ht (`int`, *optional*, defaults to 5.0):
            The weight for the dice loss.
        train_num_points (`str` or `function`, *optional*, defaults to 12544):
            Number of points used for sampling during loss calculation.
        oversample_ratio (`float`, *optional*, defaults to 3.0):
            Oversampling parameter used for calculating no. of sampled points
        importance_sample_ratio (`float`, *optional*, defaults to 0.75):
            Ratio of points that are sampled via importance sampling.
        init_std (`float`, *optional*, defaults to 0.02):
            The standard deviation of the truncated_normal_initializer for initializing all weight matrices.
        init_xavier_std (`float``, *optional*, defaults to 1.0):
            The scaling factor used for the Xavier initialization gain in the HM Attention map module.
        use_auxiliary_loss (`boolean``, *optional*, defaults to `True`):
            If `True` [`Mask2FormerForUniversalSegmentationOutput`] will contain the auxiliary losses computed using
            the logits from each decoder's stage.
        feature_strides (`List[int]`, *optional*, defaults to `[4, 8, 16, 32]`):
            Feature strides corresponding to features generated from backbone network.
        output_auxiliary_logits (`bool`, *optional*):
            Should the model output its `auxiliary_logits` or not.

    Examples:

    ```python
    >>> from transformers import Mask2FormerConfig, Mask2FormerModel

    >>> # Initializing a Mask2Former facebook/mask2former-swin-small-coco-instance configuration
    >>> configuration = Mask2FormerConfig()

    >>> # Initializing a model (with random weights) from the facebook/mask2former-swin-small-coco-instance style configuration
    >>> model = Mask2FormerModel(configuration)

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

    Z