EdjeElectronics/TensorFlow-Lite-Object-Detection-on-Android-and-Raspberry-Pi

Got my model, and deployed it and it won't detect anything. mAP score says otherwise.

slashplusdash opened this issue · 1 comments

mAP Results

Class Average mAP @ 0.5:0.95

0 83.76%

Overall 83.76%

Maybe it has something to do with the code responsible for the actual objection detection (for example the box-drawing logic). Have you modified any part of the code? Or if you used your own code, could you share it?