For EfficientNet, input preprocessing is included as part of the model (as a Rescaling layer), and thus tf.keras.applications.efficientnet.
確定! 回上一頁