guides

Use an NVIDIA GPU

Run DeepLM Ops on an NVIDIA GPU, or on the CPU.

Which build you get

Docker picks the right build for your machine.

  • Intel and AMD machines get the amd64 build, which uses an NVIDIA GPU when you pass --gpus all.
  • Apple silicon and other Arm machines get a CPU-only arm64 build, where --gpus all does nothing.

Add the GPU flag

Add --gpus all right after docker run -d, not at the end. Anything after the image name is passed to dlm-ops as its command.

bash
docker run -d --gpus all --name dlm-ops ... ghcr.io/deeplm/dlm-ops-dev:0.1.1

Requirements

  • Linux: install the NVIDIA Container Toolkit.
  • Windows: use Docker Desktop with WSL 2.