llama.cpp gets DeepSeek V4 support via community PR #24162

A pull request by contributor am17an adds support for DeepSeek V4 in llama.cpp, enabling local inference. The PR is currently open and under review.