A custom ComfyUI node for testing multiple LoRA models across different strength values. Automatically generates model/clip pairs for each LoRA-strength combination and includes an image saver node with text overlay for easy identification.
- Test multiple LoRAs with different strength values in a single workflow
- Automatic generation of model/clip pairs for each combination
- Image saver node with text overlay showing LoRA name and strength
- Optimized performance with LoRA dictionary caching
- Font caching for faster image processing
- Color dropdowns for easy customization
cd ComfyUI/custom_nodes
git clone https://github.com/Hearmeman24/ComfyUI-LoRAPlotNode.gitThen restart ComfyUI.
- Add the "LoRA Plot Node" to your workflow
- Connect your base model and CLIP
- Select up to 10 LoRAs from the dropdown menus
- Enter comma-separated strength values (e.g., "0.8,0.9,1.0")
- Connect the outputs to your KSampler nodes
The node will output:
- Multiple model outputs (one per LoRA-strength combination)
- Multiple CLIP outputs (one per LoRA-strength combination)
- Metadata strings for each combination
- Add the "LoRA Plot Image Saver" node
- Connect your decoded images
- Connect the metadata output from LoRA Plot Node
- Customize text color, background color, font size, padding, and opacity
- The node will overlay text on each image showing the LoRA name and strength
- Load Model -> LoRA Plot Node -> KSampler -> VAE Decode -> LoRA Plot Image Saver -> Save Image
The LoRA Plot Node will automatically create multiple model/clip pairs, and ComfyUI will execute the workflow once for each combination.
- LoRA files are cached in memory (up to 10 LoRAs) to avoid redundant disk I/O
- Fonts are cached by size for faster image processing
- LoRA files are loaded once per LoRA, not per strength value
- ComfyUI
- PIL/Pillow (usually included with ComfyUI)
- NumPy (usually included with ComfyUI)
- PyTorch (usually included with ComfyUI)
MIT