0.0.50

randomly pick a frame, as allowed by the masking, for both the discri…

0.0.49

just make sure unconditional and conditional video sampling during tr…

0.0.47

make some progress on phenaki trainer, allowing for all four scenario…

0.0.45

allow to sample on more than one text, also fix phenaki trainer to sa…

0.0.44

unconditional video can now be sampled during training

0.0.43

when training cvivit on video, save reconstructed tensors as gifs

0.0.42

fix all issues with cvivit training, making sure it can be trained on…

0.0.41

allow for turning off training EMA version of cvivit

0.0.40

fix issue in ema-pytorch

0.0.39

address https://github.com/lucidrains/phenaki-pytorch/issues/21