TensorFlow 1 version
|
View source on GitHub
|
tf.data.experimental.make_batched_features_dataset(
file_pattern, batch_size, features, reader=tf.compat.v1.data.TFRecordDataset,
label_key=None, reader_args=None, num_epochs=None, shuffle=True,
shuffle_buffer_size=10000, shuffle_seed=None,
prefetch_buffer_size=dataset_ops.AUTOTUNE, reader_num_threads=1,
parser_num_threads=2, sloppy_ordering=False, drop_final_batch=False
)