ProphetNet (original) (raw)

Overview

The ProphetNet model was proposed in ProphetNet: Predicting Future N-gram for Sequence-to-Sequence Pre-training, by Yu Yan, Weizhen Qi, Yeyun Gong, Dayiheng Liu, Nan Duan, Jiusheng Chen, Ruofei Zhang, Ming Zhou on 13 Jan, 2020.

ProphetNet is an encoder-decoder model and can predict n-future tokens for “ngram” language modeling instead of just the next token.

The abstract from the paper is the following:

In this paper, we present a new sequence-to-sequence pretraining model called ProphetNet, which introduces a novel self-supervised objective named future n-gram prediction and the proposed n-stream self-attention mechanism. Instead of the optimization of one-step ahead prediction in traditional sequence-to-sequence model, the ProphetNet is optimized by n-step ahead prediction which predicts the next n tokens simultaneously based on previous context tokens at each time step. The future n-gram prediction explicitly encourages the model to plan for the future tokens and prevent overfitting on strong local correlations. We pre-train ProphetNet using a base scale dataset (16GB) and a large scale dataset (160GB) respectively. Then we conduct experiments on CNN/DailyMail, Gigaword, and SQuAD 1.1 benchmarks for abstractive summarization and question generation tasks. Experimental results show that ProphetNet achieves new state-of-the-art results on all these datasets compared to the models using the same scale pretraining corpus.

The Authors’ code can be found here.

Usage tips

• ProphetNet is a model with absolute position embeddings so it’s usually advised to pad the inputs on the right rather than the left.
• The model architecture is based on the original Transformer, but replaces the “standard” self-attention mechanism in the decoder by a main self-attention mechanism and a self and n-stream (predict) self-attention mechanism.

Resources

• Causal language modeling task guide
• Translation task guide
• Summarization task guide

ProphetNetConfig

class transformers.ProphetNetConfig

< source >

( activation_dropout: typing.Optional[float] = 0.1 activation_function: typing.Union[str, typing.Callable, NoneType] = 'gelu' vocab_size: typing.Optional[int] = 30522 hidden_size: typing.Optional[int] = 1024 encoder_ffn_dim: typing.Optional[int] = 4096 num_encoder_layers: typing.Optional[int] = 12 num_encoder_attention_heads: typing.Optional[int] = 16 decoder_ffn_dim: typing.Optional[int] = 4096 num_decoder_layers: typing.Optional[int] = 12 num_decoder_attention_heads: typing.Optional[int] = 16 attention_dropout: typing.Optional[float] = 0.1 dropout: typing.Optional[float] = 0.1 max_position_embeddings: typing.Optional[int] = 512 init_std: typing.Optional[float] = 0.02 is_encoder_decoder: typing.Optional[bool] = True add_cross_attention: typing.Optional[bool] = True decoder_start_token_id: typing.Optional[int] = 0 ngram: typing.Optional[int] = 2 num_buckets: typing.Optional[int] = 32 relative_max_distance: typing.Optional[int] = 128 disable_ngram_loss: typing.Optional[bool] = False eps: typing.Optional[float] = 0.0 use_cache: typing.Optional[bool] = True pad_token_id: typing.Optional[int] = 0 bos_token_id: typing.Optional[int] = 1 eos_token_id: typing.Optional[int] = 2 **kwargs )

Parameters

• activation_dropout (float, optional, defaults to 0.1) — The dropout ratio for activations inside the fully connected layer.
• activation_function (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.
• vocab_size (int, optional, defaults to 30522) — Vocabulary size of the ProphetNET model. Defines the number of different tokens that can be represented by the inputs_ids passed when calling ProphetNetModel.
• hidden_size (int, optional, defaults to 1024) — Dimensionality of the layers and the pooler layer.
• encoder_ffn_dim (int, optional, defaults to 4096) — Dimensionality of the “intermediate” (often named feed-forward) layer in decoder.
• num_encoder_layers (int, optional, defaults to 12) — Number of encoder layers.
• num_encoder_attention_heads (int, optional, defaults to 16) — Number of attention heads for each attention layer in the Transformer encoder.
• decoder_ffn_dim (int, optional, defaults to 4096) — Dimensionality of the intermediate (often named feed-forward) layer in decoder.
• num_decoder_layers (int, optional, defaults to 12) — Number of decoder layers.
• num_decoder_attention_heads (int, optional, defaults to 16) — Number of attention heads for each attention layer in the Transformer decoder.
• attention_dropout (float, optional, defaults to 0.1) — The dropout ratio for the attention probabilities.
• dropout (float, optional, defaults to 0.1) — The dropout probability for all fully connected layers in the embeddings, encoder, and pooler.
• 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).
• init_std (float, optional, defaults to 0.02) — The standard deviation of the truncated_normal_initializer for initializing all weight matrices.
• add_cross_attention (bool, optional, defaults to True) — Whether cross-attention layers should be added to the model.
• is_encoder_decoder (bool, optional, defaults to True) — Whether this is an encoder/decoder model.
• pad_token_id (int, optional, defaults to 1) — Padding token id.
• bos_token_id (int, optional, defaults to 0) — Beginning of stream token id.
• eos_token_id (int, optional, defaults to 2) — End of stream token id.
• ngram (int, optional, defaults to 2) — Number of future tokens to predict. Set to 1 to be same as traditional Language model to predict next first token.
• num_buckets (int, optional, defaults to 32) — The number of buckets to use for each attention layer. This is for relative position calculation. See the [T5 paper](see https://arxiv.org/abs/1910.10683) for more details.
• relative_max_distance (int, optional, defaults to 128) — Relative distances greater than this number will be put into the last same bucket. This is for relative position calculation. See the [T5 paper](see https://arxiv.org/abs/1910.10683) for more details.
• disable_ngram_loss (bool, optional, defaults to False) — Whether be trained predicting only the next first token.
• eps (float, optional, defaults to 0.0) — Controls the epsilon parameter value for label smoothing in the loss calculation. If set to 0, no label smoothing is performed.
• use_cache (bool, optional, defaults to True) — Whether or not the model should return the last key/values attentions (not used by all models).

This is the configuration class to store the configuration of a ProphetNetModel. It is used to instantiate a ProphetNet 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 ProphetNetmicrosoft/prophetnet-large-uncased architecture.

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

ProphetNetTokenizer

class transformers.ProphetNetTokenizer

< source >

( vocab_file: str do_lower_case: typing.Optional[bool] = True do_basic_tokenize: typing.Optional[bool] = True never_split: typing.Optional[typing.Iterable] = None unk_token: typing.Optional[str] = '[UNK]' sep_token: typing.Optional[str] = '[SEP]' x_sep_token: typing.Optional[str] = '[X_SEP]' pad_token: typing.Optional[str] = '[PAD]' mask_token: typing.Optional[str] = '[MASK]' tokenize_chinese_chars: typing.Optional[bool] = True strip_accents: typing.Optional[bool] = None clean_up_tokenization_spaces: bool = True **kwargs )

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.
• do_basic_tokenize (bool, optional, defaults to True) — Whether or not to do basic tokenization before WordPiece.
• never_split (Iterable, optional) — Collection of tokens which will never be split during tokenization. Only has an effect whendo_basic_tokenize=True
• 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.
• x_sep_token (str, optional, defaults to "[X_SEP]") — Special second separator token, which can be generated by ProphetNetForConditionalGeneration. It is used to separate bullet-point like sentences in summarization, e.g..
• pad_token (str, optional, defaults to "[PAD]") — The token used for padding, for example when batching sequences of different lengths.
• 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.
• tokenize_chinese_chars (bool, optional, defaults to True) — Whether or not to tokenize Chinese characters.
  This should likely be deactivated for Japanese (see thisissue).
• strip_accents (bool, optional) — Whether or not to strip all accents. If this option is not specified, then it will be determined by the value for lowercase (as in the original BERT).
• clean_up_tokenization_spaces (bool, optional, defaults to True) — Whether or not to cleanup spaces after decoding, cleanup consists in removing potential artifacts like extra spaces.

Construct a ProphetNetTokenizer. Based on WordPiece.

This tokenizer inherits from PreTrainedTokenizer which contains most of the main methods. Users should refer to this superclass for more information regarding those methods.

build_inputs_with_special_tokens

< source >

( token_ids_0: typing.List[int] token_ids_1: typing.Optional[typing.List[int]] = None ) → List[int]

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.

List of input IDs with the appropriate special tokens.

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

• single sequence: [CLS] X [SEP]
• pair of sequences: [CLS] A [SEP] B [SEP]

convert_tokens_to_string

< source >

( tokens: str )

Converts a sequence of tokens (string) in a single string.

create_token_type_ids_from_sequences

< source >

( token_ids_0: typing.List[int] token_ids_1: typing.Optional[typing.List[int]] = None ) → List[int]

Parameters

• token_ids_0 (List[int]) — List of IDs.
• token_ids_1 (List[int], optional) — Optional second list of IDs for sequence pairs.

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

Create a mask from the two sequences passed to be used in a sequence-pair classification task. A ProphetNet

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).

get_special_tokens_mask

< source >

( token_ids_0: typing.List[int] token_ids_1: typing.Optional[typing.List[int]] = None already_has_special_tokens: typing.Optional[bool] = False ) → List[int]

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.

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

Retrieve 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 method.

ProphetNet specific outputs

class transformers.models.prophetnet.modeling_prophetnet.ProphetNetSeq2SeqLMOutput

< source >

( loss: typing.Optional[torch.FloatTensor] = None logits: typing.Optional[torch.FloatTensor] = None logits_ngram: typing.Optional[torch.FloatTensor] = None past_key_values: typing.Optional[typing.Tuple[torch.FloatTensor]] = None decoder_hidden_states: typing.Optional[typing.Tuple[torch.FloatTensor]] = None decoder_ngram_hidden_states: typing.Optional[typing.Tuple[torch.FloatTensor]] = None decoder_attentions: typing.Optional[typing.Tuple[torch.FloatTensor]] = None decoder_ngram_attentions: typing.Optional[typing.Tuple[torch.FloatTensor]] = None cross_attentions: typing.Optional[typing.Tuple[torch.FloatTensor]] = None encoder_last_hidden_state: typing.Optional[torch.FloatTensor] = None encoder_hidden_states: typing.Optional[typing.Tuple[torch.FloatTensor]] = None encoder_attentions: typing.Optional[typing.Tuple[torch.FloatTensor]] = None )

Parameters

• loss (torch.FloatTensor of shape (1,), optional, returned when labels is provided) — Language modeling loss.
• logits (torch.FloatTensor of shape (batch_size, decoder_sequence_length, config.vocab_size)) — Prediction scores of the main stream language modeling head (scores for each vocabulary token before SoftMax).
• logits_ngram (torch.FloatTensor of shape (batch_size, ngram * decoder_sequence_length, config.vocab_size)) — Prediction scores of the predict stream language modeling head (scores for each vocabulary token before SoftMax).
• past_key_values (List[torch.FloatTensor], optional, returned when use_cache=True is passed or when config.use_cache=True) — List of torch.FloatTensor of length config.n_layers, with each tensor of shape (2, batch_size, num_attn_heads, decoder_sequence_length, embed_size_per_head)).
  Contains pre-computed hidden-states (key and values in the attention blocks) of the decoder that can be used (see past_key_values input) to speed up sequential decoding.
• decoder_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, decoder_sequence_length, hidden_size).
  Hidden-states of main stream of the decoder at the output of each layer plus the initial embedding outputs.
• decoder_ngram_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, ngram * decoder_sequence_length, hidden_size).
  Hidden-states of the predict stream of the decoder at the output of each layer plus the initial embedding outputs.
• decoder_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_attn_heads, decoder_sequence_length, decoder_sequence_length).
  Attentions weights of the decoder, after the attention softmax, used to compute the weighted average in the self-attention heads.
• decoder_ngram_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_attn_heads, decoder_sequence_length, decoder_sequence_length).
  Attentions weights of the predict stream of the decoder, 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_attn_heads, encoder_sequence_length, decoder_sequence_length).
  Attentions weights of the cross-attention layer of the decoder, after the attention softmax, used to compute the weighted average in the
• encoder_last_hidden_state (torch.FloatTensor of shape (batch_size, encoder_sequence_length, hidden_size), optional) — Sequence of hidden-states at the output of the last layer of the encoder of the model.
• encoder_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, encoder_sequence_length, hidden_size).
  Hidden-states of the encoder at the output of each layer plus the initial embedding outputs.
• encoder_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_attn_heads, encoder_sequence_length, encoder_sequence_length). Attentions weights of the encoder, after the attention softmax, used to compute the weighted average in the self-attention heads.

Base class for sequence-to-sequence language models outputs.

class transformers.models.prophetnet.modeling_prophetnet.ProphetNetSeq2SeqModelOutput

< source >

( last_hidden_state: FloatTensor last_hidden_state_ngram: typing.Optional[torch.FloatTensor] = None past_key_values: typing.Optional[typing.Tuple[torch.FloatTensor]] = None decoder_hidden_states: typing.Optional[typing.Tuple[torch.FloatTensor]] = None decoder_ngram_hidden_states: typing.Optional[typing.Tuple[torch.FloatTensor]] = None decoder_attentions: typing.Optional[typing.Tuple[torch.FloatTensor]] = None decoder_ngram_attentions: typing.Optional[typing.Tuple[torch.FloatTensor]] = None cross_attentions: typing.Optional[typing.Tuple[torch.FloatTensor]] = None encoder_last_hidden_state: typing.Optional[torch.FloatTensor] = None encoder_hidden_states: typing.Optional[typing.Tuple[torch.FloatTensor]] = None encoder_attentions: typing.Optional[typing.Tuple[torch.FloatTensor]] = None )

Parameters

• last_hidden_state (torch.FloatTensor of shape (batch_size, decoder_sequence_length, hidden_size)) — Sequence of main stream hidden-states at the output of the last layer of the decoder 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.
• last_hidden_state_ngram (torch.FloatTensor of shape (batch_size,ngram * decoder_sequence_length, config.vocab_size), optional) — Sequence of predict stream hidden-states at the output of the last layer of the decoder of the model.
• past_key_values (List[torch.FloatTensor], optional, returned when use_cache=True is passed or when config.use_cache=True) — List of torch.FloatTensor of length config.n_layers, with each tensor of shape (2, batch_size, num_attn_heads, decoder_sequence_length, embed_size_per_head)).
  Contains pre-computed hidden-states (key and values in the attention blocks) of the decoder that can be used (see past_key_values input) to speed up sequential decoding.
• decoder_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, decoder_sequence_length, hidden_size).
  Hidden-states of main stream of the decoder at the output of each layer plus the initial embedding outputs.
• decoder_ngram_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, ngram * decoder_sequence_length, hidden_size).
  Hidden-states of the predict stream of the decoder at the output of each layer plus the initial embedding outputs.
• decoder_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_attn_heads, decoder_sequence_length, decoder_sequence_length).
  Attentions weights of the decoder, after the attention softmax, used to compute the weighted average in the self-attention heads.
• decoder_ngram_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_attn_heads, decoder_sequence_length, decoder_sequence_length).
  Attentions weights of the predict stream of the decoder, after the attention softmax, used to compute the weighted average in the
• 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_attn_heads, encoder_sequence_length, decoder_sequence_length).
  Attentions weights of the cross-attention layer of the decoder, after the attention softmax, used to compute the weighted average in the
• encoder_last_hidden_state (torch.FloatTensor of shape (batch_size, encoder_sequence_length, hidden_size), optional) — Sequence of hidden-states at the output of the last layer of the encoder of the model.
• encoder_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, encoder_sequence_length, hidden_size).
  Hidden-states of the encoder at the output of each layer plus the initial embedding outputs.
• encoder_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_attn_heads, encoder_sequence_length, encoder_sequence_length).
  Attentions weights of the encoder, after the attention softmax, used to compute the weighted average in the self-attention heads.

Base class for model encoder’s outputs that also contains : pre-computed hidden states that can speed up sequential decoding.

class transformers.models.prophetnet.modeling_prophetnet.ProphetNetDecoderModelOutput

< source >

( last_hidden_state: FloatTensor last_hidden_state_ngram: typing.Optional[torch.FloatTensor] = None past_key_values: typing.Optional[typing.Tuple[torch.FloatTensor]] = None hidden_states: typing.Optional[typing.Tuple[torch.FloatTensor]] = None hidden_states_ngram: typing.Optional[typing.Tuple[torch.FloatTensor]] = None attentions: typing.Optional[typing.Tuple[torch.FloatTensor]] = None ngram_attentions: typing.Optional[typing.Tuple[torch.FloatTensor]] = None cross_attentions: typing.Optional[typing.Tuple[torch.FloatTensor]] = None )

Parameters

• last_hidden_state (torch.FloatTensor of shape (batch_size, decoder_sequence_length, hidden_size)) — Sequence of main stream hidden-states at the output of the last layer of the decoder 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.
• last_hidden_state_ngram (torch.FloatTensor of shape (batch_size, ngram * decoder_sequence_length, config.vocab_size)) — Sequence of predict stream hidden-states at the output of the last layer of the decoder of the model.
• past_key_values (List[torch.FloatTensor], optional, returned when use_cache=True is passed or when config.use_cache=True) — List of torch.FloatTensor of length config.n_layers, with each tensor of shape (2, batch_size, num_attn_heads, decoder_sequence_length, embed_size_per_head)).
  Contains pre-computed hidden-states (key and values in the attention blocks) of the decoder 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 + one for the output of each layer) of shape (batch_size, decoder_sequence_length, hidden_size).
  Hidden-states of main stream of the decoder at the output of each layer plus the initial embedding outputs.
• ngram_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, ngram * decoder_sequence_length, hidden_size).
  Hidden-states of the predict stream of the decoder 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_attn_heads, decoder_sequence_length, decoder_sequence_length).
  Attentions weights of the decoder, after the attention softmax, used to compute the weighted average in the self-attention heads.
• ngram_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_attn_heads, decoder_sequence_length, decoder_sequence_length).
  Attentions weights of the predict stream of the decoder, after the attention softmax, used to compute the weighted average in the
• 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_attn_heads, encoder_sequence_length, decoder_sequence_length).
  Attentions weights of the cross-attention layer of the decoder, after the attention softmax, used to compute the weighted average in the

Base class for model’s outputs that may also contain a past key/values (to speed up sequential decoding).

class transformers.models.prophetnet.modeling_prophetnet.ProphetNetDecoderLMOutput

< source >

( loss: typing.Optional[torch.FloatTensor] = None logits: typing.Optional[torch.FloatTensor] = None logits_ngram: typing.Optional[torch.FloatTensor] = None past_key_values: typing.Optional[typing.Tuple[torch.FloatTensor]] = None hidden_states: typing.Optional[typing.Tuple[torch.FloatTensor]] = None hidden_states_ngram: typing.Optional[typing.Tuple[torch.FloatTensor]] = None attentions: typing.Optional[typing.Tuple[torch.FloatTensor]] = None ngram_attentions: typing.Optional[typing.Tuple[torch.FloatTensor]] = None cross_attentions: typing.Optional[typing.Tuple[torch.FloatTensor]] = None )

Parameters

• loss (torch.FloatTensor of shape (1,), optional, returned when labels is provided) — Language modeling loss.
• logits (torch.FloatTensor of shape (batch_size, decoder_sequence_length, config.vocab_size)) — Prediction scores of the main stream language modeling head (scores for each vocabulary token before SoftMax).
• logits_ngram (torch.FloatTensor of shape (batch_size, ngram * decoder_sequence_length, config.vocab_size)) — Prediction scores of the predict stream language modeling head (scores for each vocabulary token before SoftMax).
• past_key_values (List[torch.FloatTensor], optional, returned when use_cache=True is passed or when config.use_cache=True) — List of torch.FloatTensor of length config.n_layers, with each tensor of shape (2, batch_size, num_attn_heads, decoder_sequence_length, embed_size_per_head)).
  Contains pre-computed hidden-states (key and values in the attention blocks) of the decoder 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 + one for the output of each layer) of shape (batch_size, decoder_sequence_length, hidden_size).
  Hidden-states of main stream of the decoder at the output of each layer plus the initial embedding outputs.
• ngram_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, ngram * decoder_sequence_length, hidden_size).
  Hidden-states of the predict stream of the decoder 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_attn_heads, decoder_sequence_length, decoder_sequence_length).
  Attentions weights of the decoder, after the attention softmax, used to compute the weighted average in the self-attention heads.
• ngram_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_attn_heads, decoder_sequence_length, decoder_sequence_length).
  Attentions weights of the predict stream of the decoder, after the attention softmax, used to compute the weighted average in the
• 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_attn_heads, encoder_sequence_length, decoder_sequence_length).
  Attentions weights of the cross-attention layer of the decoder, after the attention softmax, used to compute the weighted average in the

Base class for model’s outputs that may also contain a past key/values (to speed up sequential decoding).

ProphetNetModel

class transformers.ProphetNetModel

< source >

( config: ProphetNetConfig )

Parameters

• config (ProphetNetConfig) — 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 thefrom_pretrained() method to load the model weights.

The bare Prophetnet Model 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.

forward

< source >

( input_ids: typing.Optional[torch.Tensor] = None attention_mask: typing.Optional[torch.Tensor] = None decoder_input_ids: typing.Optional[torch.Tensor] = None decoder_attention_mask: typing.Optional[torch.BoolTensor] = None head_mask: typing.Optional[torch.Tensor] = None decoder_head_mask: typing.Optional[torch.Tensor] = None cross_attn_head_mask: typing.Optional[torch.Tensor] = None encoder_outputs: typing.Optional[typing.Tuple] = None past_key_values: typing.Optional[typing.Tuple[typing.Tuple[torch.Tensor]]] = None inputs_embeds: typing.Optional[torch.Tensor] = None decoder_inputs_embeds: typing.Optional[torch.Tensor] = None use_cache: typing.Optional[bool] = None output_attentions: typing.Optional[bool] = None output_hidden_states: typing.Optional[bool] = None return_dict: typing.Optional[bool] = None ) → transformers.models.prophetnet.modeling_prophetnet.ProphetNetSeq2SeqModelOutput or tuple(torch.FloatTensor)

Parameters

• input_ids (torch.Tensor of shape (batch_size, sequence_length), optional) — Indices of input sequence tokens in the vocabulary. Padding will be ignored by default.
  Indices can be obtained using AutoTokenizer. See PreTrainedTokenizer.encode() andPreTrainedTokenizer.call() for details.
  What are input IDs?
• attention_mask (torch.Tensor 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?
• decoder_input_ids (torch.Tensor of shape (batch_size, target_sequence_length), optional) — Indices of decoder input sequence tokens in the vocabulary.
  Indices can be obtained using AutoTokenizer. See PreTrainedTokenizer.encode() andPreTrainedTokenizer.call() for details.
  What are decoder input IDs?
  ProphetNet uses the eos_token_id as the starting token for decoder_input_ids generation. Ifpast_key_values is used, optionally only the last decoder_input_ids have to be input (seepast_key_values).
• decoder_attention_mask (torch.BoolTensor of shape (batch_size, target_sequence_length), optional) — Default behavior: generate a tensor that ignores pad tokens in decoder_input_ids. Causal mask will also be used by default.
• head_mask (torch.Tensor 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]:
  • 1 indicates the head is not masked,
  • 0 indicates the head is masked.
• decoder_head_mask (torch.Tensor of shape (decoder_layers, decoder_attention_heads), optional) — Mask to nullify selected heads of the attention modules in the decoder. Mask values selected in [0, 1]:
  • 1 indicates the head is not masked,
  • 0 indicates the head is masked.
• cross_attn_head_mask (torch.Tensor of shape (decoder_layers, decoder_attention_heads), optional) — Mask to nullify selected heads of the cross-attention modules. Mask values selected in [0, 1]:
  • 1 indicates the head is not masked,
  • 0 indicates the head is masked.
• encoder_outputs (Tuple, optional) — Tuple consists of (last_hidden_state, optional: hidden_states, optional: attentions)last_hidden_state of shape (batch_size, sequence_length, hidden_size), optional) is a sequence of hidden-states at the output of the last layer of the encoder. Used in the cross-attention of the decoder.
• past_key_values (Tuple[Tuple[torch.Tensor]], optional) — Pre-computed hidden-states (key and values in the self-attention blocks and in the cross-attention blocks) that can be used to speed up sequential decoding. This typically consists in the past_key_valuesreturned by the model at a previous stage of decoding, when use_cache=True or config.use_cache=True.
  Two formats are allowed:
  • a Cache instance, see our kv cache guide;
  • 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)). This is also known as the legacy cache format.
    The model will output the same cache format that is fed as input. If no past_key_values are passed, the legacy cache format will be returned.
    If past_key_values are used, the user can optionally input only the last input_ids (those that don’t have their past key value states given to this model) of shape (batch_size, 1) instead of all input_idsof shape (batch_size, sequence_length).
• inputs_embeds (torch.Tensor 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.
• decoder_inputs_embeds (torch.Tensor of shape (batch_size, target_sequence_length, hidden_size), optional) — Optionally, instead of passing decoder_input_ids you can choose to directly pass an embedded representation. If past_key_values is used, optionally only the last decoder_inputs_embeds have to be input (see past_key_values). This is useful if you want more control over how to convertdecoder_input_ids indices into associated vectors than the model’s internal embedding lookup matrix.
  If decoder_input_ids and decoder_inputs_embeds are both unset, decoder_inputs_embeds takes the value of inputs_embeds.
• use_cache (bool, optional) — If set to True, past_key_values key value states are returned and can be used to speed up decoding (seepast_key_values).
• 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.

A transformers.models.prophetnet.modeling_prophetnet.ProphetNetSeq2SeqModelOutput or a tuple oftorch.FloatTensor (if return_dict=False is passed or when config.return_dict=False) comprising various elements depending on the configuration (ProphetNetConfig) and inputs.

• last_hidden_state (torch.FloatTensor of shape (batch_size, decoder_sequence_length, hidden_size)) — Sequence of main stream hidden-states at the output of the last layer of the decoder 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.
• last_hidden_state_ngram (torch.FloatTensor of shape (batch_size,ngram * decoder_sequence_length, config.vocab_size), optional) — Sequence of predict stream hidden-states at the output of the last layer of the decoder of the model.
• past_key_values (List[torch.FloatTensor], optional, returned when use_cache=True is passed or when config.use_cache=True) — List of torch.FloatTensor of length config.n_layers, with each tensor of shape (2, batch_size, num_attn_heads, decoder_sequence_length, embed_size_per_head)).
  Contains pre-computed hidden-states (key and values in the attention blocks) of the decoder that can be used (see past_key_values input) to speed up sequential decoding.
• decoder_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, decoder_sequence_length, hidden_size).
  Hidden-states of main stream of the decoder at the output of each layer plus the initial embedding outputs.
• decoder_ngram_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, ngram * decoder_sequence_length, hidden_size).
  Hidden-states of the predict stream of the decoder at the output of each layer plus the initial embedding outputs.
• decoder_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_attn_heads, decoder_sequence_length, decoder_sequence_length).
  Attentions weights of the decoder, after the attention softmax, used to compute the weighted average in the self-attention heads.
• decoder_ngram_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_attn_heads, decoder_sequence_length, decoder_sequence_length).
  Attentions weights of the predict stream of the decoder, after the attention softmax, used to compute the weighted average in the
• 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_attn_heads, encoder_sequence_length, decoder_sequence_length).
  Attentions weights of the cross-attention layer of the decoder, after the attention softmax, used to compute the weighted average in the
• encoder_last_hidden_state (torch.FloatTensor of shape (batch_size, encoder_sequence_length, hidden_size), optional) — Sequence of hidden-states at the output of the last layer of the encoder of the model.
• encoder_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, encoder_sequence_length, hidden_size).
  Hidden-states of the encoder at the output of each layer plus the initial embedding outputs.
• encoder_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_attn_heads, encoder_sequence_length, encoder_sequence_length).
  Attentions weights of the encoder, after the attention softmax, used to compute the weighted average in the self-attention heads.

The ProphetNetModel forward method, overrides the __call__ special method.

Although the recipe for forward pass needs to be defined within this function, one should call the Moduleinstance 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, ProphetNetModel

tokenizer = AutoTokenizer.from_pretrained("microsoft/prophetnet-large-uncased") model = ProphetNetModel.from_pretrained("microsoft/prophetnet-large-uncased")

input_ids = tokenizer( ... "Studies have been shown that owning a dog is good for you", return_tensors="pt" ... ).input_ids
decoder_input_ids = tokenizer("Studies show that", return_tensors="pt").input_ids
outputs = model(input_ids=input_ids, decoder_input_ids=decoder_input_ids)

last_hidden_states = outputs.last_hidden_state
last_hidden_states_ngram = outputs.last_hidden_state_ngram

ProphetNetEncoder

class transformers.ProphetNetEncoder

< source >

( config: ProphetNetConfig word_embeddings: Embedding = None )

Parameters

• config (ProphetNetConfig) — 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 thefrom_pretrained() method to load the model weights.
• word_embeddings (torch.nn.modules.sparse.Embedding of shape (config.vocab_size, config.hidden_size), optional) — The word embedding parameters. This can be used to initialize ProphetNetEncoder with pre-defined word embeddings instead of randomly initialized word embeddings.

The standalone encoder part of the ProphetNetModel.

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.

forward

< source >

( input_ids: typing.Optional[torch.Tensor] = None attention_mask: typing.Optional[torch.Tensor] = None head_mask: typing.Optional[torch.Tensor] = None inputs_embeds: typing.Optional[torch.Tensor] = None output_attentions: typing.Optional[bool] = None output_hidden_states: typing.Optional[bool] = None return_dict: typing.Optional[bool] = None ) → transformers.modeling_outputs.BaseModelOutput or tuple(torch.FloatTensor)

Parameters

• input_ids (torch.Tensor of shape (batch_size, sequence_length), optional) — Indices of input sequence tokens in the vocabulary. Padding will be ignored by default.
  Indices can be obtained using AutoTokenizer. See PreTrainedTokenizer.encode() andPreTrainedTokenizer.call() for details.
  What are input IDs?
• attention_mask (torch.Tensor 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?
• head_mask (torch.Tensor 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]:
  • 1 indicates the head is not masked,
  • 0 indicates the head is masked.
• inputs_embeds (torch.Tensor 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.

A transformers.modeling_outputs.BaseModelOutput or a tuple oftorch.FloatTensor (if return_dict=False is passed or when config.return_dict=False) comprising various elements depending on the configuration (ProphetNetConfig) 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.
• 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.

The ProphetNetEncoder forward method, overrides the __call__ special method.

Although the recipe for forward pass needs to be defined within this function, one should call the Moduleinstance 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, ProphetNetEncoder import torch

tokenizer = AutoTokenizer.from_pretrained("microsoft/prophetnet-large-uncased") model = ProphetNetEncoder.from_pretrained("patrickvonplaten/prophetnet-large-uncased-standalone") inputs = tokenizer("Hello, my dog is cute", return_tensors="pt") outputs = model(**inputs)

last_hidden_states = outputs.last_hidden_state

ProphetNetDecoder

class transformers.ProphetNetDecoder

< source >

( config: ProphetNetConfig word_embeddings: typing.Optional[torch.nn.modules.sparse.Embedding] = None )

Parameters

• config (ProphetNetConfig) — 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 thefrom_pretrained() method to load the model weights.
• word_embeddings (torch.nn.modules.sparse.Embedding of shape (config.vocab_size, config.hidden_size), optional) — The word embedding parameters. This can be used to initialize ProphetNetEncoder with pre-defined word embeddings instead of randomly initialized word embeddings.

The standalone decoder part of the ProphetNetModel.

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.

forward

< source >

( input_ids: typing.Optional[torch.Tensor] = None attention_mask: typing.Optional[torch.Tensor] = None encoder_hidden_states: typing.Optional[torch.Tensor] = None encoder_attention_mask: typing.Optional[torch.Tensor] = None head_mask: typing.Optional[torch.Tensor] = None cross_attn_head_mask: typing.Optional[torch.Tensor] = None past_key_values: typing.Optional[typing.Tuple[typing.Tuple[torch.Tensor]]] = None inputs_embeds: typing.Optional[torch.Tensor] = None use_cache: typing.Optional[bool] = None output_attentions: typing.Optional[bool] = None output_hidden_states: typing.Optional[bool] = None return_dict: typing.Optional[bool] = None ) → transformers.models.prophetnet.modeling_prophetnet.ProphetNetDecoderModelOutput or tuple(torch.FloatTensor)

Parameters

• input_ids (torch.Tensor of shape (batch_size, sequence_length), optional) — Indices of input sequence tokens in the vocabulary. Padding will be ignored by default.
  Indices can be obtained using AutoTokenizer. See PreTrainedTokenizer.encode() andPreTrainedTokenizer.call() for details.
  What are input IDs?
• attention_mask (torch.Tensor 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?
• encoder_hidden_states (torch.Tensor 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.
• encoder_attention_mask (torch.Tensor 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 tokens that are masked.
• head_mask (torch.Tensor 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]:
  • 1 indicates the head is not masked,
  • 0 indicates the head is masked.
• cross_attn_head_mask (torch.Tensor of shape (decoder_layers, decoder_attention_heads), optional) — Mask to nullify selected heads of the cross-attention modules. Mask values selected in [0, 1]:
  • 1 indicates the head is not masked,
  • 0 indicates the head is masked.
• past_key_values (Tuple[Tuple[torch.Tensor]], optional) — Pre-computed hidden-states (key and values in the self-attention blocks and in the cross-attention blocks) that can be used to speed up sequential decoding. This typically consists in the past_key_valuesreturned by the model at a previous stage of decoding, when use_cache=True or config.use_cache=True.
  Two formats are allowed:
  • a Cache instance, see our kv cache guide;
  • 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)). This is also known as the legacy cache format.
    The model will output the same cache format that is fed as input. If no past_key_values are passed, the legacy cache format will be returned.
    If past_key_values are used, the user can optionally input only the last input_ids (those that don’t have their past key value states given to this model) of shape (batch_size, 1) instead of all input_idsof shape (batch_size, sequence_length).
• inputs_embeds (torch.Tensor 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.
• use_cache (bool, optional) — If set to True, past_key_values key value states are returned and can be used to speed up decoding (seepast_key_values).
• 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.

A transformers.models.prophetnet.modeling_prophetnet.ProphetNetDecoderModelOutput or a tuple oftorch.FloatTensor (if return_dict=False is passed or when config.return_dict=False) comprising various elements depending on the configuration (ProphetNetConfig) and inputs.

• last_hidden_state (torch.FloatTensor of shape (batch_size, decoder_sequence_length, hidden_size)) — Sequence of main stream hidden-states at the output of the last layer of the decoder 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.
• last_hidden_state_ngram (torch.FloatTensor of shape (batch_size, ngram * decoder_sequence_length, config.vocab_size)) — Sequence of predict stream hidden-states at the output of the last layer of the decoder of the model.
• past_key_values (List[torch.FloatTensor], optional, returned when use_cache=True is passed or when config.use_cache=True) — List of torch.FloatTensor of length config.n_layers, with each tensor of shape (2, batch_size, num_attn_heads, decoder_sequence_length, embed_size_per_head)).
  Contains pre-computed hidden-states (key and values in the attention blocks) of the decoder 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 + one for the output of each layer) of shape (batch_size, decoder_sequence_length, hidden_size).
  Hidden-states of main stream of the decoder at the output of each layer plus the initial embedding outputs.
• ngram_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, ngram * decoder_sequence_length, hidden_size).
  Hidden-states of the predict stream of the decoder 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_attn_heads, decoder_sequence_length, decoder_sequence_length).
  Attentions weights of the decoder, after the attention softmax, used to compute the weighted average in the self-attention heads.
• ngram_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_attn_heads, decoder_sequence_length, decoder_sequence_length).
  Attentions weights of the predict stream of the decoder, after the attention softmax, used to compute the weighted average in the
• 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_attn_heads, encoder_sequence_length, decoder_sequence_length).
  Attentions weights of the cross-attention layer of the decoder, after the attention softmax, used to compute the weighted average in the

The ProphetNetDecoder forward method, overrides the __call__ special method.

Although the recipe for forward pass needs to be defined within this function, one should call the Moduleinstance 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, ProphetNetDecoder import torch

tokenizer = AutoTokenizer.from_pretrained("microsoft/prophetnet-large-uncased") model = ProphetNetDecoder.from_pretrained("microsoft/prophetnet-large-uncased", add_cross_attention=False) inputs = tokenizer("Hello, my dog is cute", return_tensors="pt") outputs = model(**inputs)

last_hidden_states = outputs.last_hidden_state

ProphetNetForConditionalGeneration

class transformers.ProphetNetForConditionalGeneration

< source >

( config: ProphetNetConfig )

Parameters

• config (ProphetNetConfig) — 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 thefrom_pretrained() method to load the model weights.

The ProphetNet Model with a language modeling head. Can be used for sequence generation tasks.

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.

forward

< source >

( input_ids: typing.Optional[torch.Tensor] = None attention_mask: typing.Optional[torch.Tensor] = None decoder_input_ids: typing.Optional[torch.Tensor] = None decoder_attention_mask: typing.Optional[torch.BoolTensor] = None head_mask: typing.Optional[torch.Tensor] = None decoder_head_mask: typing.Optional[torch.Tensor] = None cross_attn_head_mask: typing.Optional[torch.Tensor] = None encoder_outputs: typing.Optional[torch.Tensor] = None past_key_values: typing.Optional[typing.Tuple[typing.Tuple[torch.Tensor]]] = None inputs_embeds: typing.Optional[torch.Tensor] = None decoder_inputs_embeds: typing.Optional[torch.Tensor] = None labels: typing.Optional[torch.Tensor] = None use_cache: typing.Optional[bool] = None output_attentions: typing.Optional[bool] = None output_hidden_states: typing.Optional[bool] = None return_dict: typing.Optional[bool] = None ) → transformers.models.prophetnet.modeling_prophetnet.ProphetNetSeq2SeqLMOutput or tuple(torch.FloatTensor)

Parameters

• input_ids (torch.Tensor of shape (batch_size, sequence_length), optional) — Indices of input sequence tokens in the vocabulary. Padding will be ignored by default.
  Indices can be obtained using AutoTokenizer. See PreTrainedTokenizer.encode() andPreTrainedTokenizer.call() for details.
  What are input IDs?
• attention_mask (torch.Tensor 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?
• decoder_input_ids (torch.Tensor of shape (batch_size, target_sequence_length), optional) — Indices of decoder input sequence tokens in the vocabulary.
  Indices can be obtained using AutoTokenizer. See PreTrainedTokenizer.encode() andPreTrainedTokenizer.call() for details.
  What are decoder input IDs?
  ProphetNet uses the eos_token_id as the starting token for decoder_input_ids generation. Ifpast_key_values is used, optionally only the last decoder_input_ids have to be input (seepast_key_values).
• decoder_attention_mask (torch.BoolTensor of shape (batch_size, target_sequence_length), optional) — Default behavior: generate a tensor that ignores pad tokens in decoder_input_ids. Causal mask will also be used by default.
• head_mask (torch.Tensor 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]:
  • 1 indicates the head is not masked,
  • 0 indicates the head is masked.
• decoder_head_mask (torch.Tensor of shape (decoder_layers, decoder_attention_heads), optional) — Mask to nullify selected heads of the attention modules in the decoder. Mask values selected in [0, 1]:
  • 1 indicates the head is not masked,
  • 0 indicates the head is masked.
• cross_attn_head_mask (torch.Tensor of shape (decoder_layers, decoder_attention_heads), optional) — Mask to nullify selected heads of the cross-attention modules. Mask values selected in [0, 1]:
  • 1 indicates the head is not masked,
  • 0 indicates the head is masked.
• encoder_outputs (torch.Tensor, optional) — Tuple consists of (last_hidden_state, optional: hidden_states, optional: attentions)last_hidden_state of shape (batch_size, sequence_length, hidden_size), optional) is a sequence of hidden-states at the output of the last layer of the encoder. Used in the cross-attention of the decoder.
• past_key_values (Tuple[Tuple[torch.Tensor]], optional) — Pre-computed hidden-states (key and values in the self-attention blocks and in the cross-attention blocks) that can be used to speed up sequential decoding. This typically consists in the past_key_valuesreturned by the model at a previous stage of decoding, when use_cache=True or config.use_cache=True.
  Two formats are allowed:
  • a Cache instance, see our kv cache guide;
  • 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)). This is also known as the legacy cache format.
    The model will output the same cache format that is fed as input. If no past_key_values are passed, the legacy cache format will be returned.
    If past_key_values are used, the user can optionally input only the last input_ids (those that don’t have their past key value states given to this model) of shape (batch_size, 1) instead of all input_idsof shape (batch_size, sequence_length).
• inputs_embeds (torch.Tensor 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.
• decoder_inputs_embeds (torch.Tensor of shape (batch_size, target_sequence_length, hidden_size), optional) — Optionally, instead of passing decoder_input_ids you can choose to directly pass an embedded representation. If past_key_values is used, optionally only the last decoder_inputs_embeds have to be input (see past_key_values). This is useful if you want more control over how to convertdecoder_input_ids indices into associated vectors than the model’s internal embedding lookup matrix.
  If decoder_input_ids and decoder_inputs_embeds are both unset, decoder_inputs_embeds takes the value of inputs_embeds.
• labels (torch.Tensor of shape (batch_size,), optional) — Labels for computing the sequence classification/regression loss. Indices should be in [-100, 0, ..., config.vocab_size - 1]. All labels set to -100 are ignored (masked), the loss is only computed for labels in [0, ..., config.vocab_size]
• use_cache (bool, optional) — If set to True, past_key_values key value states are returned and can be used to speed up decoding (seepast_key_values).
• 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.

A transformers.models.prophetnet.modeling_prophetnet.ProphetNetSeq2SeqLMOutput or a tuple oftorch.FloatTensor (if return_dict=False is passed or when config.return_dict=False) comprising various elements depending on the configuration (ProphetNetConfig) and inputs.

• loss (torch.FloatTensor of shape (1,), optional, returned when labels is provided) — Language modeling loss.
• logits (torch.FloatTensor of shape (batch_size, decoder_sequence_length, config.vocab_size)) — Prediction scores of the main stream language modeling head (scores for each vocabulary token before SoftMax).
• logits_ngram (torch.FloatTensor of shape (batch_size, ngram * decoder_sequence_length, config.vocab_size)) — Prediction scores of the predict stream language modeling head (scores for each vocabulary token before SoftMax).
• past_key_values (List[torch.FloatTensor], optional, returned when use_cache=True is passed or when config.use_cache=True) — List of torch.FloatTensor of length config.n_layers, with each tensor of shape (2, batch_size, num_attn_heads, decoder_sequence_length, embed_size_per_head)).
  Contains pre-computed hidden-states (key and values in the attention blocks) of the decoder that can be used (see past_key_values input) to speed up sequential decoding.
• decoder_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, decoder_sequence_length, hidden_size).
  Hidden-states of main stream of the decoder at the output of each layer plus the initial embedding outputs.
• decoder_ngram_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, ngram * decoder_sequence_length, hidden_size).
  Hidden-states of the predict stream of the decoder at the output of each layer plus the initial embedding outputs.
• decoder_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_attn_heads, decoder_sequence_length, decoder_sequence_length).
  Attentions weights of the decoder, after the attention softmax, used to compute the weighted average in the self-attention heads.
• decoder_ngram_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_attn_heads, decoder_sequence_length, decoder_sequence_length).
  Attentions weights of the predict stream of the decoder, 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_attn_heads, encoder_sequence_length, decoder_sequence_length).
  Attentions weights of the cross-attention layer of the decoder, after the attention softmax, used to compute the weighted average in the
• encoder_last_hidden_state (torch.FloatTensor of shape (batch_size, encoder_sequence_length, hidden_size), optional) — Sequence of hidden-states at the output of the last layer of the encoder of the model.
• encoder_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, encoder_sequence_length, hidden_size).
  Hidden-states of the encoder at the output of each layer plus the initial embedding outputs.
• encoder_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_attn_heads, encoder_sequence_length, encoder_sequence_length). Attentions weights of the encoder, after the attention softmax, used to compute the weighted average in the self-attention heads.

The ProphetNetForConditionalGeneration forward method, overrides the __call__ special method.

Although the recipe for forward pass needs to be defined within this function, one should call the Moduleinstance 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, ProphetNetForConditionalGeneration

tokenizer = AutoTokenizer.from_pretrained("microsoft/prophetnet-large-uncased") model = ProphetNetForConditionalGeneration.from_pretrained("microsoft/prophetnet-large-uncased")

input_ids = tokenizer( ... "Studies have been shown that owning a dog is good for you", return_tensors="pt" ... ).input_ids
decoder_input_ids = tokenizer("Studies show that", return_tensors="pt").input_ids
outputs = model(input_ids=input_ids, decoder_input_ids=decoder_input_ids)

logits_next_token = outputs.logits
logits_ngram_next_tokens = outputs.logits_ngram

ProphetNetForCausalLM

class transformers.ProphetNetForCausalLM

< source >

( config: ProphetNetConfig )

Parameters

• config (ProphetNetConfig) — 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 thefrom_pretrained() method to load the model weights.

The standalone decoder part of the ProphetNetModel with a lm head on top. The model can be used for causal

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.

forward

< source >

( input_ids: typing.Optional[torch.Tensor] = None attention_mask: typing.Optional[torch.Tensor] = None encoder_hidden_states: typing.Optional[torch.Tensor] = None encoder_attention_mask: typing.Optional[torch.Tensor] = None head_mask: typing.Optional[torch.Tensor] = None cross_attn_head_mask: typing.Optional[torch.Tensor] = None past_key_values: typing.Optional[typing.Tuple[typing.Tuple[torch.Tensor]]] = None inputs_embeds: typing.Optional[torch.Tensor] = None labels: typing.Optional[torch.Tensor] = None use_cache: typing.Optional[bool] = None output_attentions: typing.Optional[bool] = None output_hidden_states: typing.Optional[bool] = None return_dict: typing.Optional[bool] = None ) → transformers.models.prophetnet.modeling_prophetnet.ProphetNetDecoderLMOutput or tuple(torch.FloatTensor)

Parameters

• input_ids (torch.Tensor of shape (batch_size, sequence_length), optional) — Indices of input sequence tokens in the vocabulary. Padding will be ignored by default.
  Indices can be obtained using AutoTokenizer. See PreTrainedTokenizer.encode() andPreTrainedTokenizer.call() for details.
  What are input IDs?
• attention_mask (torch.Tensor 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?
• encoder_hidden_states (torch.Tensor 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.
• encoder_attention_mask (torch.Tensor 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 tokens that are masked.
• head_mask (torch.Tensor 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]:
  • 1 indicates the head is not masked,
  • 0 indicates the head is masked.
• cross_attn_head_mask (torch.Tensor of shape (decoder_layers, decoder_attention_heads), optional) — Mask to nullify selected heads of the cross-attention modules. Mask values selected in [0, 1]:
  • 1 indicates the head is not masked,
  • 0 indicates the head is masked.
• past_key_values (Tuple[Tuple[torch.Tensor]], optional) — Pre-computed hidden-states (key and values in the self-attention blocks and in the cross-attention blocks) that can be used to speed up sequential decoding. This typically consists in the past_key_valuesreturned by the model at a previous stage of decoding, when use_cache=True or config.use_cache=True.
  Two formats are allowed:
  • a Cache instance, see our kv cache guide;
  • 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)). This is also known as the legacy cache format.
    The model will output the same cache format that is fed as input. If no past_key_values are passed, the legacy cache format will be returned.
    If past_key_values are used, the user can optionally input only the last input_ids (those that don’t have their past key value states given to this model) of shape (batch_size, 1) instead of all input_idsof shape (batch_size, sequence_length).
• inputs_embeds (torch.Tensor 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.
• labels (torch.Tensor 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 n [0, ..., config.vocab_size]
• use_cache (bool, optional) — If set to True, past_key_values key value states are returned and can be used to speed up decoding (seepast_key_values).
• 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.

A transformers.models.prophetnet.modeling_prophetnet.ProphetNetDecoderLMOutput or a tuple oftorch.FloatTensor (if return_dict=False is passed or when config.return_dict=False) comprising various elements depending on the configuration (ProphetNetConfig) and inputs.

• loss (torch.FloatTensor of shape (1,), optional, returned when labels is provided) — Language modeling loss.
• logits (torch.FloatTensor of shape (batch_size, decoder_sequence_length, config.vocab_size)) — Prediction scores of the main stream language modeling head (scores for each vocabulary token before SoftMax).
• logits_ngram (torch.FloatTensor of shape (batch_size, ngram * decoder_sequence_length, config.vocab_size)) — Prediction scores of the predict stream language modeling head (scores for each vocabulary token before SoftMax).
• past_key_values (List[torch.FloatTensor], optional, returned when use_cache=True is passed or when config.use_cache=True) — List of torch.FloatTensor of length config.n_layers, with each tensor of shape (2, batch_size, num_attn_heads, decoder_sequence_length, embed_size_per_head)).
  Contains pre-computed hidden-states (key and values in the attention blocks) of the decoder 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 + one for the output of each layer) of shape (batch_size, decoder_sequence_length, hidden_size).
  Hidden-states of main stream of the decoder at the output of each layer plus the initial embedding outputs.
• ngram_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, ngram * decoder_sequence_length, hidden_size).
  Hidden-states of the predict stream of the decoder 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_attn_heads, decoder_sequence_length, decoder_sequence_length).
  Attentions weights of the decoder, after the attention softmax, used to compute the weighted average in the self-attention heads.
• ngram_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_attn_heads, decoder_sequence_length, decoder_sequence_length).
  Attentions weights of the predict stream of the decoder, after the attention softmax, used to compute the weighted average in the
• 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_attn_heads, encoder_sequence_length, decoder_sequence_length).
  Attentions weights of the cross-attention layer of the decoder, after the attention softmax, used to compute the weighted average in the

The ProphetNetForCausalLM forward method, overrides the __call__ special method.

Although the recipe for forward pass needs to be defined within this function, one should call the Moduleinstance 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, ProphetNetForCausalLM import torch

tokenizer = AutoTokenizer.from_pretrained("microsoft/prophetnet-large-uncased") model = ProphetNetForCausalLM.from_pretrained("microsoft/prophetnet-large-uncased") assert model.config.is_decoder, f"{model.class} has to be configured as a decoder." inputs = tokenizer("Hello, my dog is cute", return_tensors="pt") outputs = model(**inputs)

logits = outputs.logits

from transformers import BertTokenizer, EncoderDecoderModel, AutoTokenizer import torch

tokenizer_enc = BertTokenizer.from_pretrained("google-bert/bert-large-uncased") tokenizer_dec = AutoTokenizer.from_pretrained("microsoft/prophetnet-large-uncased") model = EncoderDecoderModel.from_encoder_decoder_pretrained( ... "google-bert/bert-large-uncased", "microsoft/prophetnet-large-uncased" ... )

ARTICLE = ( ... "the us state department said wednesday it had received no " ... "formal word from bolivia that it was expelling the us ambassador there " ... "but said the charges made against him are `` baseless ." ... ) input_ids = tokenizer_enc(ARTICLE, return_tensors="pt").input_ids labels = tokenizer_dec( ... "us rejects charges against its ambassador in bolivia", return_tensors="pt" ... ).input_ids outputs = model(input_ids=input_ids, decoder_input_ids=labels[:, :-1], labels=labels[:, 1:])

loss = outputs.loss

< > Update on GitHub