microsoft / ProDA

Prototypical Pseudo Label Denoising and Target Structure Learning for Domain Adaptive Semantic Segmentation (CVPR 2021)
https://arxiv.org/abs/2101.10979
MIT License
286 stars 44 forks source link

what is the warm-up model? #12

Closed liuyk97 closed 3 years ago

liuyk97 commented 3 years ago

In the training processing of stage1, I need to download warm-up model (43.4 mIoU), and resume from it, but there seems no description of what dataset is it pretrained on. Is it pretrained on source dataset? thank you

panzhang0212 commented 3 years ago

It is pretrained by https://github.com/microsoft/ProDA/blob/9ba80c7dbbd23ba1a126e3f4003a72f27d121a1f/models/adaptation_modelv2.py#L143 function

KaiiZhang commented 3 years ago

It is pretrained by

https://github.com/microsoft/ProDA/blob/9ba80c7dbbd23ba1a126e3f4003a72f27d121a1f/models/adaptation_modelv2.py#L143

function Thanks for your great work! What initialization weights do you use in the warmup stage? Could you please upload it. Thank you very much.