dougbtv / ollama-llava-captioner

A way to caption images using LLaVa via ollama

LLaVa stable diffusion training image captioner

This script takes a directory of images and

Setup:

pip install ollama

Running it...

python llava-caption-ollama.py /path/to/images

or, specify your own prompt file:

python llava-caption-ollama.py /path/to/images /path/to/myprompt.txt

If the prompt argument is empty, uses the prompt.txt in this directory.