search

DearCaat / MHIM-MIL

[ICCV 2023 Oral] Multiple Instance Learning Framework with Masked Hard Instance Mining for Whole Slide Image Classification
49 stars 3 forks source link

EarlyStop for C16 #4

Closed HHHedo closed 6 months ago

HHHedo commented 6 months ago

Hi, Thanks for your great work. I have a question about the setting of the early stop in Camelyon16. Is there any validation set in the 3-fold CV?

Thanks!

DearCaat commented 6 months ago

Hello, thanks for attention EarlyStop for C16: max epoch: 200, stop epoch: 120, patience: 30, code. Validation in C16: No, cause the count of slides in C16 is so limited (there are only 398 in total, and only ~20 slides will be used for validation).

HHHedo commented 6 months ago

Hello, thanks for attention EarlyStop for C16: max epoch: 200, stop epoch: 120, patience: 30, code. Validation in C16: No, cause the count of slides in C16 is so limited (there are only 398 in total, and only ~20 slides will be used for validation).

Thanks for your quick response. Since no validation set is used, which metric will the EarlyStop depend on?

DearCaat commented 6 months ago

Just the test set, see codes. This is common in survival prediction tasks [1], also due to the small number of samples. [1] Multimodal Co-Attention Transformer for Survival Prediction in Gigapixel Whole Slide Images. ICCV 2021