The new Mac Studio with M3 Ultra can be configured with a max now of 512 GB unified RAM. Since the unified RAM can be used as VRAM, would Topaz VEAI be able to take advantage of this and process faster when a local Starlight model releases?
1 Like
The team is working on configuring the Starlight model to bring it down in memory resource needs so that it can be run locally. How this looks and what it will need are still being worked out as it is very early in this process.
2 Likes
Sounds good Kyle!
