Skip to content

Latest commit

 

History

History
78 lines (48 loc) · 2.41 KB

README.md

File metadata and controls

78 lines (48 loc) · 2.41 KB
Raw

SENet.pytorch

An implementation of SENet, proposed in Squeeze-and-Excitation Networks by Jie Hu, Li Shen and Gang Sun, who are the winners of ILSVRC 2017 classification competition.

Now SE-ResNet (18, 34, 50, 101, 152/20, 32) and SE-Inception-v3 are implemented.

  • python cifar.py runs SE-ResNet20 with Cifar10 dataset.

  • python imagenet.py IMAGENET_ROOT runs SE-ResNet50 with ImageNet(2012) dataset.

    • You need to prepare dataset by yourself
    • First download files and then follow the instruction.
    • The number of workers and some hyper parameters are fixed so check and change them if you need.
    • This script uses all GPUs available. To specify GPUs, use CUDA_VISIBLE_DEVICES variable. (e.g. CUDA_VISIBLE_DEVICES=1,2 to use GPU 1 and 2)

For SE-Inception-v3, the input size is required to be 299x299 as the original Inception.

Pre-requirements

  • Python>=3.6
  • PyTorch>=1.0
  • torchvision

For training

To run cifar.py or imagenet.py, you need

  • pip install git+https://github.com/moskomule/homura

hub

You can use some SE-ResNet (se_resnet{20, 56, 50, 101}) via torch.hub.

import torch.hub
hub_model = torch.hub.load(
    'moskomule/senet.pytorch',
    'se_resnet20',
    num_classes=10)

Result

SE-ResNet20/Cifar10

python cifar.py [--baseline]
ResNet20 SE-ResNet20 (reduction 4 or 8)
max. test accuracy 92% 93%

SE-ResNet50/ImageNet

The initial learning rate and mini-batch size are different from the original version because of my computational resource .

ResNet SE-ResNet
max. test accuracy(top1) 76.15 %(*) 77.06% (**)

  • (*): ResNet-50 in torchvision

  • (**): When using imagenet.py with the --distributed setting on 8 GPUs. The weight will be available soon.

senet = se_resnet50(num_classes=1000)
senet.load_state_dict(torch.load("weight.pkl"))

References

paper

authors' Caffe implementation