Spaces:
Sleeping
Sleeping
import spaces | |
from transformers import pipeline | |
import gradio as gr | |
import torch | |
asr = pipeline(model="asif00/whisper-bangla") | |
ser = pipeline("text2text-generation", model="asif00/mbart_bn_error_correction") | |
def transcribe(audio): | |
text = asr(audio)["text"] | |
print(text) | |
return text | |
def correction(text): | |
corrected_text = ser(text)[0]["generated_text"] | |
print(corrected_text) | |
return corrected_text | |
def transcribe_and_correct(audio): | |
text = transcribe(audio) | |
corrected_text = correction(text) | |
return corrected_text | |
iface = gr.Interface( | |
fn=transcribe_and_correct, | |
inputs=gr.Audio(sources="microphone", type="filepath"), | |
outputs="text", | |
title="Whisper Bangla", | |
description="Realtime demo for Bengali speech recognition using a fine-tuned Whisper small model.", | |
) | |
iface.launch() | |