DeBERTa¶

Overview¶

The DeBERTa model was proposed in DeBERTa: Decoding-enhanced BERT with Disentangled Attention by Pengcheng He, Xiaodong Liu, Jianfeng Gao, Weizhu Chen It is based on Google’s BERT model released in 2018 and Facebook’s RoBERTa model released in 2019.

It builds on RoBERTa with disentangled attention and enhanced mask decoder training with half of the data used in RoBERTa.

The abstract from the paper is the following:

Recent progress in pre-trained neural language models has significantly improved the performance of many natural language processing (NLP) tasks. In this paper we propose a new model architecture DeBERTa (Decoding-enhanced BERT with disentangled attention) that improves the BERT and RoBERTa models using two novel techniques. The first is the disentangled attention mechanism, where each word is represented using two vectors that encode its content and position, respectively, and the attention weights among words are computed using disentangled matrices on their contents and relative positions. Second, an enhanced mask decoder is used to replace the output softmax layer to predict the masked tokens for model pretraining. We show that these two techniques significantly improve the efficiency of model pre-training and performance of downstream tasks. Compared to RoBERTa-Large, a DeBERTa model trained on half of the training data performs consistently better on a wide range of NLP tasks, achieving improvements on MNLI by +0.9% (90.2% vs. 91.1%), on SQuAD v2.0 by +2.3% (88.4% vs. 90.7%) and RACE by +3.6% (83.2% vs. 86.8%). The DeBERTa code and pre-trained models will be made publicly available at https://github.com/microsoft/DeBERTa.

The original code can be found here.

DebertaConfig¶

class transformers.DebertaConfig(vocab_size=50265, 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=0, initializer_range=0.02, layer_norm_eps=1e-07, relative_attention=False, max_relative_positions=- 1, pad_token_id=0, position_biased_input=True, pos_att_type=None, pooler_dropout=0, pooler_hidden_act='gelu', **kwargs)[source]¶

This is the configuration class to store the configuration of a DebertaModel or a TFDebertaModel. It is used to instantiate a DeBERTa 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 DeBERTa microsoft/deberta-base architecture.

Configuration objects inherit from PretrainedConfig and can be used to control the model outputs. Read the documentation from PretrainedConfig for more information.

Parameters
  • vocab_size (int, optional, defaults to 30522) – Vocabulary size of the DeBERTa model. Defines the number of different tokens that can be represented by the inputs_ids passed when calling DebertaModel or TFDebertaModel.

  • 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", "gelu", "tanh", "gelu_fast", "mish", "linear", "sigmoid" 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 DebertaModel or TFDebertaModel.

  • 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.

  • relative_attention (bool, optional, defaults to False) – Whether use relative position encoding.

  • max_relative_positions (int, optional, defaults to 1) – The range of relative positions [-max_position_embeddings, max_position_embeddings]. Use the same value as max_position_embeddings.

  • pad_token_id (int, optional, defaults to 0) – The value used to pad input_ids.

  • position_biased_input (bool, optional, defaults to True) – Whether add absolute position embedding to content embedding.

  • pos_att_type (List[str], optional) – The type of relative position attention, it can be a combination of ["p2c", "c2p", "p2p"], e.g. ["p2c"], ["p2c", "c2p"], ["p2c", "c2p", 'p2p"].

  • layer_norm_eps – The epsilon used by the layer normalization layers.

DebertaTokenizer¶

class transformers.DebertaTokenizer(vocab_file, do_lower_case=False, unk_token='[UNK]', sep_token='[SEP]', pad_token='[PAD]', cls_token='[CLS]', mask_token='[MASK]', **kwargs)[source]¶

Constructs a DeBERTa tokenizer, which runs end-to-end tokenization: punctuation splitting + wordpiece

Parameters
  • vocab_file (str) – File containing the vocabulary.

  • do_lower_case (bool, optional, defaults to True) – Whether or not to lowercase the input when tokenizing.

  • unk_token (str, optional, defaults to "[UNK]") – The unknown token. A token that is not in the vocabulary cannot be converted to an ID and is set to be this token instead.

  • sep_token (str, optional, defaults to "[SEP]") – The separator token, which is used when building a sequence from multiple sequences, e.g. two sequences for sequence classification or for a text and a question for question answering. It is also used as the last token of a sequence built with special tokens.

  • pad_token (str, optional, defaults to "[PAD]") – The token used for padding, for example when batching sequences of different lengths.

  • cls_token (str, optional, defaults to "[CLS]") – The classifier token which is used when doing sequence classification (classification of the whole sequence instead of per-token classification). It is the first token of the sequence when built with special tokens.

  • mask_token (str, optional, defaults to "[MASK]") – The token used for masking values. This is the token used when training this model with masked language modeling. This is the token which the model will try to predict.

build_inputs_with_special_tokens(token_ids_0, token_ids_1=None)[source]¶

Build model inputs from a sequence or a pair of sequence for sequence classification tasks by concatenating and adding special tokens. A DeBERTa sequence has the following format:

  • single sequence: [CLS] X [SEP]

  • pair of sequences: [CLS] A [SEP] B [SEP]

Parameters
  • token_ids_0 (List[int]) – List of IDs to which the special tokens will be added.

  • token_ids_1 (List[int], optional) – Optional second list of IDs for sequence pairs.

Returns

List of input IDs with the appropriate special tokens.

Return type

List[int]

create_token_type_ids_from_sequences(token_ids_0, token_ids_1=None)[source]¶

Create a mask from the two sequences passed to be used in a sequence-pair classification task. A DeBERTa sequence pair mask has the following format:

0 0 0 0 0 0 0 0 0 0 0 1 1 1 1 1 1 1 1 1
| first sequence    | second sequence |

If token_ids_1 is None, this method only returns the first portion of the mask (0s).

Parameters
  • token_ids_0 (List[int]) – List of IDs.

  • token_ids_1 (List[int], optional) – Optional second list of IDs for sequence pairs.

Returns

List of token type IDs according to the given sequence(s).

Return type

List[int]

get_special_tokens_mask(token_ids_0, token_ids_1=None, already_has_special_tokens=False)[source]¶

Retrieves sequence ids from a token list that has no special tokens added. This method is called when adding special tokens using the tokenizer prepare_for_model or encode_plus methods.

Parameters
  • token_ids_0 (List[int]) – List of IDs.

  • token_ids_1 (List[int], optional) – Optional second list of IDs for sequence pairs.

  • already_has_special_tokens (bool, optional, defaults to False) – Whether or not the token list is already formatted with special tokens for the model.

Returns

A list of integers in the range [0, 1]: 1 for a special token, 0 for a sequence token.

Return type

List[int]

save_vocabulary(save_directory: str, filename_prefix: Optional[str] = None) → Tuple[str][source]¶

Save only the vocabulary of the tokenizer (vocabulary + added tokens).

This method won’t save the configuration and special token mappings of the tokenizer. Use _save_pretrained() to save the whole state of the tokenizer.

Parameters
  • save_directory (str) – The directory in which to save the vocabulary.

  • filename_prefix (str, optional) – An optional prefix to add to the named of the saved files.

Returns

Paths to the files saved.

Return type

Tuple(str)

DebertaModel¶

class transformers.DebertaModel(config)[source]¶

The bare DeBERTa Model transformer outputting raw hidden-states without any specific head on top. The DeBERTa model was proposed in DeBERTa: Decoding-enhanced BERT with Disentangled Attention by Pengcheng He, Xiaodong Liu, Jianfeng Gao, Weizhu Chen. It’s build on top of BERT/RoBERTa with two improvements, i.e. disentangled attention and enhanced mask decoder. With those two improvements, it out perform BERT/RoBERTa on a majority of tasks with 80GB pre-training data.

This model is also a PyTorch torch.nn.Module subclass. Use it as a regular PyTorch Module and refer to the PyTorch documentation for all matter related to general usage and behavior.```

Parameters

config (DebertaConfig) – Model configuration class with all the parameters of the model. Initializing with a config file does not load the weights associated with the model, only the configuration. Check out the from_pretrained() method to load the model weights.

forward(input_ids=None, attention_mask=None, token_type_ids=None, position_ids=None, inputs_embeds=None, output_attentions=None, output_hidden_states=None, return_dict=None)[source]¶

The DebertaModel forward method, overrides the __call__() special method.

Note

Although the recipe for forward pass needs to be defined within this function, one should call the Module instance afterwards instead of this since the former takes care of running the pre and post processing steps while the latter silently ignores them.

Parameters
  • input_ids (torch.LongTensor of shape batch_size, sequence_length) –

    Indices of input sequence tokens in the vocabulary.

    Indices can be obtained using transformers.DebertaTokenizer. See transformers.PreTrainedTokenizer.encode() and transformers.PreTrainedTokenizer.__call__() for details.

    What are input IDs?

  • attention_mask (torch.FloatTensor of shape batch_size, sequence_length, optional) –

    Mask to avoid performing attention on padding token indices. Mask values selected in [0, 1]:

    • 1 for tokens that are not masked,

    • 0 for tokens that are masked.

    What are attention masks?

  • token_type_ids (torch.LongTensor of shape batch_size, sequence_length, optional) –

    Segment token indices to indicate first and second portions of the inputs. Indices are selected in [0, 1]:

    • 0 corresponds to a sentence A token,

    • 1 corresponds to a sentence B token.

    What are token type IDs?

  • position_ids (torch.LongTensor of shape batch_size, sequence_length, optional) –

    Indices of positions of each input sequence tokens in the position embeddings. Selected in the range [0, config.max_position_embeddings - 1].

    What are position IDs?

  • inputs_embeds (torch.FloatTensor of shape (batch_size, sequence_length, hidden_size), optional) – Optionally, instead of passing input_ids you can choose to directly pass an embedded representation. This is useful if you want more control over how to convert input_ids indices into associated vectors than the model’s internal embedding lookup matrix.

  • output_attentions (bool, optional) – Whether or not to return the attentions tensors of all attention layers. See attentions under returned tensors for more detail.

  • output_hidden_states (bool, optional) – Whether or not to return the hidden states of all layers. See hidden_states under returned tensors for more detail.

  • return_dict (bool, optional) – Whether or not to return a ModelOutput instead of a plain tuple.

Returns

A SequenceClassifierOutput (if return_dict=True is passed or when config.return_dict=True) or a tuple of torch.FloatTensor comprising various elements depending on the configuration (DebertaConfig) and inputs.

  • loss (torch.FloatTensor of shape (1,), optional, returned when labels is provided) – Classification (or regression if config.num_labels==1) loss.

  • logits (torch.FloatTensor of shape (batch_size, config.num_labels)) – Classification (or regression if config.num_labels==1) scores (before SoftMax).

  • hidden_states (tuple(torch.FloatTensor), optional, returned when output_hidden_states=True is passed or when config.output_hidden_states=True) – Tuple of torch.FloatTensor (one for the output of the embeddings + one for the output of each layer) of shape (batch_size, sequence_length, hidden_size).

    Hidden-states of the model at the output of each layer plus the initial embedding outputs.

  • attentions (tuple(torch.FloatTensor), optional, returned when output_attentions=True is passed or when config.output_attentions=True) – Tuple of torch.FloatTensor (one for each layer) of shape (batch_size, num_heads, sequence_length, sequence_length).

    Attentions weights after the attention softmax, used to compute the weighted average in the self-attention heads.

Return type

SequenceClassifierOutput or tuple(torch.FloatTensor)

Example:

>>> from transformers import DebertaTokenizer, DebertaModel
>>> import torch

>>> tokenizer = DebertaTokenizer.from_pretrained('microsoft/deberta-base')
>>> model = DebertaModel.from_pretrained('microsoft/deberta-base', return_dict=True)

>>> inputs = tokenizer("Hello, my dog is cute", return_tensors="pt")
>>> outputs = model(**inputs)

>>> last_hidden_states = outputs.last_hidden_state
get_input_embeddings()[source]¶

Returns the model’s input embeddings.

Returns

A torch module mapping vocabulary to hidden states.

Return type

nn.Module

set_input_embeddings(new_embeddings)[source]¶

Set model’s input embeddings.

Parameters

value (nn.Module) – A module mapping vocabulary to hidden states.

DebertaPreTrainedModel¶

class transformers.DebertaPreTrainedModel(config: transformers.configuration_utils.PretrainedConfig, *inputs, **kwargs)[source]¶

An abstract class to handle weights initialization and a simple interface for downloading and loading pretrained models.

config_class¶

alias of transformers.configuration_deberta.DebertaConfig

DebertaForSequenceClassification¶

class transformers.DebertaForSequenceClassification(config)[source]¶

DeBERTa Model transformer with a sequence classification/regression head on top (a linear layer on top of the pooled output) e.g. for GLUE tasks.

The DeBERTa model was proposed in DeBERTa: Decoding-enhanced BERT with Disentangled Attention by Pengcheng He, Xiaodong Liu, Jianfeng Gao, Weizhu Chen. It’s build on top of BERT/RoBERTa with two improvements, i.e. disentangled attention and enhanced mask decoder. With those two improvements, it out perform BERT/RoBERTa on a majority of tasks with 80GB pre-training data.

This model is also a PyTorch torch.nn.Module subclass. Use it as a regular PyTorch Module and refer to the PyTorch documentation for all matter related to general usage and behavior.```

Parameters

config (DebertaConfig) – Model configuration class with all the parameters of the model. Initializing with a config file does not load the weights associated with the model, only the configuration. Check out the from_pretrained() method to load the model weights.

forward(input_ids=None, attention_mask=None, token_type_ids=None, position_ids=None, inputs_embeds=None, labels=None, output_attentions=None, output_hidden_states=None, return_dict=None)[source]¶

The DebertaForSequenceClassification forward method, overrides the __call__() special method.

Note

Although the recipe for forward pass needs to be defined within this function, one should call the Module instance afterwards instead of this since the former takes care of running the pre and post processing steps while the latter silently ignores them.

Parameters
  • input_ids (torch.LongTensor of shape batch_size, sequence_length) –

    Indices of input sequence tokens in the vocabulary.

    Indices can be obtained using transformers.DebertaTokenizer. See transformers.PreTrainedTokenizer.encode() and transformers.PreTrainedTokenizer.__call__() for details.

    What are input IDs?

  • attention_mask (torch.FloatTensor of shape batch_size, sequence_length, optional) –

    Mask to avoid performing attention on padding token indices. Mask values selected in [0, 1]:

    • 1 for tokens that are not masked,

    • 0 for tokens that are masked.

    What are attention masks?

  • token_type_ids (torch.LongTensor of shape batch_size, sequence_length, optional) –

    Segment token indices to indicate first and second portions of the inputs. Indices are selected in [0, 1]:

    • 0 corresponds to a sentence A token,

    • 1 corresponds to a sentence B token.

    What are token type IDs?

  • position_ids (torch.LongTensor of shape batch_size, sequence_length, optional) –

    Indices of positions of each input sequence tokens in the position embeddings. Selected in the range [0, config.max_position_embeddings - 1].

    What are position IDs?

  • inputs_embeds (torch.FloatTensor of shape (batch_size, sequence_length, hidden_size), optional) – Optionally, instead of passing input_ids you can choose to directly pass an embedded representation. This is useful if you want more control over how to convert input_ids indices into associated vectors than the model’s internal embedding lookup matrix.

  • output_attentions (bool, optional) – Whether or not to return the attentions tensors of all attention layers. See attentions under returned tensors for more detail.

  • output_hidden_states (bool, optional) – Whether or not to return the hidden states of all layers. See hidden_states under returned tensors for more detail.

  • return_dict (bool, optional) – Whether or not to return a ModelOutput instead of a plain tuple.

  • labels (torch.LongTensor of shape (batch_size,), optional) – Labels for computing the sequence classification/regression loss. Indices should be in [0, ..., config.num_labels - 1]. If config.num_labels == 1 a regression loss is computed (Mean-Square loss), If config.num_labels > 1 a classification loss is computed (Cross-Entropy).

Returns

A SequenceClassifierOutput (if return_dict=True is passed or when config.return_dict=True) or a tuple of torch.FloatTensor comprising various elements depending on the configuration (DebertaConfig) and inputs.

  • loss (torch.FloatTensor of shape (1,), optional, returned when labels is provided) – Classification (or regression if config.num_labels==1) loss.

  • logits (torch.FloatTensor of shape (batch_size, config.num_labels)) – Classification (or regression if config.num_labels==1) scores (before SoftMax).

  • hidden_states (tuple(torch.FloatTensor), optional, returned when output_hidden_states=True is passed or when config.output_hidden_states=True) – Tuple of torch.FloatTensor (one for the output of the embeddings + one for the output of each layer) of shape (batch_size, sequence_length, hidden_size).

    Hidden-states of the model at the output of each layer plus the initial embedding outputs.

  • attentions (tuple(torch.FloatTensor), optional, returned when output_attentions=True is passed or when config.output_attentions=True) – Tuple of torch.FloatTensor (one for each layer) of shape (batch_size, num_heads, sequence_length, sequence_length).

    Attentions weights after the attention softmax, used to compute the weighted average in the self-attention heads.

Return type

SequenceClassifierOutput or tuple(torch.FloatTensor)

Example:

>>> from transformers import DebertaTokenizer, DebertaForSequenceClassification
>>> import torch

>>> tokenizer = DebertaTokenizer.from_pretrained('microsoft/deberta-base')
>>> model = DebertaForSequenceClassification.from_pretrained('microsoft/deberta-base', return_dict=True)

>>> inputs = tokenizer("Hello, my dog is cute", return_tensors="pt")
>>> labels = torch.tensor([1]).unsqueeze(0)  # Batch size 1
>>> outputs = model(**inputs, labels=labels)
>>> loss = outputs.loss
>>> logits = outputs.logits
get_input_embeddings()[source]¶

Returns the model’s input embeddings.

Returns

A torch module mapping vocabulary to hidden states.

Return type

nn.Module

set_input_embeddings(new_embeddings)[source]¶

Set model’s input embeddings.

Parameters

value (nn.Module) – A module mapping vocabulary to hidden states.