Closed rlaehddls9 closed 2 years ago
Thanks for good project using Radiate dataset. But.. How to evaluate faster_rrcnn_R_50_FPN ?
First, it causes this assertion. assert torch.all(match_quality_matrix >= 0)
Therefore I removed above assertion, then I tried to evaluate above model by using RotatedCOCOEvaluator but AP was approximately 25.
How can I fix it?
Could you share your rotated faster rcnn implementation code?
Thanks for good project using Radiate dataset. But.. How to evaluate faster_rrcnn_R_50_FPN ?
First, it causes this assertion. assert torch.all(match_quality_matrix >= 0)
Therefore I removed above assertion, then I tried to evaluate above model by using RotatedCOCOEvaluator but AP was approximately 25.
How can I fix it?
Could you share your rotated faster rcnn implementation code?