Ahmed Malik
Ahmed Malik

Reputation: 5

Issue loading images for tensorfow from colab folder

I made a folder in colab called "pics" like so which contains png images and I want to make a test/train split. I thought loading them like this

data = pathlib.Path('/content/pics') 

and then using image_dataset_from_directory like this

train_ds = tf.keras.utils.image_dataset_from_directory(
  data,
  validation_split=0.3,
  subset="training",
  seed=123,
  image_size=(165, 1310),
  batch_size=30)

would work, but I get the error 'No images found in directory /content/pics.'

Upvotes: 0

Views: 745

Answers (1)

user11530462
user11530462

Reputation:

Your image folder path is not accessible by Google Colab. Try mounting your google drive in colab using:

from google.colab import drive
drive.mount('/content/GoogleDrive/')

Keep your images folder inside this googleDrive in any folder and give this path in colab to access these images as below:

data = '/content/GoogleDrive/My Drive/dataset/'

Note : (Make sure you give path till one directory above to the exact images folder)

For instance, suppose you have images inside this pics folder '/content/GoogleDrive/My Drive/dataset/pics/'

then give path till dataset folder, like this :'/content/GoogleDrive/My Drive/dataset'

Upvotes: 1

Related Questions