当数据位于s3上的多个文件中时，如何在脚本模式下训练sagemaker上的tensorflow？

import sagemaker
from sagemaker import get_execution_role
from sagemaker.tensorflow import TensorFlow

role = get_execution_role()

hyperparameters = {'epochs': 10, 'batch-size': 64}
estimator = TensorFlow(entry_point='your_script.py',
                       role=role,
                       train_instance_count=1,
                       train_instance_type='ml.p3.2xlarge',
                       framework_version='2.4.1',
                       py_version='py37',
                       hyperparameters=hyperparameters)

train_data = 's3://your-bucket/train/'
s3_train_data = sagemaker.inputs.TrainingInput(train_data, distribution='FullyReplicated')

estimator.fit({'train': s3_train_data})