A newer version of this model is available: nebulette/shared-adaln
A diffusion model. The training follows the self-flow training method with both (t, s) timestamps.
Keywords
- pixel-space model
- fully convolutional model
- Mage Flow's attention block (previously in Coloring Mage)
References
- 2607.19064
- FLUX 3 - Real World Models
- ruwww's waifu diffusion
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support