MLP-Mixer: An all-MLP Architecture for Vision

Open paper
Benchmark Model Rank Results
image-classification-on-imagenetMixer-H/14 (JFT-300M pre-train)#65Top 1 Accuracy: 87.94%
image-classification-on-imagenetViT-L/16 Dosovitskiy et al. (2021)#235Top 1 Accuracy: 85.3%
image-classification-on-imagenetMixer-B/16#855Top 1 Accuracy: 76.44%Number of params: 46M
image-classification-on-imagenet-realMixer-H/14- 448 (JFT-300M pre-train)#20Accuracy: 90.18%Params: 409M
image-classification-on-imagenet-realMixer-H/14 (JFT-300M pre-train)#30Accuracy: 87.86%Params: 409M
image-classification-on-omnibenchmarkMLP-Mixer#17Average Top-1 Accuracy: 32.2

Rank counts only results with a code link.