Dice values are so different.

NicoloPecco / Swin-Functional-Fetal-Brain-Segmentation

Swin-UNETR transformer model for Fetal rs-fMRI brain segmentation

MIT License

0 stars 0 forks source link

Dice values are so different. #1

Open WHLWJL opened 3 days ago

WHLWJL commented 3 days ago

Hello, I would like to ask why there is such a big difference between validation and test Dice values. The validation set and test set are the same dataset.

NicoloPecco commented 3 days ago

Hi @WHLWJL, thanks for your comment! I'm not sure which specific values you're referring to. The validation set is used to fine-tune the model during training, while the test set is completely unseen by the model during training. Therefore Validation set performance might appear higher compared to the test set. Hope this clears up your doubt!

WHLWJL commented 2 days ago

Thank you for your reply.