Script to transcribe audio to text using Xenova/whisper-small and transformers.js

transcribe.ts

import fs from 'fs'
import { pipeline } from '@xenova/transformers'
import { WaveFile } from 'wavefile';

const options = { //Xenova/whisper-small configurations
    chunk_length_s: 30,
    stride_length_s: 5,
    language: 'portuguese',
    task: 'transcribe',
}

export async function transcribeAudio(fileName: string) {

    let data = null;

    try {
        console.log('[STARTING TRANSCRIBE]')

        const transcriber = await pipeline('automatic-speech-recognition', 'Xenova/whisper-small') //config transformer.js

        let buffer = fs.readFileSync(fileName) //file buffer

        let wav = new WaveFile(buffer); // start .wav processing
        wav.toBitDepth('32f');
        wav.toSampleRate(16000);
        let audioData = wav.getSamples();

        if (Array.isArray(audioData)) {
            audioData = audioData[0];
        }

        let start = performance.now();

        data = await transcriber(audioData, options); // process transcribe
        
        let end = performance.now();

        console.log(`Transcribe execution duration: ${(end - start) / 1000} seconds`);
    } catch (error) {
        console.log('[ERROR TRANSCRIBE]', error)
        throw new Error('Error')
    } finally {
        return data
    }
}