Hello authors,
I am trying to reproduce the results of Strip R-CNN on the HRSC2016 dataset as mentioned in your paper.
I have already obtained the StripNet backbone pre-trained weights (stripnet_s.pth.tar) from the model zoo, but I cannot find the full detection model weights trained on HRSC2016 (i.e., strip_rcnn_s_fpn_3x_hrsc_le90.pth, which should achieve mAP 0.9054 / 0.906 on HRSC2016) in the current release.
Could you please share the HRSC2016 pre-trained weights for Strip R-CNN-S? It would be very helpful for my reproduction work.
Thank you very much for your great work and support!
Hello authors,
I am trying to reproduce the results of Strip R-CNN on the HRSC2016 dataset as mentioned in your paper.
I have already obtained the StripNet backbone pre-trained weights (stripnet_s.pth.tar) from the model zoo, but I cannot find the full detection model weights trained on HRSC2016 (i.e., strip_rcnn_s_fpn_3x_hrsc_le90.pth, which should achieve mAP 0.9054 / 0.906 on HRSC2016) in the current release.
Could you please share the HRSC2016 pre-trained weights for Strip R-CNN-S? It would be very helpful for my reproduction work.
Thank you very much for your great work and support!