search

Mukosame / Zooming-Slow-Mo-CVPR-2020

Fast and Accurate One-Stage Space-Time Video Super-Resolution (accepted in CVPR 2020)
GNU General Public License v3.0
908 stars 165 forks source link

pytorch2 #73

Closed LaurentBerger closed 5 months ago

LaurentBerger commented 6 months ago

Hi, I tried to translate model for pytorch2.

How can I translate parameter for DCN_sep(nf, nf, 3, stride=1, padding=1, dilation=1, mask=groups) to use deform_conv2d ?