Question about the Size of the Cache Model.

[TitleZ99]

Thanks for sharing such an enlightening work, I would like to ask where the 16-shot training set is reflected in the code implementation. The function shown in the figure does not have a sampling process. Looking forward to your reply！