Weijian Deng

@liangbh6 @Simon4john Thanks for your attention. The code for re-ID feature learning is mainly modified from [IDE](https://github.com/zhunzhong07/IDE-baseline-Market-1501), and the framework is Caffe.

Thanks for your question. And we conduct experiment to see that difference between pytroch and caffe, we find the BN leads to the this performance gap. I will give the...

With the help of Houjing Huang (his homepage is [here](https://github.com/huanghoujing) ), I find the performance gap on pytorch and caffe is caused by BN. I give huang's experiments here: whether...

Thanks for your attention! Due to data safety and privacy, I am not allowed to provide datasets after the challenge ended. I only provide the synthetic set (PersonX). For the...

Following up on the above, the learned coco classification model is [here](https://anu365-my.sharepoint.com/:u:/g/personal/u6854640_anu_edu_au/EVx7YpxdsD5MlhjFVHVomWsBb-oiF0CEfKintAv7kEA4Yw?e=Vwqmj0). The model structure is ``` class FT_Resnet_fea(nn.Module): def __init__(self, mode='resnet50', num_classes=12, pretrained=True): super(FT_Resnet_fea, self).__init__() if mode == 'resnet50':...

If you find our project useful, please cite our works: ``` @inproceedings{deng2020labels, author={Deng, Weijian and Zheng, Liang}, title = {Are Labels Always Necessary for Classifier Accuracy Evaluation?}, booktitle = {Proc....

@ashygsy > link has expired Thanks for the reminder. I have fixed it. OneDrive notices me _"Your organization's policy requires this link to expire after 30 days"_. I will find...

> Hi. Would it be possible to refresh the link again? It seems that the link has expired again. Thank you. Thanks. I have refreshed, I will use google drive...

Thanks for your reply. - `many factors can influence the result, I guess first the default data split is different for these two settings` Two settings have different data splits;...

when adding `--enable_cam_center`, the result is much worse (18-->9). I will check the visualization.