import torch
from diffusers import DiffusionPipeline
# switch to "mps" for apple devices
pipe = DiffusionPipeline.from_pretrained("Bl4ckSpaces/Ilustrious-Nova-V3.6-Inverse-TIES", dtype=torch.bfloat16, device_map="cuda")
prompt = "Astronaut in a jungle, cold color palette, muted colors, detailed, 8k"
image = pipe(prompt).images[0]Ilustrious-Nova-V3.6-Inverse-TIES (V3.6 Inverse TIES)
The Anatomy Transplant Edition. This version returns to the stable TIES (V3.1 Logic) method but swaps the roles to prioritize NoobAI's artstyle.
The Recipe
- BASE (Canvas): NoobAI XL v1.0 (Ensures the output looks like NoobAI).
- DONOR (Injection): IllumiYume v3.5 (Provides the skeletal structure and pose data).
Method: Surgical TIES
We calculated the delta (difference) of IllumiYume relative to NoobAI, filtered out the noise, and injected the structural data strongly into the Input Blocks while protecting the Output Blocks (Style) of NoobAI.
Recommended Settings
- Sampler: Euler a / DPM++ 2M SDE
- CFG: 5.0 - 7.0
- Steps: 28+
- V-Prediction: ENABLED
- Downloads last month
- 27