Skip to content

Latest commit

 

History

History
48 lines (29 loc) · 1.16 KB

installation_instructions.rst

File metadata and controls

48 lines (29 loc) · 1.16 KB

Installing FairScale

Installing FairScale is extremely simple with pre-built binaries (pip) that we provide. You can also build from source using the instructions below.

Requirements

  • PyTorch>= 1.8.1

Installing the pip package (stable)

pip install fairscale

Installing with conda

Fairscale is packaged by conda-forge (see here) for both linux & osx, with GPU-enabled builds available on linux.

conda install -c conda-forge fairscale

Installing from source

git clone https://github.com/facebookresearch/fairscale.git
cd fairscale
pip install -r requirements.txt
# -e signified dev mode since e stands for editable
pip install -e .

To build with GPU-support enabled, be sure to set BUILD_CUDA_EXTENSIONS=1 as well as an appropriate TORCH_CUDA_ARCH_LIST.

Note: If either of the above fails, add --no-build-isolation to the pip install command (this could be a problem with recent versions of pip).