# 1. Vision Transformers (An image is worth 16x16 words)
- Pretrained on 300+ million images
- SOTA on ImageNet (88.55%)
- Added CLS token to patch tokens => responsible for predicting true label

###### tags: `vit`