Lila [wip]

multi-task vision model(supposed end product0

(...and a tiny/partial replication of the Apple 4M-21 Any-to-Any vision model)

This repo also contains code[pytorch] implementations of various vision models like the vision transformer, masked autoencoder, etc.