Update app.py

app.py

@@ -147,6 +147,7 @@ def talk(prompt, history):
     stream = model.create_chat_completion(messages = [{"role":"system","content":SYS_PROMPT},{"role":"user","content":formatted_prompt}], max_tokens=1000, stop=["</s>"], stream=True)
 #   print(output['choices'][0]['message']['content'])
     pprint.pprint(stream)
+    text = ""
     for output in stream:
         text += output["choices"][0]["text"]
         yield text
@@ -162,10 +163,7 @@ def talk(prompt, history):
  
 #    start = time.time()
 #    NUM_TOKENS=0
-#    print('-'*4+'Start Generation'+'-'*4)
-#    for token in model.generate(input_ids):
-#    	print(model.detokenize(input_ids), end='', flush=True)
-#    	NUM_TOKENS+=1
+
 #    time_generate = time.time() - start
 #    print('\n')
 #    print('-'*4+'End Generation'+'-'*4)