Update app.py

app.py

@@ -97,9 +97,8 @@ def mode_load(path):
 
 
 @spaces.GPU()
-flag=False   
 def stream_chat(message, history: list, temperature: float, max_length: int, top_p: float, top_k: int, penalty: float):
-    if flag==False:
+    if history.size()==0:
         model = AutoModelForCausalLM.from_pretrained(
             MODEL_ID,
             torch_dtype=torch.bfloat16,