0.14.0

complete the gaussian diffusion with hybrid loss (learned variance) a…

0.12.1

calculate noise schedule with float64 for numerical accuracy

0.12.0

remove convnext blocks, they do not work well, validated in video dif…

0.11.2

groupnorm groups should be actually configurable

0.11.1

allow for configuring expansion factor in convnext

0.11.0

bring back resnet blocks, make convnext blocks an experimental option

0.10.1

give an initial conv

0.10.0

make sure time embedding dimension is kept at 4 x dimension (thanks @…

0.9.2

fix cosine beta schedule, thanks to @Zhengxinyang

0.9.1

fix bug, thanks to @jihoonerd