GodMan

Have you solved this problem?

Yes, you just need replace your softmaxloss with focalloss function. Then everything is ok.

Maybe you should use tensorflow official models at https://github.com/tensorflow/models/tree/master/research/object_detection. Then you just replace the softmax_loss with focal_loss

Hi, thanks for your praise. 1. You just need to replace softmax loss by focal loss of classification. 2. You can remove hard mining operations, but I choose to change...

Oh, my god. The latest version of models has changed the loss function return value. My code is for previous version(maybe before version 1.2). You just need return per_entry_cross_ent variable...

Hi, for Q1: what label should we set for the background, positive or negative? And where should we choose the bbox? How many should we set for background? We can...

It depends what Network you choose. For example, if you implement ssd, you may need searching anchor boxed on multi-layer features using max_iou threshold. If you implement Faster-RCNN series and...

You are welcome. I'm curious about your final training result, does it significant incrementation?

The paper of He proposed sigmoid has more robust training effect than softmax, But he didn't explain why. You can have a try. I think maybe it depends the dependent...

yes, it is.