default_float_dtype#

ivy.default_float_dtype(*, input=None, float_dtype=None, as_native=False)[source]#
Parameters:

  • input (Optional[Union[Array, NativeArray]]) – Number or array for inferring the float dtype. (default: None)

  • float_dtype (Optional[Union[FloatDtype, NativeDtype]]) – The float dtype to be returned. (default: None)

  • as_native (bool) – Whether to return the float dtype as native dtype. (default: False)

Return type:

Union[Dtype, str, NativeDtype]

Returns:

  • Return float_dtype as native or ivy dtype if provided, else

  • if input is given, return its float dtype, otherwise return the

  • global default float dtype.

Examples

>>> ivy.default_float_dtype()
'float32'

>>> ivy.set_default_float_dtype(ivy.FloatDtype("float64"))
>>> ivy.default_float_dtype()
'float64'

>>> ivy.default_float_dtype(float_dtype=ivy.FloatDtype("float16"))
'float16'

>>> ivy.default_float_dtype(input=4294.967346)
'float32'

>>> x = ivy.array([9.8,8.9], dtype="float16")
>>> ivy.default_float_dtype(input=x)
'float16'