hujie-frank/SENet

Health v2 · maintenance only

Flags: no_releases

Adoption not part of the score

3646 stars · 843 forks observed · 2026-08-28

What it is AI-extracted, prompt v1, taxonomy v1, 2026-08-29, confidence not recorded

Official Caffe/CUDA implementation of Squeeze-and-Excitation Networks (SENet), channel-attention building blocks for convolutional neural networks that won ILSVRC 2017 image classification. Includes pretrained ImageNet models (SE-ResNet, SE-ResNeXt, SENet-154) and optimized GPU layers like Axpy.

Use cases

• add channel attention to CNN architectures
• reproduce ILSVRC 2017 winning image classification results
• download pretrained SE-ResNet models for ImageNet
• speed up global average pooling on GPU in Caffe
• train image classifiers with squeeze-and-excitation blocks

When to choose

• you work in Caffe and want SE blocks or pretrained SENet weights
• you need the reference implementation of the SE paper

When to avoid

• you use PyTorch or TensorFlow - SE blocks are built into modern frameworks
• you need actively maintained code - the repo is a research artifact from 2018

Facets

library · maturity maintenance

machine-learning deep-learning image-processing computer-vision deep-learning image-processing cpp caffe cuda image-classification channel-attention pretrained-models cvpr-2018 gpu linux

Member repositories

For agents

markdown · JSON · MCP: product_card(name="hujie-frank/SENet")