Jiaming Han

建议参考：https://www.jianshu.com/p/0c9542cef58a

It is normal and will not affect the inference. The key `filter` in state_dicts is automatically generated by the equivariant networks (backbone), so we do not incldue this keys in...

I tested [it](https://pan.baidu.com/s/1RowD1GchTQNfuEGvMmH6bQ) again, and the mAP is still 80.10. ``` This is your result for task 1: mAP: 0.8010117619495942 ap of each class: plane:0.8881346311907455, baseball-diamond:0.824750605801873, bridge:0.6082657998640093, ground-track-field:0.8082372242336249, small-vehicle:0.7833569408756088, large-vehicle:0.8606413546155858,...

For your step 3, all our configs are prepared for 4GPUs training while you are using 1 single GPU. Note the learning rate schedule (as shown in our paper): *...

Hi, @nishit01. There is something wrong with our ReResNet pretrained models. We have updated it. See https://github.com/csuhan/ReDet/commit/f01eb2ac0ea90bac47626feb22d42db26b0d6da4 and https://github.com/csuhan/ReDet/commit/88f8170db12a34ec342ab61571db217c9589888d

你可以先尝试下载DOTA或HRSC2016，观察其格式，并将你的数据转换成类似于DOTA的格式。参考：https://github.com/csuhan/ReDet/blob/master/GETTING_STARTED.md 然后修改数据集中的类别： https://github.com/csuhan/ReDet/blob/0b9addf3c2734659fd6ffc7824f2e659fde4419c/mmdet/datasets/DOTA.py#L4-L14 接着修改对应config中的类别： https://github.com/csuhan/ReDet/blob/0b9addf3c2734659fd6ffc7824f2e659fde4419c/configs/ReDet/ReDet_re50_refpn_1x_dota1.py#L34-L59

Batch_size: https://github.com/csuhan/ReDet/blob/0b9addf3c2734659fd6ffc7824f2e659fde4419c/configs/ReDet/ReDet_re50_refpn_1x_dota1.py#L142 Copied from https://github.com/csuhan/ReDet/blob/master/README.md > All our models are trained on 4GPUs with a learning rate 0.01. If you train your model with more/fewer GPUs, remember to change the...

可能存在尺寸比较小的object，可以在dataset中加入尺寸判断： https://github.com/csuhan/ReDet/blob/0b9addf3c2734659fd6ffc7824f2e659fde4419c/mmdet/datasets/DOTA2.py#L130-L132 或者在dataset中判断mask_to_counter是否成功，对无法生成contour的object进行过滤，参考： https://github.com/csuhan/ReDet/blob/0b9addf3c2734659fd6ffc7824f2e659fde4419c/mmdet/core/bbox/transforms_rbbox.py#L557-L574

`val pipeline`并没有使用，只是占位

数据里旋转bbox由polygon表示，在模型内会转成(x,y,w,h,a) https://github.com/csuhan/ReDet/blob/0b9addf3c2734659fd6ffc7824f2e659fde4419c/DOTA_devkit/DOTA2COCO.py#L58 prepare时，empty image的anotation是空的。