search

aimagelab / VKD

PyTorch code for ECCV 2020 paper: "Robust Re-Identification by Multiple Views Knowledge Distillation"
MIT License
73 stars 15 forks source link

Animal Re-ID Details #22

Closed 07Agarg closed 2 years ago

07Agarg commented 2 years ago

Hi Authors, Thanks for the nice work and sharing the GitHub code!

I am trying to reproduce the results of Table 7 for Amur Tiger for animal re-identification task. The command.txt of the released code has commands for Image To Video and Video To Video settings. It can be a great help to let me know how to use the GitHub repository for the Image 2 Image setting.

Also, in the tools/ subdirectory, there are two training files, train_v2v.py, and train_distill.py. Which of these two files should I use for the Image2Image setting? If possible, please let me know the details of the hyperparameter used for the same.

Thanks again!

angpo commented 2 years ago

Hi,

To reproduce the results on ATRW, you have to train the teacher at first (hence, train_v2v.py), and then the student (train_distill.py). You can use these scripts even for the I2I setting.

I will send you an email to share the hyperparameters.

Best regards, A

07Agarg commented 2 years ago

Hi, Thanks for your quick response.

Isn't the code not integrated for Amur Tigers Dataset? The data/ subdirectory has data_manager.py for only three datasets which do not include Amur Tigers. In Line 89, data/datasets.py, the Datasets dictionary listed does not include Amur Tigers. Similarly, in Line 59, data/misc.py, get_transforms function is not customized for Amur Tigers dataset.

Thanks,

angpo commented 2 years ago

Yes, the public repo does not support the Amur Tigers dataset. However, I will try to help you by e-mail.

Thanks,