import torch
from diffusers import DiffusionPipeline
# switch to "mps" for apple devices
pipe = DiffusionPipeline.from_pretrained("Bl4ckSpaces/SpacesFusion_V1", dtype=torch.bfloat16, device_map="cuda")
prompt = "Astronaut in a jungle, cold color palette, muted colors, detailed, 8k"
image = pipe(prompt).images[0]SpacesFusion_V1
SpacesFusion_V1 is a high-precision SLERP merge focused on preserving the superior anatomy of NoobAI XL while injecting the high-resolution capabilities and newer dataset knowledge (2025 tags) from IllumiYume V3.5.
Composition Strategy (The "Nuclear" Recipe)
This model was merged using a specific block-weighted SLERP technique to ensure V-Prediction stability.
- Base Architecture: NoobAI XL v1.0 (Dominant Structure)
- Detail Injector: IllumiYume V3.5 (Dataset & Texture)
Block Weights
- Input Blocks (Anatomy/Structure): Highly biased towards NoobAI to prevent anatomical degradation.
- Middle Blocks: Balanced mix.
- Output Blocks (Texture/Lighting): Progressive injection of IllumiYume parameters to allow for higher resolution scaling (1536px+) and newer art style rendering.
Recommended Settings
- Resolution: 1024x1024 up to 1536x1536 (Native Support)
- CFG: 5-7
- Sampler: Euler a / DPM++ 2M SDE Karras
- V-Prediction: ENABLED (Critical)
Created by Bl4ckSpaces using Automated Evolutionary SLERP.
- Downloads last month
- 2