# Copyright 2023 The TensorFlow Authors. 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. # Copyright 2022 The TensorFlow Authors. 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. """Linformer model configurations and instantiation methods.""" import dataclasses import tensorflow as tf from official.modeling import tf_utils from official.modeling.hyperparams import base_config from official.nlp.configs import encoders from official.projects.lra.linformer_encoder import LinformerEncoder @dataclasses.dataclass class LinformerEncoderConfig(encoders.BertEncoderConfig): """Extra paramerters for Linformer configs. Attributes: pad_token_id: the token id for the pad token low_rank_features: number of dimensions for low-rank projection """ pad_token_id: int = 0 low_rank_features: int = 256 @base_config.bind(LinformerEncoderConfig) def get_encoder(encoder_cfg: LinformerEncoderConfig): """Gets a 'LinformerEncoder' object. Args: encoder_cfg: A 'LinformerEncoderConfig'. Returns: A encoder object. """ encoder = LinformerEncoder( vocab_size=encoder_cfg.vocab_size, hidden_size=encoder_cfg.hidden_size, num_layers=encoder_cfg.num_layers, num_attention_heads=encoder_cfg.num_attention_heads, low_rank_features=encoder_cfg.low_rank_features, inner_dim=encoder_cfg.intermediate_size, inner_activation=tf_utils.get_activation(encoder_cfg.hidden_activation), output_dropout=encoder_cfg.dropout_rate, attention_dropout=encoder_cfg.attention_dropout_rate, max_sequence_length=encoder_cfg.max_position_embeddings, type_vocab_size=encoder_cfg.type_vocab_size, initializer=tf.keras.initializers.TruncatedNormal( stddev=encoder_cfg.initializer_range ), output_range=encoder_cfg.output_range, embedding_width=encoder_cfg.embedding_size, norm_first=encoder_cfg.norm_first, ) return encoder