Skip to content

Latest commit

 

History

History
167 lines (110 loc) · 5.19 KB

README.md

File metadata and controls

167 lines (110 loc) · 5.19 KB

Dev Container Setup for ComfyStream

This guide will help you set up and run a development container for ComfyStream using Visual Studio Code (VS Code).

Prerequisites

Host Setup

Clone the Repository

First, clone the comfystream repository:

git clone https://github.com/yondonfu/comfystream.git
cd comfystream

Download or Build Base Docker Image

The livepeer/comfyui-base:latest image provides a ComfyUI workspace for ComfyStream development. You may either pull the base docker image or build it:

  • Pull from Dockerhub:
docker pull livepeer/comfyui-base:latest
  • Build the base image:
docker build -f docker/Dockerfile.base -t livepeer/comfyui-base:latest .

Host Configuration

On your host system, create directories to store models and engines:

mkdir -p ~/models/ComfyUI--models ~/models/ComfyUI--output

Note

This step should be ran on your host machine before attempting to start the container.

If you would like to use a different path to store models, open .devcontainer/devcontainer.json file and update the source to map to the correct paths to your host system. Here is an example configuration:

{
  "mounts": [
    "source=/path/to/your/model-files,target=/ComfyUI/models/ComfyUI--models,type=bind",
    "source=/path/to/your/output-files,target=/ComfyUI/models/ComfyUI--output,type=bind"
  ]
}

Replace /path/to/your/model-files and path/to/your/output-files with the path to your models and output folders on your host machine.

Dev Container Setup

  1. Open the comfystream repository in VS Code.
  2. From VS Code, reload the folder as a devcontainer:
    • Open the Command Palette (Ctrl+Shift+P or Cmd+Shift+P on macOS).
    • Select Remote-Containers: Reopen in Container.
  3. Wait for the container to build and start.

Starting ComfyUI

Start ComfyUI:

cd /workspace/comfystream/ComfyUI
conda activate comfyui
python main.py --listen

When using TensorRT engine enabled workflows, you should include the ---disable-cuda-malloc flag as shown below:

cd /workspace/comfystream/ComfyUI
conda activate comfyui
python main.py --listen --disable-cuda-malloc

Starting ComfyStream

Start ComfyStream:

cd /workspace/comfystream
conda activate comfystream
python server/app.py --workspace /workspace/ComfyUI --media-ports=5678 --host=0.0.0.0 --port 8889

Optionally, you can also start the ComfyStream UI to view the stream:

cd /workspace/comfystream/ui
npm run dev:https

Running Example Workflows

To run example workflows, you need to download models and build TensorRT engines. You can do this from within the dev container by running the following command in the terminal:

prepare_examples

Alternatively, you can follow the steps below.

Download models

From within the dev container, download models to run the example workflows:

cd /workspace/comfystream
conda activate comfystream
python src/comfystream/scripts/setup_models.py --workspace /workspace/ComfyUI

For more info about configuring model downloads, see src/comfystream/scripts/README.md

By following these steps, you should be able to set up and run your development container for ComfyStream efficiently.

Building the DepthAnything Engine

After downloading models, it is necessary to compile TensorRT engines for the example workflow.

Note

Engine files must be compiled on the same GPU hardware/architecture that they will be used on. This step must be run manually after starting the devcontainer. You may use either conda environment for this step.

  1. Run the export_trt.py script from the directory of the onnx file:

    cd /workspace/ComfyUI/models/tensorrt/depth-anything
python /workspace/ComfyUI/custom_nodes/ComfyUI-Depth-Anything-Tensorrt/export_trt.py

Debugging ComfyStream and ComfyUI

The launch.json includes sample launch configurations for ComfyStream and ComfyUI.

Setting the Python Environment

Conda is initialized in the bash shell with no environment activated to provide better interoperability with VS Code Shell Integration.

VS Code will automatically activate the comfystream environment, unless you change it:

  1. From VSCode, press Ctrl-Shift-P.
  2. Choose Python: Select Interpreter.
  3. Select comfystream or comfyui.
  4. Open a new terminal, you will see the environment name to the left of the bash terminal.

Alternatively, you may activate an environment manually with conda activate comfyui or conda activate comfystream

[!NOTE] For more information, see Python environments in VS Code

Additional Resources