Missing implementations of ContrastiveImagingAndTabularDatasets

RyanJJP / CHARMS

The code repository for ICML24 paper "Tabular Insights, Visual Impacts: Transferring Expertise from Tables to Images"

12 stars 1 forks source link

Missing implementations of ContrastiveImagingAndTabularDatasets #2

Closed MiHu773 closed 1 week ago

MiHu773 commented 1 month ago

Hello! Thank you for sharing your work. I was looking at the implementation of trainers/pretrain.py file and noticed that there is no implementation of classes:

ContrastiveImagingAndTabularDataset_SUN,
ContrastiveImagingAndTabularDataset_Celeba,
ContrastiveImagingAndTabularDataset_Pawpularity,
ContrastiveImagingAndTabularDataset_Avito,
ContrastiveImagingAndTabularDataset_DVM.

Could you share those classes as well or am I missing something?

RyanJJP commented 2 weeks ago

Hi! Thank you for your attention. These classes can be constructed in the same way as ContrastiveImagingAndTabularDataset_PetFinder. You can refer to ContrastiveImagingAndTabularDataset_PetFinder for more information.

By the way, these classes only affect MMCL methods. They have nothing to do with our CHARMS. You may refer to MMCL (https://github.com/paulhager/MMCL-Tabular-Imaging/tree/main) for these classes.