All functionality related toHugging Face Hub and libraries liketransformers,sentence transformers, anddatasets.

Hugging Face is an AI platform with all major open source models, datasets, MCPs, and demos.It supplies model inference locally and via serverlessInference Providers.

You can useInference Providers to run open source models like DeepSeek R1 on scalable serverless infrastructure.

Installation​

Most of the Hugging Face integrations are available in thelangchain-huggingface package.

pip install langchain-huggingface

Chat models​

ChatHuggingFace​

We can use theHugging Face LLM classes or directly use theChatHuggingFace class.

See ausage example.

from langchain_huggingfaceimport ChatHuggingFace

LLMs​

HuggingFaceEndpoint​

We can use theHuggingFaceEndpoint class to run open source models via serverlessInference Providers or via dedicatedInference Endpoints.

See ausage example.

from langchain_huggingfaceimport HuggingFaceEndpoint

HuggingFacePipeline​

We can use theHuggingFacePipeline class to run open source models locally.

See ausage example.

from langchain_huggingfaceimport HuggingFacePipeline

Embedding Models​

HuggingFaceEmbeddings​

We can use theHuggingFaceEmbeddings class to run open source embedding models locally.

See ausage example.

from langchain_huggingfaceimport HuggingFaceEmbeddings

HuggingFaceEndpointEmbeddings​

We can use theHuggingFaceEndpointEmbeddings class to run open source embedding models via a dedicatedInference Endpoint.

See ausage example.

from langchain_huggingfaceimport HuggingFaceEndpointEmbeddings

HuggingFaceInferenceAPIEmbeddings​

We can use theHuggingFaceInferenceAPIEmbeddings class to run open source embedding models viaInference Providers.

See ausage example.

from langchain_community.embeddingsimport HuggingFaceInferenceAPIEmbeddings

HuggingFaceInstructEmbeddings​

We can use theHuggingFaceInstructEmbeddings class to run open source embedding models locally.

See ausage example.

from langchain_community.embeddingsimport HuggingFaceInstructEmbeddings

HuggingFaceBgeEmbeddings​

BGE models on the HuggingFace are one ofthe best open-source embedding models.BGE model is created by theBeijing Academy of Artificial Intelligence (BAAI).BAAI is a private non-profit organization engaged in AI research and development.

See ausage example.

from langchain_community.embeddingsimport HuggingFaceBgeEmbeddings

Document Loaders​

Hugging Face dataset​

Hugging Face Hub is home to over 75,000datasets in more than 100 languagesthat can be used for a broad range of tasks across NLP, Computer Vision, and Audio.They used for a diverse range of tasks such as translation, automatic speechrecognition, and image classification.

We need to installdatasets python package.

pip install datasets

See ausage example.

from langchain_community.document_loaders.hugging_face_datasetimport HuggingFaceDatasetLoader

Hugging Face model loader​

Load model information fromHugging Face Hub, including README content.

This loader interfaces with theHugging Face Models API to fetchand load model metadata and README files.The API allows you to search and filter models based onspecific criteria such as model tags, authors, and more.

from langchain_community.document_loadersimport HuggingFaceModelLoader

Image captions​

It uses the Hugging Face models to generate image captions.

We need to install several python packages.

pip install transformers pillow

See ausage example.

from langchain_community.document_loadersimport ImageCaptionLoader

Tools​

Hugging Face Hub Tools​

Hugging Face Toolssupport text I/O and are loaded using theload_huggingface_tool function.

We need to install several python packages.

pip install transformers huggingface_hub

See ausage example.

from langchain_community.agent_toolkits.load_toolsimport load_huggingface_tool

Hugging Face Text-to-Speech Model Inference.​

It is a wrapper aroundOpenAI Text-to-Speech API.

from langchain_community.tools.audioimport HuggingFaceTextToSpeechModelInference