train 3 tasks but detection result are very bad

[stanny880913]

I set the default.py like below:

# Alternating optimization
_C.TRAIN.SEG_ONLY = False           # Only train two segmentation branchs
_C.TRAIN.DET_ONLY = False           # Only train detection branch
_C.TRAIN.ENC_SEG_ONLY = False       # Only train encoder and two segmentation branchs
_C.TRAIN.ENC_DET_ONLY = False       # Only train encoder and detection branch

# Single task 
_C.TRAIN.DRIVABLE_ONLY = False      # Only train da_segmentation task
_C.TRAIN.LANE_ONLY = False          # Only train ll_segmentation task
_C.TRAIN.DET_ONLY = False          # Only train detection task

but when I finished 240 epochs training, lane and driving area are better than Detection very much, how to improve it?

and Can Someone show how to training classes to 13 classes or 10 classes full pipline ?
thank you