Mtmd : add Nemotron 3.5 Transcribe support by pskrunner14 · Pull Request #31471 · ggml-org/llama.cpp

The change resolves a conflict with the main branch, with the exception of a small delta in quantization work for IQ2(NL) and IQ3(NL) optimizations. The pull request, labeled as superseding prior work, was approved by CISCappro and merged into the master branch. The project’s documentation notes the inclusion of hidden Unicode characters, which may affect code interpretation. No new features or benchmarks were explicitly detailed in the post, but the addition expands Llama.cpp’s multimodal functionality.