0.5.9

fix bug with state dictionary for vqgan vae

0.5.8

fix bug with state dictionary for vqgan vae

0.5.7

write function to convert dataset for raw videos to video indices, an…

…d allow NUWA and NUWA Video + Audio to accept video indices

0.5.6

write function to convert dataset for raw videos to video indices, an…

…d allow NUWA and NUWA Video + Audio to accept video indices

0.5.5

make sure VGG for perceptual loss is not saved in the VQGan-vae state…

… dictionary

0.5.4

complete working dataset class for specifying path to folder of video…

… as gifs + corresponding captions as txt, and show usage in readme for NUWATrainer

0.5.3

finish cycle dilation for sparse 2dna audio attention

0.5.2

todo

0.5.1

complete video + audio generation

0.5.0

complete video + audio generation