec2-xls-r-300m on the common_voice dataset.
</center>
</p>
<center>
<img src="https://huggingface.co/spaces/kingabzpro/real-time-Urdu-ASR/resolve/main/Images/cover.jpg" alt="logo" width="550"/>
</center>
"""

article = "<p style='text-align: center'><a href='https://dagshub.com/kingabzpro/Urdu-ASR-SOTA' target='_blank'>Source Code on DagsHub</a></p><p style='text-align: center'><a href='https://huggingface.co/blog/fine-tune-xlsr-wav2vec2' target='_blank'>Fine-tuning XLS-R for Multi-Lingual ASR with 🤗 Transformers</a></p></center><center><img src='https://visitor-badge.glitch.me/badge?page_id=kingabzpro/real-time-Urdu-ASR' alt='visitor badge'></center></p>"


gr.Interface(
    fn=transcribe, 
    inputs=[
        gr.Audio(source="microphone", type="filepath", streaming=True), 
        "state"
    ],
    outputs=[
        "textbox",
        "state"
    ],
    title=title,
    description=description,
    article=article,
    theme='EveryPizza/Cartoony-Gradio-Theme',
    live=True).launch()