Update README.md

README.md

@@ -71,7 +71,7 @@ model = AutoModel.from_pretrained('cmarkea/bloomz-560m-retriever-v2')
 
 def infer(txt: Union[str, List[str]]):
     tok = tokenizer(txt, padding=True, return_tensors='pt')
-    with torch.no_grad():
+    with torch.inference_mode():
         embedding = model(**tok)
     # Inportant: take only last token!
     return embedding.get('last_hidden_state')[:,-1,:].numpy()