Releases: lucidrains/magvit2-pytorch
Releases · lucidrains/magvit2-pytorch
0.1.43
protect against nans for adaptive weight
0.1.42
bgr to rgb format when using cv2
0.1.41
normalize video_to_tensor output, thanks to @iejMac
0.1.40
add image spatial squeeze excites to all residual units
0.1.39
multiscale discriminators should also be controlled by discr_start_af…
0.1.38
add discr_start_after_step for VideoTokenizerTrainer
0.1.37
turn off blur antialiasing in discriminator for now, also fix adaptiv…
0.1.36
handle rgba images edge case
0.1.35
init discriminator with correct channels
0.1.34
handle greyscale edgecase part deux