Only a single validation batch is used if `estimator.train` called with `cache_data=True` about gluonts HOT 3 CLOSED

from gluonts.itertools import Cached

module = estimator.create_lightning_module()
transformation = estimator.create_transformation()

transformed_validation_data = transformation.apply(val, is_train=True)
transformed_validation_data = Cached(transformed_validation_data)

validation_data_loader = estimator.create_validation_data_loader(
    transformed_validation_data,
    module,
)

print(len(list(validation_data_loader)))
print(len(list(validation_data_loader)))
print(len(list(validation_data_loader)))
print(len(list(validation_data_loader)))

if cache_data:
    transformed_validation_data = list(transformed_validation_data)