Skip to content

INSTALL.md

Latest commit

 

History

History
144 lines (101 loc) · 4 KB

INSTALL.md

File metadata and controls

144 lines (101 loc) · 4 KB

Prerequisites

AMD GPU

Running on AMD GPU's needs a functioning ROCm stack in order to support GPU accelaration for ML libraries such as pytorch and tensorflow. Builds of these libraries are available with the ROCm backend enabled.

Install Ubuntu 18.04.4

ROCm is not supposed to work with the prorietary AMD driver. After the operation system is installed, verify the open-source kernel module amdgpu is loaded:

kmod list | grep amdgpu

Setup ROCm

Setup ML libraries

pytorch

Docker
From source
1. Clone forked PyTorch repository on the host
cd ~
git clone https://github.com/zirkular/pytorch.git
cd pytorch
git submodule init
git submodule update
2. Build PyTorch docker image
cd pytorch/docker/caffe2/jenkins
sudo ./build.sh py3.6-gcc-rocmdeb-conda-opencv-ubuntu18.04
3. Start a docker container using the new image:
sudo docker run -it -v $HOME:/data --privileged --rm --device=/dev/kfd --device=/dev/dri --group-add video <image_id>

Reference: https://rocmdocs.amd.com/en/latest/Deep_learning/Deep-learning.html#option-2-install-using-pytorch-upstream-docker-file.

Nvidia GPU

Cuda 10.1

Follow this instruction: https://www.tensorflow.org/install/gpu#install_cuda_with_apt

Cuda 10.2

Note: CUDA 10.2 is not compatible with Tensorflow 1.14.0 Note: CUDA 10.0+ requires at least CMake 3.12.2+

wget -c "https://developer.download.nvidia.com/compute/cuda/repos/ubuntu1804/x86_64/cuda-ubuntu1804.pin"
sudo mv cuda-ubuntu1804.pin /etc/apt/preferences.d/cuda-repository-pin-600
wget http://developer.download.nvidia.com/compute/cuda/10.2/Prod/local_installers/cuda-repo-ubuntu1804-10-2-local-10.2.89-440.33.01_1.0-1_amd64.deb
sudo dpkg -i cuda-repo-ubuntu1804-10-2-local-10.2.89-440.33.01_1.0-1_amd64.deb
sudo apt-key add /var/cuda-repo-10-2-local-10.2.89-440.33.01/7fa2af80.pub
sudo apt-get update
sudo apt-get -y install cuda

CUDNN_URL="http://developer.download.nvidia.com/compute/redist/cudnn/v5.1/cudnn-8.0-linux-x64-v5.1.tgz"
wget -c ${CUDNN_URL}
sudo tar -xzf cudnn-8.0-linux-x64-v5.1.tgz -C /usr/local
rm cudnn-8.0-linux-x64-v5.1.tgz && sudo ldconfig

GPU monitoring

watch -n 0.5 nvidia-smi

High-level libraries

detectron2

Requirements

  • Linux or macOS with Python ≥ 3.6
  • PyTorch ≥ 1.4
  • torchvision that matches the PyTorch installation. You can install them together at pytorch.org to make sure of this.
  • pycocotools. Install it by pip install pycocotools>=2.0.1.
  • OpenCV, optional, needed by demo and visualization

Build Detectron2 from Source

CUDA

gcc & g++ ≥ 5 are required. ninja is recommended for faster build. After having them, run:

# Install from a local clone:
git clone https://github.com/facebookresearch/detectron2.git
python -m pip install -e detectron2

# Or if you are on macOS
CC=clang CXX=clang++ python -m pip install ......

Please refer to detectron2 for further information in the installation process.

ROCm

Need to pass WITH_HIP and WITH_OPENCL to build script.

OpenPose

sudo apt install libhdf5-dev \
    protobuf-compiler \
    libopencv-dev \
    libatlas-base-dev \
    libboost-all-dev \
    libopencv-dev

AMD GPU

git clone https://github.com/CMU-Perceptual-Computing-Lab/openpose
cd openpose
mkdir build && cd build
cmake .. -DGPU_MODE=OPENCL