mlpc-ucsd/Patch-DM

training image size

Closed this issue · 1 comments

Thank you so much for sharing the amazing work! I have a question regarding training image resolution, in img2lmdb.py I only see it converting size up to 1024. Do you support even higher resolutions such as 2k or 4k and patchify? My dataset mostly 3k resolution images. Thank you!

We didn't explore further above 1024. But yeah, the method theoretically practical for any resolution using the same method to patchify. And you may use a different base model architecture of 256or512 to improve the performance by reducing total patch numbers for higher resolutions.