Releases: Mainframework/HugstonOne
Updated Backend
Backend updated for more model support but start being fastidious, even the Cli touching in appdata, logs, nvidia etc.
It feels slower and invasive.
Eventually need to create an alternative. It works anyway but...
The Pack 1.0.8 in Hugston.com/explore is still decent.
HugstonOne Enterprise Edition 1.0.8 Mistral3
Added support for Mistral3 models:
HugstonOne Enterprise Edition
Micro update to support the latest llm models.
Running Qwen Next at 3X the speed.
Still to be tested on more models.
HugstonOne Enterprise Edition 1.0.8 with memory option
Added memory option and support for Qwen next 80b.
Full Changelog: https://github.com/Mainframework/HugstonOne/commits/HugstonOne_Enterprise_Edition_with_memory
HugstonOne Enterprise Edition 1.0.8
#1 Improved a bit more and updated some new features in the last version of HugstonOne 1.0.8.
Up to 30% faster inference, now is a real pleasure to create games, write books or code.
32k CTX as default
Took off FA as default
Tighten further safety and security (among others crucified the Geolocation in Electron).
Smarter agents, faster, safer, secure,
etc...
