AerialDetection
AerialDetection copied to clipboard
训练几次后跑到一半就报: double free or corruption(!prev) 已放弃(核心已转储) 的错误,也没有报错的地方。在跑代码就跑不起来了,重启之后才好,请问各位大佬这是为什么啊????
hi and tnx for your implemention in your paper : https://openaccess.thecvf.com/content_CVPR_2019/html/Ding_Learning_RoI_Transformer_for_Oriented_Object_Detection_in_Aerial_Images_CVPR_2019_paper.html the paper said about configs like Light-Head R-CNN OBB, deformable Position Sensitive RoI pooling and RoI Transformer where can...
在论文中只看到RoI transformer在DOTA1.0上的测试效果,请问在哪能看到DOTA1.5的RoI transformer的测试效果(mAP 和AP)呢? 谢谢
ImportError: mmdet/opts/nms/nms_cuda.cpython-36m-x86_64-linux-gnu.so:undefined symbol:THCudaFree
Can this project support extract warping rotated ROI while also keeping original aspect ratio?
Hello, thank you very much for publicizing the code. I have some doubts about the initial angle of the horizontal frame. In your paper, the angle range of the rotating...
So, if I wanted to adjust the loss function to account for class imbalance ..... where would I do that? Anyone have any thoughts on compensating for class imbalance in...
Is it some code to get result of a single image evaluation considering ground truth (loss or AP for the single image)?
Hello, Really amazing work. While doing inference, i faced the results are not as good as shown. I tested using the pre trained models: 1) faster_rcnn_RoITrans_r50_fpn_1x_dota  2) faster_rcnn_RoITrans_r50_fpn_1x_dota1_5 ...
在RoI Transformer.py文件中,为什么不直接计算rbbox,还要计算bbox的损失呢?还有就是,mask的作用是什么呢?为什么要计算mask呢? 