Some questions about the paper

[keaiwangdao]

Thank you very much for your excellent work, it is a great inspiration to me!

In Figure 3 (a), shouldn't the window attention outputs change more and the residuals change less? Does the yellow line represent the residuals, and the blue line represents the window attention outputs?

[HarrisonWu42]

In fact, Fig. 3 (a) shows the difference between window attention and full attention. The yellow line represents window attention, and the blue line represents full attention.

[Probe100]

@keaiwangdao Sorry for the late response. We have updated our paper and added the legend. Please check the latest version at https://openreview.net/pdf?id=51HQpkQy3t