Version 2 smokes version 1.They are miles apart.Realistic Vision is great for this.If you want interracial, just use the keyword in your prompt.If you don't want it, then just say woman kissing a man, and specify the races/ethnicites.Here's how you reproduce the images:1.) Call the LORA with some sort of a prompt along the lines of "woman kissing a black man". Feel free to say what her hands should be doing, or what they should be wearing.-&gt; This will give you an 80% done image2.) Then, go to img2img, and kick out the lora, and re-run the model.-&gt; This will then give you a solid finalized image.

Files with SHA256: fc8a5d5ef7d9a1911fa77803ce2600bd67228772230a1c30c88c037d7cb3a627