v0.0.31

feat: add option to provide context tokens

v0.0.30

feat: add option to change encoder num_blocks

v0.0.29

feat: add with_info during forward in autoencoder

v0.0.28

feat: add option to insert quantizer as diffusion autoencoder bottleneck

v0.0.27

feat: add convout kernels option

v0.0.26

fix: replace prod not supported py3.7

v0.0.25

feat: add option to train upsampler with multiple factors

v0.0.24

fix: add autoencoder import

v0.0.23

feat: context after downsample

v0.0.22

feat: remove unsucessful convmean, provide context only during downsa…