MaskRCNN in mmdetection only detecting 100 instances, even though the image contains more.

[av5672344]

Hello,

I am currently using MaskRCNN implemented in OpenMMLab's mmdetection library for object detection. However, I have noticed that during inference, the model is only detecting exactly 100 instances of objects, regardless of the number of objects present in the image.

I known that it was already treated in here.
I have tried adjusting the configuration settings, but the issue persists. I have also trained the model on a larger dataset, but the problem still persists. I am unsure of what could be causing this issue and would greatly appreciate any help or suggestions.

If you have any ideas or solutions to this problem, please let me know.

Thank you in advance for your help.