search

caoql98 / DPLQ

Domain Prompt Learning with Quaternion Networks （CVPR2024 Highlight)
99 stars 12 forks source link

How to get the datasets? #1

Open IHikari opened 3 months ago

IHikari commented 3 months ago

Thanks a lot for your work! How to get the datasets?

caoql98 commented 3 months ago

Hi， thanks for your attention! All datasets are open-sourced online. The medical datasets all are available in the kaggle, while the remote sensing datasets could be obtained by google or the mentioned references on the paper.

IHikari commented 3 months ago

Thank you for your reply!But what I mean is, did you segment these datasets like CoOp did?If you do, could you share the JSON file of the split dataset?Thanks again for your work.I am looking forward to your reply.

caoql98 commented 3 months ago

Hi, my code is based on the MaPLe, and the datasets would be automatically split, you could directly utilize the generated json files.

IHikari commented 3 months ago

Thanks a lot

IHikari commented 3 months ago

Does the directory structure of the RSICD dataset need to be organized like that of the Eurosat dataset? Do I need to put each image of the same category in the same folder?

caoql98 commented 1 month ago

Yes, we follow the same logic.