configuration_yar.py

# Copyright 2026 YARlabs. All rights reserved.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
#     http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
"""Yar model configuration"""

from transformers.configuration_utils import PreTrainedConfig, layer_type_validation
from transformers.modeling_rope_utils import RopeParameters
from transformers.utils import logging


logger = logging.get_logger(__name__)


class YarConfig(PreTrainedConfig):
    r"""
    This is the configuration class to store the configuration of a [`YarModel`]. It is used to instantiate a
    Yar model according to the specified arguments, defining the model architecture. Instantiating a configuration
    with the defaults will yield a similar configuration to that of
    [YARlabs/v5_Embedding-0.5B](https://huggingface.co/YARlabs/v5_Embedding-0.5B).

    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 151936):
            Vocabulary size of the Yar model. Defines the number of different tokens that can be represented by the
            `inputs_ids` passed when calling [`YarModel`]
        hidden_size (`int`, *optional*, defaults to 4096):
            Dimension of the hidden representations.
        intermediate_size (`int`, *optional*, defaults to 22016):
            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*, defaults to 32):
            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, check out [this
            paper](https://huggingface.co/papers/2305.13245). If it is not specified, will default to `32`.
        head_dim (`int`, *optional*, defaults to 128):
            The attention head dimension.
        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 32768):
            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`.
        tie_word_embeddings (`bool`, *optional*, defaults to `False`):
            Whether the model's input and output word embeddings should be tied.
        rope_parameters (`RopeParameters`, *optional*):
            Dictionary containing the configuration parameters for the RoPE embeddings. The dictionary should contain
            a value for `rope_theta` and optionally parameters used for scaling in case you want to use RoPE
            with longer `max_position_embeddings`.
        attention_bias (`bool`, defaults to `False`, *optional*, defaults to `False`):
            Whether to use a bias in the query, key, value and output projection layers during self-attention.
        use_sliding_window (`bool`, *optional*, defaults to `False`):
            Whether to use sliding window attention.
        sliding_window (`int`, *optional*, defaults to 4096):
            Sliding window attention (SWA) window size. If not specified, will default to `4096`.
        max_window_layers (`int`, *optional*, defaults to 28):
            The number of layers using full attention. The first `max_window_layers` layers will use full attention, while any
            additional layer afterwards will use SWA (Sliding Window Attention).
        layer_types (`list`, *optional*):
            Attention pattern for each layer.
        attention_dropout (`float`, *optional*, defaults to 0.0):
            The dropout ratio for the attention probabilities.
        pad_token_id (`int`, *optional*):
            Padding token id.
        bos_token_id (`int`, *optional*):
            Beginning of stream token id.
        eos_token_id (`int`, *optional*):
            End of stream token id.

    ```python
    >>> from transformers import YarModel, YarConfig

    >>> # Initializing a Yar style configuration
    >>> configuration = YarConfig()

    >>> # Initializing a model from the Yar-8B style configuration
    >>> model = YarModel(configuration)

    >>> # Accessing the model configuration
    >>> configuration = model.config
    ```"""

    model_type = "yar"
    keys_to_ignore_at_inference = ["past_key_values"]

    # Default tensor parallel plan for base model `Yar`
    base_model_tp_plan = {
        "layers.*.self_attn.q_proj": "colwise",
        "layers.*.self_attn.k_proj": "colwise",
        "layers.*.self_attn.v_proj": "colwise",
        "layers.*.self_attn.o_proj": "rowwise",
        "layers.*.mlp.gate_proj": "colwise",
        "layers.*.mlp.up_proj": "colwise",
        "layers.*.mlp.down_proj": "rowwise",
    }
    base_model_pp_plan = {
        "embed_tokens": (["input_ids"], ["inputs_embeds"]),
        "layers": (["hidden_states", "attention_mask"], ["hidden_states"]),
        "norm": (["hidden_states"], ["hidden_states"]),
    }

    def __init__(
        self,
        vocab_size: int | None = 151936,
        hidden_size: int | None = 4096,
        intermediate_size: int | None = 22016,
        num_hidden_layers: int | None = 32,
        num_attention_heads: int | None = 32,
        num_key_value_heads: int | None = 32,
        head_dim: int | None = 128,
        hidden_act: str | None = "silu",
        max_position_embeddings: int | None = 32768,
        initializer_range: float | None = 0.02,
        rms_norm_eps: int | None = 1e-6,
        use_cache: bool | None = True,
        tie_word_embeddings: bool | None = False,
        rope_parameters: RopeParameters | dict[str, RopeParameters] | None = None,
        attention_bias: bool | None = False,
        use_sliding_window: bool | None = False,
        sliding_window: int | None = 4096,
        max_window_layers: int | None = 28,
        layer_types: list[str] | None = None,
        attention_dropout: float | None = 0.0,
        pad_token_id: int | None = None,
        bos_token_id: int | None = None,
        eos_token_id: int | None = 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
        self.use_sliding_window = use_sliding_window
        self.sliding_window = sliding_window if self.use_sliding_window else None
        self.max_window_layers = max_window_layers

        # 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.head_dim = head_dim
        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.layer_types = layer_types
        if self.layer_types is None:
            self.layer_types = [
                "sliding_attention"
                if self.sliding_window is not None and i >= self.max_window_layers
                else "full_attention"
                for i in range(self.num_hidden_layers)
            ]
        layer_type_validation(self.layer_types, self.num_hidden_layers)

        self.pad_token_id = pad_token_id
        self.bos_token_id = bos_token_id
        self.eos_token_id = eos_token_id
        self.tie_word_embeddings = tie_word_embeddings
        self.rope_parameters = rope_parameters

        super().__init__(**kwargs)


__all__ = ["YarConfig"]