//=time() ?>
@TomYee_NowWhat I wanted to make it varying levels of photorealistic but sure can run next time.
I think flat anime is pretty much solved if you look at niji journey and novel ai diffusion
I'm just poking around base model a bit, pretty interesting before it gets fine tuned
@SimonGFarmer1 dunno how those come together, here are three best of eight outputs
Some more from first batch.
Lots of optimisation to do, took about an hour of playing about. Prompts here: https://t.co/4soGak9op0 negative important for 2.0 given how we flatten distribution of latents with dedupe etc
Embeddings will make it easier out of the box
So today I am seeing if #StableDiffusion2 can do character art as some have complained about this.
I think it can?
What do you think.
It could be said that diffusion models are windows into alternate realities.
“Nyan LeCun”. by @zqevans
#AIart #DALLE2 #multiverse https://t.co/grMmPBqGEo