Implementations of polygamma, lgamma, and beta functions for PyTorch. It's very hacky, but that's usually ok for research use.
To build, run:
./make.sh
You'll probably need to pass in the correct CUDA path to build.py
, which is run inside make.sh
, so modify it to instead call
python build.py --cuda-path YOUR_CUDA_PATH
Also, you'll probably need to change the architecture version/CUDA compute capability inside make.sh
, so replace sm_35
with whatever your GPU supports.
Feel free to open an issue if you run into another problem!