Attention: The deployed application uses a model that has been truncated by 75% from the originally trained version. As a result, its performance may not meet the expectations for production-level use.