Update app.py

app.py

@@ -3,8 +3,10 @@ import re
 from transformers import AutoModelForCausalLM, AutoTokenizer
 
 model_name_or_path = "teknium/OpenHermes-2-Mistral-7B"
+dtype = torch.bfloat16
 model = AutoModelForCausalLM.from_pretrained(model_name_or_path,
                                              device_map="auto",
+                                             torch_dtype=dtype,
                                              trust_remote_code=False,
                                              load_in_4bit=True,
                                              revision="main")