rocketpunch-inc / dcnv2 Goto Github PK

Tested with torch 1.11 + CUDA 11.3

• PyTorch 1.11
• Ninja build

pip install torch torchvision torchaudio

sudo apt-get install ninja-build

cd tests
python test_cuda.py  # run examples and gradient check on gpu
python test_cpu.py   # run examples and gradient check on cpu 

Now the master branch is for pytorch 1.x, you can switch back to pytorch 0.4 with,

git checkout pytorch_0.4

• Gradient check w.r.t offset (solved)
• Backward is not reentrant (minor)

This is an adaption of the official Deformable-ConvNets.

Update: all gradient check passes with double precision.

Another issue is that it raises RuntimeError: Backward is not reentrant. However, the error is very small (<1e-7 for float <1e-15 for double), so it may not be a serious problem (?)

Please post an issue or PR if you have any comments.