Mobile-Former: Bridging MobileNet and Transformer

🤖 Plain-English Summary

We present Mobile-Former, a parallel design of MobileNet and transformer with a two-way bridge in between. Additionally, we build an efficient end-to-end detector by replacing backbone, encoder and decoder in DETR with Mobile-Former, which outperforms DETR by 1.3 AP but saves 52% of computational cost and 36% of parameters.

🔑 Key Findings

• This structure leverages the advantages of MobileNet at local processing and transformer at global interaction.
• And the bridge enables bidirectional fusion of local and global features.
• Different from recent works on vision transformer, the transformer in Mobile-Former contains very few tokens (e.g.

🔗 Related Resources