About multi-scale branch

[go-ahead-maker]

Hi authors,
In MSCA, the learned multi-scale features are aggregated via adding opt rather than concatenating. Is this for the purpose of achieving a better trade-off between performance and complexity? (concat may be more expensive in terms of GPU memory or slower speed).

[MenghaoGuo]

We find add can achieve a satisfactory result and do not try to use concat operation. If you are interested in it, you can have a try. Of course, concat requires more parameters and computation.