datasets.cifar10.load_data()改成加载当前目录下的数据集

这里提供一种可能的解决方案，可以使用pickle库加载当前目录下的CIFAR-10数据集。

首先，需要将CIFAR-10数据集文件解压缩到当前目录下。可以从官方网站https://www.cs.toronto.edu/~kriz/cifar.html下载数据集文件，并解压缩到当前目录下。

接下来，可以使用以下代码加载CIFAR-10数据集：

import pickle
import numpy as np

def load_cifar10_data():
    # 加载训练数据
    with open('cifar-10-batches-py/data_batch_1', 'rb') as f:
        data = pickle.load(f, encoding='bytes')
        x_train = data[b'data']
        y_train = np.array(data[b'labels'])
    with open('cifar-10-batches-py/data_batch_2', 'rb') as f:
        data = pickle.load(f, encoding='bytes')
        x_train = np.concatenate([x_train, data[b'data']], axis=0)
        y_train = np.concatenate([y_train, np.array(data[b'labels'])], axis=0)
    with open('cifar-10-batches-py/data_batch_3', 'rb') as f:
        data = pickle.load(f, encoding='bytes')
        x_train = np.concatenate([x_train, data[b'data']], axis=0)
        y_train = np.concatenate([y_train, np.array(data[b'labels'])], axis=0)
    with open('cifar-10-batches-py/data_batch_4', 'rb') as f:
        data = pickle.load(f, encoding='bytes')
        x_train = np.concatenate([x_train, data[b'data']], axis=0)
        y_train = np.concatenate([y_train, np.array(data[b'labels'])], axis=0)
    with open('cifar-10-batches-py/data_batch_5', 'rb') as f:
        data = pickle.load(f, encoding='bytes')
        x_train = np.concatenate([x_train, data[b'data']], axis=0)
        y_train = np.concatenate([y_train, np.array(data[b'labels'])], axis=0)

    # 加载测试数据
    with open('cifar-10-batches-py/test_batch', 'rb') as f:
        data = pickle.load(f, encoding='bytes')
        x_test = data[b'data']
        y_test = np.array(data[b'labels'])

    # 调整数据格式
    x_train = x_train.reshape(-1, 3, 32, 32)
    x_train = np.transpose(x_train, (0, 2, 3, 1))
    x_test = x_test.reshape(-1, 3, 32, 32)
    x_test = np.transpose(x_test, (0, 2, 3, 1))

    return (x_train, y_train), (x_test, y_test)

# 加载数据集
(x_train, y_train), (x_test, y_test) = load_cifar10_data()

上述代码中，`load_cifar10_data`函数用于加载CIFAR-10数据集。首先，通过`open`函数打开数据集文件，然后使用pickle库的`load`函数加载数据。由于数据集文件中的数据格式为bytes类型，需要使用`encoding='bytes'`来指定编码方式。最后，通过`numpy`库进行数据格式的调整和拼接。

最后一行代码可以用来检查数据是否正确加载。