about Train "Image-level" Adaptation for Semantic Segmentation

[amiltonwong]

Hi, @jhoffman ,

Could you also provide the code about "Train Image-level Adaptation for Semantic Segmentation" as mentioned in CyCADA paper? The code found in this repo. only shows feature-level domain adaptation (ADDA) method.

THX!

[xiaoqianjia]

In the train_fcn_adda.py, it's about Segmentation part in Cycada. But, I cannot reproduce the results.