








This is Lora for Flux Fill outpainting (doesn't work with Flux Dev) to take all ??????????? from a given image.
Trained in 1024px on 65 images for 8555 steps.
To make it work do outpainting using Flux Fill model (see my workflow: https://github.com/Mozer/comfy_stuff/blob/main/workflows/flux_fill_outpaint_with_lora.json)
Works best with high resolution, make height 1024+ px, set oupaint width same as original image. Keep aspect ratio of the new part same as original, that's very important!
Set height 1500px if you have low-res artifacts on a face.
Face is cloning is not perfect. It was trained on photos, but it also works with anime and illustrations. Not tested with inpaint. Not tested with Flux Dev with Flux Fill extracted lora.
Works best with 2-panel-images: left one should be with clothes, do outpainting on the right part.
Recommended settings: height: 1024+, steps: 20, strength: 1.0.
Prompt: On/off photo of a ????? woman, full body shot, front view
1. 轉载模型僅供學習與交流分享,其版權及最終解释權归原作者。
2. 模型原作者如需認領模型,請通過官方渠道联系海藝AI工作人員進行認證。我們致力於保護每一位創作者的權益。 點擊去認領
