ActiveMLP: An MLP-like Architecture with Active Token Mixer

11 Mar 2022  ·  Guoqiang Wei, Zhizheng Zhang, Cuiling Lan, Yan Lu, Zhibo Chen ·

This paper presents ActiveMLP, a general MLP-like backbone for computer vision. The three existing dominant network families, i.e., CNNs, Transformers and MLPs, differ from each other mainly in the ways to fuse contextual information into a given token, leaving the design of more effective token-mixing mechanisms at the core of backbone architecture development. In ActiveMLP, we propose an innovative token-mixer, dubbed Active Token Mixer (ATM), to actively incorporate contextual information from other tokens in the global scope into the given one. This fundamental operator actively predicts where to capture useful contexts and learns how to fuse the captured contexts with the original information of the given token at channel levels. In this way, the spatial range of token-mixing is expanded and the way of token-mixing is reformed. With this design, ActiveMLP is endowed with the merits of global receptive fields and more flexible content-adaptive information fusion. Extensive experiments demonstrate that ActiveMLP is generally applicable and comprehensively surpasses different families of SOTA vision backbones by a clear margin on a broad range of vision tasks, including visual recognition and dense prediction tasks. The code and models will be available at https://github.com/microsoft/ActiveMLP.

PDF Abstract

Results from the Paper


Task Dataset Model Metric Name Metric Value Global Rank Result Benchmark
Semantic Segmentation ADE20K ActiveMLP-L(UperNet) Validation mIoU 51.1 # 42
Params (M) 108 # 12
Object Detection COCO minival ActiveMLP-B (Cascade Mask R-CNN) box AP 52.3 # 36
Image Classification ImageNet ActiveMLP-L Top 1 Accuracy 83.6% # 222
Number of params 76.4M # 558
GFLOPs 12.3 # 252
Image Classification ImageNet ActiveMLP-T Top 1 Accuracy 82% # 339
Number of params 27.2M # 416
GFLOPs 4 # 155

Methods


No methods listed for this paper. Add relevant methods here