Loss weighting - add weight factor to the two different types of loss

libornovax / master_thesis_code

Code for my master thesis: Vehicle Detection and Pose Estimation for Autonomous Driving

MIT License

186 stars 69 forks source link

Loss weighting - add weight factor to the two different types of loss #18

Closed libornovax closed 7 years ago

libornovax commented 7 years ago

We have coordinate and probability accumulator loss - after some time it seems that the coordinates are learned, but the probability still could train more. Therefore perhaps add more weight on the probability...