NanoLlava is the tiniest Visual Language Model

Learn how to run it with Llama.cpp

<img src='https://github.com/fabiomatricardi/NanollavaCPP/raw/main/logonanollavaSocial.png' height=200>

---

Dependencies

• create a virtual environment
• install the following packages:

pip install streamlit==1.34.0 llama-cpp-python==0.2.75

download in a local directory called nanolava the 2 files from:<br> https://huggingface.co/abetlen/nanollava-gguf

• nanollava-mmproj-f16.gguf
• nanollava-text-model-f16.gguf

Run the interface

from the terminal with the venv activated:

streamlit run st-NanoLlava.py

<img src='https://github.com/fabiomatricardi/NanollavaCPP/raw/main/nanollavaStreamlit.gif' width=800> <br><br>