Add Q4_K_M GGUF quant of fine-tuned DeepSeek Coder V2 Lite

initial commit