Fast.AI LangChain Presentation

This README provides the steps necessary to run the code presented in the LangChain introduction and code walkthrough. Follow these instructions to set up your environment and execute the notebook.

1. Create an Account at Pinecone

• Note: You will need an account, but the usage required for this notebook is well within the free tier and won't cost you anything.
• Once your account is created, create a new project. You can name this project whatever you'd like.

2. Create a New Index

• Name the index textbook-vector-store.
  • Note: You can name your index differently, but ensure you update the notebook to reflect the new name.
• Set the dimensions of the index to 768.
  • Note: The dimensions are based on the embedding model you use. For this demonstration, a local HuggingFace sentence-transformers/all-mpnet-base-v2 model was used, so the dimensions are 768. If you're using OpenAI's text-embedding-3-small embedding model, set the index dimensions to 1536.

3. Set Environment Variables

• Set the PINECONE_API_KEY environment variable to the API key for your project. This can be found on the middle left-hand side of the screen after you have selected your project from the list.
• Set the OPENAI_API_KEY environment variable to your OpenAI API key.
• (Optional) Set your ANTHROPIC_API_KEY environment variable to your AnthropIC API key. This isn't used in the actual RAG implementation but is included as an example earlier in the notebook, so it's not required.

4. Clone the Notebook to Your Local Machine

• Clone the repository containing the notebook to your local machine.

5. Create a Virtual Environment

• Run the following command to create a virtual environment:

  python -m venv venv
  

• Enter and activate your virtual environment:
  • On Windows:

    .\venv\Scripts\activate
    

  • On macOS/Linux:

    source venv/bin/activate

6. Install Required Packages

• Install the required packages using the requirements.txt file:

  pip install -r requirements.txt

7. Install PyTorch with CUDA

• Visit PyTorch's official website and get the command for your particular setup to install PyTorch with CUDA enabled in your virtual environment.
  • For example: On my setup, I used:

    pip3 install torch torchvision torchaudio --index-url https://download.pytorch.org/whl/cu118

8. Run the Notebook

• That's it! You can now run all the cells in the notebook and create your own notebook-based textbook RAG pipeline.

This presentation was presented live on June 29, 2024, and the video is available on YouTube at https://www.youtube.com/watch?v=D11C8MrFEUk.