How to cite results from this repository?

[Noixas]

I have gone through both Joint discriminative and generative learning for person re-identification and A discriminatively learned CNN embedding for person re-identification, but neither of them reports a lot of the newer trained models scores (e.g. swin transformer, HRNet-18, PCB). As far as I understand, the models provided in here are all based on the same architecture and what changes is the backbone and some parameters. For example:

• all tricks means -> --warm_epoch 5 --stride 1 --erasing_p 0.5 --batchsize 8 --lr 0.02
• circle loss
• DG

My questions are:

1. Are there any other significant changes when implementing swin transformer or other backbones?
2. Should I use Verif-Identif [55] as model name when citing the results of this repository, regardless of the architecture? Or is there a specific notation used when citing the models from this repo?

[layumi]

Hi @Noixas Sorry for the late response.

1. Swin transformer needs a fixed input size to ensure the window number.
2. You may use ResNet + Verif-Identif or Swin + Circle such citation.