Position Embedding Needs an Independent Layer Normalization

The dataset used in the paper is not explicitly described, but it is mentioned that the authors analyzed the input and output of each encoder layer in Vision Transformers (VTs) using reparameterization and visualization.

Data and Resources

Cite this as

Runyi Yu, Zhennan Wang, Yinhuai Wang, Kehan Li, Yian Zhao (2024). Dataset: Position Embedding Needs an Independent Layer Normalization. https://doi.org/10.57702/3lzeok10

DOI retrieved: December 2, 2024

Additional Info

Field Value
Created December 2, 2024
Last update December 2, 2024
Author Runyi Yu
More Authors
Zhennan Wang
Yinhuai Wang
Kehan Li
Yian Zhao
Homepage https://github.com/Ingrid725/LaPE