lucidrains/mlp-mixer-pytorch

Health v2 · maintenance only

Adoption not part of the score

1064 stars · 107 forks observed · 2026-08-28

What it is AI-extracted, prompt v1, taxonomy v1, 2026-08-30, confidence not recorded

A PyTorch implementation of Google AI's MLP-Mixer, an all-MLP architecture for image classification that uses neither convolutions nor attention. It also includes a 3D variant (MLPMixer3D) for video inputs.

Use cases

• implement mlp-mixer architecture in pytorch
• image classification without convolutions or attention
• experiment with all-mlp vision models
• classify video with an mlp-based model
• reproduce the mlp-mixer paper results

When to choose

• you want a simple, dependency-light PyTorch implementation of MLP-Mixer for research or experimentation
• you need image or video classification models based purely on MLP layers
• you want to benchmark MLP-only architectures against CNNs and vision transformers

When to avoid

• you need pretrained production-ready vision models with state-of-the-art accuracy
• you require architectures with attention or convolutions
• you work outside PyTorch, e.g. TensorFlow or JAX

Facets

library · maturity active

machine-learning deep-learning image-processing deep-learning computer-vision machine-learning python pytorch mlp-mixer vision-transformer-alternative image-classification research-code

Member repositories

For agents

markdown · JSON · MCP: product_card(name="lucidrains/mlp-mixer-pytorch")