🦙✨ llama_flow

llama_flow is an open-source application designed to develop and customize Retrieval-Augmented Generation (RAG) workflows without code. Easily run it locally using a variety of open-source and closed-source large language models, vector databases, embedding models, and chunking strategies.

🚀 Features

🖥️ Frontend

• Built with Flutter
• Available for Linux and Windows
• Source Code: llama_flow_desktop

🛠️ Backend

• Dockerized for local hosting
• Utilizes FastAPI and llama_index

Supported Components

Basic RAG Workflow

Vector DB

Embed Model Provider

Embed Models (Huggingface)

LLM Providers

Huggingface LLMs

llamacpp LLMs

ollama LLMs

Chunking Strategy

Planned Features

• Advanced RAG workflow (See illustration below)
• Custom workflow with drag-and-drop functionality

🛠️ Installation

1. Frontend UI Application

Download the desktop application for your OS:

• Linux
• Windows

2. Backend Configuration

Option 1: Dockerized Backend

1. Clone the repo

   git clone https://github.com/sabeeralikp/llama_flow.git
   cd llama_flow
   

2. Build and Run Docker Container

   docker build -t llama_flow_image . 
   docker run -d --name llama_flow_container -p 8000:8000 llama_flow_image
   

Option 2: Manual Setup with FastAPI

1. Clone the repo

   git clone https://github.com/sabeeralikp/llama_flow.git
   cd llama_flow
   

2. Create a virtual environment

   python -m venv env
   source env/bin/activate   # On Windows use `env\Scripts\activate`
   

3. Install necessary packages

   pip install -r requirements.txt
   

4. Run the backend with Uvicorn

   uvicorn main:app --workers 4
   

Option 3: Colab/Kaggle Notebooks

• llama_flow with Huggingface: Colab Notebook
• llama_flow with llamacpp: Colab Notebook
• llama_flow with Ollama: Colab Notebook

Instructions:

1. Create a Ngrok account.
2. Go to the Ngrok dashboard, create an auth tunnel, copy the auth token, and paste it into Colab.
3. Open the notebook file in Colab and run all cells.
4. Copy the external link from the cell output.
5. Open the desktop application, go to settings, change the backend to remote, and paste the copied external domain link in the base URL text field.

Additional Backend Configurations

Using llamacpp

pip install -q llama-cpp-python \
  --extra-index-url https://abetlen.github.io/llama-cpp-python/whl/cu122  # For CUDA version 12.2
pip install -q llama_index llama-index-llms-llama-cpp

Using Ollama

Follow the official Ollama installation guide for different devices.

apt install lshw
curl -fsSL https://ollama.com/install.sh | sh
pip install -q llama-index-llms-ollama

🤝 Contribution Guidelines

We welcome all contributions to improve llama_flow. Please follow these steps:

1. Fork the repository
2. Create a new branch

   git checkout -b feature-branch
   

3. Commit your changes

   git commit -m 'Add some feature'
   

4. Push to the branch

   git push origin feature-branch
   

5. Create a new Pull Request

📜 License

This project is licensed under the MIT License. See the LICENSE file for details.

---

Stay tuned for exciting updates! 🚀✨