configuration_moonshot.py

from transformers.configuration_utils import PretrainedConfig


class MoonshotConfig(PretrainedConfig):
    r"""
    This is the configuration class to store the configuration of a [`MoonshotModel`].
    It is used to instantiate an Moonshot model according to the specified arguments, defining the model
    architecture. Instantiating a configuration with the defaults will yield a similar configuration to that of
    the MSH-L architecture.

    Configuration objects inherit from  [`PretrainedConfig`] and can be used
    to control the model outputs. Read the documentation from  [`PretrainedConfig`]
    for more information.


    Args:
        vocab_size (`int`, *optional*, defaults to 32000):
            Vocabulary size of the Moonshot model. Defines the number of different tokens that can be represented by the
            `inputs_ids` passed when calling [`MoonshotModel`]
        hidden_size (`int`, *optional*, defaults to 4096):
            Dimension of the hidden representations.
        intermediate_size (`int`, *optional*, defaults to 11008):
            Dimension of the MLP representations.
        num_hidden_layers (`int`, *optional*, defaults to 32):
            Number of hidden layers in the Transformer encoder.
        num_attention_heads (`int`, *optional*, defaults to 32):
            Number of attention heads for each attention layer in the Transformer encoder.
        num_key_value_heads (`int`, *optional*):
            This is the number of key_value heads that should be used to implement Grouped Query Attention. If
            `num_key_value_heads=num_attention_heads`, the model will use Multi Head Attention (MHA), if
            `num_key_value_heads=1 the model will use Multi Query Attention (MQA) otherwise GQA is used. When
            converting a multi-head checkpoint to a GQA checkpoint, each group key and value head should be constructed
            by meanpooling all the original heads within that group. For more details checkout [this
            paper](https://arxiv.org/pdf/2305.13245.pdf). If it is not specified, will default to
            `num_attention_heads`.
        hidden_act (`str` or `function`, *optional*, defaults to `"silu"`):
            The non-linear activation function (function or string) in the decoder.
        pretraining_sequence_length (`int`, *optional*, defaults to 4096):
            The sequence length that this model was trained with
        initializer_range (`float`, *optional*, defaults to 0.02):
            The standard deviation of the truncated_normal_initializer for initializing all weight matrices.
        rms_norm_eps (`float`, *optional*, defaults to 1e-12):
            The epsilon used by the rms normalization layers.
        use_cache (`bool`, *optional*, defaults to `True`):
            Whether or not the model should return the last key/values attentions (not used by all models). Only
            relevant if `config.is_decoder=True`.
        rope_scaling (`Dict`, *optional*):
            Dictionary containing the scaling configuration for the RoPE embeddings. Currently supports three scaling
            strategies: linear and dynamic. Their scaling factor must be an float greater than 1. The expected format
            is `{"type": strategy name, "factor": scaling factor}`
            these scaling strategies behave:
            https://www.reddit.com/r/LocalLLaMA/comments/14mrgpr/dynamically_scaled_rope_further_increases/. This is an
            experimental feature, subject to breaking API changes in future versions.
        Example:

    ```python
    >>> from configuration_moonshot import MoonshotConfig

    >>> # Initializing a MSH-L style configuration
    >>> configuration = MoonshotConfig()
    ```
"""
    model_type = "moonshot"
    keys_to_ignore_at_inference = ["past_key_values"]

    def __init__(
        self,
        vocab_size=163840,
        hidden_size=4096,
        intermediate_size=11008,
        num_hidden_layers=32,
        num_attention_heads=32,
        num_key_value_heads=None,
        hidden_act="silu",
        pretraining_sequence_length=4096,
        initializer_range=0.02,
        rms_norm_eps=1e-6,
        use_cache=True,
        pad_token_id=0,
        bos_token_id=1,
        eos_token_id=2,
        rope_theta=10000.0,
        rope_scaling=None,
        attention_bias=False,
        tie_word_embeddings=False,
        attention_qknorm: dict = None,  #{'type': 'tensor'/ 'head'}
        **kwargs,
    ):
        self.vocab_size = vocab_size
        self.pretraining_sequence_length = pretraining_sequence_length
        self.hidden_size = hidden_size
        self.intermediate_size = intermediate_size
        self.num_hidden_layers = num_hidden_layers
        self.num_attention_heads = num_attention_heads

        # for backward compatibility
        if num_key_value_heads is None:
            num_key_value_heads = num_attention_heads

        self.num_key_value_heads = num_key_value_heads
        self.hidden_act = hidden_act
        self.initializer_range = initializer_range
        self.rms_norm_eps = rms_norm_eps
        self.use_cache = use_cache
        self.rope_theta = rope_theta
        self.rope_scaling = rope_scaling
        self._rope_scaling_validation()
        self.attention_bias = attention_bias
        self.attention_qknorm = attention_qknorm


        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,
        )

    def _rope_scaling_validation(self):
        """
        Validate the `rope_scaling` configuration.
        """
        if self.rope_scaling is None:
            return

        if not isinstance(self.rope_scaling, dict) or len(self.rope_scaling) != 2:
            raise ValueError(
                "`rope_scaling` must be a dictionary with with two fields, `name` and `factor`, "
                f"got {self.rope_scaling}"
            )
        rope_scaling_type = self.rope_scaling.get("type", None)
        rope_scaling_factor = self.rope_scaling.get("factor", None)
        if rope_scaling_type is None or rope_scaling_type not in ["linear", "dynamic"]:
            raise ValueError(
                f"`rope_scaling`'s name field must be one of ['linear', 'dynamic'], got {rope_scaling_type}"
            )
        if rope_scaling_factor is None or not isinstance(rope_scaling_factor, float) or rope_scaling_factor <= 1.0:
            raise ValueError(f"`rope_scaling`'s factor field must be an float > 1, got {rope_scaling_factor}")