INSTALL.md

Installation

Our installation is simple and anaconda3 based. Follow the steps below:

Requirements: NVIDIA GPU, Linux

Note: We currently do not provide support for CPU only runs except SVM trainings.

Step 1: Install Anaconda3

cd $HOME
wget https://repo.continuum.io/archive/Anaconda3-5.0.1-Linux-x86_64.sh -O anaconda3.sh
chmod +x anaconda3.sh
./anaconda3.sh -b -p $HOME/anaconda3
rm anaconda3.sh

Now add anaconda3 to your PATH so that you can use it. For that run the following command:

export PATH=$HOME/anaconda3/bin:$PATH

Now, verify your conda installation and check the version:

which conda

This command should print the path of your conda bin. If it doesn't, make sure conda is in your $PATH.

Now, let's create a conda environment which we will work in.

conda create --name ssl-benchmark python=3.6
source activate ssl-benchmark

Step 2: Install dependencies using conda

We use several conda packages that are installed easily as below:

conda install -c pytorch pytorch
conda install -yq future protobuf pyyaml six scipy pycurl opencv scikit-learn cython networkx
# To test the installation works:
python -c 'from caffe2.python import core, workspace, caffe2_pb2, scope'

Now we install the COCO API.

conda install -c conda-forge matplotlib cycler
git clone https://github.com/cocodataset/cocoapi.git $HOME/cocoapi
cd $HOME/cocoapi/PythonAPI/ && python setup.py build_ext install
# To test the installation
cd ~ && python -c 'from pycocotools.coco import COCO'

Step 3: FAIR Self-Supervision Benchmark

Now clone this repository and install using instructions:

cd $HOME && git clone --recursive https://github.com/facebookresearch/fair_self_supervision_benchmark.git
cd fair_self_supervision_benchmark
python setup.py install
# To test the installation works
python -c 'import self_supervision_benchmark'

That's it! You are now ready to use this code.