Update app.py

app.py

@@ -7,8 +7,7 @@ from transformers import AutoTokenizer, AutoModel
 model_name = "teknium/OpenHermes-2.5-Mistral-7B"
 token = os.environ['hf_token']
 
-model = AutoModel.from_pretrained(model_name).cuda()
-tokenizer = AutoTokenizer.from_pretrained(model_name)
+pipe = pipeline("text-generation", model=model_name, device="cuda")
 
 
 system_prompt = '''You are given an input text for a chat interface. Propose auto-completion to the text. You have several roles: