FLUX.1 Kontext workflow: DEMAND Any Change From Your Images (For Beginner)

This beginner-oriented FLUX.1 Kontext workflow is for changing images: it accepts an image through LoadImage and provides an image preview through PreviewImage.

The workflow includes reference-latent handling, sampling, VAE encoding and decoding, and image preview.

The named input nodes are LoadImage for an image and ReferenceLatent for reference-latent data.

The output path includes VAEDecode and PreviewImage for a decoded-image preview.

A linked video explains how to use the workflow.

The named model files are ae.safetensors, clip_l.safetensors, flux1-dev-kontext_fp8_scaled.safetensors, and t5xxl_fp8_e4m3fn_scaled.safetensors.

Named nodes include CLIPTextEncode, ConditioningZeroOut, DualCLIPLoader, EmptySD3LatentImage, FluxGuidance, ImageSharpen, and KSampler.

Other named nodes include LoadImage, PreviewImage, ReferenceLatent, UNETLoader, VAEDecode, VAEEncode, and VAELoader.

The listing names Flux.1 Kontext as the base model.

The publisher is Lordcaocao2025.

Before use, verify that your setup includes ae.safetensors, clip_l.safetensors, flux1-dev-kontext_fp8_scaled.safetensors, and t5xxl_fp8_e4m3fn_scaled.safetensors.

Vorschlag · nicht geprüft

Before use, verify that your setup includes all of the named nodes listed above.

Vorschlag · nicht geprüft