# coding=utf-8 # Copyright 2025 The Google AI Language Team Authors and The HuggingFace Inc. team. # Copyright (c) 2018, NVIDIA CORPORATION. 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. """BERT2D model configuration""" from collections import OrderedDict from typing import Mapping from transformers.configuration_utils import PretrainedConfig from transformers.onnx import OnnxConfig from transformers.utils import logging logger = logging.get_logger(__name__) class Bert2DConfig(PretrainedConfig): r""" This is the configuration class to store the configuration of a [`Bert2DModel`] or a [`TFBert2DModel`]. It is used to instantiate a BERT2D 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 BERT2D [google-bert2d/bert2d-base-uncased](https://huggingface.co/google-bert2d/bert2d-base-uncased) 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 30522): Vocabulary size of the BERT2D model. Defines the number of different tokens that can be represented by the `inputs_ids` passed when calling [`Bert2DModel`] or [`TFBert2DModel`]. hidden_size (`int`, *optional*, defaults to 768): Dimensionality of the encoder layers and the pooler layer. num_hidden_layers (`int`, *optional*, defaults to 12): Number of hidden layers in the Transformer encoder. num_attention_heads (`int`, *optional*, defaults to 12): Number of attention heads for each attention layer in the Transformer encoder. intermediate_size (`int`, *optional*, defaults to 3072): Dimensionality of the "intermediate" (often named feed-forward) layer in the Transformer encoder. hidden_act (`str` or `Callable`, *optional*, defaults to `"gelu"`): The non-linear activation function (function or string) in the encoder and pooler. If string, `"gelu"`, `"relu"`, `"silu"` and `"gelu_new"` are supported. hidden_dropout_prob (`float`, *optional*, defaults to 0.1): The dropout probability for all fully connected layers in the embeddings, encoder, and pooler. attention_probs_dropout_prob (`float`, *optional*, defaults to 0.1): The dropout ratio for the attention probabilities. max_position_embeddings (`int`, *optional*, defaults to 512): The maximum sequence length that this model might ever be used with. Typically set this to something large just in case (e.g., 512 or 1024 or 2048). type_vocab_size (`int`, *optional*, defaults to 2): The vocabulary size of the `token_type_ids` passed when calling [`Bert2DModel`] or [`TFBert2DModel`]. initializer_range (`float`, *optional*, defaults to 0.02): The standard deviation of the truncated_normal_initializer for initializing all weight matrices. layer_norm_eps (`float`, *optional*, defaults to 1e-12): The epsilon used by the layer normalization layers. position_embedding_type (`str`, *optional*, defaults to `"absolute"`): Type of position embedding. Choose one of `"absolute"`, `"relative_key"`, `"relative_key_query"`. For positional embeddings use `"absolute"`. For more information on `"relative_key"`, please refer to [Self-Attention with Relative Position Representations (Shaw et al.)](https://arxiv.org/abs/1803.02155). For more information on `"relative_key_query"`, please refer to *Method 4* in [Improve Transformer Models with Better Relative Position Embeddings (Huang et al.)](https://arxiv.org/abs/2009.13658). is_decoder (`bool`, *optional*, defaults to `False`): Whether the model is used as a decoder or not. If `False`, the model is used as an encoder. 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`. classifier_dropout (`float`, *optional*): The dropout ratio for the classification head. max_word_position_embeddings (`int`, *optional*, defaults to `512`): The maximum number of word-level positions. This is relevant for models that distinguish between word positions and subword positions within words. max_intermediate_subword_position_embeddings (`int`, *optional*, defaults to `1`): The maximum number of distinct positions for intermediate subwords within a single word. This parameter is used when subword tokenization results in multiple subwords for a single word, and the model needs to assign different positional embeddings to these subwords. For example, if a word is tokenized into three subwords (first, intermediate, last), this parameter controls how many distinct embeddings can be assigned to the "intermediate" subword(s). A value of `1` means all intermediate subwords share the same positional embedding. Examples: ```python >>> from transformers import Bert2DConfig, Bert2DModel >>> # Initializing a BERT2D google-bert2d/bert2d-base-uncased style configuration >>> configuration = Bert2DConfig() >>> # Initializing a model (with random weights) from the google-bert2d/bert2d-base-uncased style configuration >>> model = Bert2DModel(configuration) >>> # Accessing the model configuration >>> configuration = model.config ```""" model_type = "bert2d" def __init__( self, vocab_size=30522, hidden_size=768, num_hidden_layers=12, num_attention_heads=12, intermediate_size=3072, hidden_act="gelu", hidden_dropout_prob=0.1, attention_probs_dropout_prob=0.1, max_position_embeddings=512, type_vocab_size=2, initializer_range=0.02, layer_norm_eps=1e-12, pad_token_id=0, position_embedding_type="absolute", use_cache=True, classifier_dropout=None, max_word_position_embeddings=512, max_intermediate_subword_position_embeddings=1, **kwargs, ): super().__init__(pad_token_id=pad_token_id, **kwargs) self.vocab_size = vocab_size self.hidden_size = hidden_size self.num_hidden_layers = num_hidden_layers self.num_attention_heads = num_attention_heads self.hidden_act = hidden_act self.intermediate_size = intermediate_size self.hidden_dropout_prob = hidden_dropout_prob self.attention_probs_dropout_prob = attention_probs_dropout_prob self.max_position_embeddings = max_position_embeddings self.type_vocab_size = type_vocab_size self.initializer_range = initializer_range self.layer_norm_eps = layer_norm_eps self.position_embedding_type = position_embedding_type self.use_cache = use_cache self.classifier_dropout = classifier_dropout self.max_word_position_embeddings = max_word_position_embeddings self.max_intermediate_subword_position_embeddings = max_intermediate_subword_position_embeddings class Bert2DOnnxConfig(OnnxConfig): @property def inputs(self) -> Mapping[str, Mapping[int, str]]: if self.task == "multiple-choice": dynamic_axis = {0: "batch", 1: "choice", 2: "sequence"} else: dynamic_axis = {0: "batch", 1: "sequence"} return OrderedDict( [ ("input_ids", dynamic_axis), ("attention_mask", dynamic_axis), ("token_type_ids", dynamic_axis), ] ) __all__ = ["Bert2DConfig", "Bert2DOnnxConfig"]