here are some methods to use dataset with mindspore architecture.

url = 'xxx'
path = download(url, "./", kind="zip", replace=True)

load some dataset.

like  url = "https://mindspore-website.obs.cn-north-4.myhuaweicloud.com/" \
      "notebook/datasets/MNIST_Data.zip"

we can shuffle it, which means mixture to make it more random:

train_dataset = MinstDataset("MNIST_Data/train", shuffle=True)

visualize some data:

def visualize(dataset):
    figure = plt.figure(figsize=(4,4))
    cols, rows = 3,3
    plt.subplots_adjust(wspace=0.5, hspace=0.5)
    for idx, (image, label) in enumerate(dataset.create_tuple_iterator()):
        figure.add_subplot(rows, cols, idx+1)
        plt.title(int(label))
        plt.axis("off")
        plt.imshow(image.asnumpy().squeze, cmap="gray")
        if idx == cols * rows - 1:
        break
    plt.show()

enumerate(dataset.create_tuple_iterator()) supplies the api to access data by iterator.

if we run the code, the output:

visualize(train_dataset)

and most importantly, batch operation.

train_dataset = train_dataset.batch(batch_size = 64)

this is a tradeoff of your calc resource and the data size.

It would appear more than once in neural network.

Besides, your can access data with your own way:

class RandomAccessDataset:
    def __init__(self):
        self.data = np.ones((5,2))
        self.label = np.zeros ((5,1))
    def __getitem__(self,index):
        return self._data[index], self._label[index]
    def __len__(self):
        return len(self._data)

loader = RandomAccessDataset()
dataset = GeneratorDataset(source=loader, column_names=["data","label"])

for data in dataset:
    print(data)