Update app.py

app.py

@@ -38,10 +38,7 @@ model = AutoModelForCausalLM.from_pretrained(
     MODEL,
     torch_dtype=torch.bfloat16,
     device_map="auto")
-terminators = [
-    tokenizer.eos_token_id,
-    tokenizer.convert_tokens_to_ids("<|eot_id|>")
-]
+
 
 
 @spaces.GPU()
@@ -80,7 +77,7 @@ def stream_chat(
         top_p = top_p,
         top_k = top_k,
         temperature = temperature,
-        eos_token_id=terminators,
+        eos_token_id=[128001,128008,128009],
         streamer=streamer,
     )