0.0.31

from my vision transformer experience, dimension of attention head of…

0.0.28

complete contextmanager method for keeping only one unet in GPU durin…

0.0.27

give time tokens a surface area of 2 tokens as default, make it so re…

0.0.26

fix cosine  bbeta schedule, thanks to @Zhengxinyang

0.0.25

add sparse attention layers in between convnext blocks in unet (grid …

0.0.24

same for text encodings for decoder ddpm training

0.0.23

also allow for image embedding to be passed into the diffusion model,…

0.0.22

use some magic just this once to remove the need for researchers to t…

0.0.21

make kernel size and sigma for gaussian blur for cascading DDPM overr…

0.0.20

complete first pass at full cascading DDPM setup in Decoder, flexible…