search

LucasFidon / GeneralizedWassersteinDiceLoss

Official implementation of the Generalized Wasserstein Dice Loss in PyTorch
BSD 3-Clause "New" or "Revised" License
80 stars 8 forks source link

Implementation of Distributionally Robust Optimization (DRO) (Question?) #1

Open saruarlive opened 3 years ago

saruarlive commented 3 years ago

Hi Lucas, As mentioned in the paper, https://arxiv.org/pdf/2011.01614.pdf, you have used DRO (optimization) along with GWD loss. Is your DRO optimization implementation code available?

I am looking forward to hearing from you.

LucasFidon commented 3 years ago

Hi @saruarlive,

Thank you for your question.

My code for the DRO optimization is not available yet. I am planning to make it available in the future but I don't know when it will be ready.

Please note that in the paper https://arxiv.org/pdf/2011.01614.pdf, the generalized Wasserstein Dice loss and DRO are used independently for training two different networks.

I have changed the status of this issue to 'question' as it is not directly related to the generalized Wasserstein Dice loss.