sarman/dgx-spark-playbooks
1
0
Fork 0
mirror of https://github.com/NVIDIA/dgx-spark-playbooks.git synced 2026-04-25 11:23:52 +00:00
Code Issues Actions Packages Projects Releases Wiki Activity

chore: Regenerate all playbooks

Browse Source
This commit is contained in:
GitLab CI 2025-11-05 18:35:04 +00:00
parent af0833dd53
commit fc74163065
1 changed files with 119 additions and 119 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

234
nvidia/comfy-ui/README.md
View File

@ -14,185 +14,185 @@
## Basic idea ## Basic idea
ComfyUI is an open-source web server application for AI image generation using diffusion-based models like SDXL, Flux and others. ComfyUI is an open-source web server application for AI image generation using diffusion-based models like SDXL, Flux and others.
It has a browser-based UI that lets you create, edit and run image generation and editing workflows with multiple steps. It has a browser-based UI that lets you create, edit and run image generation and editing workflows with multiple steps.
Generation and editing steps (e.g. loading a model, adding text or sampling) are configurable in the UI as a node, and you connect nodes with wires to form a workflow. Generation and editing steps (e.g. loading a model, adding text or sampling) are configurable in the UI as a node, and you connect nodes with wires to form a workflow.
ComfyUI uses the host's GPU for inference, so you can install it on your Spark and do all of your image generation and editing directly on device. ComfyUI uses the host's GPU for inference, so you can install it on your Spark and do all of your image generation and editing directly on device.
Workflows are saved as JSON files, so you can version them for future work, collaboration and reproducibility. Workflows are saved as JSON files, so you can version them for future work, collaboration and reproducibility.
## What you'll accomplish # # What you'll accomplish
You'll install and configure ComfyUI on your NVIDIA DGX Spark device so you can use the unified memory to work with large models. You'll install and configure ComfyUI on your NVIDIA DGX Spark device so you can use the unified memory to work with large models.
## What to know before starting # # What to know before starting
- Experience working with Python virtual environments and package management - Experience working with Python virtual environments and package management
- Familiarity with command line operations and terminal usage - Familiarity with command line operations and terminal usage
- Basic understanding of deep learning model deployment and checkpoints - Basic understanding of deep learning model deployment and checkpoints
- Knowledge of container workflows and GPU acceleration concepts - Knowledge of container workflows and GPU acceleration concepts
- Understanding of network configuration for accessing web services - Understanding of network configuration for accessing web services
## Prerequisites # # Prerequisites
**Hardware Requirements:** **Hardware Requirements:**
- NVIDIA Spark device with Blackwell architecture - NVIDIA Spark device with Blackwell architecture
- Minimum 8GB GPU memory for Stable Diffusion models - Minimum 8GB GPU memory for Stable Diffusion models
- At least 20GB available storage space - At least 20GB available storage space
**Software Requirements:** **Software Requirements:**
- Python 3.8 or higher installed: `python3 --version` - Python 3.8 or higher installed: `python3 --version`
- pip package manager available: `pip3 --version` - pip package manager available: `pip3 --version`
- CUDA toolkit compatible with Blackwell: `nvcc --version` - CUDA toolkit compatible with Blackwell: `nvcc --version`
- Git version control: `git --version` - Git version control: `git --version`
- Network access to download models from Hugging Face - Network access to download models from Hugging Face
- Web browser access to `<SPARK_IP>:8188` port - Web browser access to `<SPARK_IP>:8188` port
## Ancillary files # # Ancillary files
All required assets can be found [in the ComfyUI repository on GitHub](https://github.com/comfyanonymous/ComfyUI) All required assets can be found [in the ComfyUI repository on GitHub](https://github.com/comfyanonymous/ComfyUI)
- `requirements.txt` - Python dependencies for ComfyUI installation - `requirements.txt` - Python dependencies for ComfyUI installation
- `main.py` - Primary ComfyUI server application entry point - `main.py` - Primary ComfyUI server application entry point
- `v1-5-pruned-emaonly-fp16.safetensors` - Stable Diffusion 1.5 checkpoint model - `v1-5-pruned-emaonly-fp16.safetensors` - Stable Diffusion 1.5 checkpoint model
## Time & risk # # Time & risk
* **Estimated time:** 30-45 minutes (including model download) * **Estimated time:** 30-45 minutes (including model download)
* **Risk level:** Medium * **Risk level:** Medium
* Model downloads are large (~2GB) and may fail due to network issues * Model downloads are large (~2GB) and may fail due to network issues
* Port 8188 must be accessible for web interface functionality * Port 8188 must be accessible for web interface functionality
* **Rollback:** Virtual environment can be deleted to remove all installed packages. Downloaded models can be removed manually from the checkpoints directory. * **Rollback:** Virtual environment can be deleted to remove all installed packages. Downloaded models can be removed manually from the checkpoints directory.
## Instructions ## Instructions
## Step 1. Verify system prerequisites ## Step 1. Verify system prerequisites
Check that your NVIDIA Spark device meets the requirements before proceeding with installation. Check that your NVIDIA Spark device meets the requirements before proceeding with installation.
```bash ```bash
python3 --version python3 --version
pip3 --version pip3 --version
nvcc --version nvcc --version
nvidia-smi nvidia-smi
``` ```
Expected output should show Python 3.8+, pip available, CUDA toolkit and GPU detection. Expected output should show Python 3.8+, pip available, CUDA toolkit and GPU detection.
## Step 2. Create Python virtual environment # # Step 2. Create Python virtual environment
You will install ComfyUI on your host system, so you should create an isolated environment to avoid conflicts with system packages. You will install ComfyUI on your host system, so you should create an isolated environment to avoid conflicts with system packages.
```bash ```bash
python3 -m venv comfyui-env python3 -m venv comfyui-env
source comfyui-env/bin/activate source comfyui-env/bin/activate
``` ```
Verify the virtual environment is active by checking the command prompt shows `(comfyui-env)`. Verify the virtual environment is active by checking the command prompt shows `(comfyui-env)`.
## Step 3. Install PyTorch with CUDA support # # Step 3. Install PyTorch with CUDA support
Install PyTorch with CUDA 12.9 support. Install PyTorch with CUDA 12.9 support.
```bash ```bash
pip3 install torch torchvision --index-url https://download.pytorch.org/whl/cu129 pip3 install torch torchvision --index-url https://download.pytorch.org/whl/cu129
``` ```
This installation targets CUDA 12.9 compatibility with Blackwell architecture GPUs. This installation targets CUDA 12.9 compatibility with Blackwell architecture GPUs.
## Step 4. Clone ComfyUI repository # # Step 4. Clone ComfyUI repository
Download the ComfyUI source code from the official repository. Download the ComfyUI source code from the official repository.
```bash ```bash
git clone https://github.com/comfyanonymous/ComfyUI.git git clone https://github.com/comfyanonymous/ComfyUI.git
cd ComfyUI/ cd ComfyUI/
``` ```
## Step 5. Install ComfyUI dependencies # # Step 5. Install ComfyUI dependencies
Install the required Python packages for ComfyUI operation. Install the required Python packages for ComfyUI operation.
```bash ```bash
pip install -r requirements.txt pip install -r requirements.txt
``` ```
This installs all necessary dependencies including web interface components and model handling libraries. This installs all necessary dependencies including web interface components and model handling libraries.
## Step 6. Download Stable Diffusion checkpoint # # Step 6. Download Stable Diffusion checkpoint
Navigate to the checkpoints directory and download the Stable Diffusion 1.5 model. Navigate to the checkpoints directory and download the Stable Diffusion 1.5 model.
```bash ```bash
cd models/checkpoints/ cd models/checkpoints/
wget https://huggingface.co/Comfy-Org/stable-diffusion-v1-5-archive/resolve/main/v1-5-pruned-emaonly-fp16.safetensors wget https://huggingface.co/Comfy-Org/stable-diffusion-v1-5-archive/resolve/main/v1-5-pruned-emaonly-fp16.safetensors
cd ../../ cd ../../
``` ```
The download will be approximately 2GB and may take several minutes depending on network speed. The download will be approximately 2GB and may take several minutes depending on network speed.
## Step 7. Launch ComfyUI server # # Step 7. Launch ComfyUI server
Start the ComfyUI web server with network access enabled. Start the ComfyUI web server with network access enabled.
```bash ```bash
python main.py --listen 0.0.0.0 python main.py --listen 0.0.0.0
``` ```
The server will bind to all network interfaces on port 8188, making it accessible from other devices. The server will bind to all network interfaces on port 8188, making it accessible from other devices.
## Step 8. Validate installation # # Step 8. Validate installation
Check that ComfyUI is running correctly and accessible via web browser. Check that ComfyUI is running correctly and accessible via web browser.
```bash ```bash
curl -I http://localhost:8188 curl -I http://localhost:8188
``` ```
Expected output should show HTTP 200 response indicating the web server is operational. Expected output should show HTTP 200 response indicating the web server is operational.
Open a web browser and navigate to `http://<SPARK_IP>:8188` where `<SPARK_IP>` is your device's IP address. Open a web browser and navigate to `http://<SPARK_IP>:8188` where `<SPARK_IP>` is your device's IP address.
## Step 9. Optional - Cleanup and rollback # # Step 9. Optional - Cleanup and rollback
If you need to remove the installation completely, follow these steps: If you need to remove the installation completely, follow these steps:
> [!WARNING] > [!WARNING]
> This will delete all installed packages and downloaded models. > This will delete all installed packages and downloaded models.
```bash ```bash
deactivate deactivate
rm -rf comfyui-env/ rm -rf comfyui-env/
rm -rf ComfyUI/ rm -rf ComfyUI/
``` ```
To rollback during installation, press `Ctrl+C` to stop the server and remove the virtual environment. To rollback during installation, press `Ctrl+C` to stop the server and remove the virtual environment.
## Step 10. Optional - Next steps # # Step 10. Optional - Next steps
Test the installation with a basic image generation workflow: Test the installation with a basic image generation workflow:
1. Access the web interface at `http://<SPARK_IP>:8188` 1. Access the web interface at `http://<SPARK_IP>:8188`
2. Load the default workflow (should appear automatically) 2. Load the default workflow (should appear automatically)
3. Click "Run" to generate your first image 3. Click "Run" to generate your first image
4. Monitor GPU usage with `nvidia-smi` in a separate terminal 4. Monitor GPU usage with `nvidia-smi` in a separate terminal
The image generation should complete within 30-60 seconds depending on your hardware configuration. The image generation should complete within 30-60 seconds depending on your hardware configuration.
## Troubleshooting ## Troubleshooting
| Symptom | Cause | Fix | | Symptom | Cause | Fix |
|---------|-------|-----| |---------|-------|-----|
| PyTorch CUDA not available | Incorrect CUDA version or missing drivers | Verify `nvcc --version` matches cu129, reinstall PyTorch | | PyTorch CUDA not available | Incorrect CUDA version or missing drivers | Verify `nvcc --version` matches cu129, reinstall PyTorch |
| Model download fails | Network connectivity or storage space | Check internet connection, verify 20GB+ available space | | Model download fails | Network connectivity or storage space | Check internet connection, verify 20GB+ available space |
| Web interface inaccessible | Firewall blocking port 8188 | Configure firewall to allow port 8188, check IP address | | Web interface inaccessible | Firewall blocking port 8188 | Configure firewall to allow port 8188, check IP address |
| Out of GPU memory errors after manually flushing buffer cache | Insufficient VRAM for model | Use smaller models or enable CPU fallback mode | | Out of GPU memory errors after manually flushing buffer cache | Insufficient VRAM for model | Use smaller models or enable CPU fallback mode |
> [!NOTE] > [!NOTE]
> DGX Spark uses a Unified Memory Architecture (UMA), which enables dynamic memory sharing between the GPU and CPU. > DGX Spark uses a Unified Memory Architecture (UMA), which enables dynamic memory sharing between the GPU and CPU.
> With many applications still updating to take advantage of UMA, you may encounter memory issues even when within > With many applications still updating to take advantage of UMA, you may encounter memory issues even when within
> the memory capacity of DGX Spark. If that happens, manually flush the buffer cache with: > the memory capacity of DGX Spark. If that happens, manually flush the buffer cache with:
```bash ```bash
sudo sh -c 'sync; echo 3 > /proc/sys/vm/drop_caches' sudo sh -c 'sync; echo 3 > /proc/sys/vm/drop_caches'
``` ```