search

valeoai / ADVENT

Adversarial Entropy Minimization for Domain Adaptation in Semantic Segmentation
https://arxiv.org/abs/1811.12833
Other
382 stars 70 forks source link

About implementation of AdvEnt+MinEnt #7

Closed changliu816 closed 5 years ago

changliu816 commented 5 years ago

Hi,

I implemented the AdvEnt+MinEnt (45.5 in paper) by combining advEnd and MinEnd together, but got 42.48 % instead.

For GTA5-> Cityscapes: I saw the result for MinEnt + ER in paper. My question is that do you use +ER or class prior techniques with AdvEnd+MinEnt experiment ?

Could you please briefly clarify how you implement AdvEnd+MinEnt experiment ?

Best, Chang

tuanhungvu commented 5 years ago

Hello @changliu816, we have not jointly trained AdvEnt + MinEnt in the paper (still you can try). The reported number is of the ensemble of the two models (see Section 4.2). To your second question, the MinEnt model used in this ensemble was (MinEnt + ER) in Table 1. Best, T-H