Upload README.md with huggingface_hub
Browse files
README.md
CHANGED
|
@@ -11,6 +11,8 @@ library_name: mdiffae
|
|
| 11 |
|
| 12 |
# mdiffae-v2
|
| 13 |
|
|
|
|
|
|
|
| 14 |
**mDiffAE v2** — **M**asked **Diff**usion **A**uto**E**ncoder v2.
|
| 15 |
A fast, single-GPU-trainable diffusion autoencoder with a **96-channel**
|
| 16 |
spatial bottleneck and optional PDG sharpening.
|
|
|
|
| 11 |
|
| 12 |
# mdiffae-v2
|
| 13 |
|
| 14 |
+
> **DEPRECATED** — This model is superseded by [**SemDisDiffAE**](https://huggingface.co/data-archetype/semdisdiffae), which offers better reconstruction quality, better downstream diffusion convergence,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.
|
| 15 |
+
|
| 16 |
**mDiffAE v2** — **M**asked **Diff**usion **A**uto**E**ncoder v2.
|
| 17 |
A fast, single-GPU-trainable diffusion autoencoder with a **96-channel**
|
| 18 |
spatial bottleneck and optional PDG sharpening.
|