A SDXL Lora model trained with photos and semi-realistic images of muscular women. Trigger words are uberfit, muscular, athletic. But other triggers should also work that describe a muscular body. If you want to give the woman a little more definition as the base model, it is enough to give the Lora a weight of 0.2-0.3. In my opinion, the look is best between 0.6 and 0.75. Depending on the prompt, 0.9 and 1 is already too strong. Sometimes it works well though. Anything above 1 will eventually explode the space-time continuum. The more weight you give the Lora, the more the semi-realistic look comes out. I feel the model works better if you don't counteract the semirealistic look with negative prompts.