Well-known member
this could have been big news: Google announce PARTI - Pathways Autoregressive Text-to-Image

but then turn around and say "you can't have it, too dangerous for you lot to play with"

Beyond the model capabilities and evaluations presented above, there are broader issues to consider with large-scale models for text-to-image generation. [...] blah..blah..blah.. [..] These considerations all contribute to our decision not to release our models, code or data at this time.

from the research paper


Well-known member