Thanks for your remarkable work.
As far as I understand, all the pedattr tasks require a pickle file for the dataset information. Do you have any documentation on how to generate the dataset.pkl for the datasets? (Or have you already shared and I missed it?) I could generate the pkl files for PA100K and RAPv2 by using the scripts (format_pa100k.py and format_rap2.py) from Rethinking_of_PAR. However, I couldn't find how to generate for HARDHC, UAV-Human, Parse27k, Market-1501 (attribute) for reproducibility of the results. Could you share how did you manage to generate those?
Thanks for your remarkable work.
As far as I understand, all the
pedattrtasks require a pickle file for the dataset information. Do you have any documentation on how to generate the dataset.pkl for the datasets? (Or have you already shared and I missed it?) I could generate the pkl files for PA100K and RAPv2 by using the scripts (format_pa100k.pyandformat_rap2.py) from Rethinking_of_PAR. However, I couldn't find how to generate forHARDHC,UAV-Human,Parse27k,Market-1501 (attribute)for reproducibility of the results. Could you share how did you manage to generate those?