YangChin

That's great! It's really helpful to set up config info when using CLIP

> Hi, Can you help with how should it be trained after applying the above code (like written in the comment at the end in the above code snippet) In...

You may need to resize the input image to a size that could be divided by 32 via stretch.

The first image is the source image, the other image also appears to have bad clothing edges.   

I agree with you that it might be one of the reasons. However, I tried another model [ComfyUI-BRIA_AI-RMBG](https://github.com/ZHO-ZHO-ZHO/ComfyUI-BRIA_AI-RMBG), which gives a smooth edge. So, there are potential problems with SAM...