from transformers import PretrainedConfig


class AgoraConfig(PretrainedConfig):
    r"""
    Configuration class for the Agora model.

    Args:
        vocab_size (`int`, *optional*, defaults to 32000):
            Vocabulary size of the Agora model.
        hidden_size (`int`, *optional*, defaults to 2048):
            Dimensionality of the embeddings and hidden states.
        intermediate_size (`int`, *optional*, defaults to 8192):
            Dimensionality of the MLP representations.
        num_hidden_layers (`int`, *optional*, defaults to 24):
            Number of hidden layers in the Transformer encoder.
        num_attention_heads (`int`, *optional*, defaults to 16):
            Number of attention heads for each attention layer.
        num_key_value_heads (`int`, *optional*, defaults to 8):
            Number of key/value heads (Grouped Query Attention).
        head_dim (`int`, *optional*, defaults to 128):
            Dimension per attention head.
        max_position_embeddings (`int`, *optional*, defaults to 4096):
            Maximum sequence length.
        rope_theta (`float`, *optional*, defaults to 10000.0):
            Base period for RoPE embeddings.
        hidden_act (`str`, *optional*, defaults to `"silu"`):
            Activation function in MLP layers.
        rms_norm_eps (`float`, *optional*, defaults to 1e-5):
            Epsilon value for RMSNorm.
        use_cache (`bool`, *optional*, defaults to `True`):
            Whether to use KV cache during generation.
        tie_word_embeddings (`bool`, *optional*, defaults to `False`):
            Whether to tie input/output embeddings.
        attention_dropout (`float`, *optional*, defaults to 0.0):
            Dropout probability for attention weights.
    """

    model_type = "agora"
    keys_to_ignore_at_inference = ["past_key_values"]

    def __init__(
        self,
        vocab_size=32000,
        hidden_size=2048,
        intermediate_size=8192,
        num_hidden_layers=24,
        num_attention_heads=16,
        num_key_value_heads=8,
        head_dim=128,
        max_position_embeddings=4096,
        rope_theta=10000.0,
        rope_scaling=None,
        hidden_act="silu",
        initializer_range=0.02,
        rms_norm_eps=1e-5,
        use_cache=True,
        tie_word_embeddings=False,
        attention_bias=False,
        attention_dropout=0.0,
        mlp_bias=False,
        pad_token_id=0,
        bos_token_id=1,
        eos_token_id=2,
        **kwargs,
    ):
        self.vocab_size = vocab_size
        self.hidden_size = hidden_size
        self.intermediate_size = intermediate_size
        self.num_hidden_layers = num_hidden_layers
        self.num_attention_heads = num_attention_heads
        self.num_key_value_heads = num_key_value_heads
        self.head_dim = head_dim
        self.max_position_embeddings = max_position_embeddings
        self.rope_theta = rope_theta
        self.rope_scaling = rope_scaling
        self.hidden_act = hidden_act
        self.initializer_range = initializer_range
        self.rms_norm_eps = rms_norm_eps
        self.use_cache = use_cache
        self.attention_bias = attention_bias
        self.attention_dropout = attention_dropout
        self.mlp_bias = mlp_bias

        super().__init__(
            pad_token_id=pad_token_id,
            bos_token_id=bos_token_id,
            eos_token_id=eos_token_id,
            tie_word_embeddings=tie_word_embeddings,
            **kwargs,
        )