| license: apache-2.0 | |
| pipeline_tag: unconditional-image-generation | |
| tags: | |
| - pixel-space | |
| Keywords: | |
| - pixel-space model | |
| - single, shared AdaLN block | |
| - min-SNR weighting | |
| - 2D RoPE with center at origin | |
| - position embedding with both cropped and non-square images | |
| Source data: | |
| - 300k images at 512x512 resolution | |
| References | |
| - 2303.09556 | |
| - 2503.10618 | |
| - 2511.13720 | |
| - [ruwww's waifu diffusion](https://ruwwww.github.io/al-folio/blog/2026/waifu-diffusion/) |