search

chuanqi305 / MobileNet-SSD

Caffe implementation of Google MobileNet SSD detection network, with pretrained weights on VOC0712 and mAP=0.727.
MIT License
2.04k stars 1.18k forks source link

How to evaluate results after running train.sh?? #188

Open tahriribraq opened 4 years ago

tahriribraq commented 4 years ago

test_eval I have attached the output in my terminal after running the test.sh script to evaluate my training results. I'm very new to all this so really not sure how to evaluate the accuracy of my model, the only thing I see could be relevant is the 3rd from last line: Test net output #0: detection_eval = 0.313658 Can anybody please help me translate the output after running test.sh to evaluate my training results?