tf.python_io

目录

Classes

Functions

tf.io.TFRecordCompressionType

Class TFRecordCompressionType

Class Members

Class TFRecordOptions

__init__

Methods

get_compression_type_string

Class Members

TFRecordWriter

tf.io.tf_record_iterator

---

Python functions for directly manipulating TFRecord-formatted files.

Classes

• class TFRecordCompressionType: The type of compression for the record.
• class TFRecordOptions: Options used for manipulating TFRecord files.
• class TFRecordWriter: A class to write records to a TFRecords file.

Functions

• tf_record_iterator(...): An iterator that read the records from a TFRecords file. (deprecated)

tf.io.TFRecordCompressionType

Class TFRecordCompressionType

The type of compression for the record.

Aliases:

• Class tf.compat.v1.io.TFRecordCompressionType
• Class tf.compat.v1.python_io.TFRecordCompressionType
• Class tf.python_io.TFRecordCompressionType

Class Members

• GZIP = 2
• NONE = 0
• ZLIB = 1

Class TFRecordOptions

Options used for manipulating TFRecord files.

Aliases:

• Class tf.compat.v1.io.TFRecordOptions
• Class tf.compat.v1.python_io.TFRecordOptions
• Class tf.compat.v2.io.TFRecordOptions
• Class tf.python_io.TFRecordOptions

__init__

View source

__init__(
    compression_type=None,
    flush_mode=None,
    input_buffer_size=None,
    output_buffer_size=None,
    window_bits=None,
    compression_level=None,
    compression_method=None,
    mem_level=None,
    compression_strategy=None
)

Creates a TFRecordOptions instance.

Options only effect TFRecordWriter when compression_type is not None. Documentation, details, and defaults can be found in zlib_compression_options.h and in the zlib manual. Leaving an option as None allows C++ to set a reasonable default.

Args:

• compression_type: "GZIP", "ZLIB", or "" (no compression).
• flush_mode: flush mode or None, Default: Z_NO_FLUSH.
• input_buffer_size: int or None.
• output_buffer_size: int or None.
• window_bits: int or None.
• compression_level: 0 to 9, or None.
• compression_method: compression method or None.
• mem_level: 1 to 9, or None.
• compression_strategy: strategy or None. Default: Z_DEFAULT_STRATEGY.

Returns:

• A TFRecordOptions object.

Raises:

• ValueError: If compression_type is invalid.

Methods

get_compression_type_string

View source

@classmethod
get_compression_type_string(
    cls,
    options
)

Convert various option types to a unified string.

Args:

• options: TFRecordOption, TFRecordCompressionType, or string.

Returns:

• Compression type as string (e.g. 'ZLIB', 'GZIP', or '').

Raises:

• ValueError: If compression_type is invalid.

Class Members

• compression_type_map

TFRecordWriter

将记录写入TFRecords文件的类。这个类实现了_enter__和_exit__，可以像普通文件一样与块一起使用。

方法:

__init__

__init__(
    path,
    options=None
)

打开文件路径并创建一个TFRecordWriter写入文件。 

参数:

• path:  TFRecords文件的路径。
• options:  (可选)TFRecordOptions对象。

可能产生的异常:

• IOError: If path cannot be opened for writing.

__enter__

__enter__()

输入一个block。

__exit__

__exit__(
    unused_type,
    unused_value,
    unused_traceback
)

退出bloack，关闭文件。

close

close()

flush

flush()

刷新文件。

write

write(record)

向文件中写入字符串记录。 

参数:

• record:  str

tf.io.tf_record_iterator

An iterator that read the records from a TFRecords file. (deprecated)

Aliases:

• tf.compat.v1.io.tf_record_iterator
• tf.compat.v1.python_io.tf_record_iterator
• tf.python_io.tf_record_iterator

tf.io.tf_record_iterator(
    path,
    options=None
)

Warning: THIS FUNCTION IS DEPRECATED. It will be removed in a future version. Instructions for updating: Use eager execution and: tf.data.TFRecordDataset(path)

Args:

• path: The path to the TFRecords file.
• options: (optional) A TFRecordOptions object.

Yields:

• Strings.

Raises:

• IOError: If path cannot be opened for reading.