Can we create a Maltese AI Lawyer? This is the code used for the 6th May 2024 GDG session about using LLMs and RAG to answer questions about Maltese legislation.
- Miniconda
- Ollama: https://github.com/ollama/ollama
- conda create -n mal
- conda activate mal
- conda install pip
- pip install llama-index (Note: llama-index was at 0.10.33 at time of writing)
- pip install llama-index-core llama-index-readers-file llama-index-llms-ollama llama-index-embeddings-huggingface
- ollama pull llama3:8b
- ollama pull gemma:7b
- ollama pull phi3:3.8b
- python3 test_e2e.py
Test and visualise embeddings generated
- pip install matplotlib
- python test_generate_embeddings.py
- python visualise_embeddings.py
- docker run --name redis-vecdb -d -p 6379:6379 -p 8001:8001 --volume ./redis_storage/:/data redis/redis-stack:latest
- pip install llama-index-vector-stores-redis redis
- python test_generate_embeddings.py
- python generate_embeddings.py
- Ensure the redis docker container from above is running
- Ensure all embeddings have been generated
- python lawyer.py