Fix tokenizer.eos_token_id

app.py

@@ -47,7 +47,7 @@ def predict(message, history, system_prompt, temperature, max_tokens):
         temperature=temperature,
         max_new_tokens=DEFAULT_MAX_NEW_TOKENS,
         use_cache=True,
-        eos_token_id=100278 # <|im_end|>
+        eos_token_id=tokenizer.eos_token_id # <|im_end|>
     )
     t = Thread(target=model.generate, kwargs=generate_kwargs)
     t.start()
@@ -61,7 +61,6 @@ def predict(message, history, system_prompt, temperature, max_tokens):
 if __name__ == "__main__":
     args = parse_args()
     tokenizer = AutoTokenizer.from_pretrained("lliu01/fortios_cli")
-    tokenizer = AutoTokenizer.from_pretrained("lliu01/fortios_cli")
     model = AutoModelForCausalLM.from_pretrained(
         "lliu01/fortios_cli",
         torch_dtype=torch.bfloat16,