adding hunyuan hf (support lora finetuning); unified hunyuan hf inference with quantization

adding hunyuan hf (support lora finetuning); unified hunyuan hf inference with quantization #46