Hi, thanks for your excellent work! However, I got unfused when checking the paper about data split. Hope for your reply!
Specially, in the paper supplementary, evaluation data was generated from the test set of Places Standard, containing 30K images. But the codes in script places_standard_evaluation_prepare_data.sh show that the evaluation data was selected from the validation set of Places Standard.
It seems that the codes mismatch the paper. So I wonder which is correct!
Hi, thanks for your excellent work! However, I got unfused when checking the paper about data split. Hope for your reply!
Specially, in the paper supplementary, evaluation data was generated from the test set of Places Standard, containing 30K images. But the codes in script
places_standard_evaluation_prepare_data.shshow that the evaluation data was selected from the validation set of Places Standard.It seems that the codes mismatch the paper. So I wonder which is correct!