Can someone reproduce the results in the paper and provide some simple guidance?

[jacktkk]

e.g., step1 train.py step2 test.py step3 evalute.py

[wenjie710]

We fixed some bugs and run the code. However, the CD we reproduced is quite different, much larger than that claimed in paper.

[xiaolul2]

I have the same question. Even if I set the threshold as [0.2, 0.5, 1]

[luca-888]

We fixed some bugs and run the code. However, the CD we reproduced is quite different, much larger than that claimed in paper.

Is CD the only parameter you changed?What order of magnitude should be set to make the effect better?Thanks a lot.