My first attempt at a size-oriented (giantess, BE, AE) LoRA. Trained off of hand-curated AI-generated images and prompts. Here is the prompt format that I used for training, with key terms separated by commas:

• badonkers - always the first tag

• one or more category tags:

  • giantess - A giant woman that cannot fit inside of a room

  • minigiantess - A large woman who can still fit inside of a room

  • hugebreasts - A woman whose breasts are larger than normal

  • breastexpansion - A woman whose breasts are growing into hugebreasts

  • hugeass - A woman whose ass is larger than normal

  • assexpansion - A woman whose ass is growing into a hugeass

• hair color (blonde, brunette, redhead, purple hair, etc.)

• hairstyle, if not long hair (ponytail, pigtails, short hair)

• clothing, starting with color, in order:

  • upper body (e.g., blue camisole)

  • lower body (e.g., black leggings)

  • feet (e.g., bare feet)

• action (standing, sitting, bent over, lying on bed, etc.)

• shot type (full body shot, upper body shot, rear shot, side-on shot)

• location (e.g., gym)

• any additional details

A couple of notes:

• It has trouble with nipples. You may want to look into chaining it with another LoRA to improve this.

• This is still a work-in-progress. I'm a newbie using FLUX and ComfyUI, so I hope to update this with better LoRAs later on.