readme update

This commit is contained in:
avtc
2026-04-24 23:55:41 +03:00
parent 8b4b46325b
commit 24248cfe05
2 changed files with 11 additions and 2 deletions
+10 -2
View File
@@ -9,19 +9,27 @@
## Implemented tiling modes
- [x] Hexagon
- [x] Rectangular (toroidal — right edge wraps to left, bottom wraps to top)
- [x] None (normal generation)
## Supported models
### UNet-based
- [x] Stable Diffusion 1.5 (also 1.4)
- [x] Stable Diffusion 2.1 (also 2.0)
- [x] Stable Diffusion XL (SDXL)
### DiT-based (via toroidal attention patching)
- [x] FLUX.2 — tested on flux.2-klein-4b
- [x] Qwen Image — tested on qwen-image-2512 (Q6 GGUF and fp16)
DiT models are automatically detected and use toroidal attention instead of latent padding for seamless tiling.
## TODO
- More tiling modes
- Optimize VAE decode (first pass is very slow)
- Support DiT based models (SD3, PixArt-Σ, FLUX.1)
## Credits
File diff suppressed because one or more lines are too long