zxiaoyusq

it fixed,thanks for your help! by the way,where can i set a separate learning rate or freeze for each layer?

OK,thanks!

when I run train report:  Will this affect the training effect?

did you modify topdown_coco_dataset.py?