Update app.py

app.py

@@ -28,6 +28,10 @@ model = AutoModelForCausalLM.from_pretrained(
     device_map="auto",
     torch_dtype=torch.bfloat16,
 )
+terminators = [
+    tokenizer.eos_token_id,
+    tokenizer.convert_tokens_to_ids("<|eot_id|>")
+]
 model.config.sliding_window = 4096
 model.eval()
 
@@ -61,6 +65,7 @@ def generate(
         temperature=temperature,
         num_beams=1,
         repetition_penalty=repetition_penalty,
+        eos_token_id=terminators,
     )
     t = Thread(target=model.generate, kwargs=generate_kwargs)
     t.start()