Sign In

VividVisions

69
551
35
Updated: Mar 26, 2024
base modelanime3dstylizedphotonsfw
Verified:
SafeTensor
Type
Checkpoint Merge
Stats
551
Reviews
Published
Mar 24, 2024
Base Model
SDXL Lightning
Hash
AutoV2
CCAC12C607

Merge of Duchaiten-NijiUncen v1Lightning and JuggernautXL V9RDPhotoLightning. Check out both these models, they are amazing.

This model produces stylized images in a variety of styles from drawings to photos. Works well with glowing effects and bright colors. Not the best at nsfw concepts, can be very hit or miss.

With many descriptive visual tags it seems to like to favor 2.5d styles so mess with the strength values and negative prompts to force it into the drawing/painting styles. (This may need more tests to get right, maybe I need to merge in another model to fix this issue)

Sampling method: Euler a, Euler a SGMUniform

Sampling steps: 12

CFG Scale: 1 - 3 (Higher CFG helps with drawn styles w higher contrast. Cfg > 3 can cause artifacting)

ImgtoImg Upscale: Same settings as above

HiRes Upscale: 4x_NMKD-Siax_200k with 2-3 Steps and 0.25 - .35 Denoise + 1.5 - 2 Upscale

Also Check out Lightning Fusion XL as this is my inspiration for this model.