llama.cpp

6a6780a2 - vulkan: Support GGML_TYPE_NVFP4 (#21455)

vulkan: Support GGML_TYPE_NVFP4 (#21455) This adds nvfp4 support for get_rows, dequant, and mul_mat(_id). For mul_mat, it does not add support for the dp4/q8_1 path, it's all via fp16/fp32.