Update app.py

app.py

@@ -31,6 +31,10 @@ model = AutoModelForCausalLM.from_pretrained(
 model.config.sliding_window = 4096
 model.eval()
 
+terminators = [
+    pipeline.tokenizer.eos_token_id,
+    pipeline.tokenizer.convert_tokens_to_ids("<|eot_id|>")
+]
 
 @spaces.GPU(duration=90)
 def generate(
@@ -62,6 +66,7 @@ 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()