For data privacy reasons and to save costs, I've created ChadGPT, the poor man's ChatGPT!

Pay attention to tokens per second to understand your hardware's performance for LLM inference.

Outside of dependencies in pyproject.toml that can be installed with poetry install based on lock file. Make sure you install Ollama and whatever model you choose. Here I'm using "llama3" but this can be easily parametrized.

streamlit run chadgpt.py