Atlas · GenAI 2026
EfficientNet
Selecting and adapting the EfficientNet convolutional-network family, with an understanding of compound scaling across network depth, width and input resolution.
Also searchable as: EfficientNets
conceptNeural ArchitecturesAI consensus: 0/3
Prerequisites
The family builds on convolutional-network representations and training.
Recommended reference
Tan and Le: EfficientNet Rethinking Model Scaling for Convolutional Neural Networks — https://arxiv.org/abs/1905.11946
Reviewed sources
Primary and first-party material reviewed for this editorial summary. These citations are separate from the AI consensus score above.
- Tan and Le: EfficientNet Rethinking Model Scaling for Convolutional Neural Networks
Compound scaling and the EfficientNet model family.
- Keras: EfficientNet models
Official B0–B7 model APIs covering input resolution, pretrained weights and the preprocessing included in the Keras implementation.
- Torchvision: EfficientNet
Official documentation for the EfficientNet B0–B7 family and builders, with or without pretrained weights, linked to the original scaling paper.
Notes from AI deep research
Related skills
- → is subcategory of: Convolutional Neural Networks