Bert Base model trained on quora question pair dataset. The task is to infer whether the two given question are similar or not. The model achieves 89% evaluation accuracy on the task.