This model was released on 2020-01-13 and added to Hugging Face Transformers on 2020-11-16.
ProphetNet
Overview
The ProphetNet model was proposed inProphetNet: Predicting Future N-gram for Sequence-to-Sequence Pre-training, by Yu Yan, Weizhen Qi, Yeyun Gong, Dayiheng Liu, Nan Duan, Jiusheng Chen, RuofeiZhang, Ming Zhou on 13 Jan, 2020.
ProphetNet is an encoder-decoder model and can predict n-future tokens for “ngram” language modeling instead of justthe 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 novelself-supervised objective named future n-gram prediction and the proposed n-stream self-attention mechanism. Instead ofthe optimization of one-step ahead prediction in traditional sequence-to-sequence model, the ProphetNet is optimized byn-step ahead prediction which predicts the next n tokens simultaneously based on previous context tokens at each timestep. The future n-gram prediction explicitly encourages the model to plan for the future tokens and preventoverfitting on strong local correlations. We pre-train ProphetNet using a base scale dataset (16GB) and a large scaledataset (160GB) respectively. Then we conduct experiments on CNN/DailyMail, Gigaword, and SQuAD 1.1 benchmarks forabstractive summarization and question generation tasks. Experimental results show that ProphetNet achieves newstate-of-the-art results on all these datasets compared to the models using the same scale pretraining corpus.
The Authors’ code can be foundhere.
Usage tips
- ProphetNet is a model with absolute position embeddings so it’s usually advised to pad the inputs on the right rather thanthe 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
ProphetNetConfig
classtransformers.ProphetNetConfig
<source>(activation_dropout: typing.Optional[float] = 0.1activation_function: typing.Union[str, collections.abc.Callable, NoneType] = 'gelu'vocab_size: typing.Optional[int] = 30522hidden_size: typing.Optional[int] = 1024encoder_ffn_dim: typing.Optional[int] = 4096num_encoder_layers: typing.Optional[int] = 12num_encoder_attention_heads: typing.Optional[int] = 16decoder_ffn_dim: typing.Optional[int] = 4096num_decoder_layers: typing.Optional[int] = 12num_decoder_attention_heads: typing.Optional[int] = 16attention_dropout: typing.Optional[float] = 0.1dropout: typing.Optional[float] = 0.1max_position_embeddings: typing.Optional[int] = 512init_std: typing.Optional[float] = 0.02is_encoder_decoder: typing.Optional[bool] = Trueadd_cross_attention: typing.Optional[bool] = Truedecoder_start_token_id: typing.Optional[int] = 0ngram: typing.Optional[int] = 2num_buckets: typing.Optional[int] = 32relative_max_distance: typing.Optional[int] = 128disable_ngram_loss: typing.Optional[bool] = Falseeps: typing.Optional[float] = 0.0use_cache: typing.Optional[bool] = Truepad_token_id: typing.Optional[int] = 0bos_token_id: typing.Optional[int] = 1eos_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 (
strorfunction,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 bytheinputs_idspassed when callingProphetNetModel. - 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 theintermediate(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 largejust 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 toTrue) —Whether cross-attention layers should be added to the model. - is_encoder_decoder (
bool,optional, defaults toTrue) —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 firsttoken. - 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](seehttps://huggingface.co/papers/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 relativeposition calculation. See the [T5 paper](seehttps://huggingface.co/papers/1910.10683) for more details. - disable_ngram_loss (
bool,optional, defaults toFalse) —Whether be trained predicting only the next first token. - eps (
float,optional, defaults to 0.0) —Controls theepsilonparameter value for label smoothing in the loss calculation. If set to 0, no labelsmoothing is performed. - use_cache (
bool,optional, defaults toTrue) —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 aProphetNetModel. It is used to instantiate aProphetNet model according to the specified arguments, defining the model architecture. Instantiating aconfiguration with the defaults will yield a similar configuration to that of the ProphetNetmicrosoft/prophetnet-large-uncased architecture.
Configuration objects inherit fromPreTrainedConfig and can be used to control the model outputs. Read thedocumentation fromPreTrainedConfig for more information.
ProphetNetTokenizer
classtransformers.ProphetNetTokenizer
<source>(vocab_file: strdo_lower_case: typing.Optional[bool] = Truedo_basic_tokenize: typing.Optional[bool] = Truenever_split: typing.Optional[collections.abc.Iterable] = Noneunk_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] = Truestrip_accents: typing.Optional[bool] = Noneclean_up_tokenization_spaces: bool = True**kwargs)
Parameters
- vocab_file (
str) —File containing the vocabulary. - do_lower_case (
bool,optional, defaults toTrue) —Whether or not to lowercase the input when tokenizing. - do_basic_tokenize (
bool,optional, defaults toTrue) —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 thistoken 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 forsequence classification or for a text and a question for question answering. It is also used as the lasttoken of a sequence built with special tokens. - x_sep_token (
str,optional, defaults to"[X_SEP]") —Special second separator token, which can be generated byProphetNetForConditionalGeneration. It isused 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 languagemodeling. This is the token which the model will try to predict. - tokenize_chinese_chars (
bool,optional, defaults toTrue) —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 thevalue forlowercase(as in the original BERT). - clean_up_tokenization_spaces (
bool,optional, defaults toTrue) —Whether or not to cleanup spaces after decoding, cleanup consists in removing potential artifacts likeextra spaces.
Construct a ProphetNetTokenizer. Based on WordPiece.
This tokenizer inherits fromPreTrainedTokenizer which contains most of the main methods. Users should refer tothis superclass for more information regarding those methods.
build_inputs_with_special_tokens
<source>(token_ids_0: listtoken_ids_1: typing.Optional[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.
Returns
List[int]
List ofinput IDs with the appropriate special tokens.
Build model inputs from a sequence or a pair of sequence for sequence classification tasks by concatenating andadding 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.
get_special_tokens_mask
<source>(token_ids_0: listtoken_ids_1: typing.Optional[list[int]] = Nonealready_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 toFalse) —Whether or not the token list is already formatted with special tokens for the model.
Returns
List[int]
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 addingspecial tokens using the tokenizerprepare_for_model method.
ProphetNet specific outputs
classtransformers.models.prophetnet.modeling_prophetnet.ProphetNetSeq2SeqLMOutput
<source>(loss: typing.Optional[torch.FloatTensor] = Nonelogits: typing.Optional[torch.FloatTensor] = Nonelogits_ngram: typing.Optional[torch.FloatTensor] = Nonepast_key_values: typing.Optional[transformers.cache_utils.Cache] = Nonedecoder_hidden_states: typing.Optional[tuple[torch.FloatTensor]] = Nonedecoder_ngram_hidden_states: typing.Optional[tuple[torch.FloatTensor]] = Nonedecoder_attentions: typing.Optional[tuple[torch.FloatTensor]] = Nonedecoder_ngram_attentions: typing.Optional[tuple[torch.FloatTensor]] = Nonecross_attentions: typing.Optional[tuple[torch.FloatTensor]] = Noneencoder_last_hidden_state: typing.Optional[torch.FloatTensor] = Noneencoder_hidden_states: typing.Optional[tuple[torch.FloatTensor]] = Noneencoder_attentions: typing.Optional[tuple[torch.FloatTensor]] = None)
Parameters
- loss (
torch.FloatTensorof shape(1,),optional, returned whenlabelsis provided) —Language modeling loss. - logits (
torch.FloatTensorof shape(batch_size, decoder_sequence_length, config.vocab_size)) —Prediction scores of the main stream language modeling head (scores for each vocabulary token beforeSoftMax). - logits_ngram (
torch.FloatTensorof shape(batch_size, ngram * decoder_sequence_length, config.vocab_size)) —Prediction scores of the predict stream language modeling head (scores for each vocabulary token beforeSoftMax). - past_key_values (
Cache,optional, returned whenuse_cache=Trueis passed or whenconfig.use_cache=True) —It is aCache instance. For more details, see ourkv cache guide.Contains pre-computed hidden-states (key and values in the attention blocks) of the decoder that can beused (see
past_key_valuesinput) to speed up sequential decoding. - decoder_hidden_states (
tuple[torch.FloatTensor],optional, returned whenoutput_hidden_states=Trueis passed or whenconfig.output_hidden_states=True) —Tuple oftorch.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 decoder at the output of each layer plus the initial embedding outputs.
- decoder_ngram_hidden_states (
tuple(torch.FloatTensor),optional, returned whenoutput_hidden_states=Trueis passed or whenconfig.output_hidden_states=True) —Tuple oftorch.FloatTensor(one for the output of the embeddings + one for the output of each layer) ofshape(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 embeddingoutputs.
- decoder_attentions (
tuple[torch.FloatTensor],optional, returned whenoutput_attentions=Trueis passed or whenconfig.output_attentions=True) —Tuple oftorch.FloatTensor(one for each layer) of shape(batch_size, num_heads, sequence_length, sequence_length).Attentions weights of the decoder, after the attention softmax, used to compute the weighted average in theself-attention heads.
- decoder_ngram_attentions (
tuple(torch.FloatTensor),optional, returned whenoutput_attentions=Trueis passed or whenconfig.output_attentions=True) —Tuple oftorch.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 theweighted average in the self-attention heads.
- cross_attentions (
tuple[torch.FloatTensor],optional, returned whenoutput_attentions=Trueis passed or whenconfig.output_attentions=True) —Tuple oftorch.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 theweighted average in the cross-attention heads.
- encoder_last_hidden_state (
torch.FloatTensorof 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 whenoutput_hidden_states=Trueis passed or whenconfig.output_hidden_states=True) —Tuple oftorch.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 encoder at the output of each layer plus the initial embedding outputs.
- encoder_attentions (
tuple[torch.FloatTensor],optional, returned whenoutput_attentions=Trueis passed or whenconfig.output_attentions=True) —Tuple oftorch.FloatTensor(one for each layer) of shape(batch_size, num_heads, sequence_length, sequence_length).Attentions weights of the encoder, after the attention softmax, used to compute the weighted average in theself-attention heads.
Base class for sequence-to-sequence language models outputs.
classtransformers.models.prophetnet.modeling_prophetnet.ProphetNetSeq2SeqModelOutput
<source>(last_hidden_state: FloatTensorlast_hidden_state_ngram: typing.Optional[torch.FloatTensor] = Nonepast_key_values: typing.Optional[transformers.cache_utils.Cache] = Nonedecoder_hidden_states: typing.Optional[tuple[torch.FloatTensor]] = Nonedecoder_ngram_hidden_states: typing.Optional[tuple[torch.FloatTensor]] = Nonedecoder_attentions: typing.Optional[tuple[torch.FloatTensor]] = Nonedecoder_ngram_attentions: typing.Optional[tuple[torch.FloatTensor]] = Nonecross_attentions: typing.Optional[tuple[torch.FloatTensor]] = Noneencoder_last_hidden_state: typing.Optional[torch.FloatTensor] = Noneencoder_hidden_states: typing.Optional[tuple[torch.FloatTensor]] = Noneencoder_attentions: typing.Optional[tuple[torch.FloatTensor]] = None)
Parameters
- last_hidden_state (
torch.FloatTensorof 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_valuesis used only the last hidden-state of the sequences of shape(batch_size, 1, hidden_size)is output. - last_hidden_state_ngram (
torch.FloatTensorof 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 (
Cache,optional, returned whenuse_cache=Trueis passed or whenconfig.use_cache=True) —It is aCache instance. For more details, see ourkv cache guide.Contains pre-computed hidden-states (key and values in the attention blocks) of the decoder that can beused (see
past_key_valuesinput) to speed up sequential decoding. - decoder_hidden_states (
tuple[torch.FloatTensor],optional, returned whenoutput_hidden_states=Trueis passed or whenconfig.output_hidden_states=True) —Tuple oftorch.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 decoder at the output of each layer plus the initial embedding outputs.
- decoder_ngram_hidden_states (
tuple(torch.FloatTensor),optional, returned whenoutput_hidden_states=Trueis passed or whenconfig.output_hidden_states=True) —Tuple oftorch.FloatTensor(one for the output of the embeddings + one for the output of each layer) ofshape(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 embeddingoutputs.
- decoder_attentions (
tuple[torch.FloatTensor],optional, returned whenoutput_attentions=Trueis passed or whenconfig.output_attentions=True) —Tuple oftorch.FloatTensor(one for each layer) of shape(batch_size, num_heads, sequence_length, sequence_length).Attentions weights of the decoder, after the attention softmax, used to compute the weighted average in theself-attention heads.
- decoder_ngram_attentions (
tuple(torch.FloatTensor),optional, returned whenoutput_attentions=Trueis passed or whenconfig.output_attentions=True) —Tuple oftorch.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 theweighted average in the
- cross_attentions (
tuple[torch.FloatTensor],optional, returned whenoutput_attentions=Trueis passed or whenconfig.output_attentions=True) —Tuple oftorch.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 theweighted average in the cross-attention heads.
- encoder_last_hidden_state (
torch.FloatTensorof 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 whenoutput_hidden_states=Trueis passed or whenconfig.output_hidden_states=True) —Tuple oftorch.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 encoder at the output of each layer plus the initial embedding outputs.
- encoder_attentions (
tuple[torch.FloatTensor],optional, returned whenoutput_attentions=Trueis passed or whenconfig.output_attentions=True) —Tuple oftorch.FloatTensor(one for each layer) of shape(batch_size, num_heads, sequence_length, sequence_length).Attentions weights of the encoder, after the attention softmax, used to compute the weighted average in theself-attention heads.
Base class for model encoder’s outputs that also contains : pre-computed hidden states that can speed up sequentialdecoding.
classtransformers.models.prophetnet.modeling_prophetnet.ProphetNetDecoderModelOutput
<source>(last_hidden_state: FloatTensorlast_hidden_state_ngram: typing.Optional[torch.FloatTensor] = Nonepast_key_values: typing.Optional[transformers.cache_utils.Cache] = Nonehidden_states: typing.Optional[tuple[torch.FloatTensor]] = Nonehidden_states_ngram: typing.Optional[tuple[torch.FloatTensor]] = Noneattentions: typing.Optional[tuple[torch.FloatTensor]] = Nonengram_attentions: typing.Optional[tuple[torch.FloatTensor]] = Nonecross_attentions: typing.Optional[tuple[torch.FloatTensor]] = None)
Parameters
- last_hidden_state (
torch.FloatTensorof 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_valuesis used only the last hidden-state of the sequences of shape(batch_size, 1, hidden_size)is output. - last_hidden_state_ngram (
torch.FloatTensorof 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 (
Cache,optional, returned whenuse_cache=Trueis passed or whenconfig.use_cache=True) —It is aCache instance. For more details, see ourkv cache guide.Contains pre-computed hidden-states (key and values in the attention blocks) of the decoder that can beused (see
past_key_valuesinput) to speed up sequential decoding. - hidden_states (
tuple[torch.FloatTensor],optional, returned whenoutput_hidden_states=Trueis passed or whenconfig.output_hidden_states=True) —Tuple oftorch.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.
- hidden_states_ngram (
tuple(torch.FloatTensor),optional, returned whenoutput_hidden_states=Trueis passed or whenconfig.output_hidden_states=True) —Tuple oftorch.FloatTensor(one for the output of the embeddings + one for the output of each layer) ofshape(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 embeddingoutputs.
- attentions (
tuple[torch.FloatTensor],optional, returned whenoutput_attentions=Trueis passed or whenconfig.output_attentions=True) —Tuple oftorch.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-attentionheads.
- ngram_attentions (
tuple(torch.FloatTensor),optional, returned whenoutput_attentions=Trueis passed or whenconfig.output_attentions=True) —Tuple oftorch.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 theweighted average in the
- cross_attentions (
tuple[torch.FloatTensor],optional, returned whenoutput_attentions=Trueis passed or whenconfig.output_attentions=True) —Tuple oftorch.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 theweighted average in the cross-attention heads.
Base class for model’s outputs that may also contain a past key/values (to speed up sequential decoding).
classtransformers.models.prophetnet.modeling_prophetnet.ProphetNetDecoderLMOutput
<source>(loss: typing.Optional[torch.FloatTensor] = Nonelogits: typing.Optional[torch.FloatTensor] = Nonelogits_ngram: typing.Optional[torch.FloatTensor] = Nonepast_key_values: typing.Optional[transformers.cache_utils.Cache] = Nonehidden_states: typing.Optional[tuple[torch.FloatTensor]] = Nonehidden_states_ngram: typing.Optional[tuple[torch.FloatTensor]] = Noneattentions: typing.Optional[tuple[torch.FloatTensor]] = Nonengram_attentions: typing.Optional[tuple[torch.FloatTensor]] = Nonecross_attentions: typing.Optional[tuple[torch.FloatTensor]] = None)
Parameters
- loss (
torch.FloatTensorof shape(1,),optional, returned whenlabelsis provided) —Language modeling loss. - logits (
torch.FloatTensorof shape(batch_size, decoder_sequence_length, config.vocab_size)) —Prediction scores of the main stream language modeling head (scores for each vocabulary token beforeSoftMax). - logits_ngram (
torch.FloatTensorof shape(batch_size, ngram * decoder_sequence_length, config.vocab_size)) —Prediction scores of the predict stream language modeling head (scores for each vocabulary token beforeSoftMax). - past_key_values (
Cache,optional, returned whenuse_cache=Trueis passed or whenconfig.use_cache=True) —It is aCache instance. For more details, see ourkv cache guide.Contains pre-computed hidden-states (key and values in the attention blocks) of the decoder that can beused (see
past_key_valuesinput) to speed up sequential decoding. - hidden_states (
tuple[torch.FloatTensor],optional, returned whenoutput_hidden_states=Trueis passed or whenconfig.output_hidden_states=True) —Tuple oftorch.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.
- hidden_states_ngram (
tuple(torch.FloatTensor),optional, returned whenoutput_hidden_states=Trueis passed or whenconfig.output_hidden_states=True) —Tuple oftorch.FloatTensor(one for the output of the embeddings + one for the output of each layer) ofshape(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 embeddingoutputs.
- attentions (
tuple[torch.FloatTensor],optional, returned whenoutput_attentions=Trueis passed or whenconfig.output_attentions=True) —Tuple oftorch.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-attentionheads.
- ngram_attentions (
tuple(torch.FloatTensor),optional, returned whenoutput_attentions=Trueis passed or whenconfig.output_attentions=True) —Tuple oftorch.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 theweighted average in the
- cross_attentions (
tuple[torch.FloatTensor],optional, returned whenoutput_attentions=Trueis passed or whenconfig.output_attentions=True) —Tuple oftorch.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 theweighted average in the cross-attention heads.
Base class for model’s outputs that may also contain a past key/values (to speed up sequential decoding).
ProphetNetModel
classtransformers.ProphetNetModel
<source>(config: ProphetNetConfig)
Parameters
- config (ProphetNetConfig) —Model configuration class with all the parameters of the model. Initializing with a config file does notload 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 fromPreTrainedModel. Check the superclass documentation for the generic methods thelibrary implements for all its model (such as downloading or saving, resizing the input embeddings, pruning headsetc.)
This model is also a PyTorchtorch.nn.Module subclass.Use it as a regular PyTorch Module and refer to the PyTorch documentation for all matter related to general usageand behavior.
forward
<source>(input_ids: typing.Optional[torch.Tensor] = Noneattention_mask: typing.Optional[torch.Tensor] = Nonedecoder_input_ids: typing.Optional[torch.Tensor] = Nonedecoder_attention_mask: typing.Optional[torch.BoolTensor] = Noneencoder_outputs: typing.Optional[tuple] = Nonepast_key_values: typing.Optional[transformers.cache_utils.Cache] = Noneinputs_embeds: typing.Optional[torch.Tensor] = Nonedecoder_inputs_embeds: typing.Optional[torch.Tensor] = Noneuse_cache: typing.Optional[bool] = Noneoutput_attentions: typing.Optional[bool] = Noneoutput_hidden_states: typing.Optional[bool] = Nonereturn_dict: typing.Optional[bool] = Nonecache_position: typing.Optional[torch.Tensor] = None)→transformers.models.prophetnet.modeling_prophetnet.ProphetNetSeq2SeqModelOutput ortuple(torch.FloatTensor)
Parameters
- input_ids (
torch.Tensorof shape(batch_size, sequence_length),optional) —Indices of input sequence tokens in the vocabulary. Padding will be ignored by default.Indices can be obtained usingAutoTokenizer. SeePreTrainedTokenizer.encode() andPreTrainedTokenizer.call() for details.
- attention_mask (
torch.Tensorof 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 arenot masked,
- 0 for tokens that aremasked.
- decoder_input_ids (
torch.LongTensorof shape(batch_size, target_sequence_length),optional) —Indices of decoder input sequence tokens in the vocabulary.Indices can be obtained usingAutoTokenizer. SeePreTrainedTokenizer.encode() andPreTrainedTokenizer.call() for details.
ProphetNet uses the
eos_token_idas the starting token fordecoder_input_idsgeneration. Ifpast_key_valuesis used, optionally only the lastdecoder_input_idshave to be input (seepast_key_values). - decoder_attention_mask (
torch.BoolTensorof shape(batch_size, target_sequence_length),optional) —Default behavior: generate a tensor that ignores pad tokens indecoder_input_ids. Causal mask will alsobe used by default. - encoder_outputs (
tuple,optional) —Tuple consists of (last_hidden_state,optional:hidden_states,optional:attentions)last_hidden_stateof shape(batch_size, sequence_length, hidden_size),optional) is a sequence ofhidden-states at the output of the last layer of the encoder. Used in the cross-attention of the decoder. - past_key_values (
~cache_utils.Cache,optional) —Pre-computed hidden-states (key and values in the self-attention blocks and in the cross-attentionblocks) that can be used to speed up sequential decoding. This typically consists in thepast_key_valuesreturned by the model at a previous stage of decoding, whenuse_cache=Trueorconfig.use_cache=True.OnlyCache instance is allowed as input, see ourkv cache guide.If no
past_key_valuesare passed,DynamicCache will be initialized by default.The model will output the same cache format that is fed as input.
If
past_key_valuesare used, the user is expected to input only unprocessedinput_ids(those that don’thave their past key value states given to this model) of shape(batch_size, unprocessed_length)instead of allinput_idsof shape(batch_size, sequence_length). - inputs_embeds (
torch.Tensorof shape(batch_size, sequence_length, hidden_size),optional) —Optionally, instead of passinginput_idsyou can choose to directly pass an embedded representation. Thisis useful if you want more control over how to convertinput_idsindices into associated vectors than themodel’s internal embedding lookup matrix. - decoder_inputs_embeds (
torch.Tensorof shape(batch_size, target_sequence_length, hidden_size),optional) —Optionally, instead of passingdecoder_input_idsyou can choose to directly pass an embeddedrepresentation. Ifpast_key_valuesis used, optionally only the lastdecoder_inputs_embedshave to beinput (seepast_key_values). This is useful if you want more control over how to convertdecoder_input_idsindices into associated vectors than the model’s internal embedding lookup matrix.If
decoder_input_idsanddecoder_inputs_embedsare both unset,decoder_inputs_embedstakes the valueofinputs_embeds. - use_cache (
bool,optional) —If set toTrue,past_key_valueskey 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. Seeattentionsunder returnedtensors for more detail. - output_hidden_states (
bool,optional) —Whether or not to return the hidden states of all layers. Seehidden_statesunder returned tensors formore detail. - return_dict (
bool,optional) —Whether or not to return aModelOutput instead of a plain tuple. - cache_position (
torch.Tensorof shape(sequence_length),optional) —Indices depicting the position of the input sequence tokens in the sequence. Contrarily toposition_ids,this tensor is not affected by padding. It is used to update the cache in the correct position and to inferthe complete sequence length.
Returns
transformers.models.prophetnet.modeling_prophetnet.ProphetNetSeq2SeqModelOutput ortuple(torch.FloatTensor)
Atransformers.models.prophetnet.modeling_prophetnet.ProphetNetSeq2SeqModelOutput or a tuple oftorch.FloatTensor (ifreturn_dict=False is passed or whenconfig.return_dict=False) comprising variouselements depending on the configuration (ProphetNetConfig) and inputs.
last_hidden_state (
torch.FloatTensorof 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_valuesis used only the last hidden-state of the sequences of shape(batch_size, 1, hidden_size)is output.last_hidden_state_ngram (
torch.FloatTensorof 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 (
Cache,optional, returned whenuse_cache=Trueis passed or whenconfig.use_cache=True) — It is aCache instance. For more details, see ourkv cache guide.Contains pre-computed hidden-states (key and values in the attention blocks) of the decoder that can beused (see
past_key_valuesinput) to speed up sequential decoding.decoder_hidden_states (
tuple[torch.FloatTensor],optional, returned whenoutput_hidden_states=Trueis passed or whenconfig.output_hidden_states=True) — Tuple oftorch.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 decoder at the output of each layer plus the initial embedding outputs.
decoder_ngram_hidden_states (
tuple(torch.FloatTensor),optional, returned whenoutput_hidden_states=Trueis passed or whenconfig.output_hidden_states=True) — Tuple oftorch.FloatTensor(one for the output of the embeddings + one for the output of each layer) ofshape(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 embeddingoutputs.
decoder_attentions (
tuple[torch.FloatTensor],optional, returned whenoutput_attentions=Trueis passed or whenconfig.output_attentions=True) — Tuple oftorch.FloatTensor(one for each layer) of shape(batch_size, num_heads, sequence_length, sequence_length).Attentions weights of the decoder, after the attention softmax, used to compute the weighted average in theself-attention heads.
decoder_ngram_attentions (
tuple(torch.FloatTensor),optional, returned whenoutput_attentions=Trueis passed or whenconfig.output_attentions=True) — Tuple oftorch.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 theweighted average in the
cross_attentions (
tuple[torch.FloatTensor],optional, returned whenoutput_attentions=Trueis passed or whenconfig.output_attentions=True) — Tuple oftorch.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 theweighted average in the cross-attention heads.
encoder_last_hidden_state (
torch.FloatTensorof 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 whenoutput_hidden_states=Trueis passed or whenconfig.output_hidden_states=True) — Tuple oftorch.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 encoder at the output of each layer plus the initial embedding outputs.
encoder_attentions (
tuple[torch.FloatTensor],optional, returned whenoutput_attentions=Trueis passed or whenconfig.output_attentions=True) — Tuple oftorch.FloatTensor(one for each layer) of shape(batch_size, num_heads, sequence_length, sequence_length).Attentions weights of the encoder, after the attention softmax, used to compute the weighted average in theself-attention heads.
TheProphetNetModel 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 whilethe latter silently ignores them.
Example:
>>>from transformersimport 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# Batch size 1>>>decoder_input_ids = tokenizer("Studies show that", return_tensors="pt").input_ids# Batch size 1>>>outputs = model(input_ids=input_ids, decoder_input_ids=decoder_input_ids)>>>last_hidden_states = outputs.last_hidden_state# main stream hidden states>>>last_hidden_states_ngram = outputs.last_hidden_state_ngram# predict hidden states
ProphetNetEncoder
classtransformers.ProphetNetEncoder
<source>(config: ProphetNetConfigword_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 notload the weights associated with the model, only the configuration. Check out thefrom_pretrained() method to load the model weights.
- word_embeddings (
torch.nn.Embeddingsof shape(config.vocab_size, config.hidden_size),optional) —The word embedding parameters. This can be used to initializeProphetNetEncoder with pre-defined wordembeddings instead of randomly initialized word embeddings.
The standalone encoder part of the ProphetNetModel.
This model inherits fromPreTrainedModel. Check the superclass documentation for the generic methods thelibrary implements for all its model (such as downloading or saving, resizing the input embeddings, pruning headsetc.)
This model is also a PyTorchtorch.nn.Module subclass.Use it as a regular PyTorch Module and refer to the PyTorch documentation for all matter related to general usageand behavior.
forward
<source>(input_ids: typing.Optional[torch.Tensor] = Noneattention_mask: typing.Optional[torch.Tensor] = Noneinputs_embeds: typing.Optional[torch.Tensor] = Noneoutput_attentions: typing.Optional[bool] = Noneoutput_hidden_states: typing.Optional[bool] = Nonereturn_dict: typing.Optional[bool] = None)→transformers.modeling_outputs.BaseModelOutput ortuple(torch.FloatTensor)
Parameters
- input_ids (
torch.Tensorof shape(batch_size, sequence_length),optional) —Indices of input sequence tokens in the vocabulary. Padding will be ignored by default.Indices can be obtained usingAutoTokenizer. SeePreTrainedTokenizer.encode() andPreTrainedTokenizer.call() for details.
- attention_mask (
torch.Tensorof 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 arenot masked,
- 0 for tokens that aremasked.
- inputs_embeds (
torch.Tensorof shape(batch_size, sequence_length, hidden_size),optional) —Optionally, instead of passinginput_idsyou can choose to directly pass an embedded representation. Thisis useful if you want more control over how to convertinput_idsindices into associated vectors than themodel’s internal embedding lookup matrix. - output_attentions (
bool,optional) —Whether or not to return the attentions tensors of all attention layers. Seeattentionsunder returnedtensors for more detail. - output_hidden_states (
bool,optional) —Whether or not to return the hidden states of all layers. Seehidden_statesunder returned tensors formore detail. - return_dict (
bool,optional) —Whether or not to return aModelOutput instead of a plain tuple.
Returns
transformers.modeling_outputs.BaseModelOutput ortuple(torch.FloatTensor)
Atransformers.modeling_outputs.BaseModelOutput or a tuple oftorch.FloatTensor (ifreturn_dict=False is passed or whenconfig.return_dict=False) comprising variouselements depending on the configuration (ProphetNetConfig) and inputs.
last_hidden_state (
torch.FloatTensorof 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 whenoutput_hidden_states=Trueis passed or whenconfig.output_hidden_states=True) — Tuple oftorch.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 whenoutput_attentions=Trueis passed or whenconfig.output_attentions=True) — Tuple oftorch.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-attentionheads.
TheProphetNetEncoder 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 whilethe latter silently ignores them.
Example:
>>>from transformersimport 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
classtransformers.ProphetNetDecoder
<source>(config: ProphetNetConfigword_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 notload the weights associated with the model, only the configuration. Check out thefrom_pretrained() method to load the model weights.
- word_embeddings (
torch.nn.Embeddingsof shape(config.vocab_size, config.hidden_size),optional) —The word embedding parameters. This can be used to initializeProphetNetEncoder with pre-defined wordembeddings instead of randomly initialized word embeddings.
The standalone decoder part of the ProphetNetModel.
This model inherits fromPreTrainedModel. Check the superclass documentation for the generic methods thelibrary implements for all its model (such as downloading or saving, resizing the input embeddings, pruning headsetc.)
This model is also a PyTorchtorch.nn.Module subclass.Use it as a regular PyTorch Module and refer to the PyTorch documentation for all matter related to general usageand behavior.
forward
<source>(input_ids: typing.Optional[torch.Tensor] = Noneattention_mask: typing.Optional[torch.Tensor] = Noneencoder_hidden_states: typing.Optional[torch.Tensor] = Noneencoder_attention_mask: typing.Optional[torch.Tensor] = Nonepast_key_values: typing.Optional[transformers.cache_utils.Cache] = Noneinputs_embeds: typing.Optional[torch.Tensor] = Noneuse_cache: typing.Optional[bool] = Noneoutput_attentions: typing.Optional[bool] = Noneoutput_hidden_states: typing.Optional[bool] = Nonereturn_dict: typing.Optional[bool] = Nonecache_position: typing.Optional[torch.Tensor] = None)→transformers.models.prophetnet.modeling_prophetnet.ProphetNetDecoderModelOutput ortuple(torch.FloatTensor)
Parameters
- input_ids (
torch.Tensorof shape(batch_size, sequence_length),optional) —Indices of input sequence tokens in the vocabulary. Padding will be ignored by default.Indices can be obtained usingAutoTokenizer. SeePreTrainedTokenizer.encode() andPreTrainedTokenizer.call() for details.
- attention_mask (
torch.Tensorof 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 arenot masked,
- 0 for tokens that aremasked.
- encoder_hidden_states (
torch.Tensorof 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-attentionif the model is configured as a decoder. - encoder_attention_mask (
torch.Tensorof shape(batch_size, sequence_length),optional) —Mask to avoid performing attention on the padding token indices of the encoder input. This mask is used inthe cross-attention if the model is configured as a decoder. Mask values selected in[0, 1]:- 1 for tokens that arenot masked,
- 0 for tokens that aremasked.
- past_key_values (
~cache_utils.Cache,optional) —Pre-computed hidden-states (key and values in the self-attention blocks and in the cross-attentionblocks) that can be used to speed up sequential decoding. This typically consists in thepast_key_valuesreturned by the model at a previous stage of decoding, whenuse_cache=Trueorconfig.use_cache=True.OnlyCache instance is allowed as input, see ourkv cache guide.If no
past_key_valuesare passed,DynamicCache will be initialized by default.The model will output the same cache format that is fed as input.
If
past_key_valuesare used, the user is expected to input only unprocessedinput_ids(those that don’thave their past key value states given to this model) of shape(batch_size, unprocessed_length)instead of allinput_idsof shape(batch_size, sequence_length). - inputs_embeds (
torch.Tensorof shape(batch_size, sequence_length, hidden_size),optional) —Optionally, instead of passinginput_idsyou can choose to directly pass an embedded representation. Thisis useful if you want more control over how to convertinput_idsindices into associated vectors than themodel’s internal embedding lookup matrix. - use_cache (
bool,optional) —If set toTrue,past_key_valueskey 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. Seeattentionsunder returnedtensors for more detail. - output_hidden_states (
bool,optional) —Whether or not to return the hidden states of all layers. Seehidden_statesunder returned tensors formore detail. - return_dict (
bool,optional) —Whether or not to return aModelOutput instead of a plain tuple. - cache_position (
torch.Tensorof shape(sequence_length),optional) —Indices depicting the position of the input sequence tokens in the sequence. Contrarily toposition_ids,this tensor is not affected by padding. It is used to update the cache in the correct position and to inferthe complete sequence length.
Returns
transformers.models.prophetnet.modeling_prophetnet.ProphetNetDecoderModelOutput ortuple(torch.FloatTensor)
Atransformers.models.prophetnet.modeling_prophetnet.ProphetNetDecoderModelOutput or a tuple oftorch.FloatTensor (ifreturn_dict=False is passed or whenconfig.return_dict=False) comprising variouselements depending on the configuration (ProphetNetConfig) and inputs.
last_hidden_state (
torch.FloatTensorof 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_valuesis used only the last hidden-state of the sequences of shape(batch_size, 1, hidden_size)is output.last_hidden_state_ngram (
torch.FloatTensorof 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 (
Cache,optional, returned whenuse_cache=Trueis passed or whenconfig.use_cache=True) — It is aCache instance. For more details, see ourkv cache guide.Contains pre-computed hidden-states (key and values in the attention blocks) of the decoder that can beused (see
past_key_valuesinput) to speed up sequential decoding.hidden_states (
tuple[torch.FloatTensor],optional, returned whenoutput_hidden_states=Trueis passed or whenconfig.output_hidden_states=True) — Tuple oftorch.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.
hidden_states_ngram (
tuple(torch.FloatTensor),optional, returned whenoutput_hidden_states=Trueis passed or whenconfig.output_hidden_states=True) — Tuple oftorch.FloatTensor(one for the output of the embeddings + one for the output of each layer) ofshape(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 embeddingoutputs.
attentions (
tuple[torch.FloatTensor],optional, returned whenoutput_attentions=Trueis passed or whenconfig.output_attentions=True) — Tuple oftorch.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-attentionheads.
ngram_attentions (
tuple(torch.FloatTensor),optional, returned whenoutput_attentions=Trueis passed or whenconfig.output_attentions=True) — Tuple oftorch.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 theweighted average in the
cross_attentions (
tuple[torch.FloatTensor],optional, returned whenoutput_attentions=Trueis passed or whenconfig.output_attentions=True) — Tuple oftorch.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 theweighted average in the cross-attention heads.
TheProphetNetDecoder 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 whilethe latter silently ignores them.
Example:
>>>from transformersimport 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
classtransformers.ProphetNetForConditionalGeneration
<source>(config: ProphetNetConfig)
Parameters
- config (ProphetNetConfig) —Model configuration class with all the parameters of the model. Initializing with a config file does notload 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 fromPreTrainedModel. Check the superclass documentation for the generic methods thelibrary implements for all its model (such as downloading or saving, resizing the input embeddings, pruning headsetc.)
This model is also a PyTorchtorch.nn.Module subclass.Use it as a regular PyTorch Module and refer to the PyTorch documentation for all matter related to general usageand behavior.
forward
<source>(input_ids: typing.Optional[torch.Tensor] = Noneattention_mask: typing.Optional[torch.Tensor] = Nonedecoder_input_ids: typing.Optional[torch.Tensor] = Nonedecoder_attention_mask: typing.Optional[torch.BoolTensor] = Noneencoder_outputs: typing.Optional[torch.Tensor] = Nonepast_key_values: typing.Optional[transformers.cache_utils.Cache] = Noneinputs_embeds: typing.Optional[torch.Tensor] = Nonedecoder_inputs_embeds: typing.Optional[torch.Tensor] = Nonelabels: typing.Optional[torch.Tensor] = Noneuse_cache: typing.Optional[bool] = Noneoutput_attentions: typing.Optional[bool] = Noneoutput_hidden_states: typing.Optional[bool] = Nonereturn_dict: typing.Optional[bool] = Nonecache_position: typing.Optional[torch.Tensor] = None**kwargs)→transformers.models.prophetnet.modeling_prophetnet.ProphetNetSeq2SeqLMOutput ortuple(torch.FloatTensor)
Parameters
- input_ids (
torch.Tensorof shape(batch_size, sequence_length),optional) —Indices of input sequence tokens in the vocabulary. Padding will be ignored by default.Indices can be obtained usingAutoTokenizer. SeePreTrainedTokenizer.encode() andPreTrainedTokenizer.call() for details.
- attention_mask (
torch.Tensorof 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 arenot masked,
- 0 for tokens that aremasked.
- decoder_input_ids (
torch.LongTensorof shape(batch_size, target_sequence_length),optional) —Indices of decoder input sequence tokens in the vocabulary.Indices can be obtained usingAutoTokenizer. SeePreTrainedTokenizer.encode() andPreTrainedTokenizer.call() for details.
ProphetNet uses the
eos_token_idas the starting token fordecoder_input_idsgeneration. Ifpast_key_valuesis used, optionally only the lastdecoder_input_idshave to be input (seepast_key_values). - decoder_attention_mask (
torch.BoolTensorof shape(batch_size, target_sequence_length),optional) —Default behavior: generate a tensor that ignores pad tokens indecoder_input_ids. Causal mask will alsobe used by default. - encoder_outputs (
torch.Tensor,optional) —Tuple consists of (last_hidden_state,optional:hidden_states,optional:attentions)last_hidden_stateof shape(batch_size, sequence_length, hidden_size),optional) is a sequence ofhidden-states at the output of the last layer of the encoder. Used in the cross-attention of the decoder. - past_key_values (
~cache_utils.Cache,optional) —Pre-computed hidden-states (key and values in the self-attention blocks and in the cross-attentionblocks) that can be used to speed up sequential decoding. This typically consists in thepast_key_valuesreturned by the model at a previous stage of decoding, whenuse_cache=Trueorconfig.use_cache=True.OnlyCache instance is allowed as input, see ourkv cache guide.If no
past_key_valuesare passed,DynamicCache will be initialized by default.The model will output the same cache format that is fed as input.
If
past_key_valuesare used, the user is expected to input only unprocessedinput_ids(those that don’thave their past key value states given to this model) of shape(batch_size, unprocessed_length)instead of allinput_idsof shape(batch_size, sequence_length). - inputs_embeds (
torch.Tensorof shape(batch_size, sequence_length, hidden_size),optional) —Optionally, instead of passinginput_idsyou can choose to directly pass an embedded representation. Thisis useful if you want more control over how to convertinput_idsindices into associated vectors than themodel’s internal embedding lookup matrix. - decoder_inputs_embeds (
torch.Tensorof shape(batch_size, target_sequence_length, hidden_size),optional) —Optionally, instead of passingdecoder_input_idsyou can choose to directly pass an embeddedrepresentation. Ifpast_key_valuesis used, optionally only the lastdecoder_inputs_embedshave to beinput (seepast_key_values). This is useful if you want more control over how to convertdecoder_input_idsindices into associated vectors than the model’s internal embedding lookup matrix.If
decoder_input_idsanddecoder_inputs_embedsare both unset,decoder_inputs_embedstakes the valueofinputs_embeds. - labels (
torch.LongTensorof 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-100are ignored (masked), the loss is only computed forlabels in[0, ..., config.vocab_size] - use_cache (
bool,optional) —If set toTrue,past_key_valueskey 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. Seeattentionsunder returnedtensors for more detail. - output_hidden_states (
bool,optional) —Whether or not to return the hidden states of all layers. Seehidden_statesunder returned tensors formore detail. - return_dict (
bool,optional) —Whether or not to return aModelOutput instead of a plain tuple. - cache_position (
torch.Tensorof shape(sequence_length),optional) —Indices depicting the position of the input sequence tokens in the sequence. Contrarily toposition_ids,this tensor is not affected by padding. It is used to update the cache in the correct position and to inferthe complete sequence length.
Returns
transformers.models.prophetnet.modeling_prophetnet.ProphetNetSeq2SeqLMOutput ortuple(torch.FloatTensor)
Atransformers.models.prophetnet.modeling_prophetnet.ProphetNetSeq2SeqLMOutput or a tuple oftorch.FloatTensor (ifreturn_dict=False is passed or whenconfig.return_dict=False) comprising variouselements depending on the configuration (ProphetNetConfig) and inputs.
loss (
torch.FloatTensorof shape(1,),optional, returned whenlabelsis provided) — Language modeling loss.logits (
torch.FloatTensorof shape(batch_size, decoder_sequence_length, config.vocab_size)) — Prediction scores of the main stream language modeling head (scores for each vocabulary token beforeSoftMax).logits_ngram (
torch.FloatTensorof shape(batch_size, ngram * decoder_sequence_length, config.vocab_size)) — Prediction scores of the predict stream language modeling head (scores for each vocabulary token beforeSoftMax).past_key_values (
Cache,optional, returned whenuse_cache=Trueis passed or whenconfig.use_cache=True) — It is aCache instance. For more details, see ourkv cache guide.Contains pre-computed hidden-states (key and values in the attention blocks) of the decoder that can beused (see
past_key_valuesinput) to speed up sequential decoding.decoder_hidden_states (
tuple[torch.FloatTensor],optional, returned whenoutput_hidden_states=Trueis passed or whenconfig.output_hidden_states=True) — Tuple oftorch.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 decoder at the output of each layer plus the initial embedding outputs.
decoder_ngram_hidden_states (
tuple(torch.FloatTensor),optional, returned whenoutput_hidden_states=Trueis passed or whenconfig.output_hidden_states=True) — Tuple oftorch.FloatTensor(one for the output of the embeddings + one for the output of each layer) ofshape(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 embeddingoutputs.
decoder_attentions (
tuple[torch.FloatTensor],optional, returned whenoutput_attentions=Trueis passed or whenconfig.output_attentions=True) — Tuple oftorch.FloatTensor(one for each layer) of shape(batch_size, num_heads, sequence_length, sequence_length).Attentions weights of the decoder, after the attention softmax, used to compute the weighted average in theself-attention heads.
decoder_ngram_attentions (
tuple(torch.FloatTensor),optional, returned whenoutput_attentions=Trueis passed or whenconfig.output_attentions=True) — Tuple oftorch.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 theweighted average in the self-attention heads.
cross_attentions (
tuple[torch.FloatTensor],optional, returned whenoutput_attentions=Trueis passed or whenconfig.output_attentions=True) — Tuple oftorch.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 theweighted average in the cross-attention heads.
encoder_last_hidden_state (
torch.FloatTensorof 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 whenoutput_hidden_states=Trueis passed or whenconfig.output_hidden_states=True) — Tuple oftorch.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 encoder at the output of each layer plus the initial embedding outputs.
encoder_attentions (
tuple[torch.FloatTensor],optional, returned whenoutput_attentions=Trueis passed or whenconfig.output_attentions=True) — Tuple oftorch.FloatTensor(one for each layer) of shape(batch_size, num_heads, sequence_length, sequence_length).Attentions weights of the encoder, after the attention softmax, used to compute the weighted average in theself-attention heads.
TheProphetNetForConditionalGeneration 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 whilethe latter silently ignores them.
Example:
>>>from transformersimport 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# Batch size 1>>>decoder_input_ids = tokenizer("Studies show that", return_tensors="pt").input_ids# Batch size 1>>>outputs = model(input_ids=input_ids, decoder_input_ids=decoder_input_ids)>>>logits_next_token = outputs.logits# logits to predict next token as usual>>>logits_ngram_next_tokens = outputs.logits_ngram# logits to predict 2nd, 3rd, ... next tokens
ProphetNetForCausalLM
classtransformers.ProphetNetForCausalLM
<source>(config: ProphetNetConfig)
Parameters
- config (ProphetNetConfig) —Model configuration class with all the parameters of the model. Initializing with a config file does notload 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 fromPreTrainedModel. Check the superclass documentation for the generic methods thelibrary implements for all its model (such as downloading or saving, resizing the input embeddings, pruning headsetc.)
This model is also a PyTorchtorch.nn.Module subclass.Use it as a regular PyTorch Module and refer to the PyTorch documentation for all matter related to general usageand behavior.
forward
<source>(input_ids: typing.Optional[torch.Tensor] = Noneattention_mask: typing.Optional[torch.Tensor] = Noneencoder_hidden_states: typing.Optional[torch.Tensor] = Noneencoder_attention_mask: typing.Optional[torch.Tensor] = Nonepast_key_values: typing.Optional[transformers.cache_utils.Cache] = Noneinputs_embeds: typing.Optional[torch.Tensor] = Nonelabels: typing.Optional[torch.Tensor] = Noneuse_cache: typing.Optional[bool] = Noneoutput_attentions: typing.Optional[bool] = Noneoutput_hidden_states: typing.Optional[bool] = Nonereturn_dict: typing.Optional[bool] = None**kwargs)→transformers.models.prophetnet.modeling_prophetnet.ProphetNetDecoderLMOutput ortuple(torch.FloatTensor)
Parameters
- input_ids (
torch.Tensorof shape(batch_size, sequence_length),optional) —Indices of input sequence tokens in the vocabulary. Padding will be ignored by default.Indices can be obtained usingAutoTokenizer. SeePreTrainedTokenizer.encode() andPreTrainedTokenizer.call() for details.
- attention_mask (
torch.Tensorof 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 arenot masked,
- 0 for tokens that aremasked.
- encoder_hidden_states (
torch.Tensorof 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-attentionif the model is configured as a decoder. - encoder_attention_mask (
torch.Tensorof shape(batch_size, sequence_length),optional) —Mask to avoid performing attention on the padding token indices of the encoder input. This mask is used inthe cross-attention if the model is configured as a decoder. Mask values selected in[0, 1]:- 1 for tokens that arenot masked,
- 0 for tokens that aremasked.
- past_key_values (
~cache_utils.Cache,optional) —Pre-computed hidden-states (key and values in the self-attention blocks and in the cross-attentionblocks) that can be used to speed up sequential decoding. This typically consists in thepast_key_valuesreturned by the model at a previous stage of decoding, whenuse_cache=Trueorconfig.use_cache=True.OnlyCache instance is allowed as input, see ourkv cache guide.If no
past_key_valuesare passed,DynamicCache will be initialized by default.The model will output the same cache format that is fed as input.
If
past_key_valuesare used, the user is expected to input only unprocessedinput_ids(those that don’thave their past key value states given to this model) of shape(batch_size, unprocessed_length)instead of allinput_idsof shape(batch_size, sequence_length). - inputs_embeds (
torch.Tensorof shape(batch_size, sequence_length, hidden_size),optional) —Optionally, instead of passinginput_idsyou can choose to directly pass an embedded representation. Thisis useful if you want more control over how to convertinput_idsindices into associated vectors than themodel’s internal embedding lookup matrix. - labels (
torch.LongTensorof 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](seeinput_idsdocstring) Tokens with indices set to-100areignored (masked), the loss is only computed for the tokens with labels n[0, ..., config.vocab_size] - use_cache (
bool,optional) —If set toTrue,past_key_valueskey 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. Seeattentionsunder returnedtensors for more detail. - output_hidden_states (
bool,optional) —Whether or not to return the hidden states of all layers. Seehidden_statesunder returned tensors formore detail. - return_dict (
bool,optional) —Whether or not to return aModelOutput instead of a plain tuple.
Returns
transformers.models.prophetnet.modeling_prophetnet.ProphetNetDecoderLMOutput ortuple(torch.FloatTensor)
Atransformers.models.prophetnet.modeling_prophetnet.ProphetNetDecoderLMOutput or a tuple oftorch.FloatTensor (ifreturn_dict=False is passed or whenconfig.return_dict=False) comprising variouselements depending on the configuration (ProphetNetConfig) and inputs.
loss (
torch.FloatTensorof shape(1,),optional, returned whenlabelsis provided) — Language modeling loss.logits (
torch.FloatTensorof shape(batch_size, decoder_sequence_length, config.vocab_size)) — Prediction scores of the main stream language modeling head (scores for each vocabulary token beforeSoftMax).logits_ngram (
torch.FloatTensorof shape(batch_size, ngram * decoder_sequence_length, config.vocab_size)) — Prediction scores of the predict stream language modeling head (scores for each vocabulary token beforeSoftMax).past_key_values (
Cache,optional, returned whenuse_cache=Trueis passed or whenconfig.use_cache=True) — It is aCache instance. For more details, see ourkv cache guide.Contains pre-computed hidden-states (key and values in the attention blocks) of the decoder that can beused (see
past_key_valuesinput) to speed up sequential decoding.hidden_states (
tuple[torch.FloatTensor],optional, returned whenoutput_hidden_states=Trueis passed or whenconfig.output_hidden_states=True) — Tuple oftorch.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.
hidden_states_ngram (
tuple(torch.FloatTensor),optional, returned whenoutput_hidden_states=Trueis passed or whenconfig.output_hidden_states=True) — Tuple oftorch.FloatTensor(one for the output of the embeddings + one for the output of each layer) ofshape(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 embeddingoutputs.
attentions (
tuple[torch.FloatTensor],optional, returned whenoutput_attentions=Trueis passed or whenconfig.output_attentions=True) — Tuple oftorch.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-attentionheads.
ngram_attentions (
tuple(torch.FloatTensor),optional, returned whenoutput_attentions=Trueis passed or whenconfig.output_attentions=True) — Tuple oftorch.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 theweighted average in the
cross_attentions (
tuple[torch.FloatTensor],optional, returned whenoutput_attentions=Trueis passed or whenconfig.output_attentions=True) — Tuple oftorch.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 theweighted average in the cross-attention heads.
TheProphetNetForCausalLM 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 whilethe latter silently ignores them.
Example:
>>>from transformersimport 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>>># Model can also be used with EncoderDecoder framework>>>from transformersimport 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