HuggingFace Inference Embeddings

​

Node Details

• Name: HuggingFaceInferenceEmbeddings
• Type: HuggingFaceInferenceEmbeddings
• Version: 1.0
• Category: Embeddings

​

Parameters

​

Credential

• Label: Connect Credential
• Name: credential
• Type: credential
• Credential Names: huggingFaceApi

​

Inputs

1. Model

   • Label: Model
   • Name: modelName
   • Type: string
   • Description: The name of the HuggingFace model to use for embeddings. Leave blank if using a custom inference endpoint.
   • Placeholder: sentence-transformers/distilbert-base-nli-mean-tokens
   • Optional: Yes

2. Endpoint

   • Label: Endpoint
   • Name: endpoint
   • Type: string
   • Description: The URL of your custom inference endpoint, if using one.
   • Placeholder: https://xyz.eu-west-1.aws.endpoints.huggingface.cloud/sentence-transformers/all-MiniLM-L6-v2
   • Optional: Yes

​

Initialization

The node initializes by creating an instance of HuggingFaceInferenceEmbeddings with the following steps:

1. Retrieves the HuggingFace API key from the provided credentials.
2. Sets up the configuration object with the API key.
3. If a model name is provided, it’s added to the configuration.
4. If a custom endpoint is provided, it’s added to the configuration.
5. Creates and returns a new HuggingFaceInferenceEmbeddings instance with the configured parameters.

​

Usage

This node is typically used in workflows where text needs to be converted into numerical vector representations. Common use cases include:

• Text similarity comparisons
• Document clustering
• Input preparation for machine learning models
• Semantic search implementations

By leveraging HuggingFace’s pre-trained models or custom-deployed endpoints, users can easily generate high-quality embeddings for a wide range of natural language processing tasks.