Kai Han

理论上可以用，已经有同行把ViG用到了Ophthalmic图像里：https://arxiv.org/pdf/2209.00773.pdf 你这些图像挺大的，直接用ViG的计算量会很大。建议切块大一些，减少node的数量。

Thanks for the correction.

@rsomani95 Good job! Thanks for the contribution.

@jinfagang The simple solution is use the resolution ratio in classification task.

Thanks for the attention. Here we reduce the number of nodes by `reduce_ratio` so as to reduce the computational cost of distance calculation.

Dilation is not always 1 since `idx` is the layer index in `min(idx // 4 + 1, max_dilation)`.

@PanXiebit The `torch.max` is applied at element-wise level, not the vector-wise.

Thanks for you attention on GhostSR. The code will be available upon acceptance of the paper.

Huawei P30Pro, CPU.

I think you have the training example for your dataset. Just replace your network with GhostNet and train it.