Repository navigation
Llama now runs decision models. Instead of chatting, they answer a typed question with a probability for each option, through the server's /v1/systemone endpoint. They carry a Decision chip, and their model page links to a guide on how to use them.
- Download DFlash draft heads when available: up to 50% faster than MTP
- Update the engine to b11429, switching right away when no model is loaded
- Merge the Downloads and Command settings tabs into a new Advanced tab
- Rename the Web UI settings tab to Chat and its setting to custom web app
- Open the quick prompt without bringing Settings to the front