The reference sampling implementation auto adjusts the shift value based on the resolution, if you don't want this you can just bypass (CTRL-B) this ModelSamplingFlux node.

If you get an error in any of the nodes above make sure the files are in the correct directories. See the top of the examples page for the links : https://comfyanonymous.github.io/ComfyUI_examples/flux/ flux1-dev.safetensors goes in: ComfyUI/models/unet/ t5xxl_fp16.safetensors and clip_l.safetensors go in: ComfyUI/models/clip/ ae.safetensors goes in: ComfyUI/models/vae/ Tip: You can set the weight_dtype above to one of the fp8 types if you have memory issues.

The redux model lets you prompt with images. It can be used with any Flux1 dev or schnell model workflow. You can chain multiple "Apply Style Model" nodes if you want to mix multiple images together.

A mysterious wolf