Don't load abnormal image

[hamac03]

When I use this command:

# Import the datamodule
from anomalib.data import Folder

# Create the datamodule
datamodule = Folder(
    name="hazelnut_toy",
    root="datasets/hazelnut_toy/hazelnut_toy",
    normal_dir="good",
    abnormal_dir="crack",
    mask_dir="mask/crack",
    normal_test_dir="../../../datasets/hazelnut_toy/hazelnut_toy/good",
    train_batch_size=16,
    eval_batch_size=16,
    normal_split_ratio=0.0,
)

# Setup the datamodule
datamodule.setup()

# Access the datasets
train_dataset = datamodule.train_data
val_dataset = datamodule.val_data
test_dataset = datamodule.test_data

# Access the dataloaders
train_dataloader = datamodule.train_dataloader()
val_dataloader = datamodule.val_dataloader()
test_dataloader = datamodule.test_dataloader()

# Print the number of samples in each dataset
len(train_dataset), len(val_dataset), len(test_dataset)

The result is:

34, 19, 20

This mean that it don't load any abnormal images from the crack dir into train_dataset.

Could you kindly explain that is the right Datasetloader or I'm wrong in some step.
Thank you,
Best regard,
Ha

[alexriedel1]

Anomaly detection is often seen as a unsupervised learning problem. The goal is to predict anomalies only from normal samples. Thats why anomalibs models dont need abnormal samples for training