ComfyUI_resolution_presets

ComfyUI Resolution Presets

A ComfyUI custom node extension that provides optimized resolution presets for various AI image generation models.

Features

• Pre-configured resolution presets for multiple AI models
• Support for various aspect ratios
• Two node variants: Basic and Advanced
• Easily extensible architecture
• Automatic validation of model/aspect ratio combinations

Supported Models

• Qwen Image - 7 aspect ratios
• SDXL - 9 aspect ratios
• Flux - 11 aspect ratios
• SD 1.5 - 7 aspect ratios

Installation

1. Navigate to your ComfyUI custom nodes directory:

   cd ComfyUI/custom_nodes/
   

2. Clone this repository:

   git clone https://github.com/yourusername/ComfyUI_resolution_presets.git
   

3. Restart ComfyUI

Nodes

Resolution Presets

The basic node that outputs resolution presets for your selected model.

Inputs:

• model - Select the AI model (Qwen Image, SDXL, Flux, SD 1.5)
• aspect_ratio - Select the desired aspect ratio

Outputs:

• ratio (STRING) - The aspect ratio (e.g., "16:9")
• width (INT) - Image width in pixels
• height (INT) - Image height in pixels
• model_name (STRING) - The selected model name

Resolution Presets (Advanced)

An enhanced version with scaling capabilities.

Inputs:

• model - Select the AI model
• aspect_ratio - Select the desired aspect ratio
• scale_multiplier - Scale the resolution (0.25x to 4.0x, default: 1.0)

Outputs:

• ratio (STRING) - The aspect ratio
• width (INT) - Scaled width (rounded to multiple of 8)
• height (INT) - Scaled height (rounded to multiple of 8)
• model_name (STRING) - The selected model name
• total_pixels (INT) - Total pixel count (width × height)

Usage Example

1. Add the "Resolution Presets" node to your workflow
2. Select your target model (e.g., "Flux")
3. Choose an aspect ratio (e.g., "16:9")
4. Connect the width and height outputs to your image generation node
5. Use the model_name output for reference or routing

Resolution Presets Reference

Qwen Image

• 1:1 → 1328×1328
• 16:9 → 1664×928
• 9:16 → 928×1664
• 4:3 → 1472×1140
• 3:4 → 1140×1472
• 3:2 → 1584×1056
• 2:3 → 1056×1584

SDXL

• 1:1 → 1024×1024
• 16:9 → 1344×768
• 9:16 → 768×1344
• 4:3 → 1152×896
• 3:4 → 896×1152
• 3:2 → 1216×832
• 2:3 → 832×1216
• 21:9 → 1536×640
• 9:21 → 640×1536

Flux

• 1:1 → 1024×1024
• 16:9 → 1344×768
• 9:16 → 768×1344
• 4:3 → 1152×896
• 3:4 → 896×1152
• 3:2 → 1216×832
• 2:3 → 832×1216
• 21:9 → 1536×640
• 9:21 → 640×1536
• 2:1 → 1408×704
• 1:2 → 704×1408

SD 1.5

• 1:1 → 512×512
• 16:9 → 682×384
• 9:16 → 384×682
• 4:3 → 608×456
• 3:4 → 456×608
• 3:2 → 624×416
• 2:3 → 416×624

Extending the Node

To add new models or resolutions, edit resolution_presets.py:

RESOLUTION_PRESETS = {
    "Your Model Name": {
        "1:1": (width, height),
        "16:9": (width, height),
        # Add more aspect ratios...
    },
    # Add more models...
}

The extension will automatically:

• Add the new model to the dropdown
• Include new aspect ratios in the sorted list
• Validate combinations at runtime

Best Practices

1. Model-specific ratios: Not all aspect ratios are available for all models. The node will show an error with available options if an invalid combination is selected.

2. Multiple of 8: The Advanced node automatically rounds dimensions to multiples of 8, which is optimal for most diffusion models.

3. Scale multiplier: Use the Advanced node's scale multiplier for quick resolution adjustments while maintaining aspect ratios.

Technical Details

• Category: image/resolution
• Return types: Strongly typed outputs (STRING, INT)
• Validation: Runtime validation with helpful error messages
• Extensibility: Simple dictionary-based configuration

License

MIT License - feel free to use and modify as needed.

Contributing

Contributions are welcome! To add support for new models:

1. Research the optimal resolutions for the model
2. Add the model to RESOLUTION_PRESETS dictionary
3. Test the presets
4. Submit a pull request

Changelog

Version 1.0.0

• Initial release
• Support for Qwen Image, SDXL, Flux, and SD 1.5
• Basic and Advanced node variants
• Automatic aspect ratio sorting