NextJS Transformers Module parse failed: Unexpected character

Question

I need to create embeddings in NextJS 15, so I started looking into @xenova/transformers and @huggingface/transformers. It was too simple to implement in a Node script (without NextJS), but I am struggling to make this work.

After the project creation, I created an API route as usual (/app/api/route.ts) and proceeded with the given example:

import { pipeline } from '@huggingface/transformers'

export async function GET() {
    const extractor = await pipeline(
        'feature-extraction',
        'Xenova/all-MiniLM-L6-v2'
    )

    const sentences = ['This is an example sentence', 'Each sentence is converted'];
    const output = await extractor(sentences, { pooling: 'mean', normalize: true });

    console.log(output)
}

But when I visit the API URL I get this error:

./node_modules/@huggingface/transformers/node_modules/onnxruntime-node/bin/napi-v3/darwin/arm64/onnxruntime_binding.node Module parse failed: Unexpected character '�' (1:0) You may need an appropriate loader to handle this file type, currently no loaders are configured to process this file. See https://webpack.js.org/concepts#loaders (Source code omitted for this binary file)

I have tried also to update the Next configuration file next.config.ts with:

import type { NextConfig } from "next";

const nextConfig: NextConfig = {
  output: 'standalone',
  serverExternalPackages: ['sharp', 'onnxruntime-node'] 
};

export default nextConfig;

I don't really know how to proceed and implement this in the server side (API routes). If I need a loader, I don't know how to implement it.

Looks like file's encoding is changed, change encoding to utf-8 in vs code and then try. — Akash Kava
– Akash Kava, Commented Dec 29, 2024 at 14:11

Rajesh Gauswami · Accepted Answer · 2025-01-05 08:21:52Z

0

This may help you

Update Next.js Configuration
Adjust your next.config.js to exclude onnxruntime-node from Webpack bundling. This ensures the binary file is treated as a native Node.js module:

const nextConfig = {
  output: 'standalone',
  webpack: (config, { isServer }) => {
    if (isServer) {
      config.externals = {
        ...config.externals,
        'onnxruntime-node': 'commonjs onnxruntime-node',
      };
    }
    return config;
  },
};

export default nextConfig;

Server-Side Execution
Ensure the embedding code runs on the server side. Update your API route to use onnxruntime-node properly in the Node.js environment:

import { pipeline } from '@huggingface/transformers';

export async function GET() {
  try {
    const extractor = await pipeline(
      'feature-extraction',
      'Xenova/all-MiniLM-L6-v2'
    );

    const sentences = ['This is an example sentence', 'Each sentence is converted'];
    const output = await extractor(sentences, { pooling: 'mean', normalize: true });

    return new Response(JSON.stringify(output), { status: 200 });
  } catch (error) {
    console.error('Error:', error);
    return new Response('Error generating embeddings', { status: 500 });
  }
}

answered Jan 5 at 8:21

Rajesh Gauswami

5727 silver badges22 bronze badges

Sign up to request clarification or add additional context in comments.

1 Comment

Tincho.UY Jan 14 at 11:47

Thanks for answering. Unfortunately, I get this error and I can't run the server.

Collectives™ on Stack Overflow

NextJS Transformers Module parse failed: Unexpected character

1 Answer 1

1 Comment

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

1 Comment

Your Answer

Sign up or log in

Post as a guest

Related