python - 错误-使用Keras的多分类神经网络

Question

我从我的NN代码中收到一个相当烦人的错误，并希望对Keras如何工作有更好了解的人可以向我解释为什么我会收到该错误。感谢您的帮助!
错误:

AttributeError: 'DirectoryIterator' object has no attribute 'ndim'

错误来自:

    Traceback (most recent call last):
    File "C:\Users\Cameron\Desktop\AI\CubeFieldNN_Train -fix.py", line 80, in <module>
    validation_steps = (validation_samples / batch_size))

码:

NN.fit(
train_set, train_labels,
batch_size = batch_size,
epochs = epochs,
validation_data = (validation_set, validation_labels),
validation_steps = (validation_samples / batch_size))

完整代码: https://pastebin.com/V1YwJW3X

完全错误:

    Traceback (most recent call last):
  File "C:\Users\Cameron\Desktop\AI\CubeFieldNN_Train -fix.py", line 80, in <module>
    validation_steps = (validation_samples / batch_size))
  File "C:\Python\lib\site-packages\keras\models.py", line 1002, in fit
    validation_steps=validation_steps)
  File "C:\Python\lib\site-packages\keras\engine\training.py", line 1630, in fit
    batch_size=batch_size)
  File "C:\Python\lib\site-packages\keras\engine\training.py", line 1476, in _standardize_user_data
    exception_prefix='input')
  File "C:\Python\lib\site-packages\keras\engine\training.py", line 76, in _standardize_input_data
    data = [np.expand_dims(x, 1) if x is not None and x.ndim == 1 else x for x in data]
  File "C:\Python\lib\site-packages\keras\engine\training.py", line 76, in <listcomp>
    data = [np.expand_dims(x, 1) if x is not None and x.ndim == 1 else x for x in data]
AttributeError: 'DirectoryIterator' object has no attribute 'ndim'

Answer 1

确实没有必要从previous question从fit过渡到fit_generator。
flow_from_directory 返回一个生成器类型的对象，该对象返回数据和标签的元组。对于validation_set同样如此。还要注意，如果指定validation_steps，还必须指定steps_per_epoch。因此，您可以使用:

NN.fit_generator(train_set,
                 steps_per_epoch=steps_per_epoch,
                 epochs=epochs,
                 validation_data=validation_set,
                 validation_steps=validation_steps)

另外，您可以一次加载所有图像，然后将其与标签一起传递给 NN.fit()函数。