Update app.py

app.py

@@ -38,7 +38,7 @@ model = AutoModelForCausalLM.from_pretrained(
     MODEL,
     torch_dtype=torch.bfloat16,
     device_map="auto")
-
+end_of_sentence = tokenizer.convert_tokens_to_ids("<|im_end|>")
 @spaces.GPU()
 def stream_chat(
     message: str, 
@@ -75,6 +75,7 @@ def stream_chat(
         top_k = top_k,
         temperature = temperature,
         repetition_penalty=penalty,
+        eos_token_id=[end_of_sentence],
         streamer=streamer,
     )