Installation¶
The PyTorch Connectomics package is mainly developed on Linux machines with NVIDIA GPUs. We recommend the users to follow the Linux Installation guidance to ensure the compatibility of latest features with your system. For Windows users, please check the Windows Installation section.
Linux Installation¶
The code is developed and tested on a machine with 8 NVIDIA GPUs with the CentOS Linux 7.4 (Core) operation system.
Tip
We do not recommend installation as root user on your system python. Please setup an Anaconda/Miniconda environment and add the required packages to the environment.
Please follow the steps below for a successful installation:
1 - Install PyTorch in a virtual environment¶
conda create -n py3_torch python=3.8
source activate py3_torch
conda install pytorch torchvision cudatoolkit=11.3 -c pytorch
More options to install PyTorch can be found here. Our package has been tested with CUDA 10.2 and 11.4. Then please ensure that at least PyTorch 1.10.0 is installed:
python -c 'import torch; print(torch.__version__)'
>>> 1.10.0
2 - Install PyTorch Connectomics¶
git clone https://github.com/zudi-lin/pytorch_connectomics.git
cd pytorch_connectomics
pip install --editable .
We install the package in editable mode by default so that there is no need to re-install it when making changes to the code.
3 - Ensure CUDA is setup correctly (optional)¶
Check that PyTorch is installed with CUDA support:
python -c 'import torch; print(torch.cuda.is_available())'
>>> True
Add CUDA to $PATH and $CPATH (note that your actual CUDA path may vary from /usr/local/cuda):
PATH=/usr/local/cuda/bin:$PATH
echo $PATH
>>> /usr/local/cuda/bin:...
CPATH=/usr/local/cuda/include:$CPATH
echo $CPATH
>>> /usr/local/cuda/include:...
Verify that nvcc is accessible from terminal:
nvcc --version
>>> nvcc: NVIDIA (R) Cuda compiler driver
>>> Copyright (c) 2005-2020 NVIDIA Corporation
>>> Built on Wed_Jul_22_19:09:09_PDT_2020
>>> Cuda compilation tools, release 11.0, V11.0.221
>>> Build cuda_11.0_bu.TC445_37.28845127_0
Ensure that PyTorch and system CUDA versions match:
python -c 'import torch; print(torch.version.cuda)'
>>> 11.0
The codebase is mainly developed and tested on the Harvard FASRC cluster. For FASRC users, please load required CUDA modules from the RC server module list during running and development on the cluster. For example:
module load cuda cudnn
If you only want to install pytorch_connectomics as a Python library without clone the repository with all the pre-defined configuration files, please
use pip to directly install it from GitHub:
pip install git+https://github.com/zudi-lin/pytorch_connectomics.git
Note
If you meet compilation errors, please open an issue and describe the steps to reproduce the errors. It is highly recommended to first play with the Jupyter notebooks to make sure that the installation is correct and also have an intial taste of the functions/modules.
Windows Installation¶
These installation instructions were tested on two different Windows 10 machines, each with 1 GPU device.
Note
These instructions were designed to be used on a Windows computer without assuming any previous software was installed, or any command-line familiarity.
Please follow the steps below for a successful installation:
1 - Install Miniconda¶
The instructions to install miniconda can be found here. Most likely you want to use the link for “Miniconda3 Windows 64-bit”
2 - Open Anaconda Prompt¶
Open the anaconda prompt. You should be able to find this in the windows start menu with your other programs. Either search for it, or look in the folder most likely called “Anaconda 3 (64-bit)” Another way to find it is by clicking the start menu / press the windows key, start typing miniconda, and select “Anaconda Prompt (Miniconda3)”
4 - Run the following commands¶
The next few commands create a virtual environment, install pytorch and pytorch_connectomics, and also some libraries that windows needs to process images properly.
conda create --name py3_torch python=3.8.11 -y
conda activate py3_torch
conda install git -y
conda install pytorch torchvision torchaudio cudatoolkit=11.3 -c pytorch -y
git clone https://github.com/zudi-lin/pytorch_connectomics.git
cd pytorch_connectomics
pip install --editable .
cd ..
conda install -c conda-forge imagecodecs -y
echo Completely finished with installation. Software is ready to use
Note
The software is now installed. When you want to use the software, you must open the anaconda prompt and type the command conda activate py3_torch.
