Missed TensorFlow Dev Summit? Check out the video playlist. Watch recordings


TensorFlow 1 version View source on GitHub

Represents the type of the elements in a Tensor.


The following DType objects are defined:

The tf.as_dtype() function converts numpy types and string type names to a DType object.


  • type_enum: A types_pb2.DataType enum value.


  • as_datatype_enum: Returns a types_pb2.DataType enum value based on this DType.
  • as_numpy_dtype: Returns a numpy.dtype based on this DType.
  • base_dtype: Returns a non-reference DType based on this DType.
  • is_bool: Returns whether this is a boolean data type.
  • is_complex: Returns whether this is a complex floating point type.
  • is_floating: Returns whether this is a (non-quantized, real) floating point type.
  • is_integer: Returns whether this is a (non-quantized) integer type.
  • is_numpy_compatible
  • is_quantized: Returns whether this is a quantized data type.
  • is_unsigned: Returns whether this type is unsigned.

    Non-numeric, unordered, and quantized types are not considered unsigned, and this function returns False.

  • limits: Return intensity limits, i.e.

    (min, max) tuple, of the dtype. Args: clip_negative : bool, optional If True, clip the negative range (i.e. return 0 for min intensity) even if the image dtype allows negative values. Returns min, max : tuple Lower and upper intensity limits.

  • max: Returns the maximum representable value in this data type.

  • min: Returns the minimum representable value in this data type.

  • name: Returns the string name for this DType.

  • real_dtype: Returns the dtype correspond to this dtype's real part.

  • size


  • TypeError: If type_enum is not a value types_pb2.DataType.



View source


Returns True iff this DType refers to the same type as other.


View source


Returns True iff self != other.


View source


Returns True if the other DType will be converted to this DType.

The conversion rules are as follows:

DType(T)       .is_compatible_with(DType(T))        == True


  • other: A DType (or object that may be converted to a DType).


True if a Tensor of the other DType will be implicitly converted to this DType.