Hugging Face's logo

LEAF-CLIP
/
OpenCLIP-ViT-H-rho50-k1

Safetensors
clip
Model card Files Community
megaelius commited on
Commit
560cb60
·
verified ·
1 Parent(s): 2623f5b

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +1 -1
README.md CHANGED
@@ -6,4 +6,4 @@ base_model:
6
  - laion/CLIP-ViT-H-14-laion2B-s32B-b79K
7
  ---
8
 
9
- Model finetuned from OpenClip ViT-H-14 with $\rho=50$ and $k=1$.
 
6
  - laion/CLIP-ViT-H-14-laion2B-s32B-b79K
7
  ---
8
 
9
+ ViT-H OpenCLIP initialized from https://huggingface.co/laion/CLIP-ViT-H-14-laion2B-s32B-b79K. The Text encoder is finetuned with the FARE loss using the charmer attack with $\rho=50$ and $k=1$.