Skip to content

ROCm Container with PyTorch 2.4 for ComfyUI with RX570/RX580/RX590 aka Polaris AMD GPU

Notifications You must be signed in to change notification settings

robertrosenbusch/gfx803_rocm61_pt24

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

13 Commits
 
 
 
 
 
 

Repository files navigation

ROCm 6.1.3, PyTorch 2.4, Torchvision 0.19 with AMD GFX803 aka AMD Polaris aka AMD RX570/RX580/RX590

This repo provides a docker buildfile based on the original ROCm-Dockerimage to compile PyTorch and Torchvision for the AMD RX570/RX580/RX590 generation. PyTorch, Torchvision and rocBLAS-Library are not compiled to use the GPU-Polaris generation in the original PIP repository. And of course not compiled too in the official ROCm-PyTorch Dockerfile. However, if Polaris 20/21 GPU support is to be used in ComfyUI, there is no way around newly compiled PyTorch and Torchvision whl/wheel python files. And in ROCm 6.X you have to recompile rocBLAS-Library too. That what this Docker Buildfile will do for you.

ROCm-6.1.3 in a Dockerfile

OS linux Python ROCm PyTorch Torchvision GPU
Ubuntu-22.04.2 6.X 3.10.10 6.1.3 2.4.0 0.19.0 RX570/580/590 aka Polaris 20/21 aka GCN 4
  • It is not necessary to install the entire ROCm-Stack on the host system. Unless you want to use something to optimize your GPU via rocm-smi. In my case, I need the rocm stuff to reduce the power consumption of my RX570 GPU to 145 watts with rocm-smi --setpoweroverdrive 145 && watch -n2 rocm-smi every time I restart the container.
  1. install the docker-subsystem / docker.io on your linux system
  2. download the latest file version of this github-repos
  3. build your Docker image via docker build . -t 'rocm61_pt24:latest'
  4. start the container via:
  5. docker run -it --device=/dev/kfd --device=/dev/dri --group-add=video --ipc=host --cap-add=SYS_PTRACE --security-opt seccomp=unconfined -p 8188:8188 rocm61_pt24:latest
  6. install ComfyUI and download a Model inside the container
  7. After installing ComfyUI reinstall pytorch and torchvision wheels into your ComfyUI-Python-Environment. you find the Polaris compiled Python-Wheel-Files into the "/pytorch/dist" and "/vision/dist" Directory

About

ROCm Container with PyTorch 2.4 for ComfyUI with RX570/RX580/RX590 aka Polaris AMD GPU

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published