Keras: Predict with generator and get result table

Question

I have a keras model trained to classify images. Since I want to predict a lot of new images (about 100k) I think the usage of a generator is the fastest way.

from keras.preprocessing.image import ImageDataGenerator

# Set up my generator
datagen = ImageDataGenerator(rescale=1./255)

test_generator = datagen.flow_from_directory(
    r"C:\new_data",
    target_size=(260, 180),
    batch_size=100,
    color_mode="grayscale",
    class_mode="categorical",
)

predicted_elements = model.predict(test_generator, batch_size=10, verbose="auto", steps=None, callbacks=None)

What I need is a table, to see which image belongs to which (predicted) class. Like:

filename	predicton
1.jpg	coffee
2.jpg	car
3.jpg	plane
...	...

Edit: My Question is: How can I create this table? The generator resorts the images anyway. Also I don't know how to get the names.