MONAI on ROCm installation

MONAI on ROCm installation#

2026-04-01

5 min read time

Applies to Linux

To install MONAI on ROCm, you have the following options:

System requirements#

ROCm version Ubuntu version Python version AMD Instinct™ GPU PyTorch for AMD ROCm version NumPy version
7.2.0, 7.0.2 24.04 3.12 MI355X, MI325X, or MI300X 2.8.0 and later No earlier than 1.24 and no later than 2.4

For the complete list of dependencies, see the requirements.txt file.

Setting up the environment#

To set up the environment for MONAI on ROCm installation, follow these steps:

  1. Set up the Docker image using the ROCm or Ubuntu Docker image from Docker Hub.

    • For ROCm 7.2.0, run:

      docker run --cap-add=SYS_PTRACE --ipc=host --privileged=true   \
--shm-size=512GB --network=host --device=/dev/kfd        \
--device=/dev/dri --group-add video -it                  \
-v $HOME:$HOME  --name ${LOGNAME}_monai                  \
                        rocm/dev-ubuntu-24.04:7.2-complete

    • For ROCm 7.0.2, run:

      docker run --cap-add=SYS_PTRACE --ipc=host --privileged=true   \
--shm-size=512GB --network=host --device=/dev/kfd        \
--device=/dev/dri --group-add video -it                  \
-v $HOME:$HOME  --name ${LOGNAME}_monai                  \
                      rocm/dev-ubuntu-24.04:7.0.2-complete

  2. Install the required system dependencies.

    apt-get update                                                        &&  \
apt-get install -y software-properties-common lsb-release gnupg wget  &&  \
apt-key adv --fetch-keys                                                  \
            https://apt.kitware.com/keys/kitware-archive-latest.asc &&  \
add-apt-repository -y "deb https://apt.kitware.com/ubuntu/ $(lsb_release -cs) main" && \
apt-get update && \
apt-get install -y --no-install-recommends \
   build-essential git gcc g++ cmake \
   ninja-build yasm python3-venv \
   openssh-client \
   libopenslide-dev libwebp-dev \
   libzstd-dev && \
rm -rf /var/lib/apt/lists/*

ROCM_VERSION=$(cat /opt/rocm/.info/version) && \
UBUNTU_CODENAME=$(lsb_release -cs) && \
echo "Detected ROCm version: ${ROCM_VERSION}, Ubuntu codename: ${UBUNTU_CODENAME}" && \
MAJOR=$(echo ${ROCM_VERSION} | cut -d. -f1) && \
MINOR=$(echo ${ROCM_VERSION} | cut -d. -f2) && \
PATCH=$(echo ${ROCM_VERSION} | cut -d. -f3) && \
PATCH=${PATCH:-0} && \
VERNUM=$((MAJOR * 10000 + MINOR * 100 + PATCH)) && \
if [ "${PATCH}" = "0" ]; then ROCM_SHORT_VERSION="${MAJOR}.${MINOR}"; else ROCM_SHORT_VERSION="${MAJOR}.${MINOR}.${PATCH}"; fi

if ! dpkg -s amdgpu-install >/dev/null 2>&1; then \
   rm -f /etc/apt/sources.list.d/amdgpu.list /etc/apt/sources.list.d/rocm.list && \
   AMDGPU_URL="https://repo.radeon.com/amdgpu-install/${ROCM_SHORT_VERSION}/ubuntu/${UBUNTU_CODENAME}/amdgpu-install_${ROCM_SHORT_VERSION}.${VERNUM}-1_all.deb" && \
   echo "Downloading: ${AMDGPU_URL}" && \
   wget "${AMDGPU_URL}" -O amdgpu-install.deb && \
   apt-get update && \
   DEBIAN_FRONTEND=noninteractive apt-get install -y ./amdgpu-install.deb && \
   rm amdgpu-install.deb; \
else \
   echo "amdgpu-install already present, skipping install"; \
fi && \
apt-get update && \
apt-get install -y --no-install-recommends amdgpu-lib && \
apt-get install -y --no-install-recommends rocjpeg rocjpeg-dev && \
rm -rf /var/lib/apt/lists/*

  3. Create the Python virtual environment.

    python3 -m venv monai_dev
source monai_dev/bin/activate
pip install --upgrade pip
pip install torch torchvision torchaudio      \
            --index-url https://download.pytorch.org/whl/rocm7.1
pip install amd-hipcim --extra-index-url=https://pypi.amd.com/rocm-${ROCM_VERSION}/simple/

  4. Set the environment variables.

    export HIP_PATH=/opt/rocm
export PATH=$HIP_PATH/bin:$PATH
export ROCM_PATH=/opt/rocm
export LD_LIBRARY_PATH=$HIP_PATH/lib:$LD_LIBRARY_PATH
export ROCM_HOME=/opt/rocm
export AMDGPU_TARGETS=gfx942
export HIP_VISIBLE_DEVICES=0

Installing using package manager#

For MONAI on ROCm installation using a package manager, follow the steps given in this section.

  1. Install the optional system dependencies depending on the workload.

    pip install ITK nibabel gdown tqdm lmdb psutil pandas einops mlflow \
            pynrrd clearml transformers pydicom fire ignite         \
            parameterized tensorboard pytorch-ignite onnx

  2. Install MONAI on ROCm from the AMD PyPI repository.

    pip install amd-monai --extra-index-url=https://pypi.amd.com/rocm-${ROCM_VERSION}/simple

Building from source#

To build MONAI on ROCm from source, follow the steps given in this section.

  1. Download the latest version of MONAI on ROCm from the GitHub repository.

    git clone git@github.com:ROCm-LS/monai.git
cd monai

  2. Create and activate the development environment for building MONAI on ROCm.

    pip install -r requirements-dev.txt -c amd-constraints.txt --build-constraint amd-constraints.txt

  3. Build and install MONAI on ROCm on a ROCm-based AMD system using the development environment.

    • To build and install the development version of MONAI on ROCm, use:

      BUILD_MONAI=1 FORCE_CUDA=1 python3 setup.py develop

    • To build and package an optimized wheel for installation, use:

      BUILD_MONAI=1 FORCE_CUDA=1 python3 setup.py develop -O1 bdist_wheel

      The preceding command builds the package in non-debug mode and the wheel file is generated under the dist directory.

Verify installation#

Use these commands to verify the MONAI on ROCm installation:

  • Print the MONAI on ROCm version.

    $ python -c "import monai; print(monai.__version__)"

1.5.2

  • Print the MONAI on ROCm package info.

    $ pip show -v amd-monai

Name: amd-monai
Version: 1.5.2
Summary: AI Toolkit for Healthcare Imaging
Home-page: https://rocm.docs.amd.com/projects/monai/en/latest/
Author: AMD Corporation
Author-email:
License: Apache License 2.0
Location: /scratch/users/souchatt/docker/souchatt_monai/monai
Editable project location: /scratch/users/souchatt/docker/souchatt_monai/monai
Requires: numpy, torch
Required-by:
Metadata-Version: 2.1
Installer:
Classifiers:
   Intended Audience :: Developers
   Intended Audience :: Education
   Intended Audience :: Science/Research
   Intended Audience :: Healthcare Industry
   Programming Language :: C++
   Programming Language :: Python :: 3
   Programming Language :: Python :: 3.12
   Topic :: Scientific/Engineering
   Topic :: Scientific/Engineering :: Artificial Intelligence
   Topic :: Scientific/Engineering :: Medical Science Apps.
   Topic :: Scientific/Engineering :: Information Analysis
   Topic :: Software Development
   Topic :: Software Development :: Libraries
   Typing :: Typed
Entry-points:
Project-URLs:
   Documentation, https://rocm.docs.amd.com/projects/monai/en/latest/
   Bug Tracker, https://github.com/ROCm-LS/monai/issues
   Source Code, https://github.com/ROCm-LS/monai/
Contents