nizhenliang

Thank you, sir! Whether the output value is directly FLOPs? Do we need to divide it by 2 to get FLOPs？

Thank you very much! According to the above settings, we still cannot achieve the experimental results in the paper. We would appreciate it if you could release the pre-trained config...

In addition, we found that the above config are consistent with the MobilenetV3 config in MMCLS. Does topformer use the same data augmentation approach as mobilenetv3?

You can remove padding in the load_dataset.

> > You can remove padding in the load_dataset. > > RuntimeError: The size of tensor a (136) must match the size of tensor b (135) at non-singleton dimension 2...

The function of "load_mask" in load_dataset.py is used for load the label.