The code was tested with Anaconda Python 3.8, CUDA 11.3, and PyTorch 1.10.1. After installing Anaconda:
-
[Optional but highly recommended] create a new conda environment.
conda create -n hier python=3.8
And activate the environment.
conda activate hier
-
Install PyTorch:
conda install pytorch=1.10.1 torchvision=0.11.2 cudatoolkit=11.3 -c pytorch
-
Clone this repo:
git clone https://github.com/anthonyweidai/HierAttn.git
-
Install the requirements
cd $HierAttn_ROOT pip install -r requirements.txt