After heavily training only the UNET of BeatriXL; we have Bigger Liminal.
This is big liminal pushed to it's utmost limits, ran with tons of epochs to heavily refine the response of models to better reflect liminal images.
Even after all the training, it seems to have unlearned the vanishing point. I'll be addressing this for the 200+ epochs.
Lets see how far down the rabbit hole goes shall we?