llama.cpp 8c2134a8755d8a9792af513dab20282affe903e4 (2026-10-14), the same pin as kernel/csrc/ggml

The PyPI gguf release lags the ggml types in this tree (gguf 0.19.0 cannot read Q2_0 tensors), so FreeToken reads GGUF files with this copy.

Copied from llama.cpp and edited in place, see docs/ggml-kernels.md.
