Description

Implementation of DSLH Algorithms. ICIP2020

How to run

Firstly, generate the sample databases owned by each class

pyhon preprocess.py

According to your data storage location, adjust the directory and hyperparameters in config.py Then, you can easily train and test just by

pyhon main.py

Dataset:

CIFAR10 (Lossless PNG format): https://drive.google.com/open?id=1NZ5QKW2zqzN-RQ4VDpuOAb-UgcsTPUJK Size: 140M

NUS-WIDE: https://drive.google.com/open?id=0B7IzDz-4yH_HMFdiSE44R1lselE Size: 5G

ImageNet: https://drive.google.com/open?id=0B7IzDz-4yH_HSmpjSTlFeUlSS00 Size: 14G

COCO: https://drive.google.com/open?id=0B7IzDz-4yH_HN0Y0SS00eERSUjQ Size: 19G

Plaform information:

python 3.6.8 numpy 1.16.5 pytorch 1.0.0 torchvision 0.2.2 pillow 6.2.0