KimSoybean

@anjingxing Sorry, recently I'm preparing for my IELTS testing and I will supplement details after I finish my test. This is really an easy code and you can refer to...

BN is evaluated by the statistics in a batch, so batchsize will impact the perfoemance. I have tried GN on mmdetection on ScratchDet, and get 1mAP lower result. I haven't...

@foralliance Hi！I think we should speak English because your questions may help the people in other countries. 1. No. When you train the model with large input image size (e.g.,...

@dby2017 Hello, I haven't tried Root-Res on faster r-cnn. Maybe it is not very effective due to the large input resolution (small objects will be larger than before).

@zimenglan-sysu-512 Hi, we will release the codes and models. But we will release them after the announcement of CVPR2019. I recommend you to reproduce the code on mmdetection, which is...

@hust-kevin Hi, the accumulate batchsize is 128, in our code for Root-ResNet34, we put 16 images on each GPU (totally 4 GPUs), and the gradient is updated by the average...

@hust-kevin I think lr=0.03 is better, but it depends on your loss curves. If you have only one GPU, I recommend you to use GN instead of BN and train...

@dby2017 Recently we have trained the 512x512 models on mmdetection which took a long time. We will release codes and models when we finish it.

@Forgetmyfairy I have just finished my graduation thesis for undergraduates. This weekend I will upload them.

@XCRobert Hello, I have tried GN on ScratchDet on mmdetection and get 1 mAP lower result than BN. Sorry I haven't found the reason but I think it is relevant...