You can download and use the model files created by training with DY-Cace and DY-BL on four different datasets for inference.
Alternatively for training the networks, download the dynamic ResNet-50 model trained on ImageNet into pretrained_models directory.
Training/Inference code and configuration files will be added.
Baseline and CaceNet person ReID models: https://github.com/TencentYoutuResearch/PersonReID-YouReID
Yu, F., Jiang, X., Gong, Y., Zhao, S., Guo, X., Zheng, W.-S., Zheng, F., Sun, X.: Devil’s in the details: Aligning visual clues for conditional embedding in person re-identification. In: Proc. IEEE CVPR (2021)
Dynamic convolution : https://github.com/liyunsheng13/dcd
Chen, Y., Dai, X., Liu, M., Chen, D., Yuan, L., Liu, Z.: Dynamic convolution: Attention over convolution kernels. In: Proc. IEEE CVPR, pp.11030–11039 (2020)