I had trouble getting images like this without img2img support, so I decided to train a model for it. LORA comes in two flavors: Dom (holding leash) and Sub (on leash).

Support me on Ko-Fi!