Closed Can-Zhao closed 1 month ago
Thanks for the pr. It seems that this PR contains a variety of transform list dependent on mri or ct, and if it's only specific in vae training, I feel that we don't need not define classes or functions for these transforms separately. But if others don't have comments, we can do that in the first version.
Yes, multi-modal training is used only for VAE for now. It might be used for diffusion model in the future.
Fixes # .
Description
MAISI VAE transforms
Checks
./figure
folder./runner.sh -t <path to .ipynb file>