Output seems truncated. Try setting max_length now that we have do_sample=True

app.py

@@ -66,6 +66,7 @@ def predict(code):
         temperature=1.2,
         repetition_penalty=1.1,
         return_full_text=False,
+        max_length=4096,
     )
     print(f"Pipe out: {repr(pipe_out)}")