llama.cpp adds support for DeepSeek V4

A pull request merging DeepSeek V4 support into llama.cpp has been accepted. Users can now pull, compile, and download GGUF models for local inference.