med-air/Endo-FM

CVC-12k training details

Closed this issue · 2 comments

Hi! Thanks for your great work!
Regarding how to fine-tune on CVC-12k dataset for segmentation, as the CVC-12k dataset contains only 612 labeled frames, how is the train, val, test data split? It would be helpful if you can link data split code. Thanks in advance!

Hi, @TianCuteQY , thanks for your interest!
The split is randomly produced, regarding the video case level.

Understood. Thanks for your reply!