Maximizing the Position Embedding for Vision Transformers with Global Average Pooling.

Wonjun Lee, Bumsub Ham, Suhyun Kim

Browse the full AAAI paper archive.