verl on ROCm installation

verl on ROCm installation#

2026-08-19

2 min read time

Applies to Linux

System requirements#

To use verl 0.7.1, you need the following prerequisites:

  • ROCm version: 7.0.2

  • Operating system: Ubuntu 22.04

  • GPU platform: AMD Instinct™ MI300X, MI325X, MI355X

  • PyTorch: 2.9.1

  • Python: 3.12

  • vLLM: 0.20.2

Install verl#

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

Use a prebuilt Docker image with verl pre-installed#

The recommended way to set up a verl environment and avoid potential installation issues is with Docker. The tested, prebuilt image includes verl, PyTorch, ROCm, and other dependencies.

Prebuilt Docker images with verl configured for ROCm are available on Docker Hub.

  1. Pull the Docker image

    docker pull rocm/verl:verl-0.7.1.amd0_rocm7.0.2_ubuntu22.04_py3.12_vllm0.20.2

  2. Launch and connect to the Docker container

    docker run -it \
   --name verl-release \
   --device /dev/kfd \
   --device /dev/dri \
   --privileged \
   --network=host \
   --group-add video \
   --cap-add SYS_PTRACE \
   --security-opt seccomp=unconfined \
   --shm-size 2048g \
   --ulimit memlock=-1 \
   --ulimit stack=67108864 \
   -w /workspace \
   rocm/verl:verl-0.7.1.amd0_rocm7.0.2_ubuntu22.04_py3.12_vllm0.20.2 \
   /bin/bash

Build your own Docker image#

  1. Download the Dockerfile from the release/0.7.1.amd0 branch:

    curl -O https://raw.githubusercontent.com/AMD-Ecosystem/verl/release/0.7.1.amd0/docker/rocm/verl0.7.1-amd0/Dockerfile

  2. Build the Docker image:

    docker build -t verl-release-v0.7.1amd0 .

  3. Run the Docker container:

    docker run -it \
   --name verl-release \
   --device /dev/kfd \
   --device /dev/dri \
   --privileged \
   --network=host \
   --group-add video \
   --cap-add=SYS_PTRACE \
   --security-opt seccomp=unconfined \
   --shm-size=2048g \
   --ulimit memlock=-1 \
   --ulimit stack=67108864 \
   -w /workspace \
   verl-release-v0.7.1amd0 \
   /bin/bash

    Note

    The --shm-size parameter allocates shared memory for the container. It can be adjusted based on your system’s resources.

Test the verl installation#

Once connected to the Docker container, verify that verl is installed:

pip list | grep verl
verl    0.7.1        /app