---
title: "NLLLoss"
sidebar_label: "NLLLoss"
description: "The clika_runtime.nn NLLLoss class."
---

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

NLLLoss(weight=None, ignore_index=None, reduction="mean")

    Negative log-likelihood over log-probabilities with the class axis last.

    The module form of ``clika_runtime.nn.functional.nll_loss``: construct it
    once with the settings, then call it like a function.

    Args:
        weight: a per-class weight ``[C]``; unset weighs every class 1; None
            takes the function's own default.
        ignore_index: a target value whose positions contribute no loss; unset
            masks nothing; None takes the function's own default.
        reduction: how the per-position losses reduce.
    

## `__init__`

```python
__init__(self, weight: 'Tensor | None' = None, ignore_index: 'int | None' = None, reduction: '_Reduction' = 'mean') -> 'None'
```

Initialize self.  See help(type(self)) for accurate signature.

## `extra_repr`

```python
extra_repr(self) -> 'str'
```

extra_repr() -> str

One line of per-class detail for :meth:`__repr__`; a layer prints
its geometry here (``in_features=64, out_features=256``).

## `forward`

```python
forward(self, input: 'Tensor', target: 'Tensor') -> 'Tensor'
```

forward(input, target) -> Tensor

Runs ``clika_runtime.nn.functional.nll_loss`` on ``input``, ``target``
with the stored settings; returns a new tensor.