Deformable-Convolution-V2-PyTorch
Deformable-Convolution-V2-PyTorch copied to clipboard
chengdazhi
Deformable ConvNets V2 (DCNv2) in PyTorch
Who can help me pytorch 1.1.0 cuda 8.0 NVRM version: NVIDIA UNIX x86_64 Kernel Module 396.45 running build running build_py creating build creating build/lib.linux-x86_64-3.6 creating build/lib.linux-x86_64-3.6/functions copying functions/__init__.py -> build/lib.linux-x86_64-3.6/functions...
Do you have any description documentation about this project,I dont know how to use it,and dont understand these files functions。thanks
Hello, I wanted to compile and test the DCN module. I'm on the master branch. with sh make.sh I got > Including CUDA code. /home/ahyun/cvlab/Deformable-Convolution-V2-PyTorch generating /tmp/tmp3y9_gfz9/_deform_conv.c setting the current...
When I run check_gradient_dconv( ), it will return this RuntimeError message. When I looked into the def gradcheck( ), I found that the output of analytical and numerical is different....
Hello Thanks for your implementation. I am trying to build your code, however I got errors in the beginning of make. deform_conv_cuda_kernel.cu(71): error: namespace "std" has no member "min" 1...
I am not familiar with CUDA programming and conv implementation in CUDA. So may I ask a stupid question? In function `deformable_im2col_gpu_kernel`, is the loop in `CUDA_KERNEL_LOOP` executed only once?...
DCNv2 is now included in the official mmdetection repo based on the master branch of this one. We rewrite it with the new cpp extension apis and it supports both...
