Q : Deeplabv3 CITYSCAPES parameters

RViMLab / ECCV2022-multi-scale-and-cross-scale-contrastive-segmentation

Implementation of the methods described in "Multi-scale and Cross-scale Contrastive Learning for Semantic Segmentation", ECCV 2022

27 stars 2 forks source link

Open junggho opened 1 year ago

junggho commented 1 year ago

Hi, Thank you for providing nice work.

I want to know loss weights information of deeplabv3. In addition, what layers did you choose for your backbone feature?

And if you don't mind me asking, may I have the config file for deeplabv3?

Thank you.