numpy.finfo#

class numpy.finfo(dtype)[source]#

Machine limits for floating point types.

Parameters:

dtypefloat, dtype, or instance: Kind of floating point or complex floating point data-type about which to get information.

See also

iinfo: The equivalent for integer data types.
spacing: The distance between a value and the nearest adjacent number
nextafter: The next floating point value after x1 towards x2

Notes

For developers of NumPy: do not instantiate this at the module level. The initial calculation of these parameters is expensive and negatively impacts import times. These objects are cached, so calling finfo() repeatedly inside your functions is not a problem.

Note that smallest_normal is not actually the smallest positive representable value in a NumPy floating point type. As in the IEEE-754 standard [1], NumPy floating point types make use of subnormal numbers to fill the gap between 0 and smallest_normal. However, subnormal numbers may have significantly reduced precision [2].

This function can also be used for complex data types as well. If used, the output will be the same as the corresponding real float type (e.g. numpy.finfo(numpy.csingle) is the same as numpy.finfo(numpy.single)). However, the output is true for the real and imaginary components.

References

[1]

IEEE Standard for Floating-Point Arithmetic, IEEE Std 754-2008, pp.1-70, 2008, https://doi.org/10.1109/IEEESTD.2008.4610935

[2]

Wikipedia, “Denormal Numbers”, https://en.wikipedia.org/wiki/Denormal_number

Examples

>>> np.finfo(np.float64).dtype
dtype('float64')
>>> np.finfo(np.complex64).dtype
dtype('float32')

Attributes:

bitsint: The number of bits occupied by the type.
dtypedtype: Returns the dtype for which finfo returns information. For complex input, the returned dtype is the associated float* dtype for its real and complex components.
epsfloat: The difference between 1.0 and the next smallest representable float larger than 1.0. For example, for 64-bit binary floats in the IEEE-754 standard, eps = 2**-52, approximately 2.22e-16.
epsnegfloat: The difference between 1.0 and the next smallest representable float less than 1.0. For example, for 64-bit binary floats in the IEEE-754 standard, epsneg = 2**-53, approximately 1.11e-16.
iexpint: The number of bits in the exponent portion of the floating point representation.
machepint: The exponent that yields eps.
maxfloating point number of the appropriate type: The largest representable number.
maxexpint: The smallest positive power of the base (2) that causes overflow.
minfloating point number of the appropriate type: The smallest representable number, typically -max.
minexpint: The most negative power of the base (2) consistent with there being no leading 0’s in the mantissa.
negepint: The exponent that yields epsneg.
nexpint: The number of bits in the exponent including its sign and bias.
nmantint: The number of bits in the mantissa.
precisionint: The approximate number of decimal digits to which this kind of float is precise.
resolutionfloating point number of the appropriate type: The approximate decimal resolution of this type, i.e., 10**-precision.
tinyfloat: Return the value for tiny, alias of smallest_normal.
smallest_normalfloat: Return the value for the smallest normal.
smallest_subnormalfloat: The smallest positive floating point number with 0 as leading bit in the mantissa following IEEE-754.