Switching to processes and whisper.cpp to avoid transformers.js memory leaks
This commit is contained in:
@@ -1,11 +1,13 @@
|
||||
import { pipeline } from '@xenova/transformers';
|
||||
import { parentPort } from 'worker_threads';
|
||||
|
||||
let embedder: any;
|
||||
const [modelDir, model] = process.argv.slice(2);
|
||||
|
||||
parentPort?.on('message', async ({text, model, modelDir }) => {
|
||||
if(!embedder) embedder = await pipeline('feature-extraction', 'Xenova/' + model, {quantized: true, cache_dir: modelDir});
|
||||
let text = '';
|
||||
process.stdin.on('data', chunk => text += chunk);
|
||||
process.stdin.on('end', async () => {
|
||||
const embedder = await pipeline('feature-extraction', 'Xenova/' + model, {quantized: true, cache_dir: modelDir});
|
||||
const output = await embedder(text, { pooling: 'mean', normalize: true });
|
||||
const embedding = Array.from(output.data);
|
||||
parentPort?.postMessage({embedding});
|
||||
console.log(JSON.stringify({embedding}));
|
||||
process.exit();
|
||||
});
|
||||
|
||||
Reference in New Issue
Block a user