How to train Centernet model for an instance segmentation?

#1

Hi all,

Can someone please show me how to train Centernet model for instance segmentation.

I have well understood the Centernet model, but I didn’t understand how to train this model for instance segmentation.

Can someone show me an example?

Thanks in advance

#2

Hi all,

Can someone explain to me how to code the training part with pytorch for the Centernet model for instance segmentation.

Thanks in advance.

#3

Hi, I’m trying to do instance segmentation on small objects like a baseball (30x30) in I’ve tried training Mask RCNN following the balloon example, but little success. CenterNet does 28.9 APs which is close to the state of the art PANet. detection, batch them and feed into a lightweight semantic segmentation model.