An old model I trained with Replicate, even though its captions were mainly just SD/Booru tags. Oh well!
Uses 53 images in dataset.
SD/Booru tags in place of captions, including some styles.
HypnoFluxDev.safetensors