VDIGPKU / M2Det

M2Det: A Single-Shot Object Detector based on Multi-Level Feature Pyramid Network
MIT License
1.45k stars 318 forks source link

Someone have test result on VOC2007 ? #109

Open nththanh opened 4 years ago

nththanh commented 4 years ago

I run train.py on VOC2007 dataset and run test.py but I don't have an other test result on VOC to compare Someone have it ?

mohanrobotics commented 4 years ago

@nththanh Can you please share your results ??

nththanh commented 4 years ago

my result

Result Final
aeroplane 0.526
bicycle 0.522
bird 0.289
boat 0.335
bottle 0.107
bus 0.557
car 0.625
cat 0.498
chair 0.254
cow 0.373
diningtable 0.445
dog 0.419
horse 0.645
motorbike 0.574
person 0.536
pottedplant 0.135
sheep 0.392
sofa 0.392
train 0.645
tvmonitor 0.376
Mean AP 0.432
mohanrobotics commented 4 years ago

Thanks @nththanh . I will update you on my results soon.

zhangke1997 commented 4 years ago

79.88% = aeroplane AP 90.25% = bicycle AP 69.46% = bird AP 62.76% = boat AP 50.18% = bottle AP 84.41% = bus AP 87.49% = car AP 83.97% = cat AP 53.00% = chair AP 79.05% = cow AP 69.53% = diningtable AP 84.65% = dog AP 85.32% = horse AP 83.41% = motorbike AP 75.41% = person AP 34.76% = pottedplant AP 69.08% = sheep AP 69.06% = sofa AP 89.85% = train AP 78.13% = tvmonitor AP mAP = 73.98%

jianguolalala commented 3 years ago

79.88% = aeroplane AP 90.25% = bicycle AP 69.46% = bird AP 62.76% = boat AP 50.18% = bottle AP 84.41% = bus AP 87.49% = car AP 83.97% = cat AP 53.00% = chair AP 79.05% = cow AP 69.53% = diningtable AP 84.65% = dog AP 85.32% = horse AP 83.41% = motorbike AP 75.41% = person AP 34.76% = pottedplant AP 69.08% = sheep AP 69.06% = sofa AP 89.85% = train AP 78.13% = tvmonitor AP mAP = 73.98%

Would you kindly provide the config file? I trying to reproduce your result.