Asking for loss value

[coolmakersss]

Hi, @rhett-chen,

Thank you very much for your excellent code.

The AP value of my own training results is only about half of what is reported in the paper and I'm debugging.

I wanted to ask if you remember the approximate values of each loss function after they converged, not too precisely. If you could provide them, I would be very grateful.

Thank you!

Best regards

[rhett-chen]

----epoch: 010 ---- batch: 5400 ---- mean graspable_count_stage1: 2026.875000 mean loss/overall_loss: 0.768805 mean loss/stage1_graspness_loss: 0.001724 mean loss/stage1_objectness_loss: 0.068601 mean loss/stage2_view_loss: 0.005997 mean loss/stage3_score_loss: 0.003021 mean loss/stage3_width_loss: 0.003791

[coolmakersss]

Thank you very much!