4 ms·

When we reach the point when average people can easily run today’s models on their own devices, today’s models will no longer be SOTA and there will still be de

by thorum 3y ago

When we reach the point when average people can easily run today’s models on their own devices, today’s models will no longer be SOTA and there will still be demand to run better models in the cloud.

brucethemoose2 3y ago
Image generation parameter count is hitting diminishing returns, going by what we've seen with SD/SDXL. The tooling around them is far more important. However, Stable diffusion already can run on mobile devices. There is already a good iOS app for it (and the dev is here on HN) but the problem seems to be that no one cares. There are 700,000 cloud imagegen apps crowding it out, because thats what's easier and more profitable to spam across the store and web.
- thorum 3y ago
  > Image generation parameter count is hitting diminishing returns, going by what we've seen with SD/SDXL. For image quality, sure - language understanding is still an issue. SDXL can generate a beautiful image, but if it doesn’t show exactly what you asked for in the prompt, on the first try, there is still room for improvement. The gap between LLMs and image generators in this regard is huge.
- Closi 3y ago
  Depends if the limit on parameter count is a 'real' limit, or just a limit based on what current-technology models can effectively use. Back in the 'Google Daydream' days, Google might have found that they didn't get any more image-generation performance by raising the parameter count - but that's just because the technology at the time couldn't effectively utilise more parameters. It's impossible to know what next-gen models might be able to use, but I suspect we will find ways to allow the models to take advantage of even higher parameter counts. Stable diffusion can run on mobile devices, but it's painful and image generation takes a fraction of the time via cloud services.

4 ms·

When we reach the point when average people can easily run today’s models on their own devices, today’s models will no longer be SOTA and there will still be de

by thorum 3y ago

brucethemoose2 3y ago
Image generation parameter count is hitting diminishing returns, going by what we've seen with SD/SDXL. The tooling around them is far more important. However, Stable diffusion already can run on mobile devices. There is already a good iOS app for it (and the dev is here on HN) but the problem seems to be that no one cares. There are 700,000 cloud imagegen apps crowding it out, because thats what's easier and more profitable to spam across the store and web.
- thorum 3y ago
  > Image generation parameter count is hitting diminishing returns, going by what we've seen with SD/SDXL. For image quality, sure - language understanding is still an issue. SDXL can generate a beautiful image, but if it doesn’t show exactly what you asked for in the prompt, on the first try, there is still room for improvement. The gap between LLMs and image generators in this regard is huge.
- Closi 3y ago
  Depends if the limit on parameter count is a 'real' limit, or just a limit based on what current-technology models can effectively use. Back in the 'Google Daydream' days, Google might have found that they didn't get any more image-generation performance by raising the parameter count - but that's just because the technology at the time couldn't effectively utilise more parameters. It's impossible to know what next-gen models might be able to use, but I suspect we will find ways to allow the models to take advantage of even higher parameter counts. Stable diffusion can run on mobile devices, but it's painful and image generation takes a fraction of the time via cloud services.