pre-trained yolo9000 bad recognition

allanzelener / YAD2K

YAD2K: Yet Another Darknet 2 Keras

Other

2.71k stars 879 forks source link

pre-trained yolo9000 bad recognition #165

Open zeevikal opened 5 years ago

zeevikal commented 5 years ago

hi, after downloaded yolo9000.weights (and yolo9000.cfg) from the original darknet repository I converted it to h5 file via yad2k.py. when running this mode; with test_yolo.py on your test images I'm getting this:

what am I doing wrong?

huangeason26 commented 4 years ago

i have the same problem, so i just wonder if you had solved it?

average-everyman commented 3 years ago

hi @zeevikal, I have no solution for you but I wanted to ask you how you managed to get to that point actually? I did the conversion of yolo9000 cfg/weights, but I am a bit unsure what to do about the class names file? the ones that were provided (coco_names.txt and pascal_names.txt) - if I remove them or change their content I am getting some errors.