Intelligent Docs

Appearance

Using HuggingFace ​

Integrating HuggingFace models into your program is simple using intelligent.

Acquiring an API key. ​

  • Create a Hugging Face account: Sign up on the Hugging Face website if you don't already have an account.
  • Access your settings: Log in to your account and go to your profile settings.
  • Navigate to Access Tokens: Within your settings, find the "Access Tokens" tab.
  • Generate a new key: Click "New token" to create a new API key.
  • Name your key: Provide a descriptive name for your API key.
  • Select access level: Choose the appropriate access level (usually "Read") for your needs.
  • Copy your key: Once generated, copy your API key to use in your applications.

Integrating a Hugging Face AI Model ​

intelligent provides a simple interface for integrating a Hugging Face AI model into your program.

Prerequisites ​

  • Node.js (version 14 or higher)

  • A Hugging Face account with a valid API key

  • intelligent installed. npm install intelligent@latest

Usage ​

Here's a step-by-step guide to using the HuggingFaceService class:

1. Import the HuggingFaceService class ​

ts
import { HuggingFaceService } from 'intelligent';

2. Create an instance of the HuggingFaceService class ​

Pass your API key, model name, and optionally, parameters to the constructor. If no model name is provided, it defaults to "gpt2". If no parameters are provided, it defaults to { max_new_tokens: 50 }.

ts
const huggingFaceService = new HuggingFaceService('API-KEY', 'gpt2', { max_new_tokens: 50 });

3. Call the response method ​

Pass a prompt to the response method to generate a response using the Hugging Face AI model.

ts
await huggingFaceService.response('Your prompt here').then((res) => {
  console.log(res);
});

Supported Models ​

The HuggingFaceService class supports any Hugging Face AI model that is compatible with the @huggingface/inference package.

Configuration Options ​

The HuggingFaceService class supports the following configuration options:

  • model: The name of the Hugging Face model to use. Defaults to "gpt2".

  • params: An object containing parameters for the model. Defaults to { max_new_tokens: 50 }.

Error Handling ​

If no API key is provided or if the API key is invalid, the response method will log an error message to the console.

Example Use Case ​

Here's an example of using the HuggingFaceService class to generate a response to a user's query:

ts
import { HuggingFaceService } from 'intelligent';

const huggingFaceService = new HuggingFaceService('YOUR_API_KEY', 'gpt2', { max_new_tokens: 50 });

async function handleUserQuery(query) {
  try {
    const response = await huggingFaceService.response(query);
    console.log(response);
  } catch (error) {
    console.error(error);
  }
}

handleUserQuery('What is the meaning of life?');

By following these steps, you can easily integrate a Hugging Face AI model into your program using the HuggingFaceService class.