Balancing Image Compression and Generation with Bootstrapped Tokenization

arXiv:2606.05552v1 Announce Type: cross Abstract: Despite progress in image tokenization, standard methods encode redundant information by mixing all granularities within each token, thus redundancy persists between tokens. The mix of information of different granularity also complicates the training of generators. This paper introduces...

Læs hele artiklen hos kilden.