Do you reproduce with the similar accuracy as the paper?If you had done it, could you impart relevant training skills?

[zengsmallqiang]

[sshkhr]

I am still training and update with results once I have them. Please leave this issue open in the meanwhile for others' reference. Thanks a lot

[NimaM96]

Hi Dr Shekhar, have you been able to reproduce their results ?

[prismformore]

This implementation is good, however, failed to reproduce the result.

[NimaM96]

Have you been able to outperform the baseline model (deep-zero padding)?

[sshkhr]

Hi all, I haven't been able to reproduce the results yet. Mostly because 2000 epochs take extremely long for the SYSU dataset. I will run it once after ICCV deadline and upload the results and trained models in April second week hopefully