Hugging Face Transformers Adds llama.cpp Quant Support

The Transformers library now supports running llama.cpp quantized models directly. This allows for more efficient local inference and easier integration of quantized weights into standard workflows.