from transformers import PretrainedConfig

class BertItalianoConfig(PretrainedConfig):
    model_type="BertItaliano"

    def __init__(
        self,
        attention_probs_dropout_prob: float = 0.1,
        gradient_checkpointing: bool = False,
        hidden_act="gelu",
        hidden_dropout_prob: float = 0.1,
        hidden_size: int = 768,
        initializer_range: float = 0.02,
        intermediate_size: int = 3072,
        layer_norm_eps: float = 1e-12,
        max_position_embeddings: int = 512,
        num_attention_heads: int = 12,
        num_hidden_layers: int = 12,
        pad_token_id: int = 0,
        position_embedding_type="absolute",
        transformers_version="4.10.3",
        torch_dtype="float32",
        type_vocab_size: int = 2,
        use_cache: bool = True,
        vocab_size: int = 32102,                
        **kwargs,
    ):
        self.attention_probs_dropout_prob = attention_probs_dropout_prob
        self.gradient_checkpointing = gradient_checkpointing
        self.hidden_act = hidden_act
        self.hidden_dropout_prob = hidden_dropout_prob
        self.hidden_size = hidden_size
        self.initializer_range = initializer_range
        self.intermediate_size = intermediate_size
        self.layer_norm_eps = layer_norm_eps
        self.max_position_embeddings = max_position_embeddings
        self.num_attention_heads = num_attention_heads
        self.num_hidden_layers = num_hidden_layers
        self.pad_token_id = pad_token_id
        self.position_embedding_type = position_embedding_type
        self.transformers_version = transformers_version
        self.torch_dtype = torch_dtype
        self.type_vocab_size = type_vocab_size
        self.use_cache = use_cache
        self.vocab_size = vocab_size        
        super().__init__(**kwargs)