Run this workflow on InstaSD
Get started in minutes! Run this ComfyUI workflow online - no setup required.
Description
Enhance your AI-generated images with Flux Dev Q8 GGUF and ControlNet Union, incorporating LORA for advanced fine-tuning. This ComfyUI workflow allows for detailed control over image generation while leveraging Flux's ControlNet union for precise image conditioning. Ideal for artists, designers, and AI enthusiasts looking to improve AI-generated compositions.
🎯 Features
- Flux Dev Q8 GGUF Support – Utilizes optimized GGUF models for efficient image generation.
- ControlNet Union – Enables precise control over image conditioning for better structure and consistency.
- LORA Integration – Fine-tune AI-generated images with additional LORA models for better style and detail.
- Customizable Preprocessors – Select from various preprocessors such as Depth Anything for better depth perception.
- Resolution Control – Easily adjust resolution and aspect ratio to match your needs.
💡 Use Cases
- Art & Illustration – Create detailed and consistent AI-assisted illustrations.
- Concept Design – Develop high-quality AI-enhanced concept art.
- AI-Assisted Creativity – Generate better AI-powered designs with refined image conditioning.
- Stylized Image Generation – Use LORA to enhance specific styles in AI-generated images.
⚙️ How It Works
- Load the Image – Upload an image to work with or start from scratch using an Empty Latent Image node.
- Select the Aspect Ratio – Choose a resolution that matches your needs.
- Configure ControlNet – Load the Flux ControlNet Union model for structure guidance.
- Apply Preprocessors – Use Depth Anything Preprocessor or other available options.
- Set Up the AI Model – Load Flux Dev Q8 GGUF with LORA fine-tuning for enhanced image generation.
- Add Prompt & Adjust Strength – Input your description and adjust the conditioning strength.
- Run the Generation – The KSampler node processes the image, applying all settings.
- Preview & Save – View the final output and save your AI-enhanced image.
Credits: pixaroma
Models
| File | Destination | Source |
|---|---|---|
| flux-dev-controlnet-union.safetensors | /ComfyUI/models/controlnet | Download |
| flux1-dev-Q8_0.gguf | /ComfyUI/models/unet | Download |
| ae.safetensors | /ComfyUI/models/vae | Download |
| t5-v1_1-xxl-encoder-Q8_0.gguf | /ComfyUI/models/clip | Download |
| clip_l.safetensors | /ComfyUI/models/clip | Download |
| aidmaHyperrealism-FLUX-V0.1.safetensors | /ComfyUI/models/loras | Download |
Nodes
VAEDecodeeasy showAnythingEmptySD3LatentImageSaveImageLabel (rgthree)PreviewImageAIO_PreprocessorControlNetLoaderLoadImageFluxResolutionNodeUnetLoaderGGUFFluxGuidanceControlNetApplyAdvancedVAELoaderDualCLIPLoaderGGUFCLIPTextEncodeKSamplerLoraLoaderModelOnly