---
title: "finfo"
sidebar_label: "finfo"
description: "The clika_runtime finfo class."
---

<!-- Generated by tools/api_reference/generate_api_docs.py. Do not edit. -->

Numeric limits of a floating-point dtype: finfo(clika_runtime.float32).eps, .max, .min, .tiny, ...

## `bits` (property)

(self) -> int

## `denorm_min` (property)

(self) -> float

## `dtype` (property)

(self) -> object

## `eps` (property)

The gap between 1.0 and the next representable value.

## `exponent_bits` (property)

(self) -> int

## `has_denorm` (property)

(self) -> bool

## `has_infinity` (property)

(self) -> bool

## `has_nan` (property)

(self) -> bool

## `is_signed` (property)

(self) -> bool

## `mantissa_bits` (property)

(self) -> int

## `max` (property)

(self) -> float

## `min` (property)

(self) -> float

## `resolution` (property)

The approximate decimal resolution, 10 to the power of minus the mantissa's decimal digits (1e-06 for float32).

## `smallest_normal` (property)

(self) -> float

## `tiny` (property)

The smallest positive normal value (the same as smallest_normal).

## `__init__`

```python
__init____init__(self, dtype: object) -> None
```

__init__(self, dtype: object) -> None

finfo(dtype) -> finfo

The limits of a floating-point dtype; an integer dtype raises InvalidArgumentError.