from transformers.configuration_utils import PretrainedConfig

class STLConfig(PretrainedConfig):

    model_type = "stldec1024"
    keys_to_ignore_at_inference = ["past_key_values"]
    attribute_map = {"num_attention_heads": "encoder_attention_heads", "hidden_size": "d_model"}

    def __init__(
        self,
        vocab_size=35,
        decoder_vocab_size=None, # unused
        max_position_embeddings=1024,
        encoder_layers=12,
        encoder_ffn_dim=4096,
        encoder_attention_heads=16,
        decoder_layers=12,
        decoder_ffn_dim=4096,
        decoder_attention_heads=16,
        encoder_layerdrop=0.0,
        decoder_layerdrop=0.0,
        use_cache=True,
        is_encoder_decoder=True,
        activation_function="gelu",
        d_model=1024,
        dropout=0.1,
        attention_dropout=0.0,
        activation_dropout=0.0,
        init_std=0.02,
        decoder_start_token_id=3,
        scale_embedding=False,
        pad_token_id=1,
        eos_token_id=3,
        bos_token_id=2,
        forced_eos_token_id=3,
        share_encoder_decoder_embeddings=True,
        **kwargs,
    ):
        self.vocab_size = vocab_size
        self.decoder_vocab_size = decoder_vocab_size or vocab_size
        self.max_position_embeddings = max_position_embeddings
        self.d_model = d_model
        self.encoder_ffn_dim = encoder_ffn_dim
        self.encoder_layers = encoder_layers
        self.encoder_attention_heads = encoder_attention_heads
        self.decoder_ffn_dim = decoder_ffn_dim
        self.decoder_layers = decoder_layers
        self.decoder_attention_heads = decoder_attention_heads
        self.dropout = dropout
        self.attention_dropout = attention_dropout
        self.activation_dropout = activation_dropout
        self.activation_function = activation_function
        self.init_std = init_std
        self.encoder_layerdrop = encoder_layerdrop
        self.decoder_layerdrop = decoder_layerdrop
        self.use_cache = use_cache
        self.num_hidden_layers = encoder_layers
        self.scale_embedding = scale_embedding  # scale factor will be sqrt(d_model) if True
        self.share_encoder_decoder_embeddings = share_encoder_decoder_embeddings
        super().__init__(
            bos_token_id=bos_token_id,
            pad_token_id=pad_token_id,
            eos_token_id=eos_token_id,
            is_encoder_decoder=is_encoder_decoder,
            decoder_start_token_id=decoder_start_token_id,
            forced_eos_token_id=forced_eos_token_id,
            **kwargs,
        )