/Adversarial-Complementary-Learning-for-Weakly-Supervised-Object-Localization

ACoL re-implementation by pytorch

Primary LanguagePythonMIT LicenseMIT

Adversarial-Complementary-Learning-for-Weakly-Supervised-Object-Localization

ACoL re-implementation by pytorch

Prerequisites: Python3.6 and Pytorch1.1

Data Preparation : We use DET as our datasets
npy file preparation：Need to write the npy format file as follows！！！ image

Before train, you need to mortify the datasets'path and some params for your own.

Train : python main.py


Due to the large pre-training model, we provide a Baidu cloud download link below
Link：https://pan.baidu.com/s/1x2wGNIM-eRPiKxYuiE05Aw Password：c0ua