The tricky part of this particular topology is that the TensorFlow generates model which is not ... logits = ResNeXt(x, training=tf.constant(False)).model.
確定! 回上一頁