# High-resolution human images

Sapiens is a vision Transformer model pre-trained on 300 million 1024×1024 resolution human images, supporting high-resolution inference and real-world scenario generalization.