Tweeted By @ak92501

Fine-tuning Image Transformers using Learnable Memory
abs: https://t.co/EysBcFM7xa

propose augmenting Vision Transformer models with learnable memory tokens. Model adapts to new tasks, using few parameters, while optionally preserving its capabilities on previously learned tasks pic.twitter.com/gJod2r1hSv

— AK (@ak92501) March 30, 2022