Question about mask inpainting with reference image
Hello! I really love your model — it is both fast and produces excellent results. However, I have encountered an issue that I hope you can help me with.
In your examples, there is a case where you use the BFS LoRA together with mask inpainting to replace a face. The description said quote "The prompt and Face picture 2 are the same as above", so I assume it used both image 1 and 2 as reference. I tried to reproduce the same workflow, but the generated face and the smoothness of the transition around the mask edges are never as good as in your example. This has been quite frustrating for me.
May I ask if this example was created using ComfyUI? If so, would it be possible for you to share the workflow? Or was it actually done using another tool?
Thank you very much for your help!
Hi, @OrvilleQ , If you want to support my model refine work, you can go to ko-fi.com/wikeeyang to get the workflow file for the full-function demo as the same as example image. Thanks a lot! 🤝