V2 has a revised dataset no longer using in-game Overwatch photos and has been replaced with Blender style images. Prompt adherence should be a bit better.
PLEASE NOTE: "3D" may be needed depending on the character or setting.
This uses my trained OverBlend lora/checkpoints from Illustrious mixed in with Base Anima! I no longer used preview 3 like the previous checkpoint models.
Works very well with natural language.
YOU NEED:
qwen base encoder & qwen image vae
All gallery images should have metadata so you can drag and drop to see the exact things you need along with prompts and settings.
SETUP:
Steps: 30 - 50
CGF: 4 - 5
Sampler: Eular/Eular A (haven't tested others)
Description
Removed and replaced about 1/3 of the images of the original dataset. No longer featuring Overwatch in-game photos and replacing it with Blender style w/ OW characters. Used an upscale on most images for a higher quality dataset. Prompt adherence should be a bit better than V1.