[Python] tensorflow/core/framework/local_rendezvous.cc:407] Local rendezvous is aborting with...

I am experiencing an issue with an image classification model.

After extracting frames from videos, saving them in a suitable folder structure for creating an image dataset from a directory using tf.keras.utils.image_dataset_from_directory, and verifying the size of the created dataset, I am encountering an error message – though I believe it might be a warning – stating: 'tensorflow/core/framework/local_rendezvous.cc:407] Local rendezvous is aborting with status: OUT_OF_RANGE: End of sequence'.

I have added programming logic to the code to print the shape of the images and the shape of the labels once the dataset is created. However, I have observed in the execution log that the aforementioned message continues to appear.

Does anyone know why this message is occurring?

Here is the programming logic code:

def create_image_dataset(directory, img_size, batch_size, seed=None):
"""Creates a tf.data.Dataset from a main directory.

Args:
directory: Main directory containing subdirectories with the images.
img_size: Size to which the images will be resized.
batch_size: Batch size.
seed: Random seed.

Returns:
A tf.data.Dataset.
"""
directory = os.path.abspath(directory)
directory = os.path.normpath(directory) # Normalize the path
print(f"Creating dataset from: {directory}")
try:
dataset = tf.keras.utils.image_dataset_from_directory(
directory,
labels='inferred',
label_mode='categorical',
image_size=img_size,
batch_size=batch_size,
shuffle=True,
seed=seed
)
print(f"Dataset created successfully from: {directory}")
return dataset
except ValueError as e:
print(f"Error creating dataset from: {directory}. Error: {e}. Directory: {directory}")
return None

# Create the dataset
print(f"Creating dataset from the directory: {DATASET_DIR}") # Path debugging
dataset = create_image_dataset(DATASET_DIR, IMG_SIZE, BATCH_SIZE, seed=SEED)

# Print the dataset size
if dataset:
print("Dataset size:", len(list(dataset.as_numpy_iterator())))
# Get one batch from the dataset to inspect its shape
for images, labels in dataset.take(1):
print("Shape of the images:", images.shape)
print("Shape of the labels:", labels.shape)


And here is a portion of the code execution log:

Found 9108 files belonging to 4 classes.
025-05-09 18:29:36.880956: I tensorflow/core/framework/local_rendezvous.cc:407] Local rendezvous is aborting with status: OUT_OF_RANGE: End of sequence
Tama�o del dataset: 285
Forma de las im�genes: (32, 224, 224, 3)
Forma de las etiquetas: (32, 4)
2025-05-09 18:29:37.102376: I tensorflow/core/framework/local_rendezvous.cc:407] Local rendezvous is aborting with status: OUT_OF_RANGE: End of sequence

Continue reading...