configuration_wedlm.py

# coding=utf-8
# Copyright 2024 The WeDLM team and the HuggingFace Inc. team. 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.
"""WeDLM model configuration"""

from transformers.configuration_utils import PretrainedConfig
from transformers.modeling_rope_utils import rope_config_validation
from transformers.utils import logging


logger = logging.get_logger(__name__)


class WeDLMConfig(PretrainedConfig):
    r"""
    This is the configuration class to store the configuration of a [`WeDLMModel`]. It is used to instantiate an
    WeDLM 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 151936):
            Vocabulary size of the WeDLM model. Defines the number of different tokens that can be represented by the
            `inputs_ids` passed when calling [`WeDLMModel`]
        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.
        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_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.
        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.
        attention_dropout (`float`, *optional*, defaults to 0.0):
            The dropout ratio for the attention probabilities.
        attention_bias (`bool`, *optional*, defaults to `True`):
            Whether to use bias in QKV projections. Set to `True` for Qwen2.5 compatibility,
            `False` for Qwen3 compatibility.
        qk_norm (`bool`, *optional*, defaults to `False`):
            Whether to use QK normalization. Set to `True` for Qwen3 compatibility.
        head_dim (`int`, *optional*):
            The dimension of each attention head. If not specified, defaults to hidden_size // num_attention_heads.
    """

    model_type = "wedlm"
    keys_to_ignore_at_inference = ["past_key_values"]

    def __init__(
        self,
        vocab_size=151936,
        hidden_size=4096,
        intermediate_size=22016,
        num_hidden_layers=32,
        num_attention_heads=32,
        num_key_value_heads=32,
        hidden_act="silu",
        max_position_embeddings=32768,
        initializer_range=0.02,
        rms_norm_eps=1e-6,
        use_cache=True,
        tie_word_embeddings=False,
        rope_theta=10000.0,
        rope_scaling=None,
        use_sliding_window=False,
        sliding_window=4096,
        max_window_layers=28,
        attention_dropout=0.0,
        attention_bias=True,  
        qk_norm=False, 
        head_dim=None,
        mask_token_id=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.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.attention_dropout = attention_dropout
        self.attention_bias = attention_bias
        self.qk_norm = qk_norm 
        self.mask_token_id = mask_token_id
        
        if head_dim is None:
            self.head_dim = hidden_size // num_attention_heads
        else:
            self.head_dim = head_dim
        
        # Validate the correctness of rotary position embeddings parameters
        # BC: if there is a 'type' field, move it to 'rope_type'.
        if self.rope_scaling is not None and "type" in self.rope_scaling:
            self.rope_scaling["rope_type"] = self.rope_scaling["type"]
        rope_config_validation(self)

        # Generate layer_types based on sliding window configuration
        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)
        ]

        super().__init__(
            tie_word_embeddings=tie_word_embeddings,
            **kwargs,
        )


__all__ = ["WeDLMConfig"]