When I create Convolutional Neural Networks e.g. with PyTorch, TensorFlow or Keras, I must always use Flatten between Convolutional layers and Linear/Dense ...
確定! 回上一頁