View from below

I noticed that, especially on Stable Diffusion, this shot is very complicated to obtain, this model tries to impose this view.

But it's still in Beta and hard to get anyway. I hope the community will help me with the generations so I can train the model better.