SIMPLE FLUX GGUF INPAINTING by weshouldhaveshotguns

Use this ComfyUI workflow for targeted inpainting edits: it takes an image and prompt text, applies the inpainting process, and saves an edited image.

The workflow loads and resizes an image, encodes prompt text, applies inpainting conditioning, samples, decodes, and saves an image.

It also includes mask blurring, Flux guidance, random noise, scheduling, GGUF model loading, and custom sampling.

The workflow names LoadAndResizeImage for the image, CLIPTextEncode for prompt text, and SaveImage for the edited output.

The workflow is tailored for inpainting in ComfyUI and uses GGUF-quantized models for low-VRAM GPUs.

Its description includes links to required and optional files for setup.

Listed model files are ae.safetensors, clip_l.safetensors, flux1-dev-Q8_0.gguf, and t5-v1_1-xxl-encoder-Q8_0 (1).gguf.

Named ComfyUI components include BasicGuider, BasicScheduler, CLIPTextEncode, DualCLIPLoaderGGUF, FluxGuidance, ImpactGaussianBlurMask, InpaintModelConditioning, and KSamplerSelect.

Other named components are LoadAndResizeImage, Power Lora Loader (rgthree), RandomNoise, SamplerCustomAdvanced, SaveImage, UnetLoaderGGUF, VAEDecode, and VAELoader.

The workflow includes an inpainting settings rundown for targeted modifications and prompt tips for blending new elements into the original image.

The listed base model is Flux.1 D.

Before using it, verify that every listed model file and ComfyUI component is available in your setup.

Vorschlag · nicht geprüft