[Question] about CNN backbone

[CaptainEven]

Is it possible to replace a ViT backbone with a regular CNN backbone like Resnet?

[gaopengpjlab]

You can check the following hybrid conv-transformer MAE architecture for your interest.

https://github.com/Alpha-VL/ConvMAE

[keyu-tian]

Is it possible to replace a ViT backbone with a regular CNN backbone like Resnet?

Perhaps this is what you're looking for: MAE on standard ResNets or ConvNeXts: "Designing BERT for Convolutional Networks: Sparse and Hierarchical Masked Modeling"

[CaptainEven]

Thanks for the advice!