A TFRecords file represents a sequence of (binary) strings. The format is not random access, so it is suitable for streaming large amounts of data but not suitable if fast sharding or other non-sequential access is desired. For more information, please refer to https://www.tensorflow.org/versions/master/api_docs/python/python_io.html#tfrecords-format-details.
Binary files are sometimes easier to use, because you don’t have to specify different directories for images and groundtruth annotations. While storing your data in binary file, you have your data in one block of memory, compared to storing each image and annotation separately. Openning a file is a considerably time-consuming operation especially if you use hdd and not ssd, because it involves moving the disk reader head and that takes quite some time. Overall, by using binary files you make it easier to distribute and make the data better aligned for efficient reading.
Term
- shuffle training examples
- shuffle v. 改变顺序，洗牌