search

Gorilla-Lab-SCUT / DADA-AAAI2020

Code release for Discriminative Adversarial Domain Adaptation (AAAI2020).
117 stars 14 forks source link

Question about the training #4

Open Carbord opened 4 years ago

Carbord commented 4 years ago

It's an impressive work and I have tried to use this code to reproduce the results in the paper. But when using the default training code such as rhe "a2w" and "d2w" setting in the "run_official31_pratial.sh", I got relatively low results. So I wonder if any settings I may ignore and really look forward to your kind reply. Thank you. result

zhouyaqian666 commented 4 years ago

I also encountered the same problem, I have tried to use this code to reproduce the results in the paper. But when using the default training code setting in the "run_official31_pratial.sh", I got relatively low results. So I wonder if any settings I may ignore and really look forward to your kind reply. Thank you.

EhsanEI commented 3 years ago

I faced the same issue. Has anyone found a solution? Thanks.

EhsanEI commented 3 years ago

In case anyone else will have this problem, the solution for me was removing the images folder. So e.g. instead of "amazon/images/back_pack", the correct path for the backpack class is "amazon/back_pack" (similar to the example in Readme file).