Update app.py

app.py

@@ -33,7 +33,7 @@ h1 {
 }
 """
 
-flag=False
+
 tokenizer = AutoTokenizer.from_pretrained(MODEL_ID, trust_remote_code=True)
 
 
@@ -97,7 +97,7 @@ 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:
         model = AutoModelForCausalLM.from_pretrained(