# 1B Parameter Vision Transformer

A 1-billion-parameter Vision Transformer model trained via masked autoencoder self-supervised learning on 2 billion web images, capable of learning visual representations without language supervision.