from transformers import PretrainedConfig

class TransformerConfig(PretrainedConfig):
    model_type = "transformer"

    def __init__(
        self,
        # === Model Architecture ===
        vocab_size=50257,
        hidden_size=1024,
        embedding_size=None,        # if None, defaults to hidden_size
        num_hidden_layers=24,
        num_attention_heads=16,
        intermediate_size=4096,
        max_position_embeddings=1024,
        max_seq_len=1024,
        # === Misc ===
        initializer_range=0.02,
        tie_word_embeddings=True,
        use_cache=False,
        use_causal_attention=True,
        bos_token_id=50256,
        eos_token_id=50256,
        pad_token_id=50256,
        truncate_activation_size=False,
        **kwargs,
    ):
        super().__init__(
            bos_token_id=bos_token_id,
            eos_token_id=eos_token_id,
            pad_token_id=pad_token_id,
            tie_word_embeddings=tie_word_embeddings,
            **kwargs,
        )
        self.vocab_size = vocab_size
        self.hidden_size = hidden_size
        self.embedding_size = embedding_size if embedding_size is not None else hidden_size
        self.num_hidden_layers = num_hidden_layers
        self.num_attention_heads = num_attention_heads
        self.intermediate_size = intermediate_size
        self.initializer_range = initializer_range
        self.max_position_embeddings = max_position_embeddings
        self.use_cache = use_cache
        self.use_causal_attention = use_causal_attention
        self.truncate_activation_size = truncate_activation_size
        self.max_seq_len = max_seq_len