# tensorplay.special._elementwise API

Source: https://www.tensorplay.cn/docs/api/tensorplay.special._elementwise.html

## Functions 4

[#](#api-tensorplay.special._elementwise.log_softmax)

### log_softmax

function[Full reference ↗](/docs/generated/tensorplay.special._elementwise.log_softmax.html)

```python
tensorplay.special._elementwise.log_softmax(input, dim=None, *, dtype=None)
```

log(softmax(input, dim)), evaluated without the intermediate exp.

[#](#api-tensorplay.special._elementwise.logsumexp)

### logsumexp

function[Full reference ↗](/docs/generated/tensorplay.special._elementwise.logsumexp.html)

```python
tensorplay.special._elementwise.logsumexp(input, dim=None, keepdim=False, *, out=None)
```

Log-sum-exp along dim, computed with the max-shift trick.

[#](#api-tensorplay.special._elementwise.round)

### round

function[Full reference ↗](/docs/generated/tensorplay.special._elementwise.round.html)

```python
tensorplay.special._elementwise.round(input)
```

Rounds to nearest even integer (native Tensor.round).

[#](#api-tensorplay.special._elementwise.softmax)

### softmax

function[Full reference ↗](/docs/generated/tensorplay.special._elementwise.softmax.html)

```python
tensorplay.special._elementwise.softmax(input, dim=None, *, dtype=None)
```

Normalized exponential along dim (default: the last dim).