This LoRA of Carmilla is trained for the Flux Dev checkpoint. It was trained against HD screencaps from the Castlevania TV series. Works well for both anime/cartoon depictions as well as photorealistic. For prompt examples check my images: https://civarchive.com/user/thebollo/images. I upload all images with full metadata attached for easy import into A111 or ComfyUI.

In my experience the best results for realism come from using the basic Flux-Dev checkpoint with a CFG between 2.8 and 3.8, and a LoRA strength/weight of 1.2.