# tensorplay.functional.argmax

Source: https://www.tensorplay.cn/docs/generated/tensorplay.functional.argmax.html

```python
tensorplay.functional.argmax(input) → LongTensor
```

Returns the indices of the maximum value of all elements in the input tensor.

This is the second value returned by [tensorplay.max()](/docs/generated/tensorplay.max.html#tensorplay.max). See its
documentation for the exact semantics of this method.

> **Note**
>
> If there are multiple maximal values then the indices of the first maximal value are returned.

Parameters:

input ([Tensor](/docs/generated/tensorplay.Tensor.html#tensorplay.Tensor)) – the input tensor.

Example:

```
>>> a = tensorplay.randn(4, 4)
>>> a
tensor([[ 1.3398,  0.2663, -0.2686,  0.2450],
        [-0.7401, -0.8805, -0.3402, -1.1936],
        [ 0.4907, -1.3948, -1.0691, -0.3132],
        [-1.6092,  0.5419, -0.2993,  0.3195]])
>>> tensorplay.argmax(a)
tensor(0)
```

```python
tensorplay.functional.argmax(input, dim, keepdim=False) → LongTensor
```

Returns the indices of the maximum values of a tensor across a dimension.

This is the second value returned by [tensorplay.max()](/docs/generated/tensorplay.max.html#tensorplay.max). See its
documentation for the exact semantics of this method.

Parameters:

- input ([Tensor](/docs/generated/tensorplay.Tensor.html#tensorplay.Tensor)) – the input tensor.

- dim – the dimension to reduce.

Example:

```
>>> a = tensorplay.randn(4, 4)
>>> a
tensor([[ 1.3398,  0.2663, -0.2686,  0.2450],
        [-0.7401, -0.8805, -0.3402, -1.1936],
        [ 0.4907, -1.3948, -1.0691, -0.3132],
        [-1.6092,  0.5419, -0.2993,  0.3195]])
>>> tensorplay.argmax(a, dim=1)
tensor([ 0,  2,  0,  1])
```