ROCm on Radeon and Ryzen

https://rocm.docs.amd.com/projects/radeon-ryzen/en/latest/
Admin
ROCm on Radeon and Ryzen enables machine learning and AI development and inference workloads on AMD...
Tokens:30,583
Snippets:556
Trust Score:10
Update:1 month ago
Show doc for...
Tokens:
Raw
### Install jaxlib and jax

Source: https://rocm.docs.amd.com/projects/radeon-ryzen/en/latest/docs/install/installrad/native_linux/install-jax.html

Install the core jaxlib and jax packages.

```bash
pip install jaxlib==0.8.2
```

```bash
pip install jax==0.8.2
```

--------------------------------

### Install HuggingFaceHub

Source: https://rocm.docs.amd.com/projects/radeon-ryzen/en/latest/docs/advanced/advancedrad/linux/llm/install-huggingface-transformers.html

Install the HuggingFaceHub client and authenticate the session.

```bash
pip install huggingface-hub
hf auth login
```

--------------------------------

### Expected verification output

Source: https://rocm.docs.amd.com/projects/radeon-ryzen/en/latest/docs/install/installrad/native_linux/install-onnx.html

Example output showing successful installation of MIGraphX and the half library.

```text
root@aus-navi3x-02:/workspace/AMDMIGraphX# dpkg -l | grep migraphx
ii  migraphx                      2.9.0                             amd64        AMD's graph optimizer
ii  migraphx-dev                  2.9.0                             amd64        AMD's graph optimizer
ii  migraphx-tests                2.9.0                             amd64        AMD's graph opt


$ dpkg -l | grep half
ii  half                          1.12.0.60000-91~20.04             amd64        HALF-PRECISION FLOATING POINT LIBRARY
```

--------------------------------

### Install the half library

Source: https://rocm.docs.amd.com/projects/radeon-ryzen/en/latest/docs/install/installrad/native_linux/install-onnx.html

Command to install the half library if it is missing from the system.

```bash
sudo apt install half
```

--------------------------------

### Install MIGraphX

Source: https://rocm.docs.amd.com/projects/radeon-ryzen/en/latest/docs/install/installrad/native_linux/install-migraphx.html

Use the system package manager to install the MIGraphX library.

```bash
$ sudo apt install migraphx
```

--------------------------------

### Install amdgpu-install script for RHEL 10.1

Source: https://rocm.docs.amd.com/projects/radeon-ryzen/en/latest/docs/install/installrad/native_linux/install-radeon.html

Download and install the repository and installer script for RHEL 10.1.

```bash
sudo yum update
wget https://repo.radeon.com/amdgpu-install/7.2.1/rhel/10.1/amdgpu-install-7.2.1.70201-1.el10.noarch.rpm
sudo yum install ./amdgpu-install-7.2.1.70201-1.el10.noarch.rpm
```

--------------------------------

### Install amdgpu-install script for Ubuntu 22.04

Source: https://rocm.docs.amd.com/projects/radeon-ryzen/en/latest/docs/install/installrad/native_linux/install-radeon.html

Download and install the repository and installer script for Ubuntu 22.04.

```bash
sudo apt update
wget https://repo.radeon.com/amdgpu-install/7.2.1/ubuntu/jammy/amdgpu-install_7.2.1.70201-1_all.deb
sudo apt install ./amdgpu-install_7.2.1.70201-1_all.deb
```

--------------------------------

### Install amdgpu-install script for Ubuntu 24.04

Source: https://rocm.docs.amd.com/projects/radeon-ryzen/en/latest/docs/install/installrad/native_linux/install-radeon.html

Download and install the repository and installer script for Ubuntu 24.04.

```bash
sudo apt update
wget https://repo.radeon.com/amdgpu-install/7.2.1/ubuntu/noble/amdgpu-install_7.2.1.70201-1_all.deb
sudo apt install ./amdgpu-install_7.2.1.70201-1_all.deb
```

--------------------------------

### Install Docker on Ubuntu

Source: https://rocm.docs.amd.com/projects/radeon-ryzen/en/latest/docs/install/installrad/native_linux/install-pytorch.html

Command to install the Docker engine on an Ubuntu system.

```bash
sudo apt install docker.io
```

--------------------------------

### Install JAX PJRT wheel

Source: https://rocm.docs.amd.com/projects/radeon-ryzen/en/latest/docs/install/installrad/native_linux/install-jax.html

Install the required PJRT wheel from the Radeon repository.

```bash
pip install https://repo.radeon.com/rocm/manylinux/rocm-rel-7.2.1/jax_rocm7_pjrt-0.8.2%2Brocm7.2.1-py3-none-manylinux_2_28_x86_64.whl
```

--------------------------------

### amd-ttm Usage Examples

Source: https://rocm.docs.amd.com/projects/radeon-ryzen/en/latest/docs/install/installryz/native_linux/install-ryzen.html

Examples for querying, setting, and clearing TTM memory settings.

```bash
amd-ttm
💻 Current TTM pages limit: 16469033 pages (62.82 GB)
💻 Total system memory: 125.65 GB
```

```bash
❯ amd-ttm --set 100
🐧 Successfully set TTM pages limit to 26214400 pages (100.00 GB)
🐧 Configuration written to /etc/modprobe.d/ttm.conf
○ NOTE: You need to reboot for changes to take effect.
Would you like to reboot the system now? (y/n): y
```

```bash
❯ amd-ttm --clear
🐧 Configuration /etc/modprobe.d/ttm.conf removed
Would you like to reboot the system now? (y/n): y
```

--------------------------------

### Install Built ONNX Runtime Wheel

Source: https://rocm.docs.amd.com/projects/radeon-ryzen/en/latest/docs/install/installrad/native_linux/install-onnx.html

Installs the locally built ONNX Runtime wheel file.

```bash
$ pip3 install /onnxruntime/build/Linux/Release/dist/*.whl
```

--------------------------------

### Prepare Python Environment

Source: https://rocm.docs.amd.com/projects/radeon-ryzen/en/latest/docs/install/installrad/native_linux/install-pytorch.html

Initial commands to install pip and upgrade the packaging tools.

```bash
sudo apt install python3-pip -y
```

```bash
pip3 install --upgrade pip wheel
```

--------------------------------

### Install transformers and accelerate

Source: https://rocm.docs.amd.com/projects/radeon-ryzen/en/latest/docs/advanced/advancedrad/linux/llm/install-huggingface-transformers.html

Install the core Huggingface libraries required for model inference.

```bash
pip install transformers
pip install accelerate
```

--------------------------------

### Run Llama-Server

Source: https://rocm.docs.amd.com/projects/radeon-ryzen/en/latest/docs/advanced/advancedrad/windows/llm/llamacpp.html

Start the OpenAI-compatible web server with GPU offloading and Flash Attention enabled.

```bash
# Start the server
# -ngl 99: Offload all layers to your AMD GPU (Crucial for performance)
# -c: Context Length
# -fa: Enable Flash Attention to reduce memory usage and increase speed
.\llama-server.exe -m test_model.gguf -c 2048 -ngl 99 -fa on --port 8080
```

--------------------------------

### Install Transformers library

Source: https://rocm.docs.amd.com/projects/radeon-ryzen/en/latest/docs/advanced/advancedrad/windows/llm/install-huggingface-transformers.html

Install the Huggingface transformers library via pip.

```bash
pip install transformers
```

--------------------------------

### Install MIGraphX node

Source: https://rocm.docs.amd.com/projects/radeon-ryzen/en/latest/docs/advanced/advancedrad/linux/comfyui/installcomfyui.html

Clone and install the MIGraphX custom node for ComfyUI.

```bash
cd ComfyUI/custom_nodes
git clone https://github.com/pnikolic-amd/ComfyUI_MIGraphX.git
cd ComfyUI_MIGraphX
pip install -r requirements.txt
```

--------------------------------

### Install PyTorch ROCm wheels

Source: https://rocm.docs.amd.com/projects/radeon-ryzen/en/latest/docs/advanced/advancedrad/linux/llm/install-huggingface-transformers.html

Install the necessary PyTorch, torchaudio, torchvision, and triton wheels within the virtual environment.

```bash
pip3 install <torch wheel> <torchaudio wheel> <torchvision wheel> <triton wheel>
```

--------------------------------

### Install JAX plugin wheel

Source: https://rocm.docs.amd.com/projects/radeon-ryzen/en/latest/docs/install/installrad/native_linux/install-jax.html

Install the specific plugin wheel for the target Python version.

```bash
pip install https://repo.radeon.com/rocm/manylinux/rocm-rel-7.2.1/jax_rocm7_plugin-0.8.2%2Brocm7.2.1-cp312-cp312-manylinux_2_28_x86_64.whl
```

```bash
pip install https://repo.radeon.com/rocm/manylinux/rocm-rel-7.2.1/jax_rocm7_plugin-0.8.2%2Brocm7.2.1-cp311-cp311-manylinux_2_28_x86_64.whl
```

--------------------------------

### Install ONNX Runtime via PIP

Source: https://rocm.docs.amd.com/projects/radeon-ryzen/en/latest/docs/install/installrad/native_linux/install-onnx.html

Recommended method to install the ONNX Runtime wheel with MIGraphX and ROCm support.

```bash
pip3 uninstall onnxruntime-migraphx
pip3 install onnxruntime-migraphx -f https://repo.radeon.com/rocm/manylinux/rocm-rel-7.2.1/
```

--------------------------------

### Install dependencies for Ubuntu

Source: https://rocm.docs.amd.com/projects/radeon-ryzen/en/latest/docs/install/installrad/native_linux/install-radeon.html

Install required Python development packages for Ubuntu systems.

```bash
sudo apt install python3-setuptools python3-wheel
```

--------------------------------

### Install ROCm SDK components

Source: https://rocm.docs.amd.com/projects/radeon-ryzen/en/latest/docs/advanced/advancedrad/windows/llm/install-huggingface-transformers.html

Install the required ROCm SDK wheels and tarball for Windows.

```cmd
pip install --no-cache-dir ^
    https://repo.radeon.com/rocm/windows/rocm-rel-7.2.1/rocm_sdk_core-7.2.1-py3-none-win_amd64.whl ^
    https://repo.radeon.com/rocm/windows/rocm-rel-7.2.1/rocm_sdk_devel-7.2.1-py3-none-win_amd64.whl ^
    https://repo.radeon.com/rocm/windows/rocm-rel-7.2.1/rocm_sdk_libraries_custom-7.2.1-py3-none-win_amd64.whl ^
    https://repo.radeon.com/rocm/windows/rocm-rel-7.2.1/rocm-7.2.1.tar.gz
```

```powershell
pip install --no-cache-dir `
    https://repo.radeon.com/rocm/windows/rocm-rel-7.2.1/rocm_sdk_core-7.2.1-py3-none-win_amd64.whl `
    https://repo.radeon.com/rocm/windows/rocm-rel-7.2.1/rocm_sdk_devel-7.2.1-py3-none-win_amd64.whl `
    https://repo.radeon.com/rocm/windows/rocm-rel-7.2.1/rocm_sdk_libraries_custom-7.2.1-py3-none-win_amd64.whl `
    https://repo.radeon.com/rocm/windows/rocm-rel-7.2.1/rocm-7.2.1.tar.gz
```

--------------------------------

### Install gcc-gfortran on RHEL

Source: https://rocm.docs.amd.com/projects/radeon-ryzen/en/latest/docs/install/installrad/native_linux/install-jax.html

Install the gcc-gfortran package to provide libgfortran.so, recommended for RHEL distributions.

```bash
sudo dnf install gcc-gfortran
```

--------------------------------

### Install and Configure Shared Memory Tools

Source: https://rocm.docs.amd.com/projects/radeon-ryzen/en/latest/docs/install/installryz/native_linux/install-ryzen.html

Installs the amd-debug-tools via pipx and configures TTM page limits.

```bash
sudo apt install pipx
```

```bash
pipx ensurepath
```

```bash
pipx install amd-debug-tools
```

```bash
amd-ttm
```

```bash
amd-ttm --set <NUM>
```

--------------------------------

### Verify Installation

Source: https://rocm.docs.amd.com/projects/radeon-ryzen/en/latest/docs/install/installryz/native_linux/install-ryzen.html

Checks group membership and verifies GPU detection via rocminfo.

```bash
groups
```

```text
<username> adm cdrom sudo dip video plugdev render lpadmin lxd sambashare
```

```bash
rocminfo
```

```text
*******
Agent 2
*******
  Name:                    gfx1151
  Uuid:                    GPU-XX
  Marketing Name:          Radeon 8050S Graphics
  Vendor Name:             AMD
  [...]
```

--------------------------------

### Import libraries

Source: https://rocm.docs.amd.com/projects/radeon-ryzen/en/latest/docs/install/installrad/native_linux/install-triton.html

Verifies the installation by importing torch and triton modules.

```python
import torch
import triton
import triton.language as tl
```

--------------------------------

### Launch ComfyUI

Source: https://rocm.docs.amd.com/projects/radeon-ryzen/en/latest/docs/advanced/advancedrad/windows/comfyui/installcomfyui.html

Start the ComfyUI application. Ensure you are in the repository's home directory before running.

```bash
python main.py
```

--------------------------------

### Verify MIGraphX and half library installation

Source: https://rocm.docs.amd.com/projects/radeon-ryzen/en/latest/docs/install/installrad/native_linux/install-onnx.html

Use dpkg to check if the required MIGraphX and half library packages are installed on the system.

```bash
$ dpkg -l | grep migraphx
$ dpkg -l | grep half
```

--------------------------------

### Install PyTorch for ROCm

Source: https://rocm.docs.amd.com/projects/radeon-ryzen/en/latest/docs/advanced/advancedrad/windows/llm/install-huggingface-transformers.html

Install PyTorch, torchvision, and torchaudio packages specifically built for ROCm support on Windows.

```cmd
pip install --no-cache-dir ^
    https://repo.radeon.com/rocm/windows/rocm-rel-7.2.1/torch-2.9.1%2Brocm7.2.1-cp312-cp312-win_amd64.whl ^
    https://repo.radeon.com/rocm/windows/rocm-rel-7.2.1/torchaudio-2.9.1%2Brocm7.2.1-cp312-cp312-win_amd64.whl ^
    https://repo.radeon.com/rocm/windows/rocm-rel-7.2.1/torchvision-0.24.1%2Brocm7.2.1-cp312-cp312-win_amd64.whl
```

```powershell
pip install --no-cache-dir `
    https://repo.radeon.com/rocm/windows/rocm-rel-7.2.1/torch-2.9.1%2Brocm7.2.1-cp312-cp312-win_amd64.whl `
    https://repo.radeon.com/rocm/windows/rocm-rel-7.2.1/torchaudio-2.9.1%2Brocm7.2.1-cp312-cp312-win_amd64.whl `
    https://repo.radeon.com/rocm/windows/rocm-rel-7.2.1/torchvision-0.24.1%2Brocm7.2.1-cp312-cp312-win_amd64.whl
```

--------------------------------

### Run Torch-MIGraphX Unit Tests

Source: https://rocm.docs.amd.com/projects/radeon-ryzen/en/latest/docs/install/installrad/native_linux/install-migraphx.html

Execute the test suite to ensure the installation is functioning correctly.

```bash
pytest ./torch_migraphx/tests
```

--------------------------------

### Create and activate virtual environment

Source: https://rocm.docs.amd.com/projects/radeon-ryzen/en/latest/docs/advanced/advancedryz/windows/comfyui/installcomfyui.html

Sets up a dedicated Python virtual environment for the ComfyUI installation.

```bash
python -m venv comfyui-venv
comfyui-venv\Scripts\activate
```

--------------------------------

### Verify ONNX Runtime installation

Source: https://rocm.docs.amd.com/projects/radeon-ryzen/en/latest/docs/install/installrad/native_linux/install-onnx.html

Use this command to check available execution providers in the current Python environment.

```bash
python3 -c "import onnxruntime as ort; print(ort.get_available_providers())"
```

--------------------------------

### Add EPEL repository for RHEL 10

Source: https://rocm.docs.amd.com/projects/radeon-ryzen/en/latest/docs/prerequisites/prerequisitesrad.html

Downloads and installs the EPEL repository package for RHEL 10 systems.

```bash
wget https://dl.fedoraproject.org/pub/epel/epel-release-latest-10.noarch.rpm
sudo rpm -ivh epel-release-latest-10.noarch.rpm
```

--------------------------------

### List available usecases

Source: https://rocm.docs.amd.com/projects/radeon-ryzen/en/latest/docs/install/installrad/native_linux/install-radeon.html

Displays the available usecase parameters for the amdgpu-install script.

```bash
sudo amdgpu-install --list-usecase
```

--------------------------------

### Install Requirements

Source: https://rocm.docs.amd.com/projects/radeon-ryzen/en/latest/docs/advanced/advancedrad/windows/comfyui/installcomfyui.html

Install the necessary Python dependencies for ComfyUI.

```bash
pip install -r requirements.txt
```

--------------------------------

### Install graphics and ROCm

Source: https://rocm.docs.amd.com/projects/radeon-ryzen/en/latest/docs/install/installrad/native_linux/install-radeon.html

Installs the graphics and ROCm components non-interactively.

```bash
amdgpu-install -y --usecase=graphics,rocm
```

--------------------------------

### Start vLLM Docker Container

Source: https://rocm.docs.amd.com/projects/radeon-ryzen/en/latest/docs/advanced/advancedrad/linux/llm/build-docker-image.html

Standard command to initialize a Docker container for vLLM with necessary device and network permissions.

```bash
docker run -it \
  --privileged \
  --device=/dev/kfd \
  --device=/dev/dri \
  --network=host \
  --group-add sudo \
  -w /app/vllm/ \
  --name <container_name> \
<image_name> \
  /bin/bash
```

--------------------------------

### Create and activate virtual environment

Source: https://rocm.docs.amd.com/projects/radeon-ryzen/en/latest/docs/advanced/advancedrad/linux/llm/install-huggingface-transformers.html

Initialize a new Python virtual environment and activate it.

```bash
python3 -m venv llm-venv
source llm-venv/bin/activate
```

--------------------------------

### Install Python venv

Source: https://rocm.docs.amd.com/projects/radeon-ryzen/en/latest/docs/advanced/advancedryz/linux/llm/install-huggingface-transformers.html

Install the Python venv package for Ubuntu 24.04.

```bash
sudo apt install python3.12-venv
```

--------------------------------

### Configure and launch with MIGraphX

Source: https://rocm.docs.amd.com/projects/radeon-ryzen/en/latest/docs/advanced/advancedrad/linux/comfyui/installcomfyui.html

Set the required environment variables and launch ComfyUI with MIGraphX support.

```bash
export PYTHONPATH=/opt/rocm-X.X.X/lib:$PYTHONPATH
```

```bash
cd ComfyUI
# For the best performance, run the following environment variable:
export MIGRAPHX_MLIR_USE_SPECIFIC_OPS="attention"
python3 main.py
```

--------------------------------

### Verify ROCm Installation

Source: https://rocm.docs.amd.com/projects/radeon-ryzen/en/latest/docs/install/installrad/native_linux/install-onnx.html

Checks that ROCm is correctly installed and detects supported GPUs.

```bash
$ rocm-smi
```

```text
======================================= ROCm System Management Interface =======================================
================================================= Concise Info =================================================
Device  [Model : Revision]    Temp    Power  Partitions      SCLK     MCLK   Fan     Perf  PwrCap  VRAM%  GPU% 
        Name (20 chars)       (Edge)  (Avg)  (Mem, Compute)                                                    
================================================================================================================
0       [0x0e0d : 0x00]       32.0°C  73.0W  N/A, N/A        1526Mhz  96Mhz  31.76%  auto  241.0W    0%   50%  
        0x7448                                                                                                 
================================================================================================================
============================================= End of ROCm SMI Log ==============================================
```

--------------------------------

### Install dependencies for RHEL

Source: https://rocm.docs.amd.com/projects/radeon-ryzen/en/latest/docs/install/installrad/native_linux/install-radeon.html

Install required Python development packages for RHEL systems.

```bash
sudo dnf install python3-setuptools python3-wheel
```

--------------------------------

### Install OEM Kernel

Source: https://rocm.docs.amd.com/projects/radeon-ryzen/en/latest/docs/install/installryz/native_linux/install-ryzen.html

Installs the required 6.14-1018 OEM kernel or newer for ROCm on Ryzen.

```bash
sudo apt update && sudo apt install linux-oem-24.04c
```

--------------------------------

### Install required libraries

Source: https://rocm.docs.amd.com/projects/radeon-ryzen/en/latest/docs/install/installrad/native_linux/install-triton.html

Installs matplotlib, pandas, and the nightly PyTorch wheels for ROCm 6.0.

```bash
pip install matplotlib pandas -q
pip install --pre torch torchvision torchaudio --index-url https://download.pytorch.org/whl/nightly/rocm6.0/ -q
```

--------------------------------

### Install Python venv package

Source: https://rocm.docs.amd.com/projects/radeon-ryzen/en/latest/docs/advanced/advancedrad/linux/llm/install-huggingface-transformers.html

Install the venv package for specific Python versions on Ubuntu.

```bash
sudo apt install python3.12-venv
```

```bash
sudo apt install python3.10-venv
```

--------------------------------

### Run MNIST Training Example

Source: https://rocm.docs.amd.com/projects/radeon-ryzen/en/latest/docs/install/installrad/native_linux/install-tensorflow.html

A script to load the MNIST dataset, build a simple neural network, and train it to verify framework functionality.

```python
import tensorflow as tf
print("TensorFlow version:", tf.__version__)
mnist = tf.keras.datasets.mnist

(x_train, y_train), (x_test, y_test) = mnist.load_data()
x_train, x_test = x_train / 255.0, x_test / 255.0
model = tf.keras.models.Sequential([
  tf.keras.layers.Flatten(input_shape=(28, 28)),
  tf.keras.layers.Dense(128, activation='relu'),
  tf.keras.layers.Dropout(0.2),
  tf.keras.layers.Dense(10)
])
predictions = model(x_train[:1]).numpy()
tf.nn.softmax(predictions).numpy()
loss_fn = tf.keras.losses.SparseCategoricalCrossentropy(from_logits=True)
loss_fn(y_train[:1], predictions).numpy()
model.compile(optimizer='adam',
              loss=loss_fn,
              metrics=['accuracy'])
model.fit(x_train, y_train, epochs=5)
model.evaluate(x_test,  y_test, verbose=2)
```

--------------------------------

### Run Torch-MIGraphX Docker Container

Source: https://rocm.docs.amd.com/projects/radeon-ryzen/en/latest/docs/install/installrad/native_linux/install-migraphx.html

Launch the container with necessary device and network permissions.

```bash
sudo docker run -it --network=host --device=/dev/kfd --device=/dev/dri --group-add=video --ipc=host --cap-add=SYS_PTRACE --security-opt seccomp=unconfined torch_migraphx
```

--------------------------------

### Download Llama.cpp Binaries

Source: https://rocm.docs.amd.com/projects/radeon-ryzen/en/latest/docs/advanced/advancedrad/windows/llm/llamacpp.html

Download the AMD-validated Windows binary package using curl.

```bash
curl.exe -o llama-bin-windows.zip "https://repo.radeon.com/rocm/llama.cpp/windows/rocm-rel-7.2.1/llama-b8407-windows-rocm-7.2.1-gfx110X-gfx115X-gfx120X-x64.zip"
```

--------------------------------

### Install ROCm Components

Source: https://rocm.docs.amd.com/projects/radeon-ryzen/en/latest/docs/install/installryz/native_linux/install-ryzen.html

Installs ROCm using the amdgpu-install script with the required --no-dkms flag for Ryzen systems.

```bash
amdgpu-install -y --usecase=rocm --no-dkms
```

--------------------------------

### Install TensorFlow Wheels

Source: https://rocm.docs.amd.com/projects/radeon-ryzen/en/latest/docs/install/installrad/native_linux/install-tensorflow.html

Commands to uninstall existing packages and install specific ROCm-compatible TensorFlow wheels for different Linux distributions.

```bash
pip3 uninstall tensorflow-rocm
pip3 install https://repo.radeon.com/rocm/manylinux/rocm-rel-7.2.1/tensorflow_rocm-2.20.0.dev0%2Bselfbuilt-cp312-cp312-manylinux_2_28_x86_64.whl
```

```bash
pip3 uninstall tensorflow-rocm
pip3 install https://repo.radeon.com/rocm/manylinux/rocm-rel-7.2.1/tensorflow_rocm-2.20.0.dev0%2Bselfbuilt-cp310-cp310-manylinux_2_28_x86_64.whl
```

--------------------------------

### Remove DKMS Packages

Source: https://rocm.docs.amd.com/projects/radeon-ryzen/en/latest/docs/install/installryz/native_linux/install-ryzen.html

Removes dkms packages if they were installed unintentionally.

```bash
sudo apt autoremove amdgpu-dkms dkms
```

--------------------------------

### Start PyTorch Docker container

Source: https://rocm.docs.amd.com/projects/radeon-ryzen/en/latest/docs/install/installrad/native_linux/install-pytorch.html

Commands to launch a container with necessary security and hardware device mappings for ROCm.

```bash
sudo docker run -it \
  --cap-add=SYS_PTRACE \
  --security-opt seccomp=unconfined \
  --device=/dev/kfd \
  --device=/dev/dri \
  --group-add video \
  --ipc=host \
  --shm-size 8G \
  rocm/pytorch:rocm7.2.1_ubuntu24.04_py3.12_pytorch_release_2.9.1
```

```bash
sudo docker run -it \
  --cap-add=SYS_PTRACE \
  --security-opt seccomp=unconfined \
  --device=/dev/kfd \
  --device=/dev/dri \
  --group-add video \
  --ipc=host \
  --shm-size 8G \
  rocm/pytorch:rocm7.2.1_ubuntu22.04_py3.10_pytorch_release_2.9.1
```

--------------------------------

### Update System Packages

Source: https://rocm.docs.amd.com/projects/radeon-ryzen/en/latest/docs/install/installryz/native_linux/install-ryzen.html

Ensures all system packages are up to date after kernel installation.

```bash
sudo apt upgrade -y
```
Tokens:
Raw