I tried training the model using his image, and I hope it works. To be honest, I'm not very familiar with the realistic aspects of Stable Diffusion, such as prompts and negatives, etc.
So, if you're using this model, please let me know your results so I can improve and update it in the future. I don't know which model is best for generating absurd or semi-realistic images. For now, I'm only using the AbsoluteReality model, which is not too bad, I guess.
However, a better model would lead to better results, especially for this purpose. So, I need your results with generative data to collect information.