Can I just change the HF path/model name to be Qwen2.5 since it is released? I think the quantization technique is the same?