<br />

|---|
| [![](https://www.tensorflow.org/images/GitHub-Mark-32px.png) View source on GitHub](https://github.com/tensorflow/models/blob/v2.20.0/official/nlp/modeling/layers/cls_head.py#L115-L210) |

Pooling heads sharing the same pooling stem.

    tfm.nlp.layers.MultiClsHeads(
        inner_dim,
        cls_list,
        cls_token_idx=0,
        activation='tanh',
        dropout_rate=0.0,
        initializer='glorot_uniform',
        **kwargs
    )

<br />

<br />

<br />

| ## Args ||
|---|---|
| `inner_dim` | The dimensionality of inner projection layer. If 0 or `None` then only the output projection layer is created. |
| `cls_list` | a list of pairs of (classification problem name and the numbers of classes. |
| `cls_token_idx` | The index inside the sequence to pool. |
| `activation` | Dense layer activation. |
| `dropout_rate` | Dropout probability. |
| `initializer` | Initializer for dense layer kernels. |
| `**kwargs` | Keyword arguments. |

<br />

<br />

<br />

<br />

| ## Attributes ||
|---|---|
| `checkpoint_items` | <br /> <br /> |

<br />

## Methods

### `call`

[View source](https://github.com/tensorflow/models/blob/v2.20.0/official/nlp/modeling/layers/cls_head.py#L162-L188)

    call(
        features: tf.Tensor, only_project: bool = False
    )

Implements call().

<br />

<br />

<br />

| Args ||
|---|---|
| `features` | a rank-3 Tensor when self.inner_dim is specified, otherwise it is a rank-2 Tensor. |
| `only_project` | a boolean. If True, we return the intermediate Tensor before projecting to class logits. |

<br />

<br />

<br />

<br />

| Returns ||
|---|---|
| If only_project is True, a Tensor with shape= \[batch size, hidden size\]. If only_project is False, a dictionary of Tensors. ||

<br />