Evaluation code

[ptrnn]

Hello, thanks for your excellent work. I was wondering whether you could release the evaluation code, especially the one related to table 4. Did you follow Pix3d to calculate EMD and CD? Though you didn't compare your work with 3d-LMNet, I've tried to evaluate their official pretrained model and found that the results lies in x 10^-2 scale instead of 10^-3 scale as shown in table 4 in your paper. I guess you may remove the square root when calculating CD, right?