Mixtral MoE support?

[stduhpf]

As of today, Mixtral models run fine on "vanilla" llama.cpp. Is there any plans for supporting this architecture on here too?

[giladgd]

@stduhpf I've just released a new version with the latest llama.cpp, so it should be supported now

[giladgd]

The next beta version will also support it when it comes out

[stduhpf]

Works great, thanks!