mangye16
mangye16
This means that you did not have the pretrained ImageNet model "resnet50-19c8e357.pth". You may mannually download it at "https://download.pytorch.org/models/resnet50-19c8e357.pth" and put it in correct path.
It is unclear about your request. Do you mean the visualization about the learned feature embeddings or the retrieved ranking list?
不太确定你的问题，是关于什么的可视化？我好像没有博客。。。
> @mangye16 作者您好，关于行人重识别检索示列结果可视化我也有不明白的地方。比如在您的文章《Bi-Directional Center-Constrained Top-Ranking for Visible Thermal Person Re-Identification》中图11的检索示列结果中的可视化结果是怎么做的？想请您解答一下，感谢~ 这个就是写了一个简单的代码把结果plot出来。。。
You may try the code in https://github.com/mangye16/Cross-Modal-Re-ID-baseline
Actually, I use the default normalization parameter of pytorch, the performance is much better for my task. I'm not sure whether it works for the ImageNet testing. You may try...
Actually, I only test it on CUB200 dataset using the pre-trained weights. Using your normalization parameter the accuracy is: recall @1 =26.5%, nmi =38.2% Using the above-mentioned parameter the accuracy...
I suppose you might have made a mistake in the visible-to-infrared setting. Just keep all the settings as default (Do not change the test_mode in [Line 69]), since this represents...
You may refer to the baseline code at [test.py in Line 254](https://github.com/mangye16/Cross-Modal-Re-ID-baseline/blob/30795ef39b3206d0c0030ae9e58b1813ef155612/test.py#L254). Other parts are kept the same.