Hugging Face's logo

BlinkDL
/
clip-guided-binary-autoencoder

Model card Files Community
1
BlinkDL commited on
Commit
7c9ec73
•
1 Parent(s): bfea08e

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +3 -1
README.md CHANGED
@@ -4,8 +4,10 @@ license: apache-2.0
4
 
5
  This model can encode 224x224 RGB image into 28x28x13bit (1274 bytes) latent. The compression rate is 28x28x13/(224x224x24)=1/118, or 0.203 bpp (same as VQGAN_f8_8192).
6
 
 
 
7
  12M params for Encoder + Decoder. Trained on LAION-Aesthetics V2 5+ for 238M images.
8
 
9
- Now with 50M and 200M params checkpoints too :) Check the files.
10
 
11
  Guided by https://huggingface.co/laion/CLIP-ViT-B-32-laion2B-s34B-b79K (it's great. better than OpenAI CLIP B/32) and https://github.com/dingkeyan93/DISTS. No GAN loss.
 
4
 
5
  This model can encode 224x224 RGB image into 28x28x13bit (1274 bytes) latent. The compression rate is 28x28x13/(224x224x24)=1/118, or 0.203 bpp (same as VQGAN_f8_8192).
6
 
7
+ Demo: https://huggingface.co/spaces/Blealtan/clip-guided-binary-autoencoder
8
+
9
  12M params for Encoder + Decoder. Trained on LAION-Aesthetics V2 5+ for 238M images.
10
 
11
+ Update: Now with 50M and 200M params checkpoints too :) Check the files.
12
 
13
  Guided by https://huggingface.co/laion/CLIP-ViT-B-32-laion2B-s34B-b79K (it's great. better than OpenAI CLIP B/32) and https://github.com/dingkeyan93/DISTS. No GAN loss.