Upload README.md with huggingface_hub
Browse files
README.md
CHANGED
|
@@ -10,6 +10,8 @@ library_name: irdiffae
|
|
| 10 |
|
| 11 |
# data-archetype/irdiffae-v1
|
| 12 |
|
|
|
|
|
|
|
| 13 |
**iRDiffAE** — **iR**epa **Diff**usion **A**uto**E**ncoder.
|
| 14 |
A fast, single-GPU-trainable diffusion autoencoder with spatially structured
|
| 15 |
latents for rapid downstream model convergence. Encoding runs ~5× faster than
|
|
|
|
| 10 |
|
| 11 |
# data-archetype/irdiffae-v1
|
| 12 |
|
| 13 |
+
> **DEPRECATED** — This model is superseded by [**SemDisDiffAE**](https://huggingface.co/data-archetype/semdisdiffae), which offers better reconstruction quality, better downstream diffusion convergence, and slightly faster inference.
|
| 14 |
+
|
| 15 |
**iRDiffAE** — **iR**epa **Diff**usion **A**uto**E**ncoder.
|
| 16 |
A fast, single-GPU-trainable diffusion autoencoder with spatially structured
|
| 17 |
latents for rapid downstream model convergence. Encoding runs ~5× faster than
|