search

harsh-99 / SCL

Implementation of "SCL: Towards Accurate Domain Adaptive Object Detection via Gradient Detach Based Stacked Complementary Losses"
89 stars 10 forks source link

Your implementation of domain adaptive faster rcnn performs better than paper values, what might be the reason? #24

Open TADanton opened 3 years ago

TADanton commented 3 years ago

Hi Zhiqiang,

I tried to run your re-implementation of domain adaptive faster rcnn in adaptation from pascal voc to clipart, and I found it reached much better mAP (30 mAP) than the value reported in paper "stong weak distribution alignment" paper. (19.8 mAP)

May I ask what's your opinion on this? Why is your re-implementation much better? Or did I probably ignore some details in training script and misuse them? (I used your trainval_net_dfrcnn.py, and didn't use any rendered datasets).

Thanks a lot for any possible help in advance!

Best, Anton

harsh-99 commented 3 years ago

Hey, It should not be the case. There must be some bug in the your code leading to this issue.