不太懂这个feature请问下载的MNIST数据集的train_feature是什么?给的样例代码是这样的,请问我是下载的MNIST数据集的话,需要怎么改写?
Its training data features a bunch of high-dimensional tensors, representing pixels in the image