Type of position embedding. Choose one of `"absolute"`, `"relative_key"`, `"relative_key_query"`. For
            positional embeddings use `"absolute"`. For more information on `"relative_key"`, please refer to
            [Self-Attention with Relative Position Representations (Shaw et al.)](https://arxiv.org/abs/1803.02155).
            For more information on `"relative_key_query"`, please refer to *Method 4* in [Improve Transformer Models
            with Better Relative Position Embeddings (Huang et al.)](https://arxiv.org/abs/2009.13658).
        is_decoder (`bool`, *optional*, defaults to `False`):
            Whether the model is used as a decoder or not. If `False`, the model is used as an encoder.
        use_cache (`bool`, *optional*, defaults to `True`):
            Whether or not the model should return the last key/values attentions (not used by all models). Only
            relevant if `config.is_decoder=True`.
        classifier_dropout (`float`, *optional*):
            The dropout ratio for the classification head.
        pre_norm (`bool`, *optional*, defaults to `False`):
            Whether to apply layer normalization before each block.
        adapter_reduction_factor (`int` or `float`, *optional*, defaults to 2):
            The factor by which the dimensionality of the adapter is reduced relative to `hidden_size`.
        adapter_layer_norm (`bool`, *optional*, defaults to `False`):
            Whether to apply a new layer normalization before the adapter modules (shared across all adapters).
        adapter_reuse_layer_norm (`bool`, *optional*, defaults to `True`):
            Whether to reuse the second layer normalization and apply it before the adapter modules as well.
        ln_before_adapter (`bool`, *optional*, defaults to `True`):
            Whether to apply the layer normalization before the residual connection around the adapter module.
        languages (`Iterable[str]`, *optional*, defaults to `["en_XX"]`):
            An iterable of language codes for which adapter modules should be initialized.
        default_language (`str`, *optional*):
            Language code of a default language. It will be assumed that the input is in this language if no language
            codes are explicitly passed to the forward method.

    Examples:

    ```python
    >>> from transformers import XmodConfig, XmodModel

    >>> # Initializing an X-MOD facebook/xmod-base style configuration
    >>> configuration = XmodConfig()

    >>> # Initializing a model (with random weights) from the facebook/xmod-base style configuration
    >>> model = XmodModel(configuration)

    >>> # Accessing the model configuration
    >>> configuration = model.config
    ```Z