AffineTok

AI research focuses on efficient visual token processing for transformers

Researchers are developing new methods to optimize the use of visual tokens in AI models, particularly for vision transformers and vision-language models. These approaches aim to reduce computational costs and improve i…