Update llama.cpp

[ysimonson]

llama.cpp recently added the qwen3vlmoe model architecture, so this update prevents the following error:

2025-11-20T14:07:26.318944Z ERROR load_from_file: llama-cpp-2: error loading model: error loading model architecture: unknown model architecture: 'qwen3vlmoe' module="llama.cpp::llama_model_load"

[MarcusDunn]

thanks for the PR. will merge and cut release if tests give the all-clear

[MarcusDunn]

looks like you'll need to update the Cargo.toml to include some new files:

https://github.com/utilityai/llama-cpp-rs/actions/runs/19540618299/job/55962123635?pr=866

[ysimonson]

Hopefully added all the missing files, though is there a way to check w/o CI? Just running cargo test doesn't seem to catch it

[MarcusDunn]

cargo publish --dry-run should do it.

[ysimonson]

nice, thanks for the pointer. Works on my end so hopefully will pass CI.

[MarcusDunn]

thanks for the PR! there's a new release with the new llama.cpp