tf.raw_ops.ParseSingleExample

Transforms a tf.Example proto (as a string) into typed tensors.

tf.raw_ops.ParseSingleExample(
    serialized, dense_defaults, num_sparse, sparse_keys, dense_keys, sparse_types,
    dense_shapes, name=None
)

Args:

  • serialized: A Tensor of type string. A vector containing a batch of binary serialized Example protos.
  • dense_defaults: A list of Tensor objects with types from: float32, int64, string. A list of Tensors (some may be empty), whose length matches the length of dense_keys. dense_defaults[j] provides default values when the example's feature_map lacks dense_key[j]. If an empty Tensor is provided for dense_defaults[j], then the Feature dense_keys[j] is required. The input type is inferred from dense_defaults[j], even when it's empty. If dense_defaults[j] is not empty, and dense_shapes[j] is fully defined, then the shape of dense_defaults[j] must match that of dense_shapes[j]. If dense_shapes[j] has an undefined major dimension (variable strides dense feature), dense_defaults[j] must contain a single element: the padding element.
  • num_sparse: An int that is >= 0. The number of sparse features to be parsed from the example. This must match the lengths of sparse_keys and sparse_types.
  • sparse_keys: A list of strings. A list of num_sparse strings. The keys expected in the Examples' features associated with sparse values.
  • dense_keys: A list of strings. The keys expected in the Examples' features associated with dense values.
  • sparse_types: A list of tf.DTypes from: tf.float32, tf.int64, tf.string. A list of num_sparse types; the data types of data in each Feature given in sparse_keys. Currently the ParseSingleExample op supports DT_FLOAT (FloatList), DT_INT64 (Int64List), and DT_STRING (BytesList).
  • dense_shapes: A list of shapes (each a tf.TensorShape or list of ints). The shapes of data in each Feature given in dense_keys. The length of this list must match the length of dense_keys. The number of elements in the Feature corresponding to dense_key[j] must always equal dense_shapes[j].NumEntries(). If dense_shapes[j] == (D0, D1, ..., DN) then the shape of output Tensor dense_values[j] will be (D0, D1, ..., DN): In the case dense_shapes[j] = (-1, D1, ..., DN), the shape of the output Tensor dense_values[j] will be (M, D1, .., DN), where M is the number of blocks of elements of length D1 * .... * DN, in the input.
  • name: A name for the operation (optional).

Returns:

A tuple of Tensor objects (sparse_indices, sparse_values, sparse_shapes, dense_values).

  • sparse_indices: A list of num_sparse Tensor objects with type int64.
  • sparse_values: A list of Tensor objects of type sparse_types.
  • sparse_shapes: A list of num_sparse Tensor objects with type int64.
  • dense_values: A list of Tensor objects. Has the same type as dense_defaults.