vllm.transformers_utils.configs.hyperclovax ¶

class HyperCLOVAXConfig(PretrainedConfig):
    r"""
    This is the configuration class to store the configuration of a
    [`HyperCLOVAXModel`]. It is used to instantiate a HyperCLOVAX model
    according to the specified arguments, defining the model 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 HyperCLOVAX model. Defines the number of
            different tokens that can be represented by the `input_ids`
            passed when calling [`HyperCLOVAXModel`]
        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 decoder.
        num_attention_heads (`int`, *optional*, defaults to 32):
            Number of attention heads for each attention layer in the
            Transformer decoder.
        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.
        max_position_embeddings (`int`, *optional*, defaults to 2048):
            The maximum sequence length that this model might ever be used
            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-06):
            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`.
        pad_token_id (`int`, *optional*):
            Padding token id.
        bos_token_id (`int`, *optional*, defaults to 1):
            Beginning of stream token id.
        eos_token_id (`int`, *optional*, defaults to 2):
            End of stream token id.
        pretraining_tp (`int`, *optional*, defaults to 1):
            Experimental feature. Tensor parallelism rank used during
            pretraining. Please refer to [this document](https://huggingface.
            co/docs/transformers/main/perf_train_gpu_many#tensor-parallelism)
            to understand more about it. This value is necessary to ensure
            exact reproducibility of the pretraining results. Please refer to
            [this issue](https://github.com/pytorch/pytorch/issues/76232).
        tie_word_embeddings (`bool`, *optional*, defaults to `False`):
            Whether to tie weight embeddings
        rope_theta (`float`, *optional*, defaults to 10000.0):
            The base period of the RoPE embeddings.
        rope_scaling (`Dict`, *optional*):
            Dictionary containing the scaling configuration for the RoPE
            embeddings. NOTE: if you apply new rope type and you expect the
            model to work on longer `max_position_embeddings`, we recommend
            you to update this value accordingly.
            Expected contents:
                `rope_type` (`str`):
                    The sub-variant of RoPE to use. Can be one of ['default',
                    'linear', 'dynamic', 'yarn', 'longrope', 'llama3'], with
                    'default' being the original RoPE implementation.
                `factor` (`float`, *optional*):
                    Used with all rope types except 'default'. The scaling
                    factor to apply to the RoPE embeddings. In most scaling
                    types, a `factor` of x will enable the model to handle
                    sequences of length x * original maximum pre-trained
                    length.
                `original_max_position_embeddings` (`int`, *optional*):
                    Used with 'dynamic', 'longrope' and 'llama3'. The
                    original max position embeddings used during pretraining.
                `attention_factor` (`float`, *optional*):
                    Used with 'yarn' and 'longrope'. The scaling factor to be
                    applied on the attention computation. If unspecified, it
                    defaults to value recommended by the implementation, using
                    the `factor` field to infer the suggested value.
                `beta_fast` (`float`, *optional*):
                    Only used with 'yarn'. Parameter to set the boundary for
                    extrapolation (only) in the linear ramp function. If
                    unspecified, it defaults to 32.
                `beta_slow` (`float`, *optional*):
                    Only used with 'yarn'. Parameter to set the boundary for
                    interpolation (only) in the linear ramp function. If
                    unspecified, it defaults to 1.
                `short_factor` (`List[float]`, *optional*):
                    Only used with 'longrope'. The scaling factor to be
                    applied to short contexts (<
                    `original_max_position_embeddings`). Must be a list of
                    numbers with the same length as the hidden size divided
                    by the number of attention heads divided by 2
                `long_factor` (`List[float]`, *optional*):
                    Only used with 'longrope'. The scaling factor to be
                    applied to long contexts (<
                    `original_max_position_embeddings`). Must be a list of
                    numbers with the same length as the hidden size divided
                    by the number of attention heads divided by 2
                `low_freq_factor` (`float`, *optional*):
                    Only used with 'llama3'. Scaling factor applied to low
                    frequency components of the RoPE
                `high_freq_factor` (`float`, *optional*):
                    Only used with 'llama3'. Scaling factor applied to high
                    frequency components of the RoPE
        attention_bias (`bool`, *optional*, defaults to `False`):
            Whether to use a bias in the query, key, value and output
            projection layers during self-attention.
        attention_dropout (`float`, *optional*, defaults to 0.0):
            The dropout ratio for the attention probabilities.
        mlp_bias (`bool`, *optional*, defaults to `False`):
            Whether to use a bias in up_proj, down_proj and gate_proj layers
            in the MLP layers.
        head_dim (`int`, *optional*):
            The attention head dimension. If None, it will default to
            hidden_size // num_heads
        embedding_multiplier (`float`, *optional*, defaults to `None`):
            Multiplier applied to the embedding weights. If `None`, it is
            equivalent to `1.0`.
        logits_scaling (`float`, *optional*, defaults to `None`):
            Scaling factor for logits. If `None`, it is equivalent to `1.0`.
        attention_multiplier (`float`, *optional*, defaults to `None`):
            Multiplier applied to the attention weights. If `None`, it is
            equivalent to `self.head_dim ** -0.5`.
        residual_multiplier (`float`, *optional*, defaults to `None`):
            Scaling factor for residual connections. If `None`, it is
            equivalent to `1.0`.
        use_post_norm (`bool`, *optional*, defaults to `True`):
            Determines whether to apply Peri-Layer Normalization. Set to
            False to disable this feature.
        rope_parameters (`dict`, *optional*):
            Dictionary containing the RoPE parameters used by vLLM's
            `get_rope`. When provided, takes precedence over `rope_theta`
            and `rope_scaling`. If `None`, it is derived from `rope_theta`
            and `rope_scaling` automatically.
    """

    model_type = "hyperclovax"
    keys_to_ignore_at_inference = ["past_key_values"]

    def __init__(
        self,
        vocab_size=32000,
        hidden_size=4096,
        intermediate_size=11008,
        num_hidden_layers=32,
        num_attention_heads=32,
        num_key_value_heads=None,
        hidden_act="silu",
        max_position_embeddings=2048,
        initializer_range=0.02,
        rms_norm_eps=1e-6,
        use_cache=True,
        pad_token_id=None,
        bos_token_id=1,
        eos_token_id=2,
        pretraining_tp=1,
        tie_word_embeddings=False,
        rope_theta=10000.0,
        rope_scaling=None,
        attention_bias=False,
        attention_dropout=0.0,
        mlp_bias=False,
        head_dim=None,
        embedding_multiplier=None,  # mup
        logits_scaling=None,  # mup
        attention_multiplier=None,  # mup
        residual_multiplier=None,  # mup
        use_post_norm=True,  # post-norm(peri-LN)
        rope_parameters=None,
        auto_map=None,
        **kwargs,
    ):
        self.vocab_size = vocab_size
        self.max_position_embeddings = max_position_embeddings
        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.pretraining_tp = pretraining_tp
        self.use_cache = use_cache
        self.rope_theta = rope_theta
        self.rope_scaling = rope_scaling
        self.attention_bias = attention_bias
        self.attention_dropout = attention_dropout
        self.mlp_bias = mlp_bias
        self.head_dim = (
            head_dim
            if head_dim is not None
            else self.hidden_size // self.num_attention_heads
        )
        # Derive rope_parameters for vLLM's get_rope() from rope_theta /
        # rope_scaling, unless the caller already provided rope_parameters.
        if rope_parameters is None:
            if rope_scaling is not None:
                # Shallow-copy to avoid mutating the caller's dict.
                rope_parameters = dict(rope_scaling)
                # BC: 'type' field -> 'rope_type', remove stale key.
                if "type" in rope_parameters:
                    rope_parameters.setdefault("rope_type", rope_parameters.pop("type"))
            else:
                rope_parameters = {"rope_type": "default"}
            if "rope_theta" not in rope_parameters:
                rope_parameters["rope_theta"] = rope_theta
        self.rope_parameters = rope_parameters

        # BC: keep self.rope_scaling consistent for HF serialization.
        if self.rope_scaling is not None and "type" in self.rope_scaling:
            self.rope_scaling["rope_type"] = self.rope_scaling["type"]

        # mup
        self.embedding_multiplier = (
            embedding_multiplier if embedding_multiplier is not None else 1.0
        )
        self.logits_scaling = logits_scaling if logits_scaling is not None else 1.0
        self.attention_multiplier = (
            attention_multiplier
            if attention_multiplier is not None
            else self.head_dim**-0.5
        )
        self.residual_multiplier = (
            residual_multiplier if residual_multiplier is not None else 1.0
        )

        # post-norm (Peri-LN)
        self.use_post_norm = use_post_norm

        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,
            auto_map=auto_map,
            **kwargs,
        )