Initialize an Out-of-Distribution Detector with Custom Prediction Arguments

Add small blurb

from daml.detectors import OOD_VAE
from daml.models.tensorflow import VAE, create_model

# instantiate an OOD detector metric
metric = OOD_VAE(create_model(VAE, dataset[0].shape))

# the training set has about 15% out-of-distribution so set the fit threshold at 85%
metric.fit(dataset, threshold_perc=85, batch_size=128, verbose=False)

# detect OOD at the 'feature' level
metric.predict(dataset, ood_type="feature")