Some issues #7
Description
Thank you so much for providing such a valuable perspective. I have some doubts about the paper can you please answer the following.
In Table III in section 3.2, I see that e.g. BYOL is not much different in time from your method in the last row. What is the significance of reducing the epoch but not the time?
The second point is that adding multiple positive samples to an image in the data enhancement phase will increase the hardware requirements, is there any solution considered in this regard?