无法重塑数组答案

【问题标题】：Cannot reshape array无法重塑数组
【发布时间】：2020-09-23 06:43:04
【问题描述】：

验证和测试数据形状：

x_train = train_data.reshape(train_data.shape[0], train_data.shape[1], train_data.shape[2], INPUT_DIMENSION)
#x_t =x_train = train_data.reshape(train_data.shape[0], train_data.shape[1], train_data.shape[2])
x_test_all = test_data.reshape(test_data.shape[0], test_data.shape[1], test_data.shape[2], INPUT_DIMENSION)

x_val = x_test_all[-VAL_SIZE:]
y_val = y_test[-VAL_SIZE:]

x_test = x_test_all[:-VAL_SIZE]
y_test = y_test[:-VAL_SIZE]


history_fdssc = model_fdssc.fit(
        [x_train.reshape(x_train.shape[0], x_train.shape[1], x_train.shape[2], x_train.shape[3], 1), 
         x_train.reshape(x_train.shape[0], x_train.shape[1], x_train.shape[2], x_train.shape[3])], [y_train, y_train, y_train],
        validation_data=(x_val.reshape(x_val.shape[0], x_val.shape[1], x_val.shape[2], x_val.shape[3], 1), y_val),
        batch_size=batch_size, epochs=nb_epoch, shuffle=True,
        callbacks=[early_Stopping, save_Best_Model, reduce_LR_On_Plateau, history, tensor_board])

当我运行程序时，我收到以下错误：

Please input the name of Dataset(IN, UP, KSC or SS):KSC
(512, 614, 176)
The class numbers of the HSI data is: 13
-----Importing Setting Parameters-----
-----Starting the  1 Iteration-----
Train size:  1048
Test size:  4163
Validation size:  524

-----Selecting Small Pieces from the Original Cube Data-----

Traceback（最近一次调用最后一次）：文件“hyb.py”，第 189 行，在 x_t =x_train = train_data.reshape(train_data.shape[0], train_data.shape[1], train_data.shape[2]) ValueError: 无法将大小为 14940288 的数组重塑为 (1048,9,9)

【问题讨论】：

标签： python tensorflow image-processing reshape conv-neural-network

【解决方案1】：

错误源于第一行，因为数组的总大小不能被给定的整形参数整除。这是一个玩具示例：：

x_train = train_data.reshape(train_data.shape[0], train_data.shape[1], train_data.shape[2], INPUT_DIMENSION)

>>> x = np.arange(14940288)
>>> x.reshape(1049,9,9)
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
ValueError: cannot reshape array of size 14940288 into shape (1049,9,9)

这是您应该尝试的。另外，请注意您没有提供变量INPUT_DIMENSION：

>>> x.reshape(-1,9,9).shape  #-1 takes the whole length
(184448, 9, 9)
>>> INPUT_DIMENSION = 16 
>>> x.reshape(-1,9,9,INPUT_DIMENSION).shape
(11528, 9, 9, 16)

【讨论】：