The BertGeneration model is a BERT model that can be leveraged for sequence-to-sequence tasks using EncoderDecoderModel as proposed in Leveraging Pre-trained Checkpoints for Sequence Generation Tasks by Sascha Rothe, Shashi Narayan, Aliaksei Severyn.
The abstract from the paper is the following:
Unsupervised pretraining of large neural models has recently revolutionized Natural Language Processing. By warm-starting from the publicly released checkpoints, NLP practitioners have pushed the state-of-the-art on multiple benchmarks while saving significant amounts of compute time. So far the focus has been mainly on the Natural Language Understanding tasks. In this paper, we demonstrate the efficacy of pre-trained checkpoints for Sequence Generation. We developed a Transformer-based sequence-to-sequence model that is compatible with publicly available pre-trained BERT, GPT-2 and RoBERTa checkpoints and conducted an extensive empirical study on the utility of initializing our model, both encoder and decoder, with these checkpoints. Our models result in new state-of-the-art results on Machine Translation, Text Summarization, Sentence Splitting, and Sentence Fusion.
This model was contributed by patrickvonplaten. The original code can be found here.
The model can be used in combination with the EncoderDecoderModel to leverage two pretrained BERT checkpoints for subsequent fine-tuning:
>>> # leverage checkpoints for Bert2Bert model...
>>> # use BERT's cls token as BOS token and sep token as EOS token
>>> encoder = BertGenerationEncoder.from_pretrained("google-bert/bert-large-uncased", bos_token_id=101, eos_token_id=102)
>>> # add cross attention layers and use BERT's cls token as BOS token and sep token as EOS token
>>> decoder = BertGenerationDecoder.from_pretrained(
... "google-bert/bert-large-uncased", add_cross_attention=True, is_decoder=True, bos_token_id=101, eos_token_id=102
... )
>>> bert2bert = EncoderDecoderModel(encoder=encoder, decoder=decoder)
>>> # create tokenizer...
>>> tokenizer = BertTokenizer.from_pretrained("google-bert/bert-large-uncased")
>>> input_ids = tokenizer(
... "This is a long article to summarize", add_special_tokens=False, return_tensors="pt"
... ).input_ids
>>> labels = tokenizer("This is a short summary", return_tensors="pt").input_ids
>>> # train...
>>> loss = bert2bert(input_ids=input_ids, decoder_input_ids=labels, labels=labels).loss
>>> loss.backward()Pretrained EncoderDecoderModel are also directly available in the model hub, e.g.:
>>> # instantiate sentence fusion model
>>> sentence_fuser = EncoderDecoderModel.from_pretrained("google/roberta2roberta_L-24_discofuse")
>>> tokenizer = AutoTokenizer.from_pretrained("google/roberta2roberta_L-24_discofuse")
>>> input_ids = tokenizer(
... "This is the first sentence. This is the second sentence.", add_special_tokens=False, return_tensors="pt"
... ).input_ids
>>> outputs = sentence_fuser.generate(input_ids)
>>> print(tokenizer.decode(outputs[0]))Tips:
EncoderDecoder.( vocab_size = 50358 hidden_size = 1024 num_hidden_layers = 24 num_attention_heads = 16 intermediate_size = 4096 hidden_act = 'gelu' hidden_dropout_prob = 0.1 attention_probs_dropout_prob = 0.1 max_position_embeddings = 512 initializer_range = 0.02 layer_norm_eps = 1e-12 pad_token_id = 0 bos_token_id = 2 eos_token_id = 1 position_embedding_type = 'absolute' use_cache = True **kwargs )
Parameters
int, optional, defaults to 50358) —
Vocabulary size of the BERT model. Defines the number of different tokens that can be represented by the
inputs_ids passed when calling BertGeneration. int, optional, defaults to 1024) —
Dimensionality of the encoder layers and the pooler layer. int, optional, defaults to 24) —
Number of hidden layers in the Transformer encoder. int, optional, defaults to 16) —
Number of attention heads for each attention layer in the Transformer encoder. int, optional, defaults to 4096) —
Dimensionality of the “intermediate” (often called feed-forward) layer in the Transformer encoder. str or function, 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. float, optional, defaults to 0.1) —
The dropout probability for all fully connected layers in the embeddings, encoder, and pooler. float, optional, defaults to 0.1) —
The dropout ratio for the attention probabilities. 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). float, optional, defaults to 0.02) —
The standard deviation of the truncated_normal_initializer for initializing all weight matrices. float, optional, defaults to 1e-12) —
The epsilon used by the layer normalization layers. int, optional, defaults to 0) —
Padding token id. int, optional, defaults to 2) —
Beginning of stream token id. int, optional, defaults to 1) —
End of stream token id. 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.).
For more information on "relative_key_query", please refer to Method 4 in Improve Transformer Models
with Better Relative Position Embeddings (Huang et al.). 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. This is the configuration class to store the configuration of a BertGenerationPreTrainedModel. It is used to
instantiate a BertGeneration 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 BertGeneration
google/bert_for_seq_generation_L-24_bbc_encoder
architecture.
Configuration objects inherit from PretrainedConfig and can be used to control the model outputs. Read the documentation from PretrainedConfig for more information.
Examples:
>>> from transformers import BertGenerationConfig, BertGenerationEncoder
>>> # Initializing a BertGeneration config
>>> configuration = BertGenerationConfig()
>>> # Initializing a model (with random weights) from the config
>>> model = BertGenerationEncoder(configuration)
>>> # Accessing the model configuration
>>> configuration = model.config( vocab_file bos_token = '<s>' eos_token = '</s>' unk_token = '<unk>' pad_token = '<pad>' sep_token = '<::::>' sp_model_kwargs: Optional = None **kwargs )
Parameters
str) —
SentencePiece file (generally has a .spm extension) that
contains the vocabulary necessary to instantiate a tokenizer. str, optional, defaults to "<s>") —
The begin of sequence token. str, optional, defaults to "</s>") —
The end of sequence 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. str, optional, defaults to "<pad>") —
The token used for padding, for example when batching sequences of different lengths. str, optional, defaults to "< --:::>"):
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. dict, optional) —
Will be passed to the SentencePieceProcessor.__init__() method. The Python wrapper for
SentencePiece can be used, among other things,
to set:
enable_sampling: Enable subword regularization.
nbest_size: Sampling parameters for unigram. Invalid for BPE-Dropout.
nbest_size = {0,1}: No sampling is performed.nbest_size > 1: samples from the nbest_size results.nbest_size < 0: assuming that nbest_size is infinite and samples from the all hypothesis (lattice)
using forward-filtering-and-backward-sampling algorithm.alpha: Smoothing parameter for unigram sampling, and dropout probability of merge operations for
BPE-dropout.
Construct a BertGeneration tokenizer. Based on SentencePiece.
This tokenizer inherits from PreTrainedTokenizer which contains most of the main methods. Users should refer to this superclass for more information regarding those methods.
( config )
Parameters
The bare BertGeneration model transformer outputting raw hidden-states without any specific head on top.
This model inherits from PreTrainedModel. Check the superclass documentation for the generic methods the library implements for all its model (such as downloading or saving, resizing the input embeddings, pruning heads etc.)
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.
The model can behave as an encoder (with only self-attention) as well as a decoder, in which case a layer of cross-attention is added between the self-attention layers, following the architecture described in Attention is all you need by Ashish Vaswani, Noam Shazeer, Niki Parmar, Jakob Uszkoreit, Llion Jones, Aidan N. Gomez, Lukasz Kaiser and Illia Polosukhin.
This model should be used when leveraging Bert or Roberta checkpoints for the EncoderDecoderModel class as described in Leveraging Pre-trained Checkpoints for Sequence Generation Tasks by Sascha Rothe, Shashi Narayan, and Aliaksei Severyn.
To behave as an decoder the model needs to be initialized with the is_decoder argument of the configuration set
to True. To be used in a Seq2Seq model, the model needs to initialized with both is_decoder argument and
add_cross_attention set to True; an encoder_hidden_states is then expected as an input to the forward pass.
( input_ids: Optional = None attention_mask: Optional = None position_ids: Optional = None head_mask: Optional = None inputs_embeds: Optional = None encoder_hidden_states: Optional = None encoder_attention_mask: Optional = None past_key_values: Optional = None use_cache: Optional = None output_attentions: Optional = None output_hidden_states: Optional = None return_dict: Optional = None ) → transformers.modeling_outputs.BaseModelOutputWithPastAndCrossAttentions or tuple(torch.FloatTensor)
Parameters
torch.LongTensor of shape (batch_size, sequence_length)) —
Indices of input sequence tokens in the vocabulary.
Indices can be obtained using AutoTokenizer. See PreTrainedTokenizer.call() and PreTrainedTokenizer.encode() for details.
torch.FloatTensor of shape (batch_size, sequence_length), optional) —
Mask to avoid performing attention on padding token indices. Mask values selected in [0, 1]:
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].
torch.FloatTensor of shape (num_heads,) or (num_layers, num_heads), optional) —
Mask to nullify selected heads of the self-attention modules. Mask values selected in [0, 1]:
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. bool, optional) —
Whether or not to return the attentions tensors of all attention layers. See attentions under returned
tensors for more detail. bool, optional) —
Whether or not to return the hidden states of all layers. See hidden_states under returned tensors for
more detail. bool, optional) —
Whether or not to return a ModelOutput instead of a plain tuple. torch.FloatTensor of shape (batch_size, sequence_length, hidden_size), optional) —
Sequence of hidden-states at the output of the last layer of the encoder. Used in the cross-attention if
the model is configured as a decoder. torch.FloatTensor of shape (batch_size, sequence_length), optional) —
Mask to avoid performing attention on the padding token indices of the encoder input. This mask is used in
the cross-attention if the model is configured as a decoder. Mask values selected in [0, 1]: 1 for
tokens that are NOT MASKED, 0 for MASKED tokens. tuple(tuple(torch.FloatTensor)) of length config.n_layers with each tuple having 4 tensors of shape (batch_size, num_heads, sequence_length - 1, embed_size_per_head)) —
Contains precomputed key and value hidden states of the attention blocks. Can be used to speed up decoding.
If past_key_values are used, the user can optionally input only the last decoder_input_ids (those that
don’t have their past key value states given to this model) of shape (batch_size, 1) instead of all
decoder_input_ids of shape (batch_size, sequence_length).
bool, optional) —
If set to True, past_key_values key value states are returned and can be used to speed up decoding (see
past_key_values). Returns
transformers.modeling_outputs.BaseModelOutputWithPastAndCrossAttentions or tuple(torch.FloatTensor)
A transformers.modeling_outputs.BaseModelOutputWithPastAndCrossAttentions or a tuple of
torch.FloatTensor (if return_dict=False is passed or when config.return_dict=False) comprising various
elements depending on the configuration (BertGenerationConfig) and inputs.
last_hidden_state (torch.FloatTensor of shape (batch_size, sequence_length, hidden_size)) — Sequence of hidden-states at the output of the last layer of the model.
If past_key_values is used only the last hidden-state of the sequences of shape (batch_size, 1, hidden_size) is output.
past_key_values (tuple(tuple(torch.FloatTensor)), optional, returned when use_cache=True is passed or when config.use_cache=True) — Tuple of tuple(torch.FloatTensor) of length config.n_layers, with each tuple having 2 tensors of shape
(batch_size, num_heads, sequence_length, embed_size_per_head)) and optionally if
config.is_encoder_decoder=True 2 additional tensors of shape (batch_size, num_heads, encoder_sequence_length, embed_size_per_head).
Contains pre-computed hidden-states (key and values in the self-attention blocks and optionally if
config.is_encoder_decoder=True in the cross-attention blocks) that can be used (see past_key_values
input) to speed up sequential decoding.
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, if the model has an embedding layer, +
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 optional 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.
cross_attentions (tuple(torch.FloatTensor), optional, returned when output_attentions=True and config.add_cross_attention=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 of the decoder’s cross-attention layer, after the attention softmax, used to compute the weighted average in the cross-attention heads.
The BertGenerationEncoder forward method, overrides the __call__ special method.
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.
Example:
>>> from transformers import AutoTokenizer, BertGenerationEncoder
>>> import torch
>>> tokenizer = AutoTokenizer.from_pretrained("google/bert_for_seq_generation_L-24_bbc_encoder")
>>> model = BertGenerationEncoder.from_pretrained("google/bert_for_seq_generation_L-24_bbc_encoder")
>>> inputs = tokenizer("Hello, my dog is cute", return_tensors="pt")
>>> outputs = model(**inputs)
>>> last_hidden_states = outputs.last_hidden_state( config )
Parameters
BertGeneration Model with a language modeling head on top for CLM fine-tuning.
This model inherits from PreTrainedModel. Check the superclass documentation for the generic methods the library implements for all its model (such as downloading or saving, resizing the input embeddings, pruning heads etc.)
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.
( input_ids: Optional = None attention_mask: Optional = None position_ids: Optional = None head_mask: Optional = None inputs_embeds: Optional = None encoder_hidden_states: Optional = None encoder_attention_mask: Optional = None labels: Optional = None past_key_values: Optional = None use_cache: Optional = None output_attentions: Optional = None output_hidden_states: Optional = None return_dict: Optional = None ) → transformers.modeling_outputs.CausalLMOutputWithCrossAttentions or tuple(torch.FloatTensor)
Parameters
torch.LongTensor of shape (batch_size, sequence_length)) —
Indices of input sequence tokens in the vocabulary.
Indices can be obtained using AutoTokenizer. See PreTrainedTokenizer.call() and PreTrainedTokenizer.encode() for details.
torch.FloatTensor of shape (batch_size, sequence_length), optional) —
Mask to avoid performing attention on padding token indices. Mask values selected in [0, 1]:
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].
torch.FloatTensor of shape (num_heads,) or (num_layers, num_heads), optional) —
Mask to nullify selected heads of the self-attention modules. Mask values selected in [0, 1]:
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. bool, optional) —
Whether or not to return the attentions tensors of all attention layers. See attentions under returned
tensors for more detail. bool, optional) —
Whether or not to return the hidden states of all layers. See hidden_states under returned tensors for
more detail. bool, optional) —
Whether or not to return a ModelOutput instead of a plain tuple. torch.FloatTensor of shape (batch_size, sequence_length, hidden_size), optional) —
Sequence of hidden-states at the output of the last layer of the encoder. Used in the cross-attention if
the model is configured as a decoder. torch.FloatTensor of shape (batch_size, sequence_length), optional) —
Mask to avoid performing attention on the padding token indices of the encoder input. This mask is used in
the cross-attention if the model is configured as a decoder. Mask values selected in [0, 1]:
torch.LongTensor of shape (batch_size, sequence_length), optional) —
Labels for computing the left-to-right language modeling loss (next word prediction). Indices should be in
[-100, 0, ..., config.vocab_size] (see input_ids docstring) Tokens with indices set to -100 are
ignored (masked), the loss is only computed for the tokens with labels in [0, ..., config.vocab_size] tuple(tuple(torch.FloatTensor)) of length config.n_layers with each tuple having 4 tensors of shape (batch_size, num_heads, sequence_length - 1, embed_size_per_head)) —
Contains precomputed key and value hidden states of the attention blocks. Can be used to speed up decoding.
If past_key_values are used, the user can optionally input only the last decoder_input_ids (those that
don’t have their past key value states given to this model) of shape (batch_size, 1) instead of all
decoder_input_ids of shape (batch_size, sequence_length).
bool, optional) —
If set to True, past_key_values key value states are returned and can be used to speed up decoding (see
past_key_values). Returns
transformers.modeling_outputs.CausalLMOutputWithCrossAttentions or tuple(torch.FloatTensor)
A transformers.modeling_outputs.CausalLMOutputWithCrossAttentions or a tuple of
torch.FloatTensor (if return_dict=False is passed or when config.return_dict=False) comprising various
elements depending on the configuration (BertGenerationConfig) and inputs.
loss (torch.FloatTensor of shape (1,), optional, returned when labels is provided) — Language modeling loss (for next-token prediction).
logits (torch.FloatTensor of shape (batch_size, sequence_length, config.vocab_size)) — Prediction scores of the language modeling head (scores for each vocabulary token 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, if the model has an embedding layer, +
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 optional 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.
cross_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).
Cross attentions weights after the attention softmax, used to compute the weighted average in the cross-attention heads.
past_key_values (tuple(tuple(torch.FloatTensor)), optional, returned when use_cache=True is passed or when config.use_cache=True) — Tuple of torch.FloatTensor tuples of length config.n_layers, with each tuple containing the cached key,
value states of the self-attention and the cross-attention layers if model is used in encoder-decoder
setting. Only relevant if config.is_decoder = True.
Contains pre-computed hidden-states (key and values in the attention blocks) that can be used (see
past_key_values input) to speed up sequential decoding.
The BertGenerationDecoder forward method, overrides the __call__ special method.
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.
Example:
>>> from transformers import AutoTokenizer, BertGenerationDecoder, BertGenerationConfig
>>> import torch
>>> tokenizer = AutoTokenizer.from_pretrained("google/bert_for_seq_generation_L-24_bbc_encoder")
>>> config = BertGenerationConfig.from_pretrained("google/bert_for_seq_generation_L-24_bbc_encoder")
>>> config.is_decoder = True
>>> model = BertGenerationDecoder.from_pretrained(
... "google/bert_for_seq_generation_L-24_bbc_encoder", config=config
... )
>>> inputs = tokenizer("Hello, my dog is cute", return_token_type_ids=False, return_tensors="pt")
>>> outputs = model(**inputs)
>>> prediction_logits = outputs.logits