app.py

import torch
import torchaudio
from transformers import Wav2Vec2ForCTC, Wav2Vec2Tokenizer
import gradio as gr

# Load pre-trained model and tokenizer
model = Wav2Vec2ForCTC.from_pretrained("facebook/wav2vec2-large-960h")
tokenizer = Wav2Vec2Tokenizer.from_pretrained("facebook/wav2vec2-large-960h")

def transcribe_speech(audio_file):
    # Load and convert audio file to waveform
    waveform, _ = torchaudio.load(audio_file)

    # Preprocess waveform
    input_values = tokenizer(waveform, return_tensors="pt").input_values

    # Perform inference
    with torch.no_grad():
        logits = model(input_values).logits

    # Get predicted transcription
    predicted_ids = torch.argmax(logits, dim=-1)
    transcription = tokenizer.batch_decode(predicted_ids)[0]

    return transcription

# Define Gradio interface
def speech_recognition(audio_file):
    transcription = transcribe_speech(audio_file)
    return transcription

inputs = gr.inputs.Audio(type="file", label="Upload Audio File")
outputs = gr.outputs.Textbox(label="Transcription")
interface = gr.Interface(fn=speech_recognition, inputs=inputs, outputs=outputs)

# Run the Gradio interface
interface.launch()