JoyCaption | Automatic Image Tagging for Training Set

2024.10.14

This workflow leverages the latest Joy_Caption_Two. Simply input the path to your image folder, and it will automatically generate image captions.

To use this workflow correctly, you need to install the following models and place them in the correct directory:

- ComfyUI/models/clip

https://hf-mirror.com/google/siglip-so400m-patch14-384/tree/main

- ComfyUI/models/LLM

https://hf-mirror.com/unsloth/Meta-Llama-3.1-8B-Instruct-bnb-4bit (low VRAM version)

https://hf-mirror.com/unsloth/Meta-Llama-3.1-8B-Instruct

- ComfyUI/models/Joy_caption_two (Create this yourself)

https://huggingface.co/spaces/fancyfeast/joy-caption-alpha-two/tree/main/cgrkzexw-599808

Download all content within the cgrkzexw-599808 folder into Joy_caption_two, ensuring the final directory structure is as follows: