| license: apache-2.0 | |
| This model can encode 224x224 RGB image into 28x28x13bit latent. The compression rate is 28x28x13/(224x224x24)=1/118, or 0.203 bpp. | |
| 12M params for Encoder + Decoder. Trained on LAION-Aesthetics V2 5+ for 60M images. | |
| (still training. final checkpt will be better) |