LLM Explorer

LLM Explorer is a powerful tool that allows you to run and test Large Language Models (LLMs) with a chat interface for free on Google Colab. This project demonstrates the capabilities of modern AI technologies and provides an accessible platform for experimentation and learning.

Features

• Free GPU Access: Utilizes Google Colab's T4 GPU resources for cost-effective LLM testing.
• Interactive Chat Interface: Streamlit-based UI for easy interaction with the LLM.
• Document Processing: Upload and analyze PDF documents to provide context for the LLM.
• Multiple LLM Support: Currently implements Llama 3.1, with potential for easy integration of other models.
• Vector Store Integration: Uses Chroma for efficient storage and retrieval of document embeddings.

Getting Started

1. Open the Chatbot.ipynb notebook in Google Colab.
2. Ensure you're using a T4 GPU runtime.
3. Run the installation cells to set up the required dependencies.
4. Follow the instructions to install and run Ollama.
5. Upload a PDF document when prompted.
6. Interact with the chatbot through the Streamlit interface.

Components

• Chatbot.ipynb: Main Jupyter notebook containing the setup and execution code.
• app.py: Streamlit application for the chat interface.
• Dependencies: langchain, llama-index, streamlit, ngrok, and others (see installation cells in the notebook).

Usage

After setting up the environment and uploading a document:

1. The system will process the document and create a vector store.
2. A chat interface will be available through a public URL (provided by ngrok).
3. Ask questions related to the uploaded document, and the LLM will provide context-aware responses.

Customization

• To use a different LLM, modify the Ollama initialization in the code.
• Adjust the embedding model by changing the SentenceTransformerEmbeddings parameters.
• Modify the app.py file to customize the Streamlit interface.

Limitations

• Requires a Google account with access to Colab.
• Performance may vary based on Colab's available resources.
• The free ngrok plan has limitations on connection duration and bandwidth.

Future Improvements

• Support for multiple document types (beyond PDF).
• Integration with more LLM models.
• Enhanced error handling and user feedback.
• Persistent storage options for chat history and document embeddings.

Contributing

Contributions to LLM Explorer are welcome! Please feel free to submit pull requests, create issues, or suggest improvements.

License

This project is open-source and available under the MIT License.

Acknowledgments

• Thanks to the teams behind Langchain, Llama Index, and Streamlit for their excellent tools.
• Gratitude to Google Colab for providing free GPU resources for education and experimentation.