README.adoc

MEGAPAK for CUDA 12.4

View on <Docker Hub>

• Derived from cu121-megapak

• Dev kits:

  • CUDA dev kit (12.4)

  • Python dev package (3.12)

  • GCC C++ (13)

  • OpenCV-devel

  • CMake, Ninja…​

• Latest stable version of xFormers + PyTorch

• Tools:

  • Vim, Fish, fd…​

Usage

Run with Docker

mkdir -p storage

docker run -it --rm \
  --name comfyui-cu124-mega \
  --gpus all \
  -p 8188:8188 \
  -v "$(pwd)"/storage:/root \
  -e CLI_ARGS="--fast" \
  yanwk/comfyui-boot:cu124-megapak

Run with Podman

mkdir -p storage

podman run -it --rm \
  --name comfyui-cu124-mega \
  --device nvidia.com/gpu=all \
  --security-opt label=disable \
  -p 8188:8188 \
  -v "$(pwd)"/storage:/root \
  -e CLI_ARGS="--fast" \
  docker.io/yanwk/comfyui-boot:cu124-megapak

Skip downloading

The container will run the download script on the first start and will create an empty file .download-complete as a marker when the download is complete.

If the download does not complete, on the next start, the download script will resume downloading (using aria2’s mechanism).

You can create the .download-complete file to skip the download script.

mkdir -p storage
touch storage/.download-complete

CLI_ARGS Reference

args	description
--lowvram	If your GPU only has 4GB VRAM.
--novram	If adding --lowvram still out-of-memory.
--cpu	Run on CPU. It’s pretty slow.
--use-pytorch-cross-attention	If you don’t want to use xFormers. This may perform well on WSL2, but significantly slower on Linux hosts.
--preview-method taesd	Enable higher-quality previews with TAESD. ComfyUI-Manager would override this (settings available in Manager UI).
--front-end-version Comfy-Org/ComfyUI_frontend@latest	Use the most up-to-date frontend version.
--fast	Enable experimental optimizations. Currently the only optimization is float8_e4m3fn matrix multiplication on 4000/ADA series Nvidia cards or later. Might break things/lower quality. See the commit.

More CLI_ARGS available at ComfyUI.

Environment Variables Reference

Variable	Example Value	Memo
HTTP_PROXY HTTPS_PROXY	http://localhost:1081 http://localhost:1081	Set HTTP proxy.
PIP_INDEX_URL	'https://pypi.org/simple'	Set mirror site for Python Package Index.
HF_ENDPOINT	'https://huggingface.co'	Set mirror site for HuggingFace Hub.
HF_TOKEN	'hf_your_token'	Set HuggingFace Access Token. More
HF_HUB_ENABLE_HF_TRANSFER	1	Enable HuggingFace Hub experimental high-speed file transfers. Only make sense if you have >1000Mbps and VERY STABLE connection (e.g. cloud server). More
TORCH_CUDA_ARCH_LIST	7.5 or '5.2+PTX;6.0;6.1+PTX;7.5;8.0;8.6;8.9+PTX'	Build target for PyTorch and its extensions. For most users, no setup is needed as it will be automatically selected on Linux. When needed, you only need to set one build target just for your GPU. More
CMAKE_ARGS	(Default) '-DBUILD_opencv_world=ON -DWITH_CUDA=ON -DCUDA_FAST_MATH=ON -DWITH_CUBLAS=ON -DWITH_NVCUVID=ON'	Build options for CMAKE for projects using CUDA.