DOC: NodeJS Example Using API Embedding

[zwilch]

Issue with current documentation:

I can not find a working nodejs API example for using Embedding. I have figured out, I should use "all-MiniLM-L6-v2-f16.gguf"

I stuck with that:

import { OpenAI } from 'openai';
const API_BASE   ="http://localhost:4891/v1"

export const run = async () => {
const openai = new OpenAI({
  apiKey: "not needed for a local LLM",
  baseURL:API_BASE+'/v1'
});

const embedding = await openai.embeddings.create({
model: "all-MiniLM-L6-v2-f16.gguf",
input: "The quick brown fox jumped over the lazy dog"
//encoding_format: “float”,
});

console.log(embedding);
}//end run();

run();//start

Idea or request for content:

Make a working example code for nodejs to use embedding API.

[jacoobes]

https://github.com/nomic-ai/gpt4all/blob/main/gpt4all-bindings/typescript/spec/embed.mjs

[jacoobes]

youll have to change the .bin ending to .gguf but essentially it is the same

[zwilch]

npm install gpt4all

import { loadModel, createEmbedding } from 'gpt4all'
const embedder = await loadModel("all-MiniLM-L6-v2-f16.gguf", { verbose: true, type: 'embedding'})
console.log(createEmbedding(embedder, "Hallo Welt!"))

it is working

any idea how to get it working with import {OpenAI} from "openai" ?

[jacoobes]

Openai and gpt4all are two different projects

[zwilch]

Ja @jacoobes richtig erkannt, es ist wirklich schlecht, dass man von seiner NodeJS APP nicht durch Ändern der URL einfach mal auf das langsame lokal laufende GPT4ALL Modell wechseln kann.