Downloads · 30 days
6
6% of all-time downloads
Ilustrious-Nova-V4.0-Trinity
Bl4ckSpaces/Ilustrious-Nova-V4.0-Trinity
Ilustrious-Nova-V4.0-Trinity is a text-to-image model from Bl4ckSpaces. Use it when you need an image from a text prompt. It is set up for diffusers. The card lists the license as openrail++.
Ilustrious-Nova V4.0 is the result of a "Trinity Merge" technique designed to stabilize high-resolution generation (1536px+) while maintaining the perfect Artstyle/Anatomy balance of V3.6.
All-time downloads
95
Public
Repo size
6.9 GB
Likes
0
Public
Repo makeup
Click a slice to open those files.
.safetensors6.9 GB · 100%
At a glance
- Task
- Text-to-Image
- Library
- diffusers
- License
- openrail++
- Access
- Public
- Created
- Jan 26, 2026
- Updated
- Jan 26, 2026
- SHA
- 4ce4d494
Model card
From the Hugging Face model README
Ilustrious-Nova-V4.0-Trinity (The Final Form)
Ilustrious-Nova V4.0 is the result of a "Trinity Merge" technique designed to stabilize high-resolution generation (1536px+) while maintaining the perfect Artstyle/Anatomy balance of V3.6.
The Problem with V3.6
While V3.6 had excellent anatomy, it suffered from "tensor drift" at extreme resolutions, causing mutations.
The V4.0 Solution (Selective Re-grounding)
We took Nova V3.6 as the center and "grounded" its tensors back to its parents specific blocks:
- Input/Middle Blocks: Mixed Nova V3.6 (60%) with IllumiYume (40%). This forces the skeletal structure to adhere to IllumiYume's stability, fixing high-res mutations.
- Output Blocks: Mixed Nova V3.6 (70%) with NoobAI (30%). This reinforces the crisp outlines and cel-shaded style, preventing the IllumiYume mix from washing out the colors.
Result
The ultimate V-Prediction Anime Model. Stable at 1536px, Pose-perfect, and Stylistically Sharp.
Recommended Settings
- Sampler: Euler a / DPM++ 2M SDE
- CFG: 5.0 - 7.0
- Steps: 28+
- V-Prediction: ENABLED
- Task
- Text-to-Image
- Library
- diffusers
- License
- openrail++
- Created
- Jan 26, 2026
- Updated
- Jan 26, 2026