Update ML Intern artifact metadata

Update README with complete setup guide

Simplify server.js - set numThreads=1 before any import

Remove preload from build - model loads at runtime instead

Fix server.js: disable threads for Alpine

Fix: disable WASM threads to avoid worker error, set single-thread from start

Update Dockerfile: use correct onnxruntime-web version path for wasm files

Fix: pin onnxruntime-web to 1.14.0 (exact version @xenova /transformers v2 expects)

Fix: copy WASM binaries to where onnxruntime-web expects them

Add onnxruntime-web as explicit dependency

Fix: create onnxruntime-node stub that redirects to onnxruntime-web

Fix: remove onnxruntime-node after install, force WASM fallback

Simplify Dockerfile

Rewrite server.js for @xenova /transformers v2

Rewrite preload.js for @xenova /transformers v2

Complete rewrite: use @xenova /transformers v2 which works out of box on Alpine

Fix server.js: set WASM paths locally

Fix: set WASM paths to local files before import, use onnxruntime-web node entry

Fix: add onnxruntime-web explicitly, override only onnxruntime-node

Fix server.js: remove broken env.backends path

Fix: remove broken env.backends path, use correct v3 API

Update Dockerfile: add npm cache clean for smaller image

Fix server.js: force WASM device

Fix preload.js: force WASM backend explicitly

Fix Alpine: exclude onnxruntime-node, force WASM backend only

Add .dockerignore

Add preload.js - downloads model at build time

Add Dockerfile for Alpine

Add server.js - main embedding API

Add package.json

Update ML Intern artifact metadata

Initial setup: Alpine + Node.js embedding server with @huggingface/transformers WASM

Update ML Intern artifact metadata

initial commit