0.0.18

0.0.18

0.0.17

allow for videos with rectangular frames, for @gmegh

0.0.16

also accept images and treat them as single framed videos

0.0.15

allow for co-training of images and video in the same batch

0.0.14

allow cvivit to be trained with variable lengthed videos or single im…

0.0.11

allow for variable lengthed video / images cotraining in cvivit by pa…

0.0.10

port over the discriminator used in stylegan2 for cvivit training

0.0.9

oops

0.0.8

make sure critic is inferenced without grad and eval is turn on

0.0.7

make it absurdly simple