RAG Chatbot with Langchain and HugginFace 🦜🔗🤗

Retrieval Augmented Generation Chatbot using Langchain 🦜🔗 and HuggingFace 🤗

Overview

The concept of Retrieval Augmented Generation (RAG) involves leveraging pre-trained Large Language Models (LLM) alongside custom data to produce responses. This approach merges the capabilities of pre-trained dense retrieval and sequence-to-sequence models. In practice, RAG models first retrieve relevant documents, then feed them into a sequence-to-sequence model, and finally aggregate the results to generate outputs. By integrating these components, RAG enhances the generation process by incorporating both the comprehensive knowledge of pre-trained models and the specific context provided by custom data.

Getting Started

Environment Setup

To get started, create a virtual environment and activate it:

virtualenv venv
source venv/bin/activate

Create a local environment file (.env) and add your huggingface API key:

HF_TOKEN=your_huggingface_api_key

Install Dependencies

Next, install the required dependencies using pip:

pip install -r requirements.txt

Run the Application

Now, you can run the application:

gradio app.py

This will start the application, allowing you to chat with the RAG model.

Usage

Once the application is up and running, you can interact with the chatbot through a web interface.

Additional Resources

• Check out the chatbot on
• Explore more about the databricks-dolly-15k dataset here
• Explore more about the Dynamic-TinyBERT model here
• Explore more about the sentence-transformers (all-MiniLM-L6-v2) model here