Download the training dataset versions attached to experiment

exp_datasets.py

attached_datasets = experiment.list_attached_dataset_versions()
train_dataset_version = experiment.get_dataset(train_version_name)
eval_dataset_version = experiment.get_dataset(eval_version_name)
images_dir = os.path.join(os.getcwd(), train_set_local_dir)

if os.path.isdir(train_set_local_dir):
  print("dataset version has already been downloaded")
else:
  train_dataset_version.download(train_set_local_dir)

if os.path.isdir(eval_set_local_dir):
  print("dataset version has already been downloaded")
else:
  eval_dataset_version.download(eval_set_local_dir)