-
Hi, i have been reading your implementation line by line. And I found a code block really strange.
In the file "master/DCNv2_op/modulated_deformable_convolution-inl.h", and in the backward function.…
-
The implementation of creating a new feature map and conv with stride=kernel_size is so cool .
Will you add deform group param next ?
-
[19:15:03] src/operator/convolution.cu:87: This convolution is not supported by cudnn, MXNET convolution is applied.
[19:15:03] src/operator/convolution.cu:87: This convolution is not supported by cu…
-
Hi, TF.
I wonder if you have any plan to add Deformable Convolution Feature.
https://arxiv.org/abs/1703.06211
https://github.com/msracver/Deformable-ConvNets
Thank you always, for nice work.
…
-
Hi, I try to reproduce the work,use config file and pre-trained model you provided and trained for 60 epoch. Using 0.1 thresh to do NMS during test.
But finally map is just 40.7,is far lower than b…
-
How to Save visualized networks in jpg/png? in python code
## Environment info
Operating System: Ubuntu 16.0.4
Compiler:
Package used Python
MXNet version: 0.9.3
Or if installed fro…
-
Hi, Mr Lin, an excellent work you have done! I have a question after I read many versions of code of deformable cnn, including the officical mxnet version. Why does the offset conv have the bias term …
-
@unsky HI
按照READEME成功将**deformable_conv**嫁接到了多个自己的caffe中,如:**py-R-FCN**.
但在进行检验,即执行`./experiments/scripts/faster_rcnn_end2end.sh 0 ResNet50 pascal_voc`的时候,总会报如下错误:
```
F0626 16:18:14.308534 28…
-
Hi team,
I met this problem when I am trying to use MKLDNN with MXNet. The code can be run with regular MKL successfully, but with MKLDNN (the latest version) and with MXNet 1.2, it has the followi…
-
Hey,
Sniper is an awesome work. However ,FPN is a standard component of object detection for scales with slightly computation increased. Have you tried to combine Sniper with FPN ?