app.py

import torch
import gradio as gr
import pytube as pt
from transformers import pipeline
from huggingface_hub import model_info
import time
import unicodedata

MODEL_NAME = "SakshiRathi77/wav2vec2-large-xlsr-300m-hi-kagglex"
lang = "hi"

device = 0 if torch.cuda.is_available() else "cpu"
pipe = pipeline(
    task="automatic-speech-recognition",
    model=MODEL_NAME,
    device=device,
)

def transcribe(microphone, file_upload):
    warn_output = ""
    if (microphone is not None) and (file_upload is not None):
        warn_output = (
            "WARNING: You've uploaded an audio file and used the microphone. "
            "The recorded file from the microphone will be used and the uploaded audio will be discarded.\n"
        )

    elif (microphone is None) and (file_upload is None):
        return "ERROR: You have to either use the microphone or upload an audio file"
    file = microphone if microphone is not None else file_upload
    text = pipe(file)["text"]
    
    return warn_output + text


# def _return_yt_html_embed(yt_url):
#     video_id = yt_url.split("?v=")[-1]
#     HTML_str = (
#         f'<center> <iframe width="500" height="320" src="https://www.youtube.com/embed/{video_id}"> </iframe>'
#         " </center>"
#     )
#     return HTML_str


# def yt_transcribe(yt_url):
#     yt = pt.YouTube(yt_url)
#     html_embed_str = _return_yt_html_embed(yt_url)
#     stream = yt.streams.filter(only_audio=True)[0]
#     stream.download(filename="audio.mp3")

#     text = pipe("audio.mp3")["text"]

#     return html_embed_str, text

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


demo = gr.Blocks()

examples=[["examples/example1.mp3"], ["examples/example2.mp3"]]

description = """
<p>
<center>
Welcome to the HindiSpeechPro, a cutting-edge interface powered by a fine-tuned version of facebook/wav2vec2-xls-r-300m on the common_voice dataset. Easily convert your spoken words to accurate text with just a few clicks.
</center>
</p>
<center>
<img src="https://huggingface.co/spaces/kingabzpro/real-time-Urdu-ASR/resolve/main/Images/cover.jpg" alt="logo" width="550"/>
</center>
"""

mf_transcribe = gr.Interface(
    fn=transcribe,
    inputs=[
        gr.inputs.Audio(source="microphone", type="filepath"),
        gr.inputs.Audio(source="upload", type="filepath"),
    ],
    outputs="text",
    theme="huggingface",
    title="HindiSpeechPro: WAV2VEC-Powered ASR Interface",
    description= description ,
    allow_flagging="never",
    examples=examples,
).launch(share=True )


gr.Interface.load("models/SakshiRathi77/wav2vec2-large-xlsr-300m-hi-kagglex").launch()