Awesome

spatially_augmented_comic_mtl

Contains the resources for "Spatially Augmented Speech Bubble to Character Association via Comic Multi-Task Learning" - MANPU 2024"

• DCM772 dataset can be reached at this google drive link => https://drive.google.com/file/d/1FGoaGZreSIYU4t01vfcc_L4_kBZu2mj-/view?usp=sharing
• fully trained model checkpoint can be found at => https://drive.google.com/file/d/12hcO89eyfj73hmysvkXvqchTvWbZ1rTu/view?usp=drive_link

TODO: @gsoykan - add citation

TODO: @gsoykan - add paper link

TODO: @gsoykan - add inference sample

TODO: @gsoykan - convert it to a package

TODO: @gsoykan - put example data folder

TODO: @gsoykan - add requirements

TODO: @gsoykan - indicate that this works in cuda only (feel free to open a pr about making the repo work on any device)

TODO: @gsoykan - import ashleve readme and give credits to that

TODO: @gsoykan - after we are done with this

re-update pytorch-lightning, torchmetrics