search

chen-judge / RGBTCrowdCounting

Official Implement of CVPR 2021 paper “Cross-Modal Collaborative Representation Learning and a Large-Scale RGBT Benchmark for Crowd Counting”
64 stars 11 forks source link

请问不需要load VGG imagenet的预训练权重吗？我见code里没用 #1

Closed Xiaoqi-Zhao-DLUT closed 3 years ago

liulingbo918 commented 3 years ago

To maintain a similar number of parameters to original models for fair comparisons, the channel number of these backbones (CSRNet , MCNN , SANet , and BL) in our framework is respectively set to 70%, 60%, 60%, and 60% of their original values. The kernel parameters are initialized by Gaussian distribution with a zero mean and a standard deviation of 1e-2.

Our BL+IADM does not adopt the original VGG as the backbone, so the pre-trained VGG model is not used and BL+IADM is trained from scratch. In fact, you can train a lightweight VGG on ImageNet as the initialization of BL+IADM，which may lead to better performance.

Xiaoqi-Zhao-DLUT commented 3 years ago

@liulingbo918 thanks!