import numpy as npimport tensorflow as tffrom keraspreprocessingimage import ImageDataGeneratorimport osfrom PIL import Imagefrom kerasutils import to_categoricalfrom sklearnmodel_selection import tra

The error occurs because the "y_train" array is being used as keys in the dictionary, but numpy arrays are not hashable and cannot be used as keys. To fix this issue, you can convert the "y_train" array to a list before creating the dictionary. Here's the modified code:

dict_train = dict(zip(y_train.tolist(), x_train))
dict_test = dict(zip(y_test.tolist(), x_test))
np.save("./lalala1.npz", dict_train)
np.save("./lalala2.npz", dict_test)

By converting the "y_train" and "y_test" arrays to lists using the .tolist() method, you can create the dictionaries without any error