def model_fn(features, labels, mode, params): ......       predictions={         'probability': tf.reshape(predictions, [1, -1])       }       export_outputs={         'predictions': tf.estimator.export.PredictOutput(predictions)       }       return tf.estimator.EstimatorSpec(mode, predictions=predictions,                        export_outputs=export_outputs) def main(_):     model=tf.estimator.Estimator(     model_fn=model_fn, ......       train_spec=tf.estimator.TrainSpec(     input_fn=lambda: input_fn(mode='train', pattern='D:/din/dataset/*'))   eval_spec=tf.estimator.EvalSpec(     input_fn=lambda: input_fn(mode='eval', pattern='D:/din/dataset/*'),     steps=100, throttle_secs=60 )   tf.estimator.train_and_evaluate(model, train_spec, eval_spec)   test_input_fn=lambda: input_fn(mode='predict', pattern='D:/din/dataset/*')   test_spec=tf.estimator.EvalSpec(input_fn=test_input_fn)   predis=model.predict(test_spec)#仅得对象<generator object Estimator.predict at 0x0000020BFD59B390>难取信息，咋办？