Closed MAX-OTW closed 1 year ago
Hey, so yes we reformatted the original dataset so that all datasets have one common template. Could have made some other choices in retrospect, but well that's what we did at the time.
If you download the original Tokyo 24/7 dataset, we provided a reformatting script in link to create the compatible version. Hope this helps.
Ok, thank you very much for your reply. I will give it a try :)
Hi, @MubarizZaffar ,I found "original_gt_filepath='/media/Tokyo247_full_dataset/tokyo247_dbStruct.npz'" in the code file. However, I did not find tokyo247_dbStruct.npz in the downloaded tokyo247 dataset, only dbfnames.mat. Does it need to be transformed? Can you provide tokyo247_dbStruct.npz? Or how do I transform?
Thanks & Best Regards!
Hey, sorry for the delay in my response, it completely skipped me. I should have uploaded this file at the time of the original commit, my bad, but well I have now uploaded it in the folder '/helper_functions/'.
Hey, sorry for the delay in my response, it completely skipped me. I should have uploaded this file at the time of the original commit, my bad, but well I have now uploaded it in the folder '/helper_functions/'.
Ok, thank you very much for your reply. I will give it a try :)
Hi, @MubarizZaffar , @oravus ,thank you very much for your work.But I have a problem,I looked at ground_truth_new.npy in the Tokyo 24/7 data set and the information provided there shows the ref range for each query. I was wondering how you named each image in the ref and query in the Tokyo 24/7 data set.(Directly number 1,2,3,4,5...... Is that so?)
Or can you provide the Tokyo24/7 version used in the paper(i.e.that's the dataset that's been reorganized)?
Thanks & Best Regards!