DCH

Class Project ：PyTorch implementation for DCH: Deep Cauchy Hashing for Hamming Space Retrieval, Yue Cao, Mingsheng Long, Bin Liu, Jianmin Wang, IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2018

Prerequisites

Linux or OSX

NVIDIA GPU + CUDA (may CuDNN) and corresponding PyTorch framework (version 0.3.1)

Python 2.7/3.5

Datasets

You can download the ImageNet dataset and NUS-WIDE dataset here. As for COCO dataset, you can downloaded here. Each line in the list file follows the following format:

<image path><space><one hot label representation>

Training

First, you can manually download the PyTorch pre-trained model introduced in `torchvision' library or if you have connected to the Internet, you can automatically downloaded them. Then, you can train the model for each dataset using the followling command.

cd src
python train.py --gpu_id 0 --dataset coco --prefix AlexNet_hashnet --hash_bit 48 --net AlexNet --lr 0.0003 --q_lambda 1.0 

Evaluation

You can evaluate the Mean Average Precision(MAP) result on each dataset using the following command.

cd src
python test.py --gpu_id 0 --dataset coco --prefix resnet50_hashnet --hash_bit 48 --snapshot iter_09000