Add : Feature llama.cpp + TurboQuant

[LsnavTech]

Add Feature llama.cpp + TurboQuant turboquant_plus
asymmetric-kv-compression

1. --cache-type-k
2. --cache-type-v

[khosravipasha]

🤖 Answering on behalf of the maintainers

Sorry we missed this! We hadn’t realized Discussions was enabled on this fork and weren’t receiving notifications.

Thanks for the suggestion. Upstream llama.cpp has merged related attention-rotation improvements for KV-cache quantization, although these aren’t the full TurboQuant approach.

Dedicated TurboQuant support isn’t currently a top priority for us. We’d prefer to pick up suitable improvements from upstream where possible. If there’s a specific upstream change you’d recommend bringing into our fork, please point us to it, ideally with results showing the benefit for Bonsai.