Questions on SmoothNet paper Table 3 experiment

[ylchiangsogood]

Hi @ailingzengzzz
I hope you don’t mind me asking, but I just wanted to clarify the evaluation setting for Table 3.
Were the results obtained using the same model across all datasets?
In my experiments, I used the h36m_fcn_3D model for inference on 3DPW and Human3.6M, and the pw3d_spin_3D model for MPII.
Based on this, I obtained the results shown in the table.
This made me wonder whether different models were used for each dataset.
Please feel free to correct me if I’m mistaken, but if that’s the case, I’m concerned the comparison might not be entirely fair.
Taking VIBE as an example,