Shortcuts

ConvMixer

Abstract

Although convolutional networks have been the dominant architecture for vision tasks for many years, recent experiments have shown that Transformer-based models, most notably the Vision Transformer (ViT), may exceed their performance in some settings. However, due to the quadratic runtime of the self-attention layers in Transformers, ViTs require the use of patch embeddings, which group together small regions of the image into single input features, in order to be applied to larger image sizes. This raises a question: Is the performance of ViTs due to the inherently-more-powerful Transformer architecture, or is it at least partly due to using patches as the input representation? In this paper, we present some evidence for the latter: specifically, we propose the ConvMixer, an extremely simple model that is similar in spirit to the ViT and the even-more-basic MLP-Mixer in that it operates directly on patches as input, separates the mixing of spatial and channel dimensions, and maintains equal size and resolution throughout the network. In contrast, however, the ConvMixer uses only standard convolutions to achieve the mixing steps. Despite its simplicity, we show that the ConvMixer outperforms the ViT, MLP-Mixer, and some of their variants for similar parameter counts and data set sizes, in addition to outperforming classical vision models such as the ResNet.

How to use it?

from mmpretrain import inference_model

predict = inference_model('convmixer-768-32_3rdparty_in1k', 'demo/bird.JPEG')
print(predict['pred_class'])
print(predict['pred_score'])

Models and results

Image Classification on ImageNet-1k

Model

Pretrain

Params (M)

Flops (G)

Top-1 (%)

Top-5 (%)

Config

Download

convmixer-768-32_3rdparty_in1k*

From scratch

21.11

19.62

80.16

95.08

config

model

convmixer-1024-20_3rdparty_in1k*

From scratch

24.38

5.55

76.94

93.36

config

model

convmixer-1536-20_3rdparty_in1k*

From scratch

51.63

48.71

81.37

95.61

config

model

Models with * are converted from the official repo. The config files of these models are only for inference. We haven’t reproduce the training results.

Citation

@misc{trockman2022patches,
      title={Patches Are All You Need?},
      author={Asher Trockman and J. Zico Kolter},
      year={2022},
      eprint={2201.09792},
      archivePrefix={arXiv},
      primaryClass={cs.CV}
}
Read the Docs v: latest
Versions
latest
stable
mmcls-1.x
mmcls-0.x
dev
Downloads
epub
On Read the Docs
Project Home
Builds

Free document hosting provided by Read the Docs.