Download
1 variant available
fp16 SafeTensor
FantasyBlendV3.safetensors
Half precision, best balance (pruned) • 1.99 GB
Verified: 3 years ago
clearvae 2.3 • ClearVAE_V2.3_fp16.safetensors
5100 1 2 3 4 5 6 7 8 90 1 2 3 4 5 6 7 8 90 1 2 3 4 5 6 7 8 9
(58)
Jan 23, 2024
Partially trained partially merged model. Can do a lot more now and many things are clearer. Style changed a fair bit and is less consistent.
Recommend using V2 inpainting for ADetailer.
Can kinda do furry images I think. There were many furry models merged in.
Show more

4.5K0 1 2 3 4 5 6 7 8 9.0 1 2 3 4 5 6 7 8 9K
46.7K0 1 2 3 4 5 6 7 8 90 1 2 3 4 5 6 7 8 9.0 1 2 3 4 5 6 7 8 9K
Recommended Usage
Sampler: Euler a | 50 Steps
Hires, Fix: 4x Ultrasharp
30 Steps | 0.4 Denoise | 2x Upscale
ADetailer (default settings):
1st Model: face_yolov8.pt (leave prompt empty) - (V2 Inpainting model for V2)
2nd Model: mediapipe_face_mesh_eyes_only (leave prompt empty) - (V2 Inpainting model for V2)
FreeU (default settings)
Also recommend using Regional Prompter with Prompt-Ex for vertical or square images, or Matrix 1,1,1 (columns) for landscape.
FreeU: https://github.com/ljleb/sd-webui-freeu

