How i can extracte x_train and y_train from train_generator?

Question

In my CNN model I want to extract X_train and y_train from train_generator. I want to use ensemble learning, bagging and boosting to evaluate the model. the main challenge is how i can extract X_train and y_train from train_generator using python language.

 history=model.fit_generator(train_generator, 
                             steps_per_epoch=num_of_train_samples // batch_size,
                             epochs=10, validation_data=validation_generator, 
                             validation_steps=num_of_val_samples // batch_size, 
                             callbacks=callbacks)

Answer 1

Well, first of all you didn't write the piece of code declaring this train_generator.

Since it seems to be a generator in the keras way, you should be accessing X_train and y_train by looping through train_generator.

This mean that train_generator[0] will give you the first batch of pairs of X_train/y_train.

x_train = []
y_train = []
for x, y in train_generator:
    x_train.append(x)
    y_train.append(y)

Answer 2

Compared to the answer above, this code is computational faster

train_generator.reset()
X_train, y_train = next(train_generator)
for i in tqdm.tqdm(range(int(train_generator.n/batch_size)-1)): 
  img, label = next(train_generator)
  X_train = np.append(X_train, img, axis=0 )
  y_train = np.append(y_train, label, axis=0)
print(X_train.shape, y_train.shape)

Note: You would have to install tqdm module as this helps us see the progress of extraction of our X_train and y_train.

pip install tqdm