A simple tool to generate image caption with LLaVA model.
Before start, make sure to install LLaVA from https://github.com/haotian-liu/LLaVA. Use the same env as the LLaVA project needs. Once you have installed LLaVA, I suggest you to manually download the pretrained model from https://huggingface.co/liuhaotian/llava-v1.5-13b or 7b and place it in a folder you like.
Start the tool by running python main.py.
Enter the path of the model and choose quatization method if you want.