from transformers import PretrainedConfig


class BinaryLLMConfig(PretrainedConfig):
    model_type = "binaryllm"

    def __init__(
        self,
        vocab_size: int = 4,
        hidden_size: int = 384,
        num_hidden_layers: int = 6,
        num_attention_heads: int = 6,
        intermediate_size: int = 1536,
        max_position_embeddings: int = 4096,
        dropout: float = 0.1,
        activation: str = "gelu",
        **kwargs,
    ):
        self.vocab_size = int(vocab_size)
        self.hidden_size = int(hidden_size)
        self.num_hidden_layers = int(num_hidden_layers)
        self.num_attention_heads = int(num_attention_heads)
        self.intermediate_size = int(intermediate_size)
        self.max_position_embeddings = int(max_position_embeddings)
        self.dropout = float(dropout)
        self.activation = str(activation)
        super().__init__(**kwargs)