How do you choose the 'early stoping' parameter ?

valeoai / ADVENT

Adversarial Entropy Minimization for Domain Adaptation in Semantic Segmentation

https://arxiv.org/abs/1811.12833

Other

382 stars 70 forks source link

How do you choose the 'early stoping' parameter ? #3

Closed idealwei closed 5 years ago

idealwei commented 5 years ago

Hi, I am very curious that how you choose the "early stoping" parameter. And it's well known that adv training is not stable, so how you choose your best segmentation model?

tuanhungvu commented 5 years ago

Hi @idealwei, regarding early stopping, we followed the one used in AdaptSegNet. To get the best model, we evaluate intermediate snapshots and keep the best one.