Support for Nemotron Models

[CharlesTheFool]

Describe the feature

Most models I've been using thus far have worked great, but NVIDIA's nemotron models either aren't recognized by the package, or crash the editor. Given NVIDIA's upcoming batch of role-playing and npc character management models that run on Nemotron architecture — such as the already available Mistral-Nemo-Minitron-8B-128k-Instruct — i think it would fit perfectly with the main goals of the package to include compatibility with these models.

[pixelogie]

The native runtime engine for nvidia models is the NeMo Framework, did you check for Llamacpp compability?
I successfully run the f16(!) model with LLMUnity on osx from bartowski:
https://huggingface.co/bartowski/Mistral-NeMo-Minitron-8B-Instruct-GGUF