Communications for Statistical Applications and Methods

Indexed in /covered by CAS, KoreaScience & DOI/Crossref:eISSN 2383-4757   pISSN 2287-7843

Fig. 7.

Download original image
Fig. 7. Structure of WRN-n-k: Expanded convolution block has similar structure to a convolution block in . The expanded convolution block has k times wider layer than that of the ResNet convolution block. Wide - dropout block has the same architecture as with widening factor k. N in the bottom is calculated as N = (n ? 4)/6.
Communications for Statistical Applications and Methods 2022;29:161-76 https://doi.org/10.29220/CSAM.2022.29.2.161
© 2022 Communications for Statistical Applications and Methods