from transformers import PretrainedConfig


class MyBertConfig(PretrainedConfig):
    model_type = "mybert"

    def __init__(
        self,
        vocab_size=28996,                 
        hidden_size=768,
        num_hidden_layers=12,
        num_attention_heads=12,
        intermediate_size=2048,           
        max_position_embeddings=2048,
        hidden_dropout_prob=0.0,
        attention_probs_dropout_prob=0.0,
        layer_norm_eps=1e-5,
        initializer_range=0.02,
        rope_theta=10000.0,
        use_bias=False,
        sparse_prediction=True,           
        pad_token_id=0,
        tie_word_embeddings=True,
        **kwargs,
    ):
        super().__init__(pad_token_id=pad_token_id,
                         tie_word_embeddings=tie_word_embeddings, **kwargs)
        assert hidden_size % num_attention_heads == 0
        self.vocab_size = vocab_size
        self.hidden_size = hidden_size
        self.num_hidden_layers = num_hidden_layers
        self.num_attention_heads = num_attention_heads
        self.intermediate_size = intermediate_size
        self.max_position_embeddings = max_position_embeddings
        self.hidden_dropout_prob = hidden_dropout_prob
        self.attention_probs_dropout_prob = attention_probs_dropout_prob
        self.layer_norm_eps = layer_norm_eps
        self.initializer_range = initializer_range
        self.rope_theta = rope_theta
        self.use_bias = use_bias
        self.sparse_prediction = sparse_prediction