BatchDataSet: get img array and labels

Question

Here is the batch data set i created before to fit in the model:

train_ds = tf.keras.preprocessing.image_dataset_from_directory(
    train_path,
    label_mode = 'categorical', #it is used for multiclass classification. It is one hot encoded labels for each class
    validation_split = 0.2,     #percentage of dataset to be considered for validation
    subset = "training",        #this subset is used for training
    seed = 1337,                # seed is set so that same results are reproduced
    image_size = img_size,      # shape of input images
    batch_size = batch_size,    # This should match with model batch size
)




valid_ds = tf.keras.preprocessing.image_dataset_from_directory(
    train_path,
    label_mode ='categorical',
    validation_split = 0.2,
    subset = "validation",      #this subset is used for validation
    seed = 1337,
    image_size = img_size,
    batch_size = batch_size,
)

if i run a for loop, i am able to access the img array and labels:

for images, labels in train_ds:
    print(labels)

But if i try to access them like this:

ATTEMPT 1)

images, labels = train_ds

I get the following value error: ValueError: too many values to unpack (expected 2)

ATTEMPT 2:

If i try to unpack it like this:

images = train_ds[:,0] # get the 0th column of all rows 
labels = train_ds[:,1] # get the 1st column of all rows 

I get the following error: TypeError: 'BatchDataset' object is not subscriptable

Is there a way for me to extract the labels and images without going trough a for loop?

Answer 1

Just unbatch your dataset and convert the data to lists:

import tensorflow as tf
import pathlib

dataset_url = "https://storage.googleapis.com/download.tensorflow.org/example_images/flower_photos.tgz" 
data_dir = tf.keras.utils.get_file('flower_photos', origin=dataset_url, untar=True)
data_dir = pathlib.Path(data_dir) 
batch_size = 32 
train_ds = tf.keras.utils.image_dataset_from_directory( 
     data_dir, validation_split=0.2, subset="training", 
     seed=123, batch_size=batch_size) 

train_ds = train_ds.unbatch()
images = list(train_ds.map(lambda x, y: x))
labels = list(train_ds.map(lambda x, y: y))

print(len(labels))
print(len(images))

Found 3670 files belonging to 5 classes.
Using 2936 files for training. 
2936 
2936