Back to researchers
Piotr Dollár
Masked autoencoders for vision (MAE)
Co-authored MAE: a strong template for scalable self-supervised vision pretraining.
Highlights
VisionSelf-supervisedTransformers
Focus: Masked autoencoders for vision (MAE)
Why it matters: Co-authored MAE: a strong template for scalable self-supervised vision pretraining.
Research Areas
VisionSelf-supervisedTransformers