I have been having a blast with DALL-E, spending about an hour a day trying out wild combinations and cracking my friends up. I cannot imagine getting bored of it; it's like getting bored with visual stimulus, or art in general. In fact, I've been glad to have a 50/day limit, because it helps me contain my hyperfocus instincts. The information about new pricing is, to me as someone just enjoying making crazy imagines…
I think people don't realize how huge these models really are. When they're free, it's pretty cool. But charge an amount where there's actual profit in the product? Suddenly seems very expensive and not economically viable for a lot of use cases. We are still in the "you need a supercomputer" phase of these models for now. Something like DALLE mini is much more accessible but the results aren't good enough. Early ear…
They really aren't that large by the contemporary scaling race standards. DALLE-2 has 3.5B parameters, which should fit on an old GPU like Nvidia RTX2080, especially if you optimize your model for inference [1][2] which is commonly done by ML engineers to minimize costs. With optimized model, your memory footprint is ~1 byte per parameter, and some less than 1 ratio (commonly ~0.2) of all parameters to store intermediate activations.
You should be able to run it on Apple M1/M2 with 16GB RAM via CoreML pretty fine, if an order of magnitude slower than on an A100.
Training isn't unreasonably costly as well: you can train a model given O(100k)$ which is less than a yearly salary of a mid-tier developer in silicon valley.
There is no reason these models shouldn't be trained cooperatively and run locally on our own machines. If someone is interested in cooperating with me on such a project, my email is in the profile.
1. https://arxiv.org/abs/2206.01861
2. https://pytorch.org/blog/introduction-to-quantization-on-pyt...