how to compute mAp about handtracking HOT 2 CLOSED

I'm glad to recieve your reply, thank you very much.

from handtracking.

Hi @Boyi92

While (or after) running your training on your data, you can run a concurrent evaluation job that evaluates your trained model on your test set. If you are using GCP ML Engine to run your train/evaluation jobs, you can follow the guide here.

As your evaluation job progresses, you can run a tensorboard instance and point it to the evaluation job directory ... at that point can now view mAP on tensorboard.

To learn more about how the evaluation job works, the metrics it provides (mAP), you can look at the code in eval.py and evaluator.py and evaluation util provided in the object detection api repo.

-V.

from handtracking.