我使用 keras.dataset 导入了数据集
fashion_mnist = keras.datasets.fashion_mnist
(X_train, y_train), (X_test, y_test) = fashion_mnist.load_data()
X_train_scaled = X_train / 255.0
X_test_scaled = X_test / 255.0
X_train_scaled.dtype
现在我需要将训练数据集拆分为在验证集中有 5,000 个观察值：
X_valid = X_train_scaled[:??]
X_train = X_train_scaled[??:]
y_valid = y_train_full[:??]
y_train = y_train_full[??:]
然后使用 plt.imshow 函数绘制训练数据集中的第一张图像并设置参数 cmap="binary"：