b10094

CUDA 13.0 arm64 build of llama.cpp-omni (feat/nemotron-video @ aece176).

Verified on NVIDIA DGX Spark (GB10, sm_121) against Nemotron-3-Nano-Omni Q4_K_M with the unified vision+audio projector: image description with OCR, speech transcription, and video description including motion — one server process, all modalities.

Contents follow the upstream release layout: binaries and shared libraries in one flat directory. Run with LD_LIBRARY_PATH pointing at the extracted directory, or from inside it.