Add support for LFM2 1.2B

[evangineer]

https://huggingface.co/LiquidAI/LFM2-1.2B-GGUF

I tried to run this with Llamafile 0.9.3.

Model wouldn't load.

Turns out this is the minimum required llama.cpp tag:
https://github.com/ggml-org/llama.cpp/releases/tag/b5921

[cjpais]

If someone wants to support it, they can submit a PR and I will try to merge it when I'm available.