Open watbywbarif opened 2 years ago
@watbywbarif Sorry for late response.
According to the error, it looks like the file format is not correct. Is GT-final_test.csv in the following format? The last column should be 'ClassId'.
$ head GT-final_test.csv
Filename;Width;Height;Roi.X1;Roi.Y1;Roi.X2;Roi.Y2;ClassId
00000.ppm;53;54;6;5;48;49;16
00001.ppm;42;45;5;5;36;40;1
00002.ppm;48;52;6;6;43;47;38
00003.ppm;27;29;5;5;22;24;33
00004.ppm;60;57;5;5;55;52;11
00005.ppm;52;56;5;5;47;51;38
00006.ppm;147;130;12;12;135;119;18
00007.ppm;32;33;5;5;26;28;12
00008.ppm;45;50;6;5;40;45;25
I stumbled over the same issue. The problem is that in the dataset there is also a file GTSRB/Final_Test/Images/GT-final_test.test.csv
which does not set the ;ClassId
column. The code is reading all .csv files and when it comes to this one it fails.
As there is also a correct GT-final_test.csv
in the dataset I deleted the other one which fixed it for me
Original links for setup on http://benchmark.ini.rub.de are not working, I found files with same name on sid.erda.dk/public/archives but I am unable to run any commands after:
python3 gen_pickle.py
fails after few minutes with error below (Ubuntu 21.04 + Python 3.9.5 + pandas-1.3.4 + tensorflow 2.7.0 - please specify if some older version of sw should be used):