Common

conda install cmake ninjapip install -r requirements.txt

On Linux

conda install mkl mkl-include# CUDA only: Add LAPACK support for the GPU if neededconda install -c pytorch magma-cuda110# or the magma-cuda* that matches your CUDA version from https://anaconda.org/pytorch/repo# (optional) If using torch.compile with inductor/triton, install the matching version of triton# Run from the pytorch directory after cloningmake triton

On MacOS

# Add this package on intel x86 processor machines onlyconda install mkl mkl-include# Add these packages if torch.distributed is neededconda install pkg-config libuv

On Windows

conda install mkl mkl-include# Add these packages if torch.distributed is needed.# Distributed package support on Windows is a prototype feature and is subject to changes.conda install -c conda-forge libuv=1.39

Next Steps