Update README.md

README.md

@@ -63,7 +63,7 @@ python setup_cuda.py install
 ```python
 >>> from transformers import AutoTokenizer, AutoModelForCausalLM
 >>> tokenizer = AutoTokenizer.from_pretrained("vivo-ai/BlueLM-7B-Chat-4bits", trust_remote_code=True, use_fast=False)
->>> model = AutoModelForCausalLM.from_pretrained("vivo-ai/BlueLM-7B-Chat-4bits", device_map="cuda:0", torch_dtype=torch.bfloat16, trust_remote_code=True)
+>>> model = AutoModelForCausalLM.from_pretrained("vivo-ai/BlueLM-7B-Chat-4bits", device_map="cuda:0", trust_remote_code=True)
 >>> model = model.eval()
 >>> inputs = tokenizer("[|Human|]:三国演义的作者是谁？[|AI|]:", return_tensors="pt")
 >>> inputs = inputs.to("cuda:0")